Two Binding Geometries for Risperidone in Dopamine D3 Receptors

Subscriber access provided by RMIT University Library

Article

Two Binding Geometries for Risperidone in Dopamine D3 Receptors: Insights on the Fast-Off Mechanism through Docking, Quantum Biochemistry and Molecular Dynamics Simulations Geancarlo Zanatta, Gustavo Della Flora Nunes, Eveline Matias Bezerra, Roner F. da Costa, Alice Maria Martins, Ewerton Wagner Santos Caetano, Valder Nogueira Freire, and Carmem Gottfried ACS Chem. Neurosci., Just Accepted Manuscript • DOI: 10.1021/acschemneuro.6b00074 • Publication Date (Web): 19 Jul 2016 Downloaded from http://pubs.acs.org on July 21, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Chemical Neuroscience is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Neuroscience

Two Binding Geometries for Risperidone in Dopamine D3 Receptors: Insights on the Fast-Off Mechanism through Docking, Quantum Biochemistry and Molecular Dynamics Simulations Geancarlo Zanattaa*, Gustavo Della Flora Nunes a, Eveline M. Bezerrab, Roner F. da Costac, Alice Martinsb, Ewerton W. S. Caetanod, Valder N. Freiree, Carmem Gottfrieda a

Department of Biochemistry at Federal University of Rio Grande do Sul, 90035-003 Porto Alegre, RS, Brazil; Post-graduate Program in Pharmaceutical Sciences, Pharmacy Faculty, Federal University of Ceará, 60430-372 Fortaleza, CE, Brazil; c Department of Physics, Universidade Federal Rural do Semi-Árido, 59780-000 Caraúbas, RN, Brazil; d Federal Institute of Education, Science and Technology, 60040-531 Fortaleza, CE, Brazil; e Department of Physics at Federal University of Ceará, 60455-760 Fortaleza, CE, Brazil. b

KEYWORDS: Quantum biochemistry; D3 Dopamine receptor; Risperidone; Fast-Off dissociation; Schizophrenia; DFT; Molecular Dynamics; Antipsychotic; Ab initio; Docking; QM/MM; ONIOM

ABSTRACT: Risperidone is an atypical antipsychotic used in the treatment of schizophrenia and of symptoms of irritability associated with autism spectrum disorder (ASD). Its main action mechanism is the blockade of D2-like receptors acting over positive and negative symptoms of schizophrenia with small risk of extrapyramidal symptoms (EPS) at doses corresponding to low/moderate D2 occupancy. Such decrease in the side effect incidence can be associated to its fast unbinding from D2 receptors in the nigrostriatal region allowing the recovery of dopamine signaling pathways. We performed docking essays using risperidone and the D3 receptor crystallographic data and results suggested two possible distinct orientations for risperidone at the binding pocket. Orientation 1 is more close to the opening of the binding site and has the 6-Fluoro-1,2 benzoxazole fragment toward the bottom of the D3 receptor cleft, while orientation 2 is deeper inside the binding pocket with the same fragment toward to the receptor surface. In order to unveil the implications of these two binding orientations, classical molecular dynamics and quantum biochemistry computations within the density functional theory formalism and the molecular fractionation with conjugate caps framework were performed. Quantum Mechanics/Molecular Mechanics suggests that orientation 2 is slightly more energetically stable than orientation 1 with the main contribution coming from residue Asp110. The residue Glu90, positioned at the opening of the binding site, is closer to orientation 1 than 2, suggesting that it may have a key role in stability through attractive interaction with risperidone. Therefore, although orientations 1 and 2 are both likely to occur, we suggest that the occurrence of the first may contribute to the reduction of side effects in patients taking risperidone due to the reduction of dopamine receptor occupancy in the nigrostriatal region through a mechanism of fast dissociation. The atypical effect may be obtained simply by either delaying D3R full blockage by spatial hindrance of orientation 1 at the binding site or through an effective blockade followed by orientation 1 fast dissociation. While the molecular interpretation suggested in this work shed some light on the potential molecular mechanisms accounting for the reduced extrapyramidal symptoms observed during risperidone treatment, further studies are necessary in order to evaluate the implications of both orientations during the receptor activation/inhibition. Altogether these data highlight important hotspots in the dopamine receptor binding site bringing relevant information for the development of novel/derivative agents with atypical profile.

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

INTRODUCTION Schizophrenia is a neuro-developmental condition affecting more than 21 million people worldwide, representing one of the leading causes of health burden with deep impact on patients, their families and society. 1 2 3 4 Its symptoms domains are classified as positive (e.g. hallucination, paranoid delusions, etc.), negative (e.g. apathy, lack of pleasure, difficulties to begin and sustain activities, social withdrawal, cognitive deficits, etc.), and cognitive (e.g. memory deficiencies, attention and executive functions). 5 Schizophrenia began to be treated with antipsychotics in the 1950s, thanks to the development of phenothiazine compounds. Latter, the mechanism of action of important therapeutic antipsychotics was related with the pharmacological antagonism in dopamine receptors; 6 the blockade of dopamine D2 receptors was identified as the main target of antipsychotic action, giving support to the dopamine hypothesis of schizophrenia. 7 After the introduction of clozapine in 1970s, the development of a new generation of antipsychotic agents with reduced side effects became the focus of research. 8 9 10 The term “atypical” was then introduced to distinguish new compounds from agents with high correlation between dosage and extrapyramidal symptoms (EPS). Clozapine, olanzapine and risperidone are examples of atypical antipsychotics which are believed/demonstrated to be more efficacious, tolerable and associated with fewer EPS. 11 12 Among the new atypical agents, risperidone (3-{2-[4-(6-Fluoro1,2-benzoxazol-3-yl)-1-piperidinyl]ethyl}-2-methyl-6,7,8,9-tetrahydro-4H-pyrido[1,2-a]pyrimidin-4-one) belongs to the chemical class of the benzisoxazole derivatives and has been used for the treatment of positive and negative symptoms of schizophrenia. Risperidone received the approval of the U.S. Food and Drug Administration (FDA) in October of 2006 for its use in the treatment of children and adolescents (age 5 to 16 years) who have schizophrenia or symptoms of irritability associated with the autism spectrum disorder (ASD). 13 14 This agent has been effective in managing behavioral problems and symptoms in children and, except for rapid weight gain, its adverse effects are, in general, manageable or rare. 14 15 Such features make risperidone an important tool agent in the pediatric treatment of these disorders. Moreover, risperidone can be used as a molecular model for the development of new agents with improved efficacy and reduced side effects. Risperidone is known to have serotonin and dopamine antagonistic properties, as it was observed earlier in a comparative study in rats and dogs, together with the reference compounds ritanserin and haloperidol. 16 Nevertheless, although antipsychotics bind to 5-HT2A receptors, 17 18 there are many objection about the effective relevance of this receptor in the treatment of psychosis, as the threshold occupancy of D2R for antipsychotic action (about 65%) or for triggering EPS (about 80%), remains the same, regardless of whether 5-HT2A receptors are blocked or not. 19 Indeed, even though 5-HT2A receptors are readily blocked at low dosages of atypical antipsychotic agents, such blockade is not enough to alleviate psychosis. For this reason, an improved understanding of the dopaminergic neurotransmission is highly important for the advance in the treatment of schizophrenic conditions. Dopamine regulates its pathway through the coordinated activation of D1-like (D1R and D5R) and D2-like (D2R, D3R and D4R) dopamine receptors. 20 21 These receptors are G-protein-coupled receptors, (GPCR, see the Nobel Prize in Chemistry lecture of 2012) 22 23 24 characterized by the presence of seven highly conserved transmembrane helices (TMH1-7), connected by extracellular (EC) and intracellular (IC) loops. 25 26Receptors belonging to the D1-like family activate adenylyl cyclase through the stimulatory G-protein alpha subunit, while receptors of the D2-like fami-

Page 2 of 30

ly are coupled to the G-protein alpha subunit, with inhibitory effect over adenylyl cyclase. 27 The analysis of in vitro and in vivo receptor binding profiles of a set of typical and atypical antipsychotics showed the predominance of D2R versus D3R antagonism (and also 5HT2A over D2R) during risperidone administration. Nevertheless, although those results are not fully conclusive for the in vivo interaction of risperidone with D3R, they clearly indicated that risperidone has also affinity for D3R which is hardly bellow its D2R affinity in both in vitro and in vivo essays. 28 Studies with SPECT and PET scanning reported the occupancies of D2-like structures in the striatum and temporal cortex, the occupancy by atypical agents clozapine, olanzapine, quetiapine and sertindole being greater in cortical regions where typical agents do not act significantly. Such differences in the level of occupancy in striatal dopamine receptors together with other additional features of the agents may account for the partial blockade of receptors in this region, leading to the reduction in EPS events, while the modulation of receptors D2R/D3R in the limbic/cortical area answers for the clinical antipsychotic effect. 9 Among the dopamine receptors, the D3R subtype has being discovered as a potential independent therapeutic target in the treatment of neurological diseases as schizophrenia, Parkinson’s disease 29 30 31, mood disorders 32 and neurocognitive disorders. 33 Such growing attention in D3 as a therapeutic target can be measured by the new 157 patents filled between 2007 and 2012 for novel antagonists. 34 While D2 blockers are associated with a wide range of side effects such as dose-dependent sedation, spontaneous motor behavior and emotional response imbalance, studies in Phase I using the selective D3 antagonist GSK598809 do not show sedation or EPS in the subjects. 35 It was showed that the D3R level was similar to or smaller in medicated schizophrenic patients than control levels, but it was doubled in patients free of antipsychotics, suggesting that D3R is related to the antipsychotic efficacy in the treatment but is not involved with extrapyramidal side effects as the blockade of D2R. 36 Analysis of postmortem tissues showed the distribution of D2R in the dorsal putamen and the dorsal caudate nucleus, but not of D3R. In the ventral putamem and in the ventral caudate and globus pallidus, a D2R:D3R distribution in a 2:1 proportion was observed. 37 The high D2R concentration in the striatum seems to be a target for antipsychotics, being related with motor side effects associated with the nigrostriatal region. In the mesolimbic region, D3R has an important role in the dopamine system, 38 and its RNA encoding is abundant in the ventral area. As a matter of fact, the understanding of the anatomical distribution of dopamine receptor subtypes in the central nervous system (SNC) is crucial for the rational development of subtype-selective agents in order to improve treatment efficacy and to reduce side effects. 39 40 41 42 43 44 In this regard, it has been recognized that antagonism in D3R can represent a novel and potent antipsychotic mechanism devoid of EPS, making D3R a good target for the improved drug treatment of schizophrenia. 38 44 Another important factor to be considered are the kinectic properties of dopamine and other antipsychotic agents at dopamine receptors. The equilibrium dissociation constant kd, which can be used to represent the inverse of the ligand’s affinity, is a function of the ratio koff/kon where kon represents the ligand association and koff the ligand dissociation rate constants. Compared to typical antipsychotic drugs, higher koff values of atypical drugs are associated with faster dissociation and lower affinity. 45 Indeed, the link between binding affinity and dissociation rates for reversible biomolecular interactions was established due to the fact that measurements of antipsychotics kon for D2R showed practically


2

Page 3 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the same values. Thus, rapid dissociation could explain the difference between distinct agent effects. 45 Raising the concentration of dopamine levels in the synaptic cleft increases the tendency for dopamine to bind more frequently to the receptors, dissociating the antipsychotic as well. Drugs with high affinity for dopamine receptors, such as haloperidol, have a low dissociation rate constant and will not dissociate rapidly from the receptor, promoting a fuller blockade and, consequently, the appearance of EPS. For agents with low affinity for the dopamine receptors, such as clozapine and quetiapine, the dissociation rate will be fast, allowing the access of dopamine to the receptors. 11 Risperidone has an atypical profile, but its binding affinity is similar to the observed for haloperidol, suggesting that a high dissociation rate may be involved in the process of fast release from the receptor in the nigrostriatal region, reducing the overall occupancy and reaching clinical efficacy with none or reduced side effects.

Risperidone is formed by a fluorine substituted benzoxazole moiety (region i) linked to an piperidinyl ring with an tertiary amine (region ii), which is connected by an ethyl linker (region iii) to a tetrahydropyridopyrimidinone moiety with a ketone group in C20 and a methyl substituent in C29 (region iv) – see Figure 1A. The protonation states of risperidone at distinct pH values is shown in Figure 1B. In the 7.2-7.4 physiological pH range there is a prevalence of the risperidone protonated state (97.34 – 95.85%) at the piperidinyl ring. The electron density distribution of the isolated DFT-optimized risperidone molecule in its most charged state is depicted in Figure 1C where the protonated tertiary amine, which is behind the strongest contributions to the risperidoneD3R binding, is represented in blue.

In this work, we describe for the first time the binding energy profile and interaction of risperidone with individual residues at the D3R binding pocket. In order to achieve this, we take full advantage of the published crystallographic data of D3R complexed with eticlopride 46 to perform docking simulations of risperidone in D3R through eticlopride replacement. Our results show the existence of two possible risperidone orientations at the binding site. Their atomic coordinates were afterwards optimized through a QM/MM protocol, followed by ab initio quantum biochemistry calculations based on the density functional theory (DFT) formalism. Drug-amino acid residue interaction energies were estimated in the framework of the molecular fractionation with conjugate caps strategy (MFCC) 47, varying the binding pocket radius r (defined as the distance of the residue to the ligand centroid) from 3.0 to 10.0 Å to take into account the most important D3R residues interacting with both risperidone configurations. To describe risperidone dissociation, classical molecular dynamics studies were performed and their results suggest that, despite the similar binding energies of both risperidone orientations at the D3 binding pocket, distinct dissociation patterns occur pointing to a fast-off behavior for orientation 2 and a persistentblocking behavior for orientation 1.

