Synthesis of Stereoenriched Piperidines via Chemo-Enzymatic Dearomatization of Activated Pyridines

The development of efficient and sustainable methods for the synthesis of nitrogen heterocycles is an important goal for the chemical industry. In particular, substituted chiral piperidines are prominent targets due to their prevalence in medicinally relevant compounds and their precursors. A potential biocatalytic approach to the synthesis of this privileged scaffold would be the asymmetric dearomatization of readily assembled activated pyridines. However, nature is yet to yield a suitable biocatalyst specifically for this reaction. Here, by combining chemical synthesis and biocatalysis, we present a general chemo-enzymatic approach for the asymmetric dearomatization of activated pyridines for the preparation of substituted piperidines with precise stereochemistry. The key step involves a stereoselective one-pot amine oxidase/ene imine reductase cascade to convert N-substituted tetrahydropyridines to stereo-defined 3- and 3,4-substituted piperidines. This chemo-enzymatic approach has proved useful for key transformations in the syntheses of antipsychotic drugs Preclamol and OSU-6162, as well as for the preparation of two important intermediates in synthetic routes of the ovarian cancer monotherapeutic Niraparib.


■ INTRODUCTION
The ubiquity of saturated nitrogen heterocycles (N-heterocycles) in natural products and pharmaceuticals continues to drive the development of innovative strategies for their efficient synthesis. 1,2In particular, chiral piperidines are much sought after structures due to their prevalence as scaffolds in a range of bioactive molecules including market-approved active pharmaceutical ingredients (APIs). 3Nature provides highly efficient biocatalysts for the biosynthesis of N-heterocycles, 4,5 offering high enantio-and regio-selectivity under benign conditions.−10 However, the translation of these methods to the corresponding stereoenriched 3substituted and 3,4-substituted scaffolds, the core of many important therapeutic compounds (Figure 1A), remains challenging due to difficulties in stereoselectivity control combined with limited availability of suitable starting materials.
Asymmetric chemical synthetic approaches for the preparation of 3-substituted and 3,4-disubstituted piperidines include those based on metalation/cross-coupling, 11−14 Grignard Michael addition, 15 ring closure, 16 and transition-metalcatalyzed dearomatization of pyridines. 17−21 However, limitations are associated with all of these approaches, including high reaction temperatures, sensitivity to moisture, lack of availability of starting materials, and the use of expensive noncommercial chiral ligands. 11,22−26 Whilst nature has yielded pyridine synthases to prepare pyridines, 27 an effective biocatalyst for their dearomatization is yet to be discovered.With this in mind, we sought to combine mild chemical reduction of pyridiniums to tetrahydropyridines (THPs) with the exquisite stereoselectivity of a biocatalytic cascade to reduce the final C�C bond as an efficient strategy for asymmetric dearomatization of activated 3-and 3,4-substituted pyridines (Figure 1C).Biocatalysts with broad substrate scope for the reduction of C�C bonds require the conjugation of the alkene to an electron-withdrawing group.Recently, C�C bonds conjugated to C�N bonds have been shown to undergo full reduction to amines through the combination of ene-reductases (EREDs) and imine reductases (IREDs), 6 as well as the newly discovered ene imine reductase (EneIR-EDs). 28We reasoned that biocatalytic oxidation, using an amine oxidase (AmOx), of the THP in situ would generate the corresponding dihydropyridiniums (DHPs), generating an activated C�C bond conjugated to the C�N bond, which could then be reduced with these biocatalysts to generate a cascade to the desired 3-and 3,4-substituted piperidines.This cascade complements a previous amine oxidase AmOx-IRED deracemization processes in which only amine oxidation and C�N bond reduction take place. 29,30RESULTS AND DISCUSSION A series of substituted N-alkyl THPs 1b-21b was prepared in good yields (50−90%) from activated pyridines (1a-21a) using NaBH 4 as previously reported. 31Initially, we explored the conversion of THPs to piperidines using AmOxs in combination with EREDs or EneIREDs (see Supporting Information 2.1.;Figures S1−S5 for the complete list of THPs screened).For the first step, we tested AmOx variants that have been shown to be effective biocatalysts for the oxidation of N-alkyl THPs. 32,33The 6-hydroxy-D-nicotine oxidase (6-HDNO) variant, E350L/E352D, 34 was found to be effective, with a broad substrate scope, including oxidation of 1b, a precursor to Preclamol.We next screened for activity for the second step, namely, reduction of the C�C bond of the α,β-unsaturated iminium ion.Whereas the panel of EREDs displayed no activity, the EneIRED from an unidentified Pseudomonas sp.(EneIRED-01), 28 in combination with 6-HDNO, was effective at reducing a number of THPs and could be used to prepare piperidine (R)-1c in good yield and with excellent enantioselectivity (see Supporting Information 2.1.,Table S1; entry 1−3, ≥42% yield, 96% ee).
