PksA, which initiates biosynthesis of the environmental carcinogen aflatoxin B1, is one of the multidomain iterative polyketide synthases (IPKSs), a large, poorly understood family of biosynthetic enzymes. We found that dissection of PksA and its reconstitution from selected sets of domains allows the accumulation and characterization of advanced octaketide intermediates bound to the enzyme, permitting the reactions controlled by individual catalytic domains to be identified. A product template (PT) domain unites with the ketosynthase and thioesterase in this IPKS system to assemble precisely seven malonyl-derived building blocks to a hexanoyl starter unit and mediate a specific cyclization cascade. Because the PT domain is common among nonreducing IPKSs, these mechanistic features should prove to be general for IPKS-catalyzed production of aromatic polyketides.Tens of thousands of natural products are known from microorganisms, plants, and animals that provide hormones, toxins, flavors and fragrances, pigments, drugs, and other materials of commercial value. A handful of biosynthetic pathways give rise to this rich diversity of useful structures. Among these, polyketides are synthesized from simple acyl-coenzyme A (acylCoA) substrates by polyketide synthases (PKSs) (1). We understand a great deal about the function of giant modular PKSs that synthesize complex products, for example, the antibiotic erythromycin and the immunosup-pressant rapamycin (2). Each catalytic domain in these multidomain (type I) megaproteins is used once in an overall "assembly-line" process as a growing intermediate is advanced along the enzyme to yield a product. In contrast to these bacterial systems, in eukaryotes iterative PKSs (IPKSs) are generally the rule where a markedly smaller number of catalytic domains are similarly fused but individually reused in multiple catalytic cycles (iteration) that are "programmed" to yield specific products. How programming is achieved is a central unanswered question of iterative catalysis.We undertook a "deconstruction" approach by domain dissection and reassembly of PksA, the nonreducing IPKS of aflatoxin biosynthesis, to reveal the global division of labor among the domains in these macromolecular machines to control polyketide chain length, cyclization of an extended poly-β-keto intermediate, and product release. A domain hypothesized to be a "product template" (PT) has been discovered to play a central role in product formation. These studies are combined with high-resolution mass spectrometry (MS) to monitor the highly †To whom correspondence should be addressed.

Polyketides are a class of natural products with diverse structures and biological activities. The structural variability of aromatic products of fungal nonreducing, multidomain iterative polyketide synthases (NR-PKS group of IPKSs) results from regiospecific cyclizations of reactive poly-β-keto intermediates [1][2][3] . How poly-β-keto species are synthesized and stabilized, how their chain lengths are determined, and, in particular, how specific cyclization patterns are controlled have been largely inaccessible and functionally unknown until recently 4 . A product template (PT) domain is responsible for controlling specific aldol cyclization and aromatization of these mature polyketide precursors, but the mechanistic basis is unknown. Here we present the 1.8 Å crystal structure and mutational studies of a dissected PT monodomain from PksA, the NR-PKS that initiates the biosynthesis of the potent hepatocarcinogen aflatoxin B 1 in Aspergillus parasiticus. Despite having minimal sequence similarity to known enzymes, the structure displays a distinct 'double hot dog' (DHD) fold. Co-crystal structures with palmitate or a bicyclic substrate mimic illustrate that PT can bind both linear and bicyclic polyketides. Docking and mutagenesis studies reveal residues important for substrate binding and catalysis, and identify a phosphopantetheine localization channel and a deep two-part interior binding pocket and reaction chamber. Sequence similarity and extensive conservation of active site residues in PT domains suggest that the mechanistic insights gleaned from these studies will prove general for this class of IPKSs, and lay a foundation for defining the molecular rules controlling NR-PKS cyclization specificity.Aflatoxin B 1 (Fig. 1a, 3) biosynthesis is initiated by the NR-PKS PksA, which accepts a hexanoyl starter unit from a dedicated fungal fatty acid synthase (FAS) and extends it through seven iterative, malonyl-derived ketide extensions to norsolorinic acid anthrone ((1), noranthrone) (Fig. 1a)5. Application of the Udwary-Merski algorithm (UMA) afforded the unanticipated insight that PksA was composed of two unrecognized domains, in additionCorrespondence and requests for materials should be addressed to C.A.T. (ctownsend@jhu.edu) or S.-C.T. (sctsai@uci.edu). * These authors contributed equally to this work.Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature.Supplementary Information is linked to the online version of the paper at www.nature.com/nature. (Fig. 1a, c) 4 . When the TE domain was absent, naphthopyrone (4) was produced by self-condensation of the stalled bicyclic PT product 7 (Fig. 1b). Author ContributionsWe crystallized PT in P2 1 2 1 2 1 and P4 1 2 1 2 space groups (Supplementary Table 1) and solved the PT crystal structure by multi-wavelength anomalous dispersion (MAD) using selenomethionine-derivatized PT followed by phase extension with the native data to 1.8 Å.The crystal structures from both crystal forms showed distinct ...

