Binding affinity prediction with different force fields: examination of the linear interaction energy method

MedChemExpress compounds prevent neuraminidase N1 via physics- and knowledge-based methods

Influenza A viruses spread out worldwide, causing several global concerns. Hence, discovering neuraminidase inhibitors to prevent the influenza A virus is of great interest. In this work, a machine learning model was employed to evaluate the ligand-binding affinity of ca. 10 000 compounds from the MedChemExpress (MCE) database for inhibiting neuraminidase. Atomistic simulations, including molecular docking and molecular dynamics simulations, then confirmed the ligand-binding affinity. Furthermore, we clarified the physical insights into the binding process of ligands to neuraminidase. It was found that five compounds, including micronomicin, didesmethyl cariprazine, argatroban, Kgp-IN-1, and AY 9944, are able to inhibit neuraminidase N1 of the influenza A virus. Ten residues, including Glu119, Asp151, Arg152, Trp179, Gln228, Glu277, Glu278, Arg293, Asn295, and Tyr402, may be very important in controlling the ligand-binding process to N1.

Dynamic Profiling and Binding Affinity Prediction of NBTI Antibacterials against DNA Gyrase Enzyme by Multidimensional Machine Learning and Molecular Dynamics Simulations

Bacterial type II topoisomerases are well-characterized and clinically important targets for antibacterial chemotherapy. Novel bacterial topoisomerase inhibitors (NBTIs) are a newly disclosed class of antibacterials. Prediction of their binding affinity to these enzymes would be beneficial for de novo design/optimization of new NBTIs. Utilizing in vitro NBTI experimental data, we constructed two comprehensive multidimensional DNA gyrase surrogate models for Staphylococcus aureus (q2 = 0.791) and Escherichia coli (q2 = 0.806). Both models accurately predicted the IC50s of 26 NBTIs from our recent studies. To investigate the NBTI's dynamic profile and binding to both targets, 10 selected NBTIs underwent molecular dynamics (MD) simulations. The analysis of MD production trajectories confirmed key hydrogen-bonding and hydrophobic contacts that NBTIs establish in both enzymes. Moreover, the binding free energies of selected NBTIs were computed by the linear interaction energy (LIE) method employing an in-house derived set of fitting parameters (α = 0.16, β = 0.029, γ = 0.0, and intercept = -1.72), which are successfully applicable to DNA gyrase of Gram-positive/Gram-negative pathogens. Both methods offer accurate predictions of the binding free energies of NBTIs against S. aureus and E. coli DNA gyrase. We are confident that this integrated modeling approach could be valuable in the de novo design and optimization of efficient NBTIs for combating resistant bacterial pathogens.

Machine learning combines atomistic simulations to predict SARS-CoV-2 Mpro inhibitors from natural compounds

To date, the COVID-19 pandemic has still been infectious around the world, continuously causing social and economic damage on a global scale. One of the most important therapeutic targets for the treatment of COVID-19 is the main protease (Mpro) of SARS-CoV-2. In this study, we combined machine-learning (ML) model with atomistic simulations to computationally search for highly promising SARS-CoV-2 Mpro inhibitors from the representative natural compounds of the National Cancer Institute (NCI) Database. First, the trained ML model was used to scan the library quickly and reliably for possible Mpro inhibitors. The ML output was then confirmed using atomistic simulations integrating molecular docking and molecular dynamic simulations with the linear interaction energy scheme. The results turned out to show that there was evidently good agreement between ML and atomistic simulations. Ten substances were proposed to be able to inhibit SARS-CoV-2 Mpro. Seven of them have high-nanomolar affinity and are very potential inhibitors. The strategy has been proven to be reliable and appropriate for fast prediction of SARS-CoV-2 Mpro inhibitors, benefiting for new emerging SARS-CoV-2 variants in the future accordingly.

Natural compounds inhibit Monkeypox virus methyltransferase VP39 in silico studies

VP39, an essential 2'-O-RNA methyltransferase enzyme discovered in Monkeypox virus (MPXV), plays a vital role in viral RNA replication and transcription. Inhibition of the enzyme may prevent viral replication. In this context, using a combination of molecular docking and molecular dynamics (MDs) simulations, the inhibitory ability of NCI Diversity Set VII natural compounds to VP39 protein was investigated. It should be noted that the computed binding free energy of ligand via molecular docking and linear interaction energy (LIE) approaches are in good agreement with the corresponding experiments with coefficients of R = 0.72 and 0.75, respectively. NSC 319990, NSC 196515 and NSC 376254 compounds were demonstrated that can inhibit MPVX methyltransferase VP39 protein with the similar affinity compared to available inhibitor sinefungin. Moreover, nine residues involving Gln39, Gly68, Gly72, Asp95, Arg97, Val116, Asp138, Arg140 and Asn156 may be argued that they play an important role in binding process of inhibitors to VP39.Communicated by Ramaswamy H. Sarma.

Calculation of binding affinity of JAK1 inhibitors via accurately computational estimation

Janus kinase 1 (JAK1) is a tyrosine kinase that is involved in the initiation of responses to a number of different cytokine receptor families. The JAK1-dependent pathway is a therapeutic target, and several JAK inhibitors have been developed thanks to intensive research. However, since the ATP binding sites of JAK family members are quite alike, JAK1 inhibitors can thus be less selective, resulting in unanticipated adverse effects. Despite this, minor variations in the ATP-binding site have been extensively used to find a variety of small compounds with different inhibitory properties. Stronger binding affinity of JAK1 inhibitors is believed to be able to reduce the negative effects, leading to better treatment results. Therefore, a thorough computational search that can effectively identify ligands with extremely high binding affinity for JAK1 to serve as promising inhibitors is required. Here, a method combining steered-molecular dynamic (SMD) simulations with a modified linear interaction energy (LIE) model has been developed to evaluate the binding affinities of known JAK1 inhibitors. The correlation coefficient between the estimated and experimental values was 0.72 and a root-mean-square error was 0.97 kcal•mol-1, revealing that the SMD/LIE method can precisely and quickly predict the binding free energies of JAK1 inhibitors. Furthermore, three marine fungus-derived compounds, namely hansforesters E, hansforesters G and tetroazolemycins B, were identified to be particularly promising JAK1 inhibitors, accordingly. These findings show that the SMD/LIE method has a lot of promise for in silico screening of possible JAK1 inhibitors from a vast number of compounds that are now accessible.Communicated by Ramaswamy H. Sarma.

