A BioBricks Metabolic Engineering Platform for the Biosynthesis of Anthracyclinones in Streptomyces coelicolor

Abstract

Actinomycetes produce a variety of clinically indispensable molecules, such as antineoplastic anthracyclines. However, the actinomycetes are hindered in their further development as genetically engineered hosts for the synthesis of new anthracycline analogues due to their slow growth kinetics associated with their mycelial life cycle and the lack of a comprehensive genetic toolbox for combinatorial biosynthesis. In this report, we tackled both issues via the development of the BIOPOLYMER (BIOBricks POLYketide Metabolic EngineeRing) toolbox: a comprehensive synthetic biology toolbox consisting of engineered strains, promoters, vectors, and biosynthetic genes for the synthesis of anthracyclinones. An improved derivative of the production host Streptomyces coelicolor M1152 was created by deleting the matAB gene cluster that specifies extracellular poly-β-1,6-N-acetylglucosamine (PNAG). This resulted in a loss of mycelial aggregation, with improved biomass accumulation and anthracyclinone production. We then leveraged BIOPOLYMER to engineer four distinct anthracyclinone pathways, identifying optimal combinations of promoters, genes, and vectors to produce aklavinone, 9-epi-aklavinone, auramycinone, and nogalamycinone at titers between 15-20 mg/L. Optimization of nogalamycinone production strains resulted in titers of 103 mg/L. We structurally characterized six anthracyclinone products from fermentations, including new compounds 9,10-seco-7-deoxy-nogalamycinone and 4-O-β-d-glucosyl-nogalamycinone. Lastly, we tested the antiproliferative activity of the anthracyclinones in a mammalian cancer cell viability assay, in which nogalamycinone, auramycinone, and aklavinone exhibited moderate cytotoxicity against several cancer cell lines. We envision that BIOPOLYMER will serve as a foundational platform technology for the synthesis of designer anthracycline analogues.

Keywords: BioBricks; Streptomyces coelicolor; anthracyclinones; anticancer; natural product biosynthesis; synthetic biology.

Figure 1

Biosynthesis of anthracyclinones. One acetyl-CoA or propionyl-CoA starter unit is condensed with nine malonyl-CoA extender units via iterative Claisen condensation reactions via the minimal PKS to form an enzyme-tethered decaketide intermediate. The polyketide ketoreductase reduces the 9-ketone to a hydroxyl group followed by C7–C12 first-ring aromatization, C5–C14 s-ring cyclization, and C3–C16 third-ring cyclization by the second-third ring cyclase. C-12 oxidation is catalyzed by an anthraquinol monooxygenase, followed by O-methylation to afford a tricyclic anthraquinone methyl ester. The key tricyclic intermediate undergoes fourth-ring cyclization via two distinct routes: SnoaL/Kyc34 cyclizes the fourth ring to afford a 9(S)-anthracyclinone and AknH/DnrD cyclizes the fourth ring to afford a 9(R)-anthracyclinone. The final step is C-7 ketoreduction by ketoreductases to afford C-21 anthracyclinones 9-epi-aklavinone or aklavinone and C-20 anthracyclinones nogalamycinone and auramycinone.

Figure 2

Structures of anthracyclin(on)es described in this work.

Figure 3

Metabolic engineering of SEK15. (A) Scheme for the biosynthesis of SEK15 via the Snoa123 minPKS. (B) Production titers of SEK15 on R5 solid agar plates from the heterologous expression of codon-optimized snoa123 in S. lividans K4–114, S. coelicolor M1146, and S. coelicolor M1152. (C) Production titers of SEK15 in SG liquid media shake flask experiments from the heterologous expression of codon-optimized snoa123 in S. coelicolor M1146 and S. coelicolor M1152. (D) Production titers of SEK15 in SG liquid media shake flask experiments from the heterologous expression of wild-type snoa123 in S. coelicolor M1152 and S. coelicolor M1152ΔmatAB. Experiments were carried out in double triplicate and the error bars reflect the standard deviation. ANOVA was carried out to determine the statistical significance between strains. The statistically significant comparisons are reflected with asterisks. The statistical significance of observed results was established with a p < 0.05. * indicates p ≤ 0.05, ** indicates p ≤ 0.01, *** indicates p ≤ 0.001, and **** indicates p ≤ 0.0001.

Figure 4

Engineering of UWM7 and WJ85 minimal aromatic polyketides. (A) Biosynthesis scheme for the synthesis of UWM7 from the DpsABCDG/AknBCDE2F minPKS and WJ85 from the OxyABCD minPKS. (B) Production titers of UWM7 resulting from the engineering of pSET-ermE*p-aknBCDE2F and pSET-kasOp*-aknBCDE2F in S. coelicolor M1152 and M1152ΔmatAB. (C) Production titers of UWM7 resulting from the engineering of pSET-ermE*p-dpsABCDG and pSET-kasOp*-dpsABCDG in S. coelicolor M1152 and M1152ΔmatAB. (D) Production titers of SEK15 and WJ85 resulting from the engineering of pSET-ermE*p-oxyABCD and pSET-kasOp*-oxyABCD in S. coelicolor M1152 and M1152ΔmatAB. ANOVA was carried out to determine the statistical significance between strains. The statistically significant comparisons are reflected with asterisks. The statistical significance of observed results was established with a p < 0.05. * indicates p ≤ 0.05, ** indicates p ≤ 0.01, *** indicates p ≤ 0.001, and **** indicates p ≤ 0.0001.

