Proteins as Macromolecular Ligands for Metal-Catalysed Asymmetric Transfer Hydrogenation of Ketones in Aqueous Medium

Article abstract of DOI:10.1002/ejic.201701359

Biohybrid catalysts resulting from the dative anchoring of half-sandwich organometallic complexes [M(arene)(H₂O)_x(Cl)_y]ⁿ⁺ (M = Ru^II, arene = η⁶-benzene, p-cymene or mesitylene; M = Ir^III, Rh^III, arene = η⁵-Cp*; x = 1–3, y = 0–2, n = 0–2) to bovine beta-lactoglobulin (βLG) or hen egg white lysozyme showed unprecedented behaviour. These constructs were shown to catalyse the asymmetric transfer hydrogenation of aryl ketones in water with sodium formate as hydrogen donor at a much faster rate than the complexes alone. Full conversion of the benchmark substrate 2,2,2-trifluoroacetophenone was reached with an ee of 86 % for the most selective biohybrid. Surprisingly, even the crude biohybrid gave a good ee despite the presence of non-protein-bound metal species in the reaction medium. Other aryl ketones were reduced in the same way, and the highest ee was obtained for ortho-substituted acetophenone derivatives. Furthermore, treatment of βLG with dimethyl pyrocarbonate resulted in a noticeable decrease of the activity and selectivity of the biohybrid, indicating that the sole accessible histidine residue (His146) was probably involved in the coordination and activation of Ru(benzene). This work underscores that protein scaffolds are efficient chiral ligands for asymmetric catalysis. The use of sodium formate instead of dihydrogen makes this approach safe, inexpensive and environmentally friendly.

Full text of DOI:10.1002/ejic.201701359

A Journal of
Accepted Article
Title: Proteins as macromolecular ligands for metal-catalysed
asymmetric transfer hydrogenation of ketones in aqueous
medium
Authors: José de Jesús Cázares-Marinero, Cédric Przybylski, and
Michèle Salmain
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Inorg. Chem. 10.1002/ejic.201701359
Link to VoR: http://dx.doi.org/10.1002/ejic.201701359

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Artificial transfer hydrogenases
Dative assembling of half-sandwich ruthenium, rhodium or iridium complexes to beta-
lactoglobulin and lysozyme gave biohybrids that catalyzed the asymmetric transfer
hydrogenation of ketones with up to 86% ee. Complementary investigations pointed out to
the involvement of beta-lactoglobulin’s His146 in ruthenium coordination.
1
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Proteins as macromolecular ligands for metal-catalysed asymmetric transfer
hydrogenation of ketones in aqueous medium
José de Jesús Cázares-Marinero, Cédric Przybylski, Michèle Salmain*
Sorbonne Université, CNRS, Institut Parisien de Chimie Moléculaire, IPCM, F-75005 Paris,
France
* to whom correspondence should be addressed. Email: michele.salmain@upmc.fr.
Homepage: http://www.ipcm.fr/presentation-682?lang=en
Abstract
We report the generation of biohybrid catalysts resulting from the dative anchoring of half-
sandwich organometallic complexes [M(arene)(H2O)x(Cl)y]n+ (M = RuII, arene = 6-benzene, p-
cymene or mesitylene; M = IrIII, RhIII, arene = 5-Cp*; x = 1-3, y = 0-2, n = 0-2) to bovine beta-
lactoglobulin (LG) or hen egg white lysozyme (HEWL) with unprecedented behaviour. These
constructs were shown to catalyse the asymmetric transfer hydrogenation (ATH) of aryl
ketones in water with sodium formate as hydrogen donor at a much faster rate compared to
the complexes alone. Full conversion of the benchmark substrate 2,2,2-
trifluoroacetophenone (TFACP) was reached with enantiomeric excess (ee) of 86% for the
most selective biohybrid. Surprisingly, even the crude biohybrid gave a good ee despite the
presence of non-protein bound metallic species in the reaction medium. Other aryl ketones
were reduced the same way, with the highest ee obtained for ortho-substituted
acetophenone derivatives. Furthermore, treatment of LG with dimethylpyrocarbonate
(DMPC) resulted in a noticeable decrease of the activity and selectivity of the biohybrid,
indicating that the sole accessible histidine (His146) was probably involved in the
coordination and activation of Ru(benzene). This piece of work underscores that protein
scaffolds are efficient chiral ligands for asymmetric catalysis. The use of sodium formate
instead of dihydrogen makes this approach safe, inexpensive and environmentally friendly.
Keywords
Beta-lactoglobulin; lysozyme; artificial metalloenzyme; mass spectrometry; ruthenium
2
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
INTRODUCTION
The design of artificial metalloenzymes (ArMs) for enantioselective catalysis has been the
subject of extensive research during the past fifteen years[1] and has now reached a state of
maturity.[2] ArMs elegantly complement wild-type enzymes by achieving high turnover
numbers and high enantioselectivity under mild conditions while expanding their application
to a wider range of reactions and substrates. Various strategies are employed to implant a
catalytically active metal centre within biomacromolecules such as DNA, native proteins or
de novo peptide scaffolds. These strategies are rationalized into three distinct approaches,
i.e. covalent, supramolecular and dative anchoring as well as a combination of them.
While the ArMs resulting from the covalent and supramolecular anchoring approaches
display good to excellent reactivity and selectivity, the dative strategy for which the protein
provides one or more amino acid residues to coordinate the metal-containing species is an
attractive alternative. In this way, metalloproteins’ active site can be efficiently mimicked by
taking into account both the first and second coordination spheres.[3] Implanting non-native
metallic species within protein scaffolds greatly expands the scope of natural enzymes by
allowing the catalysis of abiotic reactions, such as Diels-Alder cycloadditions,[4] C-H bond
activation,[5] or cyclopropanation[6].
Engineering of a metal binding site within a protein scaffold can be achieved in several ways.
One of which is to take advantage of pre-existing metal coordination sites of metalloproteins
and involves metal ion substitution or metal cofactor replacement.[7] Regarding metal ion
substitution, this approach has been employed on the zinc enzymes carboxypeptidase A[8]
and carbonic anhydrase.[9] Two prerequisites are the easy production of the apoprotein and
its sufficient stability. Another rational strategy is to engineer a metal coordination site
within a non-metallated protein scaffold[10] or to make use of latent metal binding sites in
protein scaffolds.[11] Alternatively, metal binding sites can be created by introduction of
unnatural aminoacids via genetic engineering.[12] A more serendipitous approach is to make
use of the coordination ability of proteins to bind catalytically active metal centres. This
approach has been employed with variable success on a small set of protein scaffolds
(albumin,[13] apoferritin,[14] lysozyme[15] and streptavidin[16]) and transition metal complexes
(Os, Pd, Rh, Mn, Fe, Ru and V). The best results in terms of enantioselectivity were obtained
with albumin- and streptavidin-osmium tetroxide hybrids that catalysed the cis-
dihydroxylation of -methylstyrene with up to 68%[13a] and 95%[16a] ee, respectively.
Asymmetric transfer hydrogenation (ATH) of ketones and imines is a powerful strategy to
produce chiral alcohols and amines in a safe, inexpensive and environmentally benign way.
Many half-sandwich complexes of RuII, RhIII and IrIII with phosphorous or nitrogen donor
ligands catalyse these reactions in aqueous medium using formate as hydrogen source.[17]
Occasionally, peptides have been used as ligands too. For instance, mixtures of dipeptide-
like ligands and p-cymene dichloro-bridged rutheniumII dimer catalyse the ATH of aryl
ketones in isopropanol as hydrogen donor with excellent ee values.[18] Along the same line,
mixtures of tris-aqua complex [Ir(Cp*)(H2O)3]2+ and tripeptides such as Gly-Gly-Phe catalyse
3
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
the transfer hydrogenation of a range of ketones in aqueous medium with formate as
hydrogen donor with high turnover frequencies (TOFs) but no asymmetric induction was
observed.[19] Artificial transfer hydrogenases resulting from the supramolecular assembling
of half-sandwich Ru, Rh and Ir complexes to (strept)avidin[20] and carbonic anhydrase[21] have
also been shown to catalyse the ATH of ketones and imines with very good efficiency.
Our group has also contributed to the field of ArMs by designing transfer hydrogenases
resulting from (i) covalent anchoring of achiral half-sandwich RuII and RhIII diimine complexes
to papain[22] or (ii) supramolecular assembling of related complexes to bovine beta-
lactoglobulin (LG).[23] The latter approach took advantage of the strong affinity of fatty
acids for LG allowing the most efficient construct to catalyse the ATH of the benchmark
substrate 2,2,2-trifluoroacetophenone (TFACP) with 59 % conversion and 32% ee in 72 h.
