Assembly of the BclB glycoprotein into the exosporium and evidence for its role in the formation of the exosporium ‘cap’ structure in Bacillus anthracis


For correspondence. E-mail; Tel. (+1) 573 884 2866; Fax (+1) 573 884 9395.


The outermost layer of the Bacillus anthracis spore consists of an exosporium comprised of an outer hair-like nap layer and an internal basal layer. A major component of the hair-like nap is the glycosylated collagen-like protein BclA. A second collagen-like protein, BclB, is also present in the exosporium. BclB possesses an N-terminal sequence that targets it to the exosporium and is similar in sequence to a cognate targeting region in BclA. BclB lacks, however, sequence similarity to the region of BclA thought to mediate attachment to the basal layer via covalent interactions with the basal layer protein BxpB. Here we demonstrate that BxpB is critical for correct localization of BclB during spore formation and that the N-terminal domains of the BclA and BclB proteins compete for BxpB-controlled assembly sites. We found that BclB is located principally in a region of the exosporium that excludes a short arc on one side of the exosporium (the so-called bottle-cap region). We also found that in bclB mutant spores, the distribution of exosporium proteins CotY and BxpB is altered, suggesting that BclB has roles in exosporium assembly. In bclB mutant spores, the distance between the exosporium and the coat, the interspace, is reduced.


The genus Bacillus is comprised of Gram-positive, endospore-producing bacteria, including the well-characterized species B. subtilis and the members of the B. cereus group, which includes the pathogens B. anthracis, B. cereus and B. thuringiensis (Kolsto et al., 2002). Endospores produced by Bacillus species can survive very harsh environmental conditions. The spore also serves as the infectious forms of the pathogenic species just mentioned. Spores of all the B. cereus group species possess an outermost layer, called the exosporium, which is not present on other species, including B. clausii, B. safensis and B. subtilis, among others (Driks, 2002; Traag et al., 2010). The exosporium is comprised of a paracrystalline basal layer which is covered by a hair-like nap layer (Gerhardt and Ribi, 1964; Hachisuka et al., 1966; Gerhardt, 1967; Beaman et al., 1971; Sylvestre et al., 2002; Lai et al., 2003; Steichen et al., 2003; Todd et al., 2003; Giorno et al., 2007; 2009). The exosporium, as the outermost layer of the spores of these species, likely plays important roles in interactions with surfaces in the soil. The exosporium also interacts with host cells during infection (Oliva et al., 2008; 2009). While the exosporium is not required for the disease anthrax in several animal models (Sylvestre et al., 2002; Giorno et al., 2007; Bozue et al., 2007a); the exosporium plays important roles in the initial course of infection when it is present (Brahmbhatt et al., 2007; Bozue et al., 2007b; Oliva et al., 2008; 2009).

The protein composition of the exosporium remains incompletely understood. The best-characterized exosporium protein is the collagen-like glycoprotein BclA, a major component of the exosporium nap layer (Sylvestre et al., 2002; 2003; Steichen et al., 2003; Daubenspeck et al., 2004). Although spores lacking BclA remain virulent (Sylvestre et al., 2002; Giorno et al., 2007; Bozue et al., 2007a), BclA has been shown to play a role in the infectious process (Bozue et al., 2007b; Brahmbhatt et al., 2007; Oliva et al., 2008; 2009).

In recent years, several additional exosporium proteins have been identified and partially characterized. These include BclB, a second collagen-like protein in B. anthracis (Waller et al., 2005; Thompson et al., 2007). BclB is surface-exposed (consistent with a location in the nap) and plays a role in exosporium formation or integrity as bclB mutant spores are affected specifically at one spore pole (Thompson et al., 2007). Consistent with the possibility that this phenotype reflects BclB's position, BclB is more abundant at one pole of the spore, and this bias in location is enhanced in bclA mutant spores (Thompson et al., 2007).

BxpB is a B. anthracis basal layer protein that has been implicated in serving as a foundation upon which nap proteins are assembled (Steichen et al., 2005; Sylvestre et al., 2005; Tan and Turnbough, 2010; Thompson et al., 2011b). A major part of the mechanism of localization of BclA, BclB and BetA (a recently identified protein also called BclF) appears to be a conserved N-terminal sequence (Thompson and Stewart, 2008; Leski et al., 2009; Tan and Turnbough, 2010; Tan et al., 2011; Thompson et al., 2011b,c). The N-terminal motif is adjacent to the extended collagen-like repeat regions that typify these proteins. A role for BxpB in BclA assembly is supported by several lines of evidence, including that they interact covalently (Tan et al., 2011) and that the assembly of each protein depends on the other (Steichen et al., 2005; Sylvestre et al., 2005; Thompson et al., 2011b). Importantly, these proteins interact via genetically separable motifs within the N-terminal conserved sequence directing the assembly and localization of BclA respectively (Thompson and Stewart, 2008; Tan and Turnbough, 2010; Thompson et al., 2011b).

