A Journey through Diastereomeric Space: The Design, Synthesis, In Vitro and In Vivo Pharmacological Activity, and Molecular Modeling of Novel Potent Diastereomeric MOR Agonists and Antagonists

Four sets of diastereomeric C9-alkenyl 5-phenylmorphans, varying in the length of the C9-alkenyl chain, were designed to examine the effect of these spatially distinct ligands on opioid receptors. Functional activity was obtained by forskolin-induced cAMP accumulation assays and several compounds were examined in the [35S]GTPgS assay and in an assay for respiratory depression. In each of the four sets, similarities and differences were observed dependent on the length of their C9-alkenyl chain and, most importantly, their stereochemistry. Three MOR antagonists were found to be as or more potent than naltrexone and, unlike naltrexone, none had MOR, KOR, or DOR agonist activity. Several potent MOR full agonists were obtained, and, of particular interest partial agonists were found that exhibited less respiratory depression than that caused by morphine. The effect of stereochemistry and the length of the C9-alkenyl chain was also explored using molecular modeling. The MOR antagonists were found to interact with the inactive (4DKL) MOR crystal structures and agonists were found to interact with the active (6DDF) MOR crystal structures. The comparison of their binding modes at the mouse MOR was used to gain insight into the structural basis for their stereochemically induced pharmacological differences.


Introduction
Opium from the plant Papaver somniferum has been used for millennia, and the opioids isolated from the plant have been clinically used for almost two centuries to treat acute and chronic pain. More recently, these opioids and their derivatives have become controversial marked differences from morphine-like epoxymorphinan structures in the attachment of the aromatic ring.
In our former work with N-phenethyl-3-hydroxy-5-phenylmorphans, we noted that a C9-hydroxy, methyl, and a vinyl substituent with specific stereochemistry gave moderately potent, and in some instances extremely potent MOR agonists in vitro and in vivo, and these appeared to be morphine-like in their side effects [14]. We thought that it might be of interest to synthesize diastereomeric sets of compounds with alkenyl substituents at C9 on the 5-phenylmorphan nucleus. Since there are three chiral atoms in these C9-alkenyl substituted molecules and one of them is structurally fixed, we would only need to synthesize four compounds for each C9-alkene moiety (two sets of enantiomers).

Chemistry
We sought synthetic pathways for access to the desired two-carbon C9-alkenyl 5phenylmorphans using the known intermediate 1 [14][15][16]. This intermediate underwent von Braun demethylation to form the secondary amine 2, followed by alkylation with phenethyl bromide to give intermediate 3 using optimized synthetic procedures [16]. The formation of enol ether 4 was achieved by a Wittig olefination which gave a 1:4 ratio of E/Z isomers (Scheme 1). Hydrolysis of methyl vinyl ether 4 with varying concentrations of HCl gave an epimeric mixture of aldehydes (5, Table 1). As discussed by Sulima et al., the epimeric mixture of aldehydes 5 was chromatographically unstable; the mixture was used without purification [16]. Hydrolysis of methyl vinyl ether 4 with varying concentrations of HCl gave an epimeric mixture of aldehydes (5, Table 1). As discussed by Sulima et al., the epimeric mixture of aldehydes 5 was chromatographically unstable; the mixture was used without purification [16].  The same synthetic route that was used in Schemes 2 and 3 to synthesize the diastere- The absolute configuration of 8 (1R,5S,9R) was confirmed by single-crystal X-ray diffraction analysis (Figure 1). Crystal data, atomic coordinates, etc, can be found in the Supplementary Materials. The absolute configuration of 8 (1R,5S,9R) was confirmed by single-crystal X-ray diffraction analysis ( Figure 1). Crystal data, atomic coordinates, etc, can be found in the Supplementary Materials. The same synthetic route that was used in Schemes 2 and 3 to synthesize the diastereomers 8 and 9, was also used to prepare the corresponding 1S,5R diastereomers 14 and 15. Using the known 1S,5R-ketone analogous to 3 [16], the 1S,5R-aldehyde 11 epimeric mixture was obtained in situ from 1S,5R-10) [12], leading to the phenolic methoxy analogs 12 and 13 which, on O-demethylation, gave the desired 1S,5R-alkenes, the vinyl com- As with the other alkenes, the C9-butylene compounds 23 and 24 were synthesize from the aldehyde intermediate 5, obtained in situ from 4 using propyltriphenylphosphin ium bromide with LiHMDS as the base (Scheme 7). This Wittig reaction required heatin at 45 °C for 15 h for consumption of starting material. The extended reaction times an heat resulted in more of the C9R epimer 23 to form from this reaction compared to th C9S.epimer 24. This ratio was observed by 1 H-NMR as these methoxy compounds wer not easily separable by column chromatography. The mixture of epimers underwent O demethylation using BBr3 at which point the epimers could be separated.
The propylene and the butylene series of compounds were all isolated as the Z-is mer, as indicated by the X-ray crystal structure of the propylene diastereomer 20 (Figu 2). The NMR pattern in the ca. d 5.7-5.3 region was similar for all of the Z-isomers, wi an observed ca. 10.6 coupling constant typical of Z isomers, in several of the diastereome that were not unresolved multiplets.