Presently, the same protein host LG along with hen egg white lysozyme (HEWL) were used
to build up hybrid metalloproteins by the dative anchoring approach. The resulting
constructs were shown to display transfer hydrogenase activity on a range of aromatic
ketones including TFACP. The most efficient ArM afforded full conversion of TFACP in ca. 12
h with 86% ee. This achievement compares favourably with our previous results with LG
and papain as protein hosts.
RESULTS AND DISCUSSION
Preparation and characterisation of metallated beta-lactoglobulin
Prior to metalloprotein assembling, aqueous suspensions of the dimeric complexes
[M(arene)(µ-Cl)Cl]2 (M = RuII, arene = p-cymene (p-cym), benzene (benz) or mesitylene
(mes); M = RhIII or IrIII, arene = Cp*, Fig. 1) in water were heated to 50°C for a few minutes.
Such a treatment applied to the Ru(benz) dimer in D2O has been previously shown to afford
a mixture of monomeric aqua complexes [Ru(benz)(D2O)Cl2], [Ru(benz)(D2O)2Cl]+ and
1
[Ru(benz)(D2O)3]2+ in proportion 8:28:5 as determined by H NMR analysis.[24] This finding
has been reproduced in our hands (Fig. S7A).
R''
R'
R
R''
M
Ru
Cl
Cl
Cl
Cl
Cl
Cl
Cl
M
Cl
R
Ru
R''
R'
R''
5-Cp*)Cl]2
5-Cp*)Cl]2
M = RhIII
M = IrIII
[Rh(m-Cl)(
[Ir(m-Cl)(
h
[Ru(
[Ru(
[Ru(
h
h
h
6-benz)(m-Cl)Cl]2
6-p-cym)(m-Cl)Cl]2
6-mes)(m-Cl)Cl]2
R, R', R'' = H
h
R = Me; R' = iPr; R'' = H
R, R'' = Me; R' = H
Figure 1. Dimeric half-sandwich RuII, RhIII and IrIII complexes used as precursors in the
catalysis of (A)TH of aromatic ketones.
Then, beta-lactoglobulin was incubated with hydrolysed Ru(benz) at an initial protein/Ru
ratio of 1:3 or 1:6 and submitted to gel desalting to separate the protein from unbound
4
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Ru(benz). The purified metallated protein was submitted to ICP-EOS analysis to determine
the Ru concentration (Fig. S1). This together with protein concentration determination at
280 nm, afforded a Ru/LG of 1.1 (when the initial Ru/LG ratio was 3:1) and 1.6 (when the
initial ratio was 6:1).
The same experiment was performed at an initial molar ratio of 1:1 with hydrolysed
Ru(benz), Ru(p-cym), Ru(mes), Rh(Cp*) or Ir(Cp*) and UV-visible spectra of metallated
proteins were recorded and compared to the hydrolysed complexes alone (Table 1, Fig. S2).
All the metallated protein samples (except the Ir adduct) displayed a broad band between
330 and 400 nm tentatively assigned to a metal-to-ligand-charge-transfer (MLCT) transition
of the metallic entity. The absorption maximum of this band was blue-shifted by 11 nm for
[Ru(benz)] adduct and by 25 nm for [Ru(p-cym)] adduct compared to the corresponding
hydrolysed metal complex alone. These shifts were undoubtedly assignable to a change of
the coordination sphere around the metal upon reaction with the protein. To try to
understand which amino acid(s) were involved in metal coordination, the Uv-vis spectra of
stoichiometric mixtures of complex and (i) imidazole mimicking histidine as putative N donor
ligand or (ii) methionine as S donor ligand were also recorded. Addition of imidazole shifted
the MLCT band shifted to the blue by 8 to 15 nm while addition of methionine shifted the
MLCT band to the blue by 16 to 25 nm. On the whole, the position of the MLCT band of the
biohybrids is close to that of the mixtures of complex with methionine. Nevertheless, it
would be presumptuous to assert that methionine is the amino acid involved in metal
coordination on the sole basis of absorption spectroscopy.
Table 1. Position of the MLCT band of hydrolysed M(arene), LG•M(arene) (initial M/LG =
1:1) and stoichiometric mixtures of M(arene) and imidazole or methionine in water

M(arene) +
M(arene)
M(arene)+imidazole
LG•M(arene)
methionine
M
arene
 (M-1cm-
 (M-1cm-
 (M-1cm-
 (M-1cm-

max (nm)

max (nm)

max (nm)

max (nm)
1
1
1
)
1)
630
670
500
2800
)
)
Ru
Ru
benz
mes
394
400
403
378
410
450
560
382
378
378
360
386
388
388
367
375
420
460
378
n.d.
378
356
440
n.d.
520
Ru p-cym
Rh Cp*
1700
1800
3000
Addition of hydrolysed Ru(benz) to a solution of LG in water decreased the intrinsic
fluorescence of the protein’s two tryptophan residues at 336 nm. Gradual quenching was
observed upon addition of complex until saturation was reached at ca. Ru/LG = 5:1 (Fig.
S5). Titration of LG by hydrolysed Ru(benz) was also monitored by Uv-vis spectroscopy and
the difference spectra depicted in Fig. S4 show the appearance of a band centred at ca. 325
nm upon addition of Ru(benz) which is characteristic of the biohybrid. Near Uv CD spectra of
mixtures of LG and hydrolysed Ru(benz) also showed a positive band centred at ca. 320 nm
as a result of biohybrid formation (Fig. S4).
Mass spectrometry analysis of metallated LG
5
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
The isoform A of LG (LG-A) was incubated with 0, 1, 2 or 3 mole eq. of hydrolysed
Ru(benz) complex in water for 1 h. The mixtures were next analysed by electrospray-high
resolution mass spectrometry (ESI-HRMS). The mass spectrum of native LG-A reveals a
charge states distribution from 10+ (m/z 1837.36) to 19+ (m/z 967.35) and centred around
14+ and 15+, at m/z 1225.17 and m/z 1312.62, respectively (Fig. 2A).
1237.1700
z=15
bLG-A/Ru(benz) 1/3
100
1325.4745
z=14
1159.7265
z=16
D
80
1426.9727
60
40
20
0
z=13
1546.0468
z=12
1091.6819
z=?
1044.6551
z=17
1237.0418
z=15
100
80
60
40
20
0
bLG-A/Ru(benz) 1/2
bLG-A/Ru(benz) 1/1
1325.2582
z=14
C
B
A
1248.7710
z=15
1159.7275
z=16
1427.2807
z=13
976.7176
z=19
1091.5085
z=17
1546.0483
z=12
1044.6546
z=18
1237.1105
z=15
1225.1815
z=15
1159.7285
100
80
60
40
20
0
1325.3318
z=14
z=16
1243.6393
z=15
1427.0488
z=13
1091.6270
z=17
976.7170
z=19
1545.8822
z=12
1030.9796
z=18
1686.6926
z=11
1854.7514
z=10
1225.1705
z=15
1312.6210
z=14
100
80
60
40
20
0
bLG-A/Ru(benz) 1/0
1319.6150
z=14
1148.7303
z=16
1413.5892
z=13
1231.6393
z=15
1081.1586
z=17
967.3498
z=19
1531.2175
z=12
1154.7265
z=16
1021.1482
z=18
1670.2337
z=11
1837.3581
z=10
900
1000
1100
1200
1300
1400
m/z
1500
1600
1700
1800
1900
Figure 2. ESI-HRMS mass spectra of LG-A without Ru(Benz) (A) and resulting from the
reaction of 1 (B), 2 (C) or 3 (D) mole eq. hydrolysed Ru(benz) after incubation in water for 1 h
at RT.
