Study of the Phosphoryl-Transfer Mechanism of Shikimate Kinase by NMR Spectroscopy

Article abstract of DOI:10.1002/chem.201504438

The phosphoryl-transfer mechanism of shikimate kinase from Mycobacterium tuberculosis and Helicobacter pylori, which is an attractive target for antibiotic drug discovery, has been studied by 1D ¹H and ³¹P NMR spectroscopy. Metaphosphoric acid proved to be a good mimetic of the metaphosphate intermediate and facilitated the ready and rapid evaluation by NMR spectroscopic analysis of a dissociative mechanism. The required closed form of the active site for catalysis was achieved by the use of ADP (product) or two synthetic ADP analogues (AMPNP, AMPCP). Molecular dynamics simulation studies reported here also revealed that the essential arginine (Arg116/Arg117 in H. pylori and M. tuberculosis, respectively), which activates the γ-phosphate group of ATP for catalysis and triggers the release of the product for turnover, would also be involved in the stabilisation of the metaphosphate intermediate during catalysis. We believe that the studies reported here will be helpful for future structure-based design of inhibitors of this attractive target. The approach is also expected be useful for studies on the possible dissociative mechanism of other kinase enzymes.

Full text of DOI:10.1002/chem.201504438

DOI: 10.1002/chem.201504438
Full Paper
&
Enzyme Catalysis
Study of the Phosphoryl-Transfer Mechanism of Shikimate Kinase
by NMR Spectroscopy
[
a]
[a]
[b]
[b]
[c]
Verónica Prado, Emilio Lence, Juan A. Vallejo, Alejandro Beceiro, Paul Thompson,
[c]
[a]
Alastair R. Hawkins, and Concepción Gonzµlez-Bello*
Abstract: The phosphoryl-transfer mechanism of shikimate
kinase from Mycobacterium tuberculosis and Helicobacter
pylori, which is an attractive target for antibiotic drug discov-
that the essential arginine (Arg116/Arg117 in H. pylori and M.
tuberculosis, respectively), which activates the g-phosphate
group of ATP for catalysis and triggers the release of the
product for turnover, would also be involved in the stabilisa-
tion of the metaphosphate intermediate during catalysis. We
believe that the studies reported here will be helpful for
future structure-based design of inhibitors of this attractive
target. The approach is also expected be useful for studies
on the possible dissociative mechanism of other kinase en-
zymes.
1
31
ery, has been studied by 1D H and P NMR spectroscopy.
Metaphosphoric acid proved to be a good mimetic of the
metaphosphate intermediate and facilitated the ready and
rapid evaluation by NMR spectroscopic analysis of a dissocia-
tive mechanism. The required closed form of the active site
for catalysis was achieved by the use of ADP (product) or
two synthetic ADP analogues (AMPNP, AMPCP). Molecular
dynamics simulation studies reported here also revealed
Introduction
which proceeds through a single transition state with simulta-
neous cleavage and formation of two single bonds through
[
4,5]
Phosphoryl-transfer reactions are amongst the most funda-
mental processes in biological systems. These reactions are
crucial in the biosynthesis and replication of nucleic acids, the
synthesis and assembly of key metabolites in a large variety of
metabolic pathways and in energy and signal transduction,
a pentavalent transition state as in S 2 reactions (Scheme 1).
N
[
1–7]
amongst others.
This type of reaction can proceed by three
mechanistically different nucleophilic substitution pathways:
A) dissociative, in which a trigonal metaphosphate intermedi-
ate is formed first by cleavage of the OÀP bond in the donor,
followed by subsequent nucleophilic attack of the acceptor;
B) associative, in which the nucleophile adds to the donor first
with the formation of a pentavalent phosphorane intermedi-
ate, and then the leaving group is released; and (C) concerted,
[
a] V. Prado, Dr. E. Lence, Prof. C. Gonzµlez-Bello
Centro Singular de Investigación en Química Biolóxica e
Materiais Moleculares (CIQUS)
Scheme 1. Phosphoryl-transfer reaction mechanisms.
Universidade de Santiago de Compostela
calle Jenaro de la Fuente s/n,
1
5782 Santiago de Compostela, (Spain)
These pathways differ in the timing of the cleavage of the
phosphorus–leaving group bond and the formation of the
phosphorus–nucleophile bond. The concerted process is
a one-step process with no intermediate. In contrast, associa-
tive and dissociative mechanisms are two-step processes that
do involve an intermediate, although they are quite different
E-mail: concepcion.gonzalez.bello@usc.es
[
b] Dr. J. A. Vallejo, Dr. A. Beceiro
Servicio de Microbioloxía-INIBIC
Complexo Hospitalario Universitario A CoruÇa (CHUAC)
Hospital Teresa Herrera, As Xubias, 84, 15006 A CoruÇa (Spain)
[
c] P. Thompson, Prof. A. R. Hawkins
Institute of Cell and Molecular Biosciences, Medical School
University of Newcastle upon Tyne
Catherine Cookson Building, Framlington Place
Newcastle upon Tyne NE2 4HH (UK)
(
trigonal or trigonal bipyramid, respectively). Moreover, in the
dissociative process, the acceptor is not involved in the inter-
mediate. Examples of the three possible mechanisms have
been described, although the reasons for the selection of a par-
Supporting information and ORCIDs from the authors for this article are
available on the WWW under http://dx.doi.org/10.1002/chem.201504438.
[
8–13]
ticular pathway for a given kinase are still unclear.
Chem. Eur. J. 2016, 22, 2758 – 2768
2758
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Considering that many drugs that are highly successful in
human clinical use mimic a transition state of a recognised
target enzyme, detailed knowledge of the catalytic mechanism
of a chosen enzyme can be valuable for the rational design of
mimetics (inhibitors) that can be used as drugs. Examples of
drugs that are transition-state analogues include oseltamivir,
a neuraminidase inhibitor in the treatment of the influenza
virus, and saquinavir, an HIV protease inhibitor used in the
treatment of AIDS. Structural studies have contributed greatly
to this issue with the resolution of crystal structures of diverse
kinases in complex with transition-state analogues, reaction in-
termediates or metal fluoride complexes (e.g., AlF and
3
À
[8–11,14,15]
BeF ).
However, these techniques require large quanti-
3
ties of protein, they are time-consuming and, in many cases, it
is difficult to find crystallisation conditions to obtain crystals
that provide good diffraction data. Phosphoryl-transfer reac-
tions have also been studied by hybrid QM/MM calculations,
and this approach allows the study of bond-making and bond-
breaking in a dynamic context that allows the reorganisation
of active site residues, substrate, and water molecules during
[16–18]
the whole reaction coordinate.
