复旦大学：《药物设计学》课程教学资源（教学研究）基于配体结构的药物设计_Conformational re-analysis of（+）-meptazinol_an opioid with mixed analgesic pharmacophores

Blackwell Acta Pharmacologica Sinica 2006 Sep; 27(9):1247-125 b Publishing Full-length article Conformational re-analysis of (+)-meptazinol: an opioid with mixed anal- gesic pharmacophores Wei Ll, Xing-hai WANG, Choi-wan LAU, Yun TANG, Qiong XIE, Zhui- bai QIUZ Pharmacy, East China University of Science and lech School of Pharmacy, Fudan University, Shanghai 200032, China; School of 'Department of Medicinal Chemistry: NMR Labor chnology, Shanghai 200237, China Key words Abstract (+)meptazinol; opioid analgesics; arylpiperi- Aim: To further investigate the analgesic pharmacophore of (+)-meptazinol Methods: Two different opioid pharmacophores, Pharm-l and Pharm-Il, were es ablished from structures of nine typical opiates and meperidine by using molecu- Project supported by the National Natural lar modeling approaches according to their different structure activity relation- Science Foundation of China(No 30271539,. ship properties. They were further validated by a set of conformationally con- s Correspondence to Prof Zhui-bai QIU strained arylpiperidines. Two conformers of(+)-meptazinol( Conformer-I and Con- Phn86-21-5423-7595 former-If)detected in solution were then fitted into the pharmacophores Fax86-21-5423-7264. espectively, by Fit Atoms facilities available in SYBYL, a computational modeling alysis. Results: Conformer-I fit Pharm-I fro Received 2006-0 typical opiates well. However, Conformer-ll fit none of these pharmacophores Accepted 2006-0 Instead, it was found to be similar to another potent analgesic, benzofuro[ 2, 3-c doi:10.1111/.1745-72542006.00375.x pyridin-6-ol, whose pharmacophore was suggested to hold the transitional state between the two established pharmacophores. Unlike typical analgesics derived from 4-aryl piperidine(eg, meperidine)with one conformer absolutely overwhelming the(+)-meptazinol exists in two conformers with similar amounts in solution Furthermore, both conformers can not transform to each other freely in ordinary conditions based on our NMR results. Conclusion: (+)-meptazinol was suggested to be an opioid with mixed analgesic pharmacophores, which may account for the complicated pharmacological properties of meptazinol Introduction cophore, but they failed to fit that pharmacophore well4 Meanwhile, during the investigation of (+)-meptazinol hy Meptazinol [its(+) -isomer is shown in Figure I] is an drochloride in solution, two conformers(Conformer-I and opioid with similar analgesic activity to pethidine (also known II; Figure 1)were elucidated with NMR spectroscopy, in as meperidine). Since it entered into the market in the 1980s which the phenyl group of meptazinol could take ax-or eq- as racemic mixture, it has been recognized and recommended orientations, respectively, to the azepine ring. Conformer-II as a mixed agonist-antagonist opioid It was found to an- was very similar to Conformer-III(Figure 1), which was de- tagonize morphine induced physical dependence in vivo2, termined by X-ray crystallography. Here only the two NMR however, its analgesic mechanism remains unclear. For a conformers of(+-meptazinol were discussed. It is worthy to long time meptazinol was suggested to perform its analgesic note that the relative quantity of the two conformers in solu- function through the m, opioid receptor, which was be- tion were almost the same amount(approximately 3: 2) lieved to mediate analgesia while reducing the addition and although their calculated energy was somewhat different respiratory depression potential In our previous studies, the X-ray determined enatiomers As Conformer-Ill(very similar to Conformer-Il) failed to of meptazinol were compared with typical u opiate pharma- fit the typical opiate pharmacophore), it was reasonable to C2006 CPS and sIMM

1247 Acta Pharmacologica Sinica 2006 Sep ; 27 (9): 1247–1252 ©2006 CPS and SIMM Full-lengtharticle Conformational re-analysis of (+)-meptazinol: an opioid with mixed anal￾gesic pharmacophores1 Wei LI2 , Xing-hai WANG2 , Choi-wan LAU3 , Yun TANG2,4 , Qiong XIE2 , Zhui-bai QIU2,5 2Department of Medicinal Chemistry; 3NMR Laboratory, School of Pharmacy, Fudan University, Shanghai 200032, China; 4 School of Pharmacy, East China University of Science and Technology, Shanghai 200237, China Abstract Aim: To further investigate the analgesic pharmacophore of (+)-meptazinol. Methods: Two different opioid pharmacophores, Pharm-I and Pharm-II, were es￾tablished from structures of nine typical opiates and meperidine by using molecu￾lar modeling approaches according to their different structure activity relation￾ship properties. They were further validated by a set of conformationally con￾strained arylpiperidines. Two conformers of(+)-meptazinol (Conformer-I andCon￾former-II) detected in solution were then fitted into the pharmacophores, respectively, byFit Atomsfacilities available in SYBYL, a computational modeling tool kit for molecular design and analysis. Results: Conformer-I fit Pharm-I from typical opiates well. However, Conformer-II fit none of these pharmacophores. Instead, it was found to be similar to another potent analgesic, benzofuro[2,3-c] pyridin-6-ol, whose pharmacophore was suggested to hold the transitional state between the two established pharmacophores. Unlike typical analgesics derived from4-aryl piperidine(eg,meperidine)with oneconformer absolutelyoverwhelming, the (+)-meptazinol exists in two conformers with similar amounts in solution. Furthermore, both conformers can not transform to each other freely in ordinary conditions based on our NMR results. Conclusion: (+)-meptazinol wassuggested to be an opioid with mixed analgesic pharmacophores, which may account for the complicated pharmacological properties of meptazinol. Key words (+)meptazinol; opioid analgesics; arylpiperi￾dines; Conformational Analysis, pharmacophore. 1 Project supported by the National Natural Science Foundation of China (No 30271539, 2003–2005). 5 Correspondence to Prof Zhui-bai QIU. Ph n 86-21-5423-7595. Fax 86-21-5423-7264. E-mail zbqiu@shmu.edu.cn Received 2006-03-12 Accepted 2006-04-24 doi: 10.1111/j.1745-7254.2006.00375.x Introduction Meptazinol [its (+)-isomer is shown in Figure 1] is an opioid with similar analgesic activityto pethidine (also known as meperidine). Since it entered into the market in the 1980s as a racemicmixture, it has been recognized and recommended as a mixed agonist-antagonist opioid[1] . It was found to an￾tagonize morphine induced physical dependence in vivo[2] , however, its analgesic mechanism remains unclear. For a long time meptazinol wassuggested to perform its analgesic function through the m1 opioid receptor[3] , which was be￾lieved to mediate analgesia while reducing the addition and respiratory depression potential. In our previousstudies, the X-ray determined enatiomers of meptazinol were compared with typical µ opiate pharma￾cophore, but they failed to fit that pharmacophore well[4] . Meanwhile, during the investigation of (+)-meptazinol hy￾drochloride in solution, two conformers (Conformer-I and - II; Figure 1) were elucidated with NMR spectroscopy, in which the phenyl group of meptazinol could take ax- or eq￾orientations, respectively, to the azepine ring. Conformer-II was very similar toConformer-III (Figure 1), which was de￾termined byX-ray crystallography. Here only the two NMR conformers of (+)-meptazinol were discussed. It is worthy to note that the relative quantity of the two conformersin solu￾tion were almost the same amount (approximately 3:2), although their calculated energy was somewhat different (Table 1). As Conformer-III (very similar toConformer-II) failed to fit the typical opiate pharmacophore[4] , it was reasonable to

Liw et al Acta Pharmacologica Sinica ISSN 1671-4083 户歌e (+meptazinol Conformer-lll Figure 1. (+)-Meptazinol and its expe conformations. Conformer-I and -ll were detected in d- DMSO solution, and Conformer-lll was resolved from X-ray crystal. All hy atoms in conformers were removed for clarity purposes Table 1. Difference detected between the two conformers of(+)- formers with all basic nitrogen atoms protonated. The initial structures were firstly minimized by Tripos force field with Gasteiger-Huckel charges, followed by random searches to Conformer-l Conformer-lI ensure their stable conformations by default settings. Finally, all structures were optimized with the semi-empirical quan- Aromatic orientation to piperidine Axial Equatorial tum chemical method Austin Model 1(AMl; available in ted energy(ko SYBYL), and the geometrically optimized structures were Relative content from NMR evidence 1.00 1.53 used for further structural comparisons Establishment and validation of two opioid pharmaco- consider that Conformer-l, with higher energy, was the bio- II, were defined from structures of nine typical opiates and active conformer while Conformer-ll(or Conformer-lII)was meperidine, separately, Figure 2a and 3a). Each pharma- inactive. After all, a conformer with the lowest energy is not cophore contains one aromatic region(defined by the center al ways the biological active conformer for a compound with of phenyl rings), one cation center and a hinge atom. In certain biological activities addition, two distance and one angle parameter were also In this study, two typical pharmacophores from different calculated to constrain these components'distribution in opioids were established and validated. And both conform- space(Table 2). Pharm-I was modeled from the mean values ers of (+)-meptazinol were used for structural comparisons of nine typical opiates( Figure 2b). And the root mean square to analyze their specific properties. This study provides deviations(RMSD)fitting values of nine opiates were listed ights into understanding the nature of mixed in Table 3. Pharm-lI was extracted from the well-established pharmacophores to(+)-meptazinol and the complicated phar- meperidine conformer(Figure 3b) acology of meptazinol With the defined pharmacophores, all structural compar sons were made by Fit Atoms facilities available in SYBYL Materials and methods And