FULL PAPER Zhang et al. of H285A and H328A will lead to a decrease in binding The binding mode of a small-molecule inhibitor of with value of 26.27, and 73. 47 kJ/mol respec- hSMsI These results demonstrate the rationale of our D609 is essentially the only known small-molecule inhibitor of hSMSI so far. Interestingly, dimerization of Notably, our model also suggests some key residues D609, i.e. D609 dixanthogen(Figure 2), was observed which have not been well studied before. Mutation of to have no inhibition activity at all. Based on the R342A caused great loss of binding energy with value structural model of the hSMS I/SM complex derived in of 81.79 kJ/mol. This datum indicates that the hydrogen our study, a binding mode of D609 was derived through bonding interaction between SM and R342 is important molecular docking. Our model of the ASMS1/D609 for locking the conformation of SM to facilitate the complex(Figure 9)indicates that D609 occupies the catalyzed reaction. Mutation of Y338A also causes a catalytic pocket, which blocks the entry of substrate decrease in binding energy of ca. 21.08 kJ/mol. One can molecules, such as PC and SM. The key residues in- see in Figure 5A that Y338 can form favorable hydro- volved in the binding of D609 include R342 and H285 phobic interactions with one of the carbon chain on SM. In particular, the negative charged sulfur atom on D609 In addition, Y338 can form a hydrogen bond with R342 forms a strong salt-bridge interaction with R342, which In contrast, F173A, F177A, H274A, and L281A muta- mimics the two critical hydrogen bonds between tions do not have a major impact on the binding energy SMSI and SM. d609 dixanthogen does not have the of SM(Table 4). However, the quadruple mutant F173A negatively charged sulfur atom, leading to the loss of +F177A+H274A+ L281A leads to a large decrease of the salt-bridge with R342. In addition, the steric hin- 54. 14 kJ/mol in binding energy. It indicates that the hy- drance near R342 does not allow the second D609 drophobic interactions involving these residues are co- moiety on D609 dixanthogen to go into the binding operative even though the contribution of each individ- pocket deeply. Thus, our model provides a reasonable lal residue is trivial. These residues form a hydrophobic explanation on the different hSMSl inhibition activities patch surrounding the choline group of SM, illustrating between D609 and D609 dixanthogen the importance of choline group in the PC and SM. 0 Activation mechanism of hsMsi based on nma and Conclusions We have obtained a three-dimensional structural Protein domain motion analysis indicated that the model of hSMSI through homology modeling, molecu- extracellular loops of hSMSI rotate around the hinge lar docking, and extensive molecular dynamics simula axis shown in Figure 7. In order to further detect such tions. Our model reasonably explains how hSMs conformational motions, we performed normal mode transforms a phosphocholine moiety on ceramide to analysis(NMA)on the final snapshot on the MD tra oduce sphingomyelin. It also explains the high selec jectory of the hSMSI/SM complex. The two lowest- tivity of hSMsI towards phosphocholine and sphingo- frequency normal modes(i.e. modes 7 and 8 in Table 3) myelin. Normal mode analysis on the low-frequency are illustrated in Figure 8: one mode features bending of motions of hSMSI suggests a gating mechanism of the loop towards the transmembrane domain whereas immobilizing phosphatidylcholine at the catalytic site the other features rotating of the loop on top of the and releasing sphingomyelin afterwards. Computational transmembrane domain. This motion prompts that there mutagenesis results explore the key residues for the can be two major states existing in the physiological binding of SM to hSMSl. Currently, very few state: an open state and a closed state. These two states small-molecule inhibitors of hSMs I have been reporte are connected by low-frequency conformational mo- publicly in literature. The potential value of our struc- tions. This gating mechanism of the extracellular loops tural model is that it can be applied to the discovery of on hSMSl illustrates how the reactant PC is locked after effective hSMSI inhibitors, for example, through virtual it enters the catalytic pocket and how the product SM is eleased afterwards. It also illustrates how an inhibitor may block the function of hSMSl. In particular, the References choline moiety on PC can form favorable hydrophobic I Merrill, A H Jr, Jones, DD. Biochim. Biophys. Acta 1990, interactions with F173. F177. H274. and L281 in the 1044.1 closed state, which immobilizes the choline moiety and 2 Nagao, K, Takahashi, K, Hanada, K, Kioka, N; Matsuo, locks the shape of the catalytic center(Figure 5A). The M. Ueda, K.J. Bio. Chem. 2007, 282, 14868 lowest frequency motions that induce the movement of Sano, O, Kobayashi, A, Nagao, K, Kumagai, K, Kioka F173, F177, and H274 on the loops away from the N, Hanada, K; Ueda, K, Matsuo, M. J. Lipid Res 2007, binding site will disassemble the hydrophobic patch 48,2377 originally accommodating the choline moiety in the 4 Schutze, S: Potthoff, K. Machleidt, T, Berkovic, D closed state. Consequently, SM will be released from wiegmann, K, Kronke, M. Cell 1992, 71, 765 the catalytic site in the open state 5 Luberto. C. Yoo. D. S: Suidan. H. S: Bartoli. G. M 1574www.cjc.wiley-vch.deO2011SioC,Cas,Shanghai&WileY-VchVerlagGmbh&Co.Kgaa,WeinheimChin.j.chem.2011,29,1567-1575FULL PAPER Zhang et al. 1574 www.cjc.wiley-vch.de © 2011 