Veterinary Microbiology, 27 ( 1991 ) 79-90 Elsevier Science Publishers B.V., A m s t e r d a m
79
Differentiation by western blotting of immune responses of cattle vaccinated with Brucella abortus strain 19 or infected experimentally or naturally with virulent Brucella abortus Carol A. Belzer, Louisa B. Tabatabai ~ and Billy L. Deyoe U.S. Department of Agriculture, Agricultural Research Service, National Animal Disease Center, P.O. Box 70, Ames, IA 50010, USA (Accepted 19 September 1990)
ABSTRACT Belzer, C.A., Tabatabai, L.B. and Deyoe, B.L., 1991. Differentiation by western blotting of immune responses of cattle vaccinated with Brucella abortus strain 19 or infected experimentally or naturally with virulent Brucella abortus. Vet. Microbiol., 27: 79-90.
Brucella abortus strain 19 salt-extractable proteins fractionated by differential ammonium sulfate precipitation were used in a western blotting method to detect bovine immunoglobulin G antibodies to B. abortus. Sera from infected cattle and from cattle vaccinated with strain 19 and subsequently exposed to virulent B. abortus bound to a common group of antigens ranging in molecular weights from 31 000 to 45 000 daltons. Immunoglobulin G antibodies in sera from the latter group in addition also bound to antigens with molecular weights of 66 000 to 71 000 daltons. Some sera from cattle vaccinated when sexually mature reacted similar to those from infected cattle, while immunoglobulin G antibodies in sera from Brucella-free cattle and vaccinated calves did not bind to either group of antigens. In general, fractionation of the proteins by ammonium sulfate precipitation offered no advantage for detecting differences between groups of sera. Ammonium sulfate fraction 0 to 35% reacted with a larger number of sera from a naturally infected group than fraction 0 to 70%. Both fractions reacted equally well with sera from the other groups of cattle, while fractions 35 to 70% and 70 to 100% reacted poorly in this technique. The attractive feature of the blot is that sera from calfhoodvaccinated cattle did not react.
INTRODUCTION
Standard serologic tests for diagnosis of bovine brucellosis have been in use since 1940 (Manthei et al., 1956), but the most difficult task has been in distinguishing antibodies of infected from those of vaccinated animals. To date, most diagnostic tests for Brucella abortus rely on detecting a humoral immune response, although the most definitive diagnostic test is bacterial ~To w h o m correspondence should be addressed.
0378-1135/91/$03.50
© 1991 - - Elsevier Science Publishers B.V.
80
t .'., B ~ I Z E R E T
M
culture and positive identification of B. abortus (Nicoletti, 1980). The bovine anti-Brucella antibody response is not only directed to the lipopolysaccharide (LPS) component of the cell (Saunders et al., 1977: Lamb et al., 1979: Ruppanner et al., 1980 ), but also to proteins and other macromolecular components (Stemshorn and Nielsen, 1977; Schurig et al., 1978: Nielsen et al., 1983; Tabatabai and Deyoe, 1984a,b). However, serological reactions following vaccination with strain 19 interfere with diagnosis of brucellosis unless supplemental tests are performed (Manthei et al., 1956; Alton et al., 1975 ). Recent progress in development of procedures for diagnosis of brucellosis has been reviewed ( Stemshorn, 1984 ). Specifically, a competitive enzyme linked immunosorbent assay (ELISA) procedure based on O-chain and competing monoclonal antibodies has shown promising results (Nielsen et al., 1987). It has been previously reported that B. abortus salt-extractable proteins ( BCSP ) could be used in a sensitive ELISA procedure for detecting brucellosis in cattle (Tabatabai and Deyoe, 1984b). The objective of the present study was to examine the i m m u n e response of vaccinated and infected cattle to BCSP fractionated by a m m o n i u m sulfate precipitation using the technique of sodium sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunologic detection of the antigens by western blotting (Partanen et al., 1983, Gershoni, 1985 ). MATERIALS A N D M E T H O D S
Antigen B. abortus strain 19 was grown for 48 h at 37°C in liquid culture as described (Alton et al., 1975 ). Cells were harvested by filtration through a 0.45 Ftm nitrocellulose filter (Millipore Cassette Filtration System, Millipore, Bedford, MA), and washed once with sterile 0.85% saline solution by centrifugation at 20 0 0 0 X g for 20 min at 5°C. The cells were inactivated by adding 1 volume of cell suspension at 0.2 g wet weight per ml of saline to 2 volumes of methanol and stirring gently for 4 h at 5 oC to resuspend the cells. The cell suspension was stored at 5 °C for 1 week with occasional stirring to ensure complete inactivation. Protein fractions The inactivated cells were centrifuged at 20 000 X g for 30 min at 5 ° C, and two overnight extractions were performed with 1 M NaC1-0.1 M sodium citrate at 0.2 g per ml by stirring the suspension at 5°C. The supernatants from the two extracts were combined, dialyzed against 5 m M NH4HCO3, concentrated by freeze-drying, dialyzed against 5 m M NH4HCO3 clarified by centrifugation at 20 0 0 0 X g for 15 min, and fractionated by precipitation with solid a m m o n i u m sulfate. The following a m m o n i u m sulfate fractions were prepared: 0 to 35%, 35 to 70%, 70 to 100%, and 0 to 70% by adding 0.21 g,
BOVINE ANTIBODY TO B. ABORTUS BY WESTERN BLOTTING
81
0.24 g, and 0.47 g of ammonium sulfate per ml of protein extract, for comparison with previously published results (Tabatabai and Deyoe, 1984b). These fractions were designated 0-35, 35-70, 70-100, and 0-70, respectively. The precipitates were collected by centrifugation at 20 000 X g for 10 min at 5 ° C, dissolved in 1 to 2 ml of 5 mM NH4 HCO3, and dialyzed against the same buffer until dialysate tested negative for sulfate ion.
Chemical analysis Protein content of the ammonium sulfate fractions was measured by the method of Lowry et al. ( 1951 ) total carbohydrate by the phenol-sulfuric acid method (Dubois et al., 1956), and 2-keto-3-deoxyoctulosonic acid (KDO) according to Ashwell (1966) after initial hydrolysis of the sample with 0.25 N sulfuric acid for 10 min at 100°C (Tabatabai et al., 1979). Absorbance at 532 nm (A532) due to 2-deoxy aldoses, although minimal, was adjusted as described by Warren ( 1959 ) using 2-deoxy-ribose as a standard.
SDS-PAGE SDS-PAGE was performed under denaturing conditions, using a 0.75 mm thick slab gel (10× 14 cm) containing 15% acrylamide and 1.35% bisacrylamide with a 4% acrylamide stacking gel. Buffer composition and sample were prepared as recommended (Hoeffer Scientific Instruments, San Francisco, CA). Samples containing 4 pg of protein were electrophoresed at 10 mA per gel at 15 to 18 °C for 4.5 h. Gels were stained with Coomassie Brilliant blue R-250 (Bio-Rad Laboratories, Richmond, CA) (Weber and Osborn, 1969) or with silver (Tsai and Frasch, 1982). Samples for subsequent western blotting experiments contained 4/tg of fractions 0 to 70 and 0 to 35, and 20pg of fractions 35 to 70 and 70 to 100.
