Functional Activity of Enantiomeric Oximes and Diastereomeric Amines and Cyano Substituents at C9 in 3-Hydroxy-N-phenethyl-5-phenylmorphans

The synthesis of stereochemically pure oximes, amines, saturated and unsaturated cyanomethyl compounds, and methylaminomethyl compounds at the C9 position in 3-hydroxy-N-phenethyl-5-phenylmorphans provided μ-opioid receptor (MOR) agonists with varied efficacy and potency. One of the most interesting compounds, (2-((1S,5R,9R)-5-(3-hydroxyphenyl)-2-phenethyl-2-azabicyclo[3.3.1]nonan-9-yl)acetonitrile), was found to be a potent partial MOR agonist (EC50 = 2.5 nM, %Emax = 89.6%), as determined in the forskolin-induced cAMP accumulation assay. Others ranged in potency and efficacy at the MOR, from nanomolar potency with a C9 cyanomethyl compound (EC50 = 0.85 nM) to its totally inactive diastereomer, and three compounds exhibited weak MOR antagonist activity (the primary amine 3, the secondary amine 8, and the cyanomethyl compound 41). Many of the compounds were fully efficacious; their efficacy and potency were affected by both the stereochemistry of the molecule and the specific C9 substituent. Most of the MOR agonists were selective in their receptor interactions, and only a few had δ-opioid receptor (DOR) or κ-opioid receptor (KOR) agonist activity. Only one compound, a C9-methylaminomethyl-substituted phenylmorphan, was moderately potent and fully efficacious as a KOR agonist (KOR EC50 = 18 nM (% Emax = 103%)).


Introduction
Over the past several decades, opioid abuse in the United States has steadily increased, resulting in a large number of dependent individuals and opioid-related deaths [1,2].These circumstances ultimately led to the declaration of an opioid crisis as a "public health emergency" by the United States Federal Government in 2017.We are pursuing several approaches to help combat this public health problem, one of which focuses on developing safer opioid alternatives that could also be used as medications for Opioid Use Disorder (OUD).To achieve this goal, we are synthesizing low-efficacy compounds based on the 5-phenylmorphan structure.Depending on the stereochemistry of the molecule and the specific substituent at C9 in the N-phenethyl-5-phenylmorphans (Figure 1), we found full or partial agonists with varied potency [3][4][5][6][7] at the µ-opioid receptor (MOR).Different ligands at the C9 position of the 5-phenylmorphan or compounds with a specific stereochemistry can interact with an active conformation of the receptor as a MOR agonist and turn on involved intracellular pathways [8] or act at the inactive conformation of the MOR as an opioid antagonist [5].One signaling molecule recruited by the MOR in its active conformation is β-arrestin2, which has been proposed to play a role in manifesting the negative side effects associated with opioids [9].However, the importance of β-arrestin2 has been challenged, and an attempt to reproduce studies that indicate biased ligands do not induce respiratory depression was unsuccessful [10].Gillis et al. have proposed a different theory based on intrinsic efficacy.This concept indicates that agonists with higher intrinsic efficacy will produce increased downstream signaling and show well-known opioid-like side effects.As a result, an ideal opioid would need to provide a balance between inducing sufficient signaling, which may be correlated to its efficacy in the cAMP assay, to enable significant pain reduction in vivo but not so much as to cause negative side effects [11].We attempt to measure that balance using in vitro assays (the forskolin-induced cAMP accumulation assay and the [ 35 S]GTPgS assay) [5] and in vivo assays [12,13].However, advocates for biased agonism have argued that the understanding of interactions at the MOR is still incomplete, and disregarding β-arrestin2 completely may not be correct.
