Abstact

The simple but essential azetidinone core of the β-lactam antibiotics is uniquely N-sulfonated in the monobactam subfamily. This feature confers both target binding specificity to inactivate bacterial cell wall biosynthesis (antibiosis) and structural differentiation to elude destruction by metallo-β-lactamases (MBLs). The recent FDA approval of Emblaveo to treat serious bacterial infections combines an established synthetic monobactam aztreonam and avibactam, which additionally blocks serine β-lactamases, to create a broadly effective antibacterial therapeutic. Here we report experiments to capture the native monobactam biosynthetic steps to the natural product sulfazecin with the aim of accessing new monobactams by reprogramming its biosynthetic machinery. In sulfazecin biosynthesis, the β-lactam ring is formed by a nonribosomal peptide synthetase SulM that incorporates l-2,3-diaminopropionate (Dap), which is then N-sulfonated in trans and efficiently cyclized to the fully elaborated monobactam by an unusual thioesterase (TE) domain. We describe an improved synthesis of (2S,3R)-vinylDap to support rational structure-based engineering experiments to obtain the corresponding (4R)-vinyl sulfazecin. While these experiments were initially based on an AlphaFold model of the adenylation domain that incorporates Dap (SulM A3), we further report high-resolution X-ray crystal structures with both the l-Dap substrate and an accurate analogue of the activated (3R)-methyl-Dap adenylate bound. The ligand-bound structures rationalize the inability of SulA3 to incorporate larger substrates. Comparisons with the structures of other diamino acid-activating adenylation domains identify alternate binding modes that may be more suitable for the production of sulfazecin analogues. The impact of these structures on the further engineering of the SulA3 domain and its relation to monobactam synthesis in the recently structurally characterized SulTE are discussed.