RESULTS AND DISCUSSION Since the publication of the D3R structure co-crystallized with eticlopride, 46 some in silico simulations for the docking and evaluation of distinct ligands were performed. 48 As far as we known, this is the first report of the binding orientations of risperidone in the dopamine D3 receptor through docking essays by taking advantage of the crystallographic data of D3R complexed with the antagonist eticlopride with 3.15 Å resolution, PDB ID: 3PBL. 46 Docking results suggested that risperidone may assume two distinct orientations at the binding pocket of D3 receptors. To improve the quality of our analysis both orientations were submitted to geometry optimization using classical and quantum (QM/MM) methods (see below) to obtain the best conformation to be used during the calculation of the individual amino acid residues interaction energies within the MFCC approach. Also, the binding stability was assessed through molecular dynamic simulations for both orientations. The displacement of atoms at the binding pocket is fundamental for its function and has a pivotal role in drug design 49, overcoming limitations imposed by the structural rigidity of docking procedures 50 51 52. In our study, the geometry optimization procedure provides estimates of such atomic displacements and gives a more accurate description of the pose of risperidone in its human D3R receptor binding pocket. Risperidone Protonation, Structure and Charge State

Figure 1. (a) Atom labeling of risperidone in protonated form. Region i has the 6-Fluoro-1,2 benzoxazole fragment; region ii has the piperidin-1-yl fragment with the tertiary amine protonated, as observed at physiological pH; region iii has the ethyl fragment; and region iv has the 2-methyl-6,7,8,9-tetrahydro-4H-pyrido[1,2a]pyrimidin-4-one fragment; (b) Risperidone protonation state at physiological pH; (c) DFT electron density projected onto an electrostatic potential isosurface showing negatively charged regions in red and positively charged regions in blue of the protonated risperidone state at physiological pH. Docking Results: The Existence of Two Possible Orientations Initially, the docking parameters were set up through the application of a re-docking protocol to the eticlopride-D3R system using Autodock 4.0, as performed previously by our group. 53 The


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

docking of risperidone in the rigid pocket of the crystallographic D3R generated 1000 poses. Among them, two representative cases were chosen: risperidone orientation 1 (RO1), belonging to the cluster with the most negative energy and showing the fluorine atom oriented towards TMH5 (cluster 1); and risperidone orientation 2 (RO2), from the most populated cluster, with the fluorine atom pointing towards the region of TMH1, TMH2 and TMH7 (cluster 2). The cluster distribution at a rmsd of 1.0 Å is presented in table S1 in supporting information. Among the orientations, poses belonging to clusters 1, 5, 10, 11, 13, 14, 17, 20, 21, 22, 23, 24 and 25 have orientations similar to RO1, while poses from clusters 2, 3, 4, 6, 7, 8, 9, 12, 15, 16, 18, 19 and 26 are oriented in the binding pocket as RO2. As reported for benzofuranone derivatives unsubstituted at the furan ring and free of the bulky groups, some compounds are able to bind the dopamine and 5-HT2A receptors in an alternative orientation that swaps the position of the ligand at the binding pocket, resulting in one conformation with weaker interaction. 54 Such behavior was observed previously for butyrophenones with dopaminergic and serotoninergic affinities. 55 56 In this regard, risperidone structurally resembles the antipsychotic benperidol, a butyrophenone, which may in some extension account to the observed behavior on the binding site of D3R. Nevertheless, only a detailed study comparing the physical-chemical properties of both agents in solution and bound to aminergic receptors, including D3R, will be able to address this issue. Docking Refinement Through Classical and Quantum Calculations Although docking algorithms speed up computer-aided drug design through the prediction of the binding orientation of small molecules, or a set of candidate ligands, in a protein structure 50 57 58 the search for the best orientation can be compromised by the low accuracy of scoring functions. To improve the energetic analysis of pose results, more robust calculation approaches can be applied, such as the quantum mechanical (QM) methods, 59 which are being shown to be of great importance in all phases of in silico drug design 60 and are becoming more popular due to their high accuracy to estimate (relative) binding affinities. 61 Taking into consideration the amino acid side chains reorganization in the binding pocket during ligand interaction (while the backbone displacement is more rare), 62 we performed three independent procedures of geometry optimization for each structure to improve gradually the docking results. In the first approach, hydrogen atoms were added to the molecules and their geometries were classically optimized fixing the coordinates of all nonhydrogen atoms (Crude Docking Input – CRDI procedure). In the second optimization approach, classical molecular mechanics simulations were used to improve the coordinates of all hydrogen

Page 4 of 30

atoms in D3R and all atoms of the risperidone molecule (Classical Mechanics Docking Input – CLDI procedure). The third optimization approach allows for some protein flexibility, making simulation more realistic 58 63 by employing a Quantum Mechanics/Molecular Mechanics (QM/MM) scheme using the Own Nlayer Integrated molecular Orbital molecular Mechanics (ONIOM) technique 64 65 which allows the use of QM methods in large and complex molecular systems and to promote docking refinement. 59 66 In this hybrid description, the ligand was put in the QM layer and optimized using DFT, while the receptor was put in the MM layer. During the optimization of the ligand, all hydrogen atoms in the protein and all residues up to a radius of 10 Å from the risperidone centroid were allowed to relax (Quantum Mechanics Docking Input – QMDI procedure). Such refinement leads to a reliable structure to begin the investigation of proteinligand interactions. 67 When the risperidone molecule and residues at 10 Å from its centroid were allowed to move freely during QM/MM optimization, significant differences from the original crude docking result were observed. The analysis of the final structure reveals the displacement of risperidone atoms from the original docking pose (Figure 3, A and B) to a more stable conformation (as observed by the DFT-calculated total electronic binding energy shown in Figure 4). Also, improvements in the total binding energy were observed after quantum optimization of orientation 2, promoting the rearrangement of some important residues (figure 3, C and D). Indeed, rearrangements in orientation 2 improved the attractive effect of some residues, such as Asp110, Val189 and Tyr373, but decreased the interaction strength with others, such as, Met83, Asn185, Ser192 and Tyr365. Residue Trp85 became repulsive after optimization while repulsion to His349 was slightly reduced. In general, for both orientations, an increase of the total electronic binding energy of the system was observed. Analysis of QMDI structures show that for both orientations the fluorine atom forms a hydrogen bond with serine residues. In orientation 1, risperidone is attracted by Ser193 with interaction energy of -1.00 Kcal/mol (DFT-GGA-TS functional), while in orientation 2, it binds to Ser366 with interaction energy of -6.00 kcal/mol (GGA-TS). Also, orientation 1 is stabilized through a hydrogen bond with interaction energy of -7.80 kcal/mol (GGATS) between Thr369 and the N28 atom in region iv. Indeed, a relevant role for Thr369 was previously reported in a study of Feng and colleagues 62 where, after the refinement of docking data through molecular dynamics, Thr369 was described as highly important to the selectivity of R-22 in D3R. Near to the top of the binding cleft, orientation 1 is stabilized through a Pi-cation interaction between Phe106 and the tertiary amine (N16) at risperidone region ii, while orientation 2 interacts through a Pi-sigma interaction between the side chain of Val86 and the oxazole ring at risperidone region i.


4

Page 5 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. Distinct orientations of risperidone bound to D3R after docking refinement using QM/MM optimization. (A) Stick representation of risperidone orientation 1 in the binding pocket of D3R. Risperidone’s fluorine atom is near to TMH5. (B) Stick representation of risperidone orientation 2 in the binding pocket of D3R. In orientation 2, the fluorine atom of risperidone is oriented towards a region between TMH1, TMH2 and TMH7.

The dependence of the total binding energy on the binding site radius (measured from the risperidone centroid) is shown in figure 4 and the interaction energies of risperidone with every amino acid residue at the binding site are given in tables 2 and 3. In them, one can also inspect the distance from the risperidone centroid to the nearby D3R amino acid residues after each optimization strategy. The choice of residues was based in work of Chien et. al (2010) 46 and our quantum analysis of the binding energy of eticlopride 68 and haloperidol. 53 Charges of the individual atoms of risperidone were calculated using the Hirshfeld method (HPA) 69 67 , which minimizes the loss of information related to the formation of chemical bonds between atoms in a molecule. 22 70 HPA produces improved Fukui function indices 71 72 73 capable of predicting reactivity trends within a molecule better than Mulliken population analysis, 74 natural bond orbital analysis, 75 and fitted electrostatic potentials. 76 Although its values tend to be too small, 77 78 as Hirshfeld atoms in general resemble the neutral atoms, 79 80 this limitation can be amended using the iterative Hirshfeld charge technique, 78 which has been successfully applied to the solid state 81 and the discussion of Fukui functions. 82 The GGA-PBE Hirshfeld charges of the individual atoms of risperidone are expressed in units of the fundamental charge e, and were calculated using the HPA scheme available in the DMOL3 code. 83 Detailed

atom charges and bond lengths after CRDI and QMDI optimizations obtained through DFT calculation are included in the Supplementary Material of this paper (see tables S2 and S3). According with the HPA analysis, O21 and N28 are the most negatively charged atomic species in orientation 1 (orientation 2), with Hirshfeld charges of -0.266 (-0.282) and -0.185 (-0.174), respectively. The high negative charge of O21 in orientation 2 seems to play a role in the repulsion of the negatively charged region in the His349 ring. The nitrogen atom N16(H), a tertiary amine belonging to the Piperidin-1-yl group, is responsible for a large attractive interaction with Asp110 in the binding pocket, and holds a positive charge of 0.230 (0.246) in RO1 (RO2). Net charges have not presented significant variations between CRDI and QMDI for regions ii and iii. For the RO1-CRDI case it was observed a positive net charge value for regions i (0.036) and region iv (0.028) in which the QM/MM optimization produced an increase in the positive net charge distribution for regions i (0.085) and iv (0.122). In RO2, it was observed a large charge difference between regions i (0.179) and iv (-0.022) in the CRDI structure which disappears after QM/MM optimization, leading to a neutral distribution in both regions (see Table S2).


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 30

Figure 3. D3R-risperidone spatial arrangement before and after QM/MM optimization. (A) Conformational adjustment of Phe106 and Tyr 365 residues with respect to risperidone in orientation 1; (B) Adjustment of risperidone in orientation 1, Asp110, Phe345, Phe346 and Thr369, among other residues; (C) Adjustment of risperidone in orientation 2, His349 and Tyr365; (D) Adjustment of residues Phe345, Phe346 and Tyr373 relative to risperidone in orientation 2. Docking results with the frozen receptor are colored in gray and structures generated after QM/MM optimizations are colored by atom type. Total Risperidone-D3R Electronic Binding Energy In general, for the interaction of a ligand with a protein binding pocket, the nearest residues have stronger interaction with the ligand in comparison to more distant ones (this picture may change when charged residues are present), whose effects can be therefore disregarded. Nevertheless, it is not always clear which residues are the most relevant for the binding mechanism if one considers only closest neighbors in crystallographic data or docking simulations. So, in order to find out the individual residue contributions to the binding interaction at the quantum level, we employed the molecular fraction with conjugated caps (MFCC) 47 84 85 86 technique, which is a robust approach 59 in all phases of in silico drug design 60 and it is becoming of common use due to its high accuracy to estimate (relative) binding affinities. 61 To improve the accuracy of this method, especially when residues at large distances are included, shielding effects due to neighbor amino acid residues were considered following a procedure established in previous works we published. 53 68

the Density Functional Theory (DFT) approach. 87 88 A combined DFT-MFCC scheme is a very useful approach to achieved an accurate description of biological systems quantum calculations, as demonstrated in previous publications of our group 68 89 90 (more details in the Experimental Methods section). After adding up all the ligand-residue interaction energies of risperidone bound to D3R within a given binding pocket of radius r (the geometric centroid of risperidone defines the center of the binding pocket sphere), we obtain the r-dependent total risperidone-D3R electronic binding energy E(r) for both RO1 and RO2 configurations (see Figure 4 for the GGA-TS curves). For the sake of comparison with the GGA-TS curves, LDA-OBS E(r) data are available in the Supplementary Material (Figure S1) of the paper. Any differences observed when one compares LDA and GGA occur because long range interactions tend to be overestimated when the dispersion corrected DFT-LDA approach is used, while the GGA-TS functional provides a more accurate description of Van der Waals and hydrogen bonds at the same time. 91

The computational cost to simulate systems with hundreds of atoms employing quantum mechanics can be reduced by using the electron density ρ(r) instead of the wave function, as proposed by


6

Page 7 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


4.5 and 10 Å, while it tends to oscillate in the case of the CRDI and CLDI curves. Comparing the QMDI curves for RO1 and RO2, one can note that, until 9.0 Å, RO2 shows larger binding energy than RO1, but at 9.5 Å the RO1 total energy became larger (23 kcal/mol) due to the contribution of Glu90. As mentioned above, the effect of Glu90 may be minimized by the solvent effect and it would be interesting to investigate what happens when r increases beyond 10 Å for both orientations. Nevertheless, the limitations in the available computational resources make it impossible to include such a large number of amino acid residues in the calculations, preventing us to go beyond 10 Å, but these must be overcome in the near future. Therefore, to gain more information about the real stability in the binding pocket of both risperidone orientations, it was performed molecular dynamics simulations where the effect of solvent was taken into account, as described later. BIRD panels: GGA+TS The BIRD panels (see Materials and Methods for details on the BIRD panel representation) of Figures 5 and 6 highlight some of the residue-ligand interaction energies for RO1 and RO2, respectively. Optimized structures CRDI, CLDI and QMDI were all taken into account, and distances represents those in the QMDI structure. Overall, interaction energies involving 42 residues (45 residues) were calculated, as described in Table 1 (Table 2) for the RO1-D3R (RO2-D3R) complex.

