Molecular insight from DFT computations and kinetic measurements into the steric factors influencing peptide bond hydrolysis catalyzed by a dimeric Zr(IV)-substituted keggin type polyoxometalate

Article abstract of DOI:10.1021/acs.inorgchem.6b01461

Peptide bond hydrolysis of several peptides with a Gly-X sequence (X = Gly, Ala, Val, Leu, Ile, Phe) catalyzed by a dimeric Zr(IV)- substituted Keggin type polyoxometalate (POM),(Et₂NH₂)₈[{α-PW₁₁O₃₉Zr- (μ-OH)(H₂O)}₂]·7H₂O (1), was studied by means of kinetic experiments and ¹H NMR spectroscopy. The observed rate of peptide bond hydrolysis was found to decrease with increase of the side chain bulkiness, from 4.44 × 10^-7 s^-1 for Gly-Gly to 0.81 × 10^-7 s^-1 for Gly-Ile. A thorough DFT investigation was performed to elucidate (a) the nature of the hydrolytically active species in solution, (b) the mechanism of peptide bond hydrolysis, and (c) the influence of the aliphatic residues on the rate of hydrolysis. Formation of substrate-catalyst complexes of the dimeric POM 1 was predicted as thermodynamically unlikely. Instead, the substrates prefer to bind to the monomerization product of 1, [α-PW₁₁O₃₉Zr(OH)(H₂O)]^4- (2), which is also present in solution. In the hydrolytically active complex two dipeptide ligands are coordinated to the Zr(IV) center of 2. The first ligand is bidentate-bound through its amino nitrogen and amide oxygen atoms, while the second ligand is monodentate-bound through a carboxylic oxygen atom. The mechanism of hydrolysis involves nucleophilic attack by a solvent water molecule on the amide carbon atom of the bidentate-bound ligand. In this process the uncoordinated carboxylic group of the same ligand acts as a general base to abstract a proton from the attacking water molecule. The decrease of the hydrolysis rate with an increase in the side chain bulkiness is mostly due to the increased ligand conformational strain in the rate-limiting transition state, which elevates the reaction activation energy. The conformational strain increases first upon substitution of Hα in Gly-Gly with the aliphatic α substituent and second with the β branching of the α substituent.

Full text of DOI:10.1021/acs.inorgchem.6b01461

Article
pubs.acs.org/IC
Molecular Insight from DFT Computations and Kinetic
Measurements into the Steric Factors Influencing Peptide Bond
Hydrolysis Catalyzed by a Dimeric Zr(IV)-Substituted Keggin Type
Polyoxometalate
,
†
‡
†
‡
Tzvetan T. Mihaylov,* Hong Giang T. Ly, Kristine Pierloot, and Tatjana N. Parac-Vogt
†Laboratory of Computational Coordination Chemistry, Department of Chemistry, KU Leuven, Celestijnenlaan 200F, 3001 Leuven,
Belgium
‡
Laboratory of Bioinorganic Chemistry, Department of Chemistry, KU Leuven, Celestijnenlaan 200F, 3001 Leuven, Belgium
*
S Supporting Information
ABSTRACT: Peptide bond hydrolysis of several peptides with a Gly-X
sequence (X = Gly, Ala, Val, Leu, Ile, Phe) catalyzed by a dimeric Zr(IV)-
substituted Keggin type polyoxometalate (POM), (Et NH ) [{α-PW O Zr-
2
2 8
11 39
(
μ−OH)(H O)} ]·7H O (1), was studied by means of kinetic experiments
2
2
2
1
and H NMR spectroscopy. The observed rate of peptide bond hydrolysis
was found to decrease with increase of the side chain bulkiness, from 4.44 ×
−
7
−1
−7 −1
10
s
for Gly-Gly to 0.81 × 10
s
for Gly-Ile. A thorough DFT
investigation was performed to elucidate (a) the nature of the hydrolytically
active species in solution, (b) the mechanism of peptide bond hydrolysis, and
(
c) the influence of the aliphatic residues on the rate of hydrolysis. Formation
of substrate−catalyst complexes of the dimeric POM 1 was predicted as thermodynamically unlikely. Instead, the substrates
4−
prefer to bind to the monomerization product of 1, [α-PW O Zr(OH)(H O)] (2), which is also present in solution. In the
11
39
2
hydrolytically active complex two dipeptide ligands are coordinated to the Zr(IV) center of 2. The first ligand is bidentate-bound
through its amino nitrogen and amide oxygen atoms, while the second ligand is monodentate-bound through a carboxylic oxygen
atom. The mechanism of hydrolysis involves nucleophilic attack by a solvent water molecule on the amide carbon atom of the
bidentate-bound ligand. In this process the uncoordinated carboxylic group of the same ligand acts as a general base to abstract a
proton from the attacking water molecule. The decrease of the hydrolysis rate with an increase in the side chain bulkiness is
mostly due to the increased ligand conformational strain in the rate-limiting transition state, which elevates the reaction activation
energy. The conformational strain increases first upon substitution of H in Gly-Gly with the aliphatic α substituent and second
α
with the β branching of the α substituent.
INTRODUCTION
reviews and publications focusing on the use of POMs in
■
6
7
8
heterogeneous catalysis, electrocatalysis, water oxidation, and
Selective hydrolysis of peptide bonds in proteins plays an
important role in a broad range of research fields such as
chemical biology, proteomics, and drug discovery. The
reaction is also frequently applied in structural determinations
of peptides and proteins. Proteolytic enzymes with a high
catalytic power can be used for this purpose. However, they
often operate only in a narrow range of temperature and pH. In
addition, the high cost and limited or fixed selectivity of
proteolytic enzymes has led to the search for alternative
cleavage agents for hydrolysis of peptide bonds, which are
9
oxidation of organic compounds have been published. In order
1
to improve or tune the catalytic properties of POMs, lanthanide
or transition-metal ions can be introduced into the skeleton of
vacant POMs to form metal-substituted POMs (MSPs). The
combination of the POM framework, with its negative charge
playing a role in electrostatic interaction with positively charged
substrates, and the active metal sites, which can coordinate
water or other ligands at their free coordination sites, makes
such MSPs attractive for catalytic applications. In addition, the
introduction of metal ions into the POM framework can
2
extremely stable at physiological pH and room temperature.
10
stabilize them, preventing gel formation in aqueous solution.
Polyoxometalates (POMs) are generally described as clusters
of early-transition-metal ions in their highest oxidation state
bridged by oxo anions. Owing to their fascinating properties
such as tunable acidity and redox properties, high thermal
stability, hydrolytic and oxidative stability, POMs have been
The first investigations concerned the catalytic properties of
Ti(IV)- and Zr(IV)-substituted POMs toward oxidation
11
reactions with H O . The use of MSPs as catalysts for a
2
2
broad range of chemical reactions, such as H O -based
2
2
3
4
applied in broad fields including catalysis, materials, and
5
medicine. In recent years applications of POMs for catalytic
Received: June 17, 2016
purposes have also been extensively investigated. Series of
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
oxidation of organic compounds, cyclization of citronellal
Article
12
rate is dependent on the nature and the bulkiness of the amino
acid side chain. The first mechanistic study on peptide bond
1
3
19c
derivatives, Diels−Alder reactions and aldol reactions of
1
4
15
imines, oxygenation of catechols, synthesis of cyclic
hydrolysis catalyzed by a dimeric tetrazirconium(IV) Wells−
1
6
17
23
carbonates, oximation of aldehydes and ketones, and the
splitting of water, has recently been reported.
Dowson POM was also reported recently by us. For
1
8
dipeptides with aliphatic residues the hydrolytically active
substrate−catalyst complex is formed through bidentate chelate
coordination of the ligand to a Zr(IV) center of the POM,
involving the amino nitrogen and the amide oxygen atoms,
while the terminal ligand carboxylic group remains uncoordi-
nated. In the rate-determining reaction step the latter acts as a
general base to facilitate nucleophilic attack by a solvent water
molecule on the amide carbon atom. Similar to the case for 1,
the rates of hydrolysis within a series of dipeptides were also
found with this catalyst to be dependent on both the nature and
the bulkiness of the side chains. The molecular factors
governing this selectivity, however, remain unclear. The large
variety of amino acid sequences involved in the dipeptide series
presupposes a variety of different substrate−catalyst complexes
and possible mechanistic scenarios for hydrolysis. For instance,
some of the dipeptides involve amino acids with residues that
could coordinate to Zr(IV), such as Gly-Asn, Gly-Asp, Gly-Gln,
Gly-His, Gly-Ser, and Gly-Thr, while others with positively
charged residues, such as Gly-Lys and Gly-Arg, may stabilize
the complex through electrostatic interactions with the
negatively charged POM surface. Therefore, the coordination
behavior and the mechanisms of hydrolysis should be clarified
individually in every particular case.
