Structure of the Streptococcus pneumoniae Surface Protein and Adhesin PfbA

Citation: Suits MD, Boraston AB ( Structure of the Streptococcus pneumoniae Surface Protein and Adhesin PfbA Michael D. Suits 0 1 Alisdair B. Boraston 0 1 Renwick Dobson, University of Canterbury, New Zealand 0 Funding: This work was supported by a Canadian Institutes of Health Research Operating Grant (FRN 86610). MDLS is a Michael Smith Foundation for Health Research (MSFHR) Scholar. ABB is an E.W.R Steacie Memorial Fellow and a MSFHR Scholar. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript 1 Department of Biochemistry and Microbiology, University of Victoria , Victoria, British Columbia , Canada PfbA (plasmin- and fibronectin-binding protein A) is an extracellular Streptococcus pneumoniae cell-wall attached surface protein that binds to fibronectin, plasmin, and plasminogen. Here we present a structural analysis of the surface exposed domains of PfbA using a combined approach of X-ray crystallography and small-angle X-ray scattering (SAXS). The crystal structure of the PfbA core domain, here called PfbAb, determined to 2.28 A resolution revealed an elongated 12-stranded parallel b-helix fold, which structure-based comparisons reveal is most similar to proteins with carbohydrate modifying activity. A notable feature of the PfbAb is an extensive cleft on one face of the protein with electrochemical and spatial features that are analogous to structurally similar carbohydrate-active enzymes utilizing this feature for substrate accommodation. Though this cleft displays a combination of basic amino acid residues and solvent exposed aromatic amino acids that are distinct features for recognition of carbohydrates, no obvious arrangement of amino acid side chains that would constitute catalytic machinery is evident. The pseudo-atomic SAXS model of a larger fragment of PfbA suggests that it has a relatively well-ordered structure with the N-terminal and core domains of PfbA adopting an extend organization and reveals a novel structural class of surface exposed pneumococcal matrix molecule adhesins. - Streptococcus pneumoniae (the pneumococcus) is a transient colonizer of the human nasopharynx. Though this is typically a benign relationship, the bacterium can, under appropriate but not entirely understood circumstances, switch into the role of a proficient invasive pathogen. Mediating the host-bacterium interaction is the complex extracellular landscape of the pneumococcus that includes both covalently and non-covalently linked proteins that perform a wide variety of functions [1]. Approximately fifteen of these proteins are known or predicted to be attached to the bacterial cell-wall through a sortase-dependent mechanism requiring an LPXTG sequence motif in addition to a secretion signal peptide [2]. In S. pneumoniae TIGR4 these specific cell-surface attached proteins comprise seven carbohydrate-active enzymes, four proteases, one mucin specific adhesin, two adhesins with fibronectin binding ability, and one protein of unidentified function [3]. Additionally, a key role for microbial surface components recognizing adhesive matrix molecules has emerged as a central theme for pneumococcal pathogenesis in that six proteins, PavA, PavB, PepO, PfbA, PfbB (recently reviewed in [4]), and RrgA [5] have all been identified to mediate pneumococcal attachment to plasminogen and fibronectin. Given the important biological functions of these surface exposed proteins in the hostbacterium interaction, and the potential they hold in developing vaccine components their structures and functions hold considerable practical therapeutic interest. PfbA (plasmin- and fibronectin-binding protein A) is one of the most recently identified surface attached pneumococcal proteins. The gene encoding this protein is highly conserved across all sequenced pneumococcal isolates, and PfbA was shown to be a constitutively expressed, surface-anchored protein [6]. Recombinant PfbA was found to bind to the human extracellular matrix proteins fibronectin and plasmin with high affinities (KD of 4.1 and 2.4 mM, respectively) [6], and may therefore be classified into the microbial surface cell recognition adhesion matrix molecule (MSCRAMM) family [3,4]. Mutants of S. pneumoniae R6 lacking the pfbA gene were deficient in the capacity for adherence to, and invasion of, human epithelial cells. Moreover, in keeping with the in vitro binding capacity of recombinant PfbA, the adherent and invasive potential of S. pneumoniae R6, but not the DpfbA mutant, was shown to be dependent on the presence of fibronectin. Together, these results indicate that PfbA may be considered an important adhesin in mediating the pneumococcus-host interaction. A full understanding of the biological role of this protein, however, is presently hampered by a lack of structural information. To this end, we have determined the crystal structure of the 422 amino acid core PfbAb domain (residues 139560). Greater insight into the overall structure of the protein is provided by a solution SAXS analysis of a 509 amino acid multi-domain construct of PfbA that includes the PfbAb and an N-terminal domain of unknown function. Together, the results reveal a protein with an extended architecture that contains a core parallel b-helix structure with specific molecular features that we propose Results and Discussion PfbA architecture A gene encoding PfbA is present amongst the majority of pneumococcal strains with the encoded proteins having no less than 99% amino acid sequence identity. Furthermore, homologues of this protein are distributed across a number of streptococcal and staphylococcal species. The N-terminal FSIRK and C-terminal LPXTG sequence motifs of PfbA are common Gram-positive export signal peptide and sortase-mediated cell-wall attachment motifs, respectively, consistent with the cell-surface localization of the protein [6]. To gain more insight into the architecture of the protein we initially used an approach that combined secondary structure prediction with fold recognition [7], which readily revealed a central domain in PfbA that likely adopts a parallel b-helix fold (residues 139560) (Figure 1A). This is consistent with distant amino acid sequence similarity between PfbA and various carbohydrate-active enzyme families including polysaccharide lyase families 1, 3, 6 and 9, and glycoside hydrolase families 28 and 49, and 55, all of which adopt parallel b-helix folds [8]. Because of this predicted sequence and fold similarity, PfbA and its homologues have been placed into the non-classified hydrolase category within the Carbohydrate-Active Enzyme Database [8]. In addition to the dominant b-helix domain, PfbA contains a predicted N-terminal mixed/b-domain (residues 52 138) and a predicted C-terminal, a-helical bundle (residues 561 708) (Figure 1A). Guided by this predicted domain architecture we generated constructs of the protein by dissecting a synthetic gene op (...truncated)