Figure 4. Risperidone-D3R total interaction energy as a function of the binding pocket radius using the GGA-TS exchangecorrelation functional. Blue squares, red triangles and green circles represent the CRDI, CLDI and QMDI values, respectively. Results for risperidone orientation 1 (RO1) are shown in the top panel (solid symbols), while risperidone orientation 2 (RO2) data (open symbols) are depicted in the bottom panel. For both RO1and RO2 cases, on can see (Figure 4, top) a sharp decrease of E(r) as r increases from 3.0 to 4.0 Å at the QRDI level, as the nearest amino acid residues are taken into account. For RO1, the CRDI and CLDI data follows the same trend of QMDI in this radius range, while for RO2 the CRDI total energy at 3.0 Å has a value much smaller than QRDI, decreasing smoothly as r increases to 6.0 Å. Between 6.0 and 8.0 Å, CRDI E(r) does not vary significantly, reaching a maximum near 8.5 Å and then decreasing by about 50 kcal/mol before reaching r = 10 Å. The CLDI data, on the other hand, follows closely the QMDI values in the radius interval from 4.0 to 10 Å. The QMDI E(r) has a stepwise decrease with flat regions at (a) 4.5 Å < r < 5.5 Å, (b) 6.0 Å < r < 7.0 Å, and 8.0 Å < r < 8.5 Å (c), after incorporating contributions from residues Asp 110 (a); Val86, Phe106, Val107, Ile183 (b); and Val82, Leu89, Val111, Ser366, Tyr36 (c). The beginning of a new stability region for E(r) seems to occur for r > 10 Å, after taking into account the contribution of the Glu90 residue. Nevertheless, this contribution must be cautiously considered as the dielectric constant was not considered (due to the computational cost during calculations) and Glu90 is located near the outer part of the receptor, closer to the solvent interface. Looking now to the E(r) curves for the RO2 geometry, the classically optimized structures (CRDI, CLDI) behave differently from que quantum optimized geometry (QMDI) as r increases. For the later, E(r) decreases almost linearly in the radius range between

Figure 5. BIRD panel showing the interaction energy of each amino acid residue with risperidone orientation 1 using the GGATS approximation. After the QM/MM optimization, the relative total binding energy calculated for RO1-D3R (RO2-D3R) including all residues within 10 Å was -204.40 (-181.80) kcal/mol using the GGA-TS approximation. As described in figures 5 and 6, about 20 residues (18 residues) seem to respond for the total binding energy of the ligand, as the sum of their individual interaction is very close to the total binding energy observed for RO1 (RO2) in figure 4.


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 30

pyridopyrimidinone moiety (region iv) (a detailed account of atomic charges can be found in the Supplementary Material, Table S2). An investigation of the effect of replacing Ser192 with Ala (S192A) showed a decrease of affinity for D3R agonists. 95 The interaction energy contribution of Ser192 for RO1 (-4.10 kcal/mol) is slightly more significant than in the case of RO2 (3.80 kcal/mol). In the RO1-QMDI calculation, the Ser192 residue is close to the fluorine atom in the benzoxazole moiety (region i), while in the RO2-QMDI simulation the same residue interacts with the tetrahydropyridopyrimidinone moiety (region iv) near (C23)H2. In this region, the most important role belongs to Ser193, which interacts through an hydrogen bond with energy of -1.00 kcal/mol with RO1 (12 Å binding pocket sphere), while for RO2 it is repulsive with an interaction energy of 2.30 kcal/mol (11 Å binding pocket sphere). The residue Ser366 interacts with (C23)H2 and (C24)H2 in the tetrahydropyridopyrimidinone moiety of RO1 with energy of -8.00 kcal/mol. It also interacts attractively (-6.00 kcal/mol) with RO2 through and hydrogen bond with the fluorine atom in the benzoxazole moiety (region i). Residue Thr369 has an interaction energy of -7.80 kcal/mol with RO1, forming a hydrogen bond with the nitrogen atom (N28) in the tetrahydropyridopyrimidinone moiety (region iv). In the RO2QMDI geometry, Thr369 faces the piperidinyl ring (region ii) of RO2 and interacts attractively with energy of -4.10 kcal/mol. Figure 6. BIRD panel showing the interaction energy of each amino acid residue with risperidone orientation 2 using the GGATS approximation. A comparison between structures RO1-QMDI and RO2-QMDI shows that the helix-risperidone binding strengths follow the ordering TMH3 >> TMH2 > TMH7 > TMH6 > ECL2 ~ TMH5 for RO1 and TMH3 >> TMH6 > TMH7 >ECL2 > TMH5 > TMH2 for RO2. Hjerde et al. (2005) observed that the halogen atom was oriented towards TMH5 for typical antipsychotics while, in the case of atypical ones, it was oriented towards helices 2, 3 and 7. 92 In our study, the halogen atom of RO1 points towards TMH5, resembling the binding pattern of typical antipsychotics, as observed for haloperidol in a previous study by our group. 53 Our calculations at the quantum level indicate that the interaction strength of TMH5-RO1 is weaker than TMH5-RO2 (see Tables 1 and 2), which has its substituted tetrahydropyridopyrimidinone moiety (region iv) oriented toward TMH5 and the halogen atom oriented toward TMH1, TMH2 and TMH7. On the other hand, interaction with TMH2 is stronger for RO1 (48.90 kcal/mol, with GGA-TS), due to the presence of Glu90 at 9.5 Å, than for RO2 (-6.90 kcal/mol, with GGA-TS). Results of the typical antipsychotic haloperidol, showed that the sum of the interaction energies with residues Val82, Val86 and Leu89 in TMH2 is 3.90 kcal/mol (repulsion) within the GGA-TS approach, while a value of -4.90 (-1.10) kcal/mol (attraction) was found for the same residues in the case of RO1 (RO2). For both orientations, Cys114 seems to be an important source of repulsion (see Figures 5 and 6). This residue was described as part of an important domain for ligand interactions in the D3R binding pocket 25 93, and investigation showed that it is involved in the binding of ligands with at least one N-propyl group. When compared with the wild type in a competition binding experiment with [3H]-piperone, the mutated Cys114S receptor showed 272- and 102 fold improvements in the ki values for two aminotetralin antagonists, UH-232 and AJ-76, respectively 94. Our results show that Cys114 exhibits a repulsive interaction with energy of 4.00 (4.80) kcal/mol with respect to RO1 (RO2). Due to the spatial proximity, Cys114 interacts with the oxazole ring in region I of RO1-QMDI, while for RO2-QMDI it interacts repulsively with the negatively charged nitrogen atom (N28) in the tetrahydro-

Among the residues which distinguish the binding strengths of RO1-QMDI and RO2-QMDI one can mention Trp85, Val86, Leu89, Val111, Ser182, Ser193, His349 and Tyr365. Trp85 repels both RO1 (2.00 kcal/mol of interaction energy) and RO2 (5.00 kcal/mol). Val86, which is located in the opening of the binding site cavity, repels RO1 (0.80 kcal/mol), but attracts RO2 (-3.40 kcal/mol), suggesting its involvement in the attraction of RO2 by the binding site. Leu89 interacts attractively with the tetrahydropyridopyrimidinone moiety (region iv) with an energy of -8.70 kcal/mol, but becomes repulsive to RO2 by interacting with the oxygen (O2) and nitrogen (N1) atoms in the benzoxazole moiety (region i). The residue Val111, located near to Asp110, attracts RO1 (-4.10 kcal/mol) with RO1, but repels RO2 (0.90 kcal/mol). As observed for Val86, Ser182 seems to be relevant to push ligands towards the innermost region of the binding site. The Ser182 residue tends to push back both RO1 and RO2, more intensely with RO1 (2.70 kcal/mol) than with RO2 (0.80 kcal/mol). Ser193 binds the ligand (-1.00 kcal/mol interaction energy) through a hydrogen bond with the fluorine atom in the benzoxazole moiety of RO1. In the RO2-QMDI structure, Ser193 is near the carbon (C24)H2 in the tetrahydropyridopyrimidinone moiety and interacts repulsively with RO2 (2.30 kcal/mol interaction energy), contributing to attenuate the binding. During the QM/MM optimization procedure, His349 moved towards the peptide bond of Ile183-Ser184, generating a region with larger electron density due to the presence of the oxygen atom at the carboxyl terminal of Ile183. Such reorganization creates a negatively charged region which repels RO2, probably due to the interaction with the oxygen of ketone group in the pyramidine moiety. The hydroxyl group of Tyr365 forms a hydrogen bond (1.954 Å length) with His349. After the QM/MM optimization, the distance between the oxygen in the hydroxyl group of Tyr365 and the oxygen atom (O21) in region iv of risperidone decreased from 4.720 to 4.280 Å. Such displacement also decreased the distance between the oxygen atom of Tyr365’s hydroxyl group and the ketone group in the tetrahydropyridopyrimidinone moiety (region iv), attenuating the binding strength observed in the RO2CRDI simulation. When one considers the interaction with RO1, Tyr365 is nearer to the nitrogen atom (N28) in region iv and the carbon (C13)H2 in the piperidinyl ring (region ii).


8

Page 9 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1. Individual contributions of amino acid residues to the binding of risperidone orientation 1 to D3R. The minimum binding pocket radius size which includes each residue is shown as well. Energies calculated using LDA-OBS and GGA-TS are expressed in kcal/mol. Receptor Residue CRDI segment LDA-OBS GGA-TS TMH1 Tyr36 0 6.00 Val78 -3 -2.00 Val82 5 7.00 Met83 -6 -12.00 TMH2 Trp85 -1 -1.00 Val86 -25.6 -0.90 Leu89 -10.8 -4.70 Glu90 -57.2 -42.20 Cys103 1 -1.00 Phe106 -14 -5.00 Val107 -15 -3.00 Thr108 2 1.00 TMH3 Leu109 0 -1.00 Asp110 -111.4 -96.10 Val111 -7.7 0.10 Met112 1 2.00 Cys114 6.9 7.80 TMH4 Leu168 0 0.00 Asn173 0 -1.00 Val180 0 1.00 Cys181 1 0.00 ECL2 Ser182 0.8 2.00 Ile183 -21.5 -2.40 Ser184 0 3.00 Phe188 -1.9 -1.70 Val189 -5 -2.00 TMH5 Ser192 -7.9 -6.30 Ser193 -3.8 0.10 Trp342 -2 -2.00 Phe345 -13 -3.00 Phe346 -10 -5.00 TMH6 Thr348 0 1.00 His349 -12 -4.00 Asn352 5 5.00 Thr353 4 3.00 Tyr365 -6 -5.00 Ser366 -3 -2.00 Thr368 0 -1.00 TMH7 Thr369 -14.4 -2.70 Trp370 0 0.00 Gly372 2 2.00 Tyr373 -9 5.00

CLDI QMDI r(Å) LDA-OBS GGA-TS r(Å) LDA-OBS GGA-TS 9.00 2.00 0.00 9.00 3.00 1.00 9.50 -3.00 -2.00 9.50 -1.00 -2.00 7.00 3.00 4.00 7.00 2.00 3.00 9.00 -1.00 -1.00 9.00 0.00 -1.00 9.50 1.00 0.00 9.50 2.00 2.00 6.00 -13.30 -5.30 5.50 -9.70 0.80 7.00 -11.80 -5.70 7.00 -14.00 -8.70 10.00 -53.80 -46.80 9.50 -49.00 -43.00 10.00 -4.00 -3.00 10.00 0.00 -1.00 6.00 -20.20 -5.80 5.50 -13.30 -8.10 5.50 -18.80 -6.10 5.50 -15.00 -4.90 9.00 2.00 1.00 9.50 -1.00 1.00 9.50 -0.52 -1.00 9.50 0.00 -1.00 3.50 -110.10 -97.00 4.00 -105.30 -92.80 7.00 -7.00 -1.00 7.00 -5.90 -4.10 10.00 2.40 3.70 10.00 1.60 2.70 8.50 10.10 11.00 8.50 4.00 4.00 9.00 0.00 0.00 9.00 0.00 0.00 10.00 -1.00 -2.00 10.00 -4.00 -3.00 10.50 0.60 1.40 10.00 1.10 0.20 7.00 -3.00 -3.00 7.00 -4.30 -3.50 6.00 3.90 3.50 6.00 3.00 2.70 5.50 -21.00 -5.80 5.50 -17.30 -4.70 9.50 2.00 0.00 9.50 -2.00 -1.00 9.50 -1.00 -1.00 10.00 -1.00 -2.00 9.00 -6.00 -2.00 9.00 -6.00 -2.00 10.50 -5.50 -4.90 11.00 -5.20 -4.10 12.00 -3.00 0.00 12.00 -2.00 -1.00 6.00 -1.00 -3.00 6.00 -1.00 -1.00 4.50 -16.00 -5.90 4.00 -14.00 -6.30 9.00 -8.00 -6.00 9.50 -8.00 -4.00 10.50 0.00 -1.00 10.50 1.00 1.00 5.50 -10.50 -4.30 5.50 -10.20 -5.10 10.00 3.20 2.50 9.50 1.20 1.60 10.50 0.00 2.00 10.00 1.00 0.00 4.50 -7.00 -5.00 4.50 -14.00 -8.00 8.50 -5.00 -3.00 8.50 -14.00 -8.00 8.50 -1.00 -1.00 8.50 0.00 0.00 4.50 -16.70 -0.30 4.00 -25.70 -7.80 8.50 -0.60 -1.10 8.50 0.00 0.00 8.50 1.00 1.00 8.50 7.00 5.00 4.50 -14.10 -3.50 5.00 -6.00 2.00


r(Å) 8.00 9.50 7.50 9.50 9.00 6.00 7.50 9.50 10.00 6.00 6.00 9.50 9.50 4.00 7.50 10.50 9.00 9.00 9.50 10.00 6.50 5.00 6.00 9.50 10.00 9.00 10.00 12.00 6.50 4.50 10.00 10.00 5.00 10.50 9.50 3.50 8.00 8.50 4.00 8.50 9.00 5.00

