D1 Agonist and D2 Antagonist Dual Effect of SPD 1437 molecular energies at the DFT/B3LYP/6-311G**level using helices, especially TM6 and TM7, SPD-Dl complex is quite the Gaussian 98 program. The transition state conformation relaxed. Among seven TM helices in the SPD-DI complex of SPd between the agonistic and antagonistic conforma- TM6 and TM7 have the largest outward movement toward tions was located by the synchronous transit-guided quasi- the surrounding lipids(Fig. 9 A). On the other hand, the Newton methods(46)at the same level. Fig. 7 shows that SPD-D2 complex is quite conservative(Fig 9, B and D)in there is only a small energy barrier, 1.55 kcal/mol, between the motion of its helices. The dramatic movement of the the conformation of SPD in the DI complex and the transi- TM6-TM7 helices in the SPD-DI complex appears to result tion state conformation. The energy barrier between the con- from the agonistic effect of SPD and distally propagates into formation of SPD in the D2 complex and the transition state the intracellular end. conformation is 1.02 kcal/mol. These small energy differ ences demonstrate that the transition of SPD from the agonist DISCUSSION conformation to antagonist one is relatively facile. In other words, SPD intrinsically has the capacity for dual functional In this article we integrated homology modeling, MD sim- when associating with different DRs(dI and D2) ulation, automated docking, and density function theory to understand the agonistic and antagonistic mechanism of SPD and the activation process of the DI receptor. So far, only the Dynamical helix movement of the D1 receptor structure of the bovine rhodopsin receptor was characterized upon SPD binding in the GPCR family. The identities between rhodopsin and In corresponding to the agonizing and antagonizing effects DRS DI and D2 are%. Though we carefully built DI of SPD, the TM helices of DI and D2 take different extents and D2 models and checked them by several programs of of motion. Taking the radius of gyration(Rg)as an example structural validation and all available experimental data, the (Fig. 8), the Rg of helices 1-7 in the SPD-DI complex models are still fuzzy. Furthermore, the timescale of our MD gradually increases after -3.0 ns and maintains a value of simulations is short compared to the real biological process 1.775 nm from there until the end of the MD simulations. In of DRs. However, the current models are supported by most comparison, the Rg value of the SPD-D2 complex changes known mutagenesis experiments, and our MD simulations slightly and maintains almost the same value as for the free give the trend of the motion of the proteins. Since it is very D2 receptor. Usually, the larger the Rg, the greater the extent difficult to characterize the structure of membrane protein of the range of motion of the structure. The essential dynamics and its dynamical characteristic experimentally, it is a useful analyses(47) performed on the complete MD trajectories of complement for investigation of structural and functional the complexes helped us locate the most drastic motic characteristic of the GPCR family and provides clues in among helices. Two projected structures representing the understanding the signal transduction process of DRs. In minimal and maximal amplitudes along the first eigenvector particular, the obtained agonistic and antagonistic mecha- were selected and superimposed with each other(Fig 9). The nism of SPD can provide practical guidance for the design of relative motion among helices is different before and after dual function lead compounds for Dl/D2 receptors PD binding in two receptors. In comparison with the SPD D2 complex, the binding of SPD to the DI receptor(Fig 9, A Interpretation of mutation data versus C)results in the outward movement of seven TM Without high resolution structural information for the di and D2 receptors, site-directed mutagenesis is usually employed to explore the molecular mechanism of receptor activation and signal transduction by ligand binding. To date, several residues in either the DI or D2 receptors have been revealed to be critical for ligand binding. For example, D-3.32 1.53 kcalmo (D-3 32G, D-332N, and D-332C)has been shown to be critical to SPD binding for both the DI and D2 receptors via mutagenesis studies (5,6). Consistent with experimental data, our models for both SPD complexes with the DI and D2 receptors demonstrate that D-3.32 acts as a hydrogen- bond acceptor and forms electrostatic interactions with the protonated SPD. Our complex models also show that the side-chain hydroxyl group of either S-5.42 or S-5.46 forms a hydrogen-bond with the hydroxyl group on ring D of SPD in FIGURE 7 Energy barrier between the agonistic and antagonistic con both the DI and D2 complexes. Removing the hydroxyl formers of SPD calculated with B3LYP/6311g**. The calculated energy is group of S-5.42 or S-5.46 will undoubtedly reduce the 1.55 kcal/mol after the zero-point vibrational energy (ZPVE)correction. binding affinity of SPD with these two DRs. Cox et al. (40) Biophysical Joumal 93(5)1431-1441molecular