Molecular dynamics simulations suggest Thiosemicarbazones can bind p53 cancer mutant R175H

Cancer pathologies are associated with the unfolding and aggregation of most recurring mutations in the DNA Binding Domain (DBD) of p53 that coordinate the destabilization of protein. Substitution at the 175th codon with arginine to histidine (R175H, a mutation of large to small side-chain amino acid) destabilizes the DBD by 3 kcal/mol and triggers breasts, lung cancer, etc. Stabilizing the p53 mutant by small molecules offers an attractive drug-targeted anti-cancer therapy. The thiosemicarbazone (TSC) molecules NPC and DPT are known to act as zinc-metallochaperones to reactivate p53R175H. Here, a combination of LESMD simulations for 10 TSC conformations with a p53R175H receptor, single ligand-protein conformation MD, and ensemble docking with multiple p53R175H conformations observed during simulations is suggested to identify the potential binding site of the target protein in light of their importance for the direct TSC - p53R175H binding. NPC binds mutant R175H in the loop region L2-L3, forming pivotal hydrogen bonds with HIS175, pi‑sulfur bonds with TYR163, and pi-alkyl linkages with ARG174 and PRO190, all of which are contiguous to the zinc-binding native site on p53DBD. DPT, on the other hand, was primarily targeting alternative binding sites such as the loop-helix L1/H2 region and the S8 strand. The similar structural characteristics of TSC-bound p53R175H complexes with wild-type p53DBD are thought to be attributable to involved interactions that favour binding free energy contributions of TSC ligands. Our findings may be useful in the identification of novel pockets with druggable properties.

Identifying Protein-Ligand Interactions via a Novel Distance Self-Feedback Biomolecular Interaction Network

Efficient and accurate characterizations of protein-ligand interactions are key to understanding biology at the molecular level. They are particularly useful in pharmaceutical industry applications. They are usually computationally demanding for those widely applied dynamics-based methods in identifying important residues or calculating ligand binding free energy. In this work, we proposed a graph deep learning (DL) framework, namely, the distance self-feedback biomolecular interaction network (DSBIN), in which the relationship between the complex structure and binding affinity can be established by means of a carefully designed distance self-feedback module and interaction layer. Our model can directly provide a quantitative evaluation of inhibitor binding affinities (pK_d). More importantly, the DSBIN model efficiently identifies key interactions for inhibitor binding and thus intrinsically bears the interpretability. Its generalization performance was further verified using 1405 unseen structures. The predicted binding free energies' deviations were calculated to be less than 1.37 kcal/mol for more than 55% structures. Moreover, we also compared the DSBIN model with a commonly used theoretical method in calculating the substrate binding free energy, MM/GBSA. Our results show that the current DL model has generally better performance in predicting the binding free energy. For a specific complex system, mannopentaose/TmCBM27, the DSBIN predicted binding free energy is -8.21 kcal/mol, which is very close to experimentally measured -7.76 kcal/mol and MM/GBSA calculated -7.16 kcal/mol. Meanwhile, all important aromatic residues around the binding pocket can be identified by our DL model. Considering the accuracy and efficiency of the newly developed DL model, it may be very helpful in the field of drug design and molecular recognition.

Mechanistic Insights of Polyphenolic Compounds from Rosemary Bound to Their Protein Targets Obtained by Molecular Dynamics Simulations and Free-Energy Calculations

Rosemary represents an important medicinal plant that has been attributed with various health-promoting properties, especially antioxidative, anti-inflammatory, and anticarcinogenic activities. Carnosic acid, carnosol, and rosmanol, as well as the phenolic acid ester rosmarinic acid, are the main compounds responsible for these actions. In our earlier research, we carried out an inverse molecular docking at the proteome scale to determine possible protein targets of the mentioned compounds. Here, we subjected the previously identified ligand-protein complexes with HIV-1 protease, K-RAS, and factor X to molecular dynamics simulations coupled with free-energy calculations. We observed that carnosic acid and rosmanol act as viable binders of the HIV-1 protease. In addition, carnosol represents a potential binder of the oncogene protein K-RAS. On the other hand, rosmarinic acid was characterized as a weak binder of factor X. We also emphasized the importance of water-mediated hydrogen-bond networks in stabilizing the binding conformation of the studied polyphenols, as well as in mechanistically explaining their promiscuous nature.

Computational insights for predicting the binding and selectivity of peptidomimetic plasmepsin IV inhibitors against cathepsin D

Plasmepsins (Plms) are aspartic proteases involved in the degradation of human hemoglobin by P. falciparum and are essential for the survival and growth of the parasite. Therefore, Plm enzymes are reported as an important antimalarial drug target. Herein, we have applied molecular docking, molecular dynamics (MD) simulations, and binding free energy with the Linear Interaction Energy (LIE) approach to investigate the binding of peptidomimetic PlmIV inhibitors with a particular focus on understanding their selectivity against the human Asp protease cathepsin D (CatD). The residual decomposition analysis results suggest that amino acid differences in the subsite S3 of PlmIV and CatD are responsible for the higher selectivity of the 5a inhibitor. These findings yield excellent agreement with experimental binding data and provide new details regarding van der Waals and electrostatic interactions of subsite residues as well as structural properties of the PlmIV and CatD systems.

Characterizing the ligand-binding affinity toward SARS-CoV-2 Mpro via physics- and knowledge-based approaches

Computational approaches, including physics- and knowledge-based methods, have commonly been used to determine the ligand-binding affinity toward SARS-CoV-2 main protease (Mpro or 3CLpro). Strong binding ligands can thus be suggested as potential inhibitors for blocking the biological activity of the protease. In this context, this paper aims to provide a short review of computational approaches that have recently been applied in the search for inhibitor candidates of Mpro. In particular, molecular docking and molecular dynamics (MD) simulations are usually combined to predict the binding affinity of thousands of compounds. Quantitative structure-activity relationship (QSAR) is the least computationally demanding and therefore can be used for large chemical collections of ligands. However, its accuracy may not be high. Moreover, the quantum mechanics/molecular mechanics (QM/MM) method is most commonly used for covalently binding inhibitors, which also play an important role in inhibiting the activity of SARS-CoV-2. Furthermore, machine learning (ML) models can significantly increase the searching space of ligands with high accuracy for binding affinity prediction. Physical insights into the binding process can then be confirmed via physics-based calculations. Integration of ML models into computational chemistry provides many more benefits and can lead to new therapies sooner.