9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 30

Table 2. Individual contributions of amino acid residues to the binding of risperidone orientation 2 to D3R. The minimum binding pocket radius size which includes each residue is shown as well. Energies calculated using LDA-OBS and GGA-TS are expressed in kcal/mol. Receptor Residue CRDI CLDI QMDI segment LDA GGA r(Å) LDA GGA r(Å) LDA GGA TMH1 Tyr36 2 2.00 10.00 3.00 1.00 9.50 2.00 0.00 Val78 -2 -1.00 10.00 -2.00 -3.00 10.00 0.00 -1.00 Val82 5 8.00 7.50 4.00 7.00 7.50 2.00 2.00 Met83 -5 -12.00 10.00 1.00 1.00 10.00 -2.00 -9.00 TMH2 Trp85 -3 -5.00 20.00 2.00 1.00 10.00 6.00 5.00 Val86 -16.3 -3.90 7.00 -6.80 -3.00 6.50 -8.10 -3.40 Leu89 0.3 4.10 8.50 -9.10 -6.00 8.00 -2.20 0.30 Glu90 -40.18 -35.45 11.00 -46.69 -42.97 10.50 -45.17 -42.00 Phe106 -6 -2.00 7.00 -6.00 -3.00 6.00 -6.00 -3.00 Val107 -4 -4.00 5.50 -5.00 -5.00 5.50 -6.20 -2.30 Thr108 0 -1.00 9.00 -1.00 -1.00 9.50 1.00 1.00 Leu109 -1 -1.00 9.50 -1.00 -1.00 10.00 -1.00 0.00 Asp110 -107 -96.00 4.00 -100.20 -90.00 4.00 -111.80 -106.30 TMH3 Val111 -9.8 -0.90 6.00 -9.90 -3.00 6.50 -9.80 0.90 Met112 1 0.00 9.50 1.00 2.00 10.00 2.00 2.00 Met113 0 1.00 10.00 0.00 0.00 10.50 1.00 1.00 Cys114 1.7 7.10 7.50 8.00 7.00 8.50 4.90 4.80 Thr115 1 1.00 10.00 0.00 0.00 11.00 -1.00 1.00 TMH4 Leu168 0 1.00 8.00 1.00 0.00 8.50 1.00 1.00 Asn173 0 1.00 10.00 -2.00 0.00 10.00 -3.00 -1.00 Cys181 1 -1.00 8.00 -1.00 -1.00 7.50 -2.80 -2.20 Ser182 0.8 2.00 6.50 3.60 3.00 6.00 -1.30 0.80 ECL2 Ile183 -12.7 -5.50 4.50 -15.50 -7.70 5.50 -13.60 -8.00 Ser184 0 1.00 9.00 0.00 0.00 9.00 1.00 1.00 Asn185 -4 -4.00 10.00 -4.00 -4.00 10.50 -2.00 -3.00 Phe188 -1 0.00 8.50 -1.00 -1.00 9.00 -2.00 -3.00 Val189 -22 -3.00 8.00 -17.00 -5.00 8.00 -12.00 -7.00 TMH5 Ser192 -22.6 -11.90 9.00 -8.70 -4.40 10.00 -3.70 -3.80 Ser193 -5.9 1.10 10.50 -2.70 1.20 11.00 -1.60 2.30 Ser196 0.2 -0.60 10.00 0.20 0.50 11.00 2.00 0.00 Trp342 -20 -7.00 5.50 -2.00 -5.00 6.00 -4.00 -2.00 Phe345 -19.4 -5.30 3.00 -15.60 -7.60 3.00 -18.50 -7.90 Phe346 -13.5 -6.20 8.00 -13.00 -8.00 8.50 -11.00 -11.00 Thr348 1 1.00 10.00 1.00 1.00 9.50 -1.00 0.00 TMH6 His349 -8.5 2.70 5.00 -8.50 1.70 4.50 -8.90 -0.60 Val350 -1 -2.00 9.50 1.00 -1.00 10.00 -3.00 -2.00 Asn352 5 4.00 9.50 3.10 2.40 9.00 -2.00 -1.00 Thr353 1 1.00 9.50 1.00 1.00 9.50 0.00 -2.00 Tyr365 -13 -11.00 4.50 -17.00 -14.00 4.00 -11.00 -6.00 Ser366 -3 -4.00 9.50 -9.00 -5.00 8.50 -14.00 -6.00 Thr368 -1 -2.00 8.50 -2.00 -2.00 8.50 -1.00 -2.00 Thr369 -12.4 -1.10 4.50 -11.70 -1.40 4.00 -15.90 -4.10 TMH7 Trp370 1 0.00 9.00 1.00 1.00 9.00 1.00 0.00 Leu371 0 0.00 11.00 0.00 1.00 10.50 0.00 0.00 Gly372 2 2.00 8.50 2.00 2.00 8.50 2.00 2.00 Tyr373 -10 0.00 5.50 -7.00 -1.00 5.50 -13.00 -6.00

A visual check of the electronic density distribution among risperidone and relevant residues in the binding pocket of D3R is presented in Figures 7 and 8 for RO1 and RO2 respectively. One can confirm from them that Asp110 has a high electron density concentrated around its carboxyl group, which is responsible for the interaction with the protonated tertiary amine in the piperidinyl ring (region ii) of risperidone in both structures, RO1 and RO2. The same pattern of charge distribution is observed in the carboxyl group of Glu90, which responds for a large interaction with the positively charged portion of the ring in the tetrahydropyridopyrimidinone moiety (region iv) of RO1 (see Supplementary Material, Table S2, for a list of atomic charges). We also highlight, among others, the electron density around the repelling residues Trp85 and Cys114 and around the nitrogen atom of His349. When interacting with RO1, His349 binds the carbon C8 in the aromatic ring at region i.

r(Å) 8.50 10.00 8.00 10.00 10.00 6.50 8.50 10.50 6.50 6.00 9.50 10.00 4.00 7.00 10.50 11.00 8.50 11.00 9.00 9.50 7.50 4.50 5.50 9.00 11.00 9.50 8.50 10.00 11.00 11.00 5.50 3.00 8.00 9.50 4.50 9.50 8.50 9.00 4.00 8.50 8.50 4.50 8.50 10.00 8.50 5.00

stays closer to His349 and to the oxygen atom of the carboxyl terminal region of Ile183 in the peptide bond. Such spatial reorganization leads to a repulsive environment, which prevents RO2 to reinforce its binding strength in this region, as occurs for RO1. In order to facilitate the description of the spatial arrangement of residues in the empty binding pocket, a risperidone molecule was depicted at the top panels of figures 7 and 8 with its geometry optimized, together with the electronic density and electrostatic potential environment of the empty binding pocket. The middle and bottom panels, on the other hand, show the electron density distribution affected by the binding of risperidone. From Table 2, when one compares individual energy contributions for distinct types of optimization (CLDI, CRDI and QMDI), the structural rearrangement at the binding pocket changes the pattern of binding energies and the distribution of repelling regions in the structure, leading to the reduction of the total binding energy of RO2.

When risperidone binds in the RO2 orientation, the substituted tetrahydropyridopyrimidinone moiety (region iv) of risperidone


10

Page 11 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


differences between RO1 and RO2. Ser193 interacts attractively (1.00 kcal/mol) with RO1, but repulsively (2.30 kcal/mol) with RO2. Residue His349 has a small interaction energy contribution of -0.60 kcal/mol to the binding of RO2 (in contrast with -5.10 kcal/mol, for RO1), as unveiled after QM/MM optimization. In fact, if it is accounted the contribution of Glu90 to RO2 (calculated for the QMDI structure), it would sum up – 223.8 kcal/mol (GGA-TS) of total binding energy, suggesting that RO2 is more likely to bind, while RO1 may play a role of an alternative orientation of risperidone with faster receptor dissociation.

Figure 7. DFT (GGA-TS) electrostatic potential isosurfaces for the RO1-QMDI structure. (A) Colors indicate projected electron densities for the main interacting residues at the binding pocket of D3R. The risperidone molecule is displayed but does not interact with D3R; (B) Top: attractive residues Asp110, Glu90, Phe345, Phe346, His349, Tyr365 and Ser 366. Bottom: a different view of the binding pocket. In it, one can see the repulsive residues Val82 (TMH2) and Cys114 (TMH3). The quantum calculations indicate that Glu90 plays an important role in the stability of the RO1 and RO2 bindings. This residue interacts attractively with RO1 with an energy value (-43.00 kcal/mol, GGA-TS) which is similar to the calculated for RO2 (42.00 kcal/mol, GGA-TS) and tends to repel the fluorine atom of RO2 to the inside of the pocket due to its electronegative character (see representation in figure 1C). Nevertheless, Glu90 is closer to the centroid of RO1 than RO2, suggesting a higher effect over RO1 geometry. Interesting, if one simply compares the total binding energy after QM/MM optimization without analyses the individual residue contribution it will give the false idea that RO1 is more stable than RO2. Nevertheless, a more careful reader will notice that Glu90, which responds for a huge energy value, is located on the entrance of the binding cleft and thus it attracts RO1 to its direction. On the other hand, some residues with important contributions to the binding strength of RO1, such as Leu89, Val111, Ser184, Ser193 and His349, are less important when interacting with the alternative orientation RO2, indicating that they are at least partially responsible by the binding strength

Figure 8. Electrostatic potential isosurfaces of the RO2-QMDI structure. (A) Projected electron densities for the main interacting residues in the empty binding pocket of D3R. Negative charge concentrates around Asp110, Tyr373, His349 and Ser366. Electrostatic potentials were calculated without taking into account the risperidone molecule, which was inserted only for descriptive purposes; (B) Top: at the right side attractive interactions with Asp110, Met83 and Tyr373 dominate (the last residue is shown in the reversed panel at the bottom), taking into account the presence of the risperidone molecule. The repulsive residue Cys114 is also depicted. Residue His349 is shown at the left side of the figure together with the attractive residues Ile183, Val189, Phe345 (bottom), Phe346 (bottom), Tyr365 and Ser366.


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Dynamics In order to check the hypothesis of faster dissociation, we performed a classical molecular dynamics simulation (see details in the “Experimental Methods” section). During the simulations the alpha-carbons of D3R were kept under strong restraint, allowing free movement only to the side chains and risperidone molecule. While results show stability in the D3R-RO2 complex (Figure 10), in the D3R-RO1complex analysis it was observed a change in RO1 geometry at the beginning of the simulation (Figure 9). Simulations were repeated 2 times, 40 ns each, and the distance among risperidone and some residues were used to monitor the stability of the orientation. As observed in figure S5 the distance between the carbonyl group of Asp110, which is a key residue, and the hydrogen atom in the piperidinyl fragment is more prone to oscillate in D3R-RO1 complex than in D3R-RO2 (Figure S6). The same behavior was observed when distances between residue Leu89 and carbon atom C23 (Oxygen atom O2) of RO1 (RO2) were measured (figures S7 and S8, respectively). As the simulations were performed without the presence of membrane lipid, the constraints imposed to the receptor flexibility (alpha-carbons) helped avoiding the receptor to undergo improper conformational changes. Nevertheless, important features as the induced fit expected during the interaction between D3R with both RO1 or RO2 were lost.

Page 12 of 30

Figure 10. RMSD of D3 receptor (D3R) and risperidone orientation 2 (RO2) during 40 ns of molecular dynamics. Simulation was performed in duplicate and is represented as MD1 or MD2 As reported by Kapur and Seeman (2000) while typical and atypical antipsychotics have similar association constants, the atypical mechanism may be explained by the speed of dopamine receptor dissociation 45, as observed in the fast-off D2R antagonism of clozapine, the prototype of the “atypical antipsychotics”. 97 As a matter of fact, this fast dissociation, even at equilibrium concentration, is necessary to explain the sufficient D2R stimulation by endogenous dopamine in the striatum. As pointed by Chang et al. (2008) the high frequency observed during docking indicates the binding enhancement due to favorable entropy. 96 Interesting, despite the stronger binding energy of RO1 observed from the docking simulation, the clustering strategy revealed that RO2 pose is more frequent. Moreover, RO2 is positioned deeper into the binding cleft and the attractive effect of Glu90 over RO1 suggests that the rate of dissociation for this orientation is faster than for RO2. Also, a tendency of RO1 detachment from the receptor was observed during molecular dynamic simulations (Figure S5-8), nevertheless this results must be taken carefully as the restraint forces imposed to the protein backbone limited the quality of analysis, suggesting that further simulations including membrane will be necessary to deal properly with this system. Kapur and Seeman 11 describes that D2R occupancy has distinct threshold for clinical response and EPS. While the threshold for EPS is about 80% for typical and atypical antipsychotics, the threshold for effectiveness of haloperidol (typical agent), olanzapine and risperidone (atypical agents) is 65%. Therefore, both typical and atypical antipsychotics block D2R with different occupancy kinetics and different relationships of peak to through occupancy, independently of other receptors binding profiles, ensuring that even a low affinity for D2R is sufficient for producing atypical antipsychotic activity.