Increase of Ru(Benz) from 0 to 3 eq. resulted in the progressive formation of metallated LG-
A detected as a metalloprotein with charge states ranging from 10+ (m/z 1854.75) to 19+
(m/z 977.71) (Fig. 2B-D). To confirm that the exact nature of the metal fragment bound to
the protein was a [Ru(benz)]2+ entity, a comparison between calculated and experimental
isotopic distribution patterns of the most prominent 15+ ion of LG-A and LG-A + Ru(benz)
was performed (Fig. 3)
6
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
A
1225.1705
z=15
Experimental
100
90
1225.0383
1225.3026
z=15
1225.3686
z=15
z=15
80
70
60
50
40
30
20
10
0
1224.9724
z=15
1225.4353
z=15
1225.5014
z=15
1224.9055
z=15
1225.6343
z=15
1224.7715
z=15
1226.2350
1225.7685
z=15
1226.0994
z=15
1224.4436
z=15
z=15
1225.1708
Theoretical
100
90
80
70
60
50
40
30
20
10
0
15+
[C821H1318N206O250S9 +15H]
1225.0372
1224.9702
1225.3044
1225.3711
1225.4381
1224.9036
1224.8366
1225.5047
1225.5717
1225.6384
1224.7696
1224.7030
1224.5691
1224.6
1225.7054
1225.8387
1225.9724
1226.0
1226.1728
1226.2
1224.4
1224.8
1225.0
1225.2
1225.4
1225.6
1225.8
1226.4
m/z
B
1236.0346
z=15
1237.1005
z=15
1236.9677
z=15
1237.1672
z=15
Experimental
100
90
80
70
60
50
40
30
20
10
0
1236.9014
z=15
1236.8341
z=15
1237.2990
z=15
1237.3658
z=15
1236.7677
z=15
1236.7006
1237.4999
z=15
z=15
1236.5680
z=15
1237.1008 1237.1676
Theoretical
1237.2344
100
90
80
70
60
50
40
30
20
10
0
1237.0340
1236.9672
15+
[C827H1324N206O250S9Ru+15H]
1237.3012
1237.3680
1237.4348
1236.9005
1236.8337
1236.7670
1237.5016
1237.5684
1236.7002
1236.6335
1237.6351
1237.7019
1236.5668
1237.7687
1236.0
1236.5
1237.0
1237.5
m/z
1238.0
1238.5
Figure 3. Isotopic distribution of 15+ ion for LG-A (A) and LG-A + Ru(benz) (B) with
experimental (upper) and calculated (lower) patterns.
Experimental isotopic profiles obtained for LG-A (Fig. 3A) and LG-A + Ru(benz) (Fig. 3B)
matched very well with the calculated isotopic patterns. As regards LG-A + 2[Ru(benz)], it
was quite difficult to make a straightforward assignment based only on the isotopic pattern
due to very low signal and possible resulting truncated distribution (not shown).
Furthermore, the aforementioned data i.e. multicharged spectra and isotopic profiles do not
give access to both exact masses and sample content. Consequently, to get further pieces of
information about the content of LG-A and various complexes formed, deconvolution and
deisotoping steps were achieved (Fig S6). Deconvoluted mass spectrum of LG-A only (Fig.
S6A) showed a major peak at a monoisotopic mass of 18,351.5205 which is in good
agreement with the theoretical value of LG-A (18,351.4234 g.mol-1; 5 ppm). Two other
minor species at masses of 18,449.4984 and 18675.6266 (Fig. S6A) were also observed. The
former one was assigned to a neutral adduct with phosphoric acid (mass shift of +97.98), as
7
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
described elsewhere for other proteins,[25] and the latter to a typical product of lactose
condensation (mass shift of +324.11) consecutively to the Maillard reaction.[26] As regards
the spectra corresponding to complexation of LG-A with Ru(Benz) at 1, 2 and 3 mole eq.
(Fig. S6B-D), except for a residual amount of free protein, two prominent additional species
appeared in the masses range between 18,529.4977 to 18,529.5314 and 18,707.4466 to
18,707.4488. Taking into account the mass obtained for LG-A only, these species
correspond to two consecutive average mass shifts of 177.95. Such values can certainly be
attributed to complexation of LG-A with 1 and 2 [Ru(Benz)]2+. In addition, according to
aforementioned values, a good agreement was also obtained with theoretical masses i.e.
18,529.3882 g.mol-1 (5th isotope; 5 ppm) and 18,707.3452 g.mol-1 (5th isotope; 3 ppm) for
LG-A + one Ru(benz) and LG-A + two [Ru(benz)], respectively. Assuming that all species
exhibit a similar ESI response factor, relative quantification of various species yielded ratios
of native LG-A/Ru bound forms of 21/79 and 10/90, with 1 and 2 mole eq. hydrolysed
Ru(benz), respectively. Further increase to 3 mole eq. hydrolysed Ru(benz) yielded a similar
proportion of native LG-A and Ru bound proteins (13/87). In detail, among the ruthenated
protein forms, the mono-metallated one was in all cases the most prominent with 61, 51
and 65% content for Ru(benz)/LG-A ratios of 1:1, 2:1 and 3:1, respectively. The average
number of Ru atoms bound per LG calculated from the proportion of native LG-A,
monoruthenated LG-A and diruthenated LG-A equalled 1.09 for the initial Ru(benz)/LG
ratio of 3:1, in excellent agreement with the ICP-OES data given above (1.1).
All the spectroscopic data obtained by different techniques point out to the formation of
organometallic biohybrids resulting from the coordination of beta-lactoglobulin to the metal
centre.
Transfer hydrogenation of TFACP catalysed by metallated LG
First, the ATH of TFACP in the presence of metallated LG resulting from the reaction with 1
eq. of hydrolysed half-sandwich complexes (LG•M(arene) and sodium formate as hydrogen
donor was tested (Scheme 1). In parallel, control experiments were carried out with the
hydrolysed complexes alone. Conversion and enantiomeric excess (ee) were determined by
chiral HPLC of the reaction mixtures after 33 h (Fig. 4).
Scheme 1. (Asymmetric) Transfer Hydrogenation of TFACP
Figure 4. (A)TH of TFACP (5 mM) catalysed by LG•M(arene)1 or hydrolysed M(arene) (0.1
mM). Conditions: [HCOONa] = 1 M, pH = 7.5, T = 40°C. Top: conversion and ee measured
after 33 h; bottom: turnover frequency (TOF)
8
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
100
87
93
90
82
80
70
60
50
41
38
35
40
30
20
10
0
33
30
29
28
17
16
14
13
9
Ru(benz)
Ru(p-cym)
% ee
Ru(mes)
Rh(Cp*)
Ir(Cp*)
% conversion
% conversion (complex alone)
4.4
5
4
3
2
1
0
2.3
1.6
1.5
0.2
0.13
0.1
0.1
0.6
0.2
Ru(benz)
Ru(p-cym)
Ru(mes)
Rh(Cp*)
Ir(Cp*)
bLG•M(arene)
M(arene)
Conversion of TFACP catalysed by the half-sandwich metal complexes alone clearly
depended on the metal and, in the Ru series, on the nature of the ancillary aromatic ligand.
The Ir complex alone was the most active of the series since full conversion was observed
after 33h in good agreement with Ogo and Watanabe’s findings for [Ir(Cp*)(H2O)3]2+.[27] In
accordance, our evaluations let to rank the catalytic activity of complexes alone in the
following order: Ir(Cp*) >> Ru(p-cym) > Ru(benz) > Rh(Cp*) > Ru(mes). In the case of the
Ru(benz) complex, the mechanism of transfer hydrogenation involved the transient
1
formation of a Ru-H species as observed by H NMR spectroscopy by the presence of a
singlet at -7.03 ppm (Fig. S7) in the range of previously reported (arene)(L^L)Ru-H species
(between -5 and -8 ppm depending on the L^L ligand).[28] As expected, the corresponding
alcohol was obtained as a racemic mixture in all cases where complexes were used alone.
On the other hand, all the metalloprotein hybrids were also found to catalyse the ATH of
TFACP with conversions ranging from 13% for the least active biocatalyst LG•Ru(mes) to
87% for the most active biocatalyst LG•Ru(benz). Surprisingly a completely different
9
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
ranking was established for the hybrid catalysts in comparison to that of the complexes
alone, i.e. LG•Ru(benz) >> LG•Ir(Cp*) ~ LG•Ru(p-cym) > LG•Rh(Cp*) > LG•Ru(mes).
Moreover, the (R)-enantiomer of -(trifluoromethyl)benzyl alcohol was obtained
preferentially with all the metallated LG; with ee ranging from 15% for the least selective
catalyst LG•Rh(Cp*) to 82% for the most selective catalyst LG•Ru(benz). Let us notice that
both the Rh and the Ir hybrids gave the same ee, while LG•Ru(p-cym) gave the same ee as
LG•Ru(mes). This similarity might be due to identical coordination sites for both pairs of
complexes.
The TOF representing the initial reaction rate was then calculated at 1 h reaction time for all
the LG•M(arene) and compared to that afforded by the complexes alone (Fig. 4, bottom).
The rate of TH of TFACP was markedly enhanced by coordination of LG to the Ru(benz)
complex (enhancement factor = 23). Protein coordination to the two other Ru complexes
and the Rh one enhanced the rate of reaction to a lesser extent (enhancement factor < 3 for
Ru and 12 for Rh). Interestingly, LG coordination decreased the activity of the iridium
complex by a factor of 3.
Next, LG was treated with Ru(benz) at 1:3 or 1:6 ratio as described above. The resulting
biohybrids were tested in the ATH of TFACP under standard conditions. The concentrations
of ketone and alcohol in the reaction mixtures were monitored overtime by chiral HPLC.
