II I - I’nlI I I HI .. . MIL I 'rdflullwwh 1 .lewt. MW h . .1 . .- I ‘1.IJOJ.r.l101mum4Lu,.vf. I. h”!..\‘...7. II“. ”I. HI . I, .I. .1 km... .1“... . J L. 0 1"! I l I tr. W. l III-I’V'I'I “THFCIU In“. “s! \U “in. . II I {I I IIWIIJI 1.?(1. IJII‘I IDII MI . I s 'I III . . 0 V! I § L111.) I I I zIfitqlfrfI‘DI Bun“: “Ia- i '4}! .5 . II. I” Ir)\.’l!1.IIII1|, } III \fi..I I I . I I'..UA . II.) In! 1’1"“; 1%.‘V‘tlflgknw’l‘ l I II . . it}! I. int 1 II '1 II, If 1. JII J I. - - ‘ I .’ ,rl I ‘ ’1'} ‘I1| I '(l n (I .| IgIfi'ltllliluflnII I. . Ill!" .1!!! ‘4 III III! - L %{.IIIIIIIII)II0 .lfIIIIID . II- Ill. . VJ .. . tuning; I . I .I I . llll’lllll.tl'lllll,fvw’ {Ii I'll. I , .I, it I | I ‘ [til ‘lf:"lf$dll.l|u'll.(ll1 glé'fiflltu I I- . .h V] IHII n IIIIHI. I’ll. IIIIIIII l'lfII III Lugf IIrI , . .I III 'l ‘I . Ifi‘j’} if I III I I A . |- I . . I'l I \1I DI. \ ( cluvld' I I I I I I I I13? I I D II r N\ I ll! Ill n I ’ I I I\ I‘ I '3‘ (I 0 VI II .I . . ' I .“ C I . (It!) . o .I I \I‘I' OIIIII III I 'l‘riflg‘r. IIJIJI. (lid I I ' ‘ I II A U 'I’{. l l I ,I Elli - .II. . I Iuflllmll. I." III. - D I, I O . I I I ugh.“ ”I?! I I JIlltl‘Il'lrmr‘II {fl (III. . I. IIIIII III}!!! lll’lh‘an. -m - II I. I II III THESIS This is to certify that the .- thesis entitled L A ‘ GO ’\.Nx Fox/‘10} AVC <3 VCR v Haw 'M G'\; CJKIKA ON\& A \I\ UKM :5 “\WQVB Cé VQSsi—m‘wqu hmw‘sk a“ .— presented by Muck RT! an M- GHENGIHWA has been accepted towards fulfillment of the requirements for \ .\ C), \ Mdegree mm “k ”MAM Date 2 2 0-7639 mm LHBMRY Michigan mate University ——-v I l OVERDUE FINES: 25¢ per day per item RETURNIK: LIBRARY MATERIALS: Place in book return to remve charge from ctrculetton records ' p ‘ J ‘\a‘.‘ \ A COMPARATIVE STUDY OF MAMMALIAN AND AVIAN ISOLATES OF PASTEURELLA HAEMOLYTICA BY M. M. Chengappa A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Animal Sciences 1981 G- I/f/jsa ABSTRACT A COMPARATIVE STUDY OF MAMMALIAN AND AVIAN ISOLATES OF PASTEURELLA HAEMOLYTICA BY M. M. Chengappa This investigation was carried out to characterize avian isolates of P. haemolgtica according to the following characteristics: 1) by fermentation studies, 2) by growth studies, 3) by pathogenic studies, 4) by electron microscopic studies, 5) by serological studies, and 6) by guanine-cytosine ratio estimation. The fermentation studies revealed that the avian strains may constitute a distinct biotype or new species, as there were considerable fermentative variations from those of mammalian strains of P. haemolytica. The growth studies indi- cated that the avian strain had a lag phase of 1.5 hours as compared to 3 hours for the mammalian strain, and also the former multiplied loga- rithmically for 1.5 to 6 hours, whereas the latter multiplied logarith- mically from a period of 3 to 10 hours. A 2% hemoglobin preparation was found to be a nontoxic, better enhancer of virulence than 7% swine gastric mucin. The mammalian strains were found to be more virulent for mice than the avian strains, but the difference was not statistically significant (p>0.05). The pathogenicity of avian isolates to White Leghorn birds depended on the age of the bird, the size of the inoculum and the route of inoculation. Electron microscopic studies on thin sections revealed that the cell M. M. Chengappa membrane of avian cultures was tightly attached to the cell and the membrane was ca. 62 A thicker than that of mammalian strains. The counterimmunoelectr0phoresis (CIE) test was used successfully to identify all 12 serotypes of P. haemolytica. This test was found to be superior to the rapid plate agglutination and the indirect hemag- glutination tests. Fifty-seven of the avian cultures were grouped into 6 distinct capsular types by CIE and they were designated numerically 1 through 6. The guanine-cytosine ratios in the DNA samples of Pas— teurella and Actinobacillus species were estimated by 2 different methods. The data indicated that these DNAs were genetically closely related. The results of present investigation strongly suggested that the avian cultures should be retained in the genus Pasteurella with a new species name, viz., Pasteurella avihaemolytica. ACKNWLEDGEMENTS I wish to express my appreciation to Dr. G. R. Carter and Dr. Timothy 5. Chang, my major advisors, for their guidance and critical suggestions during the course of the investigation. I would like to express my gratitude to the members of my graduate committee, Drs. Theo H. Coleman, Roberg K. Ringer and Howard D. Stowe, for their suggestions in preparing the manuscript. I am indebted to Dr. Robert C. Myers, Michigan Department of Public Health, for providing me with the rabbits for the experiments. I would like to thank Ms. Barbara Rogers for her help in electron microscopy, and members of the Clinical Microbiology Laboratory at Michigan State University for providing me with avian isolates of P. haemolytica. My thanks to Dr. John R. Beck, American Cyanamid Company, Dr. Ernst In Biberstein, University of California-Davis, and Glynn R. Frank, National Animal Disease Center, for providing me with ummmalian isolates of P. haemolytica. My special thanks to Dr. G. R. Carter and the Department of Animal Sciences for the financial arrangements for my research efforts. Finally, and most importantly, I wish to thank my wife, Muthu, for her constant support and encouragement during the course of my graduate study. ii TABLE OF LIST OF TABI-IES O O O O O O O O O O O O O 0 LIST OF FIGURES O C O O O O O O O O O O O O I NTRODUCT I ON 0 O O O O O O C O O O O O O 0 “VI Ew OF LI TERATURE O O O O O O O O O O O SYS temati CS 0 O O O O O O O O I O 0 CONTENTS Cultural and Biochemical Characteristics. . . Morphologic Features . . . . Variation. . . . . . . . . . Biochemical Activity . . . . Nutrition and Cultivation. . Typing. . . . . . . . . . . . . . . History of Typing. . . . . . Biotyping. . . . . . . . . . Serotyping . . . . . . . . . Problems in Typing . . . . . Relation of Biotypes and Serotypes. Distribution and Prevalence of Types. . . . . Geographical . . . . . . . . Zoological and Anatomical. . Antigenic Nature. . . . . . . . . . Vaccination and Immunity. . . . . . MATERIALS AND METHODS. . . . . . . . . . . I. Fermentation Studies . . . . Fermentation Tests . . . . . Different Groups. II. Growth Studies . . . . . . . Cultures . . . . . . . . . . Growth Experiment. . . . . . III. Pathogenicity Studies. . . . Pathogenicity to Mice. . . . Pathogenicity to Chickens of IV. Electron Microscopic Studies Cultures . . . . . . . . . . V. Preparation of Cultures for Ultrathin Sections . . Assay of Poly-B-hydroxybutyric Acid (PHB). . . . . Serological Studies. . . . . . . . . . . . . Cultures . . . . . . . . . . . . . . . . . . . . . Typing Antisera. . . . . . . . . . . . . . . . . Indirect Hemagglutination (IHA) Procedure. . . . . iii Page vii w wmmqmmeew 19 19 20 20 20 21 21 21 24 24 24 24 26 28 28 29 29 "A \ -.;o. VI. RESULTS. . I. II. III. IV. V. VI. DISCUSSION I. II. III. IV. V. VI. CONCLUSIONS. LITERATURE CITED APPENDIX . Rapid Plate Agglutination (RPA) Procedure. Titration of Antisera. Counterimmunoelectrophoresis (CIE) . . Staining Procedure . Agglutination Absorption Test (AAT). . Estimation of Guanine-Cytosine Cultures . Fermentation Studies Growth Studies 0 Pathogenicity Studies. Pathogenicity to Mice. Pathogenicity to Chickens of Electron Microscopic Studies Serological Studies. IHA, Estimation of Guanine-Cytosine Fermentation Studies RPA and CIE Tests Growth Studies . Pathogenicity Studies. Pathogenicity to Mice. Pathogenicity to Chickens of Electron Microscopic Studies Serological Studies. IHA, RPA and CIE Tests . Estimation of Guanine-Cytosine Ratio . iv Different (GC) Ratio. Groups. Page 30 31 31 32 32 32 32 36 36 36 40 40 43 45 53 53 63 66 66 67 69 69 71 72 73 74 76 78 81 89 Table 10 11 12 13 14 15 16 LIST OF TABLES Differential characteristics of Pasteurella species and the genus Actinobacillus. . . . . . . . . . . . . . . . . . Differential characteristics of S and R variants of P. haeml yti ca 0 I O C I O O O I O I O O O O O O O O O O O O O Differentiation of Pasteurella species. . . . . . . . . . . Differential characteristics of subtypes of P. haemolytica (Smith, 1974) o o o o o o o o o o o o o o o o o o o o o o o Isolates of P. haemolytica listed by source and origin. . . Details of the cultures used in the growth experiment . . . Origin and source of the cultures of P. haemolytica used in the pathogenicity studies . . . . . . . . . . . . . . . . . Inoculation details in chickens for the pathogenicity test Of PO haemlytica O O O O O I O I O O O O O O O O O O O O 0 Schedule for fixation and epoxy embedding for electron microscopy 0 O O O O O I O O O O O 0 O O O O O O O O O O O O Source and origin of mammalian isolates of P. haemolytica used for serological typing . . . . . . . . . . . . . . . . Source and origin of the cultures used for the estimation Of GC ratio 0 O O O O O O O O O O O O O O O O O O O O O O 0 Results of the biochemical tests on 59 avian isolates of P. haemlytica. O O O O I O O O O O O O O I O O O I O O O O Viable cell counts per milliliter of culture estimated at different intervals of incubation period. . . . . . . . . . Mouse pathogenicity of P. haemolytica cultures injected with mucin and hemoglobin . . . . . . . . . . . . . . . . . Two-way analysis of variance for mouse pathogenicity test . Results of inoculation of chicks with 3 cultures of P. haemolytica duringa postinoculation period of 21 days . . . Page 10 19 20 22 25 27 28 33 37 41 42 43 46 Table 17 18 19 20 21 22 23 24 25 Isolations of P. haemolytica from lung and intestinal samples of experimentally infected and control birds. . . . The average cell size and the cell membrane thickness of Pasteurella and Actinobacillus species as determined by electron microscopy . . . . . . . . . . . . . . . . . . . . Spectrophotometric assay of poly-B-hydroxybutyric acid. . . IHA and RPA titers of rabbit antiser against homologous P. haemlytica. O O O O O O C O O O O I O O O I O I O O O O IHA, CIE and RPA reactions of rabbit antisera against heterologous and homologous P. haemolytica serotypes before and after absorption . . . . . . . . . . . . . . . . The distribution of 50 mammalian isolates of P. haemolytica among 12 serotypes. . . . . . . . . . . . . . . . . . . . . CIE reactions of rabbit antisera against heterologous and homologous P. haemolytica . . . . . . . . . . . . . . . . . The distribution of 59 avian isolates of P. haemolytica among tentatively designated 6 serotypes. . . . . . . . . . Guanine-cytosine content of Pasteurella and Actinobacillus SpeCi es 0 O I O O O O O O O O O O O O O O O O O O O O O 0 vi Page 47 48 54 55 56 61 62 62 64 Figure 10 11 LIST OF FIGURES Growth curves of P. multocida (656) and A. lignieresii in aerated cultures . . . . . . . . . . . . . . . . . . . . Growth curves of aerated cultures of P. haemolytica of bovine and avian origin . . . . . . . . . . . . . . . . . . The number of viable organisms needed to cause 50% mor- tality among mice injected with mucin and hemoglobin preparations. . . . . . . . . . . . . . . . . . . . . . . . Electron micrograph of type B P. multocida (strain 656) grown on tryptose agar at 37 C. . . . . . . . . . . . . . . Electron micrograph of type-1 P. haemolytica (strain P1148) grown on tryptose agar at 37 C . . . . . . . . . . . Electron micrograph of avian isolate of P. haemolytica (isolate P3868) grown on tryptose agar at 37 C. . . . . . . Electron micrograph of A. lignieresii grown on tryptose agar at 37 C. O O O O O O O O O O O O O ‘0 O. O I O I O O O Counterimmunoelectrophoretic patterns of 6 avian type sera against hmologous antigens and saline controls . . . . Counterimunoelectrophoretic patterns of 12 mammalian type sera against homologous antigens. . . . . . . . . . . . . . Counterimmunoelectrophoretic patterns of homologous and heterologous reactions of serotypes 9, 4 and 1, before and after absorption (Table 19) . . . . . . . . . . . . . . Graphic depiction of results of guanine-cytosine content of the DNAs of Pasteurella species, Actinobacillus species, and E C cal «i O O O O O I O O O O I O O O O O I O O O O I O 0 Appendix Blood agar plate showing the B-hemolytic colonies of P. haemOlytiCa . O O O I O O O O O O O O O O O C O O O O O O 0 Electron micrograph of type B P. multocida (strain 656) grown on tryptose agar at 37 C. . . . . . . . . . . . . . Vii Page 39 39 44 50 50 52 52 59 59 6O 65 89 9O Figure A-3 Electron micrograph of type-l P. haemolytica (strain P1148) grown on tryptose agar at 37 C . . . . . . . . . Electron micrograph of avian isolate of P. haemolytica (isolate P3868) grown on tryptose agar at 37 C. . . . . Electron micrograph of A. lignieresii grown on tryptose agar at 37 C. . . . . . . . . . . . A standard curve for poly-B-hydroxybutyric acid assayed by Law and Slepecky method. . . . . viii Page 91 92 93 94 INTRODUCTION Pasteurella haemolytica is a gramrnegative pleomorphic nonmotile rod characterized by a fermentative carbohydrate metabolism. Concern- ing oxygen requirements, P. haemolytica is aerobic and facultatively anaerobic. The temperature range for growth is 22 to 42 C with 37 C being optimal. The occurrence and significance of P. haemolytica as a potential pathogen has only received wide recognition in recent years. Although it is commonly isolated from the nasal passages of healthy animals and of animals with respiratory diseases, its role in respiratory disease is not well defined (Carter, 1967; Collier, l968a,b; Lillie, 1974). Epidemiological studies have shown that the number of calves carrying Pasteurella in their nasopharynx increases during outbreaks of shipping fever (Hoerlein et al., 1961) and in 2 weeks after transportation (Thomson et al., 1969). Under certain stress or disease conditions, the P. haemolytica can replicate rapidly and can reach the alveoli in aerosolized droplets. The healthy host will clear the bacteria rapidly from lungs, but the diseased or stressed host may develop pneumonia. Biberstein et a1. (1960) divided isolates from cattle and sheep into 12 serotypes by an indirect hemagglutination (IHA) procedure. Smith (1959 and 1961) divided the species into biotypes A and T mainly on the basis of colonial morphology and fermentation of arabinose and trehalose. Later, when the 12 serotypes were grouped according to 2 biotype, serotypes 3, 4 and 10 were biotype T and all others were biotype A. Only a limited number of complete serotyping studies have been done on large numbers of P. haemolytica isolates owing to the time and effort involved in the IRA procedure. Thus far, serotype 1 is the predominant serotype isolated from cattle; however, serotype 2 and untypable isolates are also frequently isolated (Biberstein et al., 1960; Fox et al., 1971; Frank and Wessman, 1978; Mwangota et al., 1978). Serotypes 3, 4, 6, 7, 9 and 11 have been infrequently isolated from cattle (Frank and Wessman, 1978; Mwangota et al., 1978; Wessman and Hilker, 1968). Organisms described as P. haemolytica in the literature but derived from other hosts, including horses, swine, chickens and humans have only, in exceptional cases, been typable serologically. These isolates differ in biochemical and cultural characteristics from those of cattle and sheep (Heddleston, 1975; Biberstein et al., 1960; Biberstein, 1978). Their assignment to the species is often ques- tionable. This study was carried out to (1) develop a more sensitive mouse model to evaluate the virulence of P. haemolytica, (2) develop a more simple, fast and accurate serotyping technique for P. haemo- lytica, and (3) regroup avian isolates of P. haemolytica into a new variety or possibly a new species under the same genus Pasteurella. To accomplish this, the study was carried out in the following areas: virulence of the organisms to mice and chickens, ultrastructural analysis by electron microscopy, serologic typing by using the counter- immunoelectrophoresis technique, biochemical reactions, growth studies and guanine-cytosine ratio estimation of the organisms. REVIEW OF LITERATURE Systematics Bacteria identical to those referred to as Bacillus bovisepticus Group 1 by Jones (1921) were given the name Pasteurella haemolytica by Newsom and Cross (1932). However, some workers (Lovell and Hughes, 1935; Bosworth and Lovell, 1944) were not prepared to classify B. bovisepticus in the genus Pasteurella and preferred to use the name hemolytic coccobacilli. Marsh (1932) recovered strains from mastitis in ewes and named them Pasteurella mastitidis. An earlier examination of strains of P. mastitidis from Montana revealed that the cultural and biochemical characteristics were indistinguishable from those of typical strains of P. haemolytica (Carter, 1951). By current standards (Smith, 1974; Phillips, 1974), Actinobacillus possesses beta-galactosidase and urease activity. Pasteurella species may have one or the other enzyme, rarely both. The suggestion has been made (Bohacek and Mraz, 1967) that P. haemolytica is more closely related to the genus Actinobacillus, as it is the only species of Pas- teurella capable of some growth on MacConkey agar and harboring subtypes with beta-galactosidase capability. Mréz (1969) estimated the guanine- cytosine content and suggested that P. haemolytica be excluded from the genus Pasteurella and transferred to the genus Actinobacillus. 