Ligand Efficacy and Potency (Forskolin-Induced cAMP Accumulation Assay)
The functional activity, as determined from the forskolin-induced cAMP accumul tion assay, of three sets of compounds, each containing four diastereomers, can be seen Table 2. The sets differed in the length of the alkenyl moiety at C9 and the compound within each of the sets differed only in their stereochemistry. The vinyl compounds in th The propylene and the butylene series of compounds were all isolated as the Z-isomer, as indicated by the X-ray crystal structure of the propylene diastereomer 20 ( Figure 2). The NMR pattern in the ca. d 5.7-5.3 region was similar for all of the Z-isomers, with an observed ca. 10.6 coupling constant typical of Z isomers, in several of the diastereomers that were not unresolved multiplets. The propylene and the butylene series of compounds were all isolated as the Z-isomer, as indicated by the X-ray crystal structure of the propylene diastereomer 20 ( Figure  2). The NMR pattern in the ca. d 5.7-5.3 region was similar for all of the Z-isomers, with an observed ca. 10.6 coupling constant typical of Z isomers, in several of the diastereomers that were not unresolved multiplets.

Ligand Efficacy and Potency (Forskolin-Induced cAMP Accumulation Assay)
The functional activity, as determined from the forskolin-induced cAMP accumulation assay, of three sets of compounds, each containing four diastereomers, can be seen in Table 2. The sets differed in the length of the alkenyl moiety at C9 and the compounds  The functional activity, as determined from the forskolin-induced cAMP accumulation assay, of three sets of compounds, each containing four diastereomers, can be seen in Table 2. The sets differed in the length of the alkenyl moiety at C9 and the compounds within each of the sets differed only in their stereochemistry. The vinyl compounds in the first set had disparate activities. Diastereomer 15 with 1S,5R,9R stereochemistry was a potent MOR antagonist, more than twice as potent as naltrexone IC 50 = 3.58 nM vs. naltrexone IC 50 = 10.78). It had some DOR (IC 50 = 143.7 nM) and KOR (IC 50 = 28.4 nM) antagonist activity and was devoid of agonist activity at MOR, DOR, and KOR. The other three diastereomers in that set were morphine-like agonists in potency at MOR, two of them, 1R,5S,9S-9 and its diastereomer 1S,5R,9S-14, were partial agonists with moderate or low efficacy (%E max = 67.3 and 46.6, respectively), and the third, 1R,5S,9R-8 was a fully efficacious agonist in the cAMP assay. The 1S,5R,9R stereochemistry of 15 pertained to 20 in the second set, and in the third set to 27. These 1S,5R,9R compounds acted as potent MOR antagonists, and variably potent DOR and KOR antagonists. None of them had MOR, DOR or KOR agonist activity in the cAMP assay.
The propenyl diastereomers in the second set were also quite different and, as observed in the first set, contained a potent antagonist 20 (IC 50 = 2.34 nM), and three MOR agonists. The agonist 18 with 1R,5S,9S stereochemistry had subnanomolar potency at MOR (EC 50 = 0.07 nM) in the cAMP assay. It was 89 times more potent than morphine, and it had some DOR partial agonist activity (EC 50 = 9.69 nM, %E max = 74.5) and KOR antagonist activity with low potency (IC 50 = 237.7 nM) ( Table 2). The other two compounds in this set were MOR partial agonists, 17 and 21, had morphine-like potency (EC 50 = 2.61 and 4.66 nM). We considered 17 as worthy of further examination since it appeared, based on our previous work [12], to have efficacy that might be in the range (%E max < 90 and >65 in the cAMP assay) that we hypothesized might be necessary for a morphine-like antinociceptive with reduced side effects.
The third set of diastereomers had a C9-butenyl substituent and it contained two antagonists, 26 and 27. Diastereomer 27 had the anticipated 1S,5R,9R stereochemistry; it was twice as potent as naltrexone at MOR, with modest DOR antagonist activity and subnanomolar potency as a KOR antagonist. The 1S,5R,9S diastereomer 26 was a weak MOR antagonist. The two remaining diastereomers in that set (23 and 24) had relatively weak MOR potency. The agonist potency, but not the antagonist potency, decreased with increased bulk at C9.
The importance of a phenolic hydroxyl can be seen in the inactivity of the methoxy analog 13. Apparently, that phenolic hydroxyl is essential for interaction with opioid receptors with the 5-phenylmorphans.
The cAMP functional assay clearly showed major differences in activity between the diastereomers within a set of four compounds, and differences between the diastereomers in each of the three sets. The vinyl (15), propenyl (20), and butenyl (27) diastereomers with the same 1S,5R,9R stereochemistry had the same activity, they were all MOR, DOR, KOR antagonists with varying potencies, and all three were more potent than naltrexone. Unlike naltrexone, none of them had KOR agonist activity. These antagonists might be better able to antagonize the in vivo effects of more potent narcotics such as fentanyl and etonitazene. Two or three partial agonists, the 1R,5S,9S-vinyl diastereomer (9) and the 1R,5S,9R-propenyl diastereomer (17) appeared to have sufficient efficacy to warrant further examination for their in vivo activity. The 1S,5R,9S-propenyl diastereomer (21) appeared, in theory, to have marginal