Second generation artificial hydrogenases based on the biotin-avidin technology: Improving activity, stability and selectivity by introduction of enantiopure amino acid spacers

Article abstract of DOI:10.1002/adsc.200700022

We report on our efforts to create efficient artificial metalloenzymes for the enantioselective hydrogenation of N-protected dehydroamino acids using either avidin or streptavidin as host proteins. Introduction of chiral amino acid spacers - phenylalanine or proline - between the biotin anchor and the flexible aminodiphosphine moiety 1, combined with saturation mutagenesis at position S112X of streptavidin, affords second generation artificial hydrogenases displaying improved organic solvent tolerance, reaction rates (3-fold) and (S)-selectivities (up to 95% ee for N-acetamidoalanine and N-acetamidophenylalanine). It is shown that these artificial metalloenzymes follow Michaelis-Menten kinetics with an increased affinity for the substrate and a higher k_cat than the protein-free catalyst (compare k _cat 3.06 min^-1 and K_M 7.38 mM for [Rh(COD)Biot-1]⁺ with k_cat 12.30 min^-1 and K_M 4.36 mM for [Rh-(COD)Biot-(R)-Pro-1]⁺ ? WT Sav). Finally, we present a straightforward protocol using Biotin-Sepharose to immobilize artificial metalloenzymes (> 92% ee for N-acetamidoalanine and N-acetamidophenylalanine using [Rh(COD)Biot-(R)-Pro-1]⁺ ? Sav S112W).

Full text of DOI:10.1002/adsc.200700022

FULL PAPERS
DOI: 10.1002/adsc.200700022
Second Generation Artificial Hydrogenases Based on the Biotin-
Avidin Technology: Improving Activity, Stability and Selectivity
by Introduction of Enantiopure Amino Acid Spacers
a
a
a
a
a
Untung E. Rusbandi, Cheikh Lo, Myriem Skander, Anita Ivanova, Marc Creus,
a
a,
Nicolas Humbert, and Thomas R. Ward *
Institute of Chemistry, University of Neuchâtel, Av. Bellevaux 51, CP 158, 2009 Neuchâtel, Switzerland
a
Fax : (+41)-32-718-2511; e-mail: thomas.ward@unine.ch
Received: January 15, 2007
Supporting information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
Abstract: We report on our efforts to create efficient with an increased affinity for the substrate and a
artificial metalloenzymes for the enantioselective hy- higher kcat than the protein-free catalyst (compare
ꢀ
1
drogenation of N-protected dehydroamino acids
k
3.06 min and KM 7.38 mM for [Rh
A
H
R
U
G
cat
+
ꢀ1
using either avidin or streptavidin as host proteins. 1] with kcat 12.30 min+ and KM 4.36 mM for [Rh-
Introduction of chiral amino acid spacers – phenyl-
A
H
R
U
G
alanine or proline – between the biotin anchor and present a straightforward protocol using Biotin-Se-
the flexible aminodiphosphine moiety 1, combined pharose to immobilize artificial metalloenzymes
with saturation mutagenesis at position S112X of (>92% ee for N-acetamidoalanine and N-acetami-
+
streptavidin, affords second generation artificial hy- dophenylalanine using [Rh
drogenases displaying improved organic solvent tol- Sav S112W).
erance, reaction rates (3-fold) and (S)-selectivities
A
H
R
U
G
(
up to 95% ee for N-acetamidoalanine and N-acet- Keywords: artificial metalloenzymes; biotin-avidin;
amidophenylalanine). It is shown that these artificial chemogenetic optimization; enantioselective cataly-
metalloenzymes follow Michaelis–Menten kinetics sis; enzyme immobilization; hydrogenation
[
20–28]
[29,30]
Introduction
nation,
Alder reactions.
Based on Whitesidesꢀ early report, several groups
transfer hydrogenation
and Diels–
[
31–33]
[
20]
Today, the preparation of enantiomerically pure com-
pounds is one of the most actively pursued fields in have been exploiting the biotin-avidin technology to
synthetic chemistry. In this major effort, both homo- produce artificial hydrogenases for the enantioselec-
geneous and enzymatic catalysis occupy a place of tive reduction of N-protected dehydroamino
[
1–3]
[21–26,28]
choice.
To complement these well established acids
as well as the reduction via transfer hy-
[29,30]
methods, organocatalysis and artificial metalloen- drogenation of aromatic ketones.
zymes have witnessed a revival in recent years.
[
4]
For optimization purposes, our group relies both on
The underlying principle of artificial metalloen- chemical and on genetic strategies (i.e., chemogenet-
[
5]
zymes for enantioselective catalysis is based on the in- ic) to yield both (R)- and (S)-selective hydrogenases.
corporation of an active catalyst within a macromo- For the proof-of-principle, we focused exclusively on
lecular host (protein or oligonucleotide) which pro- achiral ligand scaffolds and spacers, thus ensuring that
vides a chiral environment, responsible for the selec- (enantio)selectivity is provided by second coordina-
[5–10]
tivity.
