C. Gyles, Guelph, Canada per lane, and the proteins were visualized by staining with ... this study were purchased from Fisher Scientific (Don Mills,. Ontario).
INFECTION AND IMMUNITY, JUlY 1987, p. 1626-1634
Vol. 55, No. 7
0019-9567/87/071626-09$02.00/0 Copyright C) 1987, American Society for Microbiology
Analysis of Major Antigens of Haemophilus (Actinobacillus) pleuropneumoniae and Related Organisms JANET
Department
of
I.
MACINNES* AND S0REN ROSENDAL
Veterinary Microbiology and Immunology, University of Guelph, Guelph, Ontario, Canada NIG 2WJ Received 20 January 1987/Accepted 4 April 1987
Outer membrane protein (OMP)-enriched extracts and whole-cell protein preparations of Haemophilus (Actinobacillus) pleuropneumoniae and related organisms were examined by polyacrylamide gel electrophoresis and immunoblotting. Both the OMP-enriched and whole-cell protein profiles of Actinobacillus suis, A. pleuropneumoniae (NAD-independent biovar), A. lignieresii, and Pasteurella haemolytica were very similar to those of H. pleuropneumoniae serotypes 1 to 8. Antisera prepared against H. pleuropneumoniae typically recognized three major OMP antigens with approximate molecular weights of 17,000 (17K), 32K, and 42K in immunoblots of H. pleuropneumoniae serotypes 1 to 8, Actinobacillus spp., and P. haemolytica. Antisera prepared against Actinobacillus spp. and Haemophilus sp. "minor group" also recognized these 17K, 32K, and 42K antigens. Using absorbed sera, we demonstrated that the 17K antigen had an epitope (or epitopes) common to all the gram-negative organisms examined, including Escherichia coli. The 32K and 42K antigens had epitopes common to members of the family Pasteurellaceae but, in the case of the 32K antigen, also contained unique epitopes. These results provide a basis for understanding the lack of specificity of serodiagnostic tests for H. pleuropneumoniae infection and provide another line of evidence for the association of H. pleuropneumoniae with the genus Actinobacillus.
Haemophilus (Actinobacillus) pleuropneumoniae is an important pathogen of swine throuthout the world (11). It can cause acute fibrinous pneumonia and pleuritis with high mortality, or chronic lung lesions resulting in growth rate depression (11, 28). H. pleuropneumoniae is host specific and is transmitted among pigs in close contact by aerosol (11; R. Nielsen, Ph.D. thesis, Royal Veterinary and Agricultural University, Copenhagen, Denmark, 1982). Pigs which survive infection may acquire protective immunity but often suffer from chronic lesions and frequently become subclinical carriers of the pathogen (11, 13, 17). When carrier animals are introduced into a herd not previously exposed, a serious outbreak of pleuropneumoniae is likely to ensue (11, 26). Current vaccines against H. pleuropneumoniae are able to reduce mortality but do not prevent chronic forms of the disease or development of the carrier state (5, 11, 13, 14, 28). To control the spread of pleuropneumonia, it is therefore necessary to detect carriers. Isolation of bacteria from the upper respiratory tract of carrier pigs is not a reliable or practical diagnostic procedure in a control program (2, 6). Pigs naturally or experimentally infected with H. pleuropneumoniae develop high antibody titers (12, 13, 28). Thus, serodiagnostic tests have been developed for the detection of H. pleuropneumoniae antibodies (3, 7, 8, 10). The sensitivity and specificity of these tests are generally good when the tests are applied to sera from animals infected in controlled environments, but the specificity is unsatisfactory when sera from animals in the field are tested (27; G. Goyette, S. Lariviere, K. R. Mittal, R. Higgins, and G. P. Martineau, Abstr. Int. Pig Vet. Soc., Barcelona, 1986, p. 258). To determine the basis of the cross-reactivity seen in field sera and to search for serotype- and species-specific antigens, we investigated whole-cell protein and outer membrane protein (OMP)-enriched fractions of H. pleuropneu*
moniae serotypes 1 to 8 and related porcine Haemophilus and Actinobacillus spp. In addition to studying Coomassie blue-stained polyacrylamide gel profiles, we analyzed protein extracts of these organisms for immunodominant molecules by immunoblotting using both rabbit and swine antisera. We demonstrated several major antigens in H. pleuropneumoniae serotypes 1 to 8: a 17,000-molecular-weight (17K) antigen, common to the gram-negative organisms used in this study, and 32K and 42K antigens which contained epitopes common to some members of the family Pasteurellaceae. In addition, we identified several other antigens which may be species or serotype specific. MATERIALS AND METHODS Bacterial strains and growth conditions. The strains used and their sources are listed in Table 1. The NADindependent strain, SW1004, is referred to as A. pleuropneumoniae, and all NAD-dependent strains are designated as H. pleuropneumoniae. NAD-requiring organisms were grown on Trypticase soy agar and 5% bovine blood, heated to 80°C for 15 min (chocolate agar) and supplemented with 0.01% (wt/vol) NAD. NAD-independent strains were cultured on fresh blood agar without NAD. For preparation of OMPenriched fractions, organisms were cultured to late logarithmic phase in tryptone-yeast extract (TYE) broth supplemented with 0.01% NAD as required (18). Analysis of whole-cell proteins by PAGE. Overnight cultures were harvested from the plates by washing with phosphate-buffered saline (pH 7.4; 8 mM Na2HPO4, 14 mM KH2PO4). After one wash with phosphate-buffered saline, the bacteria were suspended 1:1 (vol/vol) in 2x sample buffer containing 20% (vol/vol) glycerol, 10% (vol/vol) 2mercaptoethanol, 6% (wt/vol) sodium dodecyl sulfate (SDS), 0.125 M Tris (pH 6.8) and 0.1% (wt/vol) bromophenol blue. The samples were diluted 1:10 with lx sample buffer and boiled for 5 min. The undissolved material was sedimented by centrifugation. For polyacrylamide gel electrophoresis (PAGE), approximately 10 pd of the supernatant was loaded
Corresponding author. 1626
VOL. 55, 1987
MAJOR ANTIGENS OF H. PLEUROPNEUMONIAE
1627
TABLE 1. Bacteria used and sources of cultures Strain
Donor
Haemophilus pleuropneumoniae Shope 4074 (serotype 1) ......................... CM5 (serotype 1) .........................
