Aromatic amine dehydrogenase (AADH) and related enzymes are at the heart of debates on the roles of quantum tunneling and protein dynamics in catalysis. The reaction of tryptamine in AADH involves significant quantum tunneling in the rate-limiting proton transfer step, shown by large H/D primary kinetic isotope effects (KIEs), with unusual temperature dependence. We apply correlated ab initio combined quantum mechanics/molecular mechanics (QM/MM) methods, at levels up to local coupled cluster theory (LCCSD(T)/(aug)-cc-pVTZ), to calculate accurate potential energy surfaces for this reaction, which are necessary for quantitative analysis of tunneling contributions and reaction dynamics. Different levels of QM/MM treatment are tested. Multiple pathways are calculated with fully flexible transition state optimization by the climbing-image nudged elastic band method at the density functional QM/MM level. The average LCCSD(T) potential energy barriers to proton transfer are 16.7 and 14.0 kcal/mol for proton transfer to the two carboxylate atoms of the catalytic base, Asp128β. The results show that two similar, but distinct pathways are energetically accessible. These two pathways have different barriers, exothermicity and curvature, and should be considered in analyses of the temperature dependence of reaction and KIEs in AADH and other enzymes. These results provide a benchmark for this prototypical enzyme reaction and will be useful for developing empirical models, and analyzing experimental data, to distinguish between different conceptual models of enzyme catalysis.

Cdc25 phosphatase B, a potential target for cancer therapy, is inhibited by a series of quinones. The binding site and mode of quinone inhibitors to Cdc25B remains unclear, whereas this information is important for structure-based drug design. We investigated the potential binding site of NSC663284 [DA3003-1 or 6-chloro-7-(2-morpholin-4-yl-ethylamino)-quinoline-5, 8-dione] through docking and molecular dynamics simulations. Of the two main binding sites suggested by docking, the molecular dynamics simulations only support one site for stable binding of the inhibitor. Binding sites in and near the Cdc25B catalytic site that have been suggested previously do not lead to stable binding in 50 ns molecular dynamics (MD) simulations. In contrast, a shallow pocket between the C-terminal helix and the catalytic site provides a favourable binding site that shows high stability. Two similar binding modes featuring protein-inhibitor interactions involving Tyr428, Arg482, Thr547 and Ser549 are identified by clustering analysis of all stable MD trajectories. The relatively flexible C-terminal region of Cdc25B contributes to inhibitor binding. The binding mode of NSC663284, identified through MD simulation, likely prevents the binding of protein substrates to Cdc25B. The present results provide useful information for the design of quinone inhibitors and their mechanism of inhibition.

Projector-based embedding has recently emerged as a robust multiscale method for the calculation of various electronic molecular properties. We present the coupling of projector embedding with quantum mechanics/molecular mechanics modeling and apply it for the first time to an enzyme-catalyzed reaction. Using projector-based embedding, we combine coupled-cluster theory, density-functional theory (DFT), and molecular mechanics to compute energies for the proton abstraction from acetyl-coenzyme A by citrate synthase. By embedding correlated ab initio methods in DFT we eliminate functional sensitivity and obtain high-accuracy profiles in a procedure that is straightforward to apply.

The mechanism of cytochrome P450(CYP)-catalyzed hydroxylation of primary amines is currently unclear and is relevant to drug metabolism; previous small model calculations have suggested two possible mechanisms: direct N-oxidation and H-abstraction/rebound. We have modeled the N-hydroxylation of (R)-mexiletine in CYP1A2 with hybrid quantum mechanics/molecular mechanics (QM/MM) methods, providing a more detailed and realistic model. Multiple reaction barriers have been calculated at the QM(B3LYP-D)/MM(CHARMM27) level for the direct N-oxidation and H-abstraction/rebound mechanisms. Our calculated barriers indicate that the direct N-oxidation mechanism is preferred and proceeds via the doublet spin state of Compound I. Molecular dynamics simulations indicate that the presence of an ordered water molecule in the active site assists in the binding of mexiletine in the active site, but this is not a prerequisite for reaction via either mechanism. Several active site residues play a role in the binding of mexiletine in the active site, including Thr124 and Phe226. This work reveals key details of the N-hydroxylation of mexiletine and further demonstrates that mechanistic studies using QM/MM methods are useful for understanding drug metabolism.

Thymidylate synthase (TS) is a promising cancer target, due to its crucial function in thymine synthesis. It performs the reductive methylation of 2′-deoxyuridine-5′-phosphate (dUMP) to thymidine-5′-phosphate (dTMP), using N-5,10-methylene-5,6,7,8-tetrahydrofolate (mTHF) as a cofactor. After the formation of the dUMP/mTHF/TS noncovalent complex, and subsequent conformational activation, this complex has been proposed to react via nucleophilic attack (Michael addition) by Cys146, followed by methylene-bridge formation to generate the ternary covalent intermediate. Herein, QM/MM (B3LYP-D/6-31+G(d)-CHARMM27) methods are used to model the formation of the ternary covalent intermediate. A two-dimensional potential energy surface reveals that the methylene-bridged intermediate is formed via a concerted mechanism, as indicated by a single transition state on the minimum energy pathway and the absence of a stable enolate intermediate. A range of different QM methods (B3LYP, MP2 and SCS-MP2, and different basis sets) are tested for the calculation of the activation energy barrier for the formation of the methylene-bridged intermediate. We test convergence of the QM/MM results with respect to size of the QM region. Inclusion of Arg166, which interacts with the nucleophilic thiolate, in the QM region is important for reliable results; the MM model apparently does not reproduce energies for distortion of the guanidinium side chain correctly. The spin component scaled-Møller–Plessett perturbation theory (SCS-MP2) approach was shown to be in best agreement (within 1.1 kcal/mol) while the results obtained with MP2 and B3LYP also yielded acceptable values (deviating by less than 3 kcal/mol) compared with the barrier derived from experiment. Our results indicate that using a dispersion-corrected DFT method, or a QM method with an accurate treatment of electron correlation, increases the agreement between the calculated and experimental activation energy barriers, compared with the semiempirical AM1 method. These calculations provide important insight into the reaction mechanism of TS and may be useful in the design of new TS inhibitors.