To ensure localization of the catalytic tion sphere interactions between the catalyst–sub-
moiety within the macromolecular host, covalent, strate moiety and the host protein. Having produced,
dative and supramolecular anchoring strategies have purified and screened several streptavidin (Sav here-
successfully been exploited to produce enantioselec- after) mutants, we identified position S112 as a prom-
[
11]
tive hybrid catalysts for ester hydrolysis, dihydroxy- ising mutation site to perform saturation mutagene-
[12]
[13,14]
[15–19]
[23]
lation, epoxidation,
sulfoxidation,
hydroge- sis. Combining the twenty Sav S112X with eighteen
achiral biotinylated diphosphine ligands afforded ver-
Adv. Synth. Catal. 2007, 349, 1923 – 1930
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
1923

Untung E. Rusbandi et al.
FULL PAPERS
Herein we report on our efforts to produce more
active and selective artificial hydrogenases based on
the biotin-avidin technology, with an emphasis on the
use of enantiopure a-amino acid spacers (Scheme 1).
We show that this second generation of hybrid cata-
lysts displays increased stability towards organic sol-
vents, increased selectivity and increased reaction
rates. In addition, we show that, by exploiting the tet-
rameric nature of (strept)avidin [hereafter, (strept)-
avidin refers to either avidin or streptavidin], such ar-
tificial metalloenzymes can be immobilized using
commercially available Biotin-Sepharose with little
erosion of their performance. A preliminary account
[
26]
of this work has appeared recently.
Scheme 1. Second generation artificial metalloenzymes
based on the biotin-avidin technology for the reduction of
N-acetamidodehydroamino acids. The host protein [(strept)-
avidin] displays high affinity for the anchor (biotin); intro-
duction of an enantiopure a-amino acid spacer, combined
with a flexible diphosphine ligand allows us to chemically
optimize both the activity and the selectivity. Site-directed
mutagenesis further enables a genetic optimization of the
host protein to afford enantioselective artificial hydrogenas-
es.
Results and Discussion
Chemogenetic Diversity
In previous studies, we showed that the ligand scaf-
fold plays a critical role on the performance of the re-
sulting artificial hydrogenases [Rh(COD)(Biot-
A
H
R
U
G
+
spacer-diphosphine)] ꢁ (strept)avidin. For this study,
we used exclusively the most versatile ligand scaffold
1
and focused on the influence of the amino acid
satile artificial hydrogenases displaying moderate to spacer (Scheme 2).
good enantioselectivities for both enantiomers of N-
In order to exploit more efficiently the interactions
protected amino acids (N-acetamidoalanine, N- between the spacer and the (strept)avidin host pro-
[
34]
AcAla, and N-acetamidophenylalanine, N-AcPhe).
tein, phenylalanine and proline spacers were selected.
Scheme 2. Biotinylated ligands used in this study and standard operating conditions for the hydrogenation of a cocktail mix-
ture containing a-acetamidoacrylic acid and a-acetamidocinnamic acid (50 equivalents of each with respect to the biotinylat-
ed ligand).
1924
asc.wiley-vch.de
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1923 – 1930

Second Generation Artificial Hydrogenases Based on the Biotin-Avidin Technology
FULL PAPERS
Scheme 3. Fingerprint display of the results of the chemogenetic optimization for the reduction of a-acetamidoacrylic acid
top triangle) and a-acetamidocinnamic acid (bottom triangle).
(
We speculated that the phenylalanine spacer might
The results of the chemogenetic optimization are
display stabilizing p–p interactions upon interaction summarized in Scheme 3 using a fingerprint display
[25,37]
with the hydrophobic (strept)avidin binding pockets, for each substrate-protein-ligand combination.
which possess four or five aromatic residues for strep- The selectivity is colour-coded [green for (R)-selectiv-
tavidin and avidin, respectively. We reasoned that a ity and pink for (S)-selectivity] and the intensity of
proline moiety would restrict the degrees of freedom the colour codes for the conversion. Both substrates
of the biotinylated ligand, thus contributing to local- are displayed as two hypothenuse-sharing triangles to
ize the catalytic moiety within a well defined position allow a rapid identification of interesting ligand-pro-
in (strept)avidin.
tein combinations. Selected results are collected in
Both enantiomers of the Boc-protected amino acid Table 1, and a summary of all catalytic experiments is
spacers were coupled to the aminodiphosphine provided in the Supporting Information.
moiety 1. Following deprotection, (+)-biotin was in-
Inspection of the fingerprint (Scheme 3), reveals
troduced to afford Biot-(R)-Pro-1, Biot-(S)-Pro-1, several interesting trends.