SIOC, CAS, Shanghai, & WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Chin. J. Chem. 2011, 29, 1567—1575 of H285A and H328A will lead to a decrease in binding energy with value of 26.27, and 73.47 kJ/mol respec￾tively. These results demonstrate the rationale of our model further. Notably, our model also suggests some key residues which have not been well studied before. Mutation of R342A caused great loss of binding energy with value of 81.79 kJ/mol. This datum indicates that the hydrogen bonding interaction between SM and R342 is important for locking the conformation of SM to facilitate the catalyzed reaction. Mutation of Y338A also causes a decrease in binding energy of ca. 21.08 kJ/mol. One can see in Figure 5A that Y338 can form favorable hydro￾phobic interactions with one of the carbon chain on SM. In addition, Y338 can form a hydrogen bond with R342. In contrast, F173A, F177A, H274A, and L281A muta￾tions do not have a major impact on the binding energy of SM (Table 4). However, the quadruple mutant F173A ＋F177A＋H274A＋L281A leads to a large decrease of 54.14 kJ/mol in binding energy. It indicates that the hy￾drophobic interactions involving these residues are co￾operative even though the contribution of each individ￾ual residue is trivial. These residues form a hydrophobic patch surrounding the choline group of SM, illustrating the importance of choline group in the PC and SM.10 Activation mechanism of hSMS1 based on NMA and DynDom analysis Protein domain motion analysis indicated that the extracellular loops of hSMS1 rotate around the hinge axis shown in Figure 7. In order to further detect such conformational motions, we performed normal mode analysis (NMA) on the final snapshot on the MD tra￾jectory of the hSMS1/SM complex. The two lowest￾frequency normal modes (i.e. modes 7 and 8 in Table 3) are illustrated in Figure 8: one mode features bending of the loop towards the transmembrane domain whereas the other features rotating of the loop on top of the transmembrane domain. This motion prompts that there can be two major states existing in the physiological state: an open state and a closed state. These two states are connected by low-frequency conformational mo￾tions. This gating mechanism of the extracellular loops on hSMS1 illustrates how the reactant PC is locked after it enters the catalytic pocket and how the product SM is released afterwards. It also illustrates how an inhibitor may block the function of hSMS1. In particular, the choline moiety on PC can form favorable hydrophobic interactions with F173, F177, H274, and L281 in the closed state, which immobilizes the choline moiety and locks the shape of the catalytic center (Figure 5A). The lowest frequency motions that induce the movement of F173, F177, and H274 on the loops away from the binding site will disassemble the hydrophobic patch originally accommodating the choline moiety in the closed state. Consequently, SM will be released from the catalytic site in the open state. The binding mode of a small-molecule inhibitor of hSMS1 D609 is essentially the only known small-molecule inhibitor of hSMS1 so far. Interestingly, dimerization of D609, i.e. D609 dixanthogen (Figure 2), was observed to have no inhibition activity at all.13 Based on the structural model of the hSMS1/SM complex derived in our study, a binding mode of D609 was derived through molecular docking. Our model of the hSMS1/D609 complex (Figure 9) indicates that D609 occupies the catalytic pocket, which blocks the entry of substrate molecules, such as PC and SM. The key residues in￾volved in the binding of D609 include R342 and H285. In particular, the negative charged sulfur atom on D609 forms a strong salt-bridge interaction with R342, which mimics the two critical hydrogen bonds between hSMS1 and SM. D609 dixanthogen does not have the negatively charged sulfur atom, leading to the loss of the salt-bridge with R342. In addition, the steric hin￾drance near R342 does not allow the second D609 moiety on D609 dixanthogen to go into the binding pocket deeply. Thus, our model provides a reasonable explanation on the different hSMS1 inhibition activities between D609 and D609 dixanthogen. Conclusions We have obtained a three-dimensional structural model of hSMS1 through homology modeling, molecu￾lar docking, and extensive molecular dynamics simula￾tions. Our model reasonably explains how hSMS1 transforms a phosphocholine moiety on ceramide to produce sphingomyelin. It also explains the high selec￾tivity of hSMS1 towards phosphocholine and sphingo￾myelin. Normal mode analysis on the low-frequency motions of hSMS1 suggests a gating mechanism of immobilizing phosphatidylcholine at the catalytic site and releasing sphingomyelin afterwards. Computational mutagenesis results explore the key residues for the binding of SM to hSMS1. Currently, very few small-molecule inhibitors of hSMS1 have been reported publicly in literature. The potential value of our struc￾tural model is that it can be applied to the discovery of effective hSMS1 inhibitors, for example, through virtual screening. References 1 Merrill, A. H. Jr.; Jones, D. D. Biochim. Biophys. Acta 1990, 1044, 1. 2 Nagao, K.; Takahashi, K.; Hanada, K.; Kioka, N.; Matsuo, M.; Ueda, K. J. Biol. Chem. 2007, 282, 14868. 3 Sano, O.; Kobayashi, A.; Nagao, K.; Kumagai, K.; Kioka, N.; Hanada, K.; Ueda, K.; Matsuo, M. J. Lipid Res. 2007, 48, 2377. 4 Schutze, S.; Potthoff, K.; Machleidt, T.; Berkovic, D.; Wiegmann, K.; Kronke, M. Cell 1992, 71, 765. 5 Luberto, C.; Yoo, D. S.; Suidan, H. S.; Bartoli, G. M.;