The applicability of combined quantum mechanics/molecular mechanics (QM/MM) methods for the calculation of absolute binding free energies of conserved water molecules in protein/ligand complexes is demonstrated. Here, we apply QM/MM Monte Carlo simulations to investigate binding of water molecules to influenza neuraminidase. We investigate five different complexes, including those with the drugs oseltamivir and peramivir. We investigate water molecules in two different environments, one more hydrophobic and one hydrophilic. We calculate the free-energy change for perturbation of a QM to MM representation of the bound water molecule. The calculations are performed at the BLYP/aVDZ (QM) and TIP4P (MM) levels of theory, which we have previously demonstrated to be consistent with one another for QM/MM modeling. The results show that the QM to MM perturbation is significant in both environments (greater than 1 kcal mol–1) and larger in the more hydrophilic site. Comparison with the same perturbation in bulk water shows that this makes a contribution to binding. The results quantify how electronic polarization differences in different environments affect binding affinity and also demonstrate that extensive, converged QM/MM free-energy simulations, with good levels of QM theory, are now practical for protein/ligand complexes.

Combined quantum mechanics/molecular mechanics (QM/MM) simulations of the reaction catalysed by phosphite dehydrogenase (PTDH) identify Met53 as important for catalysis. This catalytic role is verified by experiments (including replacement by norleucine and selenomethionine), which show that mutation of this residue significantly affects kcat, without changing KM for phosphite. QM/MM and ab initio QM calculations show that the catalytic effect is electrostatic in nature. The side chain of Met53 specifically stabilizes the transition state for the hydride transfer step of the reaction catalysed by PTDH, forming a ‘face-on’ interaction with His292. To our knowledge, a defined catalytic role for methionine in an enzyme (as opposed to a steric or binding effect, or interaction with a metal ion) has not previously been identified. Analyses of the Protein Data Bank and Cambridge Structural Database indicate that this type of interaction may be relatively widespread, with implications for enzyme-catalysed reaction mechanisms and protein structure.

Carbapenems, ‘last resort’ antibiotics for many bacterial infections, can now be broken down by several class A β-lactamases (i.e. carbapenemases). Here, carbapenemase activity is predicted through QM/MM dynamics simulations of acyl–enzyme deacylation, requiring only the 3D structure of the apo-enzyme. This may assist in anticipating resistance and future antibiotic design.

N-Acetylneuraminic acid lyase (NAL) is a Class I aldolase that catalyzes the reversible condensation of pyruvate with N-acetyl-d-mannosamine (ManNAc) to yield the sialic acid N-acetylneuraminic acid (Neu5Ac). Aldolases are finding increasing use as biocatalysts for the stereospecific synthesis of complex molecules. Incomplete understanding of the mechanism of catalysis in aldolases, however, can hamper development of new enzyme activities and specificities, including control over newly generated stereocenters. In the case of NAL, it is clear that the enzyme catalyzes a Bi-Uni ordered condensation reaction in which pyruvate binds first to the enzyme to form a catalytically important Schiff base. The identity of the residues required for catalysis of the condensation step and the nature of the transition state for this reaction, however, have been a matter of conjecture. In order to address, this we crystallized a Y137A variant of the E. coli NAL in the presence of Neu5Ac. The three-dimensional structure shows a full length sialic acid bound in the active site of subunits A, B, and D, while in subunit C, discontinuous electron density reveals the positions of enzyme-bound pyruvate and ManNAc. These ‘snapshot’ structures, representative of intermediates in the enzyme catalytic cycle, provided an ideal starting point for QM/MM modeling of the enzymic reaction of carbon–carbon bond formation. This revealed that Tyr137 acts as the proton donor to the aldehyde oxygen of ManNAc during the reaction, the activation barrier is dominated by carbon–carbon bond formation, and proton transfer from Tyr137 is required to obtain a stable Neu5Ac-Lys165 Schiff base complex. The results also suggested that a triad of residues, Tyr137, Ser47, and Tyr110 from a neighboring subunit, are required to correctly position Tyr137 for its function, and this was confirmed by site-directed mutagenesis. This understanding of the mechanism and geometry of the transition states along the C–C bond-forming pathway will allow further development of these enzymes for stereospecific synthesis of new enzyme products.

We have analyzed the capability of state-of-the-art multiscale computational approaches to provide atomic-resolution electronic structure insights into possible catalytic scenarios of the hairpin ribozyme by evaluating potential and free energy surfaces of the reactions by various hybrid QM/MM methods. The hairpin ribozyme is a unique catalytic RNA that achieves rate acceleration similar to other small self-cleaving ribozymes but without direct metal ion participation. Guanine 8 (G8) and adenine 38 (A38) have been identified as the catalytically essential nucleobases. However, their exact catalytic roles are still being investigated. In line with the available experimental data, we considered two reaction scenarios involving protonated A38H+ as a general acid which is further assisted by either canonical G8 or deprotonated G8– forms. We used the spin-component scaled Møller–Plesset (SCS-MP2) method at the complete basis set limit as the reference method. The semiempirical AM1/d-PhoT and SCC-DFTBPR methods provided acceptable activation barriers with respect to the SCS-MP2 data but predicted significantly different reaction pathways. DFT functionals (BLYP and MPW1K) yielded the same reaction pathway as the SCS-MP2 method. The activation barriers were slightly underestimated by the GGA BLYP functional, although with accuracy comparable to the semiempirical methods. The SCS-MP2 method and hybrid MPW1K functional gave activation barriers that were closest to those derived from experimentally measured rate constants.

Adenosylcobalamin (AdoCbl) serves as a source of reactive free radicals that are generated by homolytic scission of the coenzyme’s cobalt–carbon bond. AdoCbl-dependent enzymes accelerate AdoCbl homolysis by ∼1012-fold, but the mechanism by which this is accomplished remains unclear. We have combined experimental and computational approaches to gain molecular-level insight into this process for glutamate mutase. Two residues, glutamate 330 and lysine 326, form hydrogen bonds with the adenosyl group of the coenzyme. A series of mutations that impair the enzyme’s ability to catalyze coenzyme homolysis and tritium exchange with the substrate by 2–4 orders of magnitude were introduced at these positions. These mutations, together with the wild-type enzyme, were also characterized in silico by molecular dynamics simulations of the enzyme–AdoCbl–substrate complex with AdoCbl modeled in the associated (Co–C bond formed) or dissociated [adenosyl radical with cob(II)alamin] state. The simulations reveal that the number of hydrogen bonds between the adenosyl group and the protein side chains increases in the homolytically dissociated state, with respect to the associated state, for both the wild-type and mutant enzymes. The mutations also cause a progressive increase in the mean distance between the 5′-carbon of the adenosyl radical and the abstractable hydrogen of the substrate. Interestingly, the distance between the 5′-carbon and substrate hydrogen, determined computationally, was found to inversely correlate with the log k for tritium exchange (r = 0.93) determined experimentally. Taken together, these results point to a dual role for these residues: they both stabilize the homolytic state through electrostatic interactions between the protein and the dissociated coenzyme and correctly position the adenosyl radical to facilitate the abstraction of hydrogen from the substrate.

Cytochrome P450 enzymes (P450s) are important in drug metabolism and have been linked to adverse drug reactions. P450s display broad substrate reactivity, and prediction of metabolites is complex. QM/MM studies of P450 reactivity have provided insight into important details of the reaction mechanisms and have the potential to make predictions of metabolite formation. Here we present a comprehensive study of the oxidation of three widely used pharmaceutical compounds (S-ibuprofen, diclofenac, and S-warfarin) by one of the major drug-metabolizing P450 isoforms, CYP2C9. The reaction barriers to substrate oxidation by the iron-oxo species (Compound I) have been calculated at the B3LYP-D/CHARMM27 level for different possible metabolism sites for each drug, on multiple pathways. In the cases of ibuprofen and warfarin, the process with the lowest activation energy is consistent with the experimentally preferred metabolite. For diclofenac, the pathway leading to the experimentally observed metabolite is not the one with the lowest activation energy. This apparent inconsistency with experiment might be explained by the two very different binding modes involved in oxidation at the two competing positions. The carboxylate of diclofenac interacts strongly with the CYP2C9 Arg108 side chain in the transition state for formation of the observed metabolite—but not in that for the competing pathway. We compare reaction barriers calculated both in the presence and in the absence of the protein and observe a marked improvement in selectivity prediction ability upon inclusion of the protein for all of the substrates studied. The barriers calculated with the protein are generally higher than those calculated in the gas phase. This suggests that active-site residues surrounding the substrate play an important role in controlling selectivity in CYP2C9. The results show that inclusion of sampling (particularly) and dispersion effects is important in making accurate predictions of drug metabolism selectivity of P450s using QM/MM methods.

Carbamate and urea derivatives are important classes of fatty acid amide hydrolase (FAAH) inhibitors that carbamoylate the active-site nucleophile Ser241. In the present work, the reactivation mechanism of carbamoylated FAAH is investigated by means of a quantum mechanics/molecular mechanics (QM/MM) approach. The potential energy surfaces for decarbamoylation of FAAH covalent adducts, derived from the O-aryl carbamate URB597 and from the N-piperazinylurea JNJ1661610, were calculated and compared to that for deacylation of FAAH acylated by the substrate oleamide. Calculations show that a carbamic group bound to Ser241 prevents efficient stabilization of transition states of hydrolysis, leading to large increments in the activation barrier. Moreover, the energy barrier for the piperazine carboxylate was significantly lower than that for the cyclohexyl carbamate derived from URB597. This is consistent with experimental data showing slowly reversible FAAH inhibition for the N-piperazinylurea inhibitor and irreversible inhibition for URB597.

Computational enzymology is a rapidly maturing field that is increasingly integral to understanding mechanisms of enzyme-catalyzed reactions and their practical applications. Combined quantum mechanics/molecular mechanics (QM/MM) methods are important in this field. By treating the reacting species with a quantum mechanical method (i.e., a method that calculates the electronic structure of the active site) and including the enzyme environment with simpler molecular mechanical methods, enzyme reactions can be modeled. Here, we review QM/MM methods and their application to enzyme-catalyzed reactions to investigate fundamental and practical problems in enzymology. A range of QM/MM methods is available, from cheaper and more approximate methods, which can be used for molecular dynamics simulations, to highly accurate electronic structure methods. We discuss how modeling of reactions using such methods can provide detailed insight into enzyme mechanisms and illustrate this by reviewing some recent applications. We outline some practical considerations for such simulations. Further, we highlight applications that show how QM/MM methods can contribute to the practical development and application of enzymology, e.g., in the interpretation and prediction of the effects of mutagenesis and in drug and catalyst design.

The emergence of influenza drug resistance is a major public health concern. The molecular basis of resistance to oseltamivir (Tamiflu) is investigated using a computational assay involving multiple 500 ns unrestrained molecular dynamics (MD) simulations of oseltamivir complexed with mutants of H1N1-2009 influenza neuraminidase. The simulations, accelerated using graphics processors (GPUs), and using a fully explicit model of water, are of sufficient length to observe multiple drug unbinding and rebinding events. Drug unbinding occurs during simulations of known oseltamivir-resistant mutants of neuraminidase. Molecular-level rationalizations of drug resistance are revealed by analysis of these unbinding trajectories, with particular emphasis on the dynamics of the mutant residues. The results indicate that MD simulations can predict weakening of binding associated with drug resistance. In addition, visualization and analysis of binding site water molecules reveal their importance in stabilizing the binding mode of the drug. Drug unbinding is accompanied by conformational changes, driven by the mutant residues, which results in flooding of a key pocket containing tightly bound water molecules. This displaces oseltamivir, allowing the tightly bound water molecules to be released into bulk. In addition to the role of water, analysis of the trajectories reveals novel behavior of the structurally important 150-loop. Motion of the loop, which can move between an open and closed conformation, is intimately associated with drug unbinding and rebinding. Opening of the loop occurs coincidentally with drug unbinding, and interactions between oseltamivir and the loop seem to aid in the repositioning of the drug back into an approximation of its original binding mode on rebinding. The similarity of oseltamivir to a transition state analogue for neuraminidase suggests that the dynamics of the loop could play an important functional role in the enzyme, with loop closing aiding in binding of the substrate and loop opening aiding the release of the product.