the established pharmacophores were validated by fit Material preparation All calculations were carried out ting a set of constrained arylpiperidines(Table 4) on a r14000 SGl Fuel workstation using the molecular mod- Fitting conformers of (+)-meptazinol to the pharmaco- eling software package SYBYL version 6.9(Tripos, St Louis, phores Both Conformer-land-ll of(tj-meptazinol were fit MO, USA) parately The initial pharmacophore of morphine analogs were they share similar properties. Furthermore, both conformers modeled in our previously published paper!). The conformer were superimposed with the structure of compound-4(Table 5) of meperidine was well examined by X-ray crystallographic and NMRil techniques and thus was modeled according to Results and discussion literature. The two different conformers of (+)-meptazinol Establishment and validation of opioid pharmacophores ( Conformer-I and-II)were obtained by molecular dynamics It was suggested that arylpiperidines and structural related from NOESY and TOCSY spectra. Several conformationally analgesics could be divided into two categories according constrained aromatic substituted piperidines -ll were to the orientation of the aryl groupl: ax-and eq-phenyl sketched in SYBYL according to the reported preferred con- models. The prototype analgesics of ax-and eq- phenyl 1248

1248 Li W et al Acta Pharmacologica Sinica ISSN 1671-4083 consider that Conformer-I, with higher energy, was the bio￾active conformer whileConformer-II (or Conformer-III) was inactive. After all, a conformer with the lowest energy is not always the biological active conformer for a compound with certain biological activities. In this study, two typical pharmacophoresfrom different opioids were established and validated. And both conform￾ers of (+)-meptazinol were used for structural comparisons to analyze their specific properties. This study provides new insights into understanding the nature of mixed pharmacophoresto (+)-meptazinol and the complicated phar￾macology of meptazinol. Materials and methods Material preparation All calculations were carried out on a R14000 SGI Fuel workstation using the molecular mod￾eling software package SYBYLversion 6.9 (Tripos, St Louis, MO, USA). The initial pharmacophore of morphine analogs were modeled in our previously published paper[4] . The conformer of meperidine was well examined byX-ray crystallographic[5] and NMR[6] techniques and thus was modeled according to literature. The two different conformers of (+)-meptazinol (Conformer-I and -II) were obtained by molecular dynamics from NOESY and TOCSY spectra. Several conformationally constrained aromatic substituted piperidines[6–11] were sketched in SYBYL according to the reported preferred con￾formers with all basic nitrogen atoms protonated. The initial structures were firstly minimized by Tripos force field with Gasteiger-Huckel charges, followed by random searches to ensure theirstable conformations by default settings. Finally, all structures were optimized with the semi-empirical quan￾tum chemical method Austin Model 1 (AM1; available in SYBYL)[12] , and the geometrically optimized structures were used for further structural comparisons. Establishment and validation of two opioid pharmaco￾phores Two opioid pharmacophores, namely Pharm-I and - II, were defined from structures of nine typical opiates and meperidine, separately, (Figure 2a and 3a). Each pharma￾cophore contains one aromatic region (defined by the center of phenyl rings), one cation center and a hinge atom. In addition, two distance and one angle parameter were also calculated to constrain these components’ distribution in space (Table 2). Pharm-I was modeled from the mean values of nine typical opiates(Figure 2b). And the root mean square deviations (RMSD) fitting values of nine opiates were listed in Table 3. Pharm-II was extracted from the well-established meperidine conformer (Figure 3b). With the defined pharmacophores, allstructural compari￾sons were made by Fit Atoms facilities available in SYBYL. And the established pharmacophores were validated by fit￾ting a set of constrained arylpiperidines (Table 4). Fitting conformers of (+)-meptazinol to the pharmaco￾phores Both Conformer-I and -II of (+)-meptazinol were fit￾ted into both opioid pharmacophores separately to see if they share similar properties. Furthermore, both conformers were superimposed with the structure of compound-4 (Table 5). Results and discussion Establishmentandvalidationofopioidpharmacophores It was suggested that arylpiperidines and structural related analgesics could be divided into two categories according to the orientation of the aryl group[13]: ax- and eq- phenyl models. The prototype analgesics of ax- and eq- phenyl Table 1. Difference detected between the two conformers of (+)- meptazinol. Conformer-I Conformer-II Aromatic orientation to piperidine Axial Equatorial Calculated energy (kcal/mol) 29.521 25.870 Relative content from NMR evidence 1.00 1.53 Figure 1. (+)-Meptazinol and its experimental conformations. Conformer-I and -II were detected in d6 -DMSO solution, and Conformer-III was resolved from X-ray crystal. All hydrogen atoms in conformers were removed for clarity purposes

