Bis-(-)-nor-meptazinols as Novel Nanomolar Cholinesterase Inhibitors with High Inhibitory Potency on Amyloid-beta Aggregation

J.Med.Chem.2008,51,2027-2036 Bis-()-nor-meptazinols as Novel Nanomolar Cholinesterase Inhibitors with High Inhibitory Potency on Amyloid-B Aggregation Qiong Xie, "Hao Wang, t* Zheng Xia, Meiyan Lu, t Weiwei Zhang. Xinghai Wang, t Wei Fu, t Yun Tang, Wei Sheng, Wei Li, Wei Zhou, Xu Zhu, Zhuibai Qiu, *f and Hongzhuan Chen*+ Department of Medicinal Chemistry, School of Pharmacy, Fudan University, 138 Yixueyuan Road, Shanghai 200032, P. R. China, Departmen of Pharmacology, Institute of Medical Sciences, Shanghai JiaoTong University School of Medicine, 280 South Chongqing Road, 30 Meilong Road, Shanghai 200237, P. R. China Received February 9. 2007 Bis-(-)-nor-meptazinols(bis-()-nor-MEPs)5 were designed and synthesized by connecting two MEP monomers with alkylene linkers of different lengths via the secondary amino groups acetylcholinesterase(AChE) inhibitory activities were more greatly influenced by the length of the Alkylene exhibited low-nanomolar ICso values for both ChEs, having a 10 000-fold and 1500-fold increase in inhibitic of AChE and BChE compared with(-)-MEP. Molecular docking elucidated that 5h simultaneously bound to the catalytic and peripheral sites in AChE via hydrophobic interactions with Trp86 and Trp286. In comparison, it folded in the large aliphatic cavity of BChE because of the absence of peripheral site and the enlargement of the active site. Furthermore, 5h and 5i markedly prevented the AChE-induced AB aggregation with ICso values of 16.6 and 5.8 uM, similar to that of propidium (IC50= 12. 8 uM), which suggests promising disease-modifying agents for the treatment of AD patients Introduction Alzheimer's disease(AD), which is characterized by progres- sive loss of memory and impairment in cognition, is becoming a serious threat to life expectancy for elderly people. The main pathological changes in the Ad brain are the abnormal formation Linked of extracellular senile plaques consisting of aggregated amyloid B-peptide(Ap)deposits and intracellular neurofibrillary tangles (NTFs) consisting of abnormally phosphorylated microtubule- associated protein T Current clinical therapy for Ad patients is mainly palliative = atalytic site treatment targeting acetylcholinesterase(AChE). On the basis of the cholinergic hypothesis, inhibition of AChE effectively ncreases the available acetylcholine(ACh) within cholinergic synapses, resulting in modest improvement in cognitive symp- Figure 1. Catalytic and peripheral sites of AChE active site gorge oms. Mounting evidence has indicated that AChE may be involved in several noncholinergic functions. AChE colocalizes cognitive deficit of AD patients by elevating ACh levels but with AB in senile plaques, promoting the assembly of AB into also act as disease-modifying agents delaying amyloid plaque fibrils'and accelerating AB peptide deposition. Structural formation. 9a, 13 models of the interaction between AChE and AB have recently Recently, bivalent ligand strategy has been utilized in the been explored. It has been speculated that AChE achieves design of dual binding site ache inhibitors 9-13 homobivalent aggregation-promoting action through direct binding with A or heterobivalent ligands are obtained by connecting two via the specific region of the enzyme that involves a peripheral identical or distinct moieties through a linker of suitable length binding site. Inhibition of the peripheral site might prevent A, to make contact with both the catalytic and peripheral sites. A peptide deposition induced by AChE. This enzyme has a narrow spatial 12 A distance was determined by X-ray crystal diffraction 20 A deep active site gorge, the bottom and opening regions of from Trp86(mammalian numbering), the catalytic anionic site which are known as catalytic and peripheral sites, respectively center, to Trp286(mammalian numbering), core of the periph (Figure 1). AChE inhibitors simultaneously blocking both the eral site(Figure 1). In many cases of homobivalent ligands catalytic and peripheral sites might not only alleviate the (bis-ligands ), AChE inhibitory potency and selectivity improved relative to the monomer and additional inhibition of ache- 21地0m1 hE inhibitor reported g aoli@shamu.edu.cn strategy, presenting a more than 1000-fold increase in AChE Fudan inhibiting potency and a 10000-fold increase in AChE/butyryl These authors contributed equally to this work cholinesterase (BChE) selectivity compared with tacrine Shanghai Jiao Tong University. Bis-galanthamine 2(Figure 2, n=8), bis-5-amino-5,6,7, 8- East China University of Science and Technology 10 or 12) 10. 1021/jm070154q CCC: $.75 C 2008 American Chemical Society Published on web 03/12/2008

Bis-(-)-nor-meptazinols as Novel Nanomolar Cholinesterase Inhibitors with High Inhibitory Potency on Amyloid-