energies at the DFT/B3LYP/6-311G** level using the Gaussian 98 program. The transition state conformation of SPD between the agonistic and antagonistic conforma￾tions was located by the synchronous transit-guided quasi￾Newton methods (46) at the same level. Fig. 7 shows that there is only a small energy barrier, 1.55 kcal/mol, between the conformation of SPD in the D1 complex and the transi￾tion state conformation. The energy barrier between the con￾formation of SPD in the D2 complex and the transition state conformation is 1.02 kcal/mol. These small energy differ￾ences demonstrate that the transition of SPD from the agonist conformation to antagonist one is relatively facile. In other words, SPD intrinsically has the capacity for dual functional when associating with different DRs (D1 and D2). Dynamical helix movement of the D1 receptor upon SPD binding In corresponding to the agonizing and antagonizing effects of SPD, the TM helices of D1 and D2 take different extents of motion. Taking the radius of gyration (Rg) as an example (Fig. 8), the Rg of helices 1–7 in the SPD-D1 complex gradually increases after ;3.0 ns and maintains a value of 1.775 nm from there until the end of the MD simulations. In comparison, the Rg value of the SPD-D2 complex changes slightly and maintains almost the same value as for the free D2 receptor. Usually, the larger the Rg, the greater the extent of the range of motion of the structure. The essential dynamics analyses (47) performed on the complete MD trajectories of the complexes helped us locate the most drastic motions among helices. Two projected structures representing the minimal and maximal amplitudes along the first eigenvector were selected and superimposed with each other (Fig. 9). The relative motion among helices is different before and after SPD binding in two receptors. In comparison with the SPD￾D2 complex, the binding of SPD to the D1 receptor (Fig. 9, A versus C) results in the outward movement of seven TM helices, especially TM6 and TM7, SPD-D1 complex is quite relaxed. Among seven TM helices in the SPD-D1 complex, TM6 and TM7 have the largest outward movement toward the surrounding lipids (Fig. 9 A). On the other hand, the SPD-D2 complex is quite conservative (Fig. 9, B and D) in the motion of its helices. The dramatic movement of the TM6-TM7 helices in the SPD-D1 complex appears to result from the agonistic effect of SPD and distally propagates into the intracellular end. DISCUSSION In this article we integrated homology modeling, MD sim￾ulation, automated docking, and density function theory to understand the agonistic and antagonistic mechanism of SPD and the activation process of the D1 receptor. So far, only the structure of the bovine rhodopsin receptor was characterized in the GPCR family. The identities between rhodopsin and DRs D1 and D2 are ;25%. Though we carefully built D1 and D2 models and checked them by several programs of structural validation and all available experimental data, the models are still fuzzy. Furthermore, the timescale of our MD simulations is short compared to the real biological process of DRs. However, the current models are supported by most known mutagenesis experiments, and our MD simulations give the trend of the motion of the proteins. Since it is very difficult to characterize the structure of membrane protein and its dynamical characteristic experimentally, it is a useful complement for investigation of structural and functional characteristic of the GPCR family and provides clues in understanding the signal transduction process of DRs. In particular, the obtained agonistic and antagonistic mecha￾nism of SPD can provide practical guidance for the design of dual function lead compounds for D1/D2 receptors. Interpretation of mutation data Without high resolution structural information for the D1 and D2 receptors, site-directed mutagenesis is usually employed to explore the molecular mechanism of receptor activation and signal transduction by ligand binding. To date, several residues in either the D1 or D2 receptors have been revealed to be critical for ligand binding. For example, D-3.32 (D-3.32G, D-3.32N, and D-3.32C) has been shown to be critical to SPD binding for both the D1 and D2 receptors via mutagenesis studies (5,6). Consistent with experimental data, our models for both SPD complexes with the D1 and D2 receptors demonstrate that D-3.32 acts as a hydrogen￾bond acceptor and forms electrostatic interactions with the protonated SPD. Our complex models also show that the side-chain hydroxyl group of either S-5.42 or S-5.46 forms a hydrogen-bond with the hydroxyl group on ring D of SPD in both the D1 and D2 complexes. Removing the hydroxyl group of S-5.42 or S-5.46 will undoubtedly reduce the binding affinity of SPD with these two DRs. Cox et al. (40) FIGURE 7 Energy barrier between the agonistic and antagonistic con￾formers of SPD calculated with B3LYP/6311g**. The calculated energy is 1.55 kcal/mol after the zero-point vibrational energy (ZPVE) correction. D1 Agonist and D2 Antagonist Dual Effect of SPD 1437 Biophysical Journal 93(5) 1431–1441