The role of the exosporium in the environmental distribution of anthrax



Ezra Linley, Cardiff School of Pharmacy and Pharmaceutical Sciences, Cardiff University, King Edward VII Avenue, Cardiff CF10 3NB, UK. E-mail:



To determine the contribution of the exosporium, the outer layer of the Bacillus anthracis spore, to soil attachment. Persistence of spores in soil and their ability to infect animals has been linked to a range of factors which include the presence of organic material and calcium (OMC), pH > 6·0, temperatures above 15·5°C and cycles of local flooding which are thought to transport buried spores to the surface.

Methods and Results

The ability of wild type (exosporium +ve) and sonicated (exosporium −ve) spores to bind to soils which differed in their composition was determined using a flow-through soil column-based method. A statistically significant difference (P < 0·05) in the binding of wild type spores was observed with spores adhering more firmly to the soil with the highest OMC content. We also found that the removal of the exosporium increased the ability of the spore to adhere to both soil types.


Structures within the exosporium affected the ability of B. anthracis spores to bind to different soil types. Not surprisingly, wild type spores adhered to soil which has been shown to favour the persistence of the pathogen.

Significance and Impact of the Study

The ability to persist in and colonise the soil surface is a key requirement of a pathogen which infects grazing animals. By characterising the process involved, we will be better placed to develop strategies to disrupt the infection cycle.


Bacillus anthracis, the etiological agent of anthrax forms resistant spores which enable the organism to survive the harsh conditions encountered outside of an animal host (Baillie 2009). The structure of the spore is similar to that of other members of the Bacillus cereus group which comprises of six members: Bacillus anthracis, Bacillus cereus, Bacillus mycoides, Bacillus pseudomycoides, Bacillus thuringiensis and Bacillus weihenstephanensis (Maughan and Van der Auwera 2011). Spores produced by this group possess an additional outer layer known as the exosporium which comprises a paracrystalline basal layer surrounded by an external hair-like nap comprised of several collagen-like glycoproteins (Gerhardt and Ribi 1964; Garcia-Patrone and Tandecarz 1995; Sylvestre et al. 2003; Todd et al. 2003; Plomp et al. 2005). Appendages of about 10 nm in width have also been found attached to the exosporia of all of the strains of the B. cereus group examined to date with the exception of B. anthracis (Faille et al. 2010). While the overall contribution of this complex structure to spore survival is as yet unclear, there is evidence to suggest that it plays a role in attachment, germination and immune evasion (Redmond et al. 2004; Baillie et al. 2005; Steichen et al. 2005; Bozue et al. 2007; Weaver et al. 2007; Faille et al. 2010).

Given the ability of B. anthracis spores to persist in the environment for many years and still infected susceptible grazing animals, we were keen to determine the role, if any, of the exosporium in this process. Indeed, it has recently been suggested that under alkaline conditions, the net negative charge of the exosporium causes the spores to bind to positively charged humus and limestone particles in the soil, thereby maintain a local concentration of spores potentially lethal to grazing animals (Hugh-Jones and Blackburn 2009; Chen et al. 2010). While this is an intriguing hypothesis, a recent study by Faille and colleagues found that hydrophobicity and the number of appendages had a greater impact on the adhesion of spores of the B. cereus group than net surface charge (Faille et al. 2010). An association with limestone-rich soils may have an added advantage in that it is thought to prevent the leaching of spore calcium into the surrounding soil, thus preserving the viability of the organism (Himsworth 2008). This may explain why anthrax outbreaks are commonly associated with alkaline, calciferous soils.

It has also been suggested that the high surface hydrophobicity and buoyant density of the spores causes them to rise to the surface of the soil as a consequence of local flooding where they become accessible to grazing animals (Hugh-Jones and Blackburn 2009). This might explain in part why cattle outbreaks are associated with local flooding and could represent situations in which buried spores are brought to the soil surface by the action of water where, depending on the soil type, they adhere to soil particles such that they are able to infect susceptible animals (Turner et al. 1999).

To determine the role of the exosporium in soil binding, we compared the ability of wildtype and exosporium-deficient spores of the Sterne strain of B. anthracis to bind to calcium-rich soil.

Materials and methods

Chemicals and bacterial cultures

A culture of the Sterne strain (34F2) of B. anthracis was obtained from the culture collection of the Naval Medical Research Centre, Washington DC and was stored at −20°C prior to use. Unless otherwise stated all of the chemicals used in this study were purchased from Sigma (Sigma-Aldrich Company Ltd, Gillingham, Dorset). Microbiological culture media was purchased from Fisher (Loughborough).

