Polyphasic characterization of Bacillus species from anthrax outbreaks in animals from South Africa and Lesotho

Introduction: Bacillus anthracis is the causative agent of anthrax, a disease endemic in regions of Northern Cape Province and Kruger National Park of South Africa. Accurate identification of virulent B. anthracis is essential but challenging due to its close relationship with other members of B. cereus group. This study characterized B. anthracis and Bacillus species that were recovered from animals and the environment where animals died of anthrax symptoms in southern Africa using a polyphasic approach. Methodology: For this purpose, 3 B. anthracis and 10 Bacillus isolates were subjected to microbiology tests, BiologOmniLog identification system (Biolog), 16S ribosomal RNA (rRNA) sequence analysis, polymerase chain reaction (PCR) detection of protective antigen (pag) and capsule (cap) regions, and real-time PCR using hybridization probes targeting chromosomal, pag, and capC genes. Results: The Bacillus isolates were non-hemolytic, non-motile, and susceptible to penicillin, which is typical of B. anthracis, but resistant to gamma phage, unlike typical B. anthracis. The Biolog system and 16S rRNA gene sequence analysis identified most of the Bacillus isolates as B. endophyticus (7 of 10). Conventional PCR revealed that most of the Bacillus isolates contained capBCA gene regions. This highlights the limitation of the specificity of conventional PCR and the fact that the real-time PCR is more specific and reliable for anthrax diagnosis. Conclusions: Real-time PCR, 16S rRNA sequencing, and confirmatory microbiology tests including phage resistance distinguished Bacillus isolates from B. anthracis in this study. Identification of B. anthracis should be done using a polyphasic approach.


Introduction
Bacillus cereus sensu lato group comprises six members, namely B. cereus, B. anthracis, B. thuringiensis, B. mycoides, B. pseudomycoides, and B. weihenstephanensis.The B. cereus group consists of three pathogenic species, namely B. anthracis, B. cereus, and B. thuringiensis, which share highly conserved chromosomes but differ in pathogenicity, which is mostly plasmid encoded.Bacillus cereus is a foodborne pathogen due to the production of an emetic toxin, enterotoxins, and degradative enzymes [1]; B. thuringiensis is widely used in agriculture as an insect pathogen with plasmid-encoded insecticidal crystal proteins [2], while B. anthracis is a pathogen due to the presence of toxin genes.Homologous recombination and horizontal transfer of genetic material within the B. cereus sensu lato group, including phage transmission, has been reported [3,4].
Bacillus anthracis is the causative agent of anthrax that primarily affects herbivorous animals.The plasmid-encoding toxin and capsule proteins are encoded by pXO1 and pXO2, which have been found in other members of B. cereus sensu lato and are not only restricted to B. anthracis as previously thought [5,6].The pXO1 plasmid (182 kb) encodes a tripartite protein exotoxin complex and plasmid pXO2 (95 kb) encodes the polypeptide capsule genes.The toxin genes on pXO1 consist of the protective antigen (pag), lethal factor (lef), and edema factor (cya) [7].The capsule encoded by pXO2 consists of a five-gene operon (capBCADE) that synthesizes the poly-γ-D-glutamic acid capsule of Bacillus species, which enable host immune system evasion by protecting the vegetative cells from being phagocytosed by macrophages [8].
It is paramount to provide rapid and accurate diagnosis of B. anthracis to curb the spread of this zoonotic pathogen.For this purpose, B. anthracis can be distinguished from closely related B. cereus members based on criteria that are recommended by the World Health Organization (WHO) and Centers for Disease Control (CDC).Based on these criteria, B. anthracis are non-motile, non-hemolytic, and they are sensitive to penicillin and gamma phage.Nevertheless, it is imperative to make use of DNA-based methods for consistent, accurate diagnosis of anthrax due to the challenge associated with inconsistent attributes of some isolates that resemble B. anthracis.The use of 16S ribosomal RNA sequence analysis for identification of B. anthracis revealed that the genes are homologous to the B. cereus group, and the group could be considered as a single taxon [9,10].Confirmation of B. anthracis virulence genes and specific chromosomal regions can be done by polymerase chain reaction (PCR).However, PCR presents challenges for the discrimination of B. anthracis from closely related bacteria with similar capsule genes and B. anthracis virulence gene(s) [6,11].
The Agricultural Research Council-Onderstepoort Veterinary Institute (ARC-OVI) in South Africa performs classical microbiological tests and PCR targeting pXO1 and pXO2 genes for diagnosis of suspected anthrax cases or outbreaks.Numerous samples from carcasses and the environment of animals showing symptoms that resembled anthrax were received at ARC-OVI for anthrax diagnosis (Table 1, Figure 1).In this study, the accuracy of diagnostic methods, including various microbiological assays, conventional PCR, and real-time PCR to identify B. anthracis were investigated using B. anthracis and Bacillus species isolated from suspected anthrax cases in South Africa and Lesotho.

Bacillus species isolates
Soil and tissue samples were obtained during an anthrax outbreak in 2008-2009 at Klipfontein and Kimberly in the Northern Cape Province (NCP), as well as from cases in Limpopo and Mpumalanga provinces and Maseru in Lesotho (Figure 1, Table 1).The animals showed clinical symptoms such as sudden death and bleeding from natural orifices that resembled anthrax.The samples were processed at the ARC-OVI and identified.All isolates from the suspected anthrax cases were freeze-dried and stored in the bacteriology culture collection at ARC-OVI, South Africa (Table 1).Bacillus anthracis Sterne strain, B. thuringiensis (isolated from soil by ARC-OVI) and B. cereus

Microbiology tests
Bacillus isolates were grown on 5% sheep blood agar (Onderstepoort Biological Products, Pretoria, South Africa) and incubated at 37°C for 24 hours to determine hemolytic activity.The isolates were subjected to Gram stain, motility tests in thioglycollate, gamma phage, and penicillin sensitivity according to the Office International des Epizooties (OIE) [12].Capsule visualization of the Bacillus isolates was conducted by transferring purified colonies on trypticase soy agar (Selecta-MEDIA, Johannesburg, South Africa) containing 0.8% sodium bicarbonate and incubating the colonies at 5% CO2 at 37°C for 24 hours.The capsules were stained using India ink and visualized by light microscopy.Biochemical tests such as indole, citrate, oxidase, urease, and catalase were conducted using standard protocols as described by CDC/ASM/APHL [13].Bacillus anthracis Sterne strain was used as a positive control and B. cereus and B. thuringiensis strains were included as negative controls.

API assay
The Bacillus isolates were subjected to a set of carbohydrate fermentation tests based on the API 50 CHB system.The tests were undertaken according to the manufacturer's instructions (BioMerieux, Marcy l'Etoile, France) and results were interpreted using the Analytical Profile Index (API) database (Apiweb software version 1.2.1;BioMerieux, Marcy l'Etoile, France).[14]; Ramisse et al. [17]; OIE [12]; WHO [15]; Makino et al. [32]; Beyer et al. [33]; Patra et al. [34] and S. Klee (personal communication), where cap refers to capsule genes, pag refers to protective antigens, and SASP refer to small acid-soluble protein.Melt-MAMA refers to melt analysis of mismatch amplification mutation assays; 2 Indicates probe sequences; 3 Estimated between 70-90 bp, size depends on derived or ancestral PlcR region.

BiologOmniLog identification
Suspensions of Bacillus isolates were prepared according to the manufacturer's instructions (OmniLog ID System User Guide, Biolog, Hayward, USA).Briefly, the bacterial suspensions were inoculated into wells of microplates containing 95 different carbon sources.The plates were incubated in the OmniLog incubator at 30°C for 4 to 24 hours, depending on the growth requirements of the organisms.The microplates were evaluated using the Biolog's microbial identification system software (OmniLog Data Collection), which contains biochemical fingerprints of different species (Biolog, Hayward, USA).

