doi.org/10.26434/chemrxiv.11481834.v1

Cupin Variants as Macromolecular Ligand Library for StereoselectiveMichael Addition of NitroalkanesNobutaka Fujieda, Miho Yuasa, Yosuke Nishikawa, Genji Kurisu, Shinobu Itoh, Haruna Ichihashi

Submitted date: 31/12/2019 • Posted date: 31/12/2019Licence: CC BY-NC-ND 4.0Citation information: Fujieda, Nobutaka; Yuasa, Miho; Nishikawa, Yosuke; Kurisu, Genji; Itoh, Shinobu;Ichihashi, Haruna (2019): Cupin Variants as Macromolecular Ligand Library for Stereoselective MichaelAddition of Nitroalkanes. ChemRxiv. Preprint. https://doi.org/10.26434/chemrxiv.11481834.v1

Cupin superfamily proteins (TM1459) work as a macromolecular ligand framework with a double-strandedbeta-barrel structure ligating to a Cu ion through histidine side chains. Variegating the first coordination sphereof TM1459 revealed that H52A and H54A/H58A mutants effectively catalyzed the diastereo- andenantio-selective Michael addition reaction of nitroalkanes to an α,β-unsaturated ketone. Moreover, in silicosubstrate docking signified C106N and F104W single-point mutations, which inverted the diastereoselectivityof H52A and further improved the stereoselectivity of H54A/H58A, respectively.

File list (2)

download fileview on ChemRxivManuscript_191231_NF.pdf (2.13 MiB)

download fileview on ChemRxivSI_191231_NF.pdf (3.32 MiB)

1

Cupin Variants as Macromolecular Ligand Library for Stereoselective Michael Addition of Nitroalkanes

Nobutaka Fujieda,*1,2 Haruna Ichihashi,1 Miho Yuasa,2 Yosuke Nishikawa,3 Genji Kurisu,3 and Shinobu Itoh*1

ABSTRACT: Cupin superfamily proteins (TM1459) work as a macromolecular ligand framework with a double-stranded b-barrel structure ligating to a Cu ion through histidine side chains. Variegating the first coordination sphere of TM1459 revealed that H52A and H54A/H58A mutants effectively catalyzed the diastereo- and enantio-selective Michael addition reaction of nitroalkanes to an α,β-unsaturated ketone. Moreover, in silico substrate docking signified C106N and F104W single-point mutations, which inverted the diastereoselectivity of H52A and further improved the stereoselectivity of H54A/H58A, respectively.

INTRODUCTION

In modern organic synthesis, asymmetric reactions have been carried out using transition-metal complexes with chiral ligands.1,2 However, the recognition of compounds bearing multiple chiral centers is still challenging and has garnered research interest, considering the increasing industrial and economic demands.3 Artificial metalloenzymes are emerging hybrid biocatalysts, in which intrinsic chemical reactivity and selectivity of the introduced metal ion or metal complex can be enhanced by modifying the secondary coordination sphere of the protein or DNA scaffold.4–13 By using artificial metalloenzymes, several enantioselective carbon–carbon bond forming reactions, such as Diels-Alder,5,14,15 Michael addition,16,17 and Friedel–Crafts alkylation18 reactions, were established in water by employing α,β-unsaturated ketones harboring methylimidazole or pyridine groups.

Figure 1. Protein metal ligands for Michael addition reaction of nitroalkane. (A) The subunit structure of Cu-bound wild type TM1459 (PDB code: 6L2D). The protein main chain is displayed as ribbon, key amino acid residues as sticks, Cu ion as green sphere and waters as red sphere. (B) The model structures of TM1459 mutants produced in silico and the Michael addition reaction targeted in this study.

Several highly enantioselective Michael addition reactions in water were developed using a simple DNA-based Cu-complex as a hybrid biocatalyst. Enantioselectivity of up to 99 and 94 % ee was attained by using dimethyl malonate and nitromethane as nucleophiles, respectively, and α,β-unsaturated 2-acylimidazoles as Michael acceptors.16 The transition-metal complexes have been used to develop stereoselective Michael addition reactions

by Pedro and Blay’s group.19 Recently, they synthesized pyBOX (pyridine-2,6-bis(oxazolines))-derived ligand-La(III) complex and performed the Michael addition of nitroethane or nitropropane to azachalcone ((E)-3-phenyl-1-(pyridin-2-yl)prop-2-en-1-one) in moderate enantio- and diastereoselectivity.19 Recently, a similar system involving Sc(III) complex with C2-symmetric Schiff base ligands has been developed for the catalytic asymmetric Michael addition of nitroalkanes to azachalcone N-oxides by Wang’s group with good diastereo- and enantioselectivity.20 Thus, owing to the synthetic versatility of nitro-groups, immense progress of the asymmetric Michael additions of nitroalkane has been reported, but satisfactory control of diastereoselectivity and enantioselectivity has not yet been accomplished. Additionally, in both studies, unsatisfactory catalytic efficiency was achieved [the turnover number (TON) < 10].19,20

We have developed artificial metalloenzymes using several proteins as a macromolecular metal ligand.21,22 We have also harnessed the metal-binding promiscuity of the cupin superfamily protein (TM1459) to develop an Os-cupin complex system for the regioselective cis-1,2-dihydroxylation reaction.23 TM1459 is derived from hyperthermophile, whose scaffold is very thermally stable, and has a molecular mass of only 12,977 Da (residues 1–114) with a metal-binding site consisting of a 4-His metal binding motif (Figure 1A); however, its function is unknown.24, 25 Herein, we have demonstrated that the small cupin variants of TM1459 (Figure 1B) can be effectively used to create efficient and robust biocatalysts for the stereoselective Michael addition reaction of nitroalkanes to α,β-unsaturated ketone, 2-azachalcone 1, achieving excellent diastereo- and enantioselectivity as well as high TON.

RESULTS AND DISCUSSION

The crystal structure of Cu-bound wild type TM1459 (WT) was determined (Figure 1A, Figures S1 and Table S1-S3). Considering chiral bidentate or tridentate ligands bearing N atoms as donors such as bis(oxazoline) ligands (N,N-BOX and N,N,N-PyBOX derivatives) are frequently employed in the asymmetric reactions,1,2 rationally designed mutants containing 2-His and 3-His metal-binding motifs were constructed (Figure 1B) on the basis of the crystal structure of Cu-bound WT (Figure 1 and Figure S1). In 3-His-motif-containing mutants, the metal binding sites of H52A and H58A mutants were expected to show facial coordination geometry, whereas those of H54A and H92A mutants exhibited a meridional coordination geometry (Figure

2

1B). In 2-His-motif-containing mutants, H54A/H92A presumably formed trans dyad, whereas others formed cis dyad (Figure 1B). The TM1459 mutants were prepared and purified in the same manner as WT (Figure S2 and S3 and Table S4).

We investigated the catalytic activity of a series of divalent transition metal ions (Mn, Fe, Co, Ni, Cu, and Zn) for the Michael addition reaction of nitromethane to azachalcone 1 and found that CuSO4 exhibited highest yield (Table S5), as in the case of the reported work for the Diels-Alder reaction.26 Highly selective mutants were examined for the Michael addition reaction with TM1459 variants as mini library consisting of various first coordination spheres (Table 1, entry 2-11). We found a trace amount of the racemic product of 2 with WT, H54A, and H52A/H54A (Table 1, entry 2, 4 and 8). On the other hand, other mutants showed a variety of enantiomeric excess (ee). Notably, the S enantiomer was obtained with excellent yield and enantioselectivity (ee = 99 % (S)) by employing H52A, whereas the R enantiomer was obtained with good yield and enantioselectivity (ee = 89 % (R)) with H54A/H58A (Table 1, entries 3 and 9 and Figure S4).

To propose a plausible stereochemical mechanism that can account for the observed enantioselectivity, we determined the crystal structures of the Cu-bound H52A mutant with highly reliable crystallographic statistics at the resolution of 1.20 Å (Figure 2A, Figure S5, and Table S1-S3). The coordination geometry of Cu-H52A mutant showed facial histidine triad and two water molecules located in a cis-position, completing the N3O2 square pyramidal geometry (Figure 2A and Figure S5C). It

is worth noting that cysteine 106 was posttranslationally modified to cysteine sulfinic acid, as previously observed (Figure 2A).23 The involvement of neighboring amino acid residues was evaluated by performing a docking simulation using azachalcone. In the lowest energy conformation of 1, the polar edge of the pyridine moiety pointed toward the Cu active site (calculated free energy of binding (DGb): -7.0 (kcal/mol), Figure 2B). In this azachalcone-docked structure, the b-Si-face attack (the Si-face of the double bond b-carbon of azachalcone) by nitromethane takes place at the position indicated by the dotted circle in Figure 2B. This space was tightly packed by the five amino acid residues, Phe41, Ile49, Phe94, Phe104, and Cys106 (sulfinic acid), while the opposite side is solvent accessible. Thus, we presumed that the preferential attack via b-Si-face was enforced by the hydrogen bonds exerted with cysteine sulfinic acid 106 rather than the shielding effect of protein matrix.

