Isolation of Bacillus anthracis from soil in selected high-risk areas of Zimbabwe


  • S.M. Chikerema,

    1. Department of Clinical Veterinary Studies, Faculty of Veterinary Science, University of Zimbabwe, Harare, Zimbabwe
    Search for more papers by this author
  • D.M. Pfukenyi,

    1. Department of Clinical Veterinary Studies, Faculty of Veterinary Science, University of Zimbabwe, Harare, Zimbabwe
    Search for more papers by this author
  • B.M. Hang'ombe,

    1. Microbiology Unit, Department of Paraclinical Studies, School of Veterinary Medicine, University of Zambia, Lusaka, Zambia
    Search for more papers by this author
  • T.M. L'Abee-Lund,

    1. Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, Oslo, Norway
    Search for more papers by this author
  • G. Matope

    Corresponding author
    1. Department of Paraclinical Veterinary Studies, Faculty of Veterinary Science, University of Zimbabwe, Harare, Zimbabwe
    • Department of Clinical Veterinary Studies, Faculty of Veterinary Science, University of Zimbabwe, Harare, Zimbabwe
    Search for more papers by this author


Gift Matope, Department of Paraclinical Veterinary Studies, Faculty of Veterinary Science, University of Zimbabwe, PO Box MP 167, Mount Pleasant, Harare, Zimbabwe. E-mail:



To isolate Bacillus anthracis from cattle carcass burial sites from high-risk districts in Zimbabwe.

Methods and Results

Soil samples were collected from carcass burial sites from seven areas, including two national game parks. Samples were collected from top 5–10 cm, and for spore extraction, 25 g of soil was suspended in sterile distilled water overnight. Supernatants were filtered through 0·45-μm pore cellulose nitrate, deposits suspended in 5 ml phosphate-buffered saline, aliquoted and heated at temperature regimen of 65, 70, 75 and 80°C for 15 min. Samples were plated onto PLET agar. B. anthracis isolates were identified using growth morphology and PCR detecting pXO1 and pXO2 virulence plasmids. From samples heated at 75°C for 15 min, B. anthracis were isolated from 9 of 81 (11·1%) soil samples representing five of the seven sampled areas.


We isolated B. anthracis from soil collected from carcass burial sites. PCR targeting virulence plasmids provided a rapid confirmation of B. anthracis.

Significance and Impact of the study

The positive isolation indicated that some carcass burial sites may retain viable spores for at least 12 months after the previous outbreak, which suggests that they may be important sources of B. anthracis and new disease outbreaks.


Anthrax caused by Bacillus anthracis is primarily a disease of herbivorous mammals, although all mammals, including humans and some avian species are susceptible (OIE 2004). The disease is an important zoonosis posing a public health problem to people in many regions of the world, particularly in Africa where it is currently recognized as a neglected zoonosis (WHO 2009). From early historical records until the development of an effective vaccine (Sterne 1937) and the later advent of effective antimicrobial drugs, B. anthracis was a significant cause of disease and mortality in livestock worldwide. The bacterium forms spores allowing it to remain viable in the environment for many years (Quinn et al. 1994) before coming into contact with a susceptible host (Dragon and Rennie 1995). Anthrax is endemic in Zimbabwe, with outbreaks reported annually in livestock in some areas. Outbreaks of the disease in game parks, affecting a wide variety of wild animal species are also not uncommon (Clegg et al. 2007). In the event of an outbreak, vaccination, carcass burial or burning and the application of calcium oxide (lime) on carcass burial sites are the main control measures. However, the lack of resources may hinder the application of lime or other disinfectants onto carcasses and burial sites, while some carcasses may not even be buried or burnt. The need to salvage meat from dead livestock may also result in opening of carcasses by the owners (Chikerema et al. 2011) resulting in further environmental contamination. The efficacy of lime as a disinfectant against B. anthracis spores has been questionable with recent evidence suggesting that exposure of the spores to lime may in fact aid their survival and viability (Himsworth 2008). A number of outbreaks in remote areas may also go unreported mainly due to lack of adequate veterinary coverage. A combination of these factors may result in various levels of environmental contamination by B. anthracis spores and partly explain the recurrence of outbreaks in some districts of the country.