have argued that the understanding of interactions at the MOR is still incomplete, and disregarding β-arrestin2 completely may not be correct.
The 5-phenylmorphan structure is the template for our new compounds.It contains a minimal structure of morphine that can induce antinociceptive activity [14].The 5phenylmorphans are sufficiently different from morphine and the opioid-like family of analgesics (e.g., oxymorphone, oxycodone, hydromorphone) that some of these 5phenylmorphans, those with certain C9 substituents bearing specific stereochemistry, have been found to have a different pharmacological profile, enabling antinociception with less respiratory depression and minimal gastrointestinal effects.The side effects appear to be related to the efficacy of the compounds; those with minimal efficacy do not induce respiratory depression or show GI effects, but they retain some antinociceptive activity, depending on the in vitro or in vivo assay.Previous work by our group has shown that some C9-substituted derivatives containing an N-phenethyl moiety exhibit little or no β-arrestin2 recruitment [4]; others act as potent partial MOR agonists.Because of the promise this family of molecules holds as a new type of analgesic, we have designed 5-(3-hydroxphenyl)-N-phenethyl-phenylmorphan derivatives containing varied substituents at the C9 position and have found suitable synthetic paths to obtain sufficient material for pharmacological evaluation (Figure 1).The C9 position was chosen for our initial exploration because our early work indicated that potent MOR agonists could be obtained with chiral substituents at that position [3], and there appeared to be a minimal amount of previous synthetic work based on that area of the molecule.
In this work, we discuss the synthesis and in vitro functional activity of chiral 3hydroxy-N-phenethyl-5-phenylmorphans with various nitrogen-containing substituents at C9.These substituents were oximes, amines, unsaturated and saturated cyanomethyl compounds, and methylaminomethyl substituted compounds (Figure 1).The 5-phenylmorphan structure is the template for our new compounds.It contains a minimal structure of morphine that can induce antinociceptive activity [14].The 5-phenylmorphans are sufficiently different from morphine and the opioid-like family of analgesics (e.g., oxymorphone, oxycodone, hydromorphone) that some of these 5phenylmorphans, those with certain C9 substituents bearing specific stereochemistry, have been found to have a different pharmacological profile, enabling antinociception with less respiratory depression and minimal gastrointestinal effects.The side effects appear to be related to the efficacy of the compounds; those with minimal efficacy do not induce respiratory depression or show GI effects, but they retain some antinociceptive activity, depending on the in vitro or in vivo assay.Previous work by our group has shown that some C9-substituted derivatives containing an N-phenethyl moiety exhibit little or no β-arrestin2 recruitment [4]; others act as potent partial MOR agonists.Because of the promise this family of molecules holds as a new type of analgesic, we have designed 5-(3-hydroxphenyl)-N-phenethyl-phenylmorphan derivatives containing varied substituents at the C9 position and have found suitable synthetic paths to obtain sufficient material for pharmacological evaluation (Figure 1).The C9 position was chosen for our initial exploration because our early work indicated that potent MOR agonists could be obtained with chiral substituents at that position [3], and there appeared to be a minimal amount of previous synthetic work based on that area of the molecule.
In this work, we discuss the synthesis and in vitro functional activity of chiral 3hydroxy-N-phenethyl-5-phenylmorphans with various nitrogen-containing substituents at C9.These substituents were oximes, amines, unsaturated and saturated cyanomethyl compounds, and methylaminomethyl substituted compounds (Figure 1).

