Random amplification of polymorphic DNA and phenotypic typing of Zimbabwean isolates of Pasteurella multocida

Veterinary Microbiology 82 (2001) 361±372 Random ampli®cation of polymorphic DNA and phenotypic typing of Zimbabwean isolates of Pasteurella multocida F. Dzivaa,*, H. Christensenb, J.E. Olsenb, K. Mohana a Faculty of Veterinary Science, University of Zimbabwe, P.O. Box MP167, Mount Pleasant, Harare, Zimbabwe b Department of Veterinary Microbiology, Royal Veterinary and Agricultural University, Bulowsvej 13, DK-1870, Frederiksberg C, Denmark Received 6 June 2000; received in revised form 31 May 2001; accepted 31 May 2001 Abstract Eighty-one isolates presumptively identi®ed as Pasteurella multocida from a variety of diseases in animals in Zimbabwe were subjected to biochemical characterization, capsular typing and RAPD analysis. The majority of isolates (over 80%) were assigned into named taxa and were predominantly P. multocida subsp. multocida and P. multocida subsp. septica, whilst the remainder were unassigned. Serogroup A was predominant among the three capsular types (A, B and D) of P. multocida detected. Three main RAPD clusters and three subclusters were observed among the majority of isolates (93.8%), whilst the remainder was found to be weakly related. Nine different groups of strains with similar RAPD pro®les (100% similarity) were also observed. The reference strain of capsular serogroup F clustered with the reference strain of P. multocida subsp. septica, whilst all other serogroups clustered with reference strains of subsp. multocida and gallicida. Notably, serogroups A and D were observed to be closely related to the reference strain of subsp. multocida. The relationship between biotype, capsular type, host origin and disease manifestation was not clear-cut. However, most pig isolates of subsp. multocida clustered together as did most cattle isolates of subsp. multocida. RAPD tended to separate subsp. multocida from septica. # 2001 Elsevier Science B.V. All rights reserved. Keywords: Biotype; Pasteurella multocida; RAPD; Capsular type 1. Introduction Pasteurella multocida infects a wide range of animals, where it causes a variety of disease manifestations. There is evidence that considerable diversity exists in this * Corresponding author. Tel.: ‡263-4-303211; fax: ‡263-4-333407. E-mail addresses: fdziva@compcentre.uz.ac.zw, dziva@excite.com (F. Dziva). 0378-1135/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 1 3 5 ( 0 1 ) 0 0 4 0 6 - 0 362 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 organism. Phenotypic methods, mainly biotyping and serotyping, have been widely used in the taxonomy and epidemiological studies of this species (Rimler and Brogden, 1986; Rhoades and Rimler, 1987, 1990; Mohan et al., 1994, 1997; Fegan et al., 1995). Capsular typing recognizes ®ve serogroups, designated A, B, D, E and F (Carter, 1955; Rimler and Rhoades, 1987), whereas somatic typing distinguishes 16 serotypes (Heddleston et al., 1972). However, this serotyping scheme often presents obstacles as some strains tend to represent more than one serotype (Rhoades and Rimler, 1990), whilst others may occur in acapsular states (Rhoades and Rimler, 1987). Due to shortfalls associated with phenotypic techniques, molecular typing is increasingly being applied in the study of this organism. DNA±DNA hybridization studies (Mutters et al., 1985), identi®ed three subsp. within P. multocida (multocida, gallicida and septica). Apart from hybridization studies, other methods which have been employed include (i) ribotyping (Snipes et al., 1990; Blackall et al., 1998); (ii) random ampli®cation of polymorphic DNA (Chaslus-Dancla et al., 1996); (iii) pulse ®eld gel electrophoresis (Townsend et al., 1997a); (iv) REP±PCR using oligonucleotides primers directed against repetitive extragenic palindromic (REP) sequences (Townsend et al., 1997b) and restriction enzyme analysis (Kim and Nagaraja, 1990; Wilson et al., 1992, 1993). Although molecular typing is increasingly being applied in epidemiological studies, there is evidence that serology and biotyping remain necessary for a complete study (Townsend et al., 1998). To determine whether a particular molecular typing tool could be a substitute for phenotypic typing, the present study examined the relationships between capsular serogroup, biotype and patterns of randomly ampli®ed polymorphic DNA fragments. 