In our quest to develop MSPs as potential catalysts for
peptide bond hydrolysis, Zr(IV)- and Ce(IV)-substituted
POMs have been used, because both Zr(IV) and Ce(IV)
have a high Lewis acidity, oxophilicity, and kinetic lability and
tend to form complexes with high coordination numbers. The
hydrolytic activity of a series of Zr(IV)-substituted Keggin,
Lindqvist, and Wells−Dawson POMs toward peptide bond
1
0a,b,19
20
hydrolysis in dipeptides
and oligopeptides was
discovered previously. All complexes act as homogeneous
catalysts and significantly accelerate the hydrolysis reaction
rates in mildly acidic and neutral pH environments. In addition,
we recently showed that a Ce(IV)-substituted Keggin POM
and Zr(IV)-substituted Keggin, Lindqvist, and Wells−Dawson
POMs selectively hydrolyze proteins such as hen egg white
lysozyme, albumins, and horse heart myoglobin, thus
demonstrating the potential of Ce(IV)- and Zr(IV)-substituted
10c,21
POMs as novel artificial proteases.
In aqueous solution, dimeric Zr(IV)-substituted POMs can
partially dissociate into monomeric species, which are
presumed to be the active species during the catalytic
1
0b,19a
reactions.
For example, the 1:1 Zr(IV)-Keggin POM
4
−
[
α-PW O Zr(OH)(H O)] was identified as the active
11
39
2
This study is focused on the hydrolytic activity of 1 toward a
series of dipeptides with a common Gly-X sequence, where X is
an amino acid with an aliphatic (noncoordinating) side chain,
such as Ala, Val, Leu, Ile, and Phe. In this series both the
coordination behavior and the mechanisms of hydrolysis are
expected to be rather similar, which allows us to distinguish the
role of the aliphatic side chain in the process. However, the
tendency of the parent compound 1 for partial dissociation in
solution brings additional complexity to the problem studied.
The kinetics of the Gly-X dipeptide hydrolysis in the presence
compound for the hydrolysis of DNA-model phosphoesters
by both experimental and theoretical means, even though the
22
dimeric 2:2 Zr(IV)-Keggin POM [{α-PW O Zr(μ−OH)-
1
1
39
8
−
22
(
H O)} ] (1) was used as a starting catalyst (see Figure 1).
2 2
This may be the reason for the superior hydrolytic activity of 1
toward myoglobin, in comparison to other POMs which do not
10c,23
undergo dissociation.
The selectivity of 1 toward dipeptides containing different
amino acid sequences (other than those studied here) was
recently reported, and the results indicated that the reaction
1
of 1 was studied experimentally by an H NMR method. To
explain the experimental findings, the nature of the hydrolyti-
cally active species, the mechanism of peptide bond hydrolysis,
and the effect of the aliphatic residues on the rate of hydrolysis
were investigated on a molecular level by means of density
functional methods.
EXPERIMENTAL SECTION
■
Chemicals. (Et NH ) [{α-PW O Zr(μ-OH)(H O)} ]·7H O (1)
2
2
8
11 39
2
2
2
was synthesized according to the procedure given in the literatur-
1
0b,24
e.
D O, DCl, NaOD, H PW O ·23H O, and the dipeptides
2 3 12 40 2
used in this study were purchased from Sigma-Aldrich.
NMR Measurements. 1H NMR spectra were recorded on a
Bruker Advance 400 spectrometer, and TMSP-d was used as an
4
internal reference.
Hydrolysis Studies. The hydrolysis reaction mixtures typically
contain 2.0 mM dipeptide, 2.0 mM 1, and 0.5 mM TMSP-2,2,3,3-d4
(
sodium 3-trimethylsilylpropionate) in D O. The pD of the reaction
2
mixtures was adjusted to 5.4 with minor amounts of 1.0 M DCl or 1.0
M NaOD. The pH-meter reading was corrected by the equation pD =
2
5
1
pH + 0.41. The reaction mixtures were kept at 60 °C, and H NMR
spectra were taken after mixing and after different time increments.
The hydrolysis products were identified by comparison of the H
1
NMR chemical shifts with those of the pure amino acids at the same
pD value. The rate constants (kobs) were calculated by fitting the
decrease in dipeptide concentration to a first-order monoexponential
decay function.
Figure 1. Molecular structures of 2:2 Zr(IV)-Keggin and 1:1 Zr(IV)-
Keggin POMs.
B
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 1. Structures of Gly-X Dipeptides
Computational Methods. Binding Study. DFT calculations were
applied. Due to the large size of the transition state (TS) complexes,
intrinsic reaction coordinate (IRC) calculations were performed only
in a few cases where we had doubts about the nature of the TS
obtained. Otherwise, the TS structures were judged by analyzing the
atomic movements of the corresponding imaginary frequencies. The
bulk solvent effects were included in the geometry optimizations by
26
performed with the ORCA program package, version 3.0. Molecular
geometries were fully optimized (without symmetry constraints)
27
employing the gradient-corrected BP86 functional and def2-SVP
28
basis sets.2 MWB28 and MWB60 Stuttgart−Dresden effective core
9
potentials (ECPs) were used to replace the core electrons of Zr and
4
1
W atoms, respectively. To speed up the calculations, the Split-RI-J
means of the CPCM solvation model with the default parameters
(UFF cavities). However, the final free energies of solvation were
obtained with Pauling atomic radii, taking into account the
nonelectrostatic energy contributions. The electronic energies of the
stationary points were adjusted in SP calculations utilizing the M06-
3
0
31
variant of the resolution of the identity (RI) technique with
32
auxiliary def2-SVP/J Coulomb fitting basis sets was employed
applying an increased integration grid, Grid4 in the ORCA
(
convention, and tight SCF convergence criteria). The effects of the
3
9
42
aqueous solvent were included in the geometry optimizations by using
2X-D3, M06-D3, and BMK-D3 functionals (D3 means that the
33
37b
the conductor-like screening model (COSMO), as implemented in
ORCA. The BP86 functional has already been used for geometry
original D3 damping function is used instead of the BJ function)
with def2-TZVP basis sets and MWB28/MWB60 ECPs for the Zr/W
atoms. These functionals were selected on the basis of our previous
3
4
optimization of various POMs in a number of theoretical works.
A
2
3
comparison of the structural parameters for [{α-PW O Zr(μ−
study on POM-catalyzed peptide bond hydrolysis.
1
1
39
8−
−1
OH)(H O)} ] (1) obtained in this study and those determined
A standard state of 1 mol L was used for all compounds except
water molecules, for which a 55.34 mol L standard state was used
2
2
−1
by X-ray analysis of (Et NH ) [{α-PW O Zr(μ-OH)(H O)} ]·
2
2
8
11 39
2
2
24b
43
7
H O in general showed a good agreement. Selected theoretical
2
instead.
vs experimental structural parameters may be found in Table S1 of the
Supporting Information. To find the lowest energy structures, a
conformational search was performed for all species under
consideration (except 1, for which the X-ray structure was used as
the input geometry). To verify that the stationary points found are
minima on the potential energy surface (PES) and to obtain the
thermochemical data, numerical vibrational frequencies were com-
puted using the standard rigid rotor-harmonic oscillator approxima-
tion. In order to refine the energies, single-point (SP) calculations on
the BP86 optimized structures were performed by means of the hybrid
RESULTS AND DISCUSSION
■
Hydrolysis of Gly-X Peptides. The hydrolysis of a series
of Gly-X peptides in which X = Gly, Ala, Val, Leu, Ile, Phe
Scheme 1) in the presence of 1 was examined in solution
containing equimolar amounts (2.0 mM) of the dipeptide and
(
1
1
at pD 5.4 and 60 °C. H NMR spectra of the hydrolysis
reactions were recorded at different reaction times, and the
35
density functional B3LYP and more extended basis sets, def2-
hydrolysis product glycine (Gly) was identified by a singlet
28
10b
TZVP, combined with the COSMO model. For the Zr and W atoms
the aforementioned ECPs were still used. The Split-RI-J combined
with the chain of spheres approximation to the exact exchange
peak at 3.57 ppm
(Figures S1−S6 of the Supporting
Information). The amounts of the reacted dipeptide and of the
1
free Gly were determined by integration of their respective H
3
6
(
RIJCOSX) was utilized with the auxiliary def2-TZVP/J Coulomb
32
NMR resonances, and the rate constants (kobs) were calculated
by fitting the decrease in dipeptide concentration to a
monoexponential decay function. The hydrolysis of all studied
dipeptides in the absence of 1 under the same experimental
conditions was also studied as a control reaction. Fully
homogeneous solutions were observed during the hydrolytic
reaction of all dipeptides, indicating that the presence of
fitting basis sets (applying Grid5, GridX5, and tight SCF
convergence criteria). When it is applied with the B3LYP functional,
the RIJCOSX approximation is expected to introduce a negligible error
in the reaction energies in comparison to the approximation free
3
6
calculations. To account for dispersion forces, missed in the pure
B3LYP method, Grimme’s atom-pairwise dispersion correction with
37
Becke−Johnson damping (D3BJ) was employed.
