CONSPECTUS: Enzymes are essential for all living organisms, and their effectiveness as chemical catalysts has driven more than a half century of research seeking to understand the enormous rate enhancements they provide. Nevertheless, a complete understanding of the factors that govern the rate enhancements and selectivities of enzymes remains elusive, due to the extraordinary complexity and cooperativity that are the hallmarks of these biomolecules. We have used a combination of site-directed mutagenesis, pre-steady-state kinetics, X-ray crystallography, nuclear magnetic resonance (NMR), vibrational and fluorescence spectroscopies, resonance energy transfer, and computer simulations to study the implications of conformational motions and electrostatic interactions on enzyme catalysis in the enzyme dihydrofolate reductase (DHFR). We have demonstrated that modest equilibrium conformational changes are functionally related to the hydride transfer reaction. Results obtained for mutant DHFRs illustrated that reductions in hydride transfer rates are correlated with altered conformational motions, and analysis of the evolutionary history of DHFR indicated that mutations appear to have occurred to preserve both the hydride transfer rate and the associated conformational changes. More recent results suggested that differences in local electrostatic environments contribute to finely tuning the substrate pKa in the initial protonation step. Using a combination of primary and solvent kinetic isotope effects, we demonstrated that the reaction mechanism is consistent across a broad pH range, and computer simulations suggested that deprotonation of the active site Tyr100 may play a crucial role in substrate protonation at high pH. Site-specific incorporation of vibrational thiocyanate probes into the ecDHFR active site provided an experimental tool for interrogating these microenvironments and for investigating changes in electrostatics along the DHFR catalytic cycle. Complementary molecular dynamics simulations in conjunction with mixed quantum mechanical/molecular mechanical calculations accurately reproduced the vibrational frequency shifts in these probes and provided atomic-level insight into the residues influencing these changes. Our findings indicate that conformational and electrostatic changes are intimately related and functionally essential. This approach can be readily extended to the study of other enzyme systems to identify more general trends in the relationship between conformational fluctuations and electrostatic interactions. These results are relevant to researchers seeking to design novel enzymes as well as those seeking to develop therapeutic agents that function as enzyme inhibitors.
CONSPECTUS: Enzymes are essential for all living organisms, and their effectiveness as chemical catalysts has driven more than a half century of research seeking to understand the enormous rate enhancements they provide. Nevertheless, a complete understanding of the factors that govern the rate enhancements and selectivities of enzymes remains elusive, due to the extraordinary complexity and cooperativity that are the hallmarks of these biomolecules. We have used a combination of site-directed mutagenesis, pre-steady-state kinetics, X-ray crystallography, nuclear magnetic resonance (NMR), vibrational and fluorescence spectroscopies, resonance energy transfer, and computer simulations to study the implications of conformational motions and electrostatic interactions on enzyme catalysis in the enzyme dihydrofolate reductase (DHFR). We have demonstrated that modest equilibrium conformational changes are functionally related to the hydride transfer reaction. Results obtained for mutant DHFRs illustrated that reductions in hydride transfer rates are correlated with altered conformational motions, and analysis of the evolutionary history of DHFR indicated that mutations appear to have occurred to preserve both the hydride transfer rate and the associated conformational changes. More recent results suggested that differences in local electrostatic environments contribute to finely tuning the substrate pKa in the initial protonation step. Using a combination of primary and solvent kinetic isotope effects, we demonstrated that the reaction mechanism is consistent across a broad pH range, and computer simulations suggested that deprotonation of the active site Tyr100 may play a crucial role in substrate protonation at high pH. Site-specific incorporation of vibrational thiocyanate probes into the ecDHFR active site provided an experimental tool for interrogating these microenvironments and for investigating changes in electrostatics along the DHFR catalytic cycle. Complementary molecular dynamics simulations in conjunction with mixed quantum mechanical/molecular mechanical calculations accurately reproduced the vibrational frequency shifts in these probes and provided atomic-level insight into the residues influencing these changes. Our findings indicate that conformational and electrostatic changes are intimately related and functionally essential. This approach can be readily extended to the study of other enzyme systems to identify more general trends in the relationship between conformational fluctuations and electrostatic interactions. These results are relevant to researchers seeking to design novel enzymes as well as those seeking to develop therapeutic agents that function as enzyme inhibitors.
