A strategy to engineer Bacillus subtilis spores to display heterologous antigens on the spore surface has been recently reported 35 and is illustrated in Figure 2. A spore-based display system provides several advantages with respect to systems based on the use of bacterial cells, these include the robustness of the bacterial spore allowing storage in the desiccated form, ease of production, safety and a technological platform supported by extensive tools for genetic manipulation.

In contrast to the wealth of information available on how gene expression controls differentiation of a growing cell into a dormant spore little is known about the mechanisms of protein incorporation into the coat, the nature of structural components forming the most external part of the coat and whether there are anchoring motifs. Initial attempts to expose heterologous proteins on the spore surface have been focused on two coat components, CotB and CotC. In the case of CotB this spore coat protein is known to be surface located 35 while for CotC, relative to other coat proteins, this species has a high relative abundance 36. The observation that both of these coat components was dispensable for the formation of an apparently normal spore as well as for its germination, was an additional positive feature in choosing CotB and CotC as potential carrier proteins.

Two antigens were initially selected as model proteins to display on the spore surface: i) the non-toxic 459 amino acid C-terminal fragment of the tetanus toxin (TTFC), a well characterized and highly immunogenic 51.8 kDa peptide 37, encoded by the tetC gene of Clostridium tetani; and ii) the 103 amino acid B subunit of the heat-labile toxin of enterotoxigenic strains of Escherichia coli (LTB), a 12 kDa peptide, encoded by the eltB gene 38.

Like other coat components, CotB has been associated to the outer coat layer on the basis of genetic evidence 39 and only recently an immunocytofluorimetric analysis performed on intact spores has shown that CotB is accessible to CotB-specific antibodies and therefore that it is most probably exposed on the spore surface 35.

The CotB structural gene, cotB, is under the dual transcriptional control of σ^K and the DNA-binding protein GerE. As a consequence, cotB is transcribed only in the mother cell compartment of the sporulating cell 27. Once synthesised in the mother cell cytoplasm, CotB is assembled around the forming spore in a manner somehow dependent on CotE, CotG and CotH. Therefore, CotB and the heterologous protein eventually fused to it, do not undergo a cell wall translocation step, typical of display systems in other bacteria.

CotB has a strongly hydrophilic C-terminal half formed by three 27 amino acid-repeats rich in serine, lysine and glutamine residues. Serine residues account for over 50% of the CotB C-terminal half. The lysine residues in the CotB repeats have been suggested to represent sites of intra- or inter-molecular cross-linking, by analogy to the connective tissue proteins collagen and elastin 2740. The CotB protein has a deduced molecular mass of 46 kDa, but migrates on SDS-PAGE as a 59 kDa polypeptide. Recently the discrepancy between measured and deduced molecular weight has been explained by showing that CotB is initially synthesized as a 46 kDa species, and converted into a 59 kDa homodimer 41, that retains both the N- an C-terminal ends predicted from the cotB nucleotide sequence.

The strategy to obtain recombinant B. subtilis spores expressing CotB-TTFC or CotB-LTB on their surface was based on (i) use of the cotB gene and its promoter for the construction of translational fusions and (ii) chromosomal integration of the cotB-tetC and cotB-eltB gene fusions into the coding sequence of the non-essential gene amyE (Figure 3A) 42. Placing the fusion proteins under cotB transcriptional and translational signals ensured correct timing of expression during sporulation, while its chromosomal integration guaranteed the genetic stability of the construct. Due to the lack of information on CotB coat assembly and on the requirements for anchoring motifs, initial attempts were performed by positioning the passenger protein at the C-terminal, the N-terminal or in the middle of CotB (Fig. 3B).

When TTFC and LTB were fused to the C-terminal end of CotB, the chimeric proteins failed to correctly assemble on the spore surface (Isticato and Ricca, unpublished). Such initial failures were attributed to a potential instability of the constructs, either at the DNA level (repetitive DNA sequences) or at the protein level. In order to bypass such problems TTFC and LTB were fused to the C-terminal end of a CotB form deleted of the three 27 amino acid-repeats, CotBΔ₁₀₅-TTFC (Fig. 3A). In contrast to the full length version, the CotBΔ₁₀₅-TTFC chimeric protein was correctly assembled and exposed on the spore surface 35. A quantitative dot blot showed that each recombinant spore exposed an amount of CotBΔ₁₀₅-TTFC fusion protein equal to 0.00022 pg making it possible to conclude that 1.5 × 10³ chimeric molecules are present on the surface of each recombinant spore 35.

Unlike CotBΔ₁₀₅-TTFC, CotBΔ₁₀₅-LTB was not properly assembled. The strain expressing this chimera showed reduced sporulation and germination efficiencies and its spores were not resistant to lysozyme. These observations, together with the SDS-PAGE analysis of the released coat proteins, suggested that the presence of CotBΔ₁₀₅-LTB strongly altered the spore coat layer. An in-silico analysis showed some homology between the chimeric product (in the fusion region) and LytF, a cell wall-associated endopeptidase produced by B. subtilis during vegetative growth, thus raising the possibility that the chimeric product could interfere with proper coat formation by degrading some coat components (Mauriello and Ricca, data not shown).

In addition to the C-terminal end fusion described above the model passenger protein TTFC has been fused also at the N-terminal and in the middle of CotB (Fig. 3B). In both cases the CotBΔ₁₀₅ form of CotB was used to avoid the problems experienced with the C-terminal fusion (see above). Both the N-terminal and the sandwich fusions originated chimeric products that were properly assembled in the coat structure from both the qualitative and the quantitative point of view 35. At least in the case of CotB, it was then possible to conclude that where the passenger protein is exposed it does not affect display on the spore surface.