Bacterial DNA extraction
All isolates (Table 1) were inoculated in 2 mL nutrient broth (Selecta-MEDIA, Johannesburg, South Africa) and incubated overnight at 37°C.The bacterial cells were then harvested by centrifugation at 7,500 rpm for 10 minutes.Genomic DNA was extracted from the harvested cells using the DNAeasy kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.The DNA was quantified using the Qubit fluorometric method (Life technologies, Carlsbad, Califolnia, USA) according to the manufacturer's instructions.In order to check for integrity, the DNA was electrophoresed in 0.8% ethidium bromide-stained agarose gel and visualized under UV light.
Agarose melt-mismatch amplification mutation assay (melt-MAMA) assay of the PlcR marker and the derived and ancestral controls were amplified as described by Birdsell et al. [14].The 10 μL PCR reaction contained 1x MyTaq PCR Master Mix (Bioline, Taunton, USA), 2 mM MgCl2, 0.2 µM of each primer (Table 2), and 2 ng target DNA.The PCR conditions consisted of initial denaturation at 94°C for 5 minutes, followed by 35 cycles of denaturation at 94°C for 30 seconds, annealing at 55°C for 30 seconds, extension at 72°C for 30 seconds, and a final extension at 72°C for 5 minutes.The derived and ancestral controls were amplified using the same PCR reaction and conditions as described above, except that an annealing temperature of 58°C was used.Derived and ancestral PCR amplicons were used as control templates at 1/100,000 dilution.Bacillus anthracis Sterne, Ames, and Vollum strains were used as positive controls, and B. cereus and B. thuringiensis strains were included as negative controls.The PCR amplicons were electrophoresed at 100 V for 90 minutes on a 3% ethidium bromide-stained agarose gel, and then visualized under UV light.

Real-time PCR
Real-time PCR was conducted for the detection of pag and capC regions on both virulence plasmids and a specific B. anthracis chromosomal target (small acid soluble protein, SASP) [15] using the LightCycler Nano instrument (Roche Applied Science, Mannheim, Germany).The 20 μL PCR mixture consisted of 4 mM MgCl2, 1/10 volume of FastStart Master Mix (Roche Applied Science, Mannheim, Germany), 0.5 μM of each primer, 0.2 µM of each probe, and 5 ng DNA (Table 2).Bacillus anthracis Sterne, Ames, and Vollum DNA were used as positive controls and B. cereus and B. thuringiensis strains were included as negative controls.The LightCycler experimental protocol was used as indicated by OIE experimental protocol [12], but the cutoff was after 35 cycles.The LightCycler Nano software was used to analyze the amplification of the genes.The PCR products were run on 2% agarose gels in order to confirm the amplicon sizes.

16S rRNA PCR
The 16S rRNA genes were amplified using universal primers 27F (5'-AGAGTTTGATCMTGGCTCAG-3') and 1492R (5'-CGGTTACCTTGTTACGACTT-3') [16].The 25 μL reaction contained 1x Dream Taq Green PCR Master Mix (Fermentas, Vilnius, Lithuania), 0.2 mM of each primer and 50 ng of template DNA.The PCR conditions used to amplify the 16S rRNA gene consisted of an initial denaturation at 95°C for 5 minutes, followed by 30 cycles of denaturation at 94°C for 30 seconds, annealing at 52°C for 30 seconds, and extension at 72°C for 1 minute, with a final elongation at 72°C for 7 minutes.The PCR products were electrophoresed in 1.5% ethidium bromide-stained agarose gels, followed by visualization under UV light.

