SUMMARY

INTRODUCTION

Many soil-dwelling bacteria of the phylum Firmicutes utilize sporulation as a survival mechanism whereby they cease vegetative growth and form an endospore inside the mother cell cytoplasm in times of unfavorable growth conditions, such as nutrient limitation. The spore is a survival rather than a replicative mechanism, as only a single spore develops within the cytoplasm of the sporulating cell (an endospore within a mother cell). The spore is released from the mother cell by lysis, and the mature spore exists as a dormant form of the cell that lacks metabolic activity (1, 2). Spores can persist in the environment for long periods.

Spore components include a core consisting of a cell with a dehydrated cytoplasm bearing high concentrations of calcium dipicolinate (Fig. 1). The core is enveloped in a cortex layer consisting of a modified peptidoglycan. External to the cortex is a proteinaceous shell referred to as the spore coat, often consisting of morphologically distinct inner and outer coat layers. The Bacillus subtilis coat comprises about 70 different proteins (3). Spores are highly refractive to environmental insults, such as desiccation, UV irradiation, heat, and exposure to organic solvents. When favorable growth conditions return, spores are stimulated to undergo a process of germination and outgrowth whereby the dormant spore takes up water, releases calcium ions, resumes metabolic activity, and degrades the cortex and the vegetative cell emerges from the spore coat shell. Spore germination involves signaling from small-molecule germinants that diffuse across the outer spore layers and bind to receptors in the cytoplasmic membrane (4).

Open in a separate window

FIG 1

Transmission electron micrograph of a B. anthracis spore.

Endospore-producing bacteria have substantial impacts on human and animal health. Some are human and/or animal pathogens, whereas others affect humans more indirectly, such as insect pathogens or agents of food or product spoilage. Bacillus anthracis is an important zoonotic pathogen that causes anthrax, a fatal septicemia in ruminants. Human infections result in cutaneous, gastrointestinal, or pulmonary anthrax, depending on the site of spore entry into the patient. Gastrointestinal and pulmonary anthraces are associated with high mortality rates, and anthrax spores are a concern as a weapon of bioterrorism (5). Bacillus cereus is a common cause of food poisoning (6). It is also an invasive pathogen responsible for systemic infections, including endophthalmitis, bloodstream infections, endocarditis, and respiratory tract infections (7). Severe and lethal B. cereus infections have been reported for newborn infants, especially among those born prematurely (8, 9). B. cereus infections in newborns include involvement of the central nervous system or respiratory tract, primary bacteremia, and sepsis. Bacillus thuringiensis, a species closely related to B. anthracis and B. cereus, is an insect pathogen widely utilized in agriculture as a bioinsecticide (10). Sporulating cells of B. thuringiensis produce a crystal toxin that is poisonous to insects in the orders Lepidoptera, Coleoptera, and Diptera (11).

The anaerobic spore former Clostridium difficile is the causal agent of nosocomial antibiotic-associated diarrhea and pseudomembranous colitis and is an important emerging public health concern (12,–14). Infection with C. difficile recently surpassed methicillin-resistant Staphylococcus aureus infection as the most common hospital-acquired infection in the United States (15). It is estimated that nosocomial C. difficile infections increase the cost of hospitalization 4-fold compared with matched cohorts, and these infections have been estimated to result in $4.8 billion in excess costs in U.S. acute-care facilities (12). Additionally, 20 to 27% of all C. difficile cases are community associated, with an incidence of 20 to 30 per 100,000 population (16). Clostridium perfringens is one of the most common causes of human foodborne illness, with an estimated million cases of foodborne illness each year in the United States (17, 18). C. perfringens also causes infections of skin and soft tissue, gastroenteritis, gas gangrene, necrotizing enteritis, liver abscess, bacteremia, endophthalmitis, and septic shock with acute hemolysis (19,–21). Clostridia are also the second most common cause of anaerobic bacteremia, and the 30-day mortality rate for infected patients is 27 to 44% (19).

Botulism results from exposure to a potent neurotoxin produced by the spore-forming bacterium Clostridium botulinum (22,–24). There are five main kinds of botulism. Foodborne botulism is caused by eating foods that contain the botulinum toxin. Canned (especially home-canned) and vacuum-packed foods, which provide the anaerobic environment necessary for contaminating spores to germinate and the vegetative cells to produce the toxin, are the foods usually implicated. Wound botulism is caused by wound contamination with C. botulinum spores, and currently the majority of these cases are associated with black-tar heroin injection. Infant botulism is caused by consumption of the spores of the botulinum bacteria, which then grow in the intestines and release toxin in children under the age of 1 year. Adult intestinal toxemia (adult intestinal colonization) botulism is a very rare form of botulism that occurs among adults by the same route as infant botulism. Iatrogenic botulism can occur from accidental overdose of botulinum toxin in medical or cosmetic product formulations. In the United States, an average of 145 cases of botulism are reported each year, with approximately 15% being foodborne, 65% being infant botulism, and 20% being wound associated (22).

The spore-forming process of B. subtilis is the best studied of such processes (25). Early in sporulation, the cell begins to divide asymmetrically, and the division septum undergoes remodeling, resulting in engulfment of the smaller cell compartment. This results in the formation of a smaller, double-membrane-bound forespore compartment within the mother cell. During spore maturation, the mother cell directs the synthesis and assembly of the outermost spore layers, including spore coat proteins and, depending on the bacterial species, additional outer layers, including the exosporium.

THE EXOSPORIUM LAYER OF BACTERIAL SPORES

Under higher magnification, every spore appears to be of egg-like shape and embedded within a ball-shaped glass-like mass looking like a light, narrow ring surrounding the spores.

—Robert Koch (26)

In his classic 1876 paper on the etiology of anthrax, Robert Koch first described what is now known as the exosporium layer of the Bacillus anthracis spore (26). The term “exosporium” was later coined by Flügge in 1886 (27). Bacterial endospores can be placed into two categories: those containing a distinct exosporium and those lacking this structure. The best studied of the former group are those of B. anthracis and B. cereus. Bacillus subtilis spores are the best-studied example of the exosporium-less category. Much of what we know regarding spore assembly has been learned from genetic and biochemical studies with B. subtilis. Consequently, there is a wealth of knowledge regarding initiation of sporulation through spore coat assembly. Because B. subtilis lacks an exosporium, information regarding exosporial composition and its assembly process is considerably more limited.

