Controlling Reactivity and Selectivity in the Mizoroki-Heck Reaction: High Throughput Evaluation of 1,5-Diaza-3,7-diphosphacyclooctane Ligands.

We report a high throughput evaluation of the Mizoroki-Heck reaction of diverse olefin coupling partners. Comparison of different ligands revealed the 1,5-diaza-3,7-diphosphacyclooctane (P2N2) scaffold to be more broadly applicable than common "gold standard" ligands, demonstrating that this family of readily accessible diphosphines has unrecognized potential in organic synthesis. In particular, two structurally related P2N2 ligands were identified to enable the regiodivergent arylation of styrenes. By simply altering the phosphorus substituent from a phenyl to tert-butyl group, both the linear and branched Mizoroki-Heck products can be obtained in high regioisomeric ratios. Experimental and computational mechanistic studies were performed to further probe the origin of selectivity, which suggests that both ligands coordinate to the metal in a similar manner but that rigid positioning of the phosphorus substituent forces contact with the incoming olefin in a π-π interaction (for P-Ph ligands) or with steric clash (for P-tBu ligands), dictating the regiocontrol.

A kinetic study of thiol addition to N-phenylchloroacetamide.

Irreversible enzyme inhibitors bind covalently to their target and permanently limit its function. The redox-sensitive thiol group on the side chain of cysteine (Cys) residues is often the nucleophilic group that is targeted for reaction with the electrophilic warhead of irreversible inhibitors. While the acrylamide group is the warhead applied most frequently currently in the design of inhibitors with therapeutic potential, the chloroacetamide group offers a comparable reactivity profile. In that context, we have studied the details of the mechanism of thiol addition to N-phenylchloroacetamide (NPC). A kinetic assay was developed to accurately monitor the reaction progress between NPC and a small library of thiols with varying pKa values. From these data, a Brønsted-type plot was constructed, from which a ßnucRS- value of 0.22 ± 0.07 was derived, indicative of a relatively early transition state with respect to attack by the thiolate. The halide leaving group was also varied, for the reaction with one thiol, providing rate constants consistent with a transition state that is early with respect to leaving group departure. The effects of temperature and ionic strength were also studied, and all data are consistent with an early transition state for a concerted SN2 mechanism of addition. Molecular modelling was also performed, and these calculations confirm the concerted transition state and relative reactivity of the haloacetamides. Finally, this study allows a detailed comparison of the reactivity and reaction mechanisms of the chloroacetamide group with the benchmark acrylamides used in many irreversible inhibitor drugs.

A mechanistic study of thiol addition to N-acryloylpiperidine.

Nucleophilic cysteine (Cys) residues are present in many enzyme active sites and represent the target of many different irreversible enzyme inhibitors. Given its fine balance between aqueous stability and thiolate reactivity, the acrylamide group is a particularly popular warhead pharmacophore among inhibitors designed for biological and therapeutic application. The acrylamide group is well known to undergo thiol addition, but the precise mechanism of this addition reaction has not been studied in as much detail. In this work we have focussed on the reaction of N-acryloylpiperidine (AcrPip), which represents a motif found in many targeted covalent inhibitor drugs. Using a precise HPLC-based assay, we measured the second order rate constants for the reaction of AcrPip with a panel of thiols possessing different pKa values. This allowed construction of a Brønsted-type plot that reveals the relative insensitivity of the reaction to the nucleophilicity of the thiolate. By studying temperature effects, we were able to construct an Eyring plot from which the enthalpy and entropy of activation were calculated. Ionic strength and solvent kinetic isotope effects were also studied, informing on charge dispersal and proton transfer in the transition state. DFT calculations were also performed, providing information on the potential structure of the activated complex. Taken together, these data strongly support one cohesive addition mechanism that is the microscopic reverse of the E1cb elimination, and highly relevant to the intrinsic thiol selectivity of AcrPip inhibitors and their subsequent design.

A mechanistic study of thiol addition to N-phenylacrylamide.

Cysteine (Cys) residues contain a redox-sensitive thiol and are commonly found in enzyme active sites. In recent years, the presence of a reactive thiolate group on a protein has been exploited in the development of irreversible enzyme inhibitors as therapeutic agents. Many targeted covalent inhibitors (TCIs) are designed to covalently react with a specific Cys residue on a target protein active site, irreversibly modifying the target and inhibiting its normal function. The electrophilic warhead most commonly used in this way is the acrylamide functional group. Although the acrylamide group is well known for its ability to undergo thiol-addition reactions, very few studies have been conducted to elucidate the detailed mechanism of this reaction, which inspired us to conduct a thorough kinetic investigation. First, we developed a robust kinetic assay to accurately monitor reaction progress between N-phenylacrylamide (NPA) and a small library of alkyl thiols having widely varying pKa values. This allowed us to construct a Brønsted-type plot for the thiol addition reaction, revealing a ßnucRS- value of 0.07 ± 0.04. We also studied the solvent kinetic isotope effects (SKIEs), pH dependence, and temperature dependence of the reaction, which showed that reaction has a relatively large negative ΔS‡, and a small ΔH‡. Computational studies provided a structure for the transition state that is consistent with the experimental data. All of these data are consistent with rate-limiting nucleophilic attack, followed by rapid protonation of the enolate, corresponding to the microscopic reverse of the E1revcb elimination mechanism.

Interaction between Antimicrobial Peptide Magainin 2 and Nonlipid Components in the Bacterial Outer Envelope.

Antimicrobial peptides (AMPs) offer advantages over conventional antibiotics; for example, bacteria develop more resistance to small-molecule antibiotics than to AMPs. The interaction of the AMPs with the lipopolysaccharide (LPS) layer of the Gram-negative bacteria cell envelope is not well understood. A MARTINI model was constructed of a Gram-negative bacterial outer membrane interacting with the AMP Magainin 2. In a 20 µs molecular dynamics (MD) simulation, the AMP diffused to the LPS layer of the cell envelope and remained there, suggesting interactions between the Magainin 2 and the LPS layer, causing the AMP to concentrate at that position. The free energy profile for the insertion of the Magainin 2 into the membrane was also calculated using umbrella sampling, which showed that the AMP positioned such that the cationic side chains of the AMP coordinated to the negatively charged phosphate groups of the LPS layer. These simulations indicate that the AMP Magainin 2 partition into the LPS layer of a bacterial membrane.

Modeling the Binding and Conformational Energetics of a Targeted Covalent Inhibitor to Bruton's Tyrosine Kinase.

Targeted covalent inhibitors (TCIs) bind to their targets in both covalent and noncovalent modes, providing exceptionally high affinity and selectivity. These inhibitors have been effectively employed as inhibitors of protein kinases, with Taunton and coworkers (Nat. Chem. Biol. 2015, 11, 525-531) reporting a notable example of a TCI with a cyanoacrylamide warhead that forms a covalent thioether linkage to an active-site cysteine (Cys481) of Bruton's tyrosine kinase (BTK). The specific mechanism of the binding and the relative importance of the covalent and noncovalent interactions is difficult to determine experimentally, and established simulation methods for calculating the absolute binding affinity of an inhibitor cannot describe the covalent bond-forming steps. Here, an integrated approach using alchemical free-energy perturbation and QM/MM molecular dynamics methods was employed to model the complete Gibbs energy profile for the covalent inhibition of BTK by a cyanoacrylamide TCI. These calculations provide a rigorous and complete absolute Gibbs energy profile of the covalent modification binding process. Following a classic thiol-Michael addition mechanism, the target cysteine is deprotonated to form a nucleophilic thiolate, which then undergoes a facile conjugate addition to the electrophilic functional group to form a bond with the noncovalently bound ligand. This model predicts that the formation of the covalent linkage is highly exergonic relative to the noncovalent binding alone. Nevertheless, noncovalent interactions between the ligand and individual amino acid residues in the binding pocket of the enzyme are also essential for ligand binding, particularly van der Waals dispersion forces, which have a larger contribution to the binding energy than the covalent component in absolute terms. This model also shows that the mechanism of covalent modification of a protein occurs through a complex series of steps and that entropy, conformational flexibility, noncovalent interactions, and the formation of covalent linkage are all significant factors in the ultimate binding affinity of a covalent drug to its target.

Desulfonative Suzuki-Miyaura Coupling of Sulfonyl Fluorides.

Sulfonyl fluorides have emerged as powerful "click" electrophiles to access sulfonylated derivatives. Yet, they are relatively inert towards C-C bond forming transformations, notably under transition-metal catalysis. Here, we describe conditions under which aryl sulfonyl fluorides act as electrophiles for the Pd-catalyzed Suzuki-Miyaura cross-coupling. This desulfonative cross-coupling occurs selectively in the absence of base and, unusually, even in the presence of strong acids. Divergent one-step syntheses of two analogues of bioactive compounds showcase the expanded reactivity of sulfonyl fluorides to encompass both S-Nu and C-C bond formation. Mechanistic experiments and DFT calculations suggest oxidative addition occurs at the C-S bond followed by desulfonation to form a Pd-F intermediate that facilitates transmetalation.

Benchmarking Force Field and the ANI Neural Network Potentials for the Torsional Potential Energy Surface of Biaryl Drug Fragments.

Many drug molecules contain biaryl fragments, resulting in a torsional barrier corresponding to rotation around the bond linking the aryls. The potential energy surfaces of these torsions vary significantly because of steric and electronic effects, ultimately affecting the relative stability of the molecular conformations in the protein-bound and solution states. Simulations of protein-ligand binding require accurate computational models to represent the intramolecular interactions to provide accurate predictions of the structure and dynamics of binding. In this article, we compare four force fields [generalized AMBER force field (GAFF), open force field (OpenFF), CHARMM general force field (CGenFF), optimized potentials for liquid simulations (OPLS)] and two neural network potentials (ANI-2x and ANI-1ccx) for their ability to predict the torsional potential energy surfaces of 88 biaryls extracted from drug fragments. The root mean square deviation (rmsd) over the full potential energy surface and the mean absolute deviation of the torsion rotational barrier height (MADB) relative to high-level ab initio reference data (CCSD(T1)*) were used as the measure of accuracy. Uncertainties in these metrics due to the composition of the data set were estimated using bootstrap analysis. In comparison to high-level ab initio data, ANI-1ccx was most accurate for predicting the barrier height (rmsd: 0.5 ± 0.0 kcal/mol, MADB: 0.8 ± 0.1 kcal/mol), followed closely by ANI-2x (rmsd: 0.5 ± 0.0 kcal/mol, MADB: 1.0 ± 0.2 kcal/mol), then CGenFF (rmsd: 0.8 ± 0.1 kcal/mol, MADB: 1.3 ± 0.1 kcal/mol) and OpenFF (rmsd: 0.7 ± 0.1 kcal/mol, MADB: 1.3 ± 0.1 kcal/mol), then GAFF (rmsd: 1.2 ± 0.2 kcal/mol, MADB: 2.6 ± 0.5 kcal/mol), and finally OPLS (rmsd: 3.6 ± 0.3 kcal/mol, MADB: 3.6 ± 0.3 kcal/mol). Significantly, the neural network potentials (NNPs) are systematically more accurate and more reliable than any of the force fields. As a practical example, the NNP/molecular mechanics method was used to simulate the isomerization of ozanimod, a drug used for multiple sclerosis. Multinanosecond molecular dynamics (MD) simulations in an explicit aqueous solvent were performed, as well as umbrella sampling and adaptive biasing force-enhanced sampling techniques. The rate constant for this isomerization calculated using transition state theory was 4.30 × 10-1 ns-1, which is consistent with direct MD simulations.

An efficient and accurate model for water with an improved non-bonded potential.

A molecular mechanical model for liquid water is developed that uses a physically motivated potential to represent Pauli repulsion and dispersion instead of the standard Lennard-Jones potential. The model has three atomic sites and a virtual site located on the ∠HOH bisector (i.e., a TIP4P-type model). Pauli-repulsive interactions are represented using a Buckingham-type exponential decay potential. Dispersion interactions are represented by both C6/r6 and C8/r8 terms. This higher order C8 dispersion term has been neglected by most force fields. The ForceBalance code was used to define parameters that optimally reproduce the experimental physical properties of liquid water. The resulting model is in good agreement with the experimental density, dielectric constant, enthalpy of vaporization, isothermal compressibility, thermal expansion coefficient, diffusion coefficient, and radial distribution function. A graphical processing unit-accelerated implementation of this improved non-bonded potential can be employed in OpenMM without modification by using the CustomNonBondedForce feature. The efficient and automated parameterization of these non-bonded potentials provides a rational strategy to define a new molecular mechanical force field that treats repulsion and dispersion interactions more rigorously without major modifications to the existing simulation codes or a substantially larger computational cost.

Polarisable force fields: what do they add in biomolecular simulations?

The quality of biomolecular simulations critically depends on the accuracy of the force field used to calculate the potential energy of the molecular configurations. Currently, most simulations employ non-polarisable force fields, which describe electrostatic interactions as the sum of Coulombic interactions between fixed atomic charges. Polarisation of these charge distributions is incorporated only in a mean-field manner. In the past decade, extensive efforts have been devoted to developing simple, efficient, and yet generally applicable polarisable force fields for biomolecular simulations. In this review, we summarise the latest developments in accounting for key biomolecular interactions with polarisable force fields and applications to address challenging biological questions. In the end, we provide an outlook for future development in polarisable force fields.

Simulating protein-ligand binding with neural network potentials.

Drug molecules adopt a range of conformations both in solution and in their protein-bound state. The strain and reduced flexibility of bound drugs can partially counter the intermolecular interactions that drive protein-ligand binding. To make accurate computational predictions of drug binding affinities, computational chemists have attempted to develop efficient empirical models of these interactions, although these methods are not always reliable. Machine learning has allowed the development of highly-accurate neural-network potentials (NNPs), which are capable of predicting the stability of molecular conformations with accuracy comparable to state-of-the-art quantum chemical calculations but at a billionth of the computational cost. Here, we demonstrate that these methods can be used to represent the intramolecular forces of protein-bound drugs within molecular dynamics simulations. These simulations are shown to be capable of predicting the protein-ligand binding pose and conformational component of the absolute Gibbs energy of binding for a set of drug molecules. Notably, the conformational energy for anti-cancer drug erlotinib binding to its target was found to be considerably overestimated by a molecular mechanical model, while the NNP predicts a more moderate value. Although the ANI-1ccX NNP was not trained to describe ionic molecules, reasonable binding poses are predicted for charged ligands, but this method is not suitable for modeling charged ligands in solution.

Quantum Chemical Methods for Modeling Covalent Modification of Biological Thiols.

Targeted covalent inhibitor drugs require computational methods that go beyond simple molecular-mechanical force fields in order to model the chemical reactions that occur when they bind to their targets. Here, several semiempirical and density-functional theory (DFT) methods are assessed for their ability to describe the potential energy surface and reaction energies of the covalent modification of a thiol by an electrophile. Functionals such as PBE and B3LYP fail to predict a stable enolate intermediate. This is largely due to delocalization error, which spuriously stabilizes the prereaction complex, in which excess electron density is transferred from the thiolate to the electrophile. Functionals with a high-exact exchange component, range-separated DFT functionals, and variationally optimized exact exchange (i.e., the LC-B05minV functional) correct this issue to various degrees. The large gradient behavior of the exchange enhancement factor is also found to significantly affect the results, leading to the improved performance of PBE0. While ωB97X-D and M06-2X were reasonably accurate, no method provided quantitative accuracy for all three electrophiles, making this a very strenuous test of functional performance. Additionally, one drawback of M06-2X was that molecular dynamics (MD) simulations using this functional were only stable if a fine integration grid was used. The low-cost semiempirical methods, PM3, AM1, and PM7, provide a qualitatively correct description of the reaction mechanism, although the energetics is not quantitatively reliable. As a proof of concept, the potential of mean force for the addition of methylthiolate to methylvinyl ketone was calculated using quantum mechanical/molecular mechanical MD in an explicit polarizable aqueous solvent. © 2019 Wiley Periodicals, Inc.

Suzuki-Miyaura Coupling of (Hetero)Aryl Sulfones: Complementary Reactivity Enables Iterative Polyaryl Synthesis.

Ideal organic syntheses involve the rapid construction of C-C bonds, with minimal use of functional group interconversions. The Suzuki-Miyaura cross-coupling (SMC) is a powerful way to form biaryl linkages, but the relatively similar reactivity of electrophilic partners makes iterative syntheses involving more than two sequential coupling events difficult to achieve without additional manipulations. Here we introduce (hetero)aryl sulfones as electrophilic coupling partners for the SMC reaction, which display an intermediate reactivity between those of typical aryl (pseudo)halides and nitroarenes. The new complementary reactivity allows for rapid sequential cross-coupling of arenes bearing chloride, sulfone and nitro leaving groups, affording non-symmetric ter- and quateraryls in only 2 or 3 steps, respectively. The SMC reactivity of (hetero)aryl sulfones is demonstrated in over 30 examples. Mechanistic experiments and DFT calculations are consistent with oxidative addition into the sulfone C-S bond as the turnover-limiting step. The further development of electrophilic cross-coupling partners with complementary reactivity may open new possibilities for divergent iterative synthesis starting from small pools of polyfunctionalized arenes.

Ring-opening hydroarylation of monosubstituted cyclopropanes enabled by hexafluoroisopropanol.

Ring-opening hydroarylation of cyclopropanes is typically limited to substrates bearing a donor-acceptor motif. Here, the transformation is achieved for monosubstituted cyclopropanes by using catalytic Brønsted acid in hexafluoroisopropanol (HFIP) solvent, constituting a rare example where such cyclopropanes engage in intermolecular C-C bond formation. Branched products are obtained when electron-rich arylcyclopropanes react with a broad scope of arene nucleophiles in accord with a simple SN1-type ring-opening mechanism. In contrast, linear products are obtained when cyclopropylketones react with electron-rich arene nucleophiles. In the latter case, mechanistic experiments and DFT-calculations support a homo-conjugate addition pathway.

Mechanisms of Alkyl and Aryl Thiol Addition to N-Methylmaleimide.

A mechanistic study was undertaken to elucidate the reaction pathways for thiol addition to N-methylmaleimide in water. We used linear free energy relationships, solvent kinetic isotope effects (SKIEs), activation parameters, and ionic strength effects to probe the nature of the rate-limiting transition states. Calculations were also employed and assisted in illuminating three possible mechanistic pathways: (1) stepwise addition with rate-limiting nucleophilic attack, (2) stepwise addition with rate-limiting proton transfer, and (3) concerted addition with nucleophilic attack and proton transfer occurring concurrently. Alkyl thiolate addition exhibits ßnucRS-= 0.4, small negative Δ S‡ values, prominent ionic strength effects, and no evidence of general acid catalysis, consistent with pathway 1. Aryl thiolate addition exhibited ßnucArS- = 1.0, large negative Δ S‡ values, normal primary SKIEs, general acid catalysis, and negligible sensitivity to ionic strength, consistent with pathways 2 and 3. The experimental and computational data depict an energy surface where ground state effects, namely the energy of the alkyl/aryl thiolate, play a major role in shaping the governing pathway. Application of these findings to bioconjugation chemistry is also discussed.

Simulations of lipid bilayers using the CHARMM36 force field with the TIP3P-FB and TIP4P-FB water models.

The CHARMM36 force field for lipids is widely used in simulations of lipid bilayers. The CHARMM family of force fields were developed for use with the mTIP3P water model. This water model has an anomalously high dielectric constant and low viscosity, which limits its accuracy in the calculation of quantities like permeability coefficients. The TIP3P-FB and TIP4P-FB water models are more accurate in terms of the dielectric constant and transport properties, which could allow more accurate simulations of systems containing water and lipids. To test whether the CHARMM36 lipid force field is compatible with the TIP3P-FB and TIP4P-FB water models, we have performed simulations of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine bilayers. The calculated headgroup area, compressibility, order parameters, and X-ray form factors are in good agreement with the experimental values, indicating that these improved water models can be used with the CHARMM36 lipid force field without modification when calculating membrane physical properties. The water permeability predicted by these models is significantly different; the mTIP3P-model diffusion in solution and at the lipid-water interface is anomalously fast due to the spuriously low viscosity of mTIP3P-model water, but the potential of mean force of permeation is higher for the TIP3P-FB and TIP4P-FB models due to their high excess chemical potentials. As a result, the rates of water permeation calculated the FB water models are slower than the experimental value by a factor of 15-17, while simulations with the mTIP3P model only underestimate the water permeability by a factor of 3.

How Reactive are Druggable Cysteines in Protein Kinases?

Targeted covalent inhibitors (TCIs) have been successfully developed as high-affinity and selective inhibitors of enzymes of the protein kinase family. These drugs typically act by undergoing an electrophilic addition with an active-site cysteine residue, so design of a TCI begins with the identification of a "druggable" cysteine. These electrophilic additions generally require deprotonation of the thiol to form a reactive anionic thiolate, so the acidity of the residue is a critical factor. Few experimental measurements of the p Ka's of druggable cysteines have been reported, so computational prediction could prove to be very important in selecting reactive cysteine targets. Here we report the computed p Ka's of druggable cysteines in selected protein kinases that are of clinical relevance for targeted therapies. The p Ka's of the cysteines were calculated using advanced computational methods based on all-atom replica-exchange thermodynamic integration molecular dynamics simulations in explicit solvent. We found that the acidities of druggable cysteines within protein kinases are diverse and elevated, indicating enormous differences in their reactivity. Constant-pH molecular dynamics simulations were also performed on selected protein kinases, and the results confirmed this varied range in the acidities of druggable cysteines. Many of these active-site cysteines have low exposure to solvent molecules, elevating their p Ka values. Electrostatic interactions with nearby anionic residues also elevate the p Ka's of cysteine residues in the active site. The results suggest that some cysteine residues within kinase binding sites will be slow to react with a TCI because of their low acidity. Several oncogenic kinase mutations were also modeled and found to have p Ka's similar to that of the wild-type kinase.

The hydration structure of methylthiolate from QM/MM molecular dynamics.

Thiols are widely present in biological systems, most notably as the side chain of cysteine amino acids in proteins. Thiols can be deprotonated to form a thiolate which affords a diverse range of enzymatic activity and modes for chemical modification of proteins. Parameters for modeling thiolates using molecular mechanical force fields have not yet been validated, in part due to the lack of structural data on thiolate solvation. Here, the CHARMM36 and Amber models for thiolates in aqueous solutions are assessed using free energy perturbation and hybrid quantum mechanics/molecular mechanics (QM/MM) molecular dynamics (MD) simulations. The hydration structure of methylthiolate was calculated from 1 ns of QM/MM MD (PBE0-D3/def2-TZVP//TIP3P), which shows that the water-S- distances are approximately 2 Å with a coordination number near 6. The CHARMM thiolate parameters predict a thiolate S radius close to the QM/MM value and predict a hydration Gibbs energy of -329.2 kJ/mol, close to the experimental value of -318 kJ/mol. The cysteine thiolate model in the Amber force field underestimates the thiolate radius by 0.2 Å and overestimates the thiolate hydration energy by 119 kJ/mol because it uses the same Lennard-Jones parameters for thiolates as for thiols. A recent Drude polarizable model for methylthiolate with optimized thiolate parameters also performs well. SAPT2+ [Symmetry Adapted Perturbation Theory (SAPT)] analysis indicates that exchange repulsion is larger for the methylthiolate, consistent with it having a more diffuse electron density distribution in comparison with the parent thiol. These data demonstrate that it is important to define distinct non-bonded parameters for the protonated/deprotonated states of amino acid side chains in molecular mechanical force fields.

Folding free energy landscapes of ß-sheets with non-polarizable and polarizable CHARMM force fields.

Molecular dynamics (MD) simulations of peptides and proteins offer atomic-level detail into many biological processes, although the degree of insight depends on the accuracy of the force fields used to represent them. Protein folding is a key example in which the accurate reproduction of folded-state conformations of proteins and kinetics of the folding processes in simulation is a longstanding goal. Although there have been a number of recent successes, challenges remain in capturing the full complexity of folding for even secondary-structure elements. In the present work, we have used all-atom MD simulations to study the folding properties of one such element, the C-terminal ß-hairpin of the B1 domain of streptococcal protein G (GB1). Using replica-exchange umbrella sampling simulations, we examined the folding free energy of two fixed-charge CHARMM force fields, CHARMM36 and CHARMM22*, as well as a polarizable force field, the CHARMM Drude-2013 model, which has previously been shown to improve the folding properties of α-helical peptides. The CHARMM22* and Drude-2013 models are in rough agreement with experimental studies of GB1 folding, while CHARMM36 overstabilizes the ß-hairpin. Additional free-energy calculations show that small adjustments to the atomic polarizabilities in the Drude-2013 model can improve both the backbone solubility and folding properties of GB1 without significantly affecting the model's ability to properly fold α-helices. We also identify a non-native salt bridge in the ß-turn region that overstabilizes the ß-hairpin in the C36 model. Finally, we demonstrate that tryptophan fluorescence is insufficient for capturing the full ß-hairpin folding pathway.

Evaluating the London Dispersion Coefficients of Protein Force Fields Using the Exchange-Hole Dipole Moment Model.

London dispersion is one of the fundamental interactions involved in protein folding and dynamics. The popular CHARMM36, Amber ff14sb, and OPLS-AA force fields represent these interactions through the C6/ r6 term of the Lennard-Jones potential, where the C6 parameters are assigned empirically. Here, dispersion coefficients of these three force fields are shown to be roughly 50% larger than values calculated using the quantum mechanically derived exchange-hole dipole moment (XDM) model. The CHARMM36 and Amber OL15 force fields for nucleic acids also exhibit this trend. The hydration energies of the side-chain models were calculated using REMD-TI for the CHARMM36, Amber ff14sb, and OPLS-AA force fields. These force fields predict side-chain hydration energies that are in generally good agreement with the experimental values, which suggests that the total strength of aqueous dispersion interactions is correct, despite C6 coefficients that are considerably larger than XDM predicts. An analytical expression for the dispersion hydration energy using XDM coefficients shows that higher-order dispersion terms (i.e., C8 and C10) account for roughly 37.5% of the hydration energy of methane. This suggests that the C6 dispersion coefficients used in contemporary force fields are elevated to account for the neglected higher-order terms.