The gardening soil composts used in this study are commercially available and were originally formulated by the John Innes Institute ( They comprise the following:

  • JI No.1 – seven parts loam, three parts peat and two parts sand to which 113 g John Innes Base Fertiliser and 21 g of carbonate of lime are added.

  • JI No.3 – seven parts loam, three parts peat and two parts sand, 340 g of JI base fertiliser, 64 g of carbonate of lime, 0·6 kg ground limestone, 3·6 kg hoof and horn meal (cooked ground dehydrated cattle hooves and horns), 3·6 kg superphosphate and 1·8 kg potassium sulphate.

Production of Bacillus anthracis spores

A single colony of a 24 h culture of the Sterne strain (34F2) of B. anthracis on Brain Heart Infusion agar was picked with a sterile 10 μl loop and emulsified into 9 ml of Brain Heart Infusion broth, and 5 ml of this suspension was then used to inoculate a Falcon vented flask (BD, Oxford, UK; vented cap) containing 200 ml of sterile growth medium at pH 6·7 (0·6% w/v nutrient broth, 1·2% w/v tryptone soya agar, 0·03% w/v manganese sulphate, 0·025% w/v sodium phosphate monobasic). A total of 10 flasks were inoculated per run. Each flask was incubated at 37°C for 10 days after which the spores were harvested and re-suspended into 15 ml of distilled water. The suspension was then centrifuged at 3800 g for 15 min at 4°C. The supernatant was discarded, and the pellet was re-suspended into 20 ml of 70% ethanol by agitating the tube horizontally for 60 min at 250 revolutions per min (rev min−1). The supernatant was then subjected to a further centrifugation at 3800 g for 15 min at 4°C, after which the pellet was re-suspended in 10 ml of water and heat treated at 65°C for 60 min to inactivate any vegetative organisms. Following a final centrifugation step as described previously, the spore pellet was re-suspended in 10 ml of sterile deionised water at which stage viable counts (Miles et al. 1938) and spore staining using Malachite Green and Safranin (Schaeffer and Fulton 1933) was performed to assess the concentration and purity of spores in the final suspension. Microscopic analysis confirmed that the spore preparation contained <5% vegetative cells.

Spore enumeration and viable colony counts

The number of viable spores in a sample was determined using a method based on that of Miles and Misra (Miles et al. 1938). Prior to performing counts samples were incubated at 56°C for 30 min to inactivate any vegetative bacteria. A 20 μl aliquot of heat-treated suspension was then added to 180 μl of PBS, and serial dilutions were made using a microtitre plate (Sterilin Ltd, Teddington, Middlesex, UK) until a dilution range of 10−1 to 10−10 had been achieved. Three 10-μl aliquots from each dilution were then spotted onto the surface of the predried agar plate. Once inoculated, the spots were allowed to dry at 25°C, and the plates were incubated for 48 h at 37°C. Colony forming units were then counted from each spot on the microtitre plate, and the mean number of colony forming units per microlitre (CFU ml−1) determined by multiplying the mean CFU by the dilution factor.

Removal of the exosporium by sonication

The exosporium spore layer was removed using the method of Redmond et al. (2004). A spore suspension in 50 mmol l−1 Tris/HCl, 0·5 mmol l−1 EDTA buffer (pH 7·5) containing 107 spores per ml were subjected to disruption using a Sanyo Soniprep 150 ultrasonic disintegrator for seven 1 min bursts (amplitude of 12 mm) each separated by 2 min cooling on ice. The exosporial fragments were separated from the spores by centrifuging at 9000 g for 5 min. The spore pellet was subsequently washed in 50 mmol l−1 Tris/HCl, 0·5 mmol l−1 EDTA buffer (pH 7·5) and finally re-suspended in water. The numbers of viable cells before and after sonication were determined, and the removal of exosporium was confirmed by transmission electron microscopy.

Microbial adhesion to hydrocarbon test

Spores were suspended in 9 ml of SDW to achieve an optical density of between 0·500 and 0·600 nm at OD600 (Ultrospec 1100 pro UV/Visible spectrophotometer, Biochrom, Cambridgeshire, UK). A 4-ml aliquot of spore suspension in a 50-ml Falcon container was vortex mixed with 0·4 ml of the hydrocarbon n-hexadecane at full speed for 1 min. The mixture was incubated at room temperature (22°C) for 15 min to allow the different phases to partition settle at which time the optical density of the lower aqueous phase was determined. Each assay was repeated twice. Changes in hydrophobicity were calculated as a percentage from original OD600 to the final OD600 posthexadecane exposure.