efficacy for antinociceptive activity in vivo. second set, and in the third set to 27. These 1S,5R,9R compounds acted as potent MOR antagonists, and variably potent DOR and KOR antagonists. None of them had MOR, DOR or KOR agonist activity in the cAMP assay. second set, and in the third set to 27. These 1S,5R,9R compounds acted as potent MOR antagonists, and variably potent DOR and KOR antagonists. None of them had MOR, DOR or KOR agonist activity in the cAMP assay. second set, and in the third set to 27. These 1S,5R,9R compounds acted as potent MOR antagonists, and variably potent DOR and KOR antagonists. None of them had MOR, DOR or KOR agonist activity in the cAMP assay. cious agonist in the cAMP assay. The 1S,5R,9R stereochemistry of 15 pertained to 20 in the second set, and in the third set to 27. These 1S,5R,9R compounds acted as potent MOR antagonists, and variably potent DOR and KOR antagonists. None of them had MOR, DOR or KOR agonist activity in the cAMP assay. cious agonist in the cAMP assay. The 1S,5R,9R stereochemistry of 15 pertained to 20 in the second set, and in the third set to 27. These 1S,5R,9R compounds acted as potent MOR antagonists, and variably potent DOR and KOR antagonists. None of them had MOR, DOR or KOR agonist activity in the cAMP assay.   5.53 ± 1.02 (41.3 ± 6.8%) a Inhibition of forskolin-induced cAMP accumulation; cAMP Hunter TM Chinese hamster ovary cells (CHO-K1) that express human μ-opioid receptor (OPRM1), human κ-opioid receptor (OPRK1), and human δ-opioid receptor (OPRD1) were used for the forskolin-induced cAMP accumulation assay to determine potency and efficacy of the compounds following the previously established methods; [17] to determine % efficacy in forskolin-induced cAMP assays, data were blank subtracted with the vehicle control, followed by normalization to the forskolin control. Data were then analyzed in GraphPad Prism 8 (GraphPad, LaJolla, CA, USA) using nonlinear regression; values are expressed as the mean ± SEM of at least three independent experiments; N/D = not determined. b MOR Antagonist potency (IC50) determined versus EC90 of fentanyl; Degree of antagonism (Imax) normalized to naltrexone. c DOR Antagonist potency (IC50) determined versus EC50 of SNC80; degree of antagonism (Imax) normalized to naltrexone. d KOR Antagonist potency (IC50) determined versus EC90 of U50488H; degree of antagonism (Imax) normalized to nor-BNI.
The propenyl diastereomers in the second set were also quite different and, as observed in the first set, contained a potent antagonist 20 (IC50 = 2.34 nM), and three MOR agonists. The agonist 18 with 1R,5S,9S stereochemistry had subnanomolar potency at MOR (EC50 = 0.07 nM) in the cAMP assay. It was 89 times more potent than morphine, 5.53 ± 1.02 (41.3 ± 6.8%) a Inhibition of forskolin-induced cAMP accumulation; cAMP Hunter TM Chinese hamster ovary cells (CHO-K1) that express human μ-opioid receptor (OPRM1), human κ-opioid receptor (OPRK1), and human δ-opioid receptor (OPRD1) were used for the forskolin-induced cAMP accumulation assay to determine potency and efficacy of the compounds following the previously established methods; [17] to determine % efficacy in forskolin-induced cAMP assays, data were blank subtracted with the vehicle control, followed by normalization to the forskolin control. Data were then analyzed in GraphPad Prism 8 (GraphPad, LaJolla, CA, USA) using nonlinear regression; values are expressed as the mean ± SEM of at least three independent experiments; N/D = not determined. b MOR Antagonist potency (IC50) determined versus EC90 of fentanyl; Degree of antagonism (Imax) normalized to naltrexone. c DOR Antagonist potency (IC50) determined versus EC50 of SNC80; degree of antagonism (Imax) normalized to naltrexone. d KOR Antagonist potency (IC50) determined versus EC90 of U50488H; degree of antagonism (Imax) normalized to nor-BNI.
The propenyl diastereomers in the second set were also quite different and, as observed in the first set, contained a potent antagonist 20 (IC50 = 2.34 nM), and three MOR agonists. The agonist 18 with 1R,5S,9S stereochemistry had subnanomolar potency at MOR (EC50 = 0.07 nM) in the cAMP assay. It was 89 times more potent than morphine, a Inhibition of forskolin-induced cAMP accumulation; cAMP Hunter TM Chinese hamster ovary cells (CHO-K1) that express human µ-opioid receptor (OPRM1), human κ-opioid receptor (OPRK1), and human δ-opioid receptor (OPRD1) were used for the forskolin-induced cAMP accumulation assay to determine potency and efficacy of the compounds following the previously established methods; [17] to determine % efficacy in forskolin-induced cAMP assays, data were blank subtracted with the vehicle control, followed by normalization to the forskolin control. Data were then analyzed in GraphPad Prism 8 (GraphPad, LaJolla, CA, USA) using nonlinear regression; values are expressed as the mean ± SEM of at least three independent experiments; N/D = not determined. b MOR Antagonist potency (IC 50 ) determined versus EC90 of fentanyl; Degree of antagonism (I max ) normalized to naltrexone. c DOR Antagonist potency (IC 50 ) determined versus EC50 of SNC80; degree of antagonism (I max ) normalized to naltrexone. d KOR Antagonist potency (IC 50 ) determined versus EC90 of U50488H; degree of antagonism (Imax) normalized to nor-BNI.