With effective EneIREDs for the preparation of both enantiomeric series, we probed the substrate scope of the 6-HDNO-EneIRED cascade (Table 1).Enzymes in Series A and B accepted a variety of aryl substituents at the C-3-position of the THP scaffold, affording products 1c-7c in high yields, conversion, and enantioselectivity.Five-membered heterocyclic 3-substituents such as furan 8c and thiophene 9c were also  well tolerated (≥62% yield, ≥86% ee).Sterically demanding substrates, for example, containing a 2-naphthyl substituent 10b, were also tolerated producing (R)-and (S)-10c in excellent yield and stereoselectivity (73% yield, >99% conv., ≥94% ee).Additionally, a variety of N-alkyl substituents were accepted forming the piperidine products 1c-12c in good to excellent yields.Of these, N-allyl 3c-4c and N-propargyl 12c piperidines provide useful synthetic handles that can be easily removed or further functionalized.
More hindered 3,4-disubstitituted THPs could also be reduced using the cascade, resulting in the formation of the substrate-dependent cis or trans isomer.These included substrates in which both substituents were simple alkyl (3,4dimethyl) 17b or part of a fused bi-cyclic ring system (octahydroisoquinoline) 18b.We then probed the tolerance for a combination of alkyl and aryl substituents at C-3 and C-4.3-Phenyl-4-methyl disubstituted compounds 19b and 20b provided the corresponding piperidines (A)-19c and (A)-20c in excellent yields and diastereoselectivity (cis major; >96:4 dr; ≥87% yield).The system also tolerated the isomeric 3-methyl-4-phenyl disubstituted THP 21c.In addition to the inversion of stereochemical outcome at the C-3 position (Series A vs Series B), we also discovered some EneIREDs in Series B that provided an inverted diastereomeric configuration of the 3,4disubstitituted piperidines 17c-20c compared to Series A.
To probe the mechanism of the 6-HDNO-EneIRED cascade, the conversion of THP 6b to piperidine 6c was investigated by in situ 19 F NMR reaction monitoring (Figure 2A).After <5 min, formation of piperidine 6c was apparent and after 60 min, the THP 6b was completely consumed.In the absence of the EneIRED, 6-HDNO catalyzed the previously reported aromatization of THP 6b to the corresponding pyridinium ion 6a (see Supporting Information 2.3., Figure S6, >99% conversion after 24 h). 33As expected, no direct reduction of the THP 6b to piperidine 6c was observed in the absence of 6-HDNO, suggesting that a transiently generated dihydropyridinium (DHP) is the substrate for the EneIRED reduction.Using the cascade with THP-10b, we were able to isolate the enamine 10e before full conversion to the piperidine 10c (see Supporting Information 5.1.,63% yield), which strongly suggested the participation of this compound in the reaction pathway.Accordingly, enamine 10e was converted to piperidine 10c using EneIRED-01 alone, in excellent yield and high enantioselectivity, equivalent to the full cascade with THP-10b (Figure 2B, top; 88% yield, 94% ee).Deuterium labeling experiments were also implemented to further elucidate the mechanism of the EneIRED enamine reduction step.Carrying out the reduction of enamine 10e using EneIRED-01, in the presence of GDH and D-glucose-1- d 1 to generate deuterated NAD(P)D in situ, C-2-monodeuterated 10c-d 1 was formed (80% deuterium incorporation, 82% yield).This suggests that the enamine intermediate 10e may undergo protonation to the iminium before NAD(P)H hydride delivery (Figure 2B, bottom and see Supporting Information 5.2.).
A proposed mechanism for the 6-HDNO-EneIRED cascade is outlined in Figure 2C.For illustrative purposes, we have depicted the transformation of THP-3b to (S)-3c as this product is used in the subsequent synthesis of the key intermediate for Niraparib described below.Initially, 6-HDNO oxidation of THP-3b results in the activation of the THP C� C bond for EneIRED-catalyzed asymmetric conjugate reduction of DHP-3d at the expense of NADPH to generate enamine 3e (Cycle 1; step iii).This intermediate 3e is expected to be in equilibrium with chiral iminium 3f via a nonselective protonation in solution, which has been extensively documented. 9,36,37Depending on the EneIRED employed (Series A or B), the kinetic selective reduction of one enantiomer of chiral iminium 3f affords the enantioenriched piperidine 3c as the final product via reduction with a second molecule NADPH (Cycle 2; step vii).In situ epimerization of the enantiomer in 3f via enamine 3e enables a dynamic kinetic resolution (DKR) to occur to generate piperidine (S)-3c mediated by the EneIRED.