Figure 2. (A) Crystal structures of Cu-binding site of H52A (Chain A, PDB code: 6L2E) and (C) H54A/H58A (Chain A, PDB code: 6L2F) mutants. (B) The surface models and close-up views of Cu site of H52A and (D) H54A/H58A mutants of in-silico-obtained azachalcone-TM1459 complex. 2Fo-Fc and anomalous maps contoured at 1.5 and 6.0 σ are shown in gray and magenta mesh, respectively.

To support this assumption, we prepared four H52A/C106X (A, D, N, S) variants and examined their stereoselectivity. The H52A/C106A mutant exhibited not only lower enantioselectivity but also lower yield, implying that Cys106 affects chemical reactivity as well as enantioselectivity (Table 1, entry 12). Furthermore, H52A/C106D and H52A/C106N still have excellent enantioselectivity, whereas H52A/C106S showed the lowest enantioselectivity, indicating that the spatial position of

Table 1: Addition of nitromethane to azachalcones 1 catalyzed by Cu-TM1459 and its mutants a, b

Entry TM1459 Variant Yield (%)c ee (%)

1 - 9.5 n. d.

2 Wild type 0.5 n. d.

3 H52A 91 99 (S)

4 H54A 2.7 n. d.

5 H58A 15 29 (S)

6 H92A 50 85 (S)

7 H52A/H54A 18 25 (R)

8 H52A/H58A 1.5 n. d.

9 H54A/H58A 83 89 (R)

10 H54A/H92A 23 16 (R)

11 H58A/H92A 54 55 (S)

12 H52A/C106A 31 34 (S)

13 H52A/C106S 14 54 (S)

14 H52A/C106D 47 90 (S)

15 H52A/C106N 14 85 (S)

16 I49W/H54A/H58A 45 81 (R)

17 H54A/H58A/F104W 62 97 (R)

aReaction conditions: TM1459 (0.3 mM), CuSO4 (0.3 mM), 1 (10 mM), CH3NO2 (100 equiv.), Potassium phosphate buffer (pH 6.5)/CH3CN (9:1), 20 °C, 3 h. bYields and enantiomeric excesses (ee) were determined by chiral HPLC analysis. cYields were calculated based on the total amount of (S) and (R).

NO

NO2

NO

+ CH3NO2

3 mol% CuSO43 mol%TM1459

Phosphate buffer (pH6.5),CH3CN, 20 °C1 2

3

the hydrogen-bonded atoms to nitromethane is important for enantioselectivity (Table 1, entries 13-15).

In the same manner, chiral properties of the substrate binding pockets of H54A/H58A mutant were explored. We determined the crystal structure of Cu-bound H54A/H58A mutants (1.23 Å resolution, Figure 2B, Figure S5 and Table S1-S3), whose Cu binding site has three water molecules binding to the Cu ion (major species; Cu(A), minor species; Cu(B), see Figure S5) in a facial binding mode, acquiring the N2O3 square pyramidal geometry. In the lowest energy structure docked with 1 (DGb: -6.7 (kcal/mol)), the b-Si-face was fully covered with the protein matrix, whereas the b-Re-face was exposed to the solvent (dotted circle, Figure 2D). This indicates that the Re-preferential attack of nitromethane occurs in H54A/H58A, which is consistent with the results of the catalytic reaction (Table 1, entry 9). Considering this result, two amino acids (Ile49 and Phe104) at the Si-face of azachalcone were chosen for mutation positions to strengthen the shielding effect with the protein matrix (Table 1, entry 16 and 17). As a result, H54A/H58A/F104W was found to give excellent ee of 97 % (R), which was much higher than that of H54A/H58A (89 % (R)) (Table 1, entry 17 and Figure S4).

By using mini ligand library thus obtained including the aforementioned second shell modification based on the molecular docking, the Michael addition reaction of nitroethane and nitropropane to 1 was attempted in the same manner as that of nitromethane (Table 2 and Table S6). In the absence of cupin proteins, diastereoselectivity was not clearly observed (Table 2, entry 1, 6). In contrast, in the case of nitroethane, the syn-selective product syn-3 was predominantly obtained in the presence of H52A and H54A/H58A mutants (d.r. 1:8.2 and 1:10, respectively) with excellent ee of 99 % (–) and 92 % (+), respectively (Table 2, entry 2, 3). Moreover, the second shell modification, F104W mutation, further enhanced

enantioselectivity. The H54A/H58A/F104W mutant showed improved ee from 92 % (+) to 96 % (+) and high diastereoselectivity (Table 2, entry 4). Interestingly, H52A/C106N showed slight anti-form preference, retaining excellent enantioselectivity (Table 2, entry 5), whereas H52A/C106D exhibited syn-form preference (Table S6, entry 15).

When using nitropropane, higher selectivity was observed and the syn-selective reaction proceeded giving syn-4 in the presence of H52A and H54A/H58A (d.r. 1:8.8 and 1:10, respectively, Table 2, entry 7, 8) with excellent yield and ee 99 % (–) and 95 % (+), respectively. As in the case of nitroethane, the second shell modification, F104W mutation, allowed H54A/H58A to increase ee from 95 % (+) to 99 % (+) (Table 2, entry 9). Furthermore, C106N mutation was found to cause the complete inversion of diastereo-preference to produce anti-form stereoisomers, maintaining the excellent enantioselectivity (d.r. 8.9:1, ee 99 % (–), Table 2, entry 10).

To the best of our knowledge, this study represents the first example of artificial metalloenzymes that can control anti- or syn-preference in the Michael addition reaction with 2-azachalcone with distinguished enantioselectivity and excellent TON (Table S7). The generation of the library-like mutant-series bearing various first coordination spheres and second shells in TM1459 allowed for the simultaneous optimization of the substrate binding site as well as the Cu binding site, producing efficient catalysts for the diastereo- and enantio-selective Michael addition reaction of nitroalkanes to an α, β-unsaturated ketone. Thus, the use of such a small cupin library system facilitates a promising approach for creating artificial metalloenzymes that can catalyze diastereo- and enantiodivergent reactions. Hopefully, this system based on the cupin superfamily protein will be applied various reactions that can be used in synthetic organic chemistry. ACKNOWLEDGEMENTS N.F. thanks JSPS, MEXT, Japan, (JSPS KAKENHI Grant Number JP15K13743, JP18K19151 and JSPS KAKENHI JP16H01025 and JP18H04270 in Precisely Designed Catalysts with Customized Scaffolding) and Nagase Science and Technology Foundation, and Japan Foundation for Applied Enzymology. S.I. thanks JST CREST (# JPMJCR16P1). We thank Dr. E. Yamashita, Dr. A. Higashiura, and Dr. A. Nakagawa of SPring-8 BL-44XU for their support with crystallographic data collection. AUTHOR INFORMATION 1Department of Material and Life Science, Graduate School of Engineering, Osaka University, 2-1 Yamada-oka, Suita, Osaka 565-0871, Japan 2Department of Applied Life Sciences, Graduate School of Life and Environmental Sciences, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-ku, Sakai-shi, Osaka 599-8531, Japan 3Institute for Protein Research, Osaka University, 3-2 Yamada-oka, Suita, Osaka 565-0871, Japan

Corresponding Authors

*fujieda@biochem.osakafu-u.ac.jp (NF) *shinobu@mls.eng.osaka-u.ac.jp (SI)

REFERENCES

(1) Desimoni, G.; Faita, G.; Quadrelli, P. Pyridine-2,6-Bis(Oxazolines), Helpful Ligands for Asymmetric Catalysts. Chem. Rev. 2003, 103 (8), 3119–3154. https://doi.org/10.1021/cr020004h.

(2) Desimoni, G.; Faita, G.; Jørgensen, K. A. C2-Symmetric Chiral

Table 2: Addition of nitroalkanes to azachalcones 1 catalyzed by selected Cu-TM1459 mutants. a, b

Entry -R TM1459 Variant

Yield (%)b,c

d.r. (anti:syn)

ee (%)d (anti/syn)

1 Me - 13 1:2.0 n.d.

2 Me H52A 100 1:8.2 syn, 99(–)

3 Me H54A/H58A 100 1:10 syn, 92(+)

4 Me H54A/H58A/F104W 60 1:10 syn, 96(+)

5 Me H52A/C106N 81 2.2:1 anti, 96(–)

6 Et - 34 1:2.5 n.d.

7 Et H52A 82 1:8.8 syn, 99(–)

8 Et H54A/H58A 92 1:10 syn, 95(+)

9 Et H54A/H58A/F104W 90 1:12 syn, 99(+)

10 Et H52A/C106N 59 8.9:1 anti, 99(–)

aReaction conditions: TM1459 (0.3 mM), CuSO4 (0.3 mM), 1 (10 mM), R-CH3NO2 (100 equiv.), Potassium phosphate buffer (pH 6.5)/CH3CN (9:1), 20 °C, 3 h (for nitroethane) or 12 h (for nitropropane). bYields, diastereomeric ratio (d.r.) and enantiomeric excesses (ee) were determined by chiral HPLC analysis. cYields were calculated based on the total amount of stereoisomers. d(+) or (-) was determined by polarimeter on HPLC.

NPh

O3 mol% CuSO43 mol%TM1459

Phosphate buffer (pH6.5),CH3CN, 20 °C

NO Ph

R

NO2

NO Ph

R

NO2+

+R-CH2NO2

1syn-3, R=Mesyn-4, R=Et

anti-3, R=Meanti-4, R=Et

4

Bis(Oxazoline) Ligands in Asymmetric Catalysis. Chem. Rev. 2006, 106 (9), 3561–3651. https://doi.org/10.1021/cr0505324.

(3) Zheng, K.; Liu, X.; Feng, X. Recent Advances in Metal-Catalyzed Asymmetric 1,4-Conjugate Addition (ACA) of Nonorganometallic Nucleophiles. Chem. Rev. 2018, 118 (16), 7586–7656. https://doi.org/10.1021/acs.chemrev.7b00692.

(4) Bos, J.; Roelfes, G. Artificial Metalloenzymes for Enantioselective Catalysis. Curr. Opin. Chem. Biol. 2014, 19, 135–143. https://doi.org/10.1016/j.cbpa.2014.02.002.

(5) Rioz-Martínez, A.; Roelfes, G. DNA-Based Hybrid Catalysis. Curr. Opin. Chem. Biol. 2015, 25, 80–87. https://doi.org/10.1016/j.cbpa.2014.12.033.

(6) Pordea, A. Metal-Binding Promiscuity in Artificial Metalloenzyme Design. Curr. Opin. Chem. Biol. 2015, 25, 124–132. https://doi.org/10.1016/j.cbpa.2014.12.035.

(7) Hoarau, M.; Hureau, C.; Gras, E.; Faller, P. Coordination Complexes and Biomolecules: A Wise Wedding for Catalysis Upgrade. Coord. Chem. Rev. 2016, 308, 445–459. https://doi.org/10.1016/j.ccr.2015.05.011.

(8) Lin, Y. W. Rational Design of Metalloenzymes: From Single to Multiple Active Sites. Coord. Chem. Rev. 2017, 336, 1–27. https://doi.org/10.1016/j.ccr.2017.01.001.

(9) Schwizer, F.; Okamoto, Y.; Heinisch, T.; Gu, Y.; Pellizzoni, M. M.; Lebrun, V.; Reuter, R.; Köhler, V.; Lewis, J. C.; Ward, T. R. Artificial Metalloenzymes: Reaction Scope and Optimization Strategies. Chem. Rev. 2018, 118 (1), 142–231. https://doi.org/10.1021/acs.chemrev.7b00014.

(10) Lewis, J. C. Beyond the Second Coordination Sphere: Engineering Dirhodium Artificial Metalloenzymes to Enable Protein Control of Transition Metal Catalysis. Acc. Chem. Res. 2019, 52 (3), 576–584. https://doi.org/10.1021/acs.accounts.8b00625.

(11) Natoli, S. N.; Hartwig, J. F. Noble-Metal Substitution in Hemoproteins: An Emerging Strategy for Abiological Catalysis. Acc. Chem. Res. 2019, 52, 326–335. https://doi.org/10.1021/acs.accounts.8b00586.

(12) Roelfes, G. LmrR: A Privileged Scaffold for Artificial Metalloenzymes. Acc. Chem. Res. 2019, 52 (3), 545–556. https://doi.org/10.1021/acs.accounts.9b00004.

(13) Churchfield, L. A.; Tezcan, F. A. Design and Construction of Functional Supramolecular Metalloprotein Assemblies. Acc. Chem. Res. 2019, 52 (2), 345–355. https://doi.org/10.1021/acs.accounts.8b00617.

(14) Roelfes, G.; Boersma, A. J.; Feringa, B. L. Highly Enantioselective DNA-Based Catalysis. Chem. Commun. 2006, No. 6, 635. https://doi.org/10.1039/b516552k.

(15) Ghattas, W.; Dubosclard, V.; Tachon, S.; Beaumet, M.; Guillot, R.; Réglier, M.; Simaan, A. J.; Mahy, J. CuII-Containing L-Aminocyclopropane Carboxylic Acid Oxidase Is an Efficient Stereospecific Diels-Alderase. Angew. Chemie Int. Ed. 2019, 1–6. https://doi.org/10.1002/anie.201909407.

(16) Coquière, D.; Feringa, B. L.; Roelfes, G. DNA-Based Catalytic Enantioselective Michael Reactions in Water. Angew. Chemie Int. Ed. 2007, 46 (48), 9308–9311. https://doi.org/10.1002/anie.200703459.

(17) Coquière, D.; Bos, J.; Beld, J.; Roelfes, G. Enantioselective Artificial Metalloenzymes Based on a Bovine Pancreatic Polypeptide Scaffold. Angew. Chemie Int. Ed. 2009, 48 (28), 5159–5162. https://doi.org/10.1002/anie.200901134.

(18) Boersma, A. J.; Megens, R. P.; Feringa, B. L.; Roelfes, G. DNA-Based

Asymmetric Catalysis. Chem. Soc. Rev. 2010, 39 (6), 2083. https://doi.org/10.1039/b811349c.

(19) Blay, G.; Incerti, C.; Muñoz, M. C.; Pedro, J. R. Enantioselective LaIII-PyBOX-Catalyzed Nitro-Michael Addition to (E)-2-Azachalcones. European J. Org. Chem. 2013, 2013 (9), 1696–1705. https://doi.org/10.1002/ejoc.201201579.

(20) Li, L.; Zhang, S.; Hu, Y.; Li, Y.; Li, C.; Zha, Z.; Wang, Z. Highly Diastereo- and Enantioselective Michael Addition of Nitroalkanes to 2-Enoyl-Pyridine N -Oxides Catalyzed by Scandium(III)/Copper(II) Complexes. Chem. A Eur. J. 2015, 21 (37), 12885–12888. https://doi.org/10.1002/chem.201502129.

(21) Fujieda, N.; Hasegawa, A.; Ishihama, K.; Itoh, S. Artificial Dicopper Oxidase: Rational Reprogramming of Bacterial Metallo-β-Lactamase into a Catechol Oxidase. Chem. Asian J. 2012, 7 (6), 1203–1207. https://doi.org/10.1002/asia.201101014.

(22) Fujieda, N.; Schätti, J.; Stuttfeld, E.; Ohkubo, K.; Maier, T.; Fukuzumi, S.; Ward, T. R. Enzyme Repurposing of a Hydrolase as an Emergent Peroxidase upon Metal Binding. Chem. Sci. 2015, 6 (7), 4060–4065. https://doi.org/10.1039/C5SC01065A.

(23) Fujieda, N.; Nakano, T.; Taniguchi, Y.; Ichihashi, H.; Sugimoto, H.; Morimoto, Y.; Nishikawa, Y.; Kurisu, G.; Itoh, S. A Well-Defined Osmium–Cupin Complex: Hyperstable Artificial Osmium Peroxygenase. J. Am. Chem. Soc. 2017, 139 (14), 5149–5155. https://doi.org/10.1021/jacs.7b00675.

(24) Hajnal, I.; Faber, K.; Schwab, H.; Hall, M.; Steiner, K. Oxidative Alkene Cleavage Catalysed by Manganese-Dependent Cupin TM1459 from Thermotoga Maritima. Adv. Synth. Catal. 2015, 357 (14–15), 3309–3316. https://doi.org/10.1002/adsc.201500608.

(25) Otto, S.; Bertoncin, F.; Engberts, J. B. F. N. Lewis Acid Catalysis of a Diels−Alder Reaction in Water. J. Am. Chem. Soc. 1996, 118 (33), 7702–7707. https://doi.org/10.1021/ja960318k.

download fileview on ChemRxivManuscript_191231_NF.pdf (2.13 MiB)

1

Supporting Information

for

Cupin Variants as Macromolecular Ligand Library

for Stereoselective Michael Addition of Nitroalkanes

Nobutaka Fujieda,1,2* Haruna Ichihashi,1 Miho Yuasa,2 Yosuke Nishikawa,3 Genji Kurisu,3 and

Shinobu Itoh1*

1Department of Material and Life Science, Graduate School of Engineering, Osaka University, 2-1

Yamada-oka, Suita, Osaka 565-0871, Japan, 2 Department of Applied Life Sciences, Graduate

School of Life and Environmental Sciences, Osaka Prefecture University, 1-1 Gakuen-cho, Naka-

ku, Sakai-shi, Osaka 599-8531, Japan, 3Institute for Protein Research, Osaka University, 3-2

Yamada-oka, Suita, Osaka 565-0871, Japan.

2

Experimental

General: The reagents and the solvents used in this study are commercially available and were used without

further purification. Proteins were desalted and purified by using the C18 resin packed Zip tip (Millipore). A

saturated α-cyano-4-hydroxycinnamic acid solution was used as a laser-absorption matrix. MALDI-MS

measurements were performed on a linear mode of time-of-flight mass spectrometer (Autoflex III, Bruker

Daltonics). ESI-Mass spectra were obtained on an Agilent LC/MSD (G6125B) system. NMR spectra were

recorded at 500 MHz and tetramethylsilane and CDCl3 was used as internal standard for 1H and at 125 MHz for 13C nuclei, respectively. The chemical shifts (δ) and coupling constants (J) were expressed in ppm and Hz,

respectively. Chiral HPLC analyses were performed in a chromatograph equipped with a UV diode-array

detector and polarimetric detector (OR-2090, JASCO Co.) using chiral stationary columns from YMC Co., Ltd..

