Veterinary Microbiology, 32 (1992) 327-342 Elsevier Science Publishers B.V., Amsterdam
A characterization of monoclonal antibodies prepared against Pasteurella haemolytica serotype 1 surface antigens F.W. Austin a, R.E. Corstvet a'b, K.L. Schnorr~ and W.J. Todd a'b aDepartment of Veterinary Microbiology and Parasitology, School of Veterinary Medicine, Louisiana State University, Baton Rouge, LA, USA. bDepartment of Veterinary Science and Louisiana Agricultural Experiment Station, Louisiana State University, Baton Rouge, LA, USA. (Accepted 2 April 1992 )
ABSTRACT Austin, F.W., Corstvet, R.E., Schnorr, K.L., Todd, W.J., 1992. A characterization of monoclonal antibodies prepared against Pasteurella haemolytica serotype I surface antigens. Vet. Microbiol., 32: 327-342. The production and characterization of monoclonal antibodies against Pasteurella haemolyiica serotype 1 is described. Ten monoclonal antibodies were produced and divided, on the basis of their properties, into six different groups. One produced bacteria agglutination only of P. haemolytica serotype 1. Three antibodies bound with P. haemolytica serotypes 1, 5-8 and 12 and the antigen was identified in immunoblots as lipopolysaccharide. Two antibodies bound P. haemolytica serotypes 1, 2, 5-8 and 12 and P. multocida serotypes 1-7, 9, 12, 15 and 16, recognizing an epitope present on a 29 kDa outer membrane protein. One antibody bound all P. haemolytica and P. multocida serotypes. The antigen was a hexosamine less than 30 kDa which contained a formalin sensitive epitope. One antibody bound only to P. haemolytica serotype 1 and the antigen was identified as a 66 kDa outer membrane protein. Two antibodies bound P. haemolytica serotypes 1, 2, 5-9 and 12 and the antigen, while not identified, was localized on the outer membrane. This study identified antigens which contribute to the cross-reactions among P. haemolytica and P. multocida serotypes and the antibodies may be useful in ingestigating the pathogenesis of pneumonic pasteurellosis.
Bovine pneumonic pasteurellosis or shipping fever is a severe fibrinous pneumonia causing major economic loss to the North American feedlot cattle industry Martin et al., 1981; Yates, 1982 ) Pasteurellahaemolytica serotype 1 is the bacterial pathogen most frequently isolated from cattle with the disease Correspondence to: F.W. Austin, Diagnostic Laboratory Services, College of Veterinary Medicine, Mississippi State University, P.O. Drawer V, Mississippi State, MS 39762, USA.
© 1992 Elsevier Science Publishers B.V. All rights reserved.
F.W. AUSTIN ET AL.
(Collier, 1968 ). Other serotypes ofP. haemolytica, as well as other organisms such as P. multocida, Mycoplasma spp., have been implicated in the etiology, but less frequently (Yates, 1982 ). Successful division ofP. haemolytica into greater than 12 serotypes has been accomplished using specific rabbit antisera in indirect hemagglutination (Biberstein et al., 1960) and rapid plate agglutination reactions (Frank and Wessman, 1978). Soluble capsular polysaccharide antigens are responsible for serotype specificity in agglutination reactions of P. haemolytica (Frank and Wessman, 1978; Adlam et al., 1984). In contrast, the 16 serotypes of P. muItocida are distinguished on the basis of lipopolysaccharide O-antigens in gel diffusion precipitin tests (Heddleston et al., 1972). The serotypes of P. haemoIytica can be further subdivided into 2 biotypes, A and T, respective of their ability to ferment arabinose and trehalose (Smith, 1961; Fraser, 1981 ). Serotypes 3, 4, and 10 belong to biotype T, while the remaining serotypes do not ferment trehalose and belong to biotype A. Serological cross-reactions among the various serotypes, mostly within biotypes, have been reported using indirect hemagglutination (Biberstein, 1965), rapid plate agglutination (Frank and Wessman, 1978), enzyme linked immunosorbent assay (Burrells et al., 1983), and crossed immunoelectrophoresis (Tsai et al., 1988). The capsule and outer membrane of P. haemolytica contains polysaccharides, lipopolysaccharides, and proteins thought to contribute to the cross-reactivity. However, the specific antigens responsible for these relationships remain largely undefined. Cell surface antigens of P. haemolytica have been investigated as virulence factors (Corstvet et al., 1982; Rimsay et al., 1981 ) and immunogens (Confer et al., 1985; Yates et al., 1983 ). Resistance of calves to experimental challenge with P. haemolytica serotype 1 has been obtained with a variety of bacterial surface preparations (Yates et al., 1983; Confer et al., 1985; Matsumoto et al., 1984). Crossprotection studies between P. haemolytica serotype 1 and P. multocida capsule type A serotype 8 (Mukkur, 1977) and different serotypes of P. haemolytica (Shewen and Wilkie, 1988 ) suggest the importance of surface antigens in the pathogenesis of pneumonia pasteurellosis and in the development of disease resistance. This report describes the preparation and partial characterization of monoclonal antibodies against P. haemolytica serotype 1 whole cells and capsular material. Additionally, we sought to identify the antigens recognized by the antibodies and to localize their binding on the bacterial surface. MATERIALS AND METHODS
Antigen preparations Pasteurella haemolytica biotype A, serotype 1, originally isolated from a calf with pneumonia, was maintained by lyophilization in skim milk and pe-
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
riodic animal passage as previously described (Newman et al., 1982). Agar plates containing 10% (v/v) fresh citrated bovine blood, 1% (v/v) heat inactivated horse serum, and 1% (v/v) yeast hydrolysate were used to propagate the bacteria in candle jars at 37 ° C. Capsular material antigens were extracted from confluent 6 h cultures according to the method of Gentry et al. ( 1982 ). Decapsulated bacterial cells were removed from solubilized capsular antigens by centrifugation at 12 000 g for 30 min at 4 ° C. Capsular antigens, present in the supernate, were filter sterilized (0.22 p m ) and dialyzed (8-10 kDa molecular weight cut-off, Fisher Sci. Co. ) against four changes of sterile distilled water at 4 ° C over 48 hours. The dialyzed capsular antigens were frozen at - 7 0 ° C and lyophilized prior to mouse immunization. P. haemolytica serotypes 2 through 12 and P. multocida serotypes 1 through 16 were propagated as described above and defined according to Frank and Wessman ( 1978 ) and Heddleston et al. ( 1972 ), respectively.