R. Nielsen, Copenhagen, Denmark Own collection S358 (serotype 1) ............. Own collection , MA66 (serotype 1) ............. Own collection Own collection CS177 (serotype 1) ............. WLN344 (serotype 1) ............. Own collection VSB271 (serotype 1) ............. Own collection LL164 (serotype 1) ............. Own collection S1536 (seroytpe 2) ............. R. Nielsen, Copenhagen, Denmark S1421 (serotype 3) ............. R. Nielsen, Copenhagen, Denmark M62 (serotype 4) ............. R. Nielsen, Copenhagen, Denmark K17 (serotype 5) ............. R. Nielsen, Copenhagen, Denmark R. Nielsen, Copenhagen, Denmark Fem$ (serotype 6) ............. WF83 (serotype 7) ............. Own collection 405 (serotype 8) ............. R. Nielsen, Copenhagen, Denmark Actinobacillus A. pleuropneumoniae SW1004 ............. H. Bertschinger, Zurich, Switzerland A. suis ATCC 15557 ............. American Type Culture Collection A. lignieresii ATCC 19393 ............. American Type Culture Collection A. equuli ATCC 19392 ............. American Type Culture Collection Haemophilus H. influenzae ATCC 19418 ............. American Type Culture Collection H. parasuis ATCC 19417 ............. American Type Culture Collection H. parasuis E751 ............. J. Stevens, Ottawa, Canada H. parasuis A9 ............. J. Stevens, Ottawa, Canada H. parasuis CS ............. J. Stevens, Ottawa, Canada "Minor group" and atypical Haemophilus spp. Minor group NM305 ......................... Own collection Minor group 202 ......................... M. Kilian, Aarhus, Denmark Untypable PN33 ......................... Own collection Pasteurella haemolytica serotype 1 ......................... P. Shewen, Guelph, Canada Escherichia coli V517 .......................... C. Poppe, Guelph, Canada ......................... C. Gyles, Guelph, Canada Salmonella typhimurium
lane, and the proteins were visualized by staining with Coomassie blue. The discontinuous buffer system of Laemmli as described by Rodriguez and Tait (23) was used. The separating gel contained 12.5%T-2.7%CBi, and the stacking gel contained 4.0%T-2.7%CBi,. (%T is the polyacrylamide gel concentration defined as the percentage of total monomers [i.e., acrylamide plus cross-linking agent in grams per 100 ml]; %CBiS is the percentage of bisacrylamide cross-linker.) Chemicals for PAGE were purchased from Serva Fine Chemicals (Garden City, N.Y.). Preparation of OMP-enriched extracts. Sarkosyl-insoluble OMP-enriched fractions were prepared by the method of Barenkamp et al. (1). Briefly, cells were harvested, washed, and suspended in HEPES (N-2-hydroxyethylpiperazine-N'2-ethanesulfonic acid)-buffered saline. The bacteria were disrupted by sonication, and debris and unbroken cells were removed by centrifugation. The supernatant fluid was centrifuged at 100,000 x g for 1 h. The sediment was suspended in sodium sarcosylate buffer and incubated at room temperature for 20 min. The undissolved material was sedimented at 100,000 x g and then suspended in distilled water. The proteins were dissolved in sample buffer, boiled, and loaded onto gels as described above. per
Immunoblotting. Immunoblotting
was
performed by
a
modification of the method of Towbin et al. (29). After electrophoresis of whole-cell protein preparations or OMPenriched extracts, gels were soaked in transfer buffer containing 25 mM Tris base, 192 mM glycine, and 20% (vol/vol) methanol for 30 min and then electroblotted onto nitrocellu-
lose for 12 h at 20 V or for 4 h at 60 V, using Hoeffer transfer equipment. Successful transfer was ascertained by staining part of the nitrocellulose membrane with 0.01% India ink in phosphate-buffered saline plus 0.05% Tween 20. Immunoblots were blocked with 5% gelatin (Norland, New Brunswick, N.J.) in Tris-buffered saline (TBS; 20 mM Tris, 0.5 M NaCl, pH 7.4) overnight and then incubated for 10 h with rabbit or pig serum diluted 1:100 in TBS. The excess serum was removed by four 10-min washes in TBS, and the membranes were subsequently incubated for 1 h with peroxidase-conjugated protein A diluted according to the instructions of the manufacturer (Miles Scientific, Rexdale, Ontario). The membranes were finally washed in TBS, to remove the excess protein A peroxidase, and developed with 30 mg of 4-chloro-1-naphthol dissolved in 10 ml of ice-cold methanol mixed with 30 ,ul of hydrogen peroxide in 50 ml of TBS (4). Unless specified otherwise, the chemicals used in this study were purchased from Fisher Scientific (Don Mills, Ontario). Preparation of antisera. Hyperimmune rabbit serum was prepared as described previously (14, 24, 25). Hyperimmune swine sera were prepared by injecting 106 organisms in complete Freund adjuvant subcutaneously three times at 2-week intervals. Sera were also obtained from specificpathogen-free swine housed at the Arkell Swine Research Centre of the Ontario Ministry of Agriculture and Food, from conventional herds, and from animals experimentally vaccinated and infected. In some experiments, the antiserum was absorbed with
MACINNES AND ROSENDAL
1628
2 3 4 5 6 7 8
1
A
1
10 3
5 6 7 8
2 3 4
B
-
45 0
046
_
Akwb&
C
1 2
6 778
some bands, some proteins appeared to be missing or altered. Differences among serotypes were also evident in gels of OMP-enriched fractions (Fig. 1B). From three to five
21.5
major OMPs could be detected in H. pleuropneumoniae serotypes 1 to 8. All strains contained three major OMPs
3 4
D
10
3
94.0
_$
with the approximate molecular weights of 17-18K (17K), 32K, and 42-43K (42K). In underloaded gels, two or more proteins could be visualized in the 42K range. The whole-cell protein profile (data not shown) and the major OMP-enriched fractions appeared to be shared by different strains of the same serotype, but again, small
670
differences could be detected, especially in the minor proteins. Figure 1C shows eight strains of serotype 1 of H.