4 The diagnostically important characteristics of the species of Pas- teurella and the genus Actinobacillus are summarized in Table 1 (Biberstein, 1978). Table 1. Differential characteristics of Pasteurella species and the genus Actinobacillus Growth on MacConkey Hemolysis Indole Urease ONPG agar Pasteurella multocida - + - - - Pasteurella pneumotropica - + + a - Pasteurella ureae a - + - — Pasteurella haemolytica + - - a + Actinobacillus spp. a - + + + aDifferent reactions possible. Cultural and Biochemical Characteristics Morphologic Features Species of the genus Pasteurella are gram-negative coccobacilli measuring 1.4 :_0.4 by 0.4 i 0.1 um. Organisms are capsulated especially in fresh isolates. Long filamentous forms are frequently seen in old laboratory cultures. In the Gram's-stained smear, P. haemolgtica can exhibit bipolar staining on initial isolation. Strains of P. haemolytica are morphologically indistinguishable from Pasteurella multocida. On blood agar at 37 C, 24-hour colonies are 0.5 to 1 mm in diameter, smooth, but often with concentric rings, entire and trans- lucent; older colonies are larger. A zone of hemolysis surrounds 5 colonies of freshly isolated strains but may be reduced or lost after a few subcultures (Smith, 1974). However, strains isolated from avian species possess a wider zone of hemolysis (Harbourne, 1962; Janetschke and Risk, 1970). Variation Biberstein et a1. (1958) observed that primary cultures of P. haemolytica from 3 lambs were made up of 2 different colonial variants, which he designated S and R on the basis of their behavior in the crystal violet "uptake" test. Some characteristics of these variants are summarized in Table 2. Wessman (1964) studied the interrelationships Table 2. Differential characteristics of S and R variants of P. haemolytica S Variant R Variant Dome shaped and grayish white Conical and faintly gray Did not take up crystal violet Took up crystal violet Adherent to agar after dye application Nonadherent Pathogenic for lambs Pathogenic for lambs Pathogenic for mice Less pathogenic for mice Agglutinable by 8 serum Agglutinable by S and R sera Abundant surface substance Deficient in surface substance of smooth and nonsmooth cells in the dissociation of P. haemolgtica grown in broth. The critical factor in the inhibition of the growth of smooth cells appeared to be limitation of oxygen in the medium. Selective 5 inhibition was not observed in the aerated cultures. Biochemical Activity Considerable variation has been recognized in the fermentative activity of strains of P. haemolytica. Smith (1961) observed that all 14 of his type A strains fermented arabinose in 7 days, while 14 type T strains did not. He also found with the same strains that the type T cultures fermented trehalose in 10 days, while the type A strains did not. Only 3 of the 28 isolates Split lactose and these belonged to type A. Biberstein et a1. (1960) reported that their capsular type 1 strains were, with several exceptions, lactose and catalase positive, while all of their type 2 strains were lactose negative and catalase positive. All of their type 3 cultures were lactose and catalase negative. These characteristics alone were not sufficient to categorize strains, in that the same combinations were found with other capsular types. Shreeve et a1. (1970) reported that the use of the fermentation reactions with arabinose and trehalose to distinguish the A and T biotypes was found to be more reliable for T than for A strains, since some strains of biotype A in all serotypes except serotype 11 failed to ferment arabinose within 14 days. Isolates of P- haemolytica from chickens and turkeys were found to be different in their fermentative activity from the strains isolated from cattle and sheep (Heddleston, 1975). Similar findings were reported by Janetschke and Risk (1970), who observed considerable biochemical variation among the isolates of avian origin. The Pasteurella species can be differentiated on the basis of the reactions presented in Table 3 (Carter, 1967, 1979; Smith, 1974). 7 Table 3. Differentiation of Pasteurella species Mac- Hemo- Conkey lysis Indol Urea Glucose Lactose Mannitol P. multocida - — + - A N (A) P. haemolgtica + B — — A (A) A P. ureae - a - + A N A P . pneumotropi ca - - + + A (A) N P. gallinarum - - - - A N N P. aerogenes + — + A N N (gaS) P. anatipestifer - - - - N N N A = fermentation; N = no fermentation; (A) = most ferment Nutrition and Cultivation Berkman (1942) studied the nutritional requirements of various species of Pasteurella and reported that the isolates of P. haemolytica grew well in a medium containing hydrolyzed gelatin, amino acids, glucose and inorganic salts. Wessman (1965) studied the growth pattern of P. haemolytica strain H44L under aerobic conditions in a medium containing acid-hydrolyzed casein, cysteine, inorganic salts, vitamins and carbon source. The best carbon source was found to be D-galactose or sucrose. From the same study, Wessman observed that the maximal growth resulted from an inoculum containing fewer than 10 cells per milliliter of medium. A year later, a chemically defined medium was developed by Wessman (1966) to cultivate P. haemolytica under aerobic conditions. 8 Wessman and Wessman (1970) studied the thiamine requirement of P. haemolytica and found that 2 of 11 isolates of P. haemolytica grew well with normal amounts of thiamine; the remaining 9 isolates could satisfy their thiamine requirement with free thiamine only if large amounts were provided. Typing History of Typing Two general approaches to typing of P. haemolytica have been taken. One, developed by G. R. Smith (1959, 1961), has resulted in the division of the species into 2 types, A and T. The letters stand for arabinose and trehalose fermentation, reSpectively, and this fermentation char- acteristic of the members of the respective types. The other approach is serological and, in its present form, was developed by Carter (1956) and Biberstein and co-workers (1960). Prior to the emergence of the present methods, attempts at subdi- viding the species have been more sporadic. Serological studies prior to 1960 by agglutination (Florent and Godbille, 1950; Montgomerie et al., 1938; Newsom and Cross, 1932; Tweed and Edington, 1930) and hemag- glutination (Carter, 1956) suggested the existence of no more than 3 types. One of these was undoubtedly the present type 1, which predominates in bovine infections. It is difficult to assign the cultures described in these older reports to one of the 2 biotypes A and T, as the data are not completely consistent with either. In all likelihood the overwhelming majority were of type A. Only 2 cultures studied by Newsom and Cross (1932), the proponents of the species name, appear at all compatible with type T criteria. 9 In 1959 and 1961, Smith reported the occurrence of 2 types of P. haemolytica in sheep, distinguishable by a number of cultural traits. Moreover, there were pathogenic and epidemiologic differences between the 2 types, which Smith designated A and T, after their reactivity in arabinose and trehalose broth, as previously noted. A serological study by Carter (1956) of 51 bovine isolates failed to disclose any type of diversity by IHA and agglutination procedures. Subsequently, an investigation employing similar methods on 98 isolates of P. haemolytica from sheep and cattle and some other species revealed the existence of 10 capsular types, some of which showed some distinct biochemical, ecologic, and pathogenic patterns (Biberstein et al., 1960). In 1962, an investigation (Biberstein and Gills, 1962) into the relationship of the A and T "biotypes" to the serotypes disclosed, in the sample studied, a consistent association between serotype and biotypes (Table 4). Subsequently, 2 additional serotypes were identi- fied (Biberstein and Gills, 1962; Biberstein and Thompson, 1966), so that at present 12 serological types are recognized. Biotyping G. R. Smith (1959, 1961) found that strains of P. haemolytica from ovine pneumonia could be separated into 2 major groups on the basis of different cultural, biochemical, ecological, and pathologic character- istics. The differences were not attributable to colonial variation, as was the case with the S and R variants of Biberstein et al. (1958). Smith's groups were designated types A and T. The major distinguishing features are summarized in Table 4. 10 Table 4. Differential characteristics of subtypes of P. haemolytica (Smith, 1974) Biotype A T Fermentation of arabinose + — trehalose - + salicin - + xylose +3 - lactose d - Susceptibility to penicillin high low Serotypes 1,2,5,6,7,8,9,ll,12 3,4,10 Principal localization in normal host nasopharynx tonsils Principal disease pneumonia of cattle septicemia of association and sheep; septicemia feeder lambs of nursing lambs a Serotype 2 negative, all others positive Biotype T colonies tend to be slightly larger, measuring up to 2 mm in diameter, and possess large, dark, brownish centers, whereas biotype A colonies have an even grayish color and sometimes a small, demarcated central thickening (Smith, 1961). Smith (1961) used the bromthymol blue medium of Bosworth and Lovell (1944) to study the fermentation reaction of A and T biotypes. Subse- quent workers have substituted bromcresol purple broth (Biberstein and Gills, 1962) or phenol red broth (Wessman and Hilker, 1968). The most useful substrates by most accounts are arabinose, trehalose, lactose and salicin. Arabinose is fermented only by type A strains, trehalose only by type T. Lactose is not fermented by type T strains, while salicin is usually attacked by type T, infrequently by type A. Mannose 11 has been reported to be more frequently fermented by T than A strains, while the opposite has been described for xylose (Fredriksen, 1973; Shreeve et al., 1970). Smith (1961) reported the difference of growth and death patterns between types A and T. He established,on the basis of testing 31 dif- ferent isolates, that growth curves of A and T strains were essen- tially indistinguishable at the initial stage of incubation, but subsequent incubations maintained a higher T strain population. Most of the data on antimicrobial sensitivity patterns between A and T types are of a qualitative nature. Smith (1961) observed that the type A strains were more sensitive to penicillin than type T strains. Olmos and Biberstein (1979) described a method to differen- tiate A and T strains of P. haemolytica using growth inhibitors. They found that the paper disks containing 0.3 09 of penicillin-G produced zones of inhibition larger than 10 mm with type A but not with type T strains. Also, basic fuchsin (0.2 ug/ml), brilliant green (0.005 ug/ml), and methylene blue (3.1 Ug/ml) in brain-heart infusion broth permitted the growth of type T strains but not type A. Serotyping Present—day serotyping of P. haemolytica is concerned entirely with soluble, presumably surface, antigens. These antigens have been identified as polysaccharide (Cameron, 1972) or lipopolysaccharide (Carter, 1967). The IHA procedure in one of its several modifications is usually employed to detect the antigens (Biberstein et al., 1960; Biberstein and Thompson, 1966; Biberstein et al., 1970). Recently, a rapid plate agglutination (RPA) test has been developed to detect the same surface antigens of P. haemolytica (Frank and Wessman, 1978). 12 Carter (1956) observed that 51 isolates of P. haemolgtica from pneumonic lungs of cattle were found to be serologically homogeneous by an IHA test. Because all the 51 isolates were found to be of one type, Carter (1956) mistakenly assumed a widespread homogeneity of pathogenic strains. Biberstein et a1. (1960) examined 98 strains of P- 13365101915108 by means of a modified IHA test. Ten types were identi- fied on the basis of differences in capsular substances and 18 isolates were not typable. Types were designated by arabic numbers 1 through 10. Subsequently, 2 additional serotypes (11 and 12) were identified by IHA test (Biberstein and Gills, 1962; Biberstein and Thompson, 1966). More recently, 2 new serotypes (13 and 14) were identified from sheep (Pegram et al., 1979), but their official recognition is yet to be realized. Frank and Wessman (1978) developed a simple RPA test to type P- haemolytica, based on surface antigens. Of the 103 isolates, 95 were of the same serotype,as determined by both RPA and IHA tests,and 5 were untypable. The remaining 3 isolates were typable with the RPA test but not with IHA. Biberstein et al. (1960) examined 98 isolates- of P. haemolytica by an agglutination procedure employing autoclaved bacteria and the same sera used for the IHA tests. Thirteen somatic groups were identified by this test and designated A, AB, AD, B, BD, C, D, E, F, G, H, I and V. The division into somatic groups on the basis of agglutination was not so clearcut as the division into types based on IHA. There was no reciprocity of reactions, suggesting the presence of haptens rather than complete antigens in the somatic portions. Muraschi et al. (1965), employing ether-extracted antigens by the method of Ribi et a1. (1959), were able to type strains of P. haemolytica with a modification of the Ouchterlony gel diffusion technique. 