Mice and immunizations One half mg of capsular material or 5 × 108 whole P. haemolytica serotype 1 cells from 6 h cultures in 0.5 ml Freund's complete adjuvant were used to immunize five, 6 to 8 week old, female BALB/c mice by intraperitoneal injection. One month later, mice were immunized again with either 0.5 mg capsular material or 5 X 108 whole bacterial cells in 0.5 ml phosphate buffered saline (PBS, 0.15 M, pH 7.2) by the same route. Blood samples were withdrawn from the retro-orbital plexus prior to immunization and three days following the second immunization for measurement of serum antibody. Successful immunization and selection of mice for monoclonal antibody production was established using individual mouse sera in an enzyme linked immunosorbent assay (ELISA) procedure described below. Single mice immunized with either capsular material or whole bacterial cell which produced a strong positive ELISA result were selected as a splenocyte donors for cell fusion.
Mouse plasmacytoma cells The murine plasmacytoma cell line SP2/0-Ag 14 was used as a fusion partner with mouse splenocytes in this study. The SP2/0-Ag 14 cells were maintained at 37 ° C in 5% CO2 using RPMI- 1640 m e d i u m (GIBCO Laboratories) supplemented with 15% (v/v) fetal calf serum (FCS) (Hy-Clone FCS, Sterile Systems), 2 m M L-glutamine (GIBCO Laboratories), 0.01 mM hypoxanthine and thymidine (Sigma Chemical Co.) and 100 U penicillin/streptomycin per ml (HT m e d i u m ) .
Cell fusion and hybridoma selection Four days following the second immunization, splenocytes from one mouse were fused with SP2/0-Ag 14 cells at a 4:1 ratio using 50% (v/v) polyethylene glycol (1450 MW, Kodak) in HT m e d i u m according to Bastin et al.
v.w. AUSTIN ET AL.
( 1982 ). After the fusion, cells were diluted in HT medium to 7.5 × 10 6 splenocytes/ml and 0.1 ml aliquots were delivered to the wells of four 96 well plates (Cell Wells, Corning Glass Works) which already contained 5 × 10 6 BALB/c thymocytes/well as feeder cells. The cultures were allowed to stabilize for 24 hours at 37°C in the presence of 5% CO> Hybrid cells were selected by the daily feeding of 0.1 ml volumes/well of HT medium containing 4 × 10- 4 mM aminopterin (Sigma Chemical Co.) for four consecutive days. Plates were visually assessed for clonal hybridoma growth on a daily basis during hybrid cell selection using an inverted microscope. Following hybrid cell selection, hybridomas were maintained and expanded in HT medium for assay of antibody production and cloning.
