Synthesis and pharmacological effects of the enantiomers of the N-phenethyl analogues of the ortho and para e- and f-oxide-bridged phenylmorphans

Abstract

The N-phenethyl analogues of (1R*,4aR*,9aS*)-2-phenethyl-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridin-6-ol and 8-ol and (1R*,4aR*,9aR*)-2-phenethyl-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2.3-c]pyridin-6-ol and 8-ol, the ortho- (43) and para-hydroxy e- (20), and f-oxide-bridged 5-phenylmorphans (53 and 26) were prepared in racemic and enantiomerically pure forms from a common precursor, the quaternary salt 12. Optical resolutions were accomplished by salt formation with suitable enantiomerically pure chiral acids or by preparative HPLC on a chiral support. The N-phenethyl (−)- para-e enantiomer (1S,4aS,9aR-(−)-20) was found to be a μ-opioid agonist with morphine-like antinociceptive activity in a mouse assay. In contrast, the N-phenethyl (−)-ortho-f enantiomer (1R,4aR,9aR-(−)-53) had good affinity for the μ-opioid receptor (Ki = 7 nM) and was found to be a μ-antagonist both in the [³⁵S]GTP-γ-S assay and in vivo. The molecular structures of these rigid enantiomers were energy minimized with density functional theory at the level B3LYP/6-31G* level, and then overlayed on a known potent μ-agonist. This superposition study suggests that the agonist activity of the oxide-bridged 5-phenylmorphans can be attributed to formation of a seven membered ring that is hypothesized to facilitate a proton transfer from the protonated nitrogen to a proton acceptor in the μ-opioid receptor.

Introduction

(1R*,4aR*,9aS*)-2-Methyl-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridin-8-ol (1, Chart 1) and 6-ol (2, Chart 1) and (1R*,4aR*,9aR*)-2-methyl-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2.3-c]pyridin-8-ol (3, Chart 1) and 6-ol (4, Chart 1), the ortho- and para-e and the ortho- and para-f oxide-bridged phenylmorphans,^2–6 were previously synthesized as part of our program to prepare and pharmacologically evaluate the ortho- and para-hydroxy a through f racemic oxide-bridged phenylmorphan structural class of isomers (12 racemates, Chart 1), as well as the enantiomers of those that have reasonable affinity for opioid receptors. The ortho e- (1R*,4aR*,9aS*) and the ortho f-(1R*,4aR*,9aR*), or the para e- and para f-oxide-bridged phenylmorphans are epimers and differ structurally only in their configuration at C9a.

Chart 1.

The syntheses of all of the racemic compounds have been accomplished and, except for the b-isomer, all of the syntheses have been reported.^2–11 In 1983, we noted that the N-methyl ortho-f isomer had some affinity (IC₅₀ = ca. 100 nM) for opioid receptor preparations (from whole brain homogenate),⁵ and it had intriguing, but slight, antagonist activity in the tail flick antagonism vs. morphine assay in mice.⁵ One of the reasons that we explored the synthesis of this new class of opioids, the oxide-bridged phenylmorphans, was to determine the spatial characteristics of ligands that might be capable of interacting with opioid receptors. If some of these rigid compounds could be found that interacted with opioid receptors as agonists and others as antagonists, while acting through the same opioid receptor, we might obtain information about the three-dimensional structure of a ligand that is required for interaction with a specific opioid receptor to give the determined pharmacological activity.

