Structural basis for selective inhibition of purine nucleoside phosphorylase from Schistosoma mansoni: Kinetic and structural studies

Selectivity plays a crucial role in the design of enzyme inhibitors as novel antiparasitic agents, particularly in cases where the target enzyme is also present in the human host. Purine nucleoside phosphorylase from Schistosoma mansoni (SmPNP) is an attractive target for the discovery of potential antischistosomal agents. In the present work, kinetic studies were carried out in order to determine the inhibitory potency, mode of action and enzyme selectivity of a series of inhibitors of SmPNP. In addition, crystallographic studies provided important structural insights for rational inhibitor design, revealing consistent structural differences in the binding mode of the inhibitors in the active sites of the SmPNP and human PNP (HsPNP) structures. The molecular information gathered in this work should be useful for future medicinal chemistry efforts in the design of new inhibitors of SmPNP having increased affinity and selectivity.     

Molecular modeling and dynamics studies of purine nucleoside phosphorylase from Bacteroides fragilis

Purine nucleoside phosphorylase (PNP) catalyzes the reversible phosphorolysis of N-ribosidic bonds of purine nucleosides and deoxynucleosides, except adenosine, to generate ribose 1-phosphate and the purine base. This work describes for the first time a structural model of PNP from Bacteroides fragilis (Bf). We modeled the complexes of BfPNP with six different ligands in order to determine the structural basis for specificity of these ligands against BfPNP. Comparative analysis of the model of BfPNP and the structure of HsPNP allowed identification of structural features responsible for differences in the computationally determined ligand affinities. The molecular dynamics (MD) simulation was assessed to evaluate the overall stability of the BfPNP model. The superposition of the final onto the initial minimized structure shows that there are no major conformational changes from the initial model, which is consistent with the relatively low root mean square deviation (RMSD). The results indicate that the structure of the model was stable during MD, and does not exhibit loosely structured loop regions or domain terminals.     

Discrete molecular dynamics distinguishes nativelike binding poses from decoys in difficult targets

Virtual screening is one of the major tools used in computer-aided drug discovery. In structure-based virtual screening, the scoring function is critical to identifying the correct docking pose and accurately predicting the binding affinities of compounds. However, the performance of existing scoring functions has been shown to be uneven for different targets, and some important drug targets have proven especially challenging. In these targets, scoring functions cannot accurately identify the native or near-native binding pose of the ligand from among decoy poses, which affects both the accuracy of the binding affinity prediction and the ability of virtual screening to identify true binders in chemical libraries. Here, we present an approach to discriminating native poses from decoys in difficult targets for which several scoring functions failed to correctly identify the native pose. Our approach employs Discrete Molecular Dynamics simulations to incorporate protein-ligand dynamics and the entropic effects of binding. We analyze a collection of poses generated by docking and find that the residence time of the ligand in the native and nativelike binding poses is distinctly longer than that in decoy poses. This finding suggests that molecular simulations offer a unique approach to distinguishing the native (or nativelike) binding pose from decoy poses that cannot be distinguished using scoring functions that evaluate static structures. The success of our method emphasizes the importance of protein-ligand dynamics in the accurate determination of the binding pose, an aspect that is not addressed in typical docking and scoring protocols.     

Structure based inhibitor design targeting glycogen phosphorylase b. Virtual screening,synthesis, biochemical and biological assessment of novel N-acyl-β-d-glucopyranosylamines

Glycogen phosphorylase (GP) is a validated target for the development of new type 2 diabetes treatments. Exploiting the Zinc docking database, we report the in silico screening of 1888 N-acyl-β-d-glucopyranosylamines putative GP inhibitors differing only in their R groups. CombiGlide and GOLD docking programs with different scoring functions were employed with the best performing methods combined in a ‘consensus scoring’ approach to ranking of ligand binding affinities for the active site. Six selected candidates from the screening were then synthesized and their inhibitory potency was assessed both in vitro and ex vivo. Their inhibition constants’ values, in vitro, ranged from 5 to 377 μM while two of them were effective at causing inactivation of GP in rat hepatocytes at low μM concentrations. The crystal structures of GP in complex with the inhibitors were defined and provided the structural basis for their inhibitory potency and data for further structure based design of more potent inhibitors.     

