Biosynthesis-Guided Discovery and Engineering of α-Pyrone Natural Products from Type I Polyketide Synthases

Natural products containing the α-pyrone moiety are produced by polyketide synthases (PKSs) in bacteria, fungi, and plants. The conserved biosynthetic logic for the production of the α-pyrone moiety involves the cyclization of a triketide intermediate which also off-loads the polyketide from the activating thioester. In this study, we show that truncating a tetraketide natural product producing PKS assembly line allows for a thioesterase-independent off-loading of an α-pyrone polyketide natural product, one which we find to be natively present in the extracts of the bacterium that otherwise furnishes the tetraketide natural product. By engineering the truncated PKS in vitro, we demonstrate that a ketosynthase (KS) domain with relaxed substrate selectivity when coupled with in trans acylation of polyketide extender units can expand the chemical space of α-pyrone polyketide natural products. Findings from this study point toward heterologous intermolecular protein–protein interactions being detrimental to the efficiency of engineered PKS assembly lines.


General materials and instrumentation
All chemicals, solvents, and media components were obtained commercially from Sigma-Aldrich, Fisher Scientific, and VWR, and used without further purification. Phusion high-fidelity DNA polymerase and Gibson assembly Master Mix were purchased from New England Biolabs. PrimeSTAR DNA polymerase Master Mix was purchased from Takaro Bio. Synthetic DNA fragments were ordered from Twist Biosciences. Reactions were monitored by thin layer chromatography (TLC) carried out on Supelco silica gel (60 F 254 ) glass plates visualized by UV light. Silica gel (SiliaFlash GE60, 60-200 µm) was used for flash chromatography. Nuclear magnetic resonance (NMR) spectra were recorded on Bruker Avance III HD 400, 500, or 700 MHz instruments and calibrated using residual undeuterated solvent as the internal reference (CDCl 3 δ H 7.26 and δ C 77.16, MeOD δ H 3.31 and δ C 49.00, DMSO-d 6 δ H 2.50 and δ C 39.52). The splitting patterns were reported as s=singlet, d=doublet, t=triplet, q=quartet, m=multiplet, br=broad. Mass spectra were recorded on a Bruker Impact II high resolution time of flight (ToF) mass spectrometer with an electrospray ionization (ESI) source coupled to an Agilent 1290 ultra high-performance liquid chromatography system equipped with a diode array detector.

Vector construction
Standard molecular biology techniques were used to carry out plasmid construction. Amplification of target DNA fragments were carried out with either Phusion or PrimeSTAR high-fidelity DNA polymerases. Gibson assembly was used to subclone genes encoding for proteins of interest into the target vectors. The sequences of recombinant plasmids were confirmed by Sanger sequencing at Eton Biosciences.
The construction of PltB module1 with DEBS2 C-terminal docking domain (dd) sequences in pET28(+), PltB module2 with DEBS3 N-terminal docking domain sequences in pET28(+), PltG in pET24(+), Sfp in the first multiple cloning site (MCS) of pCDFDuet-1, MatB in pET28-MBP expression vectors has been described previously. [1][2] To construct the expression vector for CalA module1 fused with DEBS2 dd, calA module1 gene was amplified from previously synthesized pET28(+)-CalA module1 vector, 1 while the pET28(+) vector backbone with DEBS2dd sequences ligated was amplified using pET28(+)-PltB module1-DEBS2dd as the template. The amplified DNA fragment encoding for CalA module1 was then inserted into pET28(+) vector containing C-terminal DEBS2dd sequences using Gibson assembly. For the construction of FabD overexpression vector, the DNA fragment encoding fabD gene was first amplified from the Escherichia coli BL21Gold(DE3) genomic DNA and then assembled into pET28(+) S3 vector. Plasmids for PltG and CalA module1 mutants were generated by site-directed mutagenesis using standard procedures.
For production of compound 6, a four-gene, three-plasmid system was designed for cooverexpression of PltB module1-DEBS2dd, DEBS3dd-PltB module2, Sfp, and MatB. The DNA fragments encoding for MatB, PltB module1 fused with DEBS2dd and PltB module2 linked to DEBS3dd were amplified using vectors constructed above as the template. The gene encoding for PltB module1-DEBS2dd was first inserted into the second MCS (between NdeI and XhoI restriction sites) of pETDuet-1 vector, followed by the introduction of the matB gene into the first MCS (between NcoI and HindIII restriction sites) to construct the recombinant pETDuet-1 vector expressing both MatB and PltB module1 DEBS2dd.
The DNA fragment containing PltB module2 DEBS3dd was inserted into the MCS2 between NdeI and XhoI restriction sites of pACYCDuet-1 vector.