Aggregation Qiong Xie,†,# Hao Wang,‡,# Zheng Xia,‡ Meiyan Lu,† Weiwei Zhang,‡ Xinghai Wang,† Wei Fu,† Yun Tang,§ Wei Sheng,† Wei Li,† Wei Zhou,‡ Xu Zhu,‡ Zhuibai Qiu,*,† and Hongzhuan Chen*,‡ Department of Medicinal Chemistry, School of Pharmacy, Fudan UniVersity, 138 Yixueyuan Road, Shanghai 200032, P. R. China, Department of Pharmacology, Institute of Medical Sciences, Shanghai JiaoTong UniVersity School of Medicine, 280 South Chongqing Road, Shanghai 200025, P. R. China, and School of Pharmacy, East China UniVersity of Science and Technology, 130 Meilong Road, Shanghai 200237, P. R. China ReceiVed February 9, 2007 Bis-(-)-nor-meptazinols (bis-(-)-nor-MEPs) 5 were designed and synthesized by connecting two (-)-nor￾MEP monomers with alkylene linkers of different lengths via the secondary amino groups. Their acetylcholinesterase (AChE) inhibitory activities were more greatly influenced by the length of the alkylene chain than butyrylcholinesterase (BChE) inhibition. The most potent nonamethylene-tethered dimer 5h exhibited low-nanomolar IC50 values for both ChEs, having a 10 000-fold and 1500-fold increase in inhibition of AChE and BChE compared with (-)-MEP. Molecular docking elucidated that 5h simultaneously bound to the catalytic and peripheral sites in AChE via hydrophobic interactions with Trp86 and Trp286. In comparison, it folded in the large aliphatic cavity of BChE because of the absence of peripheral site and the enlargement of the active site. Furthermore, 5h and 5i markedly prevented the AChE-induced A
aggregation with IC50 values of 16.6 and 5.8 µM, similar to that of propidium (IC50 ) 12.8 µM), which suggests promising disease-modifying agents for the treatment of AD patients. Introduction Alzheimer’s disease (AD), which is characterized by progres￾sive loss of memory and impairment in cognition,1 is becoming a serious threat to life expectancy for elderly people. The main pathological changes in the AD brain are the abnormal formation of extracellular senile plaques consisting of aggregated amyloid-
-peptide (A
) deposits and intracellular neurofibrillary tangles (NTFs) consisting of abnormally phosphorylated microtubule￾associated protein τ. 2 Current clinical therapy for AD patients is mainly palliative treatment targeting acetylcholinesterase (AChE). On the basis of the cholinergic hypothesis,3 inhibition of AChE effectively increases the available acetylcholine (ACh) within cholinergic synapses, resulting in modest improvement in cognitive symp￾toms. Mounting evidence has indicated that AChE may be involved in several noncholinergic functions.4 AChE colocalizes with A
in senile plaques, promoting the assembly of A
into fibrils5 and accelerating A
peptide deposition.6 Structural models of the interaction between AChE and A
have recently been explored.7 It has been speculated that AChE achieves its aggregation-promoting action through direct binding with A
via the specific region of the enzyme that involves a peripheral binding site.8 Inhibition of the peripheral site might prevent A
peptide deposition induced by AChE. This enzyme has a narrow 20 Å deep active site gorge, the bottom and opening regions of which are known as catalytic and peripheral sites, respectively (Figure 1). AChE inhibitors simultaneously blocking both the catalytic and peripheral sites might not only alleviate the cognitive deficit of AD patients by elevating ACh levels but also act as disease-modifying agents delaying amyloid plaque formation.9a,13 Recently, bivalent ligand strategy has been utilized in the design of dual binding site AChE inhibitors.9–13 Homobivalent or heterobivalent ligands are obtained by connecting two identical or distinct moieties through a linker of suitable length to make contact with both the catalytic and peripheral sites. A spatial 12 Å distance was determined by X-ray crystal diffraction from Trp86 (mammalian numbering), the catalytic anionic site center, to Trp286 (mammalian numbering), core of the periph￾eral site (Figure 1).14 In many cases of homobivalent ligands (bis-ligands), AChE inhibitory potency and selectivity improved relative to the monomer and additional inhibition of AChE￾induced A
aggregation was observed. Bis-tacrine 1 (Figure 2, n ) 7)9 was the first bivalent AChE inhibitor reported on this strategy, presenting a more than 1000-fold increase in AChE inhibiting potency and a 10000-fold increase in AChE/butyryl￾cholinesterase (BChE) selectivity compared with tacrine. Bis-galanthamine 2 (Figure 2, n ) 8),10 bis-5-amino-5,6,7,8- tetrahydroquinolinone 3 (Figure 2, n ) 10 or 12),11 and bis- * To whom correspondence should be addressed. For Z.Q.: phone, 86- 21-54237595; fax, 86-21-54237264; e-mail, zbqiu@shmu.edu.cn. For H.C.: phone, 86-21-63846590, extension 776450; fax, 86-21-64674721; e-mail, yaoli@shsmu.edu.cn. † Fudan University. # These authors contributed equally to this work. ‡ Shanghai JiaoTong University. § East China University of Science and Technology. Figure 1. Catalytic and peripheral sites of AChE active site gorge. J. Med. Chem. 2008, 51, 2027–2036 2027 10.1021/jm070154q CCC: $40.75  2008 American Chemical Society Published on Web 03/12/2008

2028 Journal of Medicinal Chemistry, 2008, VoL. 51, No. 7 Xie et al Me○ CH2) OMe 1 bis-tacrine 2 bis-galanthamine 3bs5-amin56,7,8 (CH2im N(CH2)N(CH2)m) M Figure 2. Structures of reported homobivalent AChE inhibitors and title compounds 5 Scheme 1. Synthesis of (-)-nor-MEP 8 7 8 huperzine B 4(Figure 2, m=2, n=10)(Figure 2)have also Scheme 2. Synthesis of 5a, b,e-k been reported. Our group has been interested in the study of meptazinol (MEP),a racemic marketed opioid analgesic with low addic tion liability, and its()-enantiomer, which has demonstrated moderate inhibition of AChE. We established an approach to the resolution of MEP in acceptable yields and determined the absolute configurations of(-)-MEP and(+)-MEP as S and R, respectively, by X-ray crystal structures. 