Assayfor antibody production and hybridoma cloning Evaluation of hybrid cell cultural supernates for the presence of antibody directed against P. haemolytica serotype 1 capsular material or whole bacterial cells was performed by ELISA. Briefly, 50 #1 volumes containing 5.0/~g capsular material or 3.8 × 10 7 whole bacterial cells in TEN buffer (0.05 M Tris, 0.001 M Na4 EDTA, 0.15 M NaC1, pH 7.4) were air dried overnight in 96 well plates (Immunolon # 1, Dynatech Laboratories, Inc. ). The desiccated antigens were fixed with 50 #l paraformaldehyde per well for 5 rain and washed 3 times by filling the wells with TEN buffer. Wells were blocked for nonspecific antibody binding with 200/~1 aliquots of 2% (w/v) bovine serum albumin (Sigma Chemical Co.) in phosphate buffered saline (PBS, 0.15 M, pH 7.2) for 2 hours and washed. Hybrid cell cultural supernates (50/A) were added and allowed to bind for 45 rain at room temperature with constant shaking. The wells were washed 3 times to remove cultural supernates and 50 /LI volumes of goat anti-mouse IgG and IgM horseradish peroxidase conjugated secondary antibody (Cappel Laboratories) at a 1 : 1500 dilution in TEN buffer was added to the wells and incubated as above. Unbound conjugated antibody was washed from the wells 3 times and antibody secreting hybridomas were identified using orthophenylenediamine (0.002 M orthophenylenediamine, 0.0035 M H202 in 0.15 M citric acid/disodium phosphate, pH 5.0). Plates were incubated for 20 min in the dark at room temperature and visually inspected for positive reactions. Seven wells from the capsular material hybridomas and three wells from the whole cell hybridomas which produced different visual degrees of ELISA reactivity were selected for cloning and further study. Following selection of those hybridomas secreting antibody specific for capsular material and whole bacterial cells, cells were cloned twice by limiting dilution according to Bastin et al. (1982). BALB/c thymocytes were used as feeder cells at a concentration of 5 × 106 cells/well. Cloned hybridomas were expanded in 25 cm 2 flasks (Corning Glass Works) in HT-medium and cypropreserved in liquid nitrogen using the same medium containing 20% (v/v)
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
FCS and 10% (v/v) dimethyl sulfoxide (Sigma Chemical Co.). Cloned hybridomas secreting antibody specific for P. haemolytica antigens were designed using an alpha-numeric system. Hybridomas developed against capsular material in the first fusion were designated FI while those hybridomas prepared against bacterial whole cells were referred to as F2. The second character in the naming system refers to the origin of the parent well in the 96well plates. Peritoneal ascitic fluid, for each of the cloned hybridomas, was produced in adult BALB/c mice conditioned one week in advance with 0.5 ml pristane (2,6,10,14-tetramethylpentadecane, Sigma Chemical Co. ). Preconditioned mice were inoculated in the peritoneal cavity with 106 viable cloned hybrid cells. Peritoneal ascitic fluids were harvested by periodic abdominocentesis beginning 7 days after inoculation. Peritoneal ascitic fluids were clarified by centrifugation at 2500 g for 5 min and removal of any residual pristane. Hybridoma cultural supernates and clarified peritoneal ascitic fluids were aliquoted and stored at - 70 ° C.
Identification of monoclonal antibody isotypes Isotypes of the ten selected monoclonal antibodies were identified using hybridoma cultural supernates in two commercially available kits based on immunodiffusion (Miles Scientific) and antibody capture ELISA (HyClone Laboratories Inc. ).
Assay of monoclonal antibody specificity Monoclonal antibody cross-reactivity with P. haemolytica serotypes 2 through 12 and P. multocida serotypes 1 through 16 was determined using the ELISA procedure described with slight modifications. Six hour cultures of the bacteria were suspended in PBS and adjusted to 1 × 109 C F U / m l using a spectrophotometer. Aliquots (50/~l/well) of a 1 : 26 dilution of the bacterial suspensions, determined optimal by checker board titration (data not shown), were used as antigens in the ELISA as described.
Determination of antigen sensitivity to formalin P. haemolytica serotype 1 capsular material was prepared and desiccated in 96 well plates as described. The desiccated antigen was treated with 50 #1 of 10% (v/v) buffered formalin (pH 7.0) for 5 min. The wells were washed 3 times and the ELISA completed as described.
Determination of rapid slide agglutination by peritoneal asciticfluids Fifty/zl of the clarified ascitic fluid from each cloned hybridoma was mixed with 50/A of the P. haemolytica serotypes in suspension ( 1 × 109 C F U / m l in PBS) on a cleaned glass slide. A 1:2 dilution series (in PBS), out to a 1:256 dilution, of each of the ascitic fluids were further evaluated for agglutination.
Identical suspensions of E. coli and Staphylococcus aureus were included as negative controls. The slides were slowly rocked at room temperature and observed for rapid bacterial agglutination.
Immunoblot identification of P. haemolytica antigens P. haemolytica serotype 1 whole cells (5 × 108 organisms) were suspended in 100/zl of 25 mM Tris solubilization buffer containing 1% sodium dodecyl sulfate and 1% 2-mercaptoethanol. The suspension was boiled for 5 min and the supernate was electrophoresed using a discontinuous system as described by Laemmli (1970) at 5 W. The polyacrylamide gel consisted of a 5% stacking gel and 12% resolving gel. P. haemolytica serotype 1 antigens separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis were transferred to a sheet of nitrocellulose paper by electroblotting at 30 volts for 16 hours in 25 mM Tris-192 mM glycine buffer containing 15% methanol according to Towbin et al. (1979). After electroblotting, the nitrocellulose paper was cut into strips and blocked using 1% gelatin and 0.05% Tween 20 in PBS for 30 min while rocking. The nitrocellulose paper strips were then reacted with peritoneal ascitic fluids ( 1 : 750 ) and bovine serum ( 1 : 100) diluted in the blocking buffer. U n b o u n d antibody was washed from the strips 3 times for 5 min using 0.05% Tween 20 in PBS. Affinity-purified goat anti-mouse horseradish peroxidase [F(ab')2] ( 1:750) and rabbit anti-bovine horseradish peroxidase ( 1:400, Cappel Laboratories) enzyme conjugates in wash solution were reacted with the nitrocellulose strips for 2 h with gentle rocking. U n b o u n d enzyme conjugates were removed by washing as above. P. haemolytica antigens bound by monoclonal antibodies and subsequent enzyme conjugates were visualized on the nitrocellulose strips using 4-chloro-1-naphtol as a substrate. Prestained molecular weight markers (Bio-Rad Laboratories) were included on the gels and transferred to the nitrocellulose for molecular weight determinations.
