Stereo- and Regioselectivity in Catalyzed Transformation of a 1,2-Disubstituted Vicinal Diol and the Corresponding Diketone by Wild Type and Laboratory Evolved Alcohol Dehydrogenases

Article abstract of DOI:10.1021/acscatal.8b01762

ADH-A from Rhodococcus ruber DSM 44541 catalyzes the oxidation of (S)-1-phenylethanol 3000-fold more efficiently as compared with the 2-hydroxylated derivative (R)-phenylethane-1,2-diol. The enzyme is also highly selective for sec-alcohols with comparably low activities with the corresponding primary alcohols. When challenged with a substrate containing two secondary alcohols, such as 1-phenylpropane-(1R,2S)-diol, ADH-A favors the oxidation of the benzylic carbon of this alcohol. The catalytic efficiency, however, is modest in comparison to the activity with (S)-1-phenylethanol. To investigate the structural requirements for improved oxidation of vicinal diols, we conducted iterative saturation mutagenesis combined with activity screening. A first-generation variant, B1 (Y54G, L119Y) displays a 2-fold higher k_cat value with 1-phenylpropane-(1R,2S)-diol and a shift in the cooperative behavior in alcohol binding, from negative in the wild type, to positive in B1, suggesting a shift from a less active enzyme form (T) in the wild type to a more active form (R) in the B1 variant. Also, the regiopreference changed to favor oxidation of C-2. A second-generation variant, B1F4 (F43T, Y54G, L119Y, F282W), shows further improvement in the turnover and regioselectivity in oxidation of 1-phenylpropane-(1R,2S)-diol. The crystal structures of the B1 and B1F4 variants describe the structural alterations to the active site, the most significant of which is a repositioning of a Tyr side-chain located distal to the coenzyme and the catalytic zinc ion. The links between the changes in structures and stereoselectivities are rationalized by molecular dynamics simulations of substrate binding at the respective active sites.

Full text of DOI:10.1021/acscatal.8b01762

Subscriber access provided by Technical University of Munich University Library
Article
Stereo- and Regioselectivity in Catalyzed Transformation of a
1,2-Disubstituted Vicinal Diol and the Corresponding Diketone
by Wild Type and Laboratory Evolved Alcohol Dehydrogenases
Dirk Maurer, Thilak Reddy Enugala, Emil Hamnevik, Paul Bauer, Malin Lüking, Dušan
Petrovi#, Heidi Hillier, Shina C. L. Kamerlin, Doreen Dobritzsch, and Mikael Widersten
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.8b01762 • Publication Date (Web): 05 Jul 2018
Downloaded from http://pubs.acs.org on July 7, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
Stereo- and Regioselectivity in Catalyzed Transformation of a 1,2-Disubstituted
Vicinal Diol and the Corresponding Diketone by Wild Type and Laboratory
Evolved Alcohol Dehydrogenases
†
,‡
†,‡
‡
§
§
Dirk Maurer, Thilak Reddy Enugala, Emil Hamnevik, Paul Bauer, Malin Lüking, Dušan
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
§
‡
§
‡
‡
Petrović, Heidi Hillier, Shina C. L. Kamerlin,* Doreen Dobritzsch,* and Mikael Widersten*
‡
Department of Chemistry – BMC, Uppsala University, Box 576, SE-751 23 Uppsala, Sweden
§
Department of Cell and Molecular Biology, Uppsala University, Box 596, SE-751 24 Uppsala,
Sweden
*
Corresponding authors:
S. C. L. Kamerlin (kamerlin@icm.uu.se), D. Dobritzsch (doreen.dobritzsch@kemi.uu.se) and
M. Widersten (mikael.widersten@kemi.uu.se)
Abstract
ADH-A from Rhodococcus ruber DSM 44541 catalyzes the oxidation of (S)-1-phenylethanol 3,000-
fold more efficiently as compared to the 2-hydroxylated derivative (R)-phenylethane-1,2-diol. The
enzyme is also highly selective for sec-alcohols with comparably low activities with the corresponding
primary alcohols. When challenged with a substrate containing two secondary alcohols, such as 1-
phenylpropane-(1R,2S)-diol, ADH-A favors the oxidation of the benzylic carbon of this alcohol. The
catalytic efficiency, however, is modest in comparison to the activity with (S)-1-phenylethanol. To
investigate the structural requirements for improved oxidation of vicinal diols we conducted iterative
saturation mutagenesis combined with activity screening. A first-generation variant, B1 (Y54G,
L119Y) displays a two-fold higher kcat value with 1-phenylpropane-(1R,2S)-diol and a shift in the
cooperative behavior in alcohol binding, from negative in the wild type, to positive in B1, suggesting a
shift from a less active enzyme form (T) in the wild type to a more active form (R) in the B1 variant.
Also, the regiopreference changed to favor oxidation of C-2. A second-generation variant, B1F4 (F43T,
Y54G, L119Y, F282W), shows further improvement in the turnover and regioselectivity in oxidation of
1
-phenylpropane-(1R,2S)-diol. The crystal structures of the B1 and B1F4 variants describe the
structural alterations to the active site the most significant of which is a re-positioning of a Tyr side-
chain located distal to the coenzyme and the catalytic zinc ion. The links between the changes in
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
structures and stereoselectivities are rationalized by molecular dynamics simulations of substrate
binding at the respective active sites.
Keywords Alcohol dehydrogenase; alcohol oxidation; directed evolution; enzyme engineering;
biocatalysis; stereoselectivity; crystal structure; molecular dynamics simulations
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1. INTRODUCTION
The significant contribution of biocatalytic transformations to the sustainable manufacturing of
chemicals is undisputed. Hence, a fundamental understanding of structure-activity relationships of
enzymes that catalyze relevant chemical transformations is necessary for the future development of
effective biocatalysts. Of particular importance is understanding of the structural requirements of the
active site in order to steer stereoselectivity in substrate acceptance or the resulting stereoconfiguration
of reaction products.
A cornerstone in synthetic chemistry is the oxidation/reduction of alcohols and their respective
carbonyl derivatives. These transformations are generally catalyzed by alcohol dehydrogenases in
biological systems. A class of particularly attractive synthetic building blocks are α-hydroxy ketones
(
acyloins) which are precursors in the synthesis of e.g. bio-active natural products.1,2 The biocatalytic
productions of acyloins has been approached using different enzyme systems including
3
monooxygenases introducing alcohol functions into phenyl-substituted aliphatic ketones, or by alcohol
dehydrogenase-catalyzed oxidation of 1,2-diols.4,5 Successful biocatalytic transformations of reduction
of the corresponding ketones, producing asymmetric alcohols from acyloins, diketones or cyclic
ketones, have also been reported making use of whole-cell systems,6-8 naturally occurring
dehydrogenases9,10 or in vitro engineered variants.11,12
Alcohol dehydrogenase A (ADH-A) from the bacterium Rhodococcus ruber (strain DSM 44541) is
a potential biocatalyst candidate for redox transformations. ADH-A has been shown to retain its
catalytic function in high concentrations of acetone and isopropanol,13-15 a feature that is beneficial
from different perspectives. In particular, it allows for the regeneration of the redox state of the
+
coenzyme NAD or NADH, depending on the reaction direction, simply by adding either acetone or
isopropanol as a co-substrate in the reaction mixture. Hence, only catalytic amounts of the costly
coenzyme are required with maintained good conversion into product. The possibility of adding higher
2
ACS Paragon Plus Environment