Although B. anthracis is easily isolated from clinical specimens by culture on sheep blood agar and observing the typical colonial morphology and the lack of haemolysis, isolation from environmental samples can be notoriously difficult due to contamination by other micro-organisms, including closely related species such as Bacillus cereus, Bacillus mycoides and Bacillus thuringiensis (Turnbull 1999). The high and diverse microbial load of these environmental samples and the probable presence of low numbers of B. anthracis spores necessitate the use of heat treatment or ethanol shock to eliminate other vegetative bacteria and select for spores prior to plating on media (Dragon and Rennie 2001). However, even after such treatments, the isolation of B. anthracis is still difficult because of the presence of spores of other Bacillus spp. (Marston et al. 2008). Thus, the use of selective media for B. anthracis culture from soil samples is recommended, even though in some studies nonselective media like sheep blood agar have produced better results (Dragon and Rennie 2001). Marston et al. (2008) compared the selectivity of Polymyxin B, lysozyme, EDTA, thallus acetate (PLET) and a new semi-selective medium, R&F Anthracis Chromogenic Agar (ChrA; R&F Laboratories, IL, USA), and found PLET superior to ChrA. Hence, PLET is one of the most commonly used selective media for isolation of B. anthracis from environmental samples. Dragon and Rennie (2001) noted that the addition of a combination of a detergent and high specific gravity sucrose resulted in a significantly higher numbers of spores isolated from soil than deionized water, detergent or sucrose alone. For isolation from environmental samples, heat-shocking the soil samples serves the dual purpose of inducing spore germination and destroying vegetative cells of other contaminating bacteria. The heat-shock temperature combinations vary, with evidence of about 1–4 log10 reduction of B. anthracis spores at 90°C for 10 min (Turnbull et al. 2007), although this may depend on the initial bacterial load. In most cases, lower temperature and shorter heating times may result in heavy bacterial contamination during the subsequent cultivation on agar.

Despite culture and isolation being labour intensive and more time consuming as compared to molecular methods, it is still considered by many to be the most sensitive method for detection of B. anthracis from soil (Gulledge et al. 2010). While commercial PCR kits have been developed for the detection of B. anthracis spores from environmental samples (Ryu et al. 2003), their sensitivity is mainly affected by small sample sizes required for PCR, probable presence of Taq polymerase inhibitors in soil and other hydrophobic particles that may sequester spores and hamper detection of small numbers of spores (Gulledge et al. 2010). On the other hand, following culture, the identification of B. anthracis by morphology and biochemical profiles is difficult due to similarity to other Bacillus spp. (Harrell et al. 1995). Therefore, molecular tools targeting virulence genes that are encoded on plasmids pXO1 and pXO2 (Gierczynski et al. 2004) are essential for identification of B. anthracis. The purpose of the present study was to isolate B. anthracis spores from soil collected on carcass burial sites from selected high-risk districts in Zimbabwe, compare the different heat treatment protocols for isolation of Banthracis spores from soil and to use PCR to identify B. anthracis.

Materials and methods

Soil samples collection

Administrative districts in Zimbabwe have been classified into high, medium and low-risk anthrax districts according to the mean number of outbreaks per district and mean number of years of outbreak recurrence studied over a 40-year period from 1967 to 2006 (Chikerema et al. 2011). Based on this classification, five (Murewa, Chegutu, Gokwe South, Tsholotsho and Chiredzi) of the eleven high-risk districts were randomly selected for the study. In addition, two national game parks (Gonarezhou and Mana Pools) with a history of anthrax outbreaks were also selected. From March to November 2010, field visits were made to known outbreak areas in these districts. During this period, these study areas experience climatic conditions that are predominantly characterized by cool-dry (mean daily temperature range; 14–25°C) and hot-dry (26–36°C) weather from March to August and September to November, respectively. Using information obtained from veterinary extension officers, livestock owners and game rangers, soil samples were collected from carcass burial sites. Except for Mana Pools National Park in which an outbreak had occurred within the previous 6 months, all the other anthrax outbreaks were reported to have occurred at least 1 year before commencement of this study. Samples were collected from the upper 5–10 cm of the soil profile, and each sampling site was geo-referenced. We avoided sampling surface soil because it was considered not representative of the carcass burial sites due to the possibility of including silt that may have accumulated during periodic flooding in the rain season. About 100 g soil samples were collected from each site and placed in sterile tubes, which were sealed and marked for transportation to the laboratory. Protective clothing and appropriate safety measures were taken to minimize exposure to spores.