Biot-(R)-Phe-1, Biot-(S)-Phe-1. This N!C coupling
i) The absolute configuration of the amino acid
scheme avoids epimerization – via formation of an in- spacer by-and-large determines the absolute configu-
termediate oxazolone – upon the second peptide cou- ration of the reduction product. Both Biot-(R)-Pro-1
pling step. Full experimental details for the ligands and Biot-(S)-Phe-1 favour (S)-reduction products,
synthesis and spectroscopic characterization can be while Biot-(S)-Pro-1 and Biot-(R)-Phe-1 produce
[26]
found in the preliminary account of this work.
These ligands were reacted in situ with [Rh-
preferentially (R)-reduction products.
Combining Biot-(R)-Pro-1 with Sav mutants bear-
A
C
H
T
R
E
U
N
G
(COD) ]BF and combined either with WT avidin or ing aromatic residues at position S112X affords the
2 4
with the twenty isoforms of streptavidin obtained by most (S)-selective artificial hydrogenases (up to 95%
saturation mutagenesis at position S112X. The result- ee for both substrates, Table 1, entries 3 and 4). It
ing artificial hydrogenases [Rh(COD)(Biot-spacer- should be emphasized that, in the absence of (strep-
A
H
R
N
+
1
)] ꢁ (strept)avidin were screened with a cocktail t)avidin, the biotinylated catalysts derived from the
mixture containing a-acetamidoacrylic acid and a- enantiopure ligands afford nearly racemic products
acetamidocinnamic acid (fifty equivalents of each (<10% ee in all cases).
with respect to the biotinylated ligand), Scheme 2.
ii) Depending on the amino acid residue at position
Control experiments demonstrated that both the se- S112X, the hybrid catalysts derived from Biot-(S)-
lectivity and the conversion obtained with the cocktail Phe-1 can afford both (R)- and (S)-reduction prod-
+
mixture are identical to those obtained with a single ucts. For example, [Rh
A
H
R
U
G
[
35,36]
+
substrate (i.e., no autoinduction).
S112M and [Rh
A
H
R
U
G
Adv. Synth. Catal. 2007, 349, 1923 – 1930
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
asc.wiley-vch.de
1925

Untung E. Rusbandi et al.
FULL PAPERS
Table 1. Numerical summary of selected results of the catalytic experiments.
[
a]
[a]
Entry
Ligand
Protein
ee of
Conversion of
ee of
Conversion of
N-AcPhe
N-AcPhe
N-AcAla
N-AcAla
[
b]
[b]
1
2
3
4
5
6
7
8
9
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Phe-1
Biot-(R)-Phe-1
Biot-(R)-Phe-1
Biot-(R)-Phe-1
Biot-(R)-Phe-1
Biot-(S)-Phe-1
Biot-(S)-Phe-1
WT Avi
S112K
S112W
S112Y
S112I
ꢀ89
ꢀ84
ꢀ95
ꢀ93
74
74
75
76
82
quant.
36
quant.
97
80
88
94
82
75
ꢀ87
ꢀ88
ꢀ95
ꢀ92
84
83
86
81
84
quant.[
b]
quant.
[b]
[b]
quant.[
b]
quant.
[
b]
b]
b]
b]
b]
b]
quant.[
S112L
quant.
[
S112M
S112Q
S112R
S112H-78
S112M
quant.[
quant.[
quant.[
1
1
0
1
65
-87
quant.
[b]
[b]
87
quant.
73
quant.
[
[
a]
ee in%, positive ee values in favour of the (R)-enantiomer, negative ee values in favour of the (S)-enantiomer.
Quantitative; all listed experiments were performed at least in duplicate (see Supporting Information); N-AcPhe=N-
acetamidophenylalanine, N-AcAa=N-acetamidoalanine.
b]
yield quantitatively N-AcAla in 73% ee (R) and 87% S112X Sav generates more diversity than previously
ee (S), respectively (Table 1, entries 10 and 11). For a observed with achiral spacers.
single mutation, this represents a difference in transi-
°
ꢀ1
tion state free energy dDG >2.6 kcalmol at room
temperature. Irrespective of the absolute configura- Organic Solvent Tolerance and Catalyst
tion of the phenylalanine spacer, hybrid catalysts de- Immobilization
rived from S112F, S112L, S112M, S112W and S112Y
streptavidin mutants afford preferentially the same One of the main limitations of enantioselective cataly-
enantiomer. For example, both [Rh
ACHTREUNG
+
Phe-1] ꢁ Sav S112M and [Rh
ACHTREUNG
+
1
] ꢁ Sav S112M yield quantitatively (R)-N-AcAla in erance should thus contribute to broaden the sub-
86% ee and 73% ee, respectively (Table 1, entries 7 strate scope of these artificial metalloenzymes.
and 11). We speculate that the presence of a bulky
Having identified both (R)- and (S)-selective artifi-
residue at position S112 outrules the preference im- cial hydrogenases (ee >90%), we proceeded to test
posed by the enantiopure Phe-spacer. This may be these hydrogenases in the presence of large amounts
achieved either by inverting the preferred enantioen- of organic co-solvents. For this purpose, dimethyl sulf-
riched ligand conformation or by forcing the prochiral oxide (DMSO, miscible with water) and ethyl acetate
substrate to present its si-face to the biotinylated cata- (EtOAc, non-miscible with water) were selected. The
lyst.