Terpene synthases are potentially useful biocatalysts for the synthesis of valuable compounds, such as anticancer drugs and antibiotics. The design of altered activities requires better knowledge of their mechanisms, for example, an understanding of the complex conformational changes that are part of their catalytic cycle, how they are coordinated, and what drives them. Crystallographic studies of the sesquiterpene synthase artistolochene synthase have led to a proposed sequence of ligand binding and conformational change but have provided only indirect insight. Here, we have performed extensive molecular dynamics simulations of multiple enzyme–ligand complexes (over 2 μs in total). The simulations provide clear evidence of what drives the conformational changes required for reaction. They support a picture in which the substrate farnesyl diphosphate binds first, followed by three magnesium ions in sequence, and, after reaction, the release of aristolochene and two magnesium ions followed by the final magnesium ion and diphosphate. Binding of farnesyl diphosphate leads to an increased level of sampling of open conformations, allowing the first two magnesium ions to bind. The closed enzyme conformation is maintained with a diphosphate moiety and two magnesium ions bound. The open-to-closed transition reduces flexibility around the active site entrance, partly through a lid closing over it. The simulations with all three magnesium ions and farnesyl diphosphate bound provide, for the first time, a realistic model of the Michaelis complex involved in reaction, which is inaccessible to experimental structural studies. These insights could help with the design of altered activities in a range of terpene synthases.

Fatty acid amide hydrolase (FAAH) is a member of the amidase signature family and is responsible for the hydrolytic deactivation of fatty acid amide neuromodulators, such as anandamide. FAAH carries an unusual catalytic triad consisting of Lys-Ser-Ser, which uniquely enables the enzyme to cleave amides and esters at similar rates. The acylation of 9Z-octadecenamide (oleamide, a FAAH reference substrate) has been widely investigated by computational methods, and those have shown that conformational fluctuations of the active site affect the reaction barrier. Empirical descriptors have been devised to provide a possible mechanistic explanation for such conformational effects, but a first-principles understanding is still missing. A comparison of FAAH acylation with a reference reaction in water suggests that transition-state stabilization is crucial for catalysis because the activation energy barrier falls by 6 kcal/mol in the presence of the active site. With this in mind, we have analyzed the enzymatic reaction using the differential transition-state stabilization (DTSS) approach to determine key active-site residues for lowering the barrier. We examined several QM/MM structures at the MP2 level of theory and analyzed catalytic effects with a variation–perturbation partitioning of the interaction energy into electrostatic multipole and penetration, exchange, delocalization, and correlation terms. Three residues – Thr236, Ser218, and one water molecule – appear to be essential for the stabilization of the transition state, a conclusion that is also reflected by catalytic fields and agrees with site-directed mutagenesis data. An analogous analysis for URB524, URB618, and URB694 (three potent representatives of covalent, carbamate-based FAAH inhibitors) confirms the importance of the residues involved in oleamide acylation, providing insight for future inhibitor design.

Carbapenems are the most potent β-lactam antibiotics and key drugs for treating infections by Gram-negative bacteria. In such organisms, β-lactam resistance arises principally from β-lactamase production. Although carbapenems escape the activity of most β-lactamases, due in the class A enzymes to slow deacylation of the covalent acylenzyme intermediate, carbapenem-hydrolyzing class A β-lactamases are now disseminating in clinically relevant bacteria. The reasons why carbapenems are substrates for these enzymes, but inhibit other class A β-lactamases, remain to be fully established. Here, we present crystal structures of the class A carbapenemase SFC-1 from Serratia fonticola and of complexes of its Ser70 Ala (Michaelis) and Glu166 Ala (acylenzyme) mutants with the carbapenem meropenem. These are the first crystal structures of carbapenem complexes of a class A carbapenemase. Our data reveal that, in the SFC-1 acylenzyme complex, the meropenem 6α-1R-hydroxyethyl group interacts with Asn132, but not with the deacylating water molecule. Molecular dynamics simulations indicate that this mode of binding occurs in both the Michaelis and acylenzyme complexes of wild-type SFC-1. In carbapenem-inhibited class A β-lactamases, it is proposed that the deacylating water molecule is deactivated by interaction with the carbapenem 6α-1R-hydroxyethyl substituent. Structural comparisons with such enzymes suggest that in SFC-1 subtle repositioning of key residues (Ser70, Ser130, Asn132 and Asn170) enlarges the active site, permitting rotation of the carbapenem 6α-1R-hydroxyethyl group and abolishing this contact. Our data show that SFC-1, and by implication other such carbapenem-hydrolyzing enzymes, uses Asn132 to orient bound carbapenems for efficient deacylation and prevent their interaction with the deacylating water molecule.

Density functional theory (DFT) based quantum mechanical/molecular mechanical (QM/MM) calculations have provided valuable insight into the reactivity of the cytochrome P450 family of enzymes (P450s). A failure of commonly used DFT methods, such as B3LYP, is the neglect of dispersion interactions. An empirical dispersion correction has been shown to improve the accuracy of gas phase DFT calculations of P450s. The current work examines the effect of the dispersion correction in QM/MM calculations on P450s. The hydrogen abstraction from camphor, and hydrogen abstraction and C–O addition of cyclohexene and propene by P450cam have been modeled, along with the addition of benzene to Compound I in CYP2C9, at the B3LYP-D2/CHARMM27 level of theory. Single point energy calculations were also performed at the B3LYP-D3//B3LYP-D2/CHARMM27 level. The dispersion corrections lower activation energy barriers significantly (by ∼5 kcal/mol), as seen for gas phase calculations, but has a small effect on optimized geometries.These effects are likely to be important in modeling reactions catalyzed by other enzymes also. Given the low computational cost of including such dispersion corrections, we recommend doing so in all B3LYP based QM/MM calculations.