Ultrafiltration of P. haemolytica antigens P. haemolytica serotype 1 ( 1 × 10 ~° CFU in PBS) was inoculated into 500 ml RPMI- 1640 (GIBCO Laboratories) containing 10 m M L-glutamine and 10 mM sodium bicarbonate and incubated for 5.5 h at 37°C in 5% CO2. Bacteria were pelleted by centrifugation at 12 000 g for 30 min at 4 ° C. The supernate was collected, passed through a 0.22/zm filter, and fractionated using a PM-30 membrane fitted in an ultrafiltration cell (Amicon Corporation) at 4 ° C. The ultrafiltrate containing antigens less than 30 kDa was collected and assayed for protein according to Bradford ( 1976 ) and hexosamine described by Smith and Gilkerson (1979). Cultural ultrafiltrate antigens were identified using the monoclonal antibodies in ELISA as described.
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
Determination of protein-A binding The ability of the monoclonal antibodies to bind protein-A was determined using the ELISA procedure described but protein-A conjugated horseradish peroxidase (Sigma Chemical Co.) diluted 1:500 in blocking buffer was substituted in place of the secondary antibody.
Antigen localization on the bacterial surface Colloidal gold was generated by sodium citrate reduction of gold chloride and coated with protein A (Sigma Chemical Co.) as described by Smit and Todd (1986), except that the unbound protein A was removed by ultrafiltration with an XM 300 filter (Amicon Corp.) rather than by centrifugation (Barbour et al., 1983). The success of coupling protein A to colloidal gold was determined by reacting the probe with sepharose beads which were either uncoated or coated with rabbit IgG and observing for a positive color reaction. The location of the monoclonal antibodies bound to P. haemolytica serotype 1 surface antigens was demonstrated by incubation of ascitic fluids diluted 1:500 in suspensions of the organisms at 5.0× 107 C F U / m l in PBS. As a negative control, ascitic fluids containing a protein A binding monoclonal antibody specific for an unrelated genus of bacteria was included as part of the assay protocol. Following 1 h incubation, the unbound antibodies were removed by centrifugation and the bacteria were washed twice in PBS. Antibodies bound to surface antigens were detected by adding the protein A coated colloidal gold probe to the resuspended bacteria in equal volumes and incubating for 1 h at room temperature. The unbound probe was removed by pelleting the reacted bacteria, discarding the supernate, and washing the cells twice again in PBS. For ultrastructural observations, aliquots of the reacted bacteria were absorbed onto parlodion-coated grids, air dried, and examined by electron microscopy (Zeiss EM 10). RESULTS
Ten days after the fusions, nearly every well contained rapidly growing hybridomas. Specific antibody production against either P. haemolytica serotype 1 capsular material or whole bacterial cells was identified in greater than 70% of the wells tested by ELISA. The positive wells were variable in their degree of hybrid cell growth and visual ELISA reactivity. Seven wells generated from capsular material and three wells developed from whole bacterial cells, all containing vigorously growing hybridomas, were chosen for cloning by limiting dilution and further study. The ten monoclonal antibodies were divided into six groups according to their properties (Table 1 ). Three antigens recognized by the monoclonal antibodies were identified by immunoblotting (Fig. 1, Table 2 ). More than 40 P. haemolytica antigens were identified using immune bovine serum antibodies (Lanes F and G, Fig. 1 ) in
F.W. AUSTIN ETAL.
Properties of monoclonal antibodies prepared against 1°, kaemolytica serotype 1 capsular material and whole bacterial cells Monoclonal Antibody Group a
F~D3 F~F9, FID6 FiBio FFI Ds, FID7 FIDIj b F2B2 F2B4 F2E~o
Serotype cross-reactivity P. haemolytica
Antigen formalin sensitivity
IgG2a IgM IgG2b
All 1 1,2,5-9,12
17,8,9,12,15,16 All None None
Rapid slide agglutination ~ +
aRelated monoclonal antibodies grouped according to properties. F~ Antibodies generated from capsular material. F2 Antibodies prepared from bacterial whole cells. blsotype not determined by immunodiffusion. cPerformed using monoclonal antibodies present in clarified peritoneal ascitic fluid.
