GreenChemistry Cutting-edge research for a greener sustainable future

The synthesis of chiral amines in enantioenriched form is a keystone reaction in applied chemical synthesis. There is a strong push to develop greener and more sustainable alternatives to the metal-catalysed methods currently used in the pharmaceutical, agrochemical and fine chemical industries. A biocatalytic approach using transaminase (TA or ATA) enzymes to convert prochiral ketones to chiral amines with unparalleled levels of enantioselectivity is highly appealing. However, the use of TA enzymes in synthesis is severely hampered by the unfavourable thermodynamics associated with the amine donor/acceptor equilibrium. Several ‘smart’ amine donors have been developed that leverage chemical and physical driving forces to overcome this challenging equilibrium. Alongside this strategy, enzyme engineering is typically required to develop TAs compatible with these non-physiological amine donors and the unnatural reaction conditions they require. We herein disclose N-phenylputrescine (NPP) as a readily accessible amine donor, inspired by the biosynthesis of the dipyrroloquinoline alkaloids. NPP is compatible with a broad range of synthetically useful TA biocatalysts and performs across an unparalleled variety of reaction conditions (pH and temperature). Synthetic applicability has been demonstrated through the synthesis of the anti-diabetic drug sitagliptin, delivering the product in excellent enantiopurity using just two equivalents of NPP.

The synthesis of chiral amines in enantioenriched form is a keystone reaction in applied chemical synthesis.
There is a strong push to develop greener and more sustainable alternatives to the metal-catalysed methods currently used in the pharmaceutical, agrochemical and fine chemical industries. A biocatalytic approach using transaminase (TA or ATA) enzymes to convert prochiral ketones to chiral amines with unparalleled levels of enantioselectivity is highly appealing. However, the use of TA enzymes in synthesis is severely hampered by the unfavourable thermodynamics associated with the amine donor/acceptor equilibrium.
Several 'smart' amine donors have been developed that leverage chemical and physical driving forces to overcome this challenging equilibrium. Alongside this strategy, enzyme engineering is typically required to develop TAs compatible with these non-physiological amine donors and the unnatural reaction conditions they require. We herein disclose N-phenylputrescine (NPP) as a readily accessible amine donor, inspired by the biosynthesis of the dipyrroloquinoline alkaloids. NPP is compatible with a broad range of synthetically useful TA biocatalysts and performs across an unparalleled variety of reaction conditions (pH and temperature).
Synthetic applicability has been demonstrated through the synthesis of the anti-diabetic drug sitagliptin, delivering the product in excellent enantiopurity using just two equivalents of NPP File list (2) download file view on ChemRxiv NPP-manuscript.pdf (663.07 KiB) download file view on ChemRxiv NPP-SI.pdf (5.