Sequencing
The 16S rRNA amplicons were purified using a High Pure PCR product purification kit (Roche, Basel, Switzerland) and sequenced in both directions at Inqaba Biotechnologies (Pretoria, South Africa).All the sequences were inspected and corrected using the CLC-Bio 6.0 Workbench software.In order to obtain preliminary identifications of the Bacillus spp.isolates, the basic local alignment search tool for nucleotide sequences (BLASTN) was used to compare the sequences to those in the database of nucleotides in the National Centre for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov).
The preliminary identifications were refined through phylogenetic analyses.For this purpose, the 16S rRNA sequence alignments were generated with MEGA version 5.05 software using ClustalW multiple sequence alignment.The alignments also incorporated nucleotide sequences of other relevant Bacillus species that were obtained from GenBank.The 16S rRNA phylogenetic relationships based on maximum likelihood were inferred with MEGA version 5.05 software.Nonparametric bootstrap analysis was used to estimate branch support, and it was based on 1,000 replicates.
The capC region sequences were obtained from draft whole genome sequences of B. anthracis 20SD, B. endophyticus 3618_1C and 3631_9D (unpublished data), and the whole genome sequence of B. anthracis Ames (GenBank accession number AE017335).The capC sequence alignment was generated with CLC Genomic workbench software using classical sequence alignment analysis.

Phenotypic identification of the Bacillus species
All Bacillus isolates formed white, non-hemolytic colonies on blood agar except for Bacillus isolate 8334, which was white-grey in color.These colonies appeared circular, rough, and dry.The colonies of B. anthracis isolates (20SD, 3631_1C, and 3618_2D) were distinct from other Bacillus isolates as they were characterized by a "medusa head", which appeared as curl-like projections.The colonies of Bacillus isolates were typically Gram-positive rods showing slightly variable thickness, non-motile, sensitive to penicillin like classic B. anthracis, but in contrast to the classic anthrax bacteria, the Bacillus bacteria were not lysed by gamma phage.All isolates, including B. cereus and B. thuringiensis controls, were catalase positive, oxidase  2A) and other B. anthracis isolates (20SD, 3618_2D, and 3631_1C (Figure 2B) had long square (box)-ended rods in long chains.Bacillus isolate 3618_1C (Figure 2C) and 3617_2C (Figure 2D) had small broad rods in short chains that clustered together.
The results of the API 50 CH and BiologOmniLog identification system are summarized in Table 3. Identification at a species level is assigned as an acceptable profile with a percentage greater than 75%.The BiologOmniLog system assigned most of the Bacillus isolates to B. endophyticus, except isolates 3631_9D and 7424, which were identified as B. pumilus and B. amyloliquefaciens, respectively.The Bacillus isolate 8334 from Lesotho could not be identified using the BiologOmnilog system.

Conventional PCR and sequencing analysis
The PCR assays confirmed the presence or absence of the pXO1 and pXO2 virulence genes.The Bacillus species isolates did not contain the pagA, let, and cya regions on the pXO1 plasmid, while B. anthracis isolates (3618_2D, 3631_1C and 20SD) amplified these regions (Table 4).However, non-specific binding band was observed on isolate 3631_9D on the pagA marker using primer pair pag1/4 (data not shown).The capC gene was present in the Bacillus isolates (8334, 3618_2D, 36319D, 3631_6C, 3631_10C, 3566_1B, 3566_3D, 3617_3C, 3617_2C, 3618_1C, and 7424) and B. anthracis (20SD and 3618_2D), but not in B. anthracis isolates 3631_1C and Sterne (Table 4).The Bacillus isolates produced varying results with the different regions of the capsule genes, namely capA, capB/C, and capC (Table 4).Bacillus anthracis isolates 3618_2D and 20SD amplified all the capsule regions (capA, capB/C and capC), but B. anthracis Sterne and 3631_1C lacked these regions (Table 4).

Real-time PCR
The real-time PCR assay was used for the detection of regions specific to B. anthracis in pag (pXO1), capC (pXO2), and chromosomal SASP regions.None of the B. anthracis-specific probes hybridized with the Bacillus isolates.The probes hybridized with the B. anthracis strains (Sterne, 3618_2D, 3631_1C, and  2 for primers used to amplify the target regions and probes for hybridization, (+) indicates amplification of the gene and (-) indicates no amplification; 2 Amplification with primer pairs pag1 and pag4 as well as PA8 and BAPA-R; 3 SASP: small acid-soluble proteins in B. anthracis chromosome; 4 Identified as B. anthracis based on conventional microbiology methods.20SD), but the capC probe did not hybridize with Sterne and 3631_1C (results not shown).