An In Silico Analysis of the Binding Modes and Binding Affinities of Small Molecule Modulators of PDZ-Peptide Interactions

Inhibitors of PDZ-peptide interactions have important implications in a variety of biological processes including treatment of cancer and Parkinson’s disease. Even though experimental studies have reported characterization of peptidomimetic inhibitors of PDZ-peptide interactions, the binding modes for most of them have not been characterized by structural studies. In this study we have attempted to understand the structural basis of the small molecule-PDZ interactions by in silico analysis of the binding modes and binding affinities of a set of 38 small molecules with known K_i or K_d values for PDZ2 and PDZ3 domains of PSD-95 protein. These two PDZ domains show differential selectivity for these compounds despite having a high degree of sequence similarity and almost identical peptide binding pockets. Optimum binding modes for these ligands for PDZ2 and PDZ3 domains were identified by using a novel combination of semi-flexible docking and explicit solvent molecular dynamics (MD) simulations. Analysis of the binding modes revealed most of the peptidomimectic ligands which had high K_i or K_d moved away from the peptide binding pocket, while ligands with high binding affinities remained in the peptide binding pocket. The differential specificities of the PDZ2 and PDZ3 domains primarily arise from differences in the conformation of the loop connecting βB and βC strands, because this loop interacts with the N-terminal chemical moieties of the ligands. We have also computed the MM/PBSA binding free energy values for these 38 compounds with both the PDZ domains from multiple 5 ns MD trajectories on each complex i.e. a total of 228 MD trajectories of 5 ns length each. Interestingly, computational binding free energies show good agreement with experimental binding free energies with a correlation coefficient of approximately 0.6. Thus our study demonstrates that combined use of docking and MD simulations can help in identification of potent inhibitors of PDZ-peptide complexes.     

Asymmetric structure and domain binding interfaces of human tyrosyl‐tRNA synthetase studied by molecular dynamics simulations

Human tyrosyl‐tRNA synthetase (HsTyrRS) is composed of two structural modules: N‐terminal catalytic core and an EMAP II‐like C‐terminal domain. The structures of these modules are known, but no crystal structure of the full‐length HsTyrRS is currently available. An all‐atom model of the full‐length HsTyrRS was developed in this work. The structure, dynamics, and domain binding interfaces of HsTyrRS were investigated by extensive molecular dynamics (MD) simulations. Our data suggest that HsTyrRS in solution consists of a number of compact asymmetric conformations, which differ significantly by their rigidity, internal mobility, orientation of C‐terminal modules, and the strength of interdomain binding. Interfaces of domain binding obtained in MD simulations are in perfect agreement with our previous coarse‐grained hierarchical rotations technique simulations. Formation of the hydrogen bonds between R93 residue of the ELR cytokine motif and the residues A340 and E479 in the C‐module was observed. This observation supports the idea that the lack of cytokine activity in the full‐length HsTyrRS is explained by interactions between N‐modules and C‐modules, which block the ELR motif. Copyright © 2013 John Wiley & Sons, Ltd.     

Anesthetic Binding in a Pentameric Ligand-Gated Ion Channel: GLIC

Cys-loop receptors are molecular targets of general anesthetics, but the knowledge of anesthetic binding to these proteins remains limited. Here we investigate anesthetic binding to the bacterial Gloeobacter violaceus pentameric ligand-gated ion channel (GLIC), a structural homolog of cys-loop receptors, using an experimental and computational hybrid approach. Tryptophan fluorescence quenching experiments showed halothane and thiopental binding at three tryptophan-associated sites in the extracellular (EC) domain, transmembrane (TM) domain, and EC-TM interface of GLIC. An additional binding site at the EC-TM interface was predicted by docking analysis and validated by quenching experiments on the N200W GLIC mutant. The binding affinities (K_D) of 2.3 ± 0.1 mM and 0.10 ± 0.01 mM were derived from the fluorescence quenching data of halothane and thiopental, respectively. Docking these anesthetics to the original GLIC crystal structure and the structures relaxed by molecular dynamics simulations revealed intrasubunit sites for most halothane binding and intersubunit sites for thiopental binding. Tryptophans were within reach of both intra- and intersubunit binding sites. Multiple molecular dynamics simulations on GLIC in the presence of halothane at different sites suggested that anesthetic binding at the EC-TM interface disrupted the critical interactions for channel gating, altered motion of the TM23 linker, and destabilized the open-channel conformation that can lead to inhibition of GLIC channel current. The study has not only provided insights into anesthetic binding in GLIC, but also demonstrated a successful fusion of experiments and computations for understanding anesthetic actions in complex proteins.     