10b,19c
A detailed description of the strategy used for estimation of the free
energies of complexation is given in the Computational Details section
of the Supporting Information.
dipeptides does not affect the stability of 1.
The values of calculated rate constants for Gly-X hydrolysis
are collected in Table 1. Due to the extreme stability of the
peptide bond in the absence of 1, the rate constants for the
uncatalyzed hydrolysis of dipeptides could not be determined
under these conditions. However, the amount of dipeptide
hydrolyzed in the absence of 1 after 7 months could be
Reaction Path Modeling. This part of the calculations was
3
8
performed with the Gaussian 09 Rev. D.01 suite of programs.
Geometry optimizations and frequency calculations were performed
39
with the hybrid meta-GGA functional M06-2X and 6-31+G(d,p)
40
basis sets; for the Zr and W atoms LANL2DZ ECPs/basis sets were
C
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Observed Rate Constants for the Hydrolysis of Gly-
X in the Absence and Presence of 1
binding study, we have performed a DFT investigation on the
coordination behavior of the simplest dipeptide of this series,
Gly-Gly.
a
conversion after
7 months (%) (without 1)
107kobs (s−1
(with 1)
)
Coordination Properties of Gly-Gly toward Zr(IV)-
Substituted Keggin POMs. Coordination of Gly-Gly to 2:2
Zr(IV)-Keggin POM (1). There are four different basic sites in
the Gly-Gly zwitterion suitable for coordination to Zr(IV): the
terminal amino nitrogen, the terminal carboxylic oxygens, the
amide oxygen, and the amide nitrogen. As previously
demonstrated for Cu(I), Ni(I), and Co(I) ions, tri- and
bidentate modes of Gly-Gly coordination are much more
plausible than monodentate coordination modes. In addition,
the most plausible binding mode varies depending on the metal
ion and is either bidentate (for Cu(I)), through the terminal
amino and carboxylic groups, or tridentate (for Ni(I) and
Co(I)), involving in addition the amide nitrogen or oxygen
peptide
Gly-Gly
Gly-Ala
Gly-Val
Gly-Leu
Gly-Ile
25
5
4.44 ± 0.10
3.89 ± 0.37
1.71 ± 0.19
1.72 ± 0.35
0.81 ± 0.11
2.53 ± 0.15
0
0
0
Gly-Phe
0
a
Conditions: [1] = [dipeptide] = 2.0 mM, pD 5.4, 60 °C.
estimated and is also shown in Table 1. From these results it is
clear that all dipeptides are very stable and that their hydrolysis
was significantly accelerated by 1. The mechanism of Gly-Gly
hydrolysis in the presence of Zr-POMs was previously
suggested to occur via coordination of Gly-Gly to a Zr(IV)
45
atom. It should be noted that in the case of 1 tridentate
10b,23
binding to one metal center would result in a 9-coordinated
zirconium, which is unlikel, as the coordination number typical
ion via its amide oxygen and amine nitrogen atoms.
This
interaction results in polarization of the amide CO bond,
making it more susceptible toward nucleophilic attack by a
water molecule.
46
of Zr(IV) complexes including POMs is 7 or 8. Due to steric
hindrance caused by the bulky Keggin POM ligands, our
attempts to model even bidentate coordination of Gly-Gly to
one Zr atom of 1 were unsuccessful. Therefore, the most
feasible bidentate modes of coordination were constructed,
where the dipeptide bridges the two neighboring Zr atoms, by
varying the basic sites of Gly-Gly interacting with the metals.
The optimized molecular structures, the free energies of
complexation, and their discussion can be found in Figure S7,
Table S2, and the Coordination of Gly-Gly to Compound 1
section in the Supporting Information, respectively. In all cases
positive free energies of complexation, in the range of 0.7−18.5
The dependence of the hydrolysis rate constants on the
44
volume of the amino acid side chain at the C terminus is
shown in Figure 2. From this figure it can be seen that the
−1
kcal mol , were computed. Thus, our DFT calculations predict
that the formation of a complex between Gly-Gly and 1 is
thermodynamically unfavorable. Only in the case of bidentate
bridging coordination through the terminal carboxylic group
was a free energy of complexation close to 0 obtained (0.7 kcal
−1
mol ), indicating that such a complex may occur in the
reaction mixture. This result, however, cannot explain the
13
hydrolytic activity of 1. Moreover, the C NMR data showed
−
no evidence for −COO binding but instead suggest metal−
ligand interaction involving the terminal amino and peptide
10b
carbonyl groups.
According to the calculations, however,
formation of such a complex with 1 is rather unlikely, as this
requires ∼12 kcal mol− (at near neutral pH).
1
Figure 2. Reaction rate constants for the hydrolysis of 2.0 mM Gly-X
dipeptides in the presence of 2.0 mM 1 at pD 5.4 and 60 °C as a
function of the volume of amino acid X.
These results have stimulated us to find another explanation
for the hydrolytic activity of 1.
Coordination of Gly-Gly to 1:1 Zr(IV)-Keggin POM (2).
Depending on the pH, temperature, time, concentration, and
the presence of ligands, the parent compound 1 may undergo
partial dissociation in aqueous solution leading to the formation
hydrolysis rate constant decreases when the volume of the
noncoordinating X amino acid side chain is increased, implying
that steric hindrance either impairs the interaction between the
Zr(IV) ion and the dipeptide or prevents efficient nucleophilic
attack of water or hydroxide. Interestingly, despite its higher
molecular volume in comparison to that of Gly-Val, Gly-Ile, and
Gly-Leu, a higher rate constant was obtained for Gly-Phe.
To shed light on the hydrolytic processes at the atomistic
level, we have first studied the coordination properties of the
Gly-X dipeptides to the Zr(IV) atom(s) of the POM in order to
discern the potential hydrolytically active species in solution. In
a second step, the most likely candidates were used to model
possible reaction paths of hydrolysis. Since in the series Gly-X
the substrates differ only in their aliphatic (noncoordinating)
residues, we can safely assume that their coordination
properties will be rather similar. Thus, for the purpose of the
10−
of both 1:2 species [Zr(α-PW11
O
)
]
with the lacunary
39
2
7
−
anion [α-PW11
O ]
39
and of monomeric species [α-
(
3+n)−
11c,26
b
PW O Zr(OH) (H O) ]
(n = 0, 1).
The 1:2
11
39
n
2
3‑n
complex appears in basic solution and shows a very low
10b
reactivity toward peptide bond hydrolysis,
most probably
because of the lack of free coordination sites at the Zr(IV)
center when it is sandwiched between the two Keggin moieties.
Therefore, substrate coordination to the 1:2 complex was not
examined in our DFT models. The presence of monomeric
31
species has first been suggested on the basis of P NMR signals
under strongly acidic conditions (0.1 M aqueous HCl) by
24b
11c
Nomiya et al. and in MeCN solution by Kholdeeva et al.
Recently, Jimen
́
ez-Lozano et al. have calculated the energy of
D
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 3. DFT optimized structures of the Gly-Gly complexes with 2.
4
−
the reaction 2[PW O Zr(OH)(H O)] → [{α-PW O Zr-
could play a major role in the hydrolytic process. Indeed,
1
1
39
2
11 39
8
−
(
μ-OH)(H O)} ] . With use of PCM-B3LYP (with double-ζ
basis sets and LANL2DZ ECPs for W and Zr) a dimerization
formation of a stable monomer−substrate complex may shift
2
2
48
the equilibrium in favor of the monomer.
−1
47
energy of −23.6 kcal mol was obtained. However, in terms
To further investigate this, the coordination behavior and
binding energy of Gly-Gly at the monomeric species were
investigated more closely. First, the most stable monomeric
form with respect to the coordination number and protonation
of the free energy in the aqueous phase, ΔG *, this
aq
−1
dimerization energy is reduced to −3.4 kcal mol , suggesting
22
that the monomeric species coexists in equilibrium with 1.
Finally, the presence of the monomer in solution at pH 6.4 was
state of H O molecules to Zr was identified. In order to judge
2
3
1
31
22
confirmed by means of P NMR and P DOSY experiments.
