Mutations in Two Putative Glycosyltransferase Genes in Rhizobium etli

Rhizobium elli is a microbe that can live in symbiosis with the bean plant, Phaseolus vulgaris. The lipopolysaccharide (LPS) of the bacterium is known to play a key role in this interaction. Bacteria deficient in LPS have been shown to be deficient in the symbiosis. The wildtype bacteria alter their LPS in the presence of the plant host, and one of the changes is the loss of the residue normally found at the LPS 0-antigen terminus, di-0-methylfucose (DOMfucose ). In order to study the regulatory mechanisms responsible for this host-induced alteration, the gene encoding the transferase for this terminal 0-antigen residue was desired. Recent sequencing of DNA adjacent to the R. etli lpe genetic locus revealed two putative glycosyltransferase ORFs, which were named lpeB and IpsE. It was hypothesized that one of the putative ORFs encoded the transferase responsible for adding DOMfucose to the terminus of the 0-antigen. To test this hypothesis, these ORFs were mutated by the insertion of a DNA cassette that confers kanamycin resistance. Polymerase cJtain reaction (PCR) analysis indicated that lpeB had been mutated as expected in mutant strains CE468 and CE469. These mutants produced abundant 0-antigencontaining LPS, but two lines of evidence indicated that the terminal DOMfucose residue was absent from the mutant LPS. First, the mutants failed to bind to three monoclonal antibodies used to characterize R. etli LPS. Second, analysis of the LPS sugar composition by gas chromatography revealed the absence of a peak at the elution time of DOMfucose. The lpeB mutants performed as well as wild type CE3 in symbiotic efficiency assays, which was expected because the terminal residue is absent in the wild type interaction. From these results it was concluded that lpeB was a good candidate to encode the DOMfucosyltransferase. The other ORF (IpsE) was mutated by similar means. IpsE mutants produced little or no 0-antigen-containing LPS. This result is consistent with the hypothesis that IpsE encodes a transferase which adds a fucose deeper within the 0-antigen, rather than being the transferase for the terminal residue