CotC is a 12 kDa, alkali-soluble component of the B. subtilis spore coat, previously identified by reverse genetics 32 and then associated to the outer coat layer based on genetic evidence 39. CotC was initially considered as a carrier candidate because of its relative abundance in the coat (Figure 1B). Together with CotG and CotD, CotC represents about 50% of the total solubilized coat proteins. Such relatively high amounts could allow the assembly of a significant number of CotC-based chimeras on the coat, thus ensuring an efficient heterologous display. Expression of the cotC gene is under the control of the mother cell specific σ-factor σ^K and of transcriptional regulators GerE and SpoIIID. As in the case of CotB, CotC is also transcribed in the mother cell and its assembly on the coat does not require membrane translocation. The primary product of the cotC gene is a 66 amino acid polypeptide extremely rich in tyrosine (30.3%) and lysine (28.8%) residues 32. However, it was recently shown that CotC is assembled into at least four distinct protein forms, ranging in size between 12 and 30 kDa 31. Two of these, having molecular masses of 12 and 21 kDa and corresponding most likely to a monomeric and homodimeric form of CotC respectively, are assembled on the forming spore right after their synthesis eight hours after the onset of sporulation. The other two forms, 12.5 and 30 kDa, are probably the products of post-translational modifications of the two other forms, occurring directly on the coat surface during spore maturation 31.

In the case of CotC only C-terminal fusions have been so far constructed (Figure 3B). Both CotC-TTFC and CotC-LTB gene fusions were obtained by cloning tetC or eltB in frame with the last cotC codon under the transcriptional and translational control of the cotC promoter region. The gene fusion was then integrated into the B. subtilis chromosome at the amyE locus by double cross-over recombination 36 (Figure 3A). Both these two chimeric proteins were assembled on the coat of recombinant spores without major effect on the spore structure and/or function, since they appeared identical to wild type spores in terms of efficiency of sporulation and germination and resistance properties. Western blot, cytofluorimetric analysis and, for CotC-TTFC, immunofluorescence microscopy (Figure 4) showed that both CotC-based chimeras were displayed on the surface of the recombinant spores. A quantitative determination of recombinant proteins exposed on B. subtilis spores revealed that ca. 9.7 × 10² and 2.7 × 10³ molecules of CotC-TTFC and CotC-LTB, respectively, were extracted from each spore.

Although CotC appears more abundant than CotB within the coat, comparable amounts of heterologous proteins are exposed by the CotC-based and the CotBΔ₁₀₅-based systems. This result was somewhat unexpected, since CotC appears to be much more abundant than CotB in the coat. A possible explanation comes from the recent finding that the C-terminal end of CotC is not only essential for interaction with other CotC molecules but also to other coat components (Isticato and Ricca, manuscript in preparation) and therefore shows that the use of CotC as a carrier still needs to be optimised.

One of the main reasons to propose the use of the bacterial spore as a favourable display system is its well-documented stability. Spores can be simply stored at room temperature for a long time without reduction of their resistance and stability properties. This would be an extremely useful property for a variety of biotechnological applications. As an example, if the passenger protein is an antigen, the recombinant spore could become an ideal heat-stable oral vaccine for use in developing countries, where heat-stability is of most concern due to poor distribution and storage.

However, while spore stability is well documented 23, stability of heterologous proteins exposed on the spore surface has only recently been investigated. Spores expressing CotBΔ₁₀₅-TTFC (see above) and parental spores were stored at -80°C, -20°C, +4°C and at room temperature and assayed at different storage times up to 12 weeks. In all cases the amount of heterologous protein present on the surface of recombinant spores appeared identical between freshly prepared spores and those stored for up to 12 weeks (Fig. 5). These results, indicating that heterologous proteins can be stably exposed on the surface of recombinant spores, confirm the spore-based system as a very promising display approach that could overcome some drawbacks of other systems and that could find applications in a variety of diverse biotechnological fields.

Spores expressing CotBΔ₁₀₅-TTFC have been used to immunise mice by the oral route 43. Serum IgG and faecal sIgA showed clear seroconversion to TTFC. The dosing schedule used three sets of three doses (1.67 × 10¹⁰) over 5 weeks and was based on regimes optimised for oral immunisations 4445. The titres of TTFC-specific IgG after 33 days (>10³) suggested that these were at protective levels and mice challenged with tetanus toxin corresponding to 10 LD₅₀ were fully protected. Out of eight mice challenged with a 20 LD₅₀ dose seven survived suggesting that this was the threshold for protection. A similar study was made using nasal immunisation with CotB-TTFC spores but with a lower dose and three immunisations 43. Here, TTFC-specific IgG responses were lower but still showed seroconversion. These studies show that engineered spores expressing a heterologous antigen can be used for protective immunisation. Moreover, although mucosal responses are not important for protection against Clostridium tetani (a systemic pathogen) they are obviously important for mucosal pathogens. Further studies will be needed to optimise dosing regimes (less doses and less spores) but these seminal studies have opened the way for development against specific mucosal pathogens. Although these studies are encouraging and demonstrate humoral responses there is no clear evidence yet that indicates cellular responses. However, spores have been shown to disseminate to the GALT and are found in the Peyer's Patches (PP) and mesenteric lymph nodes 46. The small size of the spore (1 μm) would allow it to be taken up by M cells and transported to the PP where it could interact with antigen presenting cells. Initial studies have shown that spores can germinate and persist for a short period of time within intestinal macrophages as well as elicit Th 1 cytokines in vivo such as IFN-γ 47.