Metagenomic ene-reductases for the bioreduction of sterically challenging enones

Ene-reductases (ERs) of the Old Yellow Enzyme family catalyse asymmetric reduction of activated alkenes providing chiral products. They have become an important method in the synthetic chemists' toolbox offering a sustainable alternative to metal-catalysed asymmetric reduction. Development of new biocatalytic alkene reduction routes, however needs easy access to novel biocatalysts. A sequence-based functional metagenomic approach was used to identify novel ERs from a drain metagenome. From the ten putative ER enzymes initially identified, eight exhibited activities towards widely accepted mono-cyclic substrates with several of the ERs giving high reaction yields and stereoselectivities. Two highly performing enzymes that displayed excellent co-solvent tolerance were used for the stereoselective reduction of sterically challenging bicyclic enones where the reactions proceeded in high yields, which is unprecedented to date with wild-type ERs. On a preparative enzymatic scale, reductions of Hajos–Parish, Wieland–Miescher derivatives and a tricyclic ketone proceeded with good to excellent yields.


Table of Contents
Synthesis of (4aR,5R,8aS)-5,8a-dimethylhexahydronaphthalene-1,6(2H,5H)-dione (12 Total DNA was isolated from a sample collected from a domestic shower drainpipe. Hair blocking the drain was removed and a surrounding gel-like liquid mass (20 mL) was added to buffer (25 mM Tris-HCl, 150 mM NaCl and 25 mM EDTA) making the total volume of 250 mL. The sample was warmed to 60 °C for few minutes and 80 mL of phenol was added (buffered phenol solution, Sigma). The sample was mixed well and left, with occasional mixing, in a 65 °C water bath for 45 min, then centrifuged for 20 min at 7,000 rpm to separate the phases. The top aqueous phase was removed into two large centrifuge bottles and an equal volume of isopropanol was added to each. Samples were left on ice for 10 min then centrifuged for 10 min at 7,000 rpm. Each pellet was resuspended in 10 mL of TE buffer (10 mM Tris-HCl pH 7.5, 1 mM Na2EDTA) and added to 10 3 Overlapping domain assignments were resolved by DomainFinder3 as previously described. 4 The completeness of the candidate gene and the corresponding protein sequence was assessed by BLASTP searches against the non-redundant GenBank database. Theoretical molecular weights and extinction coefficients were calculated with the ProtParam tool (ExPASy).

Cloning of selected genes.
Metagenomic ERs were PCR amplified directly from the drain metagenomic DNA and cloned into pET29a(+) (Novagen, Merck). Genes were cloned to start with an ATG within the NdeI restriction site and in fusion with a C-terminal Hisx6 tag. Resulting expression plasmids were verified by DNA sequencing (Eurofins Genomics). A gene encoding N-terminal Hisx6 -tagged enzyme nicotinamidedependent cyclohex-2-en-1-one reductase (NCR) from Zymomonas mobilis (UniProtKB Accession: Q5NLA1) 5 was codon optimized for expression in E.coli, synthesized and cloned into pJ411 vector by DNA2.0, USA. The gene for the cofactor recycling enzyme glucose-6-phophate dehydrogenase from Saccharomyces cerevisiae SF838 (UniProtKB Accession: P11412) was amplified from genomic DNA and cloned into pACYCDuet (Novagen, Merck) using BamHI and SalI restriction sites.

Enzyme expression and purification.
All plasmids were transformed into E. coli expression strain BL21 (DE3). Precultures were grown overnight at 37 °C in Terrific broth (TB) containing kanamycin (50 μg/mL) and were used to inoculate fresh TB media (1% v/v). Cells were grown at 37 °C under shaking to an OD600 of 0.4 -0.6. Clarified lysates were stored in small batches at −80 °C.

B
Unrooted phylogenetic tree of OYE family members was computed with MEGA7 7 (ClustalW was used for multiple sequence alignment with default parameters; maximum likelihood distance tree was run with default parameters and 100 replicates (bootstrap values are shown)); tree was visualized with iTOL. 8 Protein accessions and name abbreviations of 63 characterized eukaryotic and prokaryotic OYEs were taken from Scholtissek et al. 9 Protein sequences of PpO-ER3, Rer_ER7, and Lla-ER are from Peters et al. 10 SDS-PAGE was run on 12% Mini-PROTEAN® TGX™ Precast Gels (Bio-Rad). Gels A, D, E: protein MW marker

General considerations
All chemicals were obtained from commercial suppliers and used as received. Thin layer chromatography was carried out using Merck TLC Silica gel 60 F254 plates and products were visualised using combinations of UV light (254 nm) and potassium permanganate staining solution.
Filtrations on silica were carried out using silica gel (particle size 40-60 μm). GC analyses were

Spectrophotometric assay
ERs spectrophotometric assays were performed with purified enzyme and in 96-well plates. The assays were initiated by the addition of NADPH (160 µL, 1.25 mM) dissolved in phosphate buffer (100 mM, pH = 8) to a mixture containing the substrate (10 µL, 100 mM in DMSO), water (25 µL) and the purified enzyme (5 µL in phosphate buffer). Depletion of NADPH was monitored for 1h 30 min using a UV-vis spectrometer at λ = 340 nm. In Table S2 the qualitative results of these assays are reported.