Spore extraction, induction of germination and bacterial cultures

All the procedures for B. anthracis spore extraction, culture and isolation were carried out at the University of Zimbabwe under careful precautions in a bio-safety cabinet as recommended (OIE 2004). Sterile distilled water (SDW) was used as the spore extraction solvent, essentially using the method of Dragon and Rennie (2001) and as modified by Moazeni et al. (2007). Briefly, 25 g of soil was mixed with approx. 30 ml of SDW and shaken vigorously by hand for 1 min. The sample was then transferred into a broad-based container (10 cm base diameter) to minimize the distance travelled by the spores when migrating to the surface of the mixture. The bottles were incubated overnight at room temperature (24–26°C). After overnight incubation, the supernatant was filtered through a 0·45-μm pore cellulose nitrate filter paper (Sartorius Stedim Biotech, Goettingen, Germany). The deposit on the filter paper was suspended in 5-ml sterile phosphate-buffered saline (PBS), mixed well by vortexing, aliquoted and heated at various temperature levels to kill vegetative cells and to activate spore germination.

Due to expected high levels of vegetative cell contamination of the soil samples, each sample was subjected to four alternative temperature regimen of 65, 70, 75 and 80°C for 15 min. Each of the heated suspensions was transferred into 2-ml tubes and centrifuged at 7500 g for 10 min. The supernatant was decanted, and the pellet was reconstituted with PBS and inoculated onto duplicate PLET (Sigma-Aldrich Chemie, Steinheim, Germany) agars and the plates incubated at 37°C for 48 h. A soil sample was considered positive for B. anthracis if at least a single colony was detected and positively identified by colonial and Gram-staining morphology and PCR. As the objective of the study was to isolate B. anthracis spores, further quantitative assessment of the amount of spores per gram of soil was not pursued. Suspect B. anthracis colonies were subcultured onto sheep blood agar (Oxoid, Basingstoke, Hampshire, UK) and incubated at 37°C for 48 h. The presumptive identification of B. anthracis was done using the lack of haemolysis, characteristic colonial morphology, Gram-staining reaction and spore formation. The characteristic tailing of colonies (‘Medusa-head’) was examined using the procedure of Henry, which relies on observation of colonies under low power of a stereo-microscope illuminated by an obliquely reflected light, as outlined elsewhere (Alton et al. 1988). Cell morphology and spore formation were assessed in Gram-stained smears prepared from cultures incubated aerobically at room temperature (24–26°C) as described by Cruickshank et al. (1975). The isolates were tested for motility using the hanging drop technique and further investigated for their growth pattern in semi-solid nutrient gelatine medium. All the isolates were further subjected to the PCR for confirmation.