Even when it is present in minor concentrations, the monomer
whether a Zr−OH is deprotonated or not at the environ-
2
mental pH the pK value of a H O ligand was estimated (by
a
2
E
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Table 2. Formation Free Energy (kcal mol ) of the Coordination Compounds of Gly-Gly and 2
Article
−1
a
compl
deprot/prot
b
ΔGaqcompl(pH)
species
ΔGaq
pK /pKb
ΔG
species
a
c
2
2
2
2
2
2
a0
−1.1 (1.2)
a
−0.9 (1.6)
3.8
−3.6
→2a-H
→2b0-H
→2b-H
−4.5 (−5.6)
1.6 (0.2)
b0
b
−5.2 (−7.0)
−4.2 (−5.0)
−2.5 (−1.6)
−1.9 (−0.4)
11.3 (>NH)
10.0
6.7 (>NH)
5.0
0.8 (1.8)
c0
c
7.0 (NH2)
−0.8 (−NH2)
→2c+H
→2c-H
−2.8 (−2.2)
4.4 (12.2)
1
1.1
6.4
2
2
d
−1.9 (−0.4)
−2.5 (−1.6)
d-bi
3.7
−5.4
→2d+H
−7.9 (−12.4)
−6.8 (−10.2)
2
d+Hw
a
Coordination compounds formed according to the reaction given in the text (for more details see the Computational Details section in the
b
Supporting Information). Species obtained via deprotonation/protonation of the corresponding compounds in the first column (for more details
see the Computational Details section in the Supporting Information). Energy values in parentheses are relative to the parent compound 1,
c
assuming that both generated monomeric species 2 react with Gly-Gly ligands.
applying eq 9 in the Computational Details section of the
Supporting Information) for the structure 2+H, shown in
Figure 3. It should be noted that the accurate estimation of pKa
values of highly charged species, such as POMs, is a challenging
task by itself, implying more complicated thermodynamic cycles
and approaches. Nevertheless, the approach used here seems
sufficiently reliable for the purpose of this study, as the
NMR measurements (previously performed by us) at this
10b
pH. Coordination in the species labeled a occurs through the
terminal amino and peptide carbonyl groups, in species b
through the terminal amino and carboxyl groups, and in species
c through the carbonyl and carboxyl groups. In the three
species 2a−c, Gly-Gly is coordinated as an anion. These
4
9
structures are obtained from 2 by protonation of Zr−OH by
+
3
calculated pK = 0.6 is in good agreement with reported data
the ligand −NH group and simultaneous ligand exchange
a
50
for a Zr(IV)-bound water molecule (pK ≤ 0.6). Such a pKa
(water release). In structures 2a0−c0 the remaining H O is also
a
2
value enables the deprotonated form 2 to readily exist in neutral
media. Coordination of an additional water molecule, giving
rise to structures 2+Hw and 2+w in Figure 3, is found to be
slightly endergonic for both 2+H and 2, by 1.4 and 1.7 kcal
dissociated. As can be seen from Table 2, all three types of Gly-
Gly coordination are found to be exergonic, releasing an energy
−1
of between 0.9 and 5.2 kcal mol . In the species labeled d, Gly-
−
Gly coordination involves only the substrate −COO group,
−1
mol , respectively. Therefore, structure 2 was used to obtain
either mono- (2d) or bidentate (2d-bi). Here, Gly-Gly is
coordinated as a zwitterion and has replaced a water ligand in
structure 2. Bidentate coordination is only slightly more stable
±
the free energy change upon coordination of Gly-Gly to the
monomeric Zr-POM.
As was done for structure 1, many possible Gly-Gly
complexes of 2 in their different protonation states were
modeled. Only the lowest energy conformations are depicted in
Figure 3. Note that, with the exception of structure 2b0-H
−1
than monodentate coordination, by 0.6 kcal mol , and both
structures are similar in energy to the alternative coordination
modes a−c.
As all three species 2a−c contain a H O ligand,
2
(
where the amide N is deprotonated and acts as the third
deprotonation of this ligand might be favorable, giving rise to
ligand), either mono- or bidentate Gly-Gly coordination is
the species 2a-H−c-H. The corresponding pK values were
a
obtained for all structures. The free energies of complexation
found to be 3.8, 10.0, and 11.1. This means that further
compl
(ΔGaq
) were calculated according to the reaction
stabilization by deprotonation is only expected for species 2a,
−1
±
4−
with a calculated energy gain (at pH 6.4) of −3.6 kcal mol ,
[
Gly‐Gly] + [α‐PW O Zr(OH)(H O)]
1
1
39
2
compl
giving rise to a net formation energy (ΔG
(pH)) of −4.5
aq
4
−
−1
→
[α‐PW O Zr(OH) (H O) (Gly‐Gly)]
kcal mol for complex 2a-H. On the other hand, species 2d
and 2d-bi contain an −OH ligand. The basicity of this −OH
was also estimated. Starting from the most stable 2d-bi
1
1
39
m
2
n
+
(2 − m − n)H O
2
structure, a pK value of 3.7 was calculated, corresponding to a
where, depending on the ligand binding mode, m can be either
(where the Zr−OH ligand has received a proton from the
b
−1
gain in free energy of 5.4 kcal mol for protonation (at pH
.4). Note that protonation gives rise to a complex with a
0
+
6
substrate −NH group) or 1. For m = 0, n = 0−2, while for m
3
monocoordinated carboxylate group, 2d+H. The correspond-
=
1, n = 0, 1. Energy data are summarized in Table 2. The
ing bidentate structure was found to be less stable by 5 kcal
species shown in the left-hand column of Table 2 are formed
through the reaction above, whereas the structures at the right-
hand side are obtained after deprotonation/protonation of the
−1
mol and is not included in Figure 3. Coordination of a second
water molecule to 2d+H does not lead to additional
stabilization of this complex (structure 2d+Hw).
original reaction products. The formation free energies of the
compl
latter were obtained by adjusting the corresponding ΔG
Finally, two other acid−base reactions were found to give rise
to structures with energies similar to those already considered.
The first is structure 2b0-H, obtained from deprotonation of
the amide nitrogen in 2b0. As Figure 3 shows, 2b0-H is a
tridentate complex. However, its formation from 2b0 is
aq
values with the free energy change due to deprotonation/
compl
protonation at the environmental pH, ΔG
(pH) =
aq
compl
deprot/prot
ΔG
+ ΔG
. More details can be found in the
aq
Computational Details section of the Supporting Information.
The pH value of 6.4 was chosen for the binding studies to
enable comparison of the theoretical results with the suggested
endergonic by 6.7 kcal mol− (pK = 11.3). The second
1
a
structure is 2c+H, obtained from 2c by protonation of the
ligand −NH group. With a pK of 7.0, protonation is slightly
13
coordination mode of Gly-Gly to the POM on the basis of C
2
b
F
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 4. DFT optimized structures with two Gly-Gly ligands coordinated to 2.
−1
−1
favored, by 0.8 kcal mol , at pH 6.4. However, one can observe
Table 3. Formation Free Energy (kcal mol ) of the
that in structure 2c+H the Zr−OH bond is broken, and a
Complexes Obtained through Coordination of Two Gly-Gly
Ligands to 2
2
HO−H···O−Zr H bond is formed instead.
In order to compare the results obtained for 2 with those for
the initial compound 1 (Table S2 of the Supporting
Information), the free energy change corresponding to the
net reaction (dimer) + 2 GlyGly →2(monomer complex)
should be used. These values are provided in parentheses in
Table 2. Even though these formation energies must be
compl
/ΔGaqcompl(pH)
species
ΔGaq
a
2
(a0)2
−3.6 (−3.9)
−9.1 (−14.8)
−1.3 (0.7)
−7.6 (−11.8)
−8.2 (−13.2)
0.7 (4.8)
2
2
2
2
a0d
b0d
dc+H
dd-bi
considered more or less qualitatively (e.g., an error of 1 pK
a
−
1
2(d+H)
2
value introduces an error of 1.4 kcal mol
in the
a
compl
Energy values in parentheses are relative to the parent compound 1,
assuming that both generated monomeric species 2 react with two
ΔG
(pH) value), our calculations undoubtedly show that
aq
substrate coordination to the monomeric POM is largely
preferred in comparison to the initial dimer.
Gly-Gly ligands.
1
:1 Zr(IV)-Keggin POM Complexes with Two Gly-Gly
Ligands. In the case of 2 coordination of a second Gly-Gly
ligand to the Zr(IV) center is also possible. To examine this
scenario, a series of complexes were modeled by testing
different binding modes of Gly-Gly to the 2a0, 2b0, 2c+H, and
one or two substrate molecules may coexist in equilibrium.