Conversion plots are shown in Fig. 5 and table 2 gathers the related kinetic data.
Figure 5. Time course of ATH of TFACP (5 mM) catalysed by LG•Ru(benz) (Ru/bLG = 1:1, 3:1
or 6:1). Conditions: 2 mol% cat, [HCOONa] = 1 M, pH = 7.5, T = 40°C, V = 1 mL.
100%
80%
60%
40%
20%
0%
0
5
10
15
20
25
30
time (h)
x=1
x=3
x=3 (ee)
x=6
x=1 (ee)
x=6 (ee)
10
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Table 2. (A)TH of TFACP. TOF, conversions and ee with Ru(benz) derived catalysts.
Conditions: [TFACP] = 5 mM, 2 mol% LG, [HCOONa] = 1 M, pH 7.5, 40°C, v=1 ml
entry Catalyst
kobs (h-1)
TOF (h-1) Conversion (%) ee (%)
LG/Ru(benz)
and time (h)
ratio
1
2
3
4
5
6
1:1
0.086±0.015
0.23±0.03
0.25±0.02
n.d.
2.3
6.1
6.7
3.2
0.8
1.2
87 (21)
93 (12)
79 (10)
99 (05)
96 (25)
94 (25)
82 (R)
87 (R)
83 (R)
78 (R)
69 (R)
0
LG•Ru(benz)
LG•Ru(benz)
1:3
1:6
1:3
1:3
1:1
LG•Ru(benz)
LG + Ru(benz)
n.d.
0.083±0.015
LG-DMPC + Ru(benz)
imidazole + Ru(benz)a
a [imidazole] = [Ru(benz)] = 0.3 mM
Nearly complete conversion of TFACP was observed with the three biocatalysts. The initial
TOF (Table 2) ranged from 2.3 h-1 for LG/Ru(benz) = 1:1 to 6.7 h-1 for LG/Ru(benz) = 1:6.
The rate constant of reaction kobs (calculated assuming first order kinetics) followed the
same trend. The TOF afforded by LG/Ru(benz) = 1:3 was ca. three times that afforded by
LG/Ru(benz).= 1:1. In contrast, the TOF afforded by LG/Ru(benz) = 1:6 was only slightly
higher than that of LG/Ru(benz) = 1:3, which is in line with the similar final Ru/protein ratio
of both samples (see above). All three biocatalysts afforded constant ee values slightly above
80% over the time range which indicates that they remained stable all along.
A catalytic run was performed with a crude mixture of LG and Ru(benz) at ratio of 1:3 in the
same experimental condition (Table 2, entry 4). Conversion reached 99% in 5h and the ee of
the sec-alcohol was 78%. The presence of metal complex alone in the reaction mixture only
very slightly altered the overall selectivity. This finding can be explained by the very poor
reactivity of the Ru(benz) complex by itself compared to the protein coordinated form. Time
course of ATH of TFACP in the presence of mixture of LG and Ru(benz) at molar ratio of 1:5
1
and sodium formate was also monitored by H NMR in D2O (Fig. S8). Signals of the aromatic
protons of TFACP progressively decreased, while a multiplet and a well resolved quadruplet
1
belonging to the sec-alcohol product gradually appeared. The evolution of the H NMR
spectrum of the reaction is consistent with consumption of ketone and formation of alcohol.
A singlet at 5.56 ppm tentatively assigned to the protons of the benzene ligand of a [Ru-
formato] adduct was simultaneously seen to decrease until complete ketone conversion.
Substrate scope
The ArM formed by simply mixing LG and Ru(benz) at ratio of 1:3 was applied to the ATH of
other aromatic ketones. All the catalytic runs were performed at 1 mol% LG (3 mol%
[Ru(benz)]). Conversion and enantioselectivity were measured by chiral GC analysis of the
reaction mixtures after 40 h at 40°C (Table 3).
11
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Table 3. ATH of various aryl ketones in water.a
R1
H
H
R2
H
Me
H
H
H
H
H
H
H
Conversion (%)
ee (%)
11
23
55
17
12
41
6
74
15
75
28
73
63
81
n.d.
n.d.
82
o-Me
m-Me
p-Me
o-MeO
m-MeO
o-Br
m-Br
o-NO2
o-CF3
o-F
55
n.d.
n.d.
n.d.
52
77
80
H
H
H
a [substrate] = 10 mM, [LG] = 0.1 mM, [Ru(benz)] = 0.3 mM, [HCOONa] = 1 M, pH = 7.5; n.d:
not determined
The mixture of LG and Ru(benz) was also able to catalyse the ATH of non-activated aryl
ketones in aqueous medium. The presence of substituents in ortho position of the methyl
ketone group induced significant stereoselectivity, up to 75% ee, while substituents at the
two other positions had nearly no effect. Interestingly, the (S)-enantiomer was obtained
preferentially for all the tested ketones (and the (R)-enantiomer for the hydrogenation
product of TFACP), indicating that the nucleophilic hydride attack always occurred at the
same face of the substrate. Indeed, the R/S formalism is based on the Cahn, Ingold and
Prelog priority rules that place the CF3 group above the Ph group and the Ph group above
the CH3 group.
Transfer hydrogenation of TFACP catalysed by metallated HEWL
It has been recently shown that association of Ru(benz) and cross-linked crystals of HEWL
catalyse the TH of acetophenone derivatives with variable selectivity.[15] Thus, as a
preliminary screening, we examined the ability of crude mixtures of HEWL and hydrolysed
M(arene) complexes (M = Ru, arene = benz or p-cym; M = Rh or Ir, L = Cp*) at mole ratio 1:1
to catalyse the ATH of TFACP. For comparison, the conversion of TFACP was measured in the
presence of the complexes alone (Fig. 6).
12
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Figure 6. ATH of TFACP (10 mM) catalysed by HEWL + M(arene) (0.2 mM) or M(arene) (0.2
mM). Conditions: [HCOONa] = 1 M, pH = 7.5, T = 40°C, t = 9 h
80
64
70
60
50
40
30
20
10
0
R
51
50
24
23
21
14
4
3
3
1
Ru(benz)
Ru(p-cym)
Rh(Cp*)
Ir(Cp*)
-38
-10
-20
-30
-40
S
% conversion
% ee
% conversion (complex alone)
All the mixtures were able to catalyse the ATH of TFACP with the following ranking
HEWL+Ir(Cp*) ~ HEWL+Rh(Cp*) > HEWL+Ru(benz) > HEWL+Ru(p-cym). This is dramatically
different from the ranking established for the LG biohybrids. The presence of HEWL in the
reaction mixture markedly enhanced the activity of the metal complexes except again for
Ir(Cp*) for which a slight decrease of the conversion was observed in the presence of HEWL.
This finding and asymmetric induction observed for 3 of the 4 mixtures is in agreement with
protein coordination to the metal centre. Enantiomeric excesses were different from those
previously measured with the LG biohybrids and, quite surprisingly, the mixture of HEWL
and Ir(Cp*) afforded preferentially the (S)-enantiomer of -(trifluoromethyl)benzyl alcohol,
whereas the mixtures of HEWL and the Ru complexes gave the (R)-isomer as before. The
rather poor selectivity of the HEWL + Ru(benz) mixture is in line with previously reported
results on other acetophenone derivatives.[15]
Therefore we synthesized HEWL•Ir(Cp*) biohybrids by reacting HEWL and hydrolysed Ir(Cp*)
at initial ratios of 1:1, 1:5 or 1:10. After purification by gel filtration, time course of ATH of
TFACP in the presence of the metallated proteins HEWL•Ir(Cp*) (HEWL/Ir ratios = 1:1, 1:5 or
1:10) was monitored by chiral HPLC. As seen in Fig. 7 and Table 4, the rate of conversion and
enantiomeric excess depended on the initial HEWL/Ir ratio with the fastest rate and the
highest ee (43%) measured for ratio 1: 10. Moreover, the enantiomeric excess tended to
decrease with time, the magnitude of this decrease heightened with reaction time. This
trend was undoubtedly related to an evolution of the nature of the catalytic species during
the time course of the reaction; one possible change may be related to decomplexation of
the protein ligand. Quite unexpectedly, the catalytic performances of HEWL•Ir(Cp*) and
HEWL+Ir(Cp*) (at 1:1 ratio) differed significantly since the mixture of HEWL and Ir(Cp)* was a
more selective catalyst than HEWL•Ir(Cp*)1 (38% vs. 7 % ee). Correlatively, the gel filtration
step decreased the selectivity of the hybrid catalyst. We can reasonably assume that HEWL
13
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
contains several binding sites for Ir(Cp*), each one of these having its own binding constant
and the ability to afford a specific enantioselectivity. Gel filtration may have induced
dissociation of the iridium complex from the weakest protein binding site(s) which may also
have been those providing the highest selectivity in the ATH of TFACP.
