Modulation of Nitrile Hydratase Regioselectivity towards Dinitriles by Tailoring the Substrate Binding Pocket Residues

Article abstract of DOI:10.1002/cctc.201701170

The regioselective hydration of dinitriles is one of the most attractive approaches to prepare ω-cyanocarboxamides or diamides and such regioselectivity is often beyond the capability of chemical catalysts. The use of nitrile hydratase to biotransform dinitriles selectively would be highly desirable. Molecular docking of two aliphatic dinitriles and two aromatic dinitriles into the active site of a nitrile hydratase (NHase) from Rhodococcus rhodochrous J1 allowed the identification of proximal NHase substrate binding pocket residues. Four residues (βLeu48, βPhe51, βTyr68, and βTrp72) were selected for single- and double-point mutations to modulate the NHase regioselectivity towards dinitriles. Several NHase mutants with an altered regioselectivity were obtained, and the best one was Y68T/W72Y. Docking experiments further indicated that the poor binding affinity of aliphatic and aromatic ω-cyanocarboxamides to the NHase variants resulted in distinct regioselectivity between wild-type and mutated NHases.

Full text of DOI:10.1002/cctc.201701170

Accepted Article
Title: Modulation of NHase Regioselectivity towards Dinitriles by
Tailoring the Substrate Binding Pocket Residues
Authors: Zhongyi Cheng, Wenjing Cui, Yuanyuan Xia, Lukasz
Peplowski, Michihiko Kobayashi, and Zhemin Zhou
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: ChemCatChem 10.1002/cctc.201701170
Link to VoR: http://dx.doi.org/10.1002/cctc.201701170
A Journal of
www.chemcatchem.org

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Modulation of Nitrile Hydratase Regioselectivity towards
Dinitriles by Tailoring the Substrate Binding Pocket
Residues
Zhongyi Cheng,[a] Wenjing Cui,[a] Yuanyuan Xia,[a] Lukasz Peplowski,[b] Michihiko Kobayashi,*[c] Zhemin
Zhou*[a]
Abstract: The regioselective hydration of dinitriles is one of the
most attractive approaches to prepare ω-cyanocarboxamides or
diamides, and such regioselectivity is often beyond the capability of
chemical catalysts. Using nitrile hydratase to selectively biotransform
dinitriles would be highly desirable. Molecular docking of two
aliphatic dinitriles and two aromatic dinitriles into the active site of a
nitrile hydratase (NHase) from Rhodococcus rhodochrous J1
allowed the identification of proximal NHase substrate binding
pocket residues. Four residues (βLeu48, βPhe51, βTyr68 and
βTrp72) were selected for single and double point mutations to
modulate the NHase regioselectivity towards dinitriles. Several
NHase mutants with altered regioselectivity were obtained and the
best one was Y68T/W72Y. Docking experiments further indicated
that the poor binding affinity of aliphatic and aromatic ω-
cyanocarboxamides to the NHase variants resulted in the distinct
regioselectivity between wild-type and mutated NHases.
Scheme 1. Regioselective hydration of dinitriles to their corresponding
amides by NHase.
Introduction
nitrile biocatalysts, few such regioselective bacterial genera
containing NHase which result in ω-cyanocarboxamide
compounds have been described. Using Acinetobacter strain
(FERM P-10432), some aliphatic dinitriles (<C8) can be
Nitrile hydratase (NHase, EC 4.2.1.84) is
a
class of
metalloenzymes which catalyzes the addition of one molecule of
water to the nitrile, resulting in the formation of the
corresponding amide.[1] In the past few decades, NHase has
taken place of traditional synthetic catalysts and become
dominant in acrylamide and nicotinamide synthesis, indicating
itself as a promising biocatalyst for the production of high-value
amides with the advantages of environmental friendliness, mild
reaction conditions and high catalytic efficiency.[2]
Heterobifunctional compounds can be used as polymer
intermediates as well as intermediates in producing high value
agricultural or pharmaceutical products. From aliphatic or
aromatic α,ω-dinitriles, it is sometimes desirable to produce such
heterobifunctional compounds by hydrolyzing only one single
functional cyano group without further modification, and such
regioselectivity is often beyond the capability of chemical
catalysts.[3] However, NHase is able to recognize the subtle
structural differences and may regioselectively catalyze dinitriles
into either ω-cyanocarboxamides or α,ω-diamides. In the field of
converted
to
the
corresponding
ω-cyanocarboxamide
derivatives.[3] A newly isolated NHase-producing strain R.
aetherivorans ZJB1208 was used for regioselective hydration of
1-cyanocyclohexaneacetonitrile
to
1-
cyanocyclohexaneacetamide with a high product concentration
up to 966.7 g•L-1.[4] Furthermore, the enzymatic production of 5-
cyanovaleramide from adiponitrile has been applied in industry
with higher conversion, higher selectivity and less by-products.[5]
Although several bacterial strains have been selected and used
in industrially interesting ω-cyanocarboxamides or α,ω-diamides
production, tailoring existing NHase that can regioselectively
catalyze a wide range of α,ω-dinitriles (both aliphatic and
aromatic dinitriles) into corresponding ω-cyanocarboxamides still
remains unexplored.
Tuning a known enzyme to increase its selectivity through
protein engineering has advanced tremendously through past
decades as the general use of biocatalysis in industry has
suffered from insufficient or wrong regioselectivity.[6] Generally,
there are three approaches to protein engineering: directed
evolution, rational design and semi-rational design. The semi-
rational approach combines the advantages of directed evolution
and rational design, where only a few hotspot residues are
selected for the following smart library construction based on the
prior structural analysis. Those well-modulated enzymes
become superior to their chemical counterpart and attractive for
industrial applications. Previously, we have identified a tunnel-
forming residue that direct NHase regioselectivity towards
aliphatic dinitriles using semi-rational design that based on
amino acid sequence alignment and tunnel calculation.[7] The
[a]
Zhongyi Cheng, Prof. Wenjing Cui, Yuanyuan Xia, Prof. Zhemin
Zhou
Key laboratory of Industrial Biotechnology, Ministry of Education,
School of Biotechnology, Jiangnan University, Wuxi 214122,
Peoples Republic of China. E-mail: zhmzhou@jiangnan.edu.cn
Dr. Lukasz Peplowski
Institute of Physics, Faculty of Physics, Astronomy and Informatics,
Nicolaus Copernicus University, Grudziadzka 5, 87-100 Torun,
Poland
[b]
[c]
Prof. Michihiko Kobayashi
Institute of Applied Biochemistry and the Graduate School of Life
and Environment Sciences, the University of Tsukuba, Ibaraki, 305-
8572, Japan. E-mail: kobayashi.m.fe@u.tsukuba.ac.jp
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
β37 residue of NHase contributes to the difference in the shape
and size of the substrate access tunnel, and thus directs the
diversity of reaction regioselectivity.
