Structural Characterization of Enzymatic Interactions with Functional Nicotinamide Cofactor Biomimetics

: Synthetic nicotinamide biomimetics (NCBs) have emerged as alternatives to the use of natural cofactors. The relatively low cost and ease of manufacture of NCBs may enable the scaling of biocatalytic reactions to produce bulk chemicals (e.g., biofuels and plastics). NCBs are also recognized by only a subset of NAD(P)/NAD(P)H-dependent enzymes, which potentially allows access to orthogonal redox cascades that can be run simultaneously within a single reactor. In the work presented here, a series of NCBs was prepared and tested for activity with alcohol dehydrogenases and ene-reductases. While the NCBs did not support enzymatic activity with the alcohol dehydrogenases, the observed rate of the ene-reductases with NCBs was greater than when incubated with the natural cofactor (consistent with previous observations). We obtained the structures of an ene-reductase and an alcohol dehydrogenase with an NCB bound in their active sites. While the NCB bound to the ene-reductases in a productive position and orientation for hydride transfer to the isoalloxazine ring of the flavin cofactor, the NCB failed to adopt a catalytically competent binding mode in the alcohol dehydrogenase.


Introduction
Biocatalytic methods to produce chiral sec-alcohols, alkanes, and amines using nicotinamide-dependent enzymes such as alcohol dehydrogenases, oxidases, ene-reductases, and imine-reductases are well established [1][2][3][4] and are continuously being expanded and scaled for industrial processes [5].As we move away from petrochemical-based fuels and materials, there will be opportunities for biocatalytic processes to provide some of these products.However, at a very large scale, the use of naturally occurring NAD(P) + /NAD(P)H may not be economically viable (e.g., for the cell-free production of biofuels and chemicals for polymer manufacture) [6][7][8][9].Consequently, there is considerable interest in the use of nicotinamide cofactor analogs, also known as synthetic nicotinamide biomimetics (NCBs; Figure 1), due to their facile and inexpensive synthesis.NCBs are relatively simple small molecules that conserve the hydride transfer moiety (pyridine group) but lack other groups, such as the adenosine moiety [7,9] (Figure 1).In general, the synthesis of nicotinamide biomimetics is simple, resulting in products that can easily be isolated [7].In a review published in 2014, Hollmann and co-authors [7] discussed in detail methods for the modification of the pyridine moiety, including transformation to the reduced-form biomimetic (1,4-dihydropyridine) with sodium dithionite.
NCBs have been demonstrated to work well with ene-reductases [7,10].However, many NCBs have poor or no activity in most other classes of NAD(P) + /NAD(P)H-dependent enzymes, in which hydride transfer to/from the substrate is direct rather than via a flavin cofactor [7].Although this implies that they are only suitable for a restricted number of enzymes [8], this also opens up the possibility to implement bioorthogonality in multienzymatic cascades, wherein having a highly specific one-to-one enzyme/cofactor interaction between the NCB and one of the enzymes in the cascade allows the precise control of the cascade's reaction dynamics [10,11].
The enhanced stability of the NCBs when compared to their natural counterparts may also provide an advantage for their use in continuous flow reactors, where they may be tethered to either an enzyme or the surface of a carrier for biosynthetic processes over extended periods [12][13][14][15][16].While cofactor tethering has been achieved with NAD(P) + /NAD(P)H, the low yield from the semi-synthesis of modified natural nicotinamide cofactors has been recognized as a major impediment in the adoption of this technology by the industry [15].Including functional groups in NCBs that allow facile tethering may overcome this limitation.
This work details the preparation of a series of NCBs carrying an alcohol appended to the pyridinium moiety via a para-substituted aromatic ring and/or a short n-alkyl chain.The NCBs synthesized carry an alcohol group at the end of the molecule distal from the nicotinamide moiety to allow further modification if desired (e.g., for molecular tethering; Figure 1).We assessed the performance of the NCBs with nicotinamide-dependent NCBs are relatively simple small molecules that conserve the hydride transfer moiety (pyridine group) but lack other groups, such as the adenosine moiety [7,9] (Figure 1).In general, the synthesis of nicotinamide biomimetics is simple, resulting in products that can easily be isolated [7].In a review published in 2014, Hollmann and co-authors [7] discussed in detail methods for the modification of the pyridine moiety, including transformation to the reduced-form biomimetic (1,4-dihydropyridine) with sodium dithionite.
NCBs have been demonstrated to work well with ene-reductases [7,10].However, many NCBs have poor or no activity in most other classes of NAD(P) + /NAD(P)H-dependent enzymes, in which hydride transfer to/from the substrate is direct rather than via a flavin cofactor [7].Although this implies that they are only suitable for a restricted number of enzymes [8], this also opens up the possibility to implement bioorthogonality in multienzymatic cascades, wherein having a highly specific one-to-one enzyme/cofactor interaction between the NCB and one of the enzymes in the cascade allows the precise control of the cascade's reaction dynamics [10,11].
The enhanced stability of the NCBs when compared to their natural counterparts may also provide an advantage for their use in continuous flow reactors, where they may be tethered to either an enzyme or the surface of a carrier for biosynthetic processes over extended periods [12][13][14][15][16].While cofactor tethering has been achieved with NAD(P) + /NAD(P)H, the low yield from the semi-synthesis of modified natural nicotinamide cofactors has been recognized as a major impediment in the adoption of this technology by the industry [15].Including functional groups in NCBs that allow facile tethering may overcome this limitation.
This work details the preparation of a series of NCBs carrying an alcohol appended to the pyridinium moiety via a para-substituted aromatic ring and/or a short n-alkyl chain.The NCBs synthesized carry an alcohol group at the end of the molecule distal from the nicotinamide moiety to allow further modification if desired (e.g., for molecular tethering; Figure 1).We assessed the performance of the NCBs with nicotinamide-dependent enzymes that catalyze three important industrial reactions: the organic solvent-tolerant NAD + /NADH-dependent alcohol dehydrogenase from Rhodococcus ruber DSM 44541 (RrADH) [17,18], the thermophilic NADH-dependent ene-reductase from Thermus scotoductus SA-01 (TsER) [9,19,20], and the ene-reductase YqiM from Bacillus subtills [21].These enzymes are used (i) to produce enantiopure sec-alcohols and (ii) for the asymmetric reduction of C=C double bonds.In addition to their industrial potential, each of these enzymes has been characterized extensively, both biochemically [9,[17][18][19][20] and structurally [22][23][24][25], with TsER having been demonstrated to accept NCBs in previous studies [9].Here, the biochemical analysis of the NCBs revealed that only the ene-reductases (TsER and YqiM) can use the NCBs as reductants, and the structures of the complexes of an NCB with TsER and RrADH provide insights into the physical interactions that facilitate catalytic activity in TsER but not RrADH.

Experimental Section
Chemicals were purchased from Sigma Aldrich (Sydney, Australia), unless otherwise stated.Synthetic DNA was designed using Geneious Prime 2022.1, unless otherwise stated, and was purchased from Twist Bioscience (San Francisco, CA, USA) and Genescript Biotech Coro (Piscataway, NJ, USA).