results for the screening in the presence of the most
iii) Both substrates afford comparable enantioselec- selective catalyst ꢁ protein combinations are summar-
tivities. However, slightly better conversions are gen- ized in Table 2.
erally obtained for the smaller product N-AcAla. Sig-
As can be appreciated from these data, the catalyst
nificant substrate discrimination is found for catalysts derived from Biot-1 does not tolerate large amounts
derived from Biot-(R)-Pro-1. For example, [Rh- of either DMSO or EtOAc (Table 2, entries 1 and
+
[26]
A
C
H
T
R
E
U
N
G
(COD)Biot-(R)-Pro-1] ꢁ Sav S112K affords quanti- 2).
Introduction of an enantiopure spacer confers
tatively (S)-N-AcAla (88% ee), while the conversion significant organic-solvent tolerance in the presence
for the larger (S)-N-AcPhe remains modest (36% of streptavidin and its mutants. Both (S)- and (R)-se-
+
conversion, 84% ee, Table 1, entry 2)
iv) Combining Biot-(R)-Pro-1 with WT Avi affords, S112W and [Rh
for the first time, good levels of enantioselection perform well in the presence of either 45% DMSO or
89% ee for N-AcPhe quantitative, Table 1 entry 1). EtOAc (maximal ee erosion 12%, Table 2, entries 12
This may be of interest for future developments as and 13, 15 and 16). In contrast to streptavidin iso-
WT Avi is significantly cheaper that WT Sav. forms as host proteins, avidin-based artificial metal-
In summary and in comparison to our previous loenzymes perform poorly in the presence 45%
lective catalysts [Rh
A
H
R
U
G
+
A
H
R
U
G
(
[
22–25,34]
studies relying on achiral spacers,
the introduc- DMSO (Table 2, compare entries 4/5 and 8/9). In the
tion of enantiopure spacers yields higher (S)-selective presence of either 27% DMSO or EtOAc, the ero-
+
artificial metalloenzymes. In addition, the genetic op- sion observed with [Rh
timization by saturation mutagenesis at position WT-Avi is minimal (Table 2, entries 8 and 10).
A
H
R
U
G
ꢁ
1926
asc.wiley-vch.de
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1923 – 1930

Second Generation Artificial Hydrogenases Based on the Biotin-Avidin Technology
FULL PAPERS
Table 2. Numerical summary of selected results of the catalytic experiments performed in the presence of organic solvents or
with immobilized artificial metalloenzymes.
[
a]
[a]
Entry
Ligand
Protein
Organic solvent
ee of
Conversion of
ee of
Conversion of
N-AcPhe
N-AcPhe
N-AcAla
N-AcAla
[
[
b]
b]
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
Biot-1
Biot-1
Biot-1
WT Sav
WT Sav
WT Sav
WT Sav
WT Sav
WT Sav
WT Sav
WT Avi
WT Avi
WT Avi
WT Avi
S112W
S112W
S112W
S112M
45% DMSO
EtOAc [
24
31
93
9
5
16
30
83
71
56
d]
[c]
Immob.
78
100
94
85
36
94
34
77
32
56
86
89
68
78
50
quant.[
c]
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(R)-Pro-1
Biot-(S)-Phe-1
Biot-(S)-Phe-1
Biot-(S)-Phe-1
27% DMSO
45% DMSO
EtOAc [
ꢀ91
ꢀ86
ꢀ87
ꢀ89
ꢀ82
ꢀ15
ꢀ89
ꢀ80
ꢀ88
ꢀ94
ꢀ94
75
ꢀ87
ꢀ87
ꢀ83
ꢀ76
ꢀ86
ꢀ37
ꢀ87
ꢀ76
ꢀ85
ꢀ94
ꢀ92
69
quant.
[
[
b]
b]
[c]
quant.
90
76
d]
Immob.
[
c]
c]
c]
27% DMSO
45% DMSO
EtOAc [
quant.[
quant.
[
0
1
2
3
4
5
6
7
quant.
d]
[c]
Immob.
quant.[
c]
c]
45% DMSO
EtOAc [
quant.
quant.
[
d]
[c]
Immob.
quant.[
c]
c]
45% DMSO
EtOAc [
quant.
[
S112M
S112M
81
81
64
33
quant.
quant.
d]
[c]
Immob.
[
[
[
[
a]
ee in%, positive ee values in favour of the (R)-enantiomer, negative ee values in favour of the (S)-enantiomer.
Data from ref.
Quantitative.
b]
[26]
c]
d]
Immobilized using Biotin-Sepharose; all listed experiments were performed at least in duplicate (See Supporting Infor-
mation); N-AcPhe=N-acetamidophenylalanine, N-AcAla=N-acetamidoalanine.
In the past, we have shown that an artificial metal- offers a straightforward work-up procedure. In addi-
loenzymeꢀs selectivity is little affected upon varying tion, a second catalytic run performed with this recy-
+
the ratio of [Rh
tween one and four, suggesting that the occupancy of similar conversions and enantioselectivity compared
one binding site does not significantly affect neigh- to the first catalytic run. For [Rh(COD)Biot-(R)-Pro-
A
C
H
T
R
E
U
N
G
(COD)Biot-1] versus WT Sav be- cled immobilized artificial metalloenzyme afforded
AHCTREUNG
[24]
+
bouring sites. As streptavidin is a homo-tetrameric 1] ꢁ Sav S112W) with a-acetamidocinnamic acid, ee
protein, we thus speculated that it may be possible to 94% (S), conversion 81%, compare with Table 2
exploit one binding site to immobilize the artificial entry 14. The third catalytic run however afforded a
metalloenzyme, leaving three sites for the incorpora- significantly lower yield and enantioselectivity (ee
tion of a biotinylated catalyst.
<30%, conversion <25%).
For this purpose, one equivalent of commercially
available Biotin-Sepharose was added to tetrameric
(
strept)avidin prior to the addition of an excess bio- Michaelis–Menten Kinetics
tinylated catalyst precursor. After decanting the im-
mobilized catalyst precursors, thus removing the Having identified efficient artificial metalloenzymes,
excess biotinylated rhodium present in solution, these we proceeded to determine the Michaelis–Menten pa-
were tested using the standard reaction conditions. rameters for four catalytic systems: [Rh
ACHTREUNG
+
+
Selected results are collected in Table 2. As can be 1] , [Rh
appreciated, the immobilized artificial metalloen- (R)-Pro-1] ꢁ WT Sav and [Rh
A
H
R
U
G
ACHTREUNG
+
ACHTREUNG
+
zymes are most often nearly as selective as their ho- 1] ꢁ WT-Avi. In order to test the influence of the or-
mogeneous counterparts. However, the activity de- ganic co-solvent, catalysis with the latter two hybrid
creases, especially for the larger a-acetamidocinnamic catalysts was performed in 30% DMSO, thus ensuring
acid substrate. The most versatile immobilized hybrid good levels of selectivity both for streptavidin and for
+
catalyst is [Rh
A
C
H
T
R
E
U
N
G
(COD)Biot-(R)-Pro-1] ꢁ Sav S112W avidin based systems (compare entries 4/5 and en-
which is as active and as selective as its homogeneous tries 8/9, respectively, in Table 2). For convenience,
counterpart (compare Table 1 entry 3 and Table 2 the reactions were carried out at 2 bar hydrogen pres-
entry 14).
sure and room temperature. For solubility reasons,
In the context of artificial metalloenzymes, recover- only the more soluble substrate N-acetamidoacrylic
ing the precious protein via a simple filtration step
Adv. Synth. Catal. 2007, 349, 1923 – 1930
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
asc.wiley-vch.de
1927

Untung E. Rusbandi et al.
FULL PAPERS
acid was used. The results are displayed in Figure 1
and summarized in Table 3.
These data reveal several noteworthy features:
i) All catalytic systems display Michaelis–Menten
behaviour.
iv) Both avidin and streptavidin host proteins dis-
play similar kinetic behaviour (Table 3, entries 3 and
4).
v) The presence of large quantities of DMSO con-
tributes to significantly increase the reaction rates
ii) Compared to the protein-free catalyst (Table 3, (Table 3, entries 3 and 4).
entry 1) all artificial metalloenzymes display higher
It thus appears that both the affinity of the catalyst
affinity for the substrate (i.e., smaller K , Table 3, en- for the substrate and the turnover rates are increased
M
tries 2–4).
with respect to the protein-free reference system [Rh-
+
iii) Compared to the protein-free catalyst (Table 3,
A
H
R
U
G
entry 1), all artificial metalloenzymes display higher ence of a hydrophobic cavity provided by the protein,
pseudo-first-order reaction rates (Table 3, entries 2– which contributes to increase the local concentration
4).
of both the substrate and of the dihydrogen close to
the catalytic moiety. Accumulation within the host
protein of both substrate and dihydrogen are reflected
in lower K and higher k , provided that the H oxi-
M
cat
2
[
38]
dative addition is the rate-determining step.