Thermodynamic computational approach to capture molecular recognition in the binding of different inhibitors to the DNA gyrase B subunit from Escherichia coli

DNA gyrase subunit B, that catalyzes the hydrolysis of ATP, is an attractive target for the development of antibacterial drugs. This work is intended to rationalize molecular recognition at DNA gyrase B enzyme – inhibitor binding interface through the evaluation of different scoring functions in finding the correct pose and scoring properly 50 Escherichia coli DNA Gyrase B inhibitors belonging to five different classes. Improving the binding free energy calculation accuracy is further attempted by using rescoring schemes after short molecular dynamic simulations of the obtained docked complexes. These data are then compared with the corresponding experimental enzyme activity data. The results are analyzed from a structural point of view emphasizing the strengths and limitations of the techniques applied in the study.     

Crystal structure of HsEg5 in complex with clinical candidate CK0238273 provides insight into inhibitory mechanism, potency, and specificity

HsEg5 is an important mitotic kinesin responsible for bipolar spindle formation at early mitosis. A rich body of evidence shows that inhibition of HsEg5 can result in mitotic arrest followed by cellular apoptosis. Recently identified HsEg5 inhibitor, CK0238273, exhibits potent antitumor activity and is currently in clinical trial. Here we report the cocrystal structure of the motor domain of HsEg5 in complex with CK0238273 at a 2.15 Å resolution. Compared to the previously published HsEg5-Monastrol complex structure, CK0238273 shares the same induced-fit pocket with similar allosteric inhibitory mechanism. However, CK0238273 shows better fitting to the binding pocket with 65% increase of hydrophobic interaction area than that of Monastrol. Some unique hydrophilic interactions were also observed mostly between the phenyl ring and 8-chloro on quinazolinone of CK0238273 with ARG221 and GLY217. We believe that the combination of these interactions defines the superior potency and specificity of CK0238273.     

Virtual screening to identify novel potential inhibitors for Glutamine synthetase of Mycobacterium tuberculosis

Abstract

Glutamine synthetase (GS) of Mycobacterium tuberculosis (Mtb) is an essential enzyme which is involved in nitrogen metabolism and cell wall synthesis. It is involved in the inhibition of phagosome-lysosome fusion by preventing acidification. Targeting GS can be helpful to control the infection of Mtb. In order to identify potential inhibitors, we screened chemical libraries (56,400 compounds of ChEMBL anti-mycobacterial, 1596 FDA approved drugs, 419 Natural product and 916 phytochemical) against this target using the virtual screening approach. Screening by molecular docking identified ten top-ranked compounds as GSMtb inhibitors and they were compared with known inhibitors (as control). Since GS enzyme (GSHs) is also present in human. We have compared the protein sequence of GS from Mtb and human using the P-BLAST in NCBI. We found ～27% identity in between these two sequences, so we also compared the binding affinity of inhibitor between Mtb and human. Finally, we identified top two compounds namely CHEMBL387509, CHEMBL226198 from ChEMBL anti-mycobacterial dataset, and Eriocitrin and Malvidin from phytochemical dataset which showed lees binding affinity towards GSHs whereas Pamidronate, and Phentermine from FDA approved drugs and (-)-Quinic Acid, Hexopyranuronic acid, Quebrachit, and Castanospermine from natural product showed protein-ligand interaction with Mtb protein while no interaction with GSHs. The top two docked complexes were subjected to molecular dynamic simulation to understand the stability of the molecule. Further, we calculated the binding free energy of the docked complex and analyzed hydrogen bond, salt bridge, pie stacking, and hydrophobic interaction in the docking region. These ligands exhibited very good binding affinity GSMtb enzymes. Therefore, these ligands are novel and drug-likeness compounds, and they may be potential inhibitors of M tuberculosis.