Synthesis of Nicotinamide Biomimetics
See the Supporting Information (Sections S1 and S2; Figures S1-S6) for specific descriptions of the chemical transformations and spectral data.
Here, the biochemical analysis of the NCBs revealed that only the ene-reductases (TsER and YqiM) can use the NCBs as reductants, and the structures of the complexes of an NCB with TsER and RrADH provide insights into the physical interactions that facilitate catalytic activity in TsER but not RrADH.

Experimental Section
Chemicals were purchased from Sigma Aldrich (Sydney, Australia), unless otherwise stated.Synthetic DNA was designed using Geneious Prime 2022.1, unless otherwise stated, and was purchased from Twist Bioscience (San Francisco, CA, USA) and Genescript Biotech Coro (Piscataway, NJ, USA).

Synthesis of Nicotinamide Biomimetics
See the Supporting Information (Sections S1 and S2; Figures S1-S6) for specific descriptions of the chemical transformations and spectral data.

Determination of Molar Extinction Coefficients
The molar extinction coefficients of the natural cofactor NADH and the reduced mimetics mNADHs 3b and 4b (Figure 2) were calculated from the linear ranges of their respective calibration curves (see details in Supplementary Data Section S3), using the Beer-Lambert Law A = εcl, where 'A' is the absorbance at 365 nm (for the mimetics) or 340 nm (for natural NADH); 'Ɛ' is the molar extinction coefficient for reduced nicotinamide (l × mmol −1 × cm −1 ); 'l' is the length of optical pathway (cm); and 'c' represents the concentration of the compound (mmol L −1 ).
' is the molar extinction coefficient for reduced nicotinamide (l × mmol −1 × cm −1 ); 'l' is the length of optical pathway (cm); and 'c' represents the concentration of the compound (mmol L −1 ).

General Details
The gene encoding the protein sequence of the alcohol dehydrogenase from R. ruber DSM 44541 (RrADH; PDB: 3JV7), cloned into the expression vector pET28, was synthesized by Twist Bioscience.Plasmids encoding the ene-reductase from T. scotoductus SA-01 (TsER; PDB: 3HGJ) and flavoprotein YqjM from B. subtilis (PDB: 1Z41), cloned into the expression vector pET28, were synthesized by GenScript Biotech Coro.

General Methods for Protein Expression and Purification
For large-scale protein production, Escherichia coli BL21 (DE3) cells, transformed with the relevant expression plasmids, were incubated in Luria broth (LB), containing either kanamycin or ampicillin (50 µg mL −1 ), at 37 • C and shaken at 200 rpm overnight.Overnight cultures were transferred to 2 L flasks containing LB medium and antibiotic (see details in Supplementary Data Section S4) and incubated (37 • C, 200 rpm) until an OD 600 of 0.6 was reached.The expression of the recombinant proteins was induced with isopropyl β-D-1-thiogalactopyranoside (IPTG; 0.1-1.0mg L −1 ) at different temperatures (15-37 • C; see details in Supplementary Data Section S4).After expression, the cell cultures were harvested by centrifugation (4000× g, 4 • C, 20 min) and the pellets were kept at −20 • C until subsequent protein purification.For protein purification, the pellets were thawed and resuspended in 20 mL of Purification Buffer A (50 mM Tris, 300 mM NaCl, pH 7.5).Cell lysis was performed using the EmulsiFlex-C5 cell homogenizer (Avastin) at 20,000 psi, 4 • C, and cell debris were removed via centrifugation (50,000× g, 40 min, 10 • C; Beckman Coulter Avanti JXN-26).The lysate was filtered using a syringe and a 0.45 µm filter (Merck, Darmstadt, Germany) and applied to a Ni NTA-sepharose column (5 mL, Hi5 His-Trap™, GE Healthcare, Chicago, IL, USA) equilibrated with Purification Buffer A. Protein purification was performed using an Akta Pure FPLC instrument at 4 • C. Following the loading of the column with the lysate, it was washed with five column volumes of Purification Buffer A and Purification Buffer B (50 mM Tris, 300 mM NaCl, 500 mM imidazole, pH 7.5) at concentrations of 4% and 8% v/v.A gradient of imidazole (40-500 mM) was then applied over 15 column volumes at a flow rate of 2 mL min −1 .Fractions (5 mL) were collected and analyzed by SDS-PAGE separation (4-12% Bolt Bis-Tris Plus Polyacrylamide Gel; Invitrogen, Waltham, MA, USA), using MES SDS running buffer, and visualized with AcquaStain (Bulldog Bio, Portsmouth, NH, USA).Protein-rich fractions were pooled and concentrated using a membrane concentrator (10 kDa cutoff) or desalted via dialysis in 50 mM Tris, 100 mM NaCl, pH 7.5.The resulting protein solutions were aliquoted and stored at 4

General Details
The gene encoding the protein sequence of the alcohol dehydrogenase from R. ru DSM 44541 (RrADH; PDB: 3JV7), cloned into the expression vector pET28, was syn sized by Twist Bioscience.Plasmids encoding the ene-reductase from T. scotoductus SA (TsER; PDB: 3HGJ) and flavoprotein YqjM from B. subtilis (PDB: 1Z41), cloned into expression vector pET28, were synthesized by GenScript Biotech Coro.

General Methods for Protein Expression and Purification
For large-scale protein production, Escherichia coli BL21 (DE3) cells, transformed w the relevant expression plasmids, were incubated in Luria broth (LB), containing ei kanamycin or ampicillin (50 µg mL −1 ), at 37 °C and shaken at 200 rpm overnight.O night cultures were transferred to 2 L flasks containing LB medium and antibiotic

Measurements of Enzymatic Activity
The activity of RrADH was determined via the oxidation of 1-phenylethanol.Stock solutions of 1-phenylethanol and the oxidized cofactor biomimetics 3a and 4a were placed in HEPES buffer (25 mM), pH 7.5.The enzyme solution (0.05 mg/mL) was transferred to the same buffer.Enzyme activity assays were performed in a 96-well plate (Greiner UV-Star ® ).The reactions were initiated by adding 10 µL of enzyme solution to a reaction mix (90 µL of 1-phenylethanol stock solution and 100 µL of the cofactor analog stock solution, with final concentrations of 22.5 mM and 200 µM, respectively).Substrate controls consisted of a reaction mix in the absence of an enzyme (see details in Supplementary Data Section S5).Variations in absorbance at 365 nm were monitored for 10 min, and the slopes were determined using GraphPad Prism version 8.0.0 (San Diego, CA, USA).
The ene reduction activity of TsER and YqjM was determined via the reduction of cyclohexen-1-one.Stock solutions of 2-cyclohexen-1-one and the reduced cofactor biomimetics mNADHs 3b and 4b were placed in 25 mM HEPES buffer, pH 7.5.Enzyme solutions (0.05 mg/mL) were placed in the same buffer.Enzyme activity measurements were performed in a 96-well plate (Greiner UV-Star ® ); reactions were performed by adding 10 µL of enzyme solution to the reaction mix (90 µL of 2-cyclohexen-1-one stock solution, a final concentration of 22.5 mM; 100 µL of the cofactor analog stock solution, a final concentration of 200 µM).Substrate controls consisted of a reaction mix in the absence of an enzyme.Enzyme controls consisted of the enzyme solution in the reaction buffer (see details in Supplementary Data Section S5).Changes in absorbance at 365 nm were monitored for 10 min, and slopes were determined using GraphPad Prism version 8.0.0 (San Diego, CA, USA).
For the rate of oxygen reduction by TsER and YqjM, stock solutions of reduced cofactor biomimetics mNADHs 3b and 4b were placed in 25 mM HEPES buffer, pH 7.5.Enzyme solutions (0.05 mg/mL) were placed in 25 mM HEPES, pH 7.5.Enzyme activity assays were performed in a 96-well plate (Greiner UV-Star ® ).Reactions were performed by adding 10 µL of enzyme solution to a reaction mix (90 µL of buffer; 100 µL of cofactor analogue stock solution, final concentration of 200 µM).Substrate controls consisted of a reaction mix in the absence of an enzyme.Enzyme controls consisted of the enzyme solution in the reaction buffer (see details in SI).The variation in absorbance at 365 nm was monitored for 10 min, and slopes were determined using GraphPad Prism version 8.0.0 (San Diego, CA, USA).

Enzyme ativity U mg
where 'DA' is the change in absorbance over time (slope, min −1 ); ' Catalysts 2024, 14, x FOR PEER REVIEW enzymes that catalyze three important industrial reac NAD + /NADH-dependent alcohol dehydrogenase fro (RrADH) [17,18], the thermophilic NADH-dependent ductus SA-01 (TsER) [9,19,20], and the ene-reductase Yqi enzymes are used (i) to produce enantiopure sec-alcoh duction of C=C double bonds.In addition to their ind zymes has been characterized extensively, both bioche [22][23][24][25], with TsER having been demonstrated to acce Here, the biochemical analysis of the NCBs revealed t and YqiM) can use the NCBs as reductants, and the stru with TsER and RrADH provide insights into the physi lytic activity in TsER but not RrADH.

Experimental Section
Chemicals were purchased from Sigma Aldrich (Sy stated.Synthetic DNA was designed using Geneiou stated, and was purchased from Twist Bioscience (San script Biotech Coro (Piscataway, NJ, USA).

Synthesis of Nicotinamide Biomimetics
See the Supporting Information (Sections S1 and scriptions of the chemical transformations and spectral

Determination of Molar Extinction Coefficients
The molar extinction coefficients of the natural cof metics mNADHs 3b and 4b (Figure 2) were calculated spective calibration curves (see details in Supplementar Lambert Law A = εcl, where 'A' is the absorbance at 36 (for natural NADH); 'Ɛ' is the molar extinction coeffic mmol −1 × cm −1 ); 'l' is the length of optical pathway (cm) tion of the compound (mmol L −1 ).
' is the molar extinction coefficient for the reduced nicotinamide analog (L × mmol −1 × cm −1 ); 'l' is the length of the optical pathway (cm); and 'c' is the concentration of the enzyme (mg mL −1 ).

Protein Crystallization and Structural Determination
RrADH and TsER were expressed and purified as described above.After purification and buffer exchange into 20 mM Tris, pH 7.5, RrADH was concentrated to 13 mg/mL (the concentration was determined using the Bradford assay).Compound 1a or 4a (Figure 1) was added (each to a 50 mM final concentration) to RrADH, together with 6% isopropanol, resulting in a final protein concentration of 7.4 mg/mL.Crystallization was performed using hanging drop vapor diffusion at 20 • C with drops consisting of 400 nL of the protein solution and 400 nL of the well solution.All crystallizations were performed using crystal screening VDX plates from Hampton Research.Crystals formed overnight in 0.1 M sodium citrate tribasic dihydrate, pH 5.0, and 30% Jeffamine ED2001.Crystals were collected after three days of growth and transported to the Australian synchrotron in liquid nitrogen without added cryoprotectants.TsER was also exchanged into 20 mM Tris, pH 7.5, after purification and concentrated to 11.1 mg/mL.The enzyme could not be co-crystallized with any of the biomimetics, but crystals could be obtained either in the presence of 0.1 M succinic acid, 0.1 M bicine, pH 8.5, containing 30% MPEG550, or in 15% isopropanol, 0.1 M sodium citrate, pH 5, containing 10% w/v PEG10000.After three days, yellow crystals formed.After eight days of growth, a solution (200 nL) of 3b in 0.1 M sodium citrate, pH 5.6, containing 18% PEG3350, was added to the crystals grown in 15% isopropanol to a final concentration of 2 mM.All crystals were then cryocooled in liquid nitrogen.X-ray diffraction data were collected remotely using the MX-1 and MX-2 beamlines at the Australian Synchrotron (Melbourne) for RrADH and TsER crystals, respectively, and processed with XDS [26][27][28].Model refinement and building was carried out using Phenix v1.20.1 and Coot 0.8.9.2, respectively [29,30].The initial phasing of TsER was performed using the p-hydroxy-benzaldehyde-bound crystal structure of TsER (PDB: 3HGJ) as the template, and the phasing of RrADH was achieved using the NAD-bound crystal structure of this enzyme (PDB: 3JV7) [23,24].

Cofactor Preparation
Oxidized NCBs (1a-4a) were prepared from the reaction of nicotinamide with alkyl halides [10].The subsequent reduction of nicotinamide biomimetics 3a and 4a was achieved by following standard literature protocols that employed sodium dithionite under basic conditions, to provide dihydropyridines 3b and 4b.Attempts to prepare 2b from 2a under the same conditions were unsuccessful [31].Although modest yields of 3b and 4b were obtained (40% and 11%, respectively), sufficient amounts were afforded for biochemical characterization (Figure 3). a final concentration of 2 mM.All crystals were then cryocooled in liquid nitrogen.X-ray diffraction data were collected remotely using the MX-1 and MX-2 beamlines at the Australian Synchrotron (Melbourne) for RrADH and TsER crystals, respectively, and processed with XDS [26][27][28].Model refinement and building was carried out using Phenix v1.20.1 and Coot 0.8.9.2, respectively [29,30].The initial phasing of TsER was performed using the p-hydroxy-benzaldehyde-bound crystal structure of TsER (PDB: 3HGJ) as the template, and the phasing of RrADH was achieved using the NAD-bound crystal structure of this enzyme (PDB: 3JV7) [23,24].

Cofactor Preparation
Oxidized NCBs (1a-4a) were prepared from the reaction of nicotinamide with alkyl halides [10].The subsequent reduction of nicotinamide biomimetics 3a and 4a was achieved by following standard literature protocols that employed sodium dithionite under basic conditions, to provide dihydropyridines 3b and 4b.Attempts to prepare 2b from 2a under the same conditions were unsuccessful [31].Although modest yields of 3b and 4b were obtained (40% and 11%, respectively), sufficient amounts were afforded for biochemical characterization (Figure 3).enzymes that catalyze three important industrial reactions NAD + /NADH-dependent alcohol dehydrogenase from R (RrADH) [17,18], the thermophilic NADH-dependent ene-r ductus SA-01 (TsER) [9,19,20], and the ene-reductase YqiM fr enzymes are used (i) to produce enantiopure sec-alcohols a duction of C=C double bonds.In addition to their industria zymes has been characterized extensively, both biochemica [22][23][24][25], with TsER having been demonstrated to accept N Here, the biochemical analysis of the NCBs revealed that o and YqiM) can use the NCBs as reductants, and the structure with TsER and RrADH provide insights into the physical in lytic activity in TsER but not RrADH.