2. Materials and methods 2.1. Bacterial isolates A total of 81 clinical isolates presumptively identi®ed as P. multocida obtained from different animal hosts with a variety of disease syndromes were used in the study. The isolates had been isolated in a bacteriology diagnostic laboratory in Zimbabwe over a 10-year period as described by Mohan et al. (1994). A detailed description of the sources of the isolates is given in Table 1. Reference capsular typing strains (A1113, B925, D42, E978 and F4679) and antisera were obtained from The Royal Veterinary College (London). Subspecies type strains, P. multocida subsp. multocida (NCTC 10322T), gallicida (NCTC 10204T) and septica (NCTC 11619T) were National Collection Type Cultures (Colindale) strains. 2.2. Biochemical characterization A total of 46 tests described by Heddleston (1976), Mutters et al. (1985), Biberstein et al. (1991), including the requirement of X and V factors (Kilian, 1974), were used in the characterization of the isolates and subsequent assigning to the taxa of the genus Pasteurella. 363 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 Table 1 Sources of Pasteurella multocida isolates used in the study Disease Pneumonia Wound infection Rhinitis/bronchitis Septicaemia Tonsillitis/colonization Urogenital infections Arthritis Otherb Total a b Number of isolates 20 18 7 11 13 3 4 5 81 Host species Dogs Cats Cattle Pigs Rabbits Poultry Othera 4 11 5 0 0 3 1 2 26 1 6 1 0 3 0 0 0 11 6 0 0 5 0 0 3 0 14 5 0 0 2 10 0 0 0 17 3 0 0 2 0 0 0 0 5 0 0 0 2 0 0 0 2 4 1 1 1 0 0 0 0 1 4 Other: sheep, goat and horse. Other: conjunctivitis, keratitis, abscesses. 2.3. Capsular typing Capsular typing was performed by an adaptation of the procedure described by Sawada et al. (1982). Brie¯y, isolates were grown overnight on dextrose starch agar (Difco). An agar plate of culture was harvested into 1 ml of 0.02 M phosphate buffered saline (PBS, pH 7.2) and suspended by vortexing. The suspensions were treated with equal volumes of PBS containing 200 U of hyaluronidase (bovine testis, Sigma) for 2 h at 378C. This was followed by boiling for 1 h. Crude capsular extracts (containing antigenic capsular polysaccharide) were collected following centrifugation at 13,000  g for 20 min. The antigenic capsular polysaccharide was immobilized onto 1% glutaraldehyde-®xed sheep red blood cells by incubating at 48C for 2 h with agitation. A ®nal 0.5% (v/v) suspension of red blood cells was prepared in PBS containing 0.25% bovine serum albumin (PBS-A). Two-fold serial dilutions of reference antisera were individually prepared in PBS-A in V-shaped microtitre plates (Nuncon). Capsular antigen-coated red blood cells of each isolate were reacted with separate dilutions of reference typing antisera at room temperature. Plates were examined for haemagglutination after 1±3 h. 3. Random amplification of polymorphic DNA 3.1. Preparation of bacterial DNA Bacterial DNA was prepared following an adaptation of the procedure described by Christensen et al. (1993). Brie¯y, a total volume of 6 ml of an overnight shaken brain± heart infusion broth culture was centrifuged in Eppendorf tubes to pellet bacterial cells. The pellet was resuspended in 500 ml TE (50 mM Tris, 50 mM EDTA) buffer. Bacterial cells were lysed with 20 ml of 10% sodium dodecyl sulfate (SDS) for 30 min at 568C. 364 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 Cell proteins were precipitated by addition of 250 ml of 7.5 M ammonium acetate, followed by incubation on ice for 15 min. The lysates were ®rst extracted with equal volumes of phenol±chloroform±isoamylalcohol (25:24:1), then with chloroform± isoamylalcohol (24:1). DNA was precipitated with isopropanol. The DNA pellet was washed twice in 70% ethanol and dried in a SpeedVac. DNA pellets were resuspended in 100 ml of Milli-Q water (Millipore) and the amount estimated with GeneQuant (Pharmacia Biotech). 3.2. Polymerase chain reaction The polymerase chain reaction was performed using a ready-to-go RAPD analysis kit (Amersham Pharmacia Biotech, New York) according to the manufacturer's instructions. Preliminary experiments with all six primers were conducted on ®ve reference strains. The optimal amount of template was determined using 5, 25, 50, 100 ng DNA of reference strains. Ampli®ed fragments were resolved on a 2% agarose gel in TAE (Tris-acetate EDTA) buffer for 4 h at 120 V. The gels were stained in ethidium bromide. Stained gels were viewed, photographed and saved as TIFF ®les using a computer programme, Quantity One (Bio-Rad). 3.3. Computer analysis of RAPD fragments DNA fragments between 500 and 2000 bp were considered for analysis. Analysis was performed using a computer programme, GelCompar (Applied Maths, Kortrjilk, Belgium). The images were normalized, a similarity matrix was produced by the Dice coef®cient and a dendogram was constructed from the resulting data using a UPGMA clustering method. 4. Results 4.1. Biotypes of P. multocida Three of the isolates were found not to be P. multocida, but P. canis (one) and P. stomatis (two). The majority of the isolates (65/78) could be classi®ed into conventional taxa of the genus Pasteurella, and these were mainly from classical pasteurelloses (haemorrhagic septicaemia, pneumonia, fowl cholera, etc.). Most of the isolates belonged to the taxa P. multocida subsp. multocida and P. multocida subsp. septica (Table 2). However, 16/78 isolates could not be assigned to the known taxa, and these were designated `unassigned biotypes' as described by Biberstein et al. (1991). Nine out of a possible 24 unassigned biotypes were observed among these 16 isolates. Dogs and cats isolates presented a greater percentage (75%) of these unassigned biotypes (Table 3), as well as the taxon P. multocida subsp. septica. In contrast, cattle and pigs contributed to a greater proportion of the subsp. P. m. multocida and very little of unclassi®able isolates. Only one isolate belonged to the subsp. P. m. gallicida. 365 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 Table 2 Differential characters and biotype designations for 81 Pasteurella isolates from various animal hostsa Differential characters DUL MAN ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ SOR TRE ORN ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ ‡ Designation/biotype Number of isolates U3 U10 U11 U14 U18 P. multocida P. multocida P. multocida P. multocida U19 U20 P. stomatis P. canis U21 U24 1 1 1 1 1 24 19 18 1 1 4 2 1 2 4 subsp. subsp. subsp. subsp. multocida a multocida b septica gallicida a DUL: dulcitol; MAN: mannitol; SOR: sorbitol; TRE: trehalose; ORN: ornithine decarboxylase; a: trehalose-positive variant; b: trehalose-negative variant; U: unassigned biotype (Biberstein et al., 1991). 4.2. Capsular serogroups Of the 78 isolates of P. multocida studied, 48 were capsular group A, 6 were capsular group B, 9 were capsular group D, 10 were not capsulated, and hence not typeable and 5 were not grouped at all. Three isolates, representing the species P. stomatis (two) and P. canis (one) were also observed to be not typeable. None of the isolates belonged to either serogroup E or F. Table 4 shows the host distribution of the serogroups. Serogroup A was predominant among all the hosts, and was mainly from dogs, cats, cattle and pigs. Table 3 Distribution of biotypes of Pasteurella sp. and subsp. from different animal hostsa Host Dog Cat Cattle Pig Other Total Number of isolates P. m. m. a P. m. m. b P. m. gall. P. m. sept. P. stomatis P. canis Unassigned Total 3 1 5 10 5 24 3 1 5 6 4 19 0 0 1 0 0 1 10 4 1 1 2 18 1 1 0 0 0 2 1 0 0 0 0 1 8 4 2 0 2 16 26 11 14 17 13 81 a P. m. m. a: Pasteurella multocida subsp. multocida (trehalose-positive variant); P. m. m. b: P. multocida subsp. multocida (trehalose-negative variant); P. m. gall.: Pasteurella multocida subsp. gallicida; P. m. sept.: Pasteurella multocida subsp. septica. 