Western blotting The SDS gels were equilibrated for 30 min in blotting buffer which is composed of 25 mM Tris-base, 192 mM glycine, and 20% (v/v) of methanol (Towbin et al., 1979). Nitrocellulose membranes, BA-85, were from Schleicher and Schull (Keene, N.H.). Electrophoretic blotting was performed at 30 V and 0.11 A for 16 h using the Transblot apparatus (Bio-Rad). Following electrophoretic transfer, the nitrocellulose sheets were rinsed twice in distilled water. The sheets were then incubated for 15 min with a solution containing phosphate buffered saline (PBS)-Tween 80 containing 0.2% (w/ v) gelatin (Bacto Gelatin, Difco Products, Detroit, MI ) or 3% liquid fish gelatin (Norland Products, Inc., New Brunswick, N J). The blocking solution was removed, and bovine or rabbit antisera diluted 1 : 500 with PBS-Tween in 0.2% (w/v) gelatin (or in 3% fish gelatin) was then added to the nitrocellulose membranes and incubated for 2 h at room temperature. The membranes were washed five times for 5 min each time with PBS-Tween and sub-
~
!
\ BII/IRI
I ~.l
sequently placed in horseradish peroxidase-labeled anti-bovine immunoglobulin G ( IgG ) conjugate (heavy and light chain-specific ) ( USDA, APHIS, S&T, National Veterinary Services Laboratories, Ames, 1A ), diluted 1 : 500 in PBS-Tween-gelatin, and incubated for 2 h at room temperature. The membrane was washed as described above, the substrate solution was added, and the membrane incubated until bands were sufficiently developed. The substrate solution was composed of 0.3% (w/v) 4-chloronaphthol ( Sigma, St. Louis, MO) and 0.01% H202 in 0.1 M PBS, pH 7.2. One portion of the nitrocellulose membrane containing an identical protein blot was stained for 2 rain with an Amido Black solution (K&K Laboratories, Inc., Jamaica, NY) consisting of 0.1% (w/v) Amido Black, 25% (v/v) isopropanol, and 10% (v/v) acetic acid, and destained with a solution containing 25% (v/v) isopropanol and 10% (v/v) acetic acid. Bovine sera Bovine sera were selected to provide a spectrum of reactions that might be encountered under field conditions. A positive control serum was collected from an experimentally infected cow at necropsy 4 months after abortion (cow 47 was experimentally infected with 4 × 107 B. abortus strain 2308 ). A negative control serum was a composite of 73 pooled sera of unvaccinated brucellosis-free cattle. Table 2 shows the geometric mean titers of the serologic results obtained with 10 sera from each of the following groups of cattle: sera from calves vaccinated with 3 × 109 strain 19 collected at 3 weeks post-vaccination; sera from cattle vaccinated when sexually mature with 7 × 109 strain 19 and collected at 3 or 4 weeks post-vaccination (R.D. Angus, USDA, APHIS, S&T, National Veterinary Service Laboratories, Ames, IA); sera from calves vaccinated with 3 X 109 strain l 9, and challenge exposed with 3 X 107 B. abortus strain 2308 collected 16 weeks post-exposure (2-years post-vaccination); sera from sexually mature cows experimentally-infected with 3 × 107 B. abortus strain 2308 and collected at 4 months, post-exposure; sera from naturally infected cattle (D. Stringfellow, Department of Veterinary Microbiology, Auburn University, Auburn, AL) (B. abortus biovar was isolated from these animals periodically); and sera from brucellosis-free, non-vaccinated cattle were provided by P.J. Matthews, USDA, ARS, National Animal Disease Center, Animal Supply Services, Ames, IA. RESULTS
Characterization o f the protein fractions Protein, carbohydrate, and KDO composition of the a m m o n i u m sulfate fractions of B. abortus strain 19 salt-extractable proteins are shown in Table 1. Chemical analyses showed that fraction 0-70 contained the least KDO per mg protein, and fraction 0-35 contained the highest amount. In general, the
BOVINEANTIBODYTOB. ABORTUS BYWESTERNBLOTTING
83
TABLE 1 Carbohydrate and 2-keto-3-deoxyoctulosonic acid (KDO) ratios of the ammonium sulfate fractions of Brucella salt-extractable proteins (BCSP) Fraction
Carbohydrate/protein (mg/mg)
KDO/protein (#g/mg)
0-35 0-70 35-70 70-100
1.69 0.47 0.12 0.08
13.85 0.39 1.65 1.50
tool wt x 10.3
A
Std
I
2
6
:3
4
I
2
3
4
Fig. 1. SDS-PAGE of B. abortus protein fractions. Coomassie Brilliant Blue-stained gel (A), and silver-stained gel (B). Lane 1, fraction 0-70; lane 2, 0-35; lane 3, 35-70; and lane 4, 70100. Std indicates Mw markers.
results indicated that a considerable amount of polysaccharide a n d / o r lipopolysaccharide co-precipitated with the 0-35 and 0-70 fractions (1.69 and 0.47 mg carbohydrate/mg protein, respectively). It was not determined if the thiobarbituric acid positive tests of the 0-35 and 0-70 fractions were due solely to KDO or to other 2-keto-3-deoxy sugars such as sialic acid.
SDS electrophoresis A m m o n i u m sulfate fractionation of the extract resulted in few major differences discernible by Coomassie Brilliant Blue staining of the SDS gels of fractions 0-35, 0-70, 35-70. The major protein bands of these fractions ranged in molecular weight (Mw) from 31 000 to 45 000 (Fig. 1 ). Two low Mw bands of > 14 000 were also found. Fraction 70-100 contained proportion-
0/10 10/10
Brucella-free Vaccinated calves ( 3 weeks post-vaccination ) Vaccinated adults ( 3-4 weeks post-~ accinalion ) Vaccinated challenge exposed ( 16 weeks post-exposure ) Experimentally infected ( 16 weeks post-exposure ) Naturall} infected
t 770 (400-1600) 2000 ( 800-3200 ) 70 (0-20) 1480 (200-6400) 1510 ( 800-3200 )
SAT b
_ 140 (50-400) 400 ( 200-3200 ) 40 (1-100) 1980 (200-3200) ~>200
2ME"
10 (0-50) 350 (200-400) ~>200
_ 230 (100-400) 400
RIV d
(2/20-4/640) 4/80
(0-l/40)
(1/80-4/200)
( 1/20-1/80)
-
('F~
I0/10
I0/10
1/10
1/10
0/10
Culture results no. positive/ total
Fig. 2. panel G
Fig. 2, panel F
Fig. ,," panel E
Fig. 2, panel t)
Fig. 2, panel B Fig. 2. panel (
Western blot cross reference
~Complement fixation test: the d e n o m i n a t o r is tile highest dilution showing fixation and the numerator indicates the degree of fixation. No geometriL mean titers are calculated. "No reaction in test.
~'Titers are geometric mean titers: range o f liters is indicated in parentheses. ~'Standard agglutination tube test. ~2-mercaptoelhanol test. dRivanol precipitation plate agglutination test.
10/10
10/10
7/10
10/10
Card p o s / total
Group
Summary of serologic reactions of various groups of cattle"
TABLE 2
BOVINE ANTIBODY TO B. ABORTUS BY WESTERN BLOTTING
85
ally less of the 31 000 to 45 000 group and in addition contained a heavily stained band at a Mw of 20 000.