Oxide-bridged phenylmorphans are based on the 5-phenylmorphan structure. The N-methyl-5-(3-hydroxyphenyl)morphan (5, Chart 2) was prepared by May and Murphy¹² as a simplified analogue of morphine and had morphine-like antinociceptive activity in mouse assays.¹² The (1R,5S)-(−)-enantiomer¹³ of N-methyl-5-(3-hydroxyphenyl)morphan was reported to be morphine-like in analgesic activity and to have nalorphine-like antagonist activity, while the (1S,5R)-(+)-enantiomer was about three to five fold more potent than morphine in mice and did not appear to have antagonist activity in morphine-dependent monkeys.^14–16 The racemic N-phenethyl derivative of 5-(3-hydroxyphenyl)morphan ¹⁴ (6, Chart 2) had about half of the antinociceptive activity of the racemic N-methyl analogue and the N-phenethyl enantiomers^{17, 18} were found to be opioid antagonists in the [³⁵S]GTP-γ-S assay. The meta-hydroxy group is known to play a critical role in the potency of this class of analgesics.¹⁹ The aromatic phenolic ring is not positioned similarly with respect to the piperidine ring in the morphinans and the phenylmorphans, but despite the spatial differences in that moiety between these classes of analgesics, N-methyl compounds from both classes usually have antinociceptive activity and some are known to be partial agonists. Much greater pharmacological differences between classes of opioids are seen with N-phenethyl derivatives. These are generally potent antinociceptive agents in the rigid morphinan, epoxymorphinan, and benzomorphan classes of opioids and can have either opioid agonist¹ or antagonist activity^{18, 20} in the 5-phenylmorphans. The N-phenethyl substituted 5-(3-hydroxyphenyl)morphans with a C9S-hydroxy ((1R,5R,9S)-(−)-7, Chart 2, (1R,5R,9S)-(−)-9-hydroxy-5-(3-hydroxyphenyl)-2-phenylethyl-2-azabicyclo[3.3.1]nonane) or a methylene substituent at C9 have been found to have very high (sub-nanomolar) affinity at μ- and good affinity at δ-receptors. They had potent antinociceptive activity in vivo. The (−)-7 was found to be about 500 to 1000 fold more potent than morphine in vivo.¹ Its C9R-epimer (1R,5R,9R)-(+)-8 (Chart 2, (1R,5R,9R)-(+)-9-hydroxy-5-(3-hydroxyphenyl)-2-phenylethyl-2-azabicyclo[3.3.1]nonane) had much less affinity for opioid receptors (Ki = 59 nM at the μ receptor). N-Phenethyl substituted 5-(3-hydroxyphenyl)morphans usually,^{17, 18, 20} but not always,¹⁴ shows greater receptor affinity, efficacy, or potency in vivo than the comparable N-methyl analogue.

Chart 2.

In order to try to increase affinity in the f-series⁵ of oxide-bridged phenylmorphans, where the N-methyl ortho-hydroxy racemate had poor affinity (IC₅₀ ca. 100 nM),¹⁰ and prepare more potent opioid agonists or antagonists we decided to synthesize their N-phenethyl analogues and also to examine the effect of the N-phenethyl substituent in the ortho and para e-series. In both the racemic ortho and para e oxide-bridged phenylmorphan series, the N-methyl derivatives had little or no affinity for any opioid receptor (Ki > 1 μM, Table 1). Hashimoto et al.,³ obtained the racemic ortho- and para-e (1R*,4aR*,9aS*)-oxide-bridged phenylmorphans via intramolecular aromatic nucleophilic substitution of fluorine by alkoxide, and we used that procedure, through the common intermediate 12 (Scheme 1), to obtain both the ortho- and para-e and ortho- and para-f oxide-bridged N-substituted 5-phenylmorphan racemates, and prepared their chiral relatives. In this report we discuss the syntheses, opioid receptor affinities of the racemic and enantiopure N-phenethyl derivatives of the ortho- and para-e and the ortho-and para-f oxide-bridged phenylmorphans, and the efficacies (of the more interesting compounds) determined by the [³⁵S]GTP-γ-S assay as well as through in vivo antinociceptive assays, and rationalize the pharmacological activities by quantum chemical studies.

Table 1.

Binding affinity of the hydrobromide salts of the racemic N-methyl substituted ortho-e (1) and para-e (2) oxide-bridged phenylmorphans.^a

	Ki (nM)
Compound	μ(cells)	δ(cells)	κ(cells)
1 (1R*,4aR*,9aS* ortho-e)	5800 ± 300	> 5500	> 6500
2 (1R*,4aR*,9aS* para-e)	1250 ± 70	> 5500	1270 ± 50

^aBinding assays were performed using [³H]DAMGO and rat brain membranes for μ-receptors, [³H]DADLE and rat brain membranes for δ-receptors and [³H]U69,593 and guinea pig brains membranes for κ-opioid receptors. Ki values are derived from the pooled data of two experiments. Assays were carried out as previously noted.¹⁹

Scheme 1.

Synthesis of N-methyl para-e oxide-bridged phenylmorphan 2.³ Reagents and conditions: a) NaH, THF, Me₂NCH₂CH₂Cl; (87%), b) NaH, 5-bromovaleronitrile, THF; (69%, HCl salt), c) CF₃CO₂H, H₂O, heat; (95%, HBr salt); d) Br₂, CHCl₃; e) i-PrOH, heat; (69%, over 2 steps]; f) LiEt₃BH, THF; (74%, HBr salt); g) NO₂BF₄, sulfolane or 70% HNO₃; (91%); h) NaH, THF, reflux 3 days; (80%), i) H₂, Pd-C, MeOH; (~quant.); j) NaNO₂, H₂SO₄, Cu(NO₃)₂, Cu₂O, (25–89%).