Page 3 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
concentrations of organic solvents also facilitates solvation of non-polar substrates that may be of
synthetic interest.
ADH-A displays strong regiopreference for sec-alcohols and ketones. Primary alcohols are very
16
poor substrates in comparison. For example, (1S)-1-phenylethanol ((S)-1 in Figure 1A) is oxidized
5,000-fold more efficiently than 2-phenylethanol. Furthermore, vicinal diols are comparably poor
substrates; (1R)-1-phenylethanediol ((R)-2) is transformed with a 3,000-fold lower efficiency as
compared to (S)-1. As expected from the strong preference for sec-alcohols, given the choice of two
alcohol functions, as in (R)-2, ADH-A exclusively oxidizes the secondary alcohol to 2-
hydroxyacetophenone (5). This capacity to regioselectively oxidize 1,2-diols, producing the acyloin
products, makes ADH-A a suitable model for studies that may clarify the underlying reasons for the
observed regioselectivity in this enzyme. The dramatic differences in activity towards (S)-1 and (R)-2
further qualifies this enzyme for studies of substrate selectivity, again with special focus on
stereoselectivity. Furthermore, the established stereoselectivity of ADH-A for sec-alcohols motivated
us to study the regiopreference with a vicinal diol substrate containing two secondary alcohol
functions, (1R,2S)-3. We conducted ISM-driven laboratory evolution of the enzyme with the aim to
isolate more efficient catalysts of acyloins from (1R,2S)-3. Here, we demonstrate the successful
isolation of ADH-A variants that exhibit increased turnover numbers for the oxidation of (1R,2S)-3 and
in the reduction of the corresponding diketone (8). We are also able to rationalize the stereopreference
from molecular dynamics simulations using crystal structures of these variants as the structural basis.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
. MATERIALS AND METHODS
2.1. Chemicals and reagents. (R)-1 (≥99 %, #77848), (S)-1 (≥99 %, #77849), (R)-2 (99 %, #302163),
S)-2 (99 %, #302155), 4 (≥99 %, #00790), 5 (98 %, #445134) and 8 (99 %, #223034) were purchased
from Merck (Darmstadt, Germany). (1R,2S)-3 was produced by hydrolyzing 1-phenyl-(1S,2S)-
(
1
7
propaneoxide (98 %, #287792, Merck) using epoxide hydrolase as described before (Figure 1B). All
other reagents, solvents, substrates, coenzymes, microbial growth media, and chromatographic resins
were purchased from commercial sources at highest available purity.
2.2. Gene library construction. The design of libraries for ISM-driven18-20 evolution of ADH-A
targeted amino acid residues constituting the active-site cavity as deduced from the crystal structure
21
22
(
PDB code: 3jv7 ). The construction of library B has been described before. Library B1F was
constructed similarly with some modifications and was created from the first-generation ‘B1’ variant as
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
parent. The ‘F’ site contained two residues, F43 and F282*, where the asterisk indicates that this
residue is located in the neighboring subunit of the dimer. Codon degeneracy was minimized using
2
3
oligonucleotides containing mixtures of NDT, VMA, ATG and TGG codons, thus, the theoretical
number of unique sequences in the constructed library was 400. Detailed descriptions of the library
construction and sequences of used oligonucleotides are provided in the Supporting Information (see
also Figure S1 and Table S1). The gene constructs were subcloned into the XhoI and SpeI sites of
pGT7ADHA-5H.16 Plasmids were transformed into Escherichia coli BL21-AI (Invitrogen) already
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
4
carrying a plasmid for over-expression of chaperonins GroEL/ES, through electroporation and grown
on LB plates containing 100 mg/ml ampicillin and 30 mg/ml kanamycin.
2
.3. Library screening. Screening was performed essentially as described earlier.22 Clones were
incubated in round-bottom Nunc 96-well plates over 2 days. Day one, 350 µl 2TY (1.6 % (w/v)
tryptone, 1.0 % (w/v) yeast extract, 0.5 % (w/v) NaCl) containing 100 µg/ml ampicillin and 30 µg/ml
kanamycin was added to each well and single colonies were inoculated into designated wells. Each
plate contained four parent enzyme clones as baseline controls and four negative controls with cells
2
5
expressing an unrelated protein from the same expression construct (yeast YNR064c ) and four blank
controls containing only growth medium. Gas permeable seals were placed on the plates followed by
incubation at 30 °C for 16 hours under rotation at 150 rpm. On the following day, overnight cultures
were diluted 14-fold into new plates containing 325 µl 2TY, 50 µg/ml ampicillin and 30 µg/ml
kanamycin. Gas permeable seal was placed on the plates and the plates were incubated as described for
3
hours. After this incubation, β-D-thiogalactopyranoside (IPTG) and L-arabinose were added to each
well to final concentrations of 1 mM and 0.04 % (w/v), respectively, and the plates were incubated as
before for 16 hours. Following protein expression, cells were harvested by centrifugation at 3,000×g, at
4
°C for 20 min. Supernatants were removed and plates were dried up-side down on paper tissues for 2
min before being covered with alumina foil. Sealed plates were stored at -80 °C until analyzed. Cells
were lysed by addition of 100 µl lysis solution (10 ml B-PER (Thermo Fisher Scientific, Waltham,
MA, USA; 4 mg lysozyme, 2 mg DNase, 1 EDTA-free Complete mini-tablet, Merck; 10 µM ZnSO ),
4
then incubated for 1 h on a table-top mixer (Vortex-Genie 2, Scientific Industries, Bohemia, NY, USA)
at lowest speed. Lysed cells were diluted with 100 µl assay buffer (0.1 M sodium-phosphate, pH 8.0,
10 µM ZnSO ) and centrifuged for 60 min at 3,000×g for library in a swing-out centrifuge. The
4
oxidation activities of expressed proteins with either (1R,2S)-3 or isopropanol, as positive control, were
measured in clear flat-bottom Nunc 96-well plates in a plate reader (SpectraMAX 190, Molecular
4
ACS Paragon Plus Environment

Page 5 of 43
ACS Catalysis
1
2
3
4
5
devices, Sunnyvale, CA, USA) at 340 nm, 30 °C for 10-20 min. The conditions in each well for the
+
measurement were 0.1 M sodium-phosphate, pH 8.0, 10 µM ZnSO
4
, 0.5 mM NAD and 10 mM
(1R,2S)-3. For each activity measurement, 40 µl lysate was added in a total volume of 150 µl. Clones
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
were scored as hits dependent on the degree of increase in the reaction velocity as compared to the
parent control. A re-screen was performed on isolated hits in triplicates. Overnight cultures of 2 ml
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
TY and 100 µg/ml ampicillin were made for all hits re-screened. Plasmid DNA was extracted and the
ADH-A genes were sequenced in full. At least 1200 individual clones were screened from each library.
2
.4. Enzyme expression and purification. Enzyme variants were expressed and purified as
16
described before with the differences that expression of ADH-A variants and GroEL/ES were induced
by addition of 0.04 % (w/v) L-arabinose and 1 mM IPTG, respectively. Cultures were incubated for 16-
18 h at 30 °C. Cells were harvested by centrifugation at 5,000×g and lysed either by French press or
ultrasonication into Binding Buffer (20 mM sodium-phosphate, pH 7.5, 20 mM imidazole, 0.5 M
NaCl). Lysate was cleared by centrifugation as described before. Purification was performed by
immobilized metal ion chromatography using Ni(II) charged Chelating Sepharose (GE Healthcare,
Uppsala, Sweden). Protein was adsorbed in batch at 4 °C for 1 h and the gel/lysate slurry was
subsequently transferred to a column. The flow through was discarded and the gel was washed thrice
by addition of 10×gel volume of Washing Buffer (Binding Buffer containing 100 mM imidazole).
Protein was eluted by addition of 2×5 ml Elution Buffer (Binding Buffer containing 300 mM
imidazole). The eluted protein was desalted into Storage Buffer (0.1 M sodium-phosphate, pH 7.4, 10
µM ZnSO
ZnSO and then concentrated. Protein purity was confirmed by SDS-PAGE stained with Coomassie
Brilliant Blue-R250. The purity was in all cases deemed to be >98 %.
.5. Steady state kinetics. Kinetic parameters were determined for variants of interest by measuring
initial velocities in the presence of varying concentrations of (S)-1, (R)-(2), (S)-2, (1R,2S)-3, 4, 5 and 8
4
). For protein crystallization the protein was desalted into 30 mM Tris-HCl, pH 7.4, 10 µM
4
2
+
22
(
Figure 1A) in the presence of saturating levels of NAD or NADH as described previously. Curve
fitting and model discrimination was performed using programs in the SIMFIT package
http://www.simfit.org.uk/). The steady state parameters kcat and K were determined after fitting the
(
M
Michaelis-Menten equation to the experimental data using MMFIT. In cases where cooperative
behavior was observed in the substrate concentration dependence of the initial reaction velocities, a
2
6
modified Hill equation (Eq. 1) was fitted to the raw data, using program INRATE. The degree of
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cooperativity was defined by the Hill coefficient, n. The validity of the applied model was judged by F-
tests applying a threshold of P<0.01.
n
v0
k [ S]
cat
n
=
(1)
n
[
E]tot
K
0.5+[ S]
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2.6. Pre-steady state kinetics. Transient kinetics experiments were performed on an Applied Photo-
physics (Leatherhead, UK) SX.20MV sequential stopped-flow spectrophotometer/fluorometer. All
measurements were performed at 30 °C in 0.1 M sodium phosphate, pH 8.0, 10 µM ZnSO . Averaged
4
progression curves (≥5) were fitted to Eq. 2, a function of a single exponential with floating endpoint,
for determination of the apparent rates, kobs. F is the recorded fluorescence signal, A, the signal
amplitude, t, time and C, the floating end point.
F= Aexp(− k t)+C
obs
(2)
2
.6.1. Coenzyme binding. NADH binding to the enzymes was detected by monitoring the increase
in NADH fluorescence occurring upon binding of the coenzyme to the active site. An excitation
wavelength of 340 nm was used and fluorescent light was collected through a 420 nm cut-off filter.
+
NAD binding was detected by monitoring the quenching of the intrinsic tryptophan fluorescence.
Excitation was achieved by shining light of 290 nm, and the emitted light >320 nm was recorded.
Nucleotide concentrations were always kept at least in 10-fold excess over the enzyme concentrations.
The apparent rates, kobs, were determined as described above. Kinetic parameters were determined by
fitting the dependence of kobs on the co-enzyme concentration to a linear (Eq. 3) function using
program LINFIT.
k = k [NAD(H)]+k
obs
on
off
(3)
2
.6.2. Alcohol oxidation. The change in fluorescence from the increase in the NADH concentration,
as a result of alcohol oxidation was monitored. The oxidation reaction was measured with varying
+
concentrations of alcohol and saturating concentration of NAD (0.4 mM). The enzyme concentrations
were between 0.6 to 2.3 µM. Averaged progression curves (≥5) were fitted Eq. 2 to extract kobs. The
6
ACS Paragon Plus Environment