Experimental Section
Chemicals were purchased from Sigma Aldrich (Sydney stated.Synthetic DNA was designed using Geneious Pri stated, and was purchased from Twist Bioscience (San Fra script Biotech Coro (Piscataway, NJ, USA).

Synthesis of Nicotinamide Biomimetics
See the Supporting Information (Sections S1 and S2; F scriptions of the chemical transformations and spectral data

Determination of Molar Extinction Coefficients
The molar extinction coefficients of the natural cofactor metics mNADHs 3b and 4b (Figure 2) were calculated from spective calibration curves (see details in Supplementary Da Lambert Law A = εcl, where 'A' is the absorbance at 365 nm (for natural NADH); 'Ɛ' is the molar extinction coefficient f mmol −1 × cm −1 ); 'l' is the length of optical pathway (cm); and tion of the compound (mmol L −1 ).365 = 6607 mM −1 cm −1 ) and 4b ( , x FOR PEER REVIEW 3 of 17 enzymes that catalyze three important industrial reactions: the organic solvent-tolerant NAD + /NADH-dependent alcohol dehydrogenase from Rhodococcus ruber DSM 44541 (RrADH) [17,18], the thermophilic NADH-dependent ene-reductase from Thermus scotoductus SA-01 (TsER) [9,19,20], and the ene-reductase YqiM from Bacillus subtills [21].These enzymes are used (i) to produce enantiopure sec-alcohols and (ii) for the asymmetric reduction of C=C double bonds.In addition to their industrial potential, each of these enzymes has been characterized extensively, both biochemically [9,[17][18][19][20] and structurally [22][23][24][25], with TsER having been demonstrated to accept NCBs in previous studies [9].Here, the biochemical analysis of the NCBs revealed that only the ene-reductases (TsER and YqiM) can use the NCBs as reductants, and the structures of the complexes of an NCB with TsER and RrADH provide insights into the physical interactions that facilitate catalytic activity in TsER but not RrADH.

Experimental Section
Chemicals were purchased from Sigma Aldrich (Sydney, Australia), unless otherwise stated.Synthetic DNA was designed using Geneious Prime 2022.1, unless otherwise stated, and was purchased from Twist Bioscience (San Francisco, CA, USA) and Genescript Biotech Coro (Piscataway, NJ, USA).

Synthesis of Nicotinamide Biomimetics
See the Supporting Information (Sections S1 and S2; Figures S1-S6) for specific descriptions of the chemical transformations and spectral data.

Determination of Molar Extinction Coefficients
The molar extinction coefficients of the natural cofactor NADH and the reduced mimetics mNADHs 3b and 4b (Figure 2) were calculated from the linear ranges of their respective calibration curves (see details in Supplementary Data Section S3), using the Beer-Lambert Law A = εcl, where 'A' is the absorbance at 365 nm (for the mimetics) or 340 nm (for natural NADH); 'Ɛ' is the molar extinction coefficient for reduced nicotinamide (l × mmol −1 × cm −1 ); 'l' is the length of optical pathway (cm); and 'c' represents the concentration of the compound (mmol L −1 ).

Catalytic Competence
Like natural nicotinamide cofactors, the reduced forms 3b and 4b exhibit strong absorption between 350 and 375 nm (Figure 3).The reducing capacity of these cofactors was evaluated using the two ene-reductases TsER [24] and YqjM [21].Their oxidized counterparts 3a and 4a were tested in the oxidation of 1-phenylethanol catalyzed by the alcohol dehydrogenase RrADH [23] (Table 1).RrADH exhibited detectable levels of activity when incubated with 3a and 4a, although the enzyme was fully active when assayed in the presence of NAD + .In contrast, the flavin-dependent TsER and YqjM were able to reduce 2-cyclohexen-1-one and molecular oxygen using either 3b or 4b (Figure 4).Biomimetic 3b promoted higher ene-reductase activity (2.8 U/mg vs. 0.7 U/mg for TsER and 3.1 U/mg vs. 0.8 U/mg for YqjM) and oxygen reduction activity (1.2 U/mg vs. 0.5 U/mg for TsER and 0.8 U/mg vs. 0.4 U/mg for YqjM) than 4b for both enzymes.Notably, NCB 4b outperformed the natural NADH cofactor, where the corresponding specific activity for ene/oxygen reduction was 0.3/0.2U/mg and 0.02/0.04U/mg for TsER and YqjM, respectively (Figure 4 and Tables S2 and S3 in the SM).This observation is consistent with previous reports.Paul et al. also reported NCBs outperforming natural nicotinamide cofactors in the asymmetric reduction of a ketoisophorone using ene-reductases TsER and YqjM [10].Earlier, Kanus et al. synthesized a series of NCBs that also showed better performance than the natural nicotinamide cofactor in ene-reductase-catalyzed reactions [32].

Structural Characterization
To gain structural insight into how the NCBs developed in this study may promote enzymatic redox reactions, crystallization trials were carried out.Crystals were obtained for TsER and RrADH in complex with different ligands (succinate and 3b for TsER and 1a and 4b for RrADH).Refinement statistics for all structures are shown in Table 3.In flavin-dependent enzymes, oxygen reduction in the presence of the nicotinamide biomimetic has been reported [33,34].Therefore, we also compared the ratios of the ene and oxygen reduction rates for the biomimetics and the native cofactor (Table 2).For both enzymes, the ene reduction promoted by the NCBs significantly exceeded the oxygen reduction rates, with compound 3b having the highest ratio for both enzymes.Since the presence of hydrogen peroxide can be toxic to the enzyme, leading to enzyme inactivation and/or side reactions [35,36], the suppression of the oxygen reduction reaction in enzymes using the NCBs also contributes to their superior performance when compared to the native cofactor.

Structural Characterization
To gain structural insight into how the NCBs developed in this study may promote enzymatic redox reactions, crystallization trials were carried out.Crystals were obtained for TsER and RrADH in complex with different ligands (succinate and 3b for TsER and 1a and 4b for RrADH).Refinement statistics for all structures are shown in Table 3.
Previous structures of TsER, as well as other ene-reductases, have demonstrated that these enzymes adopt a homodimeric quaternary structure in solution [21,24,25,32,37,38].In TsER, the active site of one of the subunits also uses an arginine finger from the Cterminal end located on the opposing subunits of the dimer (Figure 5A) [21,24,25,38].While some members of this group of enzymes are more active in a homotetrameric form, the active site architecture is highly conserved (described below) [25,38].Here, two new structures of TsER have been solved.TsER bound to succinate was solved at a 2.13 Å In flavin-dependent enzymes, oxygen reduction in the presence of the nicotinamide biomimetic has been reported [33,34].Therefore, we also compared the ratios of the ene and oxygen reduction rates for the biomimetics and the native cofactor (Table 2).For both enzymes, the ene reduction promoted by the NCBs significantly exceeded the oxygen reduction rates, with compound 3b having the highest ratio for both enzymes.Since the presence of hydrogen peroxide can be toxic to the enzyme, leading to enzyme inactivation and/or side reactions [35,36], the suppression of the oxygen reduction reaction in enzymes using the NCBs also contributes to their superior performance when compared to the native cofactor.
Table 1.Reactions used to evaluate the ability of nicotinamide biomimetics to reconstitute enzymatic activity.

