Structural studies of the sugar-binding protein from the pneumococcal raffinose transport system

Abstract

Streptococcus pneumoniae (the pneumococcus) is a Gram-positive bacterium responsible for a large number of deaths annually due to pneumonia, septicaemia and meningitis, mainly among young, elderly and immunocompromised populations. The primary virulence factor is the polysaccharide capsule that surrounds the cell and confers protection from phagocytosis; in addition the organism surface is decorated with a variety of proteins attached by both covalent and non-covalent means, with many of these also being involved in virulence. The pneumococcus is highly dependent on a wide range of carbohydrates for energy and growth with both phosphotransferase systems (PTS) and adenosine triphosphate binding cassette (ABC) transport systems utilised in sugar importation. Included among these is an ABC transport system, the Raf system, responsible for the importation and initial metabolism of the trisaccharide raffinose (α-D-Galp-(1→6)-α- D-Glcp-(1→2)-β- D-Fruf) that is highly sequentially homologous to the multiple sugar metabolism (Msm) system from S. mutans. The transporter itself comprises an extracellular raffinose binding protein (RafE), two membrane permease domains (RafF and RafG) and a protein responsible for adenosine triphosphate (ATP) binding and hydrolysis that has yet to be definitively identified. The 46.6 kDa substrate binding protein from the Raf system, RafE, is attached to the surface of the cell by means of a posttranslational lipoprotein modification and is responsible for the initial detection and capture of raffinose and conveying it to the transmembrane domains. RafE has been successfully overexpressed and purified to homogeneity using a novel interaction with a gel filtration matrix. Biophysical characterisation of RafE revealed the protein to be monomeric with one raffinose binding site per molecule and an affinity of 337μM for raffinose. Affinity for melibiose (α-D-Galp-(1→6)-α- D- Glcp), a substrate of the Msm system, was determined to be 6.84mM although the Raf system is incapable melibiose transport. Purified RafE was crystallised in both native and selenomethionine-labelled forms allowing solution of the phase problem by single wavelength anomalous dispersion (SAD) to a resolution of 2.90Å. Subsequent rational truncation and a change of construct to facilitate polyhistidine tag removal has produced two different crystal forms diffracting to a resolution of 1.04Å with the purification tag in place and 1.40Å iii following tag cleavage. An original solution to ice-ring diffraction was utilised for collection of this latter dataset which obviated the need for potentially problematic cryoprotectant solution. The crystal structures of RafE reveal that the protein adopts the periplasmic binding protein-like II fold, in common with a number of other substrate binding components of ABC transport systems, comprising two α/β/α sandwich domains joined by a hinge region consisting of three cross-linking peptide chains with the active site located in the cleft formed between the domains. These structures allowed identification of key active site residues and also revealed a range of conformational motion between the two domains. The active site is formed from a large aromatic patch, formed from three tryptophan residues aligned with their aromatic faces exposed to the solvent, surrounded by polar and charged residues. To assess the interaction with raffinose, polyhistidine-tag cleaved RafE was crystallised in the presence of raffinose and diffraction data collected to a resolution of 2.80Å. Structure solution using molecular replacement of the separate domains revealed a substantial conformational shift compared to the apo structures, with the two domains rotated approximately 29o towards each other, closing the active site region and trapping raffinose. RafE forms direct hydrogen bonding contacts between nine residues and the hydroxyl groups of the carbohydrate coupled with stacking of the sugar rings to the aromatic surface of the tryptophan residues. Molecular modelling of MsmE based on the raffinose bound RafE was used to try to assess the differences contributing to different substrate specificity and revealed changes in the residues forming contacts to the galactose moiety of raffinose but these did not appear to fully explain the dissimilar specificities. As an additional project, work was carried out in an attempt to obtain a crystal structure, with a view to functional elucidation of the choline-binding protein CbpI. This protein is a member of a highly important family for pneumococcal virulence with proteins of diverse function sharing a common surface attachment domain that binds to the choline moieties of teichoic and lipoteichoic acids that intersperse the cell wall. CbpI has been successfully overexpressed, purified and crystallised with diffraction data measured to a resolution of 3.50Å. The diffraction data however display very high mosaic spread and structure solution by molecular replacement has not been successful.