Chemistry

Starting materials

We recently reported a novel approach to the racemic ortho- and para-e oxide-bridged phenylmorphan isomers through intramolecular aromatic nucleophilic substitution of fluorine by alkoxide.³ Commercially available 2-fluorophenyl acetonitrile (9, Scheme 1) was converted into quaternary salt 12 in five steps,³ establishing the desired carbon skeleton via Thorpe-Ziegler cyclization. The quaternary salt 12 was stereoselectively reduced and demethylated in a one-pot procedure and the aromatic ring was activated for the key ring-forming step by nitration.³ We have now found that 70% nitric acid also can be used instead of nitronium tetrafluroborate to give comparable yields of the nitro compound 14. Deprotonation and subsequent attack of alkoxide on the fluorinated aromatic ring resulted in the formation of the desired oxide-bridged derivative 15.³ The cyclization reaction also works well at ambient temperature in dimethylformamide. The nitro group was reduced and the resulting aniline derivative was converted into a free phenol via hydrolysis of the corresponding diazonium salt assisted by copper salts, furnishing the para-e N-methyl oxide-bridged phenylmorphan 2.³ Quaternary salt 12 was prepared in larger quantities and used as a convenient starting material for the synthesis of the N-phenethyl derivatives in the para-e oxide-bridged series (Scheme 1), as was (1R*,4aR*,9aS*)-2-methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (15) for the para-f oxide-bridged series (Scheme 2).

Fig. 4.

Overlap of ortho-f (1R,4aR,9aR-(−)-53) and para-f (1R,4aR,9aR-(−)-26) oxide-bridged phenylmorphan isomers. The phenolic hydroxyl of the ortho-f can form a seven-membered ring with the bridged oxygen atom via a water molecule while that of the para-f is not positioned to form such a seven-membered ring. Dotted lines represent hydrogen bonding.

Scheme 2.

Synthesis of N-phenethyl para-e and para–f oxide-bridged phenylmorphans 20 and 26. Reagents and conditions: a) ClCO₂Et, ClCH₂CH₂Cl, K₂CO₃; (87–89%), b) 33% HBr, AcOH, 50 °C; (79–81%), c) PhCH₂CH₂Br, KI, CH₃CN, heat; (78–80%), d) H₂, Pd-C; (92–100%), e) NaNO₂, H₂SO₄ then H₂SO₄, heat, (82–89%).

Synthesis of N-phenethyl derivatives of para-e and para-f oxide-bridged phenylmorphans

The 2-methyl-6-nitro derivative 15 (Scheme 2) was converted into the carbamate 16. Both the methyl and ethyl derivatives were prepared, and the yields from the latter compound were superior. Initial alkaline hydrolysis attempts gave complex mixtures, however the hydrolysis under acidic conditions²¹ proceeded smoothly to give the N-nor amine 17. Direct alkylation of 17 in refluxing acetonitrile²² gave good yields of the phenethyl derivative 18, which was reduced to the corresponding aniline derivative 19. Optimization of the diazotization/hydrolysis sequence, modifying conditions of Schnider and Grüssner,²³ indicated that slow addition of relatively diluted cold diazonium salt solution into boiling ca. 50% sulfuric acid (v/v) and brief refluxing reproducibly gave the desired phenolic product 20, the N-phenethyl para-e oxide-bridged phenylmorphan, in good yield. Similarly, 21 was converted to 26, the N-phenethyl para-f oxide-bridged phenylmorphan (Scheme 2).

The tertiary amine 15 was resolved via formation and recrystallization of diastereomeric salts with enantiomers of 3-bromo-8-camphorsulfonic acid. Enantiomeric purity was assessed by ¹H-NMR spectroscopy using S-(+)-1-phenyl-2,2,2-trifluoroethanol as a shift reagent. The presence of the other enantiomer was undetectable indicating >98% enantiomeric purity.²⁴ Absolute stereochemistry was determined by a single-crystal X-ray crystallographic analysis of the salt of (+)-15 with [(1S)-(endo, anti)]-(−)-3-bromocamphor-8-sulfonic acid, and it was found to be 1R,4aR,9aS (Fig. 1). The synthetic sequence shown in Scheme 2 was repeated using enantiopure intermediates to give both the (+)- and (−)-enantiomers of compounds 16 and 22. The sign of optical rotation did not change during any step of the synthetic sequence from 15 to 20 (Schemes 1 and 2), nor did it change during the synthesis of the 1R,4aR,9aS-(+)-enantiomer of 43 from the 1R,4aR,9aS-(+)-enantiomer of 15 (Scheme 5).

