Synthesis of Pharmaceutically Relevant Arylamines Enabled by a Nitroreductase from Bacillus tequilensis

Arylamines are essential building blocks for the manufacture of valuable pharmaceuticals, pigments and dyes. However, their current industrial production involves the use of chemocatalytic procedures with a significant environmental impact. As a result, flavin‐dependent nitroreductases (NRs) have received increasing attention as sustainable catalysts for more ecofriendly synthesis of arylamines. In this study, we assessed a novel NR from Bacillus tequilensis, named BtNR, for the synthesis of pharmaceutically relevant arylamines, including valuable synthons used in the manufacture of blockbuster drugs such as vismodegib, sonidegib, linezolid and sildenafil. After optimizing the enzymatic reaction conditions, high conversion of nitroaromatics to arylamines (up to 97 %) and good product yields (up to 56 %) were achieved. Our results indicate that BtNR has a broad substrate scope, including bulky nitro benzenes, nitro pyrazoles and nitro pyridines. Hence, BtNR is an interesting biocatalyst for the synthesis of pharmaceutically relevant amine‐functionalized aromatics, providing an attractive alternative to traditional chemical synthesis methodologies.


Introduction
[3][4][5] Aromatic amines are typically obtained by catalytic hydrogenation of nitro precursors, which requires the use of hydrogen gas and metal catalysts. [6,7]Current industrial methodologies, while wellestablished, present several limitations.For example, they rely on continuous gas-phase processes that typically operate at elevated pressures and temperatures, leading to high costs and potential hazard. [8,9][12][13][14] Consequently, there is a high demand for new synthetic strategies for aniline derivatives, which find extensive use in the pharmaceutical, agrochemical and textile industries.
Enzymes are gaining prominence in both industry and academia due to their ability to operate in a selective manner under mild reaction conditions.[17][18] NRs use a ping-pong bi-bi kinetic mechanism, where the nicotinamide cofactor and the substrate alternate in the enzyme active site.Each reduction step starts with the nicotinamide cofactor transferring a hydride (2e À /H + ) to the prosthetic FMN.Subsequently, the reduced FMN transfers electrons to the nitro, nitroso or hydroxylamine compound. [19]The redox potential of the substrate is influenced by the electronic effects of its substituents. [20,21]For example, electron withdrawing groups (especially in ortho and para positions) can facilitate nitroreduction, while electron donating groups have the opposite effect. [22]Additionally, the presence of large π systems in a substrate also determines its likelihood of being reduced by the enzyme. [20,22][25] The extent of these interactions depends on the size of the substrate's π system.Therefore, the ability to perform the full nitro reduction to the respective amine depends not only on the prosthetic flavin's redox potential but also on many other factors.The majority of well-characterized NRs mainly produce hydroxylamines, with a few exceptions forming amines in low yields, such as NfnB from Mycobacterium smegmatis, [16] SNR from Streptomyces mirabilis, [26] NRSal from Salmonella typhimurium, [17] Gox0834 from Gluconobacter oxydans 621H, and two NRs from Klebsiella sp.C1. [27] However, it has recently been demonstrated that several strategies can be adopted to promote the reduction of the hydroxylamine intermediates to the final anilines.For example, NRs can be used in conjunction with transition metal co-catalysts, such as vanadium, copper or iron. [28]Furthermore, the efficiency of NRs has been boosted by immobilization and application in continuous flow reactors, where reaction and workup can be continuously undertaken in a single operation. [29]More recently, our group described a flavin-dependent NR from Bacillus amyloliquefaciens, BaNTR1, that exhibits nitroreductase activity towards a plethora of substituted nitroarenes and enables their selective conversion into the corresponding amino-, azoxy-, and azo-aromatics under photobiocatalytic conditions using chlorophyll as the photocatalyst. [30]ere, we describe the identification and characterization of a NR from Bacillus tequilensis, named BtNR, that can process bulky nitroarenes and promotes arylamine formation with high conversions and in good yields.This makes BtNR an interesting biocatalyst for the sustainable synthesis of pharmaceutically relevant arylamines.