366 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 Table 4 Distribution of capsular serogroups of Pasteurella multocida from different animal hosts Host Dogs Cats Cattle Pigs Other Total Number of isolates 26 11 14 17 13 81 Capsular serogroups A B D E F NTa NDb 13 6 9 12 8 48 ± ± 5 1 ± 6 1 ± ± 4 4 9 ± ± ± ± ± 0 ± ± ± ± ± 0 9 3 ± ± 1 13c 3 2 ± ± ± 5 a NT: not typeable. ND: not determined. c Includes not typeable P. stomatis (two) and P. canis (one) isolates. b 5. Random amplified polymorphic DNA 5.1. Primer screening The ready-to-go RAPD analysis kit contains six different primers. In order to obtain a desired primer, preliminary PCR experiments involving all six primers were run on ®ve reference strains. There was a variation in the pattern of ampli®ed fragments generated by each primer on each of the ®ve reference strains. Some of the patterns generated by the six primers on two reference strains are shown in Fig. 1. For a convenient handling of data, it was decided to select primer 2 that generated few major bands, described as those with intensity equal to or greater than that of the 800 bp fragment in the 100 bp ladder. Primer 2 (sequence ˆ 50 -d(GTTTCGCTCC)-30 ) generated extra faint bands on two of the reference strains, but these bands were disregarded during analysis. 5.2. Genetic diversity of isolates as defined by RAPD patterns Fragments generated by primer 2 ranged from 300 to 2000 bp in size in all isolates, and the banding patterns obtained on some isolates are shown in Fig. 2. The number of bands including faint ones ranged from one to six, though the majority had an average of two major bands. Isolates were clustered by UPGMA based on numerical analysis of the Dice similarity coef®cient. Three major clusters were observed (Fig. 3). Cluster I was observed at 47% similarity and this was mainly comprised of P. multocida subsp. septica isolates, including the type strain. Cluster II (observed at 50% similarity) was the largest comprising of three subclusters. The ®rst subcluster (IIa) did not have any signi®cant common phenotypical similarity, but it contained the type strain P. multocida subsp. gallicida. Subcluster IIb (observed at 64% similarity) contained the majority of cattle isolates and nearly all the isolates were of subsp. multocida (8/9). The cluster also contained reference capsular serogroup B and E strains. Cluster IIc was the largest of F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 367 Fig. 1. Screening of primers from the ready-to-go RAPD analysis kit. From left to right; lane 1: 0.1 kb ladder, lanes 2±7: primers 1±6, respectively, on strain E978, lanes 9±14: primers 1±6, respectively, on strain F4679. Fig. 2. Electrophoretic profiles of some Pasteurella multocida isolates following random amplification of polymorphic DNA using primer 2 of the ready-to-go RAPD analysis kit (Amersham Pharmacia Biotech). From left to right; lane 1: 0.1 kb ladder, lanes 2±15: isolates from various animal hosts. 368 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 Fig. 3. The relationship between RAPD cluster, host, disease syndrome, capsular serogroup and phenotypic typing of Pasteurella multocida isolates from Zimbabwe. I, II and III represent clusters; IIa, IIb, IIc: subclusters; ND: not determined; NK: not known; NT: not typeable; U: unassigned; a: trehalose-positive variant; b: trehalose-negative variant; HS: haemorrhagic septicaemia. F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 369 the three subclusters, that included most of the pig isolates (11/17) and almost exclusively P. m. multocida (27/32). The subcluster also contained the reference strain of subsp. multocida and reference serogroups A and D strains, and most of the serogroup D isolates. In addition, most of the isolates in this subcluster were serologically typeable. Unassigned biotypes