INTRODUCTION
Fowl cholera is a cosmopolitan disease that affects all kinds of birds but often occurs in chickens and turkeys reared under intensive farming conditions (7). In Argentina, fowl cholera outbreaks are important in broiler breeders. The disease takes place as an acute fatal disease or most commonly, as a chronic infection. Vaccination should be considered in areas where fowl cholera is prevalent using either available bacterins and/or live vaccines, but immunization should not be substituted for good sanitary practice (7). Although these vaccines may provide some degree of protection, outbreaks in vaccinated flocks are still being reported, probably due to the lack of cross-protection among local isolates and the international recognized poultry somatic serotypes (A:1, A:3 and A:4), probably due to the local evolution of indigenous strains and the rapid adaptation of P. multocida to environmental and host changes.
Therefore, it is suggested that novel vaccines should be prepared including local or regional representative strains that should be selected for other common factors such as capsular types or virulence genes and not only according to somatic serotypes.
MATERIALS AND METHODS
Pasteurella multocida strains. A total of 101 strains of P. multocida isolated during 1980-2014 in Argentina were studied. These strains were isolated from fowl cholera outbreaks (88 strains) or from other animal species (six strains from cattle, three from sheep, two from swine, and two from rabbits).
Bacteriology. Primary species identification was done by Gram-Hucker’s staining, carbohydrate fermentation tests (4,5) (glucose, lactose, mannitol, sucrose, arabinose, sorbitol, trehalose and xylose) and conventional biochemical tests such as catalase, oxidase, and urease activity, motility, indole and hydrogen sulfide production, ONPG test and growth onto MacConkey agar .
Antibiotic resistance. All strains were examined for their susceptibility to antibiotics using the disc diffusion technique (3). All strains were grown overnight at 37°C onto Columbia Blood agar (CBA) added with 7% defibrinated bovine blood. Growth was suspended in PBS (pH 7.2) until reaching 0.5 McFarland turbidity standard. Using a sterile cotton swab, the suspension was plated onto Mueller Hinton agar plates followed by placing 3-4 discs per plate of the following antibiotics: ampicillin, chloramphenicol, colistin, enrofloxacin, streptomycin, florfenicol, gentamicin, kanamycin, neomycin, tetracycline and trimethoprim plus sulfamethoxazole (TMS). After overnight incubation at 37°C, each strain was categorized as “Susceptible” (S), “Intermediate” (I) or “Resistant” (R).
Capsular typing, virulence genes and ERIC PCR. One colony per strain was cultivated in Luria-Bertani broth and, after overnight incubation at 37°C, the broth was washed and the DNA was extracted by heating (10 min, 99ºC). Firstly, a multiplex PCR for species-specific amplification of the kmt gene and capsular typing (13) was carried out in all strains. Afterwards, another multiplex PCR assay was carried out for the detection of 4 virulence genes: dermonecrotoxin – toxA; iron acquisition system – tbpA and hgbB; and type IV fimbriae – pfhA (2). Furthermore, another PCR was done to detect Neuraminidase – nanH (4). Finally, ERIC-PCR was performed using primers ERIC1R y ERIC2 (1). Strains were considered to be identical and were allocated into the same pattern if they have 100% similarity, to the same clone if they have 90% similarity (1) or to the same cluster if they have 85% similarity (9).
Statistics. Applying the “Bionumerics” 3.5 test (Applied Maths, Kortrijk, Belgium), ERIC-PCR results were analyzed by estimating banding patterns similarity according to Dice coefficient that generated dendograms based on Unweighted Pair Group Method with Arithmatic Mean (UPGMA). The comparison of relations for P. multocida susceptibility to antibiotics was done by using UPGMA with the “R” software (10). Finally, using the Gower's General Similarity Coefficient (8) for similarity with the “R” software, an exhaustive comparison of all available data from strains: province, year of isolation, host species, biotype, virulence genes, ERIC-PCR banding pattern, and susceptibility to antibiotics.
