Sequence analyses of fimbriae subunit FimA proteins on Actinomyces naeslundii genospecies 1 and 2 and Actinomyces odontolyticus with variant carbohydrate binding specificities

Background Actinomyces naeslundii genospecies 1 and 2 express type-2 fimbriae (FimA subunit polymers) with variant Galβ binding specificities and Actinomyces odontolyticus a sialic acid specificity to colonize different oral surfaces. However, the fimbrial nature of the sialic acid binding property and sequence information about FimA proteins from multiple strains are lacking. Results Here we have sequenced fimA genes from strains of A.naeslundii genospecies 1 (n = 4) and genospecies 2 (n = 4), both of which harboured variant Galβ-dependent hemagglutination (HA) types, and from A.odontolyticus PK984 with a sialic acid-dependent HA pattern. Three unique subtypes of FimA proteins with 63.8–66.4% sequence identity were present in strains of A. naeslundii genospecies 1 and 2 and A. odontolyticus. The generally high FimA sequence identity (>97.2%) within a genospecies revealed species specific sequences or segments that coincided with binding specificity. All three FimA protein variants contained a signal peptide, pilin motif, E box, proline-rich segment and an LPXTG sorting motif among other conserved segments for secretion, assembly and sorting of fimbrial proteins. The highly conserved pilin, E box and LPXTG motifs are present in fimbriae proteins from other Gram-positive bacteria. Moreover, only strains of genospecies 1 were agglutinated with type-2 fimbriae antisera derived from A. naeslundii genospecies 1 strain 12104, emphasizing that the overall folding of FimA may generate different functionalities. Western blot analyses with FimA antisera revealed monomers and oligomers of FimA in whole cell protein extracts and a purified recombinant FimA preparation, indicating a sortase-independent oligomerization of FimA. Conclusion The genus Actinomyces involves a diversity of unique FimA proteins with conserved pilin, E box and LPXTG motifs, depending on subspecies and associated binding specificity. In addition, a sortase independent oligomerization of FimA subunit proteins in solution was indicated.

The major subunit genes of type-2 and type-1 fimbriae, fimA and fimP, respectively, have been cloned and sequenced from A. naeslundii genospecies 1 (strain 12104) and 2 (strain T14V) [22][23][24][25][26]. The deduced FimA and FimP subunit proteins are 534 and 533 amino acid proteins, respectively, with 34 % amino acid identity. FimA and FimP contain seven conserved proline-containing regions involved in folding of the two proteins and an LPXTG sorting signal followed by a N-terminal membrane spanning domain [25]. Structurally diverse fimA and fimP genes, as well as species-specific fimA gene segments, have been found for A. naeslundii genospecies 1 and 2, and linked to different coaggregation groups and types of Galβ-and PRP-related adhesion properties [14,27]. However, the fimA gene has so far only been sequenced from a single strain of both genospecies 1 (12104) and 2 (T14V) [24,25].
A. odontolyticus is a prominent member on the tongue as well as present at supra-and subgingival sites [1,4]. The fimbrial structure of A. odontolyticus, and host receptors, employed for its adhesion have not been fully investigated. However, inhibition studies show that sialylated carbohydrate structures, such as sialyl Tn and 3' sialyllac-tose structures, serve as a salivary glycoprotein gp-340 receptor for A. odontolyticus strain PK984 [28], which is a reference strain for coaggregation group E. Moreover, hybridization studies [14] have indicated FimA-or FimPrelated adhesins on A. odontolyticus, but fimA or fimP genes have not yet been identified or sequenced from A. odontolyticus.
The aim of this study was to characterize fimA genes from several strains of A. naeslundii genospecies 1 and 2 with variant Galβ binding specificities and from a strain of A. odontolyticus PK984 with a sialic acid binding specificity.