17 Continuing with our 5(a-b, e- previous research to find new AChE inhibitors through the Reagents and conditions: (i)a, a-dihaloalkane(0.5 equiv), triethylamine molecular modeling of(-)-MEP derivatives, we describe here (2 equiv), acetonitrile, reflux, 2-5h, 35-83% the design, synthesis, pharmacological evaluation, and molecular docking of a series of homobivalent(-)-N-demethylmeptazinols (bis-(-)-nor-MEPs)5. Two identical (-)-nor-MEP units are Newly synthesized compounds were tested in vitro for AChE ind BChE inhibitory potency, and their selectivity for AChE connected by alkylene linkers of different lengths via the was calculated. A molecular docking study was performed on ondary amino groups in compound 5(Figure 2) A suitable length of the alkylene linker, together with ar mouse AChE (mAChE) and human bChe (hBChE) to il luminate the binding modes of the most potent compound 5h appropriate point of the coupling position, guarantees that bis- with both enzymes. Because of the unavailability of crystal ligand will simultaneously bind to the catalytic and peripheral lographic data of mouse BChE, hBChE was used instead sites of the enzyme. According to our predicted binding mode because of a high sequence identity. The ability of compounds of(-)-MEP in AChE active site, the azepane ring is locate 5g, 5h, and 5i to inhibit the AChE-induced AB aggregation in the middle of the gorge whereas the phenolic group is oriented compared with propidium iodine and the reference compound ()-MEP derivatives were designed as bis-(-)-nor-MEPs ()-MEP, was assessed by means of a thioflavin T-based linking, via a point in the azepane ring instead of the phenolic fluorometric assay. Cell viability was tested by MTT assay group(Figure 1). As to the chain length(number of methylene nun er in the case of tacrine- based bivalent ligands was 7.However, that might not be the case in our Results and Discussion study. To find the most potent compound in our series and Chemistry. The synthetic methodology employed for the discuss the effect of linker length on inhibitory potency, preparation of bis-(--nor-MEP derivatives 5 is illustrated in compounds possessing methylene spacers varying from 2 to 12 Schemes 1-3. Key step in this route is the N-demethylation of were synthesized (-)-MEP 6 producing(-)-nor-MEP 8. A few methods have

huperzine B 4 (Figure 2, m ) 2, n ) 10)12 (Figure 2) have also been reported. Our group has been interested in the study of meptazinol (MEP),15 a racemic marketed opioid analgesic with low addic￾tion liability, and its (-)-enantiomer, which has demonstrated moderate inhibition of AChE.16 We established an approach to the resolution of MEP in acceptable yields and determined the absolute configurations of (-)-MEP and (+)-MEP as S and R, respectively, by X-ray crystal structures.17 Continuing with our previous research to find new AChE inhibitors through the molecular modeling of (-)-MEP derivatives,18 we describe here the design, synthesis, pharmacological evaluation, and molecular docking of a series of homobivalent (-)-N-demethylmeptazinols (bis-(-)-nor-MEPs) 5. Two identical (-)-nor-MEP units are connected by alkylene linkers of different lengths via the secondary amino groups in compound 5 (Figure 2). A suitable length of the alkylene linker, together with an appropriate point of the coupling position, guarantees that bis￾ligand will simultaneously bind to the catalytic and peripheral sites of the enzyme. According to our predicted binding mode of (-)-MEP in AChE active site,18 the azepane ring is located in the middle of the gorge whereas the phenolic group is oriented down into the catalytic site at the bottom. Therefore, bivalent (-)-MEP derivatives were designed as bis-(-)-nor-MEPs linking, via a point in the azepane ring instead of the phenolic group (Figure 1). As to the chain length (number of methylene units), the optimal number in the case of tacrine-based bivalent ligands was 7.9,19 However, that might not be the case in our study. To find the most potent compound in our series and discuss the effect of linker length on inhibitory potency, compounds possessing methylene spacers varying from 2 to 12 were synthesized. Newly synthesized compounds were tested in vitro for AChE and BChE inhibitory potency, and their selectivity for AChE was calculated. A molecular docking study was performed on mouse AChE (mAChE) and human BChE (hBChE) to il￾luminate the binding modes of the most potent compound 5h with both enzymes. Because of the unavailability of crystal￾lographic data of mouse BChE, hBChE was used instead because of a high sequence identity. The ability of compounds 5g, 5h, and 5i to inhibit the AChE-induced A
aggregation, compared with propidium iodine and the reference compound (-)-MEP, was assessed by means of a thioflavin T-based fluorometric assay.20 Cell viability was tested by MTT assay for 5h and 5i. Results and Discussion Chemistry. The synthetic methodology employed for the preparation of bis-(-)-nor-MEP derivatives 5 is illustrated in Schemes 1-3. Key step in this route is the N-demethylation of (-)-MEP 6 producing (-)-nor-MEP 8. A few methods have Figure 2. Structures of reported homobivalent AChE inhibitors and title compounds 5. Scheme 1. Synthesis of (-)-nor-MEP 8a a Reagents and conditions: (i) (a) ClCOOEt, KHCO3, CHCl3, reflux, 1 h; (b) K2CO3 (aq), MeOH, N2, room temp, 1 h, 95%; (ii) 50% H2SO4, N2, reflux, 4 h, 54%. Scheme 2. Synthesis of 5a,b,e-ka a Reagents and conditions: (i) R,ω-dihaloalkane (0.5 equiv), triethylamine (2 equiv), acetonitrile, reflux, 2-5 h, 35-83%. 2028 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 7 Xie et al