The
exquisite activities, specificities, and selectivities of naturally
occurring enzymes continue to inspire studies seeking to more precisely
describe the origin of the catalytic prowess of these extraordinary
biomolecules. In essence, enzymes lower the free energy barriers of
chemical reactions (i.e., increase the probabilities of sampling transition
states relative to ground states) by providing electrostatic and geometric
interactions that are complementary to the transition state.[1] This concept of transition state stabilization
generally evokes a static model with singular ground and transition
state structures. However, enzymes today are better understood in
terms of ensembles of configurations that exist on free energy landscapes[2] encompassing innumerable local minima, which
are analogous to the rugged free energy surfaces envisioned for protein
folding.[3] The current understanding of
the factors governing enzyme catalysis has been discussed extensively
in numerous recent reviews.[2,4−6]Electrostatic stabilization of transition states involves
polar and hydrogen bonding groups precisely positioned to stabilize
the migration of charge during the chemical reaction. The partly hydrophobic
character of the interiors of enzymes and the oriented nature of the
enzymes’ functional groups within the protein scaffold decrease
the energetic penalty of environmental reorganization in the protein
relative to bulk solvent.[6,7] The importance of electrostatic
interactions has led to efforts to isolate and quantify the contribution
of electrostatics to enzyme catalysis, as well as to understand the
connection between conformational and electrostatic fluctuations.[6,8−10]Motion and flexibility within enzymes play
important functional roles in substrate binding, product release,
and the formation of catalytically competent configurations.[2] In enzyme catalysis, small equilibrium conformational
changes have been shown to accompany the chemical step in many cases.
These changes are essential to stabilize geometrical variations that
are necessary for the chemical reaction (e.g., to bring a donor and
acceptor closer together or to accommodate a change in hybridization).[11,12] According to this perspective, fast thermal equilibrium fluctuations
on the femtosecond to picosecond time scale lead to overall conformational
changes that occur on the millisecond time scale to reach a state
conducive to the relatively fast formation and cleavage of chemical
bonds. However, there is no convincing evidence for specific special
promoting vibrations[13] on the femtosecond
to picosecond time scale that are directly coupled to the chemical
bond formation and cleavage. The role of protein motions in enzyme
catalysis has been demonstrated in a wide variety of enzymes, including
cyclophilin A,[14] α-amylase,[15] adenylate kinase,[16] and dihydrofolate reductase (DHFR).DHFR is a ubiquitous enzyme
that catalyzes the reduction of 7,8-dihydrofolate (DHF) to 5,6,7,8-tetrahydrofolate
(THF) using reduced nicotinamide adenine dinucleotide phosphate (NADPH)
as a cofactor (Figure 1). DHFR from Escherichia coli (ecDHFR) undergoes a large-scale
(∼8 Å) conformational change in a flexible region (residues
9–24) known as the Met20 loop during the course of its catalytic
cycle (Figure 2).[17,18] In the closed conformation, the Met20 loop interacts with the substrate,
while in the occluded conformation, the loop occupies a portion of
the cofactor binding pocket and extrudes the nicotinamide moiety of
NADP(H) from the active site. Nuclear magnetic resonance (NMR) experiments
show the presence of high-energy conformations in each state that
correspond to low-energy conformations of neighboring states in the
catalytic cycle.[19]
Figure 1
Catalytic mechanism of ecDHFR involving proton transfer from an active site water
molecule to N5 of DHF (blue) and subsequent hydride transfer from
NADPH to C6 of DHF (red).
Figure 2
(A) Ribbon structure of ecDHFR with bound NADP+ (green) and folate (sky blue) illustrating the differences
between the closed (PDB ID 3QL3; red) and occluded (PDB ID 1RX6; blue) conformations of the Met20 loop.