In B. anthracis, the assembly of the outer exosporium layer initiates at the mother cell centre facing pole of the spore (Gerhardt and Ribi, 1964; reviewed in Henriques and Moran, 2007; Steichen et al., 2007; Thompson and Stewart, 2008; Giorno et al., 2009). This results in formation of the exosporium structure known as the cap. In the completed spore, the cap ruptures during germination to release the spore (Steichen et al., 2007). Three to five hours into sporulation, several exosporium proteins, including BclA, ExsFB and BetA, colocalize at the cap, thereby initiating exosporium formation (Thompson and Stewart, 2008; Giorno et al., 2009; Thompson et al., 2011b,c). Exosporium assembly progresses by continued deposition of exosporium proteins at the leading edge of the cap, terminating at the opposite spore pole.

In this study we monitored the appearance of BclB during sporulation and its incorporation into the spore. In addition, we documented a novel phenotype for bclB mutant spores; the interspace distance is decreased. Taken together with previous data, we propose a model to explain the previously reported fragility of the exosporium of spores lacking BclB.


Accumulation of BclB and its incorporation into the spore

To monitor the timing of the appearance of BclB during sporulation and BclB incorporation into the spore, we performed immunofluorescence microscopy (IFM) on cells from a sporulating ΔSterne culture, using anti-rBclB antibodies. BclB first became detectable in the mother cell cytoplasm faintly between hours 4 and 5 after the onset of sporulation (T4–5) (Fig. 1A, B, G and H). In Fig. 1A, the forespore chromosomes appear as intense fluorescent foci and the mother cell chromosomes are more diffusely staining and rod shaped. In Fig. 1B–F, the forespore chromosome is no longer visible due to spore maturation. Over time, the BclB-associated fluorescence became more intense and appeared more discretely localized and, by T7–8, was clearly spore-localized (Fig. 1C–E and I–K). We interpret the localization at T7–8 to represent exosporium-surface-located BclB, because the exosporium should be assembled by this time (Thompson et al., 2011b). To confirm this interpretation, we reacted mature spores that had been released from the mother cell with anti-BclB antibodies and analysed them by IFM. We found that fluorescence signals were associated with the spores, consistent with a surface location for BclB (Fig. 1F and L). The fluorescence signal did not completely encircle the spore but was present in greater amounts over approximately 75% of the spore surface and absent, or in greatly reduced amounts at one pole.

Figure 1.

Synchronous sporulating cultures of B. anthracis demonstrate the timing and pattern of BclB assembly.

A–L. Immunolabelled samples from selected time points assayed with anti-rBclB polyclonal rabbit antibodies and goat anti-rabbit Alexa Fluor 488 conjugate. Cellular DNA is labelled with DAPI.

A–F. Epifluorescence demonstrates that BclB protein expression begins from T4 to T10, with localization to the exosporium at T6. BclB is completely localized to the spore by T7.

G–L. Bright-field images of corresponding (A)–(F).

M and N. Western blots of samples obtained during synchronous sporulating cultures (M) anti-rBclB polyclonal antibodies, (N) anti-rBxpB polyclonal antibodies. Sizes of selected protein markers are indicated between (M) and (N). Blot M was stripped and reprobed with anti-rBxpB antibodies to obtain blot N.

To monitor BclB production in a potentially more sensitive manner, we also performed Western blot analysis of sporulating cell extracts after fractionation by SDS-PAGE, using anti-BclB antibodies. Consistent with the immunofluorescence microscopic analysis, we found that BclB was first detectable at T4 and increased until T7 (Fig. 1M). The apparent molecular mass of BclB is ∼ 180 kDa, which is significantly larger than the apparent size of the monomer, which is 70 kDa (based on migration of recombinant BclB on SDS-PAGE, data not shown). To compare the expression of BclB with that of a different known exosporium protein, the blot was stripped and reprobed with antibodies against the exosporium basal layer protein BxpB. We detected a heterogeneous high-molecular-weight BxpB-containing population that appeared with similar timing during sporulation to the BclB-containing band (Fig. 1N). BxpB readily assembles into higher-molecular-weight complexes. The monomer form of BxpB (migrating at ∼ 18 kDa) is usually not apparent on these blots or in past experiments (Thompson and Stewart, 2008; Thompson et al., 2011b).