Introduction
The need to develop more sustainable manufacturing processes in the pharmaceutical, agrochemical, and other fine-chemical industries has led to an increased interest in biocatalytic methods. [1][2][3][4][5][6] By using biocatalysts, enantiopure compounds can often be produced in high yield under very mild aqueous reaction conditions, but they can suffer from several practical limitations and drawbacks. 7,8 For example, transaminases (TAs or ATAs) provide an environmentally friendly alternative to traditional transition-metal catalysed methods for the synthesis of chiral amines, which are ubiquitous building blocks in the pharmaceutical and agrochemical industries, and find wide application as chiral auxiliaries, catalysts, ligands, and resolving agents (Fig. 1). 9,10 Unfortunately, TA-mediated transamination reactions often suffer from unfavourable reaction equilibria, an issue that can critically limit many biocatalytic methods. 11 We herein report a biomimetic solution to this problem, which will facilitate the increased use of TA biocatalysts in sustainable synthesis. Biocatalytic transamination proceeds via a double-displacement mechanism (ping-pong bi bi mechanism), involving the cofactor pyridoxal 5´-phosphate (PLP). Overall, an amino group is transferred from an amino donor to a carbonyl acceptor (Fig. 2a). [15][16][17] The equilibrium that exists between the donor and acceptor can be problematic when applying TA biocatalysts in synthesis, particularly when the desired reaction is enthalpically unfavourable. Previous strategies for overcoming this issue have involved the use of excess amine donor and leveraging various chemical and/or physical phenomena to drive the equilibrium forward. 18,19 The most commonly employed amine donors, despite their unfavourable reaction equilibria, are L-alanine and iso-propylamine (i-PrNH2). L-Alanine is a widely accepted natural amine donor whose by-product, pyruvate, can be irreversibly consumed through coupled enzymatic processes. Although this is an effective strategy, it significantly increases production costs at scale. [20][21][22][23] i-PrNH2 is widely used in industry because evaporative loss of the acetone by-product can be used to drive the equilibrium forward. 24,25 Unfortunately, i-PrNH2 is not widely accepted by TA biocatalysts meaning costly enzyme engineering is required to enable acceptance of i-PrNH2 and to ensure the enzyme can tolerate the high concentrations of i-PrNH2/acetone and the elevated temperatures required to drive-off the acetone. 26 More complex amine donors have been developed, such as o-xylyenediamine 1, 2-(4-nitrophenyl)ethan-1-amine 2, and cadaverine 3, which can be very successfully applied in specific settings (Fig. 2b). 27-30 These 'smart' amine donors each have their own niche applications, such as in colourimetric assays or when tolerance to specific reaction conditions is required, but there remains an unmet need for a widely accepted and readily available amine donor which overcomes the donor/acceptor equilibrium problem. 18,31 In a recent review, Pavlidis and co-workers stated "…designing industrial processes with these enzymes [TAs] is still challenging, due to the fact that a universal and discernible amine donor system has not been developed". 26 We took inspiration for the design of a new amine donor from the proposed biosynthetic pathway of the plant-derived dipyrroloquinoline alkaloids, incargranine B (7) and seneciopiperidine (8) (Fig. 2c). 32 TA-mediated oxidative deamination of an N-arylputrescine 4 is proposed to give an enamine 6 and an iminium ion 5 that react together through an irreversible Povarov reaction to give the dipyrroloquinoline framework. The chemical feasibility of this pathway has been demonstrated through our biomimetic total synthesis and structural revision of incargranine B 7. 32 We herein disclose the design and synthesis of N-phenylputrescine 9 (NPP), which exploits this irreversible biosynthetic Povarov reaction to generate a dipyrroloquinoline by-product 10 that provides a thermodynamic driving force in the TA mediated synthesis of amines (Fig. 2d). As we envisaged, NPP 9 displaces the challenging transamination equilibrium when used in near equimolar quantities, is widely accepted by wild-type and engineered TA biocatalysts, and performs well under a broad range of reaction conditions.