In this study, we succeeded in engineering the substrate binding
pocket residues to modulate the regioselectivity of NHase from
Rhodococcus rhodochrous J1 (L-NHase) towards both aliphatic
and aromatic dinitriles (Scheme 1). In order to obtain valuable
ω-cyanocarboxamide intermediate,
a
semi-rational design
approach based on docking experiment was performed. Two
aliphatic α,ω-dinitriles, malononitrile (MAN) and adiponitrile
(ADN), as well as two aromatic dinitriles, terephthalonitrile
(TPTN) and phthalodinitrile (PTN) were used as substrates and
docked into the active site of L-NHase. The substrate binding
pocket residues were selected for single and double point
mutations to modulate the NHase regioselectivity. Several
NHase mutants with altered regioselectivity were obtained and
the best one was Y68T/W72Y. Such variant was able to catalyze
only one cyano group into corresponding amide group without
further modification whereas the wild-type L-NHase tended to
hydrolyze both functional cyano groups of dinitriles. These
findings clarified the role of substrate binding pocket residues on
NHase regioselectivity towards α,ω-dinitriles and may expand
the knowledge on the architecture of the NHase active site and
its relation to the regiospecific hydration of nitrile substrates.
Figure 1. Close-up view of the selected ligand surrounding residues
based on docking experiment. A) A result of docking ADN into wild-type
L-NHase. B) A result of docking MAN into wild-type L-NHase. C) A result of
docking TPTN into wild-type L-NHase. D) A result of docking PTN into wild-
type L-NHase. All the ligands are shown in ball and stick. The selected
surrounding residues are represented as yellow sticks. The active site are
shown in green sticks and the cobalt ion was shown as limon sphere.
site of L-NHase. The binding mode of the aliphatic and aromatic
substrates, catalytic residues and binding pocket residues are
shown in Figure 1. It is noteworthy that the regioselectivity of
NHase should have a close relationship with the spatial
conformation of catalytic pocket.[13] Except the catalytic residues
(Cys-Thr-Leu-Cys-Ser-Cys) of NHase, several residues located
at the α and β subunits of L-NHase (αGln90, βLeu48, βPhe51,
βArg52, βTyr68, βTrp72) are in close proximity to the docked
substrates (less than 3.5 Å in distance) and the orientations of
these amino acids match the rule that the side chains of the
amino acids for substrate binding pocket engineering should
point to the binding site of wild-type enzyme.[14] However,
αGln90 is conserved in all known NHases and involved in the
hydrogen-bond network around the catalytic center. Mutating
this residue resulted in the decrease of enzymatic activity.[15]
Furthermore, βArg52 plays an important role in the α-subunit
exchange and may have positive effect on catalytic efficiency.[16]
Therefore, αGln90 and βArg52 was excluded from the candidate
residues for subsequent mutagenesis. The rest four residues
were selected as „hot-spot‟ residues for their potential influence
on the regioselectivity of NHase. Although residues βTyr68 and
βTrp72 are conserved in all Co-type NHase, they might be
important for substrate selection and binding.[17] βLeu48 and
βPhe51 are located at the substrate access tunnel of L-NHase
and it is reported that substrate access tunnel maybe
responsible for the fine-tuning of the substrate selectivity of
NHase.[18]
Results and Discussion
Selection of Hotspot Residues for Site-directed Saturation
Mutagenesis
With the help of rapidly developed bioinformatics tools, as well
as the huge number of available crystal structure of enzymes in
the Protein Data Bank, the probability of hitting the potential „hot
spots‟ that affect enzymatic activity or selectivity has increased
in a large extent. For example, tunnel calculation together with
molecular dynamics simulation could help with the identification
of key residues directing substrate migration and enzymatic
selectivity.[8] Molecular docking also plays an important role in
identifying the residues which might influence ligand
orientation.[9] These strategies could contribute a lot to protein
engineering. A semi-rational design was chosen with the goal of
generating improvements in NHase regioselectivity for hydration
of α, ω-dinitriles. To gain insights into the binding mode of both
aliphatic and aromatic dinitries at the active site of NHase,
homology modeling and molecular docking was firstly performed.
After three iterations of PSI-BLAST search against the PDB
database,[10] the NHase from Bacillus smithii SC-J05-1 (PDB
code: 1V29) showed the highest sequence similarity to L-NHase
with 61% and 38% identities respectively for α subunit and β
subunit, which indicates these proteins are homologs.[11] The 3D
structure of L-NHase was then constructed and aligned to 1V29
crystal structure using Swiss-Model homology modeling (Figure
S1, S2).[12] RMSD calculated by Pymol between our model and
1V29 structure was 0.093, indicating the tertiary structures of
these two proteins are conversed. Subsequently, two aliphatic
dinitriles, MAN and ADN, along with two aromatic dinitriles,
TPTN and PTN were docked into the position above the active
Library construction and NHase Variants Screening
In order to examine the effect of different amino acid side chains
at the selected positions (βLeu48, βPhe51, βTyr68, βTrp72) on
NHase regioselectivity for aliphatic and aromatic dinitriles,
saturation mutagenesis was performed individually at these
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Table 1. Regioselectivity for NHases and the variants towards the aliphatic α, ω-dinitriles[a]
.
ADN hydration
MAN hydration
Regioselectivity (%)
Regioselectivity (%)
Enzyme
5-CVAM
0
ADAM
100
Conv[b]
CAAM
22.7
MAAM
77.3
Conv[b]
Wild-type
98.6
97.3
Y68T
97.6
96.8
100
2.4
3.2
0
93.1
90.4
70.5
79.5
76.2
97.1
21.5
23.8
2.9
90.6
89.3
79.5
W72Y
Y68T/W72Y
[a] Determination of regioselectivity by HPLC analysis. [b] The conversion of the substrate was measured after a 4 h reaction. Reactions were performed in 10
mM potassium phosphate buffer (KPB) at pH 7.4. The concentration of aliphatic dinitriles was 20 mM. The reaction temperature was 35 °C . The volume of the
reaction mixture was 500 μL.
amino acid positions and four single-site libraries were
generated. We then carried out the catalytic reaction for 10 min,
30 min, 1 h, 2h and 4 h using those constructed mutants. The
products were analyzed by HPLC. As mutability landscapes
have become a powerful tool to identify functional hotspots at
any place in the amino acid sequence of an enzyme, we applied
such method to the four selected residues (Figure S3-S8).[19] It is
obvious that mutations at residue βTyr68 and βTrp72 lead to the
alteration of enzymatic regioselectivity. Overall, two out of the
total seventy-six variants (nineteen variants for each library)
from the four libraries were selected because of their obviously
altered regioselectivity towards both aliphatic and aromatic
dinitriles (Table 1, 2).
