Immobilization of Lipase B from Candida antarctica on Magnetic Nanoparticles Enhances Its Selectivity in Kinetic Resolutions of Chiral Amines with Several Acylating Agents

In lipase-catalyzed kinetic resolutions (KRs), the choice of immobilization support and acylating agents (AAs) is crucial. Lipase B from Candida antarctica immobilized onto magnetic nanoparticles (CaLB-MNPs) has been successfully used for diverse KRs of racemic compounds, but there is a lack of studies of the utilization of this potent biocatalyst in the KR of chiral amines, important pharmaceutical building blocks. Therefore, in this work, several racemic amines (heptane-2-amine, 1-methoxypropan-2-amine, 1-phenylethan-1-amine, and 4-phenylbutan-2-amine, (±)-1a–d, respectively) were studied in batch and continuous-flow mode utilizing different AAs, such as diisopropyl malonate 2A, isopropyl 2-cyanoacetate 2B, and isopropyl 2-ethoxyacetate 2C. The reactions performed with CaLB-MNPs were compared with Novozym 435 (N435) and the results in the literature. CaLB-MNPs were less active than N435, leading to lower conversion, but demonstrated a higher enantiomer selectivity, proving to be a good alternative to the commercial form. Compound 2C resulted in the best balance between conversion and enantiomer selectivity among the acylating agents. CaLB-MNPs proved to be efficient in the KR of chiral amines, having comparable or superior properties to other CaLB forms utilizing porous matrices for immobilization. An additional advantage of using CaLB-MNPs is that the purification and reuse processes are facilitated via magnetic retention/separation. In the continuous-flow mode, the usability and operational stability of CaLB-MNPs were reaffirmed, corroborating with previous studies, and the results overall improve our understanding of this potent biocatalyst and the convenient U-shape reactor used.


Introduction
Chiral primary, secondary, and tertiary amines are crucial building blocks for the pharmaceutical and agrochemical industries. The necessity for these constituents in their enantiomerically pure form in several drugs requires the development of cost-effective and long-lasting catalytic procedures for their asymmetric synthesis [1,2]. Biocatalytic processes, which use different forms of enzymes to perform efficient and selective synthetic transformations, are sustainable methods characterized by excellent activity and stereo-, regio-, and chemoselectivity [3]. Among the enzymes used as biocatalysts, lipases are highlighted due to a variety of desirable properties, such as easy Lewatit VP OC 1600 resin via adsorption and interfacial activation [31]. Some of the intrinsic problems simply stem from the adsorptive immobilization, enabling enzyme desorption under aqueous conditions, and material transfer limitations within the pores. Additional challenges that are unique to the support include mechanical brittleness, a moderate hydrophilicity that allows hydrophilic molecules (such as water or glycerin) to accumulate and leads to the most serious issue: disintegration in various organic mediums [31].
A possible solution to tackle this problem is using nonporous materials such as magnetic nanoparticles (MNPs), providing a large surface area and high surface-to-volume ratio for enzymes on the supports, thereby reducing diffusion limitation issues, which makes them appealing for industrial use [32]. Their superparamagnetic properties allow magnetic decantation, so MNPs can be easily removed from the reaction mixture [32,33]. Apart from the utilization in the KR or DKR of enantiomer compounds, CaLB-MNPs are also used in biodiesel synthesis [34,35], transesterification reactions [36], or even utilized in a magnetic crosslinked inclusion body form [37]. But, to the best of our knowledge, there are no reports of the utilization of MNPs in the KR or DKR of amines, except in one study with a magnetic hybrid catalyst containing lipase and palladium applied in the DKR of (±)-1-phenylethan-1-amine [38]. Therefore, it is a topic yet to be explored.
Continuous-flow production is gaining popularity in diverse industries, including in the pharmaceutical sector, with the main benefits of this cutting-edge technology being the capacity to execute chemistry experiments that are challenging to conduct in batch mode, as well as faster and safer reactions that may be more ecologically friendly [39]. Numerous researchers combined the advantages of MNP-immobilized biocatalysts and microreactor technology. In recent studies, a simple but effective and convenient MNPbased PTFE tube microreactor with six adjusting permanent magnets underneath the tube was used for the enantiomer selective acylation of drug-like racemic alcohols such as 4-(morpholin-4-yl)butan-2-ol [8], 4-(3,4-dihydroisoquinolin-2(1H)-yl)butan-2-ol, and 4-(3,4-dihydroquinolin-1(2H)-yl)butan-2-ol [9], catalyzed using a biocatalyst obtained via the covalent immobilization of lipase B from Candida antarctica on magnetic nanoparticles (CaLB-MNPs) in a continuous-flow mode. To broaden the spectrum of substrates evaluated in this system, a U-shape reactor was used, and reactions with selected chiral amines were performed. To enhance the accuracy and controllability of the magnetic U-shape reactor, a heat-insulating box was applied to perform reactions with temperature control (Figure 1). Although CaLB has been recently studied for immobilization onto nonporous MNPs, due to having advantages over porous supports, there is a lack of studies on their behavior in the KR of racemic amines. This work reports the kinetic resolution of several chiral amines catalyzed with CaLB-MNPs utilizing different isopropyl esters as acylating agents in batch and continuous-flow modes.  Although CaLB has been recently studied for immobilization onto nonporous MNPs, due to having advantages over porous supports, there is a lack of studies on their behavior in the KR of racemic amines. This work reports the kinetic resolution of several chiral amines catalyzed with CaLB-MNPs utilizing different isopropyl esters as acylating agents in batch and continuous-flow modes.

Enzyme and Materials
Recombinant lipase B from Candida antarctica (CaLB, produced by microbial fermentation in Pichia pastoris, exhibiting a single band around 33 kDa on SDS gel EF; provided as a lyophilized powder, Lot-NO: MA-b-0002, activity: 59,900 TBU/g) for immobilization experiments was obtained from c-LEcta (Leipzig, Germany). All other reagents and solvents were products of Sigma-Aldrich (Saint Louis, MO, USA), Merck (Darmstadt, Germany), or Alfa Aesar Europe (Karlsruhe, Germany) and are listed in Supplementary Materials.

Methods
The analytical methods and calculations utilized in this study were based on previously published methods [14], with minor modifications, as detailed below.

Analytical Methods and Calculations
TLC was carried out using Kieselgel 60 F254 (Merck) sheets. Spots were visualized under UV light (Vilber Lourmat VL-6.LC, 254 nm) or after treatment with a 5% ethanolic phosphomolybdic acid solution and heating the dried plates.
Gas chromatographic (GC) analyses were performed with an Agilent 4890 gas chromatograph equipped with a flame ionization detector (FID) using H 2 as carrier gas (injector: 250 • C, detector: 250 • C, head pressure: 12 psi, split ratio: 50:1) and a Hydrodex β-6TBDM column (25 m × 0.25 mm × 0.25 µm film with heptakis-(2,3 di-O-methyl-6-O-t-butyldimethyl-silyl)-β-cyclodextrin; Macherey and Nagel (Düren, Germany)) using the temperature programs indicated in Table S1 (Supplementary Materials). GC chromatograms are given in Figures S3-S15, S19 and S23 in Supplementary Materials. Conversion (c) and enantiomeric excess (ee) values were determined via GC utilizing previously published equations, with molar response factors 1.22, 1.06, and 1.185 for amides with the different acylating agents 2A-C, respectively. The enantiomeric ratio/selectivity (E) was calculated from the enantiomeric excess of the substrate (ee s ) and product (ee p ) fractions using a previously published equation for irreversible KRs. Due to sensitivity to experimental errors, E values calculated in the 100-200 range are given as >100, those in the 200-500 range are denoted as >200, and those above 500 are denoted as 200 [40,41].

2-Ethoxyacetic Acid
Sodium (10 g, 0.434 mol) was dissolved in water-free ethanol (450 mL) at 60 • C, and then triethylbenzylammonium chloride (TEBA-Cl, 1 g, 0.036 mol) and sodium 2chloroacetate (40 g, 0.344 mol) were added to the solution. After refluxing the resulting mixture for 24 h, ethanol was removed under vacuum. The residue was dissolved in distilled water (400 mL), and pH was set to 2 via the addition of 5N HCl. After extracting the solution with ethyl acetate (3 × 100 mL), the combined organic phases were washed with brine (100 mL) and dried over anhydrous Na 2 SO 4 . After the removal of the solvent via vacuum rotary evaporation, the residue was purified with vacuum distillation to yield 2-ethoxyacetic acid as a colorless oil.

Preparation of CaLB-MNP Biocatalysts
The preparation of the MNP carriers, including the preparation of the magnetite core and its surface modifications, and further immobilization of CaLB were performed using previously published methods [8].
The surface morphology of the samples was investigated with a JEOL JSM-5500LV scanning electron microscope (SEM). An electron beam energy of 15 kV was used for the investigations. The SEM images are given in Figure S1 in the Supplementary Materials. The samples (20 µL) from KRs were diluted with ethyl acetate (480 µL), treated with acetic anhydride (10 µL) at 25 • C and 350 rpm for 30 min in a shaker (for the derivatization of unreacted amines 1a-d into their N-acetamides 1*a-d) and quenched with 20 µL of distilled water. After drying over anhydrous Na 2 SO 4 , the samples were analyzed using GC on a chiral column. A variance of less than 2% was found in the preliminary activity test in batch mode, performed in triplicate. Therefore, the optimization and time course experiments were carried out as single series. In a 4 mL screw-cap vial, racemic amine (±)-1a,b,d (57.6 mg, 44.5 mg, or 74.6 mg, respectively; 0.5 mmol), CaLB-MNPs (20 mg), and isopropyl 2-cyanoacetate 2B (0.5 mmol, 1 equiv.) were added to MTBE (200 µL), and the resulting suspension was shaken at the rate of 200 rpm for 4 h at 40 • C in a Vibramax 100 shaker (Heidolph, Schwabach, Germany). The supernatant was decanted by anchoring the CaLB-MNPs with a neodymium magnet, and the resuspended CaLB-MNPs were washed using a mixture of methyl t-butyl ether (MTBE)-hexane (1:2 ratio, 2 × 2.5 mL). The washed CaLB-MNPs were dried in a fume hood overnight at room temperature and stored in a refrigerator until further use. From the combined reaction mixture and washing solutions, the volatiles were removed via vacuum rotary evaporation, and the resulting products were separated using preparative thin layer chromatography (silica gel, eluted with CH 2 Cl 2 :MeOH = 20:1) to yield the (R)-amide (R)-3(a,b,d)B and the residual (S)-amine (S)-1a,b,d; respectively.

Design and Assembly of the Thermostatted U-Shape MNP Reactor
The U-shape MNP reactor system (Figure 1), similar to its nonthermostatted predecessor [9], was designed using AutoCAD (2020 student version) and printed with a Rankfor100 3D printer (CEI Conrad Electronic International, Ltd., New Territories, Hong Kong). In this study, we used neodymium disc magnets, 4 mm × 2 mm, N35 (Euromagnet Ltd., Budapest, Hungary) as permanent magnets and a polytetrafluoroethylene (PTFE) tube with ID 1.50 mm as the reactor body and connection parts. The U-shape reactor was a holder allowing adjustable positioning of six permanent magnets under the PTFE tube in the closest possible vicinity to the PTFE tube reactor, thereby creating six sites where the magnets could anchor the CaLB-MNP biocatalysts inside the tube ( Figure 1). A SpinSplit continuous-flow syringe pump (SpinSplit Technical Research and Development LLC, Budapest, Hungary) equipped with two glass syringes of 0-5 mL volume was coupled to the tubular reactor part of the reactor module. For heat control, a heat exchanger (water heat exchanger with aluminum heat radiator/heat sink for computer, size: 60 × 60 × 12 mm, 25 mL capacity) with a fan (size: 120 × 120 × 25 mm, 12 V DC) was connected to a Lauda Ecoline RE104 Recirculating Chiller as the heat source. The device was placed into a Styrofoam heat insulating box (270 × 235 × 145 mm internal dimension; 25 mm wall thickness), which was properly punctured for providing the inlet and outlet of the PTFE tube reactor and to allow for the installation of the tubes of the external chiller and electric wires for fan, with a Plexiglas cover (250 × 215 mm, 4 mm thickness). Pictures showing the entire reactor setup are presented in Figure S2 in Supplementary Materials.

Kinetic Resolution of the Racemic Amines (±)-1b and (±)-1c with Isopropyl 2-Ethoxyacetate 2C Using CaLB-MNPs in Continuous-Flow U-Shape Reactor
The U-shape reactor was used for the continuous-flow kinetic resolution of the designed substrate amines (±)-1b and (±)-1c. For filling the PTFE reactor tube of 1.50 mm ID as the reactor body, 6 × 1 mg of CaLB-MNPs were suspended in 6 × 1 mL of a mixture of methyl t-butyl ether (MTBE)-hexane (1:2 ratio), and the suspended CaLB-MNPs were supplied at a 50 µL min −1 flow rate into the reactor's six chambers in a countercurrent direction of the later fluid flow, filling the last chamber first and the first chamber last (by placing the permanent magnet under the tube in the actual position, from the last-to-first position).

Biocatalyst Characterization
CaLB-MNPs produced with CaLB immobilized covalently onto the MNPs were proposed as good biocatalysts for KRs of racemic compounds. To prove that the immobilization was successful, the protein concentration of the supernatant before and after immobilization was determined using the Bradford assay. The experiments indicated almost full immobilization (Y = 98%) after 120 min at a CaLB:MNP ratio of 1:15 under the conditions utilized in our previous work [8].
SEM images could not reveal the real surface morphology of the CaLB-MNPs due to the aggregation of the distinct nanoparticles ( Figure S1 in Supplementary Materials). Consequently, the difference between the MNP carrier and CaLB-MNPs could be only revealed using indirect methods based on the activity measurement of the CaLB-MNPs and immobilization yield. Furthermore, the successful immobilization of CaLB on bisepoxideactivated macroporous resins [42] and aspartase on bisepoxide-activated MNPs [43] using the same chemistry support the conclusions drawn from indirect methods.

Kinetic Resolution of Chiral Amines (±)-1a-d with Different Acylating Agents (2A-C) in Batch Mode with CaLB-MNPs and N435
Similar reaction conditions to those previously described for the KR of all racemic amines (±)-1a-d with all acylating agents (2A-C) were explored to optimize the KR processes in this work [14][15][16][17]. The first screening reactions took place at 40 °C for 6 h in MTBE, toluene, and a mixture of THF:toluene 1:1 as solvents at 12.5 mM substrate concentration. This solvent screen revealed MTBE as the most efficient solvent due to the significantly lower conversions in the other two solvents tested, and it is also the greenest alternative. Therefore, MTBE was used as a solvent in all subsequent tests (Scheme 1).
The results of the KRs in the batch mode at 45 mM substrate concentration are presented in Figure 2. This experiment was performed as detailed in Section 2.2.4., and the results were analyzed with GC on a chiral column by taking samples from the reaction at different intervals. The residual amines (S)-1a-d in KR mixtures were derivatized to their

Biocatalyst Characterization
CaLB-MNPs produced with CaLB immobilized covalently onto the MNPs were proposed as good biocatalysts for KRs of racemic compounds. To prove that the immobilization was successful, the protein concentration of the supernatant before and after immobilization was determined using the Bradford assay. The experiments indicated almost full immobilization (Y = 98%) after 120 min at a CaLB:MNP ratio of 1:15 under the conditions utilized in our previous work [8].
SEM images could not reveal the real surface morphology of the CaLB-MNPs due to the aggregation of the distinct nanoparticles ( Figure S1 in Supplementary Materials). Consequently, the difference between the MNP carrier and CaLB-MNPs could be only revealed using indirect methods based on the activity measurement of the CaLB-MNPs and immobilization yield. Furthermore, the successful immobilization of CaLB on bisepoxideactivated macroporous resins [42] and aspartase on bisepoxide-activated MNPs [43] using the same chemistry support the conclusions drawn from indirect methods. toluene, and a mixture of THF:toluene 1:1 as solvents at 12.5 mM substrate concentration. This solvent screen revealed MTBE as the most efficient solvent due to the significantly lower conversions in the other two solvents tested, and it is also the greenest alternative. Therefore, MTBE was used as a solvent in all subsequent tests (Scheme 1).
The results of the KRs in the batch mode at 45 mM substrate concentration are presented in Figure 2. This experiment was performed as detailed in Section 2.2.4, and the results were analyzed with GC on a chiral column by taking samples from the reaction at different intervals. The residual amines (S)-1a-d in KR mixtures were derivatized to their acetamides via Ac 2 O treatment to determine the conversion and enantiomeric excess of the components. Data of Table 1 indicate that with acylating agents 2A and 2B, the sixth hour of the MNP reactions was comparable to the first hour of the reactions utilizing N435 in terms of conversion, while for KRs with 2C, the first hour of both reactions was used in the comparison. In all cases, the ee of the product and the E values were equal or higher when utilizing the MNP-bound form of CaLB as compared to the forms immobilized onto It is clear from Figure 2 that the various starting amines (±)-1a-d were converted in KRs with quite different rates and levels of selectivity. In the CaLB-MNP-catalyzed KRs, the fastest reaction and highest conversion were attained during the N-acylation of the racemic 1-methoxy-2-propanamine (±)-1b, with all acylating agents 2A-C (Figure 2a,c,e). In the CaLB N435-catalyzed KRs, the acylation of racemic 1-methoxy-2-propanamine (±)-1b utilizing diisopropyl malonate 2A (Figure 2b) and isopropyl 2-ethoxyacetate 2C (Figure 2f) were faster. The CaLB-catalyzed acylation of the larger amines (±)-1a,c,d was slower usually but provided proper conversions (≥40%) after 6 h for all cases except when utilizing CaLB-MNPs and 2A as the acylating agent (Figure 2a). The conversions were usually higher in CaLB N435 experiments, often surpassing the 50% theoretical conversion limit for KRs of exclusive enantiomer selectivity.
Regarding the enantiomeric excess of the products (R)-3(a-d)(A-C)-an indicator of enantiomer selectivity-the values for the CaLB-MNP-catalyzed KRs fluctuated above 80% during the experiments, usually reaching values of ≥99%. On the other hand, in several CaLB N435-catalyzed KRs, somewhat lower enantiomeric excess values could be achieved, e.g., 40% for (R)-3bC (Figure 2f). In summary, with CaLB N435, KRs were faster and reached higher conversion rates, but the enantiomer selectivity was lower than with CaLB-MNPs exhibiting higher selectivity with sufficiently good-to-excellent conversions.
Considering all the experiments shown in Figure 2, it can be concluded that utilizing CaLB-MNPs as biocatalysts and isopropyl 2-ethoxyacetate 2C as the acylating agent (Figure 2e) provides the best balance between conversion and enantiomer selectivity. In this case, N-acylation was accomplished with a good conversion value (50%) for amines (±)-1a,b,d even after 1 h, while the KR of 1-phenylethan-1-amine (±)-1c reached this conversion after 6 h. The enantiomeric excess for amides (R)-3(a,c,d)C was >95% and above 80% for (±)-1b.
Kinetic resolutions of these racemic amines (±)-1a-d with the investigated activated acylating agents 2A-C have already been performed with various immobilized forms of CaLB, either via the covalent attachment or adsorption in various matrices such as mesoporous beads (CaLB T2-150), phenyl functionalized silica gel (CaLB G250P), and microporous acrylic resin (CaLB N435) [14][15][16][17]. In Table 1, the results of CaLB-MNPs and N435 in this work are compared with the relevant results of preceding studies.
The parameters analyzed were the conversion, enantiomeric excess (ee) of the formed amide (R)-3(a-d)(A-C) and residual amine (S)-1a-d fractions, and specific biocatalytic activity (U B ) in the CaLB-MNP-and CaLB N435-catalyzed KRs. Although the enantiomeric excess of the product (ee p ) is an important parameter for irreversible kinetic resolution, this alone cannot represent the degree of enantiomer selectivity (E) [40,41]. Since E is a constant independent of time and substrate concentrations controlled only with the ratio of the specific rate constants for the two enantiomers, k R /k S , this is an intrinsic measure of selectivity. In KRs with lower selectivity, the 50% limit of conversion could be exceeded; therefore, ee p should undergo a decrease. Thus, in a KR catalyzed using a biocatalyst of higher activity but lower selectivity, the product's ee (ee p ) should decrease above 50% conversion.
Data of Table 1 indicate that with acylating agents 2A and 2B, the sixth hour of the MNP reactions was comparable to the first hour of the reactions utilizing N435 in terms of conversion, while for KRs with 2C, the first hour of both reactions was used in the comparison. In all cases, the ee of the product and the E values were equal or higher when utilizing the MNP-bound form of CaLB as compared to the forms immobilized onto porous carriers (N453 [15], and this work, Immozyme T2-150 [16], G250P [15,16]), indicating an enhancement in the selectivity of CaLB on MNPs compared to the N435 or other forms. The specific activity U B values with 2C were also comparable for CaLB-MNPs and N435 except in the KR of 1-phenylethan-1-amine (±)-1c where U B of CaLB-MNPs remained significantly lower than that of N435. Table 1. Results of CaLB-catalyzed kinetic resolution of racemic amines (heptane-2-amine (±)-1a, 1methoxypropan-2-amine (±)-1b, 1-phenylethan-1-amine (±)-1c, and 4-phenylbutan-2-amine (±)-1d) with different acylating agents (AAs a ).

Entry
Amine AA a Amount CaLB t c ee (R)-3(a-d) An important note is that since the mass-to-mass ratio of CaLB and the MNP carrier was 1:15 in the immobilization with high immobilization yield, it could be calculated that approximately 0.8 mg of the CaLB/test was applied in this study. Although the quantity of CaLB in the N435 is not disclosed by the manufacturer, it could be estimated to be more than 10%, up to 2-3 mg [44]. This can explain why one mass unit of the CaLB-MNP was somewhat less active than the commercial N435, which is attributed to the approximate proportionality of the real CaLB content.
The diisopropyl malonate 2A as an acylating agent was previously used in the kinetic resolution of racemic amines (±)-1a-d with the reference biocatalyst Novozyme 435 (N435; from Novozym: CaLB immobilized via adsorption within macroporous acrylic resin) [14]. The results of the reference literature listed as entries 3, 10, 17, and 29 in Table 1 show good conversion values (45-52.1%) with a high ee of the products (R)-3(a-d)A (92-99.9%) and good-to-excellent enantiomeric ratio/selectivity (>100-200), which are in good agreement with our reference data obtained with N435 using a lower concentration of the substrate (45 mM) and lesser quantities of biocatalysts (50 mg/mL) and acylating agent (1 equiv.), as shown in entries 2, 9, 16, and 28. The reactions carried out with the aid of CaLB-MNPs using diisopropyl malonate 2A as an acylating agent under similar conditions are reported in entries 1, 8, 15, and 27. Moderate conversion rates were obtained for (±)-1a,c,d (21.6-33.8%), which were comparable since half of the biocatalyst amount was utilized. For (±)-1b, a quite reasonable conversion (45.9%) was attained. The ee of the product was also excellent (99.0-99.6%), and the E value was on par with or superior to that of CaLB on macroporous supports with Life 2023, 13, 1560 11 of 16 most of the investigated amines. Another advantage of the CaLB-MNP is that the filtration process to separate the biocatalyst can be eliminated due to the magnetic properties of the MNP-supported biocatalyst, enabling easy separation with magnetic decantation.
Isopropyl 2-cyanoacetate 2B (and ethyl 2-cyanoacetate 2B*) were applied as acylating agents in another study on the kinetic resolution of (±)-1c with various immobilized forms of CaLB [15]. It was observed that the conversion values employing the two acylating agents with different natures of leaving alcohol did not differ significantly. The comparison of the three different immobilized forms of CaLB (CaLB N435, CaLB Immozyme T2-150, and CaLB G250P) with ethyl 2-cyanoacetate 2B* (RT, in toluene-THF 1:1, for 24 h) revealed the best conversion with CaLB N435, and thus this form was chosen as the biocatalyst for subsequent studies. When using only 0.5 equivalent of the acylating agent, the conversions were only moderate after 24 h (31.7%, 20.0%, and 17.1%, respectively), as shown in entries 20, 22, and 23. The theoretical 50% conversion of a highly selective KR could only be achieved (50.1%), by using one equivalent of 2B* (entry 21). In contrast to other results of this work indicating significantly higher enantiomer selectivity of CaLB-MNPs when using the reference form N435 in KRs of the other three amines (±)-1a,c,d, in KRs with 1 equiv. of 2B, the selectivity in KR of 1-phenylethan-1-amine (±)-1b (entry 18) was similar to the one with N435 (entry 19), providing even a higher conversion rate (43.9%) and excellent ee of the product (R)-3cB (98.9%), which is comparable to the ones published (98.2-99.9%) [15].
The continuous-flow reactions in a thermostatted continuous-flow U-shape reactor were carried out as detailed in Section 2.2.7. To conduct the continuous-flow experiments under industrially more relevant conditions, they were performed at an elevated concentration (100 mM) of the substrates (±)-1b,c at a 1 µL min −1 flow rate at 40 • C for over two days (Figure 3).
The continuous-flow KR of 1-methoxy-2-propylamine (±)-1b reached a stationary state after 4 h, which could be maintained until 22 h. In the continuous-flow KR of 1phenylethan-1-amine (±)-1c, the stabilization took a bit longer than 4 h, but the conversion also started to decrease after 22 h. Due to activity requirements of the enzyme, the solvent could not be completely dried; therefore, the inhibition by 2-acetoxyacetic acid formed in parallel hydrolysis is the most plausible explanation for this phenomenon. Nevertheless, after 24 h, washing with a substrate-free solvent with a flow rate of 5 µL min −1 for 1 h could regain the original activity of CaLB-MNP, proving the reusability of CaLB-MNPs.
After the 48th hour, the same behavior was observed, indicating that the washing step with a substrate-free solvent is necessary after each 24-h cycle. These results improve our understanding of this potent biocatalyst and corroborate previous findings about the reusability and recyclability of CaLB-MNPs [8,9]. and convenient continuous-flow U-shape reactor inspired us to implement this methodology for several racemic amines to expand the scope of substrates analyzed in this reactor. For this investigation, the KRs of the industrially important chiral amines, (±)-1-methoxy-2-propylamine (±)-1b and (±)-1-phenylethan-1-amine (±)-1c, were selected with the most efficient isopropyl 2-ethoxyacetate 2C acylating agent in the N-acylation of racemic amines.
The results of continuous-flow reactions with other CaLB forms in PBRs cannot be directly compared with the results of the present study's U-shape reactor as a comparison of the specific reaction rate (r) for a batch process (r batch ; essentially identical to U B ) with a flow process (r flow ) is only possible under similar conditions at nearly identical conversions [45]). Therefore, the ratio of the specific reaction rate for a batch and the corresponding flow process (r flow /r batch ) should be considered. The degree of enhancement was up to r flow /r batch~2 when a PBR was used in continuous-flow mode compared with the reaction under a batch mode [45]. Regarding the specific reaction rate values (r = U b ) in KR with 2C catalyzed using CaLB-MNPs for compound (±)-1b, an enhancement of r flow = 9.6 U g −1 over r batch = 7.3 U g −1 was noted, while for compound (±)-1c, a slightly decreased r flow = 4.6 U g −1 from r batch = 6.0 U g −1 was observed. These results indicate that except for the benefit of the continuous retention of the MNP-based catalyst, the U-shape reactor does not offer further advantages over batch-mode reactors. However, other types of MNP-based flow reactors offer the possibility of larger enhancement, which could be the subject of further studies. In a recent study on agitated MNP-based flow reactors, where different types of continuous-flow reactors were compared, including a cell corresponding to a cell of the U-shape reactor used in this work, the U b value in a magnetically fixed cell could be enhanced further by three to four times in an agitated MNP flow reactor cell with rotational movement of two magnets (AFR RM ) [46]. This enhancement of up to fourfold, achievable only with MNP-bound forms, significantly exceeds the maximum twofold enhancement found thus far in PBRs as compared to batch reactions.
The results show that using CaLB-MNPs in batch mode offers identical or superior selectivity with somewhat lower productivity than CaLB on porous supports. On the other hand, with the most efficient isopropyl 2-ethoxyacetate 2C, CaLB-MNPs in the U-shape reactor proved to be efficient acylating agents, resulting in good conversion rates, and highto-excellent enantiomer selectivity. The benefits of CaLB-MNP biocatalysts can be assessed in the future with other compounds of this class, making them a viable alternative for facilitating the kinetic resolution of chiral amines with the acylating agents studied. Future research can also be carried out with other types of acylating agents and flow reactors.

Conclusions
Generally, the kinetic resolution reactions of four racemic amines (±)-1a-d with three different acylating agents, namely diisopropyl malonate 2A, isopropyl 2-cyanoacetate 2B, and isopropyl 2-ethoxyacetate 2C, were successfully performed with CaLB-MNPs as biocatalysts, with moderate-to-excellent conversion rates, providing high enantiomer selectivity in most cases. Although the processes catalyzed using the reference form of CaLB (N435) were usually more productive, in several instances, their degree of enantiomer selectivity remained lower than the CaLB-MNP form. Overall, isopropyl 2-ethoxyacetate 2C performed the best for all amines (±)-1a-d, considering the balance between conversion values and enantiomer excess in KRs with the CaLB-MNP as well as with the reference N435 form. Besides the higher selectivity in several KRs, another benefit of using MNPs as CaLB supports is the facilitated purification and reuse processes. The results of continuousflow experiments corroborate previous findings about the reusability and recyclability of CaLB-MNPs and improve our understanding of this potent biocatalyst in U-shape reactors or other MNP-based flow reactors.