Polymerase chain reaction

DNA was extracted using PureLink Quick Plasmid Miniprep Kit (Invitrogen, Carlsbad, CA) for plasmid DNA purification using centrifugation according to the manufacturer's protocol and essentially as described by Gierczynski et al. (2004). Briefly, a cell lysate was prepared by pelleting 3 ml of an overnight culture of the suspected B. anthracis isolates, suspending the pellet in 250 μl of Re-suspension Buffer (Qiagen, Hilden, Germany) and 250 μl of Lysis Buffer (Qiagen) prior to incubation at room temperature. Three hundred and fifty microlitre of Precipitation Buffer (Qiagen) was added, and the suspension was mixed and centrifuged at 12 000 g for 10 min and the supernatant loaded onto a spin column. Purification was done by addition of 500 μl of Wash Buffer (Qiagen) to the spin column and centrifuging at 12 000 g for 1 min. Seven hundred microlitre of Wash Buffer was then added to the column and centrifuged to remove residual wash buffer. The spin column was transferred to 1·5-ml recovery tubes with 75 μl of TE buffer and centrifuged to obtain pure DNA. The primers CAP1234F (5′-CTG AGC CAT TAA TCG ATA TG-3′) and CAP1301R (5′-TCC TAA CAC TAA CGA AGT CG-3′) were used for amplification of a 900 bp fragment and the PA5F (5′-TCC TAA CAC TAA CGA AGT CG-3′) and PA8R (5′-GAG GTA GAA TAT ACG GT-3′; Hokkaido System Science Pvt Ltd, Hokkaido, Japan) primers were used for amplification of a 586 bp fragment within the pXO2 and pXO1 virulence plasmid genes, respectively. The PCR was run separately for the pa (pXO1) and cap (pXO2) fragments in a T gradient thermocycler (Biometra, Goettingen, Germany) with a 25-μl final reaction mixture. WHO Guidelines (Turnbull 2008) were used for the PCR conditions. A B. anthracis isolate confirmed from an outbreak in Zambia (Hang'ombe et al. 2011) and water were used as positive and negative controls, respectively.

Detection of PCR products

Electrophoresis of 5 μl of each PCR reaction mixture was done in 1·5% w/v agarose gel in Tris-acetate-EDTA buffer with 0·5 μg/ml ethidium bromide at 100 V for 30 min. A 100-bp DNA ladder was used to compare the amplified fragments under ultraviolet illumination. Gels were photographed using a Kodak Easyshare digital camera (Eastman Kodak Co., Rochester, NY).


Isolation of Bacillus anthracis from soil samples

Table 1 shows the B. anthracis isolation results according to the location and the isolation protocol. Massive contamination, making it impossible to identify any suspect Banthracis colonies, was noted when using the 65 and 70°C/15 min temperature–time combinations. No growth was detected with the 80°C/15 min protocol. In contrast, nine positive isolates (11·1%) were obtained when using the 75°C/15 min protocol and the rate of isolation varied from 0 to 28·6%.

Table 1. Bacillus anthracis isolation results according to location and isolation protocol
LocationNo. of samplesIsolation protocol
65°C/15 min70°C/15 min75°C/15 min80°C/15 min
Positive (%)Positive (%)Positive (%)Positive (%)
  1. GNP, Gonarezhou National Park; MPNP, Mana Pools National Park.

MPNP7ContaminationContamination2 (28·6)No growth
Chegutu17ContaminationContamination3 (17·6)No growth
Murewa14ContaminationContamination2 (14·3)No growth
GNP10ContaminationContamination1 (10·0)No growth
Gokwe South13ContaminationContamination1 (7·8)No growth
Lupane11ContaminationContamination0No growth
Tsholotsho9ContaminationContamination0No growth
Total819 (11·1)

Identification of Bacillus anthracis

After 48 h of incubation on sheep blood agar, all the nine isolates produced colonies that were nonhaemolytic, flat, dry, greyish-white with a granular ‘ground glass’ appearance. When viewed under a low power of a stereoscopic microscope, the colonies produced some tailing and prominent wisps of growth trailing back towards the parent colony giving a characteristic ‘Medusa head’ appearance. The isolates produced Gram-positive rods in long chains, with spore formation evident in older cultures. All the nine isolates showed lack of motility and on nutrient gelatine, a characteristic ‘inverted fir tree’ growth pattern was evident.

Polymerase chain reaction

All the isolates had a positive PCR for the pa fragment (586 bp, Fig. 1) and the cap fragment (900 bp, Fig. 2). However, isolate 7 from Chegutu appeared to have a second amplicon of about 200 bp recognized by the same pa primer (Fig. 1).

Figure 1.

Photograph of the results of PCR amplification products of the pa (pXO1) fragments from nine soil Bacillus anthracis isolates. Arrow, 586 bp; 10, negative control; 11, positive control; 1 and 2, Murewa isolates; 3, 4 and 7, Chegutu isolates; 5, Gokwe south isolate; 6, Gonarezhou National Park isolate; 8 and 9, Mana Pools National Park isolates.