However, due to the very
demanding computational cost, the application of QM is still
limited to a relatively small number of atoms, which is an issue
for large active sites—particularly when the highest levels of
theory are employed. We report here NMR studies that allowed
the ready and rapid determination of the phosphoryl-transfer
mechanism of a recognised attractive target for antibiotic drug
Figure 1. Enzymatic conversion of shikimic acid (1) into shikimate-3-phos-
phate (2) catalyzed by SK and its Michaelis complex. The three domains of
H. pylori SK enzyme are shown: CORE (purple), LID (red) and SB (yellow).
[19]
discovery, namely shikimate kinase.
Shikimate kinase (SK, EC 2.7.1.71, aroK gene), the fifth
enzyme of the shikimic acid pathway, catalyses the stereospe-
cific phosphorylation of the C3 hydroxyl group of shikimic
acid (1) by transferring the g-phosphate group of ATP to the
hydroxyl group to provide shikimate 3-phosphate (2) and ADP
the guanidinium group of the essential arginine (Arg116/
Arg117 in H. pylori and M. tuberculosis, respectively); 2) a
closed form of the SB domain to bring the substrates together
for the phosphoryl-transfer reaction and to isolate the sub-
strate within an apolar cavity, which is controlled mainly by
two phenylalanine residues; and 3) a flexible LID and SB
domain for product release, which is triggered by the essential
(
Figure 1). SK is essential in several relevant pathogenic bacte-
ria such as Mycobacterium tuberculosis, which is responsible for
tuberculosis, Helicobacter pylori, which is the causative agent
of gastric and duodenal ulcers and has also been classified as
a type I carcinogen, and Pseudomonas aeruginosa, which is one
of the most common pathogens in healthcare-associated infec-
tions, and represents 13% of multidrug-resistant strains to
nearly all antibiotics. SK has three domains (Figure 1): 1) the
CORE domain that contains five stranded parallel b-sheets and
the P-loop, which forms the binding site for ATP and ADP;
[
14]
arginine.
Given the important features of the SK enzyme, we recently
became interested in the development of new antimicrobials
that function by inhibition of this enzyme. To this end, the
knowledge of the catalytic mechanism of this enzyme will be
helpful for the mechanism- and structure-based design of in-
hibitors. We report here NMR studies that allowed the determi-
nation of the phosphoryl-transfer mechanism of the SK
enzyme. Molecular dynamics (MD) simulation studies with the
M. tuberculosis and H. pylori SK enzymes were also carried out
to gain a better understanding of the dynamic behaviour of
the enzyme Michaelis complex and the stability of the pro-
posed metaphosphate intermediate. We believe that the ap-
proach described here would also be useful for the evaluation
of possible dissociative phosphoryl-transfer mechanisms of
other kinase enzymes.
2
) the LID domain, which has an essential arginine for the
binding of ATP (Arg116/Arg117 in H. pylori and M. tuberculosis,
respectively); and 3) the substrate binding (SB) domain, which
is responsible for the recognition and binding of shikimic acid
(
1).
Structural and molecular dynamics (MD) simulation studies
revealed that SK is exquisitely designed to recognise the unfav-
ourable conformation of shikimic acid (1), which has a diaxial
[20–22]
disposition of the C4 and C5 hydroxyl groups.
In this ar-
rangement, the enzyme enables selective phosphorylation of
the C3 hydroxyl group in 1, which is arranged in an equatorial
disposition. MD simulation studies also revealed that there are
three key structural factors for catalysis: 1) a closed form of the
LID domain to allow activation of the g-phosphate of ATP by
Results and Discussion
Based on the results of structural studies, it was suggested
that Asp33/Asp34 (in H. pylori and M. tuberculosis, respectively)
would act as a general base to deprotonate the C3 hydroxyl
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2759
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
21]
group in 1 for nucleophilic attack. More importantly, it was
proposed that the phosphoryl-transfer reaction catalysed by
SK proceeds through an associative mechanism with conse-
quent generation of a pentacovalent phosphorane intermedi-
The H. pylori enzyme used in these studies was expressed
and produced using genomic DNA from an isolated clinical
strain, courtesy of Guadalajara Hospital (Guadalajara, Spain).
[
24]
Mt-SK was produced and purified as described previously.
ADP was used in these experiments on considering previously
[
21]
ate (Scheme 1B). As a consequence, both substrates would
be involved in the reaction intermediate. However, MD simula-
tion studies carried out with the SK Michaelis complex (SK/
ATP/1) from Mycobacterium tuberculosis (Mt-SK) and Helicobact-
er pylori (Hp-SK) showed that, in both cases, the distance be-
tween the g-phosphate of ATP and the oxygen atom of the C3
hydroxyl group of 1 is over 5 Š during most of the simulation
[
25]
[26]
reported circular dichroism,
fluorescence
and structur-
[
20,24]
al
studies, which suggested that the ATP binding induces
a large movement of the LID domain over the active site to
bring two key arginines into close contact with the ATP and
[
22]
that a closed active site is required for catalysis. Specifically,
this process involved the conserved residues Arg107/Arg110
and the essential residues Arg116/Arg117 in H. pylori and M. tu-
berculosis, respectively. In this arrangement, the appropriate
disposition of ATP in the active site, which is controlled by
a cation–p interaction with Arg107/Arg110, and the activation
of the g-phosphate of ATP by the guanidinium group of the
essential arginine would occur. As shown in Figure 3, compari-
(
Figure 2 and S1). It is important to highlight that single OÀP
bonds are typically approximately 1.6 Š.
Figure 2. Relative distance between substrates in Michaelis complexes. Varia-
tion of the distance between the phosphorus atom of ATP g-phosphate and
the C3 oxygen atom of 1 during 32 ns of MD simulation of the M. tuberculo-
sis (green) and H. pylori (red) SK/ATP/1 Michaelis complexes.
These Michaelis complexes were prepared by using the
Figure 3. Superposition of the crystal structures of Mt-SK/ADP/1 (PDB
enzyme geometries found in the crystal structures of Mt-SK
[20]
[20]
2IYQ ) and Mt-SK/1 (2IYS ) complexes and Mt-SK Michaelis complex. Note
[
21]
[20]
(
PDB code 2IYQ ) and Hp-SK (PDB code 3MUF ) in complex
how in the absence of ADP or ATP the LID is in the open conformation and
the essential Arg117 is located over 20 Š away from the active site.
with ADP and shikimic acid (1) or shikimate-3-phosphate (2),
respectively, after manual replacement of the corresponding
enzyme products by their substrates. The monomer of each SK
in aqueous solution obtained by using the molecular mechan-
[
21]
son of the crystal structures of Mt-SK/ADP/1 (PDB 2IYQ ) and
[
21]
Mt-SK/1 (2IYS ) complexes shows that, in the absence of ADP,
the LID is in the open conformation and the essential Arg117
is located over 20 Š away from the active site. The Mt-SK/ADP/
1 crystal structure (PDB 2IYQ) also revealed that a closed con-
formation of the active site is achieved in the presence of ADP.