13 Problems in Typing_ Determination of biotypes may not always be clearcut and straight- forward. Differences in colonial morphology are subtle and are identified with greatest realiability when large colonies of the 2 biotypes occur side by side. Biberstein et a1. (1958) observed that colonial dissociation within the strain occurred at varying frequencies and obscured the differences further. Variable arabinose fermentation reactions have been reported for biotype A by Carter (1976), Cowan (1974) and Frederikson (1973). The variability of fermentation tests, according to the medium employed, was stressed by Wessman and Hilker (1968). Problems were encountered in biotyping P. haemolgtica isolated from poultry, as their fermentation reactions were different from those of cattle and sheep isolates (Heddleston, 1975). The main limitation of serotyping is that not all isolates identi- fiable bacteriologically as P. haemolytica can be assigned to one of the 12 serotypes. In most cases the problem was found to be due, not to their belonging to an unrecognized type, but to their lack of the surface antigen which forms the basis for typing (Biberstein, 1978). Although cross reactions have been reported between serotypes, in most cases it was possible to eliminate the problem by using the type sera at higher dilution (Biberstein, 1965). Heddleston (1975) reported that none of the 20 isolates from poultry reacted with sera prepared against 12 types of P. haemolytica. Similar problems have been encountered in the past with avian and equine isolates of P. haemolytica (Biberstein et al., 1960). Although Frank (1980) was able to identify 3 serotypes among untypable bovine isolates by the RPA test, the problem remains to be solved with avian and equine isolates. l4 Relation of Biotypes and Serotypes After the description of the A and T biotypes was published, a study was instituted to determine the distribution of these biotypes among representative strains of each of the 11 serotypes known at the time. The investigation by Biberstein and Gills (1962) revealed that all strains of types 1, 2, 5 through 9 and 11 were of type A, while all strains of types 3, 4 and 10 were of type T (Table 4). When type 12 was identified, it was found to belong to biotype A (Table 4) (BibersteinanxiThompson, 1966). These relations between serotypes and biotypes were fully confirmed by a later investigation by Shreeve et a1. (1970). Mwangota et a1. (1978) reported the occurrence of both A and T types within each of the serotypes 3, 4, 6, 10 and 12, which conflicts directly with the previous experience. The conflicting results of Mwangota were believed to be due to high frequency of the unexpected biotypes of P. haemolgtica (Biberstein, 1978). The 2 new serotypes (l3 and 14) identified in Ethiopia were recognized as biotype A (Pegram et al., 1979). Distribution and Prevalence of Types Geographical Both biotypes and all 12 serotypes have been identified wherever P. haemolytica has been studied extensively, viz.,Great Britain (Biberstein and Thompson, 1966; Thompson et al., 1977), Kenya (Mwangota et al., 1978), Ethiopia (Pegram et al., 1979) and the United States (Carter, 1956; Biberstein et al., 1960; Wessman and Hilker, 1968). In the Republic of South Africa (Cameron, 1972), 50 isolates from pneumonic sheep were found to represent 10 serotypes, all except types 15 3 and 11. Pasteurella haemolytica has been isolated from pneumonic lungs of Australian sheep kept in quarantine at the port of San Diego. The serological identity of these isolates was determined by standard IHA test (Biberstein, 1978). Zoological and Anatomical Smith (1959), in his original description of the A and T types, emphasized that all of his T types were derived from cases of septi- cemia in feeder lambs (3-12 months old), while all the A types were obtained from pneumonic sheep. Subsequently, Smith (1961) reported septicemia in infant lambs (less than 3 months old) to be associated with type A. Biberstein and Thompson (1966) suggested that type T strains, being (1) less frequent on the whole, (2) extremely rare in the normal nasopharynx, (3) usually associated with clinical infections, and (4) alone capable of causing septicemic pasteurellosis in older lambs, were of greater pathogenic potential and lesser adaptation to a commensal existence. Type A has been found in a high proportion of cases of respiratory disease, including shipping fever (Wessman and Hilker, 1968), and found frequently as part of the apparently normal nasopharyngeal popu- lation (Magwood et al., 1969). Tonsillar infection by P. haemolytica in 50 clinically normal adult sheep revealed a 3:1 preponderance of T over A strains in contrast to the usual biotype distribution in the nasopharynx of the same animal (Gilmour et al., 1974). Within biotype A, it is serotype 1 which is most often encountered and apparently the only one associated with epidemic respiratory disease (Wessman and Hilker, 1968). Of the remaining types, 7 and 11 have been identified in cattle, while all of the 3 serotypes associated 16 with biotype T (3, 4 and 10) have been infrequently reported in that host (McDonald, 1974). The relationship of serologic type to animal origin and disease, as compiled by Smith (1974), is summarized in Table 4. The report of Mwangota et al. (1978) revealed that the occurrence of P. haemolytica in goats is much like that in sheep, with all sero- types present but type 11 clearly in.the lead. Hemolytic, fermentative, oxidase-positive, gram-negative cocco- bacilli have been isolated from swine (Biberstein et al., 1960), horses (Guerrero et al., 1973), and poultry (Harbourne, 1962; Janetschke and Risk, 1970). Although these isolates were identifiable bacteriologically as P. haemolytica, the serologic identity remained unknown. Antigenic Nature The antigenic nature of P. haemolytica has received little atten- tion. Kress et al. (1964) extracted endotoxin from a bovine strain. The product was evaluated for its dermotoxic effect in rabbits and its hemodynamic effect in sheep but was not studied serologically or immuno- logically. In its general pharmacologic effects, it resembled the endotoxins of other gram-negative bacteria. Adamou et al. (1972) characterized the proteins of P. haemolytica and P. multocida by vertical polyacrylamide gel disc electrophoresis. They observed that the dif- ferentiation of cultures was possible on the basis of different combina- tions of distinct bands of proteins with an acid isoelectric point. There was a distinct conmon protein band in all the cultures of P. haemolytica and P. multocida studied by this method. Three years later, Thompson and Mould (1975) studied the protein electrophoretic patterns of 12 serotypes of P. haemolytica employing polyacrylamide gel 17 electrophoresis. Although they were able to separate the biotypes A and T, many serotypes showed only a minor or no difference in their protein electrophoretic mobility patterns. Vaccination and Immunity Bacterins containing heavy concentrations of P. multocida and P. haemolytica were employed for the prevention of shipping fever in cattle with limited success (Carter, 1970). The work of Carter (1956) and Biberstein et al. (1960) indicated that type 1 strains should be included for the prevention of P. haemolytica infections in cattle. Hamdy and Trapp (1964) showed, in vaccination and challenge experi- ments, that immunity was obtained as a result of vaccination with a preparation containing parainfluenza-B virus and strains of P. multocida and P. haemolytica. In subsequent experiments, Hamdy and associates (1965) found that the same preparations gave no protection against shipping fever in the field. Biberstein and Thompson (1965) demonstrated the importance of serotype in determining the specificity of protection of mice by using bacterins. Subsequently, it was shown that bacterins made from cultures having no demonstrable serological relationship to the eventual infect- ing strain could protect mice effectively against this strain; better, in fact, than homologous bacterins (Knight et al., 1969). Studies have been conducted on antibody response to P. haemolytica antigens. Calves in normal herds were found to have low antibody titers to several serotypes of P. haemolytica by the IHA procedure (Wray and Thompson, 1973). Intravenous and subcutaneous injections of live or killedfh.haemolytica resulted in a nonsignificant serum anti- body rise and no nasal washing antibody response (Duncan and Thomson, 18 1970a). From this same study, Duncan and Thomson found that the aerosol exposures with live P. haemolytica caused both serum antibody and nasal washing antibody responses, while aerosol exposure with heat-killed P. haemolytica resulted in lower serum antibody titers and no nasal washing antibody response. Some calves with nasal washing antibody titers to P. haemolytica serotype 1 were found to harbor serotype 1 in their nasal passages (Duncan and Thomson, 1970b). Several basic problems make testing the prophylactic value of P. haemolytica vaccines difficult. Pasteurella pneumonia cannot be produced under experimental conditions with reproducible regularity. Attempts have been made over the years to develop an effective immunogen for the prevention of P. haemolytica infection in cattle and sheep, but the basic problems remain unsolved (Cameron, 1972; Matsuoka et al., 1972; Wilkie and Norris, 1976; Lopez et al., 1976). Recently, Wilkie and associates (1980) observed the adverse response to challenge exposure with P. haemolytica in subcutaneously vaccinated calves. Although the evidence was not clear for the adverse response, based on in vitro study it was found to be due to increased bacterial-induced macrophage cytotoxicity. MATERIALS AND METHODS I. Fermentation Studies Cultures These included 49 isolates from chickens, 9 from turkeys, and 1 from a parakeet. The isolates were identified as P. haemolytica by the characteristics as described by Carter (1979) and Smith (1974). The source and origin of the isolates are listed in Table 5. Table 5. Isolates of P. haemolytica listed by source and origin Number of isolates Origin Source 2 Chicken lung National Animal Disease Center, Ames, IA . " a Ub . 21 Chicken lung VDL , MS , East LanSing, MI 26 Chicken intestine VDL, MSU, East Lansing, MI 9 Turkey lung Animal Sciences Department, MSU, East Lansing, MI 1 Parakeet lung VDL, MSU, East Lansing, MI aVDL = Veterinary Diagnostic Laboratory b . . . . MSU = Michigan State Univer51ty l9 20 Fermentation Tests Each culture was inoculated into tubes containing 1% carbohydrate substrate and 0.15% agar in phenol red broth base (Difco). The sub- strates used in the study were glucose, lactose, sucrose, arabinose, trehalose, salicin, xylose, mannitol, and sorbitol. Each isolate was also inoculated into a medium containing ornithine to detect the activity of the enzyme ornithine decarboxylase. II. Growth Studies Cultures Source and origin of the cultures used in the studies are provided in Table 6. Table 6. Details of the cultures used in the growth experiment Culture Origin Serotype Source , a P. multocnda (656) Bovine 8:2 G. R. Carter, MSU , East Lansing, MI P. haemolytica (P1148) Bovine l G. R. Carter, MSU, East Lansing, MI , b P. haemolytica (P3868) Chicken UT National Animal Disease Center, Ames, IA Actinobacillus lignieresii Bovine --- Department of Microbiology, MSU, East Lansing, MI a . . . . MSU = Michigan State UniverSity bUT = untypable by IHA 21 Growtthxperiment The experiment was performed according to the method described by Wessman (1965) with minor modifications. The seed cultures were grown in brain heart infusion broth (Difco) for 16 hours and 0.5 ml of each of these cultures was transferred into 250 m1 Erlenmeyer flasks contain- ing 25 ml of brain heart infusion broth. Flasks were incubated at 37 C in a shaker water bath (Gyrotory Waterbath Shaker, Model G76). The number of viable cells was estimated by spreading 0.1 m1 of diluted (10-7 to 10-9) cultures into petri dishes containing tryptose agar (Difco). Following incubation at 37 C for 18 hours, the colonies were counted in a colony counter. Viable counts were made at 2-hour intervals up to 16 hours of incubation, at which time the experiment was concluded. Turbidimetric measurements of growth were made by reading Optical densities on a Spectronic-ZO colorimeter (Bausch and Lomb) at 575 nm wavelength. Readings were made on cultures diluted 1:5 in distilled water with diluted medium as the blank. III. Pathogenicity Studies Pathogenicity to Mice Cultures. Two bovine cultures and 6 avian cultures were injected into mice along with a mucin or hemoglobin preparation. The prepara- tion of the mucin and hemoglobin suspensions is given below. Details of the cultures that were used in the studies are listed in Table 7. Preparation of mucin. Seven grams of swine gastric mucin (Sigma) powder were blended vigorously for 2 to 3 minutes with 93.0 ml of distilled water. Mucin powder was added slowly to the blender to 22 Table 7. Origin and source of the cultures of P. haemolytica used in the pathogenicity studies Culture Origin Serotype Source J-28 Bovine lung 2 National Animal Disease Center, Ames,IA . b P-ll48 Bov1ne lung 1 G. R. Carter, MSU, East Lansing, MI P3868 and P3873 Chicken lung UTa National Animal Disease Center, (2 cultures) Ames, IA 0 . A113 and A355 Chicken lung UT VDL, MSU, East LanSing, MI (2 cultures) A289 and A283 Chicken UT VDL, MSU, East Lansing, MI (2 cultures) aUT = untypable by IHA E Michigan State University <0 E3 II Veterinary Diagnostic Laboratory facilitate uniform mixing. The mucin mixture was placed in a heated "Magnastir" and stirred continuously until the temperature reached ca. 100 C. Finally, the mixture was autoclaved for 20 minutes at 121 C, after which the pH was adjusted to 7.2 i 0.2 with the addition of l N sodium hydroxide. Preparation of hemoglobin. One hundred milliliters of bovine blood were collected in a clean bottle and allowed to clot. After pouring the serum out, the clot was washed twice with 100 m1 of normal saline. One hundred milliliters of distilled water were then added to the clot and the clot was broken up with a glass rod. After 5 minutes, the fluid portion was poured from the bottle and filtered through 23 0.45 p (Nalge) membrane filters. The preparation was stored in the refrigerator until used. The concentration of hemoglobin in the prepa- ration was measured in a hemoglobinometer (Coulter Electronics, Inc.). Inoculation of mice. Cultures were grown in tryptose broth (Difco) for 16 hours at 37 C and 10-fold serial dilutions of the cultures were made in sterile normal saline. Dilutions ranging from 10.7 through 10.9 were used only for viable cell counts, whereas the rest of the dilutions (10.1 through 10-6) were used for the determina- tion of LDSO in mice. Viable cells were estimated by spreading 0.1 ml of diluted cultures onto the tryptose agar plates as described previously. Adult, male, Swiss Webster albino mice, 6 in each group, were selected and caged separately. Undiluted and diluted cultures (0.25 ml of each) were injected intraperitoneally to each group of mice along with 0.25 ml of mucin or hemoglobin preparation. The culture and hemoglobin preparation were drawn separately into a syringe, mixed and inoculated immediately to avoid the deleterious effect of the hemo- globin preparation on the bacteria. Two more groups of 6 mice were injected intraperitoneally with mucin and hemoglobin preparations in 0.25 ml quantities to serve as toxicity controls. Culture control mice received 0.25 ml of each of the undiluted cultures intraperitoneally. Mice were observed daily for 14 days following inoculation. The intra- peritoneal LD for each strain of P. haemolytica was calculated by 50 the method of Reed and Muench (1938). 24 Pathogenicity_to Chickens of Different Age Groups Cultures. Three isolates of P. haemolytica of avian origin were used in this study. Isolates P3868 and A113 were isolated from chicken lungs, whereas the isolate A283 was isolated frOm chicken intestine. Inoculation of chickens. Twelve-hour-old tryptose broth cultures with a viable cell count of 109 cells per milliliter were inoculated into White Leghorn males of 3 age groups. The route of inoculation, dose of inoculum, and the number of chickens per group are listed in Table 8. Chickens were observed daily for 3 weeks, at which stage the experiment was concluded. Dead birds were necropsied and lung, liver, heart blood and intestines were collected for microbiological evaluation. Those chickens that survived for 3 weeks were sacrificed by cervical dislocation and necrOpsied. Lung and intestine were collected asepti- cally for microbiological evaluation. IV. Electron Microscopic Studies Cultures The cultures that were used for growth study experiments were used for electron microscopic studies. Details of the cultures are listed in Table 6. Preparation of Cultures for Ultrathin Sections Cultures were grown on tryptose agar plates for 12 hours at 37 C. Ten to fifteen milligrams of wet bacterial mass were washed twice in normal saline and finally suspended in 5 ml of glutaraldehyde phosphate buffer (1 m1 of 25% glutaraldehyde + 9 ml of phosphate buffer) of pH 7.2. Following overnight incubation at 25 C, the buffer was removed by 25 Table 8. Inoculation details in chickens for the pathogenicity test of P. haemolytica Route of Dose of Inoculum in m1 a a Isolate inoculation Day-old 3 wk. 6 wk. No. of chickens Control 0.75 30 3O P3868 intramuscular 0.25 0.5 wing web 0.1 0.2 0.3 30 cloacal 0.25 0.5 0.75 30 oral 0.25 0.5 0.75 30 A283 intramuscular 0.25 0.5 0.75 30 30 wing web 0.1 0.2 0.3 30 cloacal 0.25 0.5 0.75 30 oral 0.25 0.5 0.75 30 A113 intramuscular 0.25 0.5 0.75 30 30 wing web 0.1 0.2 0.3 30 cloacal 0.25 0.5 0.75 30 oral ' 0.25 0.5 0.75 30 a . . Ten chickens in each age group 26 centrifugation at 10,000 x g for 5 minutes. The culture pellet was suspended in 2 ml of 1% molten Noble agar (Difco) and poured onto a 3 x 1 inch glass slide. Agar was cut in small pieces and suspended in Zetterqvist's osmium. Details of fixation and epoxy embedding pro- cedures are outlined in Table 9. Ultrathin sections were cut with a diamond knife on an LKB Ultra- tome and double stained with saturated uranyl acetate (Watson, 1958) and lead citrate (Reynolds, 1963). The grids were examined in a Zeiss-9 electron microscope at 60 KV. Assay of Poly-B-hydroxybutyric Acid (PHB) The assay of PHB was performed according to the method of Law and Slepecky (1961) using Escherichia coli and Azatobacter vinelandii as negative and positive controls, respectively. The details of the test cultures are listed in Table 6. The cells were grown in 10 m1 of tryptose broth for 10 hours and were centrifuged in polypropylene centrifuge tubes (Ivan Sorvall, Inc.) which had been previously washed thoroughly with ethanol and hot chloroform to remove plasticizers. The cell paste was resuspended in a volume of commercial sodium hypochlorite solution (Clorox) equal to the original volume of medium. After 1 hour at 37 C, the lipid granules were centrifuged, washed with water, and then washed with acetone and alcohol. Finally, the polymer was dissolved by extraction with 3 small portions of boiling chloroform, the chloro- form solution was filtered, and the filtrate was used for poly-B- hydroxybutyrate assay. For the standard, samples containing 5 to 50 ug of polymer in chloroform were transferred to clean test tubes. Ten milliliters of concentrate sulfuric acid was added to all the tubes containing known and unknown quantities of polymer. The tubes were 27 Table 9. Schedule for fixation and epoxy embedding for electron microscopy Preparation Time Period Zetterqvist's osmium 2 hr Zetterqvist's washing solution 10 min Alcohol 50% 2 5 min changes 75% 2 5 min changes 95% 2 5 min changes 100% 15 min 100% 30 min Alcohol:propylene oxide (2:1) 15 min Alcohol:propylene oxide (1:1) 15 min Alcohol:propylene oxide (1:2) 15 min Propylene oxide 100% 15 min 100% 30 min Propylene oxide:EAMa (2:1) 1 hr Propylene oxide:EAM (1:1) 1 hr Propylene oxide:EAM (1:2) 1 hr EAM overnight Tissues with EAM in blocks 1-24 hrb Blocks in oven at 60 C 48 hrC aEpon-araldite mixture (EAM): Araldite 6005 - 20 parts Epon 812 - 25 parts dodecenyl succinic anhydride - 60 parts 2,4,6 tri(dimethylamino- methy1)phenol - 2 parts. bUntil the tissue has sunk to the bottom of the EAM CUntil EAM was completely polymerized 28 capped with glass marbles and heated for 10 minutes at 100 C in a water bath. The solution was cooled and, after thorough mixing, a sample was transferred to a silica cuvette and the absorbance at 235 nm was measured in a spectrophotometer (Varian, Cary 219) against sulfuric acid blank. V. Serological Studies Cultures These included 59 avian isolates, as described previously in Table 5, and 50 manmalian isolates of P. haemolgtica. Lyophilized strains of 12 P. haemolytica serotypes were obtained from G. H. Frank, National Animal Disease Center, Ames, Iowa. Source and origin of mammalian isolates of P. haemolytica are listed in Table 10. Table 10. Source and origin of mammalian isolates of P. haemolytica used for serological typing Number of isolates Origin Source 16 Bovine J. R. Beck, American Cyanamid Company 3 Ovine J. R. Beck, American Cyanamid Company 7 Bovine G. R. Carter, Michigan State University 2 Bovine G. H. Frank, National Animal Disease Center 9 Ovine G. H. Frank, National Animal Disease Center ..9 Ovine E. L. Biberstein, University of California-Davis l Caprine E. L. Biberstein, University of California-Davis 3 Bovine E. L. Biberstein, University of California-Davis 29 Typing Antisera Antiserum to each of the 12 serotypes and to each of the 6 avian isolates was prepared in the rabbits as described by Biberstein et al. (1960), except that the tryptose agar grown cultures were used. Smooth colonies of P. haemolytica were streaked onto the tryptose agar plates. Following 18 hours of incubation at 37 C, the growth was harvested with 10 ml of normal saline. The cells were killed with 0.3% formalin and stored in the refrigerator until used. Two adult healthy New Zealand White rabbits were used for each antigenic preparation. The immunizing schedule used was that recommended for Klebsiella pneumoniae (Edwards and Ewing, 1955): 0.5 ml subcutane- ously, then 1.0, 2.0, 3.0, 3.0, 3.0 ml intravenously at 4-day intervals. Six days after the final 3 ml dose, the rabbits were trial-bled. If the titers were satisfactory (>1:200), the serum was harvested on the follow- ing day. If not, 3 additional injections of 3.0 ml were given until a satisfactory titer was Obtained. Antisera were stored at -60 C. Indirect Hemagglutination (IHA) Procedure The IHA procedure adapted to a microtiter system was used as described by Biberstein (1978). The isolate to be typed was grown in 5 m1 brain heart infusion broth at 37 C for 16 hours. The culture was heated at 56 C for 30 minutes to kill the bacteria and to release soluble antigen from the surface of the cells into the medium. Bovine red blood cells (BRBC) were washed 3 times in phosphate buffered neutral formalin- ized normal saline solution and packed after the third washing. Packed BRBC (0.5 ml) was suspended in 5 m1 of heated culture preparation. After thorough mixing, the culture-BRBC mixture was incubated in a 37 C water bath for 1 hour. At the end of the incubation period, the cells were 30 washed again 3 times in buffered formal saline. After the last washing, 10 ml of normal saline was added giving a 0.5% suspension of modified BRBC. With a microtiter set, microtiter pipettes delivering drops of 0.05 ml were used. One drop of diluted (1:50) serum and 1 drop of modified BRBC suspension were placed in each well and the mixture was incubated at room temperature for 3 hours before being examined for hemagglutination. A positive test was indicated by a smooth, uniform layer of RBC's evenly lining the bottom of the wells. Dense red buttons at the lowest point of the wells indicated a negative test. With sera of acceptable titer (>1:200), these reactions were usually quite unequivocal and only 1 serum reacted with the culture to be identified. An antigen control well contained RBC suspension and normal saline. Rapid Plate Agglutination (RPA) Procedure The RPA test was performed by the method described by Frank and Wessman (1978). A drop of antiserum (approximately 10 U1) was placed on a clean glass surface, and then a small amount of P. haemolytica colony from blood agar was picked up on an inoculating needle and mixed with the Serum. A strong positive reaction in the form of clumping and clearing occurred as the mixture was stirred with the needle. Negative reactions remained turbid. ATests were performed at room temperature All cross reactions were quantitated by titration of antisera against the cross reacting serotypes. In the control test, a drop of normal saline was mixed with a small amount of colony to be typed. 31 Titration of Antisera Two-fold serial dilutions of the sera were made in normal saline containing 1:10,000 thimerosal. The IHA and RPA tests were performed in the same manner as described previously. Titers were expressed as the reciprocal of the last serum dilution at which positive agglutina- tion occurred. Counterimmunoelectrophoresis (CIE) The CIE test was performed by the method of Cho and Greenfield (1978) as described by Carter and Chengappa (1981). Cultures were streaked to provide nearly confluent growth on fresh blood agar plates consisting of trypticase soy agar (BBL) with 6% ox blood. After incubation for 24 hours, the growth was washed off with 5 m1 of normal saline. The suspension was heated at 56 C for 30 minutes! after WhiCh the bacteria were sedimented by centrifugation. The clear supernatant fluid was used as the capsular antigen. The electrophoresis plates were prepared by precoating glass plates (10 x 8 cm) with 15 ml volumes consisting of 0.5% agarose (Seakem), 0.5% bacto agar (Difco) and 0.015% sodium azide in 0.025 M barbital buffer pH 8.8 (High Resolution Buffer; Gelman). Wells, 3 mm in diameter, were prepared with a template (Grafar). The distance between wells center to center was 7 mm. A 20 ul quantity of capsular antigen was placed in the cathodal well and an equal quantity of antiserum was placed in the anodal well. The electrophoresis tank (Gelman) contained barbital buffer pH 8.8. Controls included 0.85% sodium chloride solution against antisera and capsular antigens. The antigens and anti- sera were electrophoresed for 30 minutes at 150 V. The plates were then examined for precipitation lines and the presence of a distinct line was interpreted as positive. After electrophoresis, the plates 32 were held in a 2% saline bath overnight to wash away nonspecific pre- cipitate, dried, and stained with 0.1% amido black. Staining Procedure The plates were dried overnight at 37 C with a wet filter paper covering the agar surface. The dried plates were immersed in 0.1% amido black (0.1 g of amido black powder + 50 m1 of l M glacial acetic acid + 50 ml of 0.1 M sodium acetate) for 10 minutes and decolorized with 1 M glacial acetic acid. Decolorization step was repeated several times until a clear background was evident. Finally, the plates were rinsed in distilled water, air dried and photographed. Agglutination Absorption Test (AAT) The AAT was performed according to the method described by Frank (1980). Cells were harvested from 20-hour tryptose broth cultures, washed in 0.15 M sodium chloride, and suspended in 0.15 M sodium chloride with 1:10,000 thimerosal. Three milliliters of cell suspension equivalent to 10 times an optical density of 0.5 at 575 nm was centrifuged at 4,340 x g for 30 minutes. The cell pellet was suspended in 0.5 m1 of a 1:16 dilu- tion of the antiserum to be absorbed. After 2 hours of incubation at 37 C with frequent shaking, the mixture was refrigerated overnight, and the cells were centrifuged from the serum at 4,340 x g for 30 minutes. VI. Estimation of Guanine-Cytosine (GC) Ratio Cultures Fourteen cultures of Pasteurella sp. and 2 cultures of Actinobacillus sp. were used. Details of the cultures are listed in Table 11. The deoxyribonucleic acid (DNA) samples were isolated according to the method of Bohacek and Mraz (1973). The 24-hour cultures were washed Table 11. GC ratio 33 Source and origin of the cultures used for the estimation of Culture Number Origin Sourcea P . mul toci da 656 Bison G. R. Carter, MSU, East Lansing, MI P1235E Bovine G. R. Carter, MSU, East Lansing, MI P. haemolytica Type-1 Bovine NADC, Ames, IA Type-2 Bovine NADC, Ames, IA Actinobacillus lignieresii, Bovine Dept. of Microbiology, MSU, East Lansing, MI Actinobacillus equuli Equine Dept. of Microbiology, MSU, East P. haemolytica P3868 P3873 T3 Avian Avian Avian Lansing, MI NADC, Ames, IA NADC, Ames, IA Dept. of Animal Sciences, MSU, East Lansing, MI A164 A161 A283 A355 A113 A1681 A289 Escherichia colib Avian Avian Avian Avian Avian Avian Avian VDL, VDL, VDL, VDL, VDL, VDL, VDL, MSU, MSU, MSU, MSU, MSU, MSU, MSU, East East East East East East East Lansing, Lansing, Lansing, Lansing, Lansing, Lansing, Lansing, MI MI MI MI MI MI MI ATCC 14763 --- Dept. of Microbiology, MSU, East Lansing, MI aMSU = Michigan State University, NADC = National Animal Disease Center, VDL Veterinary Diagnostic Laboratory bControl with a GC ratio of 50% 34 by SE solution (0.15 M sodium chloride + 0.1 M ethylenediaminetetraacetic acid, pH 8), centrifuged and washed twice with SE buffer again. For further treatment, the amount of 3 to 5 g wet bacterial mass was used and suspended with careful stirring into 50 m1 of the mentioned buffer. Dodecyl sulfate was added to adjust 1% concentration and the mixture incu- bated for 15 minutes under periodic stirring in water bath at 60 C. After cooling the very viscous lysate was diluted with SE solution to make 80 m1; a 20 ml portion of 5 M sodium perchlorate was added to reach a final 1 M perchlorate concentration essential for separation of pro- teins from the DNA. At this stage, the samples were left overnight. For deproteinization, chloroform-isoamylalcohol mixture (24:1) was used at a quantity of approximately 1/3 of the original volume of the suspension. Then the treatment continued under intensive shaking for 30 minutes, performed on a mechanical shaker, and 10 minutes centrifuga- tion carried out at 5000 x 9. Finally, a fibrous DNA was precipitated with 1.5 volumes of 96% redistilled ethyl alcohol. The fibrous DNA was slightly pressured to be free from ethyl alcohol excess and immediately dissolved in a 20 ml portion of 10-fold- diluted SSC (0.15 M sodium chloride + 0.015 M sodium citrate, pH 7). After complete dissolving, the SSC concentration of a mixture was adjusted by addition of 0.1 volume of lO-fold concentrated SSC and by addition of 50 ug/ml ribonuclease (Sigma). After 30 minutes, an enzymatic cleavage was carried out at 37 C, prior to repeatedly performed depro- teinization with chloroform-isoamyl alcohol, until the protein interlayer occurring after centrifugation disappeared. Deproteinized DNA was centrifuged at 75,000 x g for 1 hour and a clear upper layer was repre- cipitated by the gradual addition of ethyl alcohol until 1.5 total volume was reached. Then the DNA sample was dissolved in lO-fold diluted 35 SSC solution. The approximate DNA concentration of 1 mg/ml, overlayered with a drop of chloroform, was kept in the refrigerator. Two methods were used for the DNA base determination. A funda- mental method used was that of thermal denaturation performed