Figure 2.

Photograph of the results of PCR amplification products of the cap (pXO2) fragments from nine soil Bacillus anthracis isolates. Arrow, 900 bp; 10, negative control; 11, positive control; 1 and 2, Murewa isolates; 3, 4 and 7, Chegutu isolates; 5, Gokwe south isolate; 6, Gonarezhou National Park isolate; 8 and 9, Mana Pools National Park isolates.


The main objective of this study was to isolate B. anthracis from carcass burial sites in order to determine the spore contamination of soils from selected high-risk districts of Zimbabwe. Anthrax in Zimbabwe is endemic with many of the high-risk districts reporting outbreaks almost every year since 1995 when more complete data from the Department of Veterinary Services became available. It is also one of the most common zoonosis in the country where people are mainly exposed through skinning of infected animals and consumption of contaminated meat as documented by studies on outbreaks of the disease in humans (Nass 1992; Mwenye et al. 1996; Chirundu et al. 2009; Gombe et al. 2010). As evidenced by the current study where B. anthracis was isolated from some carcass burial sites of more than 1-year duration, the long-term survival of spores in the environment continues to pose a risk of exposure to livestock and wildlife once the environment becomes contaminated. For this reason, it is important to determine environmental contamination by the spores so as to enable better assessment of the risk of exposure to animals. The isolation of B. anthracis from soil samples is essential in linking cases of anthrax to their source (Marston et al. 2008). This information may be useful not only in constructing models predicting future anthrax outbreaks and their spread in the country (Chen et al. 2006), but also help in focusing of control measures on the areas identified to be contaminated with B. anthracis spores. For instance, decontamination of affected areas using formaldehyde has been done successfully in the Gruinard Island (United States National Defence Council (USNDC) 2005), and a similar programme has been initiated in Zambia (B.M. Hang'ombe, personal communication).

The challenge of attempting to isolate B. anthracis from soil samples is associated with high levels of bacterial contamination. Thus, to concentrate the spores, the soil samples are suspended in SDW overnight and because of their surface hydrophobicity (Doyle et al. 1984), the spores migrate to the surface of the solution. To minimize the distance travelled by the spores, it was necessary to incubate the samples in broad-based containers. As reported in other studies (Turnbull et al. 1998; Dragon and Rennie 2001), we used the heating method to induce spore germination where the samples were heated at temperature–time combinations that were likely to destroy most vegetative bacteria without destroying the spores. Our study suggested the optimum combination to be 75°C for 15 min. A similar combination of 80°C for 10 min has also been reported as optimum (B.M. Hang'ombe, personal communication). Although lower heating temperatures of between 63 and 65°C have been shown to be effective in recovery of spores in other studies (Turnbull et al. 1998; Dragon and Rennie 2001; Moazeni et al. 2007), our results showed that heating at 65–70°C for 15 min was not effective in significantly reducing microbial contamination in grossly contaminated soil samples. However, heating at lower temperatures such as 62–65°C for a longer duration of up to 2 h needs to be explored further.

The proportion of positive isolates (11·1%) agrees with proportions reported in other studies (Vahedi et al. 2009; Dragon and Rennie 2001). Proportion of positive isolates may be influenced by the amount of viable spores in the soil sample, which in turn is influenced by the period since the last outbreak and a combination of bioclimatic factors that influence survival of spores in soil, for example, pH, calcium content, moisture, temperature and organic matter content. For instance, high calcium levels are believed to help maintain spore vitality for prolonged periods (Dragon and Rennie 1995). The ability for these spores to remain viable in soil for prolonged periods is an important factor in the epidemiology of anthrax, particularly in animals that graze in the vicinity of these carcass burial sites and drink water from nearby ‘collector’ watering points. Besides the spore concentration in soil, contaminants make it difficult to clearly distinguish colonies of B. anthracis from other bacilli. As observed in other studies (Marston et al. 2008), we found PLET medium to be suitable for isolation of B. anthracis and to be able to minimize contamination by other closely related bacilli whose growth on PLET is mostly inconsistent (Klee et al. 2006). In contrast, other studies (Dragon and Rennie 2001) have found the use of PLET selective medium to be less satisfactory compared to nonselective media such as sheep blood agar. All our nine isolates produced growth morphology and staining characteristics that were consistent with those of B. anthracis and as expected, there was little or no variability among the isolates. B. anthracis is known to be among the bacteria with least variability, presumably due to the existence of spores that lie dormant in the environment for prolonged periods and are not prone to evolutionary pressure (Keim et al. 2004).

