Synthesis of [4S‐2H]NADH, [4R‐2H]NADH, [4‐2H2]NADH and [4‐2H]NAD+ cofactors through heterogeneous biocatalysis in heavy water

This practitioner protocol describes the synthesis of a family of deuterated nicotinamide cofactors: [4S‐2H]NADH, [4R‐2H]NADH, [4‐2H2]NADH and [4‐2H]NAD+. The application of a recently developed H2‐driven heterogeneous biocatalyst enables the cofactors to be prepared with high (>90%) 2H‐incorporation with 2H2O as the only isotope source.


| INTRODUCTION
Nicotinamide cofactors in their oxidised (NAD + ) and reduced (NADH) forms ( Figure 1) mediate electron transfer in a wide range of redox enzyme (oxidoreductase)-catalysed reactions. Isotopic labelling of these molecules is a well-established technique for probing enzymatic mechanisms, 1 exemplified by the work of JW Cornforth on the stereoselectivity of biocatalysed processes, for which he was awarded the 1975 Nobel Prize in Chemistry. 2,3 More recently, deuterated NADH has been shown to be a useful tool for the preparation of synthetically challenging asymmetric heavy drug analogues. 4 Deuterated and tritiated forms of NADH may be prepared by reducing NAD + in 2 H 2 O (D 2 O)/ 3 H 2 O (T 2 O) with sodium dithionite, 2 or in 1 H 2 O with a deuteride/tritide salt. 5 Although simple, these methods lack selectivity as to which face of the nicotinamide ring the 2 H/ 3 H atom is added. Alternatively, an oxidoreductase of the desired selectivity may be supplied with NAD + and a 2 H-or 3 H-labelled substrate. 6 [7][8][9][10][11] Despite the elegance of these methods, they still require a 2 H-or 3 H-labelled organic precursor, which reduces the atom economy of the process, and may add complexity in, for example, the separation of the product.
We have recently reported a route to deuterated NADH that uses H 2 as the reducing agent and 2 H 2 O as the source of deuterium isotopes. 4 This method utilises a heterogeneous biocatalyst, whereby a hydrogenase (capable of oxidising H 2 ) is co-immobilised with an NAD + reductase (capable of reducing NAD + ) on an electronically conducting carbon support. Separation of the two enzymatic sites allows for high levels of deuterium incorporation into the cofactor, but avoids the use of a labelled reductant. The selectivity of the NAD + reductase also ensures that only [4S-2 H]NADH is formed.
In this practitioner protocol, we describe how, by extension of the heterogeneous biocatalytic method described above, [4R- 2

| Synthetic approach
The approach shown in Figure 2 was used to prepare different isotopologues of NAD + and NADH. In brief, [4S-2 H]NADH was first prepared from NAD + by using the H 2 -driven heterogeneous biocatalyst (Biocat/C) operating in 2 H 2 O (Step A). This [4S-2 H] NADH was then re-oxidised by reaction with excess acetophenone (AcPh) and a commercial ADH immobilised on carbon (ADH/C). The ADH was chosen such that it selectively removed 1

H to leave [4-2 H] NAD + (Step B).
Following the removal of the ADH/C, unreacted AcPh and phenylethanol (PhEtOH) side-product, the [ Step D). In all cases, the removal of the catalyst was easily achieved by centrifugation, owing to the carbon support employed to heterogenise the catalysts.
In the work reported here, the crude reaction mixtures were used directly after each step, and no purified or isolated cofactors were prepared at any stage. However, if further purification or isolation was required (particularly to remove buffer salts, unreacted starting material or impurities arising from cofactor degradation), we refer the reader to suitable protocols reported in this journal 8 and elsewhere. 12

| General reagents and conditions
All commercial reagents were used as received and without further purification, unless specified. NAD + and NADH were purchased from Prozomix * and all other reagents from Sigma Aldrich. † Deionised MilliQ water (Millipore, 18 MΩ cm) was used for nondeuterated solutions, and 2 H 2 O (99.98%, Sigma Aldrich) was used for deuterated solutions. When required, Trizma base was exchanged into 2 H 2 O prior to use to minimise 1 H contamination and then adjusted to the required p 2 H (pD) with 2 HCl. All aqueous solutions were sparged with dry N 2 for 60 min prior to use in order to deoxygenate them.
The heterogeneous biocatalyst (Biocat/C) comprised a NiFe-hydrogenase (Hydrogenase 1 from Escherichia coli) and an NAD + reductase (I64A variant of the soluble hydrogenase from Ralstonia eutropha) coimmobilised on carbon black particles (Black Pearls 2000 [BP2000], Cabot Corporation) ‡ according to published protocols. 4 In a typical experiment, 400 μg of carbon (supporting around 80-140 μg each of hydrogenase and NAD + reductase) was used per 1 ml of reaction mixture. The commercial ADH (ADH105, Johnson Matthey) § was immobilised on carbon particles by suspending 50 μl of concentrated solution (20 mg/ml) to an equal volume of sonicated BP2000 particles (20 mg/ml) for 30 min at 4 C. Following centrifugation (13,800 × g, 5 min), the supernatant was removed, and the particles were washed once with 100-μl 2 H 2 O.
The reactions reported here were set up in a glovebox with an N 2 atmosphere (<0.1-ppm O 2 ), and those with pressurised H 2 were conducted in a Büchi Tinyclave. ¶ It has been demonstrated elsewhere that a rigorously anaerobic setup is not required, and that conventional benchtop synthetic apparatus (N 2 -flushed round-bottom flask, H 2 balloon) can be used as an alternative. 4 Products were analysed using UV-vis and 1 H nuclear magnetic resonance (NMR) (400 MHz) spectroscopies (Figure 3), according to standard procedures. 13