tended to group together at the bottom of this subcluster. Cluster III (at 34% similarity) included 16 isolates from a variety of animals. These included dog isolates that were untypeable or were of the subsp. septica. The last ®ve isolates could not be assigned to any cluster due to their weak relatedness to others. There appeared to be a genotypic relationship as de®ned by RAPD analysis of P. multocida subsp. multocida isolates associated with respiratory tract infections of mainly pigs (subcluster IIc). In addition, RAPD tended to separate the subsp. multocida from subsp. septica. Nine different groups (each containing at least three or more isolates) of similar RAPD patterns (100% similarity) were observed, with the largest group being found in subcluster IIc. The relationship between RAPD cluster, host, disease, serogroup and biotype designation is also summarized in Fig. 3. 6. Discussion About 20% (16/78) of the isolates could not be assigned to any of the named subsp. of P. multocida. However, a previous study (Mohan et al., 1994), which included 10 isolates used in the present study, revealed that all their isolates were classi®able with the exception of two sorbitol- and dulcitol-negative isolates. In addition, no relationship between serogroup and biotype was observed. The encounter of unclassi®able isolates in the present work con®rms results obtained by Biberstein et al. (1991). More than 25% of the 356 isolates studied were found to be unclassi®able, suggesting the common occurrence of such isolates. Although capsular serotyping offers a relatively fast analysis of P. multocida, its shortfalls are well-recognized. Apart from the laborious preparation of speci®c capsular antisera against each of the known serogroups, one of the commonly encountered obstacles is the occurrence of acapsular forms of the bacteria, which consequently become nontypeable. For example, in the present study 10/78 isolates were non-groupable mainly because they were acapsular. Such problems are not encountered when genotypic methods are employed in the typing of this organism. Random ampli®ed polymorphic DNA analysis uses oligonucleotide primers that amplify certain sections of the genome by PCR to produce identi®able banding patterns, which are useful in strain differentiation (Power, 1996). The procedure has been widely used in a variety of bacteria (Lam et al., 1995; Lin et al., 1996; Chaslus-Dancla et al., 1996; Chatellier et al., 1997; Lee and Mize, 1997; Tambic et al., 1997; Zhang et al., 1997; Maurer et al., 1998; Charlton et al., 1999). Evidence suggests that it is a useful tool for strain differentiation and the availability of commercial kits makes it easy to apply this technique in molecular typing of bacterial isolates. The present study showed that no relationship exists between RAPD pattern and capsular serogroup. However, it was noted that capsular groups A and D mainly occurred in the same cluster as subsp. P. m. multocida. This suggests that the two serogroups may be 370 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 sharing similar genomic sequences. Phenotypically, some strains of these serogroups are well-recognized producers of the osteolytic toxin, the virulence attribute in the pathogenesis of porcine atrophic rhinitis. Serogroup A occurred in all clusters suggesting the heterogenous nature of the serogroup. This serogroup infects a wide range of animals causing different manifestations, whilst other serogroups tend to be speci®c. The absence of correlation between RAPD pattern and capsular group was not surprising. In a molecular typing study of P. multocida isolates, Wilson et al. (1993) observed that DNA ®ngerprint pro®les of 50 isolates failed to match those of somatic type reference strains. However, in a separate study, Wilson et al. (1992) had earlier shown that all 13 serogroup E strains possessed an identical DNA ®ngerprint pro®le, which did not match any of the reference strains. Observation of the identical pro®les in the serogroup E isolates appeared