in PE buffer medium (0.01 M sodium phosphate buffer + 0.001 M ethylenediamine- tetraacetic acid, pH 7) with a spectrophotometer (Beckman DU). The cell holder compartment was thermostated at both sides by 2 thermospacers for circulating hot water from a U-lO ultra thermostat. The tempera- ture was measured with a rod thermometer directly in one of the cuvettes. For the GC ratio calculations, the equation GC = (Tm - 51): 0.45 was employed. The width of transitional interval (AT) was recorded and the value 20 = (AT - 3) 2.5 was calculated. The value 20 served then to express graphically the heterogeneity of the sample (DeLey and VanMuylem, 1963). As a second control method for determining the percent GC content in isolated samples of DNA, the method of Fredericq et al. (1961) was used. This method is based on the fact that the extinction ratio of 260:280 nm (E ) measured in a medium of 0.1 M acetic acid (pH zeo‘Ezao 3) is dependent on the percent GC content in DNA. The samples of DNA were dissolved in a PE medium in a concentration of approximately 2 mg/ ml. Before measurement, 0.1 M acetic acid was added to the final concen- tration of DNA, 20 to 30 ug/ml. The measurements of optical density at 260 and 280 nm, respectively, were made with a type Varian, Cary 219 spectrophotometer. RESULTS I. Fermentation Studies The 59 isolates obtained from various sources (Table 5) were identified as P, haemolytica by cultural and morphological character- istics and standard biochemical tests. It was observed that all the isolates produced acid(s) from glucose, lactose, sucrose and mannitol. Eighty-seven percent of the isolates produced acid(s) from trehalose and xylose. However, arabinose and sorbitol were attacked by 67 and 78% of the isolates, respectively. The details of the results of the tests are given in Table 12. The growth of the isolates on sheep blood agar plates was much superior with 5% of the air replaced with carbon dioxide. Twenty-four-hour colonies on sheep blood agar plate ranged from 1 to 2 mm in diameter. They were smooth, opaque, and pasty in consistency, with a large zone of B-hemolysis (Appendix A1). Recovery of viable organisms from cultures stored at room temperature was not possible after 4 to 5 days. Viability of the cultures was not appreciably affected when they were stored in defibrinated sheep blood at -60 C. II. Growth Studies Growth curves of all the 4 cultures are depicted in Figures 1 and 2. Lag periods ranging from 1.5 to 3 hours were observed for these cultures under constant aeration. The growth was considerably faster 36 37 Table 12. Results of the biochemical tests on 59 avian isolates of P. haemolytica, Test Percent Positives Triple sugar iron agara 100 Oxidase 100 Nitrate reduction 100 Indol 0 Growth on MacConkey agar 100 Urease 0 Ornithine decarboxylase 0 B-hemolysis 100 Acid from glucose 100 Acid from lactose 100 Acid from sucrose 100 Acid from arabinoseb 67 Acid from trehaloseb 87 Acid from salicin 0 Acid from xylose 87 Acid from mannitol 100 Acid from sorbitol 78 a . . ACid butt, aCid slant, no hydrogen sulfide and no gas production. b Results were rec6rded after 10 days of incubation. 38 Figure 1. Growth curves of P. multocida (656) and A. lig- nieresii in aerated cultures. Figure 2. Growth curves of aerated cultures of P. haemolytica on bovine and avian origin. ABSORBENCE AT 575 nm ABSORBENCE AT 575 nm 39 1 2 _ A = P. MULTOCIDA __ B ' 8 = A. LIGNIERESII '- __ _ -.. _——- A 0.8 '- 0.4 " l I 1 l L _ _ | 0 2 4 8 8 10 12 14 18 24 TIME IN HOURS Flint: I 1 2 _ c = P. HAEMOLYTIcMBov) ' D = P. HAEMOLYTICA (AVIAN) _ - - nun-Il— C - - — —— D 0.8 '- 0.4 - I l I I l L l J _____ J 0 2 4 6 8 10 12 14 18 24 TIME IN HOURS Filure 2 40 in.P. haemolytica of avian origin than in the other 3 cultures studied. The avian isolate multiplied logarithmically for 1.5 to 6 hours, whereas the bovine isolate nmfltiplied logarithmically from a period of 3 to 10 hours. The logarithmic phase of P- multocida was similar in length to that of P. haemolytica of avian origin. The longest logarithmic period was observed in A. lignieresii, viz. 12 hours. The numbers of viable cells counted at different intervals are tabulated in Table 13. The figures are averages of counts of duplicate cultures. The maximum number of viable cells was observed at 7, 12, 10 and 6 hours of incubation of the cultures of P. multocida, A. lig- nieresii, P. haemolgtica (bovine) and P- haemolytica (avian), respec- tively. The number of viable cells gradually declined in all cultures after the respective logarithmic peaks. The viable cell counts were not estimated after 16 hours of incubation. III. Pathogenicity Studies Pathggenicity to Mice The preparation contained 2 grams of hemoglobin per 100 ml as measured in a hemoglobinometer. The results of the mouse pathogenicity study are summarized in Table 14. The hemoglobin preparation appeared to be a better enhancer of virulence of P. haemolytica than the mucin preparation, as there were more deaths of mice in hemoglobin-injected groups. However, isolate P3868 appeared to be more virulent when injected with mucin than with hemoglobin. A similar result was also observed with a hemoglobin solution of the same concentration prepared by lysing bovine red blood cells from defibrinated blood. No deaths were observed in toxicity and culture control groups. The viable cell count for each of the cultures was estimated by the spread plate 41 Table 13. Viable cell counts per milliliter of culture estimated at different intervals of incubation period Period of a . b Incubation P. multocida A. lignieresii P. haemolytica P. haemolytica 0 2.0 x 103 2.0 x 102 2.0 x 102 1.0 x 103 2 1.5 x 104 2.0 x 103 1.0 x 104 2.0 x 104 4 2.5 x 107 2.5 x 104 3.0 x 106 3.2 x 107 6 3.0 x 109 2.0 x 106 3.5 x 108 3.8 x 1010 7 1.5 x 1010 2.2 x 106 3.5 x 108 3.2 x 1010 8 2.0 x 109 4.0 x 107 2.5 x 109 2.5 x 109 10 5.0 x 107 3.5 x 109 2.5 x 1010 1.5 x 108 12 1.0 x 106 2.5 x 1010 2.5 x 109 1.2 x 106 14 1.5 x 104 1.5 x 107 1.2 x 106 2.0 x 104 16 1.8 x 104 1.5 x 105 2.2 x 104 2.2 x 104 a . . Bov1ne isolate bAvian isolate 42 Table 14. Mouse pathogenicity of P. haemolytica cultures injected with mucin and hemoglobin Viable Cell LD50 Preparation Count/mla Dilution No. organisms represented J28 + mucin 2.5 x 109 10—3'5 2.0 x 105 J28 + hemoglobin 2.5 x 109 10_4'85 8.8 x 103 P1148 + mucin 2.8 x lo9 10'?”65 1.6 x 106 P1148 + hemoglobin 2.8 x lo9 10'3°65 1.6 x lo5 P3873 + mucin 3.0 x 109 10“)”0 7.5 x 107 P3873 + hemoglobin 3.0 x lo9 10'1'36 3.3 x lo7 P3868 + mucin 3.5 x 109 10-1'5 2.8 x 107 P3868 + hemoglobin I 3.5 x 109 10-1.0 8.8 x 107 A113 + mucin 3.2 x 109 10-1'0 8.0 x 107 A113 + hemoglobin 3.2 x 109 lO-LS 2.5 x 107 A355 + mucin 2.8 x 109 10-1.2 4.4 x 10.7 A355 + hemoglobin 2.8 x 109 10-2.0 7.0 x 106 A289 + mucin 3.0 x 109 10-1'0 7.5 x 107 A289 + hemoglobin 3.0 x 109 10-1.5 2.4 x 107 A283 + mucin 2.9 x 109 10'1'0 7.3 x 107 A283 + hemoglobin 2.9 x lo9 10’2“25 4.1 x lo6 aViable cell count of the culture injected 43 technique as described previously, and the results are given in Table 14. The viable cell counts of the cultures of P. haemolytica ranged from 2.5 x 109 to 3.5 x 109 cells per ml. The total number of organisms represented by each LD was calculated and is presented in Figure 3. 50 It was observed that mice injected with the culture-mucin mixture needed more viable cells to cause 50% mortality than with the culture-hemoglobin mixture. An exception was observed with culture P3868. Mice injected with this culture-mucin mixture needed fewer viable cells. The two-way analysis of variance for the mouse pathogenicity experiment is presented in Table 15. The interaction and error were not separable. Accordingly, differences in treatment (mucin vs. hemo- globin) were not significant (p>0.25) and differences in strains (bovine vs. avian) were also not significant (p>0.50). Table 15. Two-way analysis of variance for mouse pathogenicity test Source of Variation d.f. m.s. 15a Treatments 1 2.39 x 10 15a Strains 7 1.1107 x 10 Error 7 3.6621 x 1015 ap>o.os Pathogenicity to Chickens of Different Age Groups The intramuscular route of inoculation was the most effective route, followed by the wing web route, but the cloacal and oral routes were proved to be totally ineffective in causing deaths among day-old chicks. .8080... v o o o a - - a n o o a a n 0 o o a o o a c a no ICIOIOICI‘OQOIIIOIOCJIL.IOAIAAA-IAAA-AAAAA-AA‘ . 0 0.0 a .- ooooooocououoovo o to co Ill-OIIIOIOIo0.0.0!I...0.00.0IlllJ-IIIOIIOOOOIIQC noble-00.0.0.0... o o no. 0 e c on n a o cocoa-a- - In - SISIIVSIIO :III '0! '801 I'll“ P3"! P38" IIIII I355 I"! "83 “III!!! M. The number of viable organisms needed to cause 50% mortality among mice Figure 3. 10115 . rat 0 d hemoglobin prepa inan ted with muc injec 45 The deaths in day-old chicks for a 21-day postinoculation period are given in Table 16. There were 21 deaths observed in the day-old chick group inje-ted intramuscularly, but only 8 deaths were encountered in the same age group injected by the wing web route. Three- and six- week-old chickens demonstrated a strong resistance to the cultures, as there were no deaths or sickness observed for a period of 21 days. The dead birds were necropsied and P. haemolytica was isolated from the lung and intestinal samples (Table 17). The remaining birds, including the controls, were sacrificed by cervical dislocation and necropsied. The lungs and intestines were cultured for P. haemolytica and the results are provided in Table 17. Pasteurella haemolytica was not isolated from the lungs and intestines of 91 uninfected birds and 8.3% of the 6-week-old birds carried P. haemolytica in their lungs. Although P. haemolytica was not isolated from the intestines of the 3-week-old birds, 8.3% of the intestines from the 6-week-old birds were positive for P. haemolytica. The day-old chicks that died of infection demonstrated a severe catarrhal enteritis and hemorrhages on the epi- cardia; livers and kidneys were swollen and dark red in color. The lungs were congested and tracheas were hemorrhagic in all the chicks that died of infection. The sacrificed birds, when necropsied, did not show any of the above-mentioned tissue changes. IV. Electron Microscopic Studies Electron microscopic examination of the cells grown on tryptose agar was carried out to measure the cell size and thickness of the cell membrane. The results are presented in Figures 4 through 7 and Appendices A-2 through A-5. The average cell size of each of the cultures was calculated and is listed in Table 18. It was observed that the length 46 Table 16. Results of inoculation of chicks with 3 cultures of P. haemolytica during a postinoculation period of 21 days Mortality Isolate Route of Inoculation Day-Old 3-Week 6-Week P3868 Intramuscular 6/10 0/10 0/10 Wing web 2/10 0/10 0/10 Cloacal 0/10 0/10 0/10 Oral 0/10 0/10 0/10 A283 Intramuscular 8/10 0/10 0/10 Wing web 2/10 0/10 0/10 Cloacal 0/10 0/10 0/10 Oral 0/10 0/10 0/10 A113 Intramuscular 7/10 0/10 0/10 Wing web 4/10 0/10 0/10 Cloacal 0/10 0/10 0/10 Oral 0/10 0/10 0/10 47 Table 17. Isolations of P. haemolytica from lung and intestinal samples of experimentally infected and control birds No. of Percent Recovery of P. haemolgtica Age Group Birds Lung Intestine Infected birds (succumbed) Day-old 29 100 100 Infected birds (survived) Day-old 91 0 0 3-week 120 15 0 6-week 120 8 . 33L 8 . 3a Controls Day-old 30 0 0 3-week 30 0 0 6-week 30 0 3.3 aP. haemolgtica was not always isolated from the lung and intes- tine of the same bird. 48 Table 18. The average cell size and the cell membrane thickness of Pasteurella and Actinobacillus species as determined by electron microscopy MembraneoThickness Organism Size((um) (A) P. multocida Type B (strain 656) 1.2i0.2X0.5i0.1 150 Type E (strain P1235E) 1.2i0.3X0.5i0.1 140 P. haemolytica TyPe-l (strain P1148) 1.3i0.5X0.5i0.1 120 Type-2 (strain J28) 1.3i0.4X0.5i0.l 130 A. lignieresii 1.5i0.9x0.5i0.1 150 P. haemolyticaa P3868 1.2i2.0X0.6il.0 195 P3873 1.2i2.0X0.6il.0 180 a . . AVian isolates 49 Figure 4. Electron micrograph of type B P. multocida (strain 656) grown on tryptose agar at 37 C. Arrows showing the bleb-like structures. Magnification X46,300. Figure 5. Electron micrograph of type-1 P. haemolytica (strain P1148) grown on tryptose agar at 37 C. Arrow showing the bleb-like structure. Magnification X46,300. 50 Figure 4 Figure 5 51 Figure 6. Electron micrograph of avian isolate of P- haemolytica (isolate P3868) grown on tryptose agar at 37 C. Arrow showing the bleb-like structures. Magnification X46,300. Figure 7. Electron micrograph of A. lignieresii grown on tryptose agar at 37 C. Arrows showing the bleb-like structures. Magnification X46,300. 52 Figure 6 Figure 7 53 and width of avian isolates of P. haemolytica were slightly greater than those of bovine isolates. Similarly, the cell membrane thickness was found to be greater in avian isolates than in bovine isolates of P. haemolytica. The average cell membrane thickness was measured with the aid of an ocular micrometer and ranged from 120 to 195 A. The cell membrane was tightly attached to the cell in avian isolates, but it was loosely attached to the cell in bovine isolates of P. haemolytica (Figures 5 and 6). The cell membranes of P. multocida and A. lignieresii were also found to be loosely attached (Figures 4 and 7). A few bleb- 1ike structures were observed in all the micrographs. They were pre- dominant with P. multocida, fewer with A. lignieresii and least with P. haemolytica. The bleb-like structures were fewer in avian isolates than in bovine isolates of P. haemolytica. There was no evidence for the presence of poly-B-hydroxybutyrate granules in the Pasteurella species studied (Table 19). A standard curve for poly-B-hydroxybutyric acid assay is provided in Appendix A—6. V. Serological Studies The antisera collected from the rabbits were titrated by IHA and RPA procedures and the titers are provided in Table 20. The RPA titers were lower than IHA titers for all the strains. The titers ranged from 1:256 to 1:1024 in the IHA test and 1:64 to 1:512 in the RPA test. Adverse effects during hyperimmunization of the rabbits were not observed. IHA, RPA and CIE Tests There were no cross reactions observed among 12 serotypes of P. haemolytica by the IHA test (Table 21). However, cross reactions were 54 Table 19. Spectrophotometric assay of poly-B-hydroxybutyric acid Organism Assay Result P . mul toci da Type B (strain 656) Negative Type E (strain P1235E) Negative P. haemolytica Type-l (strain P1148) Negative Type-2 (strain J28) Negative A. lignieresii Negative . a P. haemolytica P3868 Negative P3873 Negative .b . E. 0011 Negative A. vinelandil'C Positive a . . . AVian isolates b O Negative control c Positive control 55 Table 20. IHA and RPA titers of rabbit antisera against homologous P. haemolytica ‘-.—-' --‘>. Straina Capsular Type IHA Titer RPA Titer d J29 l 256 128 J28 2 256 128 863 3 256 64 S 4 512 64d 613 5 512 128 A30 6 256 64 Hl 7 512 256 H21 8 512 256 d 81 9 1024 512 JF2 10 512 128 KC282 11 256 128 d 5209 12 512 256 P3868 UTb NAC 512d P3873 UT NA 512 d A71 UT NA 512 A164 UT NA 512 A161 UT NA 512d A283 UT NA 512 a . . . . The last 6 strains are aVlan origin. b UT c . . NA = no agglutination. d Flakes observed during agglutination. chicken isolates are untypable by IHA. 56 Table 21. IHA, CIE and RPA reactions of rabbit antisera against heterologous and homologous P- haemolytica serotypes before and after absorption -4.— Anti- Antigen(s) Reacted With:a Antigen(s) Reacted With:b serum IHA CIE RPA IHA CIE RPA Type 1 l l & 4 l, 4 & 6 1 l 1, 6 & 8 Type 2 2 2 2 & 4 2 2 2 & 4 Type 3 3 3 3 3 3 3 TyPe 4 4 4 & 2 4, 2 & 8 4 4 4 & 8 Type 5 5 5 5 & 8 5 5 5 & 8 TYpe 6 6 6 6 6 6 6 Type 7 7 7 7 & 12 7 7 7 & 12 Type 8 8 8 8 8 8 8 Type 9 9 9, 2 & 6 9, 2 & 6 9 9 9 & 4 Type 10 10 10 10 10 10 10 Type 11 11 11 11 ll 11 11 Type 12 12 12 12 ‘12 12 12 aBefore absorption bAfter absorption 57 common in the RPA and CIE tests. It was possible to eliminate the cross reactions in the CIE test by serum absorption technique but not in the RPA test (Figure 10; Table 21). Figures 8 and 9 illustrate the homolo- gous counterimmunoelectrophoretic patterns of 6 avian isolates and 12 mammalian type strains of P. haemolytica, respectively. The counter- _ immunoelectrophoretic patterns of cross-reacting serotypes 1, 4, and 9, before and after the absorption of serum, are shown in Figure 10. It was not possible to eliminate the problem of flakes in the RPA test, even after the culture was passed once through a mouse to assure capsulation. Table 22 provides the distribution of 49 mammalian isolates among 12 serotypes of P. haemolytica by IHA, CIE and RPA procedures. Type-2 was found to be the predoimant serotype, followed by type-l. The remaining serotypes were less frequent. Of the 50 isolates, 49 were typable and l was untypable as determined by IHA and CIE procedures. In the RPA procedure, 40 isolates were typable and 10 were untypable. None of the 59 avian isolates reacted with the 12 type sera as tested by IHA, RPA and CIE procedures. The sera raised against 6 avian isolates were tested for cross reactions by CIE procedure and the results are presented in Table 23. No cross reactions were observed among the 6 avian isolates when the CIE test was run for 30 minutes. Strong precipi- tin lines were observed in all these tests, except with isolate A161. The band with the latter culture was much less dense. Table 24 shows the distribution of 57 avian isolates of P. haemo- lytica among 6 serotypes as determined by IHA, RPA and CIE procedures. The avian type-2 was the predominant serotype reacting with 19 of 59 avian cultures used in the study. Of the 59 isolates, 57 were typable and 2 were untypable by CIE procedure, whereas by IHA procedure all the 59 isolates were untypable. However, the RPA procedure demonstrated 58 Figure 8. CounterimmunoelectrOphoretic patterns of 6 avian type sera against homologous antigens (A) and saline controls (8). Ag = antigen; Ab = serum; C = saline. Figure 9. Counterimmunoelectrophoretic patterns of 12 mammalian type sera against homologous antigens. Ag = antigen; Ab = serum. 59 Figure 8 Figure 9 60 Figure 10, Fund... ' 'lnr'fmphnrpfin patterns of homologous and heterologous reactions of serotypes 9, 4 and 1, before (A) and after (B) absorption (Table 19). Ag = antigen; Ab = serum. Arrows show the heterologous reac- tions before the absorption of sera. 61 Table 22. The distribution of 50 mammalian isolates of P. haemolytica among 12 serotypes Test Procedure Antiseruma IHA CIE RPA Type-1 6 6 3 Type-2 21 21 16 TYpe-3 4 4 4 Type-4 l l l Type-5 4 4 4 Type-6 l l 1 Type-7 2 2 2 Type-8 3 3 3 Type-9 3 3 2 Type-10 l 1 1 Type-ll l 1 l Type-12 2 2 2 Untypable 1 l 10 a . . . . Absorbed with the corresponding cross-reacting antigen(s). 62 Table 23. CIE reactions of rabbit antisera against heterologous and homologous P. haemolytica Antigen Antiserum P3868 P3873 A164 A71 A283~ A161 P3868 + - - — - - P3873 - + - — - - A164 - - + - - - A71 - - - + - - A283 - - - - + - b A161 - - — - _ + a O I AVian isolates bWeak line of precipitin Table 24. The distribution of 59 avian isolates of P. haemolytica among tentatively designated 6 serotypes Tentative Test Procedure Antiserum Designation IHA CIE RPA P3868 Type-l 0 13 56 P3873 Type-2 0 I 19 56 A164 Type-3 0 7 56 A71 Type-4 0 l 56 A283 Type-5 0 8 56 A161 Type-6 O 9 56 Untypable --- 59 2 3 63 that 56 isolates were typable and only 3 were untypable. None of the mammalian isolates reacted with the avian type sera either in CIE or RPA procedures. The RPA test failed to recognize any serotypes, even after diluting the sera (Table 24). VI. Estimation of Guanine-cytosine Ratio For determining the percent GC content in the DNA samples of Pasteurella species and Actinobacillus species examined, 2 different methods were used. The values obtained by the 2 methods are provided in Table 25. The differences between the values obtained by the 2 methods were within the range of 0-4.2% GC (Table 25). The mean values of GC, obtained by thermal denaturation, were 39.65%, 40.85%, 41.0% and 40.25% for P. multocida, P. haemolgtica((bovine), P. haemolgtica (avian), and Actinobacillus species, respectively. Using the spectral analysis, the mean GC values were about 1% higher than the values obtained by thermal denaturation. Escherichia coli strain ATCC14763 used as control showed 49.5% GC content by thermal denaturation method and 49% GC content by spectral analysis method. The values 20 and 30 were calculated for each of the DNA samples and expressed graphically (Figure 11). The solid lines and the dotted lines correspond to 20 and 30 values, respectively. Table 25. Guanine-cytosine content of Pasteurella and Actinobacillus species Thermal Denaturation Analysis of UV-Spectra , o aev L c . Organism Tm ( C) GC% AT“ C 20 3260'3280 GC% P. multocida 656 69.0 40.0 4.2 3.0 1.48 40.0 P1235E 68.7 39.3 4.6 4.0 1.46 41.2 P. haemolytica . Type-1 69.2 40.5 4.4 3.5 1.48 40.0 Type-2 69.5 41.2 4.8 4.5 1.42 44.0 A. lignieresii 69.4 41.0 5.0 5.0 1.46 41.2 A. equuli 68.8 39.5 4.0 2.5 1.48 40.5 . d P. haemolgtlca P3868 68.6 39.0 4.0 2.5 1.49 39.4 P3873 69.0 40.0 4.2 3.0 1.48 40.0 T3 70.2 42.4 4.0 2.5 1.44 42.5 A164 68.8 39.8 4.8 4.5 1.42 44.0 A161 69.8 41.8 3.6 1.25 1.45 41.8 A283 69.6 41.5 3.5 1.25 1.42 44.0 A355 70.0 42.0 3.8 2.0 1.45 41.8 A113 70.1 42.2 4.2 3.0 1.42 44.0 A289 69.2 40.3 3.8 2.0 1.44 42.5 E. coli ATCC14763 73.2 49.5 4.0 2.5 1.38 49.0 aTemperature of melting point bTransition interval cIncludes about 95% DNA molecule dAvian isolates 65 lo go go me P. mulfocida 656 : :7 ,, a P1235E . e e 4, P. haemolytica Type-1 = :_ :fi . Type-2 : e e 4. a A. lignieresii : .; ~ t -: A. equuli H P. haemolytica P3868 WW P3873 ,....__.__.... T3 r+—-o-ra A164 knee :7 .: : A161 ”_‘_.. A283 n—o—n A355 ,,_._*‘ A113 ,__.__,_.,., A289 ' . . I . E. coli ATCC14763 r-r—-—o~c 3'0 4'0 in zsc Figure 11. Graphic depiction of results of guanine-cytosine content of the DNAs of Pasteurella species, Actinobacillus species, and E. coli. The solid horizontal lines represent percent GC :_20 and the solid lines plus dotted lines represent percent GC :_30. DISCUSSION 1. Fermentation Studies Considerable variations have been recognized in the fermentative activity of strains of P. haemolgtica. Smith (1961) observed that all 14 of his type A strains of ovine origin fermented arabinose in 7 days, while 14 type T strains of ovine origin did not. He also found that the type T cultures fermented trehalose in 10 days, while the type A strains did not. However, Shreeve et al. (1970) reported that all T strains fermented trehalose but not arabinose, whereas only a portion of A strains fermented arabinose. The use of arabinose and trehalose to distinguish Smith's biotypes seems to be more reliable for T strains than for the A strains. Fifty-two percent of the avian isolates used in the present study fermented both rehalose and arabinose in 10 days. This observation indicates that the avian isolates of P. haemolgtica may fall into a separate biotype. This also suggests that the avian strains of P. haemolytica do not fit the original biotyping concept introduced by Smith (1961). Similar fermentative variations in avian isolates of P. haemolytica have been reported by Heddleston (1975), Harbourne (1962) and Janetschke and Risk (1970). Heddleston (1975) reported that all of his 20 avian isolates fermented trehalose in 10 days but not arabinose. However, in the present study, 87% of the cultures fer- mented trehalose and 67% of the cultures fermented arabinose in 66 67 10 days. This variation in fermentation reactions may have been due to the small number of avian isolates examined by Heddleston. Arabinose positive avian strains have also been reported by Janetschke and Risk (1970). Although lactose negative A and T strains of P. haemolytica have been reported (Biberstein et al., 1960; Shreeve et al., 1970), none of the avian isolates was lactose negative. All the 59 isolates fermented lactose within 48 hours. Considerable fermentative variation from that of mammalian isolates was also observed among avian cultures in xylose, salicin and sorbitol by Shreeve et al. (1970), Biberstein and Gills (1962) and Smith (1961). Although variations were observed in fermen- tation reactions, the major genus characteristics were essentially identical. The zone of B-hemolysis produced with avian isolates was much wider than that produced with bovine or ovine isolates. This finding is in agreement with Heddleston (1975) and Janetschke and Risk (1970). The zone diameter of hemolysis produced on bovine and ovine blood agar plates was not different with the avian cultures of P. haemolytica. II. Growth Studies The growth characteristics of bovine and ovine strains of P. haemolytl'ca have been studied extensively by Wessman (1964, 1965, 1966) and Wessman and Wessman (1970), but no information is available on either the growth curves or nutritional requirements of avian isolates. Wessman (1966) reported that a strain of bovine P. haemolytica reached a count of 1010 cells per ml after 15 hours of incubation in a chemi- cally defined medium. In the present study, the bovine isolate (P1148) of P. haemolytica reached a cell count of 2.5 x 1010 per m1 after 10 68 hours of incubation in brain heart infusion broth. The difference was probably due to our richer medium. Although the culture was also dif- ferent, there is not ordinarily much difference in growth capacity of different bovine strains. We observed that the avian culture P3868 had a maximum viable cell count, 3.8 x 1010 per ml, in a shorter period of incubation, viz., 6 hours, indicating that this strain probably has a shorter doubling time than mammalian strain. There do not appear to be any reports in the literature on the growth curves of P. haemolytica; however, there have been several studies on the closely related species P. multocida (Banerji and Mukherjee, 1953; Handa, 1958; Carter and Bain, 1960). The growth curves show that the avian strain of P. haemolytica of this study had a shorter lag period (1.5 hours) than mammalian strains (3 hours). The avian strain multiplied logarithmically for 1.5 to 6 hours of incubation, whereas the mammalian strain multiplied logarithmically for 3 to 10 hours of incubation. This provides convincing evidence that the avian strain can grow faster than the mammalian strain of P. haemolytica. As with other bacteria, this would probably apply to the in vivo environment as well as the in vitro milieu. The well known hemorrhagic septicemia P. multocida culture 656 had a growth curve very similar to that of the avian P. haemolytica. The suggestion has been made that P. haemolytica is more closely related to the Actinobacillus genus than to the Pasteurella and conse- quently should be excluded from the latter genus (Mraz, 1969). Our observations do not support the view that P. haemolytica should be transferred to the Actinobacillus genus. 69 III. Pathogenicity Studies Pathogenicityyto Mice The low virulence of P- haemolytica for laboratory animals has hampered the acquisition of a suitable laboratory animal model to study infection by this organism. It has been necessary to include in the culture inoculum substances which effectively increase virulence. Gastric mucin was first used to enhance the virulence of P- haemolytica for mice by Smith (1958). The mode of action is still uncertain. Mucin has been shown to be anticomplementary in vitro by Lambert and Ritchley (1952), but a comparison of various mucins revealed little correlation between antiproperdin activity and the ability to promote infection with E. coli or Staphylococcus aureus (DeWitt, 1958). The introduction of high molecular weight dextran sulfates with the injected P. haemolytica increased the virulence to the same extent as gastric mucin, but a related compound (heparin) was less effective (wessman, 1967). Although mucin has been used extensively in challenge experi- ments as an enhancer of virulence of P- haemolytica, the problem of I nonspecific deaths due to mucin toxicity remains unsolved. In the present study, a crude hemoglobin preparation was injected along with P. haemolytica as an enhancer of virulence. Toxicity to hemoglobin was not observed in the control group, and this lack of toxicity is considered to be a major advantage over the mucin prepara- tion. Although the differences in virulence between mucin and hemo- globin injected groups were statistically not significant at the 5% confidence level, the LD50 showed a considerable difference between these 2 treatment groups. The results of the statistical analysis were not necessarily valid, as there were insufficient replications. We 70 observed that the hemoglobin preparation was a better enhancer of virulence than mucin with all the isolates studied except isolate P3868. In the latter, mucin was better. The exception may have been due to experimental error. Mammalian isolates were observed to be more virulent than the avian isolates of P. haemolytica, but the dif- ference was not statistically significant at the 5% confidence level. Iron compounds are known to increase the virulence of gram- positive and gramrnegative bacteria including P. multocida (Bornside et al., 1968; Kochen et al., 1978; Bullen et al., 1966, 1967; Bjorn et al., 1979; Perry and Brubacker, 1979; Macham et al., 1975; Smith, 1977). Kaplan and Basford (1979) observed that the iron inhibited the leukocyte's phagocytosis-associated metabolic burst and specifically inactivated an important antibacterial product, hydrogen peroxide. They thought that the impairment of intraluekocytic bacterial killing in the presence of iron occurred by at least 2 mechanisms: (1) the inactivation of cationic proteins and (2) the destruction of hydrogen peroxide. Although we do not know the exact mechanism(s) by which P. haemolytica increases its virulence, it seems likely that one of the two mechanisms might have been operating in the presence of excess iron. Pasteurella haemolytica probably produced the iron-binding proteins which enabled the organism to acquire the circulating free iron. More work needs to be done to elucidate the role of free iron in the pathogenicity of P. haemolytica. Our work demonstrated that experimentally induced hyperferremia resulted in an increase in bac- terial growth and virulence, with a marked reduction of the LDSO' 71 Pathogenicity to Chickens of Different Age Groups Pasteurella haemolytica has been isolated from chickens and turkeys with respiratory diseases and salpingitis (Heddleston, 1975). The work of Janetschke and Risk (1970) indicated that the experimental infection of hens, chickens at 10 days of age, and adult white mice depended on the number of organisms administered. Our study revealed that the experimental infection with P. haemolgtica depended on the route of inoculation and the age of the chickens. Day-old chicks were more susceptible than adult chickens. This was probably due to the poorer development of the immune system in the day-old chicks and thus a lesser ability to combat the infection. In