Fig. 1.

X-ray crystallographic structures of (1R,4aR,9aS)-2-methyl-6-nitro-1,3,4,9a- tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine•[(1S)-(endo,anti)]-(−)-3-bromocamphor-8-sulfonate ((+)-15), (1S*,4aS*,9aS*)-21, (1R,4aR,9aR )-5-(2-fluoro-5-nitrophenyl)-2-methyl-2-azabicyclo[3.3.1]nonan-9α-ol ((+)-29), and (1S,4aS,9aS)-8-hydroxy-2-(2-phenethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine•HBr ((+)-53). For all four compounds displacement ellipsoids are shown at the 50% level.

Scheme 5.

Synthesis of the N-phenethyl derivative of the ortho-e and ortho–f oxide-bridged phenylmorphans 43 and 53. Reagents and conditions: a) H₂, Pd-C, MeOH;(85–92%), b) NaNO₂, HCl, CuSO₄, Na₂S₂O₅, NaOH; (66–79%), c) NaNO₂, CF₃CO₂H; (74–84%), d) ClCO₂Et, ClCH₂CH₂Cl, K₂CO₃; (88%), e) 33% HBr, AcOH, 50 °C; (85–91%), f) PhCH₂CH₂Br, KI, CH₃CN, heat; (78–85%), g) H₂, Pd-C, EtOH, CH₃CO₂Na•3H₂O, 45–50 °C; (86–94%), h) NaNO₂, H₂SO₄ then Cu(NO₃)₂, Cu₂O, (35–61%).

New approach to para-f-oxide-bridged phenylmorphans

Our previously reported synthesis of this oxide-bridged phenylmorphan relied on cyclization of a free phenol derived from compound 27⁴ (Scheme 3). This approach was complicated by ring rearrangements, i.e., when the preparation of mesylate 28 from the corresponding equatorially-oriented (in the piperidine ring) alcohol and methanesulfonyl chloride was attempted the rearranged chloride 27 was obtained. However, upon generation of the free phenol from the ortho-methoxy moiety, compound 27 did cyclize to give the para-f-oxide-bridged compound 4.⁴ Preparation of an equatorially-oriented alkoxide from 29 should enable cyclization in the desired manner, as demonstrated earlier with the analogous axially-oriented alkoxide,³ since only the configuration of the nucleophile was reversed (Scheme 3). In the former approach to the e-bridged phenylmorphans the quaternary salt 12 (Scheme 1) was stereoselectively reduced to the β-alcohol 13 and dequaternized in a one-pot process by reduction with Superhydride.³ We theorized that compound 12 could serve as a convenient common precursor for the syntheses of both the e- and f-oxide bridged compounds by modifying the reduction protocol to selectively obtain the equatorially-oriented alcohol. Dry distillation²⁵ of compound 12 (Scheme 4) provided the corresponding tertiary amine 30 previously prepared by thermolysis of 12 in refluxing diphenyl ether and chromatographic purification.³ Fortunately, there was a precedent for the desired reduction of the ketone analogous to 30, differing only in substitution of the aromatic ring with methoxy groups.⁴ Thus, gratifyingly the hydrogenation of 30 over PtO₂ gave the desired equatorially-oriented alcohol 31, no trace of the other diastereomer was detected by NMR.

Scheme 3.

New approach to para-f-oxide-bridged phenylmorphan 4.

Scheme 4.

Synthesis of the N-methyl derivative of the para-f oxide-bridged phenylmorphan 4. Reagents and conditions: a) LiEt₃BH, THF, −78 °C; (74%), b) 250 °C, 0.3 mm Hg; (84%), c) NaBH₄, MeOH; (70%, ratio 31: 13 = 1: 1.33), d) H₂, PtO₂, EtOH; (94%), e) HCl, Et₂O then NaBH₄, MeOH; (97%), f) 70% HNO₃; (77%), g) NaH, THF, reflux 3 days; (80%), h) H₂, Pd-C; (90%), i) NaNO₂, H₂SO₄ then H₂SO₄ and heat, (75%).