Retention times are given in min.

Plasmids, Site-directed Mutagenesis, and Expression and Purification of Proteins: For expression of

TM1459 protein with Strep-tag, the pET30-based expression constructs were used as previously reported.1

Oligonucleotide-directed mutagenesis experiments for introducing site-directed mutation were performed on the

expression vector that contained the DNA fragment of TM1459. The pair of approximately 30 nucleotide long

mutagenic oligonucleotides was purchased from Macrogen Japan Co (Table S4). The site-directed mutagenesis

was carried out by inverse PCR, followed by DpnI digestion and self-ligation. For the mutant, the absence of

undesired mutations on the gene was confirmed by DNA sequencing using 3730xl DNA Analyzer (Applied

Biosystems) after the construction of the vector. The protein purification of wild type TM1459 and mutants

thereof was conducted as described previously.1 The fractions containing the TM1459 were collected and

concentrated by ultrafiltration using Vivaspin turbo 15. The protein samples were stored at –80 °C until use.

Purity was determined by SDS-PAGE (Figure S2). Protein concentration was determined by measuring the

intensity of absorbance at 280 nm.2

Crystallization: The purified apo-TM1459 protein (metal-depleted form) solution was desalted and exchanged

to 10 mM HEPES buffer (pH 7.0) using PD-10 column (GE Healthcare). Then, an aliquot of an aqueous solution

of CuSO4 (300 µM, 1.5 equiv. for protein) was added to 200 µM apo-TM1459 protein solution under N2. After

incubation at 20 °C for 16 h, the resultant solution was loaded on PD-10 columns pre-equilibrated with 10 mM

HEPES buffer (pH 7.0). The loaded sample was then eluted by the same buffer solution in order to remove excess

amount of the metal. The collected fractions were concentrated up to 20 mg/mL by ultrafiltration using Vivaspin

turbo 15 and restored −80 °C until use. The metal content was determined by an ICP-AES on ICPS-8100

3

apparatus (Shimadzu Co.). Calibration curves of the metal ion used in this study were made by using the metal

ion standard solution (10 ppm). Cu-bound TM1459 were crystallized by hanging drop vapor diffusion method

as previously described with some modification.1 Crystallization drops were prepared by mixing the 0.1 M MES

buffer at pH 6.0 containing 25 % w/v Jeffamine ED-2001 as precipitants (1 µL) and an aliquot of protein solution

(2 µL). Crystals were obtained at 20 °C within 1 week. For cryo-protection, thus obtained crystals were soaked

in the solution containing 40% w/v Jeffamine ED-2001 and 0.1 M MES at pH 6.0 for overnight.

Data Collection of X-Ray Diffraction and Structure Determination: All the diffraction data were collected

at the SPring-8 synchrotron facility (BL44XU, Harima, Hyogo, Japan). X-ray diffraction images were collected

using MX300HE CCD detectors equipped with a cryo-system at 100 K using a wavelength of 0.90000 Å. The

data were processed and scaled with the HKL-2000 program package. The data collection and refinement

statistics are summarized in Table S1. The crystal structures were solved by molecular replacement with the

Phaser program3 using the reported coordinate [apo-TM1459 (PDB code: 5WSD)] as a search model. The

structure models were refined by the program phenix.refine,4 manually rebuilding with the program Coot.5 This

procedure was iterated until the model did not further improve. Then, the structure model was further refined by

SHELXL software (version 2017/1),6 manually rebuilding with Coot. At this stage, anisotropic temperature

factors were adapted to all the atoms except the hydrogen atoms and calculated with RIGU7 and SIMU

instructions. The restraint parameters of DFIX, DANG, and FLAT instructions for the MES and PEG were

generated by using Grade Web Server (Global Phasing Ltd.). A subset of 5% of the reflections was separated to

monitor the free R factor (Rfree). In all the crystal structures, most of the residues were well defined in the final

electron density, except for some disordered loops and amino acid residues in the flexible regions (Table S2).

The distances and angles among the copper ions and their ligand atoms were completely unrestrained during

structure refinement because the resolution and the data-to-parameter ratio were high enough. The bond lengths

and angles of the metal-coordinating atoms to each metal including their estimated errors (for example, Cu–O

(water) and Cu-imidazole Ne distances) are summarized in Table S3. In the structures, Ramachandran analysis

with the Molprobity program8 showed no outliners. The atomic coordinates of the final models and experimental

structure factors have been deposited to the worldwide PDB and are accessible under the code described in Table

S1. Secondary structures were assigned by the DSSP program9 and all figures of protein structures were prepared

using PyMOL (version 2.1, Schrödinger LLC).

Substrate Docking Simulation: The azachalcone was docked into cavity around Cu using AutoDock Vina10

and PDBQT molecular structure files were made using AutoDock Tools.11 The crystal structure coordinates of

chain A of Cu-bound H52A mutant (PDB code: 6L2E), Cu-bound H54A/H58A mutant (PDB code: 6L2F) were

used in the substrate-protein docking studies. Water molecules were deleted, and hydrogens were added to the

Cu-bound H52A and Cu-bound H54A/H58A structures before creating PDBQT files. Kollman all-atom charges

4

for protein and Gasteiger−Hükel charges for the azachalcone were computed. The 2+ charge of the Cu ion was

added to AutoDock Vina’s atomic parameters. The side chains of amino acids sitting around copper binding site

(Lys24, Arg39, Phe41, Ile49, Trp56, His58, Phe94, Phe104, Cys106 (sulfinic acid), and Ile108) were allowed to

rotate during docking and all degrees of freedom were possible for the ligand. The search space was included in

a 20 Å × 20 Å × 20 Å box to cover the complete binding site of the ligand around bound copper. At least 20

docking models for each mutant were generated and analyzed on the basis of scoring values of AutoDock’s

scoring function. PyMOL software was used to visualize the docking results.The pose with the pyridine moiety

and/or keto moiety pointing to Cu was selected.

Catalytic Michael Addition of Nitroalkanes: A vial (2 mL volume) was filled with 0.3 mL reaction solution,

which is successively mixed with an aqueous solution of CuSO4 (0.09 µmol, 60 µL), CH3CN solution containing

2-azachalcone (3 µmol, 30 µL), apo-TM1459 (0.09 µmol) solution (20 mM phosphate buffer, pH 6.5) containing

nitroalkane (final 1 M). The mixture was stirred (600 rpm) in the closed vial at 20 °C for optimized hours (3, 3,

and 12 h for nitromethane, nitroethane, and nitropropane, respectively). After the reaction, 4-methoxy-1-naphthol

was added to the mixture as an internal standard. Then, the product was isolated by extraction with hexane/iPrOH

(v/v = 95/5). After drying (Na2SO4), the crude product was analyzed by HPLC equipped with a normal-phase

chiral column [CHIRAL ART Amylose-SA, CHIRAL ART Amylose-SB, and CHIRAL ART Cellulose-SC (5

µm, 4.6 ´ 250 mm, YMC Co., Ltd.)] via comparison with authentic standards to determine the yields,

enantiomeric excess (ee) and diastereomeric ratio (d.r.). The yields were calculated from a calibration curve: a

plot of mole ratio (moles of organic compound / moles of internal standard) versus area ratio (area of organic

compound/area of standard).

Synthesis of (E)-3-phenyl-1-(pyridin-2-yl)prop-2-en-1-one (1): The substrate, 1,2-unsaturated ketones 1, were

readily prepared from aldol condensation reaction according to modified reported procedures.12 1H NMR (500

MHz, CDCl3): δ = 8.76-8.74 (m, 1 H), 8.32 (d, J = 16 Hz, 1 H), 8.21-8.19 (m, 1 H), 7.95 (d, J = 16 Hz, 1 H),

7.88 (m, 1 H), 7.75-7.73 (m, 2 H), 7.50-7.41 (m, 4 H) ppm. 13C NMR (125 MHz, CDCl3): δ = 189.5, 154.2,

148.9, 144.8, 137.0, 135.1, 130.6, 128.9, 128.8, 126.9, 122.9, 120.8 ppm. MS (+ESI): m/z calcd. for C14H12NO

[M + H] 210.2; found 210.2; consistent with the 1H and 13C NMR, and mass spectroscopic data previously

reported for this compound.13,14

Synthesis of 4-Nitro-3-phenyl-1-(pyridin-2-yl)butan-1-one (2): Compound 1 for authentic standard (racemic

product) was synthesized according to reported procedures.15 Purified by column chromatography

(hexane/EtOAc, 80:20). Retention time on HPLC: tR = 27.6 (R-(+)), 30.6 (S-(–)) min (CHIRAL ART Amylose-

5

SA; hexane/iPrOH, 98.5:1.5; 1 mL/min); 1H NMR (500 MHz, CDCl3): δ = 8.67-8.59 (m, 1 H), 7.99 (dt, J = 8.0,