I*m430
pleuropneumoniae. The major 17K, 32K, and 42K OMPs
9t
*I
u4El^i";S
31.0
m 14.4 12 3 4
t
-030.
1 2 3 4 AT
_.*.
RESULTS
Analysis of Coomassie blue-stained whole-cell protein and OMP-enriched fractions. Polyacrylamide gels of Coomassie blue-stained whole-cell protein profiles of H. pleuropneumoniae serotypes 1 to 8 are shown in Fig. 1A. The protein profiles of all strains were very similar, but small differences, for example, around 27K and 35K molecular weight, could be detected. In addition to differences in the intensity of
-92.5 66.2
4l b4D4
INFECT. IMMUN.
-20.1
-1444 FIG. 1. Coomassie blue-stained SDS-PAGE profiles of wholecell proteins and OMP-enriched fractions of H. pleuropneumoniae serotypes 1 to 8. Details of the procedure are given in the text; details of the strains are given in Table 1. (A) Coomassie bluestained whole-cell proteins of strains Shope 4074 (lane 1), S1536 (lane 2), S1421 (lane 3), M62 (lane 4), K17 (lane 5), Fem0 (lane 6), WF83 (lane 7), and 405 (lane 8). (B) Coomassie blue-stained SDSPAGE profiles of OMP-enriched fractions of H. pleuropneumoniae serotypes 1 to 8. Strains Shope 4074 (lane 1), S1536 (lane 2), S1421 (lane 3), M62 (lane 4), K17 (lane 5), Fem0 (lane 6), WF83 (lane 7), and 405 (lane 8). The positions of the 17K, 32K, and 42K antigens are indicated by arrows. (C) Coomassie blue-stained SDS-PAGE profiles of OMP-enriched fractions of field and reference strains of H. pleuropneumoniae serotype 1. Strains: CM5 (lane 1), 5358 (lane 2), MA66 (lane 3), Shope 4074 (lane 4), CS177 (lane 5), WLN344 (lane 6), VSB271 (lane 7), and LL164 (lane 8). (D) Coomassie blue-stained SDS-PAGE profiles of whole-cell proteins of H. pleuropneumoniae serotypes 1 and 2 prepared from cells grown either on chocolate blood agar + NAD plates or in TYE + NAD broth: Shope 4074 plate culture (lane 1), Shope 4074 broth culture (lane 2), S1536 plate culture (lane 3), S1536 broth culture (lane 4). The molecular weights of BioRad (Fig. 1B) or Pharmacia (Fig. 1D) molecular weight markers are indicated on the right.
5
6 7 8
9 101112 13 14 15 16
2_
10 94.0
zii"
S^...
"
300
,
--
..~~~~~~~~~~~~.... .00'10 TO_ aw, _
20.1
46
0
..
A-
14.4
a X
10 3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 -16
B
--
94.0
-
67.0
43.0
--30 0 20.1
_
heterologous or homologous bacterial strains. To prepare absorbed antiserum, bacteria were suspended in phosphatebuffered saline to an optical density of 2.0 at 600 nm and broken by sonication. Equal volumes of serum and bacteria were incubated at room temperature for 2 h and then at 4°C overnight. Bacteria-antibody complexes and bacterial debris were removed by ultracentrifugation at 100,000 x g for 1 h. A second sample of sonicated bacteria was then added, and the serum was incubated and centrifuged as described above. TBS was added to the absorbed antiserum for a final dilution of 1:50.
3
14.4
FIG. 2. Coomassie blue-stained SDS-PAGE profiles of wholecell proteins (A) and OMP-enriched fractions (B) of A. equuli (lane 1), A. suis (lane 2), A. lignieresii (lane 3), A. pleuropneumoniae (lane 4); H. pleuropneumoniae strains 4074 (lane 5), S1536 (lane 6), K17 (lane 7), and 405 (lane 8), untypable strain PN33 (lane 9), "minor group" strain NM305 (lane 10), and 202 (lane 11); H. influenzae (lane 12); H. parasuis (lane 13); P. haemolytica (lane 14); E. coli (lane 15); and S. tvphimuriuin (lane 16).
MAJOR ANTIGENS OF H. PLEUROPNEUMONIAE
VOL. 55, 1987 2 3 4
1
1
5 6 7 8
A
5
2
3
4
2
3
4
6
7
6
7 8
8
B
2
3
4
5
6
7
8
1
5
c
,A
4
o
_Q
FIG. 3. Immunoblots of whole-cell proteins of H. pleuropneumoniae serotypes 1 to 8 probed with rabbit or swine sera. Strains: Shope 4074 (lane 1), S1536 (lane 2), S1421 (lane 3), M62 (lane 4), K17 (lane 5), Fem0 (lane 6), WF83 (lane 7), and 405 (lane 8). (A) Hyperimmune rabbit serum against the serotype 5 HI. pleuropneumoniae strain K17. (B) Hyperimmune rabbit serum against the serotype 6 H. pleuropneumoniae strain Fem0. (C) Hyperimmune swine serum against the serotype 2 H. pleuropneumoniae strain S1536. (D) Trivalent hyperimmune swine serum against H. parasuis strains E751, A9, and C5. The positions of 17K, 32K, and 42K OMPs are indicated by arrows.
