Mechanistic and biochemical studies of the β-lactam forming condensation domain involved in nocardicin A biosynthesis

Nocardicin A is a monocyclic β-lactam antibiotic that has demonstrated resistance to β-lactamases, an important factor in the rapid emergence of antibiotic-resistant pathogenic bacteria. In addition, recent discoveries of antiviral activity stemming from other monocyclic azetidinones have contributed to a resurgence of interest to this class of β-lactams. Nocardicin A is composed of a tripeptide core of non-ribosomal origin, in which the β-lactam ring is derived from L-serine. Contrary to the 3 module non-ribosomal peptide synthetase (NRPS) system that would be expected to underlie its biosynthesis, the nocardicin A gene cluster encodes two NRPSs, NocA and NocB, that together comprise 5 modules. In vivo mutagenesis studies demonstrated that all 5 modules are required for antibiotic production, and a series of predicted peptide thioester substrates screened against the nocardicin thioesterase established NocA and NocB product as a L-p-(hydroxy)phenylglycine[pHPG]-L-Arg-D-pHPG-L-Ser-D-pHPG β-lactam containing pentapeptide, pro-nocardicin G. Following this unexpected discovery, interest shifted towards identification of the protein responsible for β-lactam ring formation. No plausible candidate auxiliary enzyme, however, in the nocardicin A gene cluster could be identified to mediate β-lactam formation. Characterization of the termination module of NocB (M5) alone was carried out to determine its role in the biosynthesis. Pro-nocardicin G production was observed with an in vitro reconstitution system expressing M5 activity, and the C domain (C5) was identified as the likely catalytic locus of azetidinone synthesis. A mechanism was proposed in accord with C domain general acid/base chemistry in which β-elimination of the seryl hydroxyl is the first step. β-Addition then of the downstream pHPG (C-N bond formation) into the resulting dehdyroalanyl residue takes place with retention of configuration at the seryl α-carbon, but stereochemical inversion at its β-carbon, followed by cyclization and release from the peptidyl carrier protein of module 4 (PCP4) to leave a pentapeptide bearing an embedded β-lactam ring tethered to the PCP of module 5. In vitro reconstitution of M5 activity with the proposed dehydroalanyl tetrapeptide intermediate supported the proposed mechanism. Mechanistic studies described in this dissertation identified an E1cb mechanism, in which abstraction of the seryl α-hydrogen is rate-effecting. Lastly, M5 substrate promiscuity studies have demonstrated that simplified thioester surrogates of the native tetrapeptide bound to PCP4 are not sufficient for recognition by M5, indicating that binding interactions between PCP4 and C5 are essential for catalysis. In addition, the N-terminal L-pHPG-L-Arg dipeptide has also been found to be absolutely required for catalysis, despite not being retained in the final natural product. Even relatively conservative replacements of the N-terminal dipeptide are not tolerated well, leading to a drastic decrease of M5 synthetic efficiency. In contrast, C5 shows modestly efficient conversion of non-native substrates in which the internal D-pHPG residue is substituted with phenylglycine and p-fluorophenylglycine. Substitution of serine with threonine to generate a pro-nocardicin G analog having a methyl group at C-4 of the azetidinone ring failed. However, allo-threonine was successfully converted. This result indicates that elimination of the hydroxyl is stereospecific and highlights the potential of C5 for engineering of novel β-lactam containing peptides