Figure 7. Time course of ATH of TFACP (5 mM) catalysed by HEWL•Ir(Cp*) (HEWL/Ir ratio =
1:1, 1:5 or 1:10). Conditions: 2 mol% cat, 1 M HCOONa in HEPES buffer pH 7.8, T = 40°C, v=1
ml.
100%
80%
60%
40%
20%
0%
0
5
10
15
20
25
30
time (h)
x=1
x=5
x=5 (ee)
x=10
x=1 (ee)
x=10 (ee)
Table 4. (A)TH of TFACP, TOF, conversions and ee with Ir(Cp*) derived catalysts. Conditions:
[TFACP] = 5 mM, 2 mol% HEWL, 1 M HCOONa in HEPES buffer pH = 7.8, T = 40°C, V = 1 mL.
Catalyst
HEWL/Ir
ratio
kobs (h-1)
TOF (h-1)
Conversion (%)
and time (h)
ee (%)
HEWL•Ir(Cp*)
HEWL•Ir(Cp*)
HEWL•Ir(Cp*)
Ir(Cp*)a
1:1
1:5
1:10
-
0.16±0.01
0.48±0.09
0.78±0.07
0.11±0.02
n.d.
3.7
18.6
24.4
4.4
47 (24)
93 (24)
98 (8)
93 (14)
51 (9)
7 (S)
36 (S)
43 (S)
0
HEWL+Ir(Cp*)b
1:1
3.2
38 (S)
a
b
[Ir(Cp*)]=0.1 mM, 1 M HCOONa pH 7.5; [HEWL]=[Ir(Cp*)]=0.2 mM, [TFACP]=10 mM,
[HCOONa]=1 M, pH 7.5
Tentative determination of ruthenium binding site(s) on beta-lactoglobulin
A survey of the literature dealing with the interaction of half-sandwich RuII complexes with
proteins may give clues as to their preferential protein binding sites. The Ru(p-cym) and the
Ru(benz) complexes were shown to bind to crystallized HEWL at its sole His residue (His15)
with a second minor binding site at Asp101.[15,29] This behaviour was explained by the
favourable local environment around His15 and its good accessibility in the crystal state.
Moreover, His15 is surrounded by a cluster of residues creating a pocket of the right size for
14
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
the complex to fit in. When the multimeric protein apoferritin was treated with a large
excess of [Ru(p-cym)Cl2]2, three main binding sites were identified by X-ray crystallography:
(i) one involving two His and one Glu; (ii) another one involving one Glu, one Asp and one
Cys and (iii) the last one involving one Cys and one His.[14b] The anticancer complex RAPTA-C
([Ru(p-cym)(pta)Cl2]) bound preferentially to His33 of cytochrome c.[30] Another anticancer
complex namely, [Ru(p-cym)(en)Cl]PF6 (en = ethylenediamine), bound to human albumin via
coordination of its sole free cysteine residue (Cys34) as well as three accessible histidines
and one methionine.[31] The same complex was found to bind to three S-donor cysteine
residues of human glutathione-S-transferase.[32]
Beta-lactoglobulin offers several putative coordination sites (O, N and S donor ligands) for
the Ru(benz) complex. Considering that the main product of hydrolysis of the ruthenium
dimer is the bis-aqua monomeric adduct [Ru(benz)(H2O)2Cl]+, there should be at least two
positions available for coordination of LG that may act as a monodentate or a bidentate
ligand. Beta-LG contains two histidines (His146 and His161), one free cysteine (Cys121), two
methionines (Met7 and Met145), 15 glutamates and 11 aspartates. The two Met are poorly
accessible and Cys121 is completely inaccessible (located at the outer surface of the beta-
barrel on strand H under the alpha-helix).[33] From the calculation of surface accessibilities,
only one of the two histidines, namely His146, appears to be surface-exposed. Indeed, assay
of LG’s histidines with diethyl pyrocarbonate (DEPC) according to Miles[34] gave a number
of modified histidines equal to one. However, at neutral pH, LG exists as a dimer with a
dissociation constant in the micromolar range.[35] X-ray crystallographic studies showed that
dimer association occurs via a network of hydrogen bonds between the -strands I (residues
146-152) of both monomers that assemble in a head to tail fashion.[36] The 3D-structure of
LG dimer at pH 7.1 (PDB file 1BSY) is depicted in the supplementary information section
(Fig. S9). It clearly shows that His146 is positioned in a somewhat narrow cavity at the edge
of the interfacial region between the two monomers with only the N accessible to solvent.
To check whether His146 was involved in the coordination of Ru(benz), LG was treated with
excess dimethylpyrocarbonate (DMPC)[37] which has a similar reactivity as DEPC. N-
methoxycarbonylation of imidazole ring is assumed to make the remaining nitrogen of
histidine a poorer ligand and thus potentially inhibit metal coordination.[9b] After gel
filtration to remove excess DMPC, the LG-DMPC conjugate was mixed with hydrolysed
Ru(benz) at Ru/LG ratio of 3:1 with no further purification. Conversion of TFACP and
enantiomeric excess were measured at different time points (Fig. 8 and table 2). Pre-
treatment of LG with DMPC led to a dramatic decrease of the rate of ATH as compared to
untreated LG (TOF = 0.8 h-1 vs. 3.2 h-1). The enantiomeric excess was also significantly
affected, ranging from 34% to 64% during the reaction. This different behaviour of LG-
DMPC provided a solid clue as to the involvement of His146 in ruthenium coordination.
However, since some enantioselectivity was still observed, His146 is probably not the sole
residue to be involved in ruthenium coordination, which agrees well with the presence of a
doubly ruthenated protein complex identified by ESI-HRMS. Another hypothesis is that
15
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
gradual increase of the ee may be due to the progressive removal of the –C(O)OMe group
during catalysis that enables His146 to coordinate back to Ru(benz). The stability of N-
(alkoxycarbonyl)imidazole is known to be pH-dependent and dealkoxycarbonylation is
readily achieved by treatment with NH2OH at neutral pH.[34] Such a reaction may have also
occurred in the presence of 1 M HCOONa at 40°C.
Figure 8. Time course of ATH of TFACP (5 mM) catalysed by LG (0.1 mM) + Ru(benz) (0.3
mM) or LG-DMPC (0.1 mM) + Ru(benz) (0.3 mM).
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0
1
2
3
4
5
6
7
8
9
time (h)
bLG+Ru(benz)
bLG+Ru(benz) (ee)
bLG-DMCP+Ru(benz)
bLG-DMCP+Ru(benz) (ee)
Aiming at identifying amino acid residues or at least the part of sequence of LG-A involved
in the coordination to Ru(benz), two MS-based approaches i.e. MALDI-TOF and ESI-MS/MS
were attempted. No ruthenated peptides were detected by MALDI-TOF analysis performed
after disulphide bond reduction by dithiotreitol (DTT) and denaturation steps followed by
trypsin digestion (data not shown). It appeared that the ruthenated protein did not
withstand the conditions used for this type of analysis which could be due to
reductive/denaturing treatment as well as too high disrupting energy dispensed by laser as
observed elsewhere for small proteins (insulin–platinum complex).[38] Such results may
reflect the relative weakness of the bond(s) between the protein and the metal ion to be
expected if the protein coordinates to the ruthenium centre in a monodentate fashion.[39]
Such an issue has been recently pointed out by Michelucci and coll.[40] Moreover, attempts
to crystallize ruthenated LG-A or to introduce the metallic species in LG single crystals by
diffusion were also unsuccessful (data not shown).
To overcome the drawback of laser used for ionization process, ESI-top-down analysis was
performed aiming at MS/MS sequencing of ions of the whole protein. Such a strategy has
been successfully applied to determine the copper binding sites of chloroplastic protein,[41]
as well as the iridium[42] and the platinum[43] binding sites of calmodulin. The binding sites of
half-sandwich ruthenium(II) complexes on angiotensin and bombesin[44] were investigated
by the same technique as well as platinum binding sites on copper metallo-chaperone,[45]
insulin,[46] and ubiquitin[47]
.