Opioid Receptor Binding, Ligand Efficacy and Potency ([ 35 S]GTPgS Functional Assay)
The binding affinities at the MOR and functional activity in the [ 35 S]GTPgS assay (Table 3) were determined for two of the diastereomers in the first set of C9-vinyl compounds (8 and 9), and two from the second set of C9-propenyl diastereomers (17 and 18). These were chosen to compare the efficacies of full agonists (8 and 18) and partial agonists (9 and 17) in the cAMP assay to those in the GTP assay. able to antagonize the in vivo effects of more potent narcotics such as fentanyl and etonitazene. Two or three partial agonists, the 1R,5S,9S-vinyl diastereomer (9) and the 1R,5S,9Rpropenyl diastereomer (17) appeared to have sufficient efficacy to warrant further examination for their in vivo activity. The 1S,5R,9S-propenyl diastereomer (21) appeared, in theory, to have marginal efficacy for antinociceptive activity in vivo.

Opioid Receptor Binding, Ligand Efficacy and Potency ([ 35 S]GTPgS Functional Assay)
The binding affinities at the MOR and functional activity in the [ 35 S]GTPgS assay (Table 3) were determined for two of the diastereomers in the first set of C9-vinyl compounds (8 and 9), and two from the second set of C9-propenyl diastereomers (17 and 18). These were chosen to compare the efficacies of full agonists (8 and 18) and partial agonists (9 and 17) in the cAMP assay to those in the GTP assay. like naltrexone, none of them had KOR agonist activity. These antagonists might be better able to antagonize the in vivo effects of more potent narcotics such as fentanyl and etonitazene. Two or three partial agonists, the 1R,5S,9S-vinyl diastereomer (9) and the 1R,5S,9Rpropenyl diastereomer (17) appeared to have sufficient efficacy to warrant further examination for their in vivo activity. The 1S,5R,9S-propenyl diastereomer (21) appeared, in theory, to have marginal efficacy for antinociceptive activity in vivo.

Opioid Receptor Binding, Ligand Efficacy and Potency ([ 35 S]GTPgS Functional Assay)
The binding affinities at the MOR and functional activity in the [ 35 S]GTPgS assay (Table 3) were determined for two of the diastereomers in the first set of C9-vinyl compounds (8 and 9), and two from the second set of C9-propenyl diastereomers (17 and 18). These were chosen to compare the efficacies of full agonists (8 and 18) and partial agonists (9 and 17) in the cAMP assay to those in the GTP assay. like naltrexone, none of them had KOR agonist activity. These antagonists might be better able to antagonize the in vivo effects of more potent narcotics such as fentanyl and etonitazene. Two or three partial agonists, the 1R,5S,9S-vinyl diastereomer (9) and the 1R,5S,9Rpropenyl diastereomer (17) appeared to have sufficient efficacy to warrant further examination for their in vivo activity. The 1S,5R,9S-propenyl diastereomer (21) appeared, in theory, to have marginal efficacy for antinociceptive activity in vivo.

Opioid Receptor Binding, Ligand Efficacy and Potency ([ 35 S]GTPgS Functional Assay)
The binding affinities at the MOR and functional activity in the [ 35 S]GTPgS assay (Table 3) were determined for two of the diastereomers in the first set of C9-vinyl compounds (8 and 9), and two from the second set of C9-propenyl diastereomers (17 and 18). These were chosen to compare the efficacies of full agonists (8 and 18) and partial agonists (9 and 17) in the cAMP assay to those in the GTP assay. antagonists with varying potencies, and all three were more potent than naltrexone. Unlike naltrexone, none of them had KOR agonist activity. These antagonists might be better able to antagonize the in vivo effects of more potent narcotics such as fentanyl and etonitazene. Two or three partial agonists, the 1R,5S,9S-vinyl diastereomer (9) and the 1R,5S,9Rpropenyl diastereomer (17) appeared to have sufficient efficacy to warrant further examination for their in vivo activity. The 1S,5R,9S-propenyl diastereomer (21) appeared, in theory, to have marginal efficacy for antinociceptive activity in vivo.