The exosporium of the B. cereus family of bacteria consists of a basal layer surrounded by an external nap of hairlike projections (28,–30) (Fig. 1). The collagen-like glycoprotein BclA (Bacillus collagen-like protein A) is the principal component of this hairlike nap and is an immunodominant spore antigen (31,–33). The exosporium basal layer is a shell composed of a number of different proteins (33,–35).

Advances in electron microscopic imaging techniques have led to spectacular improvements in our understanding of the exosporium architecture of spores from the B. cereus family (36,–39). The basic structures of the exosporium are similar for B. anthracis, B. cereus, and B. thuringiensis. The balloon-like exosporium is deformable and irregularly shaped and possesses numerous folds and creases. The exosporium is separated from the spore coat by extensive distances (for example, 500 nm in B. anthracis [39]) yet lies in close proximity to the spore coat at other sites around the spore. The region between the basal layer of the exosporium and the outer spore coat is referred to as the interspace (35). The basal layer of the B. anthracis exosporium is approximately 12 to 16 nm thick and appears to be comprised of two, approximately 5-nm-thick sublayers (39). The basal layer has a crystalline structural organization with a 6-fold symmetry and a periodic spacing of 7 nm. The external face of this structure consists of a series of hexagonally packed concave cups in a honeycomb pattern, with the open ends oriented toward the external environment (Fig. 2) (38). The spore core-facing side of this structure consists of a honeycomb pattern of hexameric crowns with a diameter of ~7.5 to 9 nm (Fig. 2). The open ends of the crowns abut one another to form channels between the two faces of the basal layer. The channels have a diameter of ~20 to 34 Å (36, 38). A channel of this size is sufficiently large to permit entry of small-molecule germinants, such as alanine or inosine, but is too small to permit diffusion of larger molecules, such as proteins. The measured channel size is consistent with the results of spore permeation studies (40). This channel arrangement confers the semipermeable barrier properties of the exosporium.

Open in a separate window

FIG 2

Schematic diagram illustrating a model for the exosporium of the B. cereus family. The exosporium basal layer consists of a two-dimensional lattice of cups opening out to the environment on the convex outer face of the layer. There are channels penetrating the basal layer. The collagen-like fibrils of the BclA nap are shown schematically attached to the convex face, terminating in the globular tumor necrosis factor-like C-terminal domain. (Reprinted from reference 38 with permission of the publisher.)

The BclA nap filaments of the B. anthracis spore span 14 to 70 nm (32, 39) and are attached to the outer surface of the basal layer. The length of the filaments is a function of the number of collagen-like triplet amino acid repeats in the BclA protein (32). The hairlike nap covers the entire surface of the exosporium, with an interfilament spacing of approximately 7 nm, the same spacing periodicity as the basal layer crystalline array. The trimeric BclA filaments are oriented with their N termini at the basal layer and their C-terminal domains appearing as a 6-nm knob at the end of the outward-extending filaments.

The exosporium is anchored to the spore coat layer via a number of protein-protein interactions. These have been identified through the isolation of mutants that fail to stably attach the exosporium and from which, as a consequence, the exosporium is released as sheets separate from the mature spores upon lysis of the mother cells. Proteins implicated in exosporium attachment include the spore coat protein CotE and the exosporium- or interspace-localized proteins CotY, ExsA, ExsB, ExsY, and ExsM (1, 41,–45). Whether these proteins participate directly in exosporium attachment or indirectly, through proper positioning of proteins actually involved in the attachment, has not been established. CotE is likely to be the spore coat-associated part of the connector between the spore coat and the exosporium. ExsA and/or ExsB may also be part of this connection chain.

THE SPORE CRUST LAYER

B. subtilis lacks an obvious exosporium layer and has the outer spore coat layer as its surface. This is consistent with the finding that known B. cereus family exosporium genes are not present in the B. subtilis genome. However, the surfaces of B. subtilis spores may not be that dissimilar from those of B. cereus family spores. Treatment of B. subtilis spores with urea and mercaptoethanol resulted in the detachment of a layer from the underlying outer coat material that resembled the basal membrane of the exosporium of B. anthracis (46, 47). Use of a ruthenium red staining technique which has been used for visualization of carbohydrate capsules by electron microscopy resulted in the visualization of filaments on the surfaces of B. subtilis spores (Fig. 3) (48). Spore surface filaments are present and are possibly glycoprotein in nature, although this conclusion is drawn with caution, as ruthenium red staining is not totally glycoprotein specific. Spore surface-specific carbohydrates exist in B. subtilis (49). These carbohydrates display structural similarity to those of the B. cereus family spores (49, 50). Production and/or attachment of polysaccharides to the B. subtilis spore surface was shown to be spsM dependent (51). This determinant is the site of SPβ prophage integration, and functional SpsM is produced only following excision of the prophage during sporulation (51). Spores from spsM-negative mutants had altered surface properties, including increased adherence to glass and enhanced aggregation in aqueous environments. The presence of spore surface glycoprotein filaments raises the possibility that functions contributed by the BclA nap on the B. cereus family spores are provided by other surface glycoproteins in certain exosporium-less spores.

Open in a separate window

FIG 3

Electron micrographs of spores from sporulated cultures. (A) B. anthracis stained with ruthenium red. 1, glycoprotein nap; 2, basal membrane. (B) B. anthracis without ruthenium red staining. 1, glycoprotein nap; 2, basal membrane. (C) B. subtilis stained with ruthenium red. 1, glycoprotein nap; 3, outer coat; 4, inner coat. (D) B. subtilis without ruthenium red staining. 3, outer coat; 4, inner coat. (Reprinted from reference 48 with permission of the publisher [copyright Elsevier 2004].)