Results and Discussion
NPP 9 can be synthesised through a simple two-step process from commodity chemicals; 33 alkylation of aniline with 4-chlorobutyronitrile followed by reduction with borane gives NPP free base 9, as an oil, in 80% yield on a gram scale. Alternatively, the hydrochloride salt, which is an easy to handle solid, can be similarly accessed in >80% yield in >25 g batches. This has allowed us to easily prepare over 100 g of NPP 9 salt, which can be stored under ambient conditions with no special precautions required (see Supplementary Section 2.2 and 2.3 for details).
We established NPP 9 as a suitable amine donor using Halomonas elongata TA (HEWT) with pyruvate as amine acceptor. 34,35 An adaptation of the L-amino acid oxidase (L-AAO) and horse radish peroxidase (HRP) coupled colourimetric assay was used to investigate the kinetics of L-alanine formation. 36  A quantitative screen (using the recommended conditions outlined in the Codexis ® Amine Transaminase (ATA) screening kit) was carried out with 24 commercially available Codexis TAs and two wild type TAs (Chromobacterium Violaceum TA (CVTA) and HEWT) with benzaldehyde 11 as amine acceptor (Fig. 3a). 38,39 Analysis by HPLC showed all except one of the TA biocatalysts accepted NPP 9, with 12 of the 24 Codexis TAs and both wild-type TAs giving good yields of benzylamine 13 ( Fig. 3b). With this initial success, the screen was repeated with the more challenging acetophenone 12 as amine acceptor. Encouragingly, similarly effective yields of methylbenzylamine 14 product were observed (Fig. 3b). It should be noted that the reaction conditions could be optimised for each individual biocatalyst if desired. Furthermore, as new TA biocatalysts continue to be discovered through metagenomic screening and directed evolution, we envisage expanded applications of NPP 9 in the synthesis of chiral amines. 25,40,41 During this initial screen, we noticed the formation of an off-white precipitate correlated with activity.
This observation was investigated further through a scale up of the reaction using HEWT with 100 mg of NPP 9, and benzaldehyde 11 as amine acceptor. The precipitate was isolated and analysed by NMR, IR and MS (see Supplementary Section 2.3 for more details). This spectroscopic analysis confirmed the white precipitate was the expected dipyrroloquinoline by-product 10 (d.r. 2:1), a simplified analogue of incargranine B. 32 Interestingly, the diastereoselectivity of this biocatalytic reaction was similar to that of our biomimetic chemical synthesis of incargranine B aglycone, 32 suggesting the TA does not provide any additional selectivity during the Povarov reaction. The fortuitous insolubility of 10 provides an additional driving force for the TA reaction and can be used as a simple visual screening method for identifying novel TA activity (see Supplementary Fig. S1, S17).
The window of acceptable reaction conditions can limit the application of amine donors and hence we turned our attention towards screening the applicability of NPP 9 to prepare methylbenzylamine 14 over a range of pH and temperatures (Fig. 4a). Our previous screen identified ATA-234 and ATA-251 as providing the highest yields and these commercial TAs were thus selected for further study alongside HEWT and CVTA. Pleasingly, NPP 9 performed exceptionally well across a broad range of reaction conditions when used in near equimolar quantities, relative to acetophenone 12 (Fig. 4b,c). This  To further explore the effectiveness of NPP 9 as an amine donor in comparison to i-PrNH2, we investigated the synthesis of chiral amines from more challenging substrates such as p-methoxy acetophenone 15 (Fig. 5a) and 1-indanone 17 (Fig. 5b). For this screen we selected three commercial and one wild-type biocatalysts (ATA-234, ATA-238, ATA-251 and HEWT). We were pleased to observe NPP 9 outperformed i-PrNH2 in each case under identical reaction conditions (Fig. 5c,d). This reveals the potential utility of NPP in applications where i-PrNH2 does not perform well. We next selected a proof-of-concept target to highlight the practical applicability of NPP 9 at preparative scale. Merck's commercial synthesis of the antidiabetic drug, sitagliptin 20 using a latestage TA-mediated reductive amination is perhaps the most compelling demonstration of the power of modern biocatalytic methods. This drug is manufactured from ketone 19 using a highly engineered Codexis ® TA with a high molar excess of i-PrNH2 as amine donor. 25 We were intrigued as to whether this process could instead be conducted using near equimolar NPP 9 as amine donor. This was achieved by initially screening the (R)-TAs in the Codexis ® kit with ketone 19 as amine acceptor. Pleasingly five of the eleven TAs showed activity (see Supplementary Section 2.17 for more details) with ATA-025 displaying the best conversion. An optimised reaction was carried out with ATA-025 on a 100 mg scale to give (R)-sitagliptin 20 in 54% isolated yield (>99% ee), with the insoluble by-product 10 easily removed by simple filtration through Celite ® (Fig. 6 and Supplementary Section 2.22 for more details).
This showcases the significant practical advantages of using NPP 9 as a versatile amine donor in TAmediated syntheses of chiral amines.

Conclusion
The continued demand for a universal amine donor for biocatalytic transaminations prompted us to explore how nature achieves the biosynthesis of complex natural products via TA-mediated pathways.
Specifically, we took inspiration from the biosynthesis of the dipyrroloquinoline alkaloids 7 and 8 to design and develop NPP 9 as a new, broadly applicable amine donor. NPP 9 is easily prepared on multigram scale and displays excellent activity across an unprecedented range of pH and temperatures with both wild-type and commercially available, engineered TAs. NPP 9 works at near equimolar quantities even with challenging amine acceptors and industrially relevant substrates. The insolubility of the NPP by-product 10 allows for easy identification of TA activity and facilitates simple removal by filtration. The broad range of reaction conditions under which NPP 9 operates will facilitate its widespread use in the biocatalytic preparation of chiral amines. To that end, researchers interested in using NPP 9 can contact us to request a sample.

General Experimental Procedures
Commercially available chemicals were used as received. Presitagliptin ketone 19 was purchased from key organics. All buffers were prepared from standard salts and pH adjusted with NaOH or HCl before use.
Stock solutions of substrates were prepared in reaction buffer unless stated otherwise. Wild type TAs (HEWT & CVTA) were defrosted on ice and incubated with PLP (0.1 mM) for 2 h prior to assay. All commercial TAs were dissolved in reaction buffer.