For ADN hydration, the L-NHase Y68T and W72Y variants
caused a large shift in product formation and formed primarily 5-
cyanovaleramide (5-CVAM) with the minimum loss of activity,
whereas wild-type L-NHase formed only adipamide (ADAM)
after a 4 h reaction (Table 1, Figure S9). For the second round
of library construction, mutant Y68T was used as template to
build a (non-combinatorial) single-site saturation mutagenesis
library at site 72 through Iterative Saturation Mutagenesis
(ISM)[20] because of its excellent selectivity towards ω-
cyanocarboxamides, as well as the hint we obtained from
mutability landscapes. It is worth to note that one variant from
the library constructed at the second round of saturation
mutagenesis, Y68T/W72Y, inverted the regioselectivity of L-
NHase towards ADN; however, in terms of activity, the double-
point mutants diminished in a large extent. For the hydration of
the other aliphatic dinitrile MAN, the wild-type L-NHase preferred
the malonamide (MAAM) formation with 97.3% MAN conversion.
Variants Y68T and W72Y showed drastic change in
regiospecificity by producing mainly the ω-cyanocarboxamide,
cyanoacetamide (CAAM), at a relative low MAN conversion.
There was a 3.5- and 3.2-fold change in regiospecificity for
CAAM formation with variant Y68T and W72Y, respectively.
Notably, the double mutant Y68T/W72Y exhibited reversed
regioselectivity compared with that of wild-type NHase. It
produced mainly CAAM with regioselectivity up to 97.1%,
whereas the wild-type NHase only formed 22.7% CAAM;
however, diminished activity was also observed for variant
Y68T/W72Y with only 79.5% MAN conversion.
For the hydration of the two aromatic dinitriles, either Y68T or
W72Y showed significant preference for ω-cyanocarboxamide
formation (Table 2, Figure S10). Variants Y68T and W72Y
caused
a selectivity of 62.3% and 75.2% towards 4-
cyanobenzamide (4-CBAM) at 94.8% and 92.4% TPTN
conversion, respectively, whereas the wild-type L-NHase
preferred catalyzing TPTN mainly into terephthalamide with a
conversion up to 99.2%. For PTN hydration, the Y68T and
W72Y variants resulted in elevated 2-cyanobenzamide (2-
CBAM) formation as well compared with their parent enzyme.
The double mutant Y68T/W72Y caused the largest shift in ω-
cyanocarboxamide formation although the enzymatic activity
decreased in some extent. The regioselectivity of such double-
Table 2. Regioselectivity for NHases and the variants towards the aromatic α, ω-dinitriles[a]
.
TPTN hydration
PTN hydration
Regioselectivity (%)
Regioselectivity (%)
Enzyme
4-CBAM
15.7
Terephthalamide
84.3
Conv[b]
2-CBAM
0
Phthalamide
100
Conv[b]
Wild-type
99.2
96.1
Y68T
62.3
37.7
94.8
56.4
43.6
91.4
W72Y
75.2
98.2
24.8
1.8
92.4
71.1
70.7
100
29.3
0
90.2
72.5
Y68T/W72Y
[a] Determination of regioselectivity by HPLC analysis. [b] The conversion of the substrate was measured after a 4 h reaction. Reactions were performed in 10
mM potassium phosphate buffer (KPB) at pH 7.4, TPTN and PTN concentrations were 2 mM and 5 mM, respectively. The reaction temperature was 35 °C . The
volume of the reaction mixture was 500 μL.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Table 3. Kinetic parameters for NHase and the variants towards 5-cyanovaleramide and malonamide.
5-CVAM
KM [mM]
13.21.1
CAAM
Enzyme
kcat [s-1]
kcat/KM [s-1•mM-1]
KM [mM]
20.71.6
kcat [s-1]
kcat/KM [s-1•mM-1]
Wild-type
Y68T
435.77.7
33.0
0.2
417.28.6
20.2
0.4
127.63.1
113.22.2
N.A.
19.60.8
21.30.9
N.A.
88.22.5
85.33.8
N.A.
39.12.2
42.82.7
N.A.
W72Y
0.2
0.5
Y68T/W72Y
N.A.
N.A.
The 5-CVAM and CAAM concentrations were varied in the ranges of 5-600 mM and 5-200 mM respectively. N.A., not available. The reactions were performed
in 10 mM potassium phosphate buffer (KPB, pH 7.4) at 35 °C for 3 min. The volume of the reaction mixture was 500 μL.
point mutant towards TPTN and PTN was completely altered
with only 1.8% 4-CBAM and 0% 2-CBAM detected respectively.
of the aliphatic ω-cyanocarboxamide products. The catalytic
performance of NHase towards aromatic 2-CBAM and 4-CBAM
was also investigated (Table 4). The wild-type L-NHase
catalyzed 2-CBAM into phthalamide with a KM equaled to 2.5
mM and kcat equaled to 145.6 s-1. As for the hydration of 4-
CBAM, the wild-type L-NHase could biotransform 4-CBAM into
terephthalamide in a kcat/KM value equaled to 46.8 s-1•mM-1.
However, once the Y68T or W72Y mutant was introduced, it
became rather difficult to determine the exact kinetic parameters
for the 2-CBAM and 4-CBAM conversion due to the limited
solubility of the substrates. The KM values for the hydration of 2-
CBAM using Y68T and W72Y single-point mutant were more
than 10 mM while the KM values for 4-CBAM hydration using the
two variants were more than 5 mM. The kcat value for the
hydration of both 2-CBAM and 4-CBAM using Y68T and W72Y
single-point mutant also decrease in a large extent, indicating a
great loss of catalytic efficiency of these two variants towards
aromatic ω-cyanocarboxamide. Furthermore, it was clearly
Kinetic Analysis of NHase Mutants
To elucidate the reason for the difference in the regioselectivity
between the wild-type L-NHase and its variants from
a
biochemical perspective, the steady-state kinetic parameters
were measured. Aliphatic and aromatic ω-cyanocarboxamides
were used as substrates as the hydration step of ω-
cyanocarboxamides contributed mostly to the alteration of
NHase regioselectivity. For the hydration of aliphatic ω-
cyanocarboxamides 5-CVAM and CAAM, all three mutants
(Y68T, W72Y and Y68T/W72Y) exhibited poor catalytic
efficiency compared with their parent enzyme (Table 3). For the
Y68T variant, a 9.6-fold increase in KM combined with a 22.2-fold
decrease in kcat was observed while catalyzing 5-CVAM. A 50-
fold decrease in kcat/KM was also determined during the CAAM
hydration. The W72Y mutant shared the same catalytic pattern
as Y68T variant with a 165-fold decrease in kcat/KM when using
5-CVAM as the substrate. The KM for CAAM hydration equaled
to 85.3 mM, approximately a 4.1-fold increase compared with
that of wild-type L-NHase. The kcat value for the W72Y mutant
suffered an 11-fold decrease and contributed to a total 40.4-fold
decrease in catalytic efficiency. It is intriguing that introducing
W72Y to the Y68T variant makes the L-NHase almost unable to
catalyze both 5-CVAM and CAAM, resulting in the accumulation
shown in Table
4 that the double mutant Y68T/W72Y
constructed through ISM had a strong effect on preventing
aromatic ω-cyanocarboxamide from being hydrolyzed with no
activity towards 2-CBAM and 4-CBAM detected.