Results and Discussion
Based on literature data and a genome mining approach, four nitroreductase-like flavoenzymes were initially selected for this study (see Supporting Information for details).Three of these flavoenzymes are well-characterized NRs that have been reported to perform amine synthesis starting from nitro aromatic compounds, namely NfnB from Mycobacterium smegmatis, [16,22] SNR from Streptomyces mirabilis [26,31] and NfnB from Sphingopyxis sp. [32]Additionally, a putative NR from Bacillus tequilensis, named BtNR, was also selected (Figure S1).The synthetic genes were adapted with sequences encoding an Nterminal His 6 -Tag and cloned into the pET-21α vector for expression in E. coli BL21 (DE3).After expression, the cell-free lysates were supplemented with free FMN (54 μM), and the enzymes were purified with yields ranging between 40 and 180 mg/L (Figure S2).
The nitroreductase activity of the purified enzymes was evaluated using four model substrates (1 b, 1 c, 1 e and 1 g) and the transformations into the corresponding amine products were analyzed by GC-MS (Table 1).The nitroreductase BtNR enabled satisfactory conversions to the desired amine products outperforming the other three candidates and was therefore selected for further substrate scope analysis and reaction optimization.
][35][36] The ability of BtNR to convert the selected nitro compounds to the corresponding arylamines was analyzed by GC-MS using naphthalene as internal standard (Figures S7-S19).Pleasingly, the enzyme is active towards bulky nitro benzenes, nitro pyridines and nitro pyrazoles, with moderate to excellent conversions into the amine products (Table 2).Notably, no regioselectivity was observed with the dinitro substrates 1 g and 1 j, as the diamine compounds 2 g and 2 j were primarily formed with minor amounts of partially reduced side-products (Figures S13 and S16).
Informed by previous work suggesting that the presence of oxygen may be harmful for the NR-catalyzed conversion of nitro compounds into amines, [30] we decided to investigate the performance of the BtNR-catalyzed nitroreduction under both aerobic and anaerobic conditions (Figure 2B).Interestingly, the conversion to amine increased significantly under anaerobic conditions for 1 g (from 56 % to 92 %), 1 k (from 60 % to 98 %), and 1 l (from 2 % to 18 %).However, no significant differences were found for substrates 1 h (from 64 % to 57 %) and 1 j (from 84 % to 81 %).These results demonstrate that the presence of molecular oxygen can indeed hinder the conversion into the Table 1.Nitroreductase screening for the reduction of substrates 1 b, 1 c, 1 e and 1 g into the corresponding amines.

Entry
a] Substrates and corresponding amine products are shown below.
[a] Reaction conditions: 5 mM nitroarene (5 μmol), 10 μM BtNR, 50 mM glucose, 0.5 mM NAD + , 2 μM BmGDH in 0.1 M sodium phosphate buffer pH = 7.0 (1 mL total volume), containing 10 % DMSO, at rt for 24 h.Analytical yield was determined by GC-MS analysis using naphthalene (4 mM) as internal standard and given in percentage as average of two replicates.For calibration curves and quantification of the amine products, see Supporting Information for details.
[b] For amine products 2 g and 2 i, percentage (%) values are based on uncorrected area (%) measured by GC-MS.
amine.The observation that the BtNR-catalyzed aerobic biotransformations resulted in lower nitroreduction for some substrates, but not for others, suggests that the negative influence of oxygen is likely substrate and intermediate dependent.
We also tested BtNR stability in various co-solvents and at different temperatures.We selected nitro compound 1 a as a substrate for this analysis because it has poor solubility in aqueous buffers, and we aimed at optimizing its conversion into amine 2 a (Figure 2A).Notably, the conversion to the final amine underwent almost a two-fold increase with 20 % DMSO, as well as with methanol at a percentage of 10 and 20 % (Figure 2A).The optimum co-solvent was found to be DMSO at a concentration of 20 %, however the enzyme performs well also in aqueous buffer containing methanol and isopropanol.Interestingly, the yield of the amine product increased to 21 % at 37 °C, probably due to the increased solubilization of the substrate, suggesting that low solubility in water may be a limiting factor for the conversion of compound 1 a (Figure 2A).
Having optimized the reaction conditions for analytical-scale conversions, we sought to further demonstrate the synthetic usefulness of BtNR by conducting semi-preparative-scale reactions.Biotransformations were performed using the nitro substrates 1 a, 1 b, 1 c, 1 k, 1 l and 1 m, and applying anaerobic conditions to minimize the formation of side products.The substrate concentration for the semi-preparative scale reactions was increased to 8 mM, using 20 μM BtNR, 100 mM glucose, 1 mM NAD + and 4 μM GDH in 0.1 M sodium phosphate buffer at pH 7.0, containing 10 % DMSO (10 mL total volume).Reactions were performed at rt (for substrate 1 a at 37 °C) for 48 h.The enzymatically-prepared products were purified using flash chromatography and identified as the corresponding amines by 1 H-NMR and HRMS analysis (Figures S20-S31).Final isolated yields for the amine products were determined by 1 H NMR to be 6 % for product 2 a, 45 % for product 2 b, 56 % for product 2 c, 44 % for product 2 k, 18 % for product 2 l, and 13 % for product 2 m.The yields may well be further improved by optimization of the extraction and purification procedures.