Http:lwww.chinaphar.com Liw et al Figure 2. (a)The definition of Pharm-I;(b) common sul ture among nine typical opiates sed to establish Pharm-1; (c)fit. hinge atom ting Conformer-l of(+)-mepta- nine typical opiates zino to pharm.l. The st aromatic region Pharm-I with thick lines is Conformer-I of (+)-meptazinol Table 2. Parameters observed in Pharm-I and -ll from the Table 3. Root mean square deviation(RMS D) values of nine typical opiates fitting to Pharm-I Distance A(A) Distance B(A) Angle(degree) Compound RMSD(A)Compound RMSD(A) Pharm-I 2.853 Morphine 0.053 Nalbuphine 0.040 Pharm-l 151.80° O-methyletorphine 0.031 Codeine Azidomorphine 0.053 6-Hydroxylevalorphan 0.099 a Distance between the aromatic nd the hinge atom: distance Naloxone 0.085 between the hinge atom and the enter;angle formed by the yclazocine aromatic center, the hinge atom classified as ax- phenyl models not only for similar phenyl orientation but also for similar structure activity relationships Although there was evidence that analgesics in these N-CH categories were both u opioid ligands, these two models meperidine H bear different structure activity relationships properties based on the following facts: 1)N-allyl or cyclopropyl methyl ana logs in ax-phenyl conformers induced antagonist activity, but those in eq-phenyl conformers did not, 2)introduc harm of compound tion of a 3R-methyl group into the piperidine ring in analogs of eq-phenyl conformers mediates pure antagonist potencyls1 Hinge alu but not in those of ax- phenyl conformers; 3)in reversed esters of meperidine with preferred eg-phenyl conformers henolic hydroxyl insertion virtually abolishes activity 6, 71 However, in most other analogs, an m-placed phenolic hy droxyl group enhances analgesic activity. The major struc- tural difference between these two categories was the differ- Figure 3.(a) The definition of Pharm-l1; (b)structure of meperidine ent orientations of their aryl groups which mainly account used to establish Pharm-ll;(c)fitting Conformer-ll of (+)-meptazinol for different pharmacological properties. Thus Portoghese to compound-4. The structure with thick lines is conformer-ll of et al proposed different binding models to H opioid receptor (+)-meptazinol;(d)the description of compound-4 conformer: a for compounds in these categories"y transitional state between Pharm-I and Pharm-Il envi ove laps were only used to display relative positions of Here, typical opioids of ax-phenyl(nine opiates)and eq- aromatic center other than that they share aromatic phenyl(meperidine)were used to establish their correspond pharmacophores ing pharmacophores. Although the tyramine fragment was widely used as typical opioid pharmacophore in the pub- models were morphine and meperidine, respectively. And lished literaturelll as well as in our previous studiesI,the arylpiperidines with phenyl ax-conformer could also be phenyl ring of ax-arylpiperidines actually bears a relation-

Http://www.chinaphar.com Li W et al 1249 classified as ax- phenyl models not only for similar phenyl orientation but also for similarstructure activityrelationships. Although there was evidence that analgesics in these categories were both µ opioid ligands[14] , these two models bear different structure activity relationships properties based on the following facts: 1) N-allyl or cyclopropyl methyl ana￾logs in ax- phenyl conformers induced antagonist activity, but those in eq- phenyl conformers did not[13]; 2) introduc￾tion of a 3R-methyl group into the piperidine ring in analogs of eq- phenyl conformers mediates pure antagonist potency[15] but not in those of ax- phenyl conformers; 3) in reversed esters of meperidine with preferred eq- phenyl conformers, phenolic hydroxyl insertion virtually abolishes activity[16,17] . However, in most other analogs, an m-placed phenolic hy￾droxyl group enhances analgesic activity. The major struc￾tural difference between these two categories was the differ￾ent orientations of their aryl groups which mainly account for different pharmacological properties. Thus Portoghese et al proposed different binding models to m opioid receptor for compounds in these categories[16] . Here, typical opioids of ax-phenyl (nine opiates) and eq￾phenyl (meperidine) were used to establish their correspond￾ing pharmacophores. Although the tyramine fragment was widely used as typical opioid pharmacophore in the pub￾lished literature[18] as well as in our previous studies[4] , the phenyl ring of ax- arylpiperidines actually bears a relation￾Table 2. Pa ramet ers observed in Pha rm-I a nd -II from the pharmacophores defined. Distance A (Å) a Distance B (Å) b Angle (degree)c Pharm-I 2.853 2.914 107.54º Pharm-II 2.916 2.996 151.80º a Distance between the aromatic center and the hinge atom; bdistance between the hinge atom and the cation center; cangle formed by the aromatic center, the hinge atom and the cation center. Table 3. Root mean square deviation (RMSD) values of nine