Communicated by Ramaswamy H. Sarma     

Molecular modeling study for conformational changes of Sirtuin 2 due to substrate and inhibitor binding

Sirtuin is a member of NAD⁺-dependent deacetylase family. The structural details of Sirtuin 2 (SIRT2) complex will be very useful to discover the drug which might have beneficial effects on various diseases like cancer, diabetes, etc. Unfortunately, SIRT2 complex structure is not available yet, hence molecular docking was carried out to dock the substrate (NAD⁺ and acetylated lysine) and inhibitor (sirtinol) in the NAD⁺ binding site. The suitable binding orientation of substrate and inhibitor in the SIRT2 active site was selected and subjected to 5?ns molecular dynamics simulations to adjust the binding orientation of inhibitor and substrate as well as to identify the conformational changes in the active site. The result provides an insight about 3D SIRT2 structural details as well as the importance of F96 in deacetylation function. In addition, our simulations revealed the displacement of F96 upon substrate and inhibitor binding, inducing an extended conformation of loop3 and changing its interactions with the rest of SIRT2. We believe that our study could be helpful to gain a structural insight of SIRT2 and to design the receptor-based inhibitors.     

Substituting random forest for multiple linear regression improves binding affinity prediction of scoring functions: Cyscore as a case study

Background

State-of-the-art protein-ligand docking methods are generally limited by the traditionally low accuracy of their scoring functions, which are used to predict binding affinity and thus vital for discriminating between active and inactive compounds. Despite intensive research over the years, classical scoring functions have reached a plateau in their predictive performance. These assume a predetermined additive functional form for some sophisticated numerical features, and use standard multivariate linear regression (MLR) on experimental data to derive the coefficients.

Results

In this study we show that such a simple functional form is detrimental for the prediction performance of a scoring function, and replacing linear regression by machine learning techniques like random forest (RF) can improve prediction performance. We investigate the conditions of applying RF under various contexts and find that given sufficient training samples RF manages to comprehensively capture the non-linearity between structural features and measured binding affinities. Incorporating more structural features and training with more samples can both boost RF performance. In addition, we analyze the importance of structural features to binding affinity prediction using the RF variable importance tool. Lastly, we use Cyscore, a top performing empirical scoring function, as a baseline for comparison study.

Conclusions

Machine-learning scoring functions are fundamentally different from classical scoring functions because the former circumvents the fixed functional form relating structural features with binding affinities. RF, but not MLR, can effectively exploit more structural features and more training samples, leading to higher prediction performance. The future availability of more X-ray crystal structures will further widen the performance gap between RF-based and MLR-based scoring functions. This further stresses the importance of substituting RF for MLR in scoring function development.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2105-15-291) contains supplementary material, which is available to authorized users.     

Transition State Analogues of Plasmodium falciparum and Human Orotate Phosphoribosyltransferases

The survival and proliferation of Plasmodium falciparum parasites and human cancer cells require de novo pyrimidine synthesis to supply RNA and DNA precursors. Orotate phosphoribosyltransferase (OPRT) is an indispensible component in this metabolic pathway and is a target for antimalarials and antitumor drugs. P. falciparum (Pf) and Homo sapiens (Hs) OPRTs are characterized by highly dissociative transition states with ribocation character. On the basis of the geometrical and electrostatic features of the PfOPRT and HsOPRT transition states, analogues were designed, synthesized, and tested as inhibitors. Iminoribitol mimics of the ribocation transition state in linkage to pyrimidine mimics using methylene or ethylene linkers gave dissociation constants (K_d) as low as 80 nm. Inhibitors with pyrrolidine groups as ribocation mimics displayed slightly weaker binding affinities for OPRTs. Interestingly, p-nitrophenyl riboside 5′-phosphate bound to OPRTs with K_d values near 40 nm. Analogues designed with a C5-pyrimidine carbon–carbon bond to ribocation mimics gave K_d values in the range of 80–500 nm. Acyclic inhibitors with achiral serinol groups as the ribocation mimics also displayed nanomolar inhibition against OPRTs. In comparison with the nucleoside derivatives, inhibition constants of their corresponding 5′-phosphorylated transition state analogues are largely unchanged, an unusual property for a nucleotide-binding site. In silico docking of the best inhibitor into the HsOPRT active site supported an extensive hydrogen bond network associated with the tight binding affinity. These OPRT transition state analogues identify crucial components of potent inhibitors targeting OPRT enzymes. Despite their tight binding to the targets, the inhibitors did not kill cultured P. falciparum.     

Structural investigation into the inhibitory mechanisms of indomethacin and its analogues towards human glyoxalase I

In the present work, a combined study of kinetic analysis, molecular docking, and molecular dynamics simulations on indomethacin and its analogues is performed to better understand their inhibitory mechanisms towards human glyoxalase I (GLOI). A remarkable correlation (R² = 0.974) was observed for six inhibitors including indomethacin between their experimental inhibitory affinities and predicted binding free energy parameter (ΔG_bind,pred). This suggests that ΔG_bind,pred of a GLOI/inhibitor complex can be efficiently used to interpolate the experimental inhibitory affinity of a ligand of similar nature in the GLOI enzyme system. Energetic analyses revealed that electrostatic contribution plays an important role in their inhibitory mechanisms, which reflects the significant contribution of the coordination bond between zinc and ligands. The present work highlights that indomethacin is a promising lead as GLOI inhibitors for further development since it may bind all subsites in the active site pocket of GLOI and stabilize the flexible loop (152-159).     

Crystal structure and molecular dynamics studies of human purine nucleoside phosphorylase complexed with 7-deazaguanine

In humans, purine nucleoside phosphorylase (HsPNP) is responsible for degradation of deoxyguanosine, and genetic deficiency of this enzyme leads to profound T-cell mediated immunosuppression. HsPNP is a target for inhibitor development aiming at T-cell immune response modulation. Here we report the crystal structure of HsPNP in complex with 7-deazaguanine (HsPNP:7DG) at 2.75 Å. Molecular dynamics simulations were employed to assess the structural features of HsPNP in both free form and in complex with 7DG. Our results show that some regions, responsible for entrance and exit of substrate, present a conformational variability, which is dissected by dynamics simulation analysis. Enzymatic assays were also carried out and revealed that 7-deazaguanine presents a lower inhibitory activity against HsPNP (K_i = 200 μM). The present structure may be employed in both structure-based design of PNP inhibitors and in development of specific empirical scoring functions.     

Molecular Structure-Affinity Relationship of Bufadienolides and Human Serum Albumin In Vitro and Molecular Docking Analysis

The development of bufadienolides as anti-tumor agents is limited due to poor pharmacokinetic properties regarding drug half-lives and toxicity in vivo. These serious factors might be improved by increasing the drug/albumin-binding ratio. This study therefore investigated the relationship between the structural properties of nine bufadienolides and their affinities for human serum albumin (HSA) by a fluorescence spectroscopy-based analysis and molecular docking. Fluorescence quenching data showed that the interaction of each bufadienolide with HSA formed a non-fluorescent complex, while thermodynamic parameters revealed negative ΔS and ΔH values, corresponding to changes in enthalpy and entropy, respectively. The structural differences between the various bufadienolides markedly influenced their binding affinity for HSA. With the exception of a C = O bond at the C12 position that decreased the binding affinity for HSA, other polar groups tended to increase the affinity, especially a hydroxyl (OH) group at assorted bufadienolide sites. The rank order of binding affinities for drugs with tri-hydroxyl groups was as follows: 11-OH > 5-OH > 16-OH; in addition, 16-acetoxy (OAc), 10-aldehyde and 14-epoxy constituents notably enhanced the binding affinity. Among these groups, 11-OH and 16-acetyl were especially important for a seamless interaction between the bufadienolides and HSA. Furthermore, molecular docking analysis revealed that either an 11-OH or a 16-OAc group spatially close to a five-membered lactone ring significantly facilitated the anchoring of these compounds within site I of the HSA pocket via hydrogen bonding (H-bonding) with Tyr150 or Lys199, respectively. In summary, bufadienolide structure strongly affects binding with HSA, and 11-OH or 16-OAc groups improve the drug association with key amino acid residues. This information is valuable for the prospective development of bufadienolides with improved pharmacological profiles as novel anti-tumor drugs.     

Modeling of protein binary complexes using structural mass spectrometry data

In this article, we describe a general approach to modeling the structure of binary protein complexes using structural mass spectrometry data combined with molecular docking. In the first step, hydroxyl radical mediated oxidative protein footprinting is used to identify residues that experience conformational reorganization due to binding or participate in the binding interface. In the second step, a three-dimensional atomic structure of the complex is derived by computational modeling. Homology modeling approaches are used to define the structures of the individual proteins if footprinting detects significant conformational reorganization as a function of complex formation. A three-dimensional model of the complex is constructed from these binary partners using the ClusPro program, which is composed of docking, energy filtering, and clustering steps. Footprinting data are used to incorporate constraints-positive and/or negative-in the docking step and are also used to decide the type of energy filter-electrostatics or desolvation-in the successive energy-filtering step. By using this approach, we examine the structure of a number of binary complexes of monomeric actin and compare the results to crystallographic data. Based on docking alone, a number of competing models with widely varying structures are observed, one of which is likely to agree with crystallographic data. When the docking steps are guided by footprinting data, accurate models emerge as top scoring. We demonstrate this method with the actin/gelsolin segment-1 complex. We also provide a structural model for the actin/cofilin complex using this approach which does not have a crystal or NMR structure.     

A role for the human peroxisomal half-transporter ABCD3 in the oxidation of dicarboxylic acids

Peroxisomes play a major role in human cellular lipid metabolism, including fatty acid β-oxidation. Free fatty acids (FFAs) can enter peroxisomes through passive diffusion or by means of ATP binding cassette (ABC) transporters, including HsABCD1 (ALDP, adrenoleukodystrophy protein), HsABCD2 (ALDRP) and HsABCD3 (PMP70). The physiological functions of the different peroxisomal half-ABCD transporters have not been fully determined yet, but there are clear indications that both HsABCD1 and HsABCD2 are required for the breakdown of fatty acids in peroxisomes. Here we report that the phenotype of the pxa1/pxa2Δ yeast mutant, i.e. impaired oxidation of oleic acid, cannot only be partially rescued by HsABCD1, HsABCD2, but also by HsABCD3, which indicates that each peroxisomal half-transporter can function as homodimer. Fatty acid oxidation measurements using various fatty acids revealed that although the substrate specificities of HsABCD1, HsABCD2 and HsABCD3 are overlapping, they have distinctive preferences. Indeed, most hydrophobic C24:0 and C26:0 fatty acids are preferentially transported by HsABCD1, C22:0 and C22:6 by HsABCD2 and most hydrophilic substrates like long-chain unsaturated-, long branched-chain- and long-chain dicarboxylic fatty acids by HsABCD3. All these fatty acids are most likely transported as CoA esters. We postulate a role for human ABCD3 in the oxidation of dicarboxylic acids and a role in buffering fatty acids that are overflowing from the mitochondrial β-oxidation system.     

Molecular docking approaches in identification of High affinity inhibitors of Human SMO receptor

Inappropriate activation of the Hh signaling pathway has been implicated in the development of several types of cancers including prostate, lung, pancreas, breast, brain and skin. Present study identified the binding affinities of eight established inhibitors viz., Cyclopamine, Saridegib, Itraconazole, LDE-225, TAK-441, BMS-833923 (XL139), PF-04449913 and Vismodegib targeting SMO receptor - a candidate protein involved in hedgehog pathway and sought to identify the best amongst the established inhibitors through by molecular docking. Exelxis® BMS 833923 (XL 139) demonstrated superior binding affinity aided by MolDock scoring docking algorithm. Further BMS 833923 (XL 139) was evaluated for pharmacophoric features which revealed appreciable ligand receptor interactions.     

Receptor-specific scoring functions derived from quantum chemical models improve affinity estimates for in-silico drug discovery

The adaptation of forcefield-based scoring function to specific receptors remains an important challenge for in-silico drug discovery. Here we compare binding energies of forcefield-based scoring functions with models that are reparameterized on the basis of large-scale quantum calculations of the receptor. We compute binding energies of eleven ligands to the human estrogen receptor subtype alpha (ERalpha) and four ligands to the human retinoic acid receptor of isotype gamma (RARgamma). Using the FlexScreen all-atom receptor-ligand docking approach, we compare docking simulations parameterized by quantum-mechanical calculation of a large protein fragment with purely forcefield-based models. The use of receptor flexibility in the FlexScreen permits the treatment of all ligands in the same receptor model. We find a high correlation between the classical binding energy obtained in the docking simulation and quantum mechanical binding energies and a good correlation with experimental affinities R=0.81 for ERalpha and R=0.95 for RARgamma using the quantum derived scoring functions. A significant part of this improvement is retained, when only the receptor is treated with quantum-based parameters, while the ligands are parameterized with a purely classical model.