(B) The ecDHFR catalytic cycle distinguishing the
closed conformation (red) and occluded conformation (blue). Adapted
with permission from ref (10). Copyright 2014 American Chemical Society.
Catalytic mechanism of ecDHFR involving proton transfer from an active site water
molecule to N5 of DHF (blue) and subsequent hydride transfer from
NADPH to C6 of DHF (red).(A) Ribbon structure of ecDHFR with bound NADP+ (green) and folate (sky blue) illustrating the differences
between the closed (PDB ID 3QL3; red) and occluded (PDB ID 1RX6; blue) conformations of the Met20 loop.
(B) The ecDHFR catalytic cycle distinguishing the
closed conformation (red) and occluded conformation (blue). Adapted
with permission from ref (10). Copyright 2014 American Chemical Society.Herein, we describe the collaborative experimental
and theoretical research performed in our laboratories on ecDHFR catalysis. The first section summarizes the role
of conformational motions on the hydride transfer reaction and the
impact of mutations on these motions and catalysis. The second section
illustrates the importance of active site residues on finely tuning
the pKa’s within the active site,
in particular, the pKa of the substrate
N5 that is protonated prior to the hydride transfer reaction. The
subsequent sections discuss the incorporation of thiocyanate probes
into ecDHFR to examine local electrostatic microenvironments
within the enzyme and identify electrostatic interactions that directly
influence the hydride transfer reaction itself. Taken together, the
results of these inquiries highlight the importance of electrostatic
interactions and motions in enzyme catalysis.
The role of conformational
changes in enzyme catalysis is a topic of great interest in enzymology,
particularly whether conformational motions are coupled to the chemical
step in enzyme catalysis and the quantitative contribution of motions
to lowering the free energy barrier of the chemical reaction.[20−22] Empirical valence bond (EVB)[23] molecular
dynamics (MD) simulations of ecDHFR catalysis identified
several residues involved in conformational changes and indicated
that these changes occur along the collective reaction coordinate
associated with the hydride transfer reaction.[11,12] Evolutionarily conserved residues including Met42, Tyr100, and Gly121
were observed to undergo modest ≤ 1 Å conformational changes
as the reaction progressed from reactant to transition state. These
conformational changes bring the donor and acceptor closer together
in a favorable orientation and provide an appropriate electrostatic
environment, thereby facilitating the hydride movement that occurs
near the transition state. Mutations of these residues impacted the
catalytic rate constant and altered the conformational changes; furthermore,
double mutations of Met42 and Gly121 displayed nonadditive effects,
indicating cooperativity between these residues. Thus, even relatively
small conformational changes are important to the chemical step in ecDHFR.Single molecule Förster resonance energy
transfer (FRET) experiments suggested that multiple conformations
of ecDHFR exist in solution, that these conformations
interact differently with the bound ligands, and that they exhibit
different hydride transfer rates.[24,25] Furthermore,
in these experiments and more recent equilibrium FRET experiments,[26] some changes in fluorescence were associated
with hydride transfer, while other changes were unrelated to hydride
transfer. For example, FRET experiments indicated that the G121V mutation
reduced the rates of conformational fluctuations, but these rate effects
were quantitatively different from those on the hydride transfer rate,[25] indicating that the conformational motions are
not directly coupled to hydride transfer. The observation that certain
conformational changes identified by FRET can be related to the chemical
step while others are unrelated to the chemical step highlights the
importance of performing experiments designed to distinguish between
these two possibilities.[26]We recently
explored the evolutionary significance of the N23PP mutation from ecDHFR to DHFR from Homo sapiens (hsDHFR),[27] which was previously
shown to ablate millisecond time scale large-scale conformational
changes in the Met20 loop.[20] Using bioinformatics,
we identified several additional mutations between ecDHFR and hsDHFR that were found to be evolutionarily
and functionally significant.[27] Mutation
in ecDHFR of an asparagine to diproline in the flexible
Met20 loop (N23PP) ablates millisecond time scale fluctuations and
reduces the hydride transfer rate,[20] a
three residue insertion near the substrate binding pocket (G51PEKN)
has little effect in WT ecDHFR but recovers the WT
rate in the N23PP background, and a conservative mutation of an aliphatic
to an aromatic hydrophobic residue in the substrate binding pocket
(L28F) enhances the hydride transfer rate of N23PP/G51PEKN to near-mammalian
levels. EVB-MD simulations of the N23PP and N23PP/G51PEKN mutants
suggested that the recovery of the N23PP mutant by the G51PEKN mutation
was related to recovering WT-like conformational changes across the
chemical step (Figure 3). This interpretation
is broadly consistent with the kinetic isotope effect studies of Kohen
and co-workers[28] and the simulation results
of Agarwal and co-workers,[29] while Warshel
and co-workers suggested that the differences between the N23PP mutant
and WT ecDHFR are predominantly electrostatic in
nature.[30] Recognizing the inextricable
connection between conformational and electrostatic changes, these
interpretations may be related to the same phenomena rather than mutually
exclusive.