Journal of Medicinal Chemistry, 2008, Vol. 51, No. 7 2029 Scheme 3. Synthesis of 5c, d CH2)n-2 9(,d) been reported for the demethylation of tertiary amines, such as Table 1. Inhibition of AChE and BChE by Bis-(-)-nor-MEPs 5a-k ()MEP, and Rivastigmine employment of azodicarboxylic acid esters, cyanogen bro- ICso(nM) as the reagent, taking both the availability and reaction simplicity ice brain Mice serum selectivity for to account. treating MEP 6 with ethyl chloroformate in BChE the presence of KHCO3 in boiling CHCl3, followed by dealing 5a 43000±20000125±9 the resulting residue with a mild base, afforded a nonbasic 342000±14000132±510.0 carbamate intermediate, (-)-N-carboethoxy nor-MEP 7How 421400±7600104±290.0049 4000±1000192±41 0.048 ever, troubles were encountered in the hydrolysis and decar- se 1220±20 0.098 xylation of the resulting carbamate intermediate 7. Al 5f 7 270±70 102±19 reactions failed under the reported alkaline condition in KOH24 ±19 63±8 nd acidic conditions in hydrobromic acid or 25% sulfuric acid 3.9±1.3 10士3 Finally this transformation was accomplished in 50% sulfuric .5±4.5 74±11 acid under nitrogen for 4 h, and (-)-nor-MEP 8 was obtained 42±20 in a 54% yield(Scheme 1) rivastigmine 5500±15001600±300.29 Alkylation of (-)-nor-MEP 8 with a, o-dihaloalkanes(0.5 (一)MEP 41000±1400015000±40000.37 equiv) in the presence of triethylamine, followed by chromato- graphic purification, easily produced the bis-(-)-nor-MEP of AChE. Mice serum was the source of BChE. AChE was assayed compounds 5a, b e-k in 35-83% yield(Scheme 2). However. spectrophotometrically with acetylthiocholine as substrate in the presence Ikylation with 1, 4-dichlorobutane or 1, 5-dichloropentane failed with butyrylthiocholine as substrate and 1025 M BW284C51 as AChE to furnish the bivalent compounds 5c, d, since the N-4-chlo- inhibitor ICso values were computed by a nonlinear least squares regression robutyl (or 5-chloropentyl)-(-)-nor-MEP intermediate was prone program that also provided an estimate of statistical precision(standard to forming stable intramolecular five-membered or six error of the mean). Selectivity for AChE: ICso for BChE divided by ICso membered ring structures, resulting in failure to link another for AChE. )-nor-MEP unit and leading to the generation of spiro ■AChE◆BChE quaternary ammoniums Structures of two quaternary ammo- niums derived from the R enantiomer were confirmed by X-ray crystallographic diffraction. Eventually, the synthesis of the bis-ligands 5c, d was accomplished by acylation with a,a- Ikanediacyl dihalide(0.5 equiv) to form bis-amides intermedi- ates 9c, d followed by reduction using lithium aluminum hydride (LAH) in tetrahydrofuran(THF)(Scheme 3) The chemical structures of all target compounds or their ynthesized hydrochloride salts were characterized by specific rotation [a]D, IR, H NMR, and HR-ESI, as reported in the 5269只 Experimental Section. The complicated property of NMR data from bis-(-)-nor-MEP hydrochloride salts in DMSO-d6 re- sembled the case of (+)-MEP hydrochloride. It was reasonably explained by conformational switch of the azepane ring and onfigurational inversion of nitrogen AChE Inhibitory Potency and AChE/BChE Selectivity Newly synthesized compounds were tested in vitro for potency 即mh时nr and selectivity as cholinesterase( ChE)inhibitors. Extracts from mice brain and mice serum were used as sources of AChE and the length of the alkylene chain. The optimal chain length BChE, respectively. The results showed that both(-)-MEP and determined experimentally was achieved in compound 5h, with its bis-ligand analogues possessed ChE inhibitory activity. The nine methylene groups between two (-)-nor-MEP units 50 value of(-)-MEP was 41 uM, about 10 times higher than Compared with(-)-MEP (ICso=41 uM)and rivastigmine(ICso that obtained with AChE from bovine erythrocytes, and the=5.5 uM), 5h(ICs0=3.9 nM) showed a 10000-fold and 1400 testing data of the reference drug rivastigmine conformed to fold increase, respectively, in the inhibition of mice brain AChE the previous report 27 The AChE inhibitory potency within the (Table 1). Further decrease or increase of the chain length series of bis-(-)-nor-MEP derivatives was closely related to weakened the AChE inhibition(Figure 3). For instan

been reported for the demethylation of tertiary amines, such as employment of azodicarboxylic acid esters,21 cyanogen bro￾mide,22 or chloroformates.23,24 Ethyl chloroformate was chosen as the reagent, taking both the availability and reaction simplicity into account. Treating (-)-MEP 6 with ethyl chloroformate in the presence of KHCO3 in boiling CHC13, followed by dealing the resulting residue with a mild base, afforded a nonbasic carbamate intermediate, (-)-N-carboethoxy nor-MEP 7. How￾ever, troubles were encountered in the hydrolysis and decar￾boxylation of the resulting carbamate intermediate 7. All reactions failed under the reported alkaline condition in KOH24 and acidic conditions in hydrobromic acid or 25% sulfuric acid. Finally this transformation was accomplished in 50% sulfuric acid under nitrogen for 4 h, and (-)-nor-MEP 8 was obtained in a 54% yield (Scheme 1). Alkylation of (-)-nor-MEP 8 with R,ω-dihaloalkanes (0.5 equiv) in the presence of triethylamine, followed by chromato￾graphic purification, easily produced the bis-(-)-nor-MEP compounds 5a,b,e-k in 35-83% yield (Scheme 2). However, alkylation with 1,4-dichlorobutane or 1,5-dichloropentane failed to furnish the bivalent compounds 5c,d, since the N-4-chlo￾robutyl (or 5-chloropentyl)-(-)-nor-MEP intermediate was prone to forming stable intramolecular five-membered or six￾membered ring structures, resulting in failure to link another (-)-nor-MEP unit and leading to the generation of spiro quaternary ammoniums. Structures of two quaternary ammo￾niums derived from the R enantiomer were confirmed by X-ray crystallographic diffraction.25 Eventually, the synthesis of the bis-ligands 5c,d was accomplished