Organic–inorganic nanocrystal reductase to promote green asymmetric synthesis

An acetophenone reductase from Geotrichum candidum (GcAPRD) was immobilized by the organic–inorganic nanocrystal method. The GcAPRD nanocrystal presented improved stability and recyclability compared with those of the free GcAPRD. Moreover, the GcAPRD nanocrystal reduced broad kinds of ketones with excellent enantioselectivities to produce beneficial chiral alcohols such as (S)-1-(3′,4′-dichlorophenyl)ethanol with >99% yield and >99% ee. The robust and versatile properties of the GcAPRD nanocrystal demonstrated an approach to promote green asymmetric synthesis and sustainable chemistry.


Introduction
Biocatalysts have presented superior properties for green and sustainable synthesis with their high selectivity, turnover rate, and efficiency. 1,2 For example, an alcohol dehydrogenase (ADH) from Rhodococcus ruber DSM 44541 expressed in Arxula adeninivorans is able to enantioselectively reduce p-chloroacetophenone with a remarkable turnover rate (K cat ) of 432 s À1 . 3,4 Thus, the biocatalysts have played an essential role in an extensive area of industries such as ne and bulk chemical, pharmaceutical, food and beverage, cosmetic, textile, and pulp and paper. 5 Meanwhile, the needs for chiral compounds have risen continuously, especially in the pharmaceutical elds, due to the different biological activities and toxicities that may be exhibited by each enantiomer of the chiral compounds. 6,7 Asymmetric reduction of ketones by chemical and biological catalysts has been considered as a promising method to produce chiral alcohols. [8][9][10] Enzymes catalyzing asymmetric reduction reactions have been widely studied due to their high specicity property, and the well-known enzyme used for this purpose is ADH. 10 An acetophenone reductase from Geotrichum candidum NBRC 4597 (GcAPRD) is a novel ADH that reduces ketones to their corresponding (S)-alcohols. 11 The GcAPRD has versatile and robust properties such as excellent enantioselectivity, broad substrate specicity, and non-aqueous solvents tolerance that make it potential for industrial applications. [12][13][14][15][16][17][18] However, the free form of GcAPRD has limitations to recycle and operate in a ow process.
Enzyme immobilization is a method that improves stability and recyclability of the enzyme allowing the enzyme to operate under wide conditions. 2 In most cases, the enzyme is conned with the immobilization supports that limits their mobility. 19 Various techniques have been addressed to immobilize the enzyme, including the organic-inorganic nanocrystal method, rst reported by Ge et al. in 2012. 20-22 This method has been proven to be effective in immobilizing various biomolecules such as proteins, peptides, and antibodies. 21,22 Moreover, limitations faced by other immobilization methods, such as insuf-cient binding of protein on support and severe loss of activities, have been overcome by the organic-inorganic nanocrystal method. 19,23 Thereby, several studies have applied this method to immobilize broader kinds of biomolecules. 21 Enzymes have been one of the most widespread biomolecules immobilized by this method for examples a-lactalbumin, 22 laccase, 22 carbonic anhydrase, 22 lipases, 22,24-26 a-amylases, 27 papain, 28,29 and a-acetolactate decarboxylases. 30 However, only a few studies have focused on the immobilization of alcohol dehydrogenase (ADH) by this method, 31 particularly for asymmetric synthesis, 32 and there is no example yet of versatile ADH nanocrystals with broad substrate specicities.
Thus, this research aimed to immobilize the GcAPRD by using the organic-inorganic nanocrystal method. The preparation conditions of GcAPRD nanocrystal were optimized, and the properties of the GcAPRD nanocrystal were characterized. The GcAPRD nanocrystal showed improved stability and recyclability while retaining excellent enantioselectivity. To the best of our knowledge, this is the rst study to report an ADH nanocrystal with broad substrate specicity. piperazine ethanesulfonic acid (HEPES) and dithiothreitol (DTT), which were purchased from Sigma Aldrich (USA) and Wako (Japan), respectively. Protein concentration measurement reagent was supplied from Bio-Rad (USA). Commercial grade solvents, ketones, and alcohol standards (1b-3b) were purchased from Nacalai Tesque (Japan). Previously prepared alcohol (4b-7b) and chiral ester (2b-4b and 7b) standards were used in this study. [33][34][35]