PoyD is a radical S-adenosyl methionine epimerase that introduces multiple D-configured amino acids at alternating positions into the highly complex marine peptides polytheonamide A and B. This novel post-translational modification contributes to the ability of the polytheonamides to form unimolecular minimalistic ion channels and its cytotoxic activity at picomolar levels. Using a genome mining approach we have identified additional PoyD homologues in various bacteria. Three enzymes were expressed in E. coli with their cognate as well as engineered peptide precursors and shown to introduce diverse D-amino acid patterns into all-L peptides. The data reveal a family of architecturally and functionally distinct enzymes that exhibit high regioselectivity, substrate promiscuity, and irreversible action and thus provide attractive opportunities for peptide engineering.

Radical S-adenosyl-l-methionine (SAM) enzymes comprise a vast superfamily catalyzing diverse reactions essential to all life through homolytic SAM cleavage to liberate the highly reactive 5′-deoxyadenosyl radical (5′-dAdo·). Our recent observation of a catalytically competent organometallic intermediate Ω that forms during reaction of the radical SAM (RS) enzyme pyruvate formate-lyase activating-enzyme (PFL-AE) was therefore quite surprising, and led to the question of its broad relevance in the superfamily. We now show that Ω in PFL-AE forms as an intermediate under a variety of mixing order conditions, suggesting it is central to catalysis in this enzyme. We further demonstrate that Ω forms in a suite of RS enzymes chosen to span the totality of superfamily reaction types, implicating Ω as essential in catalysis across the RS superfamily. Finally, EPR and electron nuclear double resonance spectroscopy establish that Ω involves an Fe–C5′ bond between 5′-dAdo· and the [4Fe–4S] cluster. An analogous organometallic bond is found in the well-known adenosylcobalamin (coenzyme B12) cofactor used to initiate radical reactions via a 5′-dAdo· intermediate. Liberation of a reactive 5′-dAdo· intermediate via homolytic metal–carbon bond cleavage thus appears to be similar for Ω and coenzyme B12. However, coenzyme B12 is involved in enzymes catalyzing only a small number (∼12) of distinct reactions, whereas the RS superfamily has more than 100 000 distinct sequences and over 80 reaction types characterized to date. The appearance of Ω across the RS superfamily therefore dramatically enlarges the sphere of bio-organometallic chemistry in Nature.

Current textbook knowledge holds that the structural scope of ribosomal biosynthesis is based exclusively on α-amino acid backbone topology. Here we report the genome-guided discovery of bacterial pathways that posttranslationally create β-amino acid-containing products. The transformation is widespread in bacteria and is catalyzed by an enzyme belonging to a previously uncharacterized radical -adenosylmethionine family. We show that the β-amino acids result from an unusual protein splicing process involving backbone carbon-carbon bond cleavage and net excision of tyramine. The reaction can be used to incorporate diverse and multiple β-amino acids into genetically encoded precursors in In addition to enlarging the set of basic amino acid components, the excision generates keto functions that are useful as orthogonal reaction sites for chemical diversification.