The work presented in this thesis elucidates the terminal steps in the biosynthesis of the novel antifungal agent, syringomycin E. In examining these biosynthetic steps, we illustrate key strategies used by non-ribosomal peptide synthetase (NRPS) systems in generating structural diversity in peptidic natural products. Syringomycin E, a cyclic lipodepsinonapeptide produced by the plant-associated pathogen, Pseudomonas syringae pv. syringae, contains a number of non-proteinogenic amino acids that are essential for its biological activity. The terminal tripeptide of syringomycin is particularly interesting both because it contains the novel amino acids dehydro-threonine, L-threo-3-hydroxy-Asp, and 4-chlorothreonine, and because it is conserved in a number of related phytotoxic lipodepsipeptides.
Our study of the eighth step in syringomycin E biosynthesis focused on elucidating the timing and mechanism of the generation of L-threo -3-hydroxy-Asp. The syringomycin gene cluster contained no annotated aspartyl beta hydroxylase, and our search of the genome of P. syringae pv. syringae turned up only one aspartyl hydroxylase gene, aspH, located far upstream of the syringomycin gene cluster. Biochemical characterization of AspH revealed it to be a member of the iron- and alpha-ketoglutarate-dependent family of dioxygenases, and we found that it was active on a number of thioester-linked Asp-containing substrates. However, AspH generated exclusively the L-erythro diastereomer of 3-hydroxy-Asp, and therefore could not be involved in syringomycin biosynthesis. A closer investigation of the syringomycin gene cluster revealed that that syrP, a member of the syringomycin gene cluster that was previously annotated as a regulatory protein, actually encoded an enzyme with significant homology to known iron- and alpha-ketoglutarate-dependent dioxygenases. Biochemical analysis of SyrP led us to determine that it is an iron- and alpha-ketoglutarate-dependent dioxygenase that generates the L-threo diastereomer of 3-hydroxy-Asp. Furthermore, SyrP showed activity only on L-Asp tethered via a thioester linkage to a fragment of the eighth domain of the SyrE NRPS megasynthetase, upon which syringomycin is assembled. In combination, these results indicated that SyrP is the dioxygenase responsible for generation of 3-OH-Asp in syringomycin biosynthesis and that it acts via the mechanism that is common to iron- and alpha-ketoglutarate dependent dioxygenases. Furthermore, these results indicate that hydroxylation occurs while L-Asp is tethered to the SyrE megasynthetase, and therefore does not occur pre- or post-assembly line.
The generation of 4-chlorothreonine in the ninth step of syringomycin biosynthesis also involves several unusual strategies. Prior studies determined that 4-chlorothreonine is generated by the action of the iron and alpha-ketoglutarate-dependent halogenase, SyrB2, which acts on L-threonine tethered via a thioester linkage to the didomain protein SyrB1. However, the ninth module of the SyrE megasynthetase lacks an adenylation module, so it was unclear as to how 4-chlorothreonine is moved from the thiolation domain of SyrB1 to the ninth thiolation domain of the SyrE megasynthetase. The enzyme SyrC, encoded by a gene in the syringomycin cluster, was previously hypothesized to act as an acyltransfase to lipidate the N-terminus of Seri in syringomycin biosynthesis. However, using autoradiographic assays, we determined that SyrC is actually responsible for transferring 4-chlorothreonine from SyrB1 to the ninth module of SyrE. We furthermore determined that SyrC is capable of transferring non-native amino acids, such as L-Val and L-Leu onto SyrE, suggesting that SyrC could be useful in combinatorial generation of novel NRPS products.
Our examination of the tenth step of syringomycin biosynthesis focused on macrolactonization of the linear syringomycin nonapeptide catalyzed by the C-terminal TE domain of the SyrE megasynthetase. The SyrE-TE catalyzes this cyclization reaction by facilitating the formation of an ester linkage between the terminal 4-chlorothreonine residue and the hydroxyl side chain of the Ser1 residue. To better understand the substrate parameters that mediate this cyclization, we synthesized a number of substrate analogs that varied in N-terminal acylation state, peptide composition, and length and then characterized the Syr-TE activity with each of the substrates generated. The results of these studies sets the stage for comparative analysis of the Syr-TE with other known NRPS TE domains.