Conclusions
This study, based on the incorporation of enantiopure
amino acid spacers between the biotin anchor and the
diphosphine moiety, considerably extends the poten-
tial of artificial hydrogenases based on the biotin
avidin technology. The most noteworthy features re-
vealed herein include:
i) Both avidin and streptavidin can be used as host
proteins to afford selective hybrid catalysts. For the
first time, egg-white avidin – which is considerably
cheaper than streptavidin – affords reduction products
with ee >86% in favour of the (S)-enantiomer using
+
[
Rh
A
H
R
U
G
netic optimization with streptavidin mutants, affords
Figure 1. Saturation kinetics plots for the reduction of N-
hybrid catalysts for both (R)- and (S)-reduction prod-
+
acetamidoacrylic acid with [Rh
measured: squares), [Rh
(COD)Biot-1] ꢁ WT Sav (9%
DMSO, measured: triangles), [Rh
WT Sav (30% DMSO, measured: circles) and [Rh-
A
T
E
N
+
ucts with ees >90%: for example, [Rh(COD)Biot-1]
A
H
R
U
G
AHCTREUNG
+
+
Sꢁa vW. T Sav and [Rh
A
H
R
U
G
A
H
R
U
G
+
A
C
H
T
R
E
U
N
G
(COD)Biot-(R)-Pro-1] ꢁ WT Avi (30% DMSO, measured:
ii) In general, the nature of the spacer by-and-large
determines which enantiomer is preferentially pro-
duced. However, hybrid catalysts derived from Biot-
diamonds). All the fitted graphics are shown in solid lines.
(
S)-Phe-1, afford both (R)- and (S)-reduction prod-
ucts with ee >75%, depending on the nature of the
point mutation at position S112.
Table 3. Michaelis–Menten parameters for the reduction of
iii) Increased organic solvent tolerance: the use of
N-acetamidoacrylic acid with selected artificial hydrogenas- large amounts of DMSO (up to 45%) may allow one to
es at 2 bars H and room temperature.
solubilize hydrophobic substrates, while EtOAc allows a
straightforward work-up to recover the substrate.
iii) Immobilization using Biotin-Sepharose with
little erosion of the selectivity: again here, the work-
up is considerably facilitated.
2
Entry Ligand
Protein Vmax
kcat
KM
ꢀ
1
ꢀ1
[
mMmin ] (min ) [mM]
1
2
Biot-1
Biot-1
-
0.173
0.254
3.06
4.49
7.38
3.18
WT
Sav
WT
Sav
WT
Avi
iv) Michaelis–Menten behaviour with increased
rate compared to the biotinylated catalyst devoid of
3
4
Biot-(R)-
Pro-1
Biot-(R)-
Pro-1
0.695
0.634
12.30
11.23
4.36
4.80
(
strept)avidin.
Current efforts in this area are centred on the re-
duction of challenging dehydroamino acid substrates
with limited aqueous solubility) as well as increasing
(
1928
asc.wiley-vch.de
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1923 – 1930

Second Generation Artificial Hydrogenases Based on the Biotin-Avidin Technology
FULL PAPERS
the turnover numbers and turnover frequency of arti- pharose CL-4B {[105 mL, which corresponds to 0.031 mmol
biotin, capable of immobilizing 0.031 mmol tetrameric
strept)avidin], [Affiland, Belgium, suspension in TBS
buffer pH7.4, NaN 3 0.02% (w/v)]} was decanted and
washed with water to afford wet beads, which were resus-
pended in water (105 mL total volume). The protein solution
ficial hydrogenases based on the biotin-avidin tech-
nology.
(
Experimental Section
(
0.062 mmol, 2 equivs. vs. biotin sepharose, 300 mL) was
added. After ca. 10 min incubation, the immobilized protein
containing 0.031 mmol of protein, corresponding to 0.093
mmol, free active sites) was washed at least three times with
O to remove the unbound protein and flushed with nitro-
Ligand Synthesis and Characterization
(
All ligands used herein were reported (synthesis and full
characterization) in the supporting information of the pre-
H
2
[26]
liminary communication.
gen for three hours. The precatalyst solution in DMSO (100
mL, 0.062 mmol, 0.66 equivs. with respect to the free biotin-
binding sites) was added last and was incubated with the im-
mobilized protein for 10 min. The immobilized catalyst was
subjected to several cycles of washing and decanting. The
catalytic runs were performed according to the standard or-
bital shaking procedure.
Production and Purification of Streptavidin
The expression of mutant proteins Sav S112X has been de-
[34]
scribed elsewhere. WT Avi was provided by Belovo.
Hydrogenation Procedure for the Determination of Mi-
chaelis–Menten Parameters: A 15-mL glass vessel was
charged with a quantity (depending on the number of equiv-
alent of substrates used) a-acetamidoacrylic acid in MES
buffer (0.1M, pH5.5). The volume was adjusted with water
Hydrogenation Reactions
General Aspects: All aqueous and organic solvents were
flushed with nitrogen during at least three hours. All manip-
ulations were performed in a glove-box under a nitrogen at-
mosphere. The metal source [Rh
in DMSO (Acros) (4 mg, 16 mL DMSO), 1.56 mmol of this
A
H
R
N
(COD) ]BF was dissolved
to 9 mL followed by addition of (strept)avidin in H
O
2
2
4
(1 mL, 0.207 mM solution, 0.207 mmol of tetramer). The cat-
+
solution were added to the appropriate aliquoted ligand
alyst precursor [Rh
A
H
R
U
G
A
H
R
U
G
(
2.02 mmol, 1.3 equivs. with respect to the metal source).
mmol) was added last. The resulting mixture was stirred for
ca. 2 mins to ensure a homogeneous medium. The reactor
was closed and was removed from the glove box and pres-
This solution was stirred for ca. 10 min.