| Step B: Synthesis of [4-2 H]NAD + from [4S-2 H]NADH
To a deoxygenated solution of [4S-2 H]NADH (1 ml, 4 mM) in [ 2 H 5 ]-Tris. 2 HCl (100 mM, p 2 H 8.0), ADH/C particles (2,000 μg) and AcPh (10 mM) were added. The whole solution was shaken under N 2 for a period of 5 days. The ADH/C particles were removed by centrifugation (13,800 × g, 5 min), and the remaining acetophenone and formed phenylethanol were removed by extraction with CHCl 3 (3 × 500 μl). A conversion of 95% to the [4-2 H]NAD + form was observed, with full retention of 2 H at the 4-position of the nicotinamide ring from the starting material. It was possible to conduct Step B in a shorter time using the ADH in solution (rather than immobilised), though this required subsequent precipitation of the enzyme by addition of acetonitrile (1 ml) and removal by centrifugation (13,800 × g, 5 min).

| Step C: Synthesis of [4-2 H 2 ]NADH from [4-2 H]NAD +
Following on from Step B, Biocat/C were added at a loading of 200 μg ml −1 to a solution of [

| Step D: Synthesis of [4R-2 H]NADH from [4-2 H]NAD +
A solution of [4-2 H]NAD + (1 ml, 3 mM) in [ 2 H 5 ]-Tris. 2 HCl (100 mM, p 2 H 8.0) was prepared according to Step B, and an equal volume of acetonitrile was added to facilitate precipitation during the drying stage. The solution was gently evaporated to dryness under reduced pressure on a rotary evaporator (at 20 C), and the solid was redissolved in deionised 1 H 2 O. The evaporation and redissolution steps were repeated twice more to fully exchange all of the remaining 2 H to 1 H. The solution was evaporated a final time to leave an off-white solid consisting of cofactor and buffer salts. The solid was transferred to a glovebox and redissolved in deoxygenated deionised 1 H 2 O (1 ml). The cofactor solution was subsequently saturated with 1 H 2 gas. An aliquot (200 μg) of Biocat/C catalyst was added to the solution, which was subsequently sealed under 2-bar H 2 and shaken at 30 rpm for 18 h. The Biocat/C catalyst was removed by centrifugation (13,800 × g, 5 min). Subsequent analysis indicated around 50% conversion to the reduced cofactor, with a high selectivity (>90%) for the [4R-2 H]NADH isotopologue. Some degradation of the cofactor was observed during the evaporation steps, and separation by ion exchange chromatography would be needed to obtain a pure sample in this case. 8,12

| Product analysis
UV-vis and 1 H NMR (400 MHz) spectra arising from the analysis of the products in Steps A-D are shown in Figure 3 alongside commercial samples of NAD + and NADH. The ratio of oxidised to reduced cofactor in the samples was determined by comparing the ratio of the absorption at 340 nm with that at 260 nm in the UVvis spectra. 14 Analysis of the isotopic composition of the cofactors was achieved by analysing diagnostic peaks in the 1 H NMR spectra. In the case of the oxidised cofactors, the doublet at δ 8.80 ppm is assigned to the proton at the 4-position of the nicotinamide ring. As expected, the signal seen in the spectrum of NAD + ( Figure 3A) is almost absent for [4-2 H]NAD + ( Figure 3B). In the case of the reduced cofactors, the diagnostic region for the protons on the 4-position of the nicotinamide ring is between δ 2.60 and δ 2.85 ppm. 15 In the unlabelled commercial NADH sample ( Figure 3C), the two heavily roofed doublets at δ 2.80 and 2.68 ppm correspond to the pro-R and pro-S protons, respectively. Upon single deuteration at the 4-position to give either [4S-2 H] NADH ( Figure 3D) or [4R-2 H]NADH ( Figure 3E), the signals collapse into pseudo-singlets at δ 2.77 and δ 2.67 ppm, respectively. Double deuteration at the same site to give [4-2 H 2 ]NADH leads to almost the complete disappearance of peaks in this region, as expected ( Figure 3F). H]NAD + . The main advantages over previous synthetic methods arise from the atom economy of using H 2 gas as a reductant and the simplicity of using 2 H 2 O as an isotope source, avoiding the requirement for carbon-based co-reagents. Furthermore, the inherent heterogeneous nature of the H 2 -driven biocatalyst makes reaction work-up straightforward, and similar strategies have been demonstrated for translation into continuous flow. 16 Whilst the work presented here concerns cofactor deuteration, preliminary experiments in 3 H 2 O (37 MBq g −1 ) # indicate that the same procedures should be translatable for similar tritiation.