Reaction Enzyme
Catalysts 2024, 14, x FOR PEER REVIEW 7 of 17

Catalytic Competence
Like natural nicotinamide cofactors, the reduced forms 3b and 4b exhibit strong absorption between 350 and 375 nm (Figure 3).The reducing capacity of these cofactors was evaluated using the two ene-reductases TsER [24] and YqjM [21].Their oxidized counterparts 3a and 4a were tested in the oxidation of 1-phenylethanol catalyzed by the alcohol dehydrogenase RrADH [23] (Table 1).RrADH exhibited detectable levels of activity when incubated with 3a and 4a, although the enzyme was fully active when assayed in the presence of NAD + .In contrast, the flavin-dependent TsER and YqjM were able to reduce 2cyclohexen-1-one and molecular oxygen using either 3b or 4b (Figure 4).Biomimetic 3b promoted higher ene-reductase activity (2.8 U/mg vs. 0.7 U/mg for TsER and 3.1 U/mg vs. 0.8 U/mg for YqjM) and oxygen reduction activity (1.2 U/mg vs. 0.5 U/mg for TsER and 0.8 U/mg vs. 0.4 U/mg for YqjM) than 4b for both enzymes.Notably, NCB 4b outperformed the natural NADH cofactor, where the corresponding specific activity for ene/oxygen reduction was 0.3/0.2U/mg and 0.02/0.04U/mg for TsER and YqjM, respectively (Figure 4 and Tables S2 and S3 in the SM).This observation is consistent with previous reports.Paul et al. also reported NCBs outperforming natural nicotinamide cofactors in the asymmetric reduction of a ketoisophorone using ene-reductases TsER and YqjM [10].Earlier, Kanus et al. synthesized a series of NCBs that also showed better performance than the natural nicotinamide cofactor in ene-reductase-catalyzed reactions [32].
Table 1.Reactions used to evaluate the ability of nicotinamide biomimetics to reconstitute enzymatic activity.

Reaction Enzyme
TsER and YqjM

Catalytic Competence
Like natural nicotinamide cofactors, the reduced forms 3b and 4b exhibit strong absorption between 350 and 375 nm (Figure 3).The reducing capacity of these cofactors was evaluated using the two ene-reductases TsER [24] and YqjM [21].Their oxidized counterparts 3a and 4a were tested in the oxidation of 1-phenylethanol catalyzed by the alcohol dehydrogenase RrADH [23] (Table 1).RrADH exhibited detectable levels of activity when incubated with 3a and 4a, although the enzyme was fully active when assayed in the presence of NAD + .In contrast, the flavin-dependent TsER and YqjM were able to reduce 2cyclohexen-1-one and molecular oxygen using either 3b or 4b (Figure 4).Biomimetic 3b promoted higher ene-reductase activity (2.8 U/mg vs. 0.7 U/mg for TsER and 3.1 U/mg vs. 0.8 U/mg for YqjM) and oxygen reduction activity (1.2 U/mg vs. 0.5 U/mg for TsER and 0.8 U/mg vs. 0.4 U/mg for YqjM) than 4b for both enzymes.Notably, NCB 4b outperformed the natural NADH cofactor, where the corresponding specific activity for ene/oxygen reduction was 0.3/0.2U/mg and 0.02/0.04U/mg for TsER and YqjM, respectively (Figure 4 and Tables S2 and S3 in the SM).This observation is consistent with previous reports.Paul et al. also reported NCBs outperforming natural nicotinamide cofactors in the asymmetric reduction of a ketoisophorone using ene-reductases TsER and YqjM [10].Earlier, Kanus et al. synthesized a series of NCBs that also showed better performance than the natural nicotinamide cofactor in ene-reductase-catalyzed reactions [32].
Table 1.Reactions used to evaluate the ability of nicotinamide biomimetics to reconstitute enzymatic activity.

