ACS Publications. Most Trusted. Most Cited. Most Read
Quick View

OR SEARCH CITATIONS

My Activity
Publications
CONTENT TYPES

Figure 1Loading Img
Download Hi-Res ImageDownload to MS-PowerPointCite This:J. Chem. Inf. Model. 2022, 62, 22, 5645-5665
RETURN TO ISSUEPREVComputational Bioche...Computational BiochemistryNEXT

Free Energy Calculations Using the Movable Type Method with Molecular Dynamics Driven Protein–Ligand Sampling

  • Wenlang Liu
    Wenlang Liu
    School of Chemistry, Chemical Engineering and Life Science, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    More by Wenlang Liu
  • , 
  • Zhenhao Liu
    Zhenhao Liu
    School of Chemistry, Chemical Engineering and Life Science, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    More by Zhenhao Liu
  • , 
  • Hao Liu
    Hao Liu
    School of Mechanical and Electronic Engineering, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    More by Hao Liu
  • , 
  • Lance M. Westerhoff
    Lance M. Westerhoff
    QuantumBio Inc., State College, Pennsylvania16801, United States
    More by Lance M. Westerhoff
  • , and 
  • Zheng Zheng*
    Zheng Zheng
    School of Chemistry, Chemical Engineering and Life Science, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    QuantumBio Inc., State College, Pennsylvania16801, United States
    *Email: [email protected]
    More by Zheng Zheng
    Orcidhttps://orcid.org/0000-0001-5221-3209
Cite this: J. Chem. Inf. Model. 2022, 62, 22, 5645–5665
Publication Date (Web):October 25, 2022
https://doi.org/10.1021/acs.jcim.2c00278

Copyright © 2022 The Authors. Published by American Chemical Society. This publication is licensed under

CC-BY-NC-ND 4.0.
  • Open Access

Article Views

3162

Altmetric

-

Citations

3
LEARN ABOUT THESE METRICS
PDF (12 MB)
Get e-Alerts
Supporting Info (2)»
SUBJECTS:

Abstract

High Resolution Image
Download MS PowerPoint Slide

Fast and accurate biomolecular free energy estimation has been a significant interest for decades, and with recent advances in computer hardware, interest in new method development in this field has even grown. Thorough configurational state sampling using molecular dynamics (MD) simulations has long been applied to the estimation of the free energy change corresponding to the receptor–ligand complexing process. However, performing large-scale simulation is still a computational burden for the high-throughput hit screening. Among molecular modeling tools, docking and scoring methods are widely used during the early stages of the drug discovery process in that they can rapidly generate discrete receptor–ligand binding modes and their individual binding affinities. Unfortunately, the lack of thorough conformational sampling in docking and scoring protocols leads to difficulty discovering global minimum binding modes on a complicated energy landscape. The Movable Type (MT) method is a novel absolute binding free energy approach which has demonstrated itself to be robust across a wide range of targets and ligands. Traditionally, the MT method is used with protein–ligand binding modes generated with rigid-receptor or flexible-receptor (induced fit) docking protocols; however, these protocols are by their nature less likely to be effective with more highly flexible targets or with those situations in which binding involves multiple step pathways. In these situations, more thorough samplings are required to better explain the free energy of binding. Therefore, to explore the prediction capability and computational efficiency of the MT method when using more thorough protein–ligand conformational sampling protocols, in the present work, we introduced a series of binding mode modeling protocols ranging from conventional docking routines to single-trajectory conventional molecular dynamics (cMD) and parallel Monte Carlo molecular dynamics (MCMD). Through validation against several structurally and mechanistically diverse protein–ligand test sets, we explore the performance of the MT method as a virtual screening tool to work with the docking protocols and as an MD simulation-based binding free energy tool.

This publication is licensed under

CC-BY-NC-ND 4.0.
  • cc licence
  • by licence
  • nc licence
  • nd licence

Introduction

ARTICLE SECTIONS
Jump To
Successful drug discovery generally relies on receptor binding active site determination, ligand virtual screening, receptor–ligand binding mode analysis, and binding free energy estimation. (1−5) Structure based drug design (SBDD) supports this effort with atomic scale knowledge of one or more three-dimensional (3D) molecular structures coupled with receptor–ligand recognition mechanisms discovered from using computational tools. (6−8) Computational (in silico) molecular modeling protocols such as docking and scoring and Monte Carlo (MC) and molecular dynamics (MD) simulations support the needs of large-scale virtual screening during the hit-to-lead stage of drug discovery in which these methods have long been applied to help control the trial-and-error cost of the design and synthesis of drug leads. (9−12)
Docking and scoring methods are widely used during the early stages of the drug discovery process by rapidly generating discrete receptor–ligand binding modes and quickly calculating their individual binding affinities. (13−15) Such methods provide high-throughput virtual screening filters to reduce the number of compounds for synthesis and subsequent in vitro or in vivo evaluation. The goal of docking and scoring is therefore often to capture global minimum binding modes using a “one shot” pose-generation or a restricted local sampling protocol followed by estimation of receptor–ligand binding free energies using the calculated potential energies and conformational entropy-related features according to these binding modes. (16−20) However, the lack of thorough conformational sampling in the docking and scoring computations brings difficulty in exploration of the complicated energy landscape of the protein–ligand binding process. On one hand, correlated motif movements, especially at the receptor–ligand recognition interface, during the binding process compromise the accuracy of the global minimum binding mode prediction from using only local pose-sampling protocols. (21) On the other hand, binding affinity prediction based on a single binding mode conformation usually leads to ligand-size-dependence, in which larger molecules outperform smaller ones simply because there are more intermolecular pairwise contacts. (22) The binding free energy estimation is therefore contaminated by these “one shot” pose-generation strategies due to lack of knowledge of the converged conformational distributions for both bound-state and free-state protein–ligand pairs. However, notwithstanding the inherent disadvantages of the docking and scoring methods, they are still among the most popular SBDD tools because of their computational speed and reasonable binding mode prediction capabilities for many cases. (23,24)
Thorough configurational state sampling using MD simulations has been applied for decades to the estimation of the free energy change corresponding to the receptor–ligand complexing process. These methods are often quite reliable for revealing molecular movement trajectories and generating binding mode probability distributions. (25) However, even today with software optimized for fast central processing units (CPUs) and graphical processing units (GPUs), performing large-scale simulations at scale is still a significant computational burden for high-throughput hit screening. There are two primary binding free energy estimation regimes: absolute and relative, with the former term suggesting the calculation procedures directly generate the binding free energy change (ΔG) involving conformational sampling of end-point states or along the pathway collective variables, and the latter term indicating the protocols first estimating the relative binding free energy change (ΔΔG) of a new molecule referencing a known protein–ligand complex and then generating the binding free energy change (ΔG) using the thermodynamic cycle.
Regarding applicable scenarios and prediction accuracy, relative binding free energy methods benefit from using the thermodynamic cycle that common interaction energy errors among similar molecules can be canceled. Given the binding free energy of a reference ligand, relative binding free energy protocols usually have good accuracy performance from using the free energy perturbation involving alchemical structural transformations. (26,27) FEP and TI based relative binding free energy methods have been recently reported to achieve good simulation with errors around 1 kcal/mol on average against large-scale validations with test sets including congeneric ligands targeting common protein targets. However, relative binding free energy methods usually have limited application in the cases where no known ligands were discovered for the protein target, or structural differences were huge among the molecules. On the other hand, without depending on the reference molecule and structural modifications, absolute binding free energy methods are more robust against wider protein and ligand space but require more comprehensive conformational sampling for both free-state and bound-state receptor and ligand systems.
Fast, effective free energy methods are in high demand to fulfill a growing need in the fast-paced drug discovery efforts. The MovableType (MT) method is a recently introduced absolute binding free energy method which has shown itself to be competitive with other, much more computationally expensive methods by utilizing ensembles of protein–ligand structures generated using docking methods to mimic protein–ligand binding; (28,29,33) but docking is often limited to rigid or semirigid target protein structures, and therefore relying on docking alone can limit the applicability of MT to fairly rigid protein systems. In computational chemistry, there is a number of binding mode sampling methods available ranging from docking to large-scale molecular dynamics protocols which are generally distributed on a spectrum of performance between high efficiency and high accuracy. In this study, we evaluate the impact of the chosen sampling protocol on protein–ligand binding free energy estimation by challenging the MT method. Two validation benchmarks are used in this study including the well-known CASF-2016 set which covers a wide variety of protein and ligand structures (15) and the recently published Merck KGaA benchmark which consists of a set of 264 ligands with minor structural differences targeting eight pharmaceutically valuable receptors. (44) Since the MT free energy method has been developed to use bound-state and free-state conformations of protein ligand molecular systems for the evaluation of absolute binding free energies, we use the CASF and Merck KGaA sets to explore the compatibility of the MT free energy method working with different sampling methods and show that the method is applicable not only as a virtual screening algorithm by using docking generated poses but also as a flexible free energy evaluation algorithm by using protein–ligand complex trajectories from molecular dynamics simulations.

Method

ARTICLE SECTIONS
Jump To

The MT Method for Protein–Ligand Absolute Binding Free Energy Estimation

The MT method is a fast absolute binding free energy approach showing its robustness across a wide range of targets and ligands. It combines global sampling methods for generating ensembles of free-state and bound-state protein–ligand molecular conformations with a novel local sampling method which is used to numerically estimate local partition functions with respect to these sampled conformations. The current version of the MovableType software (available at http://www.quantumbioinc.com/) contains two protocols for the protein–ligand binding free energy estimation. The full protocol, denoted as MTScoreE, performs partition function estimations for the bound-state complex (holo) and free-state protein (apo) and ligand and then calculates the protein–ligand binding free energy with the ratio of the approximate partition functions. On the other hand, the simplified protocol, denoted as MTScoreES, predicts the protein–ligand binding affinity only using one single accurate complex conformation as the representative energy state. This simplified protocol is much faster than the full protocol and is positioned for higher-throughput virtual screening tasks. In this project, we applied the full MT protocol, MTScoreE, as the free energy evaluation tool coupled with different conformational sampling approaches for the estimation of the protein–ligand binding free energies. Its computational procedure is explained in the following paragraphs. The simplified protocol, MTScoreES, is also briefly summarized at the end of this subsection. The MTScoreES protocol is applied to the linear scaling parameter set generation with respect to the results rescale during the Merck KGaA set evaluation. This protocol is also applied to score the energy-minimal snapshots selected from the simulation trajectories to compare with the conformational ensemble-based binding free energy estimation results.
The full MT protocol first employs built-in or external molecular energy-state sampling methods (e.g., docking or dynamics) to generate significant free-state and bound-state protein–ligand conformations. Then, for each sampled molecular conformation, local numerical integrations are performed to generate an atomic pairwise potential ensemble which samples the local space of the atomic coordinates of the globally sampled molecular conformation. The local pairwise potential integrations are multiplied through all the atom pairs in the molecular conformer (e.g., protein, ligand, and complex) to estimate the local molecular energy states in the vicinity centered at each sampled conformation. These local energy states approximate local molecular partition functions, and when the MT method is supplied with multiple sampled conformations of the protein (apo), ligand, and complex (holo), the method sums over all the local partition functions regarding the free-state and bound-state NVT ensembles. This calculation yields an estimate of the approximated total partition functions of the protein–ligand system before and after complexation. The ratio of the partition function approximations is then used to estimate the gas-phase binding free energy as represented in eq 1:
Δ G b i n d i n g g a s ≈ − R T log ( Z P L Z P Z L )
(1)
In this equation, each partition function Z P , Z L , and Z P L for protein (apo), ligand, and complex (holo) respectively is calculated using eq 2
Z M = ∑ α N poses Z M α = Z M 1 + Z M 2 + ... + Z M N
(2)
where M indicates the sampled molecule at different states, and every Z M i is estimated using numerical integration against all the atom pairwise interactions of each sampled conformational state i. For this study, we employed DivCon Discovery Suite v.815 in which potential energies for every sampled conformation along with the partition functions represented in eqs 1 and 2 are calculated using two sets of energy functions implemented in the DivCon Discovery Suite: (1) GARF energy function developed by Zheng et al. (GARF set) (30) and (2) Amber ff19SB and General AMBER Force Field (GAFF) force field optimized for the MT method (Amber set). (31,32) When Amber ff19SB was employed, the ligand atomic charges were suppled as Mullikan charges calculated using the built-in PM6 Hamiltonian.
This MT method, referred to in the balance of this work as MTScoreE or ensemble-score, employs conformational sampling protocols to generate energy-state ensembles for a protein–ligand pair in both unbound state and bound state, and it approximates the unbound-state and bound-state NVT partition functions using the generated energy-state ensembles to derive the Helmholtz free energy using a ratio of partition functions.
In this work, the partition function approximation for the bound-state complex is calculated as
Z P L N P L = ∑ i N P L Z P L i = Z P L 1 + Z P L 2 + ... + Z P L N P L
(3)
where NPL denotes the number of sampled complex-state conformations. The conformational sampling protocols introduced in this work for the bound-state and free-state partition function approximations will be visited in the following subsection.
Similarly, the partition function approximations for the free-state protein and ligand are calculated respectively as
Z P N P = ∑ j N P Z P j = Z P 1 + Z P 2 + ... + Z P N P
(4)
and
Z L N L = ∑ k N L Z L k = Z L 1 + Z L 2 + ... + Z L N L
(5)
NP and NL in eqs 4 and 5 denote the numbers of sampled free-state protein and ligand conformations, respectively. Applying the approximated partition functions Z P L N P L , Z P N P , and Z L N L to eq 6, the gas-phase protein–ligand binding free energy is then calculated using the logarithm of the ratio of bound-state and free-state partition functions.
Δ G b i n d i n g g a s ≈ − R T log ( Z P L N P L Z P N P Z L N L )
(6)
The KMTISM solvation model is then employed for the solvation free energy change during the binding procedure. (56) KMTISM is an implicit solvation model that generates multiple solvent layers around each sampled solute conformation and calculates the solvent–solute interaction potentials by adding up the layer-based solvent–solute contact energies. Ensembles of solvent–solute interaction potentials are then generated for the protein–ligand complex and the free-state protein and ligand. Hence the solvation free energy change is calculated using the following equation
Δ G s o l = 1 − β ln [ ∑ i exp ( − β ( E s o l P L ) i ) ] − 1 − β ln [ ∑ j exp ( − β ( E s o l P ) j ) ] − 1 − β ln [ ∑ k exp ( − β ( E s o l L ) k ) ]
(7)
where (EsolPL)i, (EsolP)j, and (EsolL)k are the solvent–solute interaction potentials for the sampled complex, protein, and ligand conformations, respectively. i, j, and k are the index numbers for the corresponding conformations. The protein–ligand binding free energy is then calculated as
Δ G b i n d i n g = Δ G b i n d i n g g a s + Δ G s o l
(8)
With more careful and thorough sampling protocols, the MT method has a better chance to obtain the most significant conformations in the molecular conformational ensemble. On the other hand, we can also introduce fast and restricted sampling methods to the MT method for higher-throughput virtual screening tasks. In addition to MTScoreE, a simplified MT protocol named MTScoreES, where “ES” denotes “end-state”, has also been developed which utilizes a single binding pose as the representative energy state for the binding free energy prediction. (33) This approach only calculates the intermolecular atom pairwise potentials between the protein and ligand pair referencing the input protein–ligand binding pose. The binding free energy is then approximated in eq 9
Δ G b i n d i n g ≈ − R T log ( Z I n t e r P L 0 )
(9)
where Z I n t e r P L 0 is the protein–ligand binding pose’s local partition function considering only the intermolecular atomic pairwise interaction potential.

Conformation Sampling Protocols

When provided proper ensembles of molecular conformations for unbound-state and bound-state protein ligand systems, MT is theoretically compatible with many different conformational sampling protocols. In our previous studies, we mainly focused on introducing docking protocols to work with MT for virtual screening purposes. (33) In the present work, we further explore the performance of the MT method when it is challenged with conformers generated with different simulation-based conformational sampling protocols. While docking+MT is robust and flexible for many cases, without thorough configurational space coverage or reliable energy gradient searching, discrete conformational sampling protocols have inherent limitations locating significant molecular configurations. One would expect that introducing more globally significant conformations or landscape minima should improve the free energy prediction accuracy of the MT method with the addition of more converged energy states. Therefore, in addition to docking, this work utilizes conventional MD (cMD) and Monte Carlo MD (MCMD) simulations. Protein–ligand complex “snapshots” from these sampling simulations are provided to MT to improve binding free energy estimation accuracy. With this support in place, we evaluate the MT method as a flexible free energy tool for its compatibility with different energy functions and sampling strategies. This work open further opportunities for using this method to analyze MD trajectories and receptor–ligand complexing mechanisms.
The MTCS and MTFlex protocols are built-in modules in the DivCon Discovery Suite to generate the free-state ligand and protein conformations based on corresponding prebuilt small-molecule and amino-acid rotamer libraries, respectively, followed by local structural minimization. (34,35) Both modules contain corresponding molecular fragment libraries, in which every fragment is collected from structural fragmentation against the ligands and proteins from Protein Data Bank. Every molecular fragment contains one single rotatable bond, along which the rotational energy curve for every fragment is precalculated using the built-in energy functions, i.e., Amber ff19SB and General AMBER Force Field (GAFF) force field, or GARF energy function. When using the MTCS protocol for the conformational sampling of ligands, the energy-minimum rotamers for necessary structural fragments from the libraries are selected to construct the conformations of the corresponding molecule. Local energy minimization is then performed to generate reasonable conformations according to their energy-based scores. On the other hand, due to the vast conformational space of proteins, MTFlex performs local conformational search based on the input protein configuration, instead of constructing the protein conformations totally based on fragment rotamer assembly in a “de novo” manner. For every amino-acid conformation from the input protein conformation, the MTFlex protocol selects the close-by energy-minimum rotamers and hence assembles the protein conformation, followed by local structural minimization. In such a case, the MTFlex protocol generates a cluster of conformations, resembling the configurational features of the input protein structures.
In this project, we select the same number of free-state protein conformations as the number of bound-state complex conformations for the corresponding partition function approximations (NP = NPL), mostly because the protein usually dominates conformational space in a protein–ligand complex. To explore the convergence of the approximated partition function estimation, and hence the reliability of the ΔGbinding prediction, we validate the ΔGbinding estimation convergence by introducing different numbers of sampled conformations, to decide the proper NP and NPL numbers for sampling the bound-state complex and the free-state protein. See the detailed validation procedure and result analysis in the Supporting Information.
On the other hand, following the default setting introduced in our previous work, (34) the top five lowest-energy conformers for the free-state ligand are selected from the MTCS protocol. Compared to the number of protein conformations, much fewer free-state ligand conformations are chosen because the conformational energies for most ligands increase significantly soon after the top 3 to 5 conformations.
We employed three different protocols to generate binding modes for use in this project: high-throughput with induced-fit docking as implemented in the Molecular Operating Environment (MOE) v2019 package; (36)exhaustive with molecular dynamics simulations using the Amber18 suite; and balanced (sampling efficiency and exhaustiveness) with MCMD as captured with the parallel Consecutive Histograms Monte Carlo (CHMC) method combined with MD simulation. The following conformational sampling methods are employed for the protein–ligand binding modes generation:

MT with MOE Induced Fit Docking

The coordinate files utilized in the present study structures were downloaded from the RCSB/PDB, and preparation was performed using the Molecular Operating Environment (MOE) v2019 package (36) using the Structure Prepare application followed by the Protonate3D application to add and optimize protons. After structure preparation was complete, docking for this protocol was performed using the same version of MOE. For this calculation, the ligand parameters were calculated using the Extended Hückel Theory (Amber10:EHT) as implemented in MOE, and the “Induced Fit” docking (IFD) protocol was chosen for each protein–ligand model. Under the IFD protocol, initial placement was performed using the Triangle Matcher approach, and the London ΔG score was used as the initial score followed by the GBVI/WSA ΔG score used as the final filter. (37) In each case and in order to maximize the number of initial poses, 250 poses were provided by the Triangle Matcher, and the active site was optimized using the Amber10:EHT force field as implemented in MOE. Finally, 25 poses were passed to MTScoreE as landscape minima for ensemble scoring.

MT with the Conventional Molecular Dynamics (cMD) Simulation

Molecular dynamics simulations for all the protein–ligand complexes were performed using Amber18 with the ff99SB force field. (38,39) The Generalized AMBER Force Field (GAFF) was used for ligands and for nonstandard residues as characterized using the Antechamber program. Parameters for any metal ions (e.g., Zn2+ Mg2+, and Ca2+) were introduced from the work of Li et al., (40) and the partial charges of ligands and nonstandard residues were calculated using the Antechamber program with the AM1-BCC method. (54,55) For each molecular simulation, the initial conformation for each complex was the best MOE docking pose (determined using the GBVI/WSA ΔG score) generated from the above-noted induced-fit docking protocol. The complex was solvated in a rectangular TIP3P water box with periodic boundary conditions and neutralized with Na+ or Cl- (as required) using the tleap program. The SHAKE algorithm (41) was applied to constrain bonds containing hydrogen atoms with a 2 fs time step, and the electrostatic interactions were calculated using the PME (particle mesh Ewald) (42) with a cutoff of 10.0 Å for long-range interactions. Prior to beginning the simulation, each complex was treated with an energy minimization consisting of 10000 steps of steepest descent algorithm followed by 10000 steps conjugate gradient algorithm. Next, each molecular model was gradually heated from 0 to 300 K in the isothermal-isovolumetric (NVT) ensemble for 200 ps and then equilibrated for 1 ns in the isothermal–isobaric (NPT) ensemble (300 K and 1.01325 MPa). Finally, the molecular dynamics simulations were carried out in the NPT ensemble (300 K).

MT with the Consecutive Histograms Monte Carlo (CHMC) Protocol Followed by Parallel cMD Simulation

Recently, our laboratory developed the CHMC method employing a mixed Monte Carlo/Molecular Dynamics protocol to estimate the potential of mean force (PMF) regarding the protein–ligand binding or dissociation process. (43) CHMC first generates a series of consecutive Monte Carlo sampling units with equal volume representing the receptor–ligand association/dissociation space. These sampling units are stacked from within the receptor’s binding active site (which is designated as the ligand’s initial position), along a dissociation path vector extending to the ligand’s final location which is usually in the open space representing the fully dissociated state for the protein–ligand system.
In this study, we divided each protein model’s active site into five sampling units with equal volume. Each sampling unit was then used for modeling equivalent and comparable NVT ensembles for the estimation of the partition functions in each sampling unit. CHMC generates a set of initial binding modes within each sampling unit using a docking-like protocol, and in the present project, the top 50 binding modes (based on the protein–ligand interaction energies) per sampling unit were selected for the next calculation stage. Geometric minimization was performed against these binding modes using the Newton–Raphson method to obtain an appropriate receptor–ligand complex ensemble for each sampling unit. The top-five separate minimum-energy protein–ligand complex geometries at each of the sampling units were chosen based on the protein–ligand binding mode energies. According to the free energy estimation protocols introduced in the following subsection, Amber and GARF energy functions were both introduced in the geometric minimization and protein–ligand binding mode energy calculations. The selected five protein–ligand complex geometries were thus used as initial conformations for 50 ns cMD simulations using the same setup, structure preparation, and protocol described in the “MT with Conventional Molecular Dynamics (cMD) Simulation” subsection above.

General Setting for the MT Free Energy Method

In this work, we utilized the three sampling protocols mentioned above coupled with two pairwise potential functions available in the MT package for protein–ligand binding free energy estimation. In the following paragraphs, we use “MOE-MT-amber” to designate the MT ensemble scoring protocol challenged with MOE induced-fit docking poses in which the partition function is calculated using the Amber energy function; “MOE-MT-garf” to designate the MT protocol with MOE docking poses and the GARF energy function; “cMD-MT-amber” to designate the MT protocol performing 250 ns conventional MD simulation and calculating the MT partition functions using the Amber energy function; the term “cMD-MT-garf” denoting the MT protocol performing 250 ns conventional MD simulation and calculating the MT partition functions using the GARF energy function; the term “CHMC-cMD-MT-amber” denoting the MT protocol performing CHMC parallel sampling using the Amber energy function followed by 50 ns conventional MD simulation and calculating the MT partition functions using the Amber energy function; the term “CHMC-cMD-MT-garf” denoting the MT protocol performing CHMC parallel sampling using the GARF energy function followed by 50 ns conventional MD simulation and calculating the MT partition functions using the GARF energy function.
From the MOE docking process, 25 top ranked binding poses are selected as energy minima for the bound-state partition function estimation in the MT method. From the MD simulations, including the cMD and CHMC-cMD calculations, 500 equally spaced snapshots of every protein–ligand complex (every 10 ps) are selected from the converged region of the production MD trajectories. For each protein–ligand complex, 5 different sets of the 500 snapshots from the converged MD trajectories are selected for the estimation of error bars (standard deviations) and 90% confidence intervals.

Results and Discussion

ARTICLE SECTIONS
Jump To
We applied the MT free energy method and different conformational sampling protocols through the treatment of two validation benchmarks: (1) the industry-standard Comparative Assessment of Scoring Functions v2016 set (the CASF-2016 benchmark) containing 57 protein targets and 285 ligands (15) and (2) the validation benchmark published by Merck KGaA originally for the evaluation of Schrödinger FEP+ including 8 protein targets and 264 ligands (the Merck benchmark). (44) The CASF-2016 benchmark covers various protein and ligand structures for validating the robustness of different sampling protocols across a broad range of protein and ligand classes. On the other hand, the Merck benchmark focuses on pharmaceutically relevant targets and ligand series containing ligand structures with subtle structural differences. (45−53)Figure 1 summarizes the calculation procedures applied to the CASF-2016 benchmark and the Merck benchmark.

Figure 1

Figure 1. Illustration of the free energy estimation protocols applied to (A) the CASF-2016 benchmark and (B) the Merck benchmark.

High Resolution Image
Download MS PowerPoint Slide
For statistical analysis of the free energy calculation results compared to the experimental binding free energies, we utilized Pearson’s R and Kendall’s τ for the measure of linear and rank correlation between the predicted and experimental binding free energies and Mean Absolute Error (MAE) for measuring the calculation errors. They are calculated using the following equations respectively
R = ∑ i = 1 n ( x i − x ¯ ) ( y i − y ¯ ) ∑ i = 1 n ( x i − x ¯ ) 2 ∑ i = 1 n ( y i − y ¯ ) 2
(10)
τ = 2 ∑ i < j s g n ( x i − x j ) s g n ( y i − y j ) n ( n − 1 )
(11)
where sgn (xi – xj) = −1 if xi < xj, and sgn (xi – xj) = 1 if xi > xj, and analogously for y.
M A E = ∑ i = 1 n | ( y i − x i ) | n
(12)
To compare the error performance of the MT protocols with other published free energy methods, for all the target sets from the Merck benchmark, we also utilized Root-Mean-Square Error (RMSE) evaluated using the following equation:
R M S E = ∑ i = 1 n ( y i − x i ) 2 n
(13)
From eqs 10 to 13, xi represents a predicted binding free energy, and yi is the corresponding experimental value.

The CASF-2016 benchmark Results Analysis

The statistical results of the CASF-2016 benchmark for four MT free energy protocols (cMD-MT-amber, cMD-MT-garf, CHMC-cMD-MT-amber, and CHMC-cMD-MT-garf) are summarized in Table 1, and these results are plotted for each protocol in Figure 2.
Table 1. Results for Protein-Ligand Binding Free Energy Estimation Using Different MT-Based Protocols
MT protocols Pearson’s R Kendall’s τ MAE (kcal/mol)
cMD-MT-amber 0.66 0.49 2.10
cMD-MT-garf 0.70 0.52 1.68
CHMC-cMD-MT-amber 0.66 0.49 2.10
CHMC-cMD-MT-garf 0.71 0.52 1.67

Figure 2

Figure 2. Experimental binding free energies versus predicted binding free energies for the 285 protein–ligand complexes from the CASF-2016 test benchmark. Standard deviations for the test cases were shown as error bars.

High Resolution Image
Download MS PowerPoint Slide
As noted in the Methods section, the top-scored docked pose from MOE was chosen as the initial complex conformation for both the cMD-MT-amber and the cMD-MT-garf protocols, and the CHMC method was used to generate five different initial complex conformations to initiate the MD simulations in the CHMC-cMD-MT-amber and CHMC-cMD-MT-garf protocols. Using the MT-AMBERff19SB force field, both the cMD-MT-amber and the CHMC-cMD-MT-amber protocols achieve Pearson’s R values of 0.66, MAE values of 2.10 kcal/mol, and Kendall’s τ values of 0.49. Using the GARF energy function, the cMD-MT-garf and CHMC-cMD-MT-garf protocols generated improved results, with Pearson’s R values of 0.70 and 0.71, MAE values of 1.68 and 1.67 kcal/mol, and Kendall’s τ values of 0.52. Introducing simulation-based sampling methods in this study provided us with converged energy ensembles of each protein–ligand complex. Therefore, the initial conformations from different sampling methods were converged through the MD simulation into similar conformational ensembles, resulting in similar free energy estimation values. Figure 3 illustrated the free energies obtained from both cMD-MT and CHMC-cMD-MT against the CASF-2016 benchmark and discovered very high correlations between these two protocols when using the same energy functions.

Figure 3

Figure 3. Correlations between predicted binding free energies using cMD-MT and CHMC-cMD-MT protocols against the CASF-2016 benchmark. The correlation on the left illustrates the predicted binding free energies using cMD-MT-amber (x-axis) vs CHMC-cMD-MT-amber (y-axis). The correlation on the right illustrates the predicted binding free energies using cMD-MT-garf (x-axis) vs CHMC-cMD-MT-garf (y-axis).

High Resolution Image
Download MS PowerPoint Slide
In 2020, we reported free energy estimation results using MOE-MT-amber and MOE-MT-garf against the CASF-2016 benchmark. (33) In this prior work, based on 25 docking poses for each ligand, the MOE-MT-amber and MOE-MT-garf protocols generated Pearson’s R values of 0.63 and 0.64, respectively. In another publication, we reported the binding free energy reproduction against the CASF-2016 benchmark using the CHMC method with the GARF energy function, achieving the Pearson’s R as 0.61 and the MAE as 2.06 kcal/mol. (42) The present project shows that MD simulation-based sampling methods improve MT binding free energy prediction accuracy compared to the docking methods. Furthermore, the CHMC-cMD-MT-garf protocol shows that the free energy estimation accuracy with the parallel sampling strategy in CHMC is improved by introducing multiple short MD simulations and pairing this increased sampling with numerical integration and atom pairwise local sampling in MT.
While the present work represents the most exhaustive exploration of the impact of improved sampling on MT-based free energy predictions performed to date, there are still potential sources of error in these receptor–ligand binding free energy calculations. Specifically, the presented results utilized bound-state conformational sampling and therefore calculation of the bound, ZPL partition function; however, accurate binding free energy estimation also depends on conformational sampling for the free-state receptor and ligand molecules along with the change in solvation free energy during the binding process. Therefore, against test cases with flexible receptor and ligand structures, the lack of exhaustive conformational sampling for the free-state protein and ligand may introduce significant error in the predicted binding free energies. In the present effort, we focused on comparing the different binding mode sampling protocols by looking at the calculation results against the subsets targeting the identical protein receptor. In subsequent work, we will explore the question of whether we can further improve our predictions if we use apo as well as holo protein sampling.
The CASF-2016 benchmark contains 57 subsets, each of which has five different ligands targeting the same protein receptor. In our previous publication, we reported the binding free energy estimation results of 52 out of the 57 subsets from the CASF-2016 benchmark when the MT method was provided with 25 MOE dock-based poses for each of the protein:ligand complexes in those 52 subsets. To compare the performance of these earlier MOE-MT calculations with the present cMD-MT and CHMC-cMD-MT calculations on a case-by-case basis, Figure 4 shows the Pearson’s R values ranked according to each protocols’ performance and represented in pairs. For the cMD-MT-amber protocol, 36 subsets had Pearson’s R values higher than 0.7, and 40 subsets had Pearson’s R values higher than 0.6. Using the parallel, CHMC-cMD-MT-amber simulation protocol, 35 subsets had Pearson’s R values higher than 0.7, and 40 subsets had Pearson’s R values higher than 0.6. These results compared to the earlier MOE-MT-amber protocol in which 31 subsets had Pearson’s R values higher than 0.7, and 33 subsets had Pearson’s R values higher than 0.6. When performing the cMD-MT-garf protocol, 35 subsets had Pearson’s R values higher than 0.7 and 41 subsets had Pearson’s R values higher than 0.6. Using the parallel CHMC-cMD-MT-garf simulation protocol, 36 subsets had Pearson’s R values higher than 0.7, and 41 subsets had Pearson’s R values higher than 0.6. This compares favorably to the original MOE-MT-garf protocol for which 29 subsets had Pearson’s R values higher than 0.7, and 32 subsets had Pearson’s R values higher than 0.6.

Figure 4

Figure 4. Pearson’s R coefficients comparisons for simulation-based and docking-based protocols against 52 subsets from the CASF-2016 benchmark. Pearson’s R values for each protocol were ranked individually in their own descending orders.

High Resolution Image
Download MS PowerPoint Slide
From the calculation results comparison regarding the CASF-2016 subsets, we observe that the free energy protocols with simulation-based sampling methods generally obtain better agreement with the experimental data. When using the GARF energy function, the advantage for performing simulation-based sampling methods is even more pronounced. Moreover, the single-trajectory, exhaustive (250 ns) MD simulations (cMD-MT-amber and cMD-MT-garf) and the multitrajectory, short-time (50 ns) MD simulations (CHMC-cMD-MT-amber and CHMC-cMD-MT-garf) achieved similar free energy estimation accuracies suggesting that the free energy estimation scheme using multiple local samplings followed by parallel short-time MD simulations could obtain comparable accuracy to the free energy estimation scheme using the normal, longer running single-trajectory simulation strategy.
In the CASF-2016 benchmark, each subset contains only 5 test cases making the Pearson’s R values sensitive to protein–ligand complex selection and individual outliers, but with a sufficiently large test set, we can still obtain meaningful statistical results from the subsets study. The simulation-based methods achieved Pearson’s R values higher than 0.7 for 67% to 69% of all the 52 subsets and Pearson’s R values higher than 0.6 for 77% to 79% of the 52 subsets. The simulation-based MT free energy protocols show significant correlations with the experimental binding free energies in most of the homologous protein–ligand test sets in the CASF-2016 benchmark, and these methods show improved prediction profiles compared to the previous docking-based MT methods. To further explore the performance of the MT free energy method with different binding-mode sampling protocols, we employed another public validation benchmark, first introduced by Merck KGaA for FEP+ evaluation, which includes 8 protein targets and their corresponding congeneric ligand series.

The Merck KGaA Benchmark Results Analysis

The Merck KGaA benchmark encompasses 264 ligand structures divided between 8 protein targets including the following: 33 ligands for cyclin dependent kinase 8 (CDK8), 24 ligands for tyrosine-protein kinase met (c-MET), 28 ligands for kinesin-5 (EG5), 42 ligands for hypoxia-inducible factor 2α (HIF2α), 40 ligands for 6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 3 (PFKFB3), 26 ligands for tyrosine phosphatase (SHP2), 44 ligands for tyrosine-protein kinase (SYK), and 27 ligands for tankyrase 2 (TNKS2). The ligands for each congeneric series have high structural similarity but significantly different experimental protein–ligand binding free energies. For the most part, in preparation for this benchmark, Schindler et al. used published PDB structures for each target and coupled these models with the associated ligands. In the case of EG5, an alternative loop conformation at the binding site was also considered. This structure used different initial conformations for this loop which significantly impacted the binding affinity prediction accuracy for other free energy methods considered in the original paper. (44) We likewise applied our MT protocols to both Eg5 target sets and evaluated the results separately.
The sampling protocols and the free energy calculations using the MT method were performed following the same protocols as discussed in the Method section. To better illustrate and compare the performance of different sampling protocols applied to the receptor–ligand complex state and to compensate for the lack of apo-state conformational sampling in the absolute binding free energy protocols, the raw MT results were rescaled. We selected 8 sets of protein–ligand cocrystal structures from the PDBBind database v2019 having the same 8 corresponding protein targets as the Merck benchmark but with no overlap in the ligand series selection. This curation yielded 191 cocrystal structures with known binding affinities which were passed to the MTScoreES protocol for fast binding free energy estimation (see details in the Supporting Information). The resulting regression lines for each target-ligand series were obtained by scoring against these cocrystal structures with both the MT-AMBERFF19SB and the GARF energy functions, and the MT free energy calculation results (ΔGpred) were refitted to the corresponding regression lines as the rescaled binding free energy values (ΔGpred*) to reproduce the experimentally determined free energies magnitudes.
Δ G p r e d * = a × Δ G p r e d + b
(14)
Figure 1B illustrates the calculation procedure for all the target sets from the Merck benchmark, and this procedure yields results for the 9 target sets (including the 8 primary targets along with the EG5 target with the alternative loop) which are summarized in Tables 2, 3, and 4 (Detailed plots for each set can be found in Figure 5.). Generally, all our free energy protocols achieved acceptable to good correlations and error performance against the whole benchmark. The two MOEdock-based protocols using MT-AMBERFF19SB and GARF MT energy functions exhibit very similar correlation coefficients against the whole benchmark, with Pearson’s R as 0.551 and 0.554 and Kendall’s τ as 0.433 and 0.434 for the MOE-MT-amber and MOE-MT-garf protocols, respectively. The MD simulations bring significant improvement to the MT binding free energy estimation accuracy with Pearson’s R range from 0.620 to 0.658, and Kendall’s τ ranged from 0.459 to 0.510. Meanwhile, except for the MOE-MT-amber protocol, the rescaled results exhibited overall errors with the RMSE values which range between 1.2 and 1.5 kcal/mol.
Table 2. Pearson’s R Values with 90% Confidence Intervals for All the Target Sets Using Different Free Energy Protocols
Table 3. Kendall’s τ Values with 90% Confidence Intervals for All the Target Sets Using Different Free Energy Protocols
Table 4. RMSE Values in kcal/mol with 90% Confidence Intervals for All the Target Sets Using Different Free Energy Protocols

Figure 5

Figure 5. Correlation between predicted binding free energies and experimental values for all eight target sets using the 6 different free energy protocols (For simplicity, the results for the Eg5 target set using the alternative loop are not included.).

High Resolution Image
Download MS PowerPoint Slide
When considering each individual target subset, the MOEdock-based and simulation-based MT protocols generate medium to good correlations for the CDK8, c-Met, Eg5, and TNKS2 sets with Pearson’s R values ranging from 0.394 to 0.852. The additional sampling provided by MD leads to significant improvement in the binding free energy estimation for all target subsets. In particular, PFKFB3 benefits most from this additional sampling. For example, when considering the GARF potential, the Pearson’s R values improve from 0.056 for MOE-MT-garf to 0.840 for cMD-MT-garf and 0.741 for CHMC-cMD-MT-garf. However, HIF-2α shows uneven performance in which the cMD-MT-garf protocol obtains Pearson’s R = 0.507, while all of the other protocols’ Pearson’s R values are less than 0.3. Against the 9 target sets, the cMD-MT-amber protocol generated high correlations (Pearson’s R ≥ 0.6) for 3 sets and medium correlations (0.4 ≤ Pearson’s R ≤ 0.6) for 3 sets, and the cMD-MT-garf protocol generated high correlations (Pearson’s R ≥ 0.6) for 3 sets and medium correlations (0.4 ≤ Pearson’s R ≤ 0.6) for 6 sets. When adding the CHMC Monte Carlo protocol, the CHMC-cMD-MT-amber protocol generated high correlations (Pearson’s R ≥ 0.6) for 5 sets and medium correlations (0.4 ≤ Pearson’s R ≤ 0.6) for 1 set, and the CHMC-cMD-MT-garf protocol generated high correlations (Pearson’s R ≥ 0.6) for 4 sets and medium correlations (0.4 ≤ Pearson’s R ≤ 0.6) for 3 set. This was in stark contrast to the dock-only based versions for this benchmark in that without thorough binding mode sampling, the docking-based protocols did not achieve as good of a ranking performance. The MOE-MT-amber protocol generated high correlations (Pearson’s R ≥ 0.6) for 1 set and medium correlations (0.4 ≤ Pearson’s R ≤ 0.6) for 4 sets, and the MOE-MT-garf protocol generated high correlations (Pearson’s R ≥ 0.6) for 1 set and medium correlations (0.4 ≤ Pearson’s R ≤ 0.6) for 3 sets.
As shown in Table 5, the rescaling procedure - which was performed against 8 separately curated sets of protein–ligand complexes - made significant impact to the protein–ligand binding free energy ranking performance when evaluating the whole benchmark including all the target sets. This improvement is observed both in the correlation coefficients and in error reduction. This is especially true for ranking across different target sets. Against the whole benchmark, utilizing rescaling increased the Pearson’s R and Kendall’s τ values for all protocols except for the two MOE docking-based protocols, and all the simulation-based protocols had reduced RMSE values after introducing rescaling.
Table 5. Statistical Results for All 6 Free Energy Protocols against the Merck Benchmark before and after Introducing the Rescaling Procedure
Without rescaling, the raw estimation results showed reasonable binding free energy ranking performance for the test cases within most target sets, while struggling to rank among different target sets. Relative binding free energy estimation between protein–ligand complexes has always been a challenge for both absolute and relative binding free energy methods (especially when there were major structural differences between complex molecules). Reaction-pathway dependent free energy methods rely heavily on the free energy map simulation of the sampling windows along purposely designed alchemical structural mutation pathways or protein–ligand dissociation/binding pathways. Significant structural differences between protein–ligand molecular systems require simulations with respect to high-dimensional collective variables (CVs) which heavily burden the convergence of molecular simulations. On the other hand, end-point free energy protocols usually depend on the generation of conformational probabilities for representative energy states or numerical energy characterization of significant conformations. A major structural difference often introduces large, distinct conformational changing features among different molecules which negatively impacts the conformational state sampling thoroughness for bound-state protein, free-state protein, ligand molecules, and ultimately the conformational energy calculation accuracy.
The MT method improves accuracy by using a ratio of the bound-state and free-state partition functions for the estimation of the binding free energies. This way common computational errors within both bound-state and free-state partition functions cancel out during the free energy calculation procedure. To better capture the significant binding modes, we can utilize docking procedures (which are often limited to rigid-protein structures) or - as in the case of the present work - different MD simulation methods performed on the bound-state protein–ligand complexes to cover more sampling space. For each individual target set, we clearly observe improvement using the simulation-based MT protocols compared to the docking-based MT protocols as summarized in Tables 2, 3 and 4. However, the new MT protocols applied in this study still lack the ranking capability for the protein–ligand complexes with distinct protein targets which suggests that extra simulation procedures are needed to capture the free energy baselines introduced by different proteins. The rescaling procedure is a useful tool for the absolute binding free energy protocols to improve the binding free energies ranking capabilities across different protein targets. While simulation-based protocols all had their prediction errors reduced, the two MOEdock-based protocols had increased errors after introducing the rescaling procedure. This suggests that with the same protein and different ligand structures, compared to dock-based protocols, simulation-based protocols yield similar conformational and interaction features within and between binding site protein residues versus the corresponding protein target found in the cocrystal structures from rescaling data sets. Therefore, the energy baseline generated by the MOE IFD protocol for each protein target significantly deviates from the energy baseline from the corresponding target sets used in the rescaling procedure.
When considering the computational time for performing all the free energy protocols against every protein–ligand complex from the Merck benchmark, the MD simulation was performed on a computer with 4 × NVIDIA Tesla V100 16GB (900 GB/s) GPUs and 128GB Memory, and the MOEdock was performed on a computer with an E5-2640 20-Core v4 (2.4 GHz) and 128GB Memory. The average computing time required for every protein–ligand complex per target set in the Merck benchmark is provided in Table 6. As the fastest protocol, the IFD MOEdock procedure spent 5 to 29 min on average (per protein–ligand complex) for all the target sets. This is contrasted with the cMD-MT protocol which required roughly 1190 min or 19.83 h on average to simulate a 250 ns trajectory for each of the protein–ligand complexes from this benchmark. Finally, for CHMC sampling which also included 50 ns short-term cMD simulations, these calculations cost 560.44 min or 9.34 h on average for sampling one trajectory for each protein–ligand complex from the benchmark. Because the multithread simulations in the CHMC-cMD protocol were performed parallelly, we presented the averaged time for every trajectory as the total time for the sampling workflow for each test case in the corresponding target set, noting that it was not an apples-to-apples computational time comparison because the parallel computing required more computational resources; but, employing multiple GPU cards was a simple way to support the parallel computing efficiency, while it was more expensive to improve the computational speed for the single-trajectory simulations. Overall, the CHMC-cMD protocol achieved close or even better free energy estimation accuracy compared to the 250 ns cMD protocol but at less simulation cost.
Table 6. Averaged Computational Time Monitored for All Target Sets from the Merck Benchmark
  sampling workflow averaged time (min)
protein MOE-MT cMD-MT CHMC-cMD-MT
CDK8 10 1409 855
c-Met 9 1082 810
Eg5 12 1174 249
Eg5 with an alternative loop 29 1166 233
HIF-2α 5 742 310
PFKFB3 9 1452 786
SHP-2 23 1684 907
SYK 16 1143 532
TNKS2 6 860 362
When looking at the statistical analysis results, the simulation-based MT protocols showed obvious advantages in ranking the binding affinities for the CDK8, c-Met, HIF-2α, PFKFB3 sets and achieved better overall binding affinities ranking against the whole benchmark. On the other hand, the MOE-MT protocols using both energy functions showed acceptable ranking capabilities, with even competitive performance for a few target sets. When also comparing the computational efficiency, the MOE-MT protocols achieved a fairly good “labor productivity ratio” by generating acceptable free energy estimation accuracy with much less computing time. However, when comparing the docking poses and the corresponding conformational snapshots from MD simulations, we also discovered that particular binding modes from docking could be significantly altered through the MD simulations leading to significant differences in the binding free energy estimation accuracy for some target sets, e.g., the PFKFB2 set from this benchmark.
Binding mode generation using docking methods is largely dependent on the input structure’s conformation along with the binding cavity on the protein surface. The IFD method only modestly refines the conformation of the residues in the binding site which contains the fitted ligand structure. When trying to evaluate the binding free energy with a single complex pose, important contacts could be missing, and unstable pairwise contacts may be considered important for maintaining the protein–ligand binding modes. This computational trial showed that the simulation-based free energy protocols sometimes significantly altered the conformation of the protein–ligand binding modes compared to the original docking poses, and some significant atom-pairwise interactions recognized by the docking protocol soon disappeared in the MD trajectories. Herein, we illustrate two example cases which reveal that the MD procedure often develops more reasonable binding modes leading to better binding free energy predictions compared to using docking procedures alone (target: Eg5–ligand: CHEMBL1084431 and target: PFKFB3–ligand: ligand_41).
For the Eg5 – CHEMBL1084431 protein–ligand molecular model, Figure 6 depicts the geometric difference between a representative binding mode selected from the converged cMD trajectory (Figures 6A and 6C) and the best scored docking pose from the MOE IFD (Figures 6B and 6D). Comparing Figures 6A and 6B, we see that the binding site residues wrap tighter around the ligand molecule during the MD simulation versus the IFD protocol. Figures 6C and 6D show that both IFD and MD recognize the hydrogen bond between the ligand’s N1 amine and the Gly103 backbone carbonyl group along with the hydrogen bond between the ligand’s N2 amine and the Glu102 backbone carbonyl group. However, only the MD trajectory shows that the side chain carboxyl group of Glu104 swings down to form another hydrogen bond with the ligand’s N1 amine. By reviewing the fraction of hydrogen bonds over the last 200 ns MD trajectory as reported in Table 7, we further discover that the N1 amine group moves back and forth to form several alternating hydrogen bonds with the surrounding residues. Specifically, when the N1 group moves to the left, it forms two hydrogen bonds with the Gly103 backbone and Glu104 side chain; but when the N1 group moves to the other side of the pocket, it forms one hydrogen bond with the Glu102 side chain. As a result, though the N2 group forms a stable hydrogen bond with the Glu102 residue (89.3% of the time through the last 200 ns), the N1 group was quite “bouncy” at the entrance of the binding pocket. This suggests that the ligand molecule, CHEMBL1084431, might not bind as tightly at the Eg5 binding site as shown from the MOE docking result. Against this complex, the MOE-MT-Amber protocol generated the binding free energy as −15.02 kcal/mol. Given the unstable hydrogen bonds formed between the ligand’s N1 group and the receptor, the cMD-MT-Amber free energy protocol generated a much lower binding affinity as −11.14 kcal/mol (which agrees better with the experimental value as −9.8 kcal/mol).

Figure 6

Figure 6. Illustration of the protein–ligand binding modes for the Eg5 – CHEMBL1084431 complex. (A) CHEMBL1084431 molecule at the Eg5 binding site using a representative binding mode selected from the converged cMD trajectory. (B) CHEMBL1084431 molecule at the Eg5 binding site using the best scored docking pose from the MOE IFD protocol. The area on the protein surface covered with yellow color shows the location of the residues having close contact (<4 Å) with the ligand. (C) Important Eg5 residues interacting with CHEMBL1084431 from the cMD representative binding mode. (C) Important Eg5 residues interacting with CHEMBL1084431 from the MOE IFD protocol.

High Resolution Image
Download MS PowerPoint Slide
Table 7. Hydrogen Bonding Percentages for the Eg5 – CHEMBL1084431 and PFKFB3–ligand_41 Complexes over Their Last 200 ns MD Trajectories
acceptor donor frac/%
Eg5 – CHEMBL1084431 complex
GLU_102@O LIG@N2 89.3
GLY_103@O LIG@N1 26.7
GLU_102@OE2 LIG@N1 14.5
GLU_102@OE1 LIG@N1 3.8
GLU_104@OE2 LIG@N1 9.6
GLU_104@OE1 LIG@N1 7.8
PFKFB3–ligand_41 Complex
LIG@N1 ASN133@ND2 18.3
LIG@N3 ASN133@ND2 6.6
For the PFKFB3–ligand_41 complex, there were two major differences in the binding modes generated by the two sampling protocols. In the MOE IFD best-scored pose, we observe that two residues at the PFKFB3 binding site act like a pair of “clips” clamping the quinoxaline group of the ligand in the middle: the Arg15-Gly16 backbone amide approaches the ligand’s quinoxaline by forming a π–π stacking interaction, while Val187 provides the side-chain isopropyl group to form a C–H−π interaction against the other side of the quinoxaline group. Notably, through the 250 ns MD simulation, the ligand’s quinoxaline group still exhibits close contacts with the Arg15-Gly16 backbone amide. However, the Val187 residue moves away from the ligand binding region as the beta-sheet containing it twists during the simulation. Meanwhile, Val213 moves to contact the ligand’s quinoxaline group and form a new C–H−π interaction which replaces the Val187 residue. Another binding-mode difference is observed at the ligand’s quinoxaline binding region. For this interaction, the MOE IFD protocol places the ligand’s pyrazine ring in a location to form two hydrogen bonds with both the Ser122 and the Asn133 side chains of the receptor, while the hydrogen bond between Ser122 and the ligand’s pyrazine nitrogen is not observed in the converged MD trajectory. Instead, MD simulation shows that the Ser122 and Lys17 side chains gradually drew each other closer and form a hydrogen bond network including the Ser122 and Lys17 side chains and the Arg15 backbone. Moreover, by studying the hydrogen bond’s fraction over the last 200 ns MD trajectory, we find that the hydrogen bond between the Asn133 side chain and the ligand’s pyrazine nitrogen is not stable through the simulation, and only 18.3% of the snapshots in the last 200 ns MD trajectory had the hydrogen bond formed between Asn133 and the ligand’s pyrazine nitrogen. For binding affinity estimation, the MOE-MT-Amber protocol determined a binding free energy of −11.36 kcal/mol, and the cMD-MT-Amber free energy protocol recognized the binding affinity as −7.72 kcal/mol. This latter estimation agrees better with the experimental value of −7.39 kcal/mol (see Figure 7).

Figure 7

Figure 7. Illustration of protein–ligand binding modes for the PFKFB3–ligand_41 complex. (A) Representative binding mode selected from the converged cMD trajectory. During the MD simulation, Val213 approached the quinoxaline group of the ligand to form a C–H−π interaction, while Val187 moved away from the ligand. (B) During the MD simulation, no hydrogen bonding interaction between SER122 and the ligand is observed, while the hydrogen bonding interaction between ASN133 and the ligand was unstable. (C) The MOE IFD protocol placed the ligand at the binding site with the quinoxaline group close to the Val187 side chain, while Val213 was far from the ligand. (D) The best scored docking pose recognized two hydrogen bonding interactions between the ligand and the SER122 and ASN133 side chains.

High Resolution Image
Download MS PowerPoint Slide
The IFD protocol is focused on optimizing the intermolecular interaction potential energy by locally altering the conformations of the ligand and the receptor residues at the interface, but it is generally incapable of relaxing the intramolecular potential stress given the conformational complexity of most protein receptors. During the protein–ligand binding mechanism, certain intraprotein interactions are preferred over some possible significant intermolecular interactions to achieve lower overall potential energy for the whole complex. Moreover, sometimes a stable binding mode could only be achieved with large protein conformational changes which can scarcely be obtained from a docking protocol without the support from large-scale simulation. These two examples clearly illustrate that some significant protein–ligand interactions formed from the docking protocol are unstable or not even recognized during the MD simulation, and some significant protein–ligand interactions can only be formed through significant protein motif movement.

Free Energy Estimation Using Single Conformation vs Conformation Ensembles from Simulation Trajectories

With a more comprehensive evaluation of the intra- and intermolecular interactions, force fields usually lead to different binding modes than the docking generations alone. Moreover, simulations against many test cases in this work generated multiple converged conformational clusters, suggesting multiple energy minima during the protein–ligand recognition. From the conformational energy point of view, significant energy minima, especially the global minimal conformation, usually dominated the partition function of a canonical ensemble, and thus, we further explored the binding free energy estimation accuracy by only using the energy-minimum snapshot from the simulation trajectories of the bound-state protein–ligand complexes.
We proposed two rules for selecting the energy-minimum snapshot. From the 250 ns conventional MD trajectories of every protein–ligand complex, the first protocol directly selected the energy-minimum snapshot according to the force field computation. The logic for choosing the energy-minimum snapshot was straightforward, but the wrong snapshot might be picked due to errors in energy functions. The second protocol first chose the largest conformational cluster from the multitrajectory CHMC sampling with 50 ns MD simulations of a protein–ligand complex and then selected the energy-minimum snapshot with respect to this cluster. The parallel simulations using CHMC with MD simulation protocols were expected to cover more extensive conformational space. Assuming the multitrajectory simulation converged after 50 ns local sampling, the largest conformational cluster had a higher opportunity to contain the “global” energy minimum among all the sampled conformations. The snapshot selected from the simulations of every protein–ligand complex was then treated with the MTScoreES protocol for binding free energy estimation.
In the following analysis, we used the term “snapshot-cMD-MT-amber” denoting the binding free energy estimation using the snapshot selected from the 250 ns MD simulation and treated with the MTScoreES protocol using the Amber energy function; the term “snapshot-cMD-MT-garf” denoting the binding free energy estimation using the snapshot selected from the 250 ns MD simulation and treated with the MTScoreES protocol using the GARF energy function; the term “snapshot-CHMC-cMD-MT-amber” denoting the binding free energy estimation using the snapshot selected from the CHMC parallel sampling followed by 50 ns conventional MD simulation and treated with the MTScoreES protocol using the Amber energy function; and the term “snapshot-CHMC-cMD-MT-garf” denoting the binding free energy estimation using the snapshot selected from the CHMC parallel sampling followed by 50 ns conventional MD simulation and treated with the MTScoreES protocol using the GARF energy function.
Against the CASF benchmark, free energy estimation results using single snapshots were concluded in Table 8. It was not surprising that the prediction accuracy and ranking performance were not as good as using the conformational ensemble-based protocols (See the corresponding results in Table 1.). Moreover, the results illustrated that binding free energy estimation using more thorough sampling and converged conformational ensembles from the simulation outperformed the binding affinity prediction using single conformations or docking poses. It was also worth noting that snapshots selected from the largest conformational cluster generated better binding free energy prediction results compared to using the energy-minimum snapshot from the 250 ns single-trajectory simulation. This suggested that the high-probable conformation generated from the MD simulation may be a better representative conformational state than the energy minimum selected from the simulation trajectory.
Table 8. Statistical Results for Protein-Ligand Binding Free Energy Estimation Using Single Snapshots Treated with Different MT-Based Protocols
MT protocols Pearson’s R Kendall’s τ MAE (kcal/mol)
snapshot-cMD-MT-amber 0.59 0.43 2.22
snapshot-cMD-MT-garf 0.63 0.45 1.86
snapshot-CHMC-cMD-MT-amber 0.63 0.46 2.32
snapshot-CHMC-cMD-MT-garf 0.65 0.47 1.78
Table 9 concluded the statistical results for the free energy estimation using single snapshots for the Merck benchmark. For the eight target sets, both the global minimal snapshots and the minimal snapshots from the largest conformational clusters had comparable binding affinity prediction accuracy. This suggested that both the single-trajectory and parallel simulations converged at similar conformational states.
Table 9. Statistical Results for All the Target Sets from the Merck Benchmark Using Single Snapshots Treated with Different Free Energy Protocols
    snapshot-cMD-MT-amber snapshot-cMD-MT-garf snapshot-CHMC-cMD-MT-amber snapshot-CHMC-cMD-MT-garf
CDK8 Pearson’s R 0.448 0.552 0.391 0.452
  RMSE (kcal/mol) 1.269 1.201 1.487 1.242
  Kendall’s τ 0.379 0.456 0.340 0.404
c-MET Pearson’s R 0.715 0.724 0.719 0.757
  RMSE (kcal/mol) 1.685 1.962 1.845 1.379
  Kendall’s τ 0.667 0.539 0.596 0.554
Eg5 Pearson’s R 0.405 0.505 0.464 0.506
  RMSE (kcal/mol) 1.928 2.494 1.587 1.678
  Kendall’s τ 0.072 0.169 0.207 0.220
Eg5 with an alternative loop Pearson’s R 0.367 0.403 0.360 0.361
  RMSE (kcal/mol) 2.024 2.180 1.816 1.789
  Kendall’s τ 0.121 0.183 0.113 0.118
HIF-2α Pearson’s R 0.270 0.354 0.212 0.345
  RMSE (kcal/mol) 1.184 1.409 1.274 1.263
  Kendall’s τ 0.196 0.270 0.140 0.295
PFKFB3 Pearson’s R 0.414 0.596 0.395 0.637
  RMSE (kcal/mol) 1.176 1.624 1.054 1.249
  Kendall’s τ 0.307 0.429 0.312 0.473
SHP-2 Pearson’s R 0.150 0.357 0.056 0.334
  RMSE (kcal/mol) 1.379 1.607 1.467 1.462
  Kendall’s τ 0.102 0.290 0.022 0.203
SYK Pearson’s R 0.250 0.252 0.180 0.219
  RMSE (kcal/mol) 2.535 3.163 2.550 2.523
  Kendall’s τ 0.156 0.122 0.072 0.057
TNKS2 Pearson’s R 0.567 0.350 0.617 0.440
  RMSE (kcal/mol) 0.994 2.943 0.855 2.593
  Kendall’s τ 0.335 0.167 0.372 0.227
total Pearson’s R 0.509 0.470 0.537 0.512
  RMSE (kcal/mol) 1.626 2.171 1.628 1.765
  Kendall’s τ 0.384 0.341 0.418 0.380
When comparing to the conformational ensemble-based free energy estimation protocols, we discovered that the all protocols had comparable ranking capability against the CDK8, c-MET, Eg5, and TNKS2 sets. On the other hand, for the HIF-2α, PFKFB3, SHP-2, and SYK sets, the single-snapshot-based protocols generally had higher correlation coefficients (Pearson’s R and Kendall’s τ values) compared to the docking-based protocols but were outperformed by the conformational ensemble-based MD protocols (cMD-MT and CHMC-cMD-MT protocols). It was likely that the protein–ligand complexes in the CDK8, c-MET, Eg5, and TNKS2 sets had dominantly significant global minimal conformational states and that binding free energy estimators using single conformations or conformational ensembles from docking and MD simulations all had generally acceptable accuracy. Against the HIF-2α, PFKFB3, SHP-2, and SYK sets, the MD simulation significantly altered the protein–ligand binding modes from the docking poses. Single snapshots from the MD simulations generated more reasonable binding affinities compared to the docking predictions, and the conformational ensembles from simulation further improved the binding affinity prediction accuracy.

Conclusion

ARTICLE SECTIONS
Jump To
In this study, we introduced two simulation-based conformational sampling protocols including (1) 250 ns single-trajectory Amber molecular dynamics and (2) CHMC Monte Carlo sampling followed by 50 ns multireplica Amber MD simulation and contrasted these with the previously published MOE induced-fit docking protocol. Each ensemble of protein–ligand complexes was generated and used to explore the free energy estimation accuracy and efficiency of the Movable Type absolute binding free energy method using different conformational sampling protocols. Compared to dock-based MT protocols, simulation-based MT protocols achieve significant improvement in binding free energy estimation accuracy against both the CASF-2016 and the Merck KGaA benchmarks. The CASF-2016 benchmark contains a wide variety of test cases with different structures for both ligands and proteins. The two simulation-based protocols using the GARF energy function obtained good prediction accuracy with Pearson’s R ≥ 0.7, Kendall’s τ = 0.52, and an MAE < 2 kcal/mol, and the two simulation-based protocols using the MT-AMBERFF19SB force field had lower correlation coefficients with Pearson’s R = 0.66, Kendall’s τ = 0.49 and higher errors with an MAE = 2.10 kcal/mol. Against protein–ligand complexes with identical protein targets from the Merck benchmark, our simulation-based protocols obtained acceptable (0.4 ≤ Pearson’s R ≤ 0.6) to good (Pearson’s R ≥ 0.6) ranking performances for more than 2/3rd of all the target sets. With the help of free energy prediction rescaling using a separate set of protein–ligand cocrystal structures, Pearson’s R values for all the protocols against the whole Merck benchmark were improved from 0.195–0.372 to 0.551–0.658 after rescaling, and RMSE values for all the simulation-based protocols were within 2 kcal/mol. Conversely, the RMSE values for dock-based protocols increased when introducing rescaling suggesting that the binding-site protein residues captured by the docking protocols deviated from the cocrystal structures in the corresponding rescaling set. As an absolute binding free energy method which relies on numerical energy performed on ensembles of conformational states, the MT protocol is heavily dependent on conformational sampling for protein–ligand complex systems. The present work clearly shows that utilizing MD protocols greatly improves binding free energy estimation accuracy. However, extensive simulation did not always result in better performance, and the MT-CHMC method, which employs multiple reasonable initial poses performed with parallel short-term simulations, reached similar estimation accuracy using less than half of the computational time. Since the method is compatible with many different conformational sampling methods, the present work shows that MT is a flexible free energy tool for protein–ligand binding free energy estimation tasks.

Supporting Information

ARTICLE SECTIONS
Jump To

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.2c00278.

  • Detailed explanation of Movable Type binding free energy estimation protocol (PDF)

  • Free energy calculation results for CASF-2016 benchmark, rescale set, and Merck KGaA benchmark using cMD-MT-amber, cMD-MT-garf, CHMC-cMD-MT-amber, and CHMC-cMD-MT-garf protocols (PDF)

Terms & Conditions

Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.

Author Information

ARTICLE SECTIONS
Jump To
  • Corresponding Author
    • Zheng Zheng - School of Chemistry, Chemical Engineering and Life Science, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China;  QuantumBio Inc., State College, Pennsylvania16801, United States;  Orcidhttps://orcid.org/0000-0001-5221-3209;  Email: [email protected]
  • Authors
    • Wenlang Liu - School of Chemistry, Chemical Engineering and Life Science, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    • Zhenhao Liu - School of Chemistry, Chemical Engineering and Life Science, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    • Hao Liu - School of Mechanical and Electronic Engineering, Wuhan University of Technology, 122 Luoshi Road, Wuhan430070, PR China
    • Lance M. Westerhoff - QuantumBio Inc., State College, Pennsylvania16801, United States
  • Notes
    The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
    The authors declare no competing financial interest.

Acknowledgments

ARTICLE SECTIONS
Jump To

Z.Z. acknowledges support from the National Natural Science Foundation of China (Grant No. 21903061). L.M.W. acknowledges being supported in part by the National Institute of General Medical Sciences of the National Institutes of Health under Small Business Innovative Research (SBIR) Award R44GM134781. The authors also acknowledge the continued support of Michigan State University and MSU Technologies for licensing the MovableType technology to QuantumBio, Inc. We also thank Drs. Roger I. Martin, Oleg Borbulevych, Nupur Bansal, and Kenneth M. Merz, Jr. for their work and support in the prior study and the transfer and implementation of the MT technology in DivCon.

References

ARTICLE SECTIONS
Jump To

This article references 56 other publications.

  1. 1
    Schneider, G. Virtual screening: an endless staircase?. Nat. Rev. Drug Discovery 2010, 9 (4), 273– 276,  DOI: 10.1038/nrd3139
    Google Scholar
    1
    Virtual screening: an endless staircase?
    Schneider, Gisbert
    Nature Reviews Drug Discovery (2010), 9 (4), 273-276CODEN: NRDDAG; ISSN:1474-1776. (Nature Publishing Group)
    A review. Computational chem. - in particular, virtual screening - can provide valuable contributions in hit- and lead-compd. discovery. Numerous software tools have been developed for this purpose. However, despite the applicability of virtual screening technol. being well established, it seems that there are relatively few examples of drug discovery projects in which virtual screening has been the key contributor. Has virtual screening reached its peak. If not, what aspects are limiting its potential at present, and how can significant progress be made in the future.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXktVGjur4%253D&md5=ecb5a4aa2683d4a7d688fad3114bc31d
  2. 2
    Sliwoski, G.; Kothiwale, S.; Meiler, J.; Lowe Jr, E. W. Computational methods in drug discovery. Pharmacol. Rev. 2014, 66 (1), 334– 395,  DOI: 10.1124/pr.112.007336
    Google Scholar
    2
    Computational methods in drug discovery
    Sliwoski, Gregory; Kothiwale, Sandeepkumar; Meiler, Jens; Lowe Edward, W.
    Pharmacological Reviews (2014), 66 (1), 334-395, 62 pp.CODEN: PAREAQ; ISSN:1521-0081. (American Society for Pharmacology and Experimental Therapeutics)
    A review. Computer-aided drug discovery/design methods have played a major role in the development of therapeutically important small mols. for over three decades. These methods are broadly classified as either structure-based or ligand-based methods. Structure-based methods are in principle analogous to high-throughput screening in that both target and ligand structure information is imperative. Structure-based approaches include ligand docking, pharmacophore, and ligand design methods. The article discusses theory behind the most important methods and recent successful applications. Ligand-based methods use only ligand information for predicting activity depending on its similarity/dissimilarity to previously known active ligands. We review widely used ligand-based methods such as ligand-based pharmacophores, mol. descriptors, and quant. structure-activity relationships. In addn., important tools such as target/ligand data bases, homol. modeling, ligand fingerprint methods, etc., necessary for successful implementation of various computer-aided drug discovery/design methods in a drug discovery campaign are discussed. Finally, computational methods for toxicity prediction and optimization for favorable physiol. properties are discussed with successful examples from literature.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhsVGnu7nL&md5=3dde38a0b60c583f832c688c2d27819f
  3. 3
    Dror, O.; Schneidman-Duhovny, D.; Inbar, Y.; Nussinov, R.; Wolfson, H. J. Novel approach for efficient pharmacophore-based virtual screening: method and applications. J. Chem. Inf. Model 2009, 49 (10), 2333– 2343,  DOI: 10.1021/ci900263d
    Google Scholar
    3
    Novel Approach for Efficient Pharmacophore-Based Virtual Screening: Method and Applications
    Dror, Oranit; Schneidman-Duhovny, Dina; Inbar, Yuval; Nussinov, Ruth; Wolfson, Haim J.
    Journal of Chemical Information and Modeling (2009), 49 (10), 2333-2343CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    Virtual screening is emerging as a productive and cost-effective technol. in rational drug design for the identification of novel lead compds. An important model for virtual screening is the pharmacophore. Pharmacophore is the spatial configuration of essential features that enable a ligand mol. to interact with a specific target receptor. In the absence of a known receptor structure, a pharmacophore can be identified from a set of ligands that have been obsd. to interact with the target receptor. Here, we present a novel computational method for pharmacophore detection and virtual screening. The pharmacophore detection module is able to (i) align multiple flexible ligands in a deterministic manner without exhaustive enumeration of the conformational space, (ii) detect subsets of input ligands that may bind to different binding sites or have different binding modes, (iii) address cases where the input ligands have different affinities by defining weighted pharmacophores based on the no. of ligands that share them, and (iv) automatically select the most appropriate pharmacophore candidates for virtual screening. The algorithm is highly efficient, allowing a fast exploration of the chem. space by virtual screening of huge compd. databases. The performance of PharmaGist was successfully evaluated on a commonly used data set of G-Protein Coupled Receptor alpha1A. Addnl., a large-scale evaluation using the DUD (directory of useful decoys) data set was performed. DUD contains 2950 active ligands for 40 different receptors, with 36 decoy compds. for each active ligand. PharmaGist enrichment rates are comparable with other state-of-the-art tools for virtual screening.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD1MXht1WntrnE&md5=745357210296e3124bafa49b1f46f0de
  4. 4
    Malmstrom, R. D.; Watowich, S. J. Using free energy of binding calculations to improve the accuracy of virtual screening predictions. J. Chem. Inf. Model 2011, 51 (7), 1648– 1655,  DOI: 10.1021/ci200126v
    Google Scholar
    4
    Using Free Energy of Binding Calculations To Improve the Accuracy of Virtual Screening Predictions
    Malmstrom, Robert D.; Watowich, Stanley J.
    Journal of Chemical Information and Modeling (2011), 51 (7), 1648-1655CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    Virtual screening of small mol. databases against macromol. targets was used to identify binding ligands and predict their lowest energy bound conformation (i.e., pose). AutoDock4-generated poses were rescored using mean-field pathway decoupling free energy of binding calcns. and evaluated if these calcns. improved virtual screening discrimination between bound and nonbound ligands. Two small mol. databases were used to evaluate the effectiveness of the rescoring algorithm in correctly identifying binders of L99A T4 lysozyme. Self-dock calcns. of a database contg. compds. with known binding free energies and cocrystal structures largely reproduced exptl. measurements, although the mean difference between calcd. and exptl. binding free energies increased as the predicted bound poses diverged from the exptl. poses. In addn., free energy rescoring was more accurate than AutoDock4 scores in discriminating between known binders and nonbinders, suggesting free energy rescoring could be a useful approach to reduce false pos. predictions in virtual screening expts.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3MXos1Whtb0%253D&md5=8a5837b9e25e603960a8690c4a9834dd
  5. 5
    Klebe, G. Applying thermodynamic profiling in lead finding and optimization. Nat. Rev. Drug Discovery 2015, 14 (2), 95– 110,  DOI: 10.1038/nrd4486
    Google Scholar
    5
    Applying thermodynamic profiling in lead finding and optimization
    Klebe, Gerhard
    Nature Reviews Drug Discovery (2015), 14 (2), 95-110CODEN: NRDDAG; ISSN:1474-1776. (Nature Publishing Group)
    Small-mol. drug discovery involves the optimization of various physicochem. properties of a ligand, particularly its binding affinity for its target receptor (or receptors). In recent years, there has been growing interest in using thermodn. profiling of ligand-receptor interactions in order to select and optimize those ligands that might be most likely to become drug candidates with desirable physicochem. properties. The thermodn. of binding is influenced by multiple factors, including hydrogen bonding and hydrophobic interactions, desolvation, residual mobility, dynamics and the local water structure. This article discusses key issues in understanding the effects of these factors and applying this knowledge in drug discovery.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXhsVemtr4%253D&md5=c2e12760623de81335ce252cb804c602
  6. 6
    Albanese, S. K.; Chodera, J. D.; Volkamer, A.; Keng, S.; Abel, R.; Wang, L. Is Structure-Based Drug Design Ready for Selectivity Optimization?. J. Chem. Inf. Model 2020, 60 (12), 6211– 6227,  DOI: 10.1021/acs.jcim.0c00815
    Google Scholar
    6
    Is Structure-Based Drug Design Ready for Selectivity Optimization?
    Albanese, Steven K.; Chodera, John D.; Volkamer, Andrea; Keng, Simon; Abel, Robert; Wang, Lingle
    Journal of Chemical Information and Modeling (2020), 60 (12), 6211-6227CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    Alchem. free-energy calcns. are now widely used to drive or maintain potency in small-mol. lead optimization with a roughly 1 kcal/mol accuracy. Despite this, the potential to use free-energy calcns. to drive optimization of compd. selectivity among two similar targets has been relatively unexplored in published studies. In the most optimistic scenario, the similarity of binding sites might lead to a fortuitous cancellation of errors and allow selectivity to be predicted more accurately than affinity. Here, we assess the accuracy with which selectivity can be predicted in the context of small-mol. kinase inhibitors, considering the very similar binding sites of human kinases CDK2 and CDK9 as well as another series of ligands attempting to achieve selectivity between the more distantly related kinases CDK2 and ERK2. Using a Bayesian anal. approach, we sep. systematic from statistical errors and quantify the correlation in systematic errors between selectivity targets. We find that, in the CDK2/CDK9 case, a high correlation in systematic errors suggests that free-energy calcns. can have significant impact in aiding chemists in achieving selectivity, while in more distantly related kinases (CDK2/ERK2), the correlation in systematic error suggests that fortuitous cancellation may even occur between systems that are not as closely related. In both cases, the correlation in systematic error suggests that longer simulations are beneficial to properly balance statistical error with systematic error to take full advantage of the increase in apparent free-energy calcn. accuracy in selectivity prediction.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXitFygs7bJ&md5=08cfdf031b626801b7ba762a559d750f
  7. 7
    Raha, K.; Peters, M. B.; Wang, B.; Yu, N.; Wollacott, A. M.; Westerhoff, L. M.; Merz Jr, K. M. The role of quantum mechanics in structure-based drug design. Drug Discovery Today 2007, 12 (17–18), 725– 731,  DOI: 10.1016/j.drudis.2007.07.006
    Google Scholar
    7
    The role of quantum mechanics in structure-based drug design
    Raha, Kaushik; Peters, Martin B.; Wang, Bing; Yu, Ning; Wollacott, Andrew M.; Westerhoff, Lance M.; Merz, Kenneth M., Jr.
    Drug Discovery Today (2007), 12 (17&18), 725-731CODEN: DDTOFS; ISSN:1359-6446. (Elsevier B.V.)
    A review. Herein we will focus on the use of quantum mechanics (QM) in drug design (DD) to solve disparate problems from scoring protein-ligand poses to building QM QSAR models. Through the variational principle of QM we know that we can obtain a more accurate representation of mol. systems than classical models, and while this is not a matter of debate, it still has not been shown that the expense of QM approaches is offset by improved accuracy in DD applications. Objectively validating the improved applicability and performance of QM over classical-based models in DD will be the focus of research in the coming years along with research on the conformational sampling problem as it relates to protein-ligand complexes.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXhtVWhsrvK&md5=67fecc13843c7961ff820026c85a9290
  8. 8
    Brooijmans, N.; Kuntz, I. D. Molecular recognition and docking algorithms. Annu. Rev. Biophys. Biomol. Struct. 2003, 32, 335– 373,  DOI: 10.1146/annurev.biophys.32.110601.142532
    Google Scholar
    8
    Molecular recognition and docking algorithms
    Brooijmans, Natasja; Kuntz, Irwin D.
    Annual Review of Biophysics and Biomolecular Structure (2003), 32 (), 335-373CODEN: ABBSE4; ISSN:1056-8700. (Annual Reviews Inc.)
    A review. Mol. docking is an invaluable tool in modern drug discovery. This review focuses on methodol. developments relevant to the field of mol. docking. The forces important in mol. recognition are reviewed and followed by a discussion of how different scoring functions account for these forces. More recent applications of computational chem. tools involve library design and database screening. Last, we summarize several crit. methodol. issues that must be addressed in future developments.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXntFSgtrg%253D&md5=1db89884abfef58d8080e726cf40960c
  9. 9
    Kuntz, I. D.; Blaney, J. M.; Oatley, S. J.; Langridge, R.; Ferrin, T. E. A geometric approach to macromolecule-ligand interactions. J. Mol. Biol. 1982, 161 (2), 269– 288,  DOI: 10.1016/0022-2836(82)90153-X
    Google Scholar
    9
    A geometric approach to macromolecule-ligand interactions
    Kuntz, Irwin D.; Blaney, Jeffrey M.; Oatley, Stuart J.; Langridge, Robert; Ferrin, Thomas E.
    Journal of Molecular Biology (1982), 161 (2), 269-88CODEN: JMOBAK; ISSN:0022-2836.
    A method is described to explore geometrically feasible alignments of ligands and receptors of known structure. Algorithms are presented that examine many binding geometries and evaluate them in terms of steric overlap. The procedure uses specific mol. conformations. A method is included for finding putative binding sites on a macromol. surface. Results are reported for heme-myoglobin interaction and the binding of thyroid hormone analogs to prealbumin. In each case, the program finds structures within 1 Å of the x-ray results and also finds distinctly different geometries that provide good steric fits. The approach seems well-suited for generating conformations for energy refinement programs and interactive computer graphics routines.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaL38XmtFajsbw%253D&md5=8a4234b24356ea5340f33d906cd71d3e
  10. 10
    Kitchen, D. B.; Decornez, H.; Furr, J. R.; Bajorath, J. Docking and scoring in virtual screening for drug discovery: methods and applications. Nat. Rev. Drug Discovery 2004, 3 (11), 935– 949,  DOI: 10.1038/nrd1549
    Google Scholar
    10
    Docking and scoring in virtual screening for drug discovery: methods and applications
    Kitchen, Douglas B.; Decornez, Helene; Furr, John R.; Bajorath, Juergen
    Nature Reviews Drug Discovery (2004), 3 (11), 935-949CODEN: NRDDAG; ISSN:1474-1776. (Nature Publishing Group)
    A review. Computational approaches that 'dock' small mols. into the structures of macromol. targets and 'score' their potential complementarity to binding sites are widely used in hit identification and lead optimization. Indeed, there are now a no. of drugs whose development was heavily influenced by or based on structure-based design and screening strategies, such as HIV protease inhibitors. Nevertheless, there remain significant challenges in the application of these approaches, in particular in relation to current scoring schemes. Here, we review key concepts and specific features of small-mol.-protein docking methods, highlight selected applications and discuss recent advances that aim to address the acknowledged limitations of established approaches.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXptFemtrg%253D&md5=875a2b37a4299181509a1922b11dbd2f
  11. 11
    Sledz, P.; Caflisch, A. Protein structure-based drug design: from docking to molecular dynamics. Curr. Opin. Struct. Biol. 2018, 48, 93– 102,  DOI: 10.1016/j.sbi.2017.10.010
    Google Scholar
    11
    Protein structure-based drug design: from docking to molecular dynamics
    Sledz, Pawel; Caflisch, Amedeo
    Current Opinion in Structural Biology (2018), 48 (), 93-102CODEN: COSBEF; ISSN:0959-440X. (Elsevier Ltd.)
    Recent years have witnessed rapid developments of computer-aided drug design methods, which have reached accuracy that allows their routine practical applications in drug discovery campaigns. Protein structure-based methods are useful for the prediction of binding modes of small mols. and their relative affinity. The high-throughput docking of up to 106 small mols. followed by scoring based on implicit-solvent force field can robustly identify micromolar binders using a rigid protein target. Mol. dynamics with explicit solvent is a low-throughput technique for the characterization of flexible binding sites and accurate evaluation of binding pathways, kinetics, and thermodn. In this review we highlight recent advancements in applications of ligand docking tools and mol. dynamics simulations to ligand identification and optimization.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXhslygsbzF&md5=3a3b92c6dbd9c5075bef0bb7268cc5dc
  12. 12
    De Vivo, M.; Masetti, M.; Bottegoni, G.; Cavalli, A. Role of Molecular Dynamics and Related Methods in Drug Discovery. J. Med. Chem. 2016, 59 (9), 4035– 4061,  DOI: 10.1021/acs.jmedchem.5b01684
    Google Scholar
    12
    Role of Molecular Dynamics and Related Methods in Drug Discovery
    De Vivo, Marco; Masetti, Matteo; Bottegoni, Giovanni; Cavalli, Andrea
    Journal of Medicinal Chemistry (2016), 59 (9), 4035-4061CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    Mol. dynamics (MD) and related methods are close to becoming routine computational tools for drug discovery. Their main advantage is in explicitly treating structural flexibility and entropic effects. This allows a more accurate est. of the thermodn. and kinetics assocd. with drug-target recognition and binding, as better algorithms and hardware architectures increase their use. Here, we review the theor. background of MD and enhanced sampling methods, focusing on free-energy perturbation, metadynamics, steered MD, and other methods most consistently used to study drug-target binding. We discuss unbiased MD simulations that nowadays allow the observation of unsupervised ligand-target binding, assessing how these approaches help optimizing target affinity and drug residence time toward improved drug efficacy. Further issues discussed include allosteric modulation and the role of water mols. in ligand binding and optimization. We conclude by calling for more prospective studies to attest to these methods' utility in discovering novel drug candidates.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhtlWksLs%253D&md5=b3d8a4fb5705a3555c2bb2a962420396
  13. 13
    Cheng, T.; Li, X.; Li, Y.; Liu, Z.; Wang, R. Comparative assessment of scoring functions on a diverse test set. J. Chem. Inf. Model 2009, 49 (4), 1079– 1093,  DOI: 10.1021/ci9000053
    Google Scholar
    13
    Comparative Assessment of Scoring Functions on a Diverse Test Set
    Cheng, Tiejun; Li, Xun; Li, Yan; Liu, Zhihai; Wang, Renxiao
    Journal of Chemical Information and Modeling (2009), 49 (4), 1079-1093CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    Scoring functions are widely applied to the evaluation of protein-ligand binding in structure-based drug design. We have conducted a comparative assessment of 16 popular scoring functions implemented in main-stream com. software or released by academic research groups. A set of 195 diverse protein-ligand complexes with high-resoln. crystal structures and reliable binding consts. were selected through a systematic non-redundant sampling of the PDBbind database and used as the primary test set in our study. All scoring functions were evaluated in three aspects, i.e., "docking power", "ranking power", and "scoring power", and all evaluations were independent from the context of mol. docking or virtual screening. As for "docking power", six scoring functions, including GOLD::ASP, DS::PLP1, DrugScorePDB, GlideScore-SP, DS::LigScore, and GOLD::ChemScore, achieved success rates over 70% when the acceptance cutoff was root-mean-square deviation < 2.0 Å. Combining these scoring functions into consensus scoring schemes improved the success rates to 80% or even higher. As for "ranking power" and "scoring power", the top four scoring functions on the primary test set were X-Score, DrugScoreCSD, DS::PLP, and SYBYL::ChemScore. They were able to correctly rank the protein-ligand complexes contg. the same type of protein with success rates around 50%. Correlation coeffs. between the exptl. binding consts. and the binding scores computed by these scoring functions ranged from 0.545 to 0.644. Besides the primary test set, each scoring function was also tested on four addnl. test sets, each consisting of a certain no. of protein-ligand complexes contg. one particular type of protein. Our study serves as an updated benchmark for evaluating the general performance of today's scoring functions. Our results indicate that no single scoring function consistently outperforms others in all three aspects. Thus, it is important in practice to choose the appropriate scoring functions for different purposes.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD1MXkt1aqtLg%253D&md5=9f073faa564c30a69aa26485d4dcc7db
  14. 14
    Li, Y.; Han, L.; Liu, Z.; Wang, R. Comparative assessment of scoring functions on an updated benchmark: 2. Evaluation methods and general results. J. Chem. Inf. Model 2014, 54 (6), 1717– 1736,  DOI: 10.1021/ci500081m
    Google Scholar
    14
    Comparative Assessment of Scoring Functions on an Updated Benchmark: 2. Evaluation Methods and General Results
    Li, Yan; Han, Li; Liu, Zhihai; Wang, Renxiao
    Journal of Chemical Information and Modeling (2014), 54 (6), 1717-1736CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    Our comparative assessment of scoring functions (CASF) benchmark is created to provide an objective evaluation of current scoring functions. The key idea of CASF is to compare the general performance of scoring functions on a diverse set of protein-ligand complexes. In order to avoid testing scoring functions in the context of mol. docking, the scoring process is sepd. from the docking (or sampling) process by using ensembles of ligand binding poses that are generated in prior. Here, we describe the tech. methods and evaluation results of the latest CASF-2013 study. The PDBbind core set (version 2013) was employed as the primary test set in this study, which consists of 195 protein-ligand complexes with high-quality three-dimensional structures and reliable binding consts. A panel of 20 scoring functions, most of which are implemented in main-stream com. software, were evaluated in terms of "scoring power" (binding affinity prediction), "ranking power" (relative ranking prediction), "docking power" (binding pose prediction), and "screening power" (discrimination of true binders from random mols.). Our results reveal that the performance of these scoring functions is generally more promising in the docking/screening power tests than in the scoring/ranking power tests. Top-ranked scoring functions in the scoring power test, such as X-ScoreHM, ChemScore@SYBYL, ChemPLP@GOLD, and PLP@DS, are also top-ranked in the ranking power test. Top-ranked scoring functions in the docking power test, such as ChemPLP@GOLD, Chemscore@GOLD, GlidScore-SP, LigScore@DS, and PLP@DS, are also top-ranked in the screening power test. Our results obtained on the entire test set and its subsets suggest that the real challenge in protein-ligand binding affinity prediction lies in polar interactions and assocd. desolvation effect. Nonadditive features obsd. among high-affinity protein-ligand complexes also need attention.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXls1CrsL0%253D&md5=0d386b64f5c557667533847d694436bc
  15. 15
    Su, M.; Yang, Q.; Du, Y.; Feng, G.; Liu, Z.; Li, Y.; Wang, R. Comparative Assessment of Scoring Functions: The CASF-2016 Update. J. Chem. Inf. Model 2019, 59 (2), 895– 913,  DOI: 10.1021/acs.jcim.8b00545
    Google Scholar
    15
    Comparative Assessment of Scoring Functions: The CASF-2016 Update
    Su, Minyi; Yang, Qifan; Du, Yu; Feng, Guoqin; Liu, Zhihai; Li, Yan; Wang, Renxiao
    Journal of Chemical Information and Modeling (2019), 59 (2), 895-913CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    In structure-based drug design, scoring functions are often employed to evaluate protein-ligand interactions. A variety of scoring functions have been developed so far, and thus, some objective benchmarks are desired for assessing their strength and weakness. The comparative assessment of scoring functions (CASF) benchmark developed by us provides an answer to this demand. CASF is designed as a "scoring benchmark", where the scoring process is decoupled from the docking process to depict the performance of scoring function more precisely. Here, we describe the latest update of this benchmark, i.e., CASF-2016. Each scoring function is still evaluated by four metrics, including "scoring power", "ranking power", "docking power", and "screening power". Nevertheless, the evaluation methods have been improved considerably in several aspects. A new test set is compiled, which consists of 285 protein-ligand complexes with high-quality crystal structures and reliable binding consts. A panel of 25 scoring functions are tested on CASF-2016 as a demonstration. Our results reveal that the performance of current scoring functions is more promising in terms of docking power than scoring, ranking, and screening power. Scoring power is somewhat correlated with ranking power, so are docking power and screening power. The results obtained on CASF-2016 may provide valuable guidance for the end users to make smart choices among available scoring functions. Moreover, CASF is created as an open-access benchmark so that other researchers can utilize it to test a wider range of scoring functions. The complete CASF-2016 benchmark will be released on the PDBbind-CN web server (http://www.pdbbind-cn.org/casf.asp/) once this article is published.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXitlWhtLjM&md5=573295cfd2298a2dee933027041c3d9c
  16. 16
    Forli, S.; Huey, R.; Pique, M. E.; Sanner, M. F.; Goodsell, D. S.; Olson, A. J. Computational protein-ligand docking and virtual drug screening with the AutoDock suite. Nat. Protoc. 2016, 11 (5), 905– 919,  DOI: 10.1038/nprot.2016.051
    Google Scholar
    16
    Computational protein-ligand docking and virtual drug screening with the AutoDock suite
    Forli, Stefano; Huey, Ruth; Pique, Michael E.; Sanner, Michel F.; Goodsell, David S.; Olson, Arthur J.
    Nature Protocols (2016), 11 (5), 905-919CODEN: NPARDW; ISSN:1750-2799. (Nature Publishing Group)
    Computational docking can be used to predict bound conformations and free energies of binding for small-mol. ligands to macromol. targets. Docking is widely used for the study of biomol. interactions and mechanisms, and it is applied to structure-based drug design. The methods are fast enough to allow virtual screening of ligand libraries contg. tens of thousands of compds. This protocol covers the docking and virtual screening methods provided by the AutoDock suite of programs, including a basic docking of a drug mol. with an anticancer target, a virtual screen of this target with a small ligand library, docking with selective receptor flexibility, active site prediction and docking with explicit hydration. The entire protocol will require ∼5 h.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28Xlsl2gsLw%253D&md5=7dafc5e3acab3d38e9b33fbdaf65422d
  17. 17
    Trott, O.; Olson, A. J. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem. 2009, 31 (2), 455– 461,  DOI: 10.1002/jcc.21334
    Google Scholar
    There is no corresponding record for this reference.
  18. 18
    Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.; Taylor, R. D. Improved protein-ligand docking using GOLD. Proteins 2003, 52 (4), 609– 623,  DOI: 10.1002/prot.10465
    Google Scholar
    18
    Improved protein-ligand docking using GOLD
    Verdonk, Marcel L.; Cole, Jason C.; Hartshorn, Michael J.; Murray, Christopher W.; Taylor, Richard D.
    Proteins: Structure, Function, and Genetics (2003), 52 (4), 609-623CODEN: PSFGEY; ISSN:0887-3585. (Wiley-Liss, Inc.)
    The Chemscore function was implemented as a scoring function for the protein-ligand docking program GOLD, and its performance compared to the original Goldscore function and two consensus docking protocols, "Goldscore-CS" and "Chemscore-GS," in terms of docking accuracy, prediction of binding affinities, and speed. In the "Goldscore-CS" protocol, dockings produced with the Goldscore function are scored and ranked with the Chemscore function; in the "Chemscore-GS" protocol, dockings produced with the Chemscore function are scored and ranked with the Goldscore function. Comparisons were made for a "clean" set of 224 protein-ligand complexes, and for two subsets of this set, one for which the ligands are "drug-like," the other for which they are "fragment-like.". For "drug-like" and "fragment-like" ligands, the docking accuracies obtained with Chemscore and Goldscore functions are similar. For larger ligands, Goldscore gives superior results. Docking with the Chemscore function is up to three times faster than docking with the Goldscore function. Both combined docking protocols give significant improvements in docking accuracy over the use of the Goldscore or Chemscore function alone. "Goldscore-CS" gives success rates of up to 81% (top-ranked GOLD soln. within 2.0 Å of the exptl. binding mode) for the "clean list," but at the cost of long search times. For most virtual screening applications, "Chemscore-GS" seems optimal; search settings that give docking speeds of around 0.25-1.3 min/compd. have success rates of about 78% for "drug-like" compds. and 85% for "fragment-like" compds. In terms of producing binding energy ests., the Goldscore function appears to perform better than the Chemscore function and the two consensus protocols, particularly for faster search settings. Even at docking speeds of around 1-2 min/compd., the Goldscore function predicts binding energies with a std. deviation of ∼10.5 kJ/mol.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXmvFGrsLg%253D&md5=73a627d99dab6de1ae364c8e8e5f0fea
  19. 19
    Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.; Greenwood, J. R.; Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T. Extra precision glide: docking and scoring incorporating a model of hydrophobic enclosure for protein-ligand complexes. J. Med. Chem. 2006, 49 (21), 6177– 6196,  DOI: 10.1021/jm051256o
    Google Scholar
    19
    Extra Precision Glide: Docking and Scoring Incorporating a Model of Hydrophobic Enclosure for Protein-Ligand Complexes
    Friesner, Richard A.; Murphy, Robert B.; Repasky, Matthew P.; Frye, Leah L.; Greenwood, Jeremy R.; Halgren, Thomas A.; Sanschagrin, Paul C.; Mainz, Daniel T.
    Journal of Medicinal Chemistry (2006), 49 (21), 6177-6196CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    A novel scoring function to est. protein-ligand binding affinities has been developed and implemented as the Glide 4.0 XP scoring function and docking protocol. In addn. to unique water desolvation energy terms, protein-ligand structural motifs leading to enhanced binding affinity are included:(1) hydrophobic enclosure where groups of lipophilic ligand atoms are enclosed on opposite faces by lipophilic protein atoms, (2) neutral-neutral single or correlated hydrogen bonds in a hydrophobically enclosed environment, and (3) five categories of charged-charged hydrogen bonds. The XP scoring function and docking protocol have been developed to reproduce exptl. binding affinities for a set of 198 complexes (RMSDs of 2.26 and 1.73 kcal/mol over all and well-docked ligands, resp.) and to yield quality enrichments for a set of fifteen screens of pharmaceutical importance. Enrichment results demonstrate the importance of the novel XP mol. recognition and water scoring in sepg. active and inactive ligands and avoiding false positives.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD28XpvVGmurg%253D&md5=ea428c82ead0d8c27f8c1a7b694a1edf
  20. 20
    Jain, A. N. Surflex: fully automatic flexible molecular docking using a molecular similarity-based search engine. J. Med. Chem. 2003, 46 (4), 499– 511,  DOI: 10.1021/jm020406h
    Google Scholar
    20
    Surflex: fully automatic flexible molecular docking using a molecular similarity-based search engine
    Jain, Ajay N.
    Journal of Medicinal Chemistry (2003), 46 (4), 499-511CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    Surflex is a fully automatic flexible mol. docking algorithm that combines the scoring function from the Hammerhead docking system with a search engine that relies on a surface-based mol. similarity method as a means to rapidly generate suitable putative poses for mol. fragments. Results are presented evaluating reliability and accuracy of dockings compared with crystallog. exptl. results on 81 protein/ligand pairs of substantial structural diversity. In over 80% of the complexes, Surflex's highest scoring docked pose was within 2.5 Å root-mean-square deviation (rmsd), with over 90% of the complexes having one of the top ranked poses within 2.5 Å rmsd. Results are also presented assessing Surflex's utility as a screening tool on two protein targets (thymidine kinase and estrogen receptor) using data sets on which competing methods were run. Performance of Surflex was significantly better, with true pos. rates of greater than 80% at false pos. rates of less than 1%. Docking time was roughly linear in no. of rotatable bonds, beginning with a few seconds for rigid mols. and adding approx. 10 s per rotatable bond.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXkvFKjsA%253D%253D&md5=67a08aafeadf69101d56b2f60a922359
  21. 21
    Clark, A. J.; Tiwary, P.; Borrelli, K.; Feng, S.; Miller, E. B.; Abel, R.; Friesner, R. A.; Berne, B. J. Prediction of Protein-Ligand Binding Poses via a Combination of Induced Fit Docking and Metadynamics Simulations. J. Chem. Theory Comput. 2016, 12 (6), 2990– 2998,  DOI: 10.1021/acs.jctc.6b00201
    Google Scholar
    21
    Prediction of Protein-Ligand Binding Poses via a Combination of Induced Fit Docking and Metadynamics Simulations
    Clark, Anthony J.; Tiwary, Pratyush; Borrelli, Ken; Feng, Shulu; Miller, Edward B.; Abel, Robert; Friesner, Richard A.; Berne, B. J.
    Journal of Chemical Theory and Computation (2016), 12 (6), 2990-2998CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    Ligand docking is a widely used tool for lead discovery and binding mode prediction based drug discovery. The greatest challenges in docking occur when the receptor significantly reorganizes upon small mol. binding, thereby requiring an induced fit docking (IFD) approach in which the receptor is allowed to move in order to bind to the ligand optimally. IFD methods have had some success but suffer from a lack of reliability. Complementing IFD with all-atom mol. dynamics (MD) is a straightforward soln. in principle but not in practice due to the severe time scale limitations of MD. Here we introduce a metadynamics plus IFD strategy for accurate and reliable prediction of the structures of protein-ligand complexes at a practically useful computational cost. Our strategy allows treating this problem in full atomistic detail and in a computationally efficient manner and enhances the predictive power of IFD methods. We significantly increase the accuracy of the underlying IFD protocol across a large data set comprising 42 different ligand-receptor systems. We expect this approach to be of significant value in computationally driven drug design.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28Xnt12ktrw%253D&md5=d5a1211c0b5398ba78bab9aec4bddbeb
  22. 22
    Merz, K. M., Jr. Limits of Free Energy Computation for Protein-Ligand Interactions. J. Chem. Theory Comput. 2010, 6, 1769– 1776,  DOI: 10.1021/ct100102q
    Google Scholar
    22
    Limits of Free Energy Computation for Protein-Ligand Interactions
    Merz, Kenneth M., Jr.
    Journal of Chemical Theory and Computation (2010), 6 (5), 1769-1776CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    A detailed error anal. is presented for the computation of protein-ligand interaction energies. In particular, the authors show that it is probable that even highly accurate computed binding free energies have errors that represent a large percentage of the target free energies of binding. This is due to the observation that the error for computed energies quasi-linearly increases with the increasing no. of interactions present in a protein-ligand complex. This principle is expected to hold true for any system that involves an ever increasing no. of inter- or intramol. interactions (e.g., ab initio protein folding). The authors introduce the concept of best-case scenario errors (BCSerrors) that can be routinely applied to docking and scoring studies and that can be used to provide error bars for the computed binding free energies. These BCSerrors form a basis by which one can evaluate the outcome of a docking and scoring exercise. Moreover, the resultant error anal. enables the formation of an hypothesis that defines the best direction to proceed to improve scoring functions used in mol. docking studies.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXkslKksr4%253D&md5=e6ba0bd00675e195be2d565f2879fd1c
  23. 23
    Li, J.; Fu, A.; Zhang, L. An Overview of Scoring Functions Used for Protein-Ligand Interactions in Molecular Docking. Interdiscip. Sci. 2019, 11 (2), 320– 328,  DOI: 10.1007/s12539-019-00327-w
    Google Scholar
    23
    An Overview of Scoring Functions Used for Protein-Ligand Interactions in Molecular Docking
    Li, Jin; Fu, Ailing; Zhang, Le
    Interdisciplinary Sciences: Computational Life Sciences (2019), 11 (2), 320-328CODEN: ISCLAL; ISSN:1867-1462. (Springer GmbH)
    A review. Currently, mol. docking is becoming a key tool in drug discovery and mol. modeling applications. The reliability of mol. docking depends on the accuracy of the adopted scoring function, which can guide and det. the ligand poses when thousands of possible poses of ligand are generated. The scoring function can be used to det. the binding mode and site of a ligand, predict binding affinity and identify the potential drug leads for a given protein target. Despite intensive research over the years, accurate and rapid prediction of protein-ligand interactions is still a challenge in mol. docking. For this reason, this study reviews four basic types of scoring functions, physics-based, empirical, knowledge-based, and machine learning-based scoring functions, based on an up-to-date classification scheme. We not only discuss the foundations of the four types scoring functions, suitable application areas and shortcomings, but also discuss challenges and potential future study directions.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhtVKgur7O&md5=32d69a1e5260384d3d44d8d1e846c715
  24. 24
    Fischer, A.; Smiesko, M.; Sellner, M.; Lill, M. A. Decision Making in Structure-Based Drug Discovery: Visual Inspection of Docking Results. J. Med. Chem. 2021, 64 (5), 2489– 2500,  DOI: 10.1021/acs.jmedchem.0c02227
    Google Scholar
    24
    Decision making in structure-based drug discovery: visual inspection of docking results
    Fischer, Andre; Smiesko, Martin; Sellner, Manuel; Lill, Markus A.
    Journal of Medicinal Chemistry (2021), 64 (5), 2489-2500CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    A review. Mol. docking is a computational method widely used in drug discovery. Due to the inherent inaccuracies of mol. docking, visual inspection of binding modes is a crucial routine in the decision making process of computational medicinal chemists. Despite its apparent importance for medicinal chem. projects, guidelines for the visual docking pose assessment have been hardly discussed in the literature. Here, the authors review the medicinal chem. literature with the aim of identifying consistent principles for visual inspection, highlighting cases of its successful application, and discussing its limitations. In this context, the authors conducted a survey reaching experts in both academia and the pharmaceutical industry, which also included a challenge to distinguish native from incorrect poses. The authors were able to collect 93 expert opinions that offer valuable insights into visually supported decision-making processes. This perspective shall motivate discussions among experienced computational medicinal chemists and guide young scientists new to the field to stratify their compds.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3MXks1yjtrY%253D&md5=f5566c527c0889cdc8a6c2e4cc2c4881
  25. 25
    Shao, J.; Tanner, S. W.; Thompson, N.; Cheatham, T. E. Clustering Molecular Dynamics Trajectories: 1. Characterizing the Performance of Different Clustering Algorithms. J. Chem. Theory Comput. 2007, 3 (6), 2312– 2334,  DOI: 10.1021/ct700119m
    Google Scholar
    25
    Clustering Molecular Dynamics Trajectories: 1. Characterizing the Performance of Different Clustering Algorithms
    Shao, Jianyin; Tanner, Stephen W.; Thompson, Nephi; Cheatham, Thomas E., III
    Journal of Chemical Theory and Computation (2007), 3 (6), 2312-2334CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    Mol. dynamics simulation methods produce trajectories of at. positions (and optionally velocities and energies) as a function of time and provide a representation of the sampling of a given mol.'s energetically accessible conformational ensemble. As simulations on the 10-100 ns time scale become routine, with sampled configurations stored on the picosecond time scale, such trajectories contain large amts. of data. Data-mining techniques, like clustering, provide one means to group and make sense of the information in the trajectory. In this work, several clustering algorithms were implemented, compared, and utilized to understand MD trajectory data. The development of the algorithms into a freely available C code library, and their application to a simple test example of random (or systematically placed) points in a 2D plane (where the pairwise metric is the distance between points) provide a means to understand the relative performance. Eleven different clustering algorithms were developed, ranging from top-down splitting (hierarchical) and bottom-up aggregating (including single-linkage edge joining, centroid-linkage, av.-linkage, complete-linkage, centripetal, and centripetal-complete) to various refinement (means, Bayesian, and self-organizing maps) and tree (COBWEB) algorithms. Systematic testing in the context of MD simulation of various DNA systems (including DNA single strands and the interaction of a minor groove binding drug DB226 with a DNA hairpin) allows a more direct assessment of the relative merits of the distinct clustering algorithms. Addnl., means to assess the relative performance and differences between the algorithms, to dynamically select the initial cluster count, and to achieve faster data mining by "sieved clustering" were evaluated. Overall, it was found that there is no one perfect "one size fits all" algorithm for clustering MD trajectories and that the results strongly depend on the choice of atoms for the pairwise comparison. Some algorithms tend to produce homogeneously sized clusters, whereas others have a tendency to produce singleton clusters. Issues related to the choice of a pairwise metric, clustering metrics, which atom selection is used for the comparison, and about the relative performance are discussed. Overall, the best performance was obsd. with the av.-linkage, means, and SOM algorithms. If the cluster count is not known in advance, the hierarchical or av.-linkage clustering algorithms are recommended. Although these algorithms perform well, it is important to be aware of the limitations or weaknesses of each algorithm, specifically the high sensitivity to outliers with hierarchical, the tendency to generate homogenously sized clusters with means, and the tendency to produce small or singleton clusters with av.-linkage.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXhtFaqsLzE&md5=668cff11c29736325a4788379cc43def
  26. 26
    Wang, L.; Wu, Y.; Deng, Y.; Kim, B.; Pierce, L.; Krilov, G.; Lupyan, D.; Robinson, S.; Dahlgren, M. K.; Greenwood, J.; Romero, D. L.; Masse, C.; Knight, J. L.; Steinbrecher, T.; Beuming, T.; Damm, W.; Harder, E.; Sherman, W.; Brewer, M.; Wester, R.; Murcko, M.; Frye, L.; Farid, R.; Lin, T.; Mobley, D. L.; Jorgensen, W. L.; Berne, B. J.; Friesner, R. A.; Abel, R. Accurate and reliable prediction of relative ligand binding potency in prospective drug discovery by way of a modern free-energy calculation protocol and force field. J. Am. Chem. Soc. 2015, 137 (7), 2695– 2703,  DOI: 10.1021/ja512751q
    Google Scholar
    26
    Accurate and Reliable Prediction of Relative Ligand Binding Potency in Prospective Drug Discovery by Way of a Modern Free-Energy Calculation Protocol and Force Field
    Wang, Lingle; Wu, Yujie; Deng, Yuqing; Kim, Byungchan; Pierce, Levi; Krilov, Goran; Lupyan, Dmitry; Robinson, Shaughnessy; Dahlgren, Markus K.; Greenwood, Jeremy; Romero, Donna L.; Masse, Craig; Knight, Jennifer L.; Steinbrecher, Thomas; Beuming, Thijs; Damm, Wolfgang; Harder, Ed; Sherman, Woody; Brewer, Mark; Wester, Ron; Murcko, Mark; Frye, Leah; Farid, Ramy; Lin, Teng; Mobley, David L.; Jorgensen, William L.; Berne, Bruce J.; Friesner, Richard A.; Abel, Robert
    Journal of the American Chemical Society (2015), 137 (7), 2695-2703CODEN: JACSAT; ISSN:0002-7863. (American Chemical Society)
    Designing tight-binding ligands is a primary objective of small-mol. drug discovery. Over the past few decades, free-energy calcns. have benefited from improved force fields and sampling algorithms, as well as the advent of low-cost parallel computing. However, it has proven to be challenging to reliably achieve the level of accuracy that would be needed to guide lead optimization (∼5× in binding affinity) for a wide range of ligands and protein targets. Not surprisingly, widespread com. application of free-energy simulations has been limited due to the lack of large-scale validation coupled with the tech. challenges traditionally assocd. with running these types of calcns. Here, we report an approach that achieves an unprecedented level of accuracy across a broad range of target classes and ligands, with retrospective results encompassing 200 ligands and a wide variety of chem. perturbations, many of which involve significant changes in ligand chem. structures. In addn., we have applied the method in prospective drug discovery projects and found a significant improvement in the quality of the compds. synthesized that have been predicted to be potent. Compds. predicted to be potent by this approach have a substantial redn. in false positives relative to compds. synthesized on the basis of other computational or medicinal chem. approaches. Furthermore, the results are consistent with those obtained from our retrospective studies, demonstrating the robustness and broad range of applicability of this approach, which can be used to drive decisions in lead optimization.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXhsF2iuro%253D&md5=37a4f4a6c085f47ed531342643b6c33b
  27. 27
    Lee, T. S.; Allen, B. K.; Giese, T. J.; Guo, Z.; Li, P.; Lin, C.; McGee, T. D., Jr.; Pearlman, D. A.; Radak, B. K.; Tao, Y.; Tsai, H. C.; Xu, H.; Sherman, W.; York, D. M. Alchemical Binding Free Energy Calculations in AMBER20: Advances and Best Practices for Drug Discovery. J. Chem. Inf. Model. 2020, 60 (11), 5595– 5623,  DOI: 10.1021/acs.jcim.0c00613
    Google Scholar
    27
    Alchemical Binding Free Energy Calculations in AMBER20: Advances and Best Practices for Drug Discovery
    Lee, Tai-Sung; Allen, Bryce K.; Giese, Timothy J.; Guo, Zhenyu; Li, Pengfei; Lin, Charles; McGee Jr., T. Dwight; Pearlman, David A.; Radak, Brian K.; Tao, Yujun; Tsai, Hsu-Chun; Xu, Huafeng; Sherman, Woody; York, Darrin M.
    Journal of Chemical Information and Modeling (2020), 60 (11), 5595-5623CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    A review. Predicting protein-ligand binding affinities and the assocd. thermodn. of biomol. recognition is a primary objective of structure-based drug design. Alchem. free energy simulations offer a highly accurate and computationally efficient route to achieving this goal. While the AMBER mol. dynamics package has successfully been used for alchem. free energy simulations in academic research groups for decades, widespread impact in industrial drug discovery settings has been minimal because of the previous limitations within the AMBER alchem. code, coupled with challenges in system setup and postprocessing workflows. Through a close academia-industry collaboration we have addressed many of the previous limitations with an aim to improve accuracy, efficiency, and robustness of alchem. binding free energy simulations in industrial drug discovery applications. Here, we highlight some of the recent advances in AMBER20 with a focus on alchem. binding free energy (BFE) calcns., which are less computationally intensive than alternative binding free energy methods where full binding/unbinding paths are explored. In addn. to scientific and tech. advances in AMBER20, we also describe the essential practical aspects assocd. with running relative alchem. BFE calcns., along with recommendations for best practices, highlighting the importance not only of the alchem. simulation code but also the auxiliary functionalities and expertise required to obtain accurate and reliable results. This work is intended to provide a contemporary overview of the scientific, tech., and practical issues assocd. with running relative BFE simulations in AMBER20, with a focus on real-world drug discovery applications.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhvVahsL3L&md5=6c7385019de25c306df8c40eaf485cfc
  28. 28
    Zheng, Z.; Ucisik, M. N.; Merz Jr, K. M. The Movable Type Method Applied to Protein-Ligand Binding. J. Chem. Theory Comput. 2013, 9 (12), 5526– 5538,  DOI: 10.1021/ct4005992
    Google Scholar
    28
    The Movable Type Method Applied to Protein-Ligand Binding
    Zheng, Zheng; Ucisik, Melek N.; Merz, Kenneth M.
    Journal of Chemical Theory and Computation (2013), 9 (12), 5526-5538CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    Accurately computing the free energy for biol. processes like protein folding or protein-ligand assocn. remains a challenging problem. Both describing the complex intermol. forces involved and sampling the requisite configuration space make understanding these processes innately difficult. Herein, we address the sampling problem using a novel methodol. we term "movable type" (MT). Conceptually it can be understood by analogy with the evolution of printing and, hence, the name movable type. For example, a common approach to the study of protein-ligand complexation involves taking a database of intact drug-like mols. and exhaustively docking them into a binding pocket. This is reminiscent of early woodblock printing where each page had to be laboriously created prior to printing a book. However, printing evolved to an approach where a database of symbols (letters, numerals, etc.) was created and then assembled using a MT system, which allowed for the creation of all possible combinations of symbols on a given page, thereby, revolutionizing the dissemination of knowledge. Our MT method involves identifying all of the atom pairs seen in protein-ligand complexes and then creating two databases: one with their assocd. pairwise distant dependent energies and another assocd. with the probability of how these pairs can combine in terms of bonds, angles, dihedrals, and nonbonded interactions. Combining these two databases coupled with the principles of statistical mechanics allows us to accurately est. binding free energies as well as the pose of a ligand in a receptor. This method, by its math. construction, samples all of the configuration space of a selected region (the protein active site here) in one shot without resorting to brute force sampling schemes involving Monte Carlo, genetic algorithms, or mol. dynamics simulations making the methodol. extremely efficient. Importantly, this method explores the free energy surface eliminating the need to est. the enthalpy and entropy components individually. Finally, low free energy structures can be obtained via a free energy minimization procedure yielding all low free energy poses on a given free energy surface. Besides revolutionizing the protein-ligand docking and scoring problem, this approach can be utilized in a wide range of applications in computational biol. which involve the computation of free energies for systems with extensive phase spaces including protein folding, protein-protein docking, and protein design.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXhs1ygtr%252FF&md5=11f1ccafda165961647af0a96d17466c
  29. 29
    Stewart, J. J. Optimization of parameters for semiempirical methods V: modification of NDDO approximations and application to 70 elements. J. Mol. Model 2007, 13 (12), 1173– 1213,  DOI: 10.1007/s00894-007-0233-4
    Google Scholar
    29
    Optimization of parameters for semiempirical methods V: modification of NDDO approximations and application to 70 elements
    Stewart, James J. P.
    Journal of Molecular Modeling (2007), 13 (12), 1173-1213CODEN: JMMOFK; ISSN:0948-5023. (Springer GmbH)
    Several modifications that have been made to the NDDO core-core interaction term and to the method of parameter optimization are described. These changes have resulted in a more complete parameter optimization, called PM6, which has, in turn, allowed 70 elements to be parameterized. The av. unsigned error (AUE) between calcd. and ref. heats of formation for 4,492 species was 8.0 kcal mol-1. For the subset of 1,373 compds. involving only the elements H, C, N, O, F, P, S, Cl, and Br, the PM6 AUE was 4.4 kcal mol-1. The equivalent AUE for other methods were: RM1: 5.0, B3LYP 6-31G*: 5.2, PM5: 5.7, PM3: 6.3, HF 6-31G*: 7.4, and AM1: 10.0 kcal mol-1. Several long-standing faults in AM1 and PM3 have been cor. and significant improvements have been made in the prediction of geometries.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXhtlequr7N&md5=7d22928c8e3423f3ca8f2e7c77e6e6c9
  30. 30
    Zheng, Z.; Pei, J.; Bansal, N.; Liu, H.; Song, L. F.; Merz Jr, K. M. Generation of Pairwise Potentials Using Multidimensional Data Mining. J. Chem. Theory Comput. 2018, 14 (10), 5045– 5067,  DOI: 10.1021/acs.jctc.8b00516
    Google Scholar
    30
    Generation of Pairwise Potentials Using Multidimensional Data Mining
    Zheng, Zheng; Pei, Jun; Bansal, Nupur; Liu, Hao; Song, Lin Frank; Merz, Kenneth M.
    Journal of Chemical Theory and Computation (2018), 14 (10), 5045-5067CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    The rapid development of mol. structural databases provides the chem. community access to an enormous array of exptl. data that can be used to build and validate computational models. Using radial distribution functions collected from exptl. available X-ray and NMR structures, a no. of so-called statistical potentials have been developed over the years using the structural data mining strategy. These potentials have been developed within the context of the two-particle Kirkwood equation by extending its original use for isotropic monat. systems to anisotropic biomol. systems. However, the accuracy and the unclear phys. meaning of statistical potentials have long formed the central arguments against such methods. In this work, we present a new approach to generate mol. energy functions using structural data mining. Instead of employing the Kirkwood equation and introducing the "ref. state" approxn., we model the multidimensional probability distributions of the mol. system using graphical models and generate the target pairwise Boltzmann probabilities using the Bayesian field theory. Different from the current statistical potentials that mimic the "knowledge-based" PMF based on the 2-particle Kirkwood equation, the graphical-model-based structure-derived potential developed in this study focuses on the generation of lower-dimensional Boltzmann distributions of atoms through redn. of dimensionality. We have named this new scoring function GARF, and in this work we focus on the math. derivation of our novel approach followed by validation studies on its ability to predict protein-ligand interactions.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhs1KhtLjJ&md5=388d026039f54229c468c78e3fb68954
  31. 31
    Tian, C.; Kasavajhala, K.; Belfon, K. A. A.; Raguette, L.; Huang, H.; Migues, A. N.; Bickel, J.; Wang, Y.; Pincay, J.; Wu, Q.; Simmerling, C. ff19SB: Amino-Acid-Specific Protein Backbone Parameters Trained against Quantum Mechanics Energy Surfaces in Solution. J. Chem. Theory Comput. 2020, 16 (1), 528– 552,  DOI: 10.1021/acs.jctc.9b00591
    Google Scholar
    31
    ff19SB: Amino-Acid-Specific Protein Backbone Parameters Trained against Quantum Mechanics Energy Surfaces in Solution
    Tian Chuan; Kasavajhala Koushik; Belfon Kellon A A; Raguette Lauren; Huang He; Bickel John; Wang Yuzhang; Pincay Jorge; Simmerling Carlos; Tian Chuan; Kasavajhala Koushik; Belfon Kellon A A; Raguette Lauren; Huang He; Migues Angela N; Wang Yuzhang; Simmerling Carlos; Wu Qin
    Journal of chemical theory and computation (2020), 16 (1), 528-552 ISSN:.
    Molecular dynamics (MD) simulations have become increasingly popular in studying the motions and functions of biomolecules. The accuracy of the simulation, however, is highly determined by the molecular mechanics (MM) force field (FF), a set of functions with adjustable parameters to compute the potential energies from atomic positions. However, the overall quality of the FF, such as our previously published ff99SB and ff14SB, can be limited by assumptions that were made years ago. In the updated model presented here (ff19SB), we have significantly improved the backbone profiles for all 20 amino acids. We fit coupled φ/ψ parameters using 2D φ/ψ conformational scans for multiple amino acids, using as reference data the entire 2D quantum mechanics (QM) energy surface. We address the polarization inconsistency during dihedral parameter fitting by using both QM and MM in aqueous solution. Finally, we examine possible dependency of the backbone fitting on side chain rotamer. To extensively validate ff19SB parameters, and to compare to results using other Amber models, we have performed a total of ∼5 ms MD simulations in explicit solvent. Our results show that after amino-acid-specific training against QM data with solvent polarization, ff19SB not only reproduces the differences in amino-acid-specific Protein Data Bank (PDB) Ramachandran maps better but also shows significantly improved capability to differentiate amino-acid-dependent properties such as helical propensities. We also conclude that an inherent underestimation of helicity is present in ff14SB, which is (inexactly) compensated for by an increase in helical content driven by the TIP3P bias toward overly compact structures. In summary, ff19SB, when combined with a more accurate water model such as OPC, should have better predictive power for modeling sequence-specific behavior, protein mutations, and also rational protein design. Of the explicit water models tested here, we recommend use of OPC with ff19SB.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BB3MjnvFGisw%253D%253D&md5=7cbaecffea06e4bf7ccecb7f1d0a0f4d
  32. 32
    Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and testing of a general amber force field. J. Comput. Chem. 2004, 25 (9), 1157– 1174,  DOI: 10.1002/jcc.20035
    Google Scholar
    32
    Development and testing of a general Amber force field
    Wang, Junmei; Wolf, Romain M.; Caldwell, James W.; Kollman, Peter A.; Case, David A.
    Journal of Computational Chemistry (2004), 25 (9), 1157-1174CODEN: JCCHDD; ISSN:0192-8651. (John Wiley & Sons, Inc.)
    We describe here a general Amber force field (GAFF) for org. mols. GAFF is designed to be compatible with existing Amber force fields for proteins and nucleic acids, and has parameters for most org. and pharmaceutical mols. that are composed of H, C, N, O, S, P, and halogens. It uses a simple functional form and a limited no. of atom types, but incorporates both empirical and heuristic models to est. force consts. and partial at. charges. The performance of GAFF in test cases is encouraging. In test I, 74 crystallog. structures were compared to GAFF minimized structures, with a root-mean-square displacement of 0.26 Å, which is comparable to that of the Tripos 5.2 force field (0.25 Å) and better than those of MMFF 94 and CHARMm (0.47 and 0.44 Å, resp.). In test II, gas phase minimizations were performed on 22 nucleic acid base pairs, and the minimized structures and intermol. energies were compared to MP2/6-31G* results. The RMS of displacements and relative energies were 0.25 Å and 1.2 kcal/mol, resp. These data are comparable to results from Parm99/RESP (0.16 Å and 1.18 kcal/mol, resp.), which were parameterized to these base pairs. Test III looked at the relative energies of 71 conformational pairs that were used in development of the Parm99 force field. The RMS error in relative energies (compared to expt.) is about 0.5 kcal/mol. GAFF can be applied to wide range of mols. in an automatic fashion, making it suitable for rational drug design and database searching.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXksFakurc%253D&md5=2992017a8cf51f89290ae2562403b115
  33. 33
    Zheng, Z.; Borbulevych, O. Y.; Liu, H.; Deng, J.; Martin, R. I.; Westerhoff, L. M. MovableType Software for Fast Free Energy-Based Virtual Screening: Protocol Development, Deployment, Validation, and Assessment. J. Chem. Inf. Model 2020, 60 (11), 5437– 5456,  DOI: 10.1021/acs.jcim.0c00618
    Google Scholar
    33
    MovableType Software for Fast Free Energy-Based Virtual Screening: Protocol Development, Deployment, Validation, and Assessment
    Zheng, Zheng; Borbulevych, Oleg Y.; Liu, Hao; Deng, Jianpeng; Martin, Roger I.; Westerhoff, Lance M.
    Journal of Chemical Information and Modeling (2020), 60 (11), 5437-5456CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    For decades, the complicated energy surfaces found in macromol. protein:ligand structures, which require large amts. of computational time and resources for energy state sampling, have been an inherent obstacle to fast, routine free energy estn. in industrial drug discovery efforts. Beginning in 2013, the Merz research group addressed this cost with the introduction of a novel sampling methodol. termed "Movable Type" (MT). Using numerical integration methods, the MT method reduces the computational expense for energy state sampling by independently calcg. each at. partition function from an initial mol. conformation in order to est. the mol. free energy using ensembles of the at. partition functions. In this work, we report a software package, the DivCon Discovery Suite with the MovableType module from QuantumBio Inc., that performs this MT free energy estn. protocol in a fast, fully encapsulated manner. We discuss the computational procedures and improvements to the original work, and we detail the corresponding settings for this software package. Finally, we introduce two validation benchmarks to evaluate the overall robustness of the method against a broad range of protein:ligand structural cases. With these publicly available benchmarks, we show that the method can use a variety of input types and parameters and exhibits comparable predictability whether the method is presented with "expensive" X-ray structures or "inexpensively docked" theor. models. We also explore some next steps for the method. The MovableType software is available at http://www.quantumbioinc.com/.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhsF2nsb3L&md5=9f520f96b7938e8364bb1207ec39a417
  34. 34
    Pan, L. L.; Zheng, Z.; Wang, T.; Merz Jr, K. M. Free Energy-Based Conformational Search Algorithm Using the Movable Type Sampling Method. J. Chem. Theory Comput. 2015, 11 (12), 5853– 5864,  DOI: 10.1021/acs.jctc.5b00930
    Google Scholar
    34
    Free Energy-Based Conformational Search Algorithm Using the Movable Type Sampling Method
    Pan, Li-Li; Zheng, Zheng; Wang, Ting; Merz, Kenneth M.
    Journal of Chemical Theory and Computation (2015), 11 (12), 5853-5864CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    In this article, we extend the movable type (MT) sampling method to mol. conformational searches (MT-CS) on the free energy surface of the mol. in question. Differing from traditional systematic and stochastic searching algorithms, this method uses Boltzmann energy information to facilitate the selection of the best conformations. The generated ensembles provided good coverage of the available conformational space including available crystal structures. Furthermore, our approach directly provides the solvation free energies and the relative gas and aq. phase free energies for all generated conformers. The method is validated by a thorough anal. of thrombin ligands as well as against structures extd. from both the Protein Data Bank (PDB) and the Cambridge Structural Database (CSD). An in-depth comparison between OMEGA and MT-CS is presented to illustrate the differences between the two conformational searching strategies, i.e., energy-based vs. free energy-based searching. These studies demonstrate that our MT-based ligand conformational search algorithm is a powerful approach to delineate the conformational ensembles of mol. species on free energy surfaces.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXhvVSls7rO&md5=8a23f666b879ede91a547f1c596268fd
  35. 35
    Bansal, N.; Zheng, Z.; Merz Jr, K. M. Incorporation of side chain flexibility into protein binding pockets using MTflex. Bioorg. Med. Chem. 2016, 24 (20), 4978– 4987,  DOI: 10.1016/j.bmc.2016.08.030
    Google Scholar
    35
    Incorporation of side chain flexibility into protein binding pockets using MTflex
    Bansal, Nupur; Zheng, Zheng; Merz, Kenneth M., Jr.
    Bioorganic & Medicinal Chemistry (2016), 24 (20), 4978-4987CODEN: BMECEP; ISSN:0968-0896. (Elsevier B.V.)
    The plasticity of active sites plays a significant role in drug recognition and binding, but the accurate incorporation of receptor flexibility remains a significant computational challenge. Many approaches have been put forward to address receptor flexibility in docking studies by generating relevant ensembles on the energy surface. Herein, the authors describe the Movable Type with flexibility (MTflex) method that generates ensembles on the more relevant free energy surface in a computationally tractable manner. This novel approach enumerates conformational states based on side chain flexibility and then ests. their relative free energies using the MT methodol. The resultant conformational states can then be used in subsequent docking and scoring exercises. In particular, using the MTflex ensembles improves MT's ability to predict binding free energies over docking only to the crystal structure.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhsVOqu77E&md5=4769f5beae2495901d63ac12a493c80f
  36. 36
    Molecular Operating Environment (MOE); 2019.0102 Chemical Computing Group ULC: Montreal, QC, Canada, 2020.
    Google Scholar
    There is no corresponding record for this reference.
  37. 37
    Corbeil, C. R.; Williams, C. I.; Labute, P. Variability in docking success rates due to dataset preparation. J. Comput. Aided Mol. Des. 2012, 26 (6), 775– 786,  DOI: 10.1007/s10822-012-9570-1
    Google Scholar
    37
    Variability in docking success rates due to dataset preparation
    Corbeil, Christopher R.; Williams, Christopher I.; Labute, Paul
    Journal of Computer-Aided Molecular Design (2012), 26 (6), 775-786CODEN: JCADEQ; ISSN:0920-654X. (Springer)
    The results of cognate docking with the prepd. Astex dataset provided by the organizers of the "Docking and Scoring: A Review of Docking Programs" session at the 241st ACS national meeting are presented. The MOE software with the newly developed GBVI/WSA dG scoring function is used throughout the study. For 80 % of the Astex targets, the MOE docker produces a top-scoring pose within 2 Å of the X-ray structure. For 91 % of the targets a pose within 2 Å of the X-ray structure is produced in the top 30 poses. Docking failures, defined as cases where the top scoring pose is greater than 2 Å from the exptl. structure, are shown to be largely due to the absence of bound waters in the source dataset, highlighting the need to include these and other crucial information in future standardized sets. Docking success is shown to depend heavily on data prepn. A "dataset prepn." error of 0.5 kcal/mol is shown to cause fluctuations of over 20 % in docking success rates.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC38XhtVelu7nM&md5=8018caa912ccccc46349706526fb0485
  38. 38
    Case, D. A.; Ben-Shalom, I. Y.; Brozell, S. R.; Cerutti, D. S.; Cheatham, T. E.; Cruzeiro, V. W. D., III; Darden, T. A.; Duke, R. E.; Ghoreishi, D.; Gilson, M. K.; Gohlke, H.; Goetz, A. W.; Greene, D.; Harris, R.; Homeyer, N.; Huang, Y.; Izadi, S.; Kovalenko, A.; Kurtzman, T.; Lee, T. S.; LeGrand, S.; Li, P.; Lin, C.; Liu, J.; Luchko, T.; Luo, R.; Mermelstein, D. J.; Merz, K. M.; Miao, Y.; Monard, G.; Nguyen, C.; Nguyen, H.; Omelyan, I.; Onufriev, A.; Pan, F.; Qi, R.; Roe, D. R.; Roitberg, A.; Sagui, C.; Schott-Verdugo, S.; Shen, J.; Simmerling, C. L.; Smith, J.; Salomon-Ferrer, R.; Swails, J.; Walker, R. C.; Wang, J.; Wei, H.; Wolf, R. M.; Wu, X.; Xiao, L.; York, D. M.; Kollman, P. A. AMBER 2018; University of California: San Francisco, 2018.
    Google Scholar
    There is no corresponding record for this reference.
  39. 39
    Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.; Simmerling, C. Comparison of multiple Amber force fields and development of improved protein backbone parameters. Proteins 2006, 65 (3), 712– 725,  DOI: 10.1002/prot.21123
    Google Scholar
    39
    Comparison of multiple Amber force fields and development of improved protein backbone parameters
    Hornak, Viktor; Abel, Robert; Okur, Asim; Strockbine, Bentley; Roitberg, Adrian; Simmerling, Carlos
    Proteins: Structure, Function, and Bioinformatics (2006), 65 (3), 712-725CODEN: PSFBAF ISSN:. (Wiley-Liss, Inc.)
    The ff94 force field that is commonly assocd. with the Amber simulation package is one of the most widely used parameter sets for biomol. simulation. After a decade of extensive use and testing, limitations in this force field, such as over-stabilization of α-helixes, were reported by the authors and other researchers. This led to a no. of attempts to improve these parameters, resulting in a variety of "Amber" force fields and significant difficulty in detg. which should be used for a particular application. The authors show that several of these continue to suffer from inadequate balance between different secondary structure elements. In addn., the approach used in most of these studies neglected to account for the existence in Amber of two sets of backbone .vphi./ψ dihedral terms. This led to parameter sets that provide unreasonable conformational preferences for glycine. The authors report here an effort to improve the .vphi./ψ dihedral terms in the ff99 energy function. Dihedral term parameters are based on fitting the energies of multiple conformations of glycine and alanine tetrapeptides from high level ab initio quantum mech. calcns. The new parameters for backbone dihedrals replace those in the existing ff99 force field. This parameter set, which the authors denote ff99SB, achieves a better balance of secondary structure elements as judged by improved distribution of backbone dihedrals for glycine and alanine with respect to PDB survey data. It also accomplishes improved agreement with published exptl. data for conformational preferences of short alanine peptides and better accord with exptl. NMR relaxation data of test protein systems.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD28XhtFWqt7fM&md5=de683a26eca9e83ae524726e97ac22fa
  40. 40
    Li, P.; Roberts, B. P.; Chakravorty, D. K.; Merz Jr, K. M. Rational Design of Particle Mesh Ewald Compatible Lennard-Jones Parameters for + 2 Metal Cations in Explicit Solvent. J. Chem. Theory Comput. 2013, 9 (6), 2733– 2748,  DOI: 10.1021/ct400146w
    Google Scholar
    40
    Rational Design of Particle Mesh Ewald Compatible Lennard-Jones Parameters for +2 Metal Cations in Explicit Solvent
    Li, Pengfei; Roberts, Benjamin P.; Chakravorty, Dhruva K.; Merz, Kenneth M.
    Journal of Chemical Theory and Computation (2013), 9 (6), 2733-2748CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    Metal ions play significant roles in biol. systems. Accurate mol. dynamics (MD) simulations on these systems require a validated set of parameters. Although there are more detailed ways to model metal ions, the nonbonded model, which employs a 12-6 Lennard-Jones (LJ) term plus an electrostatic potential, is still widely used in MD simulations today due to its simple form. However, LJ parameters have limited transferability due to different combining rules, various water models, and diverse simulation methods. Recently, simulations employing a Particle Mesh Ewald (PME) treatment for long-range electrostatics have become more and more popular owing to their speed and accuracy. In the present work, we have systematically designed LJ parameters for 24 +2 metal (M(II)) cations to reproduce different exptl. properties appropriate for the Lorentz-Berthelot combining rules and PME simulations. We began by testing the transferability of currently available M(II) ion LJ parameters. The results showed that there are differences between simulations employing Ewald summation with other simulation methods and that it was necessary to design new parameters specific for PME based simulations. Employing the thermodn. integration (TI) method and performing periodic boundary MD simulations employing PME, allowed for a systematic investigation of the LJ parameter space. Hydration free energies (HFEs), the ion-oxygen distance in the first solvation shell (IOD), and coordination nos. (CNs) were obtained for various combinations of the parameters of the LJ potential for four widely used water models (TIP3P, SPC/E, TIP4P, and TIP4PEW). Results showed that the three simulated properties were highly correlated. Meanwhile, M(II) ions with the same parameters in different water models produce remarkably different HFEs but similar structural properties. It is difficult to reproduce various exptl. values simultaneously because the nonbonded model underestimates the interaction between the metal ions and water mols. at short-range. Moreover, the extent of underestimation increases successively for the TIP3P, SPC/E, TIP4PEW, and TIP4P water models. Nonetheless, we fitted a curve to describe the relationship between ε (the well depth) and radius (Rmin/2) from exptl. data on noble gases to facilitate the generation of the best possible compromise models. Hence, by targeting different exptl. values, we developed three sets of parameters for M(II) cations for three different water models (TIP3P, SPC/E, and TIP4PEW). These parameters we feel represent the best possible compromise that can be achieved using the nonbonded model for the ions in combination with simple water models. From a computational uncertainty anal. we est. that the uncertainty in our computed HFEs is on the order of ±1 kcal/mol. Further improvements will require more advanced nonbonded models likely with inclusion of polarization.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXntlGjsb0%253D&md5=e87f8df0191db081bcc7b2252806b61c
  41. 41
    Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103 (19), 8577– 8593,  DOI: 10.1063/1.470117
    Google Scholar
    41
    A smooth particle mesh Ewald method
    Essmann, Ulrich; Perera, Lalith; Berkowitz, Max L.; Darden, Tom; Lee, Hsing; Pedersen, Lee G.
    Journal of Chemical Physics (1995), 103 (19), 8577-93CODEN: JCPSA6; ISSN:0021-9606. (American Institute of Physics)
    The previously developed particle mesh Ewald method is reformulated in terms of efficient B-spline interpolation of the structure factors. This reformulation allows a natural extension of the method to potentials of the form 1/rp with p ≥ 1. Furthermore, efficient calcn. of the virial tensor follows. Use of B-splines in the place of Lagrange interpolation leads to analytic gradients as well as a significant improvement in the accuracy. The authors demonstrate that arbitrary accuracy can be achieved, independent of system size N, at a cost that scales as N log(N). For biomol. systems with many thousands of atoms and this method permits the use of Ewald summation at a computational cost comparable to that of a simple truncation method of 10 Å or less.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaK2MXptlehtrw%253D&md5=092a679dd3bee08da28df41e302383a7
  42. 42
    Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 1977, 23 (3), 327– 341,  DOI: 10.1016/0021-9991(77)90098-5
    Google Scholar
    42
    Numerical integration of the Cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes
    Ryckaert, Jean Paul; Ciccotti, Giovanni; Berendsen, Herman J. C.
    Journal of Computational Physics (1977), 23 (3), 327-41CODEN: JCTPAH; ISSN:0021-9991.
    A numerical algorithm integrating the 3N Cartesian equation of motion of a system of N points subject to holonomic constraints is applied to mol. dynamics simulation of a liq. of 64 butane mols.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaE2sXktVGhsL4%253D&md5=b4aecddfde149117813a5ea4f5353ce2
  43. 43
    Liu, H.; Deng, J.; Luo, Z.; Lin, Y.; Merz Jr, K. M.; Zheng, Z. Receptor-Ligand Binding Free Energies from a Consecutive Histograms Monte Carlo Sampling Method. J. Chem. Theory Comput. 2020, 16 (10), 6645– 6655,  DOI: 10.1021/acs.jctc.0c00457
    Google Scholar
    43
    Receptor-Ligand Binding Free Energies from a Consecutive Histograms Monte Carlo Sampling Method
    Liu, Hao; Deng, Jianpeng; Luo, Zhou; Lin, Yawei; Merz Jr., Kenneth M.; Zheng, Zheng
    Journal of Chemical Theory and Computation (2020), 16 (10), 6645-6655CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    To obtain accurate and converged free energy calcns. for ligand binding to biomol. systems requires validated force fields and extensive sampling of the energy landscape, which requires exhaustive and effective conformational searching methods. Herein, we introduce the consecutive histograms Monte Carlo (CHMC) sampling protocol that generates receptor-ligand binding modes within a series of continuously distributed sampling units ranging from placement near the geometric center of the receptor's binding site to fully unbound states. This protocol employs independent energy-state sampling for calcg. the ensemble energy within every predefined location along the receptor-ligand dissocn. pathway, without the need to traverse the energy barriers as in mol. dynamic simulations during the dissocn. procedure. We applied this method to a set of selected receptor targets with their corresponding ligands providing detailed studies of mol. binding free energy predictions. The results show that the CHMC gives an excellent accounting of the free energy surfaces and binding free energies at a reasonable computational cost.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhs12rt7%252FI&md5=2071782691a1892e807eb69d13c34c94
  44. 44
    Schindler, C. E. M.; Baumann, H.; Blum, A.; Bose, D.; Buchstaller, H. P.; Burgdorf, L.; Cappel, D.; Chekler, E.; Czodrowski, P.; Dorsch, D.; Eguida, M. K. I.; Follows, B.; Fuchss, T.; Gradler, U.; Gunera, J.; Johnson, T.; Jorand Lebrun, C.; Karra, S.; Klein, M.; Knehans, T.; Koetzner, L.; Krier, M.; Leiendecker, M.; Leuthner, B.; Li, L.; Mochalkin, I.; Musil, D.; Neagu, C.; Rippmann, F.; Schiemann, K.; Schulz, R.; Steinbrecher, T.; Tanzer, E. M.; Unzue Lopez, A.; Viacava Follis, A.; Wegener, A.; Kuhn, D. Large-Scale Assessment of Binding Free Energy Calculations in Active Drug Discovery Projects. J. Chem. Inf. Model 2020, 60 (11), 5457– 5474,  DOI: 10.1021/acs.jcim.0c00900
    Google Scholar
    44
    Large-Scale Assessment of Binding Free Energy Calculations in Active Drug Discovery Projects
    Schindler, Christina E. M.; Baumann, Hannah; Blum, Andreas; Boese, Dietrich; Buchstaller, Hans-Peter; Burgdorf, Lars; Cappel, Daniel; Chekler, Eugene; Czodrowski, Paul; Dorsch, Dieter; Eguida, Merveille K. I.; Follows, Bruce; Fuchss, Thomas; Graedler, Ulrich; Gunera, Jakub; Johnson, Theresa; Jorand Lebrun, Catherine; Karra, Srinivasa; Klein, Markus; Knehans, Tim; Koetzner, Lisa; Krier, Mireille; Leiendecker, Matthias; Leuthner, Birgitta; Li, Liwei; Mochalkin, Igor; Musil, Djordje; Neagu, Constantin; Rippmann, Friedrich; Schiemann, Kai; Schulz, Robert; Steinbrecher, Thomas; Tanzer, Eva-Maria; Unzue Lopez, Andrea; Viacava Follis, Ariele; Wegener, Ansgar; Kuhn, Daniel
    Journal of Chemical Information and Modeling (2020), 60 (11), 5457-5474CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
    Accurate ranking of compds. with regards to their binding affinity to a protein using computational methods is of great interest to pharmaceutical research. Physics-based free energy calcns. are regarded as the most rigorous way to est. binding affinity. In recent years, many retrospective studies carried out both in academia and industry have demonstrated its potential. Here, we present the results of large-scale prospective application of the FEP+ method in active drug discovery projects in an industry setting at Merck KGaA, Darmstadt, Germany. We compare these prospective data to results obtained on a new diverse, public benchmark of eight pharmaceutically relevant targets. Our results offer insights into the challenges faced when using free energy calcns. in real-life drug discovery projects and identify limitations that could be tackled by future method development. The new public data set we provide to the community can support further method development and comparative benchmarking of free energy calcns.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhs1CktrbE&md5=f2372d56cd501717435ed8095dd0dbb0
  45. 45
    Schiemann, K.; Mallinger, A.; Wienke, D.; Esdar, C.; Poeschke, O.; Busch, M.; Rohdich, F.; Eccles, S. A.; Schneider, R.; Raynaud, F. I.; Czodrowski, P.; Musil, D.; Schwarz, D.; Urbahns, K.; Blagg, J. Discovery of potent and selective CDK8 inhibitors from an HSP90 pharmacophore. Bioorg. Med. Chem. Lett. 2016, 26 (5), 1443– 1451,  DOI: 10.1016/j.bmcl.2016.01.062
    Google Scholar
    45
    Discovery of potent and selective CDK8 inhibitors from an HSP90 pharmacophore
    Schiemann, Kai; Mallinger, Aurelie; Wienke, Dirk; Esdar, Christina; Poeschke, Oliver; Busch, Michael; Rohdich, Felix; Eccles, Suzanne A.; Schneider, Richard; Raynaud, Florence I.; Czodrowski, Paul; Musil, Djordje; Schwarz, Daniel; Urbahns, Klaus; Blagg, Julian
    Bioorganic & Medicinal Chemistry Letters (2016), 26 (5), 1443-1451CODEN: BMCLE8; ISSN:0960-894X. (Elsevier B.V.)
    Here we describe the discovery and optimization of 3-benzylindazoles as potent and selective inhibitors of CDK8, also modulating CDK19, discovered from a high-throughput screening (HTS) campaign sampling the Merck compd. collection. The primary hits with strong HSP90 affinity were subsequently optimized to potent and selective CDK8 inhibitors which demonstrate inhibition of WNT pathway activity in cell-based assays. X-ray crystallog. data demonstrated that 3-benzylindazoles occupy the ATP binding site of CDK8 and adopt a Type I binding mode. Medicinal chem. optimization successfully led to improved potency, physicochem. properties and oral pharmacokinetics. Modulation of phospho-STAT1, a pharmacodynamic biomarker of CDK8, was demonstrated in an APC-mutant SW620 human colorectal carcinoma xenograft model following oral administration.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XitlCitLs%253D&md5=77e9f7705c9136557195ccfb4982cb8d
  46. 46
    Dorsch, D.; Schadt, O.; Stieber, F.; Meyring, M.; Gradler, U.; Bladt, F.; Friese-Hamim, M.; Knuhl, C.; Pehl, U.; Blaukat, A. Identification and optimization of pyridazinones as potent and selective c-Met kinase inhibitors. Bioorg. Med. Chem. Lett. 2015, 25 (7), 1597– 1602,  DOI: 10.1016/j.bmcl.2015.02.002
    Google Scholar
    46
    Identification and optimization of pyridazinones as potent and selective c-Met kinase inhibitors
    Dorsch, Dieter; Schadt, Oliver; Stieber, Frank; Meyring, Michael; Graedler, Ulrich; Bladt, Friedhelm; Friese-Hamim, Manja; Knuehl, Christine; Pehl, Ulrich; Blaukat, Andree
    Bioorganic & Medicinal Chemistry Letters (2015), 25 (7), 1597-1602CODEN: BMCLE8; ISSN:0960-894X. (Elsevier B.V.)
    In a high-throughput screening campaign for c-Met kinase inhibitors, a thiadiazinone deriv. with a carbamate group was identified as a potent in vitro inhibitor. Subsequent optimization guided by c-Met-inhibitor X-ray structures furnished new compd. classes with excellent in vitro and in vivo profiles. The thiadiazinone ring of the HTS hit was first replaced by a pyridazinone followed by an exchange of the carbamate hinge binder with a 1,5-disubstituted pyrimidine. Finally an optimized compd., 3-(1-[3-[5-(1-methyl-piperidin-4-yl-methoxy)-pyrimidin-2-yl]-benzyl]-6-oxo-1,6-dihydro-pyridazin-3-yl)-benzonitrile, (MSC2156119), with excellent in vitro potency, high kinase selectivity, long half-life after oral administration and in vivo anti-tumor efficacy at low doses, was selected as a candidate for clin. development.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXivVSqsLg%253D&md5=dadaea6c4c8a0bda928def6f356c416d
  47. 47
    Schiemann, K.; Finsinger, D.; Zenke, F.; Amendt, C.; Knochel, T.; Bruge, D.; Buchstaller, H. P.; Emde, U.; Stahle, W.; Anzali, S. The discovery and optimization of hexahydro-2H-pyrano[3,2-c]quinolines (HHPQs) as potent and selective inhibitors of the mitotic kinesin-5. Bioorg. Med. Chem. Lett. 2010, 20 (5), 1491– 1495,  DOI: 10.1016/j.bmcl.2010.01.110
    Google Scholar
    47
    The discovery and optimization of hexahydro-2H-pyrano[3,2-c]quinolines (HHPQs) as potent and selective inhibitors of the mitotic kinesin-5
    Schiemann, Kai; Finsinger, Dirk; Zenke, Frank; Amendt, Christiane; Knoechel, Thorsten; Bruge, David; Buchstaller, Hans-Peter; Emde, Ulrich; Staehle, Wolfgang; Anzali, Soheila
    Bioorganic & Medicinal Chemistry Letters (2010), 20 (5), 1491-1495CODEN: BMCLE8; ISSN:0960-894X. (Elsevier B.V.)
    Here we describe the discovery and optimization of hexahydro-2H-pyrano[3,2-c]quinolines (HHPQs) as potent and selective inhibitors of the mitotic kinesin-5 originally found during a high-throughput screening (HTS) campaign sampling our inhouse compd. collection. The compds. optimized subsequently and characterized herein were potently inhibiting the ATPase activity of Kinesin-5 and also exhibited consistent cellular activity, in that cells arrested in mitosis and apoptosis induction could be obsd. X-ray crystallog. data demonstrated that these inhibitors bind in an allosteric pocket of Kinesin-5 distant from the nucleotide and microtubule binding sites. The selected clin. candidate EMD 534085 (I) caused strong growth inhibition in human tumor xenograft models using Colo 205 colon carcinoma cells at doses below 30 mg/kg administered twice weekly without showing severe toxicity as detd. by loss of body wt.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXitleiu7s%253D&md5=33138cecfe9a0f094109185ff010c7c6
  48. 48
    Wehn, P. M.; Rizzi, J. P.; Dixon, D. D.; Grina, J. A.; Schlachter, S. T.; Wang, B.; Xu, R.; Yang, H.; Du, X.; Han, G.; Wang, K.; Cao, Z.; Cheng, T.; Czerwinski, R. M.; Goggin, B. S.; Huang, H.; Halfmann, M. M.; Maddie, M. A.; Morton, E. L.; Olive, S. R.; Tan, H.; Xie, S.; Wong, T.; Josey, J. A.; Wallace, E. M. Design and Activity of Specific Hypoxia-Inducible Factor-2alpha (HIF-2alpha) Inhibitors for the Treatment of Clear Cell Renal Cell Carcinoma: Discovery of Clinical Candidate (S)-3-((2,2-Difluoro-1-hydroxy-7-(methylsulfonyl)-2,3-dihydro-1 H-inden-4-yl)oxy)-5-fluorobenzonitrile (PT2385). J. Med. Chem. 2018, 61 (21), 9691– 9721,  DOI: 10.1021/acs.jmedchem.8b01196
    Google Scholar
    48
    Design and Activity of Specific Hypoxia-Inducible Factor-2α (HIF-2α) Inhibitors for the Treatment of Clear Cell Renal Cell Carcinoma: Discovery of Clinical Candidate (S)-3-((2,2-Difluoro-1-hydroxy-7-(methylsulfonyl)-2,3-dihydro-1H-inden-4-yl)oxy)-5-fluorobenzonitrile (PT2385)
    Wehn, Paul M.; Rizzi, James P.; Dixon, Darryl D.; Grina, Jonas A.; Schlachter, Stephen T.; Wang, Bin; Xu, Rui; Yang, Hanbiao; Du, Xinlin; Han, Guangzhou; Wang, Keshi; Cao, Zhaodan; Cheng, Tzuling; Czerwinski, Robert M.; Goggin, Barry S.; Huang, Heli; Halfmann, Megan M.; Maddie, Melissa A.; Morton, Emily L.; Olive, Sarah R.; Tan, Huiling; Xie, Shanhai; Wong, Tai; Josey, John A.; Wallace, Eli M.
    Journal of Medicinal Chemistry (2018), 61 (21), 9691-9721CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    HIF-2α, a member of the HIF family of transcription factors, is a key oncogenic driver in cancers such as clear cell renal cell carcinoma (ccRCC). A signature feature of these cancers is the overaccumulation of HIF-2α protein, often by inactivation of the E3 ligase VHL (von Hippel-Lindau). Herein the authors disclose their structure based drug design (SBDD) approach that culminated in the identification of PT2385, the first HIF-2α antagonist to enter clin. trials. Highlights include the use of a putative n → π*Ar interaction to guide early analog design, the conformational restriction of an essential hydroxyl moiety, and the remarkable impact of fluorination near the hydroxyl group. Evaluation of select compds. from two structural classes in a sequence of PK/PD, efficacy, PK, and metabolite profiling identified I (PT2385, luciferase EC50 = 27 nM) as the clin. candidate. Finally, a retrospective crystallog. anal. describes the structural perturbations necessary for efficient antagonism.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhvVKqt7%252FN&md5=fcbb754286f47e1b1df3edb62bd2381e
  49. 49
    Boutard, N.; Bialas, A.; Sabiniarz, A.; Guzik, P.; Banaszak, K.; Biela, A.; Bien, M.; Buda, A.; Bugaj, B.; Cieluch, E.; Cierpich, A.; Dudek, L.; Eggenweiler, H. M.; Fogt, J.; Gaik, M.; Gondela, A.; Jakubiec, K.; Jurzak, M.; Kitlinska, A.; Kowalczyk, P.; Kujawa, M.; Kwiecinska, K.; Les, M.; Lindemann, R.; Maciuszek, M.; Mikulski, M.; Niedziejko, P.; Obara, A.; Pawlik, H.; Rzymski, T.; Sieprawska-Lupa, M.; Sowinska, M.; Szeremeta-Spisak, J.; Stachowicz, A.; Tomczyk, M. M.; Wiklik, K.; Wloszczak, L.; Ziemianska, S.; Zarebski, A.; Brzozka, K.; Nowak, M.; Fabritius, C. H. Discovery and Structure-Activity Relationships of N-Aryl 6-Aminoquinoxalines as Potent PFKFB3 Kinase Inhibitors. ChemMedChem. 2019, 14 (1), 169– 181,  DOI: 10.1002/cmdc.201800569
    Google Scholar
    49
    Discovery and structure-activity relationships of N-aryl 6-aminoquinoxalines as potent PFKFB3 kinase inhibitors
    Boutard, Nicolas; Bialas, Arkadiusz; Sabiniarz, Aleksandra; Guzik, Pawel; Banaszak, Katarzyna; Biela, Artur; Bien, Marcin; Buda, Anna; Bugaj, Barbara; Cieluch, Ewelina; Cierpich, Anna; Dudek, Lukasz; Eggenweiler, Hans-Michael; Fogt, Joanna; Gaik, Monika; Gondela, Andrzej; Jakubiec, Krzysztof; Jurzak, Mirek; Kitlinska, Agata; Kowalczyk, Piotr; Kujawa, Maciej; Kwiecinska, Katarzyna; Les, Marcin; Lindemann, Ralph; Maciuszek, Monika; Mikulski, Maciej; Niedziejko, Paulina; Obara, Alicja; Pawlik, Henryk; Rzymski, Tomasz; Sieprawska-Lupa, Magdalena; Sowinska, Marta; Szeremeta-Spisak, Joanna; Stachowicz, Agata; Tomczyk, Mateusz M.; Wiklik, Katarzyna; Wloszczak, Lukasz; Ziemianska, Sylwia; Zarebski, Adrian; Brzozka, Krzysztof; Nowak, Mateusz; Fabritius, Charles-Henry
    ChemMedChem (2019), 14 (1), 169-181CODEN: CHEMGX; ISSN:1860-7179. (Wiley-VCH Verlag GmbH & Co. KGaA)
    Energy and biomass prodn. in cancer cells are largely supported by aerobic glycolysis in what is called the Warburg effect. The process is regulated by key enzymes, among which phosphofructokinase PFK-2 plays a significant role by producing fructose-2,6-biphosphate; the most potent activator of the glycolysis rate-limiting step performed by phosphofructokinase PFK-1. Herein, the synthesis, biol. evaluation and structure-activity relationship of novel inhibitors of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3), which is the ubiquitous and hypoxia-induced isoform of PFK-2, are reported. X-ray crystallog. and docking were instrumental in the design and optimization of a series of N-aryl 6-aminoquinoxalines. The most potent representative, N-(4-methanesulfonylpyridin-3-yl)-8-(3-methyl-1-benzothiophen-5-yl)quinoxalin-6-amine, displayed an IC50 of 14 nM for the target and an IC50 of 0.49 μM for fructose-2,6-biphosphate prodn. in human colon carcinoma HCT116 cells. This work provides a new entry in the field of PFKFB3 inhibitors with potential for development in oncol.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXisVaktbrI&md5=3bfdd82269cdaf101cf7e1a4eaceafe5
  50. 50
    Garcia Fortanet, J.; Chen, C. H.; Chen, Y. N.; Chen, Z.; Deng, Z.; Firestone, B.; Fekkes, P.; Fodor, M.; Fortin, P. D.; Fridrich, C.; Grunenfelder, D.; Ho, S.; Kang, Z. B.; Karki, R.; Kato, M.; Keen, N.; LaBonte, L. R.; Larrow, J.; Lenoir, F.; Liu, G.; Liu, S.; Lombardo, F.; Majumdar, D.; Meyer, M. J.; Palermo, M.; Perez, L.; Pu, M.; Ramsey, T.; Sellers, W. R.; Shultz, M. D.; Stams, T.; Towler, C.; Wang, P.; Williams, S. L.; Zhang, J. H.; LaMarche, M. J. Allosteric Inhibition of SHP2: Identification of a Potent, Selective, and Orally Efficacious Phosphatase Inhibitor. J. Med. Chem. 2016, 59 (17), 7773– 7782,  DOI: 10.1021/acs.jmedchem.6b00680
    Google Scholar
    50
    Allosteric Inhibition of SHP2: Identification of a Potent, Selective, and Orally Efficacious Phosphatase Inhibitor
    Garcia Fortanet, Jorge; Chen, Christine Hiu-Tung; Chen, Ying-Nan P.; Chen, Zhouliang; Deng, Zhan; Firestone, Brant; Fekkes, Peter; Fodor, Michelle; Fortin, Pascal D.; Fridrich, Cary; Grunenfelder, Denise; Ho, Samuel; Kang, Zhao B.; Karki, Rajesh; Kato, Mitsunori; Keen, Nick; LaBonte, Laura R.; Larrow, Jay; Lenoir, Francois; Liu, Gang; Liu, Shumei; Lombardo, Franco; Majumdar, Dyuti; Meyer, Matthew J.; Palermo, Mark; Perez, Lawrence; Pu, Minying; Ramsey, Timothy; Sellers, William R.; Shultz, Michael D.; Stams, Travis; Towler, Christopher; Wang, Ping; Williams, Sarah L.; Zhang, Ji-Hu; LaMarche, Matthew J.
    Journal of Medicinal Chemistry (2016), 59 (17), 7773-7782CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    SHP2 is a nonreceptor protein tyrosine phosphatase (PTP) encoded by the PTPN11 gene involved in cell growth and differentiation via the MAPK signaling pathway. SHP2 also purportedly plays an important role in the programmed cell death pathway (PD-1/PD-L1). Because it is an oncoprotein assocd. with multiple cancer-related diseases, as well as a potential immunomodulator, controlling SHP2 activity is of significant therapeutic interest. Recently in our labs., a small mol. inhibitor of SHP2 was identified as an allosteric modulator that stabilizes the autoinhibited conformation of SHP2. A high throughput screen was performed to identify progressable chem. matter, and X-ray crystallog. revealed the location of binding in a previously undisclosed allosteric binding pocket. Structure-based drug design was employed to optimize for SHP2 inhibition, and several new protein-ligand interactions were characterized. These studies culminated in the discovery of 6-(4-amino-4-methylpiperidin-1-yl)-3-(2,3-dichlorophenyl)pyrazin-2-amine (SHP099, 1), a potent, selective, orally bioavailable, and efficacious SHP2 inhibitor.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhtVKju7fP&md5=dfe43b5253e172576386f86797087f7b
  51. 51
    Chen, Y. N.; LaMarche, M. J.; Chan, H. M.; Fekkes, P.; Garcia-Fortanet, J.; Acker, M. G.; Antonakos, B.; Chen, C. H.; Chen, Z.; Cooke, V. G.; Dobson, J. R.; Deng, Z.; Fei, F.; Firestone, B.; Fodor, M.; Fridrich, C.; Gao, H.; Grunenfelder, D.; Hao, H. X.; Jacob, J.; Ho, S.; Hsiao, K.; Kang, Z. B.; Karki, R.; Kato, M.; Larrow, J.; La Bonte, L. R.; Lenoir, F.; Liu, G.; Liu, S.; Majumdar, D.; Meyer, M. J.; Palermo, M.; Perez, L.; Pu, M.; Price, E.; Quinn, C.; Shakya, S.; Shultz, M. D.; Slisz, J.; Venkatesan, K.; Wang, P.; Warmuth, M.; Williams, S.; Yang, G.; Yuan, J.; Zhang, J. H.; Zhu, P.; Ramsey, T.; Keen, N. J.; Sellers, W. R.; Stams, T.; Fortin, P. D. Allosteric inhibition of SHP2 phosphatase inhibits cancers driven by receptor tyrosine kinases. Nature 2016, 535 (7610), 148– 152,  DOI: 10.1038/nature18621
    Google Scholar
    51
    Allosteric inhibition of SHP2 phosphatase inhibits cancers driven by receptor tyrosine kinases
    Chen, Ying-Nan P.; LaMarche, Matthew J.; Chan, Ho Man; Fekkes, Peter; Garcia-Fortanet, Jorge; Acker, Michael G.; Antonakos, Brandon; Chen, Christine Hiu-Tung; Chen, Zhouliang; Cooke, Vesselina G.; Dobson, Jason R.; Deng, Zhan; Fei, Feng; Firestone, Brant; Fodor, Michelle; Fridrich, Cary; Gao, Hui; Grunenfelder, Denise; Hao, Huai-Xiang; Jacob, Jaison; Ho, Samuel; Hsiao, Kathy; Kang, Zhao B.; Karki, Rajesh; Kato, Mitsunori; Larrow, Jay; La Bonte, Laura R.; Lenoir, Francois; Liu, Gang; Liu, Shumei; Majumdar, Dyuti; Meyer, Matthew J.; Palermo, Mark; Perez, Lawrence; Pu, Minying; Price, Edmund; Quinn, Christopher; Shakya, Subarna; Shultz, Michael D.; Slisz, Joanna; Venkatesan, Kavitha; Wang, Ping; Warmuth, Markus; Williams, Sarah; Yang, Guizhi; Yuan, Jing; Zhang, Ji-Hu; Zhu, Ping; Ramsey, Timothy; Keen, Nicholas J.; Sellers, William R.; Stams, Travis; Fortin, Pascal D.
    Nature (London, United Kingdom) (2016), 535 (7610), 148-152CODEN: NATUAS; ISSN:0028-0836. (Nature Publishing Group)
    The non-receptor protein tyrosine phosphatase SHP2, encoded by PTPN11, has an important role in signal transduction downstream of growth factor receptor signalling and was the first reported oncogenic tyrosine phosphatase. Activating mutations of SHP2 have been assocd. with developmental pathologies such as Noonan syndrome and are found in multiple cancer types, including leukemia, lung and breast cancer and neuroblastoma. SHP2 is ubiquitously expressed and regulates cell survival and proliferation primarily through activation of the RAS-ERK signalling pathway2,3. It is also a key mediator of the programmed cell death 1 (PD-1) and B- and T-lymphocyte attenuator (BTLA) immune checkpoint pathways. Redn. of SHP2 activity suppresses tumor cell growth and is a potential target of cancer therapy. Here we report the discovery of a highly potent (IC50 = 0.071 μM), selective and orally bioavailable small-mol. SHP2 inhibitor, SHP099, that stabilizes SHP2 in an auto-inhibited conformation. SHP099 concurrently binds to the interface of the N-terminal SH2, C-terminal SH2, and protein tyrosine phosphatase domains, thus inhibiting SHP2 activity through an allosteric mechanism. SHP099 suppresses RAS-ERK signalling to inhibit the proliferation of receptor-tyrosine-kinase-driven human cancer cells in vitro and is efficacious in mouse tumor xenograft models. Together, these data demonstrate that pharmacol. inhibition of SHP2 is a valid therapeutic approach for the treatment of cancers.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhtVOns7bI&md5=ec8a4b612d5008a2519bd846d45ce019
  52. 52
    Currie, K. S.; Kropf, J. E.; Lee, T.; Blomgren, P.; Xu, J.; Zhao, Z.; Gallion, S.; Whitney, J. A.; Maclin, D.; Lansdon, E. B.; Maciejewski, P.; Rossi, A. M.; Rong, H.; Macaluso, J.; Barbosa, J.; Di Paolo, J. A.; Mitchell, S. A. Discovery of GS-9973, a selective and orally efficacious inhibitor of spleen tyrosine kinase. J. Med. Chem. 2014, 57 (9), 3856– 3873,  DOI: 10.1021/jm500228a
    Google Scholar
    52
    Discovery of GS-9973, a Selective and Orally Efficacious Inhibitor of Spleen Tyrosine Kinase
    Currie, Kevin S.; Kropf, Jeffrey E.; Lee, Tony; Blomgren, Peter; Xu, Jianjun; Zhao, Zhongdong; Gallion, Steve; Whitney, J. Andrew; Maclin, Deborah; Lansdon, Eric B.; Maciejewski, Patricia; Rossi, Ann Marie; Rong, Hong; Macaluso, Jennifer; Barbosa, James; Di Paolo, Julie A.; Mitchell, Scott A.
    Journal of Medicinal Chemistry (2014), 57 (9), 3856-3873CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    Spleen tyrosine kinase (Syk) is an attractive drug target in autoimmune, inflammatory, and oncol. disease indications. The most advanced Syk inhibitor, R406, (or its prodrug form fostamatinib), has shown efficacy in multiple therapeutic indications, but its clin. progress has been hampered by dose-limiting adverse effects that have been attributed, at least in part, to the off-target activities of R406. It is expected that a more selective Syk inhibitor would provide a greater therapeutic window. Herein the authors report the discovery and optimization of a novel series of imidazo[1,2-a]pyrazine Syk inhibitors. This work culminated in the identification of GS-9973, a highly selective and orally efficacious Syk inhibitor which is currently undergoing clin. evaluation for autoimmune and oncol. indications.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXls1CrsL4%253D&md5=fc53448de1f5416f2c6600876f8c6385
  53. 53
    Buchstaller, H. P.; Anlauf, U.; Dorsch, D.; Kuhn, D.; Lehmann, M.; Leuthner, B.; Musil, D.; Radtki, D.; Ritzert, C.; Rohdich, F.; Schneider, R.; Esdar, C. Discovery and Optimization of 2-Arylquinazolin-4-ones into a Potent and Selective Tankyrase Inhibitor Modulating Wnt Pathway Activity. J. Med. Chem. 2019, 62 (17), 7897– 7909,  DOI: 10.1021/acs.jmedchem.9b00656
    Google Scholar
    53
    Discovery and Optimization of 2-Arylquinazolin-4-ones into a Potent and Selective Tankyrase Inhibitor Modulating Wnt Pathway Activity
    Buchstaller, Hans-Peter; Anlauf, Uwe; Dorsch, Dieter; Kuhn, Daniel; Lehmann, Martin; Leuthner, Birgitta; Musil, Djordje; Radtki, Daniela; Ritzert, Claudio; Rohdich, Felix; Schneider, Richard; Esdar, Christina
    Journal of Medicinal Chemistry (2019), 62 (17), 7897-7909CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
    Tankyrases 1 and 2 (TNKS1/2) are promising pharmacol. targets which recently gained interest for anticancer therapy in Wnt pathway dependent tumors. 2-Aryl-quinazolinones were identified and optimized into potent tankyrase inhibitors through SAR exploration around the quinazolinone core and the 4'-position of the Ph residue. These efforts were supported by anal. of TNKS X-ray and Watermap structures and resulted in compd. 5k(I), a potent, selective tankyrase inhibitor with favorable pharmacokinetic properties. The X-ray structure of I in complex with TNKS1 was solved and confirmed the design hypothesis. Modulation of Wnt pathway activity was demonstrated with this compd. in a colorectal xenograft model in vivo.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhsFWms7bI&md5=7fcc6c91c6e614482386af2094675837
  54. 54
    Jakalian, A.; Bush, B. L.; Jack, D. B.; Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: I. Method. J. Comput. Chem. 2000, 21 (2), 132– 146,  DOI: 10.1002/(SICI)1096-987X(20000130)21:2<132::AID-JCC5>3.0.CO;2-P
    Google Scholar
    54
    Fast, efficient generation of high-quality atomic charges. AM1-BCC model: I. Method
    Jakalian, Araz; Bush, Bruce L.; Jack, David B.; Bayly, Christopher I.
    Journal of Computational Chemistry (2000), 21 (2), 132-146CODEN: JCCHDD; ISSN:0192-8651. (John Wiley & Sons, Inc.)
    The AM1-BCC method quickly and efficiently generates high-quality at. charges for use in condensed-phase simulations. The underlying features of the electron distribution including formal charge and delocalization are first captured by AM1 at. charges for the individual mol. Bond charge corrections (BCCs), which have been parameterized against the HF/6-31G* electrostatic potential (ESP) of a training set of compds. contg. relevant functional groups, are then added using a formalism identical to the consensus BCI (bond charge increment) approach. As a proof of the concept, we fit BCCs simultaneously to 45 compds. including O-, N-, and S-contg. functionalities, aroms., and heteroaroms., using only 41 BCC parameters. AM1-BCC yields charge sets of comparable quality to HF/6-31G* ESP-derived charges in a fraction of the time while reducing instabilities in the at. charges compared to direct ESP-fit methods. We then apply the BCC parameters to a small "test set" consisting of aspirin, D-glucose, and eryodictyol; the AM1-BCC model again provides at. charges of quality comparable with HF/6-31G* RESP charges, as judged by an increase of only 0.01 to 0.02 at. units in the root-mean-square (RMS) error in ESP. Based on these encouraging results, we intend to parameterize the AM1-BCC model to provide a consistent charge model for any org. or biol. mol.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3cXkt1SgsQ%253D%253D&md5=aa4774739d3491dbc8dd95ed1948d856
  55. 55
    Jakalian, A.; Jack, D. B.; Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. Parameterization and validation. J. Comput. Chem. 2002, 23 (16), 1623– 1641,  DOI: 10.1002/jcc.10128
    Google Scholar
    55
    Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. parameterization and validation
    Jakalian, Araz; Jack, David B.; Bayly, Christopher I.
    Journal of Computational Chemistry (2002), 23 (16), 1623-1641CODEN: JCCHDD; ISSN:0192-8651. (John Wiley & Sons, Inc.)
    We present the first global parameterization and validation of a novel charge model, called AM1-BCC, which quickly and efficiently generates high-quality at. charges for computer simulations of org. mols. in polar media. The goal of the charge model is to produce at. charges that emulate the HF/6-31G* electrostatic potential (ESP) of a mol. Underlying electronic structure features, including formal charge and electron delocalization, are first captured by AM1 population charges; simple additive bond charge corrections (BCCs) are then applied to these AM1 at. charges to produce the AM1-BCC charges. The parameterization of BCCs was carried out by fitting to the HF/6-31G* ESP of a training set of >2700 mols. Most org. functional groups and their combinations were sampled, as well as an extensive variety of cyclic and fused bicyclic heteroaryl systems. The resulting BCC parameters allow the AM1-BCC charging scheme to handle virtually all types of org. compds. listed in The Merck Index and the NCI Database. Validation of the model was done through comparisons of hydrogen-bonded dimer energies and relative free energies of solvation using AM1-BCC charges in conjunction with the 1994 Cornell et al. forcefield for AMBER. Homo-dimer and hetero-dimer hydrogen-bond energies of a diverse set of org. mols. were reproduced to within 0.95 kcal/mol RMS deviation from the ab initio values, and for DNA dimers the energies were within 0.9 kcal/mol RMS deviation from ab initio values. The calcd. relative free energies of solvation for a diverse set of monofunctional isosteres were reproduced to within 0.69 kcal/mol of expt. In all these validation tests, AMBER with the AM1-BCC charge model maintained a correlation coeff. above 0.96. Thus, the parameters presented here for use with the AM1-BCC method present a fast, accurate, and robust alternative to HF/6-31G* ESP-fit charges for general use with the AMBER force field in computer simulations involving org. small mols.
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD38XosF2rt74%253D&md5=3f2b738617bb17b2898f7ac4d751d7ec
  56. 56
    Zheng, Z.; Wang, T.; Li, P.; Merz, K. M., Jr. KECSA-Movable Type Implicit Solvation Model (KMTISM). J. Chem. Theory Comput. 2015, 11 (2), 667– 682,  DOI: 10.1021/ct5007828
    Google Scholar
    56
    KECSA-Movable Type Implicit Solvation Model (KMTISM)
    Zheng, Zheng; Wang, Ting; Li, Pengfei; Merz, Kenneth M.
    Journal of Chemical Theory and Computation (2015), 11 (2), 667-682CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
    Computation of the solvation free energy for chem. and biol. processes has long been of significant interest. The key challenges to effective solvation modeling center on the choice of potential function and configurational sampling. Herein, an energy sampling approach termed the "Movable Type" (MT) method, and a statistical energy function for solvation modeling, "Knowledge-based and Empirical Combined Scoring Algorithm" (KECSA) are developed and utilized to create an implicit solvation model: KECSA-Movable Type Implicit Solvation Model (KMTISM) suitable for the study of chem. and biol. systems. KMTISM is an implicit solvation model, but the MT method performs energy sampling at the atom pairwise level. For a specific mol. system, the MT method collects energies from prebuilt databases for the requisite atom pairs at all relevant distance ranges, which by its very construction encodes all possible mol. configurations simultaneously. Unlike traditional statistical energy functions, KECSA converts structural statistical information into categorized atom pairwise interaction energies as a function of the radial distance instead of a mean force energy function. Within the implicit solvent model approxn., aq. solvation free energies are then obtained from the NVT ensemble partition function generated by the MT method. Validation is performed against several subsets selected from the Minnesota Solvation Database v2012. Results are compared with several solvation free energy calcn. methods, including a one-to-one comparison against two commonly used classical implicit solvation models: MM-GBSA and MM-PBSA. Comparison against a quantum mechanics based polarizable continuum model is also discussed (Cramer and Truhlar's Solvation Model 12).
    >> More from SciFinder ®
    https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXitFehsLfJ&md5=debe378166aeaac91249dea799add7c5

Cited By

ARTICLE SECTIONS
Jump To

This article is cited by 3 publications.

  1. Xiang Wang, Yejun Deng, Yang Zhang, Caihong Zhang, Lujie Liu, Yong Liu, Jianxin Jiang, Pujun Xie, Lixin Huang. Screening and Evaluation of Novel α-Glucosidase Inhibitory Peptides from Ginkgo biloba Seed Cake Based on Molecular Docking Combined with Molecular Dynamics Simulation. Journal of Agricultural and Food Chemistry 2023, 71 (27) , 10326-10337. https://doi.org/10.1021/acs.jafc.3c00826
  2. Ze-jun Jia, Xiao-Wei Lan, Kui Lu, Xuan Meng, Wen-Jie Jing, Shi-Ru Jia, Kai Zhao, Yu-Jie Dai. Synthesis, molecular docking, and binding Gibbs free energy calculation of β-nitrostyrene derivatives: Potential inhibitors of SARS-CoV-2 3CL protease. Journal of Molecular Structure 2023, 18 , 135409. https://doi.org/10.1016/j.molstruc.2023.135409
  3. Xiao Liu, Lei Zheng, Chu Qin, Yalong Cong, John Z. H. Zhang, Zhaoxi Sun. Comprehensive Evaluation of End-Point Free Energy Techniques in Carboxylated-Pillar[6]arene Host–Guest Binding: III. Force-Field Comparison, Three-Trajectory Realization and Further Dielectric Augmentation. Molecules 2023, 28 (6) , 2767. https://doi.org/10.3390/molecules28062767
Download PDF

  • Abstract

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 1

    Figure 1. Illustration of the free energy estimation protocols applied to (A) the CASF-2016 benchmark and (B) the Merck benchmark.

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 2

    Figure 2. Experimental binding free energies versus predicted binding free energies for the 285 protein–ligand complexes from the CASF-2016 test benchmark. Standard deviations for the test cases were shown as error bars.

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 3

    Figure 3. Correlations between predicted binding free energies using cMD-MT and CHMC-cMD-MT protocols against the CASF-2016 benchmark. The correlation on the left illustrates the predicted binding free energies using cMD-MT-amber (x-axis) vs CHMC-cMD-MT-amber (y-axis). The correlation on the right illustrates the predicted binding free energies using cMD-MT-garf (x-axis) vs CHMC-cMD-MT-garf (y-axis).

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 4

    Figure 4. Pearson’s R coefficients comparisons for simulation-based and docking-based protocols against 52 subsets from the CASF-2016 benchmark. Pearson’s R values for each protocol were ranked individually in their own descending orders.

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 5

    Figure 5. Correlation between predicted binding free energies and experimental values for all eight target sets using the 6 different free energy protocols (For simplicity, the results for the Eg5 target set using the alternative loop are not included.).

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 6

    Figure 6. Illustration of the protein–ligand binding modes for the Eg5 – CHEMBL1084431 complex. (A) CHEMBL1084431 molecule at the Eg5 binding site using a representative binding mode selected from the converged cMD trajectory. (B) CHEMBL1084431 molecule at the Eg5 binding site using the best scored docking pose from the MOE IFD protocol. The area on the protein surface covered with yellow color shows the location of the residues having close contact (<4 Å) with the ligand. (C) Important Eg5 residues interacting with CHEMBL1084431 from the cMD representative binding mode. (C) Important Eg5 residues interacting with CHEMBL1084431 from the MOE IFD protocol.

    High Resolution Image
    Download MS PowerPoint Slide

    Figure 7

    Figure 7. Illustration of protein–ligand binding modes for the PFKFB3–ligand_41 complex. (A) Representative binding mode selected from the converged cMD trajectory. During the MD simulation, Val213 approached the quinoxaline group of the ligand to form a C–H−π interaction, while Val187 moved away from the ligand. (B) During the MD simulation, no hydrogen bonding interaction between SER122 and the ligand is observed, while the hydrogen bonding interaction between ASN133 and the ligand was unstable. (C) The MOE IFD protocol placed the ligand at the binding site with the quinoxaline group close to the Val187 side chain, while Val213 was far from the ligand. (D) The best scored docking pose recognized two hydrogen bonding interactions between the ligand and the SER122 and ASN133 side chains.

    High Resolution Image
    Download MS PowerPoint Slide

  • References

    ARTICLE SECTIONS
    Jump To

    This article references 56 other publications.