Predominantly, EneIREDs in Series A yielded (R)piperidines whilst enzymes in Series B such as EneIRED-07 yielded the (S)-product.In order to gain insight into the mode of substrate binding in the active site, we determined the structure of EneIRED-07 from Micromonospora sp.Rc5 to a resolution of 2.55 Å in complex with NADP + using X-ray crystallography.Crystals were obtained in the P2 1 2 1 2 space group and featured six molecules in the asymmetric unit, representing three dimers.EneIRED-07 displays the canonical fold observed for IREDs, 38,39 with two monomers associating to form two active sites between the N-terminal Rossmann domain of one subunit and the C-terminal helical bundle of its neighbor (Figure 2D).Analysis using the DALI server 40 suggested that the IRED with the most closely related structure was the IRED from Streptosporangium roseum (PDB 5OCM) with an rmsd of 1.0 Å over 288 Cα atoms.Following building and refinement of the protein atoms, clear omit density was observed in each active site corresponding to the cofactor NADP + .The iminium intermediate (S)-3f, the preferred enantiomer for EneIRED-07 imine reduction, was modeled into the active site using Autodock Vina (Figure 2E). 41The model suggests that the allyl group of (S)-3f is bound within a hydrophobic pocket formed by methionine residues M125, M183, and M214 at the rear of the active site as shown; the para-bromo-phenyl group projects toward the front of the active site bordered by L180, W184, and the NADP + cofactor.This ligand conformation places the electrophilic carbon approximately 3.6 Å from the NADP + pyridinium ring C4 atom, from which hydride is transferred.Modeling of the (R)enantiomer of 3f places the allyl group at the base of the active site with less favorable interactions with hydrophilic residues Y225 and D238 (see Supporting Information 7.4., Figure S13).
−45 This was accomplished by the comparison of experimental infrared (IR) and VCD spectra to density functional theory (DFT)-calculated spectra of a specific configuration.Because this series of molecules was of low molecular weight with a limited number of low energy conformations (fewer than 10 in each case), four different DFT methods were completed for each compound.We tested two functionals (B3LYP and B3PW91) each with two basis sets (6-31G(d) and cc-pVTZ) to see which would have the best statistical results in each case. 46As expected, all five piperidine methods yielded consistent results for each enantiomer, with the best results coming from cc-pVTZ/ B3PW91 for piperidines 6c and 9c, cc-pVTZ/B3LYP for piperidines 7c and 8c, and 6-31G(d)/B3PW91 for piperidine 10c.Neighborhood similarity values for IR and VCD, as well as confidence level (≥93%) were obtained using BioTools (Jupiter, Fl) CompareVOA software (see Supporting Information 4.3., and 12.). 47,48Because of the similarity of the chiral core, the VCD experimental spectra were very similar for all tested compounds.Future work could therefore forego the calculations in order to streamline the process.
To further demonstrate the application of the cascade, we synthesized the two intermediates 23 and 24 en route to Niraparib (Figure 3C), the first poly ADP ribose polymerase (PARP) inhibitor to be approved as a first-line monotherapeutic for the maintenance treatment of patients with advanced ovarian cancer. 49For route I, we showed that commercially available 3-(4-bromophenyl)pyridine could be efficiently converted to piperidine (S)-3c in just three steps and 61% overall yield (99% ee).This was followed by deallylation and N-Boc-protection to yield (S)-23 in 64% yield, a key intermediate in Merck's second-generation synthesis. 50lternatively, in route II, by starting from commercially available 4-(pyridin-3-yl)aniline, we converted pyridinium salt 4a to (S)-24 in 29% overall yield and with 93% ee, a key intermediate in Merck's first-generation synthesis. 51The general applicability of the method was also showcased by the preparation of the corresponding (R)-enantiomers of both 23 and 24 in good yields and high enantioselectivity (see Supporting Information 5.1.).
In summary, we report the development of a versatile and highly efficient chemo-enzymatic dearomatization of activated pyridines for the preparation of stereo-enriched 3-and 3,4disubstituted piperidines.The 6-HDNO-catalyzed oxidation of readily accessible THPs facilitates EneIRED-catalyzed conjugate reduction and iminium reduction to yield a broad range of chiral piperidines.The short syntheses of both enantiomers of Preclamol and OSU6162, as well as chiral precursors to Niraparib, highlight the flexibility and utility of the method presented, emphasizing the advantages of combining chemical synthesis with biocatalysis for developing new catalytic methods for the preparation of important chiral compounds.Furthermore, the increasing ability to systematically screen large panels of biocatalysts against new targets leads to the rapid identification of enzymes with applications in asymmetric synthesis.

Figure 2 .
Figure 2. Proposed mechanism for the 6-HDNO-EneIRED cascade.(a) Kinetic profile from in situ 19 F NMR reaction monitoring of THP-6b.a Reactions run at 1 mmol.(b) Enamine 10e is used to probe the role of this species as an intermediate in the cascade.(c) Proposed catalytic sequence for the AmOxEneIRED biocatalytic cascade.b Two equivalents of NAD(P)H consumed to form piperidine.(d) Structure of the dimer of EneIRED-07 in the ribbon format with subunits shown in green and blue.NADP + can be seen bound at the two active sites.(e) Active site of EneIRED-07 with the (S)-enantiomer of iminium ion intermediate 3f modeled into the active site.
Experimental procedures including characterization of compounds, spectroscopic data of analytical biotransformations, and control experiments (PDF) ■ AUTHOR INFORMATION Corresponding Author Nicholas J. Turner − Department of Chemistry, University of Manchester, Manchester Institute of Biotechnology, Manchester M1 7DN, United Kingdom; orcid.org/0000-