TsER and YqjM
RrADH RrADH Previous structures of TsER, as well as other ene-reductases, have demonstrated that these enzymes adopt a homodimeric quaternary structure in solution [21,24,25,32,37,38].In TsER, the active site of one of the subunits also uses an arginine finger from the C-terminal end located on the opposing subunits of the dimer (Figure 5A) [21,24,25,38].While some members of this group of enzymes are more active in a homotetrameric form, the active site architecture is highly conserved (described below) [25,38].Here, two new structures of TsER have been solved.TsER bound to succinate was solved at a 2.13 Å resolution in the space group P2 1 , while TsER bound to 3b was solved at a 2.75 Å resolution in the space group P1.The succinate-bound structure was solved with four subunits in the asymmetric unit, with the biologically active dimer being formed due to a two-fold non-crystallographic symmetry operation.The 3b-bound structure, on the other hand, was solved with eight subunits in the asymmetric unit.Despite differences in subunit organization of the crystal structures, their overall folds and active site structures are consistent with each other, as well as with the previously published structures of this enzyme.For instance, the 348 Cα atoms of a monomer from 8UAJ (succinate-bound TsER) superimpose with an RMSD of 0.202Å with the corresponding atoms of the TsER structure with the PDB code 3HGJ [24], and it is composed of eight twisted β-strands surrounded by eight α-helices, forming a TIM barrel structure (see Supplementary Data Section S6).The N-terminal end of the structure is capped by a β-hairpin loop, while the C-terminal end is open, accommodating the solvent-exposed active site.
Ene-reductases employ a bi-bi ping-pong mechanism, which can be separated into two half reactions (Figure 4) [32,38].In the first, the non-covalently bound flavin is reduced by the non-covalently bound NAD(P)H (or mNADH), and, in the second, the reduced flavin promotes the reduction of the substrate.The ene-substrate is primed for its reduction through hydrogen bonding to an active-site-conserved histidine pair (His172 and His175), which causes the polarization of the C=C bond and facilitates the hydride transfer from the N5 of the flavin cofactor to the substrate.A second proton is simultaneously donated by the highly conserved Tyr177 (Figure 5).These transfers are aided by the spatial proximity between the amide group of the reduced nicotinamide cofactor/biomimetic and the histidine pair and between the nicotinamide ring, the flavin mononucleotide and Tyr177 [24].The structure of TsER with compound 3b in the active site shows that the positioning of the NCB satisfies these requirements, suggesting that the function of the adenosine moiety of the natural cofactor is mainly the enhancement of its binding affinity, not its orientation (Figure 5B).This interpretation is supported by the superposition of the TsER structure with bound 3b and the structure of the Thermoanaerobacter pseudethanolicus old yellow enzyme (TOYE, Figure 5C) with the bound NADH mimic 1,4,5,6-tetrahydronicotinamide adenine dinucleotide (NADH 4 ) [37], which shows that while the position of the nicotinamide ring is not identical in both complexes (Figure 5D), both are consistent with intramolecular hydride transfer.While their amide groups point in opposite directions (stabilized by hydrogen bonds to Tyr27/Arg347 or Cys25 and the His172/His175 pair in TsER or TOYE, respectively; note that Arg347 in TOYE is used to stabilize the phosphate groups of the adenosine moiety), the interatomic distance between the hydride-transferring carbon on the NADH mimetic and the N5 nitrogen on FMN is between 3.56 and 3.70 Å in TOYE and between 3.45 and 3.71 Å in TsER.Although this interatomic distance appears to be slightly shorter in the 3b-bound structure, no functional conclusion may be drawn from this difference due to a coordinate error of 0.26 Å (maximum-likelihood-based).The electron density for 3b can be observed in the active sites of each chain; only in chains A, B and F could the complete biomimetic be accommodated.In the other chains, there is insufficient electron density around the phenol moiety of 3b.This observation suggests that 3b binds predominantly via its nicotinamide ring, while the phenol is more flexible.The structure of TsER was also solved with succinate bound to the active site (Figure 6), a molecule that can occupy the same location in the active site as 3b.However, while the latter only forms two hydrogen bonds between its amide group and residues Tyr27 and Arg347, succinate is held in position through a series of hydrogen bonds involving its carboxyl groups and the conserved residues Tyr27, His172, His175 and Arg347, where Tyr27, His172 and His175 are found on the same chain, while Arg347 is positioned on the neighboring protein chain.
(maximum-likelihood-based).The electron density for 3b can be observed in the act sites of each chain; only in chains A, B and F could the complete biomimetic be accomm dated.In the other chains, there is insufficient electron density around the phenol moie of 3b.This observation suggests that 3b binds predominantly via its nicotinamide rin while the phenol is more flexible.6), a molecule that can occupy the same location in the active site as 3b.However, w the latter only forms two hydrogen bonds between its amide group and residues T and Arg347, succinate is held in position through a series of hydrogen bonds involvin carboxyl groups and the conserved residues Tyr27, His172, His175 and Arg347, w Tyr27, His172 and His175 are found on the same chain, while Arg347 is positioned o neighboring protein chain.Crystal structures were also obtained for RrADH in complex with 1a or 4a at re tions of 2.2 Å and 2.99 Å, respectively (Table 3; Figure 7B,C).Interestingly, the ele density in the location of 4a better fits the reduced form of the mimetic, i.e., 4b, rather the oxidized 4a (Figure 7B).This suggests that the mimetic was reduced, either b crystallization conditions or by RrADH (although no catalytic activity was detected netic assays).RrADH has been shown to be a tetramer (composed of a dimer of dim in previous studies; however, the two structures solved in this work were found to 12 copies in an asymmetric unit (although this is likely due to crystal contacts).Both s tures were solved in the space group P21 (consistent with previously published struc of RrADH), and, despite variations in the asymmetric unit, the monomeric structur RrADH-1a and RrADH-4b closely resemble those of previously reported RrADH s tures.For example, the superposition of 345 Cα atoms of RrADH-1a and RrADH-4b a previously reported ADH from R. ruber (PDB code: 3JV7) resulted in RMSD valu 0.458 and 0.539 Å, respectively.
The ADH monomer contains two domains, a catalytic domain and a domain res sible for the binding of NAD + (Supplementary Figure S12).The NAD + -binding do Crystal structures were also obtained for RrADH in complex with 1a or 4a at resolutions of 2.2 Å and 2.99 Å, respectively (Table 3; Figure 7B,C).Interestingly, the electron density in the location of 4a better fits the reduced form of the mimetic, i.e., 4b, rather than the oxidized 4a (Figure 7B).This suggests that the mimetic was reduced, either by the crystallization conditions or by RrADH (although no catalytic activity was detected in kinetic assays).RrADH has been shown to be a tetramer (composed of a dimer of dimers) in previous studies; however, the two structures solved in this work were found to have 12 copies in an asymmetric unit (although this is likely due to crystal contacts).Both structures were solved in the space group P2 1 (consistent with previously published structures of RrADH), and, despite variations in the asymmetric unit, the monomeric structures of RrADH-1a and RrADH-4b closely resemble those of previously reported RrADH structures.For example, the superposition of 345 Cα atoms of RrADH-1a and RrADH-4b with a previously reported ADH from R. ruber (PDB code: 3JV7) resulted in RMSD values of 0.458 and 0.539 Å, respectively.
The ADH monomer contains two domains, a catalytic domain and a domain responsible for the binding of NAD + (Supplementary Figure S12).The NAD + -binding domain contains a Rossmann-fold motif, which is characteristic of proteins that bind nucleotidecontaining cofactors (such as NAD(P) and FAD) [39], while the catalytic domain contains a central β-sheet surrounded by a series of α-helices (Supplementary Figure S12) [40].The enzyme contains two zinc ions, both of which are located in the catalytic domain, one having a mostly structural function (not visible in Figure 7) and another (shown in Figure 7) that is important for the catalytic reaction (Supplementary Figure S12) [23].The NAD + cofactor binds near the catalytic zinc ion and sits in the cleft between the two domains [23,40].
Previous structural characterizations of ADH from R. ruber (PDB 3JV7 and 2XAA), horse liver (PDB 1P1R, 4XD2 and 1HLD), mouse (PDB 1E3I) and Aeropyrum pernix (PDB 1H2B) have identified relevant interactions between the enzyme and NAD(H) [23,[41][42][43].Mostly, these interactions do not involve the nicotinamide group, but instead include interactions between the enzyme and ADP moiety of the bound cofactor and the sidechains of His39, Ser40, Arg340 and Asp203, as well as backbone interactions with Val180, Gly182 and Leu183 [23,44,45].These interactions appear to be required to position the nicotinamide group for hydride transfer with the alcohol substrate.The benzyl alcohol group in 4b and OTBS group in 1a (Figure 7D) are positioned near Ser40, which places the NCB within the substrate-binding pocket, thus preventing the binding of the substrate.