typical opiates fitting to Pharm-I. Compound RMSD (Å) Compound RMSD (Å) Morphine 0.053 Nalbuphine 0.040 3-O-methyletorphine 0.031 Codeine 0.054 Azidomorphine 0.053 6-Hydroxylevalorphan 0.099 N-allynormetazocine 0.117 Naloxone 0.085 Cyclazocine 0.107 Figure 2. (a) The definition of Pharm-I; (b) common sub-struc￾ture among nine typical opiates used to establish Pharm-I; (c) fit￾ting Conformer-I of (+)-mepta￾zinol to Pharm-I. The structure with thick lines is Conformer-I of (+)-meptazinol. models were morphine and meperidine, respectively. And arylpiperidines with phenyl ax-conformer could also be Figure 3. (a) The definition of Pharm-II; (b) structure of meperidine used to establish Pharm-II; (c) fitting Conformer-II of (+)- meptazinol to compound-4. The structure with thick lines is conformer-II of (+)-meptazinol; (d) the description of compound-4 conformer: a transitional state between Pharm-I and Pharm-II. The phenyl over￾laps were only used to display relative positions of cation centers to aroma ti c cent er other than tha t they shar e simila r a romatic pharmacophores

Li w et al Acta Pharmacologica Sinica ISSN 1671-4083 Table 4. Pharmacophore validation by conformer constrained aryl-piperidine analogs Pharmacophore Fitting Pharmacophore Fitting RMSD(A) RMSD(A) Com Pharm-l Pharm-Il Pharm-ll 0.507 0.170 0.619 0.141 3 0.761 0.010 0.390 H C ④ O: CH tropane analog trans.da-arvldecahvdro m-hydroxy phenyl-35-azabie 6 benzofuro[2,3-cl aloe Table 5. Conformers of (+)-meptazinol fitting to Pharm-l and -ll ax-arylpiperidines to morphine derived tyramine fragments and compound-4 will lead to poor overlaps. Considering phenyl groups co form either parallel or T-shape T-It interactions with the hy Pharm-I(A) Pharm-1I(A)Compound 4(A) drophobic site, and hydroxyl groups in a given region could also interact with the hydrogen acceptor on the receptor, it 0.142 seemed unnecessary to consider the overall fit between the ormer-ll 0.424 0.289 0.182 phenolic groups between ax-arylpiperidine and opiates Here, we excised the tyramine fragment down to an aromatic center, a cation center and a hinge atom connecting piperi ship orthogonal to that obtained in the rigid skeleton of dine and phenyl rings to unify these structurally diverse morphinell93o). Thus not only the phenyl plane but also the opioids with ax-phenyl conformers. The hydroxyl group hydroxyl group on the aromatic ring between opiates and was excluded from our pharmacophore not only for its ax-arylpiperidines were quite different. Superimposing ambiguous function in arylpiperidines, but for poor fittir

1250 Li W et al Acta Pharmacologica Sinica ISSN 1671-4083 ship orthogonal to that obtained in the rigid skeleton of morphine[19,20] . Thus not only the phenyl plane but also the hydroxyl group on the aromatic ring between opiates and ax- arylpiperidines were quite different. Superimposing ax- arylpiperidines to morphine derived tyramine fragments will lead to poor overlaps. Considering phenyl groups could form either parallel or T-shape p-p interactions with the hy￾drophobic site, and hydroxyl groups in a given region could also interact with the hydrogen acceptor on the receptor, it seemed unnecessary to consider the overall fit between the phenolic groups between ax- arylpiperidine and opiates. Here, we excised the tyramine fragment down to an aromatic center, a cation center and a hinge atom connecting piperi￾dine and phenyl rings to unify these structurally diverse opioids with ax- phenyl conformers. The hydroxyl group was excluded from our pharmacophore not only for its ambiguous function in arylpiperidines, but for poor fitting Table 4. Pharmacophore validation by conformer constrained aryl-piperidine analogs. Pharmacophore Fitting Pharmacophore Fitting RMSD (Å) RMSD (Å) Compound Pharm-I Pharm-II Compound Pharm-I Pharm-II 1 0.507 0.170 5 0.667 0.006 2 0.082 0.619 6 0.804 0.141 3 0.091 0.761 7 0.664 0.010 4 0.390 0.297 Table 5. Conformers of (+)-meptazinol fitting to Pharm-I and -II and compound-4. Pharm-I (Å) Pharm-II (Å) Compound 4(Å) Conformer-I 0.142 0.754 0.492 Conformer-II 0.424 0.289 0.182

Http:lwww.chinaphar.com Liw et al esults among different analgesics, although it was also used Another question raised was whether these two structures for structural reference in this study. The relative orienta- shared similar properties. Considering its rigid structure and tion of the phenyl ring toward the piperidine group could be potency, compound-4 could also be employed as an eligible simply determined by the angle values defined in Table 2. template for structural comparison. With the same compo- Common model definitions for these two pharmacophores nents of pharmacophore defined previously, fitting Con were as follows: 1) The center of aryl rings to form hydro- former-ll to compound-4 led to good over lap results with a phobic interactions with the opioid receptor; 2) the pre RMSD value of0. 