Serologic reactions The geometric means of the serologic reactions of the cattle in each group are detailed in Table 2. Isolation of B. abortus from tissue or milk samples of the cattle at the termination of the vaccine experiment are also indicated. The Brucella-free cattle were seronegative by all tests. Sera of the vaccinated calves obtained 3 weeks post-vaccination were card-test positive as expected. The standard agglutination tube (SAT) test, 2-mercaptoethanol test, Rivanol, and complement fixation tests were all positive, and reactions varied considerably among the animals. The adult-vaccinated cows at 3 or 4 weeks post-vaccination varied similarly in their serologic reactions, and were approximately one dilution higher in titer compared with the vaccinated calves. The vaccinated challenge-exposed cows at 16 weeks post-exposure (2 years post-vaccination ) showed generally much lower titers than the former two groups, and were culture-negative at subsequent parturition and necropsy. Sera of the infected cattle at 16 weeks post-exposure had high titers in the serologic tests, as did those of naturally infected cows. Western blots Fig. 2 depicts the results of western blots developed with a representative cattle serum from one each of the five different groups although all sera were tested by this procedure. Fig. 2, panel A, shows the Amido Black-stained blot, followed by an immunoblot. The Amido Black-stained blot showed fewer bands than the Coomassie Brilliant Blue-stained SDS gels (Fig. 1, and panel A of Fig. 2). None of the lower Mw bands (i.e., >20 000 or those of Mw > 45 000) were observed with Amido Black. The immunoblots of the Brucella-free control cattle and the vaccinated calves showed no reaction under the condition of the procedures used (Fig. 2, panels B and C, respectively). Although seven of the adult-vaccinated cattle showed no reaction on the immuno-blot, three out of ten cattle sera showed activity primarily toward the protein bands in the Mw range of 31 000 to 45 000 and one of these toward a second group of proteins in the Mw range of 66 000 to 71 000 (Fig. 2, panel D). B. abortus strain 19 was isolated from one of the latter cows, but the bacteriologic status of the other two cows was unknown. With sera from calfhood vaccinated, and subsequently experimentally infected cattle, typically, a strong reaction with fraction 0-35 and 0-70 covering the entire Mw range from 31 000 to 45 000 was observed (lanes 1 and 2 of Fig. 2, panel E). Additional bands were observed at apparent Mw of 66 000 and 71 000 (Fig. 2, panel E, lanes 1-4). No reactions were observed with the 20 000 dalton protein under the conditions of the experiments. Six often sera
~6
~, B t - I Z I ! R [ I
i
mol wt
Z
4
3
i
2
3
4
J
2
3
4
I
2
3
\l
4
x I0 -3
71453[-
20 - 4
D
B
A
71-tl
C
D
"
4531-
4
I
20-
E
F
G
Fig. 2. Western blot of B. abortus protein fractions with cattle sera from various groups. Panel A, Amido Black stain of protein blot; panel B, Brucella-free; panel C, vaccinated;panel D, adultvaccinated; panel E, calfhood vaccinated-challengeexposed; panel F, cxperimentally infected; panel G, naturally infected. Lanes 1-4, contain fraction 0-35, fraction 0-70, fraction 35-70 and fraction 70-100, respectively. from the non-vaccinated, experimentally-infected group reacted with all four fractions (Fig. 2, panel F, lanes 1-4). Three bands could be observed corresponding to Mw to 31 000, 39 000, and 45 000. The remaining four sera reacted with fraction 0-35 and 0-70 only (data not shown). Bands of higher or lower Mw were not observed in the immunoblots. A representative i m m u n o b l o t of one of the sera from naturally infected cattle is shown in Fig. 2, panel G. Nine of ten sera of naturally infected cattle reacted weakly with the 0-35 fractions as shown in Fig. 2, panel G, lane l, and moderately with fraction 0-70 as shown in Fig. 2, lane 2. The only serum from this group which did not react on the blot (data not shown) was also tested in an ELISA procedure (Tabatabai and Deyoe, 1984b). The IgG response for this serum, measured in terms of absorbance units, was just above
BOVINE ANTIBODY TO B. ABORTUS BY WESTERN BLOTTING
87
the minimal level of 0.142 absorbance units (Tabatabai and Deyoe, 1984b), below which the IgG response is considered negative. DISCUSSION The experiments described in this paper demonstrate that the western blot technique with the BCSP antigens may distinguish the serologic reactions of calfhood-vaccinated (Fig. 2, panel C) from infected cattle (Fig. 2, panels F and G). Sera from both the experimentally-infected group and all but one of the naturally infected group reacted with the BCSP antigens, while none of the sera of vaccinated cattle reacted with the BCSP fractions. These results contrasted with those of the adult-vaccinated group. Although seven of ten sera did not react, three of the sera from the latter group reacted weakly with BCSP antigens similar to those of sera from the naturally infected cows (Fig. 2, panels D and G), and the experimentally-infected cow (Fig. 2, panel F). Considering this was a relatively short post-vaccination sampling time interval (3 or 4 weeks), it is reasonable to expect that the remaining blot-positive animals also would become negative unless a persistent infection had been established as was observed for one cow from which strain 19 had been isolated. Generally, the western blot technique detected fewer of these adultvaccinated animals than did the SAT test. The blots with sera from vaccinated challenge-exposed animals showed intense reactions to the BCSP fractions (Fig. 2, panel E), which differed qualitatively and quantitatively from the experimentally-infected group (Fig. 2, panel F) but resembled some of the blots from the naturally infected group (data not shown). These results were consistent with re-exposure of the animals to brucellae, as some of these serum samples came from animals with recurrent infections from which B. abortus biotype was periodically isolated. (D. Stringfellow, personal communication). We observed that re-exposure by experimental infection of the animals to brucellae produced a reaction with two high Mw bands (Fig. 2, panel E); the gene for one of these proteins (the 66 kDa band) has been cloned (unpublished data, Chin, Mayfield and Tabatabai, 1989). It remains to be seen how often this reaction occurs before it can be considered of diagnostic value. One serum of this group of animals showed a weak reaction on the immunoblot. This may be due to a lack of secondary i m m u n e response which reportedly had been associated with infection (Sutherland, 1985). Culture data showed that B. abortus 2308 was isolated from this animal. Although the blot was positive whenever an experimental or natural infection occurred, the blot technique failed to detect one of the naturally infected animals for which the serologic results were similar to those for the other cows in the group. When this serum was tested in the ELISA with fractions 0-35 and 0-70, it was found that the level of IgG in the serum was much lower than
NN
~ \ ~{I t Z t
R I I \t
the mean IgG level tbr the remainder of the cows (Tabatabai and Belzer, unpublished results). Based on minimal values of absorbance in the ELISA, the ELISA result for this cow serum was just below the level designated as positive (Tabatabai, 1984b). Undoubtedly, this animal would have been detected at a lower serum dilution than currently used for the blot ill order to keep the background staining acceptably low. The antibody specificity of the reactions detected in the western blots is apparently primarily directed toward IgG, because no or very faint reactions were observed when an anti-bovine IgM heavy chain-specific conjugate was used (Tabatabai and Belzer, unpublished results). Interestingly, the intensity of reaction on the blots appeared to be inversely related to the corresponding titers of the serologic reactions for the vaccinated challenge-exposed and the experimentally-infectedgroups. These results indicate that antibody specificity toward Brucella antigens may differ considerably, depending on the vaccination and/or challenge exposure history of the animal (Butler et al., 1986 ). It is reasonable to assume, based on the data in Table 1, that the observed intensity of staining of the blots was due entirely to antibody reacting with residual protein-bound LPS rather than with protein. However, the increase in intensity on the blot with certain sera does not reflect a similar increase in serologic titers. Presumably, the two methods measured antibodies of different specificities or classes, as has been observed for other serologic tests tbr bovine brucellosis ( Patterson et al., 1976 ). Fractionation of the BCSP antigens by ammonium sulfate precipitation offered no advantage for detecting differences between groups. In general, under the conditions of the test used fraction 0-35 reacted with a larger number of sera from the naturally infected group than fraction 0-70. Both fractions reacted equally well with sera from the remaining groups, while fractions 3570 and 70-100 bound bovine IgG antibody reacted poorly in the blot. The attractive feature of the blotting technique is that sera from calfhoodvaccination animals do not react. As reported for other serologic tests (Sutherland, 1984), sera from adult-vaccinated animals may pose a problem with this technique as well. However, if adult-vaccinated cattle are kept under quarantine until seronegative by the card test the adult-vaccinated cattle would be no problem with this technique.
