The Journal of Physical Chemistry B Article prefrontal cortex (mPFC), which is accompanied by D2R MATERIALS AND METHODS hyperactivity in subcortical regions such as the ventral tegmental area(VTA)and the nucleus accumbens(Nac) Homology Modeling of D1R and D2R.The quences of DiR and D2R were collected from the The dir dysfunction is believed to be responsible for the UniProtKB database (accession code: P21728 for DIR and negative symptoms of schizophrenia, whereas the D2R P14416 for D2R), and sequence similarity searches for DIR hyperactivity might lead to the positive symptoms of the and D2R were performed using the NCBI BLAST server. 26 disorder.3-15 Thus, an effective antipsychotic drug should The newly disclosed active-state structure of P2AR (PDB code simultaneously possess DiR agonistic and D2R antagonistic 3POG) was selected as the template to construct the agonistic activities . I5 conformation of DiR, while for D2R homology modeling, the ()-Stepholidine (I-SPD), an active natural compoun inactive-state structure of D3R(PDB code: 3PBL) was used isolated from the Chinese herb Stephania, is to date the only as the template as it shares a sequence identity of -78% with drug with a dual effect as a DiR agonist and D2R antagonist. [- D2R in transmembrane(Tm)regions. Sequence alignments of spd displays high affinity to DI-and D2-like receptors but low DIR and D2R with their templates were carried out us affinity to 5-HT, receptors and rare affinity to several other ClustalX 2.0.11 program. The homology modeling and loop neurotransmitter receptors. 5-17 Moreover, clinical investiga refinement were optimized with Modeler 9v4. Fifty homology tions showed that I-SPD is superior to perphenazine in models were generated for both DiR and D2R after loop antipsychotic efficacy. Unlike perphenazine, I-SPD does not refinement, and the one with the lowest dope score adopted for subsequent energy minimization. The canonical induce any enound in studying the dual action mechanism of disulfide bridge between residues C3.25(in Ballesteros- a chemical for developing novel antipsychotics. 18 The recent Weinstein numbering, a single most conserved residue among discovery that I-SPD has an antagonistic effect on D3R the class A GPCRs is designated x 50, where x is the endered both 1-SPD and D3R much more intriguing. However transmembrane helix number. all other residues on that helix are numbered relative to this conserved position)at the tip of SpD still remains obscure despite it was partly explained by our TM3, and cysteine of ECL2 in the midle of extracellular loop preliminary computational study. A more comprehensive understanding of the triple action mechanism of I-SPD against 1000 cycles of conjugate grades ycles of steepest descent and nodels were subjected to 1000 nt energy minimization. The DIR,D2R, and D3R at the atomic level is fundamental for procHEck program which analyzes the residue-by-residue developing superior antipsychotics. Obviously, the 3D geometry and overall structure quality was employed to structures of dir and D2R dispensable for computa evaluate the stereochemical quality of the energy-minimized tiona ing out the triple action mechanism. Unfortu- models of DIR and D2R. The starting structure of D3R in the nately, only the 3D structure of D3R is currently available simulation study comes from the crystal structure(PDB code It has been widely recognized that molecular modeling and 3PBL) simulation are an excellent complement to experiments Parameterization of /SPD and Molecular Docking explaining experimental results and guiding further experiment The geometry of I-SPD was built based on its crystal structure' design. Furthermore, additional information that is not using the SYBYL 6.9 program and optimized at the DFT/ accessible by experiments can be revealed by computational B3LYP/6-31G* level. It features the same configuration and approaches conformation as our previous study 20 us study 0 The partial atomic computational homology modeling methods based on high charges of I-SPD was determined with the DFT/B3LYP/6- sequence identity can be employed to predict the 3D models 31G** basis set by using the Gaussian 09 program with enough accuracy, which can subsequently be used to aid in Parameterization and topology generation of l-SPD for the understanding of protein functional mechanisms. 