Characterization of some Brucella species from Zimbabwe by biochemical profiling and AMOS-PCR

Background Bovine brucellosis caused by Brucella abortus is endemic in most large commercial and smallholder cattle farms of Zimbabwe, while brucellosis in other domestic animals is rare. The diagnosis of brucellosis is mainly accomplished using serological tests. However, some Brucella spp. have been isolated from clinical cases in the field and kept in culture collection but their biochemical profiles were not documented. We report biochemical profiling and AMOS-PCR characterization of some of these field isolates of Brucella originating from both commercial and smallholder cattle farming sectors of Zimbabwe. Findings Fourteen isolates of Brucella from culture collection were typed using biochemical profiles, agglutination by monospecific antisera, susceptibility to Brucella-specific bacteriophages and by AMOS-PCR that amplifies species- specific IS711. The results of the biochemical profiles for B. abortus biovar 1 (11 isolates) and biovar 2 (2 isolates) were consistent with those of reference strains. A single isolate from a goat originating from a smallholder mixed animal farm was identified as B. melitensis biovar 1. The AMOS-PCR produced DNA products of sizes 498 bp and 731 bp for B. abortus (biovar 1 and 2) and B. melitensis biovar 1, respectively. Conclusion We concluded that the biochemical profiles and AMOS-PCR characterization were consistent with their respective species and biovars. B. abortus biovar 1 is likely to be the predominant cause of brucellosis in both commercial and smallholder cattle farms in Zimbabwe.

Studies of the genome of Brucella spp. have demonstrated the existence of more than 70% homology [6] and based on DNA-DNA hybridisation, a single species, B. melitensis was once proposed, with the other species being biovars [7]. However, the traditional classification into nine different species is still used. This has been further supported by the recent discovery of B. ceti and B. pinnipedialis from marine mammals [5], or B. microti from a common vole Microtus arvalis [4]. The genomic similarity makes the differentiation of Brucella spp. difficult, and often a study of biological and physiological characteristics is required [3].
In Zimbabwe, only B. abortus and B. melitensis have been reported to cause animal brucellosis. B. melitensis infection was confirmed in a goat flock, believed to have been translocated from Mozambique [8,9]. Brucellosis in wildlife has been demonstrated by serology [10] and in one instance B. abortus was isolated from a Cape buffalo (Syncerus caffer) [11]. Bovine brucellosis is a problem in some commercial dairy cattle farms, while others have eradicated [12]. Previous studies showed higher seroprevelance of between 10-53% in commercial herds in different regions of the country compared to 0-16% in communal (smallholder) cattle [9,12]. The disease continued to be closely monitored by the use of the milk ring test (MRT), serological surveys and bacteriological investigations [12][13][14]. Consequently, from 1988 to 2006, Brucella spp. isolates have been collected from infected herds from different parts of the country and kept in our laboratories. Although some these isolates have been identified to the species level, the details of their biochemical profiles and biovars have not been documented. The aim of this study was to characterize all Brucella field isolates in our culture collection that originated from both commercial and smallholder cattle farms of Zimbabwe using biochemical profiles and polymerase chain reaction (AMOS-PCR). The PCR assay is based on the repetitive genetic element, the insertion sequence 711 (IS711), that is unique to Brucella spp. For most Brucella spp., multiple copies of the IS711 occur at a unique species or biovar-specific chromosomal locus [15].

