Construction of a chimeric biosynthetic pathway for the de novo biosynthesis of rosmarinic acid in Escherichia coli

Abstract

Hydroxycinnamic acid esters (HCEs) are widely-distributed phenylpropanoid-derived plant natural products. Rosmarinic acid (RA), the most well-known HCE, shows promise as a treatment for cancer and neurological disorders. In contrast to extraction from plant material or plant cell culture, microbial production of HCEs could be a sustainable, controlled means of production. Through the overexpression of a six-enzyme chimeric bacterial and plant pathway, we show the de novo biosynthesis of RA, and the related HCE isorinic acid (IA), in Escherichia coli. Probing the pathway through precursor supplementation showed several potential pathway bottlenecks. We demonstrated HCE biosynthesis using three plant rosmarinic acid synthase (RAS) orthologues, which exhibited different levels of HCE biosynthesis but produced the same ratio of IA to RA. This work serves as a proof-of-concept for a microbial production platform for HCEs by using a modular biosynthetic approach to access diverse natural and non-natural HCEs.

Keywords: biosynthesis; hydroxycinnamic acid esters; metabolic engineering; natural products; rosmarinic acid.

Figure 1. Plasmids assembled for donor substrate, acceptor substrate and HCE biosynthesis

All plasmids were assembled using the BioBrick™ system previously designed and built by our lab ^[50]. P_lac′, a modified constitutive lac promoter; P_araBAD, arabinose inducible promoter; araC, gene encoding the repressor of P_araBAD; pUCBB, high copy number plasmid backbone; pCDFBB, medium copy number plasmid backbone; pACBB, low copy number plasmid backbone. Plasmids that constitute the functional three-module RA 7 biosynthetic pathway are boxed.

Figure 2. De novo biosynthesis of 4-HPL by E. coli BW27784 cells transformed with pUCBB-hdhA

Transformants were cultured in minimal media at 30 °C; culture samples were taken over time, cleared of cells, and extracted with ethyl acetate for HPLC analysis. An analytical standard of 4-HPL 2 was used to confirm peak identity and quantification. No 4-HPL 2 was detected in cultures of cells transformed with the empty pUCBB vector.

Figure 3. Bioconversion of p-coumaric acid and 4-HPL to caffeic acid and 3,4-DHPL, respectively, by E. coli BW27784 transformants of pUCBB-hpaBC

Transformants were cultured in minimal media at 30 °C, supplemented with 1 mM p-coumaric acid or 1 mM 4-HPL 2 culture and samples were taken over time and processed for HPLC analysis. Samples were compared to standard compounds in order to identify compound peaks. Solid trace, pUCBB-hpaBC transformants; dotted trace, empty pUCBB transformants (control).

Figure 4. De novo biosynthesis of HCEs in E. coli

E. coli BW27784 was co-transformed with pUCBB-ABC containing the acceptor substrate module, pCDFBB-TBC4 containing the acceptor substrate module, and pACBB-4cbR containing the hydroxycinnamoyl transfer module (Figure 1). Transformants were cultured in minimal media at 30 °C, and at4CL2 and cbRAS expression was induced after 24 h growth. Culture samples were taken over time and processed for HPLC and LC-MS analysis. An analytical standard of RA 7 was used to confirm peak identity and mass fragmentation of RA 7, and MS/MS mass spectra confirm the presence of both RA 7 and IA 8 in culture media. Asterisks denote key mass fragments in the identification of the acceptor substrate moieties of the products.

Figure 5. Limiting factors and leaks in the E. coli HCE biosynthetic pathway

E. coli BW27784 containing the modular HCE biosynthesis plasmids were cultured in minimal media at 30 °C; after 24h, at4CL2 and cbRAS expression was induced and various supplements were added to a final concentration of 1 mM. Culture samples were taken over time and analyzed by HPLC. Asterisks denote significantly different concentrations at p < 0.05. A) HCE production is increased in the presence of different feeding precursors 48h after induction of at4CL2 and cbRAS. B) The total concentrations of donor substrates and HCE products remain stable over time. C) The total concentrations of acceptor substrates and HCE products significantly decreases after 24 h.

Figure 6. HCE biosynthesis in E. coli cultures expressing three orthologous RAS enzymes in the context of the redesigned biosynthetic pathway

Two new hydroxycinnamoyl transfer module plasmids were constructed containing laRAS and moRAS. E. coli BW27784 were co-transformed with empty pACBB, pACBB-4cbR, pACBB-4laR or pACBB-4moR, pUCBB-ABC and pCDFBB-TBC and analyzed for HCE production 72 h after induction of the at4CL2 and RAS-encoding genes. Asterisks denote significantly different concentrations at p < 0.01.

Scheme 1. Redesigned chimeric RA biosynthetic pathway

The D-hydroxyisocaproate dehydrogenase HdhA from L. delbrueckii subsp. bulgaricus catalyzes the conversion of endogenous 4-HPP 1 to 4-HPL 2, which is subsequently meta-hydroxylated by hydroxylase complex HpaBC cloned from E. coli to yield 3,4-DHPL 3, the intended acceptor substrate for rosmarinic acid synthase (RAS). Donor substrate biosynthesis is achieved through the conversion of L-tyrosine 4 to p-coumaric acid by a tyrosine ammonia lyase from R. sphaeroides (RsTAL) and hydroxylation of p-coumaric acid 5 to caffeic acid 6 by HpaBC. A 4-coumaroyl-CoA ligase from Arabidopsis thaliana (At4CL2) generates the CoA-ester of the donor substrate, allowing ester formation catalyzed by RAS. Both 3,4-DHPL 3 and 4-HPL 2 may be accepted by RAS to form both RA 7 and IA 8, respectively.