Biotransformations -general protocol with purified ERs
Purified enzymes were prepared from ammonium sulfate suspension by sampling the volume of enzyme needed into an Eppendorf tube. After centrifugation for 20 min at 4 °C at 13 000 rpm, the supernatant was discarded, and the residual pellet was dissolved in the desired volume of Tris buffer (50 mM) to reach a concentration between of 2.0 to 8.0 mg/mL depending on the enzyme.
Concentrations were measured using a NanoDrop UV-Vis spectrophotometer.   Biotransformations were carried out following the procedure described.

Biotransformations with clarified cell lysates
The bioconversion was started with the addition of 60 µL of co-expressed ene-reductase/G6PDH lysate ( Quantification of the product was performed using GC analysis (retention time and calibration curves are reported in the Analytical methods section). With S-5 the highest yields and stereoselectivities were observed with the ER from pQR1445 (0.5 mg/mL) and 10 mM substrate. These conditions were chosen for the preparative scale.  In comparison to S-5, S-8 could be used at higher concentrations (20 mM) but the best yields were observed with the ER from pQR1907 (1 mg/mL).
Integration of the signal corresponding to the starting material and product compared to the standard revealed that the starting material was converted at 60% and 50% of product was formed.  (Figures S9 -S11). † , 18,19 Once all the J couplings were determined (included in Figures S9 -S11), we then considered three possible molecular geometries together with the predicted 3 JHH values using Karplus-type equations of Haasnoot et al. 20 Compound 9 For compound 9, the values of trans-3 JHH couplings for proton 10 calculated for optimised molecular geometries of conformers using the MMX force field ‡ are shown below (cis and trans in the configuration labelling below refer to cis-and trans-fusion of the two cycles, while cis1 and cis2 indicate to the two possible conformers for the cis-fusion of the two cycles): The experimental values of trans-3 JHH couplings for proton 10 were 10.56 Hz with proton 5a and 4.75 Hz for proton 4e. These values agree best with those of 9-cis1 conformer shown above. Some     Assuming the predicted boundary values are accurate within ±0.5 Hz, 21 the uncertainty in the population determinations of two conformers is estimated to be ±4%. The preference of the 9-cis1 conformer is also confirmed by a 1D NOESY spectrum (Figure S12), which shows NOEs at 2.10 ppm for the proton pair (Me, 4a) and at 2.57 ppm for the proton pair (Me, 2a).

Compound 12
In a similar manner we have also considered predicted values of 3 J couplings in compound 12 for various plausible configurations and conformations (the first occurrence of cis and trans in the configuration labelling below refers to cis-and trans-fusion of the two cycles, while the second occurrence of cis and trans reflects the configuration of the two methyl groups; cis1 and cis2 indicate to the two possible conformers for the cis-fusion of the two cycles): The experimental values of trans-3 JHH couplings for proton 10 were 12.36 Hz with proton 5a and 2.80 Hz for proton 4e. These values agree best with those of 12-cis1-cis conformer shown above and, unlike compound 9, the presence of the second methyl group in position 5 leads to the predominance of the 12-cis1-cis conformation with a nearly 100% population, as the second possible conformer 12-cis2-cis is destabilised by interactions between two axial methyl groups. The preference of the 12-cis1-cis conformer is also confirmed by a 2D NOESY spectrum (Figure S13), which shows NOEs at (1.28 ppm, 2.14 ppm) for the proton pair (9-Me, 4a) and at (1.28 ppm, 2.66 ppm) for the proton pair  The trans-isomer 13 was also considered, which is expected to be conformationally homogeneous.
From the analysis of the NMR data (Figure S14),

Compound 15
For compound 15, the values of trans-3 JHH couplings for proton 9 calculated for optimised molecular geometries of conformers using MMX force field, § are shown below: The experimental values of trans-3 JHH couplings for the proton pairs (4a,9) was less than 6.5 Hz, hence the trans-fusion of the two cycles with the predominance of the conformer shown above as 15trans can be ruled out. This suggests that the methyl group and proton 9 have a cis-configuration. Figure S16. (a) 1D NOESY spectrum of 15 with selective excitation of the methyl protons at 1.24 ppm (mixing time 300 ms); (b) 1H NMR spectrum of 15. Spectra were recorded in CDCl3 at 25 °C using a 700 MHz spectrometer. From these spectra, protons of the 3-CH2 and 4-CH2 groups having a cisconfiguration with respect to the methyl group (as well as with proton 9) can be identified as those resonating at 2.12 ppm and 2.59 ppm, respectively.
From the 1D NOESY spectrum (Figure S16), protons of the 3-CH2 and 4-CH2 groups having a cisor trans-configuration with respect to the methyl group (hence with proton 9) can be identified, thus allowing to assign cis-and trans-3 JHH couplings of proton 9. The experimental values of trans-3 JHH couplings for proton pairs (4,9) and (3,9)

Chiral GC analysis (Cyclohexyl-derivatives)
Analyses were performed using an Agilent 7820A Gas Chromatograph equipped with a chiral column (Beta DEX 225, fused silica capillary column 30 m x 0.25 mm x 0.25 µm). The samples (5 µL) were injected with an autosampler tower G4513A and applied by split injection (ratio 20:1) at an injection temperature of 250 °C with a split flow of 36 mL/min and an oven temperature of 80 °C. For every substrate the sequence applied was as followed: