Immobilization of Old Yellow Enzymes via Covalent or Coordination Bonds

: Ene-reductases (ERs) belonging to the old yellow enzyme (OYE) family have been thoroughly investigated for the stereospeciﬁc reduction of activated prochiral C = C double bonds. In this work, OYE3 was immobilized both by covalent binding on glyoxyl-agarose (OYE3-GA), and by a ﬃ nity-based adsorption on EziG TM particles (OYE3-EziG). The immobilized OYE3-GA was demonstrated to be active (activity recovery = 52%) and to retain almost 100% of its activity under the enzymatic assay conditions (50 mM phosphate bu ﬀ er pH 7, 28 ◦ C) for six days, whereas the activity of the non-immobilized enzyme dropped to 50% after two days. In the case of EziG TM , the highest activity recovery (54%) was achieved by using the most hydrophilic carrier (EziG TM Opal) that was selected for the full characterization of this type of enzyme preparation (stability, recycling, re-use, enzyme leakage). OYE3-EziG was slightly less stable than OYE3-GA under the same experimental conditions. OYE3-GA could be recycled and re-used for up to 12 reaction cycles in the bioreduction of α -methyl- trans -cinnamaldehyde; after 12 runs, the highest conversion achieved was 40%. In the case of the co-immobilized OYE3 / GDH-EziG, the conversion dropped to 56% after two reaction cycles. No enzyme leakage was detected over 48 h for both OYE3-GA and OYE3 / GDH-EziG (50 mM phosphate bu ﬀ er pH 7, 28 ◦ C). These seed results pave the way for a true optimization of the immobilization of OYE3, as well as for the use of immobilized OYE3 for preparative applications both in batch and continuous ﬂow conditions.


Introduction
Ene-reductases (ERs) belonging to the old yellow enzyme (OYE) family (EC 1.6.99.1) are flavin mononucleotide (FMN)-containing oxidoreductases, which are able to catalyze the stereoselective reduction of C=C double bonds activated by the presence of a suitable electron-withdrawing group (EWG) as a substituent of the alkene moiety [1,2]. The reaction occurs through a stereospecific anti hydrogen addition to the C=C double bond, according to the following steps [3]. The enzyme-bound flavin (reduced by NAD(P)H cofactor to FMNH 2 ) transfers a hydride to the olefin carbon atom in β position with respect to the EWG. A proton is delivered by a tyrosine side-chain (Tyr196 for OYE1-3) of the enzyme active site to the α-carbon atom (with respect to EWG) on the opposite face of the alkene. For the application of OYE-mediated reductions in synthetic sequences of practical interest, it found to be suitable for the soluble enzyme, but not for assessing the activity of the immobilized enzyme (and, indeed, to calculate the activity recovery after immobilization) because of the interference of the solid carrier with the spectrophotometric readouts. Therefore, a GC-MS assay (see Scheme 2 and Materials and Methods, paragraph 3.3.2) based on the reduction of (R)-carvone (2) in the presence of stoichiometric NADH was used. Unless otherwise stated, this latter assay was used to determine the activity of both soluble and immobilized OYE3.
A spectrophotometric assay was also used to assess the activity of soluble GDH by measuring the formation of NADH over time at 340 nm as a result of glucose oxidation (experimental conditions: 50 mM phosphate buffer pH 7, [substrate] = 20 mM, [NAD + ] = 0.2 mM, 28 °C, reaction time = 2 min, volume = 1 mL). Glucose dehydrogenase was coupled to OYE3 for cofactor regeneration in the reduction of α-methyl-trans-cinnamaldehyde (3) (see Scheme 3 and Materials and Methods, paragraph 3.4).

Preliminary Study of OYE3 Stability at Alkaline pH
Covalent immobilization of enzymes on glyoxyl-agarose (GA) relies on the formation of imino bonds between the non-protonated ε-amino groups of enzyme Lys residues and the aldehyde groups of the carrier [19,20]. This immobilization is required to be carried out at alkaline pH (i.e., pH ≥ 10) and a final chemical reduction step is necessary for the generation of irreversible C-N bonds. Whenever compatible with the enzyme stability, sodium borohydride is used to this aim, since it transforms Schiff's bases into secondary amino bonds and the unreacted aldehyde groups of the carrier into inert hydroxyl moieties. Alternatively, the milder sodium cyanoborohydride or 2-picoline borane can be used for those enzymes which are sensitive to sodium borohydride reduction (but unreacted aldehyde groups of the carrier remain unchanged in this case) [28]. Thus, the first step before immobilization on GA was to assay the stability of OYE3 at alkaline pH. A pH "window" of 8-10 was considered and 3 h was selected as the endpoint (according to the general protocol for this type of immobilization) [19]. Enzyme activity before and after incubation at pH 8-10 (see paragraph 3.5, Materials and Methods) was assessed spectrophotometrically using the cyclohex-2-enone assay, Scheme 1). Data of the stability assay at pH 8-10 are reported in Figure 1. The residual activity of OYE3 after 3 h at pH 10 was about 70%. This result was considered acceptable for the immobilization on GA. found to be suitable for the soluble enzyme, but not for assessing the activity of the immobilized enzyme (and, indeed, to calculate the activity recovery after immobilization) because of the interference of the solid carrier with the spectrophotometric readouts. Therefore, a GC-MS assay (see Scheme 2 and Materials and Methods, paragraph 3.3.2) based on the reduction of (R)-carvone (2) in the presence of stoichiometric NADH was used. Unless otherwise stated, this latter assay was used to determine the activity of both soluble and immobilized OYE3.
A spectrophotometric assay was also used to assess the activity of soluble GDH by measuring the formation of NADH over time at 340 nm as a result of glucose oxidation (experimental conditions: 50 mM phosphate buffer pH 7, [substrate] = 20 mM, [NAD + ] = 0.2 mM, 28 °C, reaction time = 2 min, volume = 1 mL). Glucose dehydrogenase was coupled to OYE3 for cofactor regeneration in the reduction of α-methyl-trans-cinnamaldehyde (3)

Preliminary Study of OYE3 Stability at Alkaline pH
Covalent immobilization of enzymes on glyoxyl-agarose (GA) relies on the formation of imino bonds between the non-protonated ε-amino groups of enzyme Lys residues and the aldehyde groups of the carrier [19,20]. This immobilization is required to be carried out at alkaline pH (i.e., pH ≥ 10) and a final chemical reduction step is necessary for the generation of irreversible C-N bonds. Whenever compatible with the enzyme stability, sodium borohydride is used to this aim, since it transforms Schiff's bases into secondary amino bonds and the unreacted aldehyde groups of the carrier into inert hydroxyl moieties. Alternatively, the milder sodium cyanoborohydride or 2-picoline borane can be used for those enzymes which are sensitive to sodium borohydride reduction (but unreacted aldehyde groups of the carrier remain unchanged in this case) [28]. Thus, the first step before immobilization on GA was to assay the stability of OYE3 at alkaline pH. A pH "window" of 8-10 was considered and 3 h was selected as the endpoint (according to the general protocol for this type of immobilization) [19]. Enzyme activity before and after incubation at pH 8-10 (see paragraph 3.5, Materials and Methods) was assessed spectrophotometrically using the cyclohex-2-enone assay, Scheme 1). Data of the stability assay at pH 8-10 are reported in Figure 1. The residual activity of OYE3 after 3 h at pH 10 was about 70%. This result was considered acceptable for the immobilization on GA. carriers was not performed at this stage. Due to the positive result achieved with EziG TM Opal, this carrier was selected for the full characterization of this type of enzyme preparation (stability, recycling, re-use, enzyme leakage).

Stability of Immobilized OYE3
Increased stability is recognized as one of the most important advantages of immobilized enzymes [31,32]. To verify whether this "general rule" might be applied also to immobilized OYE3 preparations, the stability of both OYE3-GA and OYE3-EziG (Opal) was tested under the experimental conditions used for the reduction of (R)-carvone (2) (Scheme 2). The stability assay was carried on for six days ( Figure 2). From the stability time course, it clearly emerged that immobilization exerted a stabilizing effect on the soluble enzyme. The OYE3-GA retained almost 100% of its activity, whereas the activity of the non-immobilized enzyme dropped to 50% after two days. On the other hand, OYE3-EziG was slightly less stable than OYE3-GA, but it still retained about 90% of its starting activity after 144 h.

Preliminary Study of OYE3 Stability at Alkaline pH
Covalent immobilization of enzymes on glyoxyl-agarose (GA) relies on the formation of imino bonds between the non-protonated ε-amino groups of enzyme Lys residues and the aldehyde groups  [19,20]. This immobilization is required to be carried out at alkaline pH (i.e., pH ≥ 10) and a final chemical reduction step is necessary for the generation of irreversible C-N bonds. Whenever compatible with the enzyme stability, sodium borohydride is used to this aim, since it transforms Schiff's bases into secondary amino bonds and the unreacted aldehyde groups of the carrier into inert hydroxyl moieties. Alternatively, the milder sodium cyanoborohydride or 2-picoline borane can be used for those enzymes which are sensitive to sodium borohydride reduction (but unreacted aldehyde groups of the carrier remain unchanged in this case) [28]. Thus, the first step before immobilization on GA was to assay the stability of OYE3 at alkaline pH. A pH "window" of 8-10 was considered and 3 h was selected as the endpoint (according to the general protocol for this type of immobilization) [19]. Enzyme activity before and after incubation at pH 8-10 (see Section 3.5, Materials and Methods) was assessed spectrophotometrically using the cyclohex-2-enone assay, Scheme 1). Data of the stability assay at pH 8-10 are reported in Figure 1. The residual activity of OYE3 after 3 h at pH 10 was about 70%. This result was considered acceptable for the immobilization on GA.

Immobilization of OYE3
Immobilization of OYE3 on GA (OYE3-GA) was performed according to a standard protocol by using a 2 mg g −1 protein loading [19], as described in paragraph 3.6 (Materials and Methods). In order to extend the enzyme stability during the immobilization, the reaction was performed at 4 °C. After 3 h, almost all the protein was immobilized (immobilization yield = 95%), as revealed by protein measurements of the supernatant from the immobilization reaction. The activity recovery, after reduction of the immobilized enzyme with NaBH4, was 52% (Table 1).  1 Calculated as reported in paragraph 3.6 [29]. 2 Determined by the (R)-carvone (2) assay (see Scheme 2). 3 Calculated as reported in paragraph 3.6 [29]. Activity measurements were performed at least in duplicate. Deviation from mean values was below 5%.
For the immobilization by metal-coordination, three EziG TM resins were tested (namely Opal or EziG 1, Coral or EziG 2, and Amber or EziG 3). According to the indications of the supplier, Opal is

Immobilization of OYE3
Immobilization of OYE3 on GA (OYE3-GA) was performed according to a standard protocol by using a 2 mg g −1 protein loading [19], as described in Section 3.6 (Materials and Methods). In order to extend the enzyme stability during the immobilization, the reaction was performed at 4 • C. After 3 h, almost all the protein was immobilized (immobilization yield = 95%), as revealed by protein measurements of the supernatant from the immobilization reaction. The activity recovery, after reduction of the immobilized enzyme with NaBH 4 , was 52% (Table 1). For the immobilization by metal-coordination, three EziG TM resins were tested (namely Opal or EziG 1, Coral or EziG 2, and Amber or EziG 3). According to the indications of the supplier, Opal is a hydrophilic carrier (glass), whereas Coral and Amber are hydrophobic and semi-hydrophilic, respectively. The lower hydrophilicity of Coral and Amber in comparison with Opal is due to the presence of different polymeric coatings on the particle silica surface [30]. Specifically, EziG TM Coral is made from hybrid CPG with polyvinyl benzyl chloride, whereas EziG TM Amber is produced from hybrid CPG with a blended co-polymer. Pore size distribution of the carriers ranges from~500 Å as standard (Opal) to~300 Å (Coral and Amber, due to the polymer coating). According to the indications of the supplier, a higher protein loading (4.2 mg g −1 ) was applied to EziG TM carriers. In all cases, the immobilization was complete. The activity recovery for EziG TM Opal was 54%, thus resulting comparable to that previously obtained with GA. In the case of the hybrid CPG Coral and Amber, the activity recovery was less than half with respect to EziG TM Opal. Under the same binding chemistry, it is plausible that this result may be ascribed to the effect of the polymer coating in tuning the hydrophilicity-hydrophobicity balance of the carrier. However, a systematic study on these carriers was not performed at this stage. Due to the positive result achieved with EziG TM Opal, this carrier was selected for the full characterization of this type of enzyme preparation (stability, recycling, re-use, enzyme leakage).

Stability of Immobilized OYE3
Increased stability is recognized as one of the most important advantages of immobilized enzymes [31,32]. To verify whether this "general rule" might be applied also to immobilized OYE3 preparations, the stability of both OYE3-GA and OYE3-EziG (Opal) was tested under the experimental conditions used for the reduction of (R)-carvone (2) (Scheme 2). The stability assay was carried on for six days (Figure 2). From the stability time course, it clearly emerged that immobilization exerted a stabilizing effect on the soluble enzyme. The OYE3-GA retained almost 100% of its activity, whereas the activity of the non-immobilized enzyme dropped to 50% after two days. On the other hand, OYE3-EziG was slightly less stable than OYE3-GA, but it still retained about 90% of its starting activity after 144 h.
Catalysts 2020, 10, x FOR PEER REVIEW 5 of 12 the hydrophilicity-hydrophobicity balance of the carrier. However, a systematic study on these carriers was not performed at this stage. Due to the positive result achieved with EziG TM Opal, this carrier was selected for the full characterization of this type of enzyme preparation (stability, recycling, re-use, enzyme leakage).

Stability of Immobilized OYE3
Increased stability is recognized as one of the most important advantages of immobilized enzymes [31,32]. To verify whether this "general rule" might be applied also to immobilized OYE3 preparations, the stability of both OYE3-GA and OYE3-EziG (Opal) was tested under the experimental conditions used for the reduction of (R)-carvone (2) (Scheme 2). The stability assay was carried on for six days (Figure 2). From the stability time course, it clearly emerged that immobilization exerted a stabilizing effect on the soluble enzyme. The OYE3-GA retained almost 100% of its activity, whereas the activity of the non-immobilized enzyme dropped to 50% after two days. On the other hand, OYE3-EziG was slightly less stable than OYE3-GA, but it still retained about 90% of its starting activity after 144 h.
A catalytic amount of the cofactor (NADP + ) was used since a GDH glucose regeneration system was coupled to the main biotransformation. In the case of OYE3-GA, GDH was added as a soluble enzyme to each reaction cycle. In the case of OYE3-EziG, a co-immobilized preparation OYE3/GDH was used, taking advantage of the His-tag displayed by both enzymes. The enzymes (OYE3 and GDH) were incubated with EziG TM . Activity recovery for OYE3 was 55%. OYE3-GA could be recycled and re-used for up to 12 reaction cycles. The highest conversion achieved after 12 runs was 40% ( Figure 3). Moreover, it is worth noting that OYE3-GA retained completely its activity up to the fifth reaction cycle. In the case of OYE3/GDH-EziG, the initial conversion dropped to 56% after two reaction cycles. Interestingly, when the conversion % was almost negligible (reaction 11) and the biotransformation was supplemented with a fresh aliquot of soluble GDH, the reaction afforded a 30% conversion, thus suggesting that the gradual drop of the conversions registered for OYE3/GDH-EziG might depend either on the loss of activity of immobilized GDH or on its leakage from the carrier and/or deactivation, thereof. A catalytic amount of the cofactor (NADP + ) was used since a GDH glucose regeneration system was coupled to the main biotransformation. In the case of OYE3-GA, GDH was added as a soluble enzyme to each reaction cycle. In the case of OYE3-EziG, a co-immobilized preparation OYE3/GDH was used, taking advantage of the His-tag displayed by both enzymes. The enzymes (OYE3 and GDH) were incubated with EziG TM . Activity recovery for OYE3 was 55%. OYE3-GA could be recycled and re-used for up to 12 reaction cycles. The highest conversion achieved after 12 runs was 40% ( Figure 3). Moreover, it is worth noting that OYE3-GA retained completely its activity up to the fifth reaction cycle. In the case of OYE3/GDH-EziG, the initial conversion dropped to 56% after two reaction cycles. Interestingly, when the conversion % was almost negligible (reaction 11) and the biotransformation was supplemented with a fresh aliquot of soluble GDH, the reaction afforded a 30% conversion, thus suggesting that the gradual drop of the conversions registered for OYE3/GDH-EziG might depend either on the loss of activity of immobilized GDH or on its leakage from the carrier and/or deactivation, thereof.

Protein Leakage Assay of Immobilized OYE3
Protein leakage from the immobilization carrier is undesirable for several reasons: Loss of activity of the biocatalyst, lack of data reproducibility in the investigated biotransformation, costs and product contamination, to name a few [29]. Covalent immobilization of the enzyme to the carrier generally prevents the protein from leaching from the support surface. In this context, glyoxylagarose and other surface-functionalized carriers (i.e., epoxy carriers) are frequently preferred because of the "irreversible" nature of the bonds generated between the enzyme and the carrier. On the other hand, the high affinity and the specificity of the binding interaction based on the His-tag as in EziG TM -type carriers is plausible to generate a strong enzyme₋support interaction which might prevent enzyme leakage. However, predictions about how an enzyme will respond to an immobilization process are often hard to make. Enzyme leakage is routinely verified by protein measurements [34] of the supernatant in which the immobilized preparation is incubated under specific conditions dictated by the type of information needed (e.g., reaction or storage conditions). Both OYE3-GA and OYE3/GDH-EziG were, thus, incubated in the reaction buffer (50 mM phosphate buffer pH 7) at 28 °C under stirring for 48 h. At scheduled times, samples of the incubation buffer were collected and submitted to SDS-PAGE analysis; silver staining, due to its high sensitivity in the

Protein Leakage Assay of Immobilized OYE3
Protein leakage from the immobilization carrier is undesirable for several reasons: Loss of activity of the biocatalyst, lack of data reproducibility in the investigated biotransformation, costs and product contamination, to name a few [29]. Covalent immobilization of the enzyme to the carrier generally prevents the protein from leaching from the support surface. In this context, glyoxyl-agarose and other surface-functionalized carriers (i.e., epoxy carriers) are frequently preferred because of the "irreversible" nature of the bonds generated between the enzyme and the carrier. On the other hand, the high affinity and the specificity of the binding interaction based on the His-tag as in EziG TM -type carriers is plausible to generate a strong enzyme-support interaction which might prevent enzyme leakage. However, predictions about how an enzyme will respond to an immobilization process are often hard to make. Enzyme leakage is routinely verified by protein measurements [34] of the supernatant in which the immobilized preparation is incubated under specific conditions dictated by the type of information needed (e.g., reaction or storage conditions). Both OYE3-GA and OYE3/GDH-EziG were, thus, incubated in the reaction buffer (50 mM phosphate buffer pH 7) at 28 • C under stirring for 48 h.
At scheduled times, samples of the incubation buffer were collected and submitted to SDS-PAGE analysis; silver staining, due to its high sensitivity in the low nanogram range, was used to detect protein traces after electrophoretic separation [35]. From the inspection of the SDS-PAGE gel (data not reported), no enzyme traces were detected for either OYE3-GA or OYE3/GDH-EziG. This suggests that the loss of activity of both immobilized enzymes registered during the recycling and re-use study, although to a different extent, cannot be ascribed to protein leakage phenomena.

General
Chemicals and solvents were purchased from Merck Sigma-Aldrich ® Srl (Milano, Italy) and used without further purification. Agarose gel 6B-CL was purchased from Amersham Biosciences (Uppsala, Sweden) and activated to glyoxyl-agarose (GA), as previously reported [19]. EziG TM carriers were kindly provided by EnginZyme AB (Sweden) as free samples. GC-MS analyses were performed using a HP-5MS column (30 m × 0.25 mm × 0.25 µm, Agilent). The following temperature program was employed: 60 • C (1 min), 6 • C min −1 /150 • C (1 min), 12 • C min −1 /280 • C (5 min). Bradford assay [33] was performed by using a Shimadzu UV-1601 spectrophotometer and bovine serum albumin (BSA) was used as reference standard protein. All the enzymatic reactions were kept under stirring and controlled temperature by using a thermoshaker New Brunswick™ Excella ® E24. Reaction extraction and phase separation were accomplished through a ZX3 Advanced Vortex Mixer from VELP Scientifica ® and a Microcentrifuge ScanSpeed Mini (LaboGene™), respectively.

Enzyme Production
The enzymes OYE3 from Saccharomyces cerevisiae and GDH from Bacillus megaterium were prepared as His-tagged proteins and purified according to standard methods [36]. Concentration of the purified OYE3 was in the range 2-3 mg mL −1 , whereas the concentration of GDH was 3-4 mg mL −1 . Activity of OYE3 was measured both by the cyclohex-2-enone assay (see Section 3.3.1 and Scheme 1) and the (R)-carvone assay (see Section 3.3.2 and Scheme 2), resulting in 0.09 ± 0.008 U mg −1 and 0.13 ± 0.007 U mg −1 , respectively. Activity of GDH was measured by a spectrophotometric assay (see Section 3.3.1), resulting in 6.7 ± 0.35 U mg −1 .

Spectrophotometric Enzymatic Assay
(a) For the OYE activity assay, in a 2 mL cuvette, a solution of cyclohex-2-enone (5 µL; stock solution = 10 mM in DMSO) and a solution of NADH (20 µL; stock solution = 10 mM in water) were added to 50 mM phosphate buffer pH 7 (916 µL). The reaction was carried out at 28 • C and started by adding to the blank solution 59 µL of OYE3 (177 µg, conc. = 3 mg mL −1 ).
(b) For the GDH activity assay, in a 2 mL cuvette, a solution of glucose (205 µL; stock solution = 1 M in water) and a solution of NAD + (20 µL; stock solution = 10 mM in water) were added to 50 mM phosphate buffer pH 7 (960 µL). The reaction was carried out at 28 • C and started by adding to the blank solution 2 µL of GDH (6 µg, conc. = 3 mg mL −1 ).
Catalysts 2020, 10, 260 8 of 12 One unit (U) is defined as the amount of enzyme that converts 1 µmol of substrate in the product in 1 min under specific temperature and pH conditions.

Stability of OYE3 in the pH Range 8-10
The stability of non-immobilized OYE3 was assayed in the pH range 8-10 by incubating 100 µL of the enzyme solution (stored in 20 mM phosphate buffer pH 8.5 containing 200 mM imidazole and 500 mM NaCl) in 100 µL of 50 mM phosphate buffer pH 7.5 (final pH = 8), 50 mM Tris-HCl buffer pH 9.5 (final pH = 9), or 50 mM NaOH-glycine buffer pH 11.5 (final pH = 10) at 20 • C. After 3 h, 100 µL were withdrawn from each sample and the enzymatic activity was measured spectrophotometrically in the reduction of cyclohex-2-enone (see Section 3.3.1). The residual activity was calculated according to the following equation: Residual activity (%) = observed activity at the endpoint starting activity × 100 Catalysts 2020, 10, 260 9 of 12 3.6. Covalent Immobilization of OYE3 on Glyoxyl-Agarose (OYE3-GA) Immobilization of OYE3 on glyoxyl-agarose was performed following the protocol previously reported [19]. Briefly, glyoxyl-agarose (1.8 g = 2.52 mL) was suspended in 50 mM Na 2 CO 3 buffer (14.3 mL) at pH 10. After the addition of the desired amount of protein (1.2 mL, loading: 2 mg g −1 ), the suspension was kept under mechanical stirring for 3 h at 4 • C. Chemical reduction of Schiff bases and unreacted aldehyde groups was carried out over 30 min by adding NaBH 4 to the mixture (18 mg; 1 mg mL −1 of suspension). The immobilized enzyme was then filtered and washed with 10 mM phosphate buffer pH 7 and deionized water.
The immobilization reaction was monitored by measuring the amount of protein in the supernatant before the addition of the carrier and after 1.5 h and 3 h [34]. The immobilization yield (%, endpoint = 3 h) was calculated according to the following equation [29]: The activity of the immobilized OYE3 was measured by the GC-MS enzymatic assay reported in Section 3.3.2 (reduction of (R)-carvone, 2). The activity recovery (%) was calculated according to the following equation [29]: Activity recovery (%) = observed activity starting activity × 100

Immobilization of OYE3 by Affinity-Based Adsorption (OYE3-EziG)
Immobilization of OYE3 by affinity-based adsorption on EziG TM (Opal, Coral, and Amber) was performed following the protocol of the supplier with minor modifications. EziG TM (100 mg) was added to a 2 mL test tube containing the enzyme (140 µL, loading = 4.2 mg·g −1 of dry carrier) diluted in 50 mM phosphate buffer pH 7 (1.5 mL) at r.t. After 30 min, the test tube was centrifugated (5 min, 13,200 rpm) and the supernatant was removed by pipetting. Following supplier's instructions, the wet enzyme preparation was stored as such (when the immobilized enzyme was filtered under vacuum, it showed an almost complete loss of activity. The weight of the wet carrier was estimated to be 4.5-fold higher than the dry carrier. The activity recovery (%) was referred to the wet carrier weight).
The immobilization yield (%, endpoint = 30 min) was calculated as reported above (Section 3.6). The activity of the immobilized OYE3 was measured by the (R)-carvone assay (see Section 3.3.2.) and the activity recovery (%) was calculated as described above (Section 3.6).
Co-immobilization of OYE3 and GDH on EziG™ Opal was carried out by following the procedure applied to OYE3 immobilization. Briefly, EziG™ Opal (100 mg) was added to 50 mM phosphate buffer pH 7 (1.5 mL) containing OYE3 (140 µL, U = 0.06) and GDH (45 µL, U = 0.87). Monitoring and work-up of the immobilization mixture was performed as above reported.

Stability of Non-Immobilized OYE3, OYE3-GA, and OYE3-EziG
For each endpoint, non-immobilized or immobilized OYE3 (42 µg of protein) was incubated in a solution containing 50 mM phosphate buffer pH 7 (890 µL) and (R)-carvone (10 µL; stock solution = 100 mM in DMSO) at 28 • C. Each reaction was started by adding NADH (100 µL; stock solution: 10 mM in water). The final volume was 1 mL. The enzyme activity was assessed by the standard activity assay described in the Section 3.3.2 (reduction of (R)-carvone). The enzyme activity was compared with the starting activity which was considered as 100%.

Protein Leakage Assay
For each endpoint (1-3-6-24-48 h), 42 µg of protein (corresponding to 21 mg of OYE3-GA and 45 mg of OYE3/GDH-EziG) were incubated in 50 mM phosphate buffer pH 7 (200 µL). The suspensions were kept at 28 • C under stirring (thermoshaker). At each endpoint, a sample (15 µL) was withdrawn from the supernatant and analyzed by SDS-PAGE. GDH and OYE3 (450 ng, corresponding to 0.15 µL of each stock solution) were loaded in the gel as reference standards. Precision Plus Protein™ Unstained Protein Standards (10-250 kD) was used as molecular weight proteins size marker. PAGE analysis was performed in a Mini-Protean 3 system (Bio-Rad). SDS-PAGE was carried out on 12% polyacrylamide (running buffer = 50 mM MES, 50 mM Tris base, 0.1 % w/v SDS and 1 mM EDTA). Proteins were stained by means of a silver staining kit (SilverQuest™ kit, Thermofisher) according to supplier's protocol.

Recycling of Immobilized Enzymes
Recycling of immobilized OYE3-GA and OYE3/GDH-EziG was performed by evaluating the conversion (%) of α-methyl-trans-cinnamaldehyde (3) into 3a (see Section 3.4). At the end of the reaction, the immobilized biocatalyst was washed twice with 1 mL of 50 mM phosphate buffer pH 7, separated through centrifugation, and then re-suspended under the conditions above described for the next reaction cycle.

Conclusions
Ene-reductases are a powerful tool for the stereoselective reduction of activated C=C double bonds under mild and environmentally friendly conditions. Due to the great interest in these enzymes for preparative applications, the access to stable and reusable heterogeneous biocatalysts is being sought. Immobilization of OYE3 both on the well-known glyoxyl-agarose (GA) by covalent binding and on controlled pore glass EziG TM by affinity-based adsorption was shown to produce active, stable and reusable biocatalysts. Although activity recovery for both enzyme preparations was similar (≈50%), OYE3-GA appeared to be superior about stability and recycling properties. In this latter regard, OYE3-GA could be re-used up to five reaction cycles without any loss of activity. Still, upon a prolonged re-use (up to 12 runs), this immobilized biocatalyst allowed to obtain a 40% conversion in the reduction of α-methyl-trans-cinnamaldehyde (3). On the other hand, in the case of the co-immobilized OYE3/GDH-EziG, a drop of conversion was observed after two reaction cycles, presumably due to deactivation of GDH. Interestingly, no evidence of enzyme leaching from both the supports was registered.