More importantly, superposition of the Mt-SK Michaelis com-
plex and the Mt-SK/ADP/1 crystal structure highlights how the
essential Arg117 is pointing towards the active site. Therefore,
we consider that by using ADP (product) in the NMR experi-
ments a similar conformation of the active site for catalysis
would also be generated.
[23]
ics force field AMBER was employed. These MD simulation
studies revealed that both substrates seem to be quite far
away for an associative or concerted mechanism, in which
both donor and acceptor would be linked through a phospho-
rus atom, and the dissociative mechanism appears to be plau-
sible. If the conversion of shikimic acid (1) into shikimate 3-
phosphate (2) would take place by a dissociative mechanism,
a metaphosphate intermediate that would be recognised by
the enzyme would be generated and no covalent linkage be-
tween ATP and 1 would be required. In an effort to evaluate
this hypothesis, we explored the possible conversion of 1 into
1
31
1D H and P NMR monitoring of the formation of shikimate
3-phosphate (2) from shikimic acid (1) catalysed by Hp-SK in
the presence of ADP after the addition of metaphosphoric
2
catalysed by Hp-SK and Mt-SK in the absence of ATP and by
external addition of metaphosphoric acid as a mimetic of the
metaphosphate intermediate. The possible formation of 2 was
followed by 1D H and P NMR spectroscopy.
acid, a polymeric form of HOPO , is shown in Figure 4. The nat-
2
1
31
ural enzymatic conversion of shikimic acid (1) and ATP cata-
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2760
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
lysed by Mt-SK to give shikimate 3-phosphate (2) and ADP was
also studied by NMR spectroscopy (Figure S2). The latter ex-
periments were carried out by using very small amounts of Mt-
to H2, H5 and H4 of 2, respectively; b) one singlet (d =
P
31
0.6 ppm) in the P NMR spectra corresponding to the phos-
phate group of 2. Similar results were obtained on using the
Mt-SK enzyme (Figure S5). We assume that the singlet ob-
SK, which also has a high K , to follow easily the enzymatic
m
31
conversion. As shown in Figure S2, conversion of ATP into ADP
was not observed in the absence of 1 but it was a very fast
process after the addition of 1. Control experiments showed
that the reaction did not occur in the absence of either ADP
served in the P NMR spectra at d =À0.1 ppm (Figure 4B)
P
after the addition of metaphosphoric acid corresponds to
some hydrolysis of this acid (phosphoric acid) and it is also ob-
served in the control experiments carried out in the absence
of enzyme or ADP (Figure S4 and S3, respectively). However,
control experiments on the stability of metaphosphoric acid
showed that this compound is quite stable under the assay
conditions because significant changes were not observed
(
Figure S3) or enzyme (Figure S4). In addition, the use of other
commercially available mimetics of the metaphosphate inter-
mediate such as barium or aluminium metaphosphate failed
because of their low solubility under the reaction conditions.
1
31
31
1
D H and P NMR spectra of a mixture of 1 and ADP in the
during P NMR monitoring and it can therefore be used for
presence of a catalytic amount of the Hp-SK enzyme were re-
corded before (Figure 4A) and after the addition of metaphos-
phoric acid (Figure 4B–E). After 10 min incubation (Figure 4B),
the formation of 2 could already be detected and it was clearly
observed after 6 h (Figure 4C). The most relevant signals in the
the reported NMR studies (Figure S6).
Comparison of the experiments carried out in the presence
of ADP and with the external addition of metaphosphoric acid
1
31
(Figure 4) with H and P NMR monitoring of the natural for-
mation of shikimate 3-phosphate (2) from shikimic acid (1) and
ATP catalysed by Mt-SK (Figure S2) revealed that the same ste-
reospecific phosphorylation of the C3 hydroxyl group of
1 takes place in the reported experiments as in the natural en-
1
31
1
D H and P NMR spectra that reveal the formation of 2 are:
a) one singlet (d =6.39 ppm) and two multiplets (d =
H
H
1
3
.98 ppm and 3.76 ppm) in the H NMR spectra corresponding
1
31
Figure 4. 1D H and P NMR monitoring of the enzymatic conversion of shikimic acid (1) into shikimate 3-phosphate (2) catalysed by Hp-SK in the presence
1
31
of ADP and metaphosphoric acid. H and P NMR spectra: A) of 1, ADP and Hp-SK under the assay conditions. B) 10 min after the addition of metaphosphoric
1
31
acid (18 mm) to (A). C) after 6 h incubation. D) after 24 h incubation. E) after 96 h incubation. All experiments were recorded at 500 MHz ( H) or 202 MHz ( P)
and 298 K using ADP (4.5 mm), 1 (6.4 mm) and a catalytic amount of Hp-SK (10 units) in Tris-d11·DCl (100 mm, pD 7.7), KCl (100 mm) and MgCl (5 mm). Only
relevant regions of the H NMR spectra are shown for clarity. The most relevant signals that reveal the formation of 2 are marked with an arrow (C). The ob-
2
1
31
served P signals in the ADP solution, which corresponds to phosphate (Pi), b-phosphate of ADP (Pb) and a-phosphate of ADP (Pa), are labelled.
Chem. Eur. J. 2016, 22, 2758 – 2768 www.chemeurj.org ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2761

Full Paper
zymatic conversion. The findings that: 1) no reaction is ob-
served in the absence of enzyme, 2) a closed conformation of
the active site is required for reaction; that is, with the essen-
tial arginine in close contact with the substrates/products,
which is achieved by using ADP, and 3) nonspecific phosphory-
lation reactions of the other hydroxyl groups were detected,
all suggest that the enzyme seems to be able to recognise
metaphosphoric acid, more likely the terminal part of this poly-
mer, in a similar way to the reaction intermediate, to subse-
quently undergo a nucleophilic attack by the C3 hydroxyl
group in 1. These results also revealed that no covalent linkage
between donor and acceptor seems to be required for enzy-
matic catalysis because shikimic acid (1) can be converted into
shikimate 3-phosphate (2) by external addition of a mimetic of
the metaphosphate intermediate (second step of dissociative
process).
domain would occur. As a consequence, the substrate seems
to be less well anchored in the SB domain for the enzymatic
reaction to occur. The binding free-energy of shikimic acid (1)
in the three Hp-SK/L/1 complexes was calculated by using the
MM/PBSA approach in explicit water (GB) as implemented in
AMBER. The results suggest that 1 has approximately sixfold
higher affinity for the Hp-SK/ADP and Hp-SK/AMPNP complexes
than for the Hp-SK/AMPCP complex. We consider that this
motion might be responsible for the experimentally observed
lower efficiency of the enzymatic conversion when AMPCP is
used.