Reduction with sodium borohydride also furnished the equatorially-oriented alcohol 31. However, it was necessary to convert the amine to its hydrochloride to achieve stereoselective reduction.⁴ Reduction of the free base (Scheme 4) gave an almost equimolar mixture of equatorially-oriented and axially-oriented alcohols (31 and 13, respectively). Carefully monitored nitration of 31 in 70% nitric acid gave a good yield of derivative 29 (Scheme 4). Unlike its axially-oriented analogue 13,³ the equatorially-oriented alcohol 31 is less stable under those conditions and extensive decomposition was observed at higher temperatures and prolonged reaction times. Cyclization of the corresponding alkoxide in refluxing tetrahydrofuran provided the desired oxide-bridged phenylmorphan 21. The para hydroxy f-oxide-bridged phenylmorphan 4 was prepared in two additional steps via the amine 32 and its spectral data matched those published previously.⁴

The cyclization of 29 in dimethylformamide at ambient temperature (Scheme 4) produced a complex mixture of by-products. The cyclization of the equatorially-oriented alkoxide 29, unlike the axially-oriented alcohol 13, was concentration sensitive. At higher concentrations the yield of compound 21 was significantly decreased by the formation of two by-products, possibly axially-oriented dimers (based on ¹H-NMR and HRMS) that could have formed via intermolecular reaction, since equatorially-oriented alkoxide 29 is much more sterically accessible for the intermolecular nucleophilic reaction than corresponding axially-oriented alkoxide 13, which is hindered by the syn vicinal tertiary amine. The relative stereochemistry of 21 (1S*,4aS*,9aS*) was confirmed by single-crystal X-ray analysis (Fig. 1). Repetition of the demethylation/alkylation sequence followed by transformation of the nitro group into the free phenol via the diazonium salt, as shown in Scheme 2, provided the N-phenethyl derivative (26). <Figure 1 here>

In order to obtain enantiopure materials several chiral acids/solvent combinations were screened, but all attempts to resolve oxide-bridged compound 21 were not fruitful. However, alcohol 29 was, after some experimentation, successfully resolved through salt formation with di-O,O′-p-toluoyl-D- or L-tartaric acids in methanol-acetone mixture. The NMR chiral shift reagent used for compound 15 was not effective with compound 29, therefore its enantiomeric purity was assessed by HPLC using a Daicel Chiralcel OD column. The absolute stereochemistry was determined by an X-ray crystallographic study of the salt of (+)-29 to be 1R,4aR,9aR (Fig. 1). The synthetic sequence from 29 to the final product 26 was also carried out using enantiopure materials. The sign of optical rotation changes during cyclization of 29 to 21, then stays the same through every step of the synthetic sequence leading to 26 (Scheme 2).

Synthesis of N-phenethyl derivatives of ortho-e and ortho-f-oxide-bridged phenylmorphans

The previously published approach³ to the corresponding ortho-hydroxy N-methyl derivatives utilizing the bromo derivative 33 was modified because we found that the use of NO₂BF₄/sulfolane for nitration³ gave three products (34, 15, and 35, Chart 3). In this approach we initially planned to use the intermediate 34 (Chart 3) to obtain the desired e-isomer, and to use a comparable intermediate for the f-isomer. If compound 34 could be obtained in good yield, the synthesis of the f-isomer would be considerably improved. However, nitration of 33 gave only moderate yields of the desired compound 34 using NO₂BF₄.³ When 70% HNO₃ was used, the isomeric compound 35 along with compound 15 were the sole products in ca 6: 4 ratio (¹H-NMR). This could be explained by ipso nitration²⁶ and the resulting migration or loss of bromide giving rise to by-products 35 and 15. The nitration was studied more intensely on bromide 44 where nitration with 70% HNO₃ at 60 °C gave a mixture of two compounds 21 and 46. Lowering the reaction temperature to room temperature or to 0 °C did not change the outcome of the reaction. Nitration of 44 with NO₂BF₄ (2.0 equiv) in sulfolane at room temperature gave a mixture of starting material 44 and, based on ¹H-NMR, two compounds 21 and 46 in a ratio of 3: 2.9: 1, respectively. When more forceful conditions were employed (i.e. NaNO₂/TFA²⁷) the product of ipso attack, 21, was obtained as the sole product (Chart 3).