by acylation with R,ω- alkanediacyl dihalide (0.5 equiv) to form bis-amides intermedi￾ates 9c,d followed by reduction using lithium aluminum hydride (LAH) in tetrahydrofuran (THF) (Scheme 3). The chemical structures of all target compounds or their synthesized hydrochloride salts were characterized by specific rotation [R]D, IR, 1 H NMR, and HR-ESI, as reported in the Experimental Section. The complicated property of NMR data from bis-(-)-nor-MEP hydrochloride salts in DMSO-d6 re￾sembled the case of (+)-MEP hydrochloride.26 It was reasonably explained by conformational switch of the azepane ring and configurational inversion of nitrogen. AChE Inhibitory Potency and AChE/BChE Selectivity. Newly synthesized compounds were tested in vitro for potency and selectivity as cholinesterase (ChE) inhibitors. Extracts from mice brain and mice serum were used as sources of AChE and BChE, respectively. The results showed that both (-)-MEP and its bis-ligand analogues possessed ChE inhibitory activity. The IC50 value of (-)-MEP was 41 µM, about 10 times higher than that obtained with AChE from bovine erythrocytes,16 and the testing data of the reference drug rivastigmine conformed to the previous report.27 The AChE inhibitory potency within the series of bis-(-)-nor-MEP derivatives was closely related to the length of the alkylene chain. The optimal chain length determined experimentally was achieved in compound 5h, with nine methylene groups between two (-)-nor-MEP units. Compared with (-)-MEP (IC50 ) 41 µM) and rivastigmine (IC50 ) 5.5 µM), 5h (IC50 ) 3.9 nM) showed a 10000-fold and 1400- fold increase, respectively, in the inhibition of mice brain AChE (Table 1). Further decrease or increase of the chain length weakened the AChE inhibition (Figure 3). For instance, the Scheme 3. Synthesis of 5c,da a Reagents and conditions: (i) R,ω-alkanediacyl dihalide (0.5 equiv), triethylamine (2 equiv), dry CH2Cl2, 0 °C, 15 min, 38-41%; (ii) lithium aluminum hydride (LAH), dry THF, reflux, 1 h, 31-36%. Table 1. Inhibition of AChE and BChE by Bis-(-)-nor-MEPs 5a-k, (-)-MEP, and Rivastigminea IC50 (nM) compd chain length (n) mice brain AChE Mice serum BChE selectivity for AChEb 5a 2 43000 ( 20000 125 ( 9 0.0029 5b 3 42000 ( 14000 132 ( 51 0.0031 5c 4 21400 ( 7600 104 ( 29 0.0049 5d 5 4000 ( 1000 192 ( 41 0.048 5e 6 1220 ( 20 119 ( 20 0.098 5f 7 270 ( 70 102 ( 19 0.38 5g 8 79 ( 19 63 ( 8 0.80 5 h 9 3.9 ( 1.3 10 ( 3 2.6 5i 10 9.5 ( 4.5 17 ( 6 1.8 5j 11 24 ( 8 74 ( 11 3.1 5k 12 42 ( 20 100 ( 55 2.4 rivastigmine 5500 ( 1500 1600 ( 30 0.29 (-)-MEP 41000 ( 14000 15000 ( 4000 0.37 a Mice brain homogenate prepared in normal saline was used as a source of AChE. Mice serum was the source of BChE. AChE was assayed spectrophotometrically with acetylthiocholine as substrate in the presence of 1024 M ethopropazine as BChE inhibitor. BChE was assayed similarly with butyrylthiocholine as substrate and 1025 M BW284C51 as AChE inhibitor. IC50 values were computed by a nonlinear least squares regression program that also provided an estimate of statistical precision (standard error of the mean). b Selectivity for AChE: IC50 for BChE divided by IC50 for AChE. Figure 3. Correlation between ChE inhibitory potency (-log(IC50)) and the alkylene chain length (n) in compounds 5. Bis-(-)-nor-meptazinols as Inhibitors Journal of Medicinal Chemistry, 2008, Vol. 51, No. 7 2029

2030 Journal of Medicinal Chemistry, 2008, VoL. 51, No. 7 Xie et al inhibitory activities of 5a and 5b(43 and 42 uM, respectively) buried within the core of the enzyme binds with the catalytic are similar to that of(-)-MEP. Therefore, the effective alkylene site via face-to-face T-stacking interaction with Trp86(distance bridged bis-(-)-nor-MEP analogues required chains of suitable between the two centroids: 4.27 A). The other nor-MEP moiety length, which was 9 for the best, to bind at both the catalytic reaches the peripheral site on the surface of the enzyme by site and the peripheral site of the AChE binding pocket(gorge). cation-T and hydrophobic interactions between the seven- Compared with the AChE activity, the BChE inhibitory membered azepane ring and Trp286(distance between the two potency was less impacted by chain length(Figure 3). The centroids: 4.09 A). The spatial distance between the centroids ajority of bis-(-)-MEP analogues showed inhibition on BChE of the centric phenyl group and the peripheral azepane ring is of about 100 nM, although the highest potency (Cso=10 nM) 13.5 A, consistent with the reported distance between two was achieved in 5h. This isomer was 1500 times and 150 times tryptophane. 45h forms two hydrogen bonds, both in the more potent than(-)-MEP (ICso= 15 uM)and rivastigmine catalytic active site. The hydroxyl is hydrogen-bonded to the (Cso= 1.6 uM), respectively, but only 10 times lower than main-chain carbonyl oxygen of His447(O-O distance: 2.87 that of the majority compounds (Table 1). The reason that the A). Meanwhile, the protonated azepane amino group is hydrogen AChE inhibition seems to be the enzymic conformational A). In addition, other aliphatic and aromatic residues are difference. There is lack of a functional peripheral site involved in hydrophobic interactions BChE,2.29 and the BChE active site is wider throughout In the binding mode of 5h with the active sit Therefore, there is no restriction of linker length for bivalent of hBChE is shown in Figure 4c, d. 5h is folded in the large BChE inhibitors cavity along the aliphatic residue-dominated wall, and the Most bis-ligand analogues showed greater selectivity for separation of two terminal(-)-nor-MEP units is