Complex of AChE with a Bis-(-)-hor- meptazinol Journal of Medicinal Chemistry, 2009, Vol. 52, No 8 2547 At the PAs, as already noted, the electron density p that the MEP moiety samples more than one conformation.The principal residues surrounding it are Tyr70, Glu73, GIn74 Trp279, and gly335(Figure 6), but the conformational hetero- geneity precludes assignment of specific interactions. The possibility was considered that the apparent conformational ht reflect chemical inhomogeneity, but mass spectrometry of the compound clearly excluded this possibility (not shown). Ligplot was used to compare the principal ligand-protein interactions of 5h in the TcAChE crystal structure( Figure 7A) and in the previously published modeled complex with mAChE In this model, the MEP moieties are both in the equatorial position because this conformation was that found in the smal molecule X-ray structure of MEP3O(Figure 7B). Superposition of the crystal structures of native mAChE (PDB ode 2HA2)and of the 5h/TcAChE complex(Figure 8)reveals ructures of the Sh/TcAChE complex and of native mAChE looking no significant main chain or side chain differences that could down the gorge. No significant differences in main chain or side chai provide a basis for the differences between the latter and the conformations are observed between the TcAChE structure(blue) and the mAChE structure(red). Only the main chains are traced for clarity: model of the 5h/mAChE complex obtained using GOLD sh is shown in green There is overall similarity in the positioning of the ligand (rmsd 3. A), with hydrophobic interactions with Trp84, Phe330, and Tyr334 (or the corresponding mAChE residues)appearing In e crystal structure an computational model. However, as mentioned above there are significant differences as well. Thus, in the X-ray structure, His440 is hydrogen-bonded to 5h through the m atom of its imidazole side chain, no through its main-chain carbonyl oxygen, as was suggested by the GOLD model. Furthermore, the crystal structure does not reveal T-T stacking interactions between Trp84 and the phenyl group of the MEP group bound at the CAS nor hydrogen bonding of the protonated azepane nitrogen to Tyrl2lc (Tyr1240s in mAChE) On the basis of the 5h/TcAChE crystal structure, we decided to carry out additional GOLD simulations, with the same algorithm and parameters that we previously used, in which both MEP moieties in 5h could adopt either axial or equatorial ntations(aa and ee, respectively) or could ado orientations(ea and ae)both in the native enzyme(PDB code IEA5)and in the template of the 5h/TcAChE complex, devoid Figure 9. Comparison of the conformation of the MEP moiety in the of 5h, waters, and hydrocarbons. This was done to reconcile CAS of the 5h/TcAChE crystal structure with its conformation in two the observation that there was a difference between the axial models obtained by by docking to the protein template of the crystal orientation of the mEP moiety at the CAS determined by the (rmsd= 1. 15 A, rerank=3.2) shows good agreement between the complex X-ray structure and the equatorial orientations of the crystal structure and the model.(B)Superposition of the other model ligand used in the gOld model. while the NMR solution (rmsd= 1. 18 A, rerank= 2.4)on the crystal structure yields a totally structure of mEP detected both the axial and equatorial ncorrect orientation. The crystal structure is displayed as thick sticks conformers. Table 2 lists the conformations and scores and the models as lines obtained(models are available as Supporting Information) In this modeling study, we found rather than using either (Figure 4). Its phenolic oxygen forms a hydrogen bond wit His440N22(2.55 A), thus disrupting the latter's hydrogen bond TCAChE (EA5) or mAChE, that in general the best scores were with Ser2000 within the catalytic triad. The xi angle of the TcAChE complex structure(PDB code 2w6C) When the two crystal structure(PDB code IEA5). As a consequence, the distance between Ser2000" and His440Ne2 increases from 3.0 posed on Sh from the crystal structure of the complex(Figure ), it could be seen that the modeled binding mode of the MEP A in the native structure to 4.2 A. An analogous rupture of moiety at the CAS was quite similar to that in the crystal the catalytic triad and rotation of Ser200 was observed in the structure in the case of 2w6c_ea_220( Figure 9A), even though complex of TcAChE with a bis-tacrine inhibitor with a pen their orientations are ec in the model and ac in the experimen- tamethylene spacer. Phe330 also undergoes a conformational tal structure, respectively. However in the second model change, relative to the native structure, that involves-25and 2w6c_ee_447, the CAS MEP moiety assumes a different 33 rotations, respectively, of its %i and x2 angles. The residues orientation(Figure 9B). Both models have high ranks(Table lining the gorge between the two MEP moieties do not reveal 2), exemplifying the difficulty in choosing a sing gle model, rather any substantial changes relative to their positions and conforma- than an ensemble of models, in the absence of a crystal structure tions in native TcAChE(Figure 5) of a complex. No comparison of the MEP moieties at the PAS(Figure 4). Its phenolic oxygen forms a hydrogen bond with His440Nε2 (2.55 Å), thus disrupting the latter’s hydrogen bond with Ser200Oγ within the catalytic triad. The 1 angle of the freed Ser200 is rotated ∼120° relative to its position in the native crystal structure (PDB code 1EA5). As a consequence, the distance between Ser200Oγ and His440Nε2 increases from ∼3.0 Å in the native structure to ∼4.2 Å. An analogous rupture of the catalytic triad and rotation of Ser200 was observed in the complex of TcAChE with a bis-tacrine inhibitor with a pen￾tamethylene spacer.34 Phe330 also undergoes a conformational change, relative to the native structure, that involves ∼25° and ∼33° rotations, respectively, of its 1 and 2 angles. The residues lining the gorge between the two MEP moieties do not reveal any substantial changes relative to their positions and conforma￾tions in native TcAChE (Figure 5). At the PAS, as already noted, the electron density suggests that the MEP moiety samples more than one conformation. The principal residues surrounding it are Tyr70, Glu73, Gln74, Trp279, and Gly335 (Figure 6), but the conformational hetero￾geneity precludes assignment of specific interactions. The possibility was considered that the apparent conformational heterogeneity might reflect chemical inhomogeneity, but mass spectrometry of the compound clearly excluded this possibility (not shown). Ligplot35 was used to compare the principal ligand-protein interactions of 5h in the TcAChE crystal structure (Figure 7A) and in the previously published modeled complex with mAChE. In this model, the MEP moieties are both in the equatorial position because this conformation was that found in the small molecule X-ray structure of MEP30 (Figure 7B). Superposition of the crystal structures of native mAChE (PDB code 2HA2) and of the 5h/TcAChE complex (Figure 8) reveals no significant main chain or side chain differences that could provide a basis for the differences between the latter and the model of the 5h/mAChE complex obtained using GOLD.19 There is overall similarity in the positioning of the ligand (rmsd ) 3.01 Å), with hydrophobic interactions with Trp84, Phe330, and Tyr334 (or the corresponding mAChE residues) appearing in both the crystal structure and the computational model. However, as mentioned above, there are significant differences as well. Thus, in the X-ray structure, His440 is hydrogen-bonded to 5h through the Nε2 atom of its imidazole side chain, not through its main-chain carbonyl oxygen, as was suggested by the GOLD model. Furthermore, the crystal structure does not reveal π-π stacking interactions between Trp84 and the phenyl group of the MEP group bound at the CAS nor hydrogen￾bonding of the protonated azepane nitrogen to Tyr121O (Tyr124O in mAChE). On the basis of the 5h/TcAChE crystal structure, we decided to carry out additional GOLD simulations, with the same algorithm and parameters that we previously used, in which both MEP moieties in 5h could adopt either axial or equatorial orientations (aa and ee, respectively) or could adopt mixed orientations (ea and ae) both in the native enzyme (PDB code 1EA5) and in the template of the 5h/TcAChE complex, devoid of 5h, waters, and hydrocarbons. This was done to reconcile the observation that there was a difference between the axial orientation of the MEP moiety at the CAS determined by the complex X-ray structure and the equatorial orientations of the ligand used in the GOLD model, while the NMR solution structure of MEP detected both the axial and equatorial conformers.36 Table 2 lists the conformations and scores obtained (models are available as Supporting Information). In this modeling study, we found rather than using either TcAChE (1EA5) or mAChE, that in general the best scores were obtained for 5h modeled inside the empty template of the 5h/ TcAChE complex structure (PDB code 2W6C). When the two best models, 2w6c_ea_220 and 2w6c_ee_447, were superim￾posed on 5h from the crystal structure of the complex (Figure 9), it could be seen that the modeled binding mode of the MEP moiety at the CAS was quite similar to that in the crystal structure in the case of 2w6c_ea_220 (Figure 9A), even though their orientations are eC in the model and aC in the experimen￾tal structure, respectively. However in the second model, 2w6c_ee_447, the CAS MEP moiety assumes a different orientation (Figure 9B). Both models have high ranks (Table 2), exemplifying the difficulty in choosing a single model, rather than an ensemble of models, in the absence of a crystal structure of a complex. No comparison of the MEP moieties at the PAS Figure 8. Superposition of the active-site gorge area in the crystal structures of the 5h/TcAChE complex and of native mAChE looking down the gorge. No significant differences in main chain or side chain conformations are observed between the TcAChE structure (blue) and the mAChE structure (red). Only the main chains are traced for clarity; 5h is shown in green. Figure 9. Comparison of the conformation of the MEP moiety in the CAS of the 5h/TcAChE crystal structure with its conformation in two models obtained by docking to the protein template of the crystal structure (PDB code 2W6C). (A) Superposition of one such model (rmsd ) 1.15 Å, rerank ) 3.2) shows good agreement between the crystal structure and the model. (B) Superposition of the other model (rmsd ) 1.18 Å, rerank ) 2.4) on the crystal structure yields a totally incorrect orientation. The crystal structure is displayed as thick sticks and the models as lines. Complex of AChE with a Bis-(-)-nor-meptazinol Journal of Medicinal Chemistry, 2009, Vol. 52, No. 8 2547