Characterization of Wild-Type NHase and its beneficial
mutations
Table 4. Kinetic parameters for NHase and the variants towards 2-cyanobenzamide and 4-cyanobenzamide.
2-CBAM
KM [mM]
2.50.1
>10[a]
4-CBAM
KM [mM]
2.30.1
>5[a]
Enzyme
Wild-type
Y68T
kcat [s-1]
kcat/KM [s-1•mM-1]
kcat [s-1]
kcat/KM [s-1•mM-1]
145.64.3
>66.12.3[b]
58.2
N.A.
107.78.6
>32.11.2[b]
46.8
N.A.
W72Y
>10[a]
N.A.
N.A.
N.A.
>5[a]
N.A.
N.A.
N.A.
>42.53.0[b]
>19.20.9[b]
Y68T/W72Y
N.A.
N.A.
The 2-CBAM and 4-CBAM concentrations were varied in the ranges of 0.5-10 mM and 0.2-5 mM respectively. N.A., not available. [a] No saturation was
detected. The KM value was beyond the concentration range of the substrates due to the limited solubility of 2-CBAM and 4-CBAM. [b] Calculated based on the
highest rate detected. The reactions were performed in 10 mM potassium phosphate buffer (KPB, pH 7.4) at 35 °C for 3 min. The volume of the reaction
mixture was 500 μL.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Figure 2. The dispersion of Co-Ncyano distance and dinitrile binding energy. A) The closest distance between cobalt ion and the nitrogen atom of the
cyano group in ADN and the ADN binding energy into wild-type NHase and its variants. B) The closest distance between cobalt ion and the nitrogen atom of the
cyano group in MAN and the MAN binding energy into wild-type NHase and its variants. C) The closest distance between cobalt ion and the nitrogen atom of the
cyano group in TPTN and the TPTN binding energy into wild-type NHase and its variants. D) The closest distance between cobalt ion and the nitrogen atom of
the cyano group in PTN and the PTN binding energy into wild-type NHase and its variants. The Co-Ncyano distance longer than 10 Å represents the substrate
access tunnel.
As preferential selectivity for α, ω-dinitrile or ω-
cyanocarboxamide may be related to the substrate orientations
in NHase interior, knowledge of the binding modes of these
ligands inside the L-NHase is highly desirable. To better
illustrate the differences in regioselectivity between the wild-type
L-NHase and its variants, we firstly optimized the protein
structure of these enzymes. A 5 ns classic molecular dynamics
simulation was applied to the wild-type L-NHase as well as the
three mutants (Figure S11). Two aliphatic dinitriles and two
aromatic dinitriles (ADN, MAN, TPTN and PTN) together with
their corresponding ω-cyanocarboxamides (5-CVAM, CAAM, 4-
CBAM and 2-CBAM) were then docked above the active site
motif CYS-X-Y-CSD (cysteine-sulfinic acid) -SER-CEA (cysteine
sulfenic acid) and in total 6400 binding places were obtained
(200 docking positions for each ligand-NHase complex).
According to the proposed reaction mechanism of NHase, the
distance between the cobalt ion at the active site and the cyano
nitrogen atom in the substrate plays a crucial role in determining
the NHase catalytic activity towards the substrate.[21] In addition,
reconstructing the substrate binding pocket may cause different
ligand orientation and the binding energy of the substrate could
shift in large extent.[22] As a result, the dispersion of Co-Ncyano
distance (distance between the cobalt ion at the active site and
the cyano nitrogen in the substrate) and ligand binding energy
was analyzed.
For the docking of aliphatic dinitrile ADN, the docked substrate
represented better binding energetic scores and Co-Ncyano
distance dispersion in wild-type NHase interior than inside other
NHase variants (Figure 2A). The best docked ADN that
exhibited relative low binding energy (-4.73 kcal•mol-1) and close
Co-Ncyano distance (5.04 Å) was stabilized by hydrogen bonds
formed between the ligand and the residues αTyr115 and
αTrp117 (Figure S12A). As the residues αTyr115 and αTrp117
are in close vicinity to the catalytic motif, they might influence the
substrate recognition or orientation. Thus, two mutants, αY115A
and αW117A were introduced. These two variants exhibited
extremely low activity towards ADN, indicating residues αTyr115
and αTrp117 play important roles in directing the enzymatic
activity of NHase (Figure S13). Once the ADN was docked
inside the Y68T/W72Y variant, the binding energy increased in
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Figure 3. The dispersion of Co-Ncyano distance and ω-cyanocarboxamide binding energy. A) The distance between cobalt ion and the nitrogen atom of
the cyano group in 5-CVAM and the 5-CVAM binding energy into wild-type NHase and its variants. B) The distance between cobalt ion and the nitrogen atom of
the cyano group in CAAM and the CAAM binding energy into wild-type NHase and its variants. C) The distance between cobalt ion and the nitrogen atom of the
cyano group in 4-CBAM and the 4-CBAM binding energy into wild-type NHase and its variants. D) The distance between cobalt ion and the nitrogen atom of the
cyano group in 2-CBAM and the 2-CBAM binding energy into wild-type NHase and its variants. The Co-Ncyano distance longer than 10 Å represents the
substrate access tunnel.