Transmission electron microscopy

Spore samples (1 ml) were centrifuged at 13 000 rev min−1 for 4 min. The supernatant was removed from the pelleted spores and were fixed with 1 ml of (2% glutaraldehyde + 2% osmium tetroxide). This was left for 1 h at 22°C, and the osmium–spore mix was pelleted and removed from the spores. BHI agar was prepared (3 g to 50 ml water), and molten agar added to the stained spore pellet and left to set. The extra agar around the spores was removed, and the agar-embedded spore pellet samples were dehydrated through an ethanol series from 50% ethanol to 100% ethanol for 10 min each. The 100% stage was repeated three times and then replaced with propylene oxide (PO) for 15 min. Epoxy resin was composed of a mixture of araldite CY212, dodecyl succinic anhydride (DDSA) and benzyl dimethylamine (BDMA). An equal quantity of PO was added to the resin and used to cover each spore pellet. The pellet was left for 36 h to allow resin infiltration. The old resin: PO mixture was subsequently removed, and pure epoxy resin was added to the samples which were then embedded into a flat moulded tray left in an oven (60°C) to harden for 48 h. Sections were cut on a microtome (Reichert-Jung, Depew, NY, USA) and stained on nickel grids with uranyl acetate and lead citrate. Images were taken on the Transmission Electron Microscope (Philips TEM 208).

Contact angle measurements

To ascertain the hydrophobic characteristics of the surfaces tested in this study, we employed a contact angle test on grade 2B finish stainless steel discs (Good fellows Cambridge Ltd, Huntingdon, UK) and agar. The angle of contact between the surface of the steel and the water was measured using a horizontal projection technique. A droplet of water was put onto the stainless steel disc, and a light was used to project the image. The projection allows a crude measurement of the contact angle between the droplet and the surface using a protractor, the hypothesis being that the contact angle between the water and a hydrophobic surface (steel) would be less than the contact angle between the water and a hydrophilic surface (agar).

Spore adherence to stainless steel

In order to assess spore adherence to stainless steel, we employed a method recently developed in our laboratory (Williams et al. 2007). A stainless steel disc 2 cm in diameter with a Grade 2B finish was inoculated with 20 μl of spore suspension (1000 spores in total) and left to dry at 37°C for 20 min. The inoculated steel discs were aseptically attached to a steel rod plunger and securely fastened into the supporting metal frame (IKA, Labortechnik, Staufen, Germany) as seen in Fig. 1. An electronic weighing balance was secured to a raising platform and positioned beneath the metal plunger. Either side of the platform, two Bunsen burners created a sterile vortex of air encompassing the platform area. A BHI agar (Oxoid, Basingstoke, Hampshire, UK) plate was placed directly under the metal plunger on top of the electronic balance which was subsequently tared. The platform and balance were raised until the steel disc compressed the agar surface with a force equal to 100 g (±5 g) and held for 10 s. The balance was lowered; the agar plate replaced with another, and the compressing process repeated a total of 15 times for each inoculated disc. The agar plates were incubated overnight at 37°C, and the number of colonies recorded for each plate. Each spore sample was assayed on three separate occasions.

Figure 1.

(a) Apparatus used for spore transfer. (b) Inoculated stainless steel disc fastened to the metal plunger. (c) BHI agar plate being compressed with an inoculated steel disc.

Spore germination in soil

To determine whether spores germinated in the test soils, 5 g of each compost type (JI No.1 and 3) were added to 10 ml PBS at pH 7 and inoculated with 3·5 ml of PBS containing 3 × 107 spores per ml of the Sterne strain of B. anthracis. The mixtures were incubated at room temperature for 1 h after which time viable counts were performed. Prior to performing counts samples were incubated at 56°C for 30 min to inactivate any vegetative bacteria. Each experiment was performed in duplicate.

