Veterinary Microbiology, 31 (1992) 369-378 Elsevier Science Publishers B.V., Amsterdam
369
Factors affecting endotoxin release from the cell surface of avian strains of Pasteurella multocida M.D. Lee a, J.R.
G l i s s o n b a n d R.E. W o o l e y a
aDepartment of Medical Microbiology, College of Veterinary Medicine, University of Georgia, Athens, GA, USA bDepartment of Avian Medicine, College of Veterinary Medicine, University of Georgia, Athens, GA, USA (Accepted 28 November 1991 )
ABSTRACT Lee, M.D., Glisson, J.R. and Wooley, R.E., 1992. Factors affecting endotoxin release from the cell surface of avian strains ofPasteurella multocida. Vet. Microbiol., 31: 369-378. Two avian strains ofPasteurella multocida, a vaccine strain and a virulent field isolate, were investigated to determine their propensity to release endotoxin from the cell surface. Both organisms released comparable amounts of endotoxin when plasma complement proteins were present, however the virulent strain did so without the loss of viability that occurred in the vaccine strain. Blocking complement activity decreased the ability of plasma to elicit endotoxin release from the bacteria. When the cells were treated with divalent metal chelators such as trans-l,2-diaminocyclohexaneN,N,N~,NI-tetraacetic acid (CDTA), more endotoxin was released from the vaccine strain than from the virulent isolate. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of purified lipopolysaccharide (LPS) from both strains revealed virtually identical patterns. Both had patterns considered typical of rough LPS. Challenge studies in 8 week old turkeys showed that the field strain induced endotoxemia of longer duration than the vaccine strain and produced greater mortality.
INTRODUCTION
Pasteurella multocida is the causative agent of fowl cholera, a contagious, septicemic disease of domestic and wild fowl. The disease is especially a problem in the turkey industry where it frequently exhibits a peracute syndrome characterized by rapid onset of disease symptoms and high mortality rates. Liver lesions in affected turkeys appear to be induced by toxic effects but no secreted toxin, other than endotoxin, is commonly found in avian strains of the organism (Hofstad, 1984). The endotoxin ofPasteurella multocida is easily extracted by saline washes which indicates that it is loosely bound to the Correspondence to: Dr. R.E. Wooley, Department of Medical Microbiology. College of Veterinary Medicine, University of Georgia, Athens, GA 30602, USA.
370
M.D. LEE ET AL.
surface of the cells (Heddleston and Rebers, 1975; Ganfield and Rebers, 1976 ). The endotoxin consists of lipopolysaccharide (LPS) complexed with protein and is highly toxic for mice and turkeys (Heddleston and Rebers, 1975; Rhoades and Rimler, 1987). Although birds are much more resistant to the toxic effects of LPS than are mice, the effects elicited by LPS administration in chickens closely resemble the signs associated with peracute fowl cholera (Heddleston and Rebers, 1975). This LPS-protein complex is also highly immunogenic in low doses but is serotypically specific (Ganfield and Rebers, 1976; Rimler and Phillips, 1986 ). In Escherichia coli, LPS is released in response to serum (Demonty and Degraeve, 1982; Tesh et al., 1986; Tesh and Morrison, 1988). In serum resistant strains, the LPS release takes place in the absence of bacterial killing and is growth phase related (Tesh et al., 1986; Tesh and Morrison, 1988 ). Many pathogenic avian strains of Pasteurella multocida have been found to be serum resistant and this trait is believed to be virulence associated (Lee et al., 1988b). Since the endotoxin ofP. multocida is loosely bound, we investigated the ability of nonimmune turkey plasma to influence LPS release and the possible role LPS release might play in the pathogenesis of fowl cholera. MATERIALS AND METHODS