Synthesis of 1S,5S,9R-and 1R,5R,9S-C9 Amino and Methylamino Enantiomers
The enantiomeric (1S,5S-and 1R,5R-) 9-keto-m-methoxyphenylmorphans were obtained using the literature procedure [7,15].Reductive amination was found to allow the synthesis of C9 amino derivatives with zero carbons at C9, while the HWE reaction facilitates the formation of unsaturated and saturated cyano derivatives.A Wittig reaction to the aldehyde, followed by reductive amination and O-demethylation, gave N-methyl compounds with one carbon at C9.We began with an exploration of the use of a titaniummediated reductive amination on the enantiomeric 9-keto compounds 1S,5S-1 and 1R,5R-4 (Scheme 2).
This approach was successful with both ammonia (Scheme 2) and methyl amine (Scheme 3) as the nucleophiles.This titanium-mediated reaction gave a product with the C9 amino group cis to the tertiary amine.
While the reaction with ammonia proceeded at room temperature, methylamine required heat.In addition, we discovered that a pressure tube is needed for the latter transformation, as the methylamine is lost to the atmosphere faster than it can react with 1 if the reaction is not performed in a sealed container.
Despite the success of these reactions, they only gave one of the diastereomers for each of the 9-keto enantiomers.This was likely due to the titanium (IV) coordinating with the tertiary nitrogen in 1, effectively blocking one side of the molecule from the hydride approach during the reduction step.
Based on the success of the transformation in Schemes 2 and 3, we considered accessing the two remaining diastereomers of the C9 amino compound via the reduction of the corresponding hydrazine (Scheme 4).
We found that reacting 1 with hydrazine followed by a reduction with NaBH 4 gave a single isomer of 11, as suggested by 1 H NMR (Scheme 4).However, attempts at cleaving the nitrogen-nitrogen bond under basic conditions were unsuccessful.
We also considered the possibility of accessing other diastereomers via an azide.We began this process by performing a diastereoselective reduction of 1R,5R-4 with superhydride to form the alcohol 12, followed by mesylating the alcohol 12 to give 13 (Scheme 5).
Attempts at a substitution with sodium azide at 60 • C were unsuccessful (giving only unreacted starting material), while heating the transformation up to 100 • C was also unsuccessful, and the desired product was not obtained from this reaction.Next, we subjected the chiral alcohol 1R,5R,9S-12 to the Mitsunobu reaction in the hope of obtaining the desired 1R,5R,9R-isomer via complete inversion of stereochemistry at the C9 center.This approach was not successful either.
To access the two remaining C9 amino diastereomers, several other reductive amination conditions were attempted.Unfortunately, various Brønsted acids, reductants, and ammonia sources gave no evidence for product formation.Possibly the steric hindrance around the C9 position in 1 required a strong Lewis acid catalyst to facilitate a nucleophilic attack.We next investigated the use of benzylamine as the nucleophile, hoping that this might force the opposite configuration at C9 due to its larger size.Unfortunately, that did not give us the desired chiral product.While the reaction with ammonia proceeded at room temperature, methylamine required heat.In addition, we discovered that a pressure tube is needed for the latter transformation, as the methylamine is lost to the atmosphere faster than it can react with 1 if the reaction is not performed in a sealed container.
Despite the success of these reactions, they only gave one of the diastereomers for each of the 9-keto enantiomers.This was likely due to the titanium (IV) coordinating with the tertiary nitrogen in 1, effectively blocking one side of the molecule from the hydride approach during the reduction step.
Based on the success of the transformation in Schemes 2 and 3, we considered accessing the two remaining diastereomers of the C9 amino compound via the reduction of the corresponding hydrazine (Scheme 4).We found that reacting 1 with hydrazine followed by a reduction with NaBH4 gave a single isomer of 11, as suggested by 1 H NMR (Scheme 4).However, attempts at cleaving the nitrogen-nitrogen bond under basic conditions were unsuccessful.
We also considered the possibility of accessing other diastereomers via an azide.We began this process by performing a diastereoselective reduction of 1R,5R-4 with superhydride to form the alcohol 12, followed by mesylating the alcohol 12 to give 13 (Scheme 5). a single isomer of 11, as suggested by 1 H NMR (Scheme 4).However, attempts at cleaving the nitrogen-nitrogen bond under basic conditions were unsuccessful.
We also considered the possibility of accessing other diastereomers via an azide.We began this process by performing a diastereoselective reduction of 1R,5R-4 with superhydride to form the alcohol 12, followed by mesylating the alcohol 12 to give 13 (Scheme 5).Attempts at a substitution with sodium azide at 60 °C were unsuccessful (giving only unreacted starting material), while heating the transformation up to 100 °C was also unsuccessful, and the desired product was not obtained from this reaction.Next, we subjected the chiral alcohol 1R,5R,9S-12 to the Mitsunobu reaction in the hope of obtaining the desired 1R,5R,9R-isomer via complete inversion of stereochemistry at the C9 center.This approach was not successful either.
To access the two remaining C9 amino diastereomers, several other reductive amination conditions were attempted.Unfortunately, various Brønsted acids, reductants, and ammonia sources gave no evidence for product formation.Possibly the steric hindrance around the C9 position in 1 required a strong Lewis acid catalyst to facilitate a nucleophilic attack.We next investigated the use of benzylamine as the nucleophile, hoping that this might force the opposite configuration at C9 due to its larger size.Unfortunately, that did not give us the desired chiral product.

Synthesis of Oxime Precursors to the Remaining Set of C9 Amino Enantiomers
We then considered accessing the two remaining C9 amino compounds through their oximes.This approach was successful, and we obtained the desired oximes 15 and 17 in sufficient quantity to enable an examination of their functional activity (Scheme 6).Scheme 5. Attempted azide substitution.

Synthesis of Oxime Precursors to the Remaining Set of C9 Amino Enantiomers
We then considered accessing the two remaining C9 amino compounds through their oximes.This approach was successful, and we obtained the desired oximes 15 and 17 in sufficient quantity to enable an examination of their functional activity (Scheme 6).It was not necessary to convert the C9 keto methoxy compounds 1 and 4 to their phenolic relatives.Oximes could be prepared directly from the phenolic ether to obtain the methoxy-substituted oximes 18 and 19 (Scheme 7).The E/Z stereochemistry of the oximes 15, 17, 18, and 19 was not established.Scheme 6. Synthesis of phenolic enantiomeric oximes 1S,5S-15, and 1R,5R-17 for functional assay.
It was not necessary to convert the C9 keto methoxy compounds 1 and 4 to their phenolic relatives.Oximes could be prepared directly from the phenolic ether to obtain the methoxy-substituted oximes 18 and 19 (Scheme 7).The E/Z stereochemistry of the oximes 15, 17, 18, and 19 was not established.Scheme 6. Synthesis of phenolic enantiomeric oximes 1S,5S-15, and 1R,5R-17 for functional assay.
It was not necessary to convert the C9 keto methoxy compounds 1 and 4 to their phenolic relatives.Oximes could be prepared directly from the phenolic ether to obtain the methoxy-substituted oximes 18 and 19 (Scheme 7).The E/Z stereochemistry of the oximes 15, 17, 18, and 19 was not established.

Synthesis of the 1S,5S,9S-and 1R,5R,9R-C9 Amino Enantiomers
The reduction of the methoxy ether oxime 19 was tested under three conditions: using Red-Al, hydrogenation via an H-cube flow reactor, and using LAH.While all three conditions were found to give a mixture of diastereomers, the yields and ratios between the products varied (Table 1).

Synthesis of the 1S,5S,9S-and 1R,5R,9R-C9 Amino Enantiomers
The reduction of the methoxy ether oxime 19 was tested under three conditions: using Red-Al, hydrogenation via an H-cube flow reactor, and using LAH.While all three conditions were found to give a mixture of diastereomers, the yields and ratios between the products varied (Table 1).Despite the low yield that was observed in this initial LAH reduction, likely due to a loss of product from coordination with aluminum salt byproducts, we were able to access the desired phenolic amino compounds 22 and 23 after O-demethylation (Scheme 8).

Reductant Diastereomeric Ratio (20:5) Yield
Red-Al 1:2 15% (pure 20) Despite the low yield that was observed in this initial LAH reduction, likely due to a loss of product from coordination with aluminum salt byproducts, we were able to access the desired phenolic amino compounds 22 and 23 after O-demethylation (Scheme 8).

LAH
1.5:1 37% (pure 20) Despite the low yield that was observed in this initial LAH reduction, likely due to a loss of product from coordination with aluminum salt byproducts, we were able to access the desired phenolic amino compounds 22 and 23 after O-demethylation (Scheme 8).Scheme 8. Synthesis of enantiomeric phenolic amines 22 and 23-the diastereomers of 3 and 6.

Chirality Determined via X-ray Crystallography
The stereochemistry of these C9-substituted phenylmorphans was confirmed by Xray diffraction analysis (Figure 2).The stereochemistry of these C9-substituted phenylmorphans was confirmed by X-ray diffraction analysis (Figure 2).

Chirality Determined via NMR Analysis
The 1 H-NMR data were indicative of the stereochemistry shown in Table 1 and in the schemes.The rationale for this comes from the chemical shift of the proton at the C9 position in the formerly reported [3] diastereomeric compounds 1R,5R,9R-and 1R,5R,9S-9-hydroxy-5-(3-hydroxyphenyl-2-phenylethyl-2-azabicyclo[3.3.1]nonane, with stereochemistry determined by X-ray crystallography.The C9 hydroxyl group in the 1R,5R,9R compound was trans to the tertiary amine, and the C9 proton was found at 4.40 ppm.In comparison, in the diastereomeric 1R,5R,9S compound, where the C9 hydroxyl group was cis to the tertiary amine, the C9 proton was found upfield at 4.06 ppm.A similar effect was seen with the C9 trans and cis amino-containing compounds.When the C9 NH2 group was trans to the tertiary amine, the C9 proton was found at 3.58 ppm (in compounds 22 and 23), and when cis, it was found at 3.50 ppm (in compounds 3 and 6).The trans and cis C9 cyanomethyl compounds followed the same pattern.When the C9 cy-

Chirality Determined via NMR Analysis
The 1 H-NMR data were indicative of the stereochemistry shown in Table 1 and in the schemes.The rationale for this comes from the chemical shift of the proton at the C9 position in the formerly reported [3] diastereomeric compounds 1R,5R,9R-and 1R,5R,9S-9-hydroxy-5-(3-hydroxyphenyl-2-phenylethyl-2-azabicyclo[3.3.1]nonane, with stereochemistry determined by X-ray crystallography.The C9 hydroxyl group in the 1R,5R,9R compound was trans to the tertiary amine, and the C9 proton was found at 4.40 ppm.In comparison, in the diastereomeric 1R,5R,9S compound, where the C9 hydroxyl group was cis to the tertiary amine, the C9 proton was found upfield at 4.06 ppm.A similar effect was seen with the C9 trans and cis amino-containing compounds.When the C9 NH 2 group was trans to the tertiary amine, the C9 proton was found at 3.58 ppm (in compounds 22 and 23), and when cis, it was found at 3.50 ppm (in compounds 3 and 6).The trans and cis C9 cyanomethyl compounds followed the same pattern.When the C9 cyanomethyl was trans to the tertiary amine, the C9 proton was found at 3.43 ppm (compounds 34 and 35), and when the C9 cyanomethyl group was cis to the tertiary amine, the C9 proton was upfield at 3.31 ppm (compounds 38 and 41).
The stereochemistry of the starting materials, the 5-(3-methoxyphenyl)-9-oxo-2phenylethyl-2-azabicyclo[3.3.1]nonane(9-oxo phenylmorphans), has been established [15].There is, then, only the stereochemistry of the C9 substituent to be determined.That became known via the X-ray spectroscopic data on compound 23 (Scheme 8).Compound 22 was its enantiomer-thus establishing the stereochemistry at C9 for compound 22.The NMR chemical shifts of these known compounds were followed by all of the others.That is, the C9 substituent was trans to the nitrogen atom in the piperidine ring for the amino compounds in Scheme 8, and this gave us the chemical shift of the C9 proton in the trans series.The titanium reactions shown in Schemes 2 and 3 gave products that were cis for the C9 substituent and the nitrogen atom in the piperidine ring.This was established because the amino compounds 3 and 6 were different from 22 and 23, and given the known stereochemistry of the starting material, they had to have the cis configuration.Similarly, for the cyanomethyl compounds, the trans configuration of 34 and 35 and the cis configuration of 38 and 41 were assigned tentatively based on the chemical shift of the C9 proton in the NMR.With these compounds, a clear and consistent correlation was found with the chemical shift of the C9 proton in the NMR.The exceptions to this were the methylaminomethyl compounds.The stereochemistry of these compounds was determined via their synthesis from the aldehyde in a series of reactions that led to known stereochemistry (proven by another X-ray [5]).
To obtain the saturated nitriles, we needed to reduce the alkene in 28 and 33 while leaving the nitrile intact.Employing traditional conditions with hydrogen and palladium on carbon as the catalyst resulted in a complex mixture of both alkene and partial nitrile reduction.However, using an H-cube hydrogenation flow reactor, we were able to develop conditions that gave the desired enantiomeric saturated nitriles 1S,5R,9S-34 and 1R,5S,9R-35 in low-to-moderate yield (Scheme 10).
To access the enantiomers 38 and 41, we used the Boc-protected amines 26 and 32 (Scheme 11).
The reduction of 26 using an H-cube hydrogenation flow reactor gave rise to the saturated nitrile 36 with the opposite stereochemistry at C9 to that seen in 34 (Scheme 10).The subsequent N-alkylation of 36 gave rise to the N-phenethyl ether 37, and the Odemethylation of 37 provided the saturated cyanomethyl diastereomer 1S,5R,9R-38.Starting with 32, the enantiomer of 26, 1R,5S,9S-41, was obtained using similar procedures.
To obtain the saturated nitriles, we needed to reduce the alkene in 28 and 33 while leaving the nitrile intact.Employing traditional conditions with hydrogen and palladium on carbon as the catalyst resulted in a complex mixture of both alkene and partial nitrile reduction.However, using an H-cube hydrogenation flow reactor, we were able to develop conditions that gave the desired enantiomeric saturated nitriles 1S,5R,9S-34 and 1R,5S,9R-35 in low-to-moderate yield (Scheme 10).The reduction of 26 using an H-cube hydrogenation flow reactor gave rise to the saturated nitrile 36 with the opposite stereochemistry at C9 to that seen in 34 (Scheme 10).The subsequent N-alkylation of 36 gave rise to the N-phenethyl ether 37, and the Odemethylation of 37 provided the saturated cyanomethyl diastereomer 1S,5R,9R-38.Starting with 32, the enantiomer of 26, 1R,5S,9S-41, was obtained using similar procedures.