Dependence of BclB incorporation on BxpB

The similar timing of expression of BclB relative to BxpB and BclA, the conserved N-terminal localization domain on BclB, and the similarity in sizes of the BclB- and BxpB-containing complexes suggested the possibility that BclB assembly depends on BclA and/or BxpB. To address this, we analysed BclB assembly in bclA and bxpB mutant spores, using IFM with anti-rBclB antibodies. We detected rings of fluorescence around wild-type spores, as expected from previous experiments (Fig. 2A and B) (Thompson et al., 2007). In contrast, bxpB mutant spores exhibited only minimal, patchy labelling (Fig. 2C and D). As previously described, bclA mutant spores had enhanced BclB labelling, except at one pole, presumably at the cap region of the spore (Fig. 2E and F) (Thompson et al., 2007). No labelling was observed with bclB mutant spores (Fig. 2G and H).

Figure 2.

The dependence of BclB on BxpB for its assembly into the exosporium.

A–H. Micrographs of spores of ΔSterne (A and B), bxpB-null (C and D), bclA-null (E and F) and bclB-null (G and H) labelled for the presence of BclB in the exosporium.

I and J. Western blot analysis of spore extracts separated on 4–20% Tris-glycine and immunoblotted for BclB (I) or BxpB (J). MW markers are denoted on the right of (J).

The results just described raised the possibility that BclB assembly requires BxpB. To address this, we extracted spore proteins, fractionated them by SDS-PAGE and performed Western blot analysis using anti-BclB antibodies. The low levels of BclB detected on bxpB mutant spores by immunofluorescence were not apparent in the Western blots of spore extracts (Fig. 2I). This may be due to poorer reactivity of the anti-BclB serum in Western blots or inefficiency of the extraction procedure. However, bclB mutant spores do possess BxpB (Fig. 2J).

Competition between BclB and BclA for BxpB binding

It is possible that BclA and BclB utilize similar mechanisms to incorporate into the exosporium as they are synthesized at similar times during sporulation, possess similar N-terminal localization/attachment domains and are both dependent on BxpB for assembly. Amino acid residues critical for BclA assembly are also present in BclB, but alignment is not exact due to the presence of additional amino acid residues (FPV and HI) in the BclB sequence (Fig. 3A). The residues required for both the covalent attachment of BclA to BxpB and associated cleavage of the BclA N-terminal 19 amino acids are not conserved in the BclB N-terminal sequence (Thompson and Stewart, 2008; Thompson et al., 2011b). In keeping with this lack of a conserved cleavage/covalent attachment domain, BclB extracted from spores contains an intact N-terminus (Waller et al., 2005).

Figure 3.

Comparison of the BclA and BclB localization and targeting motifs and the competition between them.

A. Depiction of the N-terminal protein sequences of BclA and BclB and the comparison of the localization motifs. Red letters denote essential amino acids for proper incorporation of BclA and the corresponding matches in the BclB N-terminal domain. Blue letters denote the other important residues necessary for BclA incorporation.

B–M. Epifluorescence micrographs of sporulating cells expressing the BclA NTD–eGFP fusion (B–E), the BclB NTD–mCherry fusion (J–M), or both fusions (F–I). Note the lack of BclB NTD–mCherry fusion attachment to the spores when in the presence of the BclA NTD-eGFP fusion.

N. Flow cytometric quantification of overall fluorescence of the BclB NTD–mCherry fusion-bearing spores in different host backgrounds.

To at least partially address the possibility that BclA and BclB assemble into the exosporium using a similar mechanism, we asked whether overproduction of one can affect assembly of the other, thereby suggesting competition for a common incorporation mechanism. To do this, we generated strains producing the BclA–eGFP fusion protein (MUS1900), the BclB–mCherry fusion protein (MUS1770), or both fusion proteins (MUS1831). We found that both fusions were incorporated into spores during sporulation in strains producing one or the other single fusion protein (Fig. 3B–E and J–M). However, in the MUS1831 strain producing both fusion proteins, we detected BclA–eGFP but little BclB–mCherry in spores (Fig. 3F–I). The majority of the BclB–mCherry remained in the mother cell cytoplasm. We quantified the amounts of the BclB–mCherry fusion protein incorporated by flow cytometry (Fig. 3N). In the absence of native BclA (MUS8048), there was a 37% increase in the amount of the BclB–mCherry fusion incorporated when compared with wild-type ΔSterne containing the BclB–mCherry fusion (MUS1770, Fig. 3N). As a control, the lowest amount of BclB–mCherry fusion was seen in the bxpB-null strain, a 93% reduction from wild-type levels (MUS8047, Fig. 3N). When both the BclA–eGFP fusion and BclB–mCherry fusion were expressed in wild-type ΔSterne cells (MUS1831), the BclA–eGFP dominated in its incorporation, with only 16% of the BclB–mCherry levels incorporated (Fig. 3N). These data show that under our conditions, BclA–eGFP incorporates into spores more readily than does BclB–mCherry, even though separately, each protein readily assembles into the spore. This bias in incorporation was not significantly influenced by the expression vector or fusion reporter, as a combination of pMK4-bclB1-28:mCherry and pHPS-bclA1-35:eGFP yielded the same result (data not shown). These data are consistent with the possibility that these proteins share at least some machinery for incorporation into the exosporium. We cannot exclude the possibility that the positioning of the two Bcl proteins around the spore is, in fact, comparably efficient but, because BclA becomes covalently attached and BclB does not, some of the assembled BclB dissociates during exosporium assembly.