1.3 Hz, 1 H), 7.83 (td, J = 8.0, 2.0 Hz, 1 H), 7.49 (ddd, J = 7.5, 5.0, 1.3 Hz, 1 H), 7.32-7.21 (m, 5 H), 4.79 (dd, J

= 12.5, 7.0 Hz, 1 H), 4.68 (dd, J = 12.3, 8.3 Hz, 1 H), 4.28-4.23 (m, 1 H), 3.84 (dd, J = 18.3, 6.8 Hz, 1 H), 3.63

(dd, J = 18.3, 7.3 Hz, 1 H) ppm. 13C NMR (125 MHz, CDCl3): 198.7, 152.8, 149.1, 139.3, 137.1, 129.0, 127.8,

127.7, 127.7, 122.0, 80.0, 40.9, 39.3 ppm. MS (+ESI): m/z calcd. for C15H15N2O3 [M + H] 271.1; found 271.1;

consistent with the 1H and 13C NMR, and mass spectroscopic data previously reported for this compound.15

Synthesis of 4-Nitro-3-phenyl-1-(pyridin-2-yl)pentan-1-one (3): Compound 3 for authentic standard

(racemic product) was synthesized according to reported procedures.15 Retention time on HPLC: tR = 29.7 (+)-

syn, 32.3 (+)-anti, 36.1 (–)-anti, and 40.1 (–)-syn min (CHIRAL ART Amylose-SC; hexane/iPrOH, 98.5:1.5;

1 mL/min); Diastereomers were purified by column chromatography (hexane/EtOAc, 80:20). Syn-isomer: 1H

NMR (500 MHz, CDCl3): δ = 8.67-8.66 (m, 1 H), 7.93 (dt, J = 8.0, 1.0 Hz, 1 H), 7.77 (td, J = 7.5, 1.8 Hz, 1

H), 7.45 (ddd, J = 7.6, 4.9, 1.1 Hz, 1 H), 7.29-7.18 (m, 5 H), 4.96 (quint., J = 6.9 Hz, 1 H), 4.05-4.00 (m, 1 H),

3.88 (dd, J = 18, 8.5 Hz, 1H), 3.68 (dd, J = 18, 5.5 Hz, 1 H), 1.60 (d, J = 7.0 Hz, 3 H) ppm. 13C NMR (125

MHz, CDCl3): δ = 198.2, 152.6, 148.8, 138.4, 136.8, 128.7, 128.3, 127.5, 127.3, 121.7, 87.2, 45.1, 40.4, 17.6

ppm. MS (ESI): m/z calcd. for C16H17N2O3 [M + H] 285.1; found 285.1. Anti-isomer: 1H NMR (500 MHz,

CDCl3): δ = 8.67-8.66 (m, 1 H), 7.93 (dt, J = 8.0, 1.0 Hz, 1 H), 7.77 (td, J = 7.5, 1.8 Hz, 1 H), 7.45 (ddd, J =

7.6, 4.9, 1.1 Hz, 1 H),7.29-7.18 (m, 5 H), 4.96 (quint., J = 6.9 Hz, 1 H), 4.05-4.00 (m, 1 H), 3.88 (dd, J = 18,

8.5 Hz, 1H), 3.68 (dd, J = 18, 5.5 Hz, 1 H), 1.60 (d, J = 7.0 Hz, 3 H) ppm. 13C NMR (125 MHz, CDCl3): δ =

198.6, 152.7, 148.9, 138.6, 136.9, 128.5, 128.1, 127.6, 127.4, 121.8, 87.0, 44.7, 38.8, 16.8 ppm. MS (+ESI):

m/z calcd. for C16H17N2O3 [M + H] 285.1; found 285.1; consistent with the 1H and 13C NMR, and mass

spectroscopic data previously reported for these compounds.15

Synthesis of 4-Nitro-3-phenyl-1-(pyridin-2-yl)hexan-1-one (4): Compound 4 for authentic standard (racemic

product) was synthesized according to reported procedures.15 Retention time on HPLC: tR = 13.0 (+)-anti, 13.8

(–)-anti, 16.6 (–)-syn, and 17.3 (+)-syn (CHIRAL ART Amylose-SB; hexane/iPrOH, 99.0:1.0; 1 mL/min);

Diastereomers were purified by column chromatography (hexane/EtOAc, 80:20). Syn-isomer: 1H NMR (500

MHz, CDCl3): δ = 8.68-8.66 (m, 1 H), 7.94 (dt, J = 7.5, 1.1 Hz, 1 H), 7.80 (td, J = 7.8, 1.5 Hz, 1 H), 7.47 (ddd,

J = 7.5, 4.5, 1.0 Hz, 1 H), 7.28-7.19 (m, 5 H), 4.80 (ddd, J = 11.0, 8.0, 3.3 Hz, 1 H), 4.00 (dt, J = 8.5, 5.8 Hz, 1

H), 3.82 (dd, J = 18.3, 8.3 Hz, 1 H), 3.71 (dd, J = 18.0, 6.0 Hz, 1 H), 2.10-2.00 (m, 1 H), 1.97-1.89 (m, 1 H),

0.99 (t, J = 7.3 Hz, 3 H). 13C NMR (125 MHz, CDCl3): δ = 198.6, 152.8, 148.9, 138.8, 137.0, 128.9, 128.5, 127.9,

127.2, 121.9, 93.9, 43.8, 39.6, 24.7, 10.5 ppm. MS (ESI): m/z calcd. for C17H19N2O3 [M + H] 299.1; found 299.1.

Anti-isomer: 1H NMR (500 MHz, CDCl3): δ = 8.65-8.63 (m, 1 H), 7.86 (dt, J = 8.0, 1.0 Hz, 1 H), 7.75 (td, J =

7.8, 2.0 Hz, 1 H), 7.43 (ddd, J = 7.5, 5.0, 1.4 Hz, 1 H), 7.31-7.20 (m, 5 H), 4.71 (td, J = 10.5, 3.0 Hz, 1 H), 4.05

(dd, J = 17, 10.5 Hz, 1 H), 3.97 (td, J = 10.1, 3.3 Hz, 1 H), 3.34 (dd, J = 17.5, 3.0 Hz, 1 H), 1.91-1.81 (m, 1 H),

6

1.56-1.48 (m, 1 H), 0.87 (t, J = 7.5 Hz, 3 H) ppm. 125 MHz, CDCl3: δ = 198.3, 152.8, 148.9, 139.0, 136.8, 129.2,

128.6, 128.1, 127.1, 121.8, 94.3, 44.5, 40.3, 25.5, 10.3 ppm. MS (+ESI): m/z calcd. for C17H19N2O3 [M + H]

299.1; found 299.1; consistent with the 1H and 13C NMR, and mass spectroscopic data previously reported for

these compounds.15

7

Supporting References

(1) Fujieda, N.; Nakano, T.; Taniguchi, Y.; Ichihashi, H.; Sugimoto, H.; Morimoto, Y.; Nishikawa, Y.;

Kurisu, G.; Itoh, S. A Well-Defined Osmium–Cupin Complex: Hyperstable Artificial Osmium

Peroxygenase. J. Am. Chem. Soc. 2017, 139 (14), 5149–5155. https://doi.org/10.1021/jacs.7b00675.

(2) Gill, S. C.; von Hippel, P. H. Calculation of Protein Extinction Coefficients from Amino Acid Sequence

Data. Anal. Biochem. 1989, 182 (2), 319–326. https://doi.org/10.1016/0003-2697(89)90602-7.

(3) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.; Read, R. J. Phaser

Crystallographic Software. J. Appl. Crystallogr. 2007, 40 (4), 658–674.

https://doi.org/10.1107/S0021889807021206.

(4) Adams, P. D.; Afonine, P. V.; Bunkóczi, G.; Chen, V. B.; Davis, I. W.; Echols, N.; Headd, J. J.; Hung,

L.-W.; Kapral, G. J.; Grosse-Kunstleve, R. W.; et al. PHENIX : A Comprehensive Python-Based

System for Macromolecular Structure Solution. Acta Crystallogr. Sect. D 2010, 66 (2), 213–221.

https://doi.org/10.1107/S0907444909052925.

(5) Emsley, P.; Cowtan, K. Coot: Model-Building Tools for Molecular Graphics. Acta Crystallogr. Sect.

D 2004, 60 (12), 2126–2132. https://doi.org/10.1107/S0907444904019158.

(6) Sheldrick, G. M. A Short History of SHELX. Acta Crystallogr. Sect. A 2008, 64 (1), 112–122.

https://doi.org/10.1107/S0108767307043930.

(7) Thorn, A.; Dittrich, B.; Sheldrick, G. M. Enhanced Rigid-Bond Restraints. Acta Crystallogr. Sect. A

2012, 68 (4), 448–451. https://doi.org/10.1107/S0108767312014535.

(8) Chen, V. B.; Arendall, W. B.; Headd, J. J.; Keedy, D. A.; Immormino, R. M.; Kapral, G. J.; Murray, L.

W.; Richardson, J. S.; Richardson, D. C. MolProbity : All-Atom Structure Validation for

Macromolecular Crystallography. Acta Crystallogr. Sect. D 2010, 66 (1), 12–21.

https://doi.org/10.1107/S0907444909042073.

(9) Kabsch, W.; Sander, C. Dictionary of Protein Secondary Structure: Pattern Recognition of Hydrogen-

Bonded and Geometrical Features. Biopolymers 1983, 22 (12), 2577–2637.

https://doi.org/10.1002/bip.360221211.