Synthesis of the C9 Methylaminomethyl Diastereomers
Access to one pair of the C9 methylaminomethyl diastereomers was obtained from the N-phenylethyl ketone 1. 1S,5S-1 was subjected to a Witting reaction with methoxymethyl triphenylphosphonium chloride to give rise to the corresponding enol ether 42 (Scheme 12).
The two C9 methylaminomethyl analogs 44 and 45 were prepared by converting the enol ether 42 to the corresponding aldehyde 43 by treating it with 6 N HCl.The aldehyde 43 was not isolated.It was subjected to reductive amination with methylamine using titanium (IV) isopropoxide as the catalyst [16].Sodium borohydride was added at 0 • C to reduce the formed imine.The reaction was carried out in a pressure vessel to prevent the methylamine from escaping.This resulted in a 1:1 diastereomeric mixture of the amines 44 and 45 with a 75% yield (Scheme 12).The chromatographic separation of the secondary amines and the O-demethylation of the phenolic ethers gave rise to the desired methylaminomethyl diastereomers 1R,5S,9S-46 and 1S,5R,9R-47 in 65% and 81% yields, respectively (Scheme 12).Scheme 12. Synthesis of the 1S,5R,9S-C9 methylaminomethyl phenylmorphan (46) and 1S,5R,9R-C9 methylaminomethyl phenylmorphan (47) diastereomers.
The two C9 methylaminomethyl analogs 44 and 45 were prepared by converting the enol ether 42 to the corresponding aldehyde 43 by treating it with 6 N HCl.The aldehyde 43 was not isolated.It was subjected to reductive amination with methylamine using titanium (IV) isopropoxide as the catalyst [16].Sodium borohydride was added at 0 °C to reduce the formed imine.The reaction was carried out in a pressure vessel to prevent the methylamine from escaping.This resulted in a 1:1 diastereomeric mixture of the amines 44 and 45 with a 75% yield (Scheme 12).The chromatographic separation of the secondary amines and the O-demethylation of the phenolic ethers gave rise to the desired methylaminomethyl diastereomers 1R,5S,9S-46 and 1S,5R,9R-47 in 65% and 81% yields, respectively (Scheme 12).
The enantiomers of 46 and 47 were prepared using the methods described above.The C9 secondary amines 50 and 51 were O-demethylated to give rise to 1R,5S,9R-52 and 1R,5S,9S-53 in 96% and 66% yields, respectively (Scheme 13).The assigned stereochemistry of the methylaminomethyl compounds came from the known stereochemistry of C9 analogs obtained from the same common aldehyde intermediate, 49.The structure of the product of that reaction of the aldehyde was proven by X-ray crystallographic structure analysis [5,6].
The assigned stereochemistry of the methylaminomethyl compounds came from the known stereochemistry of C9 analogs obtained from the same common aldehyde intermediate, 49.The structure of the product of that reaction of the aldehyde was proven by X-ray crystallographic structure analysis [5,6].
2.8.Forskolin-Induced cAMP Accumulation Assay for In Vitro Determination of the Potency and Efficacy of the C9 Amino, Methylamino, Cyanomethyl, and Methylaminomethyl Compounds Compounds were evaluated using the inhibition of the forskolin-induced cAMP accumulation assay: functional activity for cAMP (HitHunter Chinese hamster ovary cells (CHO-K1) that express the human µ-opioid receptor (OPRM1)) (Table 2).The assigned stereochemistry of the methylaminomethyl compounds came from the known stereochemistry of C9 analogs obtained from the same common aldehyde intermediate, 49.The structure of the product of that reaction of the aldehyde was proven by X-ray crystallographic structure analysis [5,6].