Opioid Receptor Binding, Ligand Efficacy and Potency ([ 35 S]GTPgS Functional Assay)
The binding affinities at the MOR and functional activity in the [ 35 S]GTPgS assay (Table 3) were determined for two of the diastereomers in the first set of C9-vinyl compounds (8 and 9), and two from the second set of C9-propenyl diastereomers (17 and 18). These were chosen to compare the efficacies of full agonists (8 and 18) and partial agonists (9 and 17) in the cAMP assay to those in the GTP assay. The two vinyl compounds 8 and 9 had high MOR affinity in the binding assay, with 8 showing subnanomolar MOR affinity ( Table 3). The two propylene compounds 17 and 18 also had high MOR affinity, and 18 had subnanomolar MOR affinity. The diastereomeric compounds 8 and 18 were more potent than 9 and 17 in the [ 35 S]GTPgS functional assay, in accordance with the binding assay. The rank order of the potencies for the four compounds in Table 3 in the [ 35 S]GTPgS assay were in agreement with the data from the MOR binding assay. All four compounds were more potent than morphine. All of the compounds, with the exception of 18 which was morphine-like in efficacy, appeared to be partial agonists in the [ 35 S]GTPgS assay. The lower potency and efficacy of compounds in the [ 35 S]GTPgS assay, as compared with the cAMP assay, was expected due to the inherent differences between these assays [18]. The efficacies of 9 and 17 in both the cAMP ( Table 2) and [ 35 S]GTPgS (Table 3) assays were both in the range (efficacies < 20% and >10% in the [ 35 S]GTPgS assay) that we wanted to explore to determine if those partial agonists might have fewer opioid-like side effects, as postulated by theory [19].
The efficacies of the compounds in the two functional assays were somewhat different in that in the cAMP assay both 8 and 18 appeared to be morphine-like agonists. Both functional assays indicated that 9 and 17 were partial agonists. If the activity of these compounds were only related to their stereochemistry, 9 and 18 with 1R,5S,9S stereochemistry should show similarities in MOR affinity, potency and/or efficacy. They did not appear to have that relationship. Unlike that pair, the other two stereochemically similar compounds, 8 and 17, both with 1R,5S,9R stereochemistry, did appear to be somewhat similar in both the cAMP assay and in the GTP assay, although a difference was seen in their MOR binding affinity. The alkenes with 1S,5R,9R stereochemistry were not agonists. In all three sets of diasteromers that stereochemistry gave potent MOR antagonists. Obviously, within a set of C9-diastereomeric compounds, stereochemistry must be the dominant factor for their functional activity since all four compounds in a set had exactly the same 2-dimensional structure. Several compounds in the sets of diastereomers were examined by molecular modeling in an attempt to determine the differences between the C9-alkene MOR agonists and antagonists at the receptor level.

Molecular Modeling Results and Discussion
From experimental results, significant functional differences were observed between stereoisomers. To understand the structural basis of these differences and the mechanism by which they bind to the MOR, we carried out induced-fit docking in combination with MM/GBSA calculations of representative pairs of stereoisomers. We first focused on two vinyl pairs in the (1R,5S) series, the 9R-vinyl 8 and 9S-vinyl 9, and in the (1S,5R) series, 9R-vinyl 15 and 9S-vinyl 14 which are four diastereomers with the same C9-vinyl substituent in the 3-hydroxy-5-phenylmorphans.
Accommodation of the C9-Vinyl MOR Agonists 8, 9, and 14 in Both Active and Inactive Conformations of the MOR Common to both inactive and active MOR models, our docking results of pairs of vinyl analogs show that their protonated and positively charged nitrogen of the 5phenylmorphan forms an ionic interaction with D147 3.32 (superscripts denote Ballesteros-Weinstein numbering [20], the N-phenethyl is oriented towards the intracellular side of the binding pocket, and the phenol points towards the extracellular vestibule of the transmembrane domain. However, by comparing the inactive (4DKL) and active (6DDF) MOR crystal structures, we found that the side-chain orientation of N150 3.35 differs between the structures, which results in different shapes of the binding pocket in the active and inactive states. In the inactive structure 4DKL, N150 3.35 protrudes into and occludes part of the binding site that is otherwise unblocked in the active structure 6DDF. This structural difference of the binding site causes the docked ligands to adopt different binding modes in each receptor.
Specifically, the C9-vinyl agonists, 8, 9, and 14 can be easily accommodated in the binding pocket of the MOR in both active and inactive conformational states, as their docking poses do not reveal any clashes with binding site residues (Figures 3 and 4).
In the active MOR model, these ligands adopt a more linear configuration where the N-phenethyl tail extends deeper into the binding pocket and is enclosed by A117 2.53 , M151 3.36 , W293 6.48 , and Y326 7.43 , whereas the phenol moiety points towards Q124 2.60 (14), W318 7.35 (9), or Y148 3.33 (8) (Figure 3B). In the inactive model, the ligands are bent and shifted slightly, with the N-phenethyl tail in tight interactions only with W293 6.48 and Y326 7.43 , and the phenol moiety oriented more towards TM5 compared to their poses in the active state ( Figure 4B).
We then estimated the binding free energies of these docked ligands in the models by carrying out MM/GBSA calculations. In the active model, 8 has a lower binding free energy (−71.1 kcal/mol) than both 9 (−69.7 kcal/mol) and 14 (−65.3 kcal/mol) and is more favorably bound to the active receptor (Table 4). Between the active and inactive states of the MOR, the ∆∆G active-inactive of 8, 9, and 14 are −10.3, −7.2, and −3.6 kcal/mol, respectively, suggesting that these agonists may favor the active state ( Table 4).
difference of the binding site causes the docked ligands to adopt different binding modes in each receptor.
Specifically, the C9-vinyl agonists, 8, 9, and 14 can be easily accommodated in the binding pocket of the MOR in both active and inactive conformational states, as their docking poses do not reveal any clashes with binding site residues (Figures 3 and 4).