Hydrogenation in Monophasic Media (9%, 27%, 45%
DMSO): A Pyrex tube (volume ca. 3 mL) was placed in an
autoclave and charged with an N-acetamidoacrylic acid solu-
tion in MES buffer 0.38M (130 mL of a 23.85 mM solution,
surized to 2 bars H and stirred mechanically at room tem-
2
perature. Aliquots (ca. 0.5 mL) were sampled at given times.
The initial conversions were collected for each concentra-
tion of substrate. The initial rate of the reaction (n) can be
calculated by:
3
.10 mmol), and an N-acetamidocinnamic acid solution in
MES buffer 0.38M (130 mL of a 23.85 mM solution, 3.10
mmol). The volume was adjusted with water (or with
DMSO, depending on the ratio of DMSO to be used) to 900
mL. The protein solution in water was then added (100 mL
of a 0.207 mM solution, 0.0207 mmol of the tetramer). The
precatalyst solution in DMSO (100 mL, 0.062 mmol) was
added last. The resulting mixture was vortexed, removed
This value gives the observed initial rate. The calculated
initial rate can be obtained by employing the following
equation:
from the glove-box and hydrogenated at 5 bars H for 15 h.
2
Hydrogenation in Biphasic Media (EtOAc): A Pyrex tube
volume ca. 3 mL) was placed in an autoclave and charged
(
with an N-acetamidoacrylic acid solution in EtOAc (550 mL
of a 5.66 mM solution, 3.10 mmol), and an N-acetamidocin-
namic acid solution in EtOAc (550 mL of a 5.66 mM solu-
tion, 3.10 mmol). The buffer solution (580 mL of a 0.38M so-
lution) was added and the volume of the aqueous phase ad-
justed with water to 900 mL. The protein solution in water
was then added (100 mL of a 0.207 mM solution, 0.0207
mmol of the tetramer). The precatalyst solution in DMSO
(
100 mL, 0.062 mmol) was added last. The resulting mixture
was vortexed and hydrogenated at 5 bars H for 15 h using
2
orbital shaking (ca. 200 rpm).
Note: In the case of hydrogenation without protein, the
volume of the protein solution was replaced by H O.
2
All experiments were performed using MES buffer only.
This change does not influence significantly the activity and
enantioselectivity of experiments using avidin as the host
protein.
Hydrogenation Procedure for Immobilized Artificial Met-
alloenzymes: The commercial suspension of D-biotin-se-
The initial guessed values of both unknowns (KM and
Vmax) were introduced in the calculation. Then, a least
squares-minimization was performed to obtain the calculat-
ed initial rate n of the reaction and the Michaelis–Menten
parameters.
Adv. Synth. Catal. 2007, 349, 1923 – 1930
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
asc.wiley-vch.de
1929

Untung E. Rusbandi et al.
FULL PAPERS
Supporting Information
[16] M. Ohashi, T. Koshiyama, T. Ueno, M. Yanase, H.
Fujii, Y. Watanabe, Angew. Chem. Int. Ed. 2003, 42,
A summary of all catalytic runs is provided free of charge
via the internet or from the author.
1
005–1008.
[
[
17] J. R. Carey, S. K. Ma, T. D. Pfister, D. K. Garner, H. K.
Kim, J. A. Abramite, Z. Wang, Z. Guo, Y. Lu, J. Am.
Chem. Soc. 2004, 126, 10812–10813.
18] T. Ueno, T. Koshiyama, M. Ohashi, K. Kondo, M.
Kono, A. Suzuki, T. Yamane, Y. Watanabe, J. Am.
Chem. Soc. 2005, 127, 6556–6562.
Acknowledgements
This work was funded by the Swiss National Science Founda-
tion (Grants FN 200021–105192 and 200020–113348), the
Roche Foundation as well as the FP6 Marie Curie Reasearch
Training network (MRTN-CT-2003–505020) and the Canton
of Neuchâtel. We thank Belovo Egg Science and Technology
for a generous gift of egg white avidin and Umicore Precious
Metals Chemistry for a loan of rhodium. We thank C. R.
Cantor for the streptavidin gene, J.-M. Neuhaus, P. Schür-
mann (Univ. Neuchâtel) and P. Arosio (Univ. Milano) for
their help in setting up the streptavidin production, C. Malan,
J. Gradinaru and C. Schenk for synthetic help.
[
[
[
[
19] A. Mahammed, Z. Gross, J. Am. Chem. Soc. 2005, 127,
2
883–2887.
20] M. E. Wilson, G. M. Whitesides, J. Am. Chem. Soc.