(10) Trott, O.; Olson, A. J. AutoDock Vina: Improving the Speed and Accuracy of Docking with a New

Scoring Function, Efficient Optimization, and Multithreading. J. Comput. Chem. 2009, 31 (2),

455−461. https://doi.org/10.1002/jcc.21334.

(11) Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D. S.; Olson, A. J.

AutoDock4 and AutoDockTools4: Automated Docking with Selective Receptor Flexibility. J. Comput.

Chem. 2009, 30 (16), 2785–2791. https://doi.org/10.1002/jcc.21256.

(12) Otto, S.; Bertoncin, F.; Engberts, J. B. F. N. Lewis Acid Catalysis of a Diels−Alder Reaction in Water.

J. Am. Chem. Soc. 1996, 118 (33), 7702–7707. https://doi.org/10.1021/ja960318k.

(13) Gase, R. A.; Pandit, U. K. Reduced Nicotinamide Adenine Dinucleotide (NADH) Models. 14. Metal

8

Ion Catalysis of the Reduction of a,b-Unsaturated Ketones by 1,4-Dihydropyridines. A Model of D4-

3-Ketosteroid Reductases. J. Am. Chem. Soc. 1979, 101 (23), 7059–7064.

https://doi.org/10.1021/ja00517a047.

(14) Ciupa, A.; Mahon, M. F.; De Bank, P. A.; Caggiano, L. Simple Pyrazoline and Pyrazole “Turn on”

Fluorescent Sensors Selective for Cd2+ and Zn2+ in MeCN. Org. Biomol. Chem. 2012, 10 (44), 8753.

https://doi.org/10.1039/c2ob26608c.

(15) Blay, G.; Incerti, C.; Muñoz, M. C.; Pedro, J. R. Enantioselective LaIII-PyBOX-Catalyzed Nitro-

Michael Addition to (E)-2-Azachalcones. European J. Org. Chem. 2013, 2013 (9), 1696–1705.

https://doi.org/10.1002/ejoc.201201579.

9

Supporting Notes

This work was performed in part using a synchrotron beamline BL44XU at SPring-8 under the Cooperative

Research Program of Institute for Protein Research, Osaka University. Diffraction data were collected at the

Osaka University beamline BL44XU at SPring-8 (Harima, Japan) under the Proposal Numbers 2016A6500,

and 2016B6500.

10

Table S1. Data collection and refinement statistics.a

Wild type

(PDB code: 6L2D)

H52A mutant

(PDB code: 6L2E)

H54A/H58A mutant

(PDB code: 6L2F)

Dataset

X-ray source Spring-8 BL44XU Spring-8 BL44XU Spring-8 BL44XU

Space group P212121 P212121 P212121

Unit cell

a = 50.52 Å, b = 57.81 Å, c = 74.96 Å, α = β = γ = 90˚

a = 50.63 Å, b = 57.76 Å, c = 75.04 Å, α = β = γ = 90˚

a = 50.64 Å, b = 57.80 Å, c = 74.40 Å, α = β = γ = 90˚

Wavelength 0.90000 Å 0.90000 Å 0.90000 Å

Resolution 50 to 1.20 Å 50 to 1.20 Å 50 to 1.23 Å

No. reflection 497872 / 67624 501012 / 69368 462273 / 63984

Redundancyb 7.4 (7.4) 7.2 (7.1) 7.2 (7.1)

Completenessb 97.2% (95.1%) 99.9% (99.9%) 100% (99.9%)

Rmergeb 8.2% (97.6%) 7.2% (83.1%) 7.6% (97.1%)

I/σb 22.1 (2.2) 26.2 (2.9) 29.4 (3.3)

Refinement

Rwork/Rfree 0.1508/0.1857 0.1489/0.1786 0.1460/0.1881

No. of protein/solvent atomsc 1857/278 1841/294 1837/226

No. of metal ion atoms 2 3 4

B-factors of protein/solvent 16.9/33.0 14.0/19.5 14.6/30.4

B-factors of metal ions 13.20 13.05 21.7

r.m.s.d. bond/angle 0.007/1.65 0.007/1.71 0.007/1.65

Ramchandran favored/allowedd 99.1%/0.9% 98.7%/1.3% 99.1%/0.9% aA single crystal was used for all of structures. bValues in parentheses are for highest-resolution shell. cHydrogen atoms not included. dValues are calculated by Molprobity.

11

Table S2. Missing residues and atoms.

Wild type (PDB code: 6L2D)

Missing residues

Chain B, Gly(-3)–Gly(0)

Missing atoms

Chain A, Arg20 (Cg, Cd, Ne, Cz, Nh1, Nh2), Lys66 (Cd, Ce, Nz), Gln73 (Cg, Cd, Oe1, Ne2)

Chain B, Tyr7 (Cg, Od1, Ce1, Od2, Ce2, Cz, Oh), Asp17 (Cg, Od1, Od2), Lys18 (Cg, Cd, Ce, Nz), Lys31 (Cg, Cd, Ce, Nz), Lys66 (Cg, Cd, Ce, Nz),

Glu79 (Cg, Cd, Oe1, Oe2)

H52A (PDB code: 6L2E)

Missing residues

Chain B, Gly(-3)–Gly(0)

Missing atoms

Chain A, Gln12 (Cg, Cd, Oe1, Ne2), Arg20 (Cg, Cd, Ne, Cz, Nh1, Nh2), Lys66 (Cg, Cd, Ce, Nz), Gln73 (Cg, Cd, Oe1, Ne2)

Chain B, Asp17 (Cg, Od1, Od2), Lys18 (Cg, Cd, Ce, Nz), Lys31 (Cg, Cd, Ce, Nz), Lys66 (Cg, Cd, Ce, Nz), Glu79 (Cd, Oe1, Oe2)

H54A/H58A (PDB code: 6L2F)

Missing residues

Chain B, Gly(-3)–Gly(0)

Missing atoms

Chain A, Gln12 (Cg, Cd, Oe1, Ne2), Asp17 (Cg, Od1, Od2), Lys18 (Cg, Cd, Ce, Nz), Arg20 (Cg, Cd, Ne, Cz, Nh1, Nh2)

Chain B, Lys18 (Cg, Cd, Ce, Nz), Lys31 (Cg, Cd, Ce, Nz), Lys66 (Cd, Ce, Nz), Glu79 (Cg, Cd, Oe1, Oe2)

12

Table S3. Interatomic distances and bond angles in the crystal structures (Cu ion and coordinating

atoms)

Wild type (PDB code: 6L2D) Chain A Bond Lengthsa

His52(Ne) His54(Ne) His58(Ne) His92(Ne) O1 O2 Cu 2.04 (0.02) 2.14 (0.02) 1.98 (0.02) 2.07 (0.02) 1.86 (0.02) 2.76 (0.08)

Angles[a] Cu His52(Ne) His54(Ne) His58(Ne) His92(Ne) O1 O2

His52(Ne)

His54(Ne) 82.1 (1.1)

His58(Ne) 165.3 (0.8) 91.7 (0.7)

His92(Ne) 102.5 (0.8) 103.5 (0.7) 92.8 (0.7)

O1 84.5 (0.9) 96.8 (1.3) 90.5 (1.0) 159.3 (1.2)

O2 82.1 (1.1) 153.1 (1.6) 83.6 (1.2) 103.2 (1.7) 56.9 (2.0)

Chain B Bond Lengths[a]

His52(Ne) His54(Ne) His58(Ne) His92(Ne) O1 Cu 1.99 (0.02) 2.22 (0.02) 2.03 (0.02) 2.03 (0.02) 2.05 (0.03)

Angles[a]

Cu His52(Ne) His54(Ne) His58(Ne) His92(Ne) O1 His52(Ne) His54(Ne) 102.1 (0.8) His58(Ne) 166.0 (0.8) 91.1 (0.7) His92(Ne) 89.3 (0.7) 103.0 (0.8) 92.1 (0. 7)

O1 88.6 (1.1) 86.9 (1.3) 87.7 (1.1) 170.2 (1.3) aDistances and angles are given in Å and degrees (°), estimated errors are given in parentheses.

13

H52A (PDB code: 6L2E) Chain A Bond Lengthsa

His54(Ne) His58(Ne) His92(Ne) O1 O2 Cu 2.04 (0.02) 2.02 (0.02) 2.06 (0.02) 2.01 (0.03) 2.00 (0.03)

Angles[a]

Cu His54(Ne) His58(Ne) His92(Ne) O1 O2 His54(Ne) His58(Ne) 95.7 (0.8) His92(Ne) 101.8 (0.6) 99.2 (0.7)

O1 93.0 (1.2) 169.3 (1.4) 85.2 (1.1) O2 150.2 (1.7) 99.8 (1.5) 100.7 (1.4) 69.6 (1.7)

Chain B Bond Lengths[a]

His54(Ne) His58(Ne) His92(Ne) O1 O3 Cu 2.14 (0.02) 2.04 (0.02) 2.01 (0.02) 1.89 (0.03) 2.11 (0.03)

Angles[a]

Cu His54(Ne) His58(Ne) His92(Ne) O1 O3 His54(Ne) His58(Ne) 96.6 (0.8) His92(Ne) 98.7 (0.7) 98.2 (0.7)

O1 108.4 (1.7) 149.0 (2.0) 95.9 (1.3) O3 93.0 (1.1) 92.1 (1.4) 163.4 (1.1) 69.1 (1.8)

aDistances and angles are given in Å and degrees (°), estimated errors are given in parentheses.