C9-Vinyl Antagonist (15) Preference for the Inactive Conformation of the MOR
(1S,5R,9R)-15, which differs from (1S,5R,9S)-14 only in the chirality of its C9-vinyl group, has been experimentally characterized as a MOR antagonist. In the same orientation as the agonists, the docked pose of 15 in the active model shows that its phenol moiety points towards Q124 2.60 . However, its vinyl and central hexane ring groups are not compatible with the original sidechain orientations of Y148 3.33 and I296 6.51 , respectively, forcing these binding site residues to rotate away ( Figure 3C). Additionally, the relatively high binding free energy of 15 (−47.3 kcal/mol) indicates that the antagonist does not bind favorably to the active state (Table 4). In contrast, this antagonist can be feasibly accommodated in the inactive MOR conformation, showing no clashes with binding site residues. Specifically, its vinyl group occupies a relatively open space near I144 3.29 , its phenyl ring has favorable aromatic interactions with W293 6.48 , while its phenol moiety engages in hydrogen bonds with both Q124 2.60 and Y128 2.64 . In addition, the hydrophobic interactions between the central hexane ring and I322 7.39 also stabilize the ligand in the inactive model ( Figure 3C). MM/GBSA analysis reveals that 15 strongly favors the inactive state with a ∆∆G active-inactive of 23.2 kcal/mol. Thus, 15 is not compatible with the active state binding site but can be well-accommodated by the inactive state of the MOR.  In the active MOR model, these ligands adopt a more linear configuration where the N-phenethyl tail extends deeper into the binding pocket and is enclosed by A117 2.53 , M151 3.36 , W293 6.48 , and Y326 7.43 , whereas the phenol moiety points towards Q124 2.60 (14), W318 7.35 (9), or Y148 3.33 (8) (Figure 3B). In the inactive model, the ligands are bent and shifted slightly, with the N-phenethyl tail in tight interactions only with W293 6.48 and Y326 7.43 , and the phenol moiety oriented more towards TM5 compared to their poses in the active state ( Figure 4B).
We then estimated the binding free energies of these docked ligands in the models by carrying out MM/GBSA calculations. In the active model, 8 has a lower binding free energy (−71.1 kcal/mol) than both 9 (−69.7 kcal/mol) and 14 (−65.3 kcal/mol) and is more favorably bound to the active receptor (Table 4). Between the active and inactive states of the MOR, the ΔΔGactive-inactive of 8, 9, and 14 are −10.3, −7.2, and −3.6 kcal/mol, respectively, suggesting that these agonists may favor the active state (Table 4).  of compounds 15, 14, 9, and 8, respectively. Each compound of this series was treated as the Z-isomer in accordance with the X-ray spectroscopic analysis of 20 (Figure 4). We hypothesized that the extra methyl group worsens the clash of a MOR antagonist with Y148 3.33 . Indeed, the docked pose of 20 shows that the elongated substituent is more proximal to Y148 3.33 ( Figure 3E). MM/GBSA results show that 20 has a high binding free energy of −47.6 kcal/mol ( Table 4). The (9S)-isoform of 20, compound 21, which is a MOR agonist, has a similar orientation to 20; however, its propenyl tail is oriented away from Y148 3.33 where adequate space can accommodate the extra methyl ( Figure 3D). MM/GBSA results agree with these binding site differences, which show that 21 binds more favorably in the active model than 20 (Table 4). Compound 18, a propenyl analog of 9 in the vinyl series, acts as a potent agonist. Interestingly, the additional methyl of 18 overlaps with the space occupied by the vinyl group of 9, the most favorably bound agonist of the vinyl series. As such, it may be important for the alkyl substituents of these agonists to protrude into this hydrophobic space, where they can interact with I322 7.39 ( Figure 3D). MM/GBSA results indicate that 18 is favored in the active model, with the lowest binding free energy of the series (−72.3 kcal/mol). Similarly, the propenyl agonist 17, which had its C9-substituent modified from the vinyl agonist 8, has its elongated propenyl substituent overlap with that of 18 and forms a strong hydrophobic interaction with I322 7.39 as well.
Similar to the vinyl series, all of the propenyl analogs can be accommodated by the inactive model. However, the phenol moiety of the MOR antagonist 20 is able to form hydrogen bonds with Q124 2.60 and Y128 2.64 . Additionally, the central hexane group remains located between TM6 and TM7 and engaged in hydrophobic interactions with I322 7.39 ( Figure 4E). MM/GBSA calculations show that 20 has a strong preference for the inactive state with a ∆∆G active-inactive of 22.9 kcal/mol. In addition, our MM/GBSA results also indicate that 20 binds more favorably than the agonists in the inactive states. produced antinociceptive effects in some, but not all subjects [F (5,19) = 2.1, n.s.], consistent with its quite low cAMP and [ 35 S]GTPgS efficacy, and did not produce morphine-like decreases in ventilatory ratio [F (4,14) = 0.8, n.s.]. Thus, both compounds 9 and 17 had less effect on respiratory depression than those observed with morphine, which is consistent with their designation as partial agonists with sufficient intrinsic efficacy for G-protein activation. Their intrinsic efficacy in our theoretically desirable range might be one of the causes [18] for their reduced effect on respiration.

In Vivo Data
quite low cAMP and [ 35 S]GTPgS efficacy, and did not produce morphine-like decreases in ventilatory ratio [F(4,14) = 0.8, n.s.]. Thus, both compounds 9 and 17 had less effect on respiratory depression than those observed with morphine, which is consistent with their designation as partial agonists with sufficient intrinsic efficacy for G-protein activation. Their intrinsic efficacy in our theoretically desirable range might be one of the causes [18] for their reduced effect on respiration.