present in all strains, but the 25K and 67K OMPs, abundant in Shope 4047 (lane 4), were reduced or absent in some of the other strains. The protein profiles could be altered by changing the conditions of growth. Figure 1D shows serotype 1 H.
were
pleuropneumoniae harvested from plates (lane 1) or from broth culture (lane 2). Differences in intensity of some bands were detected, as indicated by the arrows. Underloaded gels showed these differences more clearly. As a result of this finding, great care was taken to ensure that the samples being compared were prepared under identical conditions. To determine whether porcine Haemophilus and Actinobacillus spp. had similar protein profiles, we examined Coomassie blue-stained whole-cell protein and OMPenriched fractions of Actinobacillus and Haemophilus spp. (Fig. 2A and B). For this analysis we also included other members of the genera Actinobacillus and Haemophilus and, for reference, Escherichia coli and Salmonella typhimurium. The profiles of Actinobacillus spp. (lanes 1 to 4), H. pleuropneumoniae serotypes 1, 2, 5, and 8 (lanes 5 to 8), and
1629
Pasteurella haemolytica (lane 14) were similar. By contrast, the protein profiles of Haemophilus sp. "minor group" strains (lanes 9 to 11), Haemophilus influenzae, Haemophilus parasuis, E. coli, and S. typhimurium (lanes 12, 13, 15, and 16) were very different from those of representative strains of H. pleuropneumoniae. These differences were also seen in the protein profiles of OMP-enriched preparations (Fig. 2B). Immunoblot analysis of whole-cell protein and OMPenriched fractions. To determine whether the detected similarities and differences in protein profiles could be correlated with the presdence or absence of specific antigens, we investigated type strains of H. pleuropneumoniae serotypes 1 to 8 by immunoblotting. Rabbit sera were prepared against each serotype and used to probe immunoblots of whole-cell and OMP-enriched fractions of H. pleuropneumoniae serotypes 1 to 8 (Fig. 3A and B). Most sera reacted more strongly with the homologous strains than with heterologous strains. Figure 3A shows whole-cell proteins of H. pleuropneumoniae serotypes 1 to 8 probed with rabbit serum against the serotype 5 strain K17. In particular, antigens with approximate molecular weights of 17K, 32K, and 42K could usually be detected in both the heterologous and homologous strains. As determined by India ink staining of nitrocellulose membranes or comparison of the Coomassie bluestained gels, the positions of these antigens was coincident with the positions of the major OMPs. In addition, 14K, 28K, and 67K antigens could often be detected, but reaction to these antigens was often weaker (Fig. 3A, B, and C). Rabbit sera prepared against serotype 6 (Fem0) crossreacted with high-molecular-weight antigens of serotypes 3 and 8 (Fig. 3B, lanes 3 and 8). Similarly, antisera prepared against serotype 4 strain M62 cross-reacted with highmolecular-weight antigens of serotype 7 strain WF83 (data not shown). Proteinase K treatment eliminated the highmolecular-weight bands as well as the 17K, 32K, and 42K antigens (data not shown). Immunoblots of H. pleuropneumoniae whole-cell proteins and OMP-enriched fractions were also probed with swine sera which had been produced by subcutaneous injection. The swine sera appeared to react to the same antigens as those recognized by the rabbit sera. In most cases, components of the homologous strains reacted more strongly than those of heterologous strains, but in some cases, all serotypes were recognized equally well (Fig. 3C). Swine antisera produced against H. parasuis (Fig. 3D) or other members of the family Pasteurellaceae (data not shown) also showed a great deal of cross-reactivity with all serotypes of H. pleuropneumoniae. As with rabbit sera, the 17K, 32K, and 42K antigens were dominant antigens recognized by the swine sera. Field sera were also used to probe immunoblots of H. pleuropneumoniae. Even in overdeveloped immunoblots, little reaction could be detected when pooled sera from specific-pathogen-free pigs were used (Fig. 4A). Pooled sera from a conventional herd without evidence of I. pleuropneumoniae infection showed some reaction with all serotypes (Fig. 4B). When pooled sera from a herd with clinical pleuropneumonia was used, a strong reaction, especially with antigens in serotype 5, was seen (Fig. 4C, lane 5). Serum from a pig vaccinated with an inactivated field strain of H. pleuropneumoniae serotype 1, WLN337, also reacted strongly with all strains (Fig. 4D). In addition to the 32K and 67K antigens, high-molecular-weight antigens were also recognized in the serotype 1 lane. The reaction to the 32K antigen was especially strong with this serum, but it did not
MACINNES AND ROSENDAL
1630
INFECT. IMMUN.
recognize the 17K and 42K antigens effectively in all serotypes. Similar reactions were detected both before and after challenge. Although variations were noted in the pattern of antigens recognized by the different sera (compare Fig. 3A, 3C, and 4D), the same major components (17K, 32K, and 42K) were typically recognized by the sera tested. To determine whether porcine Haemophilus and Actinobacillus spp. share common antigens, we probed immunoblots with both rabbit and swine sera. Whole-cell and OMP-enriched extracts of the organisms shown in Fig. 2 were used for this analysis. The whole-cell protein extracts of Actinobacillus spp. (Fig. SA, lanes 1 to 4), P. haemolytica (lane 14), and H. pleuropneumoniae serotypes 1, 2, 5, and 8 (lanes 5 to 8), and, to a lesser extent, H. pleuropneumoniaerelated "minor group" (lanes 9 to 11), shared antigens detected by probing with rabbit antisera to the H. pleuropneumoniae serotype 1 strain Shope 4074. In addition to the 17K, 32K, and 42K molecules, several other major antigens could be detected. This 4074 antiserum also reacted strongly 1
2
3
4
5
6
7
8
A
1 C
1
2
3
4
5
6
7
8
5
6
7
8
B
2 3
4
5
6
7
8
1
2 3
4
D
FIG. 4. Immunoblots of whole-cell proteins of H. pleuropneumoniae serotypes 1 to 8 probed with field swine sera. Strains: Shope 4074 (lane 1), S1536 (lane 2), S1421 (lane 3), M62 (lane 4), K17 (lane 5), Fem0 (lane 6), WF83 (lane 7), and 405 (lane 8). (A) Pooled sera from a specific-pathogen-free herd. (B) Serum from a boar in a conventional herd with no history of pleuropneumonia. (C) Pooled sera from a conventional herd with a history of pleuropneumonia. (D) Serum from a pig vaccinated against serotype 1 H. pleuropneumoniae WLN337. The positions of 17K, 32K, and 42K OMPs are indicated by arrows.