Sequencing and phylogenetic analysis
The B. anthracis Sterne and B. anthracis isolates (20SD, 3631_1C, 3618_2D) that were confirmed with conventional microbiology tests had identical 16S rRNA sequences and clustered with the B. anthracis Sterne strain (Figure 3).The 16S rRNA sequence of isolate 8334 from Lesotho clustered with the B. cereus group and was closely related to B. thuringiensis IAM 12077 in this cluster.The 16S rRNA sequence of isolate 3566_3D was closely related to Brevibacterium frigoritolerans and all the other Bacillus isolates clustered with B. endophyticus (n = 8) (Figure 3).
The capC region of B. anthracis (Ames and 20SD) and possible B. endophyticus sequences (3618_1C and 3631_9D) were aligned and showed various nucleotide substitutions with no insertion or deletion (Supplementary Figure 2S).The primers 57 and 58 designed to amplify the B. anthracis capC region [17] are present in both B. anthracis and B. endophyticus and have five to seven nucleotide differences in the primer sequences (Supplementary Figure 2S).

Discussion
In this study, B. anthracis and Bacillus isolates that were isolated from diagnostic samples from suspected anthrax cases send to ARC-OVI were identified using a polyphasic approach to investigate the accuracy of these diagnostic methods.These isolates were obtained either from animals that showed symptoms reminiscent of anthrax or from soil that was beneath animals that died of anthrax symptoms.We therefore investigated the isolates using a combination of classical microbiological techniques, including API CH50B, BiologOmniLog, conventional and real-time PCR targeting various virulence genes, and 16S rRNA gene sequencing.We determined that B. anthracis can be differentiated from other similar Bacillus species using a combination of classical microbiological techniques, real-time PCR, and sequencing.The Omnilog and 16 S rRNA sequencing identified most of the Bacillus isolates as B. endophyticus.The polyphasic approach indicated that these possible B. endophyticus isolates had similar test results with B. anthracis when using selected microbiology and molecular methods.Conventional PCR revealed the presence of the capsular or polyglutamate genes in Bacillus isolates using B. anthracis primers (Table 4, Supplementary Figure 2S).This study highlights the need to determine the significance of ruling out B. endophyticus isolates from B. anthracis anthrax outbreaks for diagnostic purposes.However, it is important to note that this study did not show evidence that the Bacillus spp.were the causative agents of anthrax, as microbiological and molecular tests results indicated that these isolates did not contain toxin genes of anthrax.
The classical microbiological tests indicated that Bacillus isolates from this study differed from the classical B. anthracis, as they were resistant to gamma phage [12,18].In this study, 20SD and 3631_1C were B. anthracis, and they were sensitive to both penicillin and gamma phage.Currently, classical B. anthracis diagnosis is based on the observation of capsule, nonmotile, non-hemolytic isolates that are sensitive to penicillin and gamma phage [18].However, the resistance to gamma phage was observed in 15% of B. anthracis in previous studies [19], and in South Africa, a survey indicated that up to 16% of B. anthracis isolates from soil and carcasses were resistant to gamma phage [20].Although it is a known fact that a single diagnostic attribute may not necessarily compromise the correct identification of classical B. anthracis [5,21], the combination of various tools used in study has shown that the non-conformance of one diagnostic trait may exclude B. anthracis.All the other identifying methods (API CH50B, BiologOmnilog, and 16SrRNA sequencing) identified isolates as B. endophyticus, B. thuringiensis, and Brevibacterium frigotoleransis despite the fact that resistance to gamma phage was the only differential microbiological criteria that excluded these isolates from being B. anthracis.
The API 50 CHB identification system for Bacillus species results lacked correlation with 16S rRNA sequences and BiologOmnilog.The latter techniques identified isolates 3617_2C, 3618_1C, 2617 3C, 3566_1B, 3631_10C, and 3631_6C (n = 6) as B. endophyticus (Table 3, Figure 3).Indeed, the API 50 CHB system was shown to be inaccurate for diagnosis of B. anthracis and Bacillus isolates in this study.This was further evidenced by the low probabilities of identities obtained (Table 3) and the inability for identification of Bacillus isolates in the study.
Bacillus cereus strains were reported from chimpanzees that died of anthrax symptoms in Tai National Park, Cote d'Ivoire (CI) and a gorilla that died in (CA) as a result of B. cereus biovar anthracis [5,6].The B. anthracis virulence genes pXO1 and pXO2 were present in the B. cereus biovar anthracis CI and CA strains [5,6].The plasmids pXO1 and/or pXO2 have also been reported in other B. cereus strains [11,22].Hoffmaster et al. [23] also identified anthrax toxin genes in B. cereus G9241 that are capable of causing a severe inhalation anthrax-like disease.The genome sequence of B. cereus G9241 revealed the plasmid pBXO1, which is 99.6% similar to the pXO1 plasmid.The plasmid pXO1 genes of B. anthracis have been found to be significantly similar to the chromosomal encoded genes from the other members of the B. cereus group, B. subtilis, B. hakodurans, Listeria spp., and Staphylococcus spp.[24].The Bacillus isolates did not amplify the lethal, edema, and protective antigen regions, and therefore these bacteria did not contain the pXO1 plasmid.The capsule region amplification of Bacillus species in this study indicated the consistent presence of capC region.Whole genome sequencing would determine the presence of B. anthracis plasmid(s) in the Bacillus species as PCR only indicated the presence of similar capsule-like regions.