Figure 3
Conformational changes measured by thermally averaged Cα–Cα distance changes from the reactant state
to the transition state of the hydride transfer reaction observed
in EVB-MD simulations of WT ecDHFR, N23PP ecDHFR, and N23PP/G51PEKN ecDHFR.[27] For clarity, only distances that increase are
shown on the left side of each plot, while only distances that decrease
are shown on the right.
Conformational changes measured by thermally averaged Cα–Cα distance changes from the reactant state
to the transition state of the hydride transfer reaction observed
in EVB-MD simulations of WT ecDHFR, N23PP ecDHFR, and N23PP/G51PEKN ecDHFR.[27] For clarity, only distances that increase are
shown on the left side of each plot, while only distances that decrease
are shown on the right.In light of the evidence for the existence of conformational
changes in enzymes and their connection to enzyme catalysis, the next
step is to determine the source of their contribution to lowering
the free energy barrier to catalysis. When polar and charged residues
are involved in conformational motions, these structural rearrangements
lead to changes in the active site microenvironments that are essential
for providing the thermodynamic stabilization of the reacting species.
As illustrated in subsequent sections, conformational fluctuations
modulate local electrostatic properties in ecDHFR
that further contribute to enzymatic activity.
Conformational
and Electrostatic Influence on Ionization and pKa
Electrostatic influence on the ionization states of catalytic
residues has been demonstrated in many enzymes, especially regarding
the role of conformational flexibility.[31,32] Conformational
and electrostatic changes are intimately coupled because structural
rearrangements of polar groups within enzymes must affect the local
electrostatic microenvironments. Thus, assessing the catalytic contribution
of enzymes based solely on electrostatics while discounting the presence
of thermal fluctuations only provides a static snapshot along the
free energy surface of an enzymatic reaction and fails to capture
the ensemble of varying conformational and electrostatic environments
that exist in solution.The large conformational changes[18,19] exhibited by ecDHFR lead to a broad spectrum of
active site electrostatic environments that can have substantial effects
on the energy landscapes of enzymatic processes. The pKa of the N5 atom of DHF, which is protonated prior to
hydride transfer (Figure 1), varies from ∼2.6
in aqueous solution[33] up to ∼9 in
the enzyme[34] depending on the conformation
of the Met20 loop. This pKa shift is likely
due to changes in the loop conformation that alter the electrostatic
environment around the nearby substrate. Because the hydride transfer
reaction follows DHF protonation, shifts in the substrate pKa determine the concentration of the reactive
species and consequently have substantial effects on the rate of the
overall reaction.Recently, we identified a reactive ecDHFR ternary species that operates under alkaline conditions
(pH > 11), in which one or more titratable residues within the
enzyme’s active site are likely to be deprotonated.[35] Using multiple kinetic isotope effects (KIEs),[36] we showed that the reaction mechanism (N5 protonation
preceding hydride transfer) remains consistent across the pH range
examined (pH 4–12), suggesting elevation of the N5 pKa beyond the previously determined value of
∼9.[34] Using free energy perturbation
MD simulations,[37] we showed that the N5
pKa can be increased by ∼5 units
when the nearby Tyr100 is deprotonated.[35] In other words, a more negatively charged active site environment
with Tyr100 or other residues deprotonated or partially deprotonated,
generated under high pH conditions, could increase the N5 pKa from 6.5 in the ternary reactant complex to
∼11 or higher, thus ensuring sufficient DHF protonation. Moreover,
thermodynamic evaluation indicates that Tyr100 and Asp27 function
synergistically through electrostatic interactions to optimize both
the substrate protonation and the subsequent hydride transfer step.