Among all calculated complexes the lowest energy structure,
compl
−1
2
a0d (ΔG
= −9.1 kcal mol ), contains two Gly-Gly
aq
molecules, one coordinated through the −CO and −NH
2
−
2
d+H species. These species were selected on the basis of their
groups and the other through the −COO group (Figure 5).
Energy contributions to the aqueous free energy of all complex
species can be found in Table S3 of the Supporting
Information.
stability, for every particular binding mode, and the availability
of free coordination sites at the Zr(IV) center (six-coordinated
zirconium ion). The resultant optimized structures and their
formation energies are given in Figure 4 and Table 3
On the basis of the results thus obtained, the species 2a0d
and 2a-H may be considered as the most probable candidates
responsible for the hydrolytic activity of 1. In both complexes
the peptide bond of the bidentate-bound Gly-Gly ligand is
activated toward nucleophilic attack through polarization of the
−CO bond by the Zr(IV) ion. On the one hand 2a0d reveals
the highest stability, while on the other hand 2a-H contains an
intramolecular −OH nucleophile, which is a good premise for
the high hydrolytic ability of this species. These two complexes
were used for the mechanistic investigations in the next section.
respectively. With the exception of 2b0d and 2(d+H) the
2
addition of a second Gly-Gly ligand to structures 2a0, 2b0, 2c
+
H, and 2d+H (with or without H O release) was found to be
2
−1
−1
exergonic by 0.4 kcal mol (2dd-bi) to 8.0 (2a0d) kcal mol .
However, comparison of the formation free energies in Tables
2
and 3 reveals that there is no substantial stabilization effect
upon coordination of a second substrate molecule to 2 instead
of a H O/OH ligand. Thus, the calculations predicted that in
2
the reaction mixture Gly-Gly complexes with 2 involving either
G
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 5. Optimized RC, TS, and INT structures involved in the rate-determining step of Gly-Gly hydrolysis via the 2a0d path. Bond lengths,
−1
dihedral angles, and relative free energies (in italics) are given in Å, deg, and kcal mol , respectively. The energy values colored in cyan indicate the
relative stability of the RC complexes, while the TS and INT energies (in black) are calculated with respect to the energies of the corresponding
starting RC conformations and a free H O molecule. All energy values were obtained with the M06 functional.
2
Mechanism of Gly-Gly Hydrolysis Catalyzed by the
Monomeric Compound 2. Free energy surfaces of the
possible mechanistic scenarios were probed by means of M06-
Scheme 2. Principal Mechanism of Nucleophilic Attack on
the Amide Carbon Atom of Gly-Gly in the 2a0d Complex
2X, M06, and BMK density functionals. The results obtained
with the three functionals are in good qualitative agreement.
Therefore, unless otherwise noted, only the M06 energy values
will be discussed below.
In a recent theoretical investigation on the mechanism of
Gly-Gly hydrolysis catalyzed by a dimeric tetrazirconium(IV)
23
Wells−Dowson POM, we have shown that in the hydrolyti-
cally active complex the substrate is bidentate-coordinated to
one Zr(IV) center, by analogy with 2a0d, and that nucleophilic
attack by a water molecule on the amide C atom is assisted by
species itself encompasses several conformations with similar
stabilities corresponding to different mutual orientations of the
two Gly-Gly ligands and the POM, and the preferential stability
order of these conformations changes with the applied
functional. Moreover, different reactant complexes (RC) are
connected to different TSs for nucleophilic attack on the PES.
To deal with this complex situation, we have considered several
initial conformations, and their associated TSs, instead of just
the one RC conformation with the lowest energy, as predicted
by a certain functional. The number of conformations was
limited by selecting only those whose relative stabilities,
calculated with the M06-2X functional, lie within 2 kcal
mol . In this way the lowest energy 2a0d conformations
predicted by the three functionals are included and examined.
Two RC conformations, G-2a0d (as also predicted by B3LYP
in the binding study) and G-2a0d1, were selected and are
depicted in Figure 5. In the lowest energy TS connected to G-
−
the ligand −COO group. This water addition step was found
23
to be the bottleneck of the whole process of hydrolysis.
A
similar mechanism could be proposed for the 2a0d species,
while in the 2a-H species the intramolecular Zr−OH attack
−
dominates over the −COO -assisted hydrolysis. There are two
−
ligand −COO groups in the 2a0d complex which potentially
could act as a general base: one is Zr-bound, while the other is
free. Since the first is a weaker proton acceptor than the latter,
the hydrolytic mechanism here most likely involves nucleo-
philic attack by a solvent water molecule with proton
−
abstraction by the free −COO group of Gly-Gly, as shown
−1
in Scheme 2. This reaction path will be further referred to as
the “2a0d path”.
Because water addition to the amide C atom is expected to
be the rate-determining step along the 2a0d path, our efforts
were mainly focused on modeling this reaction step. The 2a0d
H
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 6. Optimized RC and TS structures involved in the rate-limiting hydrolytic step of Gly-Ala and Gly-Val when bound to 2 through the 2a0d
−1
binding mode. Bond lengths, dihedral angles, and relative free energies (in italics) are given in Å, deg, and kcal mol , respectively. The energy values
colored in cyan indicate the relative stability of the RC complexes, while the TS energies (in black) are calculated with respect to the energies of the
corresponding starting RC conformation and a free H O molecule. All energy values were obtained with the M06 functional.
2
I
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
a0d on the PES, the attacking water molecule is H-bonded to
Article
−1
2
2) and 2a0d (−9.1 kcal mol , Table 3) and the possible
existence of other species more stable than 2a-H (e.g., 2d+H,
2dc+H, and 2dd-bi), it seems evident that the equilibrium
concentration of 2a-H in the reaction mixture is probably too
low to influence the hydrolysis rate.
a bridging oxygen atom of the POM skeleton while the free
−
ligand −COO group abstracts a H O proton and the nascent
2
OH nucleophile attacks the amide carbon atom. This process
requires 26.5 kcal mol . The TS relaxes to the unstable
intermediate INT with an energy lowering of only 1.3 kcal
mol (Figure 5). On the other hand, in the TS structure
connected to G-2a0d1, the attacking water molecule is H-
bonded to a Zr−O−CO oxygen atom of the neighboring
Gly-Gly ligand. The calculated energy barrier associated with
−1
Effect of the α-Substituent on the POM-Catalyzed
Hydrolysis of Dipeptides with Aliphatic Side Chains. Due
to the similarity of the Gly-X substrates and their rates of
hydrolysis, the findings obtained for Gly-Gly should also be
valid for the rest of dipeptides in the series. To trace the effects
of the aliphatic side chains on the reaction kinetics, the
hydrolysis of Gly-Ala and Gly-Val via the 2a0d path was
examined. Unfortunately, the comparatively higher side chain
flexibility of Gly-Leu, Gly-Ile, and Gly-Phe, in combination with
the variety of different mutual orientations of the two ligands
and the POM in the complexes, requires an extensive DFT
conformational search which substantially increases the
computational cost. Therefore, substrate−catalyst complexes
and TSs with these three ligands were not modeled. It should
be mentioned that the differences among the rate constants in
the Gly-X series are small, corresponding to activation free
−1
−1
this process is 27.2 kcal mol , and a similar TS → INT energy
lowering of 2.3 kcal mol was also found. Hence, on average
−1
the 2a0d path of hydrolysis requires an activation energy of
−1
around 27 kcal mol . It should be noted that the three
23
functionals used here tend to underestimate the INT stability,
which may explain the small TS−INT energy difference.
Additional energy data are provided in Table S4 of the
Supporting Information.
The hydrolytic mechanism originating from the 2a-H
species, further referred to as the 2a-H path, was also
thoroughly examined. The results are presented and discussed
in the Gly-Gly hydrolysis via 2a-H path section (Figures S8−
S10 and Tables S5 and S6) of the Supporting Information. As
expected, intramolecular OH attack on the amide C atom with
formation of a tetrahedral intermediate proceeds with a low
−1
energy differences smaller than 1 kcal mol . This is below the
accuracy of the present computational chemistry approaches.
However, since Gly-Gly, Gly-Ala ,and Gly-Val differ only in the
size of the aliphatic residue and the nature of both RCs and TSs
is the same for the three ligands, the predicted trends should be
reliable due to error cancellation. By analogy with Gly-Gly, only
RC conformations with relative stabilities less than 2 kcal mol−
were considered. For the Gly-Ala complex of 2 the three
density functionals agree that A-2a0d1 (Figure 6), which is
analogous to G-2a0d1, is the most stable conformation. The
corresponding TS is also analogous to that located for Gly-Gly,
−
1
energy barrier, estimated as 18.8 kcal mol . The rate-
determining step along this path, however, is the last reaction
step corresponding to a water-mediated proton transfer from
the −OH group to the amide nitrogen atom with a subsequent
1
−1
substrate decomposition. This process requires 23.8 kcal mol ,
which is lower than the activation energy associated with the
−
1
2
a0d path by about 3 kcal mol .
⧧
⧧
−1
The experimental ΔH and ΔS values for Gly-Gly
but the free energy barrier now amounts to 28.5 kcal mol . In
hydrolysis in the presence of 1 were previously estimated to
the case of Gly-Val three reactant complexes were selected: V-
2a0d1, V-2a0d2, and V-2a0d3 (Figure 6). Complex V-2a0d1 is
analogous to G-2a0d1 and A-2a0d1, while in the complexes V-
be 20.9 kcal mol− and −29.4 cal mol K , giving ΔG (25
1
−1
−1
⧧
−1
10b
°
C) = 28.3 kcal mol . The predicted activation energy for
−1
+
the 2a0d path (27 kcal mol ) is in better agreement with the
2a0d2 and V-2a0d3 a H bond is formed between the −NH
3
−
experiment than that for the 2a-H path, thus suggesting the
and the −COO groups of the two ligands. The calculated free
2
a0d species/mechanism to be responsible for the observed
energy barriers for nucleophilic addition of a water molecule in
peptide hydrolysis. Interestingly, the rate of Gly-Gly hydrolysis
the V-2a0d1, V-2a0d2, and V-2a0d3 forms are 28.4, 32.9, and
−1
−1
in the presence of the dimeric tetrazirconium(IV) Wells−
31.4 kcal mol , respectively (or 31 kcal mol on average).
Thus, on average the calculations predict an increase of the
activation energy (decrease in the rate constants) in the order
Gly-Gly < Gly-Ala < Gly-Val, which is in full agreement with
experiment. Activation energies calculated with the M06-2X
and BMK functionals are collected in Table S7 of the
Supporting Information.
−
7
−1
Dawson POM, k = 2.97 × 10 s , is very similar to that
obs
−
7
−1 23
determined here (4.44 × 10 s ). This POM, however, does
not contain a Zr−OH function(s) and, in contrast to the case
for 1, it is stable over a broad range of experimental
2
3
conditions. In addition, the changes in the rate constants
within the Gly-X series of dipeptides observed with the two
catalysts follow the same trend. These findings strongly suggest
that the mechanisms of Gly-X peptide bond hydrolysis
catalyzed by the two POMs are rather similar, pointing out
that the 2a0d path (which is analogous to that proposed for the
A closer look at the TS structures in Figures 5 and 6 reveals
that in all cases the aliphatic side chains do not hinder the
attacking nucleophile. Indeed, the distance between the H O
2
oxygen and the closest C atom of the side chain is always
greater than 4.4 Å. Thus, the experimentally observed reaction
rate decrease in the Gly-X series cannot be explained in terms
of steric hindrance of the attacking nucleophile. Interestingly,
we found that the decrease in the hydrolysis rate in the
presence and in the absence of 1 follows the same trend. The
uncatalyzed reactions are, however, too slow under these
experimental conditions to determine the corresponding rate
constants (Table 1). To describe the uncatalyzed hydrolysis in
terms of activation energies, the water addition step to the free
Gly-Gly, Gly-Ala, and Gly-Val dipeptides was also modeled (the
mechanism of the uncatalyzed Gly-Gly hydrolysis was
Zr -POM) is indeed the most likely mechanism of hydrolysis.
4
The activation energy difference between 2a0d and 2a-H
−1
paths (1−3 kcal mol with all three functionals; see Tables S4
and S6 of the Supporting Information) implies that hydrolysis
should occur more quickly by a few orders of magnitude along
the 2a-H path. It has been shown that Zr(IV) complexes
delivering an intramolecular OH nucleophile to the peptide
bond are capable of hydrolyzing 90% of Gly-Gly within 20 h, at
51
neutral pH and 60 °C, which is significantly faster than the
reaction rates observed here (with half-lives of more than 2
weeks). Taking into account the calculated difference in the
−1
23
stabilities of the two complexes 2a-H (−4.5 kcal mol , Table
previously reported by us ). In line with experiment the
J
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
calculated activation free energies of Gly-Gly, Gly-Ala, and Gly-
decrease in the rate constants of the POM-catalyzed reactions.
The only disagreement with the experimental trend was found
for Gly-Leu, with a rotational energy barrier lower than that of
Gly-Gly. Obviously, other factors such as weak intramolecular
ligand−ligand and/or ligand−POM interactions, which could
additionally stabilize/destabilize either the initial RC or the
activated TS states, also contribute to the barrier height.
−1
Val are 32.0, 32.6, and 34.1 kcal mol , respectively. Relevant
structures and energy data can be found in Figure S11 and
Table S8 of the Supporting Information. These results suggest
that the trend in the observed rate constants of the POM-
catalyzed reactions might be related to an inherent property of
the substrates. To check whether the α substituent can
influence the electron density distribution of the peptide
bond and potentially alter the susceptibility of the amide carbon
toward nucleophilic attack, a number of molecular descriptors,
e.g. NBO analysis, Hirshfeld charges, and atomic electrostatic
potential, were computed for all dipeptides in the Gly-X series.
The results showed insignificant changes in the peptide bond
properties when changing the aliphatic side chain; see Table S9
of the Supporting Information. A reasonable explanation,
however, can be found on the basis of the substrate
conformational changes in the complexes of Gly-Gly, Gly-Ala,
and Gly-Val upon activation. As can be seen in Figures 5 and 6,
the dihedral angle H−N−C −H (where H is the α-hydrogen
CONCLUSIONS
■
A detailed examination of the effect of the bulkiness of the
amino acid side chain on the hydrolysis rate of a series of
dipeptides with the Gly-X sequence in the presence of a dimeric
Zr(IV)-substituted Keggin POM was performed by means of a
combined theoretical−experimental approach. The observed
rate of peptide bond hydrolysis was found to decrease with an
increase in the side chain bulkiness in the order Gly-Gly > Gly-
Ala > Gly-Phe > Gly-Leu ≈ Gly-Val > Gly-Ile. Experimental
10b
data, reported previously by us, evidenced that a Gly-Gly−
catalyst complex involving coordination of both amide −CO
α
α
α
atom being substituted in the L enantiomers of the
corresponding dipeptides) decreases, in absolute values, from
and terminal −NH groups to the Zr(IV) center of the POM is
2
formed prior to hydrolysis. DFT calculations, presented here,
revealed that coordination of the substrate to the starting
compound 1 is thermodynamically not favored and that Gly-
Gly instead binds to the monomeric POM derivative 2, which is
also present in equilibrium with 1. The formation free energy
associated with the experimentally suggested coordination
−
122 to −105° in the RCs to −60 to −30° in the TS
structures. Substitution of H in Gly-Gly with a bulky aliphatic
α
substituent is expected to increase the steric hindrance between
the amide hydrogen and the α substituent in the rate-limiting
transition state, thus increasing the activation energy for
hydrolysis. To check this assumption, the energy change of
the free Gly-Gly, Gly-Ala, Gly-Val, Gly-Leu, Gly-Ile and Gly-
Phe dipeptides as a function of the H−N−C−C (or H ) angle
was examined by means of the M06-2X functional. The results
are plotted in Figure 7. As can be seen in Figure 7, within the
range −60 to −30° (TS range) the rotational barrier increases
in the order Gly-Gly < Gly-Ala < Gly-Phe < Gly-Val < Gly-Ile,
which is in perfect agreement with the experimentally observed
mode of Gly-Gly to 1 is calculated to be rather endergonic:
1
1
1.9 kcal mol− (at pD 6.4). In contrast, bidentate coordination
α
α
through the −CO and −NH groups to 2 was predicted to
2
−
1
be exergonic by −5.6 kcal mol . The stability of this species is
further increased upon binding of a second Gly-Gly ligand,
−
through the terminal −COO group, as the net complexation
−1
free energy becomes −14.8 kcal mol . The latter is considered
as the major player in the peptide bond cleavage. The
mechanism of hydrolysis involves nucleophilic attack by a
solvent water molecule on the amide carbon atom of the
bidentate-bound ligand. This process is assisted by the free
−
−
COO group of the same ligand, which acts as a general base
to abstract a proton from the attacking water molecule. Since
the uncatalyzed hydrolysis of Gly-X dipeptides is also assisted
by the carboxylic group, the main catalytic role of the POM can
be attributed to the electrophilic activation of the carbonyl
moiety by the Zr(IV) ion. The alternative mechanism involving
intramolecular nucleophilic attack by a Zr−OH group was
ruled out because this mechanism implies hydrolysis rates
significantly higher than those observed in this study.