immunoblots. All three antigens bound by the monoclonal antibodies in immunoblotting were also recognized by the bovine serum antibodies. Only monoclonal antibody FtDll produced a positive result in ELISA when cultural ultrafiltrate was used as an antigen. Antigens bound by the remaining monoclonal antibodies were not present in ultrafiltrate. Protein was completely retained in the retentate and was not detectable in ultrafiltrate. Hexosamine quantitation revealed a concentration of 0.045 # M / m l . A strong positive color reaction was observed when the protein A coated colloidal gold probe was incubated with sepharose beads bound with rabbit IgG, but not with unbound beads, indicating specific IgG binding. Use of the colloidal-gold probe and monoclonal antibodies in immune electron microscopy identified two groups based on ultrastructural location of the surface antigens. One group of monoclonal antibodies represented by F IF9, F1D6, and F~B ~o bound capsular material (Fig. 2). This reaction occurred when capsular material was dissociated from the bacterial surfaces as well as when the capsular antigen remained associated with the bacteria. Monoclonal antibody binding to antigen located within capsular material appeared to aid in dissociation of capsular material from the bacterial surface. Antibody specific for an outer membrane antigen (F2B4, F2E~o; Fig. 3 ) did not appear to cause capsular material dissociation from the bacteria. A second group of monoclonal antibodies comprising F2B4 and F2E~o bound
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
i!i!iiiiiiiiiiiii iiiii:iiii~!iii~[ -
G ~ ~ii~ ~
iii!!i ~i!! iii! ~
:............... ~ Fig. 1. Immunoblot identification P. haemolyticaserotype 1 antigens recognizedby monoclonal antibodies and bovine serum antibodies. Lanes: A and B; lipopolysaccharide recognized by antibodies FIF9, FID6, and FIB~o,C; 29 kDa protein reactive with antibodies FID5 and FID7, 19; 66 kDa protein bound by antibody F2B2, E; Enzyme-conjugate control, F and G; spectrum of antigens recognized by immune bovine serum antibodies. Antigens recognized by remaining antibodies were not identifiable by immunoblotting. only to the surface o f the bacterial outer membrane (Fig. 3 ). This group did not bind within capsular material. As demonstrated in Fig. 3, the probe reacted with monoclonal antibodies bound at the outer membrane o f the decapsulated bacterium. With the capsule intact, binding of these antibodies to the outer membrane could not be detected. Specificity o f the protein A coated colloidal-gold probe for monoclonal antibodies bound to P. haemolytica serotype 1 surface antigens was controlled by failure o f the probe to bind to the outer membrane and capsular antigens in the absence o f monoclonal antibody. Specificity was further controlled by incubation with protein A binding monoclonal antibodies not specific for P.
F.W. AUSTIN ET AL.
TABLE 2 Antigen identification and surface localization using monoclonal antibodies Monoclonal antibody group"
Electron microscopy localizalion
F~ D 6,
29 kDa protein
< 30 kDa hexosamine d 66 kDa protein -
N.D, N.D. Outer membrane
FIB,~ F~ Ds, F~D: F~Dt~ F2B2 F2B4, F2Elo
•~Relaled monoclonal antibodies grouped according to properties. Ft Antibodies generated from capsular material. F2 Antibodies prepared from bacterial whole cells. hAntigen not identifiable by immunoblotting. 'Lipopolysaccharide, diffuse zone identified-< 14.3 kDa. dDetermined using ultrafiltration and chemical methods. ~Not determined.
Fig. 2. T r a n s m i s s i o n electron microscopy localization o f lipopolysaccharide within the capsule ofP. haemolytica serotype 1 using monoclonal antibodies (FtF~, FtD6, a n d F~B~o) and proteinA coaled colloidal gold.
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
Fig. 3. Labelling of decapsulated P. haemolytica serotype 1 using monoclonal antibodies (F2B4 and F2E~o) and protein-A coated colloidal gold probes. Only bacteria with the other membrane exposed were labeled.
haemolytica, followed by the colloidal-gold probe. Therefore, the capsular and outer membrane antigens ofP. haemolytica did not non-specificallybind protein A coated colloidal gold probes. DISCUSSION
Many studies have reported on the serologic cross-reactions between P.
haemolytica and P. multocida serotypes. However, the antigens responsible for these cross-reactions have not been clearly defined. The production and characterization of monoclonal antibodies against surface antigens ofP. haemolytica was one of the first advances to define these cross-reactive antigens and to supply specific probes necessary to study the pathogenesis of pneumonic pasteurellosis and construct and efficacious component vaccine. Because monoclonal antibody F1D3 did not bind protein A, its antigen could not be identified in immunoblots or localized using electron microscopy. However, several characteristics identified herein and in other studies suggests F~D3 recognizes the P. haemolytica serotype 1 specific capsular polysaccharide. P. haemolytica serotypes have been defined by agglutination reactions of specific capsular polysaccharide antigens (Frank and Wessman, 1978; Adlam et al., 1984). Monoclonal antibodies prepared against P. haemolytica serotype 1 in another study also produced IgM serotype 1 specific agglutinating antibodies (Penaredondo et al., 1988). Those antibodies reacted specifically with highly purified serotype 1 capsular polysaccharide and were further characterized as opsonins, promoting neutrophil-mediated phagocytosis and complement-mediated immune bacteriolysis. Chemical analyses of purified capsular polysaccharide from P. haemolytica serotype 1 identified a compo-
F.W. AUSTIN ET AL.