ACKNOWLEDGMENTS
We gratefully acknowledge the expert technical assistance of Kathryn Meredith and Carol Irvin. We thank Tom Glasson for his expert photographic work, and Nancy Kilstofte for manuscript preparation.
BOVINE A N T I B O D Y T O B. ABORTUS BY WESTERN B L O T T I N G
89
REFERENCES Alton, G.G., Jones, L.M. and Pietz, D.E., 1975. Laboratory techniques in brucellosis, 2nd Edn. WHO, Washington, DC, pp. 11-63. Ashwell, G., 1966. New colorimetric methods of sugar analysis. Methods Enzymol., 8: 85-95. Butler, J.E., Seawright, G.L., McGivern, P.L. and Gildsdorf, M., 1986. Preliminary evidence for a diagnostic immunoglobulin G1 antibody response among culture-positive cows vaccinated with Brucella abortus strain 19 and challenge exposed with strain 2308. Am. J. Vet. Res., 47: 1258-1264. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A. and Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem., 28: 350-356. Gershoni, J.M., 1985. Protein blotting: development and perspectives. Trends Biochem. Sci., 10: 103-106. Lamb, V.L., Jones, L.M., Schurig, G.G. and Berman, D.T., 1979. Enzyme-linked immunosorbent assay for bovine immunoglobulin subclass-specific response to Bruce/& abortus lipopolysaccharides. Infect. Immun., 26: 240-247. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J., 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem., 193: 265-275. Manthei, C.A., Kuttler, A.K. and Goode, Jr., E.R., 1956. The Yearbook of Agriculture. Proc. Anim. Dis. 84th Congr., 2D Session, House Document No. 344, U.S. Dept. of Agric., Washington, DC, pp. 202-213. Nicoletti, P., 1980. Epidemiology of bovine brucellosis. Adv. Vet. Sci. Comp. Med., 24: 69-98. Nielsen, K.H., Heck, F.C., Stiller, J.M. and Rosenbaum, B., 1983. Interaction of specifically purified isotypes of bovine antibody to Brucella abortus in the hemolysis in gel test and enzyme-linked immunosorbent assay. Res. Vet. Sci., 35: 14-18. Nielsen, K., Wright, P.F., Cherwonogrodzky, J., Duncan, J.R. and Stemshorn, B., 1987. Enzyme immunoassay for diagnosis of bovine brucellosis. Ann. Inst. Pasteur, Microbiol., 138: 69-144. Partanen, P., Turanen, H.J., Paasivis, R., Forshblom, E., Suni, J. and Leinikki, P.O., 1983. Identification of antigenic components of Toxoplasma gondii by an immunoblotling technique. FEBS Lett., 158: 252-254. Patterson, J.M., Deyoe, B.L. and Stone, S.S., 1976. Identification ofimmunoglobulin associated with complement-fixation agglutination and low pH-buffered antigen tests for brucellosis. Am. J. Vet. Res., 37: 319-324. Ruppanner, R., Meyer, M.E., Willeberg, P. and Behymer, D.E., 1980. Comparison of the enzyme-linked immunosorbent assay with other tests for brucellosis using sera from experimentally infected heifers. Am. J. Vet. Res., 41: 1329-1332. Saunders, G.C., Clinard, E.H., Bartlett, M.L. and Sanders, W.M., 1977. Application of the indirect enzyme labeled antibody microtest to the detection and surveillance of animal diseases. J. Infect. Dis., 136: 5258-5266. Schurig, G.G., Jones, L.M., Speth, S.L. and Berman, D.T., 1978. Antibody response to antigens distinct from the smooth lipopolysaccharide complex in Brucella infection. Infect. lmmun., 21: 994-1002. Stemshorn, B.W., 1984. Recent progress in the diagnosis of brucellosis. Dev. Biol. Stand., 56: 325-340. Stemshorn, B. and Nielsen, K., 1977. The bovine immune response to B. abortus: water soluble antigen precipitated by sera of some naturally infected cattle. Can. J. Comp. Med., 41 : 152159. Sutherland, S.S., 1984. Evaluation of the enzyme-linked immunosorbent assay in the detection of cattle infected with Brucella abortus. Vet. Microbiol., 10: 23-32. Sutherland, S.S., 1985. Comparison of enzyme-linked immunosorbent assay and complement
9()
~ ~, I ~ t l Z I R I
I \1
fixation test tbr the detection of specific antibody in cattle \accmatcd and challenged wilh Ilrucella at~rlu~. J. ('lin. Microbiol., 22: 44-47. TabatabaL L.B. and Deyoe, B i . , 1984a. Biochemical and biological propcrties of soluble protern preparations from Ilrucella at~n'tu,s. Dev. Biol. Stand., 56:199-211. Tabatabai, L.B. and Deyoe, B.L., 1984b. Specific enzyme-linked immunosorbcnt assay t'or detection of bovine antibody to Brucella ab##rtus, lnfecl, lmmun., 26: 668-679. Tabatabai, L.B., Deyoe, B i . and Ritchie, A.E., 1979. Isolation and characterization of loxic fractions from Ilrucella al~orlu#. Infect. lmmun., 26: 668-679. Towbin, H., Staehelin. T. and Gordon, J,, 1979. Electrophorefic Iransfer ol~proteins from acr~,lamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA, 76: 4.350-4354. Tsai, C.M. and Frasch, C.E., 1982. A sensitive silver stain for detecting lipopolysaccharides in polyac~lamide gels. Anal. Biochem., 119; 115-119. Warren, L., 1959. The thiobarbituric acid assay of sialic acids. J. Biol. Chem., 245:197 I-1975. Weber, K. and Osborn, M., 1969. The reliability of molecular weight determinations by dodecyl sulfate polyacrylamide gel electrophoresis. J. Biol. Chem., 244:4406-4412.