16
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Thus, to favour residues accessibility upon fragmentation, a reduction step of LG-A was
carried out by DTT (0.5 mM, 1 h at 37°C). More intense, but similar multicharged ions were
obtained before (Fig. S10A) and after treatment of native LG-A (Fig. S10B), since cleavage
of the two disulphide bridges yields a neutral mass shift of only +4.0312. However, similar
treatment on LG-A/Ru(Benz) 1:1 yielded the complete dissociation of the Ru(benz) moiety
(Fig. S10C-D). Modifications of reduction conditions, i.e. temperature and/or incubation
time, from 37°C to room temperature and 1h to 2h, respectively, led to similar results (Fig
S10D-E). Such results unambiguously demonstrated that LG-A/Ru(benz) complex was
sensitive to a reduction step, and that the proximity of the disulphide bridges could be
essential to maintain a minimal structure allowing the ligation of Ru(benz) as observed for
the small chloroplastic protein with copper.[41] Herein, loss of Ru(benz) can be due to a direct
action of DTT on Ru(benz) and/or protein unfolding, resulting in disulphide bridge cleavage,
and inducing modifications of the coordination sphere. MS/MS experiments on the 19+ ion
of LG-A without (m/z 967.50, Fig. S11A) or with Ru(Benz) (m/z 977.50, Fig. S11B), revealed
abundant and numerous fragments with some of them exhibiting a clear neutral mass shift
of +177.95 when comparing native LG-A to LG-A metallated with Ru(Benz). After
identification of these fragments, such results give unambiguous evidences that the
Ru(benz) entity was located in the N-terminal region of the protein. However, as no
reduction step was possible (see above) without triggering ruthenium dissociation, it was
not possible to precisely identify the residues(s) involved in metal coordination. Moreover, it
must be kept in mind that previous literature examples of mapping of metalation sites were
performed on proteins exhibiting shorter sequence (<100 amino acids, except calmodulin
with 158 residues) and lacking disulphide bridges. Clearly, ruthenated LG-A, because of its
folding, its larger size and its two disulphide bridges appeared to be more case sensitive to
study.
CONCLUSIONS
In summary, association of organometallic [M(arene)] entities to the proteins LG and HEWL
afforded artificial metalloenzymes with transfer hydrogenase activity on various aromatic
ketones. The most active, stereoselective and stable hybrid catalyst was obtained with M =
Ru, arene = benzene and LG as protein host. Additional experiments shed light on the
nature of the hybrid species. Both mass spectrometric analysis and ICP-OES disclosed that
treatment of LG with [Ru(benz)] 2+ gave a mixture of mono- and di-ruthenated adducts.
Interestingly, the crude hybrid catalyst was nearly as effective as the purified one. The
hybrid catalysts derived from HEWL were less selective, with the iridium adduct providing an
opposite stereoselectivity. Moreover, the HEWL•Ir(Cp*) adduct appeared to be unstable in
our conditions of transfer hydrogenation; as deduced from the progressive decrease of
enantiomeric excess recorded in the course of the reaction. Eventually, a top down mass
spectrometry approach using collision-induced dissociation fragmentation enabled to
narrow the ruthenium binding sites on beta-lactoglobulin to the N-terminal region. Finally,
our approach provided new efficient metalloenzymes with rather good catalytic properties.
17
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
These might be further improved by genetic optimisation of the protein host. Computational
methods[48] along with electrochemical studies could be undertaken to get further insight on
the binding site(s) and stability of metallated protein adducts with organometallic half-
sandwich metal complexes.
Experimental section
General methods
Dichloro (arene) ruthenium, rhodium and iridium dimers and dithiothreitol (DTT) were
purchased from Strem chemicals or Sigma-Aldrich (Saint-Quentin-Fallavier, France). βLG
(mixture of both isoforms 85% purity, or isoform A), hen egg white lysozyme,
diethylpyrocarbonate (DEPC), dimethylpyrocarbonate (DMPC) and sodium formate were
1
purchased from Sigma-Aldrich and used as received. H NMR spectra were recorded either
on 300 or 400 MHz spectrometers (Bruker). Chemical shifts are quoted in parts per million
(ppm) and referenced to residual H signal of HOD (4.78 ppm).
Synthesis of LG•M(arene)
A suspension of dichloro (arene) metal dimer in water was heated to ca. 50°C for a few
minutes until complete dissolution. A solution of LG (172 µM in water, 2 mL) was mixed
with a solution of hydrolysed M(arene) (5 mM metal in water) to make a M/LG ratio of 1:1,
3:1 or 6:1. After ca. 1h at RT, the mixture was submitted to gel filtration on a prepacked
column (Pierce dextran desalting column, 10 mL, Thermo-scientific) using water as eluent.
Twelve 1-mL fractions were collected manually and analysed at 280 nm. The fractions
containing the protein were pooled and the solution concentrated by centrifugal ultra-
filtration. Protein concentration was assayed at 280 nm taking  = 17,600 M-1.cm-1.
Methoxycarbonylation of LG
Beta-LG (8.5 mg) was dissolved in 20 mM phosphate pH 7.5 (2 mL). DMPC (1 µL) was added
to the solution. After 1h at RT, the mixture was submitted to gel filtration as above. Protein
concentration was assayed at 280 nm taking  = 17,600 M-1.cm-1.
Synthesis of HEWL•Ir(Cp*)
A solution of HEWL (248 µM in water, 1.5 mL) was mixed with a solution of hydrolysed
Ir(Cp*) complex (20 mM in DMSO/H2O 1/1) to make an Ir/HEWL ratio of 1:1, 5:1 or 10:1.
After ca. 2h at RT, the mixture was submitted to gel filtration on a prepacked column (Pierce
dextran desalting column, 10 mL, Thermo scientific) using 50 mM HEPES pH 7.8 as eluent.
Twelve 1-mL fractions were collected manually and analysed at 280 nm. The fractions
containing the protein were pooled were protein concentration was assayed at 280 nm
taking  = 38,000 M-1.cm-1.
Catalytic runs with hydrolysed M(arene) and TFACP
18
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
To 1 M HCOONa (1 mL, pH 7.5) were successively added TFACP (5 µmol, 0.7 µL) and the
metal complex (0.1 µmol metal out of a 5 mM aqueous solution). Mixtures were incubated
at 40°C for up to 33 h and analysed by HPLC to determine the conversion.
Catalytic runs with LG•M(arene) and TFACP
The solution of βLG•M(arene) (0.1 µmol) was completed to 1 mL with water. Sodium
formate (68 mg, 1 M final concentration) and TFACP (5 µmol, 0.7 µL) were successively
added. The mixture was incubated at 40°C for up to 33 h and analysed by chiral HPLC over
time to determine the conversion and the enantiomeric excess.
Catalytic runs with LG + Ru(benz) and other aryl ketones
To each ketone (10 µmol) dissolved in DMSO (50 µL) was successively added 1 M HCOONa (1
mL), βLG (0.1 µmol, 2.2 mg) and hydrolysed Ru(benz) (0.3 µmol metal out of a 5 mM
solution). The mixture was incubated at 40°C for 40 h. After cooling, the aqueous solution
was extracted with isopropyl ether containing 0.1% heptane (v/v) as internal standard (1 mL).
The organic phase was analysed by chiral GC to determine the conversion and the
enantiomeric excess.
Catalytic runs with HEWL•Ir(Cp*) and TFACP
The solution of HEWL•Ir(Cp*) (0.1 µmol) was completed to 1 mL with 50 mM HEPES pH 7.8.
Sodium formate (68 mg, 1 M final concentration) and TFACP (5 µmol, 0.7 µL) were
successively added. The mixture was incubated at 40°C for up to 24 h and analysed by chiral
HPLC to determine the conversion and the enantiomeric excess at various incubation times.
Catalytic runs with HEWL + hydrolysed M(L) and TFACP
To 1 M HCOONa (1 mL, pH 7.5) were successively added TFACP (10 µmol, 1.4 µL), HEWL (0.2
µmol, 3.2 mg) and hydrolysed M(L) complex (0.2 µmol out of a 5 mM aqueous solution). The
mixture was incubated at 40°C for 9 h and analysed by chiral HPLC to determine the
conversion and the enantiomeric excess.
Titration of LG with hydrolysed Ru(benz) monitored by fluorescence spectroscopy
Fluorescence spectra were recorded on a F-6200 spectrofluorometer (JASCO) at 20°C in a 1
cm-path length quartz cuvette between 300 and 430 nm (λex = 290 nm) with 5 nm excitation
and emission bandwidth, 1 nm data pitch, 125 nm/min scanning rate. An aqueous solution
of βLG (5.4 µM) was transferred to the quartz cuvette and small aliquots of hydrolysed
Ru(benz) (2 mM in H2O) were added sequentially to achieve up to 10 equivalents of
complex. The blank experiment was performed with a solution of N-acetyl-L-
tryptophanamide ethyl ester (10 µM) instead of βLG.
