A highly productive, whole-cell DERA chemoenzymatic process for production of key lactonized side-chain intermediates in statin synthesis

Abstract

Employing DERA (2-deoxyribose-5-phosphate aldolase), we developed the first whole-cell biotransformation process for production of chiral lactol intermediates useful for synthesis of optically pure super-statins such as rosuvastatin and pitavastatin. Herein, we report the development of a fed-batch, high-density fermentation with Escherichia coli BL21 (DE3) overexpressing the native E. coli deoC gene. High activity of this biomass allows direct utilization of the fermentation broth as a whole-cell DERA biocatalyst. We further show a highly productive bioconversion processes with this biocatalyst for conversion of 2-substituted acetaldehydes to the corresponding lactols. The process is evaluated in detail for conversion of acetyloxy-acetaldehyde with the first insight into the dynamics of reaction intermediates, side products and enzyme activity, allowing optimization of the feeding strategy of the aldehyde substrates for improved productivities, yields and purities. The resulting process for production of ((2S,4R)-4,6-dihydroxytetrahydro-2H-pyran-2-yl)methyl acetate (acetyloxymethylene-lactol) has a volumetric productivity exceeding 40 g L(-1) h(-1) (up to 50 g L(-1) h(-1)) with >80% yield and >80% chromatographic purity with titers reaching 100 g L(-1). Stereochemical selectivity of DERA allows excellent enantiomeric purities (ee >99.9%), which were demonstrated on downstream advanced intermediates. The presented process is highly cost effective and environmentally friendly. To our knowledge, this is the first asymmetric aldol condensation process achieved with whole-cell DERA catalysis and it simplifies and extends previously developed DERA-catalyzed approaches based on the isolated enzyme. Finally, applicability of the presented process is demonstrated by efficient preparation of a key lactol precursor, which fits directly into the lactone pathway to optically pure super-statins.

Figure 1. Current art in the direct synthesis of optically pure super-statins from the lactol precursors.

Figure 2. The DERA-catalyzed sequential aldol condensation with 2-substituted acetaldehydes.

Several acetal/hemiacetal (9) species were found in equilibrium with 8. These may include monomeric cyclic hemiacetal as well as dimeric and trimeric cyclic acetals/hemiacetals (Information S3). Compounds 15 are obtained by chemical oxidation of the lactols 3.

Figure 3. DERA activity measurements of the whole-cell catalyst.

A. Fluorescence raw data for a DERA activity assay (dotted lines). Velocities for triplicate samples of the whole-cell catalyst were measured for 7 different loads (b–h, 3.16 µg–26.9 µg in 3.96 µg increments) of biomass. After normalization with the blank (a), maximum slopes were determined for each sample and averaged (solid lines) to yield velocity for a given biomass load. B. Velocity vs. biomass load plot. The first 5 points are taken for the specific activity calculation. Linear regression: y = 0.2366x+0.2073 R² = 0.9936 C. Comparison of velocities measured for cell-free lysate spiked with increasing loads of biomass. D. Validation of linearity of the activity assay within samples with constant biomass. The whole-cell catalyst E. coli BL21 (DE3) pET30/deoC was mixed with w.t. E. coli BL21 (DE3) biomass (•). Linear regression: y = 248.94x+1.3840, R² = 0.9995. In parallel, sonicated and cleared samples were measured (□). Linear regression: y = 235.00x+2.6433, R² = 0.9989.

Figure 4. Inactivation of DERA whole-cell catalyst and DERA cell-free lysate with various aldehydes.

Samples were treated with 75 mM, 150 mM and 225 mM substrate aldehydes for 15 minutes prior to the activity assay. The specific DERA activity was 226.8 kRFU s⁻¹ g⁻¹ and 226.6 kRFU s⁻¹ g⁻¹ for the whole-cell catalyst and for the cell-free lysate, respectively. Residual activities are given relative to non-treated whole-cell catalyst. Aldehydes used were acetaldehyde (A), 2b (B) and 2g (C).

Figure 5. Time course of whole-cell, DERA-catalyzed batch reactions.