Fluctuations in Active Site Electrostatics along the Catalytic
Cycle
The importance of electrostatic interactions for enzyme
function necessitates methodologies that are capable of monitoring
local changes in electrostatics in conjunction with catalysis. One
attractive approach is the use of non-natural amino acids containing
thiocyanate (SCN) functional groups with infrared (IR) absorption
spectroscopy.[38] These functional groups
are well-suited for studying electrostatics in enzymes due to several
properties: their absorption spectra are relatively distinct from
that of the enzyme, their stretching modes tend to decouple from other
proximal vibrational modes, and their vibrational frequencies are
sensitive to the local electrostatic environment. Furthermore, these
functional groups generally do not constitute a large perturbation
to the local protein structure, allowing the examination of a virtually
intact and functional enzyme.The application of SCN vibrational
probes to study enzyme electrostatics was extensively explored by
Boxer, Pande, and co-workers.[39] The vibrational
frequency distribution of the probe within the enzyme is often compared
with the vibrational frequency in various solvents and with the Stark
tuning rate determined in the solid state. These data can be interpreted
in terms of the Onsager reaction field or solvatochromic models,[38,40] although the heterogeneous environment of a protein broadly precludes
simple characterization via a single dielectric constant. The incorporation
of thiocyanate probes is especially powerful when combined with X-ray
crystallography or MD simulations of the enzyme that relate the electrostatic
data to atomic-level structural details.[10,39,41,42] Additional
insight into the presence of hydrogen bonding to the nitrile group
can be obtained by comparing the 13C NMR chemical shift
of the nitrilecarbon with the vibrational frequency data.[41]Theoretical methods for calculating electric
fields and IR vibrational shifts in enzymes have been developed using
classical MD[39] and mixed quantum mechanical/molecular
mechanical (QM/MM) approaches.[10,42,43] Classical MD simulations allow the electric field to be quantified
within the context of the atomic point charges in the MM force field.
Moreover, QM/MM calculations for configurations generated by classical
MD enable the calculation of the vibrational frequency shifts of nitrile
probes. Specifically, the C–N potential energy curve and associated
vibrational energy level splittings are calculated for each configuration.
Further decomposition of the electric field based on contributions
from each residue illustrates how various protein components shape
the enzyme’s electrostatic landscape.
Probing
ecDHFR Microenvironments through Vibrational Spectroscopy
The vibrational spectrum of SCN and other environmentally sensitive
vibrational probes can yield quantitative information on the changes
in the local electric field around the probes upon ligand binding,
protein folding, alteration of the surrounding hydration shell, and
protein–protein association.[8,38,39,44,45] To determine how the electrostatic microenvironments of ecDHFR respond to the conformational changes occurring along
its catalytic cycle, we recently incorporated the Cys-CN probe into
four different regions of the enzyme active site (Figure 4; Supporting Information,
Table S1). Despite the proximity of the nitrile probes to the reactive
center, these mutants retained comparable catalytic activity (∼50–110%)
to the WT enzyme, and X-ray crystallography indicated no major structural
rearrangements. The minimal catalytic and structural effects induced
by the SCN probe suggest that these constructs faithfully represent
the electrostatics of the native enzyme.[10]
Figure 4
Thiocyanate
probes in the ecDHFR active site colored according
to their FTIR maximum vibrational frequency (cm–1) for closed (red) and occluded (blue) states along the catalytic
cycle. When present, folate or THF is shown in sky blue and NADP(H)
is shown in green.