with "minor group" strains. Different sets of antigens were recognized in the H. parasuis, H. influenzae, E. coli, and S. typhimurium lanes (Fig. 5A, lanes 12, 13, 15, and 16). Similar antigens were detected when OMP-enriched fractions from the same organisms were probed with antiserum to the serotype 5 strain K17 (Fig. 5B). As with the 4047 antiserum, Actinobacillus lignieresii antigens were not recognized as well as those of other Actinobacillus spp. with the K17 antiserum. Serum to H. parasuis E751 (Fig. SC) also showed a similar pattern of recognition: predominantly the 17K, 32K, and 42K antigens in H. pleuropneumoniae, Actinobacillus spp., and P. haemolytica. Both the K17 and E751 antisera recognized high-molecular-weight antigens of the homologous strains. Some, but not all, cross-reacting antibodies could be absorbed by incubating the antisera with sonicated bacteria. The 4074 antiserum (used to probe the blot shown in Fig. SA) was absorbed with E. coli. When this absorbed serum was used to probe whole-cell extracts, the 32K and 42K antigens were still recognized in all strains of Actinobacillus, Pasteurella, and Haemophilus except H. parasuis and H. influenzae (Fig. SD, lanes 1 to 14). Reaction to the 42K antigen was reduced in many strains, and, as seen with unabsorbed antisera, the reaction with A. lignieresii was weaker (Fig. SD, lane 3). No antigens were detected in E. coli or S. typhimurium lanes with the absorbed serum (Fig. SD, lanes 15 and 16). The absorbed 4074 antisera also recognized high-molecular-weight antigens in the homologous 4074 lane. The ability of the absorbed antiserum to detect the 17K antigen was greatly reduced (Fig. SD). The 4074 antiserum was also absorbed extensively with strain 4074 bacteria. The resultant serum showed limited reaction to the 42K antigen in all the Actinobacillus, Pasteurella, and Haemophilus organisms except H. influenzae (Fig. SD, lanes 1 to 14). This serum recognized an antigen of approximately 37K molecular weight in E. coli and S. typhimurium (Fig. SD, lanes 14 and 15). The 32K antigen was also recognized with the 4047 absorbed antiserum in all Actinobacillus spp. except A. lignieresii (Fig. SD, lane 3) and in Haemophilus spp. except 4074, "minor group" strain NM305, H. influenzae, and H. parasuis (lanes 5, 10, 12, and 13). The 32K antigen was clearly recognized in P. haemolytica (lane 14). Other cross-reacting antigens demonstrated with unabsorbed sera, including the 17K antigen, were no longer detected except in Actinobacillus spp., but the reactivity was greatly reduced (Fig. 5, lanes 1 to 4). When the strain 4074 antiserum was absorbed with Actinobacillus suis, reaction to the 32K and 42K antigens of the Haemophilus and Actinobacillus strains examined was greatly reduced as compared to unabsorbed sera (Fig. SF, lanes 1 to 12). The A. suis-absorbed serum still recognized the 32K antigen of P. haemolytica (Fig. 5F, lane 14). Reaction to the 17K antigen was completely abolished; however, high-molecular-weight antigens in the strain 4074 lane were clearly detected using this antiserum (Fig. SF, lane 5). Similarly, when the strain 4047 antiserum was absorbed with "minor group" strain 202 bacteria, H. parasuis and P. haemolytica 32K antigens were detected; with H. parasuisabsorbed serum, P. haemolytica and "minor group" 32K antigens were detected (data not shown). DISCUSSION In contrast to earlier reports (9), we found that the protein profiles of reference and type strains of H. pleuropneumoniae serotypes 1 to 8, although similar, were not identical
VOL. 55, 1987
MAJOR ANTIGENS OF H. PLEUROPNEUMONIAE 1 2 3 4 5
6 7 8 9 10 11 12 13 14 15 16
1 2 3 4 5 6 7
B_D4~
AIt
M,
4m
-l",
8 9 10 1112 13 14 15 16
-_ -
.-
oft.
1631
.'
_
----.
-
-.
-
.rn..Eb
-m
A-
--O--
lifirlw .. ...iIuIb -
2 3 4 5 6 7 8 9 10 11 12 13 1415 16
1
C
1
-'-I-
D
I..--
1
. -
2 3 4 5 6 7 8 9 10 11 1213 14 15 16
2 3 4
5 6 7 8 9 10 11 12 13 14 1516
E
-
-.-
,.-
1 2 3 4 5 6 7 8 910 11 12 1314 15 16
F
FIG. 5. Immunoblots of whole-cell proteins and OMP-enriched fractions of A. equuli 19392 (lane 1), A. suis 15557 (lane 2), A. lignieresii 19393 (lane 3), and A. pleuropneumoniae SW1004 (lane 4); H. pleuropneumoniae strains Shope 4074 (lane 5), S1536 (lane 6), M62 (lane 7), and 405 (lane 8), untypable strain PN33 (lane 9), "minor group" strain NM305 (lane 10), and "minor group" strain 202 (lane 11); H. influenzae (lane 12), H. parasuis (lane 13), P. haemolytica (lane 14), E. coli (lane 15), and S. typhimurium (lane 16). (A) Whole-cell proteins probed with hyperimmune swine serum prepared against serotype 1 H. pleuropneumoniae 4074 whole-cell protein. (B) OMP-enriched fractions probed with hyperimmune rabbit serum prepared against serotype 5 H. pleuropneumoniae K17. (C through F) Whole-cell proteins probed with hyperimmune swine antiserum prepared against (C) H. parasuis E751, (D) serotype 1 H. pleuropneumoniae 4074 (as for panel A) and absorbed with E. coli, (E) strain 4074 (as for panel A) and absorbed with 4074, and (F) strain 4074 (as for panel A) and absorbed with A. suis ATCC 15557.