23 CHARMM force field were performed by using the Paratool Molecular dynamics(MD)simulations, taking advantage of plug-in of VMD 33 The automated molecular docking program iteratively tracking the trajectory of conformational change, can AutoDock4.0 with AutoDock Tools was employed to probe the ide information about the conformational properties of possible complex of 1-SPD bound with energy-minimized nolecular system and the way in which the conformation dopamine receptors(DIR, D2R, and D3R). Binding energy changes with time, such as the binding process of a ligand into a prediction was carried out using X-Score. All available protein's binding pocket, the binding conformation that the experimental data in combination with the lowest binding ligand and protein adopt, and other time-dependent properties. free energy were used to select the initial structure with the recent insightful information about the activation subsequent MD simulations mechanism of GPCR, 4 as well as the advent of the active- System Setup. The orientation of DiR, D2R, and D3R state structure of B2AR+s and the inactive-state structure of was aligned to the z-axis using Orient plug-in of VMD.Helmut D3R, we are motivated to perform a more comprehensive Grubmuller's SoLvATE program was then used to solvate the omputational study on DIR, D2R, and D3R to explore proteins followed by elimination of water molecules located underlying agonistic and antagonistic mechanisms of 1-SPD the hydrophobic protein-membrane interface. A pre-equili- which is essential for developing superior antipsychotics. Thus, brated rectangular patch of POPC membrane bilayer with homology modeling, automated molecular docking, and MD solvated lipid headgroups was generated by using the mulations are integrated in this study with the eventual aim to Membrane Builder plug-in of VMD. The partially solvated develop novel superior antipsychotic agents with less side proteins were immersed in the POPC membrane bilayer by ffects and tailor these agents to possess the designe matching the center of mass of proteins with that of membrane, properties for individual therapy espectively. The resulting placement of each receptor in 8122 dx. dolora/o.021/p30492351 Phys. Chem. B2012116,8121-813prefrontal cortex (mPFC), which is accompanied by D2R hyperactivity in subcortical regions such as the ventral tegmental area (VTA) and the nucleus accumbens (NAc). The D1R dysfunction is believed to be responsible for the negative symptoms of schizophrenia, whereas the D2R hyperactivity might lead to the positive symptoms of the disorder.13−15 Thus, an effective antipsychotic drug should simultaneously possess D1R agonistic and D2R antagonistic activities.15 (−)-Stepholidine (l-SPD), an active natural compound isolated from the Chinese herb Stephania, is to date the only drug with a dual effect as a D1R agonist and D2R antagonist. l￾SPD displays high affinity to D1- and D2-like receptors but low affinity to 5-HT2 receptors and rare affinity to several other neurotransmitter receptors.15−17 Moreover, clinical investiga￾tions showed that l-SPD is superior to perphenazine in antipsychotic efficacy.18 Unlike perphenazine, l-SPD does not induce any extrapyramidal symptoms (EPS),19 making it an attractive compound in studying the dual action mechanism of a chemical for developing novel antipsychotics.18 The recent discovery that l-SPD has an antagonistic effect on D3R rendered both l-SPD and D3R much more intriguing. However, the complete molecular basis of the dual action mechanism of l￾SPD still remains obscure despite it was partly explained by our preliminary computational study.20 A more comprehensive understanding of the triple action mechanism of l-SPD against D1R, D2R, and D3R at the atomic level is fundamental for developing superior antipsychotics. Obviously, the 3D structures of D1R and D2R are indispensable for computa￾tionally figuring out the triple action mechanism. Unfortu￾nately, only the 3D structure of D3R is currently available. It has been widely recognized that molecular modeling and simulation are an excellent complement to experiments in explaining experimental results and guiding further experiment design.20,21 Furthermore, additional information that is not accessible by experiments can be revealed by computational approaches.22 In the absence of experimental structures, computational homology modeling methods based on high sequence identity can be employed to predict the 3D models with enough accuracy, which can subsequently be used to aid in the understanding of protein functional mechanisms.23 Molecular dynamics (MD) simulations, taking advantage of iteratively tracking the trajectory of conformational change, can