Methodology-Centered Review of Molecular Modeling, Simulation, and Prediction of SARS-CoV-2

Despite tremendous efforts in the past two years, our understanding of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), virus-host interactions, immune response, virulence, transmission, and evolution is still very limited. This limitation calls for further in-depth investigation. Computational studies have become an indispensable component in combating coronavirus disease 2019 (COVID-19) due to their low cost, their efficiency, and the fact that they are free from safety and ethical constraints. Additionally, the mechanism that governs the global evolution and transmission of SARS-CoV-2 cannot be revealed from individual experiments and was discovered by integrating genotyping of massive viral sequences, biophysical modeling of protein-protein interactions, deep mutational data, deep learning, and advanced mathematics. There exists a tsunami of literature on the molecular modeling, simulations, and predictions of SARS-CoV-2 and related developments of drugs, vaccines, antibodies, and diagnostics. To provide readers with a quick update about this literature, we present a comprehensive and systematic methodology-centered review. Aspects such as molecular biophysics, bioinformatics, cheminformatics, machine learning, and mathematics are discussed. This review will be beneficial to researchers who are looking for ways to contribute to SARS-CoV-2 studies and those who are interested in the status of the field.

Fragment-based inhibitor design for SARS-CoV2 main protease

COVID-19 disease caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV2) has resulted in tremendous loss of lives across the world and is continuing to do so. Extensive work is under progress to develop inhibitors which can prevent the disease by arresting the virus in its life cycle. One such way is by targeting the main protease of the virus which is crucial for the cleavage and conversion of polyproteins into functional units of polypeptides. In this endeavor, our effort was to identify hit molecule inhibitors for SARS-CoV2 main protease using fragment-based drug discovery (FBDD), based on the available crystal structure of chromene-based inhibitor (PDB_ID: 6M2N). The designed molecules were validated by molecular docking and molecular dynamics simulations. The stability of the complexes was further assessed by calculating their binding free energies, normal mode analysis, mechanical stiffness, and principal component analysis.

Computational studies on the design of NCI natural products as inhibitors to SARS-CoV-2 main protease

The pandemic coronavirus disease (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has resulted in more than 5 million deaths globally. Currently there are no effective drugs available to treat COVID-19. The viral protease replication can be blocked by the inhibition of main protease that is encoded in polyprotein 1a and is therefore a potential protein target for drug discovery. We have carried out virtual screening of NCI natural compounds followed by molecular docking in order to identify hit molecules as probable SARS-CoV-2 main protease inhibitors. The molecular dynamics (MD) simulations of apo form in complex with N3, α-ketoamide and NCI natural products was used to validate the screened compounds. The MD simulations trajectories were analyzed using normal mode analysis and principal component analysis revealing dynamical nature of the protein. These findings aid in understanding the binding of natural products and molecular mechanisms of SARS-CoV-2 main protease inhibition.Communicated by Ramaswamy H. Sarma.

Match_Motif: A rapid computational tool to assist in protein-protein interaction design

In order to generate protein assemblies with a desired function, the rational design of protein-protein binding interfaces is of significant interest. Approaches based on random mutagenesis or directed evolution may involve complex experimental selection procedures. Also, molecular modeling approaches to design entirely new proteins and interactions with partner molecules can involve large computational efforts and screening steps. In order to simplify at least the initial effort for designing a putative binding interface between two proteins the Match_Motif approach has been developed. It employs the large collection of known protein-protein complex structures to suggest interface modifications that may lead to improved binding for a desired input interaction geometry. The approach extracts interaction motifs based on the backbone structure of short (four residues) segments and the relative arrangement with respect to short segments on the partner protein. The interaction geometry is used to search through a database of such motifs in known stable bound complexes. All matches are rapidly identified (within a few seconds) and collected and can be used to guide changes in the interface that may lead to improved binding. In the output, an alternative interface structure is also proposed based on the frequency of occurrence of side chains at a given interface position in all matches and based on sterical considerations. Applications of the procedure to known complex structures and alternative arrangements are presented and discussed. The program, data files, and example applications can be downloaded from https://www.groups.ph.tum.de/t38/downloads/.

Studying the Effect of Amino Acid Substitutions in the M2 Ion Channel of the Influenza Virus on the Antiviral Activity of the Aminoadamantane Derivative In Vitro and In Silico

Purpose: The aminoadamantane derivative of L-histidyl-1-adamantayl ethylamine hydrochloride (HCl*H-His-Rim) has showed a high inhibition level against influenza A virus strains in vitro. The aim of this work is to search and establish evidence of the direct effect of the drug on influenza A virus proton channel M2.

Methods: The compound HCl*H-His-Rim was obtained by classical peptide synthesis methods. Influenza A virus mutants of A/PuertoRico/8/34(H1N1) strain were obtained by reverse genetics methods. The mutant samples of the virus were cultured on chicken embryos with a virus titer in the hemagglutination test. ELISA was carried out on Madin-Darby canine kidney (MDCK) monolayer cells when multiplying the virus 10^-4-10^-6. The binding stability of HCl*H-His-Rim was compared to those of M2 (S31N) and M2 (S31N_A30T) channels by molecular dynamic (MD) modeling. The calculation was performed taking into account the interaction with the model lipid bilayer (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) in the presence of water molecules in accordance with the three-center model.

Results: It was found that HCl*H-His-Rim is a direct action drug against influenza A. The most likely conformation of drug binding to target protein has been shown. It has been found that the A30T mutation reduces the binding energy of the drug, and the results obtained in vitro have confirmed the data calculated in silico.

Conclusion: The mechanism of action of HCl*H-His-Rim is directly related to the suppression of the function of the proton channel M2 of influenza A virus.