Mass spectrometry
ESI-MS experiments were carried out using a LTQ-Orbitrap XL from Thermo Scientific (San
Jose, CA, USA) and operated in positive ionization mode, with a spray voltage at 4.5 kV. LG-
A
at 5.2 or 10.4 µM under native or reduced forms, respectively, in
a
19
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
water/acetonitrile/formic acid 49/50/1 (v:v:v) mixture was continuously infused using a 250
µL syringe at 5 µL/min flow. Applied voltages were 40 and 100 V for the ion transfer capillary
and the tube lens, respectively. The ion transfer capillary was held at 275°C. Sheath and
auxiliary gas were set at a flow rate of 45 and 15 arbitrary units (a.u.), respectively. Applied
voltages were 20 and 70 V for the ion transfer capillary and the tube lens, respectively. The
ion transfer capillary was held at 275°C. Detection was achieved in the Orbitrap with a
resolution set to 100,000 (at m/z 400) and a m/z range between 200-2000 in profile mode.
Spectra were analysed using the acquisition software XCalibur 2.1 (Thermo Fisher Scientific,
Courtaboeuf, France) without smoothing and background subtraction. The automatic gain
control (AGC) allowed accumulation of up to 2.105 ions for FTMS scans, Maximum injection
time was set to 300 ms and 3 µscan were acquired. During MS/MS scans, the precursor
selection window was set to 2.8 Da and collision-induced dissociation (CID) of 20% was
applied with an activation Q value of 0.25 and an activation time of 30 ms. Fragmentation
occurred in the linear ion trap analyser (LTQ) and detection in the Orbitrap in profile mode.
The automatic gain control (AGC) allowed accumulation of up to 2.105 ions for FTMS scans,
1.105 ions for FTMSn scans and 1.104 ions for ITMSn scans. Maximum injection time was set
to 300 ms and 1000 ms for FTMS and FTMSn scans, respectively. For all scan modes, 3 µscan
was acquired.
ACKNOWLEDGEMENTS
The Agence Nationale de la Recherche (ANR) is gratefully acknowledged for financial support
(project “Artzymes”, grant number ANR-11-BS07-027-01). E. Aubry is acknowledged for
carrying out the ICP-OES measurements at the METIS laboratory of the UPMC.
Supporting information: Details of chromatographic procedures, NMR and mass spectra and
fluorescence titration studies can be found in the supporting information.
References
[1]
a) T. Heinisch, M. Pellizzoni, M. Durrenberger, C. E. Tinberg, V. Kohler, J. Klehr, D.
Haussinger, D. Baker, T. R. Wardt, Journal of the American Chemical Society 2015,
137, 10414-10419; b) O. Pamies, M. Dieguez, J. E. Backvall, Advanced Synthesis &
Catalysis 2015, 357, 1567-1586; c) J. Bos, G. Roelfes, Curr. Op. Chem. Biol. 2014, 19,
135-143; d) J. C. Lewis, ACS Catal. 2013, 3, 2954-2975; e) A. Onoda, T. Hayashi, M.
Salmain, in BIoorganometallic chemistry (Eds.: G. Jaouen, M. Salmain), Wiley-VCH,
Weinheim (D), 2015, pp. 305-337.
[2]
F. Schwizer, Y. Okamoto, T. Heinisch, Y. Gu, M. M. Pellizzoni, V. Lebrun, R. Reuter, V.
Kohler, J. C. Lewis, T. R. Ward, Chem. Rev. 2017.
[3]
[4]
Y. Lu, N. Yeung, N. Sieracki, N. M. Marshall, Nature 2009, 460, 855-862.
a) J. Bos, F. Fusetti, A. J. M. Driessen, G. Roelfes, Angew. Chem. Int. Ed. 2012, 51,
7472-7475; b) D. Coquiere, J. Bos, J. Beld, G. Roelfes, Angew. Chem. Int. Ed. 2009, 48,
5159-5162; c) P. J. Deuss, G. Popa, A. M. Z. Slawin, W. Laan, P. C. J. Kamer,
ChemCatChem 2013, 5, 1184 – 1191; d) W. Ghattas, L. Cotchico-Alonso, J. D.
20
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
Marechal, A. Urvoas, M. Rousseau, J. P. Mahy, R. Ricoux, Chembiochem 2016, 17,
433-440; e) T. Himiyama, D. F. Sauer, A. Onoda, T. P. Spaniol, J. Okuda, T. Hayashi, J.
Inorg. Biochem. 2016, 158, 55-61; f) J. Podtetenieff, A. Taglieber, E. Bill, E. J. Reijerse,
M. T. Reetz, Angew. Chem. Int. Ed. 2010, 49, 5151-5155.
a) T. K. Hyster, L. Knorr, T. R. Ward, T. Rovis, Science 2012, 338, 500-503; b) P. Dydio,
H. M. Key, A. Nazarenko, J. Y. E. Rha, V. Seyedkazemi, D. S. Clark, J. F. Hartwig, Science
2016, 354, 102-106.
a) P. Srivastava, H. Yang, K. Ellis-Guardiola, J. C. Lewis, Nature Comm. 2015, 6; b) H.
M. Key, P. Dydio, Z. Liu, J. Y. E. Rha, A. Nazarenko, V. Seyedkazemi, D. S. Cark, J. F.
Hartwig, ACS Central Sci. 2017, 3, 302-308; c) M. Bordeaux, V. Tyagi, R. Fasan, Angew.
Chem.-Int. Edit. 2015, 54, 1744-1748.
[5]
[6]
[7]
[8]
[9]
Y.-W. Lin, Coord. Chem. Rev. 2017, 336, 1-27.
K. Yamamura, E. T. Kaiser, J. Chem. Soc., Chem. Commun. 1976, 830-831.
a) Q. Jing, R. J. Kazlauskas, ChemCatChem 2010, 2, 953-957; b) Q. Jing, K. Okrasa, R. J.
Kazlauskas, Chem. Eur. J. 2009, 15, 1370-1376; c) K. Okrasa, R. J. Kazlauskas, Chem.
Eur. J. 2006, 12, 1587-1596; d) H. M. Key, D. S. Clark, J. F. Hartwig, J. Am. Chem. Soc.
2015, 137, 8261-8268.
[10] A. Pordea, Curr. Opin. Chem. Biol. 2015, 25, 124-132.
[11] N. Fujieda, J. Schatti, E. Stuttfeld, K. Ohkubo, T. Maier, S. Fukuzumi, T. R. Ward, Chem.
Sci. 2015, 6, 4060-4065.
[12] a) J. C. Lewis, Current Opinion in Chemical Biology 2015, 25, 27-35; b) I. Drienovska, A.
Rioz-Martinez, A. Draksharapu, G. Roelfes, Chem. Sci. 2015, 6, 770-776; c) H. S. Lee, P.
G. Schultz, J. Am. Chem. Soc. 2008, 130, 13194-13195.
[13] a) T. Kokubo, T. Sugimoto, T. Uchida, S. Tanimoto, M. Okano, J. Chem. Soc. Chem.
Commun. 1983, 769-770; b) C. Bertucci, C. Botteghi, D. Giunta, M. Marchetti, S.
Paganelli, Adv. Synth. Catal. 2002, 344, 556-562; c) A. Mahammed, Z. Gross, J. Am.
Chem. Soc. 2005, 127, 2883-2887; d) P. Rousselot-Pailley, C. Bochot, C. Marchi-
Delapierre, A. Jorge-Robin, L. Martin, J. C. Fontecilla-Camps, C. Cavazza, S. Menage,
ChemBioChem 2009, 10, 545-552.
[14] a) S. Abe, J. Niemeyer, M. Abe, Y. Takezawa, T. Ueno, T. Hikage, G. Erker, Y.
Watanabe, J. Am. Chem. Soc. 2008, 130, 10512-10514; b) Y. Takezawa, P. Bockmann,
N. Sugi, Z. Wang, S. Abe, T. Murakami, T. Hikage, G. Erker, Y. Watanabe, S. Kitagawa,
T. Ueno, Dalton Trans. 2011, 40, 2190–2195; c) S. Abe, K. Hirata, T. Ueno, K. Morino,
N. Shimizu, M. Yamamoto, M. Takata, E. Yashima, Y. Watanabe, J. Am. Chem. Soc.
2009, 131, 6958-6960.
[15] H. Tabe, S. Abe, T. Hikage, S. Kitagawa, T. Ueno, Chem.-Asian J. 2014, 9, 1373-1378.
[16] a) V. Kohler, J. Mao, T. Heinisch, A. Pordea, A. Sardo, Y. M. Wilson, L. Knorr, M. Creus,
J.-C. Prost, T. Schirmer, T. R. Ward, Angew. Chem. Int. Ed. 2011, 50, 10863-10866; b)
A. Pordea, M. Creus, J. Panek, C. Duboc, D. Mathis, M. Novic, T. R. Ward, J. Am. Chem.
Soc. 2008, 130, 8085-8088.
[17] X. Wu, C. Wang, J. Xiao, Platinum Metals Rev. 2010, 54, 3-19.