(Fig. 1A and B). Small variations could be detected even among different strains of the same serotype (Fig. 1C). Unlike other investigators (9, 20), we found that minor changes in protein profile could be induced by altering the conditions of growth (Fig. 1D). Consistent with this finding, O'Reilly et al. have found that the synthesis of certain OMPs is influenced by the availability of NAD (T. O'Reilly, M. R. W. Brown, and D. F. Niven, Abstr. Annu. Meet. Am. Soc. Microbiol. 1987, D170, p. 100). The SDS-PAGE profiles of whole-cell protein and OMP-
enriched fractions of A. suis, Actinobacillus equuli, A. pleuropneumoniae, and the type strain of the genus Actinobacillus, A. lignieresii, were very similar to those of representative strains of H. pleuropneumoniae (Fig. 2A and B). In contrast, the SDS-PAGE profiles of H. parasuis and the type strain of the genus Haemophilus, H. influenzae, were very different from those of H. pleuropneumoniae.
Pohl and his co-workers (19) had previously demonstrated that "the Pasteurella-like organism of porcine necrotizing
1632
MACINNES AND ROSENDAL
pleuropneumoniae described by Bertschinger and Siefert in 1978 was . . . phenotypically similar to the V factorrequiring organism hitherto classified as Haemophilus pleuropneumoniae (Matthews and Pattison 1961) Shope 1964." Based on limited DNA hybridization studies, Pohl et al. proposed establishment of the genus Actinobacillus pleuropneumoniae with two biovars: an NAD-requiring biovar with strain Shope 4074 as the type strain and an NAD-independent biovar represented by the type strain Bertschinger 2008 (19). Until very recently, only NADindependent strains have been referred to as A. pleuropneumoniae, and the designation H. pleuropneumoniae has persisted in the literature. Our protein profile data support the contention of Pohl et al. that, despite their requirement for NAD, H. pleuropneumoniae strains share greater homology with members of the genus Actinobacillus than with members of the genus Haemophilus. The protein profiles of P. haemolytica also were similar to those of representative strains of H. pleuropneumoniae, again consistent with an earlier report that pleuropneumonia could be caused by NAD-independent P. haemolytica-like organisms (H. U. Bertschinger and P. Seifert, Abstr. 5th Int. Pig Vet. Soc. World Congress, Zagreb, Yugoslavia, 1978, abstr. no. M19). Additional studies are required to clarify the relationship of P. haemolytica strains with the genus Actinobacillus. The protein profiles of the "minor group" strains examined shared some similarity with H. parasuis profiles but lacked strong homology with either H. influenzae or H. pleuropneumoniae. Based on protein profile and immunoblot analysis of NM305, 202, PN33, and approximately 20 other strains, the Haemophilus sp. "minor group" isolates appear to be a very heterogeneous group. In a study of "minor group" organisms, using the complement fixation test, Rapp et al. (22) noted considerable antigenic heterogeneity among the isolates studied, but found that the organisms did not induce antibodies which cross-reacted with H. pleuropneumoniae strains 1 to 5. Swine and rabbit antisera were prepared to H. pleuropneumoniae serotypes 1 to 8 and used to probe immunoblots of whole-cell proteins and OMP-enriched fractions of various porcine Haemophilus and Actinobacillus spp. With almost all of the antisera tested, serum prepared against one serotype of H. pleuropneumoniae recognized 17K, 32K, and 42K antigens in all serotypes (for example, see Fig. 3A, 3C, 4C, SA, and 5B). A 67K antigen was also frequently detected (Fig. 4D, SA, and SB). The 17K, 32K, and 42K antigens detected by immunoblotting were coincident with the major OMPs, as judged by comparison with Coomassie bluestained gels and with India ink-stained nitrocellulose membranes. Reaction to these major antigens was abolished after proteinase K treatment. These data are consistent with the notion that the 17K, 32K, and 42K antigens are major OMPs (or molecules strongly bound to the major OMPs). The major antigens observed in this study are similar to those reported by Rapp and Ross (21). In sera from pigs infected with serotype 5 H. pleuropneumoniae, they detected antibodies to OMPs of 16.5K (our 17K), 29K/43.5K (heat modifiable) (our 32K), 38.5K (our 42K), 45K, 49.5K, and 66.5K (our 67K). In addition, they detected 54K and 95K polysaccharide antigens. The sera from serotype 5infected animals also cross-reacted with serotypes 1 and 7; however, Rapp and Ross noted that "several OMPs and lipopolysaccharide or polysaccharide determinants appeared to be type specific." Using similar, but not identical, procedures for SDS-PAGE and immunoblotting, we were unable
INFECT. IMMUN.