In the NAD + -bound structures of RrADH, the catalytic zinc ion is coordinated by three amino acid residues (Cys38, His62 and Asp153).The fourth position in the distorted tetrahedral coordination environment is occupied by a hydroxide or substrate mimics such as acetate (e.g., PDB code 3JV7; Figure 7C).In the RrADH-4b complex, Glu63 occupies the position corresponding to that of the hydroxide/substrate.This coordination arrangement has been previously observed in several bacterial class I ADHs (e.g., from E. coli, Sulfolobus solfataricus, Clostridium beijerinckii and Thermoanaerobacter brockii) [46][47][48], typically in their resting states (i.e., in the presence of the cofactor but no substrate mimic).This glutamate residue is highly conserved among class I ADHs and its substitution results in a significant decrease in catalytic efficiency [46,49].It has been proposed that Glu63 is involved in the catalytic mechanism of the enzyme by promoting the displacement of the hydroxide ligand from the catalytic Zn 2+ and its replacement by the alcohol substrate [45,50,51].Thus, the RrADH-4b complex represents an intermediate between the resting state and the catalytically competent Michaelis complex.
Catalysts 2024, 14, x FOR PEER REVIEW contains a Rossmann-fold motif, which is characteristic of proteins that bind nucl containing cofactors (such as NAD(P) and FAD)) [39], while the catalytic domain co a central β-sheet surrounded by a series of α-helices (Supplementary Figure S12) [4 enzyme contains two zinc ions, both of which are located in the catalytic doma having a mostly structural function (not visible in Figure 7) and another (shown in 7) that is important for the catalytic reaction (Supplementary Figure S12) [23].The cofactor binds near the catalytic zinc ion and sits in the cleft between the two do [23,40].
Previous structural characterizations of ADH from R. ruber (PDB 3JV7 and 2 horse liver (PDB 1P1R, 4XD2 and 1HLD), mouse (PDB 1E3I) and Aeropyrum perni 1H2B) have identified relevant interactions between the enzyme and NAD(H) [23, Mostly, these interactions do not involve the nicotinamide group, but instead incl teractions between the enzyme and ADP moiety of the bound cofactor and the side of His39, Ser40, Arg340 and Asp203, as well as backbone interactions with Val180, and Leu183 [23,44,45].These interactions appear to be required to position the ni mide group for hydride transfer with the alcohol substrate.The benzyl alcohol gr 4b and OTBS group in 1a (Figure 7D) are positioned near Ser40, which places th within the substrate-binding pocket, thus preventing the binding of the substrate.In the NAD + -bound structures of RrADH, the catalytic zinc ion is coordinate three amino acid residues (Cys38, His62 and Asp153).The fourth position in the dist tetrahedral coordination environment is occupied by a hydroxide or substrate m such as acetate (e.g., PDB code 3JV7; Figure 7C).In the RrADH-4b complex, Glu63 pies the position corresponding to that of the hydroxide/substrate.This coordinatio rangement has been previously observed in several bacterial class I ADHs (e.g., fro coli, Sulfolobus solfataricus, Clostridium beijerinckii and Thermoanaerobacter brockii) [46 typically in their resting states (i.e., in the presence of the cofactor but no substrate mi This glutamate residue is highly conserved among class I ADHs and its substitutio sults in a significant decrease in catalytic efficiency [46,49].It has been proposed that G is involved in the catalytic mechanism of the enzyme by promoting the displaceme the hydroxide ligand from the catalytic Zn 2+ and its replacement by the alcohol sub [45,50,51].Thus, the RrADH-4b complex represents an intermediate between the re state and the catalytically competent Michaelis complex. The Glu63-bound Zn 2+ geometry and the binding orientation of the cofactor mim also observed in the RrADH-1a complex (Figure 7B).However, in many of the poly tide chains, the electron density is present that extends from the catalytic Zn 2+ ion tow His39, as well as the density suggesting the presence of the cofactor binding obse above (see Supplementary Data Section S6).Although the quality of the data is insuffi to accurately model 1a, the density in this region may suggest a conformational is where the OTBS tail extends towards His39 in a conformation reminiscent of the ad sine moiety in NAD + observed in ADH structures containing the cofactor and sub mimics (i.e., structures where the substrate mimic, not Glu63, binds to the catalytic [23,[42][43][44].In addition, the density of the catalytic Zn2+-binding site suggests the pre of two alternative conformations, one where Glu63 coordinates to the metal (represe a catalytically non-competent, intermediate state) and one where Glu63 is not bou the metal ion, representing a possible geometric arrangement for the catalytically co tent Michaelis complex (Supplementary Figure S13).These Zn 2+ isomers are obse most prominently in chains displaying a stronger density in the region around His39 The Glu63-bound Zn 2+ geometry and the binding orientation of the cofactor mimic is also observed in the RrADH-1a complex (Figure 7B).However, in many of the polypeptide chains, the electron density is present that extends from the catalytic Zn 2+ ion towards His39, as well as the density suggesting the presence of the cofactor binding observed above (see Supplementary Data Section S6).Although the quality of the data is insufficient to accurately model 1a, the density in this region may suggest a conformational isomer where the OTBS tail extends towards His39 in a conformation reminiscent of the adenosine moiety in NAD + observed in ADH structures containing the cofactor and substrate mimics (i.e., structures where the substrate mimic, not Glu63, binds to the catalytic Zn 2+ ) [23,[42][43][44].In addition, the density of the catalytic Zn2+-binding site suggests the presence of two alternative conformations, one where Glu63 coordinates to the metal (representing a catalytically non-competent, intermediate state) and one where Glu63 is not bound to the metal ion, representing a possible geometric arrangement for the catalytically competent Michaelis complex (Supplementary Figure S13).These Zn 2+ isomers are observed most prominently in chains displaying a stronger density in the region around His39 (e.g., Chain A) further supporting the presence of multiple cofactor-binding conformations.These observations suggest that 1a can adopt a conformation that is suitable to promote catalytic activity, but, due to the lack of interactions between this mimic and the protein beyond the nicotinamide moiety, this conformation is not stable (or preferred).A comparison of the RrADH structures described here demonstrates that the loop extending from Ala50 to Leu56 occurs in two major conformations (Figure 8).
Chain A) further supporting the presence of multiple cofactor-binding conform These observations suggest that 1a can adopt a conformation that is suitable to pr catalytic activity, but, due to the lack of interactions between this mimic and the p beyond the nicotinamide moiety, this conformation is not stable (or preferred).A co ison of the RrADH structures described here demonstrates that the loop extending Ala50 to Leu56 occurs in two major conformations (Figure 8).In this work, we synthesized and characterized a series of nicotinamide biomim The biocatalytic capacity of these molecules was assessed using the ene-reductase and YqjM and the alcohol dehydrogenase RrADH.The biomimetics 3b and 4b (Fig outperformed the natural cofactor NADH in TsER-and YqjM-catalyzed reactions 2).Higher catalytic rates for eneoate reductases using NCBs rather than NADH a without precedence [10,32].Gueddes et al. have proposed a plausible explanation f observation based on the temperature-dependent kinetic isotope effects (KIEs) of H fer by ene-reductases in reactions involving other NCBs [52].Their studies sugge the methylbenzene and butyl moieties of the reduced NCBs BNA + , 1-benzyl-1,4 dronicotinamide and BuNA + , 1-butyl-1,4-dihydronicotinamide (Figure 1) contribute stabilization of the positive charge in the nicotinamide ring, facilitating more rap dride transfer.However, it is unclear to what extent the observed kinetic differenc be attributed to electronic effects, as other effects, such as (i) variations in the rates diffusion to and from the active site, (ii) differences in the redox potential of NAD the NCB and (iii) complex steric effects resulting in transient non-productive bindin entations of the cofactors also may play important roles.From the structural data ob In this work, we synthesized and characterized a series of nicotinamide biomimetics.The biocatalytic capacity of these molecules was assessed using the ene-reductases TsER and YqjM and the alcohol dehydrogenase RrADH.The biomimetics 3b and 4b (Figure 2) outperformed the natural cofactor NADH in TsERand YqjM-catalyzed reactions (Table 2).Higher catalytic rates for eneoate reductases using NCBs rather than NADH are not without precedence [10,32].Gueddes et al. have proposed a plausible explanation for this observation based on the temperature-dependent kinetic isotope effects (KIEs) of Htransfer by ene-reductases in reactions involving other NCBs [52].Their studies suggest that the methylbenzene and butyl moieties of the reduced NCBs BNA + , 1-benzyl-1,4dihydronicotinamide and BuNA + , 1-butyl-1,4-dihydronicotinamide (Figure 1) contribute to the stabilization of the positive charge in the nicotinamide ring, facilitating more rapid hydride transfer.However, it is unclear to what extent the observed kinetic differences can be attributed to electronic effects, as other effects, such as (i) variations in the rates for the diffusion to and from the active site, (ii) differences in the redox potential of NADH and the NCB and (iii) complex steric effects resulting in transient non-productive binding orientations of the cofactors also may play important roles.From the structural data obtained in this study, it is apparent that the NCBs are positioned in a catalytically competent conformation for hydride transfer with the isoalloxazine moiety of the flavin cofactor (Figure 5), highlighting the importance of the interactions between these cofactors to promote effective catalysis.While beyond the scope of the present study, molecular dynamics simulations may shed further light on the factors contributing to the differences in activity between the NCBs and their native counterpart.