As the unambiguous identification of B. anthracis requires the use of a series of biochemical and/or pathogenic characteristics, we further used PCR to confirm the isolates detecting the pXO1 and pXO2 virulence markers. The pathogenic characteristics of B. anthracis are endowed on two virulence determinants; the poly-γ-d-glutamic acid capsule and the tripartite protein toxins that are encoded for by capsule (capA, capB and capC) and the toxin (pag, cya and lef) genes (Hadjinicolaou et al. 2009) on plasmids pXO2 and pXO1, respectively. Our results agree with those of Klee et al. (2006) who also recommended the use of PCR to confirm B. anthracis. However, in this study, we did not attempt to use PCR to detect B. anthracis directly in the soil samples due to cost and perceived low sensitivity. The environmental detection limit of B. anthracis in soil samples by PCR methods has been estimated to range between 0·1 and 3·2× 108 CFU/g of soil, with the limits also being influenced by the pretreatment process, location and soil type (Herzog et al. 2009). Further, the use of direct PCR methods on field soil samples suffers from the presence of compounds that may inhibit restriction enzymes and Taq polymerases (Gulledge et al. 2010), presence of hydrophobic particles that may sequester spores and obviate their detection, and the high microbial DNA load, which may mask the presence of spore DNA (Kuske et al. 1998). In addition, because the ‘genetic backbone’ virulence genes (including the cap and pagA genes) of B. anthracis may be found in other Bacillus spp. (Okinaka et al. 2006), for the detection of B. anthracis to be specific, both pXO1 and pXO2 have to be detected in the same isolate. Thus, there is no guarantee that a positive PCR carried out directly on soil samples actually detects genes residing in the same isolate. Hence, a PCR done on a pure culture is more reliable than that carried out on soil samples. To improve the recovery rate of spores in soil, Dragon and Rennie (2001) made use of a nonionic detergent to disrupt spore hydrophobic interactions and buoyant concentrations of sucrose to lift the spores. Addition of these compounds could probably have improved isolation rates. The use of PCR methods on artificially spiked soil samples has been reported to have higher detection rates than culture in other studies (Fasanella et al. 2006; Marston et al. 2008; Gulledge et al. 2010), but this has not been validated for field isolates. One of our isolates (isolate 7) produced a nonspecific band of approx. 200 bp (Fig. 1). As this band was not sequenced, it is difficult to explain its exact origin, for example, as to whether or not it was due to the deletion of the original gene in one of the copies of the pXO1 plasmid. In comparison with others, this isolate did not show any atypical morphological and biochemical profiles (data not shown).

In conclusion, the results show that despite the relatively low proportion of positive isolates, there are areas in the country contaminated with viable B. anthracis spores for at least 12 months after the last outbreak. Further sampling from high-risk areas, improvements in the isolation technique may give a better picture about spore contamination, while further molecular typing techniques could yield the details of the epidemiology of B. anthracis strains in Zimbabwe. Annual vaccination in the high-risk regions coupled with decontamination of affected areas may help reduce the incidence of anthrax cases in both animals and humans.


The authors wish to thank the Norwegian Council for Higher Education (NUFU) for their financial support without which this work would not have been possible. We also thank Dr T.C. Hodobo and Mr S.T. Marambe of the Central Veterinary Laboratory, Harare and Mr M. Madyauta for their support.

Competing Interests

The authors declare that they had no financial or personal relationships that may have inappropriately influenced them in writing this manuscript.