Thiocyanate
probes in the ecDHFR active site colored according
to their FTIR maximum vibrational frequency (cm–1) for closed (red) and occluded (blue) states along the catalytic
cycle. When present, folate or THF is shown in sky blue and NADP(H)
is shown in green.After incorporating the
SCN probe into the ecDHFR active site, we examined
the changes in the CN vibrational frequency along the catalytic cycle.
Figure 4 illustrates the SCN probes colored
according to the frequency of maximum FTIR absorption, νmax, detected in each complex.[10] Shifts in this frequency reflect changes in the component of the
electric field along the CN bond of the probe. Crystal structures[18] and MD simulations[10] suggest that the L28C-CN, T46C-CN, and L54C-CN probes retain their
positions and orientations among the five catalytic complexes. This
observation implies that the changes in the electrostatic microenvironments
of these probes among the various complexes in the catalytic cycle
are due to electrostatic changes arising from ligand binding and rearrangements
of the enzyme and solvent. In contrast, the orientation and location
of the M20C-CN probe change significantly along the catalytic cycle,
relative to a fixed point in the protein, due to alterations in the
conformation of the Met20 loop.The T46C-CN probe is situated proximal to the junction between
the two reacting ligands (Figure 4),[10,18] DHF and NADPH, allowing us to monitor the microenvironment close
to the reactive center. The largest frequency shifts observed for
the T46C-CN probe occurred across the two most important steps in
the catalytic cycle: (1) upon the conversion between the model Michaelis–Menten
E:NADP+:FOL complex and the initial E:NADP+:THF
product complex (−4.1 cm–1), which mimics
the hydride transfer step, and (2) across the enzyme turnover step,
which is the release of THF (5.5 cm–1).The
L54C-CN and L28C-CN probes are situated in the folate binding pocket,
located ∼9 Å from the center of the hydride D–A
axis (Figure 4). These two probes are sensitive
to electrostatic changes upon folate binding, as well as the binding/release
of NADP(H) in the cofactor binding pocket. In general, ligand binding
induces substantial changes in the active site electrostatic microenvironments
for all four positions where the SCN probe was incorporated. For the
L54C-CN probe, on one side of the folate binding pocket, the most
significant change occurred upon the formation of the ternary E:NADP+:FOL complex (2.4 cm–1), while for the L28C-CN
probe, on the opposite side of the pocket, the largest changes occurred
during the formation of the Michaelis–Menten complex (−2.9
cm–1) and the enzyme turnover step (2.4 cm–1).[10]The M20C-CN probe is located
directly on the highly fluctuating Met20 loop, and it migrates between
different environments in the closed and occluded conformations. Significant
vibrational frequency variation was observed between all steps of
the catalytic cycle for the M20C-CN probe, with the largest changes
corresponding to steps associated with the transition between the
closed and occluded conformation. The large changes in the electric
field around the M20C-CN probe are predominantly caused by altered
hydrogen-bonding interactions between the nitrile probe and solvent
water molecules. This observation is consistent with the proposal
that interactions with the hydration shell and bulk solvent are important
for conformational changes.[46]
Residue-Based Decomposition of Active Site Microenvironments
Complementary QM/MM simulations of the SCN probes in ecDHFR provided atomic-level information regarding the major contributors
to the detected local electrostatic changes in each probe location.
In particular, we performed QM/MM simulations of the T46C-CN and L54C-CN
probes to examine changes in the electric field across the chemical
step of the ecDHFR catalytic cycle.[10] The calculated νmax of the T46C-CN probe
in the modeled Michaelis–Menten complex is 2165.8 cm–1, which is within 1.8 cm–1 of the measured peak
in the experiments, and the full width at half-maximum (fwhm) calculated
for this probe is 11 cm–1, in good agreement with
the experimental value of 14 cm–1. Furthermore,
the calculated shifts across the hydride transfer step for the T46C-CN
and L54C-CN probes are −3.7 cm–1 and −2.5
cm–1, respectively, compared with the experimentally
measured shifts of −4.1 cm–1 and −1.0
cm–1, respectively. The good agreement between the
theoretical and experimental values provides validation for the simulation
methodology.We also calculated the electric field at the midpoint
of the CN bond due to the MM partial charges in the MD simulations.