The Bacillus isolates tested positive for the three genes (capA, capB/C, and capC) using conventional PCR.This suggests the presence of either poly-glutamic acid or polyglutamate (PGA) or capsular genes (Table 4).The sequence alignment of the capsular gene capC shows that the possible B. endophyticus isolates 3631_9D and 3618_1C have single nucleotide variations with no insertions or deletions with B. anthracis Ames and 20SD in the capC region (Supplementary Figure 2S).This indicates the presence of the pgs/capC gene in these possible B. endophyticus strains.The capsular genes capBCADE play a role in the synthesis of the PGA capsule of B. anthracis, and these are located on the pXO2 plasmid.PGA confers the ability to evade the host immune system by protecting the vegetative cells from phagocytotic killing by macrophages [25,26].However, PGA is synthesized by many Bacillus species [27] and other organisms [25].Three B. anthracis genes, capB, capC, and capA, were shown to be necessary to drive the production of PGA weakly in Escherichia coli but could not promote PGA synthesis in B. subtilis [5].Candela et al. [25] investigated the five cap genes and found capA, capB, capC, and capE genes all necessary for sufficient PGA synthesis by B. anthracis.The capD gene encodes a ϒglutamyl transpeptidase or PGA depolymerase, which is important for the covalent anchoring of the PGA to the peptidoglycan in B. anthracis [26].Therefore, isolates that lack capD gene produce a loose slime layer of the PGA instead of a covalently linked capsule [25].Bacillus thuringiensis, B. subtilis 168 IFO3336, B. licheniformis ATCC 14580, and Staphylococcus epidermidis ATCC 12228 also contain the capB, capC, capA, and capE genes similar to B. anthracis.Annotations of these genes are named differently among the species, but they synthesize the PGA [25].In B. subtilis and B. licheniformis, they are named pgs (polyglutamate synthase) when PGA is released, which might help the bacterium to survive in high salt concentrations or confer resistance to adverse environment [26].In a study characterizing the capB, capC, capA, and capD of B. megaterium and other Bacillus species that produced PGA, the B. anthracisspecific primers could not amplify capsule regions from these isolates using PCR [27].These authors indicated that PGA-producing Bacillus species contain divergent capsule regions that cannot all be amplified using B. anthracis PCR-specific primers [27].Therefore, the Bacillus isolates from South Africa and Lesotho contained polyglutamate genes (Table 4), which may have contributed either to virulence or environmental survival.
The members of the B. cereus group are under the control of PlcR and its regulatory peptide PapR, which transcribes genes encoding collagenases, hemolysins, phospholipases, and enterotoxins [28].The PlcR gene product of B. cereus and B. thuringiensis is known to upregulate the production of numerous extracellular enzymes, while B. anthracis has a nonsense mutation that is responsible for non-motile and non-hemolytic phenotype [28].The PlcR melt-MAMA marker can be used as a specific marker for B. anthracis, because it differentiated B. anthracis strains from Bacillus isolates (Supplementary Figure 1S).
The SASPs found in B. anthracis spores were used as B. anthracis-specific protein markers [29].This marker has been found to be insufficient to discriminate amongst closely related organisms [30]; however, realtime PCR consists of confirmation of three targets, namely the SASP, and plasmid targets, which improves the specificity of the test.In this study, real-time PCR, 16S rRNA sequencing, and conventional microbiology tests accurately discriminated B. anthracis from closely related Bacillus species.However, closely related Bacillus species isolates with virulence genes that are similar to those of B. anthracis should continuously be investigated, as they might contribute towards anthraxassociated disease in animals.It is therefore paramount not to use only one test for diagnosis or identification of Bacillus species, as this may lead to false-positive results.Microbiological tests should be coupled with molecular tests for accurate diagnosis of anthrax (this study; [5,21]).The significant role of B. endophyticus in anthrax outbreaks is not clearly understood; however, for every anthrax case in the NCP, these species were isolated together with B. anthracis, which needs to be investigated further, as B. endophyticus has been isolated from vegetation as plant-endophytic bacteria [31].Whole genome sequencing will be subsequently used to obtain broader understanding of the genes that are shared between B. endophyticus strains and B. anthracis.This will further enhance the understanding of the B. endophyticus strains isolated in South Africa.