Organisms A virulent field isolate of P. multocida (strain 86-1913 ) and the low virulence Clemson University (CU) vaccine strain were used for this study. Both organisms are serotype 3,4:A and have been previously characterized (Lee et al., 1988a). Organisms were maintained on dextrose starch agar (Difco, Detroit, MI). Plasma Blood from 12-week-old turkey poults (reared in P. multocida-free houses) was obtained in heparinized syringes and immediately placed on ice. Red blood cells were removed by centrifugation at 1500 g for 10 min and the resulting plasma was pooled and frozen at - 70°C until use. Endotoxin assay The limulus amoebocyte colorimetric assay (Whittaker Bioproducts, Walkersville, MD) was used to detect the presence of LPS. The assay was performed using microtiter methods according to the manufacturer's directions. In vitro endotoxin release Cells from 200/tl of an overnight culture of CU or 86-1913 grown in brain heart infusion (BHI) broth (Difco, Detroit MI) were pelleted at 12 000 rpm
ENDOTOX1N RELE,~SE FROM THE CELL SURFACE OF P. MULTOCIDA
371
for 5 min in an Eppendorf centrifuge and resuspended in 200/~1 of BHI and incubated at 37°C for 2 h. Each sample ( 10 ~1) was inoculated into 1 ml of L broth (1% NaC1, 1% tryptone, and 0.5% yeast extract) or L b r o t h + 5 0 % plasma and incubated at 37°C. Aliquots were removed at 2-h intervals and cells removed by centrifugation at 12 000 rpm for 3 min in an Eppendorf centrifuge. The supernatants were frozen at - 7 0 ° C until assayed. Complement activity was blocked by treating the plasma with zymosan at 25 m g / m l and incubating at 0 ° C for 2 h (Lee, 1988b ). Trans- 1,2-diamino-cyclohexaneN,N,N~,N~-tetraacetic acid (CDTA) treated bacteria were used in some experiments to determine the effect of cation depletion on LPS release (Leive, 1965 ). The organisms were either grown in the presence of CDTA ( 10 m M in L broth ) for 6 h and aliquots removed as described above or exposed to 10 m M CDTA in L broth for 30 min after 4, 8, and 12 h of growth in L broth. For each 2-h interval, viable counts were performed. Results obtained were compared by a Student's t test to determine difference in means (Zar, 1974).
In vivo endotoxin release Eight-week-old turkey poults were challenged with a 1-ml suspension containing approximately 10 9 C F U P. multocida by the oculo/naso/oral route (Hofacre, 1989 ). The suspension was prepared by growing each strain in BHI at 37°C overnight. At 6, 24, and 48 hours, blood was obtained from half of the birds of each challenge group of 20 turkeys and pooled into 2 lots. The lots were cultured for the presence of viable P. multocida, and the plasma was collected and frozen at - 7 0 ° C until assayed for endotoxin. Mortality rates were also determined. Capsules The presence of capsules at each time period was determined by staining each strain by the Maneval m e t h o d and observing for capsular zones by light microscopy (Maneval, 1941; Gentry et al., 1987). Lipopolysaccharide analysis The LPS of each organism was isolated as described by Darveau and Hancock (1983). Briefly, lyophilized organisms were passed through a French press and the resulting solution was treated with DNase, RNase, and protease. The lipopolysaccharide was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and silver stained as described by Tsai and Frasch (1982). Smooth E. coli LPS, phenol extract of serotype 0111:B4 (Sigma, St. Louis, M O ) , was used as a standard.
372
M.D. LEE ET AL.
RESULTS
The CU vaccine strain was complement sensitive and displayed a decrease in C F U within the first 2 h of incubation with L broth + 50% fresh plasma (data not shown). This decrease was not seen when the organism was incubated in L broth alone. Strain 86-1913, a virulent field isolate, was more complement resistant and its C F U / m l increased by 1 log during this 2-h period. Both CU and 86-1913 were found to release endotoxin into the growth medium. Both organisms released comparable amounts of endotoxin in L broth while in exponential phase of growth (Fig. 1 ). However, when in stationary phase, CU released significantly more endotoxin than 86-1913, even though its cell density was 10 times less ( t = 3.99, d f = 14, P < .01 ). When exposed to plasma, total endotoxin amounts released by 86-1913 were comparable to that released by CU when in exponential or stationary phase (Figs. 2 and 3). But compared to the a m o u n t released in LB alone, 86-1913 released three times more endotoxin when under complement pressure (t = 10.6, d f = 14, P
369
Factors affecting endotoxin release from the cell surface of avian strains of Pasteurella multocida M.D. Lee a, J.R.