Figure 5

Engineering of KR/ARO/CYC cassettes with the snoa123 minPKS. (A) Chromatograms of metabolites from two-plasmid snoa123 and KR/ARO/CYC heterologous expression experiments. (B) Disposition of metabolites from two-plasmid snoa123 and KR/ARO/CYC heterologous expression experiments. (C) Chromatograms of metabolites from one-plasmid snoa123 and KR/ARO/CYC heterologous expression experiments. (D) Disposition of metabolites from one-plasmid snoa123 and KR/ARO/CYC heterologous expression experiments.

Figure 6

Engineering of aklanonic acid via combinatorial biosynthesis. (A) Scheme depicting degradation of aklanonic acid to its three detected shunt products (AA-1, AA-2, and AA-3). (B) Chromatogram of aklanonic acid degradation metabolites produced by a representative strain. (C) Aklanonic acid production titers resulting from full-factorial combinatorial biosynthesis of minPKS and KR/ARO/CYC/OXY genes. Error bars depict standard deviation. ANOVA was carried out to determine the statistical significance between strains. The statistically significant comparisons are reflected with asterisks. The statistical significance of observed results was established with a p < 0.05. * indicates p ≤ 0.05, ** indicates p ≤ 0.01, *** indicates p ≤ 0.001, and **** indicates p ≤ 0.0001.

Figure 7

Full pathway engineering of anthracyclinone polyketide synthases. Engineering of methyltransferase, fourth-ring cyclase, and ketoreductase genes furnished anthracyclinone pathways in S. coelicolor M1152ΔmatAB. (A) S. coelicolor M1152ΔmatAB::pS2S5 was complemented with different constructs to biosynthesize different amounts of 1 and 2. (B) S. coelicolor M1152ΔmatAB::pA2A5 was complemented with different constructs to biosynthesize different amounts of 7 and 8. (C) Production titer of anthracyclinones in strains expressing the entire pathway on two plasmids. Strains were grown in SG-TES liquid media for 5 days. Error bars reflect the SD of six replicates. Experimental groups were compared using a one-way ANOVA test to determine statistical significance. The statistical significance of observed results was established with a p < 0.05. * indicates p ≤ 0.05, ** indicates p ≤ 0.01, *** indicates p ≤ 0.001, and **** indicates p ≤ 0.0001. (D) HPLC chromatogram traces of different strains at 430 nm with major metabolites highlighted with an asterisk (*): (i) S. coelicolor M1152ΔmatAB::pSET152; (ii) S. coelicolor M1152ΔmatAB::A2A5 (aklanonic acid); (iii) S. coelicolor M1152ΔmatAB::S2S5 (nogalonic acid); (iv) S. coelicolor M1152ΔmatAB::S2S5A6 (auramycinone); (v) S. coelicolor M1152ΔmatAB::S2S5S6 (nogalamycinone); (vi) S. coelicolor M1152ΔmatAB::A2A5A6 (aklavinone); (vii) S. coelicolor M1152ΔmatAB::A2A5S6 (9-epi-aklavinone).

Figure 8

Production titers resulting from the expression of anthracyclinone pathways on a single plasmid. S. coelicolor M1152ΔmatAB was transformed with vectors (A) pSET154BB-kasOp*-aknBCDE2F+kasOp*-aknAE1WX+sp44-aknGHU (pEN10001, aklavinone pathway), (B) pSET154BB-kasOp*-aknBCDE2F+kasOp*-aknAE1WX+sp44-snoaC+kyc34+snoaF (pEN10002, 9-epi-aklavinone pathway), (C) pSET154BB-kasOp*-snoa123+kasOp*-snoaDEMB+sp44-aknGHU (pEN10003, auramycinone pathway), or (D) pSET154BB-kasOp*-snoa123+kasOp*-snoaDEMB+sp44-snoaC+kyc34+snoaF (pEN10004, wild-type nogalamycinone pathway) or pSET152BB-kasOp*-snoa123+kasOp*-aknAE1WX + sp44-snoaCLF (pRW10000, codon-optimized nogalamycinone pathway) and grown in shake flasks of SG-TES or E1 liquid media for 5 days. Error bars reflect the standard deviation of six replicates. Post hoc Student’s t tests were performed to determine statistical significance. The statistical significance of observed results was established with a p < 0.05. * indicates p ≤ 0.05, ** indicates p ≤ 0.01, *** indicates p ≤ 0.001, and **** indicates p ≤ 0.0001.

Figure 9

% Viability of A549 (non-small-cell lung) and PC3 (prostate) human cancer cell lines, and Merkel cells (MKL1 and MCC26) (after 72 h) at 80 μM concentration of compounds 1–7 and 9–13.