Selectivity of post-translational modification in biotinylated proteins:: The carboxy carrier protein of the acetyl-CoA carboxylase of Escherichia coli

Biotin-dependent enzymes contain a biotinyl-lysine residue in a conserved sequence motif, MKM, located in a surface hairpin turn in one of the two beta-sheets that make up the domain. A sub-gene encoding the 82-residue C-terminal biotinyl domain from the biotin carboxy carrier protein of acetyl-CoA carboxylase from Escherichia coli as a fusion protein with glutathione S-transferase was created and over-expressed in E. coli. The biotinyl domain was readily released by cleavage with thrombin. Five mutant domains were created in which the conserved MKM motif was systematically replaced: by MAK and KAM, in which the target lysine is moved one place; by KKM and MKK, in which a second potential site for biotinylation is introduced; and by DKA, the motif found in the correspondingly conserved site of lipoylation in the structurally related lipoyl domains of 2-oxo acid dehydrogenase multienzyme complexes. No biotinylation of the MAK or KAM mutants was observed in vivo or by purified biotinyl protein ligase in vitro; in the KKM and MKK mutants, only one lysine residue, presumed to be that in its native position in the hairpin turn, was found to be biotinylated in vivo and in vitro. The DKA mutant was not biotinylated in vivo, but was partly lipoylated and octanoylated. It was also a poor substrate for lipoylation in vitro catalysed by the E. coli lipoyl protein ligase encoded by the lplA gene. The flanking sequence in the MKM motif is important, but not crucial, and appears to have been conserved in part to be compatible with the subsequent carboxylation reactions of biotin-dependent enzymes. The DKA motif, displayed in the hairpin loop, is sufficient to address lipoylation in E. coli but probably by a pathway different from that mediated by the lplA-dependent ligase. The recognition of the structurally homologous lipoyl and biotinyl domains by the appropriate ligase evidently has a major structural component to it, notably the positioning of the target lysine residue in the exposed hairpin loop, but there appear to be additional recognition sites elsewhere on the domains.