to depend on the restriction enzyme used. In a characterization study of Streptococcus agalactiae, Chatellier et al. (1997) observed some form of relationship between serotype and RAPD type. In addition, no relationship between RAPD pattern and biotype could be established. Perhaps using a different primer in the present work would have yielded a better RAPD pattern. Preliminary experiments showed that each of the primers generated a different pattern on ®ve reference strains used. Though some evidence suggests that these primers (ready-to-go primer kit) show complimentary activity for each other (Charlton et al., 1999), it is yet to be determined if any of the remaining primers would give a better relationship. In conclusion, the lack of relationship between genotype, host and disease was not surprising. Host speci®city is poorly de®ned for this species and a variety of diseases are differently expressed under various conditions in the same host (Bisgaard, 1993). However, porcine isolates of the subsp. multocida tended to group together (subcluster IIc) as did those of cattle origin, in particular haemorrhagic septicaemia isolates. Lastly, RAPD tended to separate subsp. septica from multocida, but clustered together the subsp. multocida and gallicida. Acknowledgements We thank Prof. J.E. Smith and Dr. U. A®ff for supplying reference capsular strains and antisera. The study was funded under The European Union University of Zimbabwe Link Project. Part of this work was presented at the HAP99 Conference, Mabula, South Africa held in September 1999. References Biberstein, E.L., Jang, S.S., Kass, P.H., Hirsh, D.C., 1991. Distribution of indole producing urease-negative pasteurellas in animals. J. Vet. Diagn. Invest. 3, 319±323. Bisgaard, M., 1993. Ecology and significance of Pasteurellaceae in animals. Zbl. Bakt. 279, 7±29. Blackall, P.J., Fegan, N., Chew, G.T.I., Hampson, D.J., 1998. Population structure and diversity of avian isolates of Pasteurella multocida from Australia. Microbiology 144, 279±289. Carter, G.R., 1955. Studies on Pasteurella multocida. I. A haemagglutination test for the identification of serological types. Am. J. Vet. Res. 16, 481±484. F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 371 Charlton, B.R., Bickford, A.A., Walker, R.L., Yamamoto, R., 1999. Complementary randomly amplified polymorphic DNA (RAPD) analysis patterns and primer sets to differentiate Mycoplasma gallisepticum strains. J. Vet. Diagn. Invest. 11, 158±161. Chatellier, S., Ramanantsoa, C., Harriau, P., Rolland, K., Rosenau, N., Quentin, R., 1997. Characterization of Streptococcus agalactiae strains by randomly amplified polymorphic DNA analysis. J. Clin. Microbiol. 35, 2573±2579. Chaslus-Dancla, E., Lesage-Descauses, M.C., Leroy-Setrin, S., Martel, J.L., Coudert, P., Lafont, J.P., 1996. Validation of random amplified DNA assays by ribotyping as tools for epidemiological surveys of Pasteurella from animals. Vet. Microbiol. 52, 91±102. Christensen, J.P., Olsen, J.E., Bisgaard, M., 1993. Ribotypes of Salmonella enterica serovar Gallinarum biovars gallinarum and pullorum. Avian Pathol. 22, 725±738. Fegan, N., Blackall, P.J., Pahoff, J.L., 1995. Phenotypic characterization of Pasteurella multocida isolates from Australian poultry. Vet. Microbiol. 47, 281±286. Heddleston, K.L., 1976. Physiological characteristics of 1268 cultures of Pasteurella multocida strains. Am. J. Vet. Res. 7, 745±747. Heddleston, K.L., Gallagher, J.E., Rebers, P.A., 1972. Fowl cholera: gel diffusion precipitin test for serotyping Pasteurella multocida from avian species. Avian Dis. 16, 925±926. Kilian, M., 1974. A rapid method for the differentiation of Haemophilus strains: the porphyrin test. Acta Pathol. Microbiol. Scand. Sect. B 82, 835±842. Kim, C.J., Nagaraja, K.V., 1990. DNA fingerprinting for differentiation of field isolates from reference vaccine strains of Pasteurella multocida in turkeys. Am. J. Vet. Res. 51, 207±210. Lam, K.M., Yamamoto, R., DaMassa, A.J., 1995. DNA diversity among isolates of Campylobacter jejuni detected by PCR-based RAPD fingerprinting. Vet. Microbiol. 