General Information
Melting points were determined on a Mettler Toledo MP70 and are uncorrected. Proton and carbon nuclear magnetic resonance ( 1 H and 13 C NMR) spectra were recorded on a Varian Gemini-400 spectrometer in CDCl 3 (unless otherwise noted) with the values given in ppm (TMS as internal standard) and J (Hz) assignments of 1 H resonance coupling. The analyses were performed on the free base, unless otherwise noted. Mass spectra (HRMS) were recorded on a VG 7070E spectrometer or a JEOL SX102a mass spectrometer. The optical rotation data were obtained on a PerkinElmer polarimeter model 341. Thin layer chromatography (TLC) analyses were carried out on Analtech silica gel GHLF 0.25 mm plates using various gradients of CHCl 3 /MeOH containing 1% NH 4 OH or gradients of EtOAc/n-hexane. Visualization was accomplished under UV light or by staining in an iodine chamber. Flash column chromatography was performed with Fluka silica gel 60 (mesh 220 − 400). Flash column chromatography was performed using RediSep Rf normal phase silica gel cartridges. Robertson Microlit Laboratories, Ledgewood, NJ, USA, performed elemental analyses, and the results were within ±0.4% of the theoretical values.

Syntheses
(1S,5S)-5-(3-Methoxyphenyl)-2-azabicyclo [3.3.1]nonan-9-one (2): In an oven-dried flask, 1 (2.2 g, 8.3 mmol) and K 2 CO 3 (2.3 g, 16.6 mmol) in 15 mL acetonitrile were treated with cyanogen bromide (3.3 mL, 16.6 mmol). The reaction mixture was stirred under a N 2 atmosphere at room temperature for 2 h then heated to reflux. After 2 h, the reaction mixture was extracted with CHCl 3, and the organic phase was washed with brine and concentrated. The residue was taken up in a mixture of 21 mL 3 N HCl and 2.2 mL methanol and stirred at reflux for 17 h. Upon completion, the reaction mixture was cooled and quenched with 7 N NH 4 OH in methanol. The mixture was extracted with CHCl 3 and washed with water, brine, dried with Na 2 SO 4 and concentrated. Purification by flash column chromatography on silica gel (0-20% CMA in CHCl 3 ) gave the red oil 2 (1.62 g, 80% yield). The 1 H NMR of the product was identical to that of the known compound 2 [16].
The aldehyde was treated with ethyltriphenylphosphonium iodide (3 g, 3 equiv, 8 mmol) and suspended in THF (10 mL). The reaction mixture was cooled to 0 • C and treated slowly with LiHMDS (1.0 M in THF) (7 mL, 2.6 equiv, 7 mmol). After 30 min, the reaction mixture was warmed to room temperature then the mixture was heated to 45 • C for 16 h. Upon completion by TLC, the reaction mixture was quenched with MeOH and extracted with CHCl 3 . The mixture was washed with water and brine, dried with sodium sulfate and concentrated. Purification by flash column chromatography on silica gel (0-50% EtOAc in hexanes) yielded a mixture of C9 epimers 16 which was used without further purification.
In an oven-dried round-bottom flask, 16 (500 mg, 1 equiv, 1.3 mmol) was suspended in dichloromethane (15 mL) and the mixture was cooled to -78 • C. Tribromoborane (667 mg, 253 µL, 2 equiv, 2.66 mmol) was added to drop-wise and the reaction was stirred at -78 • C for 15 min. The reaction mixture was allowed to warm to room temperature and stirred 2 h. Upon completion, the reaction mixture was cooled to 0 • C and quenched with 7 mL MeOH drop wise and stirred for 30 min. subsequently, 10 mL 1 N HCl was added, and the reaction mixture was distilled at 100 • C for 1 h. The reaction mixture was then cooled to 0 • C and made basic (>10.5) with NH 4 OH and extracted with 9:1 CHCl 3 : MeOH. The combined organic layers were washed with water and brine, dried with sodium sulfate and concentrated. Purification by silica gel column chromatography 0-60% EtOAc: Hexanes. 17 was isolated as a white foam (153 mg, 32% yield) as the more polar fraction. The HCl salt of 17 was formed in iPrOH with 37% HCl (0.1 mL) and recrystallized from ethanol to give a white solid: mp 271-275   13 C under nitrogen. The reaction mixture was quenched with 7 N NH 4 OH in MeOH, extracted with CHCl 3 and washed with water and brine. The organic layer was then dried with sodium sulfate, concentrated, to yield an aldehyde intermediate. The aldehyde intermediate was treated with propyltriphenylphosphonium bromide (3.7 g, 3 equiv, 9.5 mmol) and suspended in THF (12 mL). The reaction mixture was cooled to 0 • C and treated slowly with LiHMDS (1.0 M in THF) (8.3 mL, 2.6 equiv, 8.3 mmol). After 30 min, the reaction mixture was warmed to room temperature then the mixture was heated to 45 • C for 16 h. Upon completion by TLC, the reaction mixture was quenched with MeOH and extracted with CHCl 3 . The mixture was washed with water and brine, dried with sodium sulfate and concentrated. Purification by flash column chromatography on silica gel (0-50% EtOAc in hexanes) yielded a mixture of C9 epimers 22 which were used in the next reaction without further purification or characterization.
In an oven-dried round-bottom flask, 22 (800 mg, 1 equiv, 2.05 mmol) was suspended in dichloromethane (20 mL) and the mixture was cooled to -78 • C. Tribromoborane (1.03 g, 390 µL, 2 equiv, 4.11 mmol) was added to drop-wise and the reaction was stirred at -78 • C for 15 min. The reaction mixture was allowed to warm to room temperature and stirred 2 h. Upon completion, the reaction mixture was cooled to 0 • C and quenched with 10 mL MeOH drop wise and stirred for 30 min. subsequently, 15 mL 1 N HCl was added, and the reaction mixture was distilled at 100 • C for 1 h. The reaction mixture was then cooled to 0 • C and made basic (>10.5) with NH 4 OH and extracted with 9:1 CHCl 3 : MeOH. The combined organic layers were washed with water and brine, dried with sodium sulfate and concentrated. Purification by silica gel column chromatography 0-60% EtOAc: Hexanes. 23 was isolated as a white foam (305 mg, 40% yield) as the more polar fraction. The HCl salt of 23 was formed in iPrOH (1.5 mL) with 37% HCl (0.15 mL) and recrystallized from hot ethanol (5 mL 13     In an oven-dried flask, 10 (2 g, 5.3 mmol) was suspended in THF (20 mL) and treated with HCl (6 M, 20 mL) and the reaction mixture was stirred overnight at room temperature under nitrogen. The reaction mixture was quenched with 7 N NH 4 OH in MeOH, extracted with CHCl 3 and washed with water and brine. The organic layer was then dried with sodium sulfate, concentrated, to yield the aldehyde intermediate. The aldehyde was treated with propyltriphenylphosphonium bromide (6.1 g, 3 equiv, 16 mmol) and suspended in THF (20 mL). The reaction mixture was cooled to 0 • C and treated slowly with LiHMDS (1.0 M in The centroid of the docking box was determined by the position of the bound ligand in each structure. The standard precision protocol of IFD was used, and up to 20 poses were generated for each ligand. An additional hydrogen-bond constraint was added to ensure that the docking poses maintained a conserved ionic interaction with D147 3.32 . For each ligand, the resulting docking poses were clustered using K-means clustering, and the representative pose was selected from among the best scored poses of the largest cluster. For the C9-propenyl compounds 17, 18, 20, and 21 compounds, an additional methyl moiety was grafted onto the vinyl group to generate a propenyl group in the Z-isoform of  Tail withdrawal latencies were assessed as described previously [24]. Briefly, monkeys were seated in customized Plexiglas chairs that allowed their tails to hang freely. Tail withdrawal latencies were measured by immersing the subject's tail in water held at 35 or 52 • C (temperatures were presented in a randomized order during successive test components). After obtaining a baseline tail withdrawal latency, complete dose response curves were generated in each subject using standard cumulative dosing procedures. Briefly, every 15 min after an injection tail-withdrawal latencies at each temperature were redetermined and subjects were injected with the next dose, such that the total (cumulative) dose was increased by 1 2 log 10 units in each successive cycle. This procedure was repeated until either (a) the tail-withdrawal latency from 52 • C water reached the maximum allowable latency (10 sec), or (b) tail-withdrawal latency no longer increased with increases in dose of the test drug.