182 (Figure 3c and Table 5) nated nitrogen atom to form salt bridges with anion center In addition, the angle parameter of compound-4 was on the receptor; 3)the hinge atom that held the relative ori- 129.50, a halfway transition state between Pharm-I and -ll entation of phenyl to piperidine ring. Although the two(Figure 3d). And it was 136.56 for Conformer-Il of(+)- pharmacophores shared similar distance parameters(Table meptazinol. Typical arylpiperidines generally took Pharm-l 2), the angle parameters were quite different. The angle was (ax-phenyD)or Pharm-ll(eq-phenyl)conformers to avoid 107.54 in Pharm-I(with ax- phenyl conformer) and 151. 8 in disfavored steric clashes. But the furan ring constrained Pharm-lI(with eq-phenyl conformer), which could also be compound-4 took such a distorted conformer that it belonged regarded as the main difference between the two pharmaco- to neither of these two conformers. Instead, it held a transi ores tion conformer between the two conformers. Considering Some typical constrained arylpiperidines with known other arylpiperidines with ring constrains lacked analgesi otent analgesic potencies were used for validation of the potencies 2, compound-4 may display specific pharmaco- modeled pharmacophores. And all of these compounds could logical properties other than Pharm-I and-II. Though con- be identified to their correct pharmacophores by RMSD fit- former-ll of(+)-meptazinol failed to fit Pharm-Iand Pharm-II ting values below 0.2A, as well as by determining their phe- it fit compound-4 well. So conformer-ll was also suggested nyl orientations except for compound-I and 4. Although to be an active conformer as analgesics compound-1 took ax-phenyl conformer, it seemed more rea- Although both conformers of (+)-meptazinol were sug sonable to be classified to Pharm -lI because the framework gested to be active conformers against the u opioid recepto of the tropane ring constrained the relative position of the their pharmacophores were different to each other. Con defined three components to Pharm-ll(with eq- phenyl former-I was suggested to fit typical opiate pharmacophore conformer). However, compound 4 could be classified to (Pharm-D)and conformer-II shared similar properties with neither Pharm-I nor Pharm-ll and its special pharmacophore benzofuro(2, 3-clpyridin-6-ol analogs. These conformers ex will be discussed in the next section of this paper, although ist in similar amounts in solution and could not transform to its phenyl ring took eq-conformer each other freely in ordinary conditions. Although it is also Structural comparison of (+)-meptazinol with opioid possible that (-)-meptazinol may also account for analgesics pharmacophores As a ring expanded the analog of 4- and reduced side-effects as meptazinol was marketed as a arylpiperidine, both(+)-meptazinol conformers were fitted to racemic mixture. The existence of multiple analgesic mecha- the established opioid pharmacophores. Conformer-I was nisms in(+)-meptazinol may provide further insight into found to fit Pharm-I well but not Pharm-Il(Figure 2c and meptazinol's complex pharmacological behavior However, Conformer-II fit none of these pharmacophores References (Table 5). Similar conclusions were made in our previous I Hoskin PJ, Hanks GW. Opioid agonist-antagonist drugs in acute studies based on tyramine fragment derived opiate and chronic pain states. Drugs 1991: 41: 326-44 pharmacophorel4l. Furthermore, Conformer-II was energy 2 Green D. Current concepts concerning the mode of action of favored with calculated energy 3. 651 kcal/ mol lower than Conformer-L. Should Conformer-Il also be considered as an 3 Spiegel K, Pasternak Gw. Meptazinol: a novel mu-I selective inactive conformer? opioid analgesic. J Pharmacol Exp Ther 1984: 228: 414B Surprisingly, conformer-II showed similar fitting value Li W, Hao JL, Tang Y, Chen Y, Qiu ZB. Structural comparisons to both pharmacophores(0.424 for Pharm-l and 0.289 for of meptazinol with opioid analgesics. Acta Pharmacol Sin 200 Pharm-Il)as compound-4(0.390 for Pharm-I and 0 297 for 26:334-8 Pharm-ID), although they fit neither pharmacophore(Table 4 5 Tillack JV, RC. Kennard CHL. Oh PWT pethidine hydrochloride, 4-car and Table 5). Limited studies were carried out on these ana- 4-phenylpiperidine hydrochloride, C1sH NO, HCl. Recl Trav st Chim Pays-Bas 1974: 93: 164-5

Http://www.chinaphar.com Li W et al 1251 results among different analgesics, although it was also used for structural reference in this study. The relative orienta￾tion of the phenyl ring toward the piperidine group could be simply determined by the angle values defined in Table 2. Common model definitionsfor these two pharmacophores were as follows: 1) The center of aryl rings to form hydro￾phobic interactions with the opioid receptor; 2) the proto￾nated nitrogen atom to form salt bridges with anion center on the receptor; 3) the hinge atom that held the relative ori￾entation of phenyl to piperidine ring. Although the two pharmacophores shared similar distance parameters (Table 2), the angle parameters were quite different. The angle was 107.54º in Pharm-I(with ax- phenyl conformer) and 151.8º in Pharm-II (with eq- phenyl conformer), which could also be regarded as the main difference between