Page 7 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
kinetic parameters, k
3
and K
2
were estimated by fitting the observed rates to Eq. 4 by non-linear
regression using MMFIT. K
2
is the equilibrium dissociation constant (k-2/k ) of the alcohol from the
2
ternary enzyme-substrate complex. Numbering of rate and equilibrium constants are based on the
kinetic model in Figure 2.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
k3[R-OH]
k =
(4)
obs
K +[R-OH]
2
2
2
.7. Synthesis and analysis of acyloins.
.7.1. Oxidation of (1R,2S)-3. Wild type ADH-A or B1 or B1F4 (3 µM), were mixed with 280 mM
+
(1R,2S)-3 and 16 mM NAD . For nucleotide regeneration 280 mM sodium pyruvate and 2 U/ml lactate
dehydrogenase were included. Reactions were carried out in 0.1 M sodium-phosphate, pH 8.0, with 10
µM ZnSO
.7.2. Reduction of 8. Reactions were performed as in 2.7.1 but with 300 mM 8, 0.4 mM NADH,
300 mM L-lactate (300 mM) and 2 U/ml lactate dehydrogenase.
.7.3. Product analysis. The structures of reaction products were determined by NMR
spectroscopy. Reaction mixtures were centrifuged at 5000×g to pellet debris. End-point proton NMR
400 MHz) analysis was performed on the organic phase following extraction into deuterated
chloroform (CDCl ). NMR spectra are shown in Figure S2.
.8. X-ray crystallography. The B1 and B1F4 variants were crystallized by hanging drop vapor
4
, at 30 °C in the dark for 30 h.
2
2
(
3
2
diffusion at 20 °C. The drops were placed over a 1 ml reservoir well and consisted of 1 µl protein
solution (8 mg/ml and 3 mg/ml of B1 and B1F4 respectively in 50 mM Tris-HCl, pH 8, 4 mM NAD+)
and 1 µl reservoir solution containing 18 % and 15 % (w/v) polyacrylic acid 5100 for B1 and B1F4
respectively, 100 mM Tris-HCl, pH 8.0 and 10 mM MgCl
2
. The crystals were cryo-protected by
soaking in a solution of 19 % (w/v) polyacrylic acid 5100, 100 mM Tris-HCl, pH 8.0, 10 mM MgCl
2
and 10 % (v/v) glycerol for 10 seconds and flash frozen in liquid nitrogen. Crystallographic data were
collected at 100 K at beamlines I04 or I24 of the Diamond Light Source (Didcot, UK). The data was
autoprocessed and indexed using FAST DP27 on site and further processed using COMBAT and
AIMLESS from the CCP4 software suit.28-30 The data collection statistics are shown in Table S2.
3
1
The structures of both variants were solved by molecular replacement using PHASER with the
2
1
ADH-A wild type structure (pdb ID: 3jv7 ) as a search model. Manual model building as well as the
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
2
addition of water molecules was performed with COOT and restrained and TLS refinement with
REFMAC5.33 For monitoring Rfree, a set of randomly selected reflections (5%) was used. The
crystallographic models and data were deposited in the Protein Data Bank with accession codes 5od3
for B1 and 6fg0 for B1F4.
2.9. Molecular dynamics of substrate binding. Wild type ADH-A and its B1 and B1F4 variants
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
4
35
were simulated using AMBER16 and the ff14SB force field to describe the protein. The diols (R)-
3
6
37
and (S)-2 and (1R,2S)-3 were parametrized using Antechamber and the standard RESP protocol, and
the resulting parameters are provided in Tables S3-S12 of the Supporting Information. Parameters for
+
38,39
NAD and NADH were taken from the literature.
As the system contains two divalent zinc ions per
subunit, we used a cationic dummy model to describe these metal centers, based on our success with
this model in previous studies.40-43 In the present work, dummy model parameters were obtained for a
tetrahedral zinc model, based on the previous published data.44,45 Full details of the zinc
parametrization are provided in the Supporting Information.
For wild type and B1, the B and C subunits of the tetramer in the crystal structures (3jv7 and 5od3,
respectively) and the A and D chains with B1F4 (pdb code 6fg0) were used to simulate the protein in
the dimeric form observed in solution. Where multiple rotamers of a given side chain were plausible,
the rotamer with the higher occupancy was selected for the simulations, and where the different
rotamers have equal probabilities, the first rotamer was selected. All crystallographic water molecules
were then stripped from the system, before re-solvating the protein in an octahedral box of TIP3P
4
6
water , with the closest solute atom at least 10 Å away from the box edges. Sodium and chloride ions
were added to a 0.15 M concentration, and the substrates were placed manually in the active site to
allow coordination of the hydroxyl moiety at the carbon involved in the hydride transfer towards the
catalytic zinc. All structures were minimized in three steps with decreasing restraints on the protein
system from 25 kcal/mol×Ų to zero, with 5000 minimization cycles at each step. Non-bonded
interactions were explicitly calculated for all atoms within 8 Å of each other, with interactions beyond
this cut-off being treated using the particle mesh Ewald (PME) approach. Following this, the systems
were heated up over the course of 50 ps to the final temperature of 300 K in an NVT ensemble, again
applying a modest restraint of 5 kcal/mol×Ų on the solute atoms. The step size from this point on was
4
7
set to 2 fs, with SHAKE applied to the hydrogen atoms in the system. Temperature control was
achieved using the Langevin thermostat. Following this, the system was equilibrated under NPT
conditions for additional 150 ps under the same conditions. During both steps, metal geometries where
8
ACS Paragon Plus Environment

Page 9 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
2
kept constrained to the crystal values using distance restraints of 20 kcal/mol/Å , with an equilibrium
distance of 1.8 Å, as well as angle restraints for tetrahedral coordination with force constants of 15
kcal/mol/deg. This was followed by dropping the restraints down to zero over the course of 5 additional
simulations, each 10 ps in length and reducing restraints by 1 kcal/mol each. The last equilibration step
was performed for 10 ps without any restraints on the system, including the metal center. Additional
distance restraints were applied to prevent E53 from coordinating the catalytic zinc during the during
all equilibration steps except the final one. The restraints were chosen such as to add a 20 kcal/mol/Å
penalty starting at a distance of 5.2 Å, with the well curvature set such as to increase the penalty to 50
kcal/mol/Å at 4.8 Å.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Following the equilibrations steps, the systems were further equilibrated for 40 ns in an NPT
ensemble, with the glutamate restraints alternatively activated or disabled every 5 ns to achieve a stable
coordination. After this, the systems were simulated for 60 ns without any restraints in an NPT
ensemble at 300 K. All simulations were performed in triplicate from individually equilibrated
structures.
Data analysis was performed on the snapshots captured from every 50 ps of the calculations,
excluding the restrained equilibration. Protein and substrate RMSD values, as well as protein clusters,
4
8
were calculated using CPPTRAJ after backbone alignment. Clusters were centered on the residues
surrounding the substrate bound in the active site, specifically residues at positions 383, 385, 407, 464,
4
98, 640, 694, 696 and 697, using the average linkage method and rms mass comparison with an
epsilon of 4 and random sieve of 5. The production trajectories were further analyzed using R (version
.4.4) to obtain cross-correlation information. For this, the concatenated coordinates were loaded into R
3
and analyzed using the Bio3D package, with the two simulated subunits analyzed separately by first
fitting to the C-alpha trace of the subunit, followed by PCA analysis. The script used for the analysis is
available on request. Final analysis of atomic fluctuations was performed using CPPTRAJ as used
before for RMSD analysis. For this, the coordinates were first fit to the average structure backbone
atoms, before calculating the atomic fluctuation by residue. The final numbers are fluctuations in Å.
3
. RESULTS AND DISCUSSION
.1. Stereoselectivity of ADH-A in oxidation of vicinal diols. Wild type ADH-A catalyzes the
oxidation of the benzylic carbon of vicinal diol 2 with a 14-fold preference for the R-enantiomer, as
3
judged by the corresponding kcat/K
M
values (Table 1). Given the relatively poor activity with (R)-2, as
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compared to the non-hydroxylated derivative (S)-1, it was not surprising to find that (1R,2S)-3 is
oxidized at a comparably low steady state rate. The velocity dependence on substrate concentration
with this compound exhibits clear negative cooperativity with a Hill coefficient of 0.75. This behavior
49
is indicative of stabilization of a less active enzyme form (‘T’ in classical MWC nomenclature ) with
increasing concentration of (1R,2S)-3. The enzyme preferentially oxidizes the benzylic C-1 carbon
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
producing acyloin (S)-6 in a ratio of 3.6:1 over (R)-7 as judged by NMR spectroscopy (Table 2). The
-1
catalytic activity with the 1,2-diketone 8 is high; kcat is 62 s and the kcat/K0.5 value places this
substrate as the most efficiently transformed by this enzyme so far (Table 1). The reduction of 8 is
preferentially at C-2 forming acyloin 6 (Table 2).
We complemented the kinetics analyses with molecular dynamics (MD) simulations. The 80-fold
lower kcat with (S)-2 as compared to (R)-2, can be rationalized from the simulations as the hydride to be
transferred is pointing away from the nicotinamide C-4 (Figure 3B). Furthermore, the (1R,2S)-3 diol
binds tightly within the active site in an orientation favorable for hydride transfer from the benzylic C-1
carbon, with a distance of approximately 2.7 Å in the main cluster (Figure 3C). The corresponding
distance for transfer from C-2 is similar but with increased RMSD values for the substrate (Figure 3D).
The energetic difference between reaction at either C-1 or C-2, calculated from the product ratios of
3
1
.6, is only 0.8 kcal/mol, so either binding mode may be trapped in the MD simulation. In the case of
,2-diketone 8, there was no convergence towards a stable structure cluster. Instead, none of the
obtained averaged structures contains 8 bound at the active site in manner that is compatible with
hydride transfer from C-4 of NADH to any of the carbonyl carbons of the substrate. This suggests that
this flat and rigid molecule is not bound in a specific manner in the active site and that the relationships
between rates of sampling productive binding modes will decide regioselectivity and product
configuration(s). The experimentally observed regioselectivity in reduction of 8 was, however, caught
in the simulation with one generated of the ADH-A mutants (vide infra). Our MD results suggest
therefore that different combinations of substrate and enzyme lead to distinct binding modes and
complex stabilities; subtle changes in the dynamic behavior of certain parts of the active site are
apparently causing the differences in stereopreference.
3
.2. Change in regio- and stereoselectivity by active-site substitutions. With the goal to isolate
ADH-A variants with increased catalytic activity with (1R,2S)-3 we conducted targeted saturation
mutagenesis of four active-site residues. Targeted residues were chosen based on their expected
contribution to shaping the alcohol/ketone binding compartment of the active site, thereby influencing
1
0
ACS Paragon Plus Environment

Page 11 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
substrate selectivity. To facilitate screening for active enzyme variants the number of mutated residues
per constructed library were limited to two spatially adjacent residues but probing different parts of the
substrate binding site (Figure 4). A parallel aim with the laboratory evolution effort was to test if
regioselectivity in oxidation may be linked to catalytic activity, and if so, which structural
modifications of the active site were required?
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The chosen residues were grouped in pairs; site ‘A’ consisting of Y294 and W295 and site ‘B’, of
Y54 and L119 (Figure 4A). However, no enzyme variant from the A-site library displayed increased
activity with (1R,2S)-3 under the applied screening conditions. From the B-library, however, one
mutant, dubbed ‘B1’ displayed improved catalytic activity in the screen, when compared to the parent
enzyme (Table S13). The apparent improvement in enzyme activity is due to a two-fold increase in kcat
for (1R,2S)-3 and a shift from negative to positive cooperativity with this substrate (n=1.2). Since the
screening of the constructed enzyme library was performed in the presence of relatively high
concentration of diol substrate (10 mM), the higher turnover number in combination with the positive
cooperativity, resulted in an observed velocity increase in the screening assay, although the actual
catalytic efficiency, if expressed as kcat/K0.5, is actually lower than that of the wild type (Table 1). The
kinetic parameters, however, are not directly comparable since the Y54G, L119Y substitutions also
cause the B1 variant to shift regiopreference from oxidation of the benzylic C-1 to C-2 (Figure 5). As a
consequence, B1 catalyzed oxidation of (1R,2S)-3 produces (R)-7 as the major product (3.4:1) (Table
2
). Thus, (1R,2S)-3 is treated as two different substrates by the wild type and the B1 variant.
In the MD simulations, if (1R,2S)-3 is deprotonated on the C-1 hydroxyl, it binds in an orientation in
the B1 active site that leads to a stronger interaction between the substrate phenyl ring and the
nicotinamide ring of the coenzyme, but increases the distance between the hydride and the C-4 of the
co-factor to 5 Å (Figure 6A). The binding mode for the C-2-deprotonated substrate, albeit different
from the corresponding binding in the wild type enzyme (compare Figures 3D and 6B), positions the
substrate in a manner that may allow for hydride transfer with a distance to the coenzyme of <3 Å
(Figure 6B). The MD simulations are thus in agreement with the experimental observations that show
a regiopreference for oxidation of C-2. Since none of the substituted residues can be envisioned to
interact directly in either binding mode, changes in the dynamic nature of the active site may be a
possible reason for the change in regiopreference with (1R,2S)-3.
With the mono-substituted vicinal diols 2 as substrates, the wild type’s 14-fold preference for the R-
enantiomer over (S)-2, is shifted. The B1 variant instead displays a 2-fold preference for (S)-2 as
1
1
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
substrate, mainly caused by an increase in kcat( with a parallel decrease in kcat(R)-2 (Table 1) while the
S)-2
K
M
for either enantiomer increases to similar extents (6-8 – fold). These combined changes result in the
overall differences in kcat/K . In spite of the decrease in turnover numbers with (R)-2, kcat for this
M
enantiomer still exceeds kcat(S)-2 by >four-fold. In MD simulations with these substrates the binding
behavior is similar to what is observed with the wild type (Figure S3) and largely agrees with the
observed kinetics. When including 8 and NADH in the simulation, the MD results largely follow that
of the wild type as does the activity with this substrate. However, one cluster displays 8 bound inside
the active site relatively close to the cofactor (7 Å) although outside the range expected to be
compatible with efficient hydride transfer. The regioselectivity, as determined from the product
structure, is similar to the wild type with reduction preferentially at C-2 producing 6 (Table 2).
Although the overall crystal structure of the B1 variant is essentially indistinguishable from the wild
type structure (Figure 7A), with an RMSD of 0.26 Å for superimposition of the α-carbons, the residue
changes at positions 54 and 119 affect the active-site cavity. The substitutions act truly synergistically
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
Figure 7B) in reshaping the active site distal to the reactive part of the coenzyme (Figure 7C, D).
Although the Y54G, L119Y substitutions appear to have a relatively modest effect on overall active-
site architecture, they result in local changes in intramolecular interactions: in the wild type, the phenyl
ring of Y54 interacts via π-π bonds with the backbone amide of L119. This contact is lost in the B1
mutant which instead forms a new hydrogen bond with the main chain amide of A53. It should be
mentioned that these interactions are not present in all snapshot structures of the MD simulations which
suggests dynamic formation and disruption of these bonds. Although the topology of the active site is
not altered substantially in the B1 crystal structure the functional consequences of shifts in
regioselectivity with (1R,2S)-3 and in enantioselectivity from (R)-2 to (S)-2 are clear.
The B1 gene was subsequently recruited to parent a second round of mutagenesis and selection. A
library, ‘B1F’, was constructed, encoding randomized codons for residues F43 and F282, where F282
enters the active site from the neighboring subunit (Figure 4B). The rationale behind the design of this
library was to introduce structural variability spatially opposite to the Y54G, L119Y substitutions. A
number of variants that exhibited apparently higher activity with (1R,2S)-3 under the screening
conditions were identified (Table S13). Three of these F variants, B1F1, B1F4 and B1F5 were
2
expressed at larger scale, purified and characterized. B1F1 (F43H on the B1 background), shows
comparable activity to B1 with (1R,2S)-3, and also the other tested substrates, whereas B1F5 (B1,
F43S, V194I, F282P) displays lower activities also with (1R,2S)-3 (Table 1). However, both these
1
2
ACS Paragon Plus Environment