Additional support for the presence of B. subtilis spore surface glycoproteins came from a study of the cge locus. Mutations in the cgeAB and cgeCDE operons of B. subtilis resulted in spores with altered surface properties (spores tended to clump, pelleted more tightly, and adhered more strongly to plastic and glass surfaces), and it was speculated that this was due to changes in glycosylation of surface proteins (52). In an analysis of spore coat proteins involving fluorescent protein fusions, the B. subtilis spore coat was predicted to be comprised of four distinct layers, including an outermost glycoprotein layer, named the “crust,” that is separated from the outer coat by a small space (53). Genetic and localization studies have shown that cgeA and genes in the cotVWXYZ locus are involved in spore crust formation (53, 54). Interestingly, the CotZ protein, which plays a role in the assembly of the crust layer of B. subtilis, displays sequence similarity to the ExsY protein (35% identity and 46% similarity over 141 residues), a protein involved in exosporium assembly in B. anthracis (42). B. subtilis CotE, CotV, CotY, and CotW coat proteins expressed in Escherichia coli can self-assemble intracellularly into stable complexes (55). These proteins form one-dimensional fibers, two-dimensional sheets, and three-dimensional stacks. With CotY, the mode of assembly resembles that of the exosporium basal layer described for B. anthracis and B. cereus (55). CotY is a component of the B. subtilis crust layer and the B. cereus group exosporium layer. The propensity of these Cot proteins to self-assemble may contribute to the spore coat, crust layer, or exosporium assembly processes.

The questions of whether the crust is functionally equivalent to the surface of the exosporium and whether crust-like layers are found on the outer spore coat in exosporium-containing spores remain unanswered.

THE BOTTLE CAP MODEL OF EXOSPORIUM STRUCTURE

In a study of an exsY mutant of B. anthracis, Turnbough and coworkers discovered that the exosporium layer is not a uniform structure but is comprised of two distinct entities of differing compositions (42, 56). The “bottle cap” part of the exosporium is synthesized first, originating at the mother cell central pole of the spore. After assembly of the bottle cap portion of the exosporium reaches the cap-noncap junction, the noncap portion (~75%) of the exosporium is assembled (Fig. 4). The noncap portion fails to assemble in an exsY-null mutant, yielding spores with a loosely attached, cap-only exosporium that is readily lost when the spore is released from the mother cell (42). The protein compositions of the cap and noncap regions of the exosporium differ. CotY is a cap-specific protein, whereas BclB, ExsY, and Alr are found in the noncap portion of the exosporium (56,–58). BclA is found on the exosporium surface in both cap and noncap sites, whereas BxpB is found in larger amounts, but not exclusively, in the noncap portion of the exosporium (58). The junction between the cap and noncap regions may play an important role in exosporium-containing spores during the germination process. Steichen et al. (56) demonstrated that during spore germination and outgrowth, the emerging vegetative cell escapes from its exosporium shell through the displaced cap, suggesting that the cap is designed to pop off the exosporium to permit the emergence of the outgrowing cell. Mutants of B. anthracis lacking the BclB collagen-like protein have an exosporium that exhibits damage and loss of the exosporium at one pole (59). Furthermore, the distribution of exosporium proteins, such as CotY and BxpB, is altered in this mutant, suggesting that BclB is required for correct exosporium assembly. The normally cap-associated CotY protein is distributed more completely around the spore, and BxpB is largely restricted to the noncap pole in bclB-null spores. Localization studies with the BclB-negative spores suggested that without BclB, the cap encompasses >75% of the exosporium's circumference, instead of the usual ~25%. BclB, a non-cap-region exosporium protein, may participate in the transition in exosporium composition that must take place as the assembly of the cap stops and the noncap assembly begins during the discontinuous assembly of the exosporium (56, 58). It has been postulated that in a BclB-null strain, the transition zone between the cap and noncap sections of the exosporium is either displaced to the opposite pole of the spore or enlarged to encompass the midsection of the spore (58).

Open in a separate window

FIG 4

(Top) Illustration of a mature spore in a sporulating cell of B. anthracis. The bottle cap portion of the exosporium, at the mother cell central pole of the spore, is depicted in blue, and the noncap exosporium is shown in red. (Bottom) Three distinct exosporium synthesis events which originate at the mother cell central pole of the spore. (1) The first event is the exosporium scaffold layer portion of the basal layer synthesis involving CotY and ExsY. (2) This is followed closely by outer basal layer assembly, which includes BxpB and the BclA nap layer. When the BxpB layer is then approximately 75% complete, cleavage of the BclA N terminus after amino acid 19 (arrow 3) begins at the mother cell central spore pole, and this wave of cleavage follows the assembly around the spore periphery (57).

ASSEMBLY OF THE EXOSPORIUM

Genetic studies on spore assembly in B. subtilis have ordered the steps in spore production based on the effects of particular mutations on downstream processes (60). Analysis of mutants involving exosporium genes of B. anthracis and B. cereus are more limited but do shed some light on the overall exosporium biosynthetic process. The B. cereus exosporium first appears in the sporulating mother cell as a small lamellar structure adjacent to but distinct from the outer forespore membrane (61). Electron microscopy and protein localization studies have shown that synthesis of the exosporium initiates at the mother cell central spore pole, the region of the exosporium that will become the bottle cap (35, 56, 57, 62). Exosporium assembly progresses by continued deposition of proteins at the leading edge of this nascent exosporium, terminating at the opposite spore pole. However, there is a transition in the incorporation of proteins as synthesis reaches the end of the cap region (approximately 25% of the spore circumference), as the composition of the noncap region of the exosporium differs from that of the bottle cap region. As discussed above, there are proteins involved in anchoring the exosporium to the spore coat. The exact nature of the spore coat anchor has yet to be defined. However, the spore coat protein CotE is involved in this process (43). In CotE-negative spores, the exosporium assembles in sheets which are not anchored to the spore, and mature spores have only fragments of exosporium associated with them. However, the exosporium produced in this mutant possesses the BclA nap layer, a late addition in the exosporium assembly process. Thus, anchoring of the exosporium to the underlying spore coat layer is not required for the exosporium assembly process to proceed.