provide information about the conformational properties of molecular system and the way in which the conformation changes with time, such as the binding process of a ligand into a protein’s binding pocket, the binding conformation that the ligand and protein adopt, and other time-dependent properties. With the recent insightful information about the activation mechanism of GPCR,24,25 as well as the advent of the active￾state structure of β2AR4,5 and the inactive-state structure of D3R,11 we are motivated to perform a more comprehensive computational study on D1R, D2R, and D3R to explore the underlying agonistic and antagonistic mechanisms of l-SPD, which is essential for developing superior antipsychotics. Thus, homology modeling, automated molecular docking, and MD simulations are integrated in this study with the eventual aim to develop novel superior antipsychotic agents with less side effects and tailor these agents to possess the designed properties for individual therapy. ■ MATERIALS AND METHODS Homology Modeling of D1R and D2R. The amino acid sequences of D1R and D2R were collected from the UniProtKB database (accession code: P21728 for D1R and P14416 for D2R), and sequence similarity searches for D1R and D2R were performed using the NCBI BLAST server.26 The newly disclosed active-state structure of β2AR (PDB code: 3P0G)4 was selected as the template to construct the agonistic conformation of D1R, while for D2R homology modeling, the inactive-state structure of D3R (PDB code: 3PBL)11 was used as the template as it shares a sequence identity of ∼78% with D2R in transmembrane (TM) regions. Sequence alignments of D1R and D2R with their templates were carried out using ClustalX 2.0.11 program.27 The homology modeling and loop refinement were optimized with Modeler 9v4.28 Fifty homology models were generated for both D1R and D2R after loop refinement, and the one with the lowest DOPE score28 was adopted for subsequent energy minimization. The canonical disulfide bridge between residues C3.25 (in Ballesteros− Weinstein numbering, a single most conserved residue among the class A GPCRs is designated x.50, where x is the transmembrane helix number. All other residues on that helix are numbered relative to this conserved position) at the tip of TM3, and cysteine of ECL2 in the middle of extracellular loop 2 (ECL2) was created by using SYBYL 6.9.29 The resultant models were subjected to 1000 cycles of steepest descent and 1000 cycles of conjugate gradient energy minimization. The PROCHECK30 program which analyzes the residue-by-residue geometry and overall structure quality was employed to evaluate the stereochemical quality of the energy-minimized models of D1R and D2R. The starting structure of D3R in the simulation study comes from the crystal structure (PDB code: 3PBL). Parameterization of l-SPD and Molecular Docking. The geometry of l-SPD was built based on its crystal structure31 using the SYBYL 6.9 program and optimized at the DFT/ B3LYP/6-31G** level. It features the same configuration and conformation as our previous study.20 The partial atomic charges of l-SPD was determined with the DFT/B3LYP/6- 31G** basis set by using the Gaussian 09 program.32 Parameterization and topology generation of l-SPD for CHARMM force field were performed by using the Paratool plug-in of VMD.33 The automated molecular docking program AutoDock4.0 with AutoDock Tools was employed to probe the possible complex of l-SPD bound with energy-minimized dopamine receptors (D1R, D2R, and D3R). Binding energy prediction was carried out using X-Score.34 All available experimental data in combination with the lowest binding free energy were used to select the initial structure for subsequent MD simulations. System Setup. The orientation of D1R, D2R, and D3R were adjusted using VMD, and their principal axis of symmetry was aligned to the z-axis using Orient plug-in of VMD. Helmut Grubmuller’s SOLVATE program was then used to solvate the proteins followed by elimination of water molecules located in the hydrophobic protein−membrane interface. A pre-equili￾brated rectangular patch of POPC membrane bilayer with solvated lipid headgroups was generated by using the Membrane Builder plug-in of VMD. The partially solvated proteins were immersed in the POPC membrane bilayer by matching the center of mass of proteins with that of membrane, respectively. The resulting placement of each receptor in The Journal of Physical Chemistry B Article 8122 dx.doi.org/10.1021/jp3049235 | J. Phys. Chem. B 2012, 116, 8121−8130