Respiratory Depression: Ventilatory Response to Hypercapnia (5% CO 2 in Air)
Ventilation measures were assessed as described previously [25]. Briefly, squirrel monkeys were acclimated to a customized acrylic chamber (10" d × 10" w × 10" h) that served as a whole-body plethysmograph (EMKA Technologies, Montreal, PQ, Canada). Gas (either air or a 5% CO 2 in air mixture) was introduced to and extracted from the chamber at a constant flow rate of 5 L/min. Experimental sessions consisted of 4-6 consecutive 30 min cycles, each comprising a 20 min exposure to air followed by a 10 min exposure to 5% CO 2 . Drug effects were determined using cumulative dosing procedures, and injections were administered following each exposure to 5% CO 2 . Respiratory rate and tidal volume (mL/breath) were recorded over 1 min periods and were multiplied to provide minute volumes. Data from the last three minutes of each exposure to air or CO 2 were averaged and used for analysis of drug effects on ventilation.

Data Analysis
All statistical analyses and graphic representations were completed with GraphPad Prism version 9.3.0 (GraphPad Software, San Diego, CA, USA) using log transformed values of doses. Group means ± SEM tail withdrawal latencies (in sec) and minute volume ratios are plotted as a function of drug dose. Data were analyzed using One-way ANOVA with significance set at p < 0.05, followed by Dunnett's multiple comparison test. Animals that did not receive all doses of a drug in tail withdrawal studies because they attained a maximum effect at less than the highest dose were assigned 10 s latencies for all doses higher than the last dose tested.

In Vitro Assays
In vitro binding assays were performed using monocloned mouse mu opioid receptor expressing Chinese hamster ovary (CHO) cells (mMOR-CHO). mMOR-CHO cell culture and membrane homogenate preparation were performed as previously described [26]. All assays were duplicated and repeated at least three times.