Forskolin-Induced cAMP Accumulation Assay for In Vitro Determination of the Potency and Efficacy of the C9 Amino, Methylamino, Cyanomethyl, and Methylaminomethyl Compounds
Compounds were evaluated using the inhibition of the forskolin-induced cAMP accumulation assay: functional activity for cAMP (HitHunter Chinese hamster ovary cells (CHO-K1) that express the human μ-opioid receptor (OPRM1)) (Table 2).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, background readouts of the vehicle control were subtracted from all treatment readouts and then normalized to the forskolin control.Data were then analyzed in GraphPad Prism 8 (GraphPad, La Jolla, 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 EC 90 of fentanyl; degree of antagonism (I max ) normalized to naltrexone.c DOR antagonist potency (IC 50 ) determined versus EC 50 of SNC80; degree of antagonism (I max ) normalized to naltrexone.d KOR antagonist potency (IC 50 ) determined versus EC 90 of U50488H; degree of antagonism (Imax) normalized to nor-BNI.
The C9 amino compounds 3, 6, 22, and 23 in Table 2 illustrate the importance of stereochemistry on MOR agonist activity.It is well known that enantiomers can greatly differ in their pharmacology, with one showing much more of an effect at the MOR than the other.Thus, 1R,5R,9S-6 was a moderately potent and almost fully efficacious MOR agonist (MOR EC 50 = 24 nM (% E max = 91%)), and its enantiomer, 1S,5S,9R-3 (MOR EC 50 > 10,000 nM), had no effect as an agonist and little potency as an MOR antagonist (IC 50 = 709 nM).Less difference was noted in the potencies and efficacies of the C9 amino enantiomeric pair 1S,5S,9S-22 and 1R,5R,9R-23, where 1R,5R,9R-23 was a moderately potent MOR partial agonist (EC 50 = 26 nM (% E max = 81%)) and was about three times as potent as 1S,5S,9S-22.The latter compound had much less potency as an MOR agonist and little efficacy (EC 50 = 94 nM (%E max = 53%)).There was a major difference in functional activity between 5-phenymorphans with a C9 amino moiety and those with a C9 hydroxyl substituent.The comparable chiral compound in the C9 hydroxy series, with 1R,5R,9S stereochemistry, was extremely potent in vitro and in vivo [3].Although a direct comparison cannot be made because the amino compounds were examined using the cAMP assay and the hydroxy compounds were evaluated with MOR receptor binding in a [ 35 S]GTPgS assay, it was obvious that the 1R,5R,9S-amino compound had much less potency and efficacy.Possibly the amino compound was less capable of the hydrogen bonding that apparently enabled the high potency of the hydroxy stereoisomer, as postulated through molecular modeling using energy-minimized structures [3].

General Information
All reactions were performed in oven-dried glassware under an argon atmosphere, unless otherwise noted.Proton ( 1 H NMR) and carbon ( 13 C NMR) spectra were recorded on a Varian Gemini-400 spectrometer at 400 MHz for 1 H NMR and 101 MHz for 13 C NMR.The spectra have been shown in the Supplementary Materials.Mass spectra (HMS) were recorded on a Waters (Mitford, MA, USA) Xevo G2-XS QTof.Ions were produced using positive ion electrospray (ESI) at a capillary voltage of 2.8 KV.The ESI source temperature was 280 • C. The optical rotation data were obtained on a PerkinElmer polarimeter model 341, and melting points were obtained using Thomas Hoover capillary melting point (mp) apparatus.Thin-layer chromatography (TLC) was performed on a 250 mm Analtech GHLF.Visualization was accomplished under UV or by staining in an iodine chamber.Flash column chromatography was performed with Fluka silica gel 60 (mesh 220-400).The solvents used were CHCl 3 and CMA (CHCl 3 :MeOH:NH4OH (50:45:5)), usually using a gradient of 0 → 10% or 15% CMA/CHCl 3 , or hexane and ethyl acetate at various gradients.Elemental analyses were performed by Robertson Microlit Laboratories, NJ, USA.

Forskolin-Induced cAMP Accumulation Assays
Assays were performed as previously described [4].Briefly, 10,000 cells/well of cells were plated in 384-well tissue culture plates and incubated overnight at 37 • C in 5% CO 2 .Stock solutions of the compound were prepared in DMSO at a 5 mM concentration, and then 9 to 10 concentrations of 100X working solutions were prepared by serial dilution using DMSO.Furthermore, 5X working solutions were subsequently prepared using forskolincontaining assay buffer (consisting of HBSS and HEPES).For the agonist assay, cells were incubated at 37 • C with compounds for 30 min at a 1X final concentration.For the antagonist assay [17], cells were incubated at 37 • C with compounds for 15 min before 30 min of incubation at 37 • C with the selected agonist at their EC 50 or EC 90 doses.Detection was carried out by using the HitHunter cAMP Assay for Small Molecules by DiscoverX according to the manufacturer's directions, and the BioTek Synergy H1 hybrid plate reader (BioTek, Winooski, VT, USA) and Gen5 Software version 2.01 were used to quantify luminescence (BioTek, Winooski, VT, USA).To determine the % efficacy in forskolin-induced cAMP assays, background readouts of the vehicle control were subtracted from all treatment readouts and then normalized 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.The degree of antagonism (I max ) was normalized to naltrexone (MOR, DOR) or nor-BNI (KOR).