The average projected electric fields along the T46C-CN nitrile bond
for the E:NADP+:FOL and E:NADP+:THF complexes
are −15.9 and −19.0 MV/cm, respectively. The component
of the electric field along the CN axis for each probe can be decomposed
into contributions from individual residues, ligands, ions, and solvent
molecules.[10] This procedure allows us to
assess the importance of the conformational changes of each residue
in microenvironment reorganization. For the T46C-CN probe, the NADP+ cofactor and the nearest water molecule are the dominant
contributors to the electric field change between the closed and occluded
conformations across the hydride transfer step, although Asn18 and
Ser49 also make significant contributions. Specifically, the amidenitrogen of Asn18 is ∼1.5 Å closer to the probe in the
closed conformation and the hydroxyl group of Ser49 is ∼1.0
Å closer to the nitrile probe in the closed conformation. Moreover,
following the closed to occluded transition, the CN probe becomes
more accessible to the solvent, altering the hydrogen-bonding interaction
of the probe with the nearby water molecule. More subtle effects,
such as changes in aromaticity and quadrupole moments, may also play
a role.[38] In the case of the local environment
around L54C-CN, we found that Arg57 exhibits the largest change between
the two conformational states. This effect arises because the positively
charged region of Arg57 moves away from its orthogonal orientation
with respect to the probe in the occluded conformation.
Role of Electrostatics in the Catalyzed Hydride Transfer Reaction
On the basis of the experimental validation, we used the same simulation
methodology to evaluate the component of the electric field along
the hydride D–A axis in the model Michaelis–Menten complex
and to decompose this field into contributions from each residue within
the enzyme (Figure 5). These calculations yielded
valuable information about the reactive center, which is more directly
related to catalysis than analogous information about the SCN probes
within the active site. Moreover, such information about the reactive
center is experimentally inaccessible. We found that the component
of the total electric field along the D–A axis in the model
ternary reactant complex (E:FOL:NADP+) is −48.9
MV/cm. The negative value indicates that this component of the electric
field is highly favorable for transferring a negatively charged hydride
from the cofactor to the substrate, and this magnitude could potentially
reduce the free energy barrier by several kilocalories per mole.[39] Electric field decomposition showed that the
substrate and cofactor contribute −32.4 MV/cm to the total
field along the D–A axis. The protein contributes −15.4
MV/cm, which is ∼33% of the total field, while the solvent
molecules and ions contribute very little (−1.1 MV/cm) to the
total field. The quantitative contributions from the ligands to the
electric field in the model Michaelis complex may differ from those
of the substrate and cofactor during the chemical reaction. Future
computational studies will examine the electric field during the chemical
reaction catalyzed by DHFR.
Figure 5
Ribbon structure of ecDHFR
colored according to residue-based contributions to the calculated
electric field along the hydride transfer D–A axis (depicted
as an orange arrow) for selected residues from MD simulations of the
model Michaelis complex E:FOL:NADP+. Positive values of
the electric field disfavor hydride transfer, while negative values
facilitate hydride transfer. NADP+ is colored green, and
folate is colored sky blue.
