Calculation of absolute protein-ligand binding free energy using distributed replica sampling

Calculation of binding free energies, which characterize the stability and specificity of molecular interactions (e.g., drug formulations), has been established as an important application of modern computational chemistry. Though several methodologies currently exist within the literature, there is always a need for improvements in speed and accuracy that allow for efficient and accurate chemical screening. In this paper we explore the application of an unrestrained advanced sampling method, adaptive biasing force (ABF), to the calculation of standard binding free energies for a set of molecular host-guest systems for which the binding free energy is well-known: the cucurbit[7]uril-hosted systems utilized in the HYDROPHOBE challenge [ Assaf et al. J. Phys. Chem. B 2017 , 121 , 11144 - 11162 ]. We demonstrate that the use of the ABF advanced sampling method yields values systematically closer to experiment, with a lower uncertainty in the calculation, than thermodynamic integration and restrained sampling techniques commonly used in the literature. We hypothesize that this is due to the algorithm's ability to drive multiple binding and dissociation events through applied bias forces. These results show the promise of unrestrained advanced sampling methods along well-defined reaction coordinates to the calculation of binding energies in host-guest systems.

We propose the thermodynamic integration along a spatial reaction coordinate using the molecular dynamics simulation combined with the three-dimensional reference interaction site model theory. This method provides a free energy calculation in solution along the reaction coordinate defined by the Cartesian coordinates of the solute atoms. The proposed method is based on the blue moon algorithm which can, in principle, handle any reaction coordinate as far as it is defined by the solute atom positions. In this article, we apply the present method to the complex formation process of the crown ether 18-Crown-6 (18C6) with the potassium ion in an aqueous solution. The separation between the geometric centers of these two molecules is taken to be the reaction coordinate for this system. The potential of mean force (PMF) becomes the maximum at the separation between the molecular centers being ∼4 Å, which can be identified as the free energy barrier in the process of the molecular recognition. In a separation further than the free energy barrier, the PMF is slightly reduced to exhibit a plateau. In the region closer than the free energy barrier, approach of the potassium ion to the center of 18C6 also decreases the PMF. When the potassium ion is accommodated at the center of 18C6, the free energy is lower by -5.7 ± 0.7 kcal/mol than that at the above mentioned plateau or converged state. By comparing the results with those from the free energy calculation along the coupling parameters obtained in our previous paper [T. Miyata, Y. Ikuta, and F. Hirata, J. Chem. Phys. 133, 044114 (2010)], it is found that the effective interaction in water between 18C6 and the potassium ion vanishes beyond the molecular-center-separation of 10 Å. Furthermore, the conformation of 18C6 is found to be significantly changed depending upon the 18C6-K(+) distance. A proper conformational sampling and an accurate solvent treatment are crucial for realizing the accurate PMF, and we believe that the proposed method is useful to evaluate the PMF in a solution. A discussion upon the PMF in terms of the three-dimensional distribution function for the solvent is also presented.

We recently introduced a method called conveyor belt (CB) thermodynamic integration (TI) for the calculation of alchemical free-energy differences based on molecular dynamics simulations. In the present work, the CBTI approach is generalized to conformational free-energy changes, i.e., to the determination of the potential of mean force (PMF) along a conformational coordinate ξ of interest. The proposed conveyor belt umbrella sampling (CBUS) scheme relies on the parallel simulation of K replicas k = 0,1, ..., K - 1 of the system, with K even. For each replica k, the instantaneous value of ξ is restrained to an anchor value λk. The latter anchor points are equally spaced along a forward-turn-backward-turn path (i.e., a CB) between two extreme values defining the ξ-range of interest. The rotation of the CB is controlled by a variable Λ (range from 0 to 2π) which evolves dynamically along the simulation. The evolution of Λ results from the forces exerted by the restraining potentials on the anchor points, taken equal and opposite to those they exert on the replicas. Because these forces tend to cancel out along the CB, the dynamics of Λ is essentially diffusive, and the continuous distribution of ξ-values sampled by the replica system is automatically close to homogeneous. The latter feature represents an advantage over direct counting (DCNT) and traditional umbrella sampling (TRUS), shared to some extent with replica-exchange umbrella sampling (REUS). In this work, the CBUS scheme is introduced and compared to the three latter schemes in the calculation of 45 standard absolute binding free energies. These correspond to the binding of five alkali cations to three crown ethers in three solvents. Different free-energy estimators are considered for the PMF calculation, and the calculated values are also compared to those of a previous study relying on an alchemical path, as well as to experimental data.

