Streptothricin-F (STT-F), one of the early-discovered antibiotics, consists of three components, a β-lysine homopolymer, an aminosugar D-gulosamine, and an unusual bicyclic streptolidine. The biosynthesis of streptolidine is a long-lasting but unresolved puzzle. Herein, a combination of genetic/biochemical/structural approaches was used to unravel this problem. The STT gene cluster was first sequenced from a Streptomyces variant BCRC 12163, wherein two gene products OrfP and OrfR were characterized in vitro to be a dihydroxylase and a cyclase, respectively. Thirteen high-resolution crystal structures for both enzymes in different reaction intermediate states were snapshotted to help elucidate their catalytic mechanisms. OrfP catalyzes an Fe(II) -dependent double hydroxylation reaction converting L-Arg into (3R,4R)-(OH)2 -L-Arg via (3S)-OH-L-Arg, while OrfR catalyzes an unusual PLP-dependent elimination/addition reaction cyclizing (3R,4R)-(OH)2 -L-Arg to the six-membered (4R)-OH-capreomycidine. The biosynthetic mystery finally comes to light as the latter product was incorporation into STT-F by a feeding experiment.

In the search for new efficacious antibiotics, biosynthetic engineering offers attractive opportunities to introduce minor alterations to antibiotic structures that may overcome resistance. Dbv29, a flavin-containing oxidase, catalyzes the four-electron oxidation of a vancomycin-like glycopeptide to yield A40926. Structural and biochemical examination of Dbv29 now provides insights into residues that govern flavinylation and activity, protein conformation and reaction mechanism. In particular, the serendipitous discovery of a reaction intermediate in the crystal structure led us to identify an unexpected opportunity to intercept the normal enzyme mechanism at two different points to create new teicoplanin analogs. Using this method, we synthesized families of antibiotic analogs with amidated and aminated lipid chains, some of which showed marked potency and efficacy against multidrug resistant pathogens. This method offers a new strategy for the development of chemical diversity to combat antibacterial resistance.

In biological systems, methylation is most commonly performed by methyltransferases (MTs) using the electrophilic methyl source S-adenosyl-L-methionine (SAM) via the S(N)2 mechanism. (2S,3S)-β-Methylphenylalanine, a nonproteinogenic amino acid, is a building unit of the glycopeptide antibiotic mannopeptimycin. The gene product of mppJ from the mannopeptimycin-biosynthetic gene cluster is the MT that methylates the benzylic C atom of phenylpyruvate (Ppy) to give βMePpy. Although the benzylic C atom of Ppy is acidic, how its nucleophilicity is further enhanced to become an acceptor for C-methylation has not conclusively been determined. Here, a structural approach is used to address the mechanism of MppJ and to engineer it for new functions. The purified MppJ displays a turquoise colour, implying the presence of a metal ion. The crystal structures reveal MppJ to be the first ferric ion SAM-dependent MT. An additional four structures of binary and ternary complexes illustrate the molecular mechanism for the metal ion-dependent methyltransfer reaction. Overall, MppJ has a nonhaem iron centre that bind, orients and activates the α-ketoacid substrate and has developed a sandwiched bi-water device to avoid the formation of the unwanted reactive oxo-iron(IV) species during the C-methylation reaction. This discovery further prompted the conversion of the MT into a structurally/functionally unrelated new enzyme. Through stepwise mutagenesis and manipulation of coordination chemistry, MppJ was engineered to perform both Lewis acid-assisted hydration and/or O-methyltransfer reactions to give stereospecific new compounds. This process was validated by six crystal structures. The results reported in this study will facilitate the development and design of new biocatalysts for difficult-to-synthesize biochemicals.

Mannopeptimycin, a potent drug lead, has superior activity against difficult-to-treat multidrug-resistant Gram-positive pathogens such as methicillin-resistant Staphylococcus aureus (MRSA). (2S,3S)-beta-Methylphenylalanine is a residue in the cyclic hexapeptide core of mannopeptimycin, but the synthesis of this residue is far from clear. We report here on the reaction order and the stereochemical course of reaction in the formation of (2S,3S)-beta-methylphenylalanine. The reaction is executed by the enzymes MppJ and TyrB, an S-adenosyl methionine (SAM)-dependent methyltransferase and an (S)-aromatic-amino-acid aminotransferase, respectively. Phenylpyruvic acid is methylated by MppJ at its benzylic position at the expense of one equivalent of SAM. The resulting beta-methyl phenylpyruvic acid is then converted to (2S,3S)-beta-methylphenylalanine by TyrB. MppJ was further determined to be regioselective and stereoselective in its catalysis of the formation of (3S)-beta-methylphenylpyruvic acid. The binding constant (K(D)) of MppJ versus SAM is 26 microM. The kinetic constants with respect to k(cat Ppy) and K(M Ppy), and k(cat SAM) and K(M SAM) are 0.8 s(-1) and 2.5 mM, and 8.15 s(-1) and 0.014 mM, respectively. These results suggest SAM has higher binding affinity for MppJ than Ppy, and the C--C bond formation in betamPpy might be the rate-limiting step, as opposed to the C--S bond breakage in SAM.

Lipoglycopeptide antibiotics are more effective than vancomycin against MRSA as they carry an extra aliphatic acyl side chain on glucosamine (Glm) at residue 4 (r4). The biosynthesis of the r4 N-acyl Glc moiety at teicoplanin (Tei) or A40926 has been elucidated, in which the primary amine nucleophile of Glm is freed from the r4 GlcNac pseudo-Tei precursor by Orf2* for the subsequent acylation reaction to occur. In this report, two Orf2* structures in complex with β-D-octyl glucoside or Tei were solved. Of the complexed structures, the substrate binding site and a previously unknown hydrophobic cavity were revealed, wherein r4 GlcNac acts as the key signature for molecular recognition and the cavity allows substrates carrying longer acyl side chains in addition to the acetyl group. On the basis of the complexed structures, a triple-mutation mutant S98A/V121A/F193Y is able to regioselectively deacetylate r6 GlcNac pseudo-Tei instead of that at r4. Thereby, novel analogs can be made at the r6 sugar moiety.