to detect the 54K and 94K polysaccharide antigens with any of the sera which we tested. This discrepancy may be due to minor technical variations or strain differences. In most cases, antisera recognized high-molecular-weight antigens in the homologous lanes in addition to the 17K, 32K, and 42K antigens (e.g., Fig. 3A and 5B). Proteinase K treatment abolished reaction to these antigens, suggesting that, in addition to specific polysaccharide antigens, there may be species- or serotype-specific protein antigens. With serum prepared against serotype 3, these high-molecularweight antigens could be detected in whole-cell protein preparations of serotypes 3, 6, and 8. Cross-reactivity among these strains has been reported previously (15, 16). In addition to cross-reacting with all serotypes of H. pleuropneumoniae, the H. pleuropneumoniae antisera also reacted with 17K, 32K, and 42K antigens of Actinobacillus spp. and P. haemolytica (Fig. SB, lanes 1 through 8 and 14). The reaction with A. lignieresii antigens was weaker, especially with the 32K antigen. These antisera recognized antigens in "minor group" strains, H. parasuis, H. influenzae, and, to a lesser extent, E. coli and S. typhimurium (Fig. SA, lanes 9 to 13, 15, 16). The 17K, 32K, and 42K antigens were also recognized by antisera prepared against H. parasuis, A. suis, various '6minor group" strains, and P. haemolytica. Thus, in addition to sharing comigrating proteins, A. suis, A. pleuropneumoniae, A. equuli, P. haemolytica, and A. lignieresii shared common epitopes which are likely on the major OMPs. Insofar as many of the cross-reacting strains may be present in conventional swine, these data illustrate why serological tests for H. pleuropneumoniae lack specificity when crude antigen preparations are employed (27). The data also suggest that major OMPs would be poor candidates for the development of a specific diagnostic test. From analysis of absorption experiments it was possible to determine some relationships among the major antigens. The 17K antigen was recognized by all antisera tested. Antibodies to the 17K antigen could be largely absorbed with all strains of Haemophilus (including "minor group"), Actinobacillus, and even E. coli (Fig. 5D, E, and F). The limited reaction seen after absorption was probably due to the fact that new epitopes are revealed in SDS-denatured proteins. The data suggest that the 17K antigen contains an epitope (or epitopes) common to many gram-negative organisms. Cross-reactivity within and even between bacterial families has been noted before (30; M. T. Collins, L. H. Tsai, and N. H0iby, Conference of Research Workers in Animal Disease, Chicago, Ill., 1986, abstr. no. 20). Epitopes on the 32K antigen appeared to be shared by some members of the family Pasteurellaceae. Antibodies to the 32K antigen persisted after antiserum to strain 4047 was absorbed with E. coli. Antibodies to the 32K antigens of H. pleuropneumoniae were removed, however, when antiserum was absorbed with H. pleuropneumoniae bacteria, but this antiserum still reacted strongly with the 32K antigen of Actinobacillus spp. and P. haemolytica. On the other hand, the reaction of strain 4047 antiserum absorbed with A. suis (or "minor group") to both Actinobacillus spp. and Haemophilus spp. was markedly reduced. A. suis-absorbed antisera still recognized the 32K antigen of P. haemolytica (Fig. SF, lane 13). These data suggest that there is an epitope common to some strains of Actinobacillus, Haemophilus, and Pasteurella. The persistence of cross-reactivity, however, suggests that there are, in addition, unique epitopes. Furthermore, antibodies to these unique epitopes are commonly present in hyperimmune swine sera. A. suis, A.
VOL. 55, 1987
MAJOR ANTIGENS OF H. PLEUROPNEUMONIAE
equuli, and A. pleuropneumoniae appear to share one class of such unique epitopes. H. pleuropneumoniae and "minor group" strains appear to share a second class, and P. haemolytica represents a third. The 42K antigen was recognized by virtually all of the sera tested. Antibodies to this antigen could largely be removed by absorbing the serum with Actinobacillus spp., Haemophilus spp., or P. haemolytica, but not with E. coli. The residual reaction seen to the 42K antigen may have been the result of the presence of additional minor proteins in the region of the 42K, or due to an epitope(s) on the 42K which is recognized poorly by antibodies against other antigens, or due to new epitopes in the SDS-denatured proteins. Although the cross-reactivity noted makes serodiagnosis difficult, it may be possible to take advantage of common antigens for the development of improved vaccines. Previous studies have shown that parenteral vaccination is effective only against challenge with the homologous serotype and may cause severe localized lesions (13, 14). Animals infected naturally or intranasally, however, can develop protective immunity against all serotypes (13). Our data suggest that it might be possible to prepare vaccines which could be given intranasally and afford some degree of protective immunity from comparably benign organisms such as "minor group" strains or from cross-reacting subunits of virulent organisms. This study provides a basis for understanding the problems of cross-reactivity seen in serodiagnostic tests of H. pleuropneumoniae and supplies another line of evidence for the association of H. pleuropneumoniae with the genus Actinobacillus.
8. Mittal, K. R., R. Higgins, S. Lariviere, and D. Leblanc. 1984. A 2-mercaptoethanol tube agglutination test for diagnosis of Haemophilus pleuropneumoniae infection in pigs. Am. J. Vet. Res. 45:715-719. 9. Nicolet, J., P. Paroz, and M. Krawinkler. 1980. Polyacrylamide gel electrophoresis of whole-cell proteins of porcine strains of Haemophilus. Int. J. Syst. Bacteriol. 30:69-76. 10. Nicolet, J., P. Paroz, M. Krawinkler, and A. Baumgartner. 1981. An enzyme-linked immunosorbent assay, using an EDTA-extracted antigen for the serology of Haemophilus pleuropneumoniae. Am. J. Vet. Res. 42:2139-2142. 11. Nicolet, J., and E. Scholi. 1981. Haemophilus infections, p. 368-377. In A. D. Leman, R. D. Glock, W. L. Mengeling, R. H. C. Penny, E. Scholl, and B. Straw (ed.), Diseases of swine, 5th ed. Iowa State University Press, Ames. 12. Nielsen, R. 1974. Serological and immunological studies of pleuropneumonia of swine caused by Haemophilus parahaemolyticus. Acta Vet. Scand. 15:80-89. 13. Nielsen, R. 1976. Pleuropneumonia of swine caused by Haemophilus parahaemolyticus: studies on the protection obtained by vaccination. Nord. Veterinaermed. 28:337-348. 14. Nielsen, R. 1984. Haemophilus pleuropneumoniae serotypescross protection experiments. Nord. Veterinaermed. 36:221-
ACKNOWLEDGMENTS This work was supported by a Natural Sciences and Engineering Research Council of Canada operating grant to J.M., by an Agriculture Canada operating grant to S.R., and by an Ontario Pork Producers Marketing Board grant to S.R. and J.M. We gratefully acknowledge the receipt of strains from those listed in Table 1, and we thank Sheila Edwards for her expert technical assistance.