sition of mannosamine monomer (Adlam et al., 1984). In these studies, reduction, oxidation, and de O-acetylation did not alter the polymer's immunologic properties, but the-N-acetylation destroyed both erythrocyte adherence and immune precipitation. Fox et al. (1985) reported the most formalin reactive sites are primary amines and amides which are readily cross-linked. The loss of antigenicity following strong formalin treatment of capsular polysaccharide in this study is consistent with reports by Adlam et al. (1984), and Fox et al. ( 1985 ), and may represent chemical alteration of the antibody binding site. P. haemolytica serotype 1 specific agglutinating antibody may be necessary for resistance to pneumonic pasteurellosis (Shewen and Wilkie, 1988). Immunization of cattle with crude capsular extracts from P. haemolytica serotype 1 has afforded some degree of resistance to experimental challenge (Confer et al., 1985; Yates et al., 1983), however, formalinized bacterins have been shown to have a neutral or deleterious effect (Friend et al., 1977; Confer et al., 1985). The results reported herein indicate FID3 recognizes an epitope on the serotype 1 specific capsular polysaccharide of P. haemolytica which is formalin-sensitive and may account, at least in part, for the ineffectiveness of formalinized bacterins to impact disease resistance. Monoclonal antibodies produced against P. haemolytica serotype 1 lipopolysaccharide in other studies were also reported as IgG3 in isotype (Durham et al., 1988; Penaredondo et al., 1988). Durham et al. (1988) reported similar reactivity with P. haemolytica serotypes 1, 5, 6, 7, 8 and 12 using formalinized capsular material and whole bacterial cells. Our antibodies reacted with the same serotypes and the antigen was not determined to be formalin sensitive. Durham et al. ( 1988 ) detected a loss oflipopolysaccharide antigenicity following periodate treatment. The epitope bound by those antibodies was localized in the core polysaccharide-O-antigen following acid hydrolysis of lipopolysaccharide (Durham et al., 1988). Monoclonal antibodies F IFg, F tD6 and F1B~0, by virtue of their isotype, P. haemolytica serotype cross-reactivity, lack of antigen formalin sensitivity, inability to cause bacterial agglutination, and immunoblot antigen identification appear identical to those antibodies previously produced against P. haemolytica lipopolysaccharide. The antigen recognized by monoclonal antibodies F ID5 and F ID7 was identified as a 29 kDa protein which could not be localized on the bacterial surface due to a lack of protein A binding. The origin of the 29 kDa protein is most likely to be the outer membrane because the antibodies were produced from and reacted with saline extractable capsular material constituting cell surface antigens. Furthermore, similar 29 and 30 kDa proteins have been reported from the outer membrane of P. haemolytica serotype 1 (Squire et al., 1984; Donachie et al., 1984; Craven et al., 1991 ). It is interesting to note that Donachie et al. (1984) demonstrated P. haemolytica serotype 6 cross-protected mice experimentally challenged with P. haemolytica serotype 1. Our studies suggest P. haemolytica serotypes 1 and 6 share epitopes present on LPS and
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
the 29 kDa protein, which may account for the ability of cross-protection. High antibody responses against the 30 kDa protein of P. haemolytica has been significantly correlated with resistance in cattle to experimental challenge (Craven et al., 1991 ). It is unclear whether our antibodies recognize the 30 kDa protein cloned by Craven et al. ( 1991 ) or other similar proteins. Antiserum raised against the 30 kDa protein of Craven et al. ( 1991 ) reacted with all P. haemolytica serotypes including a 15 kDa protein. Our antibodies cross reacted with a limited number ofP. haemolytica serotypes and most P. multocida serotypes suggesting the antigen bound may be different from that cloned by Craven et al. ( 1991 ). It is possible that serum antibodies bind with other epitopes present on the 30 kDa protein of all P. haemolytica serotypes. Monoclonal antibody F~D~ was isotyped in the IgG2a subclass and reacted with all P. haemolytica and P. multocida serotypes. It is unknown whether this antibody reacts with other species of the genus Pasteurella or other closely related genera. Common antigens of Gram negative bacteria have been demonstrated in P. haernolytica (Tsai et al., 1988 ) and E. coli (Makela and Mayer, 1976; Mayer and Schmidt, 1979 ). However, this antibody did not react with E. coli used as a control in ELISA (data not shown). The enterobacterial common antigen has been shown to share an epitope with capsular polysaccharide in E. coli (Peters et al., 1985 ). It is unclear if similar epitope sharing occurs in the genus Pasteurella because the antigens bound by F~Dll could not be resolved by immunoblotting. Monoclonal antibodies recognizing all P. haemolytica and P. multocida serotypes have not been previously reported, perhaps due to antigen formalin sensitivity. The identification ofa P. haemolytica serotype 1 specific protein or protein epitope, by monoclonal antibody F2B2, suggests antigens other than carbohydrates may contribute to the individuality of P. haemolytica serotype 1. Serotype specific proteins or protein epitopes in the genus Pasteurella have not been previously reported. Unique proteins or protein epitopes of the serotypes in the genus Pasteurella need to be further defined because they may play a role in the specific diseases produced by the individual serotypes. Lu et al. ( 1991 ) demonstrated antibodies to outer membrane proteins, but not to LPS inhibited pulmonary proliferation of P. multocida in mice indicating immunity to surface proteins may be more important in disease resistance than immunity against LPS. To date, only two somatic proteins ofP. haemolytica have been extensively characterized (Craven et al., 1991; Nelson and Frank, 1989). Further work is required to determine the identity of the antigen recognized by monoclonal antibodies F2B4 and F2E~0. Monoclonal antibodies F2B4 and F2EIo were similar to the group of antibodies recognizing LPS in their serotype cross-reactivity, antigen formalin sensitivity and inability to cause bacterial agglutination. However, F2B4 and F2EIo bound P. haemolytica serotype 9, the antigen could not be identified in immunoblots and the antigen
F.W. A U S T I N E~ AL.
was localized entirely on the outer membrane allowing unequivocal differentiation. By process of elimination, since the antigen is not capsular polysaccharide or LPS, it is likely that F2B4 and F2EIo bind a surface outer membrane protein. Failure to identify this putative outer membrane protein in immunoblots could be attributable to antigen denaturation under reducing conditions of electrophoresis. The ability of different species of Pasteurella and serotypes of P. haemolytica to impart variable degrees of cross protection in experimental challenges (Mukkur, 1977; Shewen and Wilkie, 1988 ) may be due to common or shared antigens or epitopes among the various strains. The elicidation of the differences and similarities among the P. haemolytica serotypes allow for determination of the relative contribution of the antigens in the pathogenesis in and immunity against pneumonic pasteurellosis. The results of this study define several surface antigens which contribute to the serological similarity and differences among the numerous types of P. haemolytica and P. multocida. Moreover, specific antibody probes are supplied to further study surface antigens thought to be important in pneumonic pasteurellosis. A deeper understanding of the pathogenesis in pneumonic pasteurellosis and the rational construction of an efficacious component vaccine remain as future challenges. ACKNOWLEDGEMENTS
This report represents a portion of a dissertation submitted by the senior author (F.W.A.) to the Graduate School as partial fulfillment of the requirements for the PhD degree. Supported in part by U.S. Department of Agriculture special grant 85-CRSR-2-2675 and Louisiana Agricultural Experiment Station regional grant 903-64-3117. Special thanks are extended to D. Nobles for technical assistance and B. Lambert for manuscript preparation.
REFERENCES Adlam, C., Knights, J.M., Mugridge, A., Lindon, J.C., Baker, P.R.W., Beesley, J.E., Spacey, B., Craig, G.R. and Nagy, L.K., 1984. Purification, characterization and immunological properties of the serotype-specific capsular polysaccharide of Pasteurella haemolytica (serotype AI ) organisms. J. Gen. Microbiol., 130: 2415-2426. Barbour, A.C., Tessier, S.L. and Todd, W.J., 1983. Lyme disease spirochetes and ixodid tick spirochetes share a common surface antigenic determinant defined by a monoclonal antibody. Infect. lmmun., 41: 795-804. Bastin, J.M., Kirkley, J. and McMichael, A.J., 1982. In: A.J. Michael and J.W. Fabre (Editors), Monoclonal Antibodies in Clinical Medicine, Academic Press, pp 511. Biberstein, E.L., Gills, M. and Knight, H., 1960. Serological types of Pasteurella haemolytica. Cornell Vet., 50: 283-300. Biberstein, E.L., 1965. Cross-reactions between types of Pasteurella haemolytica. Cornell Vet., 55: 495-499.
CHARACTERIZATION OF MONOCLONAL ANTIBODIES
Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem., 72: 248-254. Burrells, C., Evens, H.B. and Daeson, A.M., 1983. Antigenic relationships between the serovars of Pasteurella haemolytica demonstrable by enzyme linked immunosorbent assay (ELISA). Vet. Microbiol., 8: 187-198. Collier, J.R., 1968. Pasteurellae in bovine respiratory disease. J. Am. Vet. Med. Assoc., 152: 824-828. Confer, A.W., Lessley, B.A., Panciera, R.J., Fulton, R.W. and Kreps, J.A., 1985. Serum antibodies to antigens derived from a saline extract of Pasteurella haemolytica: Correlation with resistance to experimental bovine pneumonic pasteurellosis. Vet. Immunol. Immunopathol., 10: 265-278. Corstvet, R.E., Gentry, M.J., Newman, P.R., Rummage, J.A. and Confer, A.W., 1982. Demonostration of age-dependent capsular material on Pasteurella haemolytica serotype 1. J. Clin. Microbiol., 16:1123-1126. Craven, R.C., Confer, A.W. and Gentry, M.J., 1991. Cloning and expression of a 30 kDa surface antigen ofPasteurella haemolytica. Vet. Microbiol., 27: 63-78. Donachie, W., Gilmour, N.J.L., Mould, D.L. and Poxton, I.R., 1984. Comparison of cell surface antigen extracts from two serotypes ofPasteurella haemolytica. J. Gen. Microbiol., 130:12091216. Durham, J.A., Antone, J.M., Cunningham, M.W. and Confer, A.W., 1988. Monoclonal antibodies to Pasteurella haernolytica serotype 1 lipopolysaccharide: Demonstration of antigenic similarities among several serotypes. J. clin. Microbiol., 26: 885-889. Faser, J., Laird, S., Gilmour and N.J.L., 1981. A new serovar (biovar T) of Pasteurella haernolytica. Res. Vet. Sci., 32: 127-128. Fox, C.H., Johnson, F.B., Whiting, J. and Roller, P.P., 1985. Formaldehyde fixation. J. Histochem, Cytochem., 33: 845-853. Frank, G.H. and Wessman, G.E., 1978. Rapid plate agglutination procedure for serotyping Pasteurella haemolytica. J. Clin. Microbiol., 7: 142-145. Friend, S.C.E., Wilkie, B.N., Thompson, R.G. and Barnuum, D.A., 1977. Bovine pneumonic pasteurellosis: experimental induction in vaccinated and nonvaccinated calves. Can. J. Comp. Med., 41: 77-83. Gentry, M., Corstvet, R.E. and Panciera, R.J., 1982. Extraction of capsular material from Pasteurella haemolytica. Am. J. Vet. Res., 43: 2070-2073. Heddleston, K.L., Gallagher, J.E. and Rebers, P.A., 1972. Fowl cholera: gel diffusion precipitin test for serotyping Pasteurella multocida from avian species. Avian Dis., 16: 925-936. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227: 680-685. Lu, Y.-S., Aguila, H.N., Lai, W.C. and Pakes, S.P., 1991. Antibodies to outer membrane proteins but not to lipopolysaccharide inhibit pulmonary proliferation ofPasteurella multocida in mice. Infect. Immun., 59: 1470-1475. Makela, P.H. and Mayer, H., 1976. Enterobacterial common antigen. Bacteriol. Rev., 40:591632. Martin, S.W., Meek, A.H., Davis, D.G., Thompson, R.G., Johnson, J.A., Lopez, A., Stephens, L., Curtis, R.A., Prescott, J.F., Rosendal, S., Savin, M., Zubaidy, A.J. and Bolton, M.R., 1981. Factors associated with morbidity and mortality in feedlot calves: The Bruce County Beef Project, year two. Can. J. Comp. Med., 45:103-112. Matsumoto, M., Schmitz, J.A., Syuto, B., Watrous, B.J. and Mattson, D.E., 1984. Immunogenicity of a soluble antigen against Pasteurella haernolytica-associated pneumonia in calves. Vet. Res. Comm., 8:117-130. Mayer, H., and Schmidt, G., 1979. Chemistry and bilogy of the enterobacterial common antigen (ECA). Curr. Top. Microbiol. Immunol., 85: 99-153.
F.W. AUSTIN ET AL.
Mukkur, T.K.S., 1977. Demonstration of cross-protection between Pasteurella multocida type A and Pasteurella haemolytica serotype 1. Infect. Immun., 18: 583-585. Nelson, S.L. and Frank, G.H., 1989. Purification and characterization of a 94-kDa Pasteurella haemolytica antigen. Vet. Microbiol., 21: 57-66. Newman, P.R., Corstvet, R.E., and Panciera, R.J., 1982. Distribution of Pasteurella haemolytica and Pasteurella multicida in the bovine lung following vaccination and challenge as an indicator of lung resistance. Am. J. Vet. Res., 43: 417-422. Penaredondo, M.V., Maheswaran, S.K., Whiteley, L.O., Ames, T.R. and Leiniger, J.R., 1988. Production and partial characterization of monoclonal antibodies to Pasteurella haemolytica A 1 capsular polysaccharide and lipopolysaccharide. Am. J. Vet. Res., 49:1962-1968. Peters, H., Juts, M., Jann, B., Jann, K., Timmis, K.N. and Biitter-Suermann, D., 1985. Monoclonal antibodies to enterobacterial common antigen and to Escherichia coli lipopolysaccharide outer core: Demonstration of an antigenic determinant shared by enterobacterial common antigen and E. coli K5 capsular polysaccharide. Infect. Immun., 50: 459-466. Rimsay, R.L., Coyle-Dennis, J.E., Lauerman, L.H. and Squire, P.G., 1981. Purification and biological characterization of endotoxin fractions from Pasteurella haemolytica. Am. J. Vet. Res., 42: 2134-2138. Shewen, P.E. and Wilkie, B.N., 1988. Vaccination of calves with leukotoxic culture supernate from Pasteurella haemolytica. Can. J. Vet. Res., 52: 30-36. Smit, J., and Todd, W.J., 1986. Colloidal gold labels for immunochemical analysis of microbes. In: Aldrich, H.C. and Todd, W.J. (Editors): Ultrastructural Techniques for Microorganisms, Plenear Publishing Corp, New York, pp 469-516. Smith, G.R., 1961. The characteristics of two types of PasteureUa haemolytica associated with different pathological conditions in sheep. J. Pathol. Bacteriol., 81:431-440. Smith, R.L. and Gilkerson, E., 1979. Quantitation of glycosaminoglycan hexosamine using 3methyl-2-benzothiazolone hydrazone hydrochloride. Anal. Biochem., 98: 478-480. Squire, P.G., Smiley, D.W. and Croskell, R.B., 1984. Identification and extraction of Pasteurella haemolytica membrane proteins. Infect. Immun., 45: 667-673. Towbin, H., Staehelin, T. and Gordon, J., 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci., USA, 76: 4350-4354. Tsai, L.-H., Collins, M.T. and Hoiby, N., 1988. Antigenic analysis of Pasteurella haemolytica serovars 1 through 15 by crossed immunoelectrophoresis. Am. J. Vet. Res., 49:213-222. Yates, W.D.G., 1982. A review of infectious bovine rhinotracheitis, shipping fever pneumonia, and viral-bacterial synergism in respiratory disease of cattle. Can. J. Comp. Med., 46: 225263. Yates, W.D.G., Stockdale, D.G.H. and Babuik, L.A., 1983. Prevention of experimental bovine pneumonic pasteurellosis with an extract of Pasteurella haemolytica. Can. J. Comp. Med., 47: 250-256.