Sequence diversity in the A domain of Staphylococcus aureus fibronectin-binding protein A

Abstract

Background: Fibronectin-binding protein A (FnBPA) mediates adhesion of Staphylococcus aureus to fibronectin, fibrinogen and elastin. We previously reported that S. aureus strain P1 encodes an FnBPA protein where the fibrinogen/elastin-binding domain (A domain) is substantially divergent in amino acid sequence from the archetypal FnBPA of S. aureus NCTC8325, and that these variations created differences in antigenicity. In this study strains from multilocus sequence types (MLST) that spanned the genetic diversity of S.aureus were examined to determine the extent of FnBPA A domain variation within the S. aureus population and its effect on ligand binding and immuno-crossreactivity.

Results: Seven different isotype forms (I - VII) of the FnBPA A domain were identified which were between 66 to 76% identical in amino acid sequence in any pair-wise alignment. The fnbA allelic variants in strains of different multilocus sequence type were identified by DNA hybridization using probes specific for sequences encoding the highly divergent N3 sub-domain of different isotypes. Several isotypes were not restricted to specific clones or clonal complexes but were more widely distributed. It is highly likely that certain fnbA genes have been transferred horizontally. Residues lining the putative ligand-binding trench were conserved, which is consistent with the ability of each A domain isotype to bind immobilized fibrinogen and elastin by the dock-latch-lock mechanism. Variant amino acid residues were mapped on a three-dimensional model of the FnBPA A domain and were predicted to be surface-exposed. Polyclonal antibodies raised against the recombinant isotype I A domain bound that protein with a 4 - 7 fold higher apparent affinity compared to the A domains of isotypes II - VII, while some monoclonal antibodies generated against the isotype I A domain showed reduced or no binding to the other isotypes.

Conclusion: The FnBPA A domain occurs in at least 7 different isotypes which differ antigenically and exhibit limited immuno-crossreactivity, yet retain their ligand-binding functions. Antigenic variation of the FnBPA A domain may aid S. aureus to evade the host's immune responses. These findings have implications for the development of vaccines or immunotherapeutics that target FnBPA.

Figure 1

Schematic representation of FnBPA from S. aureus 8325-4. (a) The N-terminus of FnBPA contains a signal sequence (S) followed by a 475 amino acid A domain, part of which is the binding site for fibrinogen and elastin. Adjacent to the A domain are eleven fibronectin-binding motifs that span residues 512–878 [27]. The C-terminus of FnBPA contains a proline-rich repeat (PRR), wall (W) and membrane (M)-spanning domains, and the sortase recognition motif LPETG. (b) The FnBPA A domain is predicted to comprise three independently-folded sub-domains N1 (37–193), N2 (194–337), and N3 (338–511) [15]. The His-tagged recombinant A domain isotype proteins (I – VII) analyzed during this study comprise the N2 and N3 sub-domains. Type-specific DNA probes used for typing fnbA genes from clinical isolates by DNA hybridization contain approximately 300 nucleotides that encode the central portion of the highly divergent N3 sub-domain. The diagram also indicates positions of general primers used for amplifying fnbA gene sequences that encode the entire A domain plus flanking sequence. These fnbA gene fragments were used as templates in DNA hybridization experiments with type specific probes, and for sequencing from a select number of strains.

Figure 2

FnBPA A domain typing of S. aureus strains by dot blot hybridization. Genomic DNA from S. aureus isolates was purified, and the fnbA gene fragment encoding the entire A domain was amplified by PCR. Five ng purified fnbA DNA from each isolate was spotted onto nitrocellulose membranes and probed with DIG-labelled fnbA type-specific probes corresponding to isotype I (a), isotype II (b), isotype III (c), isotype IV (d) and isotype V (e). fnbA DNA from isolates 8325-4, MRSA252, MSSA476, P1 and 3110 (top row of blots) were used as controls. The MLST genotypes of strains typed here are given in Table 1.

Figure 3

Evolutionary relationships of fnbA N1, N2 and N3 allele sequences. fnbA gene segments encoding the entire FnBPA A domain were sequenced from 14 isolates from divserse MLST genotypes. Strain P1 (ST-973)fnbA was sequenced previously [15]. In addition, the fnbA gene sequences from MSSA 476 (ST-1); N315 (ST-5); MRSA252 (ST-36); COL (ST-250); RF122 (ST-151) as well as the archetype fnbA sequence from 8325-4 (ST-8) were included. The evolutionary history was inferred using the Neighbor-Joining method [43]. Neighbour-joining trees were constructed using fnbA DNA sequences encoding the sub-domains N1 (a), N2 (b) and N3 (c) using MEGA 3. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (500 replicates) are shown next to the branches [44]. The phylogenetic tree was linearized assuming equal evolutionary rates in all lineages [45]. The evolutionary distances were computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site.

Figure 4

Neighbour-joining phylogentic tree of concatenated MLST allele sequences. Concatenated MLST allele sequences representing each strain were downloaded from the MLST database [39] and used to generate a neighbour joining tree using MEGA 3 [40]. The A domain isotypes carried by strains of each MLST genotype, determined by sequencing and hybridization analysis, are indicated. FnBPA A domain types from strains of ST-80 and ST-45 were determined from sequence data from a previous study of polymorphic genes in S. aureus [21]. The dashed line indicates the separation of the MLST genotypes into Groups 1 and 2, which is based on sequence data from MLST alleles and other unlinked loci [23].

Figure 5

Binding of recombinant A domain isoypes I – VII to immobilized fibrinogen and elastin. Microtitre dishes were coated with human fibrinogen (a) or human aortic elastin (b). Wells were blocked and then incubated with recombinant A domain proteins isotype I (closed squares), isotype II (closed circles), isotype III (open squares), isoype IV (open circles), isoype V (closed triangles), isoype VI (open diamonds), and isoype VII (open triangles) at the indicated concentrations. Bound A domain proteins were detected by incubation with mouse anti-hexahistidine monoclonal antibody 7E8, then HRP-conjugated goat anti-mouse IgG antibodies followed by TMB substrate. Graphs are representative of three separate experiments.

Figure 6

Binding of polyclonal and monoclonal anti-isotype I antibodies to A domain isotypes I – VII. Microtitre dishes were coated with A domains isoype I (closed squares), isotype II (closed circles), isotype III (open squares), isotype IV (open circles), isotype V (closed triangles), isotype VI (open diamonds), and isotype VII (open triangles) at the indicated concentrations. Wells were then blocked and then incubated with (a) polyclonal rabbit anti-isotype I A domain antibodies, or mouse monoclonal anti-isotype I A domain antibodies 1F9 (b), 1G8 (c), 7C5 (d), 1E6 (e), and 7B7 (f). Bound antibodies were detected with either HRP-conjugated goat anti-rabbit IgG antibodies (a) or HRP-conjugated goat anti-mouse IgG antibodies (b-f), followed by TMB substrate. Graphs are representative of three separate experiments.