Spore adherence to soil

The method employed was based on that of Riley et al. (2001). Bio-Rad Econocolumns (5 × 1·5 cm inner diameter) were packed with 5·5 g of sand [−50 + 70 mesh sand (Sigma S-9887)] and 5·5 g of soil (either JI No.1 or 3), rinsed with 15 ml PBS at pH 7 and allowed to drain under gravity. Spore suspensions of 3·5 ml containing approximately 2·0 × 105 spores per ml in PBS pH 7 were loaded onto the top of the column and allowed to pass into the matrix (3·5 ml void volume) at which point the flow was stopped. The loaded column was incubated at room temperature for 1 h to allow the spores to bind to the soil : sand matrix. The column was then eluted with 40 ml PBS, and 10 flow-through fractions were collected in 4-ml aliquots. Viable counts were performed in duplicate on each fraction, and each individual experiment was performed in triplicate. The percentage of spores which remained attached to the column was calculated by subtracting the sum of the current and previous fractions from the total added to the column following the removal of each fraction.

Statistical analysis

Statistical analyses were performed using the General Linear Model function in Minitab ver. 16 (Minitab Solutions, Coventry, UK). Post hoc comparisons were performed using Tukey's post hoc test assuming equal variances. A P value of <0·05 was considered significant.


Spore structure of wildtype and sonicated spores

To study the contribution of the exosporium to the physical properties of the spores of the Sterne strain of B. anthracis spore, we employed a previously optimised sonication protocol to physically remove the exosporium (Redmond et al. 2004). Electron microscopy studies confirmed that the sonication removed the exosporium from 95% of the spores (data not shown). As can be seen in Fig. 2, sonicated spores lack the additional layer which comprises the exosporium.

Figure 2.

Transmission Electron Microscopy images of wildtype (a) and sonicated (b) Bacillus anthracis spores showing the presence and absence of the exosporium (50 000× magnification- representative examples).

While the removal of the exosporium had no effect on the viability of the spores, it did have an effect on their appearance and on their hydrophobicity. Using a Microbial Adhesion to Hydrocarbon Test with hexadecane as the organic solvent, we observed that the removal of the exosporium markedly reduced relative hydrophobicity from 74·3% ± 1·67 (mean of three replicates ± SD) for wildtype spores to 44·4% ± 3·94. Given that the ability of a spore to adhere to a surface is influenced by its hydrophobicity (Andersson et al. 1998; Faille et al. 2010), these results suggest that the exosporium may play an important role in surface attachment.

Adhesion to stainless steel

To further characterise spore surface adhesion, we compared the ability of wildtype and exosporium-deficient spores to remain attached to the surface of a hydrophobic (>45 degree contact angle) steel disc following impression on the surface of 15 hydrophilic (0 degree contact angle) BHI agar plates. Stainless steel was selected as it has previously been used to characterise the adherence of B. cereus group spores (Faille et al. 2010; Lequette et al. 2011; Mercier-Bonin et al. 2011). As can be seen in Fig. 3, the two spore types differed (P < 0·05) in their ability to transfer from the steel onto the surface of an agar plate.

Figure 3.

A comparison of the relative ability of wild type (×) and sonicated (○) spores of the sterne strain of Bacillus anthracis to adhere to stainless steel. The results are expressed as the number of colony forming units in each fraction subtracted from the number of spores added to column as a percentage calculated by subtracting the sum of the current and previous fractions from the total added to the column. Each data point is the mean of three separate experiments. A significant difference (P < 0·05) was seen between the wildtype and sonicated spores.

Binding of Bacillus anthracis spores to soil

While the binding of spores to steel provides a useful tool with which to compare the properties of different spore variants and to compare the results generated in these studies to those of others, it bears little relevance to the natural ecology of B. anthracis which is an organism which spends the vast majority of its time in soil. For this reason, we sought to characterise the ability of spores to adhere to different commercially available garden soils.

Two soils, JI No.1 and JI No.3 which differ in their nutrient content, were examined. Prior to soil binding studies, the ability of each spore type to germinate in these soils under the test conditions employed to determine soil binding was assessed. Neither soil type supported germination (data not shown), indeed both proved to be sterile on routine culture.

To determine the effect of soil type on the retention of wild type and sonicated spores, we employed a soil column to which 2 × 105 spores per ml were added. Each soil was mixed with an equal mass of sand to improve water passage. The failure of the spores used in this study to bind to the sand was determined using soil columns packed only with sand. Spore loaded columns were subsequently washed with 40 ml of PBS at pH 7, and each sequential 4-ml fraction was collected and assayed for the presence of spores.

When we compared the ability of wild type spores to adhere to different soil types, we observed a statistically significant (P < 0·05) difference in binding with spores adhering more strongly to JI No.3 soil than to JI No.1 soil (Fig. 4). This difference between binding to soil types was eliminated when the exosporium was removed – sonicated spores bound similarly to both soil types (data not shown).

Figure 4.

