Biosynthetic pathway for the cyanide-free production of phenylacetonitrile in Escherichia coli by utilizing plant cytochrome P450 79A2 and bacterial aldoxime dehydratase

Abstract

The biosynthetic pathway for the production of phenylacetonitrile (PAN), which has a wide variety of uses in chemical and pharmaceutical industries, was constructed in Escherichia coli utilizing enzymes from the plant glucosinolate-biosynthetic and bacterial aldoxime-nitrile pathways. First, the single-step reaction to produce E,Z-phenylacetaldoxime (PAOx) from l-Phe was constructed in E. coli by introducing the genes encoding cytochrome P450 (CYP) 79A2 and CYP reductase from Arabidopsis thaliana, yielding the E,Z-PAOx-producing transformant. Second, this step was expanded to the production of PAN by further introducing the aldoxime dehydratase (Oxd) gene from Bacillus sp. strain OxB-1, yielding the PAN-producing transformant. The E,Z-PAOx-producing transformant also produced phenethyl alcohol and PAN as by-products, which were suggested to be the metabolites of E,Z-PAOx produced by E. coli enzymes, while the PAN-producing transformant accumulated only PAN in the culture broth, which suggested that the CYP79A2 reaction (the conversion of l-Phe to E,Z-PAOx) was a potential bottleneck in the PAN production pathway. Expression of active CYP79A2 and concentration of biomass were improved by the combination of the autoinduction method, coexpression of groE, encoding the heat shock protein GroEL/GroES, N-terminal truncation of CYP79A2, and optimization of the culture conditions, yielding a >60-fold concentration of E,Z-PAOx (up to 2.9 mM). The concentration of PAN was 4.9 mM under the optimized conditions. These achievements show the potential of this bioprocess to produce nitriles and nitrile derivatives in the absence of toxic chemicals.

FIG 1

Plant cyanogenic glycoside and glucosinolate synthetic pathway (top) and bacterial aldoxime-nitrile pathway (bottom), initiating from l-Phe. E,Z-PAOx, E,Z-phenylacetaldoxime; PAN, phenylacetonitrile; MAN, R,S-mandelonitrile; FMO, flavin-containing monooxygenase. The involvement of FMO (bottom, dashed arrow) was predicted by the recent studies of S. coelicolor FMO acting on l-Trp (24). The pathways used for the production of PAN in this study are shown in gray boxes.

FIG 2

Schematic representations of N-terminal sequences in ATR (A) and CYP79A2 (B). (A) N-terminal sequence of ATR containing 2 to 44 residues of the transmembrane sequence (gray box), followed by a functional domain. (B) N-terminal membrane-binding domain of CYP79A2 constituting a transmembrane region, basic region, and Pro-rich region (gray boxes), followed by a functional domain. The N-terminal sequences of Δ14CYP79A2, Δ23CYP79A2, Δ14CYP79A2-2A₃, and Δ14CYP79A2-K2A are also shown. Sites of point mutations are indicated by underlining.

FIG 3

GC-MS analysis of the products by the transformants. (A) Total ion chromatograms for the hexane extracts from the culture broth of the transformants with pETDuet-Δ44atr-cyp79A2 (top), pETDuet-Δ44atr (middle), and pETDuet-Δ44atr-cyp79A2 and pCDF-oxd (bottom). (B) MS spectra of E,Z-PAOx (peak 1), PAN (peak 2), and phenethyl alcohol (peak 3).

FIG 4

Time course of the concentrations of biomass, E,Z-PAOx, and PAN. (A) Time courses of E,Z-PAOx concentration (closed symbols) and OD₆₀₀ (open symbols). Circles, transformant with pETDuet-Δ44atr-cyp79A2 grown in LB, followed by the induction with IPTG; diamonds, transformant with pETDuet-Δ44atr-cyp79A2 grown in the LB-based autoinduction medium; triangles, transformant with pETDuet-Δ44atr-cyp79A2 and pGro7 grown in the LB-based autoinduction medium containing 2 mg/ml arabinose. (B) Time courses of PAN concentration (closed symbols) and OD₆₀₀ (open symbols). Triangles, transformant with pETDuet-Δ44atr-cyp79A2 and pGro7; squares, transformant with pETDuet-Δ44atr-cyp79A2, pCDF-oxd, and pGro7. Both transformants were grown in LB-based autoinduction medium containing 2 mg/ml arabinose. Values are means and error bars show 95% confidence limits from replicate assays.

FIG 5

Effects of the cultivation temperature shift on the concentrations of E,Z-PAOx (A) and PAN (B). The transformants with pETDuet-Δ44atr-cyp79A2 and pGro7 (A) and pETDuet-Δ44atr-cyp79A2, pCDF-oxd, and pGro7 (B) were grown in optimized 2% Casamino Acids autoinduction medium at 26°C for the first 24 h (until the time indicated by the dashed line), and the cultivation temperatures were then shifted to 16°C (closed circles), 20°C (closed triangles), 26°C (open triangles), 30°C (open squares), and 37°C (open circles). Values are means and error bars show 95% confidence limits from replicate assays.

FIG 6

Inhibition of E. coli growth by E,Z-PAOx and PAN. The culture broth of E. coli C41(DE3) (without any plasmid) cultivated in LB medium at 37°C for 12 h with shaking at 300 rpm was diluted in fresh LB medium containing different concentrations of E,Z-PAOx and PAN to an OD₆₀₀ of 0.01. OD₆₀₀ was measured after another 12-h cultivation under the same conditions.