Apparatus
Ultraviolet-visible spectroscopic analysis was conducted in a UV-1900-UV-Visible spectrophotometer from Shimadzu (Japan). Chiral gas chromatography (GC) was performed on a GC-14B equipped with ame ionization detector and a CP-Chirasil-Dex-CB column (Varian 0.32 mm Â 0.25 mm Â 50 m) using He carrier gas (5 mL min À1 , head pressure: 274 kPa, injector: 180 C, detector: 180 C) (Shimadzu, Japan). The morphology and energy dispersive X-ray spectroscopy (EDX) analysis were studied by Bench-top Scanning Electron Microscope proX supplied by Phenom-World (Netherlands). Thermo Gravimetric Analysis (TGA) was performed on TG-DTA, DTG-60 (Shimadzu, Japan). The optical rotation value was measured by JASCO P-2200 Polarimeter with a 10 cm path-length cell. The 1 H-NMR analysis was performed on a Bruker Biospin AVANCE III 400 spectrometer at 400 MHz in CDCl 3 .

GcAPRD nanocrystal preparation
The phosphate-buffered saline (PBS) containing GcAPRD was mixed with a metal ion by pipetting for 5 times gently, incubated at 4 C for 24 h, and centrifuged at 4 C, 12 000 rpm for 5 min. The detailed experimental conditions are shown in the ESI (Table S1 †). The supernatant was removed to measure the residual protein amount by Bradford method, 36 and to calculate the immobilization yield (1). The precipitant was washed by distilled water and resuspended in PBS (GcAPRD nanocrystal).

Immobilization yield ð%Þ
[Protein] I ¼ initial protein concentration (mg mL À1 ), [Protein] R ¼ concentration of residual protein in the supernatant aer the GcAPRD nanocrystal formation and centrifugation (mg mL À1 ).

Relative reaction yield and activity measurement of enzyme
The relative reaction yield measurement to investigate the optimum immobilization conditions was conducted by mixing acetophenone 1a Activity measurement to investigate the temperature and pH prole, and the stability of GcAPRD nanocrystal were assayed by using UV-spectrophotometer. All assays were performed at 37 C in 1.0 mL HEPES-NaOH buffer (0.10 M, pH 7.2), consisting of 1a (5.0 mM) except for the pH prole experiment (5.6 mM), NADH (0.28 mM), and GcAPRD nanocrystal (0.80-4.00 mg of protein per mL). Initial velocity was determined by detecting the decrease of NADH at 340 nm for 132 s.
The recyclability was assessed by batch process. The reaction mixture consisted of 1a (3.0 mM), 2-propanol (15% v/v), GcAPRD nanocrystal (630 mg of protein), NAD + (0.20 mM), and HEPES-NaOH buffer (0.10 M, pH 7.2) up to 1.5 mL. The reaction in each cycle was conducted at 30 C, 200 rpm for up to 180 min. Aerward, the reaction mixture was centrifuged at 4 C, 16 000 rpm for 5 min. The supernatant was removed, extracted with diethyl ether, and analyzed by GC. The precipitant was washed by the reaction mixture with the absence of NAD + , and proceeded the next reaction cycle.

Micromole scale asymmetric reduction of ketones by GcAPRD nanocrystal
The experiment was conducted based on the previously reported procedure. 13 Micromole scale reductions were performed in a 3.0 mL reaction mixture, which consists of 1a-7a (5.0 mM), 2-propanol (15% v/v), NAD + (5.0 mM), HEPES-NaOH buffer (0.10 M, pH 7.2), and GcAPRD nanocrystal with the amount needed to achieve 1.0 mmol min À1 mL À1 reduction of each substrate to its corresponding alcohol for 1a-6a, and 350 mg of protein/mL for 7a. The mixture was incubated at 30 C, 200 rpm for 3 h. A portion of the mixture was extracted with diethyl ether to determine the reaction yield by GC analysis. The enantioselectivity excess (ee) values and the absolute congurations of 1b, 5b, and 6b were determined by comparing their GC retention times with the corresponding authentic samples prepared before. 35 Meanwhile, for 2b-4b and 7b, the reaction mixtures were extracted by dichloromethane, dried over MgSO 4 , and used for propionylation by propionyl chloride reaction. ee values and absolute congurations were determined by comparing the GC retention times of propionates of 2b-4b and 7b with the corresponding authentic samples prepared before. 33,34 The chiral GC retention time of ketone, alcohol, and ester standards are given in Table S2. †