Materials and methods
The details of the identity and origin of all the Brucella spp. isolates, including the reference strains, used in this study are listed in Table 1. Prior to use, all strains from culture collection were stored either as lyophilized or in a -80°C deep freezer. Lyophilized isolates were re-constituted and cultured in tryptone soya broth (TSB) (Oxoid) and subsequently sub-cultured onto Farrell's medium (Oxoid) and assessed for purity on bovine blood agar (Oxoid). All plates were incubated at 37°C under 10% CO 2 . Isolates were then inoculated into TSB with 5% glycerol, frozen and exported to Norway for further characterization.
For the observation of colonial morphology, Brucella spp. isolates were cultured on Mueller-Hinton agar (Oxoid) and single colonies were examined using a low power stereoscopic microscope illuminated by obliquely reflected light as described [3]. The tests for production of urease, catalase, oxidase, H 2 S and indole; sensitivity to dyes (thionin and basic fuchsin) were carried out as described by Alton and co-workers [3]. Further tests for CO 2 requirement, sensitivity to dyes, lysis by bacteriophages (Tbilisi, Tb; Berkeley, Bk 2 ; Firenze, Fi; Izatnagar, Iz1; R/C) and agglutination by A, M and R monospecific antisera were carried out at the Central Veterinary Laboratory, UK, using the procedures described [3].
The Brucella spp. isolates were grown on Farrell's agar (Oxoid) and incubated for 48 hours at 37°C under 10% CO 2 . To yield DNA, a few colonies from a pure culture were harvested and suspended in 200 μl of sterile distilled water in Ependorf tubes. A homogeneous suspension was made by stirring with the inoculation loop. Bacterial cells were inactivated by heating the tubes at 100°C for 10 minutes on a heating block (Grant Instruments, UK). To separate the DNA, killed bacterial cells were centrifuged at 15, 700 × g for 10 minutes. The supernatant containing crude DNA template was pipetted into new sterile Ependorf tubes and the sediment discarded. The concentration of the extracted crude DNA was measured using a ND-1000 V3.0 spectrophotometer (NanoDrop ® Technologies Inc., USA). The DNA was stored at -20°C until further tests were carried out.
The AMOS-PCR (B. abortus, B. melitensis, B. ovis and B. suis-Polymerase Chain Reaction) was done as described previously [15] but with minor modifications of the assay environment. Briefly, PCR assay reaction mixture consisted of the following: 1 × PCR buffer (Applied Biosystems), 3 mM MgCl 2 , 200 μM (each) of the four deoxynucleotide triphosphates (dNTPs) (Finnzymes Oy, Espoo, Finland), and the 5 sets of primers (0.2 μM each) of B. abortus, B. melitensis, B. ovis, B. suis and IS711-specific primer. One and half units (1.5 U) of AmpliTaq Gold ® DNA polymerase (Applied Biosystems) per 45 μl reaction mixture was added and then dispensed into MicroAmp vials (Applied Biosystems). A total of 5 μl DNA template of killed bacteria (estimated at 200 ng/ml) was added per 45 μl reaction mixture. The PCR was performed with a PTC-200 Peltier Thermocycler (Roche Molecular Systems Inc, Almelda, USA). Amplification was performed for 35 cycles, each cycle comprised of denaturation at 95°C for 1 minute and 15 seconds, annealing at 60°C for 2 minutes, and extension at 72°C for 2 minutes. The PCR products were incubated for a further 5 minutes at 72°C to allow elongation of products before storage at 4°C. The PCR products were separated by electrophoresis using 1.5% agarose gel (w/v) (BDH Electran ® ) at 100 V for 1.5 hours. Gels were stained with ethidium bromide and photographed using a gene snap camera (Syngene Pvt Ltd, UK).

Results
All 14 Brucella spp. isolates characterized in this study (Table 1) yielded the following results that are typical of the genus; Gram-negative coccobacilli, non-motile, positive for modified Ziehl-Neelsen staining, oxidase and catalase production, and negative for indole production ( Table 2). Their growth on Mueller-Hinton agar produced colonies that were convex, with entire edges and a smooth shiny consistency. The morphological, growth characteristics and biochemical profiles of the field isolates were similar to their respective reference Brucella species and biovars (data not shown).
Isolates belonging to the same biovars showed consistently similar results, except for their CO 2 -dependence for growth ( Table 3). Regardless of the biovar type, seven of the 13 B. abortus isolates were CO 2 -independent, while the remaining six strains were CO 2 -dependent. The B. abortus isolates were lysed by phages Tb, Fi, Bk 2 , Iz1 and resistant to R/C. Only one isolate was lysed by the R/C phage. The single B. melitensis isolate was resistant to all phages but showed partial lysis to Bk 2 (Table 3). B. abortus isolates were agglutinated by A-antiserum and B. melitensis by the M-antiserum, but all were not agglutinated by the R-antiserum (data not shown). Eleven and two of the 13 B. abortus isolates were identified to be biovars 1 and 2, respectively ( Table 3).
The Brucella isolates were detected by the AMOS-PCR and produced predicted amplicons of sizes 498 bp and 731 bp for B. abortus and B. melitensis, respectively. Similar DNA products were produced for the reference B. abortus biovar 1 and B. melitensis biovar 1, respectively.