79
Differentiation by western blotting of immune responses of cattle vaccinated with Brucella abortus strain 19 or infected experimentally or naturally with virulent Brucella abortus Carol A. Belzer, Louisa B. Tabatabai ~ and Billy L. Deyoe U.S. Department of Agriculture, Agricultural Research Service, National Animal Disease Center, P.O. Box 70, Ames, IA 50010, USA (Accepted 19 September 1990)
ABSTRACT Belzer, C.A., Tabatabai, L.B. and Deyoe, B.L., 1991. Differentiation by western blotting of immune responses of cattle vaccinated with Brucella abortus strain 19 or infected experimentally or naturally with virulent Brucella abortus. Vet. Microbiol., 27: 79-90.
Brucella abortus strain 19 salt-extractable proteins fractionated by differential ammonium sulfate precipitation were used in a western blotting method to detect bovine immunoglobulin G antibodies to B. abortus. Sera from infected cattle and from cattle vaccinated with strain 19 and subsequently exposed to virulent B. abortus bound to a common group of antigens ranging in molecular weights from 31 000 to 45 000 daltons. Immunoglobulin G antibodies in sera from the latter group in addition also bound to antigens with molecular weights of 66 000 to 71 000 daltons. Some sera from cattle vaccinated when sexually mature reacted similar to those from infected cattle, while immunoglobulin G antibodies in sera from Brucella-free cattle and vaccinated calves did not bind to either group of antigens. In general, fractionation of the proteins by ammonium sulfate precipitation offered no advantage for detecting differences between groups of sera. Ammonium sulfate fraction 0 to 35% reacted with a larger number of sera from a naturally infected group than fraction 0 to 70%. Both fractions reacted equally well with sera from the other groups of cattle, while fractions 35 to 70% and 70 to 100% reacted poorly in this technique. The attractive feature of the blot is that sera from calfhoodvaccinated cattle did not react.
INTRODUCTION
Standard serologic tests for diagnosis of bovine brucellosis have been in use since 1940 (Manthei et al., 1956), but the most difficult task has been in distinguishing antibodies of infected from those of vaccinated animals. To date, most diagnostic tests for Brucella abortus rely on detecting a humoral immune response, although the most definitive diagnostic test is bacterial ~To w h o m correspondence should be addressed.
0378-1135/91/$03.50
© 1991 - - Elsevier Science Publishers B.V.
80
t .'., B ~ I Z E R E T
M
culture and positive identification of B. abortus (Nicoletti, 1980). The bovine anti-Brucella antibody response is not only directed to the lipopolysaccharide (LPS) component of the cell (Saunders et al., 1977: Lamb et al., 1979: Ruppanner et al., 1980 ), but also to proteins and other macromolecular components (Stemshorn and Nielsen, 1977; Schurig et al., 1978: Nielsen et al., 1983; Tabatabai and Deyoe, 1984a,b). However, serological reactions following vaccination with strain 19 interfere with diagnosis of brucellosis unless supplemental tests are performed (Manthei et al., 1956; Alton et al., 1975 ). Recent progress in development of procedures for diagnosis of brucellosis has been reviewed ( Stemshorn, 1984 ). Specifically, a competitive enzyme linked immunosorbent assay (ELISA) procedure based on O-chain and competing monoclonal antibodies has shown promising results (Nielsen et al., 1987). It has been previously reported that B. abortus salt-extractable proteins ( BCSP ) could be used in a sensitive ELISA procedure for detecting brucellosis in cattle (Tabatabai and Deyoe, 1984b). The objective of the present study was to examine the i m m u n e response of vaccinated and infected cattle to BCSP fractionated by a m m o n i u m sulfate precipitation using the technique of sodium sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by immunologic detection of the antigens by western blotting (Partanen et al., 1983, Gershoni, 1985 ). MATERIALS A N D M E T H O D S
Antigen B. abortus strain 19 was grown for 48 h at 37°C in liquid culture as described (Alton et al., 1975 ). Cells were harvested by filtration through a 0.45 Ftm nitrocellulose filter (Millipore Cassette Filtration System, Millipore, Bedford, MA), and washed once with sterile 0.85% saline solution by centrifugation at 20 0 0 0 X g for 20 min at 5°C. The cells were inactivated by adding 1 volume of cell suspension at 0.2 g wet weight per ml of saline to 2 volumes of methanol and stirring gently for 4 h at 5 oC to resuspend the cells. The cell suspension was stored at 5 °C for 1 week with occasional stirring to ensure complete inactivation. Protein fractions The inactivated cells were centrifuged at 20 000 X g for 30 min at 5 ° C, and two overnight extractions were performed with 1 M NaC1-0.1 M sodium citrate at 0.2 g per ml by stirring the suspension at 5°C. The supernatants from the two extracts were combined, dialyzed against 5 m M NH4HCO3, concentrated by freeze-drying, dialyzed against 5 m M NH4HCO3 clarified by centrifugation at 20 0 0 0 X g for 15 min, and fractionated by precipitation with solid a m m o n i u m sulfate. The following a m m o n i u m sulfate fractions were prepared: 0 to 35%, 35 to 70%, 70 to 100%, and 0 to 70% by adding 0.21 g,
BOVINE ANTIBODY TO B. ABORTUS BY WESTERN BLOTTING
81
0.24 g, and 0.47 g of ammonium sulfate per ml of protein extract, for comparison with previously published results (Tabatabai and Deyoe, 1984b). These fractions were designated 0-35, 35-70, 70-100, and 0-70, respectively. The precipitates were collected by centrifugation at 20 000 X g for 10 min at 5 ° C, dissolved in 1 to 2 ml of 5 mM NH4 HCO3, and dialyzed against the same buffer until dialysate tested negative for sulfate ion.
Chemical analysis Protein content of the ammonium sulfate fractions was measured by the method of Lowry et al. ( 1951 ) total carbohydrate by the phenol-sulfuric acid method (Dubois et al., 1956), and 2-keto-3-deoxyoctulosonic acid (KDO) according to Ashwell (1966) after initial hydrolysis of the sample with 0.25 N sulfuric acid for 10 min at 100°C (Tabatabai et al., 1979). Absorbance at 532 nm (A532) due to 2-deoxy aldoses, although minimal, was adjusted as described by Warren ( 1959 ) using 2-deoxy-ribose as a standard.
SDS-PAGE SDS-PAGE was performed under denaturing conditions, using a 0.75 mm thick slab gel (10× 14 cm) containing 15% acrylamide and 1.35% bisacrylamide with a 4% acrylamide stacking gel. Buffer composition and sample were prepared as recommended (Hoeffer Scientific Instruments, San Francisco, CA). Samples containing 4 pg of protein were electrophoresed at 10 mA per gel at 15 to 18 °C for 4.5 h. Gels were stained with Coomassie Brilliant blue R-250 (Bio-Rad Laboratories, Richmond, CA) (Weber and Osborn, 1969) or with silver (Tsai and Frasch, 1982). Samples for subsequent western blotting experiments contained 4/tg of fractions 0 to 70 and 0 to 35, and 20pg of fractions 35 to 70 and 70 to 100.
Western blotting The SDS gels were equilibrated for 30 min in blotting buffer which is composed of 25 mM Tris-base, 192 mM glycine, and 20% (v/v) of methanol (Towbin et al., 1979). Nitrocellulose membranes, BA-85, were from Schleicher and Schull (Keene, N.H.). Electrophoretic blotting was performed at 30 V and 0.11 A for 16 h using the Transblot apparatus (Bio-Rad). Following electrophoretic transfer, the nitrocellulose sheets were rinsed twice in distilled water. The sheets were then incubated for 15 min with a solution containing phosphate buffered saline (PBS)-Tween 80 containing 0.2% (w/ v) gelatin (Bacto Gelatin, Difco Products, Detroit, MI ) or 3% liquid fish gelatin (Norland Products, Inc., New Brunswick, N J). The blocking solution was removed, and bovine or rabbit antisera diluted 1 : 500 with PBS-Tween in 0.2% (w/v) gelatin (or in 3% fish gelatin) was then added to the nitrocellulose membranes and incubated for 2 h at room temperature. The membranes were washed five times for 5 min each time with PBS-Tween and sub-
~
!
\ BII/IRI
I ~.l
sequently placed in horseradish peroxidase-labeled anti-bovine immunoglobulin G ( IgG ) conjugate (heavy and light chain-specific ) ( USDA, APHIS, S&T, National Veterinary Services Laboratories, Ames, 1A ), diluted 1 : 500 in PBS-Tween-gelatin, and incubated for 2 h at room temperature. The membrane was washed as described above, the substrate solution was added, and the membrane incubated until bands were sufficiently developed. The substrate solution was composed of 0.3% (w/v) 4-chloronaphthol ( Sigma, St. Louis, MO) and 0.01% H202 in 0.1 M PBS, pH 7.2. One portion of the nitrocellulose membrane containing an identical protein blot was stained for 2 rain with an Amido Black solution (K&K Laboratories, Inc., Jamaica, NY) consisting of 0.1% (w/v) Amido Black, 25% (v/v) isopropanol, and 10% (v/v) acetic acid, and destained with a solution containing 25% (v/v) isopropanol and 10% (v/v) acetic acid. Bovine sera Bovine sera were selected to provide a spectrum of reactions that might be encountered under field conditions. A positive control serum was collected from an experimentally infected cow at necropsy 4 months after abortion (cow 47 was experimentally infected with 4 × 107 B. abortus strain 2308 ). A negative control serum was a composite of 73 pooled sera of unvaccinated brucellosis-free cattle. Table 2 shows the geometric mean titers of the serologic results obtained with 10 sera from each of the following groups of cattle: sera from calves vaccinated with 3 × 109 strain 19 collected at 3 weeks post-vaccination; sera from cattle vaccinated when sexually mature with 7 × 109 strain 19 and collected at 3 or 4 weeks post-vaccination (R.D. Angus, USDA, APHIS, S&T, National Veterinary Service Laboratories, Ames, IA); sera from calves vaccinated with 3 X 109 strain l 9, and challenge exposed with 3 X 107 B. abortus strain 2308 collected 16 weeks post-exposure (2-years post-vaccination); sera from sexually mature cows experimentally-infected with 3 × 107 B. abortus strain 2308 and collected at 4 months, post-exposure; sera from naturally infected cattle (D. Stringfellow, Department of Veterinary Microbiology, Auburn University, Auburn, AL) (B. abortus biovar was isolated from these animals periodically); and sera from brucellosis-free, non-vaccinated cattle were provided by P.J. Matthews, USDA, ARS, National Animal Disease Center, Animal Supply Services, Ames, IA. RESULTS
Characterization o f the protein fractions Protein, carbohydrate, and KDO composition of the a m m o n i u m sulfate fractions of B. abortus strain 19 salt-extractable proteins are shown in Table 1. Chemical analyses showed that fraction 0-70 contained the least KDO per mg protein, and fraction 0-35 contained the highest amount. In general, the
BOVINEANTIBODYTOB. ABORTUS BYWESTERNBLOTTING
83
TABLE 1 Carbohydrate and 2-keto-3-deoxyoctulosonic acid (KDO) ratios of the ammonium sulfate fractions of Brucella salt-extractable proteins (BCSP) Fraction
Carbohydrate/protein (mg/mg)
KDO/protein (#g/mg)
0-35 0-70 35-70 70-100
1.69 0.47 0.12 0.08
13.85 0.39 1.65 1.50
tool wt x 10.3
A
Std
I
2
6
:3
4
I
2
3
4
Fig. 1. SDS-PAGE of B. abortus protein fractions. Coomassie Brilliant Blue-stained gel (A), and silver-stained gel (B). Lane 1, fraction 0-70; lane 2, 0-35; lane 3, 35-70; and lane 4, 70100. Std indicates Mw markers.
results indicated that a considerable amount of polysaccharide a n d / o r lipopolysaccharide co-precipitated with the 0-35 and 0-70 fractions (1.69 and 0.47 mg carbohydrate/mg protein, respectively). It was not determined if the thiobarbituric acid positive tests of the 0-35 and 0-70 fractions were due solely to KDO or to other 2-keto-3-deoxy sugars such as sialic acid.
SDS electrophoresis A m m o n i u m sulfate fractionation of the extract resulted in few major differences discernible by Coomassie Brilliant Blue staining of the SDS gels of fractions 0-35, 0-70, 35-70. The major protein bands of these fractions ranged in molecular weight (Mw) from 31 000 to 45 000 (Fig. 1 ). Two low Mw bands of > 14 000 were also found. Fraction 70-100 contained proportion-
0/10 10/10
Brucella-free Vaccinated calves ( 3 weeks post-vaccination ) Vaccinated adults ( 3-4 weeks post-~ accinalion ) Vaccinated challenge exposed ( 16 weeks post-exposure ) Experimentally infected ( 16 weeks post-exposure ) Naturall} infected
t 770 (400-1600) 2000 ( 800-3200 ) 70 (0-20) 1480 (200-6400) 1510 ( 800-3200 )
SAT b
_ 140 (50-400) 400 ( 200-3200 ) 40 (1-100) 1980 (200-3200) ~>200
2ME"
10 (0-50) 350 (200-400) ~>200
_ 230 (100-400) 400
RIV d
(2/20-4/640) 4/80
(0-l/40)
(1/80-4/200)
( 1/20-1/80)
-
('F~
I0/10
I0/10
1/10
1/10
0/10
Culture results no. positive/ total
Fig. 2. panel G
Fig. 2, panel F
Fig. ,," panel E
Fig. 2, panel t)
Fig. 2, panel B Fig. 2. panel (
Western blot cross reference
~Complement fixation test: the d e n o m i n a t o r is tile highest dilution showing fixation and the numerator indicates the degree of fixation. No geometriL mean titers are calculated. "No reaction in test.
~'Titers are geometric mean titers: range o f liters is indicated in parentheses. ~'Standard agglutination tube test. ~2-mercaptoelhanol test. dRivanol precipitation plate agglutination test.
10/10
10/10
7/10
10/10
Card p o s / total
Group
Summary of serologic reactions of various groups of cattle"
TABLE 2
BOVINE ANTIBODY TO B. ABORTUS BY WESTERN BLOTTING
85
ally less of the 31 000 to 45 000 group and in addition contained a heavily stained band at a Mw of 20 000.
Serologic reactions The geometric means of the serologic reactions of the cattle in each group are detailed in Table 2. Isolation of B. abortus from tissue or milk samples of the cattle at the termination of the vaccine experiment are also indicated. The Brucella-free cattle were seronegative by all tests. Sera of the vaccinated calves obtained 3 weeks post-vaccination were card-test positive as expected. The standard agglutination tube (SAT) test, 2-mercaptoethanol test, Rivanol, and complement fixation tests were all positive, and reactions varied considerably among the animals. The adult-vaccinated cows at 3 or 4 weeks post-vaccination varied similarly in their serologic reactions, and were approximately one dilution higher in titer compared with the vaccinated calves. The vaccinated challenge-exposed cows at 16 weeks post-exposure (2 years post-vaccination ) showed generally much lower titers than the former two groups, and were culture-negative at subsequent parturition and necropsy. Sera of the infected cattle at 16 weeks post-exposure had high titers in the serologic tests, as did those of naturally infected cows. Western blots Fig. 2 depicts the results of western blots developed with a representative cattle serum from one each of the five different groups although all sera were tested by this procedure. Fig. 2, panel A, shows the Amido Black-stained blot, followed by an immunoblot. The Amido Black-stained blot showed fewer bands than the Coomassie Brilliant Blue-stained SDS gels (Fig. 1, and panel A of Fig. 2). None of the lower Mw bands (i.e., >20 000 or those of Mw > 45 000) were observed with Amido Black. The immunoblots of the Brucella-free control cattle and the vaccinated calves showed no reaction under the condition of the procedures used (Fig. 2, panels B and C, respectively). Although seven of the adult-vaccinated cattle showed no reaction on the immuno-blot, three out of ten cattle sera showed activity primarily toward the protein bands in the Mw range of 31 000 to 45 000 and one of these toward a second group of proteins in the Mw range of 66 000 to 71 000 (Fig. 2, panel D). B. abortus strain 19 was isolated from one of the latter cows, but the bacteriologic status of the other two cows was unknown. With sera from calfhood vaccinated, and subsequently experimentally infected cattle, typically, a strong reaction with fraction 0-35 and 0-70 covering the entire Mw range from 31 000 to 45 000 was observed (lanes 1 and 2 of Fig. 2, panel E). Additional bands were observed at apparent Mw of 66 000 and 71 000 (Fig. 2, panel E, lanes 1-4). No reactions were observed with the 20 000 dalton protein under the conditions of the experiments. Six often sera
~6
~, B t - I Z I ! R [ I
i
mol wt
Z
4
3
i
2
3
4
J
2
3
4
I
2
3
\l
4
x I0 -3
71453[-
20 - 4
D
B
A
71-tl
C
D
"
4531-
4
I
20-
E
F
G
Fig. 2. Western blot of B. abortus protein fractions with cattle sera from various groups. Panel A, Amido Black stain of protein blot; panel B, Brucella-free; panel C, vaccinated;panel D, adultvaccinated; panel E, calfhood vaccinated-challengeexposed; panel F, cxperimentally infected; panel G, naturally infected. Lanes 1-4, contain fraction 0-35, fraction 0-70, fraction 35-70 and fraction 70-100, respectively. from the non-vaccinated, experimentally-infected group reacted with all four fractions (Fig. 2, panel F, lanes 1-4). Three bands could be observed corresponding to Mw to 31 000, 39 000, and 45 000. The remaining four sera reacted with fraction 0-35 and 0-70 only (data not shown). Bands of higher or lower Mw were not observed in the immunoblots. A representative i m m u n o b l o t of one of the sera from naturally infected cattle is shown in Fig. 2, panel G. Nine of ten sera of naturally infected cattle reacted weakly with the 0-35 fractions as shown in Fig. 2, panel G, lane l, and moderately with fraction 0-70 as shown in Fig. 2, lane 2. The only serum from this group which did not react on the blot (data not shown) was also tested in an ELISA procedure (Tabatabai and Deyoe, 1984b). The IgG response for this serum, measured in terms of absorbance units, was just above
BOVINE ANTIBODY TO B. ABORTUS BY WESTERN BLOTTING
87
the minimal level of 0.142 absorbance units (Tabatabai and Deyoe, 1984b), below which the IgG response is considered negative. DISCUSSION The experiments described in this paper demonstrate that the western blot technique with the BCSP antigens may distinguish the serologic reactions of calfhood-vaccinated (Fig. 2, panel C) from infected cattle (Fig. 2, panels F and G). Sera from both the experimentally-infected group and all but one of the naturally infected group reacted with the BCSP antigens, while none of the sera of vaccinated cattle reacted with the BCSP fractions. These results contrasted with those of the adult-vaccinated group. Although seven of ten sera did not react, three of the sera from the latter group reacted weakly with BCSP antigens similar to those of sera from the naturally infected cows (Fig. 2, panels D and G), and the experimentally-infected cow (Fig. 2, panel F). Considering this was a relatively short post-vaccination sampling time interval (3 or 4 weeks), it is reasonable to expect that the remaining blot-positive animals also would become negative unless a persistent infection had been established as was observed for one cow from which strain 19 had been isolated. Generally, the western blot technique detected fewer of these adultvaccinated animals than did the SAT test. The blots with sera from vaccinated challenge-exposed animals showed intense reactions to the BCSP fractions (Fig. 2, panel E), which differed qualitatively and quantitatively from the experimentally-infected group (Fig. 2, panel F) but resembled some of the blots from the naturally infected group (data not shown). These results were consistent with re-exposure of the animals to brucellae, as some of these serum samples came from animals with recurrent infections from which B. abortus biotype was periodically isolated. (D. Stringfellow, personal communication). We observed that re-exposure by experimental infection of the animals to brucellae produced a reaction with two high Mw bands (Fig. 2, panel E); the gene for one of these proteins (the 66 kDa band) has been cloned (unpublished data, Chin, Mayfield and Tabatabai, 1989). It remains to be seen how often this reaction occurs before it can be considered of diagnostic value. One serum of this group of animals showed a weak reaction on the immunoblot. This may be due to a lack of secondary i m m u n e response which reportedly had been associated with infection (Sutherland, 1985). Culture data showed that B. abortus 2308 was isolated from this animal. Although the blot was positive whenever an experimental or natural infection occurred, the blot technique failed to detect one of the naturally infected animals for which the serologic results were similar to those for the other cows in the group. When this serum was tested in the ELISA with fractions 0-35 and 0-70, it was found that the level of IgG in the serum was much lower than
NN
~ \ ~{I t Z t
R I I \t
the mean IgG level tbr the remainder of the cows (Tabatabai and Belzer, unpublished results). Based on minimal values of absorbance in the ELISA, the ELISA result for this cow serum was just below the level designated as positive (Tabatabai, 1984b). Undoubtedly, this animal would have been detected at a lower serum dilution than currently used for the blot ill order to keep the background staining acceptably low. The antibody specificity of the reactions detected in the western blots is apparently primarily directed toward IgG, because no or very faint reactions were observed when an anti-bovine IgM heavy chain-specific conjugate was used (Tabatabai and Belzer, unpublished results). Interestingly, the intensity of reaction on the blots appeared to be inversely related to the corresponding titers of the serologic reactions for the vaccinated challenge-exposed and the experimentally-infectedgroups. These results indicate that antibody specificity toward Brucella antigens may differ considerably, depending on the vaccination and/or challenge exposure history of the animal (Butler et al., 1986 ). It is reasonable to assume, based on the data in Table 1, that the observed intensity of staining of the blots was due entirely to antibody reacting with residual protein-bound LPS rather than with protein. However, the increase in intensity on the blot with certain sera does not reflect a similar increase in serologic titers. Presumably, the two methods measured antibodies of different specificities or classes, as has been observed for other serologic tests tbr bovine brucellosis ( Patterson et al., 1976 ). Fractionation of the BCSP antigens by ammonium sulfate precipitation offered no advantage for detecting differences between groups. In general, under the conditions of the test used fraction 0-35 reacted with a larger number of sera from the naturally infected group than fraction 0-70. Both fractions reacted equally well with sera from the remaining groups, while fractions 3570 and 70-100 bound bovine IgG antibody reacted poorly in the blot. The attractive feature of the blotting technique is that sera from calfhoodvaccination animals do not react. As reported for other serologic tests (Sutherland, 1984), sera from adult-vaccinated animals may pose a problem with this technique as well. However, if adult-vaccinated cattle are kept under quarantine until seronegative by the card test the adult-vaccinated cattle would be no problem with this technique.
ACKNOWLEDGMENTS
We gratefully acknowledge the expert technical assistance of Kathryn Meredith and Carol Irvin. We thank Tom Glasson for his expert photographic work, and Nancy Kilstofte for manuscript preparation.
BOVINE A N T I B O D Y T O B. ABORTUS BY WESTERN B L O T T I N G
89
REFERENCES Alton, G.G., Jones, L.M. and Pietz, D.E., 1975. Laboratory techniques in brucellosis, 2nd Edn. WHO, Washington, DC, pp. 11-63. Ashwell, G., 1966. New colorimetric methods of sugar analysis. Methods Enzymol., 8: 85-95. Butler, J.E., Seawright, G.L., McGivern, P.L. and Gildsdorf, M., 1986. Preliminary evidence for a diagnostic immunoglobulin G1 antibody response among culture-positive cows vaccinated with Brucella abortus strain 19 and challenge exposed with strain 2308. Am. J. Vet. Res., 47: 1258-1264. Dubois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A. and Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem., 28: 350-356. Gershoni, J.M., 1985. Protein blotting: development and perspectives. Trends Biochem. Sci., 10: 103-106. Lamb, V.L., Jones, L.M., Schurig, G.G. and Berman, D.T., 1979. Enzyme-linked immunosorbent assay for bovine immunoglobulin subclass-specific response to Bruce/& abortus lipopolysaccharides. Infect. Immun., 26: 240-247. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J., 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem., 193: 265-275. Manthei, C.A., Kuttler, A.K. and Goode, Jr., E.R., 1956. The Yearbook of Agriculture. Proc. Anim. Dis. 84th Congr., 2D Session, House Document No. 344, U.S. Dept. of Agric., Washington, DC, pp. 202-213. Nicoletti, P., 1980. Epidemiology of bovine brucellosis. Adv. Vet. Sci. Comp. Med., 24: 69-98. Nielsen, K.H., Heck, F.C., Stiller, J.M. and Rosenbaum, B., 1983. Interaction of specifically purified isotypes of bovine antibody to Brucella abortus in the hemolysis in gel test and enzyme-linked immunosorbent assay. Res. Vet. Sci., 35: 14-18. Nielsen, K., Wright, P.F., Cherwonogrodzky, J., Duncan, J.R. and Stemshorn, B., 1987. Enzyme immunoassay for diagnosis of bovine brucellosis. Ann. Inst. Pasteur, Microbiol., 138: 69-144. Partanen, P., Turanen, H.J., Paasivis, R., Forshblom, E., Suni, J. and Leinikki, P.O., 1983. Identification of antigenic components of Toxoplasma gondii by an immunoblotling technique. FEBS Lett., 158: 252-254. Patterson, J.M., Deyoe, B.L. and Stone, S.S., 1976. Identification ofimmunoglobulin associated with complement-fixation agglutination and low pH-buffered antigen tests for brucellosis. Am. J. Vet. Res., 37: 319-324. Ruppanner, R., Meyer, M.E., Willeberg, P. and Behymer, D.E., 1980. Comparison of the enzyme-linked immunosorbent assay with other tests for brucellosis using sera from experimentally infected heifers. Am. J. Vet. Res., 41: 1329-1332. Saunders, G.C., Clinard, E.H., Bartlett, M.L. and Sanders, W.M., 1977. Application of the indirect enzyme labeled antibody microtest to the detection and surveillance of animal diseases. J. Infect. Dis., 136: 5258-5266. Schurig, G.G., Jones, L.M., Speth, S.L. and Berman, D.T., 1978. Antibody response to antigens distinct from the smooth lipopolysaccharide complex in Brucella infection. Infect. lmmun., 21: 994-1002. Stemshorn, B.W., 1984. Recent progress in the diagnosis of brucellosis. Dev. Biol. Stand., 56: 325-340. Stemshorn, B. and Nielsen, K., 1977. The bovine immune response to B. abortus: water soluble antigen precipitated by sera of some naturally infected cattle. Can. J. Comp. Med., 41 : 152159. Sutherland, S.S., 1984. Evaluation of the enzyme-linked immunosorbent assay in the detection of cattle infected with Brucella abortus. Vet. Microbiol., 10: 23-32. Sutherland, S.S., 1985. Comparison of enzyme-linked immunosorbent assay and complement
9()
~ ~, I ~ t l Z I R I
I \1
fixation test tbr the detection of specific antibody in cattle \accmatcd and challenged wilh Ilrucella at~rlu~. J. ('lin. Microbiol., 22: 44-47. TabatabaL L.B. and Deyoe, B i . , 1984a. Biochemical and biological propcrties of soluble protern preparations from Ilrucella at~n'tu,s. Dev. Biol. Stand., 56:199-211. Tabatabai, L.B. and Deyoe, B.L., 1984b. Specific enzyme-linked immunosorbcnt assay t'or detection of bovine antibody to Brucella ab##rtus, lnfecl, lmmun., 26: 668-679. Tabatabai, L.B., Deyoe, B i . and Ritchie, A.E., 1979. Isolation and characterization of loxic fractions from Ilrucella al~orlu#. Infect. lmmun., 26: 668-679. Towbin, H., Staehelin. T. and Gordon, J,, 1979. Electrophorefic Iransfer ol~proteins from acr~,lamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA, 76: 4.350-4354. Tsai, C.M. and Frasch, C.E., 1982. A sensitive silver stain for detecting lipopolysaccharides in polyac~lamide gels. Anal. Biochem., 119; 115-119. Warren, L., 1959. The thiobarbituric acid assay of sialic acids. J. Biol. Chem., 245:197 I-1975. Weber, K. and Osborn, M., 1969. The reliability of molecular weight determinations by dodecyl sulfate polyacrylamide gel electrophoresis. J. Biol. Chem., 244:4406-4412.
