Assessing and harnessing updated polyketide synthase modules through combinatorial engineering.

The modular nature of polyketide assembly lines and the significance of their products make them prime targets for combinatorial engineering. While short synthases constructed using the recently updated module boundary have been shown to outperform those using the traditional boundary, larger synthases constructed using the updated boundary have not been investigated. Here we describe our design and implementation of a BioBricks-like platform to rapidly construct 5 triketide, 25 tetraketide, and 125 pentaketide synthases from the updated modules of the Pikromycin synthase. Every combinatorial possibility of modules 2-6 inserted between the first and last modules of the native synthase was constructed and assayed. Anticipated products were observed from 60% of the triketide synthases, 32% of the tetraketide synthases, and 6.4% of the pentaketide synthases. Ketosynthase gatekeeping and module-skipping were determined to be the principal impediments to obtaining functional synthases. The platform was also used to create functional hybrid synthases through the incorporation of modules from the Erythromycin, Spinosyn, and Rapamycin assembly lines. The relaxed gatekeeping observed from a ketosynthase in the Rapamycin synthase is especially encouraging in the quest to produce designer polyketides.

Boosting titers of engineered triketide and tetraketide synthases to record levels through T7 promoter tuning.

Modular polyketide synthases (PKS's) are promising platforms for the rational engineering of designer polyketides and commodity chemicals, yet their low productivities are a barrier to the practical biosynthesis of these compounds. Previously, we engineered triketide lactone synthases such as Pik167 using the recently updated module definition and showed they generate hundreds of milligrams of product per liter of Escherichia coli K207-3 shake flask culture. As the molar ratio between the 2 polypeptides of Pik167 is highly skewed, we sought to attenuate the strength of the T7 promoter controlling the production of the smaller, better-expressing polypeptide and thereby increase production of the first polypeptide under the control of an unoptimized T7 promoter. Through this strategy, a 1.8-fold boost in titer was obtained. After a further 1.5-fold boost obtained by increasing the propionate concentration in the media from 20 to 80 mM, a record titer of 791 mg L-1 (627 mg L-1 isolated) was achieved, a 2.6-fold increase overall. Spurred on by this result, the tetraketide synthase Pik1567 was engineered and the T7 promoter attenuation strategy was applied to its second and third genes. A 5-fold boost, from 20 mg L-1 to 100 mg L-1, in the titer of its tetraketide product was achieved.

Crystal structures reveal the framework of cis -acyltransferase modular polyketide synthases.

Although the domains of cis -acyltransferase ( cis -AT) modular polyketide synthases (PKS's) have been understood at atomic resolution for over a decade, the domain-domain interactions responsible for the architectures and activities of these giant molecular assembly lines remain largely uncharacterized. The multimeric structure of the α 6 ß 6 fungal fatty acid synthase (FAS) provides 6 equivalent reaction chambers for its acyl carrier protein (ACP) domains to shuttle carbon building blocks and the growing acyl chain between surrounding, oriented enzymatic domains. The presumed homodimeric oligomerization of cis -AT assembly lines is insufficient to provide similar reaction chambers; however, the crystal structure of a ketosynthase (KS)+AT didomain presented here and three already reported show an interaction between the AT domains appropriate for lateral multimerization. This interaction was used to construct a framework for the pikromycin PKS from its KS, AT, and docking domains that contains highly-ordered reaction chambers. Its AT domains also mediate vertical interactions, both with upstream KS domains and downstream docking domains.

Priming enzymes from the pikromycin synthase reveal how assembly-line ketosynthases catalyze carbon-carbon chemistry.