FFENCODER-PL: Pair Wise Energy Descriptors for Protein-Ligand Pose Selection

Scoring functions are the essential component in molecular docking methods. An accurate scoring function is expected to distinguish the native ligand pose from decoy poses. Our previous experience (Pei et al. J. Chem. Inf. Model. 2019, 59 (7), 3305-3315) proved that combining the random forest (RF) algorithm with knowledge-based potential functions can emphasize germane pair wise interactions and improve the performance of original knowledge-based potential functions on protein-ligand decoy detection. One of the most important potential function classes is the force field (FF) potential with one example being the Amber collection of FFs, which are widely available in the AMBER suite of simulation programs. However, for use in RF modeling studies, one needs pair wise energies that are hard to directly extract from Amber. To address this issue, FFENCODER-PL was constructed to calculate the pair wise energies based on the FF14SB and GAFF2 FFs in Amber. FFENCODER-PL was validated using 275 ligand and 21 protein-ligand structures. RF models were built by combining an RF classification algorithm with the pair wise energies calculated from FFENCODER-PL. CASF-2016 (Su et al. J. Chem. Inf. Model. 2019, 59, 895-913) was employed to test the performance of the resultant RF models, which outperformed 33 scoring functions on accuracy and native ranking tests. For the best decoy RMSD test, RF models give a best decoy with an RMSD of around 2 Å from the native pose after including the best decoy-decoy comparisons in the RF model. The relative importance of the RF algorithm and force field potentials was also tested with the results suggesting that both the RF algorithm and force field potentials are important and combining them is the only way to achieve high accuracy. Finally, FFENCODER-PL makes force field-based pair wise energies available for further development of machine learning-based scoring functions. The codes and data used in this paper can be found at https://github.com/JunPei000/Amber_protein_ligand_encoding.

Recent Developments in Free Energy Calculations for Drug Discovery

The grand challenge in structure-based drug design is achieving accurate prediction of binding free energies. Molecular dynamics (MD) simulations enable modeling of conformational changes critical to the binding process, leading to calculation of thermodynamic quantities involved in estimation of binding affinities. With recent advancements in computing capability and predictive accuracy, MD based virtual screening has progressed from the domain of theoretical attempts to real application in drug development. Approaches including the Molecular Mechanics Poisson Boltzmann Surface Area (MM-PBSA), Linear Interaction Energy (LIE), and alchemical methods have been broadly applied to model molecular recognition for drug discovery and lead optimization. Here we review the varied methodology of these approaches, developments enhancing simulation efficiency and reliability, remaining challenges hindering predictive performance, and applications to problems in the fields of medicine and biochemistry.

Benchmark of Popular Free Energy Approaches Revealing the Inhibitors Binding to SARS-CoV-2 Mpro

The COVID-19 pandemic has killed millions of people worldwide since its outbreak in December 2019. The pandemic is caused by the SARS-CoV-2 virus whose main protease (Mpro) is a promising drug target since it plays a key role in viral proliferation and replication. Currently, developing an effective therapy is an urgent task, which requires accurately estimating the ligand-binding free energy to SARS-CoV-2 Mpro. However, it should be noted that the accuracy of a free energy method probably depends on the protein target. A highly accurate approach for some targets may fail to produce a reasonable correlation with the experiment when a novel enzyme is considered as a drug target. Therefore, in this context, the ligand-binding affinity to SARS-CoV-2 Mpro was calculated via various approaches. The molecular docking approach was manipulated using Autodock Vina (Vina) and Autodock4 (AD4) protocols to preliminarily investigate the ligand-binding affinity and pose to SARS-CoV-2 Mpro. The binding free energy was then refined using the fast pulling of ligand (FPL), linear interaction energy (LIE), molecular mechanics-Poisson-Boltzmann surface area (MM-PBSA), and free energy perturbation (FEP) methods. The benchmark results indicated that for docking calculations, Vina is more accurate than AD4, and for free energy methods, FEP is the most accurate method, followed by LIE, FPL, and MM-PBSA (FEP > LIE ≈ FPL > MM-PBSA). Moreover, atomistic simulations revealed that the van der Waals interaction is the dominant factor. The residues Thr26, His41, Ser46, Asn142, Gly143, Cys145, His164, Glu166, and Gln189 are essential elements affecting the binding process. Our benchmark provides guidelines for further investigations using computational approaches.

Structural Basis of Inhibition of Human Insulin-Regulated Aminopeptidase (IRAP) by Benzopyran-Based Inhibitors

Inhibition of the insulin-regulated aminopeptidase (IRAP) improves memory and cognition in animal models. The enzyme has recently been crystallized and several series of inhibitors reported. We herein focused on one series of benzopyran-based inhibitors of IRAP known as the HFI series, with unresolved binding mode to IRAP, and developed a robust computational model to explain the structure-activity relationship (SAR) and potentially guide their further optimization. The binding model here proposed places the benzopyran ring in the catalytic binding site, coordinating the Zn²⁺ ion through the oxygen in position 3, in contrast to previous hypothesis. The whole series of HFI compounds was then systematically simulated, starting from this binding mode, using molecular dynamics and binding affinity estimated with the linear interaction energy (LIE) method. The agreement with experimental affinities supports the binding mode proposed, which was further challenged by rigorous free energy perturbation (FEP) calculations. Here, we found excellent correlation between experimental and calculated binding affinity differences, both between selected compound pairs and also for recently reported experimental data concerning the site directed mutagenesis of residue Phe544. The computationally derived structure-activity relationship of the HFI series and the understanding of the involvement of Phe544 in the binding of this scaffold provide valuable information for further lead optimization of novel IRAP inhibitors.

Recent Developments in Linear Interaction Energy Based Binding Free Energy Calculations

The linear interaction energy (LIE) approach is an end-point method to compute binding affinities. As such it combines explicit conformational sampling (of the protein-bound and unbound-ligand states) with efficiency in calculating values for the protein-ligand binding free energy ΔG _bind . This perspective summarizes our recent efforts to use molecular simulation and empirically calibrated LIE models for accurate and efficient calculation of ΔG _bind for diverse sets of compounds binding to flexible proteins (e.g., Cytochrome P450s and other proteins of direct pharmaceutical or biochemical interest). Such proteins pose challenges on ΔG _bind computation, which we tackle using a previously introduced statistically weighted LIE scheme. Because calibrated LIE models require empirical fitting of scaling parameters, they need to be accompanied with an applicability domain (AD) definition to provide a measure of confidence for predictions for arbitrary query compounds within a reference frame defined by a collective chemical and interaction space. To enable AD assessment of LIE predictions (or other protein-structure and -dynamic based ΔG _bind calculations) we recently introduced strategies for AD assignment of LIE models, based on simulation and training data only. These strategies are reviewed here as well, together with available tools to facilitate and/or automate LIE computation (including software for combined statistically-weighted LIE calculations and AD assessment).