relatively short. BChE because of their low affinity for AChE. Only four Three hydrogen bonds are found: (i)between the phenolic compounds, 5h, 51, 5j, and 5k, demonstrated slightly more hydroxyl within the core of the enzyme and the carboxylic acid selectivity for AChE. Recent evidence suggests that both AChE xygen of Asp70(O-O distance, 2.49 A);(ii) between the R四hh图 nd progression of Carboxylic acid ox gen of GI.a21 the activity of AChE decreases progressively in certain regions and (ii) between the oxygen on the phenolic hydroxyl at the to reach 10-15% of normal values, whereas the activity of BChE entrance and the side chain amide nh of Ginl19(O.N stays unchanged or is even increased by 20%.30 Thus, it may distance,3.22 A). Ala 277 and Ala328 are not so importantly not be an advantage for a ChE inhibitor to be considerably more involved in the hydrophobic interactions with 5h, unlike the elective for AChE; on the contrary, a good balance between way their counterparts Try286 and Tyr337 in the mAChE AChE and BChE may result in higher efficacy. As a dual catalytic site behave. On the contrary, some residues unique to inhibitor of both AChE and BChE, rivastigmine appears to be hBChE, such as GInI 19, Leu286, and Val288, form hydrophobic beneficial for people with mild to moderate AD, and BChE contacts with 5h. nhibition correlates significantly with cognitive improvement The difference in the binding mode as well as in pharmace in these patients. In our study, although 5h is slightly more logical activities between 5h with AChE and BChE is funda selective for AChE than BChE (2.6), it had the greatest mentally caused by conformational differences between the two inhibition at a nanomolar lever on both enzymes. Therefore, it enzymes. One of the most important differences is the lack of as suggested that it was a promising drug candidate worthy peripheral site in BChE. Residues responsible for J-T or of further investigations. cation-T interactions at AChE peripheral site are replaced by Molecular Docking Studies. Molecular docking study was aliphatic residues in BChE. For example, Trp286, Tyr72, and performed to ascertain the possibility for the most potent Tyrl24 in mAChE are the counterparts of Ala277, Asn68, and ompound Sh to bind at both the catalytic and peripheral sites GInl1g in hBChE, respectively. In addition, some bulky of AChE and to explore the difference in the interactions of Sh aromatic residues in AChE active site have been replaced by with AChE and BChE. Mammalian enzymes were used in small aliphatic ones in BChE. As a result, the active site of docking, compatible with the pharmacological test. Because of BChE is wider and able to accommodate bis-ligands with linkers the unavailability of crystallographic data of mouse BChE. of wider-ranging lengths. These might be the main reasons that hBChE was used instead because there is a high sequence the bche inhibition is less sensitive to the linker length identity (82%)especially betwe ouse and human bche and Inhibition of A ChE-lnduced AB Aggregation. Three com- the residues in active sites are highly conservative. Here, we pounds, 5g, 5h, and 5i, were selected to assess their abilities to chose recently resolved X-ray crystal structures of the mAChE inhibit AB aggregation induced by AChE using a thioflavin complex with succinylcholine'with a high resolution of 2.05 T-based fluorometric assay, compared with the reference A and of native hBChE28 with a 2.0 A resolution. GOLD compound propidium iodine(Sigma-Aldrich), a known specific docking protocol was employed because it has been proved by peripheral site-binding inhibitor, and the monomer(-)-MEP our previous study to be accurate and reliable for reproducing (Table 2). Results showed that Sh and 5i markedly prevented the binding modes of seven AChE inhibitors in their X-ray the AChE-induced AB aggregation with ICso values of 16.6 and rystal structures of Torpedo californica AChE (TcAChE) 5.8 uM, similar to that of propidium(IC50= 12. 8 uM). With a complexes. An advanced consensus scoring technology was used small ICso value and a higher efficiency of inhibition, 5i and to guide the selection of the most reliable conformation from a 5h were the wonderful compounds that inhibited the aggregation set of candidate conformations that GOLD generated of AB induced by AChE. In contrast, (-)-MEP and 5g showed Our results show that Sh is able to simultaneously make fairly low inhibitory activity (Table 2), which indicated their ontact with both the catalytic and peripheral sites of mAChE, limited ability to interact with the peripheral site of the enzyme s illustrated in Figure 4a. The key interactions of these dimeric Different behaviors of these four compounds were attributed inhibitors with the catalytic and peripheral sites are T-stacking to different lengths of linker, demonstrating that a linker no and cation- interactions. The phenyl group of the nor-MEP horter than 9 was necessary to inhibit AChE-induced A