We propose a free energy calculation method for receptor-ligand binding, which have multiple binding poses that avoids exhaustive enumeration of the poses. For systems with multiple binding poses, the standard procedure is to enumerate orientations of the binding poses, restrain the ligand to each orientation, and then, calculate the binding free energies for each binding pose. In this study, we modify a part of the thermodynamic cycle in order to sample a broader conformational space of the ligand in the binding site. This modification leads to more accurate free energy calculation without performing separate free energy simulations for each binding pose. We applied our modification to simple model host-guest systems as a test, which have only two binding poses, by using a single decoupling method (SDM) in implicit solvent. The results showed that the binding free energies obtained from our method without knowing the two binding poses were in good agreement with the benchmark results obtained by explicit enumeration of the binding poses. Our method is applicable to other alchemical binding free energy calculation methods such as the double decoupling method (DDM) in explicit solvent. We performed a calculation for a protein-ligand system with explicit solvent using our modified thermodynamic path. The results of the free energy simulation along our modified path were in good agreement with the results of conventional DDM, which requires a separate binding free energy calculation for each of the binding poses of the example of phenol binding to T4 lysozyme in explicit solvent. © 2019 Wiley Periodicals, Inc.

Molecular dynamics (MD) simulations are widely applied to estimate absolute binding free energies of protein-ligand and protein-protein complexes. A routinely used method for binding free energy calculations with MD is umbrella sampling (US), which calculates the potential of mean force (PMF) along a single reaction coordinate. Surprisingly, in spite of its widespread use, few validation studies have focused on the convergence of the free energy computed along a single path for specific cases, not addressing the reproducibility of such calculations in general. In this work, we therefore investigate the reproducibility and convergence of US along a standard distance-based reaction coordinate for various protein-protein and protein-ligand complexes, following commonly used guidelines for the setup. We show that repeating the complete US workflow can lead to differences of 2-20 kcal/mol in computed binding free energies. We attribute those discrepancies to small differences in the binding pathways. While these differences are unavoidable in the established US protocol, the popularity of the latter could hint at a lack of awareness of such reproducibility problems. To test if the convergence of PMF profiles can be improved if multiple pathways are sampled simultaneously, we performed additional simulations with an adaptive-biasing method, here the accelerated weight histogram (AWH) approach. Indeed, the PMFs obtained from AHW simulations are consistent and reproducible for the systems tested. To the best of our knowledge, our work is the first to attempt a systematic assessment of the pitfalls in one the most widely used protocols for computing binding affinities. We anticipate therefore that our results will provide an incentive for a critical reassessment of the validity of PMFs computed with US, and make a strong case to further benchmark the performance of adaptive-biasing methods for computing binding affinities.

A practical approach that enables one to calculate the standard free energy of binding from a one-dimensional potential of mean force (PMF) is proposed. Umbrella sampling and the weighted histogram analysis method are used to generate a PMF along the reaction coordinate of binding. At each point, a restraint is applied orthogonal to the reaction coordinate to make possible the determination of the volume sampled by the ligand. The free energy of binding from an arbitrary unbound volume to the restrained bound form is calculated from the ratio of the PMF integrated over the bound region to that of the unbound. Adding the free energy changes from the standard-state volume to the unbound volume and from the restrained to the unrestrained bound state gives the standard free energy of binding. Exploration of the best choice of binding paths is also made. This approach is first demonstrated on a model binding system and then tested on the benzamidine-trypsin system for which reasonable agreement with experiment is found. A comparison is made with other methods to obtain the standard free energy of binding from the PMF.

Changes in free energy provide valuable information for molecular recognition, including both ligand-receptor binding thermodynamics and kinetics. Umbrella sampling (US), a widely used free energy calculation method, has long been used to explore the dissociation process of ligand-receptor systems and compute binding free energy. In existing publications, the binding free energy computed from the potential of mean force (PMF) with US simulation mostly yielded "ball park" values with experimental data. However, the computed PMF values are highly influenced by factors such as initial conformations and/or trajectories provided, the reaction coordinate, and sampling of conformational space in each US window. These critical factors have rarely been carefully studied. Here we used US to study the guest aspirin and 1-butanol dissociation processes of β-cyclodextrin (β-CD) and an inhibitor SB2 dissociation from a p38α mitogen-activated protein kinase (MAPK) complex. For β-CD, we used three different β-CD conformations to generate the dissociation path with US windows. For p38α, we generated the dissociation pathway by using accelerated molecular dynamics followed by conformational relaxing with short conventional MD, steered MD, and manual pulling. We found that, even for small β-CD complexes, different β-CD conformations altered the height of the PMF, but the pattern of PMF was not affected if the MD sampling in each US window was well-converged. Because changing the macrocyclic ring conformation needs to rotate dihedral angles in the ring, a bound ligand largely restrains the motion of cyclodextrin. Therefore, once a guest is in the binding site, cyclodextrin cannot freely change its initial conformation, resulting in different absolute heights of the PMF, which cannot be overcome by running excessively long MD simulations for each US window. Moreover, if the US simulations were not converged, the important barrier and minimum were missed. For ligand-protein systems, our studies also suggest that the dissociation trajectories modeled by an enhanced sampling method must maintain a natural molecular movement to avoid biased PMF plots when using US simulations.

We introduce a hybrid projection scheme that combines linear Mori projection and conditional Zwanzig projection techniques and use it to derive a generalized Langevin equation(GLE) for a general interacting many-body system. The resulting GLE includes (i) explicitly the potential of mean force (PMF) that describes the equilibrium distribution of the system in the chosen space of reaction coordinates, (ii) a random force term that explicitly depends on the initial state of the system, and (iii) a memory friction contribution that splits into two parts: a part that is linear in the past reaction-coordinate velocity and a part that is in general nonlinear in the past reaction coordinates but does not depend on velocities. Our hybrid scheme thus combines all desirable properties of the Zwanzig and Mori projection schemes. The nonlinear memory friction contribution is shown to be related to correlations between the reaction-coordinate velocity and the random force. We present a numerical method to compute all parameters of our GLE, in particular the nonlinear memory friction function and the random force distribution, from a trajectory in reaction coordinate space. We apply our method on the dihedral-angle dynamics of a butane molecule in water obtained from atomistic molecular dynamics simulations. For this example, we demonstrate that nonlinear memory friction is present and that the random force exhibits significant non-Gaussian corrections. We also present the derivation of the GLE for multidimensional reaction coordinates that are general functions of all positions in the phase-space of the underlying many-body system; this corresponds to a systematic coarse-graining procedure that preserves not only the correct equilibrium behavior but also the correct dynamics of the coarse-grained system.

We present an extension of the alchemical transfer method (ATM) for the estimation of relative binding free energies of molecular complexes applicable to conventional, as well as scaffold-hopping, alchemical transformations. Named ATM-RBFE, the method is implemented in the free and open-source OpenMM molecular simulation package and aims to provide a simpler and more generally applicable route to the calculation of relative binding free energies than what is currently available. ATM-RBFE is based on sound statistical mechanics theory and a novel coordinate perturbation scheme designed to swap the positions of a pair of ligands such that one is transferred from the bulk solvent to the receptor binding site while the other moves simultaneously in the opposite direction. The calculation is conducted directly in a single solvent box with a system prepared with conventional setup tools, without splitting of electrostatic and nonelectrostatic transformations, and without pairwise soft-core potentials. ATM-RBFE is validated here against the absolute binding free energies of the SAMPL8 GDCC host-guest benchmark set and against protein-ligand benchmark sets that include complexes of the estrogen receptor ERα and those of the methyltransferase EZH2. In each case the method yields self-consistent and converged relative binding free energy estimates in agreement with absolute binding free energies and reference literature values, as well as experimental measurements.

A method is proposed for the estimation of absolute binding free energy of interaction between proteins and ligands. The linear interaction energy method is combined with atom-bond electronegativity equalization method at σπ level Force field (fused into molecular mechanics) and generalized Born continuum model calculation of electrostatic solvation for the estimation of the absolute free energy of binding. The parameters of this method are calibrated by using a training set of 24 HIV-1 protease-inhibitor complexes (PDB entry 1AAQ). A correlation coefficient of 0.93 was obtained with a root mean square deviation of 0.70 kcal mol(-1) . This approach is further tested on seven inhibitor and protease complexes, and it provides small root mean square deviation between the calculated binding free energy and experimental binding free energy without reparametrization. By comparing the radii of gyration and the hydrogen bond distances between ligand and protein of three training model molecules, the consistent comparison result of binding free energy is obtained. It proves that this method of calculating the binding free energy with appropriate structural analysis can be applied to quickly assess new inhibitors of HIV-1 proteases. To test whether the parameters of this method can apply to other drug targets, we have validated this method for the drug target cyclooxygenase-2.

According to implicit ligand theory, the standard binding free energy is an exponential average of the binding potential of mean force (BPMF), an exponential average of the interaction energy between the unbound ligand ensemble and a rigid receptor. Here, we use the fast Fourier transform (FFT) to efficiently evaluate BPMFs by calculating interaction energies when rigid ligand configurations from the unbound ensemble are discretely translated across rigid receptor conformations. Results for standard binding free energies between T4 lysozyme and 141 small organic molecules are in good agreement with previous alchemical calculations based on (1) a flexible complex ( R≈0.9 for 24 systems) and (2) flexible ligand with multiple rigid receptor configurations ( R≈0.8 for 141 systems). While the FFT is routinely used for molecular docking, to our knowledge this is the first time that the algorithm has been used for rigorous binding free energy calculations. © 2017 Wiley Periodicals, Inc.

β-amyloid cleaving enzyme 1 (BACE1) is regarded as an important target of drug design toward the treatment of Alzheimer's disease (AD). In this study, three separate molecular dynamics (MD) simulations and calculations of binding free energies were carried out to comparatively determine the identification mechanism of BACE1 for three inhibitors, 60W, 954 and 60X. The analyses of MD trajectories indicated that the presence of three inhibitors influences the structural stability, flexibility and internal dynamics of BACE1. Binding free energies calculated by using solvated interaction energy (SIE) and molecular mechanics generalized Born surface area (MM-GBSA) methods reveal that the hydrophobic interactions provide decisive forces for inhibitor-BACE1 binding. The calculations of residue-based free energy decomposition suggest that the sidechains of residues L91, D93, S96, V130, Q134, W137, F169 and I179 play key roles in inhibitor-BACE1 binding, which provides a direction for future drug design toward the treatment of AD.

A calculation of the binding free energy for the dimerization of insulin has been performed using the molecular mechanics-generalized Born surface area approach. The calculated absolute binding free energy is -11.9 kcal/mol, in approximate agreement with the experimental value of -7.2 kcal/mol. The results show that the dimerization is mainly due to nonpolar interactions. The role of the hydrogen bonds between the 2 monomers appears to give the direction of the interactions. A per-atom decomposition of the binding free energy has been performed to identify the residues contributing most to the self association free energy. Residues B24-B26 are found to make the largest favorable contributions to the dimerization. Other residues situated at the interface between the 2 monomers were found to make favorable but smaller contributions to the dimerization: Tyr B16, Val B12, and Pro B28, and to an even lesser extent, Gly B23. The energy decomposition on a per-residue basis is in agreement with experimental alanine scanning data. The results obtained from a single trajectory (i.e., the dimer trajectory is also used for the monomer analysis) and 2 trajectories (i.e., separate trajectories are used for the monomer and dimer) are similar.

Since Hansch's extra thermodynamic multi-parameter approach, originally coined as Linear Free Energy Relationship, great efforts in medicinal chemistry have been made to properly estimate the binding free energy. Despite the often small amount, its value is however very critical in determining a successful binding. As a result, its correct estimation may provide a guide for a prospective rational drug design. The calculation of the absolute binding free energies is however a very challenging task as it requires a rigorous treatment of a number of physical terms that are both very time demanding and to some extent not immediately interpretable. In view of this, the introduction of some numerical approximations has permitted to develop the so called Linear Interaction Energy method that, at present, constitutes the best compromise among accuracy, speed of computation and easy interpretation. The initially developed Linear Interaction Energy method was subsequently revisited and several important improvements have been made. Significant examples are the Extended Linear Response, the surface generalized Born LIE, the molecular mechanics generalized Born surface area, the linear interaction energy in continuum electrostatics as well as its quantum mechanics variant. Principles and selected applications of these methods will be herein reviewed.

Rigorous physics-based methods to calculate binding free energies of protein-ligand complexes have become a valued component of structure-based drug design. Relative and absolute binding free energy calculations have been deployed prospectively in support of solving diverse drug discovery challenges. Here we review recent applications of binding free energy calculations to fragment growing and linking, scaffold hopping, binding pose validation, virtual screening, covalent enzyme inhibition, and positional analogue scanning. Furthermore, we discuss the merits of using protein models and highlight recent efforts to replace costly binding free energy calculations with predictions from machine learning models trained on a limited number of free energy perturbation or thermodynamic integration calculations thereby allowing for extended chemical space exploration.