Percentage binding of wild type spores to JI No.1 type soil (×) and JI No.3 type soil (○) calculated by subtracting the sum of the current and previous fractions from the total added to the column. The results show the mean of three experiments.

We then compared the ability of intact and sonicated spores to adhere to JI No.1 soil. As can be seen from Fig. 5a, sonicated spores bound more firmly (P < 0·001) than their intact counterparts. A similar pattern was observed with JI No.3 soil (Fig. 5b), although the difference was not as pronounced, it was still statistically significant (P < 0·05).

Figure 5.

Percentage binding of wild type (×) and sonicated, exosporium-deficient (○) spores of the Sterne strain of Bacillus anthracis to (a) JI No.1. type soil and (b) JI No.3 type soil calculated by subtracting the sum of the current and previous fractions from the total added to the column. The results show the mean of three experiments. Asterisks show individual points that are significantly different from one another (P < 0·05).


On the basis of published data B. anthracis spores appear to survive best in soils rich in organic matter with calcareous or gypsum-rich subsoil and above neutral pH (Smith et al. 1999; Himsworth 2008; Hugh-Jones and Blackburn 2009). The ability of spores to attach to the many components that make up these soils is thought to be influenced by a number of factors, including surface hydrophobicity and the presence of surface structures (Hugh-Jones and Blackburn 2009). This is perhaps not surprising given that hydrophobicity has been shown to play a key role in mediating the binding of Gram negative and positive bacteria to diverse surfaces such as soil mineral particles, stainless steel and human gut cells (Stenström 1989; Andersson et al. 1998; Peng et al. 2001; Faille et al. 2010). Indeed, it has been observed that the hydrophobicity of B. cereus spores, a close relative of B. anthracis which also expresses an exosporium, contributes to the ability of this organism to adhere to stainless steel and human colon carcinoma cells where they germinate and express exotoxins (Andersson et al. 1998; Peng et al. 2001; Faille et al. 2010).

To determine the role that the surface of the B. anthracis spore plays in mediating binding to stainless steel and different types of soil, we characterized the ability of wild type and exosporium-deficient spores to bind to these surfaces. Intact spores differed markedly from exosporium-deficient variants in terms of their hydrophobicity, which was reflected in a difference in their ability to adhere to stainless steel, suggesting a possible role for the exosporium in surface binding. We are not the only group to report a difference in the surface properties of exosporium-deficient B. anthracis spores. In a recent study, Chen et al. (2010) observed that the hydrophilicity of exosporium-deficient variants of the Sterne strain differed from that of wild type spores.

While binding to stainless steel provided a useful tool with which to characterise different spore variants, the main thrust of this work was to determine whether spores differed in their ability to adhere to soil. The ability of spores to bind to soil varied depending on the characteristics of the soil and the presence or absence of the exosporium. While wild type spores bound strongly to JI No.3, the soil with the highest organic and calcium content, we saw a ~25% reduction in binding to JI No.1. Interestingly, the exosporium-deficient spores bound in a similar manner to both soil types and demonstrated higher percentage retention than the wild type in both cases.

Taken together, these results suggest that structures within the exosporium affected the ability of B. anthracis spores to bind to different soil types and persist in different micro-environments. Indeed, the persistence of spores in soil and their ability to infect susceptible animals has been linked to a range of factors including the presence of organic material, calcium, a pH > 6·0, ambient temperature above 15·5°C and cycles of local flooding in which soil water transports buried spores to the surface (Dragon and Rennie 1995; Hugh-Jones and Blackburn 2009).

Thus, it is tempting to speculate that the surface hydrophobicity and buoyant density of B. anthracis spores cause them to remain suspended in soil water for extended periods. As the water table retreats deeper into the soil, the spores are transported away from the damaging effect of sunlight and become attached to subterranean soil particles under conditions similar to those found in the calcium-rich JI No.3 soil. In addition to protecting spores, soil binding could also cause an increase in local spore numbers and could be responsible for the hot spots that have been reported by some groups (Hugh-Jones and Blackburn 2009). Periods of local flooding would cause the water table to rise which would detach the spores from their refuge and transport them back to the surface to contaminate the vegetation grazed by susceptible herbivores (Hugh-Jones and Blackburn 2009). This would explain why outbreaks of infection occur in cattle which graze water meadows subject to spring flooding (Turner et al. 1999).


This work was supported by a Marie Curie International Reintegration Grants (IRG) FP7-PEOPLE-2007-4-3-IRG to Les Baillie.