978, 100, 306–307.
1
21] C.-C. Lin, C.-W. Lin, A. S. C. Chan, Tetrahedron Asym-
metry 1999, 10, 1887–1893.
22] J. Collot, J. Gradinaru, N. Humbert, M. Skander, A.
Zocchi, T. R. Ward, J. Am. Chem. Soc. 2003, 125,
9
030–9031.
[
23] M. Skander, N. Humbert, J. Collot, J. Gradinaru, G.
Klein, A. Loosli, J. Sauser, A. Zocchi, F. Gilardoni,
T. R. Ward, J. Am. Chem. Soc. 2004, 126, 14411–14418.
24] J. Collot, N. Humbert, M. Skander, G. Klein, T. R.
Ward, J. Organomet. Chem. 2004, 689, 4868–4871.
25] T. R. Ward, Chem. Eur. J. 2005, 11, 3798–3804.
26] M. Skander, C. Malan, A. Ivanova, T. R. Ward, Chem.
Commun. 2005, 4815–4817.
[
References
[
[
[
1] E. N. Jacobsen, A. Pfaltz, H. Yamamoto (Eds.), Com-
prehensive Asymmetric Catalysis, Vols. 1–3, Springer,
Berlin, 1999.
th
[27] H. Yamaguchi, T. Hirano, H. Kiminami, D. Taura, A.
[2] K. Faber, Biotransformations in Organic Chemistry, 5
Harada, Org. Biomol. Chem. 2006, 4, 3571–3573.
28] M. T. Reetz, J. J.-P. Peyeralans, A. Maichele, Y. Fu, M.
Maywald, Chem. Commun. 2006, 4318–4320.
29] C. Letondor, N. Humbert, T. R. Ward, Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 4683–4687.
edn., Springer, Berlin, 2004.
3] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Kess-
[
[
[
[
eler, R. Stürmer, T. Zelinski, Angew. Chem. Int. Ed.
2
004, 43, 788–824.
[
[
[
4] R. Berkessel, H. Grçger, Metal-free organic catalysts in
asymmetric synthesis, Wiley-VCH, Weinheim, 2005.
5] D. Qi, C.-M. Tann, D. Haring, M. D. Distefano, Chem.
Rev. 2001, 101, 3081–3111.
30] C. Letondor, A. Pordea, N. Humbert, A. Ivanova, S.
Mazurek, M. Novic, T. R. Ward, J. Am. Chem. Soc.
2
006, 128, 8320–8328.
31] G. Roelfes, B. L. Feringa, Angew. Chem. Int. Ed. 2005,
4, 3230–3232.
[
[
[
[
6] C. M. Thomas, T. R. Ward, Chem. Soc. Rev. 2005, 34,
4
3
37–346.
32] G. Roelfes, A. J. Boersma, B. L. Feringa, Chem.
Commun. 2006, 635–637.
33] M. T. Reetz, N. Jiao, Angew. Chem. Int. Ed. 2006, 45,
[7] Y. Lu, Angew. Chem. Int. Ed. 2006, 45, 5588–5601.
[8] R. Krꢂmer, Angew. Chem. Int. Ed. 2006, 45, 858–860.
[9] C. Letondor, T. R. Ward, ChemBioChem. 2006, 7,
2
416–2419.
1
845–1852.
34] G. Klein, N. Humbert, J. Gradinaru, A. Ivanova, F. Gi-
lardoni, U. E. Rusbandi, T. R. Ward, Angew. Chem. Int.
Ed. 2005, 44, 7764–7767.
[10] Y. Lu, Inorg. Chem. 2006, 45, 9930–9940.
[11] R. R. Davies, M. D. Distefano, J. Am. Chem. Soc. 1997,
1
19, 11643–11652.
[
[
[
[
12] T. Kokubo, T. Sugimoto, T. Uchida, S. Tanimoto, M.
Okano, J. Chem. Soc.Chem. Commun. 1983, 769–770.
13] K. Okrasa, R. J. Kazlauskas, Chem. Eur. J. 2006, 12,
[35] T. Satyanarayana, H. B. Kagan, Adv. Synth. Catal. 2005,
347, 737–748.
[36] C. Gennari, U. Piarulli, Chem. Rev. 2003, 103, 3071–
3100.
1
587–1596.
14] A. Fernandez-Gacio, A. Codina, J. Fastrez, O. Riant, P.
[37] D. Wahler, F. Badalassi, P. Crotti, J.-L. Reymond,
Soumillon, ChemBioChem. 2006, 7, 1013–1016.
15] F. van de Velde, L. Kçnemann, F. van Rantwijk, R. A.
Sheldon, Chem. Commun. 1998, 1891–1892.
Chem. Eur. J. 2002, 8, 3211–3228.
[38] C. R. Landis, J. Halpern, J. Am. Chem. Soc. 1987, 109,
1746–1754.
1930
asc.wiley-vch.de
ꢁ 2007 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1923 – 1930