Agriculture Canada, Animal Diseases Research Institute, Nepean, P.O. Box 11300, Station H, Ottawa, ... relate the host response to defined components of the.
JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 1991, p. 1659-1664 0095-1137/91/081659-06$02.00/0 Copyright © 1991, American Society for Microbiology
Vol. 29, No. 8
Chromatographic Purification and Characterization of Antigens A and D from Mycobacterium paratuberculosis and Their Use in Enzyme-Linked Immunosorbent Assays for Diagnosis of Paratuberculosis in Sheep E. A. SUGDEN,1* B. W. BROOKS,' N. M. YOUNG,2 D. C. WATSON,2 K. H. NIELSEN,' A. H. CORNER,1 C. TURCOTTE,1 A. MICHAELIDES,l AND R. B. STEWART' Agriculture Canada, Animal Diseases Research Institute, Nepean, P.O. Box 11300, Station H, Ottawa, Ontario, Canada K2H 8P9,1 and Division of Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada KIA 0R62 Received 28 November 1990/Accepted 23 May 1991 The protein antigens A and D were purified from culture filtrates and sonic extracts of laboratory strains of
Mycobacterium paratuberculosis by salt precipitation and chromatography. The characterization of antigen A is shown here, and both antigens were evaluated along with lipoarabinomannan antigen in indirect enzyme-linked immunosorbent assays (ELISA) for the serodiagnosis of ovine paratuberculosis. After anionexchange (DEAE-5PW) and hydrophobic (phenyl-5PW) chromatography using high-performance liquid chromatography, antigen A showed a prominant band in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) at 31 kDa with small amounts of low-molecular-mass proteins but with no evidence of antigen D. A single precipitin arc was evident with purified antigen A in crossed immunoelectrophoresis. The determination of the N-terminal amino acid sequence showed a high degree of homology between the 31-kDa component of antigen A and antigens of the BCG85 complex of Mycobacterium bovis BCG, a total of 24 of 26 residues being identical to those of BCG85C. A prominant SDS-PAGE band at 400 kDa and a single crossed-immunoelectrophoresis arc was also evident for antigen D after gel filtration (Sephacryl S-200), anion-exchange (DEAE-Sephacel), and concanavalin A-Sepharose affinity chromatography. By ELISA, purified antigen A detected antibody in the sera of 18 of 22 paratuberculosis-infected sheep (82% sensitivity), whereas the purified antigen D detected antibody in all 22 infected animals (100% sensitivity). Combined ELISA results showed increased specificity with some loss in sensitivity.
In view of a need for more sensitive and specific serological protocols for detection of ruminant paratuberculosis (7), this laboratory has focused on the use of purified antigens such as lipoarabinomannan (LAM) (19) in enzyme-linked immunosorbent assays (ELISA) (17). Use of ELISA for the serodiagnosis of paratuberculosis in cattle has proven to show higher sensitivity as compared with the complement fixation test (1) and agar gel immunodiffusion (AGID) (26). Additional advantages of this approach lie in an improved quality control of antigen preparations and the ability to relate the host response to defined components of the infecting microorganism. Two precipitin lines arising from protein components designated A and D of Mycobacterium paratuberculosis have been assessed in the serodiagnosis of ovine paratuberculosis by using sonic extracts of the bacilli as the antigen in the AGID (A-AGID and D-AGID) and crossed-immunoelectrophoresis (CIE) tests (5), resulting in a high sensitivity for D-AGID as compared with that of the complement fixation test using polysaccharide antigen (5, 17). In this work, antigens A and D were chromatographically purified and evaluated in ELISA. Furthermore, the N-terminal amino acid sequence was determined on a prominent protein in purified antigen A.
*
MATERIALS AND METHODS
Preparation of culture filtrates and sonic extracts. Laboratory strains V and C-286 of M. paratuberculosis were grown in modified Long's synthetic liquid medium for 12 weeks at 37°C as previously detailed (5, 6). The cultures were filtered through Whatman filter paper no. 2, the bacilli being retained for the preparation of sonic extracts (see below). The culture filtrate was passed through a D-6 Seitz filter pad and a 0.22-p.m-pore-size Millipore filter (Millipak 60; Millipore, Bedford, Mass.) and was concentrated to approximately 4 mg/ml with a hollow fiber device (HlOP3-20; Amicon, Danvers, Mass.) and a TCF-10 concentrator with PM-10 ultrafilters (Amicon). Protein content (Bio-Rad Laboratories, Richmond, Calif.) (4) and carbohydrate content (9) were also determined. For the preparation of sonic extract, the bacilli were suspended in saline (0.67 g of bacilli [wet weight] per ml) and sonicated for 22 min at 90% pulse and 90 W (sonicator model W-375; Heat Systems-Ultrasonics, Farmingdale, N.Y.) in a water-cooled chamber (5). The sonicated material was clarified and concentrated as for the culture filtrate to approximately 4 mg/ml with an Amicon PM-10 ultrafilter. Purification procedures; purification of antigens A and D from culture filtrates and sonic extracts. Figure 1 summarizes the purification protocol for A and D antigens. The presence of these antigens was monitored by immunoprecipitation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Protein was precipitated by adding 0.25 g of
Corresponding author. 1659
1660
SUGDEN ET AL.
J. CLIN. MICROBIOL.
Culture Filtrate or Sonicate
Ultrafilter Concentrate
Ammonium Sulphate Precipitate
l
l
DEAE-5PW HPLC (2x)
Sephacryl S-200
Phenyl-5PW HPLC
DEAE Sephacel
I l
DEAE-5PW HPLC
I A antigen
I I
Con A Sepharose 4B
Il
D antigen
FIG. 1. Purification protocol for antigens A and D from cultures of M. paratuberculosis C-286.
(NH4)2SO4 to each ml of the clarified concentrated culture filtrate or sonic extract. The precipitate was redissolved in 0.01 M Tris-hydrochloride, pH 7.5, and dialyzed overnight versus 0.15 M NaCl by using tubing with a molecular mass cutoff of 12 to 14 kDa (Spectrum Medical Industries, Inc., Los Angeles, Calif.). The solution was clarified, reprecipitated, and then reconstituted in a minimum volume of 0.05 M Tris-hydrochloride, pH 8.0, for antigen A, or 0.01 M Trishydrochloride, pH 7.5, for antigen D (approximately 50 ml if the concentrate volume was 500 ml). Residual (NH4)2SO4 was removed by dialysis versus these buffers, and the solutions were centrifuged. The resulting sediments were resuspended in 20 ml of fresh buffer to facilitate thorough dissolution of antigens and, after clarification, were pooled with the initial supernatants. The reconstituted material for antigen A purification was chromatographed by using an LKB high-pressure liquid chromatography (HPLC) system on a DEAE-5PW HPLC column (21.5 by 150 mm; Pharmacia Fine Chemicals, Uppsala, Sweden) and eluting with a composite sodium chloride gradient. A typical gradient at 4.0 ml/min was no NaCl until 15 min, 0.2 M NaCl at 75 min, 0.5 M NaCl at 150 min, and 2.0 M NaCI at 200 min. The material in that section of the major peak containing the most A antigen with the least interference was dialyzed versus 0.05 M Tris-hydrochloride, pH 8.0, and rechromatographed on the DEAE-SPW column. The resultant single peak pool was adjusted to 20% saturated (NH4)2SO4-0.05 M Tris-hydrochloride, pH 8.0, and was chromatographed on a phenyl-5PW column (7.5 by 75 mm), eluting with a decreasing gradient of (NH4)2SO4. A typical gradient at 1 ml/min was 20% saturated (NH4)2SO4 until 10 min, 12% saturated (NH4)2SO4 at 30 min, 6% saturated (NH4)2SO4 at 50 min, 4% saturated (NH4)2SO4 at 90 min, and no (NH4)2SO4 at 115 min. The antigen A-enriched pool was rechromatographed on the DEAE-5PW column. This
was dialyzed versus water and lyophilized in 0.1-mg aliquots. The reconstituted material for antigen D purification was concentrated to 10 ml by ultrafiltration and chromatographed on a column (85 by 1.6 cm) of Sephacryl S-200 (Pharmacia) preequilibrated with 0.1 M NaCI-0.01 M Tris-hydrochloride, pH 7.5. The void pool containing antigen D was dialyzed versus 0.05 M Tris-hydrochloride, pH 8.0, applied to a column (2.0 by 1.6 cm) of DEAE-Sephacel (Pharmacia), and eluted with a linear NaCl gradient (8). The appropriate pool was dialyzed versus 0.1 M Tris-hydrochloride, pH 7.2, 1 mM CaCl2, 1 mM MnCl2, and 1 mM MgCl2 (concanavalin A [ConA] buffer [10]) and passed through a column of ConASepharose 4B (Pharmacia) equilibrated with ConA buffer. This material was dialyzed versus water and lyophilized in 0.1-mg aliquots. Occasionally a predigestion of (NH4)2SO2precipitated culture filtrates and sonic extracts with proteinase K (Sigma Chemical Co., St. Louis, Mo.) was also done
(6).