some extent. The binding energy of the close lying docked ADN
was -3.55 kcal•mol-1 and the Co-Ncyano distance was 4.09 Å
whereas the ADN docked in the substrate tunnel represented
lower binding energy that equaled to -3.73 kcal•mol-1 (Figure
S12B). Although the Co-Ncyano distance was closer inside the
Y68T/W72Y variant than that inside the wild-type NHase, the
docking of ADN was less stable. Besides, it is noteworthy that
the ADN had the potential to be docked into the substrate
access tunnel with relative low binding energy (Figure S14). The
substrate tunnel was calculated based on previous studies and
was located at the interface of α and β subunits (Figure S15).[7,
23] The substrate could bind to the region above the bottleneck of
the substrate tunnel. As the bottleneck residue at βsite37 might
was quite similar to that of ADN docking (Figure 2B). All the
docked MAN inside the wild-type NHase were close to the cobalt
ion with low binding energy whereas the docked substrate inside
the Y68T/W72Y mutant was less stable above the active site
and tended to interact with the substrate access tunnel. Two
aromatic dinitriles, TPTN and PTN, were also docked into the
NHase and its mutants (Figure S17). Similarly, the aromatic
ligands represented stable binding mode inside the wild-type
NHase with low binding energy and modest Co-Ncyano distance,
whereas the docking of the aromatic dinitriles inside the three
mutants exhibited high energetic scores and most of the
substrate docking positions were remote to the active-site center
(Figure 2C, 2D). Such binding energetic scores and Co-Ncyano
play an important role in directing the regioselectivity of NHase,[7, distance dispersion of both aliphatic and aromatic dinitriles
18]
we performed saturation mutagenesis at this site. However,
only one mutant, F37P, changed the selectivity for ADN slightly
(Figure S16). For the docking of the other aliphatic dinitrile MAN,
the binding energetic scores and Co-Ncyano distance dispersion
inside the NHase variants, especially inside the Y68T/W72Y
mutant, indicated that it may be difficult for them to enter the
enzyme active site and thus could not be easily catalyzed. This
hypothesis coincides well with the relative low conversion rate of
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Figure 4. Molecular docking of aliphatic ω-cyanocarboxamides into wild-type NHase and the Y68T/W72Y variant. A) The best docked 5-CVAM into the
active site of wild-type NHase. B) The best docked 5-CVAM into the active site of the Y68T/W72Y variant. C) The best docked CAAM into the active site of wild-
type NHase. D) The best docked CAAM into the active site of the Y68T/W72Y variant. The βTyr68 and βTrp72 residues in wild-type NHase are shown in yellow
sticks. The βThr68 and βTyr72 residues in the Y68T/W72Y variant are shown as purpleblue sticks. The active site and the residues involving in hydrogen bond
formation are shown in green sticks. The cobalt ion is represented as limon sphere. The orange dashed line stands for the Co-Ncyano distance. The yellow
dashed line indicates the hydrogen bonds.
the Y68T/W72Y variant towards the aliphatic and aromatic
dinitriles.
7.93 Å from the cobalt ion and the binding energy was -3.74 kcal
•mol-1. The amide group could only create two hydrogen bonds
with residue αCys112 and αSer114, resulting in a less stable
binding mode compared with the docking of 5-CVAM inside the
wild-type NHase (Figure 4B). It is intriguing that besides the
substrate binding pocket, 5-CVAM could interact with two other
areas that are adjacent to the active site: one is the substrate
access tunnel and the other is a potential tunnel that may
connect the solvent with the NHase active site. The docked 5-
CVAM inside the potential tunnel of the Y68T/W72Y variant was
located at the horizontal plane of the active site (Figure S14D),
which could not match the proposed mechanism that nitriles
should coordinate with the cobalt ion above the active site plane
to initiate hydration process.[13] The Co-Ncyano distance of such
docked 5-CVAM was 6.29 Å and the binding energy was -3.95
kcal•mol-1. Those docked 5-CVAMs inside the substrate access
tunnel were even more remote from the active site (> 10 Å) and
therefore could not be catalyzed into corresponding diamide.
The weak dispersion of the binding energy and Co-Ncyano
distance of 5-CVAM inside the Y68T/W72Y variant
As the two cyano groups of the dinitriles should be hydrolyzed
one after the other during catalytic process,[24] the ability of
catalyzing the corresponding ω-cyanocarboxamides would make
key contributions to the regioselectivity of NHase. Consequently,
the four ω-cyanocarboxamides (5-CVAM, CAAM, 4-CBAM and
2-CBAM) were docked inside the wild-type NHase and its
mutants. With respect to the docking of 5-CVAM, the ligand
showed the lowest energetic scores inside the active site of the
wild-type NHase (Figure 3A). The best docked 5-CVAM was
4.94 Å from the cobalt ion with the binding energy equaled to -
5.03 kcal•mol-1 (Figure 4A). The amide group of 5-CVAM was
stabilized by the formation of hydrogen bonds with αSer114,
αTyr115 and αGln90. In contrast, the dockings of 5-CVAM inside
the three NHase variants represented higher binding energy and
the Co-Ncyano distance was longer than that of wild-type NHase.
Most of the docked 5-CVAMs inside the Y68T/W72Y variant had
long Co-Ncyano distance that was too remote to initiate the
hydration of the cyano group. The best docked 5-CVAM was
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
demonstrated that the double-point mutant could hardly catalyze
5-CVAM and thus the intermediate product was accumulated,
which was in good agreement with the measured KM values.
As for the docking of the rest three ω-cyanocarboxamides
(Figure 4C, 4D, S18), it is obvious that when the ligands were
docked inside the wild-type NHase, they exhibited the most
stable binding modes above the active site with appropriate Co-
Ncyano distance. Once the ligands were docked into the three
NHase mutants, particularly the binding mode inside the
Y68T/W72Y variant, the binding energy increased and in some
cases the ligands could interact with the substrate access tunnel.
The high energetic scores and the remote distance between the
docked ligands and the cobalt ion implied that the binding affinity
of the NHase variants to ω-cyanocarboxamides has been
significantly weakened, which fitted well with the analysis of
kinetics parameters and therefore the regioselectivity of NHase
towards α,ω-dinitriles was modulated.
compared with its parent enzyme, but improving the catalytic
efficiency remains to be solved. This study adds to the list of
research in which substituting a single amino acid for another
one can alter regiospecificity of certain enzymes and it would be
interesting to examine the effect of combining different beneficial
variants and extend the work to other substrates to improve the
versatility of NHase.
Experimental Section
Strains, Plasmids and Chemical Reagents
Plasmid pET-24a, which carries the NHase genes from Rhodococcus
rhodochrous J1, was synthesized by the Shanghai Gene Core
Biotechnologies Co., Ltd. The 99% pure 5-cyanovaleramide was bought
from Toronto Research Chemicals Inc., Canada. Adiponitrile, adipoamide,
malononitrile and malonamide were purchased from Sigma-Aldrich Co.,
USA. Terephthalonitrile, phthalodinitrile, 4-cyanobenzamide, 2-
cyanobenzamide, phthalamide and terephthaldiamide were obtained
from Tokyo Chemical Industry., Japan.
Although our newly engineered variants showed distinct
regioselectivity towards dinitriles compared with their parent
enzyme, their catalytic efficiency towards the substrates was
compromised, especially for the double mutant Y68T/W72Y.