Figure 9. RMSD of D3 receptor (D3R) and risperidone orientation 1 (RO1) during 40 ns of molecular dynamics. Simulation was performed in duplicate and is represented as MD1 or MD2.

Binding affinity is a dynamic process, with simultaneous association and dissociation stages, being represented by rate constants of association dissociation. It was already demonstrated that typical and atypical antipsychotics attach to D2R with similar rate constants of association, but differ in how fast they dissociate from the receptor. 45 This result is useful to understand how antipsychotics and endogenous dopamine compete in the brain, where low affinity drugs, in equilibrium, decrease their occupancy much faster and provide much more access to surges of dopamine. 98 Here, we propose that the atypical behavior observed for risperidone is due to the combination of the binding profile of two orientations, where the fast dissociation of one orientation (RO1) is responsible for the reduction in the receptor D2 occupancy in the nigrostriatal region, avoiding EPS triggering. The existence of two alternative orientations of risperidone in the binding site of D3R may be a confusing factor for structural characterization through methods such as X-ray diffraction. The results reported in this work should stimulate careful studies to elucidation the risperidone binding configuration in dopamine receptors to confirm the existence of RO1 and RO2. Longer molecular dynamics with protein embedded into membrane plus essays of receptor occupancy and measurements of the dissociation constant must be employed in order to analyze the dissociation profile of risperidone based in our hypothesis of alternative orientations.


12

Page 13 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


CONCLUSIONS Our simulations highlight hot spots in the dopamine receptor binding site and suggests two distinct orientations (RO1 and RO2) for the binding of risperidone in the D3R receptor. Classical and quantum simulations point to a mechanism suggesting the reduction in receptor occupancy through the fast dissociation of RO1. These results were obtained through refinement of the docking calculations by employing a QM/MM protocol to estimate drugresidue interaction energies. The number of residues interacting with the drug was investigated by increasing the binding pocket radius r (measured from the drug centroid) from 3.0 to 10 Å. Classical molecular dynamics was employed to compare the receptor dissociation rates of both drug orientations. Docking procedures followed by QM/MM refinement showed that the largest populated cluster, although is not the cluster with larger binding energy, represents the best orientation. The conserved residue Asp110 has a pivotal role binding RO1 and RO2, exhibiting a strong interaction with the tertiary amine of the ligand. RO1 showed a stable binding geometry, with relevant attractive interactions involving Asp110, Glu90, Leu89, Phe106, Tyr365, Ser366, Thr369, Phe345, and His349 among others. Among those, Glu90 seems to attract risperidone to the opening of the binding cleft, corroborating to its dissociation. Also, repulsive interactions were found for Gly372, Met112, and Cys114. RO1 interacts with the D3R helices in the following order of strength: TMH3 >> TMH2 > TMH7 > TMH6 > ECL2 ~ TMH5. Although at 10 Å radius the total binding energy of RO2 was smaller than the value for RO1, when the interaction of Glu90 is accounted RO2 come up to be more stable. RO2 is located deeper into the binding pocket and interacts attractively with the following residues: Asp110, Phe346, Met83, Ile183, Phe345, and Val189. Binding energies of RO2 with the D3R helices follow the order TMH3 >> TMH6 > TMH7 >ECL2 > TMH5 > TMH2. Trp85 and Cys114 repel RO2, with a small attractive interaction involving His349, which seems to weaken the binding of this configuration. After performing classical molecular dynamics, it was shown that RO1 is likely to dissociate faster from the receptor. As risperidone binds strongly to D3R, like typical agents do, the presence of a secondary binding conformation able to dissociate more quickly while the primary remains stable at the binding site can aid in the understanding of this drug atypical profile. Such rapid dissociation may account for the D2R occupancy at nigrostriatal region in levels below to those of EPS incidence. Overall, our results demonstrate that computational techniques at the quantum level can be very useful for the development of novel/derivative atypical antipsychotics. In the future, crystallography studies should search for the alternative binding orientations proposed in this work and their implications for the binding stability and occupancy of human dopamine receptors D2/D3.

EXPERIMENTAL METHODS Structural data The calculations performed in this study used the X-ray crystal structure of human dopamine D3 receptor in complex with eticlopride (PDB ID: 3PBL) with a resolution of 3.15 Å. 46 The crystal asymmetric unit cell contains two receptors (A and B) in an antiparallel orientation, both exhibiting slight shape differences. We arbitrarily chose receptor A to use in our study. The eticlopride molecule was removed and the remaining receptor geometry was adopted during the docking approach. The drawing process and the evaluation of the protonation state at physiological pH of risperidone were carried out using the Marvin Sketch code version

5.5.0.1 (Marvin Beans Suite – ChemAxon). The so-built risperidone molecule was subjected to classical annealing and quantumlevel geometry optimizations to find the conformation with smallest total energy. To adjust the molecular structure to the protonation state at physiological pH, a single hydrogen atom was added to the amine group of risperidone and charged to +1 (electron charge = -1). The D3R structure was protonated as prescribed by the analysis performed through the PROPKA 3.1 web server (http://propka.ki.ku.dk) and the Protonation tool in the Discovery Studio package. Molecular Docking Molecular docking in this work was performed using the Autodock4 code. 63 99 100 The docking protocol adopted here follows the same steps of our previous work involving the binding of haloperidol in D3R 53, including the application of the Lamarckian genetic algorithm (GA). Docking simulations were performed 20 times using the optimized structure of risperidone as the input file, GA with 25,000,000 energy evaluations per run, population size set to 150 and a maximum of 27,000 generations per run. In the end, one thousand poses were obtained (50 poses per output) and clustered using a RMSD tolerance of 1.0 Å employing Autodock Tools. 63 101 Construction of the D3R-risperidone complex The D3R-risperidone complex was built using the dopamine D3 receptor structure after removing the ligand eticlopride. The molecule of risperidone, previously submitted to total energy minimization, was docked to the binding pocket using a rigid-protein protocol. Results were clustered and two representative poses (one from the largest binding energy cluster and another from the most populated cluster) were scrutinized. ONIOM (QM/MM) Optimization The preparation of input structure was carried out using the TAO package 102. Quantum mechanics / molecular mechanics (QM/MM) geometry optimizations were performed using the 2layer ONIOM framework 64 available in the Gaussian code 103. For the quantum region, the hybrid meta-GGA exchangecorrelation functional M06-2X 104 105 was chosen, together with a 6-311G(d,p) basis set to expand the electronic orbitals. The AMBER force field was adopted for the MM part using an electronic embedding scheme. Ligand charges were assigned using the AMBER force field and all amino acid residues at a radius of 10.0 Å from the ligand centroid were allowed to move freely during the geometry optimization steps. Only the ligand was included in the QM layer, while the receptor was put in the MM layer. Classical and DFT simulations for the MFCC procedure Hydrogen atoms were added to the D3 X-ray structure and their positions were optimized classically fixing the non-hydrogen atoms. The geometry optimization procedure was performed using the Forcite code with convergence tolerances set to 2.0 x 105 kcal/mol (total energy variation), 0.001 kcal/mol.Å (maximum force per atom) and 1.0 x 10-5 Å (maximum atomic displacement). The DFT calculations were carried out using two distinct exchange-correlation functionals: the first within the local density approximation (LDA) using the parameters obtained by Cerpeley, Alder, Perdew, and Zunger 106 107 (CAPZ), and the second within the generalized gradient approximation (GGA) employing the Perdew-Burke-Ernzerhof (PBE) 108 parametrization with the dispersion corrected energy proposed by Tkatchenko and Scheffler 109 (GGA+TS). While the PBE functional produce results close to those of the PW91 functional 110, dispersion correction terms avoid the need to use high-level quantum methods to


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

describe van der Waals interactions. The DMOL3 code 83 111 was used and a Double Numerical plus Polarization (DNP) basis set was chosen to expand the Kohn-Sham orbitals. The orbital cutoff, which controls the quality of the numerical basis set and of the numerical integrations performed during the computations, was set to ensure a good balance between accuracy and computational time. The total energy variation to achieve self-consistent field (SCF) convergence was set to 10-6 Ha. Molecular Fractionation with Conjugate Caps (MFCC) and Shielding Effect Correction The MFCC scheme is a very useful method to achieve an accurate description of biological systems through quantum calculations 47 84 85 86 112 with reduced computational cost. We represent the risperidone molecule by M and use R(i) to indicate the i-th amino acid residue interacting with M. The C(i-1) [C(i+1)] cap is made from the residue covalently bound to the amine (carboxyl) group of R(i) with hydrogen atoms being added to passivate any dangling bonds. To improve the accuracy of the MFCC method, we have took into account electrostatic shielding effects, as previously described 68 89 90 . The shielding effect is due to the presence of charged residues R(b) placed between M and R(i), which contribute significantly to affect their interaction (Figure S2). In this situation, the interaction energy EI[M-R(i)] is calculated in two steps. First, the energy, taking into account both the R(b) and R(i) contributions (Figure S3), EI[M-R(b)R(i)), is obtained: EI[M-R(b)R(i)] = E[M+C(b-1)R(b)C(b+1)C(i-1)R(i)C(i+1)] – E[C(b-1)R(b)C(b+1)C(i-1)R(i)C(i+1)] - E[M+C(b-1)C(b+1)C(i1)C(i+1)] + E[C(b-1)C(b+1)C(i-1)C(i+1)] (1) Afterwards, the EI(M-Rb) interaction energy (Figure S4) is calculated: EI[M-R(b)] = E[M+C(b-1)R(b)C(b+1)] – E[C(b-1)R(b)C(b+1)] E[M+C(b-1)C(b+1)] + E[C(b-1)C(b+1)] (2) Finally, the corresponding individual interaction (binding) energy EI[M-R(i)] is obtained through: EI[M-R(i)] = EI[M- R(b)R(i)] - EI[M-R(b)] (3) At the right side of Eq. (1), the first term E[M+C(b1)R(b)C(b+1)C(i-1)R(i)C(i+1)] is the total energy of the system formed by M, R(i), the shielding residues R(b) and the capped residues; the second term, E[C(b-1)R(b)C(b+1)C(i-1)R(i)C(i+1)], gives the total energy of the capped residues alone, while the third term, E[M+C(b-1)C(b+1)C(i-1)C(i+1)] is the total energy of the system formed by the set of caps and M; finally, E[C(b1)C(b+1)C(i-1)C(i+1)] is the total energy of the system formed by the isolated caps. In Eq. (2), the first term E[M+C(b1)R(b)C(b+1)] is the total energy of the system formed by M, the shielding residues R(b) and the capped shielded residues; the second term, E[C(b-1)R(b)C(b+1)], gives the total energy of the capped residues alone, while the third term, E[M+C(b-1)C(b+1)] is the total energy of the system formed by the set of caps and M; finally, E[C(b-1)C(b+1)] is the total energy of the system formed by the isolated caps. Eq. (3) gives to the difference between the interaction energies of Eq. (1) and Eq. (2), resulting in the total interaction energy of the residue investigated R(i) and the shielded risperidone molecule M. Total interaction energy as a function of the binding pocket radius

Page 14 of 30

The total interaction energy of the drug-receptor system depends on the size of the binding pocket considered in the simulations. By increasing its radius, one includes more residues which contribute to the binding. For the risperidone-D3 complex, we increased the binding pocket radius for the complexes obtained after (i) docking, (ii) classical molecular mechanics optimization, and (iii) QM/MM optimization. Using an arbitrary binding pocket size would risk missing important residues, so the binding radius was varied until the total binding energy variation was very small. The total binding energy as a function of the binding site radius r was obtained for a 3.0-10 Å binding pocket radius range with incremental steps of 0.5 Å. Only residues with at least one atom inside an imaginary sphere centered at the drug centroid with radius r were taken in to account to compute the total binding energy. The BIRD Panel Interaction energies of individual amino acid residues were plotted in a so-called BIRD panel. BIRD is an acronym for Binding site, Interaction energy and Residues Domain, and it shows: (i) the interaction energy (in kcal/mol) of the drug to each residue at the binding site using horizontal bars, from which one can assess visually the relevance and role of each residue, whether it attracts or repels the drug; (ii) the most important residues contributing to the binding in the column of residues at the left side; (iii) the region (i, ii, iii or iv) of risperidone closest to each residue; and (iv) the distance of each residue to the risperidone centroid, at the right side. Molecular Dynamics Molecular Dynamics (MD) simulations were performed using GROMACS v. 5.1.2 113, adopting the SPC water model and the GROMOS 53a6 force ﬁeld 114 in a box containing 65226 water molecules. Chloride and sodium counter-ions were added to neutralize the system maintaining a final salt concentration of 0.15 mol/L. Simulations were performed using alpha-carbon constraint to avoid improper conformational changes in the absence of membrane. Before carrying out the MD simulations, total energy minimization was accomplished by combining the steepest-descent algorithm and the conjugate gradient method in sequence. This initial optimization was followed by the equilibration where the temperature of the system was increased to 300K. At the total, 40 ns of simulation time were performed for each one of two runs (MD1 and MD2) for both risperidone orientations. The “distance” code was employed to calculate minimum interatomic distances while the “rms” code evaluated the distance root mean square deviations of the ligand and the receptor amino acid residues. Molecular Drawing and Image Acquisition The Marvin Sketch code version 5.5.0.1-2011 (ChemAxon, http://www.chemaxon.com) was used to draw the 2D structure of risperidone and to predict its protonation state at physiological pH. Images of protein-ligand structures were rendered using PyMol 1.3 115. Supporting Information Available Information regarded to calculation using LDA-OBS functional is depicted in the BIRD graph. Graphical description of MFCC scheme including the shielding effect is shown. Molecular dynamic analysis is described in terms of RMSD of receptor and risperidone orientations and in terms of residue-ligand distance. Tables showing the docking clusters, calculated atom charges and bond distances are available. This material is available free of charge via the Internet at http://pubs.acs.org.


14

Page 15 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


AUTHORS INFORMATIONS

New Hypothesis. Am. J. Psychiatry 158, 360–369.

Corresponding Author

(12) Chien, W. T., and Yip, A. L. (2013) Current approaches to treatments for schizophrenia spectrum disorders, part I: an overview and medical treatments. Neuropsychiatr Dis Treat 9, 1311–1332.

Geancarlo Zanatta E-mail: [email protected] Departamento de Bioquímica Rua Ramiro Barcelos, 2600 – anexo Bairro Santa Cecília 90035-000 - Porto Alegre – RS Phone: +55 51 3308-5551

(13) Yan, J. (2007) Risperidone approved to treat schizophrenia in children. Am. Psychiatr. News 42. (14) Sharma, A., and Shaw, S. R. (2012) Efficacy of Risperidone in Managing Maladaptive Behaviors for Children With Autistic Spectrum Disorder: A MetaAnalysis. J. Pediatr. Heal. Care 26, 291–299.

Author Contributions G.Z. performed the molecular docking and prepared the Risperidone-D3R complexes. G.Z. performed the structural refinements under the supervision of E.M.B., A.M. and R.F.C. G.N. and G.Z. prepared the input files required to apply the MFCC scheme. G.Z. did the quantum calculations under supervision of V.N.F. and E.W.S.C. G.Z. performed molecular dynamics simulations. G.Z, V.N.F., C.G. and E.W.S.C. wrote the initial version of the manuscript and all authors discussed, corrected and approved the final version.

ACKNOWLEDGMENT The authors would like to thank the Brazilian System of High Performance Processing (Sistema Nacional de Processamento de Alto Desempenho – SINAPAD), specially CESUP-RS and CENAPAD-UFC. A.M., V.N.F., and C. G. are senior researchers from the Brazilian National Research Council (CNPq), and acknowledge the financial support received during the development of this work from the Brazilian Research Agency CNPq-INCT-Nano(Bio)Simes, project 573925/2008-9. E.W.S.C. received financial support from CNPq project 307843/2013-0. C.G. received financial support from CNPq project 478916/2010-8.

REFERENCES (1) Organization, W. H. (2016) Schizophrenia - Fact Sheet. April 2016. (2) WHO, W. H. O. –. (2001) Nations for Mental Health - Schizophrenia and public health. Geneva. (3) Whiteford, H. A., Degenhardt, L., Rehm, J., Baxter, A. J., Ferrari, A. J., Erskine, H. E., Charlson, F. J., Norman, R. E., Flaxman, A. D., Johns, N., Burstein, R., Murray, C. J. L., and Vos, T. (2013) Global burden of disease attributable to mental and substance use disorders: Findings from the Global Burden of Disease Study 2010. Lancet 382, 1575–1586. (4) Whiteford, H. A., Ferrari, A. J., Degenhardt, L., Feigin, V., and Vos, T. (2015) The global burden of mental, neurological and substance use disorders: An analysis from the global burden of disease study 2010. PLoS One 10. (5) Andreasen, N. C., and Olsen, S. (1982) Negative V Positive Schizophrenia - Definition and Validation. Arch. Gen. Psychiatry 39, 789– 794. (6) Seeman, P., Lee, T., Chau-Wong, M., and Wong, K. (1976) Antipsychotic drug doses and neuroleptic/dopamine receptors. Nature 261, 717–719. (7) Seeman, P. (2001) Antipsychotic drugs, dopamine receptors, and schizophrenia. Clin. Neurosci. Res. 1, 53–60. (8) Carlsson, A. (2001) A Paradigm Shift in Brain Research. Science (80-. ). 294, 1021–1024. (9) Strange, P. G. (2001) Antipsychotic Drugs: Importance of Dopamine Receptors for Mechanisms of Therapeutic Actions and Side Effects. Pharmacol. Rev. 53, 119–134. (10) Stahl, S. M. (2003) Describing an Atypical Antipsychotic: Receptor Binding and Its Role in Pathophysiology. J. Clin. Psychiatry 5, 9–13. (11) Kapur, S., and Seeman, P. (2001) Does Fast Dissociation From the Dopamine D2 Receptor Explain the Action of Atypical Antipsychotics?: A

(15) Shea, S., Turgay, A., Carroll, A., Schulz, M., Orlik, H., Smith, I., and Dunbar, F. (2004) Risperidone in the Treatment of Disruptive Behavioral Symptoms in Children With Autistic and Other Pervasive Developmental Disorders. Pediatrics 114, e634–e641. (16) Janssen, P. A., Niemegeers, C. J., Awouters, F., Schellekens, K. H., Megens, A. A., and Meert, T. F. (1988) Pharmacology of risperidone (R 64 766), a new antipsychotic with serotonin-S2 and dopamine-D2 antagonistic properties. J. Pharmacol. Exp. Ther. 244, 685–693. (17) Meltzer, H. Y., Matsubara, S., and Lee, J. C. (1989) Classification of typical and atypical antipsychotic drugs on the basis of dopamine D-1, D-2 and serotonin2 pKi values. J. Pharmacol. Exp. Ther. 251, 238–246. (18) Meltzer, H. Y. (1999) The Role of Serotonin in Antipsychotic Drug Action. Neuropsychopharmacology 21, 106S–115S. (19) Seeman, P. (2002) Atypical antipsychotics: mechanism of action. Can. J. Psychiatry. 47, 27–38. (20) Kebabian, J. W., and Calne, D. B. (1979) Multiple receptors for dopamine. Nature 277, 93–96. (21) Jaber, M., Robinson, S. W., Missale, C., and Caron, M. G. (1997) Dopamine receptors and brain function. Neuropharmacology 35, 1503–1519. (22) Ayers, P. W., Morrison, R. C., and Roy, R. K. (2002) Variational principles for describing chemical reactions: Condensed reactivity indices. J. Chem. Phys. 116, 8731–8744. (23) De Lean, A., Stadel, J. M., and Lefkowitz, R. J. (1980) A ternary complex model explains the agonist-specific binding properties of the adenylate cyclase-coupled beta-adrenergic receptor. J. Biol. Chem. 255, 7108–7117. (24) Rasmussen, S. G., Choi, H. J., Rosenbaum, D. M., Kobilka, T. S., Thian, F. S., Edwards, P. C., Burghammer, M., Ratnala, V. R., Sanishvili, R., Fischetti, R. F., Schertler, G. F., Weis, W. I., and Kobilka, B. K. (2007) Crystal structure of the human beta2 adrenergic G-protein-coupled receptor. Nature 450, 383–387. (25) Ostrowski, J., Kjelsberg, M. A., Caron, M. G., and Lefkowitz, R. J. (1992) Mutagenesis of the beta2-Adrenergic Receptor: How Structure Elucidates Function. Annu. Rev. Pharmacol. Toxicol. 32, 167–183. (26) Sibley, D. R., and Monsma, F. J. (1992) Molecular biology of dopamine receptors. Trends Pharmacol. Sci. 13, 61–69. (27) Missale, C., Nash, S. R., Robinson, S. W., Jaber, M., and Caron, M. G. (1998) Dopamine Receptors: From Structure to Function. Physiol. Rev. 78, 189–225. (28) Schotte, A., Janssen, P. F. M., Gommeren, W., Luyten, W. H. M. L., Gompel, P., Lesage, A. S., Loore, K., and Leysen, J. E. (1996) Risperidone compared with new and reference antipsychotic drugs: in vitro and in vivo receptor binding. Psychopharmacology (Berl). 124, 57–73. (29) Joyce, J. N. (2001) Dopamine D3 receptor as a therapeutic target for antipsychotic and antiparkinsonian drugs. Pharmacol. Ther. 90, 231–259. (30) Prasad, S. S., Kojic, L. Z., Li, P., Mitchell, D. E., Hachisuka, A., Sawada, J., Gu, Q., and Cynader, M. S. (2002) Gene expression patterns during enhanced periods of visual cortex plasticity. Neuroscience 111, 35–45. (31) Shimohama, S., Sawada, H., Kitamura, Y., and Taniguchi, T. (2003) Disease model: Parkinson’s disease. Trends Mol. Med. 9, 360–365. (32) Zarate, C. a, Payne, J. L., Singh, J., Quiroz, J. a, Luckenbaugh, D. a, Denicoff, K. D., Charney, D. S., and Manji, H. K. (2004) Pramipexole for bipolar II depression: a placebo-controlled proof of concept study. Biol. Psychiatry 56, 54–60. (33) Nakajima, S., Gerretsen, P., Takeuchi, H., Caravaggio, F., Chow, T., Le Foll, B., Mulsant, B., Pollock, B., and Graff-Guerrero, A. (2013) The potential role of dopamine D3 receptor neurotransmission in cognition. Eur. Neuropsychopharmacol. 23, 799–813. (34) Micheli, F., and Heidbreder, C. (2013) Dopamine D3 receptor


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

antagonists: a patent review (2007 - 2012). Expert Opin. Ther. Pat. 23, 363– 81. (35) te Beek, E. T., Moerland, M., de Boer, P., van Nueten, L., de Kam, M. L., Burggraaf, J., Cohen, A. F., and van Gerven, J. M. a. (2012) Pharmacokinetics and central nervous system effects of the novel dopamine D2 receptor antagonist JNJ-37822681. J. Psychopharmacol. 26, 1119–27. (36) Joyce, J. N., and Meador-Woodruff, J. H. (1997) Linking the Family of D2 Receptors to Neuronal Circuits in Human Brain: Insights into Schizophrenia. Neuropsychopharmacology 16, 375–384. (37) Seeman, P., Wilson, A., Gmeiner, P., and Kapur, S. (2006) Dopamine D2 and D3 receptors in human putamen, caudate nucleus, and globus pallidus. Synapse 60, 205–211. (38) Sokoloff, P., Giros, B., Martres, M.-P., Bouthenet, M.-L., and Schwartz, J.-C. (1990) Molecular cloning and characterization of a novel dopamine receptor (D3) as a target for neuroleptics. Nature 347, 146–151. (39) Chen, J., Levant, B., and Wang, S. (2012) High-affinity and selective dopamine D3 receptor full agonists. Bioorganic Med. Chem. Lett. 22, 5612– 5617. (40) Ji, M., Chen, J., Ding, K., Wu, X., Varady, J., Levant, B., and Wang, S. (2005) Design, synthesis and structure–activity relationship studies of hexahydropyrazinoquinolines as a novel class of potent and selective dopamine receptor 3 (D3) ligands. Bioorganic Med. Chem. Lett. 15, 1701– 1705. (41) Campiani, G., Butini, S., Fattorusso, C., Catalanotti, B., Gemma, S., Nacci, V., Morelli, E., Cagnotto, A., Mereghetti, I., Mennini, T., Carli, M., Minetti, P., Di Cesare, M. A., Mastroianni, D., Scafetta, N., Galletti, B., Stasi, M. A., Castorina, M., Pacifici, L., Vertechy, M., Serio, S. Di, Ghirardi, O., Tinti, O., and Carminati, P. (2003) Pyrrolo[1,3]benzothiazepine-Based Serotonin and Dopamine Receptor Antagonists. Molecular Modeling, Further Structure−Activity Relationship Studies, and Identification of Novel Atypical Antipsychotic Agents. J. Med. Chem. 47, 143–157.

Page 16 of 30

(54) Aranda, R., Villalba, K., Raviña, E., Masaguer, C. F., Brea, J., Areias, F., Domínguez, E., Selent, J., López, L., Sanz, F., Pastor, M., and Loza, M. I. (2008) Synthesis, Binding Affinity, and Molecular Docking Analysis of New Benzofuranone Derivatives as Potential Antipsychotics. J. Med. Chem. 51, 6085–6094. (55) Dezi, C., Brea, J., Alvarado, M., Raviña, E., Masaguer, C. F., Loza, M. I., Sanz, F., and Pastor, M. (2007) Multistructure 3D-QSAR Studies on a Series of Conformationally Constrained Butyrophenones Docked into a New Homology Model of the 5-HT2A Receptor. J. Med. Chem. 50, 3242–3255. (56) Raviña, E., Negreira, J., Cid, J., Masaguer, C. F., Rosa, E., Rivas, M. E., Fontenla, J. A., Loza, M. I., Tristán, H., Cadavid, M. I., Sanz, F., Lozoya, E., Carotti, A., and Carrieri, A. (1999) Conformationally Constrained Butyrophenones with Mixed Dopaminergic (D2) and Serotoninergic (5-HT2A, 5-HT2C) Affinities: Synthesis, Pharmacology, 3D-QSAR, and Molecular Modeling of (Aminoalkyl)benzo- and -thienocycloalkanones as Putative Atypical Antipsy. J. Med. Chem. 42, 2774–2797. (57) Abagyan, R., and Totrov, M. (2001) High-throughput docking for lead generation. Curr. Opin. Chem. Biol. 5, 375–382. (58) Brooijmans, N., and Kuntz, I. D. (2003) MOLECULAR RECOGNITION AND DOCKING ALGORITHMS. Annu. Rev. Biophys. Biomol. Struct. 32, 335–373. (59) Cho, A. E., Chung, J. Y., Kim, M., and Park, K. (2009) Quantum mechanical scoring for protein docking. J. Chem. Phys. 131, 134108. (60) Raha Peters, M. B., Wang, B., Yu, N., Wollacott, A. M., Westerhoff, L. M., and Merz, K. M., Jr., K. (2007) The role of quantum mechanics in structure-based drug design. Drug Discov. Today 12, 725–731. (61) Zhou Huang, D., and Caflisch, A., T. (2010) Quantum Mechanical Methods for Drug Design. Curr. Top. Med. Chem. 10, 33–45. (62) Boström, J., Hogner, A., and Schmitt, S. (2006) Do Structurally Similar Ligands Bind in a Similar Fashion? J. Med. Chem. 49, 6716–6725.

(42) Löber, S., Hübner, H., Tschammer, N., and Gmeiner, P. (2011) Recent advances in the search for D3- and D4-selective drugs: probes, models and candidates. Trends Pharmacol. Sci. 32, 148–157.

(63) Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S., and Olson, A. J. (2009) AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J. Comput. Chem. 30, 2785–2791.

(43) Carlsson, J., Coleman, R. G., Setola, V., Irwin, J. J., Fan, H., Schlessinger, A., Sali, A., Roth, B. L., and Shoichet, B. K. (2011) Ligand discovery from a dopamine D3 receptor homology model and crystal structure. Nat Chem Biol 7, 769–778.

(64) Vreven, T., Morokuma, K., Farkas, Ö., Schlegel, H. B., and Frisch, M. J. (2003) Geometry optimization with QM/MM, ONIOM, and other combined methods. I. Microiterations and constraints. J. Comput. Chem. 24, 760–769.

(44) Park, W.-K., Jeong, D., Yun, C.-W., Lee, S., Cho, H., Kim, G.-D., Koh, H. Y., Pae, A. N., Cho, Y. S., Choi, K. Il, Jung, J. Y., Jung, S. H., and Kong, J. Y. (2003) Pharmacological actions of a novel and selective dopamine D3 receptor antagonist, KCH-1110. Pharmacol. Res. 48, 615–622. (45) Kapur, S., and Seeman, P. (2000) Antipsychotic agents differ in how fast they come off the dopamine D2 receptors. Implications for atypical antipsychotic action. J. Psychiatry Neurosci. 25, 161–166. (46) Chien, E. Y. T., Liu, W., Zhao, Q., Katritch, V., Won Han, G., Hanson, M. A., Shi, L., Newman, A. H., Javitch, J. A., Cherezov, V., and Stevens, R. C. (2010) Structure of the Human Dopamine D3 Receptor in Complex with a D2/D3 Selective Antagonist. Science (80-. ). 330, 1091–1095. (47) Gao, A. M., Zhang, D. W., Zhang, J. Z. H., and Zhang, Y. (2004) An efficient linear scaling method for ab initio calculation of electron density of proteins. Chem. Phys. Lett. 394, 293–297. (48) Platania, C. B. M., Salomone, S., Leggio, G. M., Drago, F., and Bucolo, C. (2012) Homology Modeling of Dopamine D2 and D3 Receptors: Molecular Dynamics Refinement and Docking Evaluation. PLoS One 7, e44316. (49) Martin, Y., and Muchmore, S. (2012) Frozen out: molecular modeling in the age of cryocrystallography. J. Comput. Aided. Mol. Des. 26, 91–92. (50) Sousa, S. F., Fernandes, P. A., and Ramos, M. J. (2006) Protein–ligand docking: Current status and future challenges. Proteins Struct. Funct. Bioinforma. 65, 15–26. (51) Król, M., Chaleil, R. A. G., Tournier, A. L., and Bates, P. A. (2007) Implicit flexibility in protein docking: Cross-docking and local refinement. Proteins Struct. Funct. Bioinforma. 69, 750–757. (52) Teague, S. J. (2003) Implications of protein flexibility for drug discovery. Nat Rev Drug Discov 2, 527–541. (53) Zanatta, G., Nunes, G., Bezerra, E. M., da Costa, R. F., Martins, A., Caetano, E. W. S., Freire, V. N., and Gottfried, C. (2014) Antipsychotic Haloperidol Binding to the Human Dopamine D3 Receptor: Beyond Docking Through QM/MM Refinement Toward the Design of Improved Schizophrenia Medicines. ACS Chem. Neurosci.

(65) Chung, L. W., Hirao, H., Li, X., and Morokuma, K. (2012) The ONIOM method: its foundation and applications to metalloenzymes and photobiology. Wiley Interdiscip. Rev. Comput. Mol. Sci. 2, 327–350. (66) Burger, S. K., Thompson, D. C., and Ayers, P. W. (2010) Quantum Mechanics/Molecular Mechanics Strategies for Docking Pose Refinement: Distinguishing between Binders and Decoys in Cytochrome c Peroxidase. J. Chem. Inf. Model. 51, 93–101. (67) Hirshfeld, F. L. (1977) Bounded-atom fragments for describing molecular charge densities. Theor. Chim. Acta 44, 129–138. (68) Zanatta, G., Barroso-Neto, I. L., Bambini-Junior, V., Dutra, M. F., Bezerra, E. M., da Costa, R. F., Caetano, E. W. S., Cavada, B. S., Freire, V. N., and Gottfried, C. (2012) Quantum Biochemistry Description of the Human Dopamine D3 Receptor in Complex with the Selective Antagonist Eticlopride. Proteomics Bioinforma. 5, 155–162. (69) Silva, A. M., Silva, B. P., Sales, F. A. M., Freire, V. N., Moreira, E., Fulco, U. L., Albuquerque, E. L., Maia Jr., F. F., and Caetano, E. W. S. (2012) Optical absorption and DFT calculations in L-aspartic acid anhydrous crystals: Charge carrier effective masses point to semiconducting behavior. Phys. Rev. B 86, 195201. (70) Nalewajski, R. F., and Parr, R. G. (2000) Information theory, atoms in molecules, and molecular similarity. Proc. Natl. Acad. Sci. 97, 8879–8882. (71) Kinkar Roy, R., Hirao, K., Krishnamurty, S., and Pal, S. (2001) Mulliken population analysis based evaluation of condensed Fukui function indices using fractional molecular charge. J. Chem. Phys. 115, 2901–2907. (72) Roy, R. K., Pal, S., and Hirao, K. (1999) On non-negativity of Fukui function indices. J. Chem. Phys. 110, 8236–8245. (73) Parr, R. G., and Yang, W. (1984) Density functional approach to the frontier-electron theory of chemical reactivity. J. Am. Chem. Soc. 106, 4049– 4050. (74) Mulliken, R. S. (1955) Electronic Population Analysis on LCAO–MO Molecular Wave Functions. I. J. Chem. Phys. 23, 1833–1840.


16

Page 17 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(75) Foster, J. P., and Weinhold, F. (1980) Natural hybrid orbitals. J. Am. Chem. Soc. 102, 7211–7218. (76) Bonaccorsi, R., Scrocco, E., and Tomasi, J. (1970) Molecular SCF Calculations for the Ground State of Some Three‐Membered Ring Molecules: (CH2)3, (CH2)2NH, (CH2)2NH2+, (CH2)2O, (CH2)2S, (CH)2CH2, and N2CH2. J. Chem. Phys. 52, 5270–5284. (77) Davidson, E., and Chakravorty, S. (1992) A test of the Hirshfeld definition of atomic charges and moments. Theor. Chim. Acta 83, 319–330. (78) Bultinck, P., Van Alsenoy, C., Ayers, P. W., and Carbó-Dorca, R. (2007) Critical analysis and extension of the Hirshfeld atoms in molecules. J. Chem. Phys. 126, -. (79) Ayers, P. (2006) Information Theory, the Shape Function, and the Hirshfeld Atom. Theor. Chem. Acc. 115, 370–378. (80) Parr, R. G., Ayers, P. W., and Nalewajski, R. F. (2005) What Is an Atom in a Molecule? J. Phys. Chem. A 109, 3957–3959. (81) Verstraelen, T., Sukhomlinov, S. V, Van Speybroeck, V., Waroquier, M., and Smirnov, K. S. (2011) Computation of Charge Distribution and Electrostatic Potential in Silicates with the Use of Chemical Potential Equalization Models. J. Phys. Chem. C 116, 490–504. (82) Bultinck, P., Fias, S., Van Alsenoy, C., Ayers, P. W., and Carbó-Dorca, R. (2007) Critical thoughts on computing atom condensed Fukui functions. J. Chem. Phys. 127, -. (83) Delley, B. (2000) From molecules to solids with the DMol3 approach. J. Chem. Phys. 113, 7756–7764. (84) Chen, X., Zhang, Y., and Zhang, J. Z. H. (2005) An efficient approach for ab initio energy calculation of biopolymers. J. Chem. Phys. 122, 184105. (85) He, X., and Zhang, J. Z. H. (2005) A new method for direct calculation of total energy of protein. J. Chem. Phys. 122, 31103–31104. (86) Zhang, D. W., and Zhang, J. Z. H. (2003) Molecular fractionation with conjugate caps for full quantum mechanical calculation of protein--molecule interaction energy. J. Chem. Phys. 119, 3599–3605. (87) Hohenberg, P., and Kohn, W. (1964) Inhomogeneous Electron Gas. Phys. Rev. 136, B864. (88) Kohn, W., and Sham, L. J. (1965) Self-Consistent Equations Including Exchange and Correlation Effects. Phys. Rev. 140, A1133. (89) da Costa, R. F., Freire, V. N., Bezerra, E. M., Cavada, B. S., Caetano, E. W. S., de Lima Filho, J. L., and Albuquerque, E. L. (2012) Explaining statin inhibition effectiveness of HMG-CoA reductase by quantum biochemistry computations. Phys. Chem. Chem. Phys. 14, 1389–1398. (90) Barroso-Neto, I. L., Marques, J. P. C., da Costa, R. F., Caetano, E. W. S., Cavada, B. S., Gottfried, C., and Freire, V. N. (2012) Inactivation of Ovine Cyclooxygenase-1 by Bromoaspirin and Aspirin: A Quantum Chemistry Description. J. Phys. Chem. B 116, 3270–3279. (91) Reckien, W., Janetzko, F., Peintinger, M. F., and Bredow, T. (2012) Implementation of empirical dispersion corrections to density functional theory for periodic systems. J. Comput. Chem. 33, 2023–2031. (92) Hjerde, E., Dahl, S. G., and Sylte, I. (2005) Atypical and typical antipsychotic drug interactions with the dopamine D2 receptor. Eur. J. Med. Chem. 40, 185–194. (93) Malmberg, A., Nordvall, G., Johansson, A. M., Mohell, N., and Hacksell, U. (1994) Molecular basis for the binding of 2-aminotetralins to human dopamine D2A and D3 receptors. Mol. Pharmacol. 46, 299–312. (94) Alberts, G. L., Pregenzer, J. F., and Bin Im, W. (1998) Contributions of cysteine 114 of the human D3 dopamine receptor to ligand binding and sensitivity to external oxidizing agents. Br. J. Pharmacol. 125, 705–710.

Drugs From Dopamine D2 Receptors: An Explanation for Low Receptor Occupancy and Early Clinical Relapse Upon Withdrawal of Clozapine or Quetiapine. Am. J. Psychiatry 156, 876–884. (99) Huey, R., Morris, G. M., Olson, A. J., and Goodsell, D. S. (2007) A semiempirical free energy force field with charge-based desolvation. J. Comput. Chem. 28, 1145–1152. (100) Morris, G. M., Goodsell, D. S., Halliday, R. S., Huey, R., Hart, W. E., Belew, R. K., and Olson, A. J. (1998) Automated docking using a Lamarckian genetic algorithm and an empirical binding free energy function. J. Comput. Chem. 19, 1639–1662. (101) Sanner, M. F. (1999) Python: A Programming Language for Software Integration and Development. J. Mol. Graph. Mod. 17, Februa, 57–61. (102) Tao, P., Fisher, J. F., Shi, Q., Vreven, T., Mobashery, S., and Schlegel, H. B. (2009) Matrix Metalloproteinase 2 Inhibition: Combined Quantum Mechanics and Molecular Mechanics Studies of the Inhibition Mechanism of (4-Phenoxyphenylsulfonyl)methylthiirane and Its Oxirane Analogue. Biochemistry 48, 9839–9847. (103) Frisch, M. J. ., Trucks, G. W. ., Schlegel, H. B. ., Scuseria, G. E. ., Robb, M. A. ., Cheeseman, J. R. ., Scalmani, G. ., Barone, V. ., Mennucci, B. ., Petersson, G. A. ., Nakatsuji, H. ., Caricato, M. ., Li H. P.;, X. . H., Izmaylov, A. F. ., Bloino G.;, J. . Z., Sonnenberg, J. L. ., Hada M.;, M. . E., Toyota, K. ., Fukuda, R. ., Hasegawa, J. ., Ishida, M. ., Nakajima, T. ., Honda, Y. ., Kitao, O. ., Nakai, H. ., Vreven, T. ., Montgomery J. A.;, J., Peralta, J. E. ., Ogliaro, F. ., Bearpark, M. ., Heyd, J. J. ., Brothers, E. ., Kudin, K. N. ., Staroverov, V. N. ., Kobayashi, R. ., Normand, J. ., Raghavachari, K. ., Rendell, A. ., Burant, J. C. ., Iyengar, S. S. ., Tomasi, J. ., Cossi, M. ., Rega, N. ., Millam, J. M. ., Klene, M. ., Knox, J. E. ., Cross, J. B. ., Bakken, V. ., Adamo, C. ., Jaramillo, J. ., Gomperts, R. ., Stratmann, R. E. ., Yazyev, O. ., Austin, A. J. ., Cammi, R. ., Pomelli, C. ., Ochterski, J. W. ., Martin, R. L. ., Morokuma, K. ., Zakrzewski, V. G. ., Voth, G. A. ., Salvador, P. ., Dannenberg, J. J. ., Dapprich, S. ., Daniels, A. D. ., Farkas, Ö. ., Foresman, J. B. ., Ortiz, J. V. ., Cioslowski, J. ., and Fox, D. J. Gaussian 09, Revision A.1. Gaussian, Inc., Wallingford CT, 2009. (104) Zhao, Y., and Truhlar, D. G. (2008) Exploring the Limit of Accuracy of the Global Hybrid Meta Density Functional for Main-Group Thermochemistry, Kinetics, and Noncovalent Interactions. J. Chem. Theory Comput. 4, 1849– 1868. (105) Zhao, Y., and Truhlar, D. (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other function. Theor. Chem. Acc. 120, 215–241. (106) Ceperley, D. M., and Alder, B. J. (1980) Ground State of the Electron Gas by a Stochastic Method. Phys. Rev. Lett. 45, 566–569. (107) Perdew, J. P., and Zunger, A. (1981) Self-interaction correction to density-functional approximations for many-electron systems. Phys. Rev. B 23, 5048–5079. (108) Perdew, J. P., Burke, K., and Ernzerhof, M. (1996) Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 77, 3865–3868. (109) Tkatchenko, A., and Scheffler, M. (2009) Accurate Molecular Van Der Waals Interactions from Ground-State Electron Density and Free-Atom Reference Data. Phys. Rev. Lett. 102, 73005. (110) Perdew, J. P., and Wang, Y. (1992) Accurate and simple analytic representation of the electron-gas correlation energy. Phys. Rev. B 45, 13244–13249. (111) Delley, B. (1990) An all-electron numerical method for solving the local density functional for polyatomic molecules. J. Chem. Phys. 92, 508–517. (112) Gordon, M. S., Fedorov, D. G., Pruitt, S. R., and Slipchenko, L. V. (2011) Fragmentation Methods: A Route to Accurate Calculations on Large Systems. Chem. Rev. 112, 632–672.

(95) Kortagere, S., Cheng, S.-Y., Antonio, T., Zhen, J., Reith, M. E. A., and Dutta, A. K. (2011) Interaction of novel hybrid compounds with the D3 dopamine receptor: Site-directed mutagenesis and homology modeling studies. Biochem. Pharmacol. 81, 157–163.

(113) Van Der Spoel, D., Lindahl, E., Hess, B., Groenhof, G., Mark, A. E., and Berendsen, H. J. C. (2005) GROMACS: Fast, flexible, and free. J. Comput. Chem.

(96) Chang, M. W., Belew, R. K., Carroll, K. S., Olson, A. J., and Goodsell, D. S. (2008) Empirical entropic contributions in computational docking: Evaluation in APS reductase complexes. J. Comput. Chem. 29, 1753–1761.

(114) Oostenbrink, C., Soares, T. A., Van Der Vegt, N. F. A., and Van Gunsteren, W. F. (2005) Validation of the 53A6 GROMOS force field. Eur. Biophys. J. 34, 273–284.

(97) Vauquelin, G., Bostoen, S., Vanderheyden, P., and Seeman, P. (2012) Clozapine, atypical antipsychotics, and the benefits of fast-off D2 dopamine receptor antagonism. Naunyn. Schmiedebergs. Arch. Pharmacol. 385, 337– 372.

(115) Schrodinger Version 1.3r1.

LLC. (2010) The PyMOL Molecular Graphics System,

(98) Seeman, P., and Tallerico, T. (1999) Rapid Release of Antipsychotic


17


Page 18 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

18

Page 19 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Captions Figure 1. (a) Atom labeling of risperidone in protonated form. Region i has the 6-Fluoro-1,2 benzoxazole fragment; region ii has the piperidin-1-yl fragment with the tertiary amine protonated, as observed at physiological pH; region iii has the ethyl fragment; and region iv has the 2-methyl-6,7,8,9-tetrahydro-4H-pyrido[1,2-a]pyrimidin-4-one fragment; (b) Risperidone protonation state at physiological pH; (c) DFT electron density projected onto an electrostatic potential isosurface showing negatively charged regions in red and positively charged regions in blue of the protonated risperidone state at physiological pH. Figure 2. Distinct orientations of risperidone bound to D3R after docking refinement using QM/MM optimization. (A) Stick representation of risperidone orientation 1 in the binding pocket of D3R. Risperidone’s fluorine atom is near to TMH5. (B) Stick representation of risperidone orientation 2 in the binding pocket of D3R. In orientation 2, the fluorine atom of risperidone is oriented towards a region between TMH1, TMH2 and TMH7. Figure 3. D3R-risperidone spatial arrangement before and after QM/MM optimization. (A) Conformational adjustment of Phe106 and Tyr 365 residues with respect to risperidone in orientation 1; (B) Adjustment of risperidone in orientation 1, Asp110, Phe345, Phe346 and Thr369, among other residues; (C) Adjustment of risperidone in orientation 2, His349 and Tyr365; (D) Adjustment of residues Phe345, Phe346 and Tyr373 relative to risperidone in orientation 2. Docking results with the frozen receptor are colored in gray and structures generated after QM/MM optimizations are colored by atom type. Figure 4. Risperidone-D3R total interaction energy as a function of the binding pocket radius using the GGA-TS exchange-correlation functional. Blue squares, red triangles and green circles represent the CRDI, CLDI and QMDI values, respectively. Results for risperidone orientation 1 (RO1) are shown in the top panel (solid symbols), while risperidone orientation 2 (RO2) data (open symbols) are depicted in the bottom panel. Figure 5. BIRD panel showing the interaction energy of each amino acid residue with risperidone orientation 1 using the GGA-TS approximation. Figure 6. BIRD panel showing the interaction energy of each amino acid residue with risperidone orientation 2 using the GGA-TS approximation. Figure 7. DFT (GGA-TS) electrostatic potential isosurfaces for the RO1-QMDI structure. (A) Colors indicate projected electron densities for the main interacting residues at the binding pocket of D3R. The risperidone molecule is displayed but does not interact with D3R; (B) Top: attractive residues Asp110, Glu90, Phe345, Phe346, His349, Tyr365 and Ser 366. Bottom: a different view of the binding pocket. In it, one can see the repulsive residues Val82 (TMH2) and Cys114 (TMH3). Figure 8. Electrostatic potential isosurfaces of the RO2-QMDI structure. (A) Projected electron densities for the main interacting residues in the empty binding pocket of D3R. Negative charge concentrates around Asp110, Tyr373, His349 and Ser366. Electrostatic potentials were calculated without taking into account the risperidone molecule, which was inserted only for descriptive purposes; (B) Top: at the right side attractive interactions with Asp110, Met83 and Tyr373 dominate (the last residue is shown in the reversed panel at the bottom), taking into account the presence of the risperidone molecule. The repulsive residue Cys114 is also depicted. Residue His349 is shown at the left side of the figure together with the attractive residues Ile183, Val189, Phe345 (bottom), Phe346 (bottom), Tyr365 and Ser366. Figure 9. RMSD of D3 receptor (D3R) and risperidone orientation 1 (RO1) during 50 ns of molecular dynamics. Simulation was performed in duplicate and is represented as MD1 or MD2. Figure 10. RMSD of D3 receptor (D3R) and risperidone orientation 2 (RO2) during 50 ns of molecular dynamics. Simulation was performed in duplicate and is represented as MD1 or MD2 Table 1. Individual contributions of amino acid residues to the binding of risperidone orientation 1 to D3R. The minimum binding pocket radius size which includes each residue is shown as well. Energies calculated using LDA-OBS and GGA-TS are expressed in kcal/mol. Table 2. Individual contributions of amino acid residues to the binding of risperidone orientation 2 to D3R. The minimum binding pocket radius size which includes each residue is shown as well. Energies calculated using LDA-OBS and GGA-TS are expressed in kcal/mol.


19


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. (a) Atom labeling of risperidone in protonated form. Region i has the 6-Fluoro-1,2 benzoxazole fragment; region ii has the piperidin-1-yl fragment with the tertiary amine protonated, as observed at physiological pH; region iii has the ethinyl fragment; and region iv has the 2-methyl-6,7,8,9-tetrahydro-4Hpyrido[1,2-a]pyrimidin-4-one fragment; (b) Risperidone protonation state at physiological pH; (c) DFT electron density projected onto an electrostatic potential isosurface showing negatively charged regions in red and positively charged regions in blue of the protonated risperidone state at physiological pH. Figure 1 90x142mm (300 x 300 DPI)


Page 20 of 30

Page 21 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2. Distinct orientations of risperidone bound to D3R after docking refinement using QM/MM optimization. (A) Stick represen-tation of risperidone orientation 1 in the binding pocket of D3R. Risperidone’s fluorine atom is near to TMH5. (B) Stick representa-tion of risperidone orientation 2 in the binding pocket of D3R. In orientation 2, the fluorine atom of risperidone is oriented towards a region between TMH1, TMH2 and TMH7. Figure 2 266x93mm (72 x 72 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 3. D3R-risperidone spatial arrangement before and after QM/MM optimization. (A) Conformational adjustment of Phe106 and Tyr 365 residues with respect to risperidone in orientation 1; (B) Adjustment of risperidone in orientation 1, Asp110, Phe345, Phe346 and Thr369, among other residues; (C) Adjustment of risperidone in orientation 2, His349 and Tyr365; (D) Adjustment of residues Phe345, Phe346 and Tyr373 relative to risperidone in orientation 2. Docking results with the frozen receptor are colored in gray and structures generated after QM/MM optimizations are colored by atom type. Figure 3 170x139mm (300 x 300 DPI)


Page 22 of 30

Page 23 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. Risperidone-D3R total interaction energy as a function of the binding pocket radius using the GGATS exchange-correlation functional. Blue squares, red triangles and green circles represent the CRDI, CLDI and QMDI values, respective-ly. Results for risperidone orientation 1 (RO1) are shown in the top panel (solid symbols), while risperidone orientation 2 (RO2) data (open symbols) are depicted in the bottom panel. Figure 4 85x121mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 5. BIRD panel showing the interaction energy of each amino acid residue with risperidone orientation 1 using the GGA-TS approximation. Figure 5 89x96mm (300 x 300 DPI)


Page 24 of 30

Page 25 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 6. BIRD panel showing the interaction energy of each amino acid residue with risperidone orientation 2 using the GGA-TS approximation. Figure 6 89x94mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 7. DFT (GGA-TS) electrostatic potential isosurfaces for the RO1-QMDI structure. (A) Colors indicate projected elec-tron densities for the main interacting residues at the binding pocket of D3R. The risperidone molecule is displayed but does not interact with D3R; (B) Top: attractive residues Asp110, Glu90, Phe345, Phe346, His349, Tyr365 and Ser 366. Bottom: a different view of the binding pocket. In it, one can see the repulsive residues Val82 (TMH2) and Cys114 (TMH3). Figure 7 90x144mm (300 x 300 DPI)


Page 26 of 30

Page 27 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8. Electrostatic potential isosurfaces of the RO2-QMDI structure. (A) Projected electron densities for the main interact-ing residues in the empty binding pocket of D3R. Negative charge concentrates around Asp110, Tyr373, His349 and Ser366. Electrostatic potentials were calculated without taking into account the risperidone molecule, which was inserted only for descriptive purposes; (B) Top: at the right side attractive interactions with Asp110, Met83 and Tyr373 dominate (the last residue is shown in the reversed panel at the bottom), taking into account the presence of the risperidone molecule. The repulsive residue Cys114 is also depicted. Residue His349 is shown at the left side of the figure together with the attractive residues Ile183, Val189, Phe345 (bottom), Phe346 (bottom), Tyr365 and Ser366. Figure 8 90x153mm (300 x 300 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 9. RMSD of D3 receptor (D3R) and risperidone orienta-tion 1 (RO1) during 40 ns of molecular dynamics. Simulation was performed in duplicate and is represented as MD1 or MD2. Figure 9 416x306mm (72 x 72 DPI)


Page 28 of 30

Page 29 of 30

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 10. RMSD of D3 receptor (D3R) and risperidone orien-tation 2 (RO2) during 40 ns of molecular dynamics. Simulation was performed in duplicate and is represented as MD1 or MD2 Figure 10 416x298mm (72 x 72 DPI)



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For Table of Contents Use Only Two Binding Geometries for Risperidone in Dopamine D3 Receptors: Insights on the Fast-Off Mechanism Through Docking, Quantum Biochemistry and Molecular Dynamics Simulations” by Geancarlo Zanatta, Gustavo Della Flora Nunes, Eveline M. Bezerra, Roner F. da Costa, Alice Martins, Ewerton W. S. Caetano, Valder N. Freire, Carmem Gottfried For Table of Contents Use Only 247x79mm (72 x 72 DPI)


Page 30 of 30

Two Binding Geometries for Risperidone in Dopamine D3 Receptors

Recommend Documents