A. naeslundii genospecies 1 and 2 and A. odontolyticus PK984 display deviating cell binding properties
A. naeslundii genospecies 1 and 2 express type-2 fimbriae with variant types of Galβ binding specificity and A. odontolyticus express a sialic acid binding specificity potentially related to type-2 fimbriae (Table 1). Accordingly, while A. odontolyticus PK984 agglutinated untreated but not sialic acid-depleted red blood cells, A. naeslundii genospecies 1 and 2 (strains 12104 and LY7, respectively) agglutinated only sialic acid depleted cells strongly, due to exposure of penultimate Galβ-residues (Table 1). a Subtypes of type-2 fimbriae, types-2:1 to 2:3, among Actinomyces species based on different hybridization patterns of strains with fimAderived DNA probes and binding patterns to carbohydrates, red blood cells and bacterial partners [14]. b Saccharide specificity and receptor ligand (GL= glycolipid, GP = glycoprotein) for each fimbriae subtype or reference strain [20,28]. c Hemagglutination (HA) patterns for each fimbriae subtype or reference strain. Score 0 marks no HA and score 4 strong HA. Similar HA results were obtained with goat and chicken erythrocytes (data not shown).
Structurally variant FimA proteins were found in A. naeslundii genospecies 1 and 2 and A. odontolyticus, respectively ( Table 2, Fig. 1). The FimA sequence identity between the three species was in the 62.8-66.4% range, while the sequence identity between strains of the same species was 88.6-99.6 %. Accordingly, sequence analyses of the FimA proteins, clustered FimA proteins from genospecies 1 and 2 and A. odontolyticus into separate groups, but the two latter species more closely (Fig. 1). In addition, the SpaH fimbriae protein from Corynebacterium diphteriae clustered more closely to FimA than did FimP from A. naeslundii.

Structural features of the novel FimA protein in A. odontolyticus with a sialic acid specificity
The fimA gene of A.odontolyticus PK984 encodes a 535 amino acid protein (Fig. 2). The FimA protein contains: i) an N-terminal signal peptide with a signal peptidase cleavage site, ii) a pilin motif for polymerisation of subunit monomers, iii) a proline-rich segment, iv) an E box motif, v) an LPXTG sorting motif, and vi) a C-terminal cell membrane spanning domain.
Cleavage of the 535 amino acid FimA protein of A.odontolyticus between residues 32 and 33 would generate a fimbrial subunit protein of 503 residues. This gives a theoretical subunit molecular weight of 52.6 kDa.

Species-specific and conserved segments in FimA proteins from A. naeslundii genospecies 1 and 2 and A. odontolyticus
The FimA proteins from strains of A. naeslundii genospecies 1 (n = 4) and 2 (n = 4) and A. odontolyticus (n = 1) were compared for FimA sequences or segments conserved between the species or unique to each species (Fig. 3A). The generally high sequence identity (>97.2%) within a genospecies (Table 3) revealed species specific sequences or segments that coincided with binding specificity.

Dendrogram of FimA proteins
Multiple FimA segments showed a low sequence identity between the species.

Structural comparison of FimA and FimP proteins from A. naeslundii, A. odontolyticus and A. viscosus
The FimP (from A. naeslundii genospecies 2, n = 2, and A. viscosus, n = 1) and FimA protein families contained conserved N-terminal signal peptides and C-terminal membrane spanning segments, but with low sequence identity to each other (~24 %) (Fig 3A). High identity sequences in both FimA and FimP are the proline-containing domains 2, 3, 4 (pilin motif), 6 (E box) and 7 (LPXTG). In contrast, the proline-containing regions 1 and 5 are conserved domains in FimP but show only conserved proline residues in FimA.

FimA from A. naeslundii genospecies 1 and 2 and A. odontolyticus display different antigenic properties
Strains of A. naeslundii genospecies 1 and 2 and A. odontolyticus were tested for reactivity with antisera R70-3 specific to FimA from type-2 fimbriae of genospecies 1 strain 12104 using whole cell agglutination and Western blot analyses with whole cell proteins (Fig. 4). While all strains of A. naeslundii genospecies 1, except for strain Pn-6-N, were agglutinated by the antisera, neither strains of genospecies 2 nor A. odontolyticus PK984 were agglutinated (Fig. 4A). Thus, the native FimA protein variants possess different antigenic properties and potentially different overall folding patterns.
Moreover, only genospecies 1 strains possessed positive FimA protein bands in Western blot analyses with type 2 fimbriae specific antisera (Fig. 4). Besides the FimA monomer (no. 1) and polymers (no. 3) detected in all genospecies 1 strains, di-to oligomers of FimA (no.2) were also detactable, suggesting either oligomerization of FimA  subunits or the presence of polymeric fragments of covalently tathered FimA subunits.

Oligomerization of recombinant FimA protein
To further explore the possible oligomerization of FimA in solution, we expressed and purified the FimA protein from A. naeslundii strain 12104 as a recombinant protein and analysed its ability to oligomerize by gel electrophoresis and Western blot analyses using FimA specific antisera (Fig 4C). Gels of the recombinant FimA protein revealed monomers (no. 1) and oligomers (no. 2) of FimA, as confirmed by Western blotting, verifying the possibility of FimA di-to oligomerization in solution but dependence on whole cells for fimbriae polymerization.

Discussion
The present study shows three unique subgroups of FimA proteins present in A. naeslundii genospecies 1 and 2 and A. odontolyticus with different glycoconjugate receptors. It therefore supports our hypothesis that commensal micro organisms, like the genus Actinomyces, exhibit complex and divergent mosaics of adhesin molecules related to species or subpopulations with different tropism and ecological niches. Notably, A. naeslundii genospecies 1 and 2 and A. odontolyticus are members of coaggregation groups A/F, B/C/D and E, respectively, and differ in a variety of type-1, type-2 and other adhesive properties [14]. The FimA protein, which contains both sequences unique to and conserved between the species, may have evolved to match specific niches in supra-or subgingival plaque or in buccal or tongue epithelial tissues. The novel FimA protein from A. odontolyticus strain PK984, a reference strain for Actinomyces-Streptococcus coaggregations typical of sub-gingival plaque, may mediate coaggregations or adhesive interactions typical of subgingival plaque. Strains of A. odontolyticus from the tongue display FimA and FimP hybridization patterns somewhat different from that observed for strain PK984 [14].
All three FimA protein variants contained a pilin motif for polymerization, E box for associated proteins and LPXTG motif for cell surface sorting and anchoring. The pilin, E box and LPXTG motifs were highly conserved among the three species, and present among the proline-containing domains suggested by Yeung and Cisar to account for folding or intermolecular interactions of FimA and FimP proteins [25]. Interestingly, the pilin, E box and LPXTG motifs are present in fimbriae proteins from C. diphtheriae, which have been used to express type-2 fimbriae from A. naeslundii, and in various other Gram-positive bacteria expressing pili-like structures [29,30]. This suggests that Gram-positive bacteria may have evolved related proteins and pathways for pili formation and function [31]. Moreover, the serologically different properties of the FimA proteins present in genospecies 1 and 2 and A. odontolyticus could imply that the overall folding of the FimA protein could form different functionalities or binding specificities. Actually, the variant FimA proteins did not cross-react with antisera despite many conserved motifs, which consequently may be hidden within the subunit or by intermolecular interactions. Apart from the sequences unique to each species, we could not link any FimA motifs or substitutions to the species-specific binding specificities or variant HA patterns of each genospecies. However, receptor-binding FimA subunit domains or tip-localised adhesins other than the FimA subunit remains to be iden- tified. In this respect the unique presence of a proline-rich segment in FimA but not in FimP proteins is interesting, but of yet unknown biological significance.
Another interesting finding of the present study is the possible sortase-independent oligomerization of FimA in solution. Western blot analyses revealed FimA oligomers in whole cell protein extracts and, more importantly, mono to oligomers of FimA in purified recombinant FimA preparations. Sortase, which is absent in Gram-negative E. coli used for expression of recombinant FimA protein, catalyses the covalent tethering of FimA subunits through pilin and LPXTG motifs when expressed in Grampositive C. diphtheriae [29]. It remains, however, to be determined whether the pilin motif and/or other con-served subunit motifs are involved in this sortase-independent ability of FimA to oligomerize in solution.
Finally, it is reasonable to assume that the spontaneous ability of FimA to di-to oligomerize acts in conjunction with sortase in the whole cell-dependent process of fimbriae assembly and polymerization.
The unique and conserved nature of FimA for each Actinomyces species or subspecies reinforces the important role of FimA in selection of ecological niches. Moreover, the highly conserved nature of FimA within a subspecies could also indicate an immunological tolerance to this protein on Actinomyces species that early colonize the oral cavity of infants [3]. The conserved nature of FimA is different to the antigenic variation found in P-fimbriae-associated PapG adhesins on uropathogenic Escherichia coli [32]. Based on its unique and conserved nature, we have previously designed DNA probes from particular FimA segments to distinguish between clinical isolates of genospecies 1 and 2. We assume that corresponding FimA segments unique to A. odontolyticus could be used in a similar way to generate probes specific to particular receptorbinding subtypes of this organism.
While different Galβ specificities of genospecies 1 and 2 target the two species to glycolipid or glycoprotein receptors [20], respectively, A. odontolyticus PK984 recognizes sialic acid residues on glycoproteins. Both glycolipids and glycoproteins are capable of mediating adhesion and intra-cellular signalling by epithelial cells. The host responses mediated by a glycolipid or a glycoprotein receptor may be different, and hypothetically relate to the commensal or pathogenic potential of Actinomyces subtypes. Moreover, the Galβ and sialic acid binding specificities may target A. naeslundii and A. odontolyticus to different bacterial partners, in particular since they belong to different coaggregation groups. In this respect, it is notable that sialidase, which modifies the hemagglutination properties of A. naeslundii and A. odontolyticus in the opposite way, is produced by Actinomyces and other plaque bacteria [33]. However, whether A. naeslundii or related coaggregation communities use sialidase to compete with A. odontolyticus in vivo is unknown. Finally, A. naeslundii and A. odontolyticus are interesting model bacteria for studying the role of fimbriae proteins and their receptors specificities in microbial colonisation and hostmicrobe interactions.