[
27]
In an effort to gain insight into the stability and the binding
mode of the proposed metaphosphate intermediate in the
active site of Hp-SK, MD simulation studies were performed.
The Hp-SK/ADP/PO /1 complex was then subjected to 50 ns of
3
dynamic simulation. As shown in Figure 8, the metaphosphate
intermediate remains trapped between the ADP and the shiki-
mic acid (1) during the simulation, mainly due to the strong
electrostatic interactions with the guanidinium group of the
To prove that the presence of ADP is required to achieve the
appropriate arrangement for catalysis but that it does not par-
ticipate in the reaction, two synthetic analogues of ADP,
namely AMPNP and AMPCP, were also employed (Figure 5 and
Figure 6). Adenosine-5’-[(a,b)-imido]diphosphate (AMPNP) and
adenosine-5’-[(a,b)-methyleno]diphosphate (AMPCP), are ADP
analogues in which the b-phosphate is replaced by imidophos-
phate and methylenophosphate derivatives, respectively. As
shown in Figure 5, the NMR results indicate that Hp-SK also
catalyses the stereospecific phosphorylation of the C3 hydroxyl
group of 1 by addition of metaphosphoric acid in the presence
of AMPNP. The latter experiments also highlight the required
closed form of the LID for catalysis. Remarkably, this process
was much faster than observed when using ADP. Thus, to ach-
ieve approximately 50% conversion, incubation for 24 h is re-
quired when using ADP vs. only 2 h for AMPNP. In contrast, ex-
periments carried out using AMPCP showed that the conver-
sion of 1 into 2 is much slower (Figure 6). In fact, the enzymat-
ic reaction required the use of much larger quantities of meta-
phosphoric acid and, under these conditions, only
approximately 20% conversion was observed.
2
+
essential Arg116, the side chain of Lys14, and Mg . This pro-
cess can be clearly visualised by analysing the variation of the
À
distances involving the O1 oxygen atom of PO3 and the NZ
À
2+
atom of Lys14 (red), the O2 oxygen atom of PO3 and Mg
À
(purple), the O3 oxygen atom of PO3 , and the NH2 and NE
atoms of the essential Arg116 (green and blue, respectively)
during the whole simulation (Figure 8C). It has been previously
suggested that Arg116/Arg117 might also act by activating the
g-phosphate group of ATP for catalysis and also trigger the re-
[
22]
lease of the product from the active site. These studies re-
vealed that Arg116 might also act by activating and position-
ing the metaphosphate intermediate for subsequent nucleo-
philic attack by the C3 hydroxyl group in 1. Moreover, consid-
ering that the conversion of ATP into ADP is exothermic
À1
(À7.3 kcalmol ), one would expect that the rate-limiting step
of the reaction catalysed by SK would be the nucleophilic
attack by the C3 hydroxyl group on the metaphosphate inter-
mediate to afford a tetrahedral transition state. Potential inhibi-
tors that target the SK enzyme might be mimetics of this tetra-
hedral species.
Considering that the differences observed on using ADP,
AMPNP and AMPCP would be due to some differences in ach-
ieving a closed arrangement of the active site, MD simulation
studies were performed. These studies were also carried out in
aqueous solution using the molecular mechanics force field
AMBER. The three possible Hp-SK/L/1 complexes were then
subjected to 50 ns of dynamic simulation (L=AMPNP, AMPCP
or ADP). As shown in Figure 7, the use of AMPNP as ligand led
to the most efficient closed form of the active site, with the es-
sential arginine side chain in close contact with the phosphate
groups of the ligand and with the shikimic acid (1) anchored in
A dissociative mechanism is supported by the findings:
1) the distance observed between the phosphorus atom of
ATP g-phosphate and the C3 hydroxyl atom of 1 during 32 ns
of simulation of both SK/ATP/1 Michaelis complexes is over 5 Š
during most of the simulation. Therefore, both substrates
seem to be too far away to form a pentavalent phosphorane
intermediate in which the two substrates are linked through
a phosphorus atom with distances between O and P atoms
close to a single OÀP bond (ca. 1.6 Š); and 2) the results from
our NMR studies show that the enzyme seems to be able to
recognise metaphosphoric acid to perform the stereospecific
phosphorylation of the C3 hydroxyl group in 1, as occurs on
using ATP as the phosphorus source and when a closed form
of the active site is achieved either using ADP, AMPNP or
AMPCP. Hybrid QM/MM calculations are required to gain fur-
ther insights into the catalytic mechanism in atomic detail and
this study is in progress in our laboratories.
[
22]
the SB domain in the appropriate conformation for reaction.
Significant changes were not observed during the whole simu-
lation and a tidy closed active site seems to be achieved (Fig-
ure 7A). A similar situation was obtained with the Hp-SK/ADP/
1
complex. In contrast, for the Hp-SK/AMPCP/1 complex, a less
stable closed arrangement of the active site was obtained (Fig-
ure 7B).
Our simulation studies indicate that significant motion in
the substrate covering-loop, the essential arginine, and SB
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2762
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
31
Figure 5. 1D H and P NMR monitoring of the enzymatic conversion of shikimic acid (1) into shikimate 3-phosphate (2) catalysed by Hp-SK in the presence
1
31
of AMPNP and metaphosphoric acid. H and P NMR spectra: A) 1, AMPNP and Hp-SK under the assay conditions. B) after addition to (A) of metaphosphoric
acid (18 mm). C) after 20 min incubation. D) after 1 h incubation. E) after 2 h incubation. F) after 5 h incubation. G) after 24 h incubation. All experiments were
1
31
recorded at 500 MHz ( H) or 202 MHz ( P) and 298 K using shikimic acid (3.2 mm), AMPNP (1.8 mm) and a catalytic amount of Hp-SK (10 units) in Tris-d11·DCl
1
(
2
100 mm pD 7.7), KCl (100 mm) and MgCl (5 mm). Only relevant regions of the H NMR spectra are shown for clarity. The most relevant signals that reveal the
31
1
formation of 2 are marked with an arrow (D). The P and H signals shown in Panel A corresponding to relevant peaks of 1 and AMPNP are labelled. Pb, Pa
and Pi correspond to b- and a-phosphate of AMPNP and phosphate, respectively. Ad=adenosine moiety.
Conclusion
sential arginine (Arg116/Arg117 in H. pylori and M. tuberculosis,
The results reported here indicate that the reaction catalysed
by SK involves a dissociative phosphoryl-transfer mechanism
via a metaphosphate intermediate. Commercially available
metaphosphoric acid proved to be a good mimetic of the
latter intermediate, the transformation of which by the
respectively) in close contact with the substrates and products.