The first domain of modular polyketide synthases (PKSs) is most commonly a ketosynthase (KS)-like enzyme, KSQ, that primes polyketide synthesis. Unlike downstream KSs that fuse α-carboxyacyl groups to growing polyketide chains, it performs an extension-decoupled decarboxylation of these groups to generate primer units. When Pik127, a model triketide synthase constructed from modules of the pikromycin synthase, was studied by cryoelectron microscopy (cryo-EM), the dimeric didomain comprised of KSQ and the neighboring methylmalonyl-selective acyltransferase (AT) dominated the class averages and yielded structures at 2.5- and 2.8-Å resolution, respectively. Comparisons with ketosynthases complexed with their substrates revealed the conformation of the (2S)-methylmalonyl-S-phosphopantetheinyl portion of KSQ and KS substrates prior to decarboxylation. Point mutants of Pik127 probed the roles of residues in the KSQ active site, while an AT-swapped version of Pik127 demonstrated that KSQ can also decarboxylate malonyl groups. Mechanisms for how KSQ and KS domains catalyze carbon-carbon chemistry are proposed.

Preparative production of an enantiomeric pair by engineered polyketide synthases.

Using the updated module boundary of polyketide assembly lines, modules from the pikromycin synthase were recombined into engineered synthases that furnish an enantiomeric pair of 2-stereocenter triketide lactones at >99% ee with yields up to 0.39 g per liter of E. coli K207-3 in shake flasks.

Insights into modular polyketide synthase loops aided by repetitive sequences.

The loops of modular polyketide synthases (PKSs) serve diverse functions but are largely uncharacterized. They frequently contain amino acid repeats resulting from genetic events such as slipped-strand mispairing. Determining the tolerance of loops to amino acid changes would aid in understanding and engineering these multidomain molecule factories. Here, tandem repeats in the DNA encoding 949 modules within 129 cis-acyltransferase PKSs were cataloged, and the locations of the corresponding amino acids within the module were identified. The most frequently inserted interdomain loop corresponds with the updated module boundary immediately downstream of the ketosynthase (KS), while the loops bordering the dehydratase are nearly intolerant to such insertions. From the 949 modules, no repetitive sequence loop insertions are located within ACP, and only 2 reside within KS, indicating the sensitivity of these domains to alteration.

An in vitro platform for engineering and harnessing modular polyketide synthases.

To harness the synthetic power of modular polyketide synthases (PKSs), many aspects of their biochemistry must be elucidated. A robust platform to study these megadalton assembly lines has not yet been described. Here, we in vitro reconstitute the venemycin PKS, a short assembly line that generates an aromatic product. Incubating its polypeptides, VemG and VemH, with 3,5-dihydroxybenzoic acid, ATP, malonate, coenzyme A, and the malonyl-CoA ligase MatB, venemycin production can be monitored by HPLC and NMR. Multi-milligram quantities of venemycin are isolable from dialysis-based reactors without chromatography, and the enzymes can be recycled. Assembly line engineering is performed using pikromycin modules, with synthases designed using the updated module boundaries outperforming those using the traditional module boundaries by over an order of magnitude. Using combinations of VemG, VemH, and their engineered derivatives, as well as the alternate starter unit 3-hydroxybenzoic acid, a combinatorial library of six polyketide products is readily accessed.

An integrated screening system for the selection of exemplary substrates for natural and engineered cytochrome P450s.

Information about substrate and product selectivity is critical for understanding the function of cytochrome P450 monooxygenases. In addition, comprehensive understanding of changes in substrate selectivity of P450 upon amino acid mutation would enable the design and creation of engineered P450s with desired selectivities. Therefore, systematic methods for obtaining such information are required. Herein, we developed an integrated P450 substrate screening system for the selection of "exemplary" substrates for a P450 of interest. The established screening system accurately selected the known exemplary substrates and also identified previously unknown exemplary substrates for microbial-derived P450s from a library containing sp3-rich synthetic small molecules. Synthetically potent transformations were also found by analyzing the reactions and oxidation products. The screening system was applied to analyze the substrate selectivity of the P450 BM3 mutants F87A and F87A/A330W, which acquired an ability to hydroxylate non-natural substrate steroids regio- and stereoselectively by two amino acid mutations. The distinct transition of exemplary substrates due to each single amino acid mutation was revealed, demonstrating the utility of the established system.

A crotonyl-CoA reductase-carboxylase independent pathway for assembly of unusual alkylmalonyl-CoA polyketide synthase extender units.