LITERATURE CITED 1. Barenkamp, S. J., R. S. Munson, Jr., and D. M. Granoff. 1981. Subtyping isolates of Haemophilus influenzae type b by outermembrane protein profiles. J. Infect. Dis. 143:668-676. 2. Gilbride, K. A., and S. Rosendal. 1983. Evaluation of a selective medium for isolation of Haemophilus pleuropneumoniae. Can. J. Comp. Med. 47:445-450. 3. Gunnarsson, A. 1979. Evaluation of different antigens in the complement fixation test for diagnosis of Haemophilus pleuropneumoniae (parahaemolyticus) infections in swine. Am. J. Vet. Res. 40:1564-1567. 4. Hawkes, R. 1982. Identification of concanavalin A-binding proteins after sodium dodecyl sulfate-gel electrophoresis and protein blotting. Anal. Biochem. 123:143-146. 5. Higgins, R., S. Lariviere, K. R. Mittal, G. P. Martineau, P. Rousseau, and J. Cameron. 1985. Evaluation of killed vaccine against porcine pleuropneumonia due to Haemophilus pleuropneumoniae. Can. Vet. J. 26:86-89. 6. Kume, K., T. Nakai, and A. Sawata. 1984. Isolation of Haemnophilus pleuropneumoniae from the nasal cavities of healthy pigs. Jpn. J. Vet. Sci. 46:641-647. 7. Lombin, L. H., S. Rosendal, and W. R. Mitchell. 1982. Evaluation of the complement fixation test for the diagnosis of pleuropneumonia of swine caused by Haemophilus pleuropneumoniae. Can. J. Comp. Med. 46:109-114.
1633
234. 15. Nielsen, R. 1985. Haemophilus pleuropneumoniae (Actinobacillus pleuropneumoniae) serotypes 8, 3, and 6. Serological response and cross immunity in pigs. Nord. Veterinaermed. 37:217-227. 16. Nielsen, R., and P. J. O'Connor. 1984. Serological characterization of 8 Haemophilus pleuropneumoniae strains and proposal of a new serotype, serotype 8. Acta Vet. Scand. 25:1-11. 17. Nielsen, R., A. D. Thomsen, and S. D. Versterlund. 1976. Pleuropneumonia caused by Haemophilus pleuropneumoniae. An attempt to control the disease at two progeny testing stations by serological blood testing followed by removal of the seropositive animals and their litter mates. Nord. Veterinaermed. 28:349-352. 18. O'Reilly, T., and D. F. Niven. 1986. Tryptone-yeast extract broth as a culture medium for Haemophilus pleuropneumonia and Haemophilus parasuis to be used as challenge inocula. Can. J. Vet. Res. 50:441-443. 19. Pohl, S., H. U. Bertschinger, W. Frederiksen, and W. Mannheim. 1983. Transfer of Haemophilus pleuropneumoniae and the Pasteurella haemolytica-like organism causing porcine necrotic pleuropneumonia to the genus Actinobacillus (Actinobacillus pleuropneumoniae comb. nov.) on the basis of phenotypic and deoxyribonucleic acid relatedness. Int. J. Syst. Bacteriol. 33:510-514. 20. Rapp, V. J., R. S. Munson, Jr., and R. F. Ross. 1986. Outer membrane protein profiles of Haemophilus pleuropneumoniae. Infect. Immun. 52:414-420. 21. Rapp, V. J., and R. F. Ross. 1986. Antibody response to outer membrane components of Haemophilus pleuropneumoniae during infection. Infect. Immun. 54:751-760. 22. Rapp, V. J., R. F. Ross, and T. F. Young. 1985. Characterization of Haemophilus spp. isolated from healthy swine and evaluation of cross-reactivity of complement-fixing antibodies to Haemophilus pleuropneumoniae and Haemophilus taxon "minor group," J. Clin. Microbiol. 22:945-950. 23. Rodriguez, R. L., and R. C. Tait. 1983. Recombinant DNA techniques: an introduction, p. 197-199. Addison-Wesley Publishing Co., Don Mills, Ontario. 24. Rosendal, S., and D. A. Boyd. 1982. Haemophilus pleuropneumoniae serotyping. J. Clin. Microbiol. 16:840-843. 25. Rosendal, S., L. Lombin, and J. Demoor. 1981. Serotyping and detection of Haemophilus pleuropneumoniae by indirect fluorescent antibody technique. Can. J. Comp. Med. 45:271274. 26. Rosendal, S., and W. R. Mitchell. 1983. Epidemiology of Haemophilus pleuropneumoniae infection in pigs: a survey of Ontario pork producers, 1981. Can. J. Comp. Med. 47:1-5. 27. Rosendal, S., and K. R. Mittal. 1985. Serological cross-reactivity between a porcine Actinobacillus strain and Haemophilus
1634
MACINNES AND ROSENDAL
pleuropneumoniae. Can. J. Comp. Med. 49:164-170. 28. Shope, R. E. 1964. Porcine contagious pleuropneumonia. I. Experimental transmission, etiology and pathology. J. Exp. Med. 119:357-368. 29. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose
INFECT. IMMUN.
sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76:4350-4354. 30. van Alphen, L., T. Riemens, and H. C. Zenen. 1983. Antibody response against outer membrane components of Haemophilus influenzae type b strains in patients with meningitis. FEMS Microbiol. Lett. 18:189-195.