    1. 1
      Schneider, G. Virtual screening: an endless staircase?. Nat. Rev. Drug Discovery 2010, 9 (4), 273– 276,  DOI: 10.1038/nrd3139
      1
      Virtual screening: an endless staircase?
      Schneider, Gisbert
      Nature Reviews Drug Discovery (2010), 9 (4), 273-276CODEN: NRDDAG; ISSN:1474-1776. (Nature Publishing Group)
      A review. Computational chem. - in particular, virtual screening - can provide valuable contributions in hit- and lead-compd. discovery. Numerous software tools have been developed for this purpose. However, despite the applicability of virtual screening technol. being well established, it seems that there are relatively few examples of drug discovery projects in which virtual screening has been the key contributor. Has virtual screening reached its peak. If not, what aspects are limiting its potential at present, and how can significant progress be made in the future.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXktVGjur4%253D&md5=ecb5a4aa2683d4a7d688fad3114bc31d
    2. 2
      Sliwoski, G.; Kothiwale, S.; Meiler, J.; Lowe Jr, E. W. Computational methods in drug discovery. Pharmacol. Rev. 2014, 66 (1), 334– 395,  DOI: 10.1124/pr.112.007336
      2
      Computational methods in drug discovery
      Sliwoski, Gregory; Kothiwale, Sandeepkumar; Meiler, Jens; Lowe Edward, W.
      Pharmacological Reviews (2014), 66 (1), 334-395, 62 pp.CODEN: PAREAQ; ISSN:1521-0081. (American Society for Pharmacology and Experimental Therapeutics)
      A review. Computer-aided drug discovery/design methods have played a major role in the development of therapeutically important small mols. for over three decades. These methods are broadly classified as either structure-based or ligand-based methods. Structure-based methods are in principle analogous to high-throughput screening in that both target and ligand structure information is imperative. Structure-based approaches include ligand docking, pharmacophore, and ligand design methods. The article discusses theory behind the most important methods and recent successful applications. Ligand-based methods use only ligand information for predicting activity depending on its similarity/dissimilarity to previously known active ligands. We review widely used ligand-based methods such as ligand-based pharmacophores, mol. descriptors, and quant. structure-activity relationships. In addn., important tools such as target/ligand data bases, homol. modeling, ligand fingerprint methods, etc., necessary for successful implementation of various computer-aided drug discovery/design methods in a drug discovery campaign are discussed. Finally, computational methods for toxicity prediction and optimization for favorable physiol. properties are discussed with successful examples from literature.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXhsVGnu7nL&md5=3dde38a0b60c583f832c688c2d27819f
    3. 3
      Dror, O.; Schneidman-Duhovny, D.; Inbar, Y.; Nussinov, R.; Wolfson, H. J. Novel approach for efficient pharmacophore-based virtual screening: method and applications. J. Chem. Inf. Model 2009, 49 (10), 2333– 2343,  DOI: 10.1021/ci900263d
      3
      Novel Approach for Efficient Pharmacophore-Based Virtual Screening: Method and Applications
      Dror, Oranit; Schneidman-Duhovny, Dina; Inbar, Yuval; Nussinov, Ruth; Wolfson, Haim J.
      Journal of Chemical Information and Modeling (2009), 49 (10), 2333-2343CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      Virtual screening is emerging as a productive and cost-effective technol. in rational drug design for the identification of novel lead compds. An important model for virtual screening is the pharmacophore. Pharmacophore is the spatial configuration of essential features that enable a ligand mol. to interact with a specific target receptor. In the absence of a known receptor structure, a pharmacophore can be identified from a set of ligands that have been obsd. to interact with the target receptor. Here, we present a novel computational method for pharmacophore detection and virtual screening. The pharmacophore detection module is able to (i) align multiple flexible ligands in a deterministic manner without exhaustive enumeration of the conformational space, (ii) detect subsets of input ligands that may bind to different binding sites or have different binding modes, (iii) address cases where the input ligands have different affinities by defining weighted pharmacophores based on the no. of ligands that share them, and (iv) automatically select the most appropriate pharmacophore candidates for virtual screening. The algorithm is highly efficient, allowing a fast exploration of the chem. space by virtual screening of huge compd. databases. The performance of PharmaGist was successfully evaluated on a commonly used data set of G-Protein Coupled Receptor alpha1A. Addnl., a large-scale evaluation using the DUD (directory of useful decoys) data set was performed. DUD contains 2950 active ligands for 40 different receptors, with 36 decoy compds. for each active ligand. PharmaGist enrichment rates are comparable with other state-of-the-art tools for virtual screening.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD1MXht1WntrnE&md5=745357210296e3124bafa49b1f46f0de
    4. 4
      Malmstrom, R. D.; Watowich, S. J. Using free energy of binding calculations to improve the accuracy of virtual screening predictions. J. Chem. Inf. Model 2011, 51 (7), 1648– 1655,  DOI: 10.1021/ci200126v
      4
      Using Free Energy of Binding Calculations To Improve the Accuracy of Virtual Screening Predictions
      Malmstrom, Robert D.; Watowich, Stanley J.
      Journal of Chemical Information and Modeling (2011), 51 (7), 1648-1655CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      Virtual screening of small mol. databases against macromol. targets was used to identify binding ligands and predict their lowest energy bound conformation (i.e., pose). AutoDock4-generated poses were rescored using mean-field pathway decoupling free energy of binding calcns. and evaluated if these calcns. improved virtual screening discrimination between bound and nonbound ligands. Two small mol. databases were used to evaluate the effectiveness of the rescoring algorithm in correctly identifying binders of L99A T4 lysozyme. Self-dock calcns. of a database contg. compds. with known binding free energies and cocrystal structures largely reproduced exptl. measurements, although the mean difference between calcd. and exptl. binding free energies increased as the predicted bound poses diverged from the exptl. poses. In addn., free energy rescoring was more accurate than AutoDock4 scores in discriminating between known binders and nonbinders, suggesting free energy rescoring could be a useful approach to reduce false pos. predictions in virtual screening expts.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3MXos1Whtb0%253D&md5=8a5837b9e25e603960a8690c4a9834dd
    5. 5
      Klebe, G. Applying thermodynamic profiling in lead finding and optimization. Nat. Rev. Drug Discovery 2015, 14 (2), 95– 110,  DOI: 10.1038/nrd4486
      5
      Applying thermodynamic profiling in lead finding and optimization
      Klebe, Gerhard
      Nature Reviews Drug Discovery (2015), 14 (2), 95-110CODEN: NRDDAG; ISSN:1474-1776. (Nature Publishing Group)
      Small-mol. drug discovery involves the optimization of various physicochem. properties of a ligand, particularly its binding affinity for its target receptor (or receptors). In recent years, there has been growing interest in using thermodn. profiling of ligand-receptor interactions in order to select and optimize those ligands that might be most likely to become drug candidates with desirable physicochem. properties. The thermodn. of binding is influenced by multiple factors, including hydrogen bonding and hydrophobic interactions, desolvation, residual mobility, dynamics and the local water structure. This article discusses key issues in understanding the effects of these factors and applying this knowledge in drug discovery.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXhsVemtr4%253D&md5=c2e12760623de81335ce252cb804c602
    6. 6
      Albanese, S. K.; Chodera, J. D.; Volkamer, A.; Keng, S.; Abel, R.; Wang, L. Is Structure-Based Drug Design Ready for Selectivity Optimization?. J. Chem. Inf. Model 2020, 60 (12), 6211– 6227,  DOI: 10.1021/acs.jcim.0c00815
      6
      Is Structure-Based Drug Design Ready for Selectivity Optimization?
      Albanese, Steven K.; Chodera, John D.; Volkamer, Andrea; Keng, Simon; Abel, Robert; Wang, Lingle
      Journal of Chemical Information and Modeling (2020), 60 (12), 6211-6227CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      Alchem. free-energy calcns. are now widely used to drive or maintain potency in small-mol. lead optimization with a roughly 1 kcal/mol accuracy. Despite this, the potential to use free-energy calcns. to drive optimization of compd. selectivity among two similar targets has been relatively unexplored in published studies. In the most optimistic scenario, the similarity of binding sites might lead to a fortuitous cancellation of errors and allow selectivity to be predicted more accurately than affinity. Here, we assess the accuracy with which selectivity can be predicted in the context of small-mol. kinase inhibitors, considering the very similar binding sites of human kinases CDK2 and CDK9 as well as another series of ligands attempting to achieve selectivity between the more distantly related kinases CDK2 and ERK2. Using a Bayesian anal. approach, we sep. systematic from statistical errors and quantify the correlation in systematic errors between selectivity targets. We find that, in the CDK2/CDK9 case, a high correlation in systematic errors suggests that free-energy calcns. can have significant impact in aiding chemists in achieving selectivity, while in more distantly related kinases (CDK2/ERK2), the correlation in systematic error suggests that fortuitous cancellation may even occur between systems that are not as closely related. In both cases, the correlation in systematic error suggests that longer simulations are beneficial to properly balance statistical error with systematic error to take full advantage of the increase in apparent free-energy calcn. accuracy in selectivity prediction.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXitFygs7bJ&md5=08cfdf031b626801b7ba762a559d750f
    7. 7
      Raha, K.; Peters, M. B.; Wang, B.; Yu, N.; Wollacott, A. M.; Westerhoff, L. M.; Merz Jr, K. M. The role of quantum mechanics in structure-based drug design. Drug Discovery Today 2007, 12 (17–18), 725– 731,  DOI: 10.1016/j.drudis.2007.07.006
      7
      The role of quantum mechanics in structure-based drug design
      Raha, Kaushik; Peters, Martin B.; Wang, Bing; Yu, Ning; Wollacott, Andrew M.; Westerhoff, Lance M.; Merz, Kenneth M., Jr.
      Drug Discovery Today (2007), 12 (17&18), 725-731CODEN: DDTOFS; ISSN:1359-6446. (Elsevier B.V.)
      A review. Herein we will focus on the use of quantum mechanics (QM) in drug design (DD) to solve disparate problems from scoring protein-ligand poses to building QM QSAR models. Through the variational principle of QM we know that we can obtain a more accurate representation of mol. systems than classical models, and while this is not a matter of debate, it still has not been shown that the expense of QM approaches is offset by improved accuracy in DD applications. Objectively validating the improved applicability and performance of QM over classical-based models in DD will be the focus of research in the coming years along with research on the conformational sampling problem as it relates to protein-ligand complexes.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXhtVWhsrvK&md5=67fecc13843c7961ff820026c85a9290
    8. 8
      Brooijmans, N.; Kuntz, I. D. Molecular recognition and docking algorithms. Annu. Rev. Biophys. Biomol. Struct. 2003, 32, 335– 373,  DOI: 10.1146/annurev.biophys.32.110601.142532
      8
      Molecular recognition and docking algorithms
      Brooijmans, Natasja; Kuntz, Irwin D.
      Annual Review of Biophysics and Biomolecular Structure (2003), 32 (), 335-373CODEN: ABBSE4; ISSN:1056-8700. (Annual Reviews Inc.)
      A review. Mol. docking is an invaluable tool in modern drug discovery. This review focuses on methodol. developments relevant to the field of mol. docking. The forces important in mol. recognition are reviewed and followed by a discussion of how different scoring functions account for these forces. More recent applications of computational chem. tools involve library design and database screening. Last, we summarize several crit. methodol. issues that must be addressed in future developments.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXntFSgtrg%253D&md5=1db89884abfef58d8080e726cf40960c
    9. 9
      Kuntz, I. D.; Blaney, J. M.; Oatley, S. J.; Langridge, R.; Ferrin, T. E. A geometric approach to macromolecule-ligand interactions. J. Mol. Biol. 1982, 161 (2), 269– 288,  DOI: 10.1016/0022-2836(82)90153-X
      9
      A geometric approach to macromolecule-ligand interactions
      Kuntz, Irwin D.; Blaney, Jeffrey M.; Oatley, Stuart J.; Langridge, Robert; Ferrin, Thomas E.
      Journal of Molecular Biology (1982), 161 (2), 269-88CODEN: JMOBAK; ISSN:0022-2836.
      A method is described to explore geometrically feasible alignments of ligands and receptors of known structure. Algorithms are presented that examine many binding geometries and evaluate them in terms of steric overlap. The procedure uses specific mol. conformations. A method is included for finding putative binding sites on a macromol. surface. Results are reported for heme-myoglobin interaction and the binding of thyroid hormone analogs to prealbumin. In each case, the program finds structures within 1 Å of the x-ray results and also finds distinctly different geometries that provide good steric fits. The approach seems well-suited for generating conformations for energy refinement programs and interactive computer graphics routines.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaL38XmtFajsbw%253D&md5=8a4234b24356ea5340f33d906cd71d3e
    10. 10
      Kitchen, D. B.; Decornez, H.; Furr, J. R.; Bajorath, J. Docking and scoring in virtual screening for drug discovery: methods and applications. Nat. Rev. Drug Discovery 2004, 3 (11), 935– 949,  DOI: 10.1038/nrd1549
      10
      Docking and scoring in virtual screening for drug discovery: methods and applications
      Kitchen, Douglas B.; Decornez, Helene; Furr, John R.; Bajorath, Juergen
      Nature Reviews Drug Discovery (2004), 3 (11), 935-949CODEN: NRDDAG; ISSN:1474-1776. (Nature Publishing Group)
      A review. Computational approaches that 'dock' small mols. into the structures of macromol. targets and 'score' their potential complementarity to binding sites are widely used in hit identification and lead optimization. Indeed, there are now a no. of drugs whose development was heavily influenced by or based on structure-based design and screening strategies, such as HIV protease inhibitors. Nevertheless, there remain significant challenges in the application of these approaches, in particular in relation to current scoring schemes. Here, we review key concepts and specific features of small-mol.-protein docking methods, highlight selected applications and discuss recent advances that aim to address the acknowledged limitations of established approaches.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXptFemtrg%253D&md5=875a2b37a4299181509a1922b11dbd2f
    11. 11
      Sledz, P.; Caflisch, A. Protein structure-based drug design: from docking to molecular dynamics. Curr. Opin. Struct. Biol. 2018, 48, 93– 102,  DOI: 10.1016/j.sbi.2017.10.010
      11
      Protein structure-based drug design: from docking to molecular dynamics
      Sledz, Pawel; Caflisch, Amedeo
      Current Opinion in Structural Biology (2018), 48 (), 93-102CODEN: COSBEF; ISSN:0959-440X. (Elsevier Ltd.)
      Recent years have witnessed rapid developments of computer-aided drug design methods, which have reached accuracy that allows their routine practical applications in drug discovery campaigns. Protein structure-based methods are useful for the prediction of binding modes of small mols. and their relative affinity. The high-throughput docking of up to 106 small mols. followed by scoring based on implicit-solvent force field can robustly identify micromolar binders using a rigid protein target. Mol. dynamics with explicit solvent is a low-throughput technique for the characterization of flexible binding sites and accurate evaluation of binding pathways, kinetics, and thermodn. In this review we highlight recent advancements in applications of ligand docking tools and mol. dynamics simulations to ligand identification and optimization.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXhslygsbzF&md5=3a3b92c6dbd9c5075bef0bb7268cc5dc
    12. 12
      De Vivo, M.; Masetti, M.; Bottegoni, G.; Cavalli, A. Role of Molecular Dynamics and Related Methods in Drug Discovery. J. Med. Chem. 2016, 59 (9), 4035– 4061,  DOI: 10.1021/acs.jmedchem.5b01684
      12
      Role of Molecular Dynamics and Related Methods in Drug Discovery
      De Vivo, Marco; Masetti, Matteo; Bottegoni, Giovanni; Cavalli, Andrea
      Journal of Medicinal Chemistry (2016), 59 (9), 4035-4061CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      Mol. dynamics (MD) and related methods are close to becoming routine computational tools for drug discovery. Their main advantage is in explicitly treating structural flexibility and entropic effects. This allows a more accurate est. of the thermodn. and kinetics assocd. with drug-target recognition and binding, as better algorithms and hardware architectures increase their use. Here, we review the theor. background of MD and enhanced sampling methods, focusing on free-energy perturbation, metadynamics, steered MD, and other methods most consistently used to study drug-target binding. We discuss unbiased MD simulations that nowadays allow the observation of unsupervised ligand-target binding, assessing how these approaches help optimizing target affinity and drug residence time toward improved drug efficacy. Further issues discussed include allosteric modulation and the role of water mols. in ligand binding and optimization. We conclude by calling for more prospective studies to attest to these methods' utility in discovering novel drug candidates.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhtlWksLs%253D&md5=b3d8a4fb5705a3555c2bb2a962420396
    13. 13
      Cheng, T.; Li, X.; Li, Y.; Liu, Z.; Wang, R. Comparative assessment of scoring functions on a diverse test set. J. Chem. Inf. Model 2009, 49 (4), 1079– 1093,  DOI: 10.1021/ci9000053
      13
      Comparative Assessment of Scoring Functions on a Diverse Test Set
      Cheng, Tiejun; Li, Xun; Li, Yan; Liu, Zhihai; Wang, Renxiao
      Journal of Chemical Information and Modeling (2009), 49 (4), 1079-1093CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      Scoring functions are widely applied to the evaluation of protein-ligand binding in structure-based drug design. We have conducted a comparative assessment of 16 popular scoring functions implemented in main-stream com. software or released by academic research groups. A set of 195 diverse protein-ligand complexes with high-resoln. crystal structures and reliable binding consts. were selected through a systematic non-redundant sampling of the PDBbind database and used as the primary test set in our study. All scoring functions were evaluated in three aspects, i.e., "docking power", "ranking power", and "scoring power", and all evaluations were independent from the context of mol. docking or virtual screening. As for "docking power", six scoring functions, including GOLD::ASP, DS::PLP1, DrugScorePDB, GlideScore-SP, DS::LigScore, and GOLD::ChemScore, achieved success rates over 70% when the acceptance cutoff was root-mean-square deviation < 2.0 Å. Combining these scoring functions into consensus scoring schemes improved the success rates to 80% or even higher. As for "ranking power" and "scoring power", the top four scoring functions on the primary test set were X-Score, DrugScoreCSD, DS::PLP, and SYBYL::ChemScore. They were able to correctly rank the protein-ligand complexes contg. the same type of protein with success rates around 50%. Correlation coeffs. between the exptl. binding consts. and the binding scores computed by these scoring functions ranged from 0.545 to 0.644. Besides the primary test set, each scoring function was also tested on four addnl. test sets, each consisting of a certain no. of protein-ligand complexes contg. one particular type of protein. Our study serves as an updated benchmark for evaluating the general performance of today's scoring functions. Our results indicate that no single scoring function consistently outperforms others in all three aspects. Thus, it is important in practice to choose the appropriate scoring functions for different purposes.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD1MXkt1aqtLg%253D&md5=9f073faa564c30a69aa26485d4dcc7db
    14. 14
      Li, Y.; Han, L.; Liu, Z.; Wang, R. Comparative assessment of scoring functions on an updated benchmark: 2. Evaluation methods and general results. J. Chem. Inf. Model 2014, 54 (6), 1717– 1736,  DOI: 10.1021/ci500081m
      14
      Comparative Assessment of Scoring Functions on an Updated Benchmark: 2. Evaluation Methods and General Results
      Li, Yan; Han, Li; Liu, Zhihai; Wang, Renxiao
      Journal of Chemical Information and Modeling (2014), 54 (6), 1717-1736CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      Our comparative assessment of scoring functions (CASF) benchmark is created to provide an objective evaluation of current scoring functions. The key idea of CASF is to compare the general performance of scoring functions on a diverse set of protein-ligand complexes. In order to avoid testing scoring functions in the context of mol. docking, the scoring process is sepd. from the docking (or sampling) process by using ensembles of ligand binding poses that are generated in prior. Here, we describe the tech. methods and evaluation results of the latest CASF-2013 study. The PDBbind core set (version 2013) was employed as the primary test set in this study, which consists of 195 protein-ligand complexes with high-quality three-dimensional structures and reliable binding consts. A panel of 20 scoring functions, most of which are implemented in main-stream com. software, were evaluated in terms of "scoring power" (binding affinity prediction), "ranking power" (relative ranking prediction), "docking power" (binding pose prediction), and "screening power" (discrimination of true binders from random mols.). Our results reveal that the performance of these scoring functions is generally more promising in the docking/screening power tests than in the scoring/ranking power tests. Top-ranked scoring functions in the scoring power test, such as X-ScoreHM, ChemScore@SYBYL, ChemPLP@GOLD, and PLP@DS, are also top-ranked in the ranking power test. Top-ranked scoring functions in the docking power test, such as ChemPLP@GOLD, Chemscore@GOLD, GlidScore-SP, LigScore@DS, and PLP@DS, are also top-ranked in the screening power test. Our results obtained on the entire test set and its subsets suggest that the real challenge in protein-ligand binding affinity prediction lies in polar interactions and assocd. desolvation effect. Nonadditive features obsd. among high-affinity protein-ligand complexes also need attention.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXls1CrsL0%253D&md5=0d386b64f5c557667533847d694436bc
    15. 15
      Su, M.; Yang, Q.; Du, Y.; Feng, G.; Liu, Z.; Li, Y.; Wang, R. Comparative Assessment of Scoring Functions: The CASF-2016 Update. J. Chem. Inf. Model 2019, 59 (2), 895– 913,  DOI: 10.1021/acs.jcim.8b00545
      15
      Comparative Assessment of Scoring Functions: The CASF-2016 Update
      Su, Minyi; Yang, Qifan; Du, Yu; Feng, Guoqin; Liu, Zhihai; Li, Yan; Wang, Renxiao
      Journal of Chemical Information and Modeling (2019), 59 (2), 895-913CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      In structure-based drug design, scoring functions are often employed to evaluate protein-ligand interactions. A variety of scoring functions have been developed so far, and thus, some objective benchmarks are desired for assessing their strength and weakness. The comparative assessment of scoring functions (CASF) benchmark developed by us provides an answer to this demand. CASF is designed as a "scoring benchmark", where the scoring process is decoupled from the docking process to depict the performance of scoring function more precisely. Here, we describe the latest update of this benchmark, i.e., CASF-2016. Each scoring function is still evaluated by four metrics, including "scoring power", "ranking power", "docking power", and "screening power". Nevertheless, the evaluation methods have been improved considerably in several aspects. A new test set is compiled, which consists of 285 protein-ligand complexes with high-quality crystal structures and reliable binding consts. A panel of 25 scoring functions are tested on CASF-2016 as a demonstration. Our results reveal that the performance of current scoring functions is more promising in terms of docking power than scoring, ranking, and screening power. Scoring power is somewhat correlated with ranking power, so are docking power and screening power. The results obtained on CASF-2016 may provide valuable guidance for the end users to make smart choices among available scoring functions. Moreover, CASF is created as an open-access benchmark so that other researchers can utilize it to test a wider range of scoring functions. The complete CASF-2016 benchmark will be released on the PDBbind-CN web server (http://www.pdbbind-cn.org/casf.asp/) once this article is published.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXitlWhtLjM&md5=573295cfd2298a2dee933027041c3d9c
    16. 16
      Forli, S.; Huey, R.; Pique, M. E.; Sanner, M. F.; Goodsell, D. S.; Olson, A. J. Computational protein-ligand docking and virtual drug screening with the AutoDock suite. Nat. Protoc. 2016, 11 (5), 905– 919,  DOI: 10.1038/nprot.2016.051
      16
      Computational protein-ligand docking and virtual drug screening with the AutoDock suite
      Forli, Stefano; Huey, Ruth; Pique, Michael E.; Sanner, Michel F.; Goodsell, David S.; Olson, Arthur J.
      Nature Protocols (2016), 11 (5), 905-919CODEN: NPARDW; ISSN:1750-2799. (Nature Publishing Group)
      Computational docking can be used to predict bound conformations and free energies of binding for small-mol. ligands to macromol. targets. Docking is widely used for the study of biomol. interactions and mechanisms, and it is applied to structure-based drug design. The methods are fast enough to allow virtual screening of ligand libraries contg. tens of thousands of compds. This protocol covers the docking and virtual screening methods provided by the AutoDock suite of programs, including a basic docking of a drug mol. with an anticancer target, a virtual screen of this target with a small ligand library, docking with selective receptor flexibility, active site prediction and docking with explicit hydration. The entire protocol will require ∼5 h.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28Xlsl2gsLw%253D&md5=7dafc5e3acab3d38e9b33fbdaf65422d
    17. 17
      Trott, O.; Olson, A. J. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem. 2009, 31 (2), 455– 461,  DOI: 10.1002/jcc.21334
      There is no corresponding record for this reference.
    18. 18
      Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.; Taylor, R. D. Improved protein-ligand docking using GOLD. Proteins 2003, 52 (4), 609– 623,  DOI: 10.1002/prot.10465
      18
      Improved protein-ligand docking using GOLD
      Verdonk, Marcel L.; Cole, Jason C.; Hartshorn, Michael J.; Murray, Christopher W.; Taylor, Richard D.
      Proteins: Structure, Function, and Genetics (2003), 52 (4), 609-623CODEN: PSFGEY; ISSN:0887-3585. (Wiley-Liss, Inc.)
      The Chemscore function was implemented as a scoring function for the protein-ligand docking program GOLD, and its performance compared to the original Goldscore function and two consensus docking protocols, "Goldscore-CS" and "Chemscore-GS," in terms of docking accuracy, prediction of binding affinities, and speed. In the "Goldscore-CS" protocol, dockings produced with the Goldscore function are scored and ranked with the Chemscore function; in the "Chemscore-GS" protocol, dockings produced with the Chemscore function are scored and ranked with the Goldscore function. Comparisons were made for a "clean" set of 224 protein-ligand complexes, and for two subsets of this set, one for which the ligands are "drug-like," the other for which they are "fragment-like.". For "drug-like" and "fragment-like" ligands, the docking accuracies obtained with Chemscore and Goldscore functions are similar. For larger ligands, Goldscore gives superior results. Docking with the Chemscore function is up to three times faster than docking with the Goldscore function. Both combined docking protocols give significant improvements in docking accuracy over the use of the Goldscore or Chemscore function alone. "Goldscore-CS" gives success rates of up to 81% (top-ranked GOLD soln. within 2.0 Å of the exptl. binding mode) for the "clean list," but at the cost of long search times. For most virtual screening applications, "Chemscore-GS" seems optimal; search settings that give docking speeds of around 0.25-1.3 min/compd. have success rates of about 78% for "drug-like" compds. and 85% for "fragment-like" compds. In terms of producing binding energy ests., the Goldscore function appears to perform better than the Chemscore function and the two consensus protocols, particularly for faster search settings. Even at docking speeds of around 1-2 min/compd., the Goldscore function predicts binding energies with a std. deviation of ∼10.5 kJ/mol.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXmvFGrsLg%253D&md5=73a627d99dab6de1ae364c8e8e5f0fea
    19. 19
      Friesner, R. A.; Murphy, R. B.; Repasky, M. P.; Frye, L. L.; Greenwood, J. R.; Halgren, T. A.; Sanschagrin, P. C.; Mainz, D. T. Extra precision glide: docking and scoring incorporating a model of hydrophobic enclosure for protein-ligand complexes. J. Med. Chem. 2006, 49 (21), 6177– 6196,  DOI: 10.1021/jm051256o
      19
      Extra Precision Glide: Docking and Scoring Incorporating a Model of Hydrophobic Enclosure for Protein-Ligand Complexes
      Friesner, Richard A.; Murphy, Robert B.; Repasky, Matthew P.; Frye, Leah L.; Greenwood, Jeremy R.; Halgren, Thomas A.; Sanschagrin, Paul C.; Mainz, Daniel T.
      Journal of Medicinal Chemistry (2006), 49 (21), 6177-6196CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      A novel scoring function to est. protein-ligand binding affinities has been developed and implemented as the Glide 4.0 XP scoring function and docking protocol. In addn. to unique water desolvation energy terms, protein-ligand structural motifs leading to enhanced binding affinity are included:(1) hydrophobic enclosure where groups of lipophilic ligand atoms are enclosed on opposite faces by lipophilic protein atoms, (2) neutral-neutral single or correlated hydrogen bonds in a hydrophobically enclosed environment, and (3) five categories of charged-charged hydrogen bonds. The XP scoring function and docking protocol have been developed to reproduce exptl. binding affinities for a set of 198 complexes (RMSDs of 2.26 and 1.73 kcal/mol over all and well-docked ligands, resp.) and to yield quality enrichments for a set of fifteen screens of pharmaceutical importance. Enrichment results demonstrate the importance of the novel XP mol. recognition and water scoring in sepg. active and inactive ligands and avoiding false positives.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD28XpvVGmurg%253D&md5=ea428c82ead0d8c27f8c1a7b694a1edf
    20. 20
      Jain, A. N. Surflex: fully automatic flexible molecular docking using a molecular similarity-based search engine. J. Med. Chem. 2003, 46 (4), 499– 511,  DOI: 10.1021/jm020406h
      20
      Surflex: fully automatic flexible molecular docking using a molecular similarity-based search engine
      Jain, Ajay N.
      Journal of Medicinal Chemistry (2003), 46 (4), 499-511CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      Surflex is a fully automatic flexible mol. docking algorithm that combines the scoring function from the Hammerhead docking system with a search engine that relies on a surface-based mol. similarity method as a means to rapidly generate suitable putative poses for mol. fragments. Results are presented evaluating reliability and accuracy of dockings compared with crystallog. exptl. results on 81 protein/ligand pairs of substantial structural diversity. In over 80% of the complexes, Surflex's highest scoring docked pose was within 2.5 Å root-mean-square deviation (rmsd), with over 90% of the complexes having one of the top ranked poses within 2.5 Å rmsd. Results are also presented assessing Surflex's utility as a screening tool on two protein targets (thymidine kinase and estrogen receptor) using data sets on which competing methods were run. Performance of Surflex was significantly better, with true pos. rates of greater than 80% at false pos. rates of less than 1%. Docking time was roughly linear in no. of rotatable bonds, beginning with a few seconds for rigid mols. and adding approx. 10 s per rotatable bond.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3sXkvFKjsA%253D%253D&md5=67a08aafeadf69101d56b2f60a922359
    21. 21
      Clark, A. J.; Tiwary, P.; Borrelli, K.; Feng, S.; Miller, E. B.; Abel, R.; Friesner, R. A.; Berne, B. J. Prediction of Protein-Ligand Binding Poses via a Combination of Induced Fit Docking and Metadynamics Simulations. J. Chem. Theory Comput. 2016, 12 (6), 2990– 2998,  DOI: 10.1021/acs.jctc.6b00201
      21
      Prediction of Protein-Ligand Binding Poses via a Combination of Induced Fit Docking and Metadynamics Simulations
      Clark, Anthony J.; Tiwary, Pratyush; Borrelli, Ken; Feng, Shulu; Miller, Edward B.; Abel, Robert; Friesner, Richard A.; Berne, B. J.
      Journal of Chemical Theory and Computation (2016), 12 (6), 2990-2998CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      Ligand docking is a widely used tool for lead discovery and binding mode prediction based drug discovery. The greatest challenges in docking occur when the receptor significantly reorganizes upon small mol. binding, thereby requiring an induced fit docking (IFD) approach in which the receptor is allowed to move in order to bind to the ligand optimally. IFD methods have had some success but suffer from a lack of reliability. Complementing IFD with all-atom mol. dynamics (MD) is a straightforward soln. in principle but not in practice due to the severe time scale limitations of MD. Here we introduce a metadynamics plus IFD strategy for accurate and reliable prediction of the structures of protein-ligand complexes at a practically useful computational cost. Our strategy allows treating this problem in full atomistic detail and in a computationally efficient manner and enhances the predictive power of IFD methods. We significantly increase the accuracy of the underlying IFD protocol across a large data set comprising 42 different ligand-receptor systems. We expect this approach to be of significant value in computationally driven drug design.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28Xnt12ktrw%253D&md5=d5a1211c0b5398ba78bab9aec4bddbeb
    22. 22
      Merz, K. M., Jr. Limits of Free Energy Computation for Protein-Ligand Interactions. J. Chem. Theory Comput. 2010, 6, 1769– 1776,  DOI: 10.1021/ct100102q
      22
      Limits of Free Energy Computation for Protein-Ligand Interactions
      Merz, Kenneth M., Jr.
      Journal of Chemical Theory and Computation (2010), 6 (5), 1769-1776CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      A detailed error anal. is presented for the computation of protein-ligand interaction energies. In particular, the authors show that it is probable that even highly accurate computed binding free energies have errors that represent a large percentage of the target free energies of binding. This is due to the observation that the error for computed energies quasi-linearly increases with the increasing no. of interactions present in a protein-ligand complex. This principle is expected to hold true for any system that involves an ever increasing no. of inter- or intramol. interactions (e.g., ab initio protein folding). The authors introduce the concept of best-case scenario errors (BCSerrors) that can be routinely applied to docking and scoring studies and that can be used to provide error bars for the computed binding free energies. These BCSerrors form a basis by which one can evaluate the outcome of a docking and scoring exercise. Moreover, the resultant error anal. enables the formation of an hypothesis that defines the best direction to proceed to improve scoring functions used in mol. docking studies.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXkslKksr4%253D&md5=e6ba0bd00675e195be2d565f2879fd1c
    23. 23
      Li, J.; Fu, A.; Zhang, L. An Overview of Scoring Functions Used for Protein-Ligand Interactions in Molecular Docking. Interdiscip. Sci. 2019, 11 (2), 320– 328,  DOI: 10.1007/s12539-019-00327-w
      23
      An Overview of Scoring Functions Used for Protein-Ligand Interactions in Molecular Docking
      Li, Jin; Fu, Ailing; Zhang, Le
      Interdisciplinary Sciences: Computational Life Sciences (2019), 11 (2), 320-328CODEN: ISCLAL; ISSN:1867-1462. (Springer GmbH)
      A review. Currently, mol. docking is becoming a key tool in drug discovery and mol. modeling applications. The reliability of mol. docking depends on the accuracy of the adopted scoring function, which can guide and det. the ligand poses when thousands of possible poses of ligand are generated. The scoring function can be used to det. the binding mode and site of a ligand, predict binding affinity and identify the potential drug leads for a given protein target. Despite intensive research over the years, accurate and rapid prediction of protein-ligand interactions is still a challenge in mol. docking. For this reason, this study reviews four basic types of scoring functions, physics-based, empirical, knowledge-based, and machine learning-based scoring functions, based on an up-to-date classification scheme. We not only discuss the foundations of the four types scoring functions, suitable application areas and shortcomings, but also discuss challenges and potential future study directions.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhtVKgur7O&md5=32d69a1e5260384d3d44d8d1e846c715
    24. 24
      Fischer, A.; Smiesko, M.; Sellner, M.; Lill, M. A. Decision Making in Structure-Based Drug Discovery: Visual Inspection of Docking Results. J. Med. Chem. 2021, 64 (5), 2489– 2500,  DOI: 10.1021/acs.jmedchem.0c02227
      24
      Decision making in structure-based drug discovery: visual inspection of docking results
      Fischer, Andre; Smiesko, Martin; Sellner, Manuel; Lill, Markus A.
      Journal of Medicinal Chemistry (2021), 64 (5), 2489-2500CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      A review. Mol. docking is a computational method widely used in drug discovery. Due to the inherent inaccuracies of mol. docking, visual inspection of binding modes is a crucial routine in the decision making process of computational medicinal chemists. Despite its apparent importance for medicinal chem. projects, guidelines for the visual docking pose assessment have been hardly discussed in the literature. Here, the authors review the medicinal chem. literature with the aim of identifying consistent principles for visual inspection, highlighting cases of its successful application, and discussing its limitations. In this context, the authors conducted a survey reaching experts in both academia and the pharmaceutical industry, which also included a challenge to distinguish native from incorrect poses. The authors were able to collect 93 expert opinions that offer valuable insights into visually supported decision-making processes. This perspective shall motivate discussions among experienced computational medicinal chemists and guide young scientists new to the field to stratify their compds.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3MXks1yjtrY%253D&md5=f5566c527c0889cdc8a6c2e4cc2c4881
    25. 25
      Shao, J.; Tanner, S. W.; Thompson, N.; Cheatham, T. E. Clustering Molecular Dynamics Trajectories: 1. Characterizing the Performance of Different Clustering Algorithms. J. Chem. Theory Comput. 2007, 3 (6), 2312– 2334,  DOI: 10.1021/ct700119m
      25
      Clustering Molecular Dynamics Trajectories: 1. Characterizing the Performance of Different Clustering Algorithms
      Shao, Jianyin; Tanner, Stephen W.; Thompson, Nephi; Cheatham, Thomas E., III
      Journal of Chemical Theory and Computation (2007), 3 (6), 2312-2334CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      Mol. dynamics simulation methods produce trajectories of at. positions (and optionally velocities and energies) as a function of time and provide a representation of the sampling of a given mol.'s energetically accessible conformational ensemble. As simulations on the 10-100 ns time scale become routine, with sampled configurations stored on the picosecond time scale, such trajectories contain large amts. of data. Data-mining techniques, like clustering, provide one means to group and make sense of the information in the trajectory. In this work, several clustering algorithms were implemented, compared, and utilized to understand MD trajectory data. The development of the algorithms into a freely available C code library, and their application to a simple test example of random (or systematically placed) points in a 2D plane (where the pairwise metric is the distance between points) provide a means to understand the relative performance. Eleven different clustering algorithms were developed, ranging from top-down splitting (hierarchical) and bottom-up aggregating (including single-linkage edge joining, centroid-linkage, av.-linkage, complete-linkage, centripetal, and centripetal-complete) to various refinement (means, Bayesian, and self-organizing maps) and tree (COBWEB) algorithms. Systematic testing in the context of MD simulation of various DNA systems (including DNA single strands and the interaction of a minor groove binding drug DB226 with a DNA hairpin) allows a more direct assessment of the relative merits of the distinct clustering algorithms. Addnl., means to assess the relative performance and differences between the algorithms, to dynamically select the initial cluster count, and to achieve faster data mining by "sieved clustering" were evaluated. Overall, it was found that there is no one perfect "one size fits all" algorithm for clustering MD trajectories and that the results strongly depend on the choice of atoms for the pairwise comparison. Some algorithms tend to produce homogeneously sized clusters, whereas others have a tendency to produce singleton clusters. Issues related to the choice of a pairwise metric, clustering metrics, which atom selection is used for the comparison, and about the relative performance are discussed. Overall, the best performance was obsd. with the av.-linkage, means, and SOM algorithms. If the cluster count is not known in advance, the hierarchical or av.-linkage clustering algorithms are recommended. Although these algorithms perform well, it is important to be aware of the limitations or weaknesses of each algorithm, specifically the high sensitivity to outliers with hierarchical, the tendency to generate homogenously sized clusters with means, and the tendency to produce small or singleton clusters with av.-linkage.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXhtFaqsLzE&md5=668cff11c29736325a4788379cc43def
    26. 26
      Wang, L.; Wu, Y.; Deng, Y.; Kim, B.; Pierce, L.; Krilov, G.; Lupyan, D.; Robinson, S.; Dahlgren, M. K.; Greenwood, J.; Romero, D. L.; Masse, C.; Knight, J. L.; Steinbrecher, T.; Beuming, T.; Damm, W.; Harder, E.; Sherman, W.; Brewer, M.; Wester, R.; Murcko, M.; Frye, L.; Farid, R.; Lin, T.; Mobley, D. L.; Jorgensen, W. L.; Berne, B. J.; Friesner, R. A.; Abel, R. Accurate and reliable prediction of relative ligand binding potency in prospective drug discovery by way of a modern free-energy calculation protocol and force field. J. Am. Chem. Soc. 2015, 137 (7), 2695– 2703,  DOI: 10.1021/ja512751q
      26
      Accurate and Reliable Prediction of Relative Ligand Binding Potency in Prospective Drug Discovery by Way of a Modern Free-Energy Calculation Protocol and Force Field
      Wang, Lingle; Wu, Yujie; Deng, Yuqing; Kim, Byungchan; Pierce, Levi; Krilov, Goran; Lupyan, Dmitry; Robinson, Shaughnessy; Dahlgren, Markus K.; Greenwood, Jeremy; Romero, Donna L.; Masse, Craig; Knight, Jennifer L.; Steinbrecher, Thomas; Beuming, Thijs; Damm, Wolfgang; Harder, Ed; Sherman, Woody; Brewer, Mark; Wester, Ron; Murcko, Mark; Frye, Leah; Farid, Ramy; Lin, Teng; Mobley, David L.; Jorgensen, William L.; Berne, Bruce J.; Friesner, Richard A.; Abel, Robert
      Journal of the American Chemical Society (2015), 137 (7), 2695-2703CODEN: JACSAT; ISSN:0002-7863. (American Chemical Society)
      Designing tight-binding ligands is a primary objective of small-mol. drug discovery. Over the past few decades, free-energy calcns. have benefited from improved force fields and sampling algorithms, as well as the advent of low-cost parallel computing. However, it has proven to be challenging to reliably achieve the level of accuracy that would be needed to guide lead optimization (∼5× in binding affinity) for a wide range of ligands and protein targets. Not surprisingly, widespread com. application of free-energy simulations has been limited due to the lack of large-scale validation coupled with the tech. challenges traditionally assocd. with running these types of calcns. Here, we report an approach that achieves an unprecedented level of accuracy across a broad range of target classes and ligands, with retrospective results encompassing 200 ligands and a wide variety of chem. perturbations, many of which involve significant changes in ligand chem. structures. In addn., we have applied the method in prospective drug discovery projects and found a significant improvement in the quality of the compds. synthesized that have been predicted to be potent. Compds. predicted to be potent by this approach have a substantial redn. in false positives relative to compds. synthesized on the basis of other computational or medicinal chem. approaches. Furthermore, the results are consistent with those obtained from our retrospective studies, demonstrating the robustness and broad range of applicability of this approach, which can be used to drive decisions in lead optimization.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXhsF2iuro%253D&md5=37a4f4a6c085f47ed531342643b6c33b
    27. 27
      Lee, T. S.; Allen, B. K.; Giese, T. J.; Guo, Z.; Li, P.; Lin, C.; McGee, T. D., Jr.; Pearlman, D. A.; Radak, B. K.; Tao, Y.; Tsai, H. C.; Xu, H.; Sherman, W.; York, D. M. Alchemical Binding Free Energy Calculations in AMBER20: Advances and Best Practices for Drug Discovery. J. Chem. Inf. Model. 2020, 60 (11), 5595– 5623,  DOI: 10.1021/acs.jcim.0c00613
      27
      Alchemical Binding Free Energy Calculations in AMBER20: Advances and Best Practices for Drug Discovery
      Lee, Tai-Sung; Allen, Bryce K.; Giese, Timothy J.; Guo, Zhenyu; Li, Pengfei; Lin, Charles; McGee Jr., T. Dwight; Pearlman, David A.; Radak, Brian K.; Tao, Yujun; Tsai, Hsu-Chun; Xu, Huafeng; Sherman, Woody; York, Darrin M.
      Journal of Chemical Information and Modeling (2020), 60 (11), 5595-5623CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      A review. Predicting protein-ligand binding affinities and the assocd. thermodn. of biomol. recognition is a primary objective of structure-based drug design. Alchem. free energy simulations offer a highly accurate and computationally efficient route to achieving this goal. While the AMBER mol. dynamics package has successfully been used for alchem. free energy simulations in academic research groups for decades, widespread impact in industrial drug discovery settings has been minimal because of the previous limitations within the AMBER alchem. code, coupled with challenges in system setup and postprocessing workflows. Through a close academia-industry collaboration we have addressed many of the previous limitations with an aim to improve accuracy, efficiency, and robustness of alchem. binding free energy simulations in industrial drug discovery applications. Here, we highlight some of the recent advances in AMBER20 with a focus on alchem. binding free energy (BFE) calcns., which are less computationally intensive than alternative binding free energy methods where full binding/unbinding paths are explored. In addn. to scientific and tech. advances in AMBER20, we also describe the essential practical aspects assocd. with running relative alchem. BFE calcns., along with recommendations for best practices, highlighting the importance not only of the alchem. simulation code but also the auxiliary functionalities and expertise required to obtain accurate and reliable results. This work is intended to provide a contemporary overview of the scientific, tech., and practical issues assocd. with running relative BFE simulations in AMBER20, with a focus on real-world drug discovery applications.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhvVahsL3L&md5=6c7385019de25c306df8c40eaf485cfc
    28. 28
      Zheng, Z.; Ucisik, M. N.; Merz Jr, K. M. The Movable Type Method Applied to Protein-Ligand Binding. J. Chem. Theory Comput. 2013, 9 (12), 5526– 5538,  DOI: 10.1021/ct4005992
      28
      The Movable Type Method Applied to Protein-Ligand Binding
      Zheng, Zheng; Ucisik, Melek N.; Merz, Kenneth M.
      Journal of Chemical Theory and Computation (2013), 9 (12), 5526-5538CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      Accurately computing the free energy for biol. processes like protein folding or protein-ligand assocn. remains a challenging problem. Both describing the complex intermol. forces involved and sampling the requisite configuration space make understanding these processes innately difficult. Herein, we address the sampling problem using a novel methodol. we term "movable type" (MT). Conceptually it can be understood by analogy with the evolution of printing and, hence, the name movable type. For example, a common approach to the study of protein-ligand complexation involves taking a database of intact drug-like mols. and exhaustively docking them into a binding pocket. This is reminiscent of early woodblock printing where each page had to be laboriously created prior to printing a book. However, printing evolved to an approach where a database of symbols (letters, numerals, etc.) was created and then assembled using a MT system, which allowed for the creation of all possible combinations of symbols on a given page, thereby, revolutionizing the dissemination of knowledge. Our MT method involves identifying all of the atom pairs seen in protein-ligand complexes and then creating two databases: one with their assocd. pairwise distant dependent energies and another assocd. with the probability of how these pairs can combine in terms of bonds, angles, dihedrals, and nonbonded interactions. Combining these two databases coupled with the principles of statistical mechanics allows us to accurately est. binding free energies as well as the pose of a ligand in a receptor. This method, by its math. construction, samples all of the configuration space of a selected region (the protein active site here) in one shot without resorting to brute force sampling schemes involving Monte Carlo, genetic algorithms, or mol. dynamics simulations making the methodol. extremely efficient. Importantly, this method explores the free energy surface eliminating the need to est. the enthalpy and entropy components individually. Finally, low free energy structures can be obtained via a free energy minimization procedure yielding all low free energy poses on a given free energy surface. Besides revolutionizing the protein-ligand docking and scoring problem, this approach can be utilized in a wide range of applications in computational biol. which involve the computation of free energies for systems with extensive phase spaces including protein folding, protein-protein docking, and protein design.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXhs1ygtr%252FF&md5=11f1ccafda165961647af0a96d17466c
    29. 29
      Stewart, J. J. Optimization of parameters for semiempirical methods V: modification of NDDO approximations and application to 70 elements. J. Mol. Model 2007, 13 (12), 1173– 1213,  DOI: 10.1007/s00894-007-0233-4
      29
      Optimization of parameters for semiempirical methods V: modification of NDDO approximations and application to 70 elements
      Stewart, James J. P.
      Journal of Molecular Modeling (2007), 13 (12), 1173-1213CODEN: JMMOFK; ISSN:0948-5023. (Springer GmbH)
      Several modifications that have been made to the NDDO core-core interaction term and to the method of parameter optimization are described. These changes have resulted in a more complete parameter optimization, called PM6, which has, in turn, allowed 70 elements to be parameterized. The av. unsigned error (AUE) between calcd. and ref. heats of formation for 4,492 species was 8.0 kcal mol-1. For the subset of 1,373 compds. involving only the elements H, C, N, O, F, P, S, Cl, and Br, the PM6 AUE was 4.4 kcal mol-1. The equivalent AUE for other methods were: RM1: 5.0, B3LYP 6-31G*: 5.2, PM5: 5.7, PM3: 6.3, HF 6-31G*: 7.4, and AM1: 10.0 kcal mol-1. Several long-standing faults in AM1 and PM3 have been cor. and significant improvements have been made in the prediction of geometries.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2sXhtlequr7N&md5=7d22928c8e3423f3ca8f2e7c77e6e6c9
    30. 30
      Zheng, Z.; Pei, J.; Bansal, N.; Liu, H.; Song, L. F.; Merz Jr, K. M. Generation of Pairwise Potentials Using Multidimensional Data Mining. J. Chem. Theory Comput. 2018, 14 (10), 5045– 5067,  DOI: 10.1021/acs.jctc.8b00516
      30
      Generation of Pairwise Potentials Using Multidimensional Data Mining
      Zheng, Zheng; Pei, Jun; Bansal, Nupur; Liu, Hao; Song, Lin Frank; Merz, Kenneth M.
      Journal of Chemical Theory and Computation (2018), 14 (10), 5045-5067CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      The rapid development of mol. structural databases provides the chem. community access to an enormous array of exptl. data that can be used to build and validate computational models. Using radial distribution functions collected from exptl. available X-ray and NMR structures, a no. of so-called statistical potentials have been developed over the years using the structural data mining strategy. These potentials have been developed within the context of the two-particle Kirkwood equation by extending its original use for isotropic monat. systems to anisotropic biomol. systems. However, the accuracy and the unclear phys. meaning of statistical potentials have long formed the central arguments against such methods. In this work, we present a new approach to generate mol. energy functions using structural data mining. Instead of employing the Kirkwood equation and introducing the "ref. state" approxn., we model the multidimensional probability distributions of the mol. system using graphical models and generate the target pairwise Boltzmann probabilities using the Bayesian field theory. Different from the current statistical potentials that mimic the "knowledge-based" PMF based on the 2-particle Kirkwood equation, the graphical-model-based structure-derived potential developed in this study focuses on the generation of lower-dimensional Boltzmann distributions of atoms through redn. of dimensionality. We have named this new scoring function GARF, and in this work we focus on the math. derivation of our novel approach followed by validation studies on its ability to predict protein-ligand interactions.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhs1KhtLjJ&md5=388d026039f54229c468c78e3fb68954
    31. 31
      Tian, C.; Kasavajhala, K.; Belfon, K. A. A.; Raguette, L.; Huang, H.; Migues, A. N.; Bickel, J.; Wang, Y.; Pincay, J.; Wu, Q.; Simmerling, C. ff19SB: Amino-Acid-Specific Protein Backbone Parameters Trained against Quantum Mechanics Energy Surfaces in Solution. J. Chem. Theory Comput. 2020, 16 (1), 528– 552,  DOI: 10.1021/acs.jctc.9b00591
      31
      ff19SB: Amino-Acid-Specific Protein Backbone Parameters Trained against Quantum Mechanics Energy Surfaces in Solution
      Tian Chuan; Kasavajhala Koushik; Belfon Kellon A A; Raguette Lauren; Huang He; Bickel John; Wang Yuzhang; Pincay Jorge; Simmerling Carlos; Tian Chuan; Kasavajhala Koushik; Belfon Kellon A A; Raguette Lauren; Huang He; Migues Angela N; Wang Yuzhang; Simmerling Carlos; Wu Qin
      Journal of chemical theory and computation (2020), 16 (1), 528-552 ISSN:.
      Molecular dynamics (MD) simulations have become increasingly popular in studying the motions and functions of biomolecules. The accuracy of the simulation, however, is highly determined by the molecular mechanics (MM) force field (FF), a set of functions with adjustable parameters to compute the potential energies from atomic positions. However, the overall quality of the FF, such as our previously published ff99SB and ff14SB, can be limited by assumptions that were made years ago. In the updated model presented here (ff19SB), we have significantly improved the backbone profiles for all 20 amino acids. We fit coupled φ/ψ parameters using 2D φ/ψ conformational scans for multiple amino acids, using as reference data the entire 2D quantum mechanics (QM) energy surface. We address the polarization inconsistency during dihedral parameter fitting by using both QM and MM in aqueous solution. Finally, we examine possible dependency of the backbone fitting on side chain rotamer. To extensively validate ff19SB parameters, and to compare to results using other Amber models, we have performed a total of ∼5 ms MD simulations in explicit solvent. Our results show that after amino-acid-specific training against QM data with solvent polarization, ff19SB not only reproduces the differences in amino-acid-specific Protein Data Bank (PDB) Ramachandran maps better but also shows significantly improved capability to differentiate amino-acid-dependent properties such as helical propensities. We also conclude that an inherent underestimation of helicity is present in ff14SB, which is (inexactly) compensated for by an increase in helical content driven by the TIP3P bias toward overly compact structures. In summary, ff19SB, when combined with a more accurate water model such as OPC, should have better predictive power for modeling sequence-specific behavior, protein mutations, and also rational protein design. Of the explicit water models tested here, we recommend use of OPC with ff19SB.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A280%3ADC%252BB3MjnvFGisw%253D%253D&md5=7cbaecffea06e4bf7ccecb7f1d0a0f4d
    32. 32
      Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D. A. Development and testing of a general amber force field. J. Comput. Chem. 2004, 25 (9), 1157– 1174,  DOI: 10.1002/jcc.20035
      32
      Development and testing of a general Amber force field
      Wang, Junmei; Wolf, Romain M.; Caldwell, James W.; Kollman, Peter A.; Case, David A.
      Journal of Computational Chemistry (2004), 25 (9), 1157-1174CODEN: JCCHDD; ISSN:0192-8651. (John Wiley & Sons, Inc.)
      We describe here a general Amber force field (GAFF) for org. mols. GAFF is designed to be compatible with existing Amber force fields for proteins and nucleic acids, and has parameters for most org. and pharmaceutical mols. that are composed of H, C, N, O, S, P, and halogens. It uses a simple functional form and a limited no. of atom types, but incorporates both empirical and heuristic models to est. force consts. and partial at. charges. The performance of GAFF in test cases is encouraging. In test I, 74 crystallog. structures were compared to GAFF minimized structures, with a root-mean-square displacement of 0.26 Å, which is comparable to that of the Tripos 5.2 force field (0.25 Å) and better than those of MMFF 94 and CHARMm (0.47 and 0.44 Å, resp.). In test II, gas phase minimizations were performed on 22 nucleic acid base pairs, and the minimized structures and intermol. energies were compared to MP2/6-31G* results. The RMS of displacements and relative energies were 0.25 Å and 1.2 kcal/mol, resp. These data are comparable to results from Parm99/RESP (0.16 Å and 1.18 kcal/mol, resp.), which were parameterized to these base pairs. Test III looked at the relative energies of 71 conformational pairs that were used in development of the Parm99 force field. The RMS error in relative energies (compared to expt.) is about 0.5 kcal/mol. GAFF can be applied to wide range of mols. in an automatic fashion, making it suitable for rational drug design and database searching.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD2cXksFakurc%253D&md5=2992017a8cf51f89290ae2562403b115
    33. 33
      Zheng, Z.; Borbulevych, O. Y.; Liu, H.; Deng, J.; Martin, R. I.; Westerhoff, L. M. MovableType Software for Fast Free Energy-Based Virtual Screening: Protocol Development, Deployment, Validation, and Assessment. J. Chem. Inf. Model 2020, 60 (11), 5437– 5456,  DOI: 10.1021/acs.jcim.0c00618
      33
      MovableType Software for Fast Free Energy-Based Virtual Screening: Protocol Development, Deployment, Validation, and Assessment
      Zheng, Zheng; Borbulevych, Oleg Y.; Liu, Hao; Deng, Jianpeng; Martin, Roger I.; Westerhoff, Lance M.
      Journal of Chemical Information and Modeling (2020), 60 (11), 5437-5456CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      For decades, the complicated energy surfaces found in macromol. protein:ligand structures, which require large amts. of computational time and resources for energy state sampling, have been an inherent obstacle to fast, routine free energy estn. in industrial drug discovery efforts. Beginning in 2013, the Merz research group addressed this cost with the introduction of a novel sampling methodol. termed "Movable Type" (MT). Using numerical integration methods, the MT method reduces the computational expense for energy state sampling by independently calcg. each at. partition function from an initial mol. conformation in order to est. the mol. free energy using ensembles of the at. partition functions. In this work, we report a software package, the DivCon Discovery Suite with the MovableType module from QuantumBio Inc., that performs this MT free energy estn. protocol in a fast, fully encapsulated manner. We discuss the computational procedures and improvements to the original work, and we detail the corresponding settings for this software package. Finally, we introduce two validation benchmarks to evaluate the overall robustness of the method against a broad range of protein:ligand structural cases. With these publicly available benchmarks, we show that the method can use a variety of input types and parameters and exhibits comparable predictability whether the method is presented with "expensive" X-ray structures or "inexpensively docked" theor. models. We also explore some next steps for the method. The MovableType software is available at http://www.quantumbioinc.com/.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhsF2nsb3L&md5=9f520f96b7938e8364bb1207ec39a417
    34. 34
      Pan, L. L.; Zheng, Z.; Wang, T.; Merz Jr, K. M. Free Energy-Based Conformational Search Algorithm Using the Movable Type Sampling Method. J. Chem. Theory Comput. 2015, 11 (12), 5853– 5864,  DOI: 10.1021/acs.jctc.5b00930
      34
      Free Energy-Based Conformational Search Algorithm Using the Movable Type Sampling Method
      Pan, Li-Li; Zheng, Zheng; Wang, Ting; Merz, Kenneth M.
      Journal of Chemical Theory and Computation (2015), 11 (12), 5853-5864CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      In this article, we extend the movable type (MT) sampling method to mol. conformational searches (MT-CS) on the free energy surface of the mol. in question. Differing from traditional systematic and stochastic searching algorithms, this method uses Boltzmann energy information to facilitate the selection of the best conformations. The generated ensembles provided good coverage of the available conformational space including available crystal structures. Furthermore, our approach directly provides the solvation free energies and the relative gas and aq. phase free energies for all generated conformers. The method is validated by a thorough anal. of thrombin ligands as well as against structures extd. from both the Protein Data Bank (PDB) and the Cambridge Structural Database (CSD). An in-depth comparison between OMEGA and MT-CS is presented to illustrate the differences between the two conformational searching strategies, i.e., energy-based vs. free energy-based searching. These studies demonstrate that our MT-based ligand conformational search algorithm is a powerful approach to delineate the conformational ensembles of mol. species on free energy surfaces.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXhvVSls7rO&md5=8a23f666b879ede91a547f1c596268fd
    35. 35
      Bansal, N.; Zheng, Z.; Merz Jr, K. M. Incorporation of side chain flexibility into protein binding pockets using MTflex. Bioorg. Med. Chem. 2016, 24 (20), 4978– 4987,  DOI: 10.1016/j.bmc.2016.08.030
      35
      Incorporation of side chain flexibility into protein binding pockets using MTflex
      Bansal, Nupur; Zheng, Zheng; Merz, Kenneth M., Jr.
      Bioorganic & Medicinal Chemistry (2016), 24 (20), 4978-4987CODEN: BMECEP; ISSN:0968-0896. (Elsevier B.V.)
      The plasticity of active sites plays a significant role in drug recognition and binding, but the accurate incorporation of receptor flexibility remains a significant computational challenge. Many approaches have been put forward to address receptor flexibility in docking studies by generating relevant ensembles on the energy surface. Herein, the authors describe the Movable Type with flexibility (MTflex) method that generates ensembles on the more relevant free energy surface in a computationally tractable manner. This novel approach enumerates conformational states based on side chain flexibility and then ests. their relative free energies using the MT methodol. The resultant conformational states can then be used in subsequent docking and scoring exercises. In particular, using the MTflex ensembles improves MT's ability to predict binding free energies over docking only to the crystal structure.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhsVOqu77E&md5=4769f5beae2495901d63ac12a493c80f
    36. 36
      Molecular Operating Environment (MOE); 2019.0102 Chemical Computing Group ULC: Montreal, QC, Canada, 2020.
      There is no corresponding record for this reference.
    37. 37
      Corbeil, C. R.; Williams, C. I.; Labute, P. Variability in docking success rates due to dataset preparation. J. Comput. Aided Mol. Des. 2012, 26 (6), 775– 786,  DOI: 10.1007/s10822-012-9570-1
      37
      Variability in docking success rates due to dataset preparation
      Corbeil, Christopher R.; Williams, Christopher I.; Labute, Paul
      Journal of Computer-Aided Molecular Design (2012), 26 (6), 775-786CODEN: JCADEQ; ISSN:0920-654X. (Springer)
      The results of cognate docking with the prepd. Astex dataset provided by the organizers of the "Docking and Scoring: A Review of Docking Programs" session at the 241st ACS national meeting are presented. The MOE software with the newly developed GBVI/WSA dG scoring function is used throughout the study. For 80 % of the Astex targets, the MOE docker produces a top-scoring pose within 2 Å of the X-ray structure. For 91 % of the targets a pose within 2 Å of the X-ray structure is produced in the top 30 poses. Docking failures, defined as cases where the top scoring pose is greater than 2 Å from the exptl. structure, are shown to be largely due to the absence of bound waters in the source dataset, highlighting the need to include these and other crucial information in future standardized sets. Docking success is shown to depend heavily on data prepn. A "dataset prepn." error of 0.5 kcal/mol is shown to cause fluctuations of over 20 % in docking success rates.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC38XhtVelu7nM&md5=8018caa912ccccc46349706526fb0485
    38. 38
      Case, D. A.; Ben-Shalom, I. Y.; Brozell, S. R.; Cerutti, D. S.; Cheatham, T. E.; Cruzeiro, V. W. D., III; Darden, T. A.; Duke, R. E.; Ghoreishi, D.; Gilson, M. K.; Gohlke, H.; Goetz, A. W.; Greene, D.; Harris, R.; Homeyer, N.; Huang, Y.; Izadi, S.; Kovalenko, A.; Kurtzman, T.; Lee, T. S.; LeGrand, S.; Li, P.; Lin, C.; Liu, J.; Luchko, T.; Luo, R.; Mermelstein, D. J.; Merz, K. M.; Miao, Y.; Monard, G.; Nguyen, C.; Nguyen, H.; Omelyan, I.; Onufriev, A.; Pan, F.; Qi, R.; Roe, D. R.; Roitberg, A.; Sagui, C.; Schott-Verdugo, S.; Shen, J.; Simmerling, C. L.; Smith, J.; Salomon-Ferrer, R.; Swails, J.; Walker, R. C.; Wang, J.; Wei, H.; Wolf, R. M.; Wu, X.; Xiao, L.; York, D. M.; Kollman, P. A. AMBER 2018; University of California: San Francisco, 2018.
      There is no corresponding record for this reference.
    39. 39
      Hornak, V.; Abel, R.; Okur, A.; Strockbine, B.; Roitberg, A.; Simmerling, C. Comparison of multiple Amber force fields and development of improved protein backbone parameters. Proteins 2006, 65 (3), 712– 725,  DOI: 10.1002/prot.21123
      39
      Comparison of multiple Amber force fields and development of improved protein backbone parameters
      Hornak, Viktor; Abel, Robert; Okur, Asim; Strockbine, Bentley; Roitberg, Adrian; Simmerling, Carlos
      Proteins: Structure, Function, and Bioinformatics (2006), 65 (3), 712-725CODEN: PSFBAF ISSN:. (Wiley-Liss, Inc.)
      The ff94 force field that is commonly assocd. with the Amber simulation package is one of the most widely used parameter sets for biomol. simulation. After a decade of extensive use and testing, limitations in this force field, such as over-stabilization of α-helixes, were reported by the authors and other researchers. This led to a no. of attempts to improve these parameters, resulting in a variety of "Amber" force fields and significant difficulty in detg. which should be used for a particular application. The authors show that several of these continue to suffer from inadequate balance between different secondary structure elements. In addn., the approach used in most of these studies neglected to account for the existence in Amber of two sets of backbone .vphi./ψ dihedral terms. This led to parameter sets that provide unreasonable conformational preferences for glycine. The authors report here an effort to improve the .vphi./ψ dihedral terms in the ff99 energy function. Dihedral term parameters are based on fitting the energies of multiple conformations of glycine and alanine tetrapeptides from high level ab initio quantum mech. calcns. The new parameters for backbone dihedrals replace those in the existing ff99 force field. This parameter set, which the authors denote ff99SB, achieves a better balance of secondary structure elements as judged by improved distribution of backbone dihedrals for glycine and alanine with respect to PDB survey data. It also accomplishes improved agreement with published exptl. data for conformational preferences of short alanine peptides and better accord with exptl. NMR relaxation data of test protein systems.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD28XhtFWqt7fM&md5=de683a26eca9e83ae524726e97ac22fa
    40. 40
      Li, P.; Roberts, B. P.; Chakravorty, D. K.; Merz Jr, K. M. Rational Design of Particle Mesh Ewald Compatible Lennard-Jones Parameters for + 2 Metal Cations in Explicit Solvent. J. Chem. Theory Comput. 2013, 9 (6), 2733– 2748,  DOI: 10.1021/ct400146w
      40
      Rational Design of Particle Mesh Ewald Compatible Lennard-Jones Parameters for +2 Metal Cations in Explicit Solvent
      Li, Pengfei; Roberts, Benjamin P.; Chakravorty, Dhruva K.; Merz, Kenneth M.
      Journal of Chemical Theory and Computation (2013), 9 (6), 2733-2748CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      Metal ions play significant roles in biol. systems. Accurate mol. dynamics (MD) simulations on these systems require a validated set of parameters. Although there are more detailed ways to model metal ions, the nonbonded model, which employs a 12-6 Lennard-Jones (LJ) term plus an electrostatic potential, is still widely used in MD simulations today due to its simple form. However, LJ parameters have limited transferability due to different combining rules, various water models, and diverse simulation methods. Recently, simulations employing a Particle Mesh Ewald (PME) treatment for long-range electrostatics have become more and more popular owing to their speed and accuracy. In the present work, we have systematically designed LJ parameters for 24 +2 metal (M(II)) cations to reproduce different exptl. properties appropriate for the Lorentz-Berthelot combining rules and PME simulations. We began by testing the transferability of currently available M(II) ion LJ parameters. The results showed that there are differences between simulations employing Ewald summation with other simulation methods and that it was necessary to design new parameters specific for PME based simulations. Employing the thermodn. integration (TI) method and performing periodic boundary MD simulations employing PME, allowed for a systematic investigation of the LJ parameter space. Hydration free energies (HFEs), the ion-oxygen distance in the first solvation shell (IOD), and coordination nos. (CNs) were obtained for various combinations of the parameters of the LJ potential for four widely used water models (TIP3P, SPC/E, TIP4P, and TIP4PEW). Results showed that the three simulated properties were highly correlated. Meanwhile, M(II) ions with the same parameters in different water models produce remarkably different HFEs but similar structural properties. It is difficult to reproduce various exptl. values simultaneously because the nonbonded model underestimates the interaction between the metal ions and water mols. at short-range. Moreover, the extent of underestimation increases successively for the TIP3P, SPC/E, TIP4PEW, and TIP4P water models. Nonetheless, we fitted a curve to describe the relationship between ε (the well depth) and radius (Rmin/2) from exptl. data on noble gases to facilitate the generation of the best possible compromise models. Hence, by targeting different exptl. values, we developed three sets of parameters for M(II) cations for three different water models (TIP3P, SPC/E, and TIP4PEW). These parameters we feel represent the best possible compromise that can be achieved using the nonbonded model for the ions in combination with simple water models. From a computational uncertainty anal. we est. that the uncertainty in our computed HFEs is on the order of ±1 kcal/mol. Further improvements will require more advanced nonbonded models likely with inclusion of polarization.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3sXntlGjsb0%253D&md5=e87f8df0191db081bcc7b2252806b61c
    41. 41
      Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103 (19), 8577– 8593,  DOI: 10.1063/1.470117
      41
      A smooth particle mesh Ewald method
      Essmann, Ulrich; Perera, Lalith; Berkowitz, Max L.; Darden, Tom; Lee, Hsing; Pedersen, Lee G.
      Journal of Chemical Physics (1995), 103 (19), 8577-93CODEN: JCPSA6; ISSN:0021-9606. (American Institute of Physics)
      The previously developed particle mesh Ewald method is reformulated in terms of efficient B-spline interpolation of the structure factors. This reformulation allows a natural extension of the method to potentials of the form 1/rp with p ≥ 1. Furthermore, efficient calcn. of the virial tensor follows. Use of B-splines in the place of Lagrange interpolation leads to analytic gradients as well as a significant improvement in the accuracy. The authors demonstrate that arbitrary accuracy can be achieved, independent of system size N, at a cost that scales as N log(N). For biomol. systems with many thousands of atoms and this method permits the use of Ewald summation at a computational cost comparable to that of a simple truncation method of 10 Å or less.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaK2MXptlehtrw%253D&md5=092a679dd3bee08da28df41e302383a7
    42. 42
      Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 1977, 23 (3), 327– 341,  DOI: 10.1016/0021-9991(77)90098-5
      42
      Numerical integration of the Cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes
      Ryckaert, Jean Paul; Ciccotti, Giovanni; Berendsen, Herman J. C.
      Journal of Computational Physics (1977), 23 (3), 327-41CODEN: JCTPAH; ISSN:0021-9991.
      A numerical algorithm integrating the 3N Cartesian equation of motion of a system of N points subject to holonomic constraints is applied to mol. dynamics simulation of a liq. of 64 butane mols.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADyaE2sXktVGhsL4%253D&md5=b4aecddfde149117813a5ea4f5353ce2
    43. 43
      Liu, H.; Deng, J.; Luo, Z.; Lin, Y.; Merz Jr, K. M.; Zheng, Z. Receptor-Ligand Binding Free Energies from a Consecutive Histograms Monte Carlo Sampling Method. J. Chem. Theory Comput. 2020, 16 (10), 6645– 6655,  DOI: 10.1021/acs.jctc.0c00457
      43
      Receptor-Ligand Binding Free Energies from a Consecutive Histograms Monte Carlo Sampling Method
      Liu, Hao; Deng, Jianpeng; Luo, Zhou; Lin, Yawei; Merz Jr., Kenneth M.; Zheng, Zheng
      Journal of Chemical Theory and Computation (2020), 16 (10), 6645-6655CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      To obtain accurate and converged free energy calcns. for ligand binding to biomol. systems requires validated force fields and extensive sampling of the energy landscape, which requires exhaustive and effective conformational searching methods. Herein, we introduce the consecutive histograms Monte Carlo (CHMC) sampling protocol that generates receptor-ligand binding modes within a series of continuously distributed sampling units ranging from placement near the geometric center of the receptor's binding site to fully unbound states. This protocol employs independent energy-state sampling for calcg. the ensemble energy within every predefined location along the receptor-ligand dissocn. pathway, without the need to traverse the energy barriers as in mol. dynamic simulations during the dissocn. procedure. We applied this method to a set of selected receptor targets with their corresponding ligands providing detailed studies of mol. binding free energy predictions. The results show that the CHMC gives an excellent accounting of the free energy surfaces and binding free energies at a reasonable computational cost.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhs12rt7%252FI&md5=2071782691a1892e807eb69d13c34c94
    44. 44
      Schindler, C. E. M.; Baumann, H.; Blum, A.; Bose, D.; Buchstaller, H. P.; Burgdorf, L.; Cappel, D.; Chekler, E.; Czodrowski, P.; Dorsch, D.; Eguida, M. K. I.; Follows, B.; Fuchss, T.; Gradler, U.; Gunera, J.; Johnson, T.; Jorand Lebrun, C.; Karra, S.; Klein, M.; Knehans, T.; Koetzner, L.; Krier, M.; Leiendecker, M.; Leuthner, B.; Li, L.; Mochalkin, I.; Musil, D.; Neagu, C.; Rippmann, F.; Schiemann, K.; Schulz, R.; Steinbrecher, T.; Tanzer, E. M.; Unzue Lopez, A.; Viacava Follis, A.; Wegener, A.; Kuhn, D. Large-Scale Assessment of Binding Free Energy Calculations in Active Drug Discovery Projects. J. Chem. Inf. Model 2020, 60 (11), 5457– 5474,  DOI: 10.1021/acs.jcim.0c00900
      44
      Large-Scale Assessment of Binding Free Energy Calculations in Active Drug Discovery Projects
      Schindler, Christina E. M.; Baumann, Hannah; Blum, Andreas; Boese, Dietrich; Buchstaller, Hans-Peter; Burgdorf, Lars; Cappel, Daniel; Chekler, Eugene; Czodrowski, Paul; Dorsch, Dieter; Eguida, Merveille K. I.; Follows, Bruce; Fuchss, Thomas; Graedler, Ulrich; Gunera, Jakub; Johnson, Theresa; Jorand Lebrun, Catherine; Karra, Srinivasa; Klein, Markus; Knehans, Tim; Koetzner, Lisa; Krier, Mireille; Leiendecker, Matthias; Leuthner, Birgitta; Li, Liwei; Mochalkin, Igor; Musil, Djordje; Neagu, Constantin; Rippmann, Friedrich; Schiemann, Kai; Schulz, Robert; Steinbrecher, Thomas; Tanzer, Eva-Maria; Unzue Lopez, Andrea; Viacava Follis, Ariele; Wegener, Ansgar; Kuhn, Daniel
      Journal of Chemical Information and Modeling (2020), 60 (11), 5457-5474CODEN: JCISD8; ISSN:1549-9596. (American Chemical Society)
      Accurate ranking of compds. with regards to their binding affinity to a protein using computational methods is of great interest to pharmaceutical research. Physics-based free energy calcns. are regarded as the most rigorous way to est. binding affinity. In recent years, many retrospective studies carried out both in academia and industry have demonstrated its potential. Here, we present the results of large-scale prospective application of the FEP+ method in active drug discovery projects in an industry setting at Merck KGaA, Darmstadt, Germany. We compare these prospective data to results obtained on a new diverse, public benchmark of eight pharmaceutically relevant targets. Our results offer insights into the challenges faced when using free energy calcns. in real-life drug discovery projects and identify limitations that could be tackled by future method development. The new public data set we provide to the community can support further method development and comparative benchmarking of free energy calcns.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BB3cXhs1CktrbE&md5=f2372d56cd501717435ed8095dd0dbb0
    45. 45
      Schiemann, K.; Mallinger, A.; Wienke, D.; Esdar, C.; Poeschke, O.; Busch, M.; Rohdich, F.; Eccles, S. A.; Schneider, R.; Raynaud, F. I.; Czodrowski, P.; Musil, D.; Schwarz, D.; Urbahns, K.; Blagg, J. Discovery of potent and selective CDK8 inhibitors from an HSP90 pharmacophore. Bioorg. Med. Chem. Lett. 2016, 26 (5), 1443– 1451,  DOI: 10.1016/j.bmcl.2016.01.062
      45
      Discovery of potent and selective CDK8 inhibitors from an HSP90 pharmacophore
      Schiemann, Kai; Mallinger, Aurelie; Wienke, Dirk; Esdar, Christina; Poeschke, Oliver; Busch, Michael; Rohdich, Felix; Eccles, Suzanne A.; Schneider, Richard; Raynaud, Florence I.; Czodrowski, Paul; Musil, Djordje; Schwarz, Daniel; Urbahns, Klaus; Blagg, Julian
      Bioorganic & Medicinal Chemistry Letters (2016), 26 (5), 1443-1451CODEN: BMCLE8; ISSN:0960-894X. (Elsevier B.V.)
      Here we describe the discovery and optimization of 3-benzylindazoles as potent and selective inhibitors of CDK8, also modulating CDK19, discovered from a high-throughput screening (HTS) campaign sampling the Merck compd. collection. The primary hits with strong HSP90 affinity were subsequently optimized to potent and selective CDK8 inhibitors which demonstrate inhibition of WNT pathway activity in cell-based assays. X-ray crystallog. data demonstrated that 3-benzylindazoles occupy the ATP binding site of CDK8 and adopt a Type I binding mode. Medicinal chem. optimization successfully led to improved potency, physicochem. properties and oral pharmacokinetics. Modulation of phospho-STAT1, a pharmacodynamic biomarker of CDK8, was demonstrated in an APC-mutant SW620 human colorectal carcinoma xenograft model following oral administration.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XitlCitLs%253D&md5=77e9f7705c9136557195ccfb4982cb8d
    46. 46
      Dorsch, D.; Schadt, O.; Stieber, F.; Meyring, M.; Gradler, U.; Bladt, F.; Friese-Hamim, M.; Knuhl, C.; Pehl, U.; Blaukat, A. Identification and optimization of pyridazinones as potent and selective c-Met kinase inhibitors. Bioorg. Med. Chem. Lett. 2015, 25 (7), 1597– 1602,  DOI: 10.1016/j.bmcl.2015.02.002
      46
      Identification and optimization of pyridazinones as potent and selective c-Met kinase inhibitors
      Dorsch, Dieter; Schadt, Oliver; Stieber, Frank; Meyring, Michael; Graedler, Ulrich; Bladt, Friedhelm; Friese-Hamim, Manja; Knuehl, Christine; Pehl, Ulrich; Blaukat, Andree
      Bioorganic & Medicinal Chemistry Letters (2015), 25 (7), 1597-1602CODEN: BMCLE8; ISSN:0960-894X. (Elsevier B.V.)
      In a high-throughput screening campaign for c-Met kinase inhibitors, a thiadiazinone deriv. with a carbamate group was identified as a potent in vitro inhibitor. Subsequent optimization guided by c-Met-inhibitor X-ray structures furnished new compd. classes with excellent in vitro and in vivo profiles. The thiadiazinone ring of the HTS hit was first replaced by a pyridazinone followed by an exchange of the carbamate hinge binder with a 1,5-disubstituted pyrimidine. Finally an optimized compd., 3-(1-[3-[5-(1-methyl-piperidin-4-yl-methoxy)-pyrimidin-2-yl]-benzyl]-6-oxo-1,6-dihydro-pyridazin-3-yl)-benzonitrile, (MSC2156119), with excellent in vitro potency, high kinase selectivity, long half-life after oral administration and in vivo anti-tumor efficacy at low doses, was selected as a candidate for clin. development.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2MXivVSqsLg%253D&md5=dadaea6c4c8a0bda928def6f356c416d
    47. 47
      Schiemann, K.; Finsinger, D.; Zenke, F.; Amendt, C.; Knochel, T.; Bruge, D.; Buchstaller, H. P.; Emde, U.; Stahle, W.; Anzali, S. The discovery and optimization of hexahydro-2H-pyrano[3,2-c]quinolines (HHPQs) as potent and selective inhibitors of the mitotic kinesin-5. Bioorg. Med. Chem. Lett. 2010, 20 (5), 1491– 1495,  DOI: 10.1016/j.bmcl.2010.01.110
      47
      The discovery and optimization of hexahydro-2H-pyrano[3,2-c]quinolines (HHPQs) as potent and selective inhibitors of the mitotic kinesin-5
      Schiemann, Kai; Finsinger, Dirk; Zenke, Frank; Amendt, Christiane; Knoechel, Thorsten; Bruge, David; Buchstaller, Hans-Peter; Emde, Ulrich; Staehle, Wolfgang; Anzali, Soheila
      Bioorganic & Medicinal Chemistry Letters (2010), 20 (5), 1491-1495CODEN: BMCLE8; ISSN:0960-894X. (Elsevier B.V.)
      Here we describe the discovery and optimization of hexahydro-2H-pyrano[3,2-c]quinolines (HHPQs) as potent and selective inhibitors of the mitotic kinesin-5 originally found during a high-throughput screening (HTS) campaign sampling our inhouse compd. collection. The compds. optimized subsequently and characterized herein were potently inhibiting the ATPase activity of Kinesin-5 and also exhibited consistent cellular activity, in that cells arrested in mitosis and apoptosis induction could be obsd. X-ray crystallog. data demonstrated that these inhibitors bind in an allosteric pocket of Kinesin-5 distant from the nucleotide and microtubule binding sites. The selected clin. candidate EMD 534085 (I) caused strong growth inhibition in human tumor xenograft models using Colo 205 colon carcinoma cells at doses below 30 mg/kg administered twice weekly without showing severe toxicity as detd. by loss of body wt.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC3cXitleiu7s%253D&md5=33138cecfe9a0f094109185ff010c7c6
    48. 48
      Wehn, P. M.; Rizzi, J. P.; Dixon, D. D.; Grina, J. A.; Schlachter, S. T.; Wang, B.; Xu, R.; Yang, H.; Du, X.; Han, G.; Wang, K.; Cao, Z.; Cheng, T.; Czerwinski, R. M.; Goggin, B. S.; Huang, H.; Halfmann, M. M.; Maddie, M. A.; Morton, E. L.; Olive, S. R.; Tan, H.; Xie, S.; Wong, T.; Josey, J. A.; Wallace, E. M. Design and Activity of Specific Hypoxia-Inducible Factor-2alpha (HIF-2alpha) Inhibitors for the Treatment of Clear Cell Renal Cell Carcinoma: Discovery of Clinical Candidate (S)-3-((2,2-Difluoro-1-hydroxy-7-(methylsulfonyl)-2,3-dihydro-1 H-inden-4-yl)oxy)-5-fluorobenzonitrile (PT2385). J. Med. Chem. 2018, 61 (21), 9691– 9721,  DOI: 10.1021/acs.jmedchem.8b01196
      48
      Design and Activity of Specific Hypoxia-Inducible Factor-2α (HIF-2α) Inhibitors for the Treatment of Clear Cell Renal Cell Carcinoma: Discovery of Clinical Candidate (S)-3-((2,2-Difluoro-1-hydroxy-7-(methylsulfonyl)-2,3-dihydro-1H-inden-4-yl)oxy)-5-fluorobenzonitrile (PT2385)
      Wehn, Paul M.; Rizzi, James P.; Dixon, Darryl D.; Grina, Jonas A.; Schlachter, Stephen T.; Wang, Bin; Xu, Rui; Yang, Hanbiao; Du, Xinlin; Han, Guangzhou; Wang, Keshi; Cao, Zhaodan; Cheng, Tzuling; Czerwinski, Robert M.; Goggin, Barry S.; Huang, Heli; Halfmann, Megan M.; Maddie, Melissa A.; Morton, Emily L.; Olive, Sarah R.; Tan, Huiling; Xie, Shanhai; Wong, Tai; Josey, John A.; Wallace, Eli M.
      Journal of Medicinal Chemistry (2018), 61 (21), 9691-9721CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      HIF-2α, a member of the HIF family of transcription factors, is a key oncogenic driver in cancers such as clear cell renal cell carcinoma (ccRCC). A signature feature of these cancers is the overaccumulation of HIF-2α protein, often by inactivation of the E3 ligase VHL (von Hippel-Lindau). Herein the authors disclose their structure based drug design (SBDD) approach that culminated in the identification of PT2385, the first HIF-2α antagonist to enter clin. trials. Highlights include the use of a putative n → π*Ar interaction to guide early analog design, the conformational restriction of an essential hydroxyl moiety, and the remarkable impact of fluorination near the hydroxyl group. Evaluation of select compds. from two structural classes in a sequence of PK/PD, efficacy, PK, and metabolite profiling identified I (PT2385, luciferase EC50 = 27 nM) as the clin. candidate. Finally, a retrospective crystallog. anal. describes the structural perturbations necessary for efficient antagonism.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXhvVKqt7%252FN&md5=fcbb754286f47e1b1df3edb62bd2381e
    49. 49
      Boutard, N.; Bialas, A.; Sabiniarz, A.; Guzik, P.; Banaszak, K.; Biela, A.; Bien, M.; Buda, A.; Bugaj, B.; Cieluch, E.; Cierpich, A.; Dudek, L.; Eggenweiler, H. M.; Fogt, J.; Gaik, M.; Gondela, A.; Jakubiec, K.; Jurzak, M.; Kitlinska, A.; Kowalczyk, P.; Kujawa, M.; Kwiecinska, K.; Les, M.; Lindemann, R.; Maciuszek, M.; Mikulski, M.; Niedziejko, P.; Obara, A.; Pawlik, H.; Rzymski, T.; Sieprawska-Lupa, M.; Sowinska, M.; Szeremeta-Spisak, J.; Stachowicz, A.; Tomczyk, M. M.; Wiklik, K.; Wloszczak, L.; Ziemianska, S.; Zarebski, A.; Brzozka, K.; Nowak, M.; Fabritius, C. H. Discovery and Structure-Activity Relationships of N-Aryl 6-Aminoquinoxalines as Potent PFKFB3 Kinase Inhibitors. ChemMedChem. 2019, 14 (1), 169– 181,  DOI: 10.1002/cmdc.201800569
      49
      Discovery and structure-activity relationships of N-aryl 6-aminoquinoxalines as potent PFKFB3 kinase inhibitors
      Boutard, Nicolas; Bialas, Arkadiusz; Sabiniarz, Aleksandra; Guzik, Pawel; Banaszak, Katarzyna; Biela, Artur; Bien, Marcin; Buda, Anna; Bugaj, Barbara; Cieluch, Ewelina; Cierpich, Anna; Dudek, Lukasz; Eggenweiler, Hans-Michael; Fogt, Joanna; Gaik, Monika; Gondela, Andrzej; Jakubiec, Krzysztof; Jurzak, Mirek; Kitlinska, Agata; Kowalczyk, Piotr; Kujawa, Maciej; Kwiecinska, Katarzyna; Les, Marcin; Lindemann, Ralph; Maciuszek, Monika; Mikulski, Maciej; Niedziejko, Paulina; Obara, Alicja; Pawlik, Henryk; Rzymski, Tomasz; Sieprawska-Lupa, Magdalena; Sowinska, Marta; Szeremeta-Spisak, Joanna; Stachowicz, Agata; Tomczyk, Mateusz M.; Wiklik, Katarzyna; Wloszczak, Lukasz; Ziemianska, Sylwia; Zarebski, Adrian; Brzozka, Krzysztof; Nowak, Mateusz; Fabritius, Charles-Henry
      ChemMedChem (2019), 14 (1), 169-181CODEN: CHEMGX; ISSN:1860-7179. (Wiley-VCH Verlag GmbH & Co. KGaA)
      Energy and biomass prodn. in cancer cells are largely supported by aerobic glycolysis in what is called the Warburg effect. The process is regulated by key enzymes, among which phosphofructokinase PFK-2 plays a significant role by producing fructose-2,6-biphosphate; the most potent activator of the glycolysis rate-limiting step performed by phosphofructokinase PFK-1. Herein, the synthesis, biol. evaluation and structure-activity relationship of novel inhibitors of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3), which is the ubiquitous and hypoxia-induced isoform of PFK-2, are reported. X-ray crystallog. and docking were instrumental in the design and optimization of a series of N-aryl 6-aminoquinoxalines. The most potent representative, N-(4-methanesulfonylpyridin-3-yl)-8-(3-methyl-1-benzothiophen-5-yl)quinoxalin-6-amine, displayed an IC50 of 14 nM for the target and an IC50 of 0.49 μM for fructose-2,6-biphosphate prodn. in human colon carcinoma HCT116 cells. This work provides a new entry in the field of PFKFB3 inhibitors with potential for development in oncol.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1cXisVaktbrI&md5=3bfdd82269cdaf101cf7e1a4eaceafe5
    50. 50
      Garcia Fortanet, J.; Chen, C. H.; Chen, Y. N.; Chen, Z.; Deng, Z.; Firestone, B.; Fekkes, P.; Fodor, M.; Fortin, P. D.; Fridrich, C.; Grunenfelder, D.; Ho, S.; Kang, Z. B.; Karki, R.; Kato, M.; Keen, N.; LaBonte, L. R.; Larrow, J.; Lenoir, F.; Liu, G.; Liu, S.; Lombardo, F.; Majumdar, D.; Meyer, M. J.; Palermo, M.; Perez, L.; Pu, M.; Ramsey, T.; Sellers, W. R.; Shultz, M. D.; Stams, T.; Towler, C.; Wang, P.; Williams, S. L.; Zhang, J. H.; LaMarche, M. J. Allosteric Inhibition of SHP2: Identification of a Potent, Selective, and Orally Efficacious Phosphatase Inhibitor. J. Med. Chem. 2016, 59 (17), 7773– 7782,  DOI: 10.1021/acs.jmedchem.6b00680
      50
      Allosteric Inhibition of SHP2: Identification of a Potent, Selective, and Orally Efficacious Phosphatase Inhibitor
      Garcia Fortanet, Jorge; Chen, Christine Hiu-Tung; Chen, Ying-Nan P.; Chen, Zhouliang; Deng, Zhan; Firestone, Brant; Fekkes, Peter; Fodor, Michelle; Fortin, Pascal D.; Fridrich, Cary; Grunenfelder, Denise; Ho, Samuel; Kang, Zhao B.; Karki, Rajesh; Kato, Mitsunori; Keen, Nick; LaBonte, Laura R.; Larrow, Jay; Lenoir, Francois; Liu, Gang; Liu, Shumei; Lombardo, Franco; Majumdar, Dyuti; Meyer, Matthew J.; Palermo, Mark; Perez, Lawrence; Pu, Minying; Ramsey, Timothy; Sellers, William R.; Shultz, Michael D.; Stams, Travis; Towler, Christopher; Wang, Ping; Williams, Sarah L.; Zhang, Ji-Hu; LaMarche, Matthew J.
      Journal of Medicinal Chemistry (2016), 59 (17), 7773-7782CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      SHP2 is a nonreceptor protein tyrosine phosphatase (PTP) encoded by the PTPN11 gene involved in cell growth and differentiation via the MAPK signaling pathway. SHP2 also purportedly plays an important role in the programmed cell death pathway (PD-1/PD-L1). Because it is an oncoprotein assocd. with multiple cancer-related diseases, as well as a potential immunomodulator, controlling SHP2 activity is of significant therapeutic interest. Recently in our labs., a small mol. inhibitor of SHP2 was identified as an allosteric modulator that stabilizes the autoinhibited conformation of SHP2. A high throughput screen was performed to identify progressable chem. matter, and X-ray crystallog. revealed the location of binding in a previously undisclosed allosteric binding pocket. Structure-based drug design was employed to optimize for SHP2 inhibition, and several new protein-ligand interactions were characterized. These studies culminated in the discovery of 6-(4-amino-4-methylpiperidin-1-yl)-3-(2,3-dichlorophenyl)pyrazin-2-amine (SHP099, 1), a potent, selective, orally bioavailable, and efficacious SHP2 inhibitor.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhtVKju7fP&md5=dfe43b5253e172576386f86797087f7b
    51. 51
      Chen, Y. N.; LaMarche, M. J.; Chan, H. M.; Fekkes, P.; Garcia-Fortanet, J.; Acker, M. G.; Antonakos, B.; Chen, C. H.; Chen, Z.; Cooke, V. G.; Dobson, J. R.; Deng, Z.; Fei, F.; Firestone, B.; Fodor, M.; Fridrich, C.; Gao, H.; Grunenfelder, D.; Hao, H. X.; Jacob, J.; Ho, S.; Hsiao, K.; Kang, Z. B.; Karki, R.; Kato, M.; Larrow, J.; La Bonte, L. R.; Lenoir, F.; Liu, G.; Liu, S.; Majumdar, D.; Meyer, M. J.; Palermo, M.; Perez, L.; Pu, M.; Price, E.; Quinn, C.; Shakya, S.; Shultz, M. D.; Slisz, J.; Venkatesan, K.; Wang, P.; Warmuth, M.; Williams, S.; Yang, G.; Yuan, J.; Zhang, J. H.; Zhu, P.; Ramsey, T.; Keen, N. J.; Sellers, W. R.; Stams, T.; Fortin, P. D. Allosteric inhibition of SHP2 phosphatase inhibits cancers driven by receptor tyrosine kinases. Nature 2016, 535 (7610), 148– 152,  DOI: 10.1038/nature18621
      51
      Allosteric inhibition of SHP2 phosphatase inhibits cancers driven by receptor tyrosine kinases
      Chen, Ying-Nan P.; LaMarche, Matthew J.; Chan, Ho Man; Fekkes, Peter; Garcia-Fortanet, Jorge; Acker, Michael G.; Antonakos, Brandon; Chen, Christine Hiu-Tung; Chen, Zhouliang; Cooke, Vesselina G.; Dobson, Jason R.; Deng, Zhan; Fei, Feng; Firestone, Brant; Fodor, Michelle; Fridrich, Cary; Gao, Hui; Grunenfelder, Denise; Hao, Huai-Xiang; Jacob, Jaison; Ho, Samuel; Hsiao, Kathy; Kang, Zhao B.; Karki, Rajesh; Kato, Mitsunori; Larrow, Jay; La Bonte, Laura R.; Lenoir, Francois; Liu, Gang; Liu, Shumei; Majumdar, Dyuti; Meyer, Matthew J.; Palermo, Mark; Perez, Lawrence; Pu, Minying; Price, Edmund; Quinn, Christopher; Shakya, Subarna; Shultz, Michael D.; Slisz, Joanna; Venkatesan, Kavitha; Wang, Ping; Warmuth, Markus; Williams, Sarah; Yang, Guizhi; Yuan, Jing; Zhang, Ji-Hu; Zhu, Ping; Ramsey, Timothy; Keen, Nicholas J.; Sellers, William R.; Stams, Travis; Fortin, Pascal D.
      Nature (London, United Kingdom) (2016), 535 (7610), 148-152CODEN: NATUAS; ISSN:0028-0836. (Nature Publishing Group)
      The non-receptor protein tyrosine phosphatase SHP2, encoded by PTPN11, has an important role in signal transduction downstream of growth factor receptor signalling and was the first reported oncogenic tyrosine phosphatase. Activating mutations of SHP2 have been assocd. with developmental pathologies such as Noonan syndrome and are found in multiple cancer types, including leukemia, lung and breast cancer and neuroblastoma. SHP2 is ubiquitously expressed and regulates cell survival and proliferation primarily through activation of the RAS-ERK signalling pathway2,3. It is also a key mediator of the programmed cell death 1 (PD-1) and B- and T-lymphocyte attenuator (BTLA) immune checkpoint pathways. Redn. of SHP2 activity suppresses tumor cell growth and is a potential target of cancer therapy. Here we report the discovery of a highly potent (IC50 = 0.071 μM), selective and orally bioavailable small-mol. SHP2 inhibitor, SHP099, that stabilizes SHP2 in an auto-inhibited conformation. SHP099 concurrently binds to the interface of the N-terminal SH2, C-terminal SH2, and protein tyrosine phosphatase domains, thus inhibiting SHP2 activity through an allosteric mechanism. SHP099 suppresses RAS-ERK signalling to inhibit the proliferation of receptor-tyrosine-kinase-driven human cancer cells in vitro and is efficacious in mouse tumor xenograft models. Together, these data demonstrate that pharmacol. inhibition of SHP2 is a valid therapeutic approach for the treatment of cancers.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC28XhtVOns7bI&md5=ec8a4b612d5008a2519bd846d45ce019
    52. 52
      Currie, K. S.; Kropf, J. E.; Lee, T.; Blomgren, P.; Xu, J.; Zhao, Z.; Gallion, S.; Whitney, J. A.; Maclin, D.; Lansdon, E. B.; Maciejewski, P.; Rossi, A. M.; Rong, H.; Macaluso, J.; Barbosa, J.; Di Paolo, J. A.; Mitchell, S. A. Discovery of GS-9973, a selective and orally efficacious inhibitor of spleen tyrosine kinase. J. Med. Chem. 2014, 57 (9), 3856– 3873,  DOI: 10.1021/jm500228a
      52
      Discovery of GS-9973, a Selective and Orally Efficacious Inhibitor of Spleen Tyrosine Kinase
      Currie, Kevin S.; Kropf, Jeffrey E.; Lee, Tony; Blomgren, Peter; Xu, Jianjun; Zhao, Zhongdong; Gallion, Steve; Whitney, J. Andrew; Maclin, Deborah; Lansdon, Eric B.; Maciejewski, Patricia; Rossi, Ann Marie; Rong, Hong; Macaluso, Jennifer; Barbosa, James; Di Paolo, Julie A.; Mitchell, Scott A.
      Journal of Medicinal Chemistry (2014), 57 (9), 3856-3873CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      Spleen tyrosine kinase (Syk) is an attractive drug target in autoimmune, inflammatory, and oncol. disease indications. The most advanced Syk inhibitor, R406, (or its prodrug form fostamatinib), has shown efficacy in multiple therapeutic indications, but its clin. progress has been hampered by dose-limiting adverse effects that have been attributed, at least in part, to the off-target activities of R406. It is expected that a more selective Syk inhibitor would provide a greater therapeutic window. Herein the authors report the discovery and optimization of a novel series of imidazo[1,2-a]pyrazine Syk inhibitors. This work culminated in the identification of GS-9973, a highly selective and orally efficacious Syk inhibitor which is currently undergoing clin. evaluation for autoimmune and oncol. indications.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXls1CrsL4%253D&md5=fc53448de1f5416f2c6600876f8c6385
    53. 53
      Buchstaller, H. P.; Anlauf, U.; Dorsch, D.; Kuhn, D.; Lehmann, M.; Leuthner, B.; Musil, D.; Radtki, D.; Ritzert, C.; Rohdich, F.; Schneider, R.; Esdar, C. Discovery and Optimization of 2-Arylquinazolin-4-ones into a Potent and Selective Tankyrase Inhibitor Modulating Wnt Pathway Activity. J. Med. Chem. 2019, 62 (17), 7897– 7909,  DOI: 10.1021/acs.jmedchem.9b00656
      53
      Discovery and Optimization of 2-Arylquinazolin-4-ones into a Potent and Selective Tankyrase Inhibitor Modulating Wnt Pathway Activity
      Buchstaller, Hans-Peter; Anlauf, Uwe; Dorsch, Dieter; Kuhn, Daniel; Lehmann, Martin; Leuthner, Birgitta; Musil, Djordje; Radtki, Daniela; Ritzert, Claudio; Rohdich, Felix; Schneider, Richard; Esdar, Christina
      Journal of Medicinal Chemistry (2019), 62 (17), 7897-7909CODEN: JMCMAR; ISSN:0022-2623. (American Chemical Society)
      Tankyrases 1 and 2 (TNKS1/2) are promising pharmacol. targets which recently gained interest for anticancer therapy in Wnt pathway dependent tumors. 2-Aryl-quinazolinones were identified and optimized into potent tankyrase inhibitors through SAR exploration around the quinazolinone core and the 4'-position of the Ph residue. These efforts were supported by anal. of TNKS X-ray and Watermap structures and resulted in compd. 5k(I), a potent, selective tankyrase inhibitor with favorable pharmacokinetic properties. The X-ray structure of I in complex with TNKS1 was solved and confirmed the design hypothesis. Modulation of Wnt pathway activity was demonstrated with this compd. in a colorectal xenograft model in vivo.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC1MXhsFWms7bI&md5=7fcc6c91c6e614482386af2094675837
    54. 54
      Jakalian, A.; Bush, B. L.; Jack, D. B.; Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: I. Method. J. Comput. Chem. 2000, 21 (2), 132– 146,  DOI: 10.1002/(SICI)1096-987X(20000130)21:2<132::AID-JCC5>3.0.CO;2-P
      54
      Fast, efficient generation of high-quality atomic charges. AM1-BCC model: I. Method
      Jakalian, Araz; Bush, Bruce L.; Jack, David B.; Bayly, Christopher I.
      Journal of Computational Chemistry (2000), 21 (2), 132-146CODEN: JCCHDD; ISSN:0192-8651. (John Wiley & Sons, Inc.)
      The AM1-BCC method quickly and efficiently generates high-quality at. charges for use in condensed-phase simulations. The underlying features of the electron distribution including formal charge and delocalization are first captured by AM1 at. charges for the individual mol. Bond charge corrections (BCCs), which have been parameterized against the HF/6-31G* electrostatic potential (ESP) of a training set of compds. contg. relevant functional groups, are then added using a formalism identical to the consensus BCI (bond charge increment) approach. As a proof of the concept, we fit BCCs simultaneously to 45 compds. including O-, N-, and S-contg. functionalities, aroms., and heteroaroms., using only 41 BCC parameters. AM1-BCC yields charge sets of comparable quality to HF/6-31G* ESP-derived charges in a fraction of the time while reducing instabilities in the at. charges compared to direct ESP-fit methods. We then apply the BCC parameters to a small "test set" consisting of aspirin, D-glucose, and eryodictyol; the AM1-BCC model again provides at. charges of quality comparable with HF/6-31G* RESP charges, as judged by an increase of only 0.01 to 0.02 at. units in the root-mean-square (RMS) error in ESP. Based on these encouraging results, we intend to parameterize the AM1-BCC model to provide a consistent charge model for any org. or biol. mol.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD3cXkt1SgsQ%253D%253D&md5=aa4774739d3491dbc8dd95ed1948d856
    55. 55
      Jakalian, A.; Jack, D. B.; Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. Parameterization and validation. J. Comput. Chem. 2002, 23 (16), 1623– 1641,  DOI: 10.1002/jcc.10128
      55
      Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. parameterization and validation
      Jakalian, Araz; Jack, David B.; Bayly, Christopher I.
      Journal of Computational Chemistry (2002), 23 (16), 1623-1641CODEN: JCCHDD; ISSN:0192-8651. (John Wiley & Sons, Inc.)
      We present the first global parameterization and validation of a novel charge model, called AM1-BCC, which quickly and efficiently generates high-quality at. charges for computer simulations of org. mols. in polar media. The goal of the charge model is to produce at. charges that emulate the HF/6-31G* electrostatic potential (ESP) of a mol. Underlying electronic structure features, including formal charge and electron delocalization, are first captured by AM1 population charges; simple additive bond charge corrections (BCCs) are then applied to these AM1 at. charges to produce the AM1-BCC charges. The parameterization of BCCs was carried out by fitting to the HF/6-31G* ESP of a training set of >2700 mols. Most org. functional groups and their combinations were sampled, as well as an extensive variety of cyclic and fused bicyclic heteroaryl systems. The resulting BCC parameters allow the AM1-BCC charging scheme to handle virtually all types of org. compds. listed in The Merck Index and the NCI Database. Validation of the model was done through comparisons of hydrogen-bonded dimer energies and relative free energies of solvation using AM1-BCC charges in conjunction with the 1994 Cornell et al. forcefield for AMBER. Homo-dimer and hetero-dimer hydrogen-bond energies of a diverse set of org. mols. were reproduced to within 0.95 kcal/mol RMS deviation from the ab initio values, and for DNA dimers the energies were within 0.9 kcal/mol RMS deviation from ab initio values. The calcd. relative free energies of solvation for a diverse set of monofunctional isosteres were reproduced to within 0.69 kcal/mol of expt. In all these validation tests, AMBER with the AM1-BCC charge model maintained a correlation coeff. above 0.96. Thus, the parameters presented here for use with the AM1-BCC method present a fast, accurate, and robust alternative to HF/6-31G* ESP-fit charges for general use with the AMBER force field in computer simulations involving org. small mols.
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BD38XosF2rt74%253D&md5=3f2b738617bb17b2898f7ac4d751d7ec
    56. 56
      Zheng, Z.; Wang, T.; Li, P.; Merz, K. M., Jr. KECSA-Movable Type Implicit Solvation Model (KMTISM). J. Chem. Theory Comput. 2015, 11 (2), 667– 682,  DOI: 10.1021/ct5007828
      56
      KECSA-Movable Type Implicit Solvation Model (KMTISM)
      Zheng, Zheng; Wang, Ting; Li, Pengfei; Merz, Kenneth M.
      Journal of Chemical Theory and Computation (2015), 11 (2), 667-682CODEN: JCTCCE; ISSN:1549-9618. (American Chemical Society)
      Computation of the solvation free energy for chem. and biol. processes has long been of significant interest. The key challenges to effective solvation modeling center on the choice of potential function and configurational sampling. Herein, an energy sampling approach termed the "Movable Type" (MT) method, and a statistical energy function for solvation modeling, "Knowledge-based and Empirical Combined Scoring Algorithm" (KECSA) are developed and utilized to create an implicit solvation model: KECSA-Movable Type Implicit Solvation Model (KMTISM) suitable for the study of chem. and biol. systems. KMTISM is an implicit solvation model, but the MT method performs energy sampling at the atom pairwise level. For a specific mol. system, the MT method collects energies from prebuilt databases for the requisite atom pairs at all relevant distance ranges, which by its very construction encodes all possible mol. configurations simultaneously. Unlike traditional statistical energy functions, KECSA converts structural statistical information into categorized atom pairwise interaction energies as a function of the radial distance instead of a mean force energy function. Within the implicit solvent model approxn., aq. solvation free energies are then obtained from the NVT ensemble partition function generated by the MT method. Validation is performed against several subsets selected from the Minnesota Solvation Database v2012. Results are compared with several solvation free energy calcn. methods, including a one-to-one comparison against two commonly used classical implicit solvation models: MM-GBSA and MM-PBSA. Comparison against a quantum mechanics based polarizable continuum model is also discussed (Cramer and Truhlar's Solvation Model 12).
      >> More from SciFinder ®
      https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2cXitFehsLfJ&md5=debe378166aeaac91249dea799add7c5

  • Supporting Information

    Supporting Information

    ARTICLE SECTIONS
    Jump To

    The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.2c00278.

    • Detailed explanation of Movable Type binding free energy estimation protocol (PDF)

    • Free energy calculation results for CASF-2016 benchmark, rescale set, and Merck KGaA benchmark using cMD-MT-amber, cMD-MT-garf, CHMC-cMD-MT-amber, and CHMC-cMD-MT-garf protocols (PDF)


    Terms & Conditions

    Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html.

PDF [ 12MB]

Pair your accounts.

Export articles to Mendeley

Get article recommendations from ACS based on references in your Mendeley library.

Pair your accounts.

Export articles to Mendeley

Get article recommendations from ACS based on references in your Mendeley library.

You’ve supercharged your research process with ACS and Mendeley!

STEP 1:
Click to create an ACS ID

Please note: If you switch to a different device, you may be asked to login again with only your ACS ID.

Please note: If you switch to a different device, you may be asked to login again with only your ACS ID.

Please note: If you switch to a different device, you may be asked to login again with only your ACS ID.

MENDELEY PAIRING EXPIRED
Your Mendeley pairing has expired. Please reconnect