Known exosporium basal layer proteins include ExsY, CotY, BxpB (also known as ExsFA), and ExsFB. The former two proteins are involved in an earlier stage of exosporium synthesis than the latter pair. Loss of ExsY in B. anthracis results in the production of only the cap region of the exosporium, with no noncap region of the exosporium evident in either late-stage sporulating cells or released mature spores (42). The exosporium cap region is not firmly anchored to the spore and is lost, presumably due to fluid shear forces, when spores are prepared in a liquid medium (42, 44). B. cereus mutants devoid of both ExsY and CotY lack an exosporium layer (42, 44), whereas loss of ExsY results in spores with only fragments of exosporium-like material adherent to the spores. The absence of exosporium caps on B. cereus exsY mutant spores was likely the result of shear forces in place during preparation of the spores (44). B. cereus single cotY mutants produce an intact exosporium. With the B. anthracis exsY mutant, no fragments of apparent exosporium remain attached to the released spore as observed with B. cereus. ExsY appears to be essential for noncap exosporium assembly, whereas its paralog CotY is not required for production of an intact exosporium (44). The cap region exosporium produced in the B. anthracis exsY mutant possesses the BclA nap layer (42). Since BclA incorporation occurs later in the exosporium biosynthesis process, this indicates that later-stage assembly events occur in the cap independently of production of the noncap region.

BxpB (ExsFA) is an exosporium basal layer protein that, along with its paralog ExsFB, is required for assembly of the BclA nap layer. Loss of both BxpB and ExsFB is required for complete loss of BclA. Without BxpB, only about 25% of BclA is incorporated onto the spore surface. Loss of ExsFB results in only a very modest reduction in BclA spore levels (63, 64). Mutants lacking both BxpB and ExsFA produce an intact exosporium encircling the spore. However, this structure is fragile, is easily damaged, and is lost from the spores over time (63). This observation suggests that BxpB and ExsFB are assembled into the growing exosporium at a later stage than that involving ExsY and CotY, since loss of the latter pair results in no visible exosporium structure. Thus, the exosporium is likely assembled in two or more distinct layers, consistent with the two layers of exosporium basal layer evident by electron microscopy, as described above. CotY (at the cap) and ExsY (at the noncap region) are incorporated into an early scaffold layer (the sac sublayer of Henriques and Moran [1]). Subsequently, BxpB and ExsFB are added in a layer that is likely to be positioned externally relative to the ExsY/CotY-containing scaffold layer (Fig. 4).

Much of what is known regarding exosporium synthesis concerns incorporation of the nap glycoprotein BclA and the basal layer protein BxpB into the developing exosporium of B. anthracis. Both proteins appear in the mother cell cytoplasm approximately 4 to 5 h into sporulation (in broth cultures shaken at 37°C). The majority of each of these proteins first appears in the form of higher-molecular-weight complexes rather than monomers, indicating that stable complex formation occurs in the mother cell cytoplasm well in advance of placement of these proteins at the site of the developing exosporium. Appearance of BclA and BxpB in the sporulating cell initiates an hour prior to evidence of initial assembly on the developing spore (57). The BxpB-containing and BclA-containing complexes observed in Western blots are similar in size, suggesting that interactions between these two proteins may occur. The protein-protein interactions in these complexes are stable to boiling in the presence of 1% SDS and 8 M urea and in the presence of reducing agents.

An N-terminal targeting sequence of BclA and certain other collagen-like proteins is responsible for positioning these proteins at the site of the developing exosporium (57, 58, 65,–67). No such localization domain is present in the BxpB sequence. Therefore, my colleagues and I have proposed a model whereby BxpB is delivered to the spore by virtue of forming a complex with BclA and possibly, to a lesser extent, with other exosporium-targeting motif sequences (58). In keeping with this, we observed a reduction in the amount of BxpB extractable from spores of a BclA-negative strain, a finding similar to that observed by others (65, 66, 68). However, this model is disputed by Rodenburg et al. (39), based on their analysis of B. anthracis spore structure by high-resolution electron microscopy. They propose that loss of BclA has no effect on BxpB spore incorporation. The assignment of these structures as BxpB, however, was not validated either biochemically or immunologically, so the data remain speculative. How their results can be reconciled with the reduction in BxpB levels in spores in which BclA failed to localize, as previously reported by Tan and Turnbough (65), was not addressed in their more recent report.

Exosporium assembly initiates at the mother cell central pole of the spore (the cap region of the exosporium). Assembly begins with the incorporation of an as yet undefined set of proteins (but likely including CotY and ExsY) which provides the scaffold for assembly of the basal layer. After scaffold assembly has initiated, the assembly of the BclA- and BxpB-containing complexes occurs at this mother cell central spore pole and follows scaffold assembly around the spore (57, 66). When positioning of BclA-containing complexes has occurred over approximately 75% of the spore, proteolytic cleavage of the BclA N terminus after amino acid 19 occurs, beginning at the cap end of the exosporium, and the N-terminal peptide is released into the mother cell cytoplasm (57). This cleavage event may be associated with covalent attachment of BclA to an exosporium basal layer protein, thus anchoring the nap glycoprotein to the exosporium surface. Late in the assembly process, after BclA has been incorporated onto the spore surface, glycosylation of BclA occurs (66).

BACILLUS EXOSPORIUM PROTEINS

Exosporium proteins have principally been defined for strains of B. anthracis and B. cereus. Except where noted below, exosporium proteins of the B. cereus family (B. anthracis, B. cereus, and B. thuringiensis), other than collagen-like proteins, tend not to be found, at least as conserved homologs, in other species of Bacillus or in Clostridia. Therefore, an absence of these proteins in genomic sequences does not rule out the presence of an exosporium. The exosporium-associated Cot proteins display sequence similarity with spore coat proteins of other spore formers, and their locations within spores in the non-B. cereus-family spore-forming bacteria have not been defined. The chromosomal locations of putative exosporium determinants in B. anthracis are shown in Fig. 5. The determinants map to multiple sites in the genome rather than being clustered in a single or small number of loci. The proteins indicated below have been shown to be either exosporium associated or possibly so, in cases where the actual location has not been determined precisely. The BAS numbers indicate the Sterne strain coding sequence (CDS) numbers (NCBI accession numbers NC_005945.1 and AE017225.1).

Open in a separate window

FIG 5

B. anthracis Sterne chromosome with map positions of exosporium determinants.

BclA (BAS1130)

BclA was the first spore surface glycoprotein discovered in B. anthracis and is the prominent protein component of the exosporium nap layer (31, 33). The filaments of the nap are composed of this trimeric collagen-like protein. BclA is comprised of three domains. The N-terminal domain (NTD) contains the targeting and attachment domains whereby the protein is apparently covalently attached to the basal layer of the exosporium (57, 65, 66, 75). Stable attachment requires proteolytic cleavage of the NTD between residues serine 19 and alanine 20. The NTD possesses a targeting sequence which is required for the positioning of BclA around the developing exosporium and the attachment domain containing the proteolytic cleavage site (Fig. 6) (31, 57, 65, 66). The targeting sequence is conserved in a subset of other exosporium-associated collagen-like proteins, whereas the attachment domain, which lies upstream of the targeting sequence, is not identifiable in these proteins (Fig. 6) (57).