45, 269±274. Lee, M.D., Mize, M.S., 1997. RAPD analysis of serotype 3, 4 isolates of Pasteurella multocida. Poult. Sci. 76 (Suppl. 1), 147. Lin, A.W., Usera, M.A., Barrett, T.J., Goldsby, R.A., 1996. Application of random amplified polymorphic DNA analysis to differentiate strains of Salmonella enteritidis. J. Clin. Microbiol. 34, 870±876. Maurer, J.J., Lee, M.D., Lobsinger, C., Brown, T., Maier, M., Thayer, S.G., 1998. Molecular typing of avian Escherichia coli isolates by random amplification of polymorphic DNA. Avian Dis. 42, 431±451. Mohan, K., Sadza, M., Madsen, M., Hill, F.W.G., Pawandiwa, A., 1994. Phenotypic characterization of Zimbabwean isolates of Pasteurella multocida. Vet. Microbiol. 38, 351±357. Mohan, K., Kelly, P.J., Hill, F.W.G., Muvavarirwa, P., Pawandiwa, A., 1997. Phenotype and serotype of Pasteurella multocida isolates from diseases of dogs and cats in Zimbabwe. Comp. Immun. Microbiol. Infect. Dis. 20, 29±34. Mutters, R., Ihm, P., Pohl, W., Frederiksen, W., Mannheim, W., 1985. Reclassification of the genus Pasteurella (Trevisan, 1887) on the basis of deoxyribonucleic acid homology, with proposals for the new species Pasteurella dagmatis, Pasteurella canis, Pasteurella stomatis, Pasteurella anatis and Pasteurella langaa. Int. J. Syst. Bacteriol. 35, 309±322. Power, E.G.M., 1996. RAPD typing in microbiology: a technical review. J. Hosp. Infect. 34, 247±265. Rhoades, K.R., Rimler, R.B., 1987. Capsular serogroups of Pasteurella multocida isolated from avian hosts. Avian Dis. 31, 895±898. Rhoades, K.R., Rimler, R.B., 1990. Somatic serotypes of Pasteurella multocida strains isolated from avian hosts (1976±1988). Avian Dis. 34, 193±195. Rimler, R.B., Brogden, K.A., 1986. Pasteurella multocida isolated from rabbits and swine: serologic types and toxin production. Am. J. Vet. Res. 47, 730±737. Rimler, R.B., Rhoades, K.R., 1987. Serogroup F, a new capsular serogroup of Pasteurella multocida. J. Clin. Microbiol. 25, 615±618. Sawada, T., Rimler, R.B., Rhoades, K.R., 1982. Indirect haemagglutination test that uses glutaraldehyde-fixed sheep erythrocytes sensitized with extract antigen for the detection of Pasteurella multocida antibody. J. Clin. Microbiol. 15, 752±756. Snipes, K.P., Hirsh, D.C., Kasten, R.W., Carpenter, T.E., Hird, D.W., McCapes, R.H., 1990. Differentiation of field isolates of Pasteurella multocida serotype 3, 4 from live vaccine strain by genotypic characterization. Avian Dis. 34, 419±424. 372 F. Dziva et al. / Veterinary Microbiology 82 (2001) 361±372 Tambic, A., Power, E.G.M., Talsania, H., Anthony, R.M., French, G.L., 1997. Analysis of an outbreak of non-phage typeable methicillin-resistant Staphylococcus aureus by using a randomly amplified polymorphic DNA assay. J. Clin. Microbiol. 35, 3092±3097. Townsend, K.M., Dawkins, H.J.S., Papadimitriou, J.M., 1997a. Analysis of haemorrhagic septicaemia-causing isolates of Pasteurella multocida by ribotyping and field alternation gel electrophoresis (FAGE). Vet. Microbiol. 57, 383±395. Townsend, K.M., Papadimitriou, J.M., Dawkins, H.J.S., 1997b. REP±PCR analysis of Pasteurella multocida isolates that cause haemorrhagic septicaemia. Res. Vet. Sci. 63, 151±155. Townsend, K.M., O'Boyle, D., Phan, T.T., Hanh, T.X., Wijewardana, T.G., Wilke, I., Trung, N.T., Frost, A.J., 1998. Acute septicaemic pasteurellosis in Vietnamese pigs. Vet. Microbiol. 63, 205±215. Wilson, M.A., Rimler, R.B., Hoffman, L.J., 1992. Comparison of DNA fingerprints and somatic serotypes of serogroup B and E Pasteurella multocida isolates. J. Clin. Microbiol. 30, 1518±1524. Wilson, M.A., Morgan, M.J., Barger, G.E., 1993. Comparison of DNA fingerprinting and serotyping for identification of avian Pasteurella multocida isolates. J. Clin. Microbiol. 31, 255±259. Zhang, Y., Rajagopalan, M., Brown, B.A., Wallace Jr., R.J., 1997. Randomly amplified polymorphic DNA PCR for comparison of Mycobacterium abscessus strains from nosocomial outbreaks. J. Clin. Microbiol. 35, 3132±3139.