Such phenomenon is often observed in protein engineering or
directed evolution campaigns.[25] As discussed in previous
review, accuracy (selectivity) of an enzyme can be tailored with
Library Construction
a
significant cost, a
cost in catalytic rate.[26] Enzymatic
The residues βLeu48, βPhe51, βTyr68 and βTrp72 were chosen to
create saturation mutagenesis libraries. Site-directed mutagenesis of
amino acid residue substitution in these positions was performed by
means of a whole plasmid PCR protocol.[29] PrimeSTAR HS DNA
Polymerase from TAKARA BIO INC was used. The mixture (50 μL)
contains: pET24a-L-NHase 1 μL (20-50 ng/μL), forward primer 1 μL,
reverse primer 1 μL, 5×PCR buffer 10 μL, dNTP Mixture 4 μL (2.5 mM),
Primer STAR Enzyme 0.5 μL (2.5 U/μL), ddH2O 32.5 μL. The PCR cycle
condition was listed as follows: 18 cycles; Pre-denaturation at 95 °C for 5
min, denaturation at 94 °C for 1 min, annealing at 56 °C for 40 sec,
extension at 68 °C for 7 min. The PCR products were digested for 5
hours at 37 °C using 10 units of QuickCutTM Dpn I from TAKARA BIO INC.
to remove the parent plasmid. The reaction mixtures were transformed
into E. coli strain BL21 (DE3). We created each of the 19 single-point
variants at each of the 4 sites (76 variants) as well as the 19 double
mutants at site 72 by using Y68T as temple. All the corresponding 190
forward primers and reverse primers are shown in Table S1-S5. We
picked three clones for each single and double mutant and totally 285
transformants were picked. All the picked clones were verified one by
one via DNA sequencing. Each variant was successfully created. All
clones were grown in 2YT medium containing CoCl2∙6H2O (0.05 g/L) and
kanamycin (50 mg/mL) at 37 °C . Isopropyl β-D-thiogalactopyranside
(IPTG) was used as the inducer and added to a final concentration of 0.4
mM after the A600 reached 0.8. The cells were then cultivated at 24 °C
for 16 h. Cells were harvested by centrifugation (4 ℃, 10000 × rpm, 5
min).
selectivity-activity tradeoffs needs deeper understanding and in
our case, further efforts should be made to improve the
enzymatic activity with enzymatic selectivity remains excellent.
On one hand, tunnel residues that are remote to the active site
might have effects on enzymatic activity and selectivity. For
example, a residue D285, located at the tunnel entrance of
toluene 4-monooxygenase, distant from the active site was
randomized and the rate of catalysis was enhanced.[27] In our
study, the substrates could bind to the tunnel regions and such
circumstance might influence the enzymatic activity or selectivity.
Although we performed saturation mutagenesis on three tunnel
forming residues (βPhe37, βLeu48 and βPhe51), no excellent
results were obtained. Further site-directed mutagenesis could
be performed on other tunnel residues such as residues located
at the entrance of substrate access tunnel. On the other hand,
we only focus on the residues with close distance to the active
site (less than 3.5 Å). First and second sphere residues,
however, should also be noticed for further protein engineering.
Tailoring these residues might result in the fine tuning of the
specificity or activity of a certain enzyme.[28] It might be a good
choice to engineer the tunnel residues located within the
first/second sphere of L-NHase.
Conclusions
In summary, the regioselectivity of NHase towards both aliphatic
and aromatic α,ω-dinitriles could be improved by a structure-
based semi-rational design approach. Several NHase mutants
with altered or even inversed regioselectivity were identified
based on iterative saturation mutagenesis in conjunction with
mutability landscapes and their effects were rationalized using
MDs and in silico docking. The key substrate binding pocket
residues (βTyr68 and βTrp72) which determine regioselectivity
towards α,ω-dinitriles were found. The best mutant Y68T/W72Y
displays totally different regioselectivity towards dinitriles
Expression and Purification of Enzymes
Recombinant E. coli carrying the plasmid pET-24a that contains the wild-
type NHase or a mutant NHase gene with a Strep tag at the N terminal of
β subunit, was grown in 2YT medium with CoCl2∙6H2O (0.05 g/L) and
kanamycin (50 mg/mL) inside at 200 rpm and 37 °C . The expression of
NHase was induced by adding IPTG to a final concentration of 0.4 mM
when the optical density reached 0.8 and was continuously incubated at
24 °C for 16 h. Cells used for activity and enantioselectivity assessments
were harvested by centrifugation at 6,500
×
g for 10 min and
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
resuspended in 10 mM potassium phosphate buffer (KPB) (containing
0.5 mM dithiothreitol, pH 7.4). For the purification of positive mutants,
cells mentioned previously were distrupted at 4 °C with ultrasonic cell
disruptor (Scientz Biotechnology, JY92-II, Ningbo, China), which was
operated at 200 W for 200 cycles (3 s on 7 s off). The cell debris was
removed via centrifugation at 12000 × g and 4 °C for 10 min. All
purification steps were carried out at 4 °C using the AKTA purifier (GE
Healthcare UK Ltd). A StrepTrap HP column (GE Healthcare UK Ltd.)
equilibrated with binding buffer (20 mM Na2HPO4•12H2O, 280 mM NaCl,
6 mM KCl, pH 7.4) was used. The active fractions were collected,
concentrated to 1 ml by ultrafiltration and applied to a SuperdexTM 200
10/300 GL column (GE Healthcare UK Ltd.) (Figure S19).
was constructed using “mutate” option in psfgen tool (part of NAMD
package). RMSD calculated by Pymol between the wild-type L-NHase
and Y68T/W72Y mutant was 1.945. The substrate access tunnels of
NHases were calculated and presented by CAVER analyst 1.0.8.[32] The
Min. probe radius was 0.9 Å and the shell depth was 4 Å. The tunnel
starting point was the geometric center of the NHase active site (C-T-L-
C-S-C). We aim to identify the exact tunnel for ligand migration.
Experimental X-ray data indicated that the migration of substrates or
products occurs through the tunnel located at the interface of α and β
23]
subunits of NHase.[21a,
The potential tunnels are calculated and it
demonstrated that there are several tunnels inside L-NHase (Figure 15A)
and we have identified one tunnel that fits well with previous X-ray data
and SMD simulation from all those calculated tunnels (Figure 15B).[18]
While the other possible ligand/product transfer routes cannot definitively
be ruled out, both X-ray analysis[33] and previous Locally Enhanced
Sampling MD simulations of the related Fe–NHase[34] indicate that this
tunnel plays the major role in the NHase functioning. Phe37 and Leu48 at
the β subunit of NHase formed the bottleneck of the substrate access
tunnel. The MD simulations were performed using NAMD 2.11 with the
Charmm27 force field.[35] Each protein model was put into a cubic water
box 70 × 70 × 70 Å3 with a layer of water of at least 7 Å in each
dimension. A cutoff of 12 Å for non-bonded interactions was applied. We
performed 1 ns of water equilibration, 1,000 steps of minimization and 60
ps of heating from the 0 K up to 300 K before each main MD. Then a 5
ns classical MD simulation was carried out. Because the nonstandard
amino acids and cobalt ion exist in the active site of NHase, standard
Charmm 27 forcefield cannot be used for active centre. Thus we
developed new parameters and topology compatible with
posttranslationally modified cysteines, amino acids involved in cobalt
coordination, and for Co (III) based on DFT/B3LYP/6-31G(d,p) and HF/6-
31G* quantum calculations (results unpublished). The dockings of
ligands into the optimized Rhodococcus rhodochrous J1 nitrile hydratase
and its variants were performed using AutoDock 4.2.[36]. The Charmm27
compatible topology and parameters for all the ligands were obtained
using CGenFF program[37]. Geometries of ligands were optimized using
Gaussian 98 code with the HF method.[38] The docked ligands were
flexible. Hydrogen atoms were added using psfgen tool from NAMD 2.7
suite. For each ligand-enzyme complex, 200 starts of Lamarckian genetic
algorithm were performed. In total 6400 binding positions were obtained.