Open in a separate window

FIG 6

N-terminal amino acid sequences of B. anthracis collagen-like proteins possessing the exosporium assembly domain. The proteolytic cleavage site of BclA is denoted by an arrow. Attachment motif sequences are shown in red, and localization motif sequences are shown in green.

The NTD is followed by a collagen repeat region consisting of GX₁X₂ (primarily GPT) triplet amino acid repeats. These repeats are the primary attachment point for rhamnose-oligosaccharides within BclA (76). The number of glycine-containing triplet amino acid repeats in BclA varies among strains (17 to 91 triplet repeats), and this variation is responsible for the different lengths of the hairlike nap observed on spores of different B. anthracis strains (32, 33). The 134-residue carboxy-terminal domain is organized into an all-β structure with a jelly roll topology (77, 78). It is structurally similar (but dissimilar in sequence) to the globular domain of the human C1q complement protein, which belongs to the tumor necrosis factor-like family of proteins. The trimeric globular head formed by the C-terminal domain is thought to be the nucleation site for the formation of the collagen-like triple helix, as in several mammalian collagens (77).

The polysaccharide composition of BclA includes l-rhamnose, N-acetylgalactosamine, 3-O-methyl rhamnose, and anthrose [2-O-methyl-4-(3-hydroxy-3-methylbutamido)-4,6-dideoxyglucose; found at the nonreducing terminus of a tetrasaccharide] (50, 76). Daubenspeck et al. (76) identified a 715-Da tetrasaccharide [2-O-methyl-4-(3-hydroxy-3-methylbutamido)-4,6-dideoxy-β-d-glucopyranosyl-(1→3)-α-l-rhamnopyranosyl-(1→3)-α-l-rhamnopyranosyl-(1→2)-l-rhamnopyranose] and a 324-Da disaccharide attached to BclA, likely through N-acetylgalactosamine residues at the reducing ends, making the polysaccharides actually pentasaccharides and trisaccharides. Multiple copies of the likely pentasaccharides are O-linked to the collagen-like region of BclA, apparently through threonine residues present in the many GXX repeats in this region. The attachment site of the trisaccharides was predicted to be outside the collagen-like repeat region. The BclA glycoprotein is present in strains of B. cereus and B. thuringiensis, but the presence of anthrose appears to be specific only to BclA from B. anthracis (76). However, antisera raised against anthrose cross-reacted with spores of certain B. cereus strains (79). The C-terminal domain of BclA from B. cereus ATCC 14579 was reported to be glycosylated, with 2-O-methyl-rhamnose and 2,4-O-methyl-rhamnose substitutions found in the C-terminal domain (80).

BclA has also been shown to be an immunodominant protein on the B. anthracis spore surface and contributes to the overall spore hydrophobicity (33, 81). This immunogenicity provides a potential benefit for anthrax vaccine development, as immunization with protective antigen (PA) and BclA protected mice and guinea pigs from challenge with B. anthracis Sterne (pXO1⁺/pXO2⁻) or Ames strains better than immunization with PA alone (68, 81,–83). Loss of BclA has little, if any, effect on sporulation rates of B. anthracis. However, germination of BclA-null spores was reported to occur more rapidly than that of wild-type spores in vitro (84).

The BclA protein has a predicted molecular weight (MW) of 37,000, but it first appears as higher-molecular-weight forms in sporulating cells (57, 66, 67). BclA extracted from spores is found in a heterogeneous band of complexes centered with an apparent MW of >250,000. The higher-molecular-weight forms are likely the result of trimerization of this collagen-like protein, its glycosylation, and its capacity to form covalent and noncovalent but SDS- and urea-resistant complexes with other exosporium proteins.

The bclA determinant (bas1130) is carried within a cluster of genes (bas1127 to bas1145) which is organized as shown in Fig. 7. It possesses a putative promoter element recognized by the σ^K-bearing form of RNA polymerase, resulting in its expression approximately 3 h into sporulation in the mother cell compartment of the sporulating cell. The bclA determinant is flanked by genes predicted to be involved in glycosylation of this, and perhaps other, spore proteins. The rmlABCD genes are predicted to encode enzymes for rhamnose production, and inactivation of the rmlA determinant resulted in a loss of rhamnose incorporation into spores (85). rmlA mutant spores exhibited decreased binding to macrophages, suggesting a role for the carbohydrate moiety of the BclA glycoprotein in host cell interactions (85). Genes known to be associated with glycosylation of BclA are not found at a single locus on the B. anthracis genome. The anthrose component of the BclA glycoprotein is synthesized by the products encoded within a four-gene operon (bas3322 to bas3319) not linked to bclA (86,–88). Furthermore, the bas5304 determinant was shown to be the epimerase required to produce N-acetylgalactosamine for BclA oligosaccharide biosynthesis (86).

Open in a separate window

FIG 7

bclA locus of the B. anthracis Sterne chromosome. Exosporium determinants are shown in black, rhamnose biosynthetic genes important for glycosylation of BclA are shown in light blue, the gene for the spore coat protein CotO is shown in dark blue, putative exosporium determinants are shown in brown, and genes encoding proteins which may be involved in glycosylation of BclA, and perhaps other spore collagen-like proteins, are shown in red (putative methyltransferases) and purple (putative glycosyltransferases).

FUNCTIONS OF THE EXOSPORIUM IN THE ENVIRONMENT

Spores present in soil environments constitute the source of animal infections by B. anthracis and other soil-associated pathogens. Furthermore, it has been shown that under the right conditions, B. anthracis spores can germinate on and around plant roots and that genetic exchange between strains of Bacillus is possible (110). Because of the spore mode of infection, it is important for spores to persist at their site of deposition and not to get washed too deeply into the soil and thus be inaccessible to grazing ruminants in amounts constituting an infectious dose (111). The exosporium affects the ability of B. anthracis spores to bind to different soil types. Wild-type spores of B. anthracis adhere more strongly than exosporium-deficient spores to soils with high organic material and calcium levels, i.e., soils which have been shown to favor the persistence of this pathogen (112). The presence of BclA on the surfaces of spores resulted in better retention of spores in a porous medium of silica sand (113).