Type I modular polyketide synthases assemble diverse bioactive natural products. Such multienzymes typically use malonyl and methylmalonyl-CoA building blocks for polyketide chain assembly. However, in several cases more exotic alkylmalonyl-CoA extender units are also known to be incorporated. In all examples studied to date, such unusual extender units are biosynthesized via reductive carboxylation of α, ß-unsaturated thioesters catalysed by crotonyl-CoA reductase/carboxylase (CCRC) homologues. Here we show using a chemically-synthesized deuterium-labelled mechanistic probe, and heterologous gene expression experiments that the unusual alkylmalonyl-CoA extender units incorporated into the stambomycin family of polyketide antibiotics are assembled by direct carboxylation of medium chain acyl-CoA thioesters. X-ray crystal structures of the unusual ß-subunit of the acyl-CoA carboxylase (YCC) responsible for this reaction, alone and in complex with hexanoyl-CoA, reveal the molecular basis for substrate recognition, inspiring the development of methodology for polyketide bio-orthogonal tagging via incorporation of 6-azidohexanoic acid and 8-nonynoic acid into novel stambomycin analogues.

Identification of Middle Chain Fatty Acyl-CoA Ligase Responsible for the Biosynthesis of 2-Alkylmalonyl-CoAs for Polyketide Extender Unit.

Understanding the biosynthetic mechanism of the atypical polyketide extender unit is important for the development of bioactive natural products. Reveromycin (RM) derivatives produced by Streptomyces sp. SN-593 possess several aliphatic extender units. Here, we studied the molecular basis of 2-alkylmalonyl-CoA formation by analyzing the revR and revS genes, which form a transcriptional unit with the revT gene, a crotonyl-CoA carboxylase/reductase homolog. We mainly focused on the uncharacterized adenylate-forming enzyme (RevS). revS gene disruption resulted in the reduction of all RM derivatives, whereas reintroduction of the gene restored the yield of RMs. Although RevS was classified in the fatty acyl-AMP ligase clade based on phylogenetic analysis, biochemical characterization revealed that the enzyme catalyzed the middle chain fatty acyl-CoA ligase (FACL) but not the fatty acyl-AMP ligase activity, suggesting the molecular evolution for acyl-CoA biosynthesis. Moreover, we examined the in vitro conversion of fatty acid into 2-alkylmalonyl-CoA using purified RevS and RevT. The coupling reaction showed efficient conversion of hexenoic acid into butylmalonyl-CoA. RevS efficiently catalyzed C8-C10 middle chain FACL activity; therefore, we speculated that the acyl-CoA precursor was truncated via ß-oxidation and converted into (E)-2-enoyl-CoA, a RevT substrate. To determine whether the ß-oxidation process is involved between the RevS and RevT reaction, we performed the feeding experiment using [1,2,3,4-(13)C]octanoic acid. (13)C NMR analysis clearly demonstrated incorporation of the [3,4-(13)C]octanoic acid moiety into the structure of RM-A. Our results provide insight into the role of uncharacterized RevS homologs that may catalyze middle chain FACL to produce a unique polyketide extender unit.

Structure-function analyses of cytochrome P450revI involved in reveromycin A biosynthesis and evaluation of the biological activity of its substrate, reveromycin T.