Recombinant protein expression and purification
For overexpression of holo-PKSs, the pET28(+) vector carrying PKS modules and the pCDFDuet-1 vector carrying Sfp were co-transformed into E. coli BL21Gold(DE3). Overnight cultures were inoculated into 2-4 L terrific both media supplemented with appropriate antibiotics. The cells were grown at 30 ˚C until OD 600 reached 0.4-0.5 at which time growth temperature was reduced to 18 ˚C. When OD 600 reached 0.7-0.8, protein expression was induced by the addition of 0.05-0.1 mM IPTG together with 0.25 mM calcium pantothenate. Bacterial cells were cultured at 18 ˚C for an addition of 18 h before harvested by centrifugation at 2,000×gfor 25 min. Cell pellets were stored at −80 ˚C until purification. For MatB, PltG (wild type and mutants), and FabD, expression vectors encoding for proteins of interest were transformed into E. coli BL21Gold(DE3). Similar growth condition was used as described above, except that 1 L terrific broth medium was used for cell culture and protein expression was induced with 0.2 mM IPTG.
All steps for protein purifications were performed at 4 ˚C or on ice. Cell pellets were resuspended in binding buffer (20 mM Tris-HCl pH=8.0, 500 mM NaCl, 10% glycerol) and lysed by sonication. The lysate was clarified by centrifugation at 25,000×g for 45 min, and then applied to a 5 mL HisTrap HP column. The column was washed extensively with wash buffer (20 mM Tris-HCl pH=8.0, 30 mM imidazole, 500 mM NaCl, 10% glycerol) till a stable UV-absorbance base line was observed, and then eluted with a linear gradient to 100 % of elution buffer (20 mM Tris-HCl pH=8.0, 250 mM imidazole, 500 mM NaCl, 10% glycerol) over 8 column volumes using ÄKTAprime plus FPLC system. Purity of eluent protein fractions were checked by SDS-PAGE, and fractions containing desired purified proteins were combined.
For holo-PKSs, pooled protein solutions were concentrated using 50 kDa Amicon centrifugal filters and S4 desalted into binding buffer with PD-10 columns. Combined FabD protein solution was dialyzed overnight in 2 L binding buffer before storage. For PltG wild type and mutants, combined protein fractions were dialyzed in 2 L buffer composed of 20 mM Tris-HCl pH=8.9, 50 mM KCl and 10% glycerol overnight.
The protein samples after dialysis were loaded to a 5 mL Hi-Trap Q column, washed with 5 column volumes of buffer A (20 mM Tris-HCl pH=8.9, 50 mM KCl), and then eluted with buffer B (20 mM Tris-HCl pH=8.9, 1 M KCl). Eluted protein fractions were checked by SDS-PAGE. Fractions containing protein of interest were combined, glycerol was added to a final concentration of 10% v/v and stored in small aliquots at −80˚C. Fresh aliquots were used each time for enzyme assays.
The reactions were incubated at 30 ˚C for 4.5 h before being quenched with 75 µL MeOH containing 10% v/v formic acid and 0.03 mM internal standard (4-hydroxy-6-methyl-2-pyrone). For time-course assay, the reactions were performed in a total volume of 400 µL; 70 µL aliquots were withdrawn at 1,2,4,6,8 h and added to 30 µL MeOH with 10% v/v formic acid. Each experiment was conduct in triplicate. The quenched reactions were analyzed by LC-MS or HPLC.