Preparative scale asymmetric reduction of a ketone by GcAPRD nanocrystal
The reaction was performed in a 100 mL scale consisting of 6a (197 mg, 10.0 mM), 2-propanol (3.0% v/v), NAD + (1.4 mM), HEPES-NaOH buffer (0.10 M, pH 7.2), and GcAPRD nanocrystal (2 mg of protein). The mixture was incubated at 30 C, 130 rpm for 18 h, and extracted by diethyl ether for 3 times, dried over MgSO 4 , ltered, and evaporated. The corresponding alcohol was puried by silica gel column chromatography (hexane : ethyl acetate, 3 : 1) to afford (S)-6b (165 mg, 0.87 mmol, 84%), and characterized by 1 H-NMR analysis. The 1 H-NMR spectra of alcohol (S)-6b was in agreement with that reported in the literature. 37 ee was determined by chiral GC analysis, and the absolute conguration was determined by comparing the optical rotation value of product with the literature. 38

2.7
GcAPRD nanocrystal characterization by scanning electron microscopy (SEM) and EDX analysis, and TGA The prepared GcAPRD nanocrystal was washed by distilled water several times to remove the residual phosphate and metal ion prior to drying at room temperature. The dried GcAPRD nanocrystal was used for SEM and EDX analysis, and TGA.

Optimization of GcAPRD nanocrystal synthesis
Several parameters for the GcAPRD nanocrystal preparation, such as kinds of metal ions, metal concentration, PBS concentration, PBS pH, and protein concentration, were varied. The immobilization yields and relative reaction yields are shown in Fig. 1.
Different amino acid residues in each biomolecule present different affinities with various metal ions, and the appropriate metal ion concentration is necessary for the nucleation. 21 In this research, 6 metal ions were used to prepare the GcAPRD nanocrystal (Fig. 1a). It was found that the immobilization yield was high when using Co 2+ or Ni 2+ as a metal ion. Nonetheless, only the GcAPRD nanocrystal synthesized by Co 2+ retained sufficient activity. It was assumed that the His-tag located in the C-terminus of the GcAPRD acted as a binder with Co 2+ or Ni 2+ ion, which was comparable to the discovery of López-Gallego and Yate's. 31 They reported that the presence of His-tagged protein is mandatory as the driving force for the crystal formation. The His-tag binds Co 2+ ions via the coordination bond and forms a nucleation site, which initiates the crystal formation. 31 Although Ni 2+ could form the crystal with Histagged protein, high concentration of nickel could cause the enzyme deactivation (Fig. S1 †). To investigate the importance of Co 2+ and His-tagged GcAPRD bonding, the formed GcAPRD nanocrystal was suspended in 500 mM imidazole in 5 mM PBS to break the Co 2+ and His-tag bond, thus eluting the protein.
Aer the elution, the GcAPRD nanocrystal was apparently detached, suggested that Co 2+ and His-tag bonding played an important role in the GcAPRD nanocrystal formation (Fig. S2 †). Additionally, the metal concentration was optimized to 10 mM (Fig. 1b), comparable with Kim et al.'s study (10 mM of cobalt for the formation of BSA-inorganic hybrid nanoower). 39 PBS concentration is another important parameter for the nanocrystal formation. 28 Thereby, we investigated the effect of PBS concentration on the GcAPRD nanocrystal formation (Fig. 1c). The prepared GcAPRD nanocrystal performed the highest relative reaction yield at 3.75 mM PBS. However, the optimum PBS concentration is different depending on the biomolecule and other synthesis parameters. 40,41 Synthesis pH is an essential parameter inuencing the nanocrystal formation. 29,41,42 Thus, this study investigated the effect of PBS pH on the GcAPRD nanocrystal formation (Fig. 1d). The immobilization yield of GcAPRD was low in acidic conditions (pH 5.5), and higher in higher pH (pH 7.4, 9.2, and 11.0). It was proposed that in acidic conditions, the biomolecule is highly protonated, causing strong repulsions with the positive charges of the transition metal (i.e., Cu 2+ and Co 2+ ). 43 Joshua et al., reported that protonating the imidazole nitrogen atom of the histidine residue (pKa 6.0) by lowering the pH to 6.0 can disrupt the coordination bond between histidine and Co 2+ . 44 Altinkaynak et al. also reported an easier synthesis of lactoperoxidase hybrid nanoower at pH 9.0 and 10.0. 19 The GcAPRD nanocrystal showed the best relative reaction yield in pH 9.2, and pH 11.0.
The presence of biomolecule is mandatory for the hybrid nanocrystal formation. 21,22 In this study, the protein concentration in the GcAPRD nanocrystal formation was varied to 0.02, 0.10, 0.05, 1.00, and 2.00 mg mL À1 (Fig. 1e). The experiments demonstrated that at the protein concentration of 0.05 mg mL À1 and below, the enzyme was entirely immobilized (immobilization yield >99%). On the other hand, the protein concentration of more than 1.00 mg mL À1 resulted in a decrease in immobilization yield, but protein immobilized per metal amount was high. Nevertheless, the relative reaction yield result showed that the reaction proceeded better when the synthesized protein concentrations were more than 0.50 mg mL À1 , and the greatest relative reaction yield was found on the 2.00 mg mL À1 of protein. Ge et al. reported that different concentration of protein has a different nanocrystal formation mechanism. The lower concentration of protein causes fewer nucleation sites, leading to a larger nanocrystal with a more complex structure. 22 At last, it was conrmed that the active GcAPRD was mandatory to the GcAPRD nanocrystal activity (ESI, section 4 †). To understand this phenomenon clearly, SEM analysis on GcAPRD nanocrystal synthesized with various protein concentrations were performed as described in section 3.2.