NMR Spectroscopy
1 H NMR spectra were recorded at 400 and 600 MHz using a Bruker Ascend 400 or Bruker Ultrashield 600 spectrometer.
The spectra have been referenced with the appropriate residual solvent peaks (CDCl3 7.26 ppm, CH3OD 3.31 ppm) and the coupling constants are reported to the nearest 0.1 Hz. 13 C NMR spectra were taken at 126 Hz or 151 Hz using the spectrometers mentioned above and referenced from solvent peaks (CDCl3 77.16 ppm, CH3OD 49.00 ppm). 2D NMR spectra such as COSY, HSQC and HMBC were used to assist with proton signal assignments.

Mass Spectrometry (MS)
Mass spectrometry analysis of small molecules was completed using a Bruker microTOF instrument using electrospray ionisation (ESI+).

Infrared Spectroscopy (IR)
A Shimadzu IR Affinity-1 Fourier transform IR spectrometer was used to record IR spectra of neat samples using a Pike MIRacle ATR accessory.

Protein Purification
All protein purification steps unless otherwise stated were completed on AKTA purifier Cytiva Lifesciences using a Frac-920 fraction collector.

Centrifugation
Thermoscientific Heraeus Multifuge X3R was used during protein growth steps with 8 × 50 mL rotor and 6 × 250 mL rotor at 4 °C. Benchtop centrifuge is VWR Microstar 17 and is used at rt.

Plate Reader
For colourimetric assays a Biotek Synergy HT plate reader was used.

Experimental procedure for the small-scale synthesis of NPP 9
Procedure adapted from Orelli and co-workers. 1 Step The relevant fractions were combined and concentrated under reduced pressure to give NPP 9 as an amber oil (1.27 g, 7.75 mmol, 80% yield). All spectroscopic data matched that reported in the literature. [1] Rf 0.38 (10:1:0.

HEWT Growth & Expression
The clone that expresses Halomonas elongata TA (HEWT) was kindly gifted by the Paradisi research group in a pHESPUC plasmid and has UNIPROT code E1V9I3. Growth, expression and purification procedures were adapted from their initial published characterization of HEWT which also details DNA sequence and molecular biology protocols. 2 The plasmid (2 µL) was transformed into BL21 (DE3) competent cells (10 µL

Activity Assay
The activity of HEWT was confirmed using the assay described by Schätzle et

HEWT binding UV/Vis scans
The UV/visible spectrum of HEWT was recorded after purification by desalting the enzyme using a disposable PD 10 desalting column. The assay was carried out in a CARY UV/Vis spectrometer at 37 °C. An initial spectrum was taken of HEWT (35 µM) in HEPES buffer (100 mM, pH 7.5). A second spectrum was recorded upon addition of NPP 9 (6 µL, 0.25 M stock) which was incubated for a further 15 min, after which a third spectrum was recorded. A final spectrum was recorded upon addition of sodium pyruvate (6 µL, 0.25 M).

CVTA Growth & Expression
The clone expressing Chromobacterium violaceum TA (CVTA) was kindly gifted by the O'Reilly group in a pet28a plasmid with UNIPROT code Q7NW64. The PDB code for the holo structure is 4AH3. 4 The growth, expression and purification procedures were adapted from CVTA initial characterisation publication which also contains bioinformatics details. 5

Activity Assay
The activity if CVTA was confirmed by the assay described by Schätzle

CVTA NPP binding assay
The UV/visible spectrum of CVTA was recorded after purification by desalting the enzyme using a disposable PD 10 desalting column. The assay was carried out in a CARY UV/Vis spectrometer at rt. An initial spectrum was taken of CVTA (36 µM) in KPhos buffer (100 mM, pH 8). A second spectrum was recorded upon addition of NPP 9 (60 µL, 25 mM stock) which was incubated for a further 15 min, after which a third spectrum was recorded. A final spectrum was recorded upon addition of sodium pyruvate (60 µL, 25 mM).

General Procedure for L-AAO coupled assay of amine donor with HEWT
We  The assay was also carried out with methylbenzylamine 14 (0-60 mM) as amine donor as a comparison of acceptance of NPP 9 by HEWT.
This demonstrates that only the desired enantiomer was observed in each biocatalytic reaction.