inhibitory activities of 5a and 5b (43 and 42 µM, respectively) are similar to that of (-)-MEP. Therefore, the effective alkylene￾bridged bis-(-)-nor-MEP analogues required chains of suitable length, which was 9 for the best, to bind at both the catalytic site and the peripheral site of the AChE binding pocket (gorge). Compared with the AChE activity, the BChE inhibitory potency was less impacted by chain length (Figure 3). The majority of bis-(-)-MEP analogues showed inhibition on BChE of about 100 nM, although the highest potency (IC50 ) 10 nM) was achieved in 5h. This isomer was 1500 times and 150 times more potent than (-)-MEP (IC50 ) 15 µM) and rivastigmine (IC50 ) 1.6 µM), respectively, but only 10 times lower than that of the majority compounds (Table 1). The reason that the BChE inhibition is less sensitive to the linker length than the AChE inhibition seems to be the enzymic conformational difference. There is lack of a functional peripheral site in BChE,28,29 and the BChE active site is wider throughout. Therefore, there is no restriction of linker length for bivalent BChE inhibitors. Most bis-ligand analogues showed greater selectivity for BChE because of their low affinity for AChE. Only four compounds, 5h, 5i, 5j, and 5k, demonstrated slightly more selectivity for AChE. Recent evidence suggests that both AChE and BChE may play roles in the etiology and progression of AD beyond regulation of synaptic ACh levels. In the AD brain, the activity of AChE decreases progressively in certain regions to reach 10–15% of normal values, whereas the activity of BChE stays unchanged or is even increased by 20%.30 Thus, it may not be an advantage for a ChE inhibitor to be considerably more selective for AChE; on the contrary, a good balance between AChE and BChE may result in higher efficacy. As a dual inhibitor of both AChE and BChE, rivastigmine appears to be beneficial for people with mild to moderate AD, and BChE inhibition correlates significantly with cognitive improvement in these patients.31 In our study, although 5h is slightly more selective for AChE than BChE (2.6), it had the greatest inhibition at a nanomolar lever on both enzymes. Therefore, it was suggested that it was a promising drug candidate worthy of further investigations. Molecular Docking Studies. Molecular docking study was performed to ascertain the possibility for the most potent compound 5h to bind at both the catalytic and peripheral sites of AChE and to explore the difference in the interactions of 5h with AChE and BChE. Mammalian enzymes were used in docking, compatible with the pharmacological test. Because of the unavailability of crystallographic data of mouse BChE, hBChE was used instead because there is a high sequence identity (82%) especially between mouse and human BChE and the residues in active sites are highly conservative. Here, we chose recently resolved X-ray crystal structures of the mAChE complex with succinylcholine32 with a high resolution of 2.05 Å and of native hBChE28 with a 2.0 Å resolution. GOLD33 docking protocol was employed because it has been proved by our previous study18 to be accurate and reliable for reproducing the binding modes of seven AChE inhibitors in their X-ray crystal structures of Torpedo californica AChE (TcAChE) complexes. An advanced consensus scoring technology was used to guide the selection of the most reliable conformation from a set of candidate conformations that GOLD generated. Our results show that 5h is able to simultaneously make contact with both the catalytic and peripheral sites of mAChE, as illustrated in Figure 4a. The key interactions of these dimeric inhibitors with the catalytic and peripheral sites are π-stacking and cation-π interactions. The phenyl group of the nor-MEP buried within the core of the enzyme binds with the catalytic site via face-to-face π-stacking interaction with Trp86 (distance between the two centroids: 4.27 Å). The other nor-MEP moiety reaches the peripheral site on the surface of the enzyme by cation-π and hydrophobic interactions between the seven￾membered azepane ring and Trp286 (distance between the two centroids: 4.09 Å). The spatial distance between the centroids of the centric phenyl group and the peripheral azepane ring is 13.5 Å, consistent with the reported distance between two tryptophanes.14 5h forms two hydrogen bonds, both in the catalytic active site. The hydroxyl is hydrogen-bonded to the main-chain carbonyl oxygen of His447 (O ··· O distance: 2.87 Å). Meanwhile, the protonated azepane amino group is hydrogen￾bonded to the hydroxyl oxygen of Tyr124 (N···O distance: 3.22 Å). In addition, other aliphatic and aromatic residues are involved in hydrophobic interactions, as shown in Figure 4b. In comparison, the binding mode of 5h with the active site of hBChE is shown in Figure 4c,d. 5h is folded in the large cavity along the aliphatic residue-dominated wall, and the separation of two terminal (-)-nor-MEP units is relatively short. Three hydrogen bonds are found: (i) between the phenolic hydroxyl within the core of the enzyme and the carboxylic acid oxygen of Asp70 (O ··· O distance, 2.49 Å); (ii) between the phenolic hydroxyl at the entrance of the enzyme and the carboxylic acid oxygen of Glu276 (O ··· O distance, 2.61 Å); and (iii) between the oxygen on the phenolic hydroxyl at the entrance and the side chain amide NH of Gln119 (O ··· N distance, 3.22 Å). Ala 277 and Ala328 are not so importantly involved in the hydrophobic interactions with 5h, unlike the way their counterparts Try286 and Tyr337 in the mAChE catalytic site behave. On the contrary, some residues unique to hBChE, such as Gln119, Leu286, and Val288, form hydrophobic contacts with 5h. The difference in the binding mode as well as in pharmaco￾logical activities between 5h with AChE and BChE is funda￾mentally caused by conformational differences between the two enzymes. One of the most important differences is the lack of peripheral site in BChE. Residues responsible for π-π or cation-π interactions at AChE peripheral site are replaced by aliphatic residues in BChE. For example, Trp286, Tyr72, and Tyr124 in mAChE are the counterparts of Ala277, Asn68, and Gln119 in hBChE, respectively. In addition, some bulky aromatic residues in AChE active site have been replaced by small aliphatic ones in BChE. As a result, the active site of BChE is wider and able to accommodate bis-ligands with linkers of wider-ranging lengths. These might be the main reasons that the BChE inhibition is less sensitive to the linker length. Inhibition of AChE-Induced A