SDS-PAGE and CIE determinations. SDS-PAGE in slabs was performed in a vertical electrophoresis unit (Protean II; Bio-Rad Laboratories (Canada) Ltd., Mississauga, Ontario, Canada) by a modification of the discontinuous system described by Laemmli (13). The stacking gel contained 6% acrylamide, and the separating gel was a 10 to 15% linear acrylamide gradient. No SDS was included in the stacking or separating gels. The SDS-PAGE sample buffer contained SDS (1.25% [wt/vol]), glycerol (12.5% [vol/vol]), 0.0625 M Tris-hydrochloride (pH 6.8), 2-mercaptoethanol (1.25% [vol/ vol]), and bromophenol blue. Electrophoresis was conducted at 20 mA of constant current per gel. Proteins were visualized with Coomassie blue, and the silver stain method of Tsai and Frasch (20) was used for detection of carbohydrates. CIE was performed, as previously described, in an LKB Multiphor (Pharmacia) electrophoresis system (5) using pooled sera from one herd of paratuberculosis-infected goats. In the case of tandem CIE, two 4-mm-diameter wells with centers 0.8 cm apart were used. Determination of peroxidase activity. The substrate-chromogen contained 4.0 mM H202 plus 1.0 mM 2,2'-azino-di(3-ethylbenzthiazoline sulfonic acid) in 0.05 M sodium citrate, pH 5.0 (9). To 2.0 ml of this freshly made mixture was added 0.2 ml of a sample followed by a 1-h incubation at room temperature and evaluation at 414 nm. Characterization of antigen A. For N-terminal amino acid sequencing, the 31-kDa component of the antigen A preparation was separated by SDS-PAGE and transferred electrophoretically to a polyvinylidene difluoride membrane (14). The membrane segment bearing the antigen was cut out and sequenced with a model 475A sequencer (Applied Biosystems Inc., Foster City, Calif.) and on-line identification of the phenylthiohydantoin derivatives. Carbohydrate compositions were determined with a BioLC analyzer with a pulsed amperometric detector (Dionex Corp., Sunnyside, Calif.) for samples hydrolyzed for 16 h at 100°C in 0.5 M trifluoroacetic acid (12). For amino acid composition, a sample of antigen A was purified by size exclusion HPLC on a TSK-2000 column run in 6 M guanidinium hydrochloride-1 M Na2SO4-0.05 M NaH2PO4, pH 5.9. After exhaustive dialysis against water, duplicate aliquots of the purified antigen were hydrolyzed in vacuo with 6 N HCI at 110°C for 22 h. The amino acid contents of the hydrolysates were measured with a Durrum D-500 analyzer. Indirect ELISA. Serum was taken from 22 paratuberculosis-infected sheep culled from a flock on the basis of performance and/or serology (complement fixation test and
1661
PROTEIN ANTIGENS OF M. PARATUBERCULOSIS
VOL. 29, 1991
D-AGID) and evaluated for paratuberculosis and caseous lymphadenitis (infection with Corynebacterium pseudotuberculosis) by histopathology and culture as previously described (5, 11, 17). Serum was also taken from 58 minimal disease sheep maintained at this institute. Polystyrene microtiter plates (Nunc-ImmunoPlate IF, no. 439454; GIBCO/ BRL, Burlington, Ontario, Canada) were coated with LAM antigen, antigen A, and antigen D at 1.0, 1.0, and 0.1 ,ug/ml, respectively, in 0.06 M sodium carbonate buffer, pH 9.6, overnight. After four washes with 0.01 M Tris-hydrochloride, pH 8.0, containing 0.15 M NaCl and 0.05% Tween 20 (TBS-T), ovine sera, prediluted 1/200 for LAM antigen and 1/50 for antigens A and D with TBS-T, were incubated in the plates for 3 h at room temperature in diagonal quadrants (24). After another wash, diluted alkaline phosphatase-labelled monoclonal anti-bovine immunoglobulin Gl antibody (M-23) was added to the plates and incubated for 2 h. This conjugate, which cross-reacts with ovine or caprine sera, was made according to the one-step glutaraldehyde method (2). After another wash, the hydrolysis of the p-nitrophenyl phosphate substrate dissolved in diethanolamine buffer (15) was monitored at 405 nm with a microtiter plate reader (Flow Laboratories, McLean, Va.) as previously described (25) and following a kinetic approach (3). The maximum acceptable allowable coefficient of variation between duplicate wells is 21.7 + (2.65/average optical density [OD]). The cutoff for establishing positive and negative OD values was taken as the ninth decile (16) of the minimal disease flock serum OD values. Sensitivity and specificity are defined in the footnotes to Table 2 (see below). RESULTS
Figure 2a shows the final step in the chromatographic purification of antigen A from the culture filtrate of strain V by using a DEAE-5PW column and showing a single peak. As in the first two elutions from this column, antigen A eluted at approximately 0.1 M NaCl. The second chromatography from DEAE-5PW resulted in a single peak as well. Elution from phenyl-5PW used a decreasing gradient of (NH4)2SO4 in the third chromatography, resulting in further purification, the single peak from DEAE-5PW being resolved into antigen A-rich material eluting with 5 to 0% saturated (NH4)2SO4 from this hydrophobic column, after other material eluted earlier. Figure 2b shows the DEAESephacel step in the chromatographic isolation of antigen D from the culture filtrate of strain C-286 with antigen eluting at 0.4 to 1.0 M NaCl. The first step involved chromatography through Sephacryl S-200 whereby the antigen D was present in the void pool. The final step involved passage through ConA-Sepharose, which removed extraneous carbohydrate and allowed antigen D to pass through (data not shown). Peroxidase activity was also determined in chromatographic fractions and was associated closely with antigen D. Indeed, antigen D has recently been characterized as a peroxidase-active bacterioferritin (6). Figure 3 shows an SDS-PAGE evaluation of antigen A and D isolation from M. paratuberculosis C-286 revealing that the antigen A preparation is free of antigen D and vice versa and that low-molecular-mass (