Morphology analysis by SEM
SEM analysis on GcAPRD nanocrystal synthesized with various protein concentrations was performed to understand its effect in the immobilization yield and relative reaction yield. As shown in Fig. 2, it was found that without enzyme, the cobaltphosphate crystal was formed ( Fig. 2a and b). In the presence of 0.02 mg mL À1 (Fig. 2c and d) and 0.10 mg mL À1 (Fig. 2e and f) of protein, the surfaces of the crystals were similar to the control, assuming that the small amount of protein was buried underneath the cobalt-phosphate crystals. Therefore, the mass transfer between the hindered enzyme and the substrate was limited, and low relative reaction yields were observed for the nanocrystals synthesized with 0.02 and 0.10 mg mL À1 of protein. The surfaces of GcAPRD nanocrystals synthesized with 0.50 mg mL À1 (Fig. 2g-h) and 2.00 mg mL À1 (Fig. 2i-j) protein were different from the control and displaying a parallel hexahedron structure assembled to a ower-like structure mixed with sponge-like structure. The proposed formation mechanism of the GcAPRD nanocrystal synthesized with 0.50 and 2.00 mg mL À1 of protein is presented in Fig. 3. There may be 3 steps, comprised of nucleation, aggregation, and anisotropic growth. A similar structure was reported previously by Wang et al. in the formation of nanocrystals with the presence of 2.00 mg mL À1 protein. 27,31 The high relative reaction yield suggested that the GcAPRD was gathered on the surface of the nanostructure, and the enzymes were not buried. In addition, the GcAPRD nanocrystal was characterized by EDX analysis and TGA to examine the chemical composition and structure of the GcAPRD nanocrystal as in Fig. S3-S5 (ESI, † section 5). Fig. 1 Optimization of GcAPRD nanocrystal preparation (a) metal ions, (b) metal ion (Co 2+ ) concentration, (c) PBS concentration, (d) PBS pH, and (e) protein concentration. Left and right column depict the immobilization yields and relative reaction yields, respectively. The reaction yield of free GcAPRD was set to be 100%. *Very small precipitation was found. The preparation and reaction conditions are described in the section 2.4.