Page 13 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
variants, and especially B1F5, are produced at higher-than-average levels which can explain why they
were scored as hits during the screen in which the measured reaction velocities were not normalized for
protein expression levels.
Variant B1F4 (B1, F43T, F282W) displays improved activity with (1R,2S)-3 also after its
purification. The turnover number is increased 2.3-fold as compared to the parent B1 and 4.8-fold as
compared to wild type ADH-A (Table 1). The activity with (S)-2 is comparable to B1 while kcat( is
increased to a similar degree as with (1R,2S)-3. B1F4 also displays positive cooperativity in the
oxidation of (1R,2S)-3 (n=1.4). The regiopreference of B1F4 mimics that of the parent B1 but with a
further accentuated preference for oxidation at C-2 (Table 2 and Figure 5).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
R)-2
When analyzing the primary cluster obtained by clustering the structures obtained from the MD
simulations (based on occupancies of greater than 90 %), (1R,2S)-3 is bound in such a conformation as
+
to facilitate hydride transfer primarily from C-1 to NAD , although also the C-2 hydride is within 3.5 Å
from the coenzyme (Figure 6C, D). Thus, the MD simulations, in this case, cannot capture the shift in
regioselectivity observed experimentally. This is possibly due to the fact that the adjustments in
substrate binding mode that are required to achieve the observed shift in regiopreference from C-1 to
the closely positioned C-2, can be accommodated by the changes in the overall structure dynamics.
-1
The catalyzed reduction of diketone 8 is especially effective with a kcat of 160 s , a 2.5-fold increase
as compared to wild type ADH-A. The regiopreference in the catalyzed reductive reaction is, similar to
both other enzymes, towards C-2 (Table 2). Again, the primary cluster found when analyzing the MD
simulations as above shows that with B1F4 and 8 the substrate is bound in a conformation that would
preferentially favor hydride transfer to C-2, in accordance with the experimental results (Figure 8).
The crystal structure of B1F4 is overall very similar to the wild type and B1 structures (Figure 7A)
with an RMSD of the corresponding subunit’s alpha carbons of compared chains of 0.18-0.26 Å for
superimposition with B1, and 0.27-0.33 Å for superimposition with the wild type. The F43T and
F282W substitutions creates a somewhat wider entrance into the active site (Figure 7E), but the
differences are minor. T43 engages in a new water-linked hydrogen bond network involving the
carbonyl oxygen of I271 and the 2’ and 3’ hydroxyls of the ribose of the nicotinamide moiety of the
coenzyme (Figure 9). A similar interaction network has been observed in an F43S ADH-A variant
isolated in another directed evolution effort22 and the same F43S substitution is also present in a
number of F
2
variants isolated in this work (Table S13). The F43H substitution in B1F1-B1F3 has
been observed before12,22 and the H43 always hydrogen-bonds to the same ribose hydroxyl groups. The
1
3
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
significance of these new protein-coenzyme interactions is not yet clear but could possibly compensate
for other destabilizing mutations.
3
.3. Rate limitation for kcat and destabilization of nonproductive substrate binding. The two-
and five-fold increases in turnover of (1R,2S)-3 by B1 and B1F4, respectively, are, paradoxically, not
due to changes in rate-limiting reaction steps. We have previously shown that nucleotide release is rate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6
determining in the ADH-A catalyzed transformations of the preferred substrates (S)-1 and 4. The
deuterium kinetic isotope effect for kcat in the oxidation of isopropanol is significantly smaller
(D
k
cat=1.5) than that of the oxidation step (Dk =3.6)22 which further suggests that (a) step(s)
3
downstream of the chemical transformation is(are) rate determining.
Here, the oxidation rates of (1R,2S)-3) are also considerably faster than the corresponding turnover
numbers (Table 3) as are the release rates of NADH, which are 65 and 24-fold faster than kcat(1R,2S)-3 for
B1 and B1F4, respectively. The comparable rates of alcohol oxidation and NADH release result in
lower degrees of accumulation of the binary E•NADH complexes, and indeed, the amplitude of the
kinetic ‘burst’ of NADH formation during the pre-steady state phase of the reaction is significantly
lower as compared to e.g. (S)-1 (Figure S4).
Since both k
changes in the kinetic parameters must be caused also by other mechanisms. The parallel increases in
cat and K0.5 in the B1 and B1F4 catalyzed oxidation reactions suggest two things: (1) the frequency of
3
and k
5
exceeds the turnover rates by more than one order of magnitude, the observed
k
binding modes that allow for formation of productive ternary enzyme-substrate complexes has
increased at the expense of binding affinity for unproductive ground-state ternary complex(es), as
reflected in higher K0.5 values. This destabilization increases the probability of formation of productive
binding modes, observed as elevations of kcat. (2) productive binding modes require a shift in the
+
positioning of the diol substrate so that the C-2 hydride can now react with NAD . In contrast, in the
wild type enzyme the substrate is bound in a less reactive manner in which the benzylic (C-1) hydride
is preferentially transferred.
The nature of the productive ternary complex is illusive; is the increase in kcat simply a reflection of
a larger proportion of enzyme-substrate complexes with geometries that facilitates the climb to the
5
0
transition state, similar to the proximity effect of intramolecular catalysis or does it represent an
increase in ternary ‘near attack complexes’ (NACs) where the alcohol and coenzyme are bound in a
TS-like geometry, thereby lowering the energetic barrier for transformation into product?51-59 Improved
stabilization of a TS-like intermediate, such as a NAC, is expected to increase the rate of oxidation. The
1
4
ACS Paragon Plus Environment

Page 15 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
rates of oxidation (k
) are indeed, albeit moderately, faster in the B1 and B1F4 variants (Table 3). The
3
fact that these enzymes favor oxidation at C-2, however, complicates the comparison since the
measured rates describe different reactions.
The underlying mechanism responsible for the increase in turnover numbers is not known but the
phenomenon can be explained by the theory of conformational sampling and selection in catalysis.
Here, the enzyme structure undergoes dynamic changes between substrates that may exhibit distinct
properties regarding catalytic efficiencies and substrate selectivities.60,61 Introduced mutations (or
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
2
changes in environmental conditions ) can shift the equilibria between substrates to unveil new
functional properties.63,64 The increase in kcat and the shifts in regioselectivities in the oxidation of
(1R,2S)-3) displayed here by the isolated ADH-A variants are due to the structural changes caused by
the introduced substitutions but the distal location of these modifications in relation to the site of
hydride transfer makes it unlikely that they are directly causing the observed changes in
regioselectivities. An alternative suggestion could be that the equilibrium mixtures of the
conformational substrates of these enzymes have been shifted to instead favor oxidation of C-2 rather
than C-1, which is in itself a very subtle change. Thus, the increase in productively bound ternary
enzyme-substrate complexes may be a reflection of a shift in the distribution of conformational
substrates. The analysis of structural fluctuations during the MD simulation shows that the regions of
the tested enzymes that display the largest differences in fluctuations of the C-α carbons between
oxidation of C-1 or C-2 in (1R,2S)-3) include the mutated residues 43, 54 and 119 (Figure 10). We
note also that the issue of non-productive binding conformations in enzyme catalysis has been taken up
by several other workers in different contexts, in addition to the references mentioned above (for some
examples see e.g. refs. 65-68).
4
. CONCLUSIONS
Nonproductive binding of substrate in the ground state has not been widely presented as an intrinsic
feature of enzyme catalysis and substrate discrimination although it may be a general phenomenon.69
Increased understanding of its role in enzyme selectivity will be important for efficient engineering of
biocatalysts and for the fundamental understanding of enzyme function. In this study, relatively minor
changes in the topology of the ADH-A active site, primarily caused by the altered position of a tyrosine
side chain, affects the substrate selectivity: the regiopreference towards a vicinal diol containing two
1
5
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
sec-alcohol functionalities is shifted and the enantiodiscrimination between (R) and (S)-1-
phenylethane-1,2-diol is affected.
Competing interests
The authors declare no competing interests.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Funding information
This work was supported by Stiftelsen Olle Engkvist Byggmästare (MW). The support from COST
action 1303 Systems Biocatalysis is also gratefully acknowledged, as well as the Swedish Research
Council (2015-04928) and the Knut and Alice Wallenberg Foundation (KAW 2013.0124, Wallenberg
Academy Fellowship to SCLK). The authors also thank Diamond Light Source for beamtime (proposal
nr. Mx11171 and mx15868), and the staff of beamlines I04 and I24 for assistance with crystal testing
and data collection. Access was supported in part by the EU FP7 infrastructure grant BIOSTRUCTX
(contract no. 283570). All computational work was performed on the Abisko and Kebnekaise clusters
at HPC2N, through the generous allocation of computational time by the Swedish National
Infrastructure for Computing (SNIC, 2015/16-12 and 2016/34-27).
Author contributions
†
DM and TRE contributed equally to this work.
TRE, EH, HH and MW conducted enzyme kinetic analyses. DM and HH conducted the protein
crystallography. DM and DD solved the protein crystal structures. PB, ML, DP and SCLK conducted
and analyzed the computational simulations. All the authors discussed the results and contributed to
writing of the manuscript.
Supporting Information
(1) Description of strategy and oligonucleotide sequences used in the construction of library ‘B1F’.
(2) Description of the parametrization of the catalytic zinc ion for the molecular dynamics simulations.
(3) Supporting figures: Figure S1, cloning strategy; Figure S2, Extended NMR information and
spectra of compound mixtures following ADH-A-catalyzed oxidation of (1R,2S)-3 and reduction of 8.
Figure S3, Highest populated clusters after MD simulations with either (R)-2 or (S)-2 bound at the
1
6
ACS Paragon Plus Environment

