In Vivo Synthesis of Polyhydroxylated Compounds from a “Hidden Reservoir” of Toxic Aldehyde Species

Abstract

Synthetic enzyme cascades in living cells often lack efficiency owing to the formation of byproducts by endogenous enzymes or toxicity of the cascade intermediates. Highly reactive aldehyde species can trigger a metabolic stress response, and this leads to undesired side reactions and decreased yields. Owing to the metabolic background of Escherichia coli (E. coli), aldehydes may be irreversibly oxidized to carboxylic acids or reduced to the corresponding alcohols. Herein, we applied an approach to equilibrate the aldehyde concentration in vivo. We oxidized primary alcohols to the corresponding aldehydes by AlkJ, an alcohol dehydrogenase from Pseudomonas putida. Introduction of a carboxylic acid reductase from Nocardia iowensis allowed the target compound to be retrieved from the carboxylate sink. Further reduction of the aldehydes to alcohols by endogenous E. coli enzymes completed the equilibration between alcohols, aldehydes, and carboxylic acids. Thus, the aldehyde concentrations remained below nonviable concentrations. We demonstrated the concept on several primary alcohols, which reached the redox equilibrium within 6 h and persisted up to 24 h. Subsequent combination with a dihydroxyacetone‐dependent aldolase (Fsa1‐A129S, E. coli) demonstrated that the reactive aldehyde species were freely available and gave the aldol product, (3S,4R)‐1,3,4‐trihydroxy‐5‐phenylpentan‐2‐one, in 70 % yield within short reaction times.

Cell Factory Design and Optimization for the Stereoselective Synthesis of Polyhydroxylated Compounds

“…Longer reaction times led to significant overoxidation to the carboxylic acids (up to 80 %). Control experiments with E. coli BL21(DE3) in the presence of 1 a showed no endogenous oxidation activity under standard reaction conditions . Apart from the beneficial substrate uptake velocity of membrane‐associated AlkJ, this flavin adenine dinucleotide (FAD)‐dependent ADH promotes thermodynamically disfavored alcohol oxidation and takes advantage of irreversible O 2 reduction in the electron transport chain .…”

“…Two methods for vector construction were used: FastCloning (FC) and the application of a seamless and ligation‐independent cloning extract (SLiCE) . The previously constructed pKA1/ alkJ plasmid served as the parent plasmid backbone for the insertion of the fsa1‐A129S gene in different genetic architectures (Figure ). In the first round of cloning, the OPE plasmid was constructed by SLiCE to assemble the pKA1/ alkJ backbone and fsa1‐A129S fragments in the operon configuration.…”

“…In our operon, the expression of the two genes is controlled by only one T7 promoter (P T7 ) in front of the alkJ gene and one T7 terminator (T T7 ) downstream of the aldolase‐coding region (Figure A). The previously FC‐assembled POP plasmid contains the two genes in pseudo‐operon configuration, including an individual P T7 in front of the fsa1‐A129S gene (Figure B). Subsequently, a vector with monocistronic arrangement was constructed by molecular cloning (Figure C).…”

“…Based on the results presented above (e.g., growth rate and enzyme expression levels; Table ), the pseudo‐operon configuration was selected for further experiments (BL21(DE3) and RARE, 5 m m 1 a ). As previously demonstrated, increasing concentrations of DHA (5 and 20 equiv) pushed the equilibrium of the aldol reaction toward the aldol adduct site and stably maintained target compounds, such as 2 d , over long reactions times (24 h) in resting cells (RCs) of E. coli BL21(DE3) harboring the POP plasmid (Figure B and C) . In the absence of DHA, primary alcohol 1 a was almost completely overoxidized to the corresponding carboxylic acid, 1 c (Figure A).…”

“…Pathway optimization was performed at the genetic level through the inclusion of different regulatory elements (e.g., promoters, terminators). Through this study, we extended our previous proof‐of‐concept work on the synthesis of chiral aldol products in vivo and provided in‐depth optimization at genetic and process levels (Figure ). Finally, we investigated a fast and very efficient solid‐phase extraction (SPE) purification protocol for the isolation of all target molecules in good to excellent yields (see Table , below; 1 – 4 d , e ).…”