Of the two synthetic analogues of ADP used in these studies,
AMPNP proved to be more efficient than AMPCP. The results of
the MD simulation studies suggested that AMPNP and ADP
provide a better disposition of shikimic acid in the active site
for catalysis than AMPCP, which might account for their higher
efficiency. The results of the simulation studies also suggest
that the essential arginine would also be involved in the stabi-
lisation of the metaphosphate intermediate during the cataly-
1
31
enzyme can be easily followed by 1D H and P NMR spectros-
copy. These experiments require the presence of ADP or a syn-
thetic analogue of ADP, such as AMPNP or AMPCP to achieve
a closed form of the LID with the guanidinium group of the es-
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2763
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1
31
Figure 6. 1D H and P NMR monitoring of the enzymatic conversion of shikimic acid (1) to shikimate 3-phosphate (2) catalysed by Hp-SK in the presence of
1
31
AMPCP and metaphosphoric acid. H and P NMR spectra: A) 1, AMPCP and Hp-SK under the assay conditions. B) After addition to (A) of metaphosphoric
acid (36 mm). C) After 18 h incubation. D) After addition to (C) of metaphosphoric acid (total concentration 54 mm) and 33 h incubation. All experiments were
1
31
recorded at 500 ( H) or 202 MHz ( P) and 298 K using shikimic acid (3.2 mm), AMPCP (4.5 mm), and a catalytic amount of Hp-SK (10 units) in Tris-d11·DCl
1
(
2
100 mm pD 7.7), KCl (100 mm) and MgCl (5 mm). Only relevant regions of the H NMR spectra are shown for clarity. The most relevant signals that reveal the
31
1
formation of 2 are marked with an arrow (C). The P and H signals shown in Panel A corresponding to relevant peaks of 1 and AMPCP are labeled. Pb, Pa
and Pi correspond to b- and a-phosphate of AMPCP and phosphate, respectively. Ad=adenosine moiety.
sis. We believe that the NMR studies reported here will be
helpful for mechanism- and structure-based design of inhibi-
tors of this attractive target for antibiotic drug discovery as
well as for the ready and rapid evaluation of possible dissocia-
tive phosphoryl-transfer mechanisms of other kinase enzymes.
(Figure S7). A similar comparison with AROK_HELPJ, Q9ZMS3
strain J99/ATCC 700824) showed two changes, namely M99I and
T123A.
(
The nucleotide sequence encoding the Hp-SK enzyme was sub-
cloned into the Escherichia coli expression plasmid pET3a via 5/
/
Nde1 and 3 BamH1 sites. The resulting plasmid was designated
pMUT319. E. coli strain BL21-AI was transformed with pMUT319,
and transformants were selected at 378C on Luria Broth (LB) agar
supplemented with 100 mgmL ampicillin. Transformed cells were
Experimental Section
À1
Helicobacter pylori SK cloning, expression and purification: The
aroK gene, which encodes Hp-SK, was PCR-amplified using genom-
ic DNA from an isolated clinical strain, courtesy of the Guadalajara
Hospital (Guadalajara, Spain) as template with the Expand High Fi-
delity PCR System (Roche). H. pylori strain was cultured in Colum-
bia agar plates in microaerobic conditions (GasPak EZ CampyPouch
System, Becton and Dickinson). The forward primer included a rec-
ognition site for the restriction enzyme NdeI (5’-CGCATCCATATG-
CAGCATTTAGTTTTAATCG-3’) and the reverse primer for the restric-
tion enzyme BamHI (5’-CGGGATCCTTATGCGATGAATTGTAGCAC-3’).
The PCR-amplified fragment, corresponding to aroK gene, was di-
gested with NdeI and BamHI and ligated to the expression vector
pET3a previously digested with the same restriction enzymes.
Comparison of the amino acid sequence in Hp-SK enzymes from
AROK_HELPY, P56073 (strain ATCC 700392/26695) with clinically
isolated strain from Guadalajara Hospital (Spain) revealed that the
latter strain has three changes, specifically V75I, M99I and T123A
grown in drug supplemented LB at 258C to an attenuance of 0.8
at 550 nm, at which time the temperature was reduced to 158C.
The culture medium was then made 0.2% (w/v) with arabinose
and incubation was continued at 158C for a further 15 h. Cells
were harvested by centrifugation (48C, 10000”g, 10 min) and
stored at À208C until needed. Cell paste (50 g) was sonicated at
48C in 50 mm potassium phosphate buffer pH 7.2, 150 mm NaCl,
1 mm DTT, 1 mm benzamidine (Buffer 1). Insoluble material was re-
moved at 48C by centrifugation at 21860”g for 1 h. The clarified
supernatant was then applied to a Q sepharose column (5”15 cm)
equilibrated in 50 mm potassium phosphate buffer pH 7.2, 150 mm
NaCl, 1 mm DTT (Buffer 2). The column flowthrough and a 1 L
wash of Buffer 2 were combined and made 35% saturated with
ammonium sulfate. Precipitated material was removed at 48C by
centrifugation at 13690”g for 1 h. The supernatant was made
55% saturated with ammonium sulfate and the precipitated mate-
rial recovered by centrifugation at 13690”g for 1 h. The pellet was
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2764
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 7. Comparison of several snapshots of the position of shikimic acid (1), ADP and the two ADP analogues in the Hp-SK/AMPNP/1 (A), Hp-SK/AMPCP/
1
(B,D), Hp-SK/ADP/1 (C) complexes during 50 ns of MD simulations. Note how in the Hp-SK/AMPNP/1 and Hp-SK/ADP/1 complexes the position of the essen-
tial arginine side chain and the substrate covering-loop remain stable during the simulation. In contrast, for the Hp-SK/AMPCP/1, a less stable closed arrange-
ment of the active site is obtained. Only relevant side chain residues are shown and labelled.
recovered, dissolved in 170 mL of 50 mm potassium phosphate
buffer pH 7.2, 150 mm NaCl, 1 mm DTT, 1.0m ammonium sulfate
absorbance at 340 nm in the UV spectrum due to the absorbance
À1
À1
of NADH (e, 6200 M cm ) in a coupled assay format, wherein
ADP formed after the formation of shikimate-3-phosphate (2) was
detected by using pyruvate kinase (PK) and lactate dehydrogenase
(LDH). Oxidation of NADH to NAD during PK-LDH activity was
monitored at 340 nm. Standard assay conditions for shikimate
kinase were 100 mm Tris·HCl pH 7.7, 100 mm NaCl, 5 mm MgCl2,
2.5 mm ATP, 1 mm shikimic acid (1), 0.20 mm NADH, ca. 2.8 units of
PK-LDH and 0.03 unit of SK protein. One unit of enzyme is defined
as the amount of enzyme required to convert 1 mmol of substrate
into product in 1 min. Each assay was initiated by addition of shiki-
mic acid (1). Solutions of ATP and NADH were calibrated by meas-
uring the absorbance at 259 and 340 nm in the UV spectrum, re-
spectively. Solutions of shikimic acid (1) were calibrated by equili-
bration with SK and measurement of the change in the UV absorb-
ance at 340 nm due to the disappearance of NADH. The initial
(
Buffer 3) and loaded onto a phenyl Sepharose column (5”15 cm)
equilibrated in Buffer 3. Following a 1.0 L wash with Buffer 3, the
column was eluted with a 1 L ammonium sulfate gradient (1.0m to
0
.0m in Buffer 2) collecting 1 mL fractions. At the end of the gradi-
ent, the column was washed with a further 500 mL of Buffer 3 col-
lecting 10 mL fractions. Hp-SK containing fractions were identified
by SDS-PAGE (12% gel) and pooled appropriately. Hp-SK protein
was then dialysed against 3”5 L changes of 25 mm potassium
phosphate buffer pH 6.6, 150 mm NaCl, 1 mm DTT (Buffer 4) and
loaded onto a hydroxyapatite column (5”7.5 cm) equilibrated in
Buffer 4. Following a 500 mL wash with Buffer 4, the column was
then eluted with a gradient of potassium phosphate buffer pH 6.6
(
1 L, 25 to 400 mm) containing 150 mm NaCl and 1 mm DTT collect-
ing 10 mL fractions. Hp-SK containing fractions were identified by
SDS-PAGE (12% gel) and pooled appropriately. The pooled protein
rates at fixed enzyme and substrate concentrations (0.2–5 K ) were
m
(
8
100 mL) was dialysed against 5 L Buffer 2, yielding approximately
20 mg of Hp-SK at greater than 95% purity.
measured. Kinetic constants were evaluated by using the GraFit 5
program (Erithacus Software Ltd.). The determined kinetic parame-
ters for Hp-SK and Mt-SK were K (1)=39Æ8 mm; k =116 Æ4
Shikimate kinase assay:
(
(
Shikimate kinase from M. tuberculosis (Mt-SK) was purified as de-
scribed previously and the concentrated solution (1.5 mgmL )
A
concentrated solution of Hp-SK
m
cat
À3 À1
À1
À1
10
s
and Km (1)=544Æ14 mm; kcat =295Æ8 s , respectively
7.4 mgmL ) in potassium phosphate buffer (50 mm, pH 7.2), DTT
(
Table S1).
1 mm) and NaCl (150 mm), which was stored at À808C, was used.
[24]
À1
Molecular Dynamics Simulations
was stored in potassium phosphate buffer (50 mm, pH 7.2), DTT
(
1 mm) and NaCl (150 mm) at À808C. When required for assays, ali-
Ligand preparation: Ligand geometries were minimised by using
quots of the enzyme stocks were diluted into buffer and stored on
ice. Enzyme activity was measured by monitoring the decrease in
a restricted Hartree–Fock (RHF) method and a 6–31G(d) basis set,
[28]
as implemented in the ab initio program Gaussian 09. Ligands
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2765 ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 8. MD simulation studies of Hp-SK/ADP/PO
and the substrate in the Hp-SK/ADP/PO
3
/1 complex. A) Comparison of several snapshots of the position of ADP, the metaphosphate intermediate
/1 complex during 50 ns of MD simulations. Note how the metaphosphate intermediate remains trapped between
3
the ADP and the substrate during the simulation, mainly due to the strong electrostatic interaction with the guanidinium group of the essential Arg116.
B) Detailed view of the relevant binding interactions of the metaphosphate intermediate in the active site of Hp-SK after 10 ns of MD simulation. C) Variation
À
À
2+
of the distance between the O1 oxygen atom of PO
3
and the NZ atom of Lys14 (red), the O2 oxygen atom of PO
3
and Mg (purple), the O3 oxygen atom
/1 com-
À
of PO
3
and the NH2 and NE atoms of the essential Arg116 (green and blue, respectively) during 50 ns of MD simulation of the two Hp-SK/ADP/PO
3
plexes. Relevant side-chain residues are shown and labelled.
[
21]
were manually docked into the active site. The resulting wavefunc-
tions were used to calculate electrostatic potential-derived (ESP)
ADP and shikimic acid (1) (PDB code 2IYQ ) and after manual re-
placement of ADP by ATP.
[23]
charges employing the restrained electrostatic potential (RESP)
methodology, as implemented in the assisted model building with
(
b) Hp-SK model: These studies were carried out using the enzyme
coordinates found in the crystal structure of Hp-SK/ADP/2 complex
PDB code 3MUF ) after manual replacement of both products by
[29]
energy refinement (AMBER)
suite of programs. The missing
[20]
(
bonded and nonbonded parameters were assigned, by analogy or
through interpolation, from those already present in the AMBER
the substrates.
[30]
Crystallographic water molecules were maintained in both com-
plexes except for those coordinated with the b-phosphate of ADP
in PDB 2IYQ. For the two Michaelis complexes, the crystallographic
water molecules coordinated with the g-phosphate of ATP and
Mg in PDB 2IYW were also added. Protein N- and C-terminal
groups were modified as N-acetyl and N-methyl amide derivatives,
database (GAFF).
Generation and minimization of the SK Michaelis complexes:
a) Mt-SK model: Simulations were carried out using the enzyme
coordinates found in the crystal structure of Mt-SK in complex with
2
+
(
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2766
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
respectively. Computation of the protonation state of titratable
of ADP or the enzyme, the corresponding amount was replaced by
deuterated buffer.
+
+
[31]
groups at pH 7.0 were carried out using the H Web server. Ad-
dition of hydrogen and molecular mechanics parameters from the
ff03 and GAFF force fields, respectively, were assigned to the pro-
NMR monitoring of the enzymatic conversion of shikimic acid (1)
and ATP to provide shikimate 3-phosphate (2) and ADP was carried
out using ATP magnesium salt (4.5 mm), shikimic acid (1) (6.4 mm),
[32]
tein and the ligands using the LEaP module of AMBER 12. ATP,
2
+
ADP and Mg parameters used with the AMBER force field were
KCl (100 mm) and MgCl2 (5 mm) in Tris-d ·DCl (100 mm pD 7.7).
11
[33–35]
included.
model was then carried out in stages, starting with ligand
1000 steps, half of them steepest descent, the other half conju-
Energy minimization using the implicit solvent GB
NMR spectra were recorded prior to and after the addition of 1 to
detect possible changes in the ATP. The reaction was then initiated
with the addition of a catalytic amount of Mt-SK (0.3 units). NMR
spectra were immediately recorded after the addition of the
enzyme and over an 18 h period.
(
gate gradient), protein side chains (1000 steps, idem) and finally
the entire complex (1000 steps, idem). A positional restraint force
À1 À2
constant of 50 kcalmol
Š
was applied to those unminimised
atoms in each step during all calculations. The complex was neu-
tralised by addition of chloride ions and immersed in a truncated
Acknowledgements
[29,36,37]
octahedron of ca. 5300 TIP3P water molecules.
Simulations: MD simulations were performed using the sander
Financial support from the Spanish Ministry of Economy and
Competiveness (SAF2013-42899-R), Xunta de Galicia (GRC2013-
041) and the European Regional Development Fund (ERDF) is
gratefully acknowledged. V.P. and E.L. thank the Spanish Minis-
try of Economy and Competiveness and Xunta de Galicia for
their respective FPI and postdoctoral fellowships. J.A.V. and
A.B. thank the Sara Borrell Programme (CD13/00373) and the
ISCIII General Subdirection of Assesment and Promotion of the
Research (PI14/00059), respectively, for financial support. We
are grateful to Dr. Trinidad Parra Cid from Guadalajara Hospital
[38]
module from the AMBER 12 suite of programs. Periodic boun-
dary conditions were applied, and electrostatic interactions were
[39]
treated using the smooth particle mesh Ewald method (PME)
with a grid spacing of 1 Š. The cutoff distance for the nonbonded
[40]
interactions was 9 Š. The SHAKE algorithm was applied to all
À5
bonds containing hydrogen, using a tolerance of 10 Š and an in-
tegration step of 2.0 fs. Minimization was carried out in three
steps, starting with the octahedron water hydrogen atoms, fol-
lowed by solvent molecules and counterions, and finally the entire
system. The minimised system was heated at 300 K (1 atm, 200 ps,
À1 À2
a positional restraint force constant of 50 kcalmol Š ). These ini-
(
Guadalajara, Spain) for kindly providing the isolated clinical
tial harmonic restraints applied to all a carbon atoms were gradu-
À1 À2
strain of Helicobacter pylori and to the Centro de Supercompu-
ally reduced from 5 to 1 mol
Š
(six steps, 100 ps each), and the
resulting systems were allowed to equilibrate further (100 ps). Un-
restrained MD simulations were carried out for 32 ns. System coor-
dinates were collected every 10 ps for further analysis. Figures de-
tación de Galicia (CESGA) for use of the Finis Terrae computer.
Keywords: drug discovery · enzyme catalysis · molecular
dynamics · molecular modelling · reaction mechanism
[41]
picting the structures were prepared by using PYMOL.
[
[
[
1] J. R. Knowles, Annu. Rev. Biochem. 1980, 49, 877–919.
2] F. H. Westheimer, Science 1987, 235, 1173–1178.
3] W. W. Cleland, A. C. Hengge, Chem. Rev. 2006, 106, 3252–3278.
NMR Experiments
All NMR experiments were performed with a Varian Inova 500
[4] J. K. Lassila, J. G. Zalatan, D. Herschlag, Annu. Rev. Biochem. 2011, 80,
1
31
spectrometer at 500 MHz ( H) or 202 MHz ( P) and 298 K using shi-
kimic acid (1) (6.4 mm), ADP sodium salt (4.5 mm), KCl (100 mm)
and MgCl (5 mm) in Tris-d ·DCl (100 mm, pD 7.7) and a catalytic
669–702.
[5] K. D. Allen, D. Dunaway-Mariano, Trends Biochem. Sci. 2004, 29, 495–
503.
2
11
[
[
6] A. C. Hengge, Adv. Phys. Org. Chem. 2005, 40, 49–108.
amount of Hp-SK (10 units) or Mt-SK (3 units). The reaction was ini-
tiated by the addition of a solution of metaphosphoric acid
7] N. Sträter, W. N. Lipscomb, T. Klabunde, B. Krebs, Angew. Chem. Int. Ed.
Engl. 1996, 35, 2024–2055; Angew. Chem. 1996, 108, 2158–2191.
8] S. Liu, Z. Lu, Y. Jia, D. Dunaway-Mariano, O. Herzberg, Biochemistry
(
18 mm) to a solution of 1, ADP, KCl, MgCl in deuterated buffer.
2
[
The stock samples of 1 (50 mm), ADP (50 mm) and metaphosphoric
acid (1m) were prepared in deuterated buffer. The total volume of
each NMR sample was 550 mL. Solvent suppression of the residual
HDO peak was achieved by continuous low-power presaturation
pulse during the relaxation delay. NMR spectra were recorded prior
to and after the addition of metaphosphoric acid over a 96 h
period, unless indicated otherwise.
2
002, 41, 10270–10276.
[9] K. Parang, J. H. Till, A. J. Ablooglu, R. A. Kohanski, S. R. Hubbard, P. A.
Cole, Nat. Struct. Biol. 2001, 8, 37–41.
[10] S. D. Lahiri, G. Zhang, D. Dunaway-Mariano, K. D. Allen, Science 2003,
2
99, 2067–2071.
11] S. T. Mundle, J. C. Delaney, J. M. Essigmann, P. R. Strauss, Biochemistry
009, 48, 19–26.
[
2
[
[
12] M. C. Hutter, V. Helms, Protein Sci. 2000, 9, 2225–2231.
13] S. Loverix, A. Winqvist, R. Stromberg, J. Steyaert, Chem. Biol. 2000, 7,
651–658.
[14] A. M. Sudom, L. Prasad, H. Goldie, L. T. J. Delbaere, J. Mol. Biol. 2001,
314, 83–92.
The NMR experiments conducted with adenosine-5’-[(a,b)-imido]di-
phosphate sodium salt (AMPNP) and adenosine-5’-[(a,b)-methyl-
eno]diphosphate sodium salt (AMPCP), purchased from Jena Bio-
science, were performed as described for those with ADP (see
above). It reaction involved shikimic acid (1) (3.2 mm), AMPNP
[
15] Y.-W. Xu, S. Morera, J. Janin, J. Cherfils, Proc. Natl. Acad. Sci. USA 1997,
4, 3579–3583.
16] M. Valiev, R. Kawai, J. A. Adams, J. H. Weare, J. Am. Chem. Soc. 2003,
25, 9926–9927.
9
(
1.8 mm), KCl (100 mm) and MgCl (5 mm) in Tris-d ·DCl (100 mm,
2 11
[
pD 7.7), metaphosphoric acid (18 mm) and a catalytic amount of
Hp-SK (10 units). For those experiments carried out with AMPCP
1
[
[
17] G. K. Smith, Z. Ke, H. Guo, J. Phys. Chem. A 2011, 115, 13713–13722.
18] B. Elsaesser, S. Dohmeier-Fischer, G. Fels, J. Mol. Model. 2012, 18, 3169–
(
4.5 mm), a higher concentration of metaphosphoric acid (36 mm)
was employed. After 18 h incubation, more metaphosphoric acid
was added (10 mL of 1m stock) to achieve a total concentration of
3179.
3
1
[19] For other P NMR studies on kinases, see: a) M. Cohn, Curr. Top. Cell.
Regul. 1984, 24, 1–13; b) N. M. Doliba, A. M. Babsky, N. M. Doliba, S. L.
5
4 mm. For those control experiments carried out in the absence
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org 2767
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Wehrli, M. D. Osbakken, Cell. Biochem. Funct. 2015, 33, 67–72; c) C. Ga-
bellieri, M. Beloueche-Babari, Y. Jamin, G. S. Payne, M. O. Leach, T. R.
Eykyn, NMR Biomed. 2009, 22, 456–461; d) G. Fayat, S. Blanquet, B. D. N.
Rao, M. Cohn, J. Biol. Chem. 1980, 255, 8164–8169; e) B. D. Rao, D. H.
Buttlaire, M. Cohn, J. Biol. Chem. 1976, 251, 6981–6986; f) A. Lecroisey,
I. Lascu, A. Bominaar, M. Veron, M. Delepierre, Biochemistry 1995, 34,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
[29] D. A. Case, T. E. Cheatham, T. Darden, H. Gohlke, R. Luo, K. M. Merz, O.
Onufriev, C. Simmerling, B. Wang, R. J. Woods, J. Comput. Chem. 2005,
26, 1668–1688.
1
2445–12450; g) M. Shimane, Y. Sugai, R. Kainuma, M. Natsume, H. Ka-
waide, Bios. Biotech. Biochem. 2012, 76, 1558–1560; h) H.-W. Kim, B.
Yeh, J.-W. Choi, Y.-S. Jeong, J.-H. Sohn, J.-H. Kim, Bull. Korean Chem. Soc.
[30] a) J. Wang, W. Wang, P. A. Kollman, D. A. Case, J. Mol. Graphics Modell.
2006, 25, 247–260; b) http://q4md-forcedfieldtools.org/RED/resp/.
[31] a) J. C. Gordon, J. B. Myers, T. Folta, V. Shoja, L. S. Heath, A. Onufriev, Nu-
cleic Acids Res. 2005, 33, W368–W371; b) http://biophysics.cs.vt.edu/
H+ +.
2012, 33, 2419–2422.
[
[
[
20] W. C. Cheng, Y. F. Chen, H. J. Wang, K. C. Hsu, S. C. Lin, T. J. Chen, T. J.
Yang, W. C. Wang, Plos One 2012, 7, e33481.
21] M. D. Hartmann, G. P. Bourenkov, A. Oberschall, N. Strizhov, H. D. Bartu-
nik, J. Mol. Biol. 2006, 364, 411 –423.
22] B. Blanco, V. Prado, E. Lence, J. M. Otero, C. García-Doval, M. J. van Raaij,
A. L. Llamas-Saiz, H. Lamb, A. R. Hawkins, C. Gonzµlez-Bello, J. Am.
Chem. Soc. 2013, 135, 12366–12376.
23] a) W. D. Cornell, P. Cieplak, C. I. Bayly, I. R. Gould, K. M. Merz, D. M. Fergu-
son, D. C. Spellmeyer, T. Fox, J. W. Caldwell, P. A. Kollman, J. Am. Chem.
Soc. 1995, 117, 5179–5197; b) http://q4md-forcedfieldtools.org/RED/
resp/.
24] T. Krell, J. Maclean, D. J. Boam, A. Cooper, M. Resmini, K. Brocklehurst,
S. M. Kelly, N. C. Price, A. J. Lapthorn, J. R. Coggins, Protein Sci. 2001, 10,
[32] AMBER 12, Case, D. A., Darden, T. A., Cheatham III, T. E., Simmerling,
C. L., Wang, J., Duke, R. E., Luo, R., Walker, R. C., Zhang, W., Merz, K. M.,
Roberts, B., Hayik, S., Roitberg, A., Seabra, G., Swails, J., Goetz, A. W.,
Kolossvµry, I., Wong, K. F. , Paesani, F. , Vanicek, J., Wolf, R. M., Liu, J., Wu,
X., Brozell, S. R., Steinbrecher, T., Gohlke, H., Cai, Q., Ye, X., Wang, J.,
Hsieh, M.-J., Cui, G., Roe, D. R., Mathews, D. H., Seetin, M. G., Salomon-
Ferrer, R., Sagui, C., Babin, V. , Luchko, T. , Gusarov, S., Kovalenko, A.,
Kollman, P. A. University of California, San Francisco, 2012.
[33] O. AllnØr, L. Nilsson, A. Villa, J. Chem. Theory Comput. 2012, 8, 1493–
1502.
[34] Mg parameters used with the AMBER force field were downloaded
from http://www.pharmacy.manchester.ac.uk/bryce/amber/.
[35] K. L. Meagher, L. T. Redman, H. A. Carlson, J. Comput. Chem. 2003, 24,
1016–1025.
[36] J. Aqvist, J. Phys. Chem. 1990, 94, 8021–8024.
[37] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, J. Chem. Phys. 1983, 79,
926–935.
[38] S. Le Grand, A. W. Goetz, R. C. Walker, Comp. Phys. Comm. 2013, 184,
374–380.
[39] R. Salomon-Ferrer, A. W. Goetz, D. Poole, S. Le Grand, R. C. Walker, J.
Chem. Theory Comput. 2013, 9, 3878–3888.
[40] J.-P. Ryckaert, G. Ciccotti, H. J. C. Berendsen, J. Comput. Phys. 1977, 23,
327–341.
[41] http://www.pymol.org.
[
[
2
+
1
137–1149.
25] Y. Gu, L. Reshentnikova, Y. Li, Y. Wu, H. Yan, S. Singh, X. Ji, J. Mol. Biol.
002, 319, 779–789.
[
[
[
[
2
26] B. Dhaliwal, C. E. Nichols, J. Ren, M. Lockyer, I. Charles, A. R. Hawkins,
D. K. Stammers, FEBS Lett. 2004, 574, 49–54.
27] B. R. Miller III, T. D. McGee Jr., J. M. Swails, N. Homeyer, H. Gohlke, A. E.
Roitberg, J. Chem. Theory Comput. 2012, 8, 3314–3321.
28] Gaussian 09, revision A.2, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
Received: November 4, 2015
Published online on January 21, 2016
Chem. Eur. J. 2016, 22, 2758 – 2768
www.chemeurj.org
2768
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim