Abstract

Lipoic acid is a covalently bound disulfide-containing cofactor required for function of the pyruvate dehydrogenase, alpha-ketoglutarate dehydrogenase, and glycine cleavage enzyme complexes of Escherichia coli. Recently we described the isolation of the lplA locus, the first gene known to encode a lipoyl-protein ligase for the attachment of lipoyl groups to lipoate-dependent apoenzymes (T. W. Morris, K. E. Reed, and J. E. Cronan, Jr., J. Biol. Chem. 269:16091-16100, 1994). Here, we report an unexpected redundancy between the functions of lplA and lipB, a gene previously identified as a putative lipoate biosynthetic locus. First, analysis of lplA null mutants revealed the existence of a second lipoyl ligase enzyme. We found that lplA null mutants displayed no growth defects unless combined with lipA (lipoate synthesis) or lipB mutations and that overexpression of wild-type LplA suppressed lipB null mutations. Assays of growth, transport, lipoyl-protein content, and apoprotein modification demonstrated that lplA encoded a ligase for the incorporation of exogenously supplied lipoate, whereas lipB was required for function of the second lipoyl ligase, which utilizes lipoyl groups generated via endogenous (lipA-mediated) biosynthesis. The lipB-dependent ligase was further shown to cause the accumulation of aberrantly modified octanoyl-proteins in lipoate-deficient cells. Lipoate uptake assays of strains that overproduced lipoate-accepting apoproteins also demonstrated coupling between transport and the subsequent ligation of lipoate to apoprotein by the LplA enzyme. Although mutations in two genes (fadD and fadL) involved in fatty acid failed to affect lipoate utilization, disruption of the smp gene severely decreased lipoate utilization. DNA sequencing of the previously identified slr1 selenolipoate resistance mutation (K. E. Reed, T. W. Morris, and J. E. Cronan, Jr., Proc. Natl. Acad. Sci. USA 91:3720-3724, 1994) showed this mutation (now called lplA1) to be a G76S substitution in the LplA ligase. When compared with the wild-type allele, the cloned lplA1 allele conferred a threefold increase in the ability to discriminate against the selenium-containing analog. These results support a two-pathway/two-ligase model of lipoate metabolism in E. coli.