RESULTS
All strains were identified as P. multocida: nonmotile Gram-negative rods, unable to grow onto MacConkey agar; produced catalase, oxidase and indol but did not produce urease and hydrogen sulfide; were negative to ONPG test; fermented glucose, mannitol and sucrose but did not ferment lactose. According to the differences in the fermentation of arabinose, sorbitol, trehalose, and xylose, the strains were further classified: 75 strains were allocated within subsp. multocida and 25 strains within subsp. gallicida. Biotyping was referred to Fegan et al. (5) where the subsp. multocida was distributed into four biovars: biovar one (13 strains); biovar two (16 strains); biovar three (44 strains); and biovar four (2 strains). One strain was provisionally allocated into biovar five (undefined subsp.). Furthermore, 22 strains were classified as subspecies gallicida biovar seven while two other strains were sorbitol positive variants of biovar seven and were regarded as biovar 8. Another strain was a sucrose negative variant of the newly mentioned biovar eight and consequently was considered to be biovar nine. Biovars six and 10 correspond to subsp. septica and tigris, which have not been found in the poultry strains.
Eighty P. multocida strains from poultry have a capsule type A as well as six strains from bovines, two strains from rabbits and two of the three strains from sheep. The third strain from sheep belonged to the capsule type D as well as the two strains from swine. One strain from poultry belonged to the capsule type F and it was not possible to allocate seven strains from poultry in any known capsular type.
All P. multocida strains were susceptible to florfenicol. Most of the strains were susceptible to chloramphenicol, TMS or tetracycline while more than 80% of the strains were resistant to neomycin, gentamicin or streptomycin. Susceptibility to ampicillin, colistin, enrofloxacin, and kanamycin was variable among strains. Taking in account the susceptibility results of strains, a dendrogram was constructed and the strains were grouped into nine clusters. The biggest cluster included 57 strains and four clusters included eight, nine, 10 or 12 strains. One cluster included two strains from bovines. Finally, three strains represented individually three clusters.
It was found that 73, 59, and 79 P. multocida strains from poultry were carrying pfhA, hgbB, and nanH genes, respectively. All non-poultry strains carried nanH gene. On the other hand, none of the poultry strains was carrying either toxA or tbpA gene, whereas only one strain from sheep carried toxA gene and seven strains from cattle, sheep and rabbit were carrying tbpA gene.
Fingerprinting by ERIC-PCR resulted in the presentation of between 7 and 14 bands, sized 200- 1000bp. According to the banding patterns, a dendrogram was constructed grouping these strains into 78 patterns. These patterns were further gathered into 41 clones or 25 clusters (1,9).
DISCUSSION
The use of either live vaccines or inactivated bacterins exclusively based on the international somatic serotypes (A:1, A:3 and A:4) is widely used but afford variable protection (7). The present work was performed to obtain information about similarities and differences among P. multocida strains isolated from poultry in Argentina, in order to identify suitable potential cross-protecting strains, which could be candidates for a novel vaccine.
Most poultry strains belong to the subsp. multocida commonly biotypes two and three. The subsp. multocida is very common in poultry (5,12) whereas the subsp. gallicida was barely found and the subsp. septica and tigris (11) have not been described in poultry. In this work, three novel variants of the subsp. gallicida biotype seven have been described: two strains that fermented sorbitol and sucrose and one strain only sorbitol.
Using PCR, most poultry strains were allocated as capsular type A. The single type F strain has been isolated from a free-range layer flock from a rural school; although capsular type F have been associated with fowl cholera (7,11), in this case, it is most likely that infection of these hens have been originated from an animal species different from poultry, as in this farm biosecurity measures were poor.
Comparison of susceptibility to antibiotics of all strains resulted in 45 different patterns that were grouped into 9 clusters, which had no epidemiological interpretation or meaning. Nevertheless, strains isolated from the same outbreaks were allocated into the same cluster, showing good discriminatory ability for epidemiological comparisons.
Five virulence genes were studied: pfhA, hgbB, and nanH varied among strains while none of the strains that were isolated from poultry carried either toxA or tbpA. Similar variability was previously reported among P. multocida strains from poultry (4,6).
Using ERIC-PCR, no relationship of the patterns/clones/clusters with the origin of the strains could be found. However, when strains were isolated from two outbreaks occurring at the same poultry farm, with a few years of difference, ERIC-PCR was capable to clearly differentiate the strains from each outbreak. Therefore, ERIC-PCR may be a suitable technique for studying host adaptation of P. multocida and epidemiology of fowl cholera (11).
Applying an exhaustive analysis that included all phenotypic and genotypic assays, it was not possible to establish a typical pattern to select candidate representative strains, which would be able to afford cross-protection between different P. multocida strains, due to the great variation among them.
Presented at the 67th Western Poultry Disease Conference 2018, April 16-18, Salt Lake City, UT, USA.