Figures were prepared using PyMOL.[39]
Enzymatic assay
The regioselectivity of NHase towards aliphatic dinitriles was assayed in
a reaction mixture (500 μL) containing 10 mM KPB (pH 7.4), 20 mM
adiponitrile/malononitrile, and 10 μL of the purified enzyme with a
concentration of 0.2 mg/mL. The reaction was carried out at 35 °C for 4 h
and stopped by adding 0.5 mL of methanol. The products were
determined using high-pressure liquid chromatography (HPLC) (Hitachi,
Tokyo, Japan) with a HITACHI LACHROM C18 (5UM) column. The
following solvent system was used: 25 mM H3PO4 buffer (pH 2.5) and
methanol (89.01:10.09 vol/vol), at the flow rate of 1 mL/min. The
wavelength of 210 nm was used for monitoring.[5a] The kinetic
parameters for 5-CVAM and CAAM hydration by NHases were
determined in 10 mM KPB at 35 °C for 3 min, and the concentration of
the enzyme was 0.3 mg/mL. The concentration of 5-CVAM and CAAM
were 5-600 mM and 5-200 mM respectively. The results were calculated
by Michaelis-Menten equation model. The regioselectivity of NHase
towards the aromatic dinitriles were determined by the addition of 2 mM
terephthalonitrile/5 mM phthalodinitrile to the reaction mixture described
above,
the
following
HPLC
solvent
system
was
used:
water/acetonitrile/acetic acid
=
2250/750/15 (ml), sodium 1-
octanesulfonate 6.45 g. The wavelength of 254nm was used for
monitoring.[7] The concentration of 4-CBAM and 2-CBAM for kinetic
measurement were 0.5-10 mM and 0.2-5 mM respectively. The reaction
was carried out in 10 mM KPB at 35 °C for 3 min.
Computational design methods
Using homology model which is a state of the art in protein engineering is
often a good choice as sometimes it is rather difficult to obtain the crystal
structure of a certain protein. To find the best template for homology
modelling of wild-type L-NHase, three iterations of PSI-BLAST[10] with
standard settings against PDB database were performed. Two of the
Acknowledgements
This work is financially supported by the National natural science
foundation of China (31400078, 31400058), a Project Funded by
the Priority Academic Program Development of Jiangsu Higher
Education Institutions, the 111 Project (No. 111-2-06), the
Jiangsu province "Collaborative Innovation Center for Advanced
Industrial Fermentation" industry development program, the
International S&T Innovation Cooperation Key Project
(2016YFE0127400), and the Fundamental Research Funds for
the Central Universities (JUSRP51713B).
most similar sequences were found: 1V29 chain B with E-value 2e-107
,
38% of identities, 56% of positives compared with the chain B of L-
NHase; 1V29 chain A with E-value 3e-107, 38% of identities, 56% of
positives compared with the chain A of L-NHase. As a result, NHase from
Bacillus smithii SC-J05-1 (PDB code: 1V29) was chosen as the template
in Swiss-Model[12b, 30]. For each chain, the best model showing GMQE
and QMEAN values equal to 0.40 and -1.25 respectively in case of
homology model based on chain A of 1V29 structure and in case of
homology model based on chain B, the values equal to 0.41 and -1.54.
In order to obtain proper structure of our constructed model, minimization
of potential energy (10000 steps) using Charmm27 force field was used
for whole protein. Constructed model was evaluated using Verify 3D
online tool[31] where 96.56% of the residues had an averaged 3D-1D
score bigger than 0.2. RMSD calculated by Pymol between our model
and 1V29 structure was 0.093. The βY68T and βW72Y point mutation
Keywords: Nitrile hydratase • Regioselectivity • Semi-rational
design
References
[1]
M. Kobayashi, S. Shimizu, Nat. Biotechnol. 1998, 16, 733.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
[2]
A. S. R. Banerjee, U. C. Banerjee, Appl. Microbiol. Biotechnol.
R. L. Dunbrack, J. D. Evanseck, M. J. Field, S. Fischer, J. Gao, H.
Guo, S. Ha, J. Phys. Chem. B. 1998, 102, 3586.
2016, 60, 1-1.
[3]
[4]
R. DiCosimo, B. Stieglitz, R. D. Fallon, US 5728556 A[P]. 1998.
R. C. Zheng, X. J. Yin, Y. G. Zheng, J. Chem. Technol. Biotechnol.
2016, 91, 1314-1319.
a) Y. Shen, M. Wang, X. Li, J. Zhang, H. Sun, J. Luo, J. Chem.
Technol. Biotechnol. 2012, 87, 1396–1400; b) E. C. Hann, A.
Eisenberg, S. K. Fager, N. E. Perkins, F. G. Gallagher, S. M.
Cooper, J. E. Gavagan, B. Stieglitz, S. M. Hennessey, R. Dicosimo,
Bioorg. Med. Chem. 1999, 7, 2239.
[35]
[36]
G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart,
R. K. Belew, A. J. Olson, J. Comput. Chem. 2015, 19, 1639-1662.
K. Vanommeslaeghe, E. Hatcher, C. Acharya, S. Kundu, S. Zhong,
J. Shim, E. Darian, O. Guvench, P. Lopes, I. Vorobyov, J. Comput.
Chem. 2010, 31, 671-690.
Gaussian 98 (RevisionA.9): M. J. Frisch et al.
E. Lilkova, P. Petkov, N. Ilieva et al. The PyMOL Molecular
Graphics System, Gaussian, Inc. 2015:10919-10925.
[5]
[37]
[38]
[6]
[7]
[8]
M. T. Reetz, Directed evolution of selective enzymes : catalysts for
organic chemistry and biotechnology, Wiley-VCH, Weinheim, 2016.
Z. Cheng, W. Cui, Z. Liu, L. Zhou, M. Wang, M. Kobayashi, Z.
Zhou, Catal. Sci. Technol. 2016, 6, 1292-1296.
a) Q. A. Li, D. V. Mavrodi, L. S. Thomashow, M. Roessle, W.
Blankenfeldt, J. Biol. Chem. 2011, 286, 18213; b) J. Carere, P.
Baker, S. Y. K. Seah, Biochemistry 2011, 50, 8407.
a) D. Zhu, Y. Yang, S. Majkowicz, H. Y. Pan, K. Kantardjieff, L.
Hua, Org. Lett. 2008, 10, 525-528; b) M. Anzini, M. Rovini, A.