Discussion
This paper provides the first detailed biochemical profiling and AMOS-PCR characterization of some Brucella spp. isolates from Zimbabwe. The phage sensitivity patterns of all the Brucella spp. isolates were consistent with what has  been reported [16]. However, a single isolate of B. abortus (B15) was lysed by R/C phage and this susceptibility could be indicative of the presence of phage attachment sites which are present in the non-smooth phases of brucellae [16]. The B. melitensis (B10) also showed an atypical reaction because it was not lysed by the Iz1 phage which normally lyses smooth strains of this species. However, the examination of single colonies on microscopy by Henry illumination [3] showed no sign of dissociation and none of the isolates were agglutinated by the R-monospecific antiserum which is an indicator of dissociation. The agglutination reaction by monospecific antisera   showed the predominance of A-specific and M-specific epitopes in our B. abortus and the B. melitensis isolate, respectively. All smooth strains of Brucella may possess either the A, M or both A and M antigenic epitopes on the O chains of the lipopolysaccharides [16].
Although phage typing is used primarily for identification at the nomen species level, some Brucella strains, especially B. melitensis may show deviation from the standard pattern of susceptibility to Bk 2 , Iz1 and Wb phages [16].
The use of phage typing as a means of differentiating Brucella spp. has become less discriminatory as a typing tool because of the discovery of new strains with atypical sensitivity patterns [17].
The growth characteristics and the biochemical profiles of the field Brucella spp. isolates (Table 2) were similar to those of the reference strains used in this study. In addition, the results were consistent with what is reported for Brucella spp. and biovars [2,3,16]. However, the requirement for CO 2 for growth was at variance with reports from literature [3]. Although most strains of B. abortus biovars 1-4 require CO 2 for primary isolation, this attribute is quickly lost on repeated subcultures and such isolates will adapt to growing in atmospheres without CO 2 [3,16].
The use of the AMOS-PCR results were consistent with those reported elsewhere [15]. These results confirmed the identity of the Brucella spp. that was obtained using biochemical profiles. The IS711 analysis using AMOS-PCR can identify only three B. abortus biovars, 1, 2 and 4; all three biovars of B. melitensis; biovar 1 of B. suis and B. ovis, but the individual biovars within a species are not differentiated [15]. Therefore, further DNA fingerprinting methods such as the variable number of tandem repeat analysis (VNTR) [15] could be used to investigate the molecular epidemiology of these Brucella isolates.
Although the B. abortus used in this study originated from five of the eight geographical provinces of Zimbabwe (Table 1), it is difficult to conclude on the spatial distribution due to the limited number of isolates used. These isolates could possibly be restricted to one or a few geographical regions of Zimbabwe from where they have spread through movement of infected cattle. A study of more isolates is required to determine the spatial distribution of B. abortus in Zimbabwe. However, the predominance of B. abortus biovar 1 over biovar 2 suggested that it is the major cause of bovine brucellosis in both commercial and smallholder cattle farms. Another study which used fewer Brucella isolates from commercial dairy farms reported similar findings [12]. Although B. abortus biovar 2 was also detected and originating from both the commercial and smallholder cattle farms, its distribution could be limited to a few isolated areas. Elsewhere in South Africa biovar 1 had been reported to account for about 90% while biovar 2 contributed 10% of all the B. abortus isolates [18]. South Africa, to a large extent, shares similar geographic, climatic and livestock husbandry systems with Zimbabwe. While it is difficult to explain the reasons for the distribution of these B. abortus biovars in the cattle farming sectors, this could largely be influenced by movement of infected cattle between farms. Some farms often purchase cattle from other farms for the purpose of improving the genetics of their herds [9] or in the case of smallholder farms, to restock their herds which are continuously lost due to infectious diseases and lack of adequate grazing, especially during the drought seasons.
Despite the relatively few isolates studied, our results suggested that B. abortus biovar 3 and indeed other biovars may be rare in Zimbabwe, but this requires further study. B. abortus biovar 3 has been infrequently reported in South Africa, East and North Africa, while there seems to be no reports of isolation of the other biovars [2,18]. World wide, in countries where bovine brucellosis is endemic, B. abortus biovar 1 is predominant and B. abortus biovar 2 occurs less frequently while the other biovars are rare [2,18].

Conclusion
We concluded that the biochemical profiles and AMOS-PCR characterization were consistent with their respective species and biovars. B. abortus biovar 1 is likely to be the predominant cause of brucellosis in both commercial and smallholder cattle farms in Zimbabwe. Further studies are required that will apply DNA-fingerprinting to study distribution patterns of B. abortus biovars in Zimbabwe.
in phenotypic and molecular characterization of isolates, result interpretation, supervision of laboratory work and helped in drafting of the manuscript. All authors have read and approved the final manuscript.