the two pharmaco￾phores. Some typical constrained arylpiperidines with known potent analgesic potencies were used for validation of the modeled pharmacophores. And all of these compounds could be identified to their correct pharmacophores by RMSD fit￾ting values below 0.2Å, as well as by determining their phe￾nyl orientations except for compound-1 and 4. Although compound-1 took ax- phenyl conformer, it seemed more rea￾sonable to be classified to Pharm-II because the framework of the tropane ring constrained the relative position of the defined three components to Pharm-II (with eq- phenyl conformer). However, compound 4 could be classified to neither Pharm-I nor Pharm-II and itsspecial pharmacophore will be discussed in the next section of this paper, although its phenyl ring took eq- conformer. Structural comparison of (+)-meptazinol with opioid pharmacophores As a ring expanded the analog of 4- arylpiperidine, both (+)-meptazinol conformers were fitted to the established opioid pharmacophores. Conformer-I was found to fit Pharm-I well but not Pharm-II (Figure 2c and Table 5). However, Conformer-II fit none ofthese pharmacophores (Table 5). Similar conclusions were made in our previous studies based on tyramine fragment derived opiate pharmacophore[4] . Furthermore, Conformer-II was energy favored with calculated energy 3.651 kcal/mol lower than Conformer-I. Should Conformer-II also be considered as an inactive conformer? Surprisingly, conformer-II showed similar fitting values to both pharmacophores (0.424 for Pharm-I and 0.289 for Pharm-II) as compound-4 (0.390 for Pharm-I and 0.297 for Pharm-II), although they fit neither pharmacophore (Table 4 and Table 5). Limited studies were carried out on these ana￾logs[10,11] , but compound-4 was a potent µ opioid agonist. Another question raised was whether these two structures shared similar properties. Considering itsrigid structure and potency, compound-4 could also be employed as an eligible template for structural comparison. With the same compo￾nents of pharmacophore defined previously, fitting Con￾former-II to compound-4 led to good overlap results with a RMSD value of 0.182 (Figure 3c and Table 5). In addition, the angle parameter of compound-4 was 129.50º , a halfway transition state between Pharm-I and -II (Figure 3d). And it was 136.56º for Conformer-II of (+)- meptazinol. Typical arylpiperidines generally took Pharm-I (ax- phenyl) or Pharm-II (eq- phenyl) conformers to avoid disfavored steric clashes. But the furan ring constrained compound-4 took such a distorted conformer that it belonged to neither of these two conformers. Instead, it held a transi￾tion conformer between the two conformers. Considering other arylpiperidines with ring constrains lacked analgesic potencies[21] , compound-4 may display specific pharmaco￾logical properties other than Pharm-I and -II. Though con￾former-II of(+)-meptazinolfailed to fitPharm-I and Pharm-II, it fit compound-4 well. So conformer-II was also suggested to be an active conformer as analgesics. Although both conformers of (+)-meptazinol were sug￾gested to be active conformers against the µ opioid receptor, their pharmacophores were different to each other. Con￾former-I was suggested to fit typical opiate pharmacophore (Pharm-I) and conformer-II shared similar properties with benzofuro[2,3-c]pyridin-6-ol analogs. These conformers ex￾ist in similar amounts in solution and could not transform to each other freely in ordinary conditions. Although it is also possible that (-)-meptazinol may also account for analgesics and reduced side-effects as meptazinol was marketed as a racemic mixture. The existence of multiple analgesic mecha￾nisms in (+)-meptazinol may provide further insight into meptazinol’s complex pharmacological behavior. References 1 Hoskin PJ, Hanks GW. Opioid agonist-antagonist drugs in acute and chronic pain states. Drugs 1991; 41: 326–44. 2 Green D. Current concepts concerning the mode of action of meptazinol as an analgesic. Postgrad Med J 1983; 59: 9–12. Supp1. 3 Spiegel K, Pasternak GW. Meptazinol: a novel mu-1 selective opioid analgesic. J Pharmacol Exp Ther 1984; 228: 414B. 4 Li W, Hao JL, Tang Y, Chen Y, Qiu ZB. Structural comparisons of meptazinol with opioid analgesics. Acta Pharmacol Sin 2005; 26: 334–8. 5 Tillack JV, Seccombe RC, Kennard CHL, Oh PWT. Analgetics. I. Crystal structure of pethidine hydrochloride, 4-carbethoxy-1- 4-phenylpiperidine hydrochloride, C15H21NO2 ·HCl. Recl Trav Chim Pays-Bas 1974; 93:164–5