Conclusions
In summary, we have discovered a nitroreductase from Bacillus tequilensis, named BtNR, that can provide new opportunities for the practical synthesis of arylamines starting from diverse nitroarenes.This flavoenzyme displays a broad substrate scope, including bulky nitro benzenes, nitro pyridines, and nitro pyrazoles, achieving high conversions to the corresponding amines.While the enzymatic nitroreductions can be performed under aerobic conditions, we have applied an N 2 atmosphere to avoid the potential oxidation of reaction intermediates and enhance the yield of the desired amine product.With isolated yields ranging from 6 % to 56 %, the product yields might be further improved by the optimization of the extraction and purification procedures.Our results highlight the potential of nitroreductases for the environmentally friendly production of synthetically useful arylamines, and pave the way (by expanding the available toolbox of NRs) for the application of nitroreductases in chemoenzymatic cascade synthesis of pharmaceutically active products. [37,38]e previously demonstrated that an homologous enzyme, BaNTR1, promotes the selective photoenzymatic conversion of nitroarenes into amino-, azoxy-and azo-aromatics. [30]These compounds are versatile synthons that are extensively produced in the textile, food, cosmetic and pharmaceutical industries.Current work in our group is therefore focused on  exploring the usefulness of BtNR, and closely related nitroreductases, for developing photoenzymatic systems capable of synthesizing bulky azoxy-and azo-aromatics, as well as valuable heterocyclic compounds.Furthermore, we are conducting a comprehensive biochemical characterization of BtNR, including structural and mechanistic analysis, docking and molecular dynamics studies, to better understand its catalytic mechanism and set the stage for engineering of its catalytic properties.

Figure 2 .
Figure2.Optimization of the enzymatic reaction conditions.(A) Effect of temperature and different co-solvents on the conversions of nitro substrate 1 a to amine 2 a.All reactions were performed twice and the conversions are reported as single measurements (black dots), with the grey bar representing the average of the replicates.The conversions were determined by GC-MS analysis using naphthalene (4 mM) as internal standard and based on a calibration curve generated with authentic 2 a (Suppl.FigureS4).(B) Comparison of BtNR-catalyzed nitroreduction under aerobic and anaerobic conditions for substrates 1 g, 1 h, 1 j, 1 k and 1 l.All reactions were performed in two replicates and the conversions to the corresponding amine are reported as single GC-MS measurements (black dots), being the grey bar the average of the replicates.Percentage (%) values are based on uncorrected area (%) measured by GC-MS.
Figure2.Optimization of the enzymatic reaction conditions.(A) Effect of temperature and different co-solvents on the conversions of nitro substrate 1 a to amine 2 a.All reactions were performed twice and the conversions are reported as single measurements (black dots), with the grey bar representing the average of the replicates.The conversions were determined by GC-MS analysis using naphthalene (4 mM) as internal standard and based on a calibration curve generated with authentic 2 a (Suppl.FigureS4).(B) Comparison of BtNR-catalyzed nitroreduction under aerobic and anaerobic conditions for substrates 1 g, 1 h, 1 j, 1 k and 1 l.All reactions were performed in two replicates and the conversions to the corresponding amine are reported as single GC-MS measurements (black dots), being the grey bar the average of the replicates.Percentage (%) values are based on uncorrected area (%) measured by GC-MS.