Page 17 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
active site of variant B1. Figure S4, Comparison of signal amplitudes after rapid mixing of enzyme and
substrate during the pre-steady state phase of catalyzed reactions.
(4) All non-standard force field parameters used in the simulations.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Abbreviations ADH, alcohol dehydrogenase; ISM, iterative saturation mutagenesis; IPTG, β-D-
thiogalactopyranoside; HPLC, high-performance liquid chromatography; NMR, nuclear magnetic
resonance; RMSD, root mean square deviation; MD, molecular dynamics.
Acknowledgements
We are grateful to Andreas Söderlind for his contribution to the work.
References
(
(
(
1) Davis, F. A.; Chen, B.-C. Asymmetric Hydroxylation of Enolates with N-Sulfonyloxaziridines.
Chem. Rev. 1992 92, 919-934.
,
2) Wessjohann, L. Epothilones: Promising Natural Products with Taxol-Like Activity. Angew.
Chem. Int. Ed. 1997, 36, 715-718.
3) Agudo, R.; Roiban, G.-D.; Lonsdale, R.; Ilie, A.; Reetz, M. T. Biocatalytic Route to Chiral
Acyloins: P450-Catalyzed Regio- and Enantioselective α-Hydroxylation of Ketones. J. Org.
Chem. 2015, 80, 950-956.
(4) Bortolino, O.; Casanova, E.; Fantin, G.; Medici, A.; Poli, S.; Hanau, S. Kinetic Resolution of vic-
Diols by Bacillus stearothermophilusdiacetyl Reductase. Tetrahedron Asym. 1998, 9, 647-651.
(
5) Zhang, J.; Xu, T; Li, Z. Enantioselective Biooxidation of Racemic trans-Cyclic Vicinal Diols:
One-Pot Synthesis of Both Enantiopure (S,S)-Cyclic Vicinal Diols and (R)-α-Hydroxy Ketones.
Adv. Synth. Catal. 2013, 355, 3147-3153.
(
6) Souza Brenelli, E. C.; Moran, P. J. S.; Rodrigues, J. A. R. Preparation of (-)-(1R,2S)-1-
Phenylpropane-1,2-Diol by Fermenting Baker's Yeast Reduction of 1-Phenyl-1,2-propanedione.
Synth. Commun. 1990, 20, 261-266.
1
7
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(7) Takeshita, M.; Sato, T. Synthesis of Optically-Active 1-Phenyl-1,2-Propanediol by Use of
Baker’s-Yeast. Chem. Pharm. Bull. 1989, 37, 1085-1086.
(
8) Martins, R. S.; Zampieri, D. S.; Rodrigues, J. A. R.; Carvalho, P. O.; Moran, P. J. S.
Stereoselective Oxidation of 1-Phenyl-1,2-Propanediol Mediated by Microorganisms.
ChemCatChem 2011, 3, 1469-1473.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
9) Kulig, J.; Simon, R. C.; Rose, C. A.: Masood Husain, S.: Häckh, M.: Lüdeke, S.: Zeitler, K.;
Kroutil, W.; Pohl, M.; Rother, D. Stereoselective Synthesis of Bulky 1,2-Diols With Alcohol
Dehydrogenases. Catal. Sci. Technol. 2012, 2, 1580-1589.
(10) Chen, X.; Mei, T.; Cui, Y.; Chen, Q.; Liu, X.; Feng, J.; Wu, Q.; Zhu, D. Highly Efficient
Synthesis of Optically Pure (S)-1-Phenyl-1,2-Ethanediol by a Self-Sufficient Whole Cell
Biocatalyst. Chem. Open 2015, 4, 483-488.
(11) Sun, Z.; Lonsdale, R.; Ilie, A.; Li, G.; Zhou, J.; Reetz, M. T. Catalytic Asymmetric Reduction of
Difficult-to-Reduce Ketones: Triple-Code Saturation Mutagenesis of an Alcohol Dehydrogenase.
ACS Catal. 2016, 6, 1598-1605.
(
12) Hamnevik, E.; Maurer, D.; Enugala, T. R.; Chu, T.; Löfgren, R.; Dobritzsch, D.; Widersten, M.
Directed Evolution of Alcohol Dehydrogenase for Improved Stereoselective Redox
Transformations of 1-Phenylethane-1,2-Diol and Its Corresponding Acyloin. Biochemistry 2018,
5
7, 1059-1062.
(
13) Stampfer, W.; Kosjek, B.; Moitzi, C.; Kroutil, W.; Faber, K. Biocatalytic Asymmetric Hydrogen
Transfer. Angew. Chem. Int. Ed. 2002, 41, 1014-1017.
(
14) Stampfer, W.; Kosjek, B.; Faber, K.; Kroutil, W. Biocatalytic Asymmetric Hydrogen Transfer
Employing Rhodococcus ruber DSM 44541. J. Org. Chem. 2003, 68, 402-406.
(15) Kosjek, B.; Stampfer, W.; Pogorevc, M.; Goessler, W.; Faber, K.; Kroutil, W. Purification and
Characterization of a Chemotolerant Alcohol Dehydrogenase Applicable to Coupled Redox
Reactions. Biotechnol. Bioengineer. 2004, 86, 55-62.
(16) Hamnevik, E.; Blikstad, C.; Norrehed, S.; Widersten, M. Kinetic Charecterization of
Rhodococcus ruber DSM 44541 Alcohol Dehydrogenase A. J. Mol. Catal. B. Enzym. 2014, 99,
6
8-78.
(17) Lindberg, D.; de la Fuente Revenga, M.; Widersten, M. Deep Eutectic Solvents (DESs) Are
Viable CoSolvents for Enzyme-Catalyzed Epoxide Hydrolysis. J. Biotechnol. 2010, 147, 169-
1
71.
1
8
ACS Paragon Plus Environment

Page 19 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(18) Reetz, M. T.; Bocola, M.; Carballeira, J. D.; Zha, D.; Vogel, A. Expanding the Range of Substrate
Acceptance of Enzymes: Combinatorial Active-Site Saturation Test. Angew. Chemie Int. Ed.
2
005, 44, 4192-4196.
(
(
(
19) Reetz, M. T.; Carballeira, J. D. Iterative Saturation Mutagenesis (ISM) for Rapid Directed
Evolution of Functional Enzymes. Nat. Prot. 2007, 2, 891-903.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
20) Reetz, M. T.; Kahakeaw, D.; Sanchis, Shedding Light on the Efficacy of Laboratory Evolution
Based on Iterative Saturation Mutagenesis. J. Mol. BioSyst. 2009, 5, 115-122.
21) Karabec, M.; Łyskowski, A.; Tauber, K. C.; Steinkellner, G.; Kroutil, W.; Grogan, G.; Gruber, K.
Structural Insights into Substrate Specificity and Solvent Tolerance in Alcohol Dehydrogenase ADH-
‘A’ from Rhodococcus ruber DSM 44541. Chem. Commun. 2010, 46, 6314-6316.
(
22) Hamnevik, E.; Enugala, T. R.; Maurer, D.; Ntuku, S.; Oliveira, A.; Dobritzsch, D.; Widersten, M.
Relaxation of Nonproductive Binding and Increased Rate of Coenzyme Release in an Alcohol
Dehydrogenase Increases Turnover with a Nonpreferred Alcohol Enantiomer. FEBS J. 2017, 284,
3
895-3914.
(
(
(
23) Tang, L.; Gao, H.; Zhu, X.; Wang, X.; Zhou, M.; Jiang, R. Construction of “Small-Intelligent”
Focused Mutagenesis Libraries Using Well-Designed Combinatorial Degenerate Primers.
Biotechniques 2012, 52, 149-158.
24) Dale, G. E.; Schönfeld, H.-J.; Langen, H.; Stieger, M. Increased Solubility of Trimethoprim-
Resistant Type SI DHFR from Staphylococcus aureus in Escherichia coli Cells Overproducing
the Chaperonins GroEL and GroES. Protein Eng. 1994, 7, 925-931.
25) Elfström, L. T.; Widersten, M. The Saccharomyces cerevisiae ORF YNR064c Protein Has
Characteristics of an ‘Orphaned’ Epoxide Hydrolase. Biochim. Biophys. Acta 2005, 1748, 213-
221.
(
26) Hill, A. V. The Possible Effects of the Aggregation of the Molecules of Haemoglobin on its
Dissociation Curves. J. Physiol. 1910, 40, iv–vii.
(
27) Winter, G.; McAuley, K. E. Automated Data Collection for Macromolecular Crystallography.
Methods 2011, 55, 81-93.
(
28) Evans, P. Scaling and Assessment of Data Quality. Acta Cryst. 2006, D62, 72-82.
29) Collaborative Computational Project, Number 4. The CCP4 Suite: Programs for Protein
Crystallography. Acta Cryst. 1994, D50, 760-763.
(
1
9
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(30) Winn, M. D.; Ballard, C. C.; Cowtan, K. D.; Dodson, E. J.; Emsley, P.; Evans, P. R.; Keegan, R.
M.; Krissinel, E. B.; Leslie, A. G.; McCoy, A.; McNicholas, S. J.; Murshudov, G. N.; Pannu, N.
S.; Potterton, E. A.; Powell, H. R.,;Read, R. J.; Vagin, A.; Wilson, K. S. Overview of the CCP4
Suite and Current Developments. Acta Cryst. 2011, D67, 235-242.
(31) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.; Read, R. J.
Phaser Crystallographic Software. J. Appl. Cryst. 2007, 40, 658-674.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(32) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features and Development of Coot. Acta
Cryst. 2010, D66, 486-501.
(33) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of Macromolecular Structures by the
Maximum-Likelihood Method. Acta Cryst. 1997, D53, 240-255.
(
34) Case, D. A.; Betz, R. M.; Botello-Smith, W.; Cerutti, D. S.; Cheatham, III, T. E.; Darden, T. A.;
Duke, R. E.; Giese, T. J.; Gohlke, H.; Goetz, A. W.; Homeyer, N.; Izadi, S.; Janowski, P.; Kaus,
J.; Kovalenko, A.; Lee, T. S.; LeGrand, S.; Li, P.; Lin, C.; Luchko, T.; Luo, R.; Madej, B.;
Mermelstein, D.; Merz, K. M.; Monard, G.; Nguyen, H.; Nguyen, H. T.; Omelyan, I.; Onufriev,
A.; Roe, D. R.; Roitberg, A.; Sagui, C.; Simmerling, C. L.; Swails, J.; Walker, R. C.; Wang, J.;
Wolf, R. M.; Wu, X.; Xiao, L.; York, D. M.; Kollman, P. A. 2016, AMBER 2016 Reference
Manual (Covers Amber16 and AmberTools16), University of California, San Francisco.
35) Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L; Hauser, K. E.; Simmerling, C. ff14SB:
Improving the Accuracy of Protein Side Chain and Backbone Parameters from ff99SB. J. Chem.
Theory Comput. 2015, 11, 3696-3713.
(
(
36) Wang, J.; Wang, W.; Kollman, P. A.; Case, D. A. Automatic Atom Type and Bond Type
Perception in Molecular Mechanical Calculations. J. Mol. Graph. Modell. 2006, 25, 247-260.
(37) Cieplak, P.; Cornell, W. D.; Bayly, C.; Kollman, P. A. Application of the Multimolecule and
Multiconformational RESP Methodology to Biopolymers: Charge Derivation for DNA, RNA,
and Proteins. J. Comp. Chem. 1995, 16, 1357-1377.
(38) Pavelites, J. J.; Bash, P. A.; Mackerell, Jr., A. D. A Molecular Mechanics Force Field for
+
NAD NADH, and the Pyrophosphate Groups of Nucleotides. J. Comput. Chem. 1997, 18, 221-
2
39.
(
39) Walker, R.C.; de Souza, M. M.; Mercer, I. P.; Gould, I. R.; Klug, D. R. Large and Fast
Relaxations Inside a Protein:ꢀCalculation and Measurement of Reorganization Energies in
Alcohol Dehydrogenase. J. Phys. Chem. B 2002, 106, 11658-11665.
2
0
ACS Paragon Plus Environment