The role of BclB in the composition of the exosporium cap and non-cap

The protein composition of the exosporium is not uniform across its surface. Steichen and colleagues showed that alanine racemase is biased towards the mother cell-distal portion of the exosporium (a region known as the ‘non-cap’) (Steichen et al., 2007). In addition, they demonstrated that spore outgrowth and germination occur at the more fragile ‘cap’ region of the exosporium (Steichen et al., 2007). To determine whether BclB has a similarly biased localization, we visualized BclB in wild-type spores using IFM and anti-BclB antibodies. We also visualized the locations of BclB–mCherry and BclB NTD–mCherry in spores using epifluorescence (in strains MUS1741 and MUS1770 respectively). These experiments showed that BclB is largely excluded from the mother cell-proximal region of the exosporium (Fig. 4A–L).

Figure 4.

Protein locations of the BclB protein in intact spores assayed via fusion expression and immunolabelling with anti-rBclB antibodies.

A–C. BclB full-length–mCherry fusion-bearing cells in late sporulation at 400×.

D–F. BclB NTD–mCherry fusion-bearing cells in late sporulation at 400×.

G–I. Wild-type spores labelled with anti-BclB antibodies followed by goat anti-rabbit Ig-Alexa Fluor 488 at 400×. Arrowheads denote the area designated as the ‘cap’ which appear as the unlabelled exosporium pole of the release spores.

J–L. Enlarged panels of wild-type free spores labelled with anti-rBclB antibodies at 600× (J) and 1000× (K and L). Arrowheads again designate the lack of BclB at one pole of the spore, likely the ‘cap’ region of the exosporium.

BclB could affect deposition of other exosporium proteins. One of these is CotY, which plays a role in exosporium formation (Johnson et al., 2006). To address this, we used fluorescence microscopy to analyse the incorporation of a CotY–mCherry fusion in strains that are otherwise wild type or which harbour a bclB mutation (strains MUS1910 and MUS1912 respectively). In otherwise wild-type cells, we found that CotY–mCherry localized to the cap (Fig. 5A–D). This is the first description of a ‘cap’ region localization of the CotY protein. In contrast, in the presence of the bclB mutation, CotY–mCherry fluorescence encompassed the majority of the exosporium (Fig. 5E and F). Complementation of the bclB mutation restored the wild-type pattern of fluorescence (Fig. 5G and H). Therefore, the presence of BclB likely affects CotY assembly. To provide more direct evidence that the locations of CotY and BclB are largely mutually exclusive, we used IFM with anti-rBclB antibodies to analyse the strain bearing the CotY fusion (Fig. 5I–L). These data showed that CotY–mCherry is located predominantly at the cap and that BclB occupies the remaining mother cell-distal portion of the exosporium (Fig. 5I–L). The effect of BclB on CotY assembly is unlikely to be due solely to an effect of BclB on CotY levels, because we found CotY levels to be similar between wild-type and bclB mutant spores, using flow cytometry (data not shown).

Figure 5.

Examination of ‘cap’ and ‘non-cap’ protein localizations shifting in the absence of BclB.

A and B. Late stage sporulating cells of B. anthracis demonstrating the concentration of CotY–mCherry fluorescence at the mother cell central ‘cap’ of the exosporium (arrowheads).

C–H. Free CotY–mCherry fusion bearing spores in wild-type (C and D), bclB-null (E and F) and complemented bclB-null backgrounds (G and H).

I–L. Wild-type spores bearing the CotY–mCherry fusions immunolabelled with anti-BclB antibodies.

M–R. Free BxpB–mCherry fusion bearing spores in wild-type (M and N), bclB-null (O and P) and complemented bclB-null spores (Q and R).

S–V. Wild-type spores bearing the BxpB–mCherry fusions immunolabelled with anti-rBclB antibodies.

Assembly of the exosporium protein basal layer protein BxpB could also be affected by BclB. To address this possibility, we used epifluorescence microscopy to visualize a BxpB–mCherry fusion in wild-type and bclB mutant strains (MUS1893 and MUS1906 respectively). In wild-type spores, we detected BxpB over the entire surface, but the intensity was greatest in the ‘non-cap’ portion of the exosporium (Fig. 5M and N). In bclB mutant spores, fluorescence was largely restricted to the non-cap pole (Fig. 5O and P). The effect of the bclB mutation was largely restored by complementation with a plasmid-borne copy of wild-type bclB (Fig. 5Q and R). To confirm that BxpB–mCherry and BclB colocalize, we performed IFM with anti-rBclB antibodies. This experiment showed the colocalization of BclB and BxpB–mCherry in individual spores (Fig. 5S–V).