the wing web method of inoculation, the dose of the inoculum was probably not enough to cause a high mortality among the day-old chicks. This was in agreement with the report of Janetschke and Risk (1970), who found that the infection was dose related. Although Harbourne (1962) reported experi- mental infection with P. haemolytica in lS-week-old Rhode Island Red chickens, we were unable to infect even 3-week-old White Leghorn chickens. These different results may have been due to the difference in the size of inoculum and route of inoculation used in these 2 studies. Also, the virulence of the organisms injected and the breed of chickens used for the studies might have added to this variation. The incidence of isolation of P3 haemolythmafrom the lungs and intes- tines of surviving birds was not high. This was further evident from the absence of gross pathologic lesions in surviving birds. It was not possible to confirm that those cultures isolated were the same as the original injected cultures. Our results indicated that the injected organisms were not eliminated from the body and that possibly there was a tendency for a carrier state being established in survivors. The 72 percent isolation of P. haemolytica from the control birds was con- siderably less as compared to the percent isolations from birds of the infected survival groups. This strongly suggests that a carrier state for P. haemolytica can be experimentally established in chickens. IV. Electron Microscopic Studies There do not appear to be any reports in the literature on the ultrastructure of P. haemolytica. The present study revealed few important structural differences between the avian and mammalian strains of P. haemolytica. The cell membrane of avian P. haemolytica was very tightly attached to the cell and was much thicker than that of mammalian P. haemolytica. The cell membranes of P. multocida and A. lignieresii were loosely attached like that of mammalian P. haemolytica. The cells of avian strains were found to be shorter and wider than those of mammalian strains of P. haemolytica. The results of this ultrastruc- tural analysis suggested that the avian strains are slightly different from mammalian strains of P. haemolytica. The results also indi-' cated tflfifll there were no major ultrastructural differences between P. haemolytica and A. lignieresii. Projections of bleb-like material from the outer membrane were observed in the micrographs of all the cultures studied. The grams negative bacteria that have been shown to release endotoxin in the form of cell wall blebs originating from theouter membrane include E. coli (Work et al., 1966), Vibrio cholerae (Chatterjee and Das, 1967) and Neisseria meningitidis (Devoe and Gilchrist, 1973). It has been shown that the pathogenicity of certain gram-negative bacteria depended on the release of endotoxin (Devoe and Gilchrist, 1973). Since P. haemo- lytica, P. multocida and A. lignieresii are known to produce these 73 bleb-like structures, it seems possible that the pathogenicity of these organisms may be related to the active extrusion of endotoxin as in other gram-negative bacteria. Poly-B-hydroxybutyric acid has been recognized in many grams positive and gram-negative bacteria (Duodoro-f and Stanier, 1959; Forsyth et al., 1958; Kallio and Harrington, 1960; Morris and Roberts, 1959). This polymer is thought to be the storage granule of some bacteria. Pasteurella haemolytica, P. multocida and A. lignieresii did not seem to produce this polymer. Its absence was confirmed by a spectrophotometric assay. The clear circular spaces observed in the cells of P. haemolytica of mammalian origin which were originally thought to be storage granules were probably artifacts produced as a result of the fixation. V. Serological Studies The titers presented in Table 20 suggest that the rapid plate agglutination test is less sensitive than indirect hemagglutination test. This is in agreement with Frank and Wessman (1978), the developers of the RPA test. Although Frank and Wessman (1978) did not apparently encounter the problem of "flakes" during the RPA test, we observed flakes in 7 of 18 homologous reactions. Since the elimination of the flakes was not possible, even after the cultures were passed once through mice, it was thought that the flake formations were due to partial autoagglutination. It was not possible to estimate the IHA titers of the sera of the avian strains, as there were no reactions observed. This is in agreement with Biberstein et al. (1960), who serologically examined 3 avian strains. The reasons for the negative IHA tests with the avian strains have not been elucidated and, as a 74 result, the taxonomic relationship of these strains to those of mammalian strains has been questioned (Biberstein, 1978). IHA, RPA and CIE Tests Biberstein et al. (1960) divided mammalian P. haemolytica into 10 serotypes with an IHA procedure. Later, this was expanded to 12 serotypes (Biberstein and Gills, 1962; Bibersten and Thompson, 1966). Recently, a rapid plate agglutination procedure for serotyping mammalian P. haemolgtica was developed and the test yielded essentially the same results as the IHA procedure (Frank and Wessman, 1978). The CIE test has been used effectively to detect a variety of microbial antigens, including the capsular polysaccharides of Haemophilus influenzae (Ingram et al., 1979), pneumococci and meningococci (Shackelford et al., 1974) and P. multocida (Carter and Chengappa, 1981). In the present study it was found that the CIE test is a rapid, specific method for serotyping mammalian P. haemolytica. The procedure worked well with older laboratory cultures as well as with those freshly recovered from clinical material. The CIE procedure yielded essentially the same result as the IHA procedure. The advantages of the CIE test over the IHA test were found to be as follows: (1) it was less time consuming; (2) there was no problem of hemolysis of red blood cells as experienced in the IHA test; and (3) it was a more sensitive test. The CIE test was also found to be better than the RPA test, as more mammalian isolates were typable. There were also fewer cross reactions and it was more sensitive than the RPA test. The cross reactions observed in the CIE test were eliminated after the sera were absorbed with the corresponding antigen(s), but it was not possible to eliminate the cross reactions in the RPA test. The cross reactions 75 were probably due to repeated injedtions of antigens during hyperimmuni- zation procedures. We think that the problem of cross reaction in the CIE test might be eliminated by injecting rabbits with a single dose of antigen instead of 6 doses. Since the CIE test proved to be more sensitive, the antibody response produced by a single dose of antigen would probably be sufficient to yield a clear-cut specific line. The results of the RPA test suggested that the somatic antigens might have contributed to the reactions, as the size of the capsule of Pasteurella species is thought to be a phenotypic character and in some instances may have been deficient. The inconsistent reactions observed in the RPA test before and after the serum absorption are further supportive of the effect of somatic antigens on the test. The possible involvement of somatic antigens in the RPA test has been suggested by Frank and Wessman (1978). Of the 50 mammalian isolates, only one was not typable by the IHA and CIE procedures, whereas in the RPA procedure 10 isolates were not typable. The distinct advantage of the CIE procedure, besides being rapid and specific, was that more untypable isolates would be typed. The serotyping of avian strains of P. haemolytica by the IHA test was found to be unsuccessful (Biberstein et al., 1960). Frank and wessman (1978) have not included the avian strains in their RPA pro- cedure. The CIE procedure identified 6 different capsular types among the 59 cultures examined. In contrast, the 59 avian strains were not typable by the RPA or IHA procedures. Whether or not a modification of the IHA procedure will identify the 6 capsular types remains to be explored. Although agglutination was observed in the RPA procedure with avian strains, the test was not specific enough to identify the capsular 76 substances. The lack of specificity may have been due in part to the possible interference of somatic antigens. VI. Estimation of Guanine:§ytosine Ratio The GC ratios of Pasteurella and Actinobacillus species have been studied by Bohacek and Mraz (1967, 1973) and the results indicated that species of these 2 genera were closely related genetically. The genetic relatedness of Pasteurella species was also studied by DNA hybridization technique (Ritter and Gerloff, 1966), and the results indicated that the Pasteurella genus was made up of several subgroups whose members were closely related to each other. The results of our studies were in agreement with the results of previous workers. The present study indicated that the avian strains were genetically closely related to P. multocida, mammalian P. haemolytica, and Actinobacillus species. This further suggested that the avian isolates should be retained in the genus Pasteurella. The estimation of the GC ratio is thought to be a useful tool in taxonomic problems at the genus level but not at the species level. The differences between the results obtained by the 2 methods indicated the presence of impurities in the DNA preparations. In our Opinion, the method recommended by Frederieq et al. (1961) is useful for a rapid and orienting determination of percent GC. However, since this method does not furnish data on the heterogeneity of DNA molecules, it is advisable to use in parallel method of Tm determinations to obtain more detailed information. Figure 11 illustrates graphically the base composition and exponen- tial distribution of DNA molecules. The solid horizontal lines represent percent GC :.20 and the solid lines plus dotted lines 77 represent percent GC 1.30 of the DNA molecules. When comparing the heterogeneity of molecular distribution of individual DNA samples, it is apparent that there is a remarkable genetic relationship among the strains studied. CONCLUSIONS The P. haemolytica cultures isolated from avian species were found to be different in their carbohydrate fermentive reactions from those of mammalian cultures. Fifty-two percent of the avian isolates fermented arabinose and trehalose within 10 days, indicating that they may represent a distinct biotype within the avian strains. The growth studies revealed that the avian strain had shorter lag and logarithmic phases than that of the mammalian strain._ A crude 2% hemoglobin preparation injected into mice along with P. haemolytica was found to be a better enhancer of virulence than 7% swine gastric mucin. The lack of toxicity of the hemoglobin preparation was considered a major advantage over the mucin preparation in challenge experiments. The avian cultures appeared to be less pathogenic to mice than mammalian cultures. The difference in mouse virulence between the avian and mammalian isolates was not statistically significant, as there were insufficient replications. Pasteurella haemolytica was found to be pathogenic to day-old chicks and the pathogenicity depended on the dose and route of inoculation. The ultrastructure of P. haemolytica indicated that the cell membrane of avian isolates was ca. 62 A thicker than that of mammalian isolates. The study also indicated that the cell membrane of avian strains was loosely attached to the cell but in mammalian strains, in contrast, the cell membrane was tightly attached to the cell. The 78 79 bleb-like structures observed on the cells may have a direct relation- ship with the virulence of the organism, as these structures are known to extrude endotoxin. The counterimmunoelectrophoresis (CIE) procedure was found to be less time consuming and more sensitive than the indirect hemagglutina- tion (IHA) test. The CIE test yielded essentially the same result as the IHA test. The results indicated that the CIE test was a much better test than the rapid plate agglutination (RPA) test, as more mammalian isolates were typable and fewer cross reactions were observed. It was possible to serotype 57 avian isolates into 6 distinct capsular types by CIE. The IHA and RPA tests failed to identify these specific capsular substances. The guanine-cytosine (GC) ratios of the DNA samples studied indicated that the avian strains were genetically closely related to P. multocida, P. haemolytica and Actinobacillus species. The pooled results of all the studies strongly support the contention that the avian isolates are different from the mammalian isolates of P. haemolytica and should perhaps be placed in a new group, viz., Pasteurella avihaemolytica. Further research should be directed toward serological and fermen- tative studies with a larger number of avian strains isolated from different geographic regions. More work is also needed to elucidate the pathogenic mechanisms of avian strains under various stress situa- tions. Studies on the effect of different iron compounds as enhancers of virulence of P. haemolytica could yield information which may also be helpful in understanding the in vivo pathogenic mechanisms. Research should also be directed to exploring the role of capsules and possibly the role of pili or other adherence structures in the attachment and colonization ofdP. haemolytica in the intestine and lungs. This 80 information may be useful in reducing the incidence and severity of P. haemolytica infection in domestic animals and poultry. LITERATURE CI TED LITERATURE CITED Adamou, L., J. Bruckler, and H. Blobel, 1972. Further Analysis of Pasteurella Proteins by Vertical Polyacrylamide Gel Disk Electrophoresis. Zentralblatt fur Veterinarmedizin 19B: 412-415. Banerji, T. P., and R. Mukherjee, 1953. Nutritional Requirements of Pasteurella septica. Current Sci. 22: 177-178. Berkman, S., 1942. Accessory Growth Factor Requirements of the Members of the Genus Pasteurella. J. Infect. Dis. 71: 201-211. Biberstein, E. L., 1965. Cross-Reactions Between Types of Pasteurella haemolytica. Cornell Vet. 55: 495-499. Biberstein, E. L., 1978. Biotyping and Serotyping of Pasteurella haemo- lytica. In: Methods in Microbiology, Chapter Ix, pp. 253-269, Vol. 10. Ed. by T. Bergman and J. R. Norris. Academic Press, Inc., New York. Biberstein, E. L., and M. Gills, 1962. The Relation of the Antigenic Types of the A and T Types of Pasteurella haemolytica. J. Comp. Path. 72: 316-320. Biberstein, E. L., M. Gills, and H. Knight, 1960. Serological Types of Pasteurella haemolytica. Cornell Vet. 50: 283-300. Biberstein, E.L., M. E. Meyer, and P. C. Kennedy, 1958. Colonial Variations of Pasteurella haemolytica Isolated from Sheep. J. Bact. 76: 445-452. Biberstein, E. L., B. J. Shreeve, and D. A. Thompson, 1970. Variation in Carrier Rates of Pasteurella haemolytica in Sheep Flocks. I. Normal Flocks. J. Comp. Path. 80: 499-507. Biberstein, E. L., and D. A. Thompson, 1965. Type Specificity of Immunity to Pasteurella haemolytica Infection in Mice. J. Comp. Path. 75: 331-337. Biberstein, E. L., and D. A. Thompson, 1966. Epidemiological Studies on Pasteurella haemolytica in Sheep. J. Comp. Path. 76: 83-94. Bjorn, M. J., P. A. Sokol, and B. H. Iglewski, 1979. Influence of Iron on Yields of Extracellular Products in Pseudomonas aeruginosa Cultures. J. Bact. 138: 193-200. 