Polysialic acids are synthesized by pathogenic bacteria as capsular polysaccharides. These polymers have been implicated in the virulence of some strains of Escherichia coli, which cause neonatal meningitis and urinary tract infections. Polysialic acid is also present in the developing human brain. There has been significant progress in identifying the gene necessary for capsular polysaccharide biosynthesis in gram negative bacteria. The objective of this project is to determine the mechanism of capsular polysaccharide biosynthesis in virulent encapulated bacteria. Our approach is to characterize the structure and function of the enzymes in the pathway and to use them as tools for understanding sialylation in bacteria and humans. Much of the enzymology of polysialic acid capsular polysaccharide synthesis has been done with the a(2-8) polysialyltransferase complex of E. coli K1. Bacteria containing DNA fragments encoding several capsule related genes have been used as a source of enzyme activity. As a model system for investigating the mechanism of capsular glycosyltransferases we have chosen to investigate the K92 a(2-8)(2-9) polysialyltransferase in a genetic background lacking other capsule related genes. The neuS gene encodes this glycosyltransferase and is the only glycosyltransferase to date identified with synthesis of this polymer. We have shown that the K92 neuS gene product can synthesize both a(2,8) and a(2,9) neuNAc linkages in vitro in a background free of other capsule related gene products and confirmed in vivo synthesis of this polymer by 13C-NMR. The K92 neuS polysialyltransferase is associated with the membrane in lysates of cells harboring the neuS gene in expression vectors, but is not active in a detergent solubilized form suggesting that K92 polysialyltransferase requires association with a membrane for activity. Although the enzyme can transfer sialic acid to the nonreducing end of oligosaccharides with either linkage it is unable to initiate chain synthesis without exogenously added polysialic acid. Thus, the polysialyltransferases encoded by neuS is not sufficient for de novo synthesis of polysaccharide but requires another membrane component for initiation. The acceptor specificity of this polysialyltransferase was studied using sialic acid oligosaccharides of various structures as exogenous acceptors. The enzyme can transfer to the non reducing end of all bacteria polysialic acids, but has a definite preference of a(2,8) acceptors. The minimum acceptor is an oligosaccharide containing a disaccharide of neuNAc a(2,8)neuNAc. Gangliosides containing this disialylated oligosaccharide are elongated, while monsialylated gangliosides are not. Disialylgangliosides are better acceptors than short oligosaccharides suggesting a role for lipid in the elongation reaction. The idea of a lipid linked acceptor is further supported by the elongation of a disialyloligosaccharide possessing a hydrophobic aglycon. In order to determine the nature of the acceptor molecule we have used E. coli K1 constructs defective in sialic acid synthesis and lacking a function sialyltransferase as starting material. The state of membranes in such strains should represent the polysaccharide mechinary just prior to the addition fo the first sialic acid. We have extracted membranes from such strains and demonstrated the presence of an acceptor in these extracts. The acceptor is not a protein and behaves chromatographically like a glycolipid. This acceptor will support the formation of polymer by the K92 neuS polysialyltransferase. Our future experiments are directed at characterizing this molecule and determining whihc gene products are responsible for its synthesis.