14

H54A/H58A (PDB code: 6L2F) Chain A Bond Lengtha

His52(Ne) His92(Ne) O1 O2 O3 Cu(major) 2.00 (0.03) 2.13 (0.03) 2.22 (0.04) 2.42 (0.03) 1.80 (0.03) Cu(minor) 1.95 (0.05) 2.52 (0.07) 1.65 (0.05)

Angles[a]

Cu(major) His52(Ne) His92(Ne) O1 O2 O3 His52(Ne) His92(Ne) 96.9 (1.0)

O1 161.0 (1.4) 83.6 (1.3) O2 96.4 (1.1) 96.0 (1.0) 102.5 (1.3) O3 91.8 (1.6) 139.4 (1.9) 76.2 (1.8) 122.4 (1.7)

Cu(minor) His52(Ne) His92(Ne) O1 O2 O3 His92(Ne) 67.4 (1.7)

O1 91.3 (1.8) 79.8 (2.3) O2 77.0 (2.2) 140.3 (3.5) 119.2 (2.8)

Chain B Bond Lengths[a]

His52(Ne) His92(Ne) O1 O2 O3 Cu(major) 1.96 (0.03) 2.16 (0.03) 2.09 (0.03) 2.37 (0.03) 2.14 (0.04) Cu(minor) 1.83 (0.03) 2.33 (0.06) 1.88 (0.04)

Angles[a]

Cu(major) His52(Ne) His92(Ne) O1 O2 O3 His52(Ne) His92(Ne) 95.1 (1.0)

O1 158.5 (1.9) 93.6 (1.4) O2 93.5 (1.0) 100.7 (1.1) 104.1 (1.8) O3 95.2 (1.7) 119.9 (2.2) 63.4 (2.1) 137.3 (2.0)

Cu(minor) His52(Ne) His92(Ne) O1 O2 O3 His92(Ne) 74.0 (1.4)

O1 95.9 (1.8) 95.6 (2.0) O2 77.5 (1.5) 141.3 (2.4) 113.1 (2.1)

aDistances and angles are given in Å and degrees (°), estimated errors are given in parentheses.

15

Table S4. Oligonucleotides Used for Site-directed Mutagenesis of TM1459.a

No. Sequence (5'-3') Length Use

1 AGCCATCCGTGGGAACATGAAATTTTC 27 Forward primer for H52X, H54X

2 CGGATGGCTTGCGCGATCAATCAGACCACCCGGTT 27 Reverse primer for H52A

3 CGGTGCGCTATGGCGATCAATCAGACCACCCGGTT 35 Reverse primer for H54A

4 TGGGAAGCAGAAATTTTCGTGCTGAAAGGCAAAC 34 Forward primer for H58A

5 CGGATGGCTATGGCGATCAATCAGACCACCCGGTT 35 Reverse primer for H58X

6 GGCTTTCGCAACGATACCGATAGCGAAG 28 Forward primer for H92X

7 TGCGATTTCATTCGGTTCCACGAAGATG 28 Reverse primer for H92A

8 CGGTGCGCTTGCGCGATCAATCAGACCACCCGGTT 35 Reverse primer for H52A/H54A

9 TGGGAACATGAAATTTTCGTGCTGAAAGGCAAAC 34 Forward primer for I49X

10 CGGCGCGCTATGGCGATCCCACAGACCACCCGG 33 Reverse primer for I49W/H54A

11 CTGTGTCTGATTCCGAAAGAAGGCGGAG 28 Forward primer for F104X

12 CCATTCCACTTCGCTATCGGTATCGTTGC 29 Reverse primer for F104W

13 CTGATTCCGAAAGAAGGCGGAGAATG 26 Forward primer for C106X

14 TGCCAGAAATTCCACTTCGCTATCGGTATC 30 Reverse primer for C106A

15 ATCCAGAAATTCCACTTCGCTATCGGTATC 30 Reverse primer for C106D

16 ATTCAGAAATTCCACTTCGCTATCGGTATC 30 Reverse primer for C106N

17 GATTCCGAAAGAAGGCGGAGAATG 24 Forward primer for C106S

18 AGGCTCAGAAATTCCACTTCGCTATCGGTA 30 Reverse primer for C106S

aThe mutagenic nucleotides are underlined.

16

Table S5. Screening transition metals with the H52A mutant in the Michael addition. a, b

Entry Metal Yeild (%)[c] ee (%)[d]

1 MnSO4 1.1 n.d.

2 FeSO4 1.2 n.d.

3 CoSO4 0.9 n.d.

4 NiSO4 1.8 n.d.

5 CuSO4 91 99 (S)

6 ZnSO4 4.7 n.d. aReaction conditions: TM1459 (0.3 mM), meltal(II) sulfate (0.3 mM), 1 (10 mM), CH3NO2 (100 equiv.), Potassium phosphate buffer (pH 6.5)/CH3CN (9:1), 20 °C, 3 h. bYields and enantiomeric excesses (ee) were determined by chiral HPLC analysis. cYields were calculated based on the total amount of (S) and (R).

NO

NO2

NO

+ CH3NO2

3 mol% MSO43 mol% H52A

Phosphate buffer (pH6.5),CH3CN, 20 °C1 2

17

Table S6. Addition of nitroalkanes to azachalcones catalyzed by Cu-TM1459 and its mutants. a, b

Entry -R TM1459

Variant Yield (%)[b,c]

d.r. (anti:syn)

ee (%)[d] (anti/syn)

1 Me - 13 1:2.0 n.d. 2 Me WT 4.2 1:2.5 25(–)/24(–)

3 Me H52A 100 1:8.2 88(–)/99(–)

4 Me H54A 14 1:3.5 16(–)/55(–)

5 Me H58A 12 1:3.3 1.5(–)/3.4(–)

6 Me H92A 54 1:4.6 25(+)/80(–)

7 Me H52A/H54A 44 1:4.0 7.4(–)/49(+)

8 Me H52A/H58A 23 1:2.7 24(–)/40(+)

9 Me H54A/H58A 100 1:10 78(+)/92(+)

10 Me H54A/H92A 36 1:7.4 0.4(+)/38(+)

11 Me H58A/H92A 68 1:3.0 6.2(+)/31(–)

12 Me H52A/C106A 73 1.6:1 58(–)/45(–)

13 Me H52A/C106S 30 1.7:1 68(–)/61(–)

14 Me H52A/C106D 100 1:1.8 68(–)/89(–)

15 Me H52A/C106N 81 2.2:1 96(–)/93(–)

16 Me I49W/H54A/H58A 84 1:8.4 6(+)/85(+)

17 Me H54A/H58A/F104W 60 1:10 79(+)/96(+)

18 Et - 34 1:2.5 n.d.

19 Et WT 2.1 1:2.0 n.d.

20 Et H52A 82 1:8.8 93(–)/99(–)

21 Et H54A 31 1:5.3 37(–)/19(–)

22 Et H58A 23 1:3.4 14(–)/4.9(+)

23 Et H92A 21 1:3.3 23(+)/75(–)

24 Et H52A/H54A 48 1:1.0 74(–)/53(+)

25 Et H52A/H58A 75 1:4.0 68(–)/75(+)

26 Et H54A/H58A 92 1:10 95(+)/95(+)

27 Et H54A/H92A 49 1:4.9 1.0(+)/48(+)

28 Et H58A/H92A 43 1:2.9 5.9(–)/28(–)

29 Et H52A/C106A 37 2.4:1 88(–)/54(–)

30 Et H52A/C106S 10 2.1:1 44(–)/55(–)

31 Et H52A/C106D 64 1:1.6 88(–)/92(–)

32 Et H52A/C106N 59 8.9:1 99(–)/89(–)

33 Et I49W/H54A/H58A 69 1:6.7 16(–)/92(+)

34 Et H54A/H58A/F104W 90 1:12 94(+)/99(+)

aReaction conditions: TM1459 (0.3 mM), CuSO4 (0.3 mM), 1 (10 mM), R-CH3NO2 (100 equiv.), Potassium phosphate buffer (pH 6.5)/CH3CN (9:1), 20 °C, 3 h (for nitroethane) or 12 h (for nitropropane). bYields, diastereomeric ratio (d.r.) and enantiomeric excesses (ee) were determined by chiral HPLC analysis. cYields were calculated based on the total amount of stereoisomers. d(+) or (-) was determined by polarimeter on HPLC.