Page 21 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(40) Barrozo, A.; Duarte, F.; Bauer, P.; Carvalho, A. T. P.; Kamerlin, S. C. L. Cooperative
Electrostatic Interactions Drive Functional Evolution in the Alkaline Phosphatase Superfamily. J.
Am. Chem. Soc. 2015, 137, 9061-9076.
(
41) Purg, M.; Pabis, A.; Baier, F.; Tokuriki, N.; Jackson, C.; Kamerlin, S. C. L. Probing the
Mechanisms for the Selectivity and Promiscuity of Methyl Parathion Hydrolase. Phil. Trans. R.
Soc. A 2016, 374, 20160150.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
42) Wallin, C.; Kulkarni, Y. S.; Abelein, A.; Jarvet, J.; Liao, Q.; Strodel, B.; Olsson, L.; Luo, J.;
Abrahams, J. P.; Sholts, S. B.; Roos, P. M.; Kamerlin, S. C. L.; Gräslund, A.; Wärmländer, S. K.
T. S. Characterization of Mn(II) Ion Binding to the Amyloid- Peptide in Alzheimer’s Disease. J.
Trace Elem. Med. Biol. 2016, 38, 183-193.
(
43) Blaha-Nelson, D.; Krüger, D. M.; Szeler, K.; Ben-David, M.; Kamerlin, S. C. L. Active Site
Hydrophobicity and the Convergent Evolution of Paraoxonase Activity in Structurally Divergent
Enzymes: The Case of Serum Paraoxonase 1. J. Am. Chem. Soc. 2017, 139, 1155-1167.
44) Duarte, F.; Bauer, P.; Barrozo, A.; Amrein, B. A.; Purg, M.; Åqvist, J.; Kamerlin, S. C. L. Force
Field Independent Metal Parameters Using a Nonbonded Dummy Model. J. Phys. Chem. B 2014,
118, 4351-4362.
(
(45) Liao, Q. H.; Kamerlin, S. C. L.; Strodel, B. Development and Application of a Nonbonded
Cu2+ Model that Includes the Jahn–Teller Effect. J. Phys. Chem. Lett. 2015, 6, 2657-2662.
(46) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. Comparison of
Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926-935.
(47) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. Numerical Integration of the Cartesian Equations
of Motion of a System with Constraints: Molecular Dynamics of n-Alkanes. J. Comput. Phys.
1977, 23, 327-341.
(
(
(
48) Roe, D. R.; Cheatham, III, T. E. PTRAJ and CPPTRAJ: Software for Processing and Analysis of
Molecular Dynamics Trajectory Data. J. Chem. Theory Comput. 2013, 9, 3084–3095.
49) Monod, J.; Wyman, J.; Changeux, J.-P. On the Nature of Allosteric Transitions: A Plausible
Model. J. Mol. Biol. 1965, 12, 88-118.
50) Bruice, T. C.; Pandit, U. K. Intramolecular Models Depicting the Kinetic Importance of “Fit” in
Enzymatic Catalysis. Proc. Natl. Acad. Sci. U.S.A. 1960, 46, 402-404.
(51) Lightstone, F. C.; Bruice, T.C. Ground State Conformations and Entropic and Enthalpic Factors
in the Efficiency of Intramolecular and Enzymatic Reactions. 1. Cyclic Anhydride Formation by
2
1
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Substituted Glutarates, Succinate, and 3,6-Endoxo-∆4-tetrahydrophthalate Monophenyl Esters. J.
Am. Chem. Soc. 1996, 118, 2595-2605.
(
52) Torres, R. A.; Bruice, T. C. Molecular Dynamics Study Displays Near In-Line Attack
Conformations in the Hammerhead Ribozyme Self-Cleavage reaction. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 11077-11082.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
(
(
53) Rajagapolan, P. T. R.; Benkovic, S. J. Preorganization and Protein Dynamics in Enzyme
Catalysis. Chem. Record 2002, 2, 24-36.
54) Hur, S.; Bruice, T. C. The Near Attack Conformation Approach to the Study of the Chorismate to
Prephenate Reaction. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 12015-12020.
55) Giraldo, J.; Roche, D.; Rovira, X.; Serra, J. The Catalytic Power of Enzymes: Conformational
Selection or Transition State Stabilization? FEBS Lett. 2006, 580, 2170-2177.
(56) Griffin, J. L.; Bowler, M. W.; Baxter, N. J.; Leigh, K. N.; Dannatt, H. R. W.; Hounslow, A. M.;
Blackburn, G. M.; Webster, C. E.; Cliff, M. J.; Waltho, J. P. Near Attack Conformers Dominate
β-Phosphoglucomutase Complexes Where Geometry and Charge Distribution Reflect Those of
Substrate. Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 6910-6915.
(57) Burchowsky, D.; van Erde, A.; Ökvist, M.; Kienhöfer, A.; Kast, P.; Hilvert, D.; Krengel, U.
Electrostatic Transition State Stabilization Rather Than Reactant Destabilization Provides the
Chemical Basis for Efficient Chorismate Mutase Catalysis. Proc. Natl. Acad. Sci. U.S.A. 2014,
1
11, 17516-17521.
(
(
(
(
58) Sadiq, S. K.; Coveney, P. V. Computing the Role of Near Attack Conformations in an Enzyme-
Catalyzed Nucleophilic Bimolecular Reaction. J. Chem. Theory Comput. 2015, 11, 316-324.
59) Warshel, A.; Bora, R. P. Defining and Quantifying the Role of Dynamics in Enzyme Catalysis. J.
Chem. Phys. 2016, 144, 180901.
60) Ma, B.; Nussinov, R. Enzyme Dynamics Point to Stepwise Conformational Selection in catalysis.
Curr. Opin. Chem. Biol. 2010, 14, 652-659.
61) Ramanathan, A.; Savol, A.; Burger, V.; Chennubhotla, C. S.; Agarwal, P. K. Protein
Conformational Populations and Functionally Relevant Substates. Acc. Chem. Res. 2014, 47, 149-
1
56.
(
62) Klinman, J. P.; Kohen, A. Evolutionary Aspects of Enzyme Dynamics. J. Biol. Chem. 2014, 289,
30205-30212.
2
2
ACS Paragon Plus Environment