81 82 Bohaéek, J., and 0. Mraz, 1967. Base Composition of the DNA in Pasteurella haemolytica, Actinobacillus lignieresii and Actinobacillus equuli. Zentbl. Bakt. Parasitkde I. 202: 468-478. Bohacek, J., and O. Mréz, 1973. Deoxyribonucleic Acid Base Composi- tion in Genus Pasteurella. Acta Veterinaria, Czechoslovakia 42: 399-405. Bornside, G. H., P. J. Bouis, Jr., and I. Cohn, Jr., 1968. Hemoglobin and Escherichia coli, a Lethal Intraperitoneal Combination. J. Bact. 95: 1567-1571. Bosworth, T. J., and R. Lovell, 1944. The Occurrence of Haemolytic Coccobacilli in the Nose of Normal Sheep and Cattle. J. Comp. Path. 34: 168-171. Bullen, J. J., G. H. Cushnie, and H. J. Rogers, 1966. The Abolition of the Protective Effect of Clostridium welchii Type A Antiserum by Ferric Iron. Immunology 12: 303-312. Bullen, J. J., A. B. Wilson, G. H. Cushnie, and H. J. Rogers, 1967. The Abolition of the Protective Effect of Pasteurella septica Antiserum by Iron Compounds. Immunology 14: 889-898. Cameron, C. M., 1972. Immunization of Sheep Against Pasteurellosis. J. South African Vet. Assoc. 43: 77-81. Carter, G. R., 1951. Unpublished observations. Carter, G. R., 1956. A Serological Study of Pasteurella haemolytica. Can. J. Microbiol. 2: 483-488. Carter, G. R., 1957, A Bacterin for the Prevention of Shipping Fever in Canada. Vet. Med. 52: 254-255. Carter, G. R., 1967. Pasteurellosis: Pasteurellanmltocida and Pasteurella haemolytica. In: Advances in Veterinary Science, Vol. 11, pp. 321-379. Ed. by C. A. Brandly and C. Cornelius. Academic Press, New York and London. Carter, G. R., 1976. Disease Caused by Pasteurella and Francisella. In: Essentials of Veterinary Bacteriology and Mycology, Chapter 19, pp. 165-172. Michigan State University Press, East Lansing, Michigan. Carter, G- R., 1979. Pasteurella, Yersinia and Francisella. In: Diagnostic Procedures in Veterinary Bacteriology and Mycology, 3rd Ed., PP. 99-107. Charles C. Thomas, Publisher, Springfield, Illinois. Carter, G. R., and R. V. S. Bain, 1960. Pasteurellosis (Pasteurella multocida) A Review Stressing Recent Developments. Vet. Rev. Annotations 6: 105-128. 83 Carter, G. R., and M. M. Chengappa, 1981. Identification of Types B and E Pasteurella multocida by CounterimmunoelectrophoreSis. Chatterjee, S. N., and J. Das, 1967. Electron Microscopic Observations on the Excretion of Cell-Wall Material of Vibrlo cholerae. J. Gen. Microbiol. 49: 1-11. Cho, H. J., and J. Greenfield, 1978. Eradication of Aleutian Disease of Mink by Eliminating Positive Counterimmunoelectrophoresis Test Reactions. J. Clinical Microbiol. 7: 18-22. Collier, J. R., 1968a. Pasteurellae in Bovine Respiratory Disease. J. Am. Vet. Med. Assoc. 152: 824-828. Collier, J. R., 1968b. Significance of Bacteria in Bovine Respiratory Disease. J. Am. Vet. Med. Assoc. 153: 1645-1651. Cowan, S. T., 1974. Identification of Medical Bacteria, 2nd Ed. London University Press, Cambridge. DeLey, J., and J. VanMuylem, 1963. Some Applications of Deoxyribo- nucleic Acid Base Composition in Bacterial Taxonomy. Antonie Van Leeuwenhoek 29: 344-358. Devoe, I. W., and J. E. Gilchrist, 1973. Release of Endotoxin in the Form of Cell Wall Blebs During in vitro Growth of Neisseria meningitides. J. Exp. Med. 138: 1156-1167. DeWitt, C. W., 1958. Differential Effect of Hog Gastric Mucin on PrOperdin and Host Resistance to Infection. J. Bact. 76: 631- 639. Doudoroff, M., and R. Y. Stanier, 1959. Role of Poly-B-hydroxybutyric Acid in the Assimilation of Organic Carbon by Bacteria. Nature 183: 1440-1442. Duncan, J. R., and R. G. Thomson, 1970a. I. Preliminary Observations on the Effect of Specific Immunity on Nasal Flora. Can. J. Comp. Med. 34: 90-93. Duncan, J. R., and R. G. Thomson, 1970b. 11. Influence of Immunization Procedures on Upper Respiratory Tract Immunity in Cattle. Can. J. Comp. Med. 34: 94-100. Edwards, R. R., and W. H. Ewing, 1955. In: Identification of the Enterobacteriaceae. Burgess Publishing Co., Minneapolis, Minnesota. Florent, A., and M. Godbille, 1950. Les Pasteurellas (Coccobacilles) hemolytiques et leur importance dans La Pathogic Animale en Belgique. Leur Sensibilité a La Penicilline. Ann. de Méd Vét 94: 337-361. 84 Forsyth, W. G. C., A. C. Howard, and J. B. Roberts, 1958. Occurrence of Poly-B-hydroxybutyric Acid in Aerobic Gram-Negative Bacteria. Nature 182: 800-801. Fox, M. L., R. G. Thomson, and S. E. Magwood, 1971. Pasteurella haemo- lytica of Cattle Serotype, Production of B—galactosidase and Antibacterial Sensitivity. .Can. J. Comp. Med. 35: 313-317. Frank, G. H., 1980. Serological Groups Among Untypable Bovine Isolates of Pasteurella haemolytica. J. Clinical Microbiol. 12: 579-582. Frank, G. H., and G. E. Wessman, 1978. Rapid Plate Agglutination Procedure for Typing Pasteurella haemolytica. J. Clinical Microbiol. 7: 142-145. Frederiksen, W., 1973. Pasteurella Taxonomy and Nomenclature. Contri- butions of Microbiology and Immunology 2: 170-176. Fredericq, E., A. 0th, and F. Fontaine, 1961. The Ultraviolet Spectrum of DNA and Their Constituents. J. Mol. Biol. 3: 11-17. Gilmour, N. J. L., D. A. Thompson, and J. Fraser, 1974. The Recovery of Pasteurella haemolytica from the Tonsils of Adult Sheep. Res. Vet. Sci. 17: 413-414. Guerrero, R. J., E. L. Biberstein, and S. Jang, 1973. Anales del Congreso Nacional de Medicina Veterinaria Y Zootenia 17-23 de Mayo de 1970, Lima, Peru, pp. 96-97. Hamdy, A. H., N. E. King, and A. L. Trapp: 1965. Attempted Immuniza- tion of Cattle Against Shipping Fever: A Field Trial. Am. J. Hamdy, A. H., and A. L. Trapp, 1964. Immunization of Cattle Against Shipping Fever: Experimental Exposure. Am. J. Vet. Res. 25: 1648-1652. Handa, R. R., 1958. Studies on Pasteurella septica (P. 52) (1) Optimum Requirements. J. Vet. Animal Husbandry Res. 3: 40-53. Harbourne, J. F., 1962. A Haemolytic Cocco-bacillus Recovered from Poultry. Vet. Rec. 74: 566-567. Heddleston, K. L., 1975. Pasteurellosis. In: Isolation and Identifi- cation of Avian Pathogens, Chapter 5, pp. 38-51. Ed. by S. B. Hitchner, C. H. Domermuth, H. G. Purchase, and J. E. Williams. Am. Assoc. Avian Pathologists. Hoerlein, A. B., S. P. Saxena, and M. E. Mansfield, 1961. Studies on Shipping Fever of Cattle. II. Prevalence of Pasteurella Species in Nasal Secretions from Normal Calves and Calves with Shipping Fever. Am. J. Vet. Res. 22: 470-472. 85 Ingram, D. L., A. M. Collier, E. Pendergrass, and S. H. King, 1979. Method for Serotyping Nasopharyngeal Isolates of Haemophilus influenzae: Slide Agglutination, Quelling Reaction Counter- current ImmunoelectrOphoresis, Latex Agglutination and Antiserum Agar. J. Clinical Microbiol. 9: 570-574. Janetschke, P., and G. Risk, 1970. High Incidence of Pasteurella haemolytica in Fowls in Syria. Mh. Vet. Med. 25: 23-27. Jones, F. 5., 1921. A Study of Bacillus bovisepticus. J. Exp. Med. 34: 561-577. Kallio, R. E., and A. A. Harrington, 1960. Sudanophilic Granules and Lipid of Pseudomonas methanica. J. Bact. 80: 321-324. Kaplan, S. S., and R. E. Basford, 1979. Exogenous Iron and Impairment of Intraleukocytic Bacterial Killing, or the Leukocyte Tells Her Story. In: Microbiology, pp. 92-95. Ed. by D. Schlessinger. American Society for Microbiology, Washington, D.C. Knight, H. H., E. L. Biberstein, and M. Allison, 1969. The Role of Capsular and Somatic Antigens in Immunization of Mice Against Pasteurella haemolytica Infection. Cornell Vet. 59: 55-64. Kochen, I., J. Wasynczuk, and M. A. MoCabe, 1978. Effect of Injected Iron and Siderophores on Infection in Normal and Immune Mice. Inf. Immunity 22: 560-567. Kress, R. E., D. H. Will, and J. R. Collier, 1964. Skin Toxicity and Hemodynamic Properties of Endotoxin Derived from Pasteurella haemolytica. Am. J. Vet. Res. 25: 925-942. Lambert, H. P., and J. Ritchley, 1952. The Action of Mucin in Promoting Infections: The Anticomplementary Effect of Mucin Extracts and Certain Other Substrates. Brit. J. Exp. Path. 33: 327-339. Law, J. H., and R. A. Slepecky, 1961. Assay of Poly-B-hydroxybutyric Acid. J. Bact. 82: 33—36. Lillie, L. E., 1974. The Bovine Respiratory Disease Complex. Can. Vet. J. 15: 233-242. Lopez, A., R. G. Thomson and M. Savan, 1976. The Pulmonary Clearance of Pasteurella haemolytica in Calves Infected with Bovine Parainfluenza-3 Virus. Can. J. Comp. Med. 40: 385-391. Lovell, R., and D. L. Hughes, 1935. Diseases of Young Calves: A Bacteriological Examination of 100 Calves. J. Comp. Path. 48: 267-284. Macham, L. P., C. Ratledge, and J. C. Nocton, 1975. Extracellular Iron Acquisition by Mycobacteria: Role of the Exochelins Against the Participation of Mycobactin. Inf. Immunity 12: 1242-1251. 86 Magwood, S. E., D. A. Barnum, and R. G. Thomson, 1960. Nasal Bacterial Flora of Calves in Healthy and in Pneumonia—Prone Herds. Can. J. Comp. Med. 33: 237-243. Marsh, H., 1932. Mastitis in Ewes, Caused by an Infection with a Pasteurella. J. Am. Vet. Med. Assoc. 81: 376-382. Matsuoka, T., M. Folkerts, and C. Gale, 1972. Evaluation in Calves of an Inactivated Bovine Rhinotracheitis and PI-3 Vaccines Combined with Pasteurella Bacterin. J. Am. Vet. Med. Assoc. 160: 333-337. Montgomerie, R. F., T. J. Bosworth, and R. E. Glover, 1938. Enzootic Pneumonia of Sheep. J. Comp. Path. and Ther. 51:-87-107. Morris, M. B., and J. B. Roberts, 1959. A Group of Pseudomonads Able to Synthesize Poly-B-hydroxybutyric Acid. Nature 183: 1538-1539. Mraz, 0., 1969. Comparative Study of the Species Actinobacillus lignieresii and Pasteurella haemolytica. III. Actinobacillus haemolyticus (Newsom and Cross 1932). Comb. Nov. Zentbl. Bakt. Parasitkde I 209: 336-349. Muraschi, T. F., M. Lindsay, and D. Bolles, 1965. Identification and Typing of Strains of Flavobacterium meningosepticum and Pas- teurella haemolytica by the Gel Diffusion Technique. J. Inf. Dis. 115: 100—104. Mwangota, A. U., S. I. Muhammed and R. G. Thomson, 1978. Serological Types of Pasteurella haemolytica in Kenya. Cornell Vet. 68: 84-93. McDonald, M., 1974. A Study of Pasteurella haemolytica Isolated from Feedlot Cattle. M.A. Thesis, University of California, Davis. Newsome, I. E., and F. Cross, 1932. Some Bipolar Organisms Found in Pneumonia in Sheep. J. Am. Vet. Med. Assoc. 80: 711-719. Olmos, A., and E. L. Biberstein, 1979. Differentiation of Pasteurella haemolytica Biotypes A and T with Growth Inhibitions. J. Clinical Micrdbiol. 10: 231-234. Pegram, R. G., P. L. Roeder, and J. M. Scott, 1979. Two New Serotypes of Pasteurella haemolytica from Sheep in Ethiopia. TrOpical Ani. Health and Production 11: 29-30. Perry, R. B., and R. R. Brubacker, 1979. 'Accumulation of Iron in Yersinae. J. Bact. 137: 1290-1298. Phillips, J. E., 1974. Genus Actinobacillus. In: Bergey's Manual of Determinative Bacteriology, 8th Ed., pp. 373-377. Ed. by R. E. Buchanan and N. E. Gibbons. Williams and Wilkins Co., Baltimore, Maryland. Reed, L. J., and H. Muench, 1938. A Simple Method of Estimating Fifty Percent Endpoints. Am. J. Hygiene 27: 493-497. 87 Reynolds, E. S., 1963. The Use of Lead Citrate at High pH as an Electron-Opaque Stain in Electron Microscopy. J. Cell Biology 17: 208-213. Ribi, E., K. C. Milner, and T. D. Perrine, 1959. Endotoxic and Anti- genic Fractions from the Cell Wall of Salmonella enteritidis. Methods for Separation and Some Biologic Activities. J. Immunol. 82: 75-84. Ritter, D. B., and R. K. Gerloff, 1966. Deoxyribonucleic Acid Hybri- dization Among Some Species of the Genus Pasteurella. J. Bact. 92: 1838-1839. Schakleford, C. P., J. Campbell, and R. D. Feigin, 1974. Counter- Current Imunoelectrophoresis in the Evaluation of Childhood Infection. J. Pediatrics 85: 478-481. Shreeve, B. J., I. N.Ivanov, and D. A. Thompson, 1970. Biochemical Reactions of Different Serotypes of Pasteurella haemolytica. J. Med. Micrdbiol. 3: 356-358. Smith, G. R., 1958. Experimental Infections of Pasteurella haemolytica in Mice and Their Use in Demonstrating Passive Immunity. J. Comp. Path. 68: 455-468. Smith, G. R., 1959. Isolation of Two Types of Pasteurella haemolytica from Sheep. Nature (London) 188: 1132-1133. Smith, G. R., 1961. The Characteristics of 2 Types of Pasteurella haemolytica Associated with Different Pathological Conditions in Sheep. J. Pathol. Bacteriol. 81: 431-440. Smith, H., 1977. Microbial Surface in Relation to Pathogenicity. Bact. Rev. 41: 475-500. Smith, J. E., 1974. Genus Pasteurella. In: Bergey's Manual of Determinative Bacteriology, 8th Ed., pp. 370-373. Ed. by R. E. Buchanan and N. E. Gibbons. Williams and Wilkins Co., Baltimore, Maryland. Thompson, D.A., J. Fraser, and N. J. L. Gilmour, 1977. Serotypes of Pasteurella haemolytica in Ovine Pasteurellosis. Res. Vet. Sci. 22: 130-131. Thompson, D. A., and D. L. Mould, 1975. Protein Electrophoretic Pattern of Pasteurella haemolytica. Res. Vet. Sci. 18: 342-343. Thomson, R. G., M. L. Benson, and M. Savan, 1969. Pneumonic Pasteurello- sis of Cattle: Microbiology and Immunology. Can. J. Comp. Med. 33: 194-206. TWeed, W., and J. W. Edington, 1930. Pneumonia of Bovines Due to Pasteurella boviseptica. J. Comp. Path. and Ther. 43: 234-252. 88 Watson, M. L., 1958. Staining of Tissue Sections for Electron Microscopy with Heavy Metals. J. Biophysical and Biochemical Cytology 4: 475-478. Wessman, G. E., 1964. Interrelationships of Smooth and Nonsmooth Variants in the Dissociation of Pasteurella haemolytica. J. Bact. 88: 356-360. Wessman, G. E., 1965. Cultivation of Pasteurella haemolytica in a Casein Hydrolysate Medium. Applied Microbiol. 13: 426-431. Wessman, G. E., 1966. Cultivation of Pasteurella haemolytica in a Chemically Defined Medium. Applied Microbiol. 14: 597-602. Wessman, G. E., 1967. Susceptibility of Mice, Guinea Pigs and Hamsters to Challenge With Pasteurella haemolytica and Its Enhancement by Microbial Polysaccharides and Related Compounds. J. Inf. Dis. 117: 421-428. Wessman, G. E., and G. Hilker, 1968. Characterizations of Pasteurella haemolytica Isolated from the Respiratory Tract of Cattle. Can. J. Comp. Med. 32: 498-504. Wessman, G. E., and G. wessman, 1970. Chemically Defined Media for Pasteurella multocida and Pasteurella ureae and a Comparison of Their Thiamine Requirements with Those of Pasteurella haemo- lytica. Can. J. Microbiol. 16: 751-757. Wilkie, B. M., R. J. F. Markham, and P. E. Shewen, 1980. Response of Calves to Lung Challenge Exposure with Pasteurella haemolytica after Parenteral or Pulmonary Immunization. Am. J. Vet. Res. 41: 1773-1778. Wilkie, B. N., and A. M. Norris, 1976. The Effect of Route of Immuniza- tion on the Lapine Immune Response to Killed Pasteurella haemo- lytica and the Influence of Aerosol Challenge. Can. J. Comp. Med. 40: 346-349. WOrk, E., K. W. Knok, and M. Vesk, 1966. The Chemistry and Electron Microscopy of an Extracellular Lipopolysaccharide from Escherichia coli. Annals of the New York Academy of Sciences 133: 438-449. - Wray, C., and D. A. Thompson, 1973. An Epidemiological Study of Pasteurella haemolytica in Calves. Brit. Vet. J. 129: 116-123. APPENDIX 89 Figure A-l. Blood agar plate showing the B-hemolytic colonies of P. haemolytica. A = avian isolate: B = mammalian isolate. 90 Figure A-2. Electron micrograph of type B P. mul- tocida (strain 656) grown on tryptose agar at 37 C. Magnification X22,500. 91 Figure A-3. Electron micrograph of type-1 P. haemo- lytica (strain P1148) grown on tryptose agar at 37 C. Magnification X22,500. 92 Figure A-4. Electron micrograph of avian isolate of P. haemolytica (isolate P3868) grown on tryptose agar at 37 C. Magnification X22,500. 93 Figure A-S. Electron micrograph of A. lignieresii grown on tryptose agar at 37 C. Magnification x22,soo. 94 .poouoa axommwam can sue an commune oaom owuausomxouomolmuaaom now o>uso pudendum d .o|< ousowm on as 8 _ .53: 6.823323... Q 53.. S _ 8 — 88832 _. _ .— mu 9:: w BOU‘BQJOSQV