Reactions were performed using E. coli BL21 (DE3) pET30/deoC fermentation cultures directly (DERA specific activity = 232 kRFU s⁻¹ g⁻¹, WCW = 207 g L⁻¹). Results are given as mass concentrations obtained from GC-FID analysis. The measured quantity of a particular compound, with the exception of the stable 6-ring hemiacetals (3), represents the sum of the corresponding equilibrium forms (hydrate, aldehyde and acetal/hemiacetal), which exist under the reaction conditions. A: Reaction species data from reactions using 400 mmol L⁻¹ of 2g and 840 mmol L⁻¹ of 1 are shown. 1 (▪, black), 3a (▴, blue) 3g (♦, green), 8g (•, red), 10g (Δ, orange) and 2g (◊, brown). B: Reaction species data from reactions using 400 mmol L⁻¹ of 2b and 840 mmol L⁻¹ of 1 are shown. 1 (▪, black), 3a (▴, blue) 3b (♦, green), 8b (•, red), 10b (Δ, orange), 2b (◊, brown), 2,6-chloro-2,4-dideoxyhexose (□, grey). Concentration of the latter (Information S8) is evaluated based on the assumption, that the GC-FID response factor is similar to that of 3b.

Figure 6. Compounds influencing the production of acetiloxy-lactol (3g).

Hydrate forms of the aldehydes (11, 12 and 13) are not depicted here. A: Reaction species arising from acetaldehyde alone. B: Reaction species arising from acetaldehyde 1 and 2g. Several acetal/hemiacetal (9g) species were found in equilibrium with 8g. These may include monomeric cyclic hemiacetal as well as dimeric and trimeric cyclic acetals/hemiacetals (Information S3).

Figure 7. Time course of whole-cell, DERA-catalyzed, fed-batch reactions yielding 3g.

Reaction species data from three independent experiments using (in total) 550 mmol L⁻¹ of 2g and 1200 mmol L⁻¹ of 1 are shown. Whole-cell catalyst (E. coli BL21 (DE3) pET30/deoC high-density culture) with 217 kRFU s⁻¹ g⁻¹ DERA specific activity and 182 g L⁻¹ WCW was used. Results are given as molar concentrations obtained from GC-FID analysis. The measured quantity of a particular compound, with the exception of the stable 6-ring hemiacetals (3), represents the sum of the corresponding equilibrium forms (hydrate, aldehyde and acetal/hemiacetal) which exist under the reaction conditions. 1 (▪, black), 3a (▴, blue) 3g (♦, green), 8g (•, red), 10g (Δ, orange), 2g (◊, brown), cumulative molarity of reaction species originating from 2g (□, grey; sum of 2g, 8g, 10g and 3g ). Secondary vertical axis shows in %: residual DERA activity (□, violet), cumulative feed of 2g (dotted line), cumulative feed of 1 (dashed line).

Figure 8. Time course of exemplary whole-cell, DERA-catalyzed, fed-batch reactions with ∼50 g L⁻¹ h⁻¹ volumetric productivity.

Whole-cell catalyst (E. coli BL21 (DE3) pET30/deoC high-density culture) with 247 kRFU s⁻¹ g⁻¹ DERA specific activity and 215 g L⁻¹ WCW was used. Results are given as mass concentrations obtained from GC-FID analysis. The measured quantity of a particular compound, with the exception of the stable 6-ring hemiacetals (3), represents the sum of the corresponding equilibrium forms (hydrate, aldehyde and acetal/hemiacetal) which exist under the reaction conditions. A: Reaction species data from reaction using (in total) 700 mmol L⁻¹ of 2g and 1540 mmol L⁻¹ of 1 are shown. 1 (▪, black), 3a (▴, blue) 3g (♦, green), 8g (•, red), 10g (Δ, orange), 2g (◊, brown), acetic acid (□, grey). Secondary vertical axis shows in %: cumulative feed of 2g (dotted line), cumulative feed of 1 (dashed line). B: Reaction species data from reaction using (in total) 700 mmol L⁻¹ of 2b and 1540 mmol L⁻¹ of 1 are shown. 1 (▪, black), 3a (▴, blue) 3b (♦, green), 8b (•, red), 10b (Δ, orange), 2b (◊, brown), acetic acid (□, grey), 2,6-chloro-2,4-dideoxyhexose (□, purple). Secondary vertical axis shows in %: cumulative feed of 2b (dotted line), cumulative feed of 1 (dashed line).