Organic extraction of Pseudomonas protegens Pf-5
The bacterium P. protegens Pf-5 was inoculated into 50 mL LB media from glycerol stock and grown at 30 ˚C for 24 h. Cell pellets and liquid media were separated by centrifugation at 2,000×g for 20 min. The supernatant was acidified with 3 M HCl to pH = 2.0, and then extracted with EtOAc (50 mL, 3).
The organic layers were combined, dried with anhydrous Na 2 SO 4 , and concentrated under vacuum. The concentrated residue was dissolved in 1 mL MeOH and analyzed by HPLC-HRMS in the negative ionization mode.

LC-MS and HPLC analysis
High resolution mass spectrometry (HRMS) data for synthesized acyl-SNACs were collected on an Agilent 1290 Infinity II UHPLC system coupled to a Bruker Impact II ToF mass spectrometer operating at room temperature with Kinetex 1.7 µm C18 100 Å column (502.1 mm). Mixture of water and acetonitrile with 0.1% formic acid was used as the mobile phase at a flow rate of 0.5 mLmin -1 . All MS data for acyl-SNACs were collected in positive mode.
Production of pyrone analogs were monitored by either LC-MS or HPLC. HPLC analyses to monitor the production of 6 and 7 were carried out using Phenomenex Luna 5 µm C8(2) 100 Å LC column (2504.6 mm) using Agilent 1260 Infinity II HPLC system. Water Poroshell 120 EC-C18 column (2.7 µm, 4.6  100 mm) using Agilent 1260 Infinity HPLC coupled to a Bruker amaZon SL operating in the negative ionization mode. Water (solvent A) and MeCN (solvent B) with 0.1 % formic acid were used as the mobile phase. A flow rate of 0.5 mLmin -1 was used with the following gradient: 0-3 min: 5% B, 3-16 min: linear gradient to 100% B, 16-20 min: 100% B, 20-21 min: linear gradient to 5% B, 21-22 min: 5% B, 22-23 min: linear gradient to 100% B, 23-24 min: 100% B, 24-25 min: linear gradient to 5% B, 25-27 min: 5% B. Ions corresponding to pyrones and internal standard were extracted, and peak areas for extracted ion counts were integrated. Ion count ratio between pyrone product and internal standard was calculated, normalized based on monoisotopic abundance, and averaged between three triplicate experiments. To represent relative activity, activity of each Plt PKS system (native or engineered) towards dichloropyrrole-2-carbonyl-SNAC is set as 1. Production level of other pyrone analogs by native or engineered PKSs was normalized to the production level of 6 by each PKS system.
The same UHPLC-HRMS system was used to collect HRMS data for pyrone analogs except that the MS data for pyrone analogs were collected in the negative mode.
After 24 h induction, E. coli cells were harvested by centrifugation at 2,000×g for 25 min. Cell pellets collected from 2 L culture were combined, resuspended in 60 mL buffer containing 400 mM potassium phosphate pH=7.5 and 10% glycerol, and sonicated for 40 min. The lysate was clarified by centrifugation S7 and used directly for large scale enzyme assays which was composed of 10 mM ATP, 10 mM MgCl 2 , 10 mM Na-malonate, 0.2 mM coenzyme A, 0.4 mM dichloropyrrole-2-carbonyl-SNAC, 5 mM TCEP, 5 mM calcium pantothenate, 400 mM potassium phosphate pH=7.5 and 10% glycerol in a total volume of 100 mL. ATP, Na-malonate and dichloropyrrole-2-carbonyl-SNAC were divided into 4 batches and added every 2 h. The reaction mixture was shaken at 60 rpm at 30 ˚C for 24 h before quenched with TFA to pH = 2.0. The acidified mixture with precipitate was extracted with EtOAC (100 mL, 3). The organic and aqueous layers were separated by centrifugation at 1,000×g for 5 min. The organic layers were combined, dried with anhydrous Na 2 SO 4 , and concentrated under vacuum. The crude extract was first purified by silica flash column (0-100% EtOAc in hexane). Fractions contain desired products (checked by LC-MS) were pooled, concentrated, and further purified by two rounds of preparative HPLC. For the first round of preparative HPLC purification, it was carried out on Luna 5 µm C18 (2) (2) 100 Å LC column (2504.6 mm) with isocratic elution at 50% B using the same solvent system at a flow rate of 0.5 mLmin -1 . The product was monitored at UV 352 and 254 nm. Purified molecule was analyzed by HRMS and NMR spectra.