The decrease of the hydrolysis rate with increase of the side
chain bulkiness is mostly due to the increased ligand
conformational strain in the rate-limiting transition state
which elevates the reaction activation energy. This conforma-
tional strain is imposed by the steric hindrance between the
amide hydrogen atom and the α substituent. It increases first
upon substitution of Hα in Gly-Gly with an aliphatic α
substituent and second with the β branching of the α
substituent. This explanation should also be valid for the
hydrolysis of Gly-X in the presence of a dimeric tetrazirconium-
Figure 7. Energy profile of the H−N−C−C (or H ) dihedral angle
α
α
(
IV) Wells−Dawson POM, where the same trend in the rate
rotation in Gly-Gly, Gly-Ala, Gly-Val, Gly-Leu, Gly-Ile, and Gly-Phe
dipeptides obtained with relaxed PES scans at the CPCM-M06-2X/6-
constants was observed. This work demonstrates the power of a
combined theoretical and experimental study to unravel the
molecular origin of the catalytic activity of metal-substituted
POMs toward organic transformations, especially since
reactivity studies on these types of compounds are often
3
11+G(2df,2p) level of theory. The ranges of calculated dihedral
angles for both RCs and TSs of the Gly-Gly, Gly-Ala, and Gly-Val
complexes with 2 (2a0d binding mode) are indicated as “RC range”
and “TS range”, respectively.
K
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
hampered by the rich solution chemistry of metal-substituted
POMs.
T.; Absillis, G.; Parac-Vogt, T. N. Dalton Trans. 2013, 42, 10929−
1
0938. (c) Ly, H. G. T.; Absillis, G.; Janssens, R.; Proost, P.; Parac-
Vogt, T. N. Angew. Chem., Int. Ed. 2015, 54, 7391−7394.
(11) (a) Kato, C. N.; Negishi, S.; Yoshida, K.; Hayashi, K.; Nomiya,
ASSOCIATED CONTENT
Supporting Information
■
K. Appl. Catal., A 2005, 292, 97−104. (b) Kholdeeva, O. A.;
*
S
Maksimovskaya, R. I. J. Mol. Catal. A: Chem. 2007, 262, 7−24.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.6b01461.
(c) Kholdeeva, O. A.; Maksimov, G. M.; Maksimovskaya, R. I.; Vanina,
M. P.; Trubitsina, T. A.; Naumov, D. Y.; Kolesov, B. A.; Antonova, N.
S.; Carbo, J. J.; Poblet, J. M. Inorg. Chem. 2006, 45, 7224−7234.
1H NMR spectra, additional computational details, a
discussion of Gly-Gly binding to 1, energy contributions
to the aqueous free energies of the Gly-Gly-POM
complexes, a discussion of the 2a-H path of hydrolysis,
energy contributions to the relative free energies of the
stationary points along the reaction paths calculated with
the M06, M06-2X, and BMK functionals, and molecular
structures of relevant species in Cartesian coordinates
(12) (a) Mizuno, N.; Yamaguchi, K.; Kamata, K. Coord. Chem. Rev.
2
005, 249, 1944−1956. (b) Mizuno, N.; Kamata, K.; Yamaguchi, K.
Top. Catal. 2010, 53, 876−893. (c) Kato, C. N.; Hayashi, K.; Negishi,
S.; Nomiya, K. J. Mol. Catal. A: Chem. 2007, 262, 25−29. (d) Estrada,
A. C.; Santos, I. C. M. S.; Simoes, M. M. Q.; Neves, M. G. P. M. S.;
Cavaleiro, J. A. S.; Cavaleiro, A. M. V. Appl. Catal., A 2011, 392, 28−
3
5.
13) Kikukawa, Y.; Yamaguchi, S.; Nakagawa, Y.; Uehara, K.; Uchida,
S.; Yamaguchi, K.; Mizuno, N. J. Am. Chem. Soc. 2008, 130, 15872−
5878.
14) Boglio, C.; Lemiere, G.; Hasenknopf, B.; Thorimbert, S.; Lacote,
E.; Malacria, M. Angew. Chem., Int. Ed. 2006, 45, 3324−3327.
15) Sartorel, A.; Carraro, M.; Scorrano, G.; Bassil, B. S.; Dickman,
M. H.; Keita, B.; Nadjo, L.; Kortz, U.; Bonchio, M. Chem. - Eur. J.
(
(
PDF)
1
(
AUTHOR INFORMATION
Corresponding Author
E-mail for T.T.M.: tzvetan.mihaylov@kuleuven.be.
■
*
(
2
009, 15, 7854−7858.
Notes
(16) Yasuda, H.; He, L. N.; Sakakura, T.; Hu, C. W. J. Catal. 2005,
The authors declare no competing financial interest.
233, 119−122.
(
17) Zhao, S.; Huang, L.; Song, Y.-F. Eur. J. Inorg. Chem. 2013, 2013,
1
(
659−1663.
18) Orlandi, M.; Argazzi, R.; Sartorel, A.; Carraro, M.; Scorrano, G.;
Bonchio, M.; Scandola, F. Chem. Commun. 2010, 46, 3152−3154.
ACKNOWLEDGMENTS
■
T.T.M. thanks the FWO Flanders for financial support under
project G.0260.12. H.G.T.L. thanks the Vietnamese Govern-
ment and KU Leuven for a doctoral fellowship. T.N.P.-V.
thanks KU Leuven for financial support (OT/13/060). The
computational resources and services used in this work were
provided by the VSC (Flemish Supercomputer Center), funded
by the Hercules Foundation and the Flemish Government−
department EWI. The calculations were performed on the HPC
cluster Tier-2 of KU Leuven.
(
9
2
19) (a) Absillis, G.; Parac-Vogt, T. N. Inorg. Chem. 2012, 51, 9902−
910. (b) Vanhaecht, S.; Absillis, G.; Parac-Vogt, T. N. Dalton Trans.
013, 42, 15437−15446. (c) Ly, H. G. T.; Absillis, G.; Parac-Vogt, T.
N. New J. Chem. 2016, 40, 976−984.
(20) (a) Ly, H. G. T.; Absillis, G.; Parac-Vogt, T. N. Eur. J. Inorg.
Chem. 2015, 2015, 2206−2215. (b) Stroobants, K.; Absillis, G.;
Shestakova, P. S.; Willem, R.; Parac-Vogt, T. N. J. Cluster Sci. 2014, 25,
855−866.
(
21) (a) Stroobants, K.; Moelants, E.; Ly, H. G. T.; Proost, P.; Bartik,
K.; Parac-Vogt, T. N. Chem. - Eur. J. 2013, 19, 2848−2858.
b) Stroobants, K.; Goovaerts, V.; Absillis, G.; Bruylants, G.;
Moelants, E.; Proost, P.; Parac-Vogt, T. N. Chem. - Eur. J. 2014, 20,
567−9577. (c) Stroobants, K.; Absillis, G.; Moelants, E.; Proost, P.;
Parac-Vogt, T. N. Chem. - Eur. J. 2014, 20, 3894−3897.
22) Luong, T. K. N.; Shestakova, P.; Mihaylov, T. T.; Absillis, G.;
Pierloot, K.; Parac-Vogt, T. N. Chem. - Eur. J. 2015, 21, 4428−4439.
23) Ly, H. G. T.; Mihaylov, T.; Absillis, G.; Pierloot, K.; Parac-Vogt,
T. N. Inorg. Chem. 2015, 54, 11477−11492.
24) (a) Kato, C. N.; Shinohara, A.; Hayashi, K.; Nomiya, K. Inorg.
REFERENCES
■
(
(
1) (a) Aebersold, R.; Mann, M. Nature 2003, 422, 198−207.
b) Gallop, M. A.; Barrett, R. W.; Dower, W. J.; Fodor, S. P. A.;
Gordon, E. M. J. Med. Chem. 1994, 37, 1233−1251. (c) Cruz-Migoni,
A.; Fuentes-Fernandez, N.; Rabbitts, T. H. MedChemComm 2013, 4,
(
9
(
1
(
6
(
218−1221.
2) Radzicka, A.; Wolfenden, R. J. Am. Chem. Soc. 1996, 118, 6105−
109.
3) (a) Kozhevnikov, I. V. Chem. Rev. 1998, 98, 171−198.
b) Kozhevnikov, I. V. Catalysis by Polyoxometalates; Wiley: Chi-
chester, U.K., 2002; Vol. 2. (c) Izarova, N. V.; Pope, M. T.; Kortz, U.