Temperature and pH prole, and stability of GcAPRD nanocrystal
Immobilization has been reported to improve the temperature and pH prole, as well as the stability of the enzyme. 2,31,40,41 The temperature prole of GcAPRD nanocrystal and free GcAPRD was examined (Fig. 4a). Both free GcAPRD and GcAPRD nanocrystal demonstrated poor activity at low temperatures (20-30 C). The activities of both enzymes were increased in higher temperatures. The free GcAPRD performed the highest relative activity at 50 C (170%), and the activity gradually decreased at the temperature above 60 C. To our delight, the GcAPRD nanocrystal presented the improved temperature prole as the highest relative activity was observed at 60 C (189%). At the same time, the effect of pH on the GcAPRD nanocrystal activity was investigated, and it showed less sensitivity than the free GcAPRD in acidic conditions (Fig. 4b). The relative activity of the free GcAPRD dramatically decreased from 99% to 18% when the pH was lowered from 5.0 to 4.5, while the GcAPRD nanocrystal's  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 30953-30960 | 30957 relative activity moderately decreased from 132% to 89%. The storage stability of the enzymes at 4 C was investigated. As shown in Fig. 4c, aer 7 days, the residual activity of the free GcAPRD was 68%, while the GcAPRD nanocrystal still retained 100% activity. Also, aer 18 days, the free GcAPRD's residual activity was 61%, while the GcAPRD nanocrystal had 76% of residual activity.

Recyclability of GcAPRD nanocrystal
We investigated the recyclability of the GcAPRD nanocrystal by performing the reduction of 1a in batch process. The GcAPRD  nanocrystal was able to be recycled with noticeable high reduction yield (>99%) and excellent enantioselectivity (>99%) up to 7 reaction cycles (Fig. 5). This result proved the superior recyclable properties of the GcAPRD nanocrystal, which is a promising approach to further applications.

Substrate specicity of GcAPRD nanocrystal
Excellent enantioselectivity and broad substrate specicity are one of the most crucial aspects of the enzyme utilization for industrial processes. 10 The GcAPRD nanocrystal was tested to reduce broad kinds of ketones (Table 1). 1a, as a model substrate for the GcAPRD reduction, was reduced by GcAPRD nanocrystal with >99% yield and >99% ee (S). The GcAPRD nanocrystal also reduced challenging aliphatic ketones 2a-4a with high yields and enantioselectivities (up to >99% ee (S)). Moreover, the GcAPRD nanocrystal was able to enantioselectively reduce "difficult to resolve" ketone 3a with only one carbon difference between side chains adjacent to the carbonyl carbon. The enantioselectivities reported for the free enzyme were retained perfectly. 9,12,13 The GcAPRD nanocrystal produced benecial chiral alcohols with excellent enantioselectivities (ee up to >99% (S)) such as (S)-5b (intermediate of EGFR and HER2 kinase inhibitors, a potential cancer treatment agent), 45 (S)-6b (intermediate of sertraline, an obsessive-compulsive disorder medication), 46 and (S)-7b (intermediate of aminotetralin, which can be used to stimulate dopamine, serotonin, and melatonin receptors). 47,48 We were also able to isolate (S)-6b with 84% isolated yield, in a preparative scale reaction catalyzed by the GcAPRD nanocrystal. To the best of our knowledge, this is the rst study to report an ADH nanocrystal with broad substrate specicity.

Conclusions
We successfully immobilized a novel ADH, GcAPRD, by the organic-inorganic nanocrystal method. The GcAPRD nanocrystal presented improved temperature and pH prole, and stability compared with those of the free GcAPRD. The GcAPRD nanocrystal overcame the recyclability limitation of the free GcAPRD. It could be recycled up to 7 cycles with remarkably high yield and excellent enantioselectivity. The GcAPRD nanocrystal reduced broad kinds of ketones to produce valuable chiral alcohols. The robust and versatile properties of the GcAPRD nanocrystal made it a promising and sustainable catalyst for industrial applications. Future studies will investigate the reaction of GcAPRD nanocrystal in non-aqueous solvent and perform cascade reactions to allow the production of broader kinds of benecial compounds.

Conflicts of interest
There are no conicts to declare.