Figure 2 .
Figure 2. Synthesis of oxidized and reduced nicotinamide biomimetics.Experimental details are described in the SI.

Figure 4 .
Figure 4. NCB-promoted reactions catalyzed by the flavin-dependent enzymes TsER and YqjM.(Left) The ene reduction activity was evaluated using the substrate 2-cyclohexen-1-one.The reaction with the substrate oxygen generates hydrogen peroxide.(Right) Relative ene and oxygen reduction activity for TsER and YqjM with biomimetics, mNADHs 3b and 4b, and the native cofactor, NADH.In flavin-dependent enzymes, oxygen reduction in the presence of the nicotinamide biomimetic has

Figure 4 .
Figure 4. NCB-promoted reactions catalyzed by the flavin-dependent enzymes TsER and YqjM.(Left) The ene reduction activity was evaluated using the substrate 2-cyclohexen-1-one.The reaction with the substrate oxygen generates hydrogen peroxide.(Right) Relative ene and oxygen reduction activity for TsER and YqjM with biomimetics, mNADHs 3b and 4b, and the native cofactor, NADH.In flavin-dependent enzymes, oxygen reduction in the presence of the nicotinamide biomimetic has been reported[33,34].Therefore, we also compared the ratios of the ene and oxygen reduction rates for the biomimetics and the native cofactor (Table2).For both enzymes, the ene reduction promoted by the NCBs significantly exceeded the oxygen reduction rates, with compound 3b having the highest ratio for both enzymes.Since the presence of hydrogen peroxide can be toxic to the enzyme, leading to enzyme inactivation and/or side reactions[35,36], the suppression of the oxygen reduction reaction in enzymes using the NCBs also contributes to their superior performance when compared to the native cofactor.

Figure 5 .Figure 5 .
Figure 5. Crystal structures of compound 3b bound to TsER and an NADH analog, NADH4, bou to TOYE.(A) The overall structure of the catalytic dimer of TsER (3b-bound structure).Individ monomers are shown in light orange and teal.(B) The active site structure of chain A of TsER w Figure 5. Crystal structures of compound 3b bound to TsER and an NADH analog, NADH 4 , bound to TOYE.(A) The overall structure of the catalytic dimer of TsER (3b-bound structure).Individual monomers are shown in light orange and teal.(B) The active site structure of chain A of TsER with reduced NCB 3b bound.The 2F o -F c electron density for 3b is shown as blue chicken wire and is contoured to 0.8 σ.The bound substrate and relevant amino acid residues discussed in the text are shown in teal.Oxygen is shown in red, nitrogen in blue and the peptide backbone in light orange.Ligand-protein interactions are shown as grey dotted lines.Note that Arg347 is fit as a conformational isomer.(C) The active site structure of TOYE with NADH 4 bound (PDB 3KRZ).(D) The superimposition of the bound 3b structure of TsER and the NADH 4 -bound structure of TOYE (PDB 3KRZ).The 3b-bound structure is shown as teal sticks, and the NADH 4 -bound structure is shown as light orange wires.In both structures, the relevant amino acid residues are shown in light orange, while oxygens are shown in red and nitrogen in blue.Note that Arg347 in the TOYE structure is only partially fit.Water molecules close to the active site are not shown for clarity.

Figure 6 .
Figure 6.The structure of TsER with succinate bound to the active site.The Fo-Fc electron densi the bound succinate is shown as blue chicken wire and contoured to 3.5 σ.The bound substrat relevant amino acid residues discussed in the text are shown in teal.Oxygen is shown in red, gen in blue and the peptide backbone in light orange.Ligand-protein interactions are shown a dotted lines.Note that Arg347 is fit as a conformational isomer.

Figure 6 .
Figure 6.The structure of TsER with succinate bound to the active site.The F o -F c electron density for the bound succinate is shown as blue chicken wire and contoured to 3.5 σ.The bound substrate and relevant amino acid residues discussed in the text are shown in teal.Oxygen is shown in red, nitrogen in blue and the peptide backbone in light orange.Ligand-protein interactions are shown as grey dotted lines.Note that Arg347 is fit as a conformational isomer.

Figure 7 .
Figure 7.The crystal structures of 4b, 1a and NAD + bound to RrADH.(A) The overall structure of RrADH (4a bound) with α-helicies shown in teal, β-sheets shown in maroon and loops shown in beige.(B) Active site of RrADH with 4b bound.The F o -F c density is shown for the bound biomimetic in blue chicken wire contoured to 2.75 σ. (C) Active site RrADH with 1a bound.The F o -F c density is shown for the bound biomimetic in blue chicken wire contoured to 3 σ.(D) The active site of NAD +bound RrADH (PDB 3JV7), where ACY is acetic acid and MPD is a pentanediol substrate.(E) The superposition of the active sites of the 4b-bound RrADH structure and the NAD + -bound structure.In the superposition, the 4b-bound structure is shown in teal and the NAD + -bound structure in light orange.In all other structures, relevant amino acid residues are shown in teal.The peptide backbone is shown in light orange.In all structures, metal coordination bonds are shown as grey dotted lines, oxygen is shown in red, nitrogen is shown in blue, sulfur is shown in yellow and silicon is shown in beige.

Figure 8 .
Figure 8. Structure of the 1a-bound RrADH displaying the multiple conformations of the Leu56 loop.(A) Chain A loop, 2Fo-Fc electron density shown as blue mesh contoured to 1 Chain B loop, 2Fo-Fc electron density shown as blue mesh contoured to 1 σ.(C) Superimpos Chain A (shown in light orange) and Chain B (shown in teal).

Figure 8 .
Figure 8. Structure of the 1a-bound RrADH displaying the multiple conformations of the Ala50-Leu56 loop.(A) Chain A loop, 2F o -F c electron density shown as blue mesh contoured to 1 σ.(B) Chain B loop, 2F o -F c electron density shown as blue mesh contoured to 1 σ.(C) Superimposition of Chain A (shown in light orange) and Chain B (shown in teal).

Table 2 .
Ratios of the rates of ene and oxygen reduction catalyzed by the flavoproteins of this study as a function of different cofactors.

Table 2 .
Ratios of the rates of ene and oxygen reduction catalyzed by the flavoproteins of this study as a function of different cofactors.

Table 3 .
Data collection parameters, refinement and Ramachandran plot statistics for each of the solved crystal structures.Statistics for the highest-resolution shell are shown in brackets.