The exosporium layer and the BclA glycoprotein, in spores possessing it, contribute to the overall hydrophobicity of the spore (80, 84, 99). The hydrophobic character contributes to the binding properties of the spore in the environment. Binding to food preparation surfaces is important to the foodborne pathogen B. cereus. Spore adherence to surfaces, such as stainless steel, contributes to the environmental persistence of the spores and ultimately to infection of foods by this bacterium.

Studies have attributed roles for the exosporium (114, 115) or BclA (78, 99) in binding to stainless steel surfaces. In contrast to the results obtained with B. anthracis, a study with a strain of B. cereus (ATCC 14579) indicated an increase in spore hydrophobicity in bclA deletion mutants, and this coincided with a reduction in spore binding to stainless steel (99). However, overall spore hydrophobicity does not adequately explain the differences in stainless steel adhesion properties. Other reports have concluded that the overall hydrophobicity of the spore and the presence of surface appendages by these spores have the greatest impacts on stainless steel binding (116,–118).

FUNCTIONS OF THE EXOSPORIUM IN AN INFECTED HOST

The majority of studies concerned with spore infectious processes have focused on the zoonotic pathogen B. anthracis. This bacterium has an unusual host-pathogen interaction. The bacterium, in order to disseminate as many infectious spores into the soil as possible, has evolved mechanisms to gain access to the host's bloodstream, where it replicates to large numbers and kills the host through elaboration of toxins. The sporulation process does not occur to any significant extent in the infected host until after the animal dies and postmortem carcass decomposition provides the oxygen-replete environment needed to induce sporulation. The large bacterial numbers then translate to large numbers of spores released into the soil around the decomposing carcass. To accomplish this infectious process, the infecting spores must transit from the site of infection to the draining lymph node, germinate to the vegetative bacterial form, and initiate invasion of the bloodstream and rapid replication. The exosporium, in particular the surface-exposed BclA protein, participates in many of the initial spore-host interactions.

Studies with animal models of disease and tissue culture-based infections have provided clues to the nature of the initial events in the infectious process. These studies indicate that when spores are introduced into an animal host, they are taken up by macrophages and dendritic cells (119,–123). Spores in the lung are phagocytosed by alveolar macrophages or dendritic cells, which initiate the processes leading to pulmonary anthrax (120, 122, 123). For pulmonary infections, it has also been shown that B. anthracis spores are capable of entering epithelial cells of the lung and crossing this cell layer without apparent disruption of the barrier integrity (124, 125). The gastrointestinal form of anthrax ensues when spores are ingested and migrate to the intestinal tract, where they pass across the gut epithelial cell layer without germination. Once outside the intestines, the spores are taken up by the macrophages of the Peyer's patches (119). In either form of anthrax, the phagocytes migrate to the draining lymph nodes. The spores germinate en route within the macrophage or dendritic cell and begin to replicate and produce toxins. At the lymph node, the bacterium-containing phagocytes lyse, releasing the bacteria. Invasion of the bloodstream, replication, toxin elaboration, and death of the susceptible host ensue.

Uptake of spores by phagocytic host cells has been shown to involve interactions between BclA and the host integrin Mac-1 (CR3) (126). Spores lacking the BclA glycoprotein do not specifically target these cells but bind more generally to epithelial cells (127). CD14 was shown to bind to rhamnose residues of BclA and to act as a coreceptor for spore binding by Mac-1 (128). Oliva et al. proposed a model whereby binding of rhamnose to CD14 triggers signaling through Toll-like receptor 2 (TLR-2), leading to activation of phosphatidylinositide 3-kinases and converting Mac-1 into a more active receptor for B. anthracis spores (128). Survival studies using C57BL/6 and CD11b^−/− mice revealed that CD11b^−/− mice lacking Mac-1 were more resistant to infection with wild-type but not BclA-deficient spores (126, 128).

These results suggest that the spore surface BclA fibers may act to promote uptake by professional phagocytes and to inhibit nonspecific interactions between B. anthracis spores and nonprofessional phagocytic cells. Gu et al. reported that activation of the classical complement pathway is a primary mechanism of spore phagocytosis by murine macrophages (129). Phagocytic uptake of spores was found to be reduced significantly in the absence of complement component C1q or C3. C1q recruitment to the spore surface was shown to be mediated by BclA, which can directly bind C1q in a dose-dependent and saturable manner (129, 130). Furthermore, C1q acted as a bridging molecule between BclA and integrin α₂β₁ to mediate B. anthracis spore entry into epithelial cells in a complement activation-independent manner. A binding protein for the globular head domains of complement component C1q, designated gC1qR, was also reported to be involved in at least the initial stages of B. cereus spore attachment and/or entry (131).

The above findings suggest a central role for BclA in B. anthracis spore infection of hosts. However, loss of BclA has been shown to have no effect on virulence of B. anthracis in animal models of infection, as measured by 50% lethal dose (LD₅₀) values (84, 132). Additionally, the mean time to death following subcutaneous or intranasal inoculation of ΔbclA mutant spores in an A/J mouse model of infection was shorter than that with wild-type spores (84). Oliva et al. found that C57BL/6 mice were more susceptible to ΔbclA spores than to wild-type spores (the LD₅₀ of ΔbclA spores was approximately 5-fold lower than that of wild-type Sterne spores in a subcutaneous challenge model, and the mean time to death was shorter with the BclA-lacking spores) (126). They also reported similar results with A/J mice after intrathecal spore administration. A better understanding of the early events in the infectious process is needed before we can explain the discrepancy in results reported for animal disease models and why loss of the BclA spore surface glycoprotein either has a minimal impact on the lethal dose or creates an increase in virulence, depending on the animal model of infection utilized.

Spores of B. anthracis can persist in the lungs of infected hosts following aerosol exposure. For example, viable spores were found in the lungs of rhesus macaques at least 42 days following aerosol exposure (133,–135) and in mouse lungs for more than 25 to 30 days (136, 137). Spores of B. anthracis were significantly better at persisting in the lung than spores of a non-exosporium-possessing B. subtilis strain, and some of the persisting spores were found within epithelial cells (138). Treatment of mice with an Src family kinase inhibitor significantly reduced B. anthracis dissemination from the lung to distal organs and prolonged the median survival time for mice compared to those for the untreated control group (139). There appears to be a signaling pathway specifically required for spore entry into epithelial cells, and these studies provided evidence suggesting that this pathway is important for dissemination and virulence in vivo (139).