Chart 3.

Reagents and conditions: a) NO₂BF₄, sulfolane; (54% of 34), b) 70% HNO₃, reflux; c) 70% HNO₃, 0 °C, 20 °C, or 60 °C; (44 % of 35), d) NO₂BF₄ (2.0 equiv), sulfolane, 20 °C; e) NaNO₂, TFA.

We therefore decided to replace the bromo-substituent in 33 and 44 with a more stable halide to prevent ipso nitration and its resultant by-products. The f- and e-oxide-bridged chloro analogues 37 and 47 (Scheme 5) were prepared from the corresponding aniline derivatives 36 and 32, and they proved to be stable under these reaction conditions. The optimal reagent system for the nitration was found to be NaNO₂/TFA,²⁷ providing high yields of the desired 6-chloro-8-nitro-substituted derivatives 38 and 48. The ipso nitration-derived by-products were not observed. With 48, two re-crystallizations were necessary to remove a minor impurity (ca. 6% by HPLC), which could not be separated by chromatography. The N-methyl derivatives were converted to the corresponding ethylcarbamates 39 and 49, and these were hydrolyzed in 33% hydrobromic acid to give the N-nor amines 40 and 50.

Direct alkylation of 40 and 50 gave the desired N-phenethyl derivatives 41 and 51. Each of these was reduced over Pd-C in ethanol at ca. 50–60 °C (dechlorination is very slow at room temperature) in the presence of sodium acetate trihydrate, to provide the corresponding dehalogenated aniline derivatives 42 and 52. Attempted use of the conditions for the conversion to the ortho-phenols that had been optimized with the corresponding para-derivatives, failed to yield the desired products (possibly due to the decomposition of the diazonium salt in hot sulfuric acid). Instead, a milder method²⁸ was adapted utilizing copper salts to decompose the diazonium salt at or below room temperature. This method gave products 43 and 53, albeit in modest yields, and their isolation required repeated chromatographic purification. The racemates were resolved by HPLC on a Daicel Chiralcel OD semi-preparative column to give the enantiomers of 43 and 53 in sufficient amounts for initial pharmacological testing. The absolute stereochemistry of the desired N-phenethyl derivative of the f-isomer ((+)-53) was determined by X-ray diffraction analysis of its HBr salt to be 1S,4aS,9aS (Fig. 1). The attempts to grow suitable crystals of (+)- or (−)-43•HBr were not successful, therefore the sequence from resolved (+)-15 to (+)-43 was repeated, the sign of optical rotations stays the same throughout the sequence. Proton NMR and chiral HPLC analysis confirmed the identity and purity of this material, which co-eluted with the previously obtained sample of (+)-43. The absolute stereochemistry of (+)-43 was, then, assigned to be 1R,4aR,9aS from the synthetic methodology used and the X-ray diffraction analyses on 1R,4aR,9aS-(+)-15.

Results and Discussion

The affinity of the racemic N-methyl compounds are listed in Table 1, and the racemic and enantiomeric N-phenethyl compounds in Table 2. The functional activity of the compounds that were found to have significant affinity at μ-, δ- and/or κ-receptors is shown in Table 3. The racemic N-phenethyl analogues showed considerably enhanced receptor affinity compared with the racemic ortho-and para-e N-methyl compounds (43 and 20) that had slight or no affinity for opioid receptors. Two of the N-phenethyl enantiomers were especially interesting. The (−)-para-e and the (−)-ortho-f N-phenethyl compounds (1S,4aS,9aR-(−)-20 and 1R,4aR,9aR-(−)-(−)-53) had fair (Ki = 23 nM) and good (Ki = 7 nM) affinity for the μ-opioid receptor, respectively. The (−)-para e enantiomer (−)-20 had weak agonist activity in the [³⁵S] GTP-γ-S assay, and the (−)-ortho f enantiomer (−)-53) had fairly potent μ-antagonist activity in that assay (Ke = 1.4), four times more potent than naloxone (Ke = 6.4). The (−)-ortho f isomer ((−)-53) was also naloxone-like as an antagonist at the κ-receptor. Presumably, the N-phenethyl (−)-para e compound (−)-20) had good intrinsic efficacy, because it was found to be more potent than expected in vivo. It was morphine-like as an antinociceptive in the mouse tail-flick assay (Table 4). Corroboration of the antagonist activity of the N-phenethyl (−)-ortho f enantiomer