Li w et al Acta Pharmacologica Sinica ISSN 1671-4083 6 Casy AF, Dewar GH, Al-Deeb OAA. Conformational equilibra of Plenum Press: New York and London, 1986, p485 hy drochloride salts of pethidine, ketobemidone, and related cen- 14 Tyers MB. A classification of opiate receptors that mediate al analgesics of the 4-arylpiperidine class. J Chem Soc Perkin antinociception in animals. Br J Pharmacol 1980: 69: 503-12 Trans21989;2:1243-7 15 Zimmerman DM, Nickander R, Horng JS, Wong DT. New struc- 7 Daum SJ, Martini CM, Kullnig RK, Clarke RL. Analgesic activity tural concepts for narcotic-antagonists defined in a 4-phenylpi- of the epimeric tropane analogs of meperidine. Physical and peridine series. Nature 1978: 275: 332-4 pharmacological study. J Med Chem 1975: 18: 496-501 1 6 Portoghese PS, Alreja BD, Larson DL. Allylprodine analogs as 8 Zimmerman DM. Cantrell BE Swartzendruber JK Jones ND receptor probes-evidence that phenolic and nonphenolic ligands Mendelsohn LG, Leander JD, et al. Synthesis and analgesic prop- nteract with different subsites on identical opioid Receptors. J erties of N-substituted trans-4a-aryldecahydro isoquinolines. J Med Chem1981;24:782-1 Med Chen1988;31:555-60 Casy aF, Ogungbamila FO. Phenolic analogs of reversed esters 9 Froimowitz M, Salva P, Hite GJ, Gianntsos G, Suzdak P, Heyman of pethidine. J Pharm Pharmacol 1985: 37: 121-3 R. Conformational properties of a- and B-azabicyclane opiates. 18 Horn AS, Rodgers JR. The enkephalins and opiates: structure- The effect of conformation on pharmacological activity. J activity relations. J Pharm Pharmacol 1977; 29: 257-65 Comput Chem 1 19 Hodgson DJ, Rychlewska U, Eliel EL, Manoharan M, Knox DE, 10 Hutchison A. de Jesus r. wiliams M. Simke JP. Neale RF. Jack- Olefirowicz EM. Rotational conformation of the phenyl moi son RH, et al. Benzofuro[2, 3-cIpyridine-6-ols: synthesis, affin- ety in geminally substituted phenylcyclohexane with equatorial ity for opioid subtypes, and antinociceptive activity. J Med phenyl. J Org Chem 1985: 50: 4838-43 Chem1989;32:2221-6 20 Allinger NL, Tribble MT. Conformational analysis LxXVIll I 1 Froimowitz M, Pick CG, Pasternak Gw. Phenylmorphans and he conformation of phenyl cyclohexane and related molecules analogs-opioid receptor subtype selectivity and effect of con mation on activity. J Med Chem 1992: 35: 1521-5 21 Burke TR, Bajwa BS, Jacobson AE, Rice KC, Streaty RA, Klee 12 Dewar MJS, Zoebisch EG, Healy EF, Stewart JJP. AM1: a new WA. Probes for narcotic receptor mediated phenomena 7. Syn- neral purpose quantum mechanical model. J Am Chem Soc thesis and pharmacological properties of irreversible ligands spe- 1985;107:3902-9. cific for mu and delta opiate receptors. J Med Chem 1984: 27 14 Casy A F, Parfitt RT. Opioid analgesics: Chemistry and receptor 1570-4

1252 Li W et al Acta Pharmacologica Sinica ISSN 1671-4083 6 Casy AF, Dewar GH, Al-Deeb OAA. Conformational equilibra of hydrochloride salts of pethidine, ketobemidone, and related cen￾tral analgesics of the 4-arylpiperidine class. J Chem Soc Perkin Trans 2 1989; 2: 1243–7. 7 Daum SJ, Martini CM, Kullnig RK, Clarke RL. Analgesic activity of the epimeric tropane analogs of meperidine. Physical and pharmacological study. J Med Chem 1975; 18: 496–501. 8 Zimmerman DM, Cantrell BE, Swartzendruber JK, Jones ND, Mendelsohn LG, Leander JD, et al. Synthesis and analgesic prop￾erties of N-substituted trans-4a-aryldecahydro isoquinolines. J Med Chem 1988; 31: 555–60. 9 Froimowitz M, Salva P, Hite GJ, Gianntsos G, Suzdak P, Heyman R. Conformational properties of a- and b-azabicyclane opiates. The effect of conformation on pharmacological activity. J Comput Chem 1984; 5; 291–8. 1 0 Hutchison AJ, de Jesus R, Wiliams M, Simke JP, Neale RF, Jack￾son RH, et al. Benzofuro[2,3-c]pyridine-6-ols: synthesis, affin￾ity for opioid subtypes, and antinociceptive activity. J Med Chem 1989; 32: 2221–6. 1 1 Froimowitz M, Pick CG, Pasternak GW. Phenylmorphans and analogs-opioid receptor subtype selectivity and effect of confor￾mation on activity. J Med Chem 1992; 35: 1521–5. 1 2 Dewar MJS, Zoebisch EG, Healy EF, Stewart JJP. AM1: a new general purpose quantum mechanical model. J Am Chem Soc 1985; 107: 3902–9. 1 4 Casy A F, Parfitt RT. Opioid analgesics: Chemistry and receptor. Plenum Press: New York and London, 1986, p485. 1 4 Tyers MB. A classification of opiate receptors that mediate antinociception in animals. Br J Pharmacol 1980; 69: 503–12. 1 5 Zimmerman DM, Nickander R, Horng JS, Wong DT. New struc￾tural concepts for narcotic-antagonists defined in a 4-phenylpi￾peridine series. Nature 1978; 275: 332–4. 1 6 Portoghese PS, Alreja BD, Larson DL. Allylprodine analogs as receptor probes - evidence that phenolic and nonphenolic ligands interact with different subsites on identical opioid Receptors. J Med Chem 1981; 24: 782–7. 1 7 Casy AF, Ogungbamila FO. Phenolic analogs of reversed esters of pethidine. J Pharm Pharmacol 1985; 37: 121–3. 1 8 Horn AS, Rodgers JR. The enkephalins and opiates: structure￾activity relations. J Pharm Pharmacol 1977; 29: 257–65. 1 9 Hodgson DJ, Rychlewska U, Eliel EL, Manoharan M, Knox DE, Olefirowicz EM. Rotational conformation of the phenyl moi￾ety in geminally substituted phenylcyclohexanes with equatorial phenyl. J Org Chem 1985; 50: 4838–43. 2 0 Allinger NL, Tribble MT. Conformational analysis LXXVIII. The conformation of phenyl cyclohexane and related molecules. Tetrahedron Lett 1971; 12: 3259–62. 2 1 Burke TRJ, Bajwa BS, Jacobson AE, Rice KC, Streaty RA, Klee WA. Probes for narcotic receptor mediated phenomena 7. Syn￾thesis and pharmacological properties of irreversible ligands spe￾cific for mu and delta opiate receptors. J Med Chem 1984; 27: 1570–4