Another possible mechanism whereby spores persist in the lungs involves innate immune system evasion. Interference with the innate immune system may initially occur upon introduction of spores into a susceptible host. Plasminogen efficiently binds to spores in an exosporium- and BclA-dependent manner (140). Plasminogen-bound spores are capable of exhibiting antiopsonic properties by cleaving C3b molecules, resulting in a decrease in macrophage phagocytosis, possibly contributing to a longer persistence of free spores in infected lung tissue.

Perhaps as a consequence of the decreased hydrophobicity of BclA-deficient spores, the extracellular matrix proteins laminin and fibronectin bound significantly better to ΔbclA mutant spores than to wild-type spores (84). BclA-negative spores were also found to adhere better to epithelial cells, fibroblasts, and endothelial cells but not macrophages. Spores lacking the exosporium layer were found to elicit a higher cytokine response (beta interferon, interleukin-1β [IL-1β], tumor necrosis factor alpha [TNF-α], and IL-6) in mouse macrophages than that in response to exosporium-bearing spores (141). This suggests a role for the exosporium in avoidance of innate immune signaling during the early stages of spore infection.

Studies have suggested that the timing of spore germination within phagocytic cells is regulated to prevent premature germination of spores. The timing of germination is thought to be critical to the outcome of the spore-macrophage interaction (70). The exosporium is involved in the initial spore interactions with the intracellular compartment of phagocytes following spore uptake. Enzymes are present in the outer spore layers, apparently to prevent premature germination of the internalized spores and to provide protection to the emerging vegetative cell following germination of the spore (34). These spore-associated enzymes include alanine racemase to convert l-alanine (a germinant) to d-alanine (an inhibitor of l-alanine-induced germination) and at least two, and possibly more, inosine-preferring nucleoside hydrolases that may function to inactivate inosine (another germinant species) (35, 69, 70, 101, 142).

The B. anthracis exosporium contains an arginase that metabolizes l-arginine to l-ornithine and urea, thereby reducing the levels of arginine, the substrate for the host cell nitric oxide synthase (71, 72). This decreases the amount of NO radicals produced in spore-infected macrophages. Counterintuitively, B. anthracis spores have their own NO synthase. Spores lacking this enzyme were attenuated for virulence in an A/J mouse model of infection and were more susceptible to killing when germinating within macrophages (143). Resistance to macrophage killing is thought to be dependent on NO-mediated activation of bacterial catalase and suppression of the Fenton reaction. Additionally, a potential exosporium-associated virulence factor is superoxide dismutase, which may provide protection to the germinating spores against reactive oxygen species-mediated defense systems of phagocytic cells (107, 108). These spore-associated enzymes may protect the newly emerging vegetative cells at a time when they are very sensitive to killing by reactive oxygen species, until toxin production by the newly emerging vegetative bacteria can ensue and ultimately kill the host cell.

CONTRIBUTIONS OF EXOSPORIUM PROTEINS TO PROTECTIVE IMMUNITY

The first demonstration of an effective vaccine against anthrax in animals was conducted by Louis Pasteur at the famous public experiment at Pouilly-le-Fort, in May 1881, to demonstrate his concept of vaccination. The vaccine consisted of B. anthracis spores that were chemically attenuated for virulence. The anthrax vaccine widely used in veterinary medicine today is a live spore vaccine and involves the use of the Sterne strain, which lacks the pXO2 plasmid and thus cannot produce the bacterial cell capsule, an important virulence factor for this bacterium. This vaccine was originally developed at the Onderstepoort Laboratory, Pretoria, South Africa. Human vaccines against anthrax, in contrast, are cell-free preparations whose efficacy is considered to be due primarily to antibodies generated against the protective antigen component of the anthrax lethal and edema toxins. These include the American anthrax vaccine adsorbed (AVA; BioThrax) and the British anthrax vaccine precipitated (AVP). Animal studies of the efficacy of the human vaccine found that protection against the most virulent strains of B. anthracis required the addition of spore antigens to the vaccination regimen (144,–147). Improved vaccine protection was observed when the added spore antigens included inactivated whole spores (147,–150), live attenuated spores (146, 151,–154), BclA (81, 83, 148, 155), BxpB (68, 148), and BAS5303 (68, 148). Antibodies directed against whole spores or, specifically, BclA have been shown to be inhibitory to spore germination in vitro (148, 156). Studies involving passive immunization with spore-specific antisera have produced mixed results. Protection has been reported in some studies but could not be demonstrated in others (148, 157,–159). The distinct difference between the unambiguous benefit obtained with active immunization with PA and spore components and the less clear-cut results obtained via passive immunization involving anti-spore-component antibodies underscores the likely involvement of cellular immunity in protection (159).

BACTERIA WITH EXOSPORIUM-CONTAINING SPORES

Several endospore-producing bacteria have been reported to possess exosporia. Examples are listed in Table 1. In addition, genes similar to bclA of B. anthracis are present in the genome sequences of Clostridium autoethanogenum DSM 10061, Clostridium beijerinckii NCIMB 8052, Clostridium ljungdahlii DSM 13528, Clostridium saccharolyticum WM1, and Clostridium ramosum.

Clostridium difficile

Several articles have been published that refer to an exosporium layer of C. difficile (recently proposed to be reclassified as Peptoclostridium difficile) (184). In only one publication is there electron microscopic evidence of a possible traditional exosporium layer with a defined basal layer and interspace region (185). In this report, as in others, micrographs indicate an external crust-like layer (such as that shown in Fig. 8) or fragments of material adherent to the spore coat in strains of C. difficile (13, 174, 185, 186). Joshi et al. also indicated a pronounced variation in spore hydrophobicity among strains, with the more hydrophobic strains possessing a more organized exosporium (173). These studies suggest that possession of an exosporium may not be a feature of C. difficile spores or may be a strain-variable trait in this bacterium. An alternative possibility is that the exosporium layer is extremely fragile and readily lost from spores (13). The more likely possibility is that C. difficile spores lack a true exosporium layer (as proposed previously [13]) and possess a crust layer, albeit one with a pronounced hairlike nap. The bacterium exhibits strain variability in the appearance of the crust layer nap. A proteomic study did not reveal exosporium proteins similar to the known B. cereus family exosporium proteins, except for the presence of collagen-like proteins (187). However, the lack of B. cereus family exosporium homologs does not rule out the existence of C. difficile exosporium structures, since genomic analyses of known exosporium producers outside the B. cereus family suggest that exosporium proteins differ markedly in various species.