Page 23 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(63) James, L. C.; Tawfik, D. S. Conformational Diversity and Protein Evolution – a 60-Year-Old
Hypothesis Revisited. Trends Biochem. Sci. 2003, 28, 361-368.
(
64) Tokuriki, N.; Tawfik, D. S. Protein Dynamism and Evolvability. Science 2009, 324, 203-207.
65) Glick, B. R.; Brubacher, L. J. Evidence for Nonproductive Binding Subsites Within the Active
Site of Papain. Can. J. Biochem. 1974, 52, 877-883.
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
(
(
66) Natarajan, S. K.; Sierks, M. R. Minimizing Nonproductive Substrate Binding: A New Look at
Glucoamylase Subsite Affinities. Biochemistry 1997, 36, 14946-14955.
67) Stojan, J. Non-Productive Binding of Butyryl(thio)choline in the Active Site of Vertebrate
Acetylcholinesterase. Chem. Biol. Interact. 2010, 187, 128-134.
68) Gao, D.; Chundawat, S. P. S.; Sethi, A.; Balan, V.; Gnanakran, S.; Dale, B. E. Increased Enzyme
Binding to Substrate is Not Necessary for More Efficient Cellulose Hydrolysis. Proc. Natl. Acad.
Sci. U.S.A. 2013, 110, 10922-10927.
(
69) Bar-Even, A.; Milo, R.; Noor, E.; Tawfik, D. S. The Moderately Efficient Enzyme: Futile
Encounters and Enzyme Floppiness. Biochemistry 2015, 54, 4969-4977.
(70) The PyMOL Molecular Graphics System, Version 1.8 Schrödinger,
LLC.
2
3
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
TABLES
Table 1. Steady state kinetic parameters
cat (s-1)
-1
-1
a
cat (s-1)
-1
-1
k
K
0.5 (mM)
k
cat/K0.5 (s M )
n
k
K
0.5 (mM)
kcat/K0.5 (s M ) n
substrate wild type
B1 (Y54G, L119Y)
b
b
b
(
(
(
(
S)-1
R)-2
S)-2
80±20
0.63±0.05 130 000±30 000
(1)
11±1
32±6
340±30
(1)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
b
b
b
0.73±0.01
17±0.6
44±8
0.17±0.008
0.039±0.001
0.98±0.03
7.3±0.6
130±10
17±2
1.3±0.06
2.3±0.2
(1)
b
b
b
0.0094±0.0005
3.0±0.7
11±1
3.1±0.6
43±5
(1)
1R,2S)-3 0.48±0.01
0.75±0.03
(1)
33±2
30±2
1.2±0.06
(1)
b
b
b
4
5
8
36±0.8
2.0±0.08
62±0.6
1.2±0.09
30 000±2 000
4.8±1
1 500±300
36±1
b
b
b
3.7±0.4
550±30
0.54±0.02
85±0.5
15±0.9
0.32±0.009
(1)
0.31±0.01
200 000±9 000
1.3±0.06
260 000±8 000
0.93±0.02
c
B1F1 (F43H, Y54G, L119Y)
B1F4 (F43T, Y54G, L119Y, F282W* )
(
(
(
(
S)-1
22±0.8
47±3
460±40
1.0±0.05
1.3±0.3
19±0.6
1.1±0.02
(1)
37±3
73±10
140±4
60±20
35±1
500±100
3.2±0.004
1.2±0.4
0.77±0.02
(1)
R)-2
S)-2
0.96±0.09
0.022±0.002
940±100
28±5
0.45±0.006
0.072±0.008
2.3±0.04
26±0.6
1.3±0.2
1.3±0.03
1.5±0.09
1.1±0.03
1.2±0.07
0.79±0.06
1.4±0.05
1.2±0.05
(1)
1R,2S)-3 0.76±0.01
41±1
66±3
4
5
8
5.7±0.03
0.62±0.03
84±2
5.9±0.1
39±3
980±20
16±2
8.5±0.5
25±1
3 000±200
38±1
0.98±0.2
160±2
1.8±0.1
48 000±3 000
0.51±0.02
320 000±2 000
(1)
B1F5 (F43S, Y54G, L119Y, V194I, F282P*)
(1R,2S)-3 0.41±0.009
49±0.7
an, Hill coefficient. K0.5 = K
48±2
8.4±0.4
1.4±0.05
(1)
8
1.8±0.09
28 000±1 000
b
c
M
if n=1. Data from ref. 16. *, Residue from the neighboring subunit.
ACS Paragon Plus Environment

Page 25 of 43
1
ACS Catalysis
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. Relative amounts of formed products following oxidation of (1R,2S)-3 or reduction of 8.
1H NMR integrals (%)a
oxidation of (1R,2S)-3
reduction of 8
wild type
B1
B1F4
wild type
B1
B1F4
71
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
2
2
2
70
32
12
55
-
78
21
1
82
(1R,2S)-3
8
6
7
3
7
24
-
17
1
27
2
(S)-6
6
(R)-7
-b
-
-
-
8
a
b
Calculated from the relative abundance of the distinguishing methyl protons. -, not detected. NMR
spectra are shown in Figure S2.
2
5
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 3. Kinetic rates and equilibrium dissociation constants
-
1
-1
-1 (s-1)
a
b
-1
b
-1
-1
-1
c
Substrate
k
1
(s µM )
k
K
1
(µM)
k
3
(s )
K
2
(mM)
k
5
(s )
k
-5 (s µM )
5
K (µM)
wild type
6.8±0.5
NAD+
NADH
(S)-1
110±10
17±2
51±6d
5.1±0.1d
10±0.1d
630±40e
42±8
2.5±0.5e
87±30
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
R)-2
(1R,2S)-3
20±9
62±40
B1 (Y54G, L119Y)
9.7±0.4 190±20
NAD+
20±2
NADH
64±20
56±30
12±0.5
9.6±1
5.3±2
(
1R,2S)-3
77±20
91±30
f
B1F4 (F43T, Y54G, L119Y, F282W* )
NAD+
6.9±0.3
110±9
16±1
NADH
5.9±3
(
1R,2S)-3
68±20
41±20
a
. b From fitting Eq. 4 to the observed transient rates. Calculated
Data from ref. 22. Data from ref. 16. *, Residue from the neighboring
c
Calculated from the ratios of k-1/k
1
d
e
f
from the ratios of k
5
/k-5
.
subunit.
2
6
ACS Paragon Plus Environment

Page 27 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figures
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. (A) Alcohols and ketones 1-5 and 8 were tested as substrates. Acyloins 6 and 7 are the
oxidation products from (1R,2S)-3 following hydride transfer from either carbon 1 or 2, respectively.
(B) (1R,2S)-3 is produced by epoxide hydrolase catalyzed hydrolysis of the corresponding (1S,2S)-
epoxide.
Figure 2. Model of the kinetic mechanism for alcohol oxidation by ADH-A.
2
7
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 43
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Structures of MD snapshots of most populated clusters obtained from the calculations with
wild type ADH-A and (R)-2 (A), (S)-2 (B) and (1R,2S)-3 (C, D). (R)-2 binds in a configuration that can
facilitate hydride transfer from C-1 of the diol to the nicotinamide C-4 (A) whereas with (S)-2, the
corresponding hydrogen points away from the nicotinamide ring (B, hydrogen highlighted by a red
dotted circle). (1R,2S)-3 is bound in an orientation for hydride transfer from the benzylic C-1 carbon
with a distance of approximately 2.7 Å in the main cluster (C). The corresponding distance for transfer
from C-2 is similar (D) but with slightly increased RMSD values for the substrate. See the Materials
and Methods section for details on the methodology and the Supporting Information for details on
2
+
parametrization of the active-site tetrahedral Zn .
2
8
ACS Paragon Plus Environment

Page 29 of 43
ACS Catalysis
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Residues exposed to saturation mutagenesis. The surface shows the volume of the active site
cavities. (A) In the wild type enzyme (pink), Y294 and W295 make up site ‘A’ and Y54 and L119, site
‘B’. (B) In the B1 variant (Y54G, L119Y), F43 and F282* (from the neighboring subunit) make up site
7
0
22
‘
F’. Image created with PyMOL ver. 1.8.7 with the atomic coordinates in 3jv7 (wild type) and 5od3
(B1).
Figure 5. Outcome of oxidation of (1R, 2S)-3 by wild type ADH-A or the B1 or B1F4 variants. The
introduced substitution changes the regiopreference from C-1 to C-2.
2
9
ACS Paragon Plus Environment