Effect of bclB on the volume of the interspace

Using TEM, we found that bclB mutant spores (from strain MUS1691) have significant defects in exosporium morphology. About 40% of bclB mutant spores have severely damaged or missing exosporia (in contrast to about 5% in wild-type spores) (Thompson et al., 2007). We observed that the interspace, the region between the exosporium and the coat (Giorno et al., 2007; 2009), appeared to exhibit a smaller volume in bclB mutant spores than in wild type (Fig. 6). In addition, the percentage of spores in either strain in which the exosporium appeared to abut the coat was higher in the mutant spore population. We found that this occurred in 86% of the bclB mutant spores but in only 31% of wild-type spores. With images of 100 spores of each type examined, it is probable that this reduction in interspace volume represented a real structural difference and was not an artefact of spore orientation relative to the section plane.

Figure 6.

Transmission electron microscopy of wild-type and bclB-null spores and quantization of the area of the interspace region. TEM micrographs of (A) wild-type spores and bclB-null spores (B). Bars represent 1 μm. Data are the mean plus the standard error of the mean. P-values are from a two-tailed analysis of the data points.


In this study, we further characterized the incorporation of BclB into the exosporium, and its role in exosporium assembly. There are two major conclusions from this work. First, we found that BclA incorporates into the exosporium more readily than does BclB. Both proteins assemble into the exosporium in a BxpB-dependent manner, both possess an N-terminal region required for exosporium incorporation, and the two proteins can compete for their incorporation into the exosporium. However, the mechanisms directing BclA or BclB into the exosporium are very likely to differ, as the N-terminal sequences differ between the two proteins, and the interaction between BclA and BxpB is covalent (Tan et al., 2011), unlike the interaction between BclB and BxpB. We propose that differences between the N-termini outside of the consensus targeting sequence of BclA and BclB are largely responsible for the differences in their degree of incorporation of the two proteins. Similarly, these differences may account for the differences in effects of mutations in bclA and bclB on incorporation of other exosporium proteins. One example is the assembly of BxpB into the exosporium, which is mostly dependent on BclA, but not BclB.

The presence of additional spore proteins with similar N-terminal exosporium targeting domains may explain why incorporation of BxpB into the spore is not entirely BclA-dependent. Residual BxpB incorporation into BclA-negative spores may result from BxpB's interactions with the other targeting domain possessing proteins such as BclB and BetA. However, because the BxpB–BclA interaction is with higher affinity, BxpB incorporation in the absence of BclA is substantially reduced.

The second major conclusion from this work is that without BclB, the cap encompasses > 75% of the exosporium's circumference, instead of the usual ∼ 25%. Because we found BclB to be present largely in the non-cap region, we speculate that during exosporium assembly, BclB participates in the transition in exosporium composition that must take place as the assembly of the cap stops and the non-cap assembly begins.

To illustrate the potential implications of BclB in the cap-to-non-cap transition, we propose the following expansion of the bottle cap model of exosporium formation involving BclB (Fig. 7) (Steichen et al., 2007). CotY first localizes to the mother cell proximal pole of the spore, where it polymerizes, becoming a major component of the cap. As CotY polymerization reaches an area we call the transition zone (TZ), CotY assembly arrests. The TZ is a distinct locale on the exosporium in which both cap proteins and non-cap proteins are incorporated as the protein composition of the exosporium transitions from incorporation of cap-specific proteins to that of non-cap exosporium components.

Figure 7.

Model for the bottle-cap assembly of the exosporium.

A. Illustration demonstrating the organization of the exosporium into the ‘cap’ region (blue), found on the mother cell proximal pole of the spore, and the ‘non-cap’ region (red), found on the mother cell distal pole of the spore. Proteins that appear to segregate or concentrate in the cap region are listed in blue, and those that segregate or concentrate in the non-cap region are listed in red. The exosporium formation initiates in the cap region and assembly proceeds around the spore until it reached the transition zone (TZ, in grey). In the transition zone, levels of cap proteins diminish and levels of non-cap proteins increase. Outside of the transition zone, the non-cap proteins predominate. Certain proteins, such as BclA and BetA, are found in approximately equal distribution throughout the exosporium.

B. In the absence of BclB, assembly of the cap region of the exosporium extends beyond the normal TZ, leading to the cap region encompassing approximately 75% of the spore periphery, with a corresponding reduction of the non-cap region of the exosporium. This could result from the loss of a signal specifying the normal site of the TZ.