G l i s s o n b a n d R.E. W o o l e y a
aDepartment of Medical Microbiology, College of Veterinary Medicine, University of Georgia, Athens, GA, USA bDepartment of Avian Medicine, College of Veterinary Medicine, University of Georgia, Athens, GA, USA (Accepted 28 November 1991 )
ABSTRACT Lee, M.D., Glisson, J.R. and Wooley, R.E., 1992. Factors affecting endotoxin release from the cell surface of avian strains ofPasteurella multocida. Vet. Microbiol., 31: 369-378. Two avian strains ofPasteurella multocida, a vaccine strain and a virulent field isolate, were investigated to determine their propensity to release endotoxin from the cell surface. Both organisms released comparable amounts of endotoxin when plasma complement proteins were present, however the virulent strain did so without the loss of viability that occurred in the vaccine strain. Blocking complement activity decreased the ability of plasma to elicit endotoxin release from the bacteria. When the cells were treated with divalent metal chelators such as trans-l,2-diaminocyclohexaneN,N,N~,NI-tetraacetic acid (CDTA), more endotoxin was released from the vaccine strain than from the virulent isolate. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of purified lipopolysaccharide (LPS) from both strains revealed virtually identical patterns. Both had patterns considered typical of rough LPS. Challenge studies in 8 week old turkeys showed that the field strain induced endotoxemia of longer duration than the vaccine strain and produced greater mortality.
INTRODUCTION
Pasteurella multocida is the causative agent of fowl cholera, a contagious, septicemic disease of domestic and wild fowl. The disease is especially a problem in the turkey industry where it frequently exhibits a peracute syndrome characterized by rapid onset of disease symptoms and high mortality rates. Liver lesions in affected turkeys appear to be induced by toxic effects but no secreted toxin, other than endotoxin, is commonly found in avian strains of the organism (Hofstad, 1984). The endotoxin ofPasteurella multocida is easily extracted by saline washes which indicates that it is loosely bound to the Correspondence to: Dr. R.E. Wooley, Department of Medical Microbiology. College of Veterinary Medicine, University of Georgia, Athens, GA 30602, USA.
370
M.D. LEE ET AL.
surface of the cells (Heddleston and Rebers, 1975; Ganfield and Rebers, 1976 ). The endotoxin consists of lipopolysaccharide (LPS) complexed with protein and is highly toxic for mice and turkeys (Heddleston and Rebers, 1975; Rhoades and Rimler, 1987). Although birds are much more resistant to the toxic effects of LPS than are mice, the effects elicited by LPS administration in chickens closely resemble the signs associated with peracute fowl cholera (Heddleston and Rebers, 1975). This LPS-protein complex is also highly immunogenic in low doses but is serotypically specific (Ganfield and Rebers, 1976; Rimler and Phillips, 1986 ). In Escherichia coli, LPS is released in response to serum (Demonty and Degraeve, 1982; Tesh et al., 1986; Tesh and Morrison, 1988). In serum resistant strains, the LPS release takes place in the absence of bacterial killing and is growth phase related (Tesh et al., 1986; Tesh and Morrison, 1988 ). Many pathogenic avian strains of Pasteurella multocida have been found to be serum resistant and this trait is believed to be virulence associated (Lee et al., 1988b). Since the endotoxin ofP. multocida is loosely bound, we investigated the ability of nonimmune turkey plasma to influence LPS release and the possible role LPS release might play in the pathogenesis of fowl cholera. MATERIALS AND METHODS
