Abstract
Early steps in the biosynthesis of lipopolysaccharide (LPS) involve the transfer of 3-deoxy-α-D-manno-oct-2-ulopyranosonic acid (Kdo) to lipid A. Whereas Kdo transferases (WaaA) of Escherichia coli generate a (2 → 4)-linked Kdo disaccharide, Chlamydiae contain tri- or tetra-functional WaaA generating oligosaccharides with (2 → 8)- and (2 → 4)-linkages between Kdo. It has been suggested that the transfer of L-glycero-α-D-manno-heptose (Hep) to Kdo by an E. coli WaaC may not be possible in the presence of (2 → 8)-linked Kdo. E. coli double-mutants deficient in heptosyltransferases I (waaC) and II (waaF) and expressing waaA of Chlamydiae instead of their own, make Chlamydia-type Kdo oligosaccharides which are attached to an E. coli lipid A. Using such strains expressing waaA of Chlamydophila pneumoniae, Chlamydophila psittaci, or Chlamydia trachomatis, we have studied the effect of E. coli waaC gene expression on LPS structure. Structural analyses revealed the formation of two novel oligosaccharides Hep-(1 → 5)[Kdo-(2 → 4)]-Kdo and Hep-(1 → 5)[Kdo-(2 → 8)-Kdo-(2 → 4)]-Kdo showing that Hep is transferred in the presence of (2 → 8)-linked Kdo. Surprisingly, the transfer of Hep onto Kdo-(2 → 4)-Kdo-(2 → 4)-Kdo did not occur, despite the fact that Hep-(1 → 5)[Kdo-(2 → 4)-Kdo-(2 → 4)]-Kdo is found in nature as a partial structure of E. coli LPS. The premature end of the biosynthesis and incorporation of Hep into the LPS indicated that WaaC had access to the substrate before Kdo transfer was completed. We have observed differences between WaaA of C. trachomatis, C. pneumoniae and C. psittaci which indicate mechanistic differences between these Kdo transferases.