Leucoanthocyanidin reductase (LAR) catalyzes the synthesis of catechin, an initiating monomer of condensed tannin or proanthocyanidin (PA) synthesis, from 3,4-cis-leucocyanidin and thus is the first committed step in PA biosynthesis. The enzyme was purified to near homogeneity from PA-rich leaves of the legume Desmodium uncinatum (Jacq.) DC, partially sequenced and the corresponding cDNA cloned. The identity of the enzyme was confirmed by expressing active recombinant LAR in Escherichia coli and in tobacco and white clover. The enzyme is a monomer of 43 kDa (382 amino acids) and is most active synthesizing catechin (specific activity of approximately 10 micromol min+1 mg of protein+1) but also synthesizes afzelechin and gallocatechin. LAR is most closely related to the isoflavone reductase group of plant enzymes that are part of the Reductase-Epimerase-Dehydrogenase (RED) family of proteins. Unlike all other plant isoflavone reductase homologues that are about 320 amino acids long, LAR has an additional 65-amino acid C-terminal extension whose function is not known. Curiously, although Arabidopsis makes PA, there is no obvious LAR orthologue in the Arabidopsis genome. This may be because Arabidopsis seems to produce only an epicatechin, rather than a dual catechin/epicatechin-based PA similar to many other plants.

We cured a strain of Salmonella typhimurium of its cryptic plasmid and confirmed that orally administered cured strains lost virulence for mice. Loss of the cryptic plasmid rendered the S. typhimurium strain sensitive to the bactericidal action of normal human serum. However, loss of the plasmid did not change the ability of the strain to associate with HeLa cells in tissue culture. Furthermore, when administered orally to mice, both the plasmid-containing and plasmid-free strains invaded the Peyer's patches of the small intestine to the same extent, and both were capable of inducing resistance to oral challenge with virulent S. typhimurium. When injected intraperitoneally, the cured strain was eliminated rapidly, whereas the parental strain persisted. We also showed that the cured strain did not contain a plasmid copy in the chromosome. We propose that although the plasmid-cured strain of S. typhimurium is able to colonize Peyer's patches, it cannot survive when administered intraperitoneally because it is susceptible to elimination by macrophages.

On initial isolation of Aeromonas sobria 3767 from a diarrhoeal stool specimen, two colony types were obtained: opaque (3767O) and translucent (3767T). Strain 3767O consistently produced lipopolysaccharide (LPS) core and O-antigen side chain, detectable by SDS-PAGE and by Western blotting with an O-antigen-specific monoclonal antibody. Strain 3767T produced LPS core but the amount of O-antigen was dependent on factors including growth medium and bacterial growth phase. Strain 3767T exhibited significantly lower levels of adhesion to HEp-2 cells than 3767O and this correlated with the level of LPS expression, with the greatest reduction (61%) at stationary phase when no LPS was detectable. The results implicate LPS as an adhesin for A. sobria 3767.