NPh

O3 mol% CuSO43 mol%TM1459

Phosphate buffer (pH6.5),CH3CN, 20 °C

NO Ph

R

NO2

NO Ph

R

NO2+

+R-CH2NO2

1syn-3, R=Mesyn-4, R=Et

anti-3, R=Meanti-4, R=Et

18

Table S7. Dependence of catalyst amount in the Michael addition reaction catalyzed by Cu-

H52A or Cu-H54A/H58A.a, b

It is worth noting that unprecedented catalytic efficiency was observed for Cu-H52A. As the concentration of Cu-H54A/H58A complex was reduced from 3 to 0.5 mol % catalyst, TON barely changed from 35, 36, and 60 (0.5 mol % catalyst) for nitromethane, nitroethane, and nitropropane, respectively. In contrast, in the case of Cu-H52A, TON increased to 106, 152 [250 (0.1 mol % catalyst)], and 95 for nitromethane, nitroethane, and nitropropane, respectively, thus retaining stereoselectivity. Therefore, the maximum TON of Cu-H52A is significantly larger than that of Cu-H54A/H58A, implying that cysteine sulfinic acid 106 (sulfinate form) plays a significant role in the activation of the donor as a base as well in the enhancement of the chemical reactivity within H52A variants.

Entry -R TM1459 Variant

Cu-TM1459 (mol%) TON Yeild

(%)[c] d.r. (anti/syn)

ee (%)[d] (anti/syn)

1 H H52A 3 30 91 - 99 (S)

2 H H52A 1 72 72 - 96 (S)

3 H H52A 0.5 106 53 - 95 (S)

4 H H52A 0.1 n.d. n.d. - n.d.

5 H H54A/H58A 3 28 83 - 89 (R)

6 H H54A/H58A 1 40 40 - 93 (R)

7 H H54A/H58A 0.5 35 17 - 92 (R)

8 H H54A/H58A 0.1 n.d. n.d. - n.d.

9 Me H52A 3 33 100 1:8 82(–)/99(–)

10 Me H52A 1 92 92 1:7 85(–)/99(–)

11 Me H52A 0.5 152 76 1:7 83(–)/98(–)

12 Me H52A 0.1 250 25 1:7 78(–)/97(–)

13 Me H54A/H58A 3 33 100 1:10 78(+)/92(+)

14 Me H54A/H58A 1 37 37 1:11 62(+)/88(+)

15 Me H54A/H58A 0.5 36 18 1:14 97(+)/94(+)

16 Me H54A/H58A 0.1 n.d. n.d. n.d. n.d.

17 Et H52A 3 27 82 1:9 93(–)/99(–)

18 Et H52A 1 64 64 1:8 85(–)/98(–)

19 Et H52A 0.5 95 47 1:8 80(–)/98(–)

20 Et H52A 0.1 n.d. n.d. n.d. n.d.

21 Et H54A/H58A 3 31 92 1:10 95(+)/95(+)

22 Et H54A/H58A 1 59 59 1:9 86(+)/96(+)

23 Et H54A/H58A 0.5 60 30 1:7 86(+)/94(+)

24 Et H54A/H58A 0.1 n.d. n.d. n.d. n.d. aReaction conditions: TM1459 (0.001-0.3 mM), CuSO4 (0.001-0.3 mM), 1 (10 mM), R-CH3NO2 (100 equiv.), Potassium phosphate buffer (pH 6.5)/CH3CN (9:1), 20 °C, 3 h (for nitromethane and nitroethane) or 12 h (for nitropropane). bYields, diastereomeric ratio (d.r.) and enantiomeric excesses (ee) were determined by chiral HPLC analysis. cYields were calculated based on the total amount of stereoisomers. d(+) or (-) was determined by polarimeter on HPLC.

NO

NO2

NO

+ CH3NO2

0.1-3 mol% CuSO40.1-3 mol% H52A or H54A/H58A

Phosphate buffer (pH6.5),CH3CN, 20 °C1 2

N Ph

O

+ R-CH3NO2

0.1-3 mol% CuSO40.1-3 mol% H52A or H54A/H58A

Phosphate buffer (pH6.5),CH3CN, 20 °C1

NO Ph

R

NO2syn-3, R=Mesyn-4, R=Et

19

Figure S1. The crystal structure of Cu-bound TM1459 at the resolution of 1.2 Å (PDB code: 6L2D). (A)

Superimposed overall structures of Cu-bound one and Cu-depleted one (PDB code: 5WSD), (B) Cu-binding

site of Chain A, and (C) Cu-binding site of Chain B. The protein main chains are displayed as ribbons, and the

key amino acid residues are highlighted as sticks. The 2Fo-Fc and anomalous maps are contoured at 1.5 and

6.0 and shown in gray and magenta mesh, respectively.

Apo-form of wild type TM1459 was incubated with 1.5 equiv of CuSO4 for 16 h at 20 °C under nitrogen gas.

After desalting of the sample, the copper content of thus obtained Cu-bound wild type was determined to be 1.0

per protein subunits by the ICP-AES analysis. Overall structure of Cu-TM1459 is almost the same to that of the

reported apo-TM1459 (rmsd = 0.166 Å, 222 Cα, PDB code: 5WSD). In the crystal of Cu-bound wild type, strong

anomalous peaks were observed at the 4-His metal-binding site, whereas no additional anomalous electron

density due to metal ions was seen elsewhere in the protein structures, indicative of no non-specific binding

metals. The coordination geometry showed N4O2 distorted octahedral structure in chain A (B), whereas N3O2

square pyramidal structure in chain B (C). It is reasonable to assume that axial coordination bonds (Cu-His54 Ne

(2.14 Å) and Cu-O2 (2.76 Å) in chain A and Cu-His54 Ne (2.22 Å) in chain B) were elongated due to Jahn-

Teller distortion (Table S3).

20

Figure S2. SDS-PAGE (14 % acrylamide) analysis of the apo-TM1459 isoforms used in this study. (A) H52A;

(B) H54A; (C) H58A; (D) H92A; (E) H52A/H54A; (F) H52A/H58A; (G) H54A/H58A; (H) H54A/H92A; (I)

H58A/H92A; (J) H52A/C106A; (K) H52A/C106S; (L) H52A/C106D (M) H52A/C106N; (N)

I49W/H54A/H58A; (O) H54A/H58A/F104W. The TM1459 mutants were prepared and purified in the same

manner as WT, except the H52A/H92A mutant, which formed inclusion bodies in E.coli cells (Figures S2 and

S3).

21

Figure S3. MALDI-TOF/MS spectra of the apo-TM1459 isoforms used in this study. (A) H52A; (B) H54A;

(C) H58A; (D) H92A; (E) H52A/H54A; (F) H52A/H58A; (G) H54A/H58A; (H) H54A/H92A; (I)

H58A/H92A; (J) H52A/C106A; (K) H52A/C106S; (L) H52A/C106D (M) H52A/C106N; (N)

I49W/H54A/H58A; (O) H54A/H58A/F104W. The observed mass numbers were almost identical to the

calculated masses derived from the corresponding amino acid sequence except H52A, H54A, H58A, and

H54A/H58A mutants, C106 of which are oxidized (Figure 2AB and Figure S5CD).

22

Figure S4. Separation of product stereoisomers for the Michael addition of nitroalkane catalyzed by selected

mutant on chiral normal-phase chromatography. (A) Separation of product 2, a, (R)-(+)-isomer; b, (S)-(–)-

isomer. (B) Separation of product 3, a, (+)-syn-isomer; b, (+)-anti-isomer; c, (–)-anti-isomer; d, (–)-syn-isomer.

(C) Separation of product 4, a, (+)-anti-isomer; b, (–)-anti-isomer; c, (–)-syn-isomer; d, (+)-syn-isomer.

23

Figure S5. The crystal structure of Cu-bound H52A (PDB code: 6L2E) and Cu-bound H54A/H58A (PDB

code: 6L2F). (A) Superimposed overall structures of Cu-bound WT, H52A, and H54A/H58A, (B) close-up

views of superimposed Cu-binding site, (C) Cu-binding site of H52A (Chain B), (D) Cu-binding site of

H54A/H58A (Chain B), (E) unspecific Cu-binding site of H52A (Chain A), (F) close-up views of unspecific

Cu-binding site The 2Fo-Fc and anomalous maps are contoured at 1.5 and 6.0 and shown in gray and magenta

mesh, respectively.

In the same manner of wild type, Cu-bound H52A and Cu-bound H54/H58A were prepared. The

copper content of Cu-bound H52A and Cu-bound H54/H58A were determined to be 0.9 and 0.8 per protein

subunits, respectively, by the ICP-AES analysis. Overall structures of Cu-bound H52A and Cu-bound

H54/H58A are almost the same to that of the Cu-bound wild type (WT vs. H52A, rmsd = 0.100 Å, 224 Cα;

WT vs. H54A/H58A, rmsd = 0.167 Å, 227 Cα). Although the coordination geometry of Cu-H52A mutant in

chain B was also square pyramidal structure, the axial ligand was His54 unlike chain A (His92 in chain A,

Figure 2A and Table S3). Additionally, an unspecific bound copper was observed only in chain A (the

occupancy < 0.3, Figure S5EF). In the case of H54A/H58A, the minor copper species (Cu(B), their

occupancies are 16 % in chain A and 23 % in chain B) were observed in the proximity of major copper species

(Cu(A), their occupancies are 73 % in chain A and 65 % in chain B) within the distance of 1.78 Å in chain A

and 1.46 Å in chain B. These results suggested that the Cu binding affinity of these two mutants are lower than

that of WT.

download fileview on ChemRxivSI_191231_NF.pdf (3.32 MiB)