[18] a) A. Bogevig, I. M. Pastor, H. Adolfsson, Chem. Eur. J. 2004, 10, 294-302; b) I. M.
Pastor, P. Vastila, H. Adolfsson, Chem. Eur. J. 2003, 9, 4031-4045; c) J. Wettergren, E.
Buitrago, P. Ryberg, H. Adolfsson, Chem. Eur. J. 2009, 15, 5709-5718; d) J.
Wettergren, A. B. Zaitsev, H. Adolfsson, Adv. Synth. Catal. 2007, 349, 2556-2562.
[19] C. Mayer, D. Hilvert, Eur. J. Org. Chem. 2013, 3427-3431.
21
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
[20] a) M. Durrenberger, T. Heinisch, Y. M. Wilson, T. Rossel, E. Nogueira, L. Knorr, A.
Mutschler, K. Kersten, M. J. Zimbron, J. Pierron, T. Schirmer, T. R. Ward, Angew.
Chem. Int. Ed. 2011, 50, 3026-3029; b) A. Pordea, M. Creus, C. Letondor, A. Ivanova,
T. R. Ward, Inorg. Chim. Acta 2010, 363, 601-604; c) M. Creus, A. Pordea, T. Rossel, A.
Sardo, C. Letondor, A. Ivanova, I. Le Trong, R. E. Stenkamp, T. R. Ward, Angew. Chem.
Int. Ed. 2008, 47, 1400-1404; d) C. Letondor, N. Humbert, T. R. Ward, Proc. Natl.
Acad. Sci. USA 2005, 102, 4683-4887; e) C. Letondor, A. Pordea, N. Humbert, A.
Ivanovna, S. Mazurek, M. Novic, T. R. Ward, J. Am. Chem. Soc. 2006, 128, 8320-8328.
[21] F. W. Monnard, E. S. Nogueira, T. Heinisch, T. Schirmer, T. R. Ward, Chem. Sci. 2013,
4, 3269-3274.
[22] a) N. Madern, N. Queyriaux, A. Chevalley, M. Ghasemi, O. Nicolotti, I. Ciofini, G. F.
Mangiatordi, M. Salmain, J.. Mol. Cat. B 2015, 122, 314-322; b) N. Madern, B. Talbi,
M. Salmain, Appl. Organomet. Chem. 2013, 27, 6-12.
[23] a) M. V. Cherrier, S. Engilberge, P. Amara, A. Chevalley, M. Salmain, J. C. Fontecilla-
Camps, Eur. J. Inorg. Chem. 2013, 3596-3600; b) A. Chevalley, M. V. Cherrier, J. C.
Fontecilla-Camps, M. Ghasemi, M. Salmain, Dalton Trans. 2014, 43, 5482 - 5489; c) A.
Chevalley, M. Salmain, Chem. Commun. 2012, 48, 11984-11986.
[24] G. Suss-Fink, J. Organomet. Chem. 2013, 751, 2-19.
[25] S. K. Chowdhury, V. Katta, R. C. Beavis, B. T. Chait, J. Am. Soc. Mass Spectrom. 1990,
1, 382-388.
[26] a) J. Leonil, D. Molle, J. Fauquant, J. L. Maubois, R. J. Pearce, S. Bouhallab, J. Dairy Sci.
1997, 80, 2270-2281; b) J. Metltretter, J. Wurst, M. Pischetrieder, J. Agric. Food
Chem. 2013, 61, 6971-6981.
[27] S. Ogo, N. Makihara, Y. Watanabe, Organometallics 1999, 18, 5470-5474.
[28] a) K. Everaere, A. Mortreux, M. Bulliard, J. Brussee, A. V. d. Gen, G. Nowogrocki, J. F.
Carpentier, Eur. J. Org. Chem. 2001, 275–291; b) K. J. Haack, S. Hashiguchi, A. Fujii, T.
Ikariya, R. Noyori, Angew. Chem., Int. Ed. Engl. 1997, 36, 285–288; c) F. K. Cheung, A.
J. Clarke, G. J. Clarkson, D. J. Fox, M. A. Graham, C. Lin, A. Lorente Crivillé, M. Wills,
Dalton Trans. 2010, 39, 1395–1402; d) N. Pannetier, J. B. Sortais, P. S. Dieng, L.
Barloy, C. Sirlin, M. Pfeffer, Organometallics 2008, 27, 5852–5859; e) S. Y. Liu, L. Y.
Xu, C. Y. Liu, Z. G. Ren, D. J. Young, J. P. Lang, Tetrahedron 2017, 73, 2374–2381.
[29] a) I. W. McNae, K. Fishburne, A. Habtemariam, T. M. Hunter, M. Melchart, F. Wang,
M. D. Walkinshaw, P. J. Sadler, Chem. Commun. 2004, 1786-1787; b) M. P. Sullivan,
M. Groessl, S. M. Meier, R. L. Kingston, D. C. Goldstone, C. G. Hartinger, Chem.
Commun. 2017, 53, 4246-4249.
[30] A. Casini, G. Mastrobuoni, W. H. Ang, C. Gabbiani, G. Pieraccini, G. Moneti, P. J.
Dyson, L. Messori, ChemMedChem 2007, 2, 631-635.
[31] W. Hu, Q. Luo, X. Ma, K. Wu, J. Liu, Y. Chen, S. Xiong, J. Wang, P. J. Sadler, F. Wang,
Chem. Eur. J. 2009, 15, 6586-6594.
[32] Y. Lin, Y. Huang, W. Zheng, F. Wang, A. Habtemariam, Q. Luo, X. Li, K. Wu, P. J. Sadler,
S. Xiong, J. Inorg. Biochem. 2013, 128, 77-84.
[33] G. Kontopidis, C. Holt, L. Sawyer, J. Dairy Sci. 2004, 87, 785-796.
[34] E. Wilson Miles, Methods Enzymol. 1977, 47, 431-442.
[35] D. Mercadante, L. D. Melton, G. E. Norris, T. S. Loo, M. A. Williams, R. C. Dobson, G. B.
Jameson, Biophys. J. 2012, 103, 303-312.
[36] B. Y. Qin, L. K. Creamer, E. N. Baker, G. B. Jameson, FEBS Lett. 1998, 438, 272-278.
[37] K. C. Cheng, T. Nowak, J. Biol. Chem. 1989, 264, 19666-19676.
22
This article is protected by copyright. All rights reserved.

10.1002/ejic.201701359
European Journal of Inorganic Chemistry
[38] E. Moreno-Gordaliza, B. Cañas, M. A. Palacios, M. M. Gómez-Gómez, Anal. Chem.
2009, 81, 3507-3516.
[39] J. Will, A. Kyas, W. S. Sheldrick, D. Wolters, J. Biol. Inorg. Chem. 2007, 12, 883-894.
[40] E. Michelucci, G. Pieraccini, G. Moneti, C. Gabbiani, A. Pratesi, L. Messori, Talanta
2017, 167, 30-38.
[41] J. Erales, B. Gontero, J. Whitelegge, F. Halgand, Biochem. J. 2009, 419, 75-86.
[42] Y. Qi, Z. Liu, H. Li, P. J. Sadler, P. B. O'Connor, Rapid Commun. Mass Spectrom. 2013,
27, 2028-2032.
[43] H. Li, T.-Y. Lin, S. L. Van Orden, Y. Zhao, M. P. Barrow, A. M. Pizarro, Y. Qi, P. J. Sadler,
P. B. O’Connor, Anal. Chem. 2011, 83, 9507-9515.
[44] R. H. Wills, A. Habtemariam, A. F. Lopez-Clavijo, M. P. Barrow, P. J. Sadler, P. B.
O’Connor, J. Am. Soc. Mass Spectrom. 2014, 25, 662-672.
[45] C. M. Sze, Z. Shi, G. N. Khairallah, L. Feketeova, R. A. J. O'Hair, Z. Xiao, P. S. Donnelly,
A. G. Wedd, Metallomics 2013, 5, 946-954.
[46] H. L. Li, J. R. Snelling, M. P. Barrow, J. H. Scrivens, P. J. Sadler, P. B. O'Connor, J. Am.
Soc. Mass Spectrom. 2014, 25, 1217-1227.
[47] a) S. M. Meier, Y. O. Tsybin, P. J. Dyson, B. K. Keppler, C. G. Hartinger, Anal. Bioanal.
Chem. 2012, 402, 2655-2662; b) R. F. S. Lee, L. Menin, L. Patiny, D. Ortiz, P. J. Dyson,
Anal. Chem. 2017.
[48] D. Das, A. Dutta, P. Mondal, RSC Adv. 2014, 4, 60548-60556.
23
This article is protected by copyright. All rights reserved.