Conclusion
This report shows three unique subgroups of FimA proteins in A. naeslundii genospecies 1 and 2 and A. odontolyticus, and that each subgroup coincides with a unique carbohydrate binding specificity. All FimA proteins contained a pilin motif for polymerization, E box, and LPXTG motif for cell surface sorting and anchoring. Finally, a sor-

Cloning and sequencing of fimA genes
Chromosomal DNA was isolated from bacteria and purified as described [14]. Gene segments were amplified by PCR, using standard protocols, by use of primers both inside and outside the fimA open reading frame (primer sequences are available upon request

Cloning, expression and purification of recombinant FimA
The fimA gene was PCR-amplified from genomic DNA from strain 12104 and cloned into the expression vector pETM11 (EMBL, Hamburg, Germany). The resulting construct encodes a protein (recombinant FimA or rFimA) with an N-terminal hexa-histidine tag, an 18 residue long linker and the FimA protein excluding the N-terminal signal sequence and the C-terminal transmembrane helix. E. coli BL21 (DE3) (Novagen, Madison, WI) was transformed with the pETM11-rFimA construct and grown at 37°C to optical density (OD 600 ) of 0,5. Protein expression was induced with 0.5 mM isopropyl β-D-thiogalactoside for four hours at 30°C. Cultures were harvested and the cells lysed by sonication. rFimA was purified by Ni-agarose chromatography (Quiagen) and elution with imidazole. The protein was further purified by a size exclusion on a Superdex 26/60 column (Amersham, Uppsala, Sweden). The rFimA fractions were analyzed by SDS-PAGE and by Western blot.

Hemagglutination
Equal volumes (10 µml of each) of suspensions of bacterial cells (3 × 10 9 cells/ml PBS or reciprocal dilutions) and human, goat or chicken erythrocytes (4 % erythrocyte suspension in PBS) were mixed and agitated gently for 1 min on a glas slide. Erythrocytes were depleted of sialic acid by incubation with 1 unit/ml Clostridium perfringens neuraminidase (Sigma Chemical Co, St Louis, MO) for 30 min. at 37°C. Agglutination was scored visually; 0 = no agglutination, 1 = weak agglutination, 2 = moderate agglutination, 3 = strong agglutination, and 4 = very strong agglutination.

Authors' contributions
MD, KH and UÖ performed laboratory analyses, analysed data and participated in writing the manuscript. AB sequenced and analysed the fimA gene from PK984 and KP expressed the rFimA protein. IJ and NS designed and planned the project, as well as were responsible for writing the manuscript, at an overall level.
Publish with Bio Med Central and every scientist can read your work free of charge