Aggregation. Three com￾pounds, 5g, 5h, and 5i, were selected to assess their abilities to inhibit A
aggregation induced by AChE using a thioflavin T-based fluorometric assay,20 compared with the reference compound propidium iodine (Sigma-Aldrich), a known specific peripheral site-binding inhibitor, and the monomer (-)-MEP (Table 2). Results showed that 5h and 5i markedly prevented the AChE-induced A
aggregation with IC50 values of 16.6 and 5.8 µM, similar to that of propidium (IC50 ) 12.8 µM). With a small IC50 value and a higher efficiency of inhibition, 5i and 5h were the wonderful compounds that inhibited the aggregation of A
induced by AChE. In contrast, (-)-MEP and 5g showed fairly low inhibitory activity (Table 2), which indicated their limited ability to interact with the peripheral site of the enzyme. Different behaviors of these four compounds were attributed to different lengths of linker, demonstrating that a linker no shorter than 9 was necessary to inhibit AChE-induced A
2030 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 7 Xie et al

Journal of Medicinal Chemistry, 2008, Vol. 51, No. 7 2031 Gly 1a277 TYr 33 His 4 -Asp 74 33l288 115 如 48 b d Ligand bond His 53 Non-hgand residues involved in hydrophobic Non-lisand bond Om contact(s) ●-● Hvdrogen bond and its length Figure 4. Representation of 5h colored green) docked into the binding sites of mAChe (a) and hbChe (c). The binding site surfaces are colored according to the vacuum statics protein contact potential, calculated by PyMoL 0.99rc2(DeLano Scientific LLC, San Carlos, CA) Crucial catalytic and peripheral he counterpart)residues are colored yellow. Hydrogen bonds and hydrophobic contacts between 5h and the rotein residues of mAChE (b) I site (or the ChE(d)are shown by Lig Plot 4.4.2. 41 aggregation. These findings agree with the results from enzy-(--)-nor-MEPs 5. Their AChE inhibitory activities were closely natic test and molecular docking, which indicated that a linker related to the length of the alkylene chain, whereas BChE of 9 or 10 methylenes would help to reach the peripheral site inhibition was less influenced. The optimal chain length for of achE AChE and BChE inhibition was achieved with 9 in 5h. which Cell Viability. The toxicity of the most po ent two bis 9(-) showed a 10000-fold and 1500-fold increase in the inhibition nor an neuroblastoma cell line of mice brain AChE and mice serum BChE, respectively SH-SY5Y. Cell viability was not affected for 5h and 5i at compared with(-)-MEP. Molecular docking elucidated that Sh concentrations of 1-100 uM(higher than the ICso values of 5h simultaneously bound to the catalytic and peripheral sites via and 5i against B-amyloid aggregation inhibition(around 80 uM) hydrophobic interactions with Trp86 and Trp286 in mAChE. and 10000 times higher than their AChE inhibiting ICso values) In comparison, it folded in the large aliphatic cavity of hBChE Conclusion The differences were explained by the absence of peripheral We have discovered novel nanomolar ChE inhibitors with site and the enlargement of the active site gorge in BChE h inhibitory potency on AB aggregation, i.e., a series of bis- Furthermore, 5h and 5i markedly prevented the AChE-induced

aggregation. These findings agree with the results from enzy￾matic test and molecular docking, which indicated that a linker of 9 or 10 methylenes would help to reach the peripheral site of AChE. Cell Viability. The toxicity of the most potent two bis-(-)- nor-MEPs was determined in human neuroblastoma cell line SH-SY5Y. Cell viability was not affected for 5h and 5i at concentrations of 1–100 µM (higher than the IC50 values of 5h and 5i against
-amyloid aggregation inhibition (around 80 µM) and 10000 times higher than their AChE inhibiting IC50 values). Conclusion We have discovered novel nanomolar ChE inhibitors with high inhibitory potency on A
aggregation, i.e., a series of bis- (-)-nor-MEPs 5. Their AChE inhibitory activities were closely related to the length of the alkylene chain, whereas BChE inhibition was less influenced. The optimal chain length for AChE and BChE inhibition was achieved with 9 in 5h, which showed a 10000-fold and 1500-fold increase in the inhibition of mice brain AChE and mice serum BChE, respectively, compared with (-)-MEP. Molecular docking elucidated that 5h simultaneously bound to the catalytic and peripheral sites via hydrophobic interactions with Trp86 and Trp286 in mAChE. In comparison, it folded in the large aliphatic cavity of hBChE. The differences were explained by the absence of peripheral site and the enlargement of the active site gorge in BChE. Furthermore, 5h and 5i markedly prevented the AChE-induced Figure 4. Representation of 5h (C atoms colored green) docked into the binding sites of mAChE (a) and hBChE (c). The binding site surfaces are colored according to the vacuum electrostatics protein contact potential, calculated by PyMOL 0.99rc2 (DeLano Scientific LLC, San Carlos, CA). Crucial catalytic and peripheral site (or the counterpart) residues are colored yellow. Hydrogen bonds and hydrophobic contacts between 5h and the protein residues of mAChE (b) and hBChE (d) are shown by LigPlot 4.4.2.41 Bis-(-)-nor-meptazinols as Inhibitors Journal of Medicinal Chemistry, 2008, Vol. 51, No. 7 2031