Organisms A virulent field isolate of P. multocida (strain 86-1913 ) and the low virulence Clemson University (CU) vaccine strain were used for this study. Both organisms are serotype 3,4:A and have been previously characterized (Lee et al., 1988a). Organisms were maintained on dextrose starch agar (Difco, Detroit, MI). Plasma Blood from 12-week-old turkey poults (reared in P. multocida-free houses) was obtained in heparinized syringes and immediately placed on ice. Red blood cells were removed by centrifugation at 1500 g for 10 min and the resulting plasma was pooled and frozen at - 70°C until use. Endotoxin assay The limulus amoebocyte colorimetric assay (Whittaker Bioproducts, Walkersville, MD) was used to detect the presence of LPS. The assay was performed using microtiter methods according to the manufacturer's directions. In vitro endotoxin release Cells from 200/tl of an overnight culture of CU or 86-1913 grown in brain heart infusion (BHI) broth (Difco, Detroit MI) were pelleted at 12 000 rpm
ENDOTOX1N RELE,~SE FROM THE CELL SURFACE OF P. MULTOCIDA
371
for 5 min in an Eppendorf centrifuge and resuspended in 200/~1 of BHI and incubated at 37°C for 2 h. Each sample ( 10 ~1) was inoculated into 1 ml of L broth (1% NaC1, 1% tryptone, and 0.5% yeast extract) or L b r o t h + 5 0 % plasma and incubated at 37°C. Aliquots were removed at 2-h intervals and cells removed by centrifugation at 12 000 rpm for 3 min in an Eppendorf centrifuge. The supernatants were frozen at - 7 0 ° C until assayed. Complement activity was blocked by treating the plasma with zymosan at 25 m g / m l and incubating at 0 ° C for 2 h (Lee, 1988b ). Trans- 1,2-diamino-cyclohexaneN,N,N~,N~-tetraacetic acid (CDTA) treated bacteria were used in some experiments to determine the effect of cation depletion on LPS release (Leive, 1965 ). The organisms were either grown in the presence of CDTA ( 10 m M in L broth ) for 6 h and aliquots removed as described above or exposed to 10 m M CDTA in L broth for 30 min after 4, 8, and 12 h of growth in L broth. For each 2-h interval, viable counts were performed. Results obtained were compared by a Student's t test to determine difference in means (Zar, 1974).
In vivo endotoxin release Eight-week-old turkey poults were challenged with a 1-ml suspension containing approximately 10 9 C F U P. multocida by the oculo/naso/oral route (Hofacre, 1989 ). The suspension was prepared by growing each strain in BHI at 37°C overnight. At 6, 24, and 48 hours, blood was obtained from half of the birds of each challenge group of 20 turkeys and pooled into 2 lots. The lots were cultured for the presence of viable P. multocida, and the plasma was collected and frozen at - 7 0 ° C until assayed for endotoxin. Mortality rates were also determined. Capsules The presence of capsules at each time period was determined by staining each strain by the Maneval m e t h o d and observing for capsular zones by light microscopy (Maneval, 1941; Gentry et al., 1987). Lipopolysaccharide analysis The LPS of each organism was isolated as described by Darveau and Hancock (1983). Briefly, lyophilized organisms were passed through a French press and the resulting solution was treated with DNase, RNase, and protease. The lipopolysaccharide was subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and silver stained as described by Tsai and Frasch (1982). Smooth E. coli LPS, phenol extract of serotype 0111:B4 (Sigma, St. Louis, M O ) , was used as a standard.
372
M.D. LEE ET AL.
RESULTS
The CU vaccine strain was complement sensitive and displayed a decrease in C F U within the first 2 h of incubation with L broth + 50% fresh plasma (data not shown). This decrease was not seen when the organism was incubated in L broth alone. Strain 86-1913, a virulent field isolate, was more complement resistant and its C F U / m l increased by 1 log during this 2-h period. Both CU and 86-1913 were found to release endotoxin into the growth medium. Both organisms released comparable amounts of endotoxin in L broth while in exponential phase of growth (Fig. 1 ). However, when in stationary phase, CU released significantly more endotoxin than 86-1913, even though its cell density was 10 times less ( t = 3.99, d f = 14, P < .01 ). When exposed to plasma, total endotoxin amounts released by 86-1913 were comparable to that released by CU when in exponential or stationary phase (Figs. 2 and 3). But compared to the a m o u n t released in LB alone, 86-1913 released three times more endotoxin when under complement pressure (t = 10.6, d f = 14, P