How Computational Chemistry and Drug Delivery Techniques Can Support the Development of New Anticancer Drugs

The early and late development of new anticancer drugs, small molecules or peptides can be slowed down by some issues such as poor selectivity for the target or poor ADME properties. Computer-aided drug design (CADD) and target drug delivery (TDD) techniques, although apparently far from each other, are two research fields that can give a significant contribution to overcome these problems. Their combination may provide mechanistic understanding resulting in a synergy that makes possible the rational design of novel anticancer based therapies. Herein, we aim to discuss selected applications, some also from our research experience, in the fields of anticancer small organic drugs and peptides.

Optimal designs for pairwise calculation: An application to free energy perturbation in minimizing prediction variability

Pairwise-based methods such as the free energy perturbation (FEP) method have been widely deployed to compute the binding free energy differences between two similar host-guest complexes. The calculated pairwise free energy difference is either directly adopted or transformed to absolute binding free energy for molecule rank ordering. We investigated, through both analytic derivations and simulations, how the selection of pairs in the experiment could impact the overall prediction precision. Our studies showed that (1) the estimated absolute binding free energy ( Δ G ^ ) derived from calculated pairwise differences (ΔΔG) through weighted least squares fitting is more precise in prediction than the pairwise difference values when the number of pairs is more than the number of ligands and (2) prediction precision is influenced by both the total number of pairs and the specifically selected pairs, the latter being critically important when the number of calculated pairs is limited. Furthermore, we applied optimal experimental design in pair selection and found that the optimally selected pairs can outperform randomly selected pairs in prediction precision. In an illustrative example, we showed that, upon weighing ligand structure similarity into design optimization, the weighted optimal designs are more efficient than the literature reported designs. This work provides a new approach to assess retrospective pairwise-based prediction results, and a method to design new prospective pairwise-based experiments for molecular lead optimization. © 2019 Wiley Periodicals, Inc.

Macrocyclic peptidomimetics as inhibitors of insulin-regulated aminopeptidase (IRAP)

Macrocyclic analogues of the linear hexapeptide, angiotensin IV (AngIV) have proved to be potent inhibitors of insulin-regulated aminopeptidase (IRAP, oxytocinase, EC 3.4.11.3). Along with higher affinity, macrocycles may also offer better metabolic stability, membrane permeability and selectivity, however predicting the outcome of particular cycle modifications is challenging. Here we describe the development of a series of macrocyclic IRAP inhibitors with either disulphide, olefin metathesis or lactam bridges and variations of ring size and other functionality. The binding mode of these compounds is proposed based on molecular dynamics analysis. Estimation of binding affinities (ΔG) and relative binding free energies (ΔΔG) with the linear interaction energy (LIE) method and free energy perturbation (FEP) method showed good general agreement with the observed inhibitory potency. Experimental and calculated data highlight the cumulative importance of an intact N-terminal peptide, the specific nature of the macrocycle, the phenolic oxygen and the C-terminal functionality.

Effective estimation of the inhibitor affinity of HIV-1 protease via a modified LIE approach

The inhibition of the Human Immunodeficiency Virus Type 1 Protease (HIV-1 PR) can prevent the synthesis of new viruses. Computer-aided drug design (CADD) would enhance the discovery of new therapies, through which the estimation of ligand-binding affinity is critical to predict the most efficient inhibitor. A time-consuming binding free energy method would reduce the usefulness of CADD. The modified linear interaction energy (LIE) approach emerges as an appropriate protocol that performs this task. In particular, the polar interaction free energy, which is obtained via numerically resolving the linear Poisson–Boltzmann equation, plays as an important role in driving the binding mechanism of the HIV-1 PR + inhibitor complex. The electrostatic interaction energy contributes to the attraction between two molecules, but the vdW interaction acts as a repulsive factor between the ligand and the HIV-1 PR. Moreover, the ligands were found to adopt a very strong hydrophobic interaction with the HIV-1 PR. Furthermore, the results obtained corroborate the high accuracy and precision of computational studies with a large correlation coefficient value R = 0.83 and a small RMSE δ_RMSE = 1.25 kcal mol⁻¹. This method is less time-consuming than the other end-point methods, such as the molecular mechanics Poisson–Boltzmann surface area (MM/PBSA) and free energy perturbation (FEP) approaches. Overall, the modified LIE approach would provide ligand-binding affinity with HIV-1 PR accurately, precisely, and rapidly, resulting in a more efficient design of new inhibitors.

The inhibition of the Human Immunodeficiency Virus Type 1 Protease (HIV-1 PR) can prevent the synthesis of new viruses.

Study of SHMT2 Inhibitors and Their Binding Mechanism by Computational Alanine Scanning

Mitochondrial serine hydroxymethyl transferase isoform 2 (SHMT2) has attracted increasing attention as a pivotal catalyzing regulator of the serine/glycine pathway in the one-carbon metabolism of cancer cells. However, few inhibitors that target this potential anticancer target have been discovered. Quantitative characterization of the interactions between SHMT2 and its known inhibitors should benefit future discovery of novel inhibitors. In this study, we employed a recently developed alanine-scanning-interaction-entropy method to quantitatively calculate the residue-specific binding free energy of 28 different SHMT2 inhibitors that originate from the same skeleton. Major contributing residues from SHMT2 and chemical groups from the inhibitors were identified, and the binding energy of each residue was quantitatively determined, revealing essential features of the protein-inhibitor interaction. The most important contributing residue is Y105 of the B chain followed by L166 of the A chain. The calculated protein-ligand binding free energies are in good agreement with the experimental results and showed better correlation and smaller errors compared with those obtained using the conventional MM/GBSA with the normal mode method. These results may aid the rational design of more effective SHMT2 inhibitors.

DLIGAND2: an improved knowledge-based energy function for protein-ligand interactions using the distance-scaled, finite, ideal-gas reference state

Performance of structure-based molecular docking largely depends on the accuracy of scoring functions. One important type of scoring functions are knowledge-based potentials derived from known three-dimensional structures of proteins and/or protein–ligand complex structures. This study seeks to improve a knowledge-based protein–ligand potential based on a distance-scale finite ideal-gas reference (DFIRE) state (DLIGAND) by expanding the representation of protein atoms from 13 mol2 atom types to 167 residue-specific atom types, and employing a recently updated dataset containing 12,450 monomer protein chains for training. We found that the updated version DLIGAND2 has a consistent improvement over DLIGAND in predicting binding affinities for either native complex structures or docking-generated poses. More importantly, DLIGAND2 has a 52% increase over DLIGAND in enrichment factors in top 1% predictions based on the DUD-E decoy set, and consistently improves over Autodock Vina and other statistical energy functions in all three benchmark tests. We further found that DLIGAND2 outperforms empirical and machine-learning methods compared for virtual screening on new targets that are not homologous to the DUD-E training set. Given the best performance as a parameter-free statistical potential and among the best in all performance measures, DLIGAND2 should be useful for re-assessing the poses generated by docking software, or acting as one term in other scoring functions. The program is available at https://github.com/sysu-yanglab/DLIGAND2.