Numerous cytochrome P450s are involved in secondary metabolite biosynthesis. The biosynthetic gene cluster for reveromycin A (RM-A), which is a promising lead compound with anti-osteoclastic activity, also includes a P450 gene, revI. To understand the roles of P450revI, we comprehensively characterized the enzyme by genetic, kinetic, and structural studies. The revI gene disruptants (ΔrevI) resulted in accumulation of reveromycin T (RM-T), and revI gene complementation restored RM-A production, indicating that the physiological substrate of P450revI is RM-T. Indeed, the purified P450revI catalyzed the C18-hydroxylation of RM-T more efficiently than the other RM derivatives tested. Moreover, the 1.4 Å resolution co-crystal structure of P450revI with RM-T revealed that the substrate binds the enzyme with a folded compact conformation for C18-hydroxylation. To address the structure-enzyme activity relationship, site-directed mutagenesis was performed in P450revI. R190A and R81A mutations, which abolished salt bridge formation with C1 and C24 carboxyl groups of RM-T, respectively, resulted in significant loss of enzyme activity. The interaction between Arg(190) and the C1 carboxyl group of RM-T elucidated why P450revI was unable to catalyze both RM-T 1-methyl ester and RM-T 1-ethyl ester. Moreover, the accumulation of RM-T in ΔrevI mutants enabled us to characterize its biological activity. Our results show that RM-T had stronger anticancer activity and isoleucyl-tRNA synthetase inhibition than RM-A. However, RM-T showed much less anti-osteoclastic activity than RM-A, indicating that hemisuccinate moiety is important for the activity. Structure-based P450revI engineering for novel hydroxylation and subsequent hemisuccinylation will help facilitate the development of RM derivatives with anti-osteoclast activity.

The Bacillus subtilis essential gene dgkB is dispensable in mutants with defective lipoteichoic acid synthesis.

The dgkB gene is essential for the growth of Bacillus subtilis. It encodes a diacylglycerol (DG) kinase that converts DG to phosphatidic acid to reintroduce it into the phospholipid synthesis pathway. Repression of the dgkB gene placed under a regulatable promoter causes accumulation of DG and leads to lethality. DG is formed as a byproduct of the synthesis of lipoteichoic acid (LTA), a polyanionic component of the cell envelope. B. subtilis synthesizes LTA by polymerizing the glycerophosphate moiety of phosphatidylglycerol (PG) onto a glucolipid membrane anchor, and releasing the DG moiety of PG. B. subtilis has four genes homologous to Staphylococcus aureus ltaS, which encodes LTA synthase. Disruption of either or both of two genes, yflE and yfnI, whose products show higher homology with S. aureus LtaS among the four homologues, suppressed the lethality caused by dgkB repression. In cells with dgkB repression, DG was accumulated to 43 ± 3% of total lipids, about three times the content of wild type cells (13 ± 1%). Disruption of yfnI in the dgkB-repressed cells reduced the DG content to 15 ± 2%, but yflE-disruption did not (42 ± 1%); this was probably due to efficient LTA synthesis by YfnI in the yflE-disrupted cells. Further introduction of a disrupted allele of ugtP, encoding glucolipid synthase that consumes DG as a substrate, partially lowered the colony forming capacity in strains with yflE-disruption. A disrupted dgkB allele was successfully introduced into strains disrupted for either or both of yflE and yfnI, indicating that the essential gene dgkB is dispensable in mutants defective in LTA synthesis.

Stress distribution patterns at the coracoacromial arch in rotator cuff tear measured by computed tomography osteoabsorptiometry.

When a rotator cuff tear occurs, forces compressing the humeral head toward the glenoid are disturbed, and the kinematics of the glenohumeral joint change. Therefore, stress distributions at the coracoacromial arch in cuff tear shoulders should differ from those in normal shoulders. To investigate this hypothesis, we studied stress distribution patterns at the coracoacromial arch in normal and cuff tear shoulders using a computed tomography (CT) osteoabsorptiometry method, in which bone density correlates directly with long-term physiologic loading. Eight normal subjects and 11 patients with cuff tear were examined. The stress distributions at the undersurface of the acromion and the posterolateral surface of the coracoid process differed markedly between normal and cuff tear shoulders. In cuff tear shoulders, a high-density area was located at the anterior or the anterolateral part of the undersurface of the acromion, while it was located at the posterior part in all but one normal shoulder. Additionally, a high-density area was located at the superior or the lateral part of the coracoid process in most of the cuff tear shoulders; on the other hand, it was located at the base in all but one normal shoulder. We believe that the differences in stress distribution patterns are due to impingement at the coracoacromial arch in cuff tear shoulders. CT osteoabsorptiometry can provide useful information in performing coracoacromial arch decompression for cuff tear shoulders.