Soluble epoxide hydrolase (sEH) is an enzyme involved in drug metabolism that catalyzes the hydrolysis of epoxides to form their corresponding diols. sEH has a broad substrate range and shows high regio- and enantioselectivity for nucleophilic ring opening by Asp333. Epoxide hydrolases therefore have potential synthetic applications. We have used combined quantum mechanics/molecular mechanics (QM/MM) umbrella sampling molecular dynamics (MD) simulations (at the AM1/CHARMM22 level) and high-level ab initio (SCS-MP2) QM/MM calculations to analyze the reactions, and determinants of selectivity, for two substrates: trans-stilbene oxide (t-SO) and trans-diphenylpropene oxide (t-DPPO). The calculated free energy barriers from the QM/MM (AM1/CHARMM22) umbrella sampling MD simulations show a lower barrier for phenyl attack in t-DPPO, compared with that for benzylic attack, in agreement with experiment. Activation barriers in agreement with experimental rate constants are obtained only with the highest level of QM theory (SCS-MP2) used. Our results show that the selectivity of the ring-opening reaction is influenced by several factors, including proximity to the nucleophile, electronic stabilization of the transition state, and hydrogen bonding to two active site tyrosine residues. The protonation state of His523 during nucleophilic attack has also been investigated, and our results show that the protonated form is most consistent with experimental findings. The work presented here illustrates how determinants of selectivity can be identified from QM/MM simulations. These insights may also provide useful information for the design of novel catalysts for use in the synthesis of enantiopure compounds.

Multidrug resistance of the pandemic H1N1-2009 strain of influenza has been reported due to widespread treatment using the neuraminidase (NA) inhibitors, oseltamivir (Tamiflu), and zanamivir (Relenza). From clinical data, the single I223R (IR(1)) mutant of H1N1-2009 NA reduced efficacy of oseltamivir and zanamivir by 45 and 10 times,(1) respectively. More seriously, the efficacy of these two inhibitors against the double mutant I223R/H275Y (IRHY(2)) was significantly reduced by a factor of 12 374 and 21 times, respectively, compared to the wild-type.(2) This has led to the question of why the efficacy of the NA inhibitors is reduced by the occurrence of these mutations and, specifically, why the efficacy of oseltamivir against the double mutant IRHY was significantly reduced, to the point where oseltamivir has become an ineffective treatment. In this study, 1 μs of molecular dynamics (MD) simulations was performed to answer these questions. The simulations, run using graphical processors (GPUs), were used to investigate the effect of conformational change upon binding of the NA inhibitors oseltamivir and zanamivir in the wild-type and the IR and IRHY mutant strains. These long time scale dynamics simulations demonstrated that the mechanism of resistance of IRHY to oseltamivir was due to the loss of key hydrogen bonds between the inhibitor and residues in the 150-loop. This allowed NA to transition from a closed to an open conformation. Oseltamivir binds weakly with the open conformation of NA due to poor electrostatic interactions between the inhibitor and the active site. The results suggest that the efficacy of oseltamivir is reduced significantly because of conformational changes that lead to the open form of the 150-loop. This suggests that drug resistance could be overcome by increasing hydrogen bond interactions between NA inhibitors and residues in the 150-loop, with the aim of maintaining the closed conformation, or by designing inhibitors that can form a hydrogen bond to the mutant R223 residue, thereby preventing competition between R223 and R152.

QM/MM modelling of FAAH inactivation by O-biphenyl-3-yl carbamates identifies the deprotonation of Ser241 as the key reaction step, explaining why FAAH is insensitive to the electron-donor effect of conjugated substituents; this may aid design of new inhibitors with improved selectivity and in vivo potency.

Chorismate mutase is at the centre of current controversy about fundamental features of biological catalysts. Some recent studies have proposed that catalysis in this enzyme does not involve transition state (TS) stabilization but instead is due largely to the formation of a reactive conformation of the substrate. To understand the origins of catalysis, it is necessary to compare equivalent reactions in different environments. The pericyclic conversion of chorismate to prephenate catalysed by chorismate mutase also occurs (much more slowly) in aqueous solution. In this study we analyse the origins of catalysis by comparison of multiple quantum mechanics/molecular mechanics (QM/MM) reaction pathways at a reliable, well tested level of theory (B3LYP/6-31G(d)/CHARMM27) for the reaction (i) in Bacillus subtilischorismate mutase (BsCM) and (ii) in aqueous solvent. The average calculated reaction (potential energy) barriers are 11.3 kcal mol−1 in the enzyme and 17.4 kcal mol−1 in water, both of which are in good agreement with experiment. Comparison of the two sets of reaction pathways shows that the reaction follows a slightly different reaction pathway in the enzyme than in it does in solution, because of a destabilization, or strain, of the substrate in the enzyme. The substrate strain energy within the enzyme remains constant throughout the reaction. There is no unique reactive conformation of the substrate common to both environments, and the transition state structures are also different in the enzyme and in water. Analysis of the barrier heights in each environment shows a clear correlation between TS stabilization and the barrier height. The average differential TS stabilization is 7.3 kcal mol−1 in the enzyme. This is significantly higher than the small amount of TS stabilization in water (on average only 1.0 kcal mol−1 relative to the substrate). The TS is stabilized mainly by electrostatic interactions with active site residues in the enzyme, with Arg90, Arg7 and Glu78 generally the most important. Conformational effects (e.g. strain of the substrate in the enzyme) do not contribute significantly to the lower barrier observed in the enzyme. The results show that catalysis is mainly due to better TS stabilization by the enzyme.

Family 18 chitinases catalyze the hydrolysis of β-1,4-glycosidic bonds in chitin. The mechanism has been proposed to involve the formation of an oxazolinium ion intermediate via an unusual substrate-assisted mechanism, in which the substrate itself acts as an intramolecular nucleophile (instead of an enzyme residue). Here, we have modeled the first step of the chitin hydrolysis catalyzed by Serratia marcescens chitinase B for the first time using a combined quantum mechanics/molecular mechanics approach. The calculated reaction barriers based on multiple snapshots are 15.8–19.8 kcal mol–1 [B3LYP/6-31+G(d)//AM1-CHARMM22], in good agreement with the activation free energy of 16.1 kcal mol–1 derived from experiment. The enzyme significantly stabilizes the oxazolinium intermediate. Two stable conformations (4C1-chair and B3,O-boat) of the oxazolinium ion intermediate in subsite −1 were unexpectedly observed. The transition state structure has significant oxacarbenium ion-like character. The glycosyl residue in subsite −1 was found to follow a complex conformational pathway during the reaction (1,4B → [4H5/4E]⧧ → 4C1 ↔ B3,O), indicating complex conformational behavior in glycoside hydrolases that utilize a substrate-assisted catalytic mechanism. The D142N mutant is found to follow the same wild-type-like mechanism: the calculated barriers for reaction in this mutant (16.0–21.1 kcal mol–1) are higher than in the wild type, in agreement with the experiment. Asp142 is found to be important in transition state and intermediate stabilization.

The effects of structural fluctuations, due to protein dynamics, on enzyme activity are at the heart of current debates on enzyme catalysis. There is evidence that fatty acid amide hydrolase (FAAH) is an enzyme for which reaction proceeds via a high-energy, reactive conformation, distinct from the predominant enzyme−substrate complex (Lodola et al. Biophys. J. 2007, 92, L20−22). Identifying the structural causes of differences in reactivity between conformations in such complex systems is not trivial. Here, we show that multivariate analysis of key structural parameters can identify structural determinants of barrier height by analysis of multiple reaction paths. We apply a well-tested quantum mechanics/molecular mechanics (QM/MM) method to the first step of the acylation reaction between FAAH and oleamide substrate for 36 different starting structures. Geometrical parameters (consisting of the key bond distances that change during the reaction) were collected and used for principal component analysis (PCA), partial least-squares (PLS) regression analysis, and multiple linear regression (MLR) analysis. PCA indicates that different “families” of enzyme−substrate conformations arise from QM/MM molecular dynamics simulation and that rarely sampled, catalytically significant conformational states can be identified. PLS and MLR analyses allowed the construction of linear regression models, correlating the calculated activation barriers with simple geometrical descriptors. These analyses reveal the presence of two fully independent geometrical effects, explaining 78% of the variation in the activation barrier, which are directly correlated with transition-state stabilization (playing a major role in catalysis) and substrate binding. These results highlight the power of statistical approaches of this type in identifying crucial structural features that contribute to enzyme reactivity.

Transfer of hydrogen as a proton, hydride or hydrogen atom is an important step in many enzymic reactions. Experiments show kinetic isotope effects (KIEs) for some enzyme-catalysed hydrogen transfer reactions that deviate significantly from the limits imposed by considering the differences in mass of the isotopes alone (i.e. the semiclassical limit). These KIEs can be explained if the transfer of the hydrogen species occurs via a quantum mechanical tunnelling mechanism. The unusual temperature dependence of some KIEs has led to suggestions that enzymes have evolved to promote tunnelling through dynamics — a highly controversial hypothesis. Molecular simulations have a vital role in resolving these questions, providing a level of detail of analysis not possible through experiments alone. Here, we review computational molecular modelling studies of quantum tunnelling in enzymes, in particular focusing on the enzymes soybean lipoxygenase-1 (SLO-1), dihydrofolate reductase (DHFR), methylamine dehydrogenase (MADH) and aromatic amine dehydrogenase (AADH) to illustrate the current controversy regarding the importance of quantum effects in enzyme catalysis.

Combined quantum mechanics/molecular mechanics (QM/MM) calculations with high levels of correlated ab initio theory can now provide benchmarks for enzyme-catalyzed reactions. Here, we use such methods to test various QM/MM methods and the sensitivity of the results to details of the models for an important enzyme reaction, proton abstraction from acetyl-coenzyme A in citrate synthase. We calculate multiple QM/MM potential energy surfaces up to the local coupled cluster theory (LCCSD(T0)) level, with structures optimized at hybrid density functional theory and Hartree−Fock levels. The influence of QM methods, basis sets, and QM region size is shown to be significant. Correlated ab initio QM/MM calculations give barriers in agreement with experiment for formation of the acetyl-CoA enolate intermediate. In contrast, B3LYP fails to identify the enolate as an intermediate, whereas BH&HLYP does. The results indicate that QM/MM methods and setup should be tested, ideally using high-level calculations, to draw reliable mechanistic conclusions.

High-level (MP2, HF, and BLYP with the aug-cc-pVDZ basis set) quantum mechanics/molecular mechanics (QM/MM) Monte Carlo free energy simulations of liquid water are used here to test the compatibility of various QM methods with four standard empirical “molecular mechanics” (MM) water models. Consistency of QM methods with water models is of particular importance, given the aqueous environment of many of the systems of interest for QM/MM modeling (e.g., biological systems). The results show that treating a single water molecule using a QM method in bulk TIP3P can induce solvent structuring consistent with experiment. The results also show that the TIP4P model is the most suitable water model of those tested for such QM/MM simulations, while the TIP5P model is not well suited. The findings have important implications for future QM/MM method development and applications. They indicate that the choice of MM models should be made carefully for consistency and compatibility in QM/MM simulations.Keywords (keywords): biomolecular simulation; density functional theory; free energy calculations; quantum mechanics/molecular mechanics; thermodynamic integration;

Quantum mechanics/molecular mechanics (QM/MM) molecular dynamics simulations indicate that the reaction of native HEWL with its natural substrate involves a covalent intermediate, in contrast to the ‘textbook’ mechanism for this seminal enzyme.

High-level ab initio quantum mechanical/molecular mechanical (QM/MM) modelling of citryl-CoA formation in citrate synthase reveals that an arginine residue acts as the proton donor; this proposed new mechanism helps to explain how chemical and large scale conformational changes are coupled in this paradigmatic enzyme.

Modelling of the mechanism of covalent adduct formation by the inhibitor O-arylcarbamate URB524 in FAAH shows that only one of the two possible inhibitor binding orientations is consistent with the experimentally observed irreversible carbamoylation of the nucleophile serine: this is a potentially crucial insight for designing new covalent inhibitors of this promising drug target.

The mechanism of benzene hydroxylation was investigated in the realistic enzyme environment of the human CYP 2C9 by using quantum mechanical/molecular mechanical (QM/MM) calculations of the whole reaction profile using the B3LYP method to describe the QM region. The calculated QM/MM barriers for addition of the active species Compound I to benzene are consistent with experimental rate constants for benzene metabolism in CYP 2E1. In contrast to gas-phase model calculations, our results suggest that competing side-on and face-on geometries of arene addition may both occur in the case of aromatic ring oxidation in cytochrome P450s. QM/MM profiles for three different rearrangement pathways of the initially formed σ-adduct, leading to formation of epoxide, ketone, and an N-protonated porphyrin species, were calculated. Our results suggest that epoxide and ketone products form with comparable ease in the face-on pathway, whereas epoxide formation is preferred in the side-on pathway. Additionally, rearrangement to the N-protonated porphyrin species was found to be competitive with side-on epoxide formation. This suggests that overall, the competition between formation of epoxide and phenol final products in P450 oxidation of aromatic substrates is quite finely balanced.

The first step of the reaction catalysed by the enzyme citrate synthase is studied here with high level combined quantum mechanical/molecular mechanical (QM/MM) methods (up to the MP2/6-31+G(d)//6-31G(d)/CHARMM level). In the first step of the reaction, acetyl-CoA is deprotonated by Asp375, producing an intermediate, which is the nucleophile for attack on the second substrate, oxaloacetate, prior to hydrolysis of the thioester bond of acetyl-CoA and release of the products. A central question has been whether the nucleophilic intermediate is the enolate of acetyl-CoA, the enol, or an ‘enolic’ intermediate stabilized by a ‘low-barrier’ hydrogen bond with His274 at the active site. The imidazole sidechain of His274 is neutral, and donates a hydrogen bond to the carbonyl oxygen of acetyl-CoA in substrate complexes. We have investigated the identity of the nucleophilic intermediate by QM/MM calculations on the substrate (keto), enolate, enol and enolic forms of acetyl-CoA at the active site of citrate synthase. The transition states for proton abstraction from acetyl-CoA by Asp375, and for transfer of the hydrogen bonded proton between His274 and acetyl-CoA have been modelled approximately. The effects of electron correlation are included by MP2/6-31G(d) and MP2/6-31+G(d) calculations on active site geometries produced by QM/MM energy minimization. The results do not support the hypothesis that a low-barrier hydrogen bond is involved in catalysis in citrate synthase, in agreement with earlier calculations. The acetyl-CoA enolate is identified as the only intermediate consistent with the experimental barrier for condensation, stabilized by conventional hydrogen bonds from His274 and a water molecule.

The fundamental nature of reactivity in cytochrome P450 enzymes is currently controversial. Modelling of bacterial P450cam has suggested an important role for the haem propionates in the catalysis, though this finding has been questioned. Understanding the mechanisms of this enzyme family is important both in terms of basic biochemistry and potentially in the prediction of drug metabolism. We have modelled the hydroxylation of camphor by P450cam, using combined quantum mechanics/molecular mechanics (QM/MM) methods. A set of reaction pathways in the enzyme was determined. We were able to pinpoint the source of the discrepancies in the previous results. We show that when a correct ionization state is assigned to Asp297, no spin density appears on the haem propionates and the protein structure in this region remains preserved. These results indicate that the haem propionates are not involved in catalysis.

Fatty acid amide hydrolase (FAAH), a promising target for the treatment of several central and peripheral nervous system disorders, such as anxiety, pain and hypertension, has an unusual catalytic site, and its mechanism has been uncertain; hybrid quantum mechanics/molecular mechanics (QM/MM) calculations reveal a new mechanism of nucleophile activation (involving a Lys–Ser–Ser catalytic triad), with potentially crucial insights for the design of potent and selective inhibitors.

The controversial ‘near attack conformation’ (NAC) effect in the important model enzyme chorismate mutase is calculated to be 3.8–4.6 kcal mol−1 by QM/MM free energy perturbation molecular dynamics methods, showing that the NAC effect by itself does not account for catalysis in this enzyme.

To investigate fundamental features of enzyme catalysis, there is a need for high-level calculations capable of modelling crucial, unstable species such as transition states as they are formed within enzymes. We have modelled an important model enzyme reaction, the Claisen rearrangement of chorismate to prephenate in chorismate mutase, by combined ab initio quantum mechanics/molecular mechanics (QM/MM) methods. The best estimates of the potential energy barrier in the enzyme are 7.4–11.0 kcal mol−1 (MP2/6-31+G(d)//6-31G(d)/CHARMM22) and 12.7–16.1 kcal mol−1 (B3LYP/6-311+G(2d,p)//6-31G(d)/CHARMM22), comparable to the experimental estimate of ΔH‡ = 12.7 ± 0.4 kcal mol−1. The results provide unequivocal evidence of transition state (TS) stabilization by the enzyme, with contributions from residues Arg90, Arg7, and Arg63. Glu78 stabilizes the prephenate product (relative to substrate), and can also stabilize the TS. Examination of the same pathway in solution (with a variety of continuum models), at the same ab initio levels, allows comparison of the catalyzed and uncatalyzed reactions. Calculated barriers in solution are 28.0 kcal mol−1 (MP2/6-31+G(d)/PCM) and 24.6 kcal mol−1 (B3LYP/6-311+G(2d,p)/PCM), comparable to the experimental finding of ΔG‡ = 25.4 kcal mol−1 and consistent with the experimentally-deduced 106-fold rate acceleration by the enzyme. The substrate is found to be significantly distorted in the enzyme, adopting a structure closer to the transition state, although the degree of compression is less than predicted by lower-level calculations. This apparent substrate strain, or compression, is potentially also catalytically relevant. Solution calculations, however, suggest that the catalytic contribution of this compression may be relatively small. Consideration of the same reaction pathway in solution and in the enzyme, involving reaction from a ‘near-attack conformer’ of the substrate, indicates that adoption of this conformation is not in itself a major contribution to catalysis. Transition state stabilization (by electrostatic interactions, including hydrogen bonds) is found to be central to catalysis by the enzyme. Several hydrogen bonds are observed to shorten at the TS. The active site is clearly complementary to the transition state for the reaction, stabilizing it more than the substrate, so reducing the barrier to reaction.

A key step in the cyclooxygenase reaction cycle of cyclooxygenase 1 (COX-1) is abstraction of the pro-S hydrogen atom of the arachidonic acid by a radical that is formed at the protein residue Tyr-385. Here we investigate this reaction step by a quantum-mechanics/molecular-mechanics approach in combination with molecular-dynamics simulations. The simulations identify the hydrogen abstraction angle as a crucial geometric determinant of the reaction, thus revealing the importance of the cyclooxygenase active site for calculating the potential energy surface of the reaction.

Quantum mechanics/molecular mechanics and molecular dynamics simulations of fatty acid amide hydrolase show that reaction (amide hydrolysis) occurs via a distinct, high energy conformation. This unusual finding has important implications for fatty acid amide hydrolase, a key enzyme in the endocannabinoid system. These results demonstrate the importance of structural fluctuations and the need to include them in the modeling of enzyme reactions. They also show that approaches based simply on studying enzyme-substrate complexes can be misleading for understanding biochemical reactivity.

Carbapenems, ‘last resort’ antibiotics for many bacterial infections, can now be broken down by several class A β-lactamases (i.e. carbapenemases). Here, carbapenemase activity is predicted through QM/MM dynamics simulations of acyl–enzyme deacylation, requiring only the 3D structure of the apo-enzyme. This may assist in anticipating resistance and future antibiotic design.

QM/MM modelling of FAAH inactivation by O-biphenyl-3-yl carbamates identifies the deprotonation of Ser241 as the key reaction step, explaining why FAAH is insensitive to the electron-donor effect of conjugated substituents; this may aid design of new inhibitors with improved selectivity and in vivo potency.

Modelling of the mechanism of covalent adduct formation by the inhibitor O-arylcarbamate URB524 in FAAH shows that only one of the two possible inhibitor binding orientations is consistent with the experimentally observed irreversible carbamoylation of the nucleophile serine: this is a potentially crucial insight for designing new covalent inhibitors of this promising drug target.

High-level ab initio quantum mechanical/molecular mechanical (QM/MM) modelling of citryl-CoA formation in citrate synthase reveals that an arginine residue acts as the proton donor; this proposed new mechanism helps to explain how chemical and large scale conformational changes are coupled in this paradigmatic enzyme.

Quantum mechanics/molecular mechanics (QM/MM) molecular dynamics simulations indicate that the reaction of native HEWL with its natural substrate involves a covalent intermediate, in contrast to the ‘textbook’ mechanism for this seminal enzyme.

Chorismate mutase is at the centre of current controversy about fundamental features of biological catalysts. Some recent studies have proposed that catalysis in this enzyme does not involve transition state (TS) stabilization but instead is due largely to the formation of a reactive conformation of the substrate. To understand the origins of catalysis, it is necessary to compare equivalent reactions in different environments. The pericyclic conversion of chorismate to prephenate catalysed by chorismate mutase also occurs (much more slowly) in aqueous solution. In this study we analyse the origins of catalysis by comparison of multiple quantum mechanics/molecular mechanics (QM/MM) reaction pathways at a reliable, well tested level of theory (B3LYP/6-31G(d)/CHARMM27) for the reaction (i) in Bacillus subtilischorismate mutase (BsCM) and (ii) in aqueous solvent. The average calculated reaction (potential energy) barriers are 11.3 kcal mol−1 in the enzyme and 17.4 kcal mol−1 in water, both of which are in good agreement with experiment. Comparison of the two sets of reaction pathways shows that the reaction follows a slightly different reaction pathway in the enzyme than in it does in solution, because of a destabilization, or strain, of the substrate in the enzyme. The substrate strain energy within the enzyme remains constant throughout the reaction. There is no unique reactive conformation of the substrate common to both environments, and the transition state structures are also different in the enzyme and in water. Analysis of the barrier heights in each environment shows a clear correlation between TS stabilization and the barrier height. The average differential TS stabilization is 7.3 kcal mol−1 in the enzyme. This is significantly higher than the small amount of TS stabilization in water (on average only 1.0 kcal mol−1 relative to the substrate). The TS is stabilized mainly by electrostatic interactions with active site residues in the enzyme, with Arg90, Arg7 and Glu78 generally the most important. Conformational effects (e.g. strain of the substrate in the enzyme) do not contribute significantly to the lower barrier observed in the enzyme. The results show that catalysis is mainly due to better TS stabilization by the enzyme.

Combined quantum mechanics/molecular mechanics (QM/MM) simulations of the reaction catalysed by phosphite dehydrogenase (PTDH) identify Met53 as important for catalysis. This catalytic role is verified by experiments (including replacement by norleucine and selenomethionine), which show that mutation of this residue significantly affects kcat, without changing KM for phosphite. QM/MM and ab initio QM calculations show that the catalytic effect is electrostatic in nature. The side chain of Met53 specifically stabilizes the transition state for the hydride transfer step of the reaction catalysed by PTDH, forming a ‘face-on’ interaction with His292. To our knowledge, a defined catalytic role for methionine in an enzyme (as opposed to a steric or binding effect, or interaction with a metal ion) has not previously been identified. Analyses of the Protein Data Bank and Cambridge Structural Database indicate that this type of interaction may be relatively widespread, with implications for enzyme-catalysed reaction mechanisms and protein structure.