Cappelli, S. Vomero, F. Manetti, M. Botta, L. Sautebin, A. Rossi, C.
Pergola, C. Ghelardini, J. Med. Chem. 2008, 51, 4476-4481.
S. F. Altschul, T. L. Madden, A. A. Schäffer, J. Zhang, Z. Zhang, W.
Miller, D. J. Lipman, Nucleic Acids Res. 1997, 25, 3389-3402.
W. R. Pearson, An Introduction to Sequence Similarity
(“Homology”) Searching. Current Protocols in Bioinformatics. 2009,
42:3.1:3.1.1–3.1.8.
[9]
[10]
[11]
[12]
a) M. Biasini, S. Bienert, A. Waterhouse, K. Arnold, G. Studer, T.
Schmidt, F. Kiefer, C. T. Gallo, M. Bertoni, L. Bordoli, Nucleic
Acids Res. 2014, 42, W252; b) F. Kiefer, K. Arnold, M. Künzli, L.
Bordoli, T. Schwede, Nucleic Acids Res. 2009, 37, 387-392; c) K.
Arnold, L. Bordoli, J. Kopp, T. Schwede, Bioinformatics 2006, 22,
195-201; d) N. Guex, M. C. Peitsch, T. Schwede, Electrophoresis
2009, 30, S162–S173.
[13]
[14]
[15]
L. Peplowski, K. Kubiak, W. Nowak, J. Mol. Model. 2007, 13, 725-
730.
M. T. Reetz, M. Bocola, J. D. Carballeira, D. Zha, A. Vogel, Angew.
Chem. 2005, 44, 4192-4196.
H. Takarada, Y. Kawano, K. Hashimoto, H. Nakayama, S. Ueda, M.
Yohda, N. Kamiya, N. Dohmae, M. Maeda, M. Odaka, Biosci.,
Biotechnol., Biochem. 2006, 70, 881.
[16]
[17]
[18]
[19]
[20]
[21]
Z. Zhou, Y. Hashimoto, K. Shiraki, M. Kobayashi, Proc. Natl. Acad.
Sci. U. S. A. 2008, 105, 14849-14854.
A. O. Tastan Bishop, T. Sewell, Bioche.m Biophys. Res. Commun.
2006, 343, 319-325.
L. Peplowski, K. Kubiak, W. Nowak, Chem. Phys. Lett. 2008, 467,
144-149.
J.-Y. van der Meer, L. Biewenga, G. J. Poelarends, ChemBioChem
2016, 17, 1792-1799.
a) M. T. Reetz, J. D. Carballeira, Nat. Protocols 2007, 2, 891-903;
b) M. T. Reetz, Angew. Chem. 2011, 50, 138–174.
a) W. Huang, J. Jia, J. Cummings, M. Nelson, G. Schneider, Y.
Lindqvist, Structure 1997, 5, 691; b) K. H. Hopmann, Inorg. Chem.
2014, 53, 2760-2762.
[22]
[23]
B. Sönmez, K. C. Yanık-Yıldırım, T. K. Wood, G. Vardar-Schara,
Biotechnol. Bioeng. 2014, 111, 1506-1512.
L. Song, M. Wang, J. Shi, Z. Xue, M. X. Wang, S. Qian, Bioche.m
Biophys. Res. Commun. 2007, 362, 319.
[24]
[25]
F. Effenberger, B. W. Graef, J. Biotechnol. 1998, 60, 165-174.
a) A. R. Fersht, J. S. Shindler, W. C. Tsui, Biochemistry 1980, 19,
5520-5524; b) R. F. Albu, M. Zacharias, T. P. Jurkowski, A. Jeltsch,
ChemBioChem 2012, 13, 1304–1311; c) D. M. Matje, H. Zhou, D.
A. Smith, R. K. Neely, D. T. Dryden, A. C. Jones, F. W. Dahlquist,
N. O. Reich, Biochemistry 2013, 52, 1677.
[26]
[27]
D. S. Tawfik, Curr. Opin. Chem. Biol. 2014, 21, 73.
M. Brouk, N.-L. Derry, J. Shainsky, Z. B.-B. Zelas, Y. Boyko, K.
Dabush, A. Fishman, J. Mol. Catal. B: Enzym. 2010, 66, 72-80.
a) O. Dubreuil, M. Bossus, M. Graille, M. Bilous, A. Savatier, M.
Jolivet, A. Ménez, E. Stura, F. Ducancel, J. Biol. Chem. 2005, 280,
24880-24887; b) P. Xiao, X. Wang, H. M. Wang, X. L. Fu, F. A. Cui,
X. Yu, S. S. Wen, W. X. Bi, J. P. Sun, Int. J. Biochem. Cell Biol.
2014, 57, 84.
[28]
[29]
[30]
M. Laible, K. Boonrod, J. Visualized Exp. 2008, 11, e1135-e1135.
a) M. Biasini, S. Bienert, A. Waterhouse, K. Arnold, G. Studer, T.
Schmidt, F. Kiefer, T. G. Cassarino, M. Bertoni, L. Bordoli, Nucleic
Acids Res. 2014, 42, W252; b) K. Arnold, L. Bordoli, T. Schwede,
Bioinformatics 2008, 22, 195-201; c) N. Guex, M. C. Peitsch, T.
Schwede, Electrophoresis 2009, 30, S162–S173.
[31]
[32]
a) D. B. J. U. Eisenberg, R. Luthy, 1993, pp. 164-170; b) D.
Eisenberg, R. Lüthy, J. U. Bowie, Methods Enzymol. 1992, 356,
83-85.
A. Miyanaga, S. Fushinobu, K. Ito, T. Wakagi, Biochem. Biophys.
Res. Commun. 2001, 288, 1169-1174.
[33]
[34]
K. Kubiak, W. Nowak, Biophys. J. 2008, 94, 3824.
a) J. C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E.
Villa, C. Chipot, R. D. Skeel, L. Kalé, K. Schulten, J. Comput.
Chem. 2005, 26, 1781; b) A. D. Mackerell, D. Bashford, M. Bellott,
This article is protected by copyright. All rights reserved.

10.1002/cctc.201701170
ChemCatChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Nitrile hydratase from Rhodococcus
rhodochrous J1 is able to catalyse α,
ω-dinitrile into corresponding final
product, diamide. Introduction of
Y68T/W72Y at the βsubunit inverses
the regioselectivity of NHase towards
dinitrile and results in the formation of
Zhongyi Cheng, Wenjing Cui, Yuanyuan
Xia, Lukasz Peplowski, Michihiko
Kobayashi,* Zhemin Zhou*
Page No. – Page No.
Modulation of NHase Regioselectivity
towards Dinitriles by Tailoring the
Substrate Binding Pocket Residues
valuable
cyanocarboxamide
intermediate.
This article is protected by copyright. All rights reserved.