X-ray Crystal Data
Single-crystal X-ray diffraction data on compound 23 were collected using Cu Kα radiation and a Bruker SMART APEX II CCD area detector.The crystal was prepared for data collection by coating it with high-viscosity microscope oil.The oil-coated crystal was mounted on a micromesh mount (MiTeGen, Inc., Ithaca NY, USA) and transferred to the diffractometer, where a dataset was collected at 100(2) K.The 0.341 × 0.046 × 0.030 mm 3 crystal was orthorhombic in space group P2 1 2 1 2 1 .The structure was solved by direct methods and refined by full-matrix least-squares refinement on F 2 values using the programs found in the SHELXL suite (Bruker, SHELXL v2014.7,2014, Bruker AXS Inc., Madison, WI, USA).Corrections were applied for Lorentz, polarization, and absorption effects.The parameters refined included atomic coordinates and anisotropic thermal parameters for all non-hydrogen atoms.The H atoms were included using riding models and direct assignment.The hydrogen atoms could not be found when the solvent, methanol, was used for recrystallization.This is likely due to the positional disorder that was modeled.The stereochemistry of C1 (R), C5 (R), and C9 (S) was determined via the orientation of the molecules.It is worth noting that there are A and B alerts observed in the cif.Each will be addressed here: The rotational disorder of a solvent water molecule resulted in an "alert A", a CCDC qualifier which, in this molecule, suggests short hydrogen bonding distances to the protonated amine.The hydrogen bonding is likely due to rotations of the amine as well.The residual electron density observed in alert B is due to a further positional disorder of the methanol mentioned above.Attempts to resolve this disorder with additional parts were not successful.The proton on the water without an acceptor is due to the water molecule having a rotational disorder between two molecules that it interacts with, as discussed in alert A. The outlier reflections were checked, and there were no major missing or overly intense peaks (~10 sigma).Nothing is out of the ordinary with each reflection, and no changes were made.None of the alert C or G errors were a cause for concern, but they were checked on a case-by-case basis to verify there were no major problems.Tables of the X-ray spectroscopic data were included in the Supplementary Materials.
The atomic coordinates for 23 have been deposited with the Cambridge Crystallographic Data Centre, deposition number 2336627.Copies of the data can be obtained, free of charge, upon sending an application to CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK (fax: +44(0)-1223-336033 or e-mail: deposit@ccdc.cam.ac.uk).

Conclusions
None of the C9 amino compounds were as potent as the formerly reported stereoisomers bearing a 1R,5R,9S-OH or a 1R,5S,9R-methyl substituent at C9 in 5-phenylmorphan [3].Only the cyanomethyl compound 1R,5S,9R-35 had nanomolar potency among the various compounds with a nitrogen atom; it was about 7x more potent than morphine and was found to have high MOR efficacy.An amino substituent directly attached at C9 had lowered potency in comparison with the generally more potent methylaminomethyl compounds with a secondary amino function situated one carbon away from C9.Both the stereochemistry of the molecule and that of the C9 substituent were found to affect potency and efficacy.Most of these new compounds were either fully efficacious or had little or no efficacy at the MOR, with two exceptions: the primary amine partial agonist 1R,5R,9R-23 and the saturated cyanomethyl compound 1S,5R,9R-38.The former was only moderately potent, while the latter was twice as potent as morphine.That compound, 1S,5S,9R-38, had potency and efficacy like those of a compound that had been previously noted, bearing a C9 hydroxyethyl substituent at C9.That compound was found to have fewer adverse side effects than those displayed by analgesics like fentanyl [13].Among the enantiomers, it was common to have a ten-fold difference in potency.Only a few compounds were weak MOR antagonists (compounds 3, 8, and 41), and only one compound, 1S,5R,9S-46, was found to be a potent KOR agonist (KOR EC 50 = 18 nM (% E max = 103%)).Two compounds, 33 and 35, were potent DOR agonists, but neither was efficacious at DOR (DOR % E max = 69% and 35%, respectively).

Table 2 .
Opioid receptor activity measured in the forskolin-induced cAMP accumulation assay a .

Table 2 .
Opioid receptor activity measured in the forskolin-induced cAMP accumulation assay a .