Conclusions
The BiologOmnilog system and 16S rRNA gene sequencing could identify most of the isolates as B. endophyticus.These techniques can identify Bacillus to species level (therefore differentiate B. anthracis from Bacillus species).However, the BiologOmnilog system may not be sufficient for diagnosis of Bacillus species isolates on its own, and neither of these methods would detect the virulence genes.The presence of virulence genes could be detected using PCR.However, PCR analysis provides limited information about the genetic basis or virulence of isolates, as it relies on previous known genetic sequences.Therefore, whole genome sequencing is needed to resolve the variability within species and sub-species groups that are closely related amongst the B. cereus/subtilis group.This is important for comparative analysis of the virulence genes that might be associated with anthrax symptoms.

Figure 1 .
Figure 1.South African map indicating the locations of Bacillus anthracis and Bacillus species used in this study.

Figure 3 .
Figure 3. Maximum likelihood (ML) phylogeny for the southern African Bacillus anthracis and Bacillus species isolates based on 16S rRNA sequences with other Bacillus group obtained from GenBank.Bootstrap values > 60% are indicated at the internodes.Geobacillus thermoglucosidasius and Alicyclobacillus acidocaldarius were used as outgroups to root the tree.Scale bar indicates nucleotide substitutions per site.

Table 1 .
Bacillus isolates from South Africa and Lesotho that where isolated from animals with anthrax clinical symptoms.

Table 2 .
Summary of the primer and probe sequences, gene targets, and expected PCR product sizes 1 .

Table 3 .
Identification of the Bacillus isolates by means of API 50 CHB biochemical test, BiologOmniLog system, and 16S ribosomal RNA sequencing.

Table 4 .
Summary of the PCR results of Bacillus anthracis and Bacillus isolates for pXO1, pXO2 plasmids, and/or chromosomal genes 1 .