Ribbon structure of ecDHFR
colored according to residue-based contributions to the calculated
electric field along the hydride transfer D–A axis (depicted
as an orange arrow) for selected residues from MD simulations of the
model Michaelis complex E:FOL:NADP+. Positive values of
the electric field disfavor hydride transfer, while negative values
facilitate hydride transfer. NADP+ is colored green, and
folate is colored sky blue.Further residue-based decomposition of the protein electric
field (−15.4 MV/cm) showed that there is a strong degree of
overlap between the residues that contribute significantly to the
field along the D–A axis and the network of coupled motions
that has been identified for DHFR catalysis.[10,12] This network of coupled motions corresponds to equilibrium conformational
changes that occur as the reaction evolves from the reactant to the
transition state along the collective reaction coordinate associated
with hydride transfer. Such motions facilitate the chemical reaction
by bringing the substrate and cofactor closer together in a favorable
orientation and providing a suitable electrostatic environment for
hydride transfer. The residues Ile14, Tyr100, and Asp122 contribute
−1.2, −4.2, and −2.4 MV/cm, respectively, and
they were all implicated in the network of coupled motions.[12] Additional residues that significantly contribute
to the electric field along the D–A axis in a manner that facilitates
hydride transfer are Lys32 (−3.2 MV/cm), Arg52 (−3.9
MV/cm), and Arg57 (−4.2 MV/cm). The residues that contribute
to the field along the D–A axis in a manner that disfavors
hydride transfer include Asp27 (5.4 MV/cm), Arg98 (3.4 MV/cm), and
His124 (3.0 MV/cm). All of these residues represent potential targets
to alter the electrostatic contribution to catalysis. Interestingly
(see section 3), two of the major contributors
(Asp27 and Tyr100) were found to be important catalytic residues that
operate synergistically to affect various aspects of the enzymatic
reaction, such as the pKa of N5 on DHF,
proton transfer from a solvent molecule to DHF, and hydride transfer
from NADPH, via electrostatic interactions.[35]
Conclusion, Implications, and Future Prospects
The results from recent studies of DHFR are consistent with the view
that a series of rearrangements brings the substrates into a conformation
that realizes the highly favorable electrostatic field for facilitating
the hydride transfer reaction.[2,4,10] Protein motions are particularly important in enzymes with a high
degree of conformational flexibility due to a greater number of accessible
structural and electrostatic states. An integrated computational–experimental
approach provides a very powerful tool for probing the detailed active
site microenvironments, active site hydration states, and electrostatic
contributions along the catalytic cycle.With scientific disciplines
presently shifting in ways that favor the translational value of research,
fundamental enzymology queries into enzyme function and catalysis
still present important potential for discovery. Enzymes are highly
flexible entities, especially under biological conditions with significant
thermal energy available, and it is logical that conformational fluctuations
will generate an ensemble of species covering a wide range of electrostatic
variation. Since the electrostatic environment can have a dramatic
effect on enzyme catalysis, a more dynamic perspective must be adopted
when considering the source of catalytic efficacy. This approach could
be helpful in guiding the development of novel small molecules or
peptides for inhibiting protein activity[47,48] as well as the design of catalysts with activities comparable to
those of naturally occurring enzymes.[49]
Authors: Hans Frauenfelder; Guo Chen; Joel Berendzen; Paul W Fenimore; Helén Jansson; Benjamin H McMahon; Izabela R Stroe; Jan Swenson; Robert D Young Journal: Proc Natl Acad Sci U S A Date: 2009-02-27 Impact factor: 11.205
Authors: Gira Bhabha; Jeeyeon Lee; Damian C Ekiert; Jongsik Gam; Ian A Wilson; H Jane Dyson; Stephen J Benkovic; Peter E Wright Journal: Science Date: 2011-04-08 Impact factor: 47.728
Authors: Zhiquan Zhang; P T Ravi Rajagopalan; Tzvia Selzer; Stephen J Benkovic; Gordon G Hammes Journal: Proc Natl Acad Sci U S A Date: 2004-02-20 Impact factor: 11.205
Authors: C Tony Liu; Joshua P Layfield; Robert J Stewart; Jarrod B French; Philip Hanoian; John B Asbury; Sharon Hammes-Schiffer; Stephen J Benkovic Journal: J Am Chem Soc Date: 2014-07-11 Impact factor: 15.419
Authors: Donald Gagné; Chitra Narayanan; Nhung Nguyen-Thi; Louise D Roux; David N Bernard; Joseph S Brunzelle; Jean-François Couture; Pratul K Agarwal; Nicolas Doucet Journal: Biochemistry Date: 2016-07-21 Impact factor: 3.162