C. An alternative hypothesis is that in the absence of BclB, the transition zone is enlarged, and covered the entire central region of the spore. This enlarged transition zone includes proteins from both the cap and the non-cap portions of the spore.

We previously reported a cap-enhanced localization of ExsFB in the developing spore, with increased incorporation of ExsFB in a bxpB-null mutant (Thompson et al., 2011b). Based on a high degree of similarity with BxpB, ExsFB's potential binding partners may include BclA and BetA, both found in the cap. BxpB is found at reduced concentrations at the cap, relative to the non-cap portion of the exosporium. Partnering with both BxpB and ExsFB would provide a mechanism for uniform incorporation of BclA and BetA on the exosporium surface.

This study extends previous work indicating that BclB, alanine racemase (Alr), BxpB and likely ExsY predominate in the TZ and non-cap regions of the spore exosporium (Boydston et al., 2006; Steichen et al., 2007; Thompson et al., 2011b). In an exsY mutant, the exosporium assembly is defective with only a small piece of exosporium attached to the spore. The likely explanation is that in the absence of ExsY, the cap forms but the progression of assembly through the TZ into the non-cap portion of the exosporium fails to occur (Boydston et al., 2006). ExsY either is likely essential for the construction of the non-cap region or is required for the transition to non-cap assembly. BclB also appears to function in directing the transition. Interestingly, a B. cereus cotY mutant (CotY being a paralogue of ExsY) possesses an intact exosporium (although its structure may be altered), suggesting that although CotY is a predominant cap protein, in its absence ExsY can still create the cap and the non-cap portions of the exosporium (Johnson et al., 2006).

In a bclB mutant, the transition zone either is displaced to the opposite pole of the spore (Fig. 7B, Model 1), or is enlarged to encompass the midsection of the spore (Fig. 7C, Model 2). The role that BclB plays in the placement, maintenance or construction of the transition zone is unknown, as are its potential protein binding partners other than BxpB. The transition defect found in the bclB mutant spores does not involve BxpB, as bxpB-null spores still contain both a cap and non-cap region, albeit without the BclA hair-like nap layer (Steichen et al., 2005; Sylvestre et al., 2005). We have demonstrated that BxpB can interact with BclA and/or BclB depending on the availability of either partner but, in wild-type spores, BclA is the preferred choice for incorporation into the cap structure. We do not know the cause of this underlying bias, but it could be due to the binding affinities of these proteins or to other proteins involved in exosporium assembly. The placement of BclB in the TZ and non-cap regions could affect the transition zone formation by bringing in non-cap or transition proteins that recognize BclB complexes. BclB could directly interact, stabilize or bridge the connection between cap and non-cap proteins in the transition zone.

Although the BclB protein plays a role in the transition from cap to non-cap assembly of the exosporium, it does not greatly alter the overall composition of the exosporium. For example, the increased polymerization of CotY into the transition zone in the bclB mutant is accompanied by a decrease in the polymerization of CotY in the cap region of the exosporium rather than a gross change in exosporium composition. Likewise, the concentration of BxpB at the non-cap pole of the exosporium is increased in the bclB mutant, shifting the BxpB presence from the central region of the spore to an increased concentration at the non-cap pole of the spore.

The apparent functional redundancy of proteins involved in exosporium assembly may be a clue that exosporium assembly has complexities beyond those suggested by our model. For example, possibly the paralogue pairs BxpB and ExsFB, and CotY and ExsY may allow for more discrimination and control in the assembly process than is revealed by current IFM methodologies.

The interface between the cap and the non-cap regions might be the most fragile area of the exosporium, as newly germinating spores rupture the exosporium at this interface, with the removal of the cap as one of the early stages of spore outgrowth (Steichen et al., 2007). In bclB mutant spores, the increased fragility of the exosporium (Thompson et al., 2007) is likely due to TZ defects. The tendency for bclB mutant spores to rupture at one pole provides a measure of support for Model 1 of Fig. 7.

This study points to the striking complexity of exosporium assembly, and the need for a deeper understanding of its regulation. We anticipate that a detailed understanding of BclB and its interactions with other exosporium proteins will reveal important features of the mechanisms controlling formation of the cap as well as how the need for a stable outer layer is balanced with the requirement for a mechanism to liberate the newly revived cell after germination.

Experimental procedures

Strains and growth conditions

Strains utilized in this study are derivatives of B. anthracis Sterne, the veterinary vaccine strain 34F2, or ΔSterne (the pXO1-free derivative of Sterne), and are listed in Table 1. The B. anthracis plasmids have not been shown to affect exosporium structure or function. No spore-associated functions are known to be encoded on pXO2. The pXO1 plasmid harbours a class II transposon which encodes the gerX operon and a sporulation sensor histidine kinase (Guidi-Rontani et al., 1999; White et al., 2006; Scaramozzino et al., 2009). The former encodes putative membrane localized germination proteins and the latter a factor potentially affecting initial events in the sporulation process. No exosporium-specific functions are known to be plasmid-associated, no plasmid effects on exosporium structure have been observed (Ball et al., 2008; Thompson and Stewart, unpubl. data). Escherichia coli was grown at 37°C with shaking (225 r.p.m.) in Luria–Bertani broth. B. anthracis was grown at 37°C with shaking (225 r.p.m.) in brain heart infusion broth (Difco). Sporulation was induced by growth in Tiger Broth at 37°C as previously described (Thompson et al., 2011b). Tiger broth is a modified version of Mod G medium (Bergman et al., 2006; Thompson and Stewart, 2008) that permits better synchronous growth and sporulation in liquid culture. Tiger broth has the same recipe as Mod G but with the addition of 1.44 g l−1 Na2HPO4 and 0.48 g l−1 KH2PO4, pH 7.3.

Table 1. Bacterial strains and plasmids used in the study
Strain or plasmidRelevant characteristicsReference or source
  1. Cmr, chloramphenicol resistance; Eryr, erythromycin resistance; Kmr, kanamycin resistance; Spcr, spectinomycin resistance.
E. coli strains  
DH5αHost for cloning48
GM48dam-minus host for electroporation plasmid preparation41
B. anthracis strains  
SternepX01-positive, pX02-negativeLaboratory strain
ΔSternepX01-negative, pX02-negativeLaboratory strain
MUS1691ΔSterne bclB null; KmrThompson et al. (2007)
MUS1741ΔSterne pBT3955This study
MUS1770ΔSterne pBT4065This study
MUS1776ΔSterne bclA null; KmrThompson et al. (2011b)
MUS1831ΔSterne pBT4433, pBT4065This study
MUS1893ΔSterne pBT4130Thompson et al. (2011b)
MUS1900ΔSterne pBT4433Thompson et al. (2011b)
MUS1906MUS1691 pBT4130This study
MUS1910ΔSterne pBT4434This study
MUS1912MUS1691 pBT4434This study
MUS8047RG124 pBT4065This study
MUS8048MUS1776 pBT4065This study
MUS8049MUS1691 pBT4434 pBT4432This study
MUS8050MUS1691 pBT4130 pBT4432This study
RG124Sterne bxpB null; KmrGiorno et al. (2009)
pBT3955pHP13 bclB:mCherry, Cmr, EryrThis study
pBT4065pMK4 bclB1-32:mCherry, CmrThis study
pBT4130pMK4 bxpB:mCherry, CmrThompson et al. (2011b)
pBT4431pHPS bclB1-32:mCherry, Spcr EryrThis study
pBT4432pHPS bclB, complement, Spcr EryrThis study
pBT4433pHPS bclA1-35:eGFP, Spcr EryrThis study
pBT4434pMK4 cotY:mCherry, CmrThis study
pHPSE. coli–Bacillus shuttle plasmid; Eryr, SpcrThompson et al. (2011b)
pMK4E. coli–Bacillus shuttle plasmid; CmrSullivan et al. (1984)
pHP13E. coli–Bacillus shuttle plasmid; Cmr, EryrHaima et al. (1987)

Overnight cultures were grown in 5 ml of BHI broth with the addition of appropriate antibiotics. One ml of the overnight culture was used to inoculate 50 ml of pre-warmed Tiger broth cultures to achieve a starting OD600 of less than 0.1. Cultures were measured spectrophotometrically at A600 until the plot of the growth kinetics deviated from linearity (defined as T0). When required, media were supplemented with 100 μg ml−1 ampicillin, 5 μg ml−1 erythromycin, 10 μg ml−1 chloramphenicol or 100 μg ml−1 spectinomycin.

Immunolabelling of spores and vegetative cells

A total of 1 × 107 purified spores or sporulating cells (1 ml of cells at 1.3 A595) were resuspended in 500 μl of 4% paraformaldehyde in PBS and incubated for 2 h at room temperature with 300 nM DAPI (24 h at 4°C for sporulating cells). Cells or spores were washed 4× with PBS and resuspended in blocking buffer, (StartingBlock, Thermo Fisher) and incubated with mixing at room temperature for 45 min. Cells or spores were then pelleted and resuspended in StartingBlock. Rabbit polyclonal antiserum against rBclB (Thompson et al., 2007; 1:500 dilution) or rBxpB (Thompson et al., 2011b; 1:1000 dilution) was added and incubated with mixing at room temperature for 45 min. The cells/spores were then washed 3× in StartingBlock, incubated with mixing with goat anti-rabbit Alexa Fluor 488 (Invitrogen) and incubated for 45 min at room temperature. The spores were then washed 3× with StartingBlock, followed by 2× with PBS and resuspended in 10 μl of PBS containing 0.4% DABCO (Diazabicyclooctane, Acros Organics) anti-fade reagent and, finally, examined by epifluorescence microscopy using a Nikon E1000 microscope.

Western blot analysis

Protein samples for Western blot analysis were collected in two ways. To analyse entire cell populations in a sporulating cell culture, 1 ml of sporulating culture was harvested by centrifugation and resuspended in 100 μl of 8 M urea-containing SDS sample buffer (Thompson et al., 2007; 2011a) and immediately boiled for 5 min. To analyse protein extracts from purified spores, 107 spores were pelleted and resuspended in 100 μl of 8 M urea-containing SDS sample buffer and boiled for 5 min. All samples were microcentrifuged (13 000 g for 1 min) and 50 μl of extract were loaded into each well. Samples were electrophoresed in 4–12% or 4–20% Tris-glycine Fast gels (Bio-Rad). Blots were reacted with primary antibodies as follows: polyclonal anti-rBxpB (1:100 000) or polyclonal anti-BclB (1:25 000). A goat anti-rabbit HRP conjugate (Thermo Scientific) was utilized to expose the blots using a Fujifilm LAS-3000 Intelligent Dark Box imager (Fujifilm).

Creation of fusion constructs and flow cytometric analysis

Fusion constructs were created by PCR amplification of the intact ORF and the native promoter elements followed by splicing by overlapping extension to attach the ORF in-frame to the fluorescent reporter ORF (Thompson et al., 2011b). Correct fusions contained intact promoter regions including putative σK elements, as well as the native RBS and start codons. Fusion constructs were then subcloned into the shuttle plasmid pMK4 (Sullivan et al., 1984), propagated in a dam-minus E. coli host, followed by electroporation into the various strains of B. anthracis. Correct transformants were obtained by antibiotic selection, followed by DNA sequencing of plasmids isolated from vegetative cells. Additionally, when two fusions constructs were utilized, the second fusion (BclB NTD–mCherry in Fig. 4 and BclB complement in Fig. 5) was cloned into pHPS. pHPS is a derivative of pHP13 with the chloramphenicol cassette replaced by a spectinomycin cassette (Thompson et al., 2011b). pHPS and pMK4 are compatible plasmids in B. anthracis. Experiments involving both vectors in a single host were repeated with the fusions in the opposite vectors to eliminate concerns over copy number differences having effects on the results.

To perform flow-cytometry, fusion-bearing spores were prepared as above in Tiger Broth at 37°C. A total of 1 × 107 purified spores were then resuspended in 500 μl of 4% paraformaldehyde in PBS and incubated for 2 h at room temperature. The spores were then washed 3× in PBS and processed on a FACScan flow cytometer (Beckton Dickinson Biosciences) or a MoFlo XDP Flow Cytometer (Beckton Dickinson Biosciences). Data were analysed using Cell Quest or Summit 5.2 (Beckton Dickinson Biosciences) analysis software.

Transmission electron microscopy

To PBS-washed spores, 1 ml of a 2% glutaraldehyde, 0.1 M sodium cacodylate containing 0.1% ruthenium red (Electron Microscopy Sciences, Fort Washington, PA) was added and incubated for 1 h at 37°C. Each pellet was then washed in cacodylate buffer and fixed for 3 h at room temperature in a 1% osmium tetroxide (Electron Microscopy Sciences), 0.1 M sodium cacodylate solution containing 0.1% ruthenium red to allow for the visualization of the exosporium nap-like layer (Waller et al., 2004). Spores were washed in buffer and embedded in 3% agar (EM Science, Gibbstown, NJ). Dehydration was performed with sequential treatment with 25%, 50%, 75%, 95% and 100% acetone. Polymerization was carried out at 60°C in Epon/araldite resin. Electron microscopy sections were cut at 85 nm thickness and put on 200 mesh carbon-coated copper grids and then stained with a 2% uranyl acetate solution (Electron Microscopy Sciences) for 40 min at 37°C. The sections were then treated with Sato's Triple Lead for 3 min, washed in ultrapure water and stained again for 18 min in 5% uranyl acetate, followed by one final wash and were observed by transmission electron microscopy with a JEOL 1200EX electron microscope at the University of Missouri Electron Microscopy Core Facility. Statistical analysis was performed using the unpaired Student's t-test.


We thank Chris Lorson for use of his microscopy facility, Katie Thompson for assistance with the Figure 7 illustration and the University of Missouri EM Core for their assistance with the electron microscopy. Supported in part by NIH Grant AI101093 to G.S.