Open in a separate window

FIG 8

Transmission electron micrographs of C. difficile strain ATCC 700057 spores. The spores were stained with ruthenium red, a carbohydrate stain to identify the glycoprotein layer of the spore surface. The spores feature a prominent cortex region (the white area surrounding the core). An interspace region, the area between the exosporium and the outer spore coat, is not an obvious feature in C. difficile spores. The ruthenium red-stained putative glycoprotein layer is denoted by arrows.

Three genes encoding collagen-like glycoproteins have been identified in the C. difficile genome (bclA1, bclA2, and bclA3), and the encoded proteins are expressed during sporulation and localize to the spore, with BclA1 likely present in the outermost “exosporium/crust” layer of the spore (13, 188). BclA1, unlike BclA of B. anthracis, was reported to be poorly immunogenic in this study. ClosTron-generated mutants of the bclA1 and bclA2 determinants produced spores with altered surfaces, including release of “sheets of coat-like material” from the spores (188). BclA3 was identified as a glycosylated protein in extracts of C. difficile spores, and the glycosyltransferase gene sgtA, carried upstream of bclA3, is at least partly responsible for glycosylation of the spore surface (189).

A cysteine-rich protein (CdeC) localized to the outermost spore layer was found to be important for the formation of the nap layer of the C. difficile spore and might be the anchoring protein linking the outermost layer to the spore coat (190). The CotA protein was shown to be important for proper assembly of the outermost spore layer, and the CotA, CotE (a bifunctional peroxiredoxin reductase and chitinase), CotF, and SodA (superoxide dismutase) proteins were found to be associated with the spore surface (185). The C. difficile bclA1, bclA2, bclA3, cdeC, cotA, cotCB, cotD, and cotE genes were determined to be dependent on σ^K for expression; this is the same sigma factor involved in exosporium gene expression in the B. cereus family of organisms (191).

A proteomic analysis of proteins extracted from the outermost layers of C. difficile strain 630 identified 184 distinct proteins (192). Proteins were extracted from spores by use of trypsin treatment or sonication methods. In this report, Flag fusion studies indicated an “exosporium” location for the BclA1, BclA2, BclA3, CdeA, CdeB, CdeC, and CdeM proteins.

The outermost layer of C. difficile spores has been reported to interact with a receptor or receptors on intestinal epithelial cells (193) and may have an adverse effect on macrophage phagosomal membranes (194). Proteolytic treatment of spores decreased adherence to the cells. Removal of the outermost spore layer(s) was also reported to reduce adherence to colonic cells, decrease overall spore hydrophobicity, and increase spore colony-forming frequencies in vitro (13, 173). The more pronounced exosporium-like layers may correlate with increased spore hydrophobicity and better adherence to gut epithelial cell monolayers and environmental stainless steel surfaces (174). Loss of each of the BclA1–3 glycoproteins resulted in decreased spore hydrophobicity. Relative to the wild-type parent strain, mutant spores lacking BclA1 were shown to have a 2-log higher infectious dose requirement in a mouse model of infection and displayed reduced virulence in a hamster infection model (188). BclA1 is thought to be important in the initial colonization of spores in the gut epithelium of the infected host.

THE EXOSPORIUM AS A SPORE DISPLAY PLATFORM

The BclA collagen-like glycoprotein is abundantly expressed on the surfaces of spores of B. anthracis, B. cereus, and B. thuringiensis. The N-terminal 35 amino acids of BclA have been shown to be sufficient for localization of the protein around the spore and its stable incorporation onto the exosporium basal layer surface (57, 65). It is possible to display foreign proteins on the surfaces of B. anthracis, B. cereus, or B. thuringiensis spores by incorporating the 35 BclA N-terminal codons at the N-terminal end of the foreign gene open reading frame and expressing it in sporulating cells by utilizing the bclA promoter and ribosome binding site. The foreign protein is synthesized by the sporulating cells and stably incorporated in large numbers at the spore surface. The recombinant protein-decorated spores are easily purified and can be employed as microparticles for industrial, bioremediation, or vaccine applications (H. Y. Hsieh, G. C. Stewart, B. M. Thompson, and C. H. Lin, unpublished data). Enzymes that retain activity when tethered at the N terminus are ideal for these biocatalytic processes. The spore display system is ideally suited for soil environments, where spores can remain in place following application, whereas free enzymes tend to wash out and are more readily inactivated.

CONCLUSIONS

Our understanding of the composition, assembly process, and function of the spore exosporium layer remains incomplete. The presence of exosporium proteins in large SDS-resistant complexes hampers studies of specific protein-protein interactions. As with other aspects of endospore formation, genetic studies will likely be the best route to a more complete understanding of the composition, assembly pathway, and function of this outermost spore layer. The distribution of sugar residues on the spore surface glycoproteins and the roles that they may play are understudied aspects of spore biology. Atomic force microscopy was recently utilized to begin to define the distribution of sugars on the surfaces of B. cereus spores (205). Regulation of exosporium formation warrants further investigation. Recently, it was discovered that production of spores of a particular strain of B. cereus under conditions of anaerobiosis resulted in defects in the exosporium structure (206). Based on what we know regarding spore formation by B. anthracis, it could be anticipated that reduced oxygen tension might result in reduced sporulation efficiencies. However, structural defects in the exosporia of the formed spores were unexpected. How exosporia evolved in different species by using unique building blocks will be a fascinating field of study as our understanding of non-B. cereus group exosporium composition improves. It is apparent that the studies begun in the 19th century by Robert Koch still have a lot to disclose regarding the biology of endospores.

ACKNOWLEDGMENTS

Notes

This review is dedicated to the memory of Alvin Fox, who started with me down the road of B. anthracis spore research.

REFERENCES