Angew. Chem., Int. Ed. 2012, 51, 9492−9510. (d) Sartorel, A.; Bonchio,
M.; Campagna, S.; Scandola, F. Chem. Soc. Rev. 2013, 42, 2262−2280.
4) (a) Muller, A.; Peters, F.; Pope, M. T.; Gatteschi, D. Chem. Rev.
998, 98, 239−272. (b) Yamase, T. Chem. Rev. 1998, 98, 307−326.
5) (a) Stephan, H.; Kubeil, M.; Emmerling, F.; Muller, C. E. Eur. J.
(
(
(
Chem. 2006, 45, 8108−8119. (b) Nomiya, K.; Saku, Y.; Yamada, S.;
Takahashi, W.; Sekiya, H.; Shinohara, A.; Ishimaru, M.; Sakai, Y.
Dalton Trans. 2009, 5504−5511.
(
(
(
25) Glasoe, P. K.; Long, F. A. J. Phys. Chem. 1960, 64, 188−190.
(
1
(
26) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73−78.
27) (a) Becke, A. D. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38,
3
098−3100. (b) Perdew, J. P. Phys. Rev. B: Condens. Matter Mater.
Phys. 1986, 33, 8822−8824.
28) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
Inorg. Chem. 2013, 2013, 1585−1594. (b) Rhule, J. T.; Hill, C. L.;
Judd, D. A. Chem. Rev. 1998, 98, 327−358.
(
(
(
(
6) Mizuno, N.; Misono, M. Chem. Rev. 1998, 98, 199−218.
7) Sadakane, M.; Steckhan, E. Chem. Rev. 1998, 98, 219−237.
8) (a) Lv, H.; Geletii, Y. V.; Zhao, C.; Vickers, J. W.; Zhu, G.; Luo,
3297−3305.
(29) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123−141.
Z.; Song, J.; Lian, T.; Musaev, D. G.; Hill, C. L. Chem. Soc. Rev. 2012,
1, 7572−7589. (b) Sumliner, J. M.; Lv, H.; Fielden, J.; Geletii, Y. V.;
Hill, C. L. Eur. J. Inorg. Chem. 2014, 2014, 635−644.
9) (a) Ishimoto, R.; Kamata, K.; Mizuno, N. Angew. Chem., Int. Ed.
012, 51, 4662−4665. (b) Kamata, K.; Yonehara, K.; Sumida, Y.;
Yamaguchi, K.; Hikichi, S.; Mizuno, N. Science 2003, 300, 964−966.
10) (a) Ly, H. G. T.; Absillis, G.; Bajpe, S. R.; Martens, J. A.; Parac-
Vogt, T. N. Eur. J. Inorg. Chem. 2013, 2013, 4601−4611. (b) Ly, H. G.
4
(30) Neese, F. J. Comput. Chem. 2003, 24, 1740−1747.
̈
̈
(31) (a) Eichkorn, K.; Treutler, O.; Ohm, H.; Haser, M.; Ahlrichs, R.
(
2
Chem. Phys. Lett. 1995, 240, 283−290. (b) Eichkorn, K.; Weigend, F.;
Treutler, O.; Ahlrichs, R. Theor. Chem. Acc. 1997, 97, 119−124.
(32) Weigend, F. Phys. Chem. Chem. Phys. 2006, 8, 1057−1065.
(
(33) Klamt, A.; Schu
799−805.
̈
u
̈
rmann, G. J. Chem. Soc., Perkin Trans. 2 1993, 5,
L
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
34) (a) Yan, L.; Lop
Poblet, J. J. Am. Chem. Soc. 2008, 130, 8223−8233. (b) Mal, S. S.;
Troppner, O.; Ivanovic-Burmazovic, I.; Burger, P. Eur. J. Inorg. Chem.
013, 2013, 1960−1967. (c) Lopez, X.; de Graaf, C.; Maestre, J. M.;
Benard, M.; Rohmer, M.-M.; Bo, C.; Poblet, J. M. J. Chem. Theory
Comput. 2005, 1, 856−861. (d) Lopez, X.; Fernandez, J.; Romo, S.;
Paul, J.; Kazansky, L.; Poblet, J. J. Comput. Chem. 2004, 25, 1542−
549. (e) Fernandez, J.; Lopez, X.; Bo, C.; de Graaf, C.; Baerends, E.;
Poblet, J. J. Am. Chem. Soc. 2007, 129, 12244−12253. (f) Fernandez, J.
A.; Lopez, X.; Poblet, J. M. J. Mol. Catal. A: Chem. 2007, 262, 236−
42.
35) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter Mater. Phys.
988, 37, 785−789. (c) Stephens, P. J.; Devlin, F. J.; Chabalowski, C.
F.; Frish, M. J. J. Phys. Chem. 1994, 98, 11623−11627.
36) Neese, F.; Wennmohs, F.; Hansen, A.; Becker, U. Chem. Phys.
009, 356, 98−109.
37) (a) Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput. Chem. 2011,
2, 1456−1465. (b) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J.
Chem. Phys. 2010, 132, 154104−19.
38) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
́
ez, X.; Carbo,
́
J.; Sniatynsky, R.; Duncan, D.;
(51) (a) Kassai, M.; Ravi, R. G.; Shealy, S. J.; Grant, K. B. Inorg.
Chem. 2004, 43, 6130−6132. (b) Kassai, M.; Grant, K. B. Inorg. Chem.
Commun. 2008, 11, 521−525.
̈
́
́
2
́
́
́
́
1
́
́
́
2
(
1
(
2
(
3
(
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Keith, T.; Kobayashi, R.; Normand, J.;
Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.;
Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.;
Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;
Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador,
P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision D.01; Gaussian, Inc., Wallingford, CT, 2013.
(39) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241.
(40) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299−310.
(41) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995−
2
001. (b) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput.
Chem. 2003, 24, 669−681.
42) Boese, A. D.; Martin, J. M. L. J. Chem. Phys. 2004, 121, 3405−
6.
43) (a) Pliego, J. R., Jr.; Riveros, J. M. J. Phys. Chem. A 2001, 105,
241−7247. (b) Bryantsev, V. S.; Diallo, M. S.; Goddard, W. A., III J.
(
1
(
7
Phys. Chem. B 2008, 112, 9709−9719.
(
(
44) Zamyatnin, A. A. Prog. Biophys. Mol. Biol. 1972, 24, 107−123.
45) Constantino, E.; Rimola, A.; Rodríguez-Santiago, L.; Sodupe, M.
New J. Chem. 2005, 29, 1585−1593.
46) (a) Bassil, B. S.; Dickman, M. H.; Kortz, U. Inorg. Chem. 2006,
5, 2394−2396. (b) Fang, X.; Anderson, T. M.; Hill, C. L. Angew.
(
4
Chem., Int. Ed. 2005, 44, 3540−3544. (c) Gaunt, A. J.; May, I.;
Collison, D.; Helliwell, M. Acta Crystallogr., Sect. C: Cryst. Struct.
Commun. 2003, 59, i65−i66. (d) Gaunt, A. J.; May, I.; Collison, D.;
Holman, K. T.; Pope, M. T. J. Mol. Struct. 2003, 656, 101−106.
(
e) Villanneau, R.; Carabineiro, H.; Carrier, X.; Thouvenot, R.;
Herson, P.; Lemos, F.; Ribeiro, F. R.; Che, M. J. Phys. Chem. B 2004,
08, 12465−12471. (f) Carabineiro, H.; Villanneau, R.; Carrier, X.;
Herson, P.; Lemos, F.; Ramoa Ribeiro, F.; Proust, A.; Che, M. Inorg.
Chem. 2006, 45, 1915−1923.
47) Jimenez-Lozano, P.; Carbo,
Rodríguez-Fortea, A.; Wipff, G. Inorg. Chem. 2014, 53, 778−786.
48) Boglio, C.; Hasenknopf, B.; Lenoble, G.; Remy, P.; Gouzerh, P.;
Thorimbert, S.; Lacote, E.; Malacria, M.; Thouvenot, R. Chem. - Eur. J.
008, 14, 1532−1540.
1
̂
(
́
́
J. J.; Chaumont, A.; Poblet, J. M.;
(
́
̂
2
(
(
49) Ho, J.; Coote, M. L. Theor. Chem. Acc. 2010, 125, 3−21.
50) Burgess, J. Metal Ions in Solution; Halsted Press (Wiley): New
York, 1978.
M
DOI: 10.1021/acs.inorgchem.6b01461
Inorg. Chem. XXXX, XXX, XXX−XXX