Adequate prediction for inhibitor affinity of Aβ40 protofibril using the linear interaction energy method

The search for efficient inhibitors targeting Aβ oligomers and fibrils is an important issue in Alzheimer's disease treatment. As a consequence, an accurate and computationally cheap approach to estimate the binding affinity for many ligands interacting with Aβ peptides is very important. Here, the calculated binding free energies of 30 ligands interacting with 12Aβ11-40 peptides using the linear interaction energy (LIE) approach are found to be in good correlation with experimental data (R = 0.79). The binding affinities of these complexes are also calculated by using free energy perturbation (FEP) and molecular mechanic/Poisson-Boltzmann surface area (MM/PBSA) methods. The time-consuming FEP method provides results with similar correlation (R = 0.72), whereas MM/PBSA calculations show very low correlation with experimental data (R = 0.27). In all complexes, van der Waals interactions contribute much more than electrostatic interactions. The LIE model, which is much less time-consuming than both the FEP and MM/PBSA methods, opens the door to accurate and rapid affinity prediction of ligands with Aβ peptides and the design of new ligands.

Structural Basis of Inhibition of Human Insulin-Regulated Aminopeptidase (IRAP) by Aryl Sulfonamides

The insulin-regulated aminopeptidase (IRAP) is a membrane-bound zinc metallopeptidase with many important regulatory functions. It has been demonstrated that inhibition of IRAP by angiotensin IV (Ang IV) and other peptides, as well as more druglike inhibitors, improves cognition in several rodent models. We recently reported a series of aryl sulfonamides as small-molecule IRAP inhibitors and a promising scaffold for pharmacological intervention. We have now expanded with a number of derivatives, report their stability in liver microsomes, and characterize the activity of the whole series in a new assay performed on recombinant human IRAP. Several compounds, such as the new fluorinated derivative 29, present submicromolar affinity and high metabolic stability. Starting from the two binding modes previously proposed for the sulfonamide scaffold, we systematically performed molecular dynamics simulations and binding affinity estimation with the linear interaction energy method for the full compound series. The significant agreement with experimental affinities suggests one of the binding modes, which was further confirmed by the excellent correlation for binding affinity differences between the selected pair of compounds obtained by rigorous free energy perturbation calculations. The new experimental data and the computationally derived structure-activity relationship of the sulfonamide series provide valuable information for further lead optimization of novel IRAP inhibitors.

Identification of a less toxic vinca alkaloid derivative for use as a chemotherapeutic agent, based on in silico structural insights and metabolic interactions with CYP3A4 and CYP3A5

Vinca alkaloids are chemotherapeutic agents used in the treatment of both pediatric and adult cancer patients. Cytochrome P450 3A5 (CYP3A5) is 9- to 14-fold more efficient at clearing vincristine than cytochrome P450 3A4 (CYP3A4) is. However, patients who express an inactive form of the polymorphic CYP3A5 enzyme suffer from severe neurotoxicity during vincristine treatment, resulting in chemotherapy failure. Previous studies have found that the addition of new features to the parent drug can enhance its binding affinity to tubulin manyfold and could therefore yield novel anticancer drugs. However, there is no report of any study of the metabolic activities of CYP3A4 and CYP3A5 with respect to vincristine and vinblastine, so we studied the interactions of these two drugs and 15 vinca derivatives with CYP3A4 and CYP3A5 by performing docking studies using GOLD. Six of the vinca derivatives in complexes with CYP3A4 and CYP3A5 were further investigated in 100-ns molecular dynamic simulations. Interaction energies, hydrogen bonds, and linear interaction energies were calculated and principal component analysis was carried out to visualize the binding interface in each complex. The results indicate that the addition of dimethylurea at the C20' position in vincristine may increase its binding affinity and lead to enhanced interactions with the less polymorphic CYP3A4 rather than CYP3A5. Thus, dimethylurea vincristine may be a useful drug in cancer chemotherapy treatment as it should be significantly less likely than vincristine to induce severe neurotoxicity in patients. Graphical Abstract Proposed modification of Vinca alkaloid derivatives to decrease the neurotoxicity level in cancer patients exhibiting CYP3A4 gene rather than polymorphic CYP3A5 gene.

Probing the Time Dependency of Cyclooxygenase-1 Inhibitors by Computer Simulations

Time-dependent inhibition of the cyclooxygenases (COX) by a range of nonsteroidal anti-inflammatory drugs has been described since the first experimental assays of COX were performed. Slow tight-binding inhibitors of COX-1 bind in a two-step mechanism in which the EI → EI* transition is slow and practically irreversible. Since then, various properties of the inhibitors have been proposed to cause or affect the time dependency. Conformational changes in the enzyme have also been proposed to cause the time dependency, but no particular structural feature has been identified. Here, we investigated a series of inhibitors of COX-1 that are either time-independent or time-dependent using a combination of molecular dynamics simulations, binding free energy calculations, and potential of mean force calculations. We find that the time-dependent inhibitors stabilize a conformational change in the enzyme mainly identified by the rotation of a leucine side chain adjacent to the binding pocket. The induced conformation has been previously shown to be essential for the high binding affinities of tight-binding inhibitors in COX-1. The results of this work show that the structural features of the enzyme involved in both time-dependent and tight-binding inhibition are identical and further identify a structural mechanism responsible for the transition between the two enzyme-inhibitor complexes characteristic of slow tight-binding COX-1 inhibitors.

Aryl Sulfonamide Inhibitors of Insulin-Regulated Aminopeptidase Enhance Spine Density in Primary Hippocampal Neuron Cultures

The zinc metallopeptidase insulin regulated aminopeptidase (IRAP), which is highly expressed in the hippocampus and other brain regions associated with cognitive function, has been identified as a high-affinity binding site of the hexapeptide angiotensin IV (Ang IV). This hexapeptide is thought to facilitate learning and memory by binding to the catalytic site of IRAP to inhibit its enzymatic activity. In support of this hypothesis, low molecular weight, nonpeptide specific inhibitors of IRAP have been shown to enhance memory in rodent models. Recently, it was demonstrated that linear and macrocyclic Ang IV-derived peptides can alter the shape and increase the number of dendritic spines in hippocampal cultures, properties associated with enhanced cognitive performance. After screening a library of 10 500 drug-like substances for their ability to inhibit IRAP, we identified a series of low molecular weight aryl sulfonamides, which exhibit no structural similarity to Ang IV, as moderately potent IRAP inhibitors. A structural and biological characterization of three of these aryl sulfonamides was performed. Their binding modes to human IRAP were explored by docking calculations combined with molecular dynamics simulations and binding affinity estimations using the linear interaction energy method. Two alternative binding modes emerged from this analysis, both of which correctly rank the ligands according to their experimental binding affinities for this series of compounds. Finally, we show that two of these drug-like IRAP inhibitors can alter dendritic spine morphology and increase spine density in primary cultures of hippocampal neurons.

Improved QM/MM Linear-Interaction Energy Model for Substrate Recognition in Zinc-Containing Metalloenzymes

One of the essential challenges in the description of receptor-drug interactions in the presence of various polyvalent cations (such as zinc, magnesium, or iron) is the accurate assessment of the electronic effects due to cofactor binding. The effects can range from partial electronic polarization of the proximal atoms in a receptor and bound substrate to long-range effects related to partial charge transfer and electronic delocalization effects between the cofactor and the drug. Here, we examine the role of the explicit account for electronic effects for a panel of small-molecule inhibitors binding to the zinc-aminopeptidase PfA-M1, an essential target for antimalarial drug development. Our study on PfA-M1:inhibitor interactions at the QM level reveals that the partial charge and proton transfer due to bound zinc ion are important mechanisms in the inhibitors' recognition and catalysis. The combination of classical MD simulations with a posteriori QM/MM corrections with novel DFTB parameters for the zinc cation and the linear-interaction energy (LIE) approach offers by far the most accurate estimates for the PfA-M1:inhibitor binding affinities, opening the door for future inhibitor design.

Binding to and Inhibition of Insulin-Regulated Aminopeptidase by Macrocyclic Disulfides Enhances Spine Density

Angiotensin IV (Ang IV) and related peptide analogs, as well as nonpeptide inhibitors of insulin-regulated aminopeptidase (IRAP), have previously been shown to enhance memory and cognition in animal models. Furthermore, the endogenous IRAP substrates oxytocin and vasopressin are known to facilitate learning and memory. In this study, the two recently synthesized 13-membered macrocyclic competitive IRAP inhibitors HA08 and HA09, which were designed to mimic the N terminus of oxytocin and vasopressin, were assessed and compared based on their ability to bind to the IRAP active site, and alter dendritic spine density in rat hippocampal primary cultures. The binding modes of the IRAP inhibitors HA08, HA09, and of Ang IV in either the extended or γ-turn conformation at the C terminus to human IRAP were predicted by docking and molecular dynamics simulations. The binding free energies calculated with the linear interaction energy method, which are in excellent agreement with experimental data and simulations, have been used to explain the differences in activities of the IRAP inhibitors, both of which are structurally very similar, but differ only with regard to one stereogenic center. In addition, we show that HA08, which is 100-fold more potent than the epimer HA09, can enhance dendritic spine number and alter morphology, a process associated with memory facilitation. Therefore, HA08, one of the most potent IRAP inhibitors known today, may serve as a suitable starting point for medicinal chemistry programs aided by MD simulations aimed at discovering more drug-like cognitive enhancers acting via augmenting synaptic plasticity.

Improving the Accuracy of the Linear Interaction Energy Method for Solvation Free Energies

A linear response method for estimating the free energy of solvation is presented and validated using explicit solvent molecular dynamics, thermodynamic perturbation calculations, and experimental data. The electrostatic contribution to the solvation free energy is calculated using a linear response estimate, which is obtained by comparison to the free energy calculated using thermodynamic perturbation. Systematic deviations from the value of (1)/2 in the potential energy scaling factor are observed for some types of compounds, and these are taken into account by introducing specific coefficients for different chemical groups. The derived model reduces the rms error of the linear response estimate significantly from 1.6 to 0.3 kcal/mol on a training set of 221 molecules used to parametrize the model and from 3.7 to 1.3 kcal/mol on a test set of 355 molecules that were not used in the derivation of the model. The total solvation free energy is estimated by combining the derived model with an empirical size dependent term for predicting the nonpolar contribution. Using this model, the experimental hydration free energies for 192 molecules are reproduced with an rms error of 1.1 kcal/mol. The use of LIE in simplified binding free energy calculations to predict protein-ligand binding free energies is also discussed.

Molecular simulations study of novel 1,4-dihydropyridines derivatives with a high selectivity for Cav3.1 calcium channel

1,4-Dihydropyridines (DHPs) have been developed to treat hypertension, angina, and nerve system disease. They are thought to mainly target the L-type calcium channels, but low selectivity prompts them to block Cav1.2 and Cav3.1 channels simultaneously. Recently, some novel DHPs with different hydrophobic groups have been synthesized and among them M12 has a higher selectivity for Cav3.1. However, the structural information about Cav3.1-DHPs complexes is not available in the experiment. Thus, we combined homology modeling, molecular docking, molecular dynamics simulations, and binding free energy calculations to quantitatively elucidate the inhibition mechanism of DHPs. The calculated results indicate that our model is in excellent agreement with experimental results. On the basis of conformational analysis, we identify the main interactions between DHPs and calcium channels and further elaborate on the different selectivity of ligands from the micro perspective. In conjunction with energy distribution, we propose that the binding sites of Cav3.1-DHPs is characterized by several interspersed hydrophobic amino acid residues on the IIIS6 and IVS6 segments. We also speculate the favorable function groups on prospective DHPs. Besides, our model provides important information for further mutagenesis experiments.

Identification of chikungunya virus nsP2 protease inhibitors using structure-base approaches

The nsP2 protease of chikungunya virus (CHIKV) is one of the essential components of viral replication and it plays a crucial role in the cleavage of polyprotein precursors for the viral replication process. Therefore, it is gaining attention as a potential drug design target against CHIKV. Based on the recently determined crystal structure of the nsP2 protease of CHIKV, this study identified potential inhibitors of the virus using structure-based approaches with a combination of molecular docking, virtual screening and molecular dynamics (MD) simulations. The top hit compounds from database searching, using the NCI Diversity Set II, with targeting at five potential binding sites of the nsP2 protease, were identified by blind dockings and focused dockings. These complexes were then subjected to MD simulations to investigate the stability and flexibility of the complexes and to gain a more detailed insight into the interactions between the compounds and the enzyme. The hydrogen bonds and hydrophobic contacts were characterized for the complexes. Through structural alignment, the catalytic residues Cys1013 and His1083 were identified in the N-terminal region of the nsP2 protease. The absolute binding free energies were estimated by the linear interaction energy approach and compared with the binding affinities predicted with docking. The results provide valuable information for the development of inhibitors for CHIKV.

Protein modeling and molecular dynamics simulation of the two novel surfactant proteins SP-G and SP-H

Surfactant proteins are well known from the human lung where they are responsible for the stability and flexibility of the pulmonary surfactant system. They are able to influence the surface tension of the gas–liquid interface specifically by directly interacting with single lipids. This work describes the generation of reliable protein structure models to support the experimental characterization of two novel putative surfactant proteins called SP-G and SP-H. The obtained protein models were complemented by predicted posttranslational modifications and placed in a lipid model system mimicking the pulmonary surface. Molecular dynamics simulations of these protein-lipid systems showed the stability of the protein models and the formation of interactions between protein surface and lipid head groups on an atomic scale. Thereby, interaction interface and strength seem to be dependent on orientation and posttranslational modification of the protein. The here presented modeling was fundamental for experimental localization studies and the simulations showed that SP-G and SP-H are theoretically able to interact with lipid systems and thus are members of the surfactant protein family.

Ligand-receptor affinities computed by an adapted linear interaction model for continuum electrostatics and by protein conformational averaging

Accurate calculations of free energies involved in small-molecule binding to a receptor are challenging. Interactions between ligand, receptor, and solvent molecules have to be described precisely, and a large number of conformational microstates has to be sampled, particularly for ligand binding to a flexible protein. Linear interaction energy models are computationally efficient methods that have found considerable success in the prediction of binding free energies. Here, we parametrize a linear interaction model for implicit solvation with coefficients adapted by ligand and binding site relative polarities in order to predict ligand binding free energies. Results obtained for a diverse series of ligands suggest that the model has good predictive power and transferability. We also apply implicit ligand theory and propose approximations to average contributions of multiple ligand-receptor poses built from a protein conformational ensemble and find that exponential averages require proper energy discrimination between plausible binding poses and false-positives (i.e., decoys). The linear interaction model and the averaging procedures presented can be applied independently of each other and of the method used to obtain the receptor structural representation.

Reranking docking poses using molecular simulations and approximate free energy methods

Fast and accurate identification of active compounds is essential for effective use of virtual screening workflows. Here, we have compared the ligand-ranking efficiency of the linear interaction energy (LIE) method against standard docking approaches. Using a trypsin set of 1549 compounds, we performed 12,250 molecular dynamics simulations. The LIE method proved effective but did not yield results significantly better than those obtained with docking codes. The entire database of simulations is released.

Toward an optimal docking and free energy calculation scheme in ligand design with application to COX-1 inhibitors

Cyclooxygenase-1 (COX-1) is one of the main targets of most pain-relieving pharmaceuticals. Although the enzyme is well characterized, it is known to be a difficult target for automated molecular docking and scoring. We collected from the literature a structurally diverse set of 45 nonsteroidal anti-inflammatory drugs (NSAIDs) and COX-2-selective inhibitors (coxibs) with a wide range of binding affinities for COX-1. The binding of this data set to a homology model of human COX-1 was analyzed with different combinations of molecular docking algorithms, scoring functions, and the linear interaction energy (LIE) method for estimating binding affinities. It is found that the computational protocols for estimation of binding affinities are extremely sensitive to the initial orientations of the ligands in the binding pocket. To overcome this limitation, we propose a systematic exploration of docking poses using the LIE calculations as a postscoring function. This scheme yields predictions in excellent agreement with experiment, with a mean unsigned error of 0.9 kcal/mol for binding free energies and structures of high quality. A significant improvement of the results is also seen when averaging over experimental data from several independent measurements.

Accessing the applicability of polarized protein-specific charge in linear interaction energy analysis

The reliability of the linear interaction energy (LIE) depends on the atomic charge model used to delineate the Coulomb interaction between the ligand and its environment. In this work, the polarized protein-specific charge (PPC) implementing a recently proposed fitting scheme has been examined in the LIE calculations of the binding affinities for avidin and β-secretase binding complexes. This charge fitting scheme, termed delta restrained electrostatic potential, bypasses the prevalent numerical difficulty of rank deficiency in electrostatic-potential-based charge fitting methods via a dual-step fitting strategy. A remarkable consistency between the predicted binding affinities and the experimental measurement has been observed. This work serves as a direct evidence of PPC's applicability in rational drug design.

Effects of water models on binding affinity: evidence from all-atom simulation of binding of tamiflu to A/H5N1 neuraminidase

The influence of water models SPC, SPC/E, TIP3P, and TIP4P on ligand binding affinity is examined by calculating the binding free energy ΔG_bind of oseltamivir carboxylate (Tamiflu) to the wild type of glycoprotein neuraminidase from the pandemic A/H5N1 virus. ΔG_bind is estimated by the Molecular Mechanic-Poisson Boltzmann Surface Area method and all-atom simulations with different combinations of these aqueous models and four force fields AMBER99SB, CHARMM27, GROMOS96 43a1, and OPLS-AA/L. It is shown that there is no correlation between the binding free energy and the water density in the binding pocket in CHARMM. However, for three remaining force fields ΔG_bind decays with increase of water density. SPC/E provides the lowest binding free energy for any force field, while the water effect is the most pronounced in CHARMM. In agreement with the popular GROMACS recommendation, the binding score obtained by combinations of AMBER-TIP3P, OPLS-TIP4P, and GROMOS-SPC is the most relevant to the experiments. For wild-type neuraminidase we have found that SPC is more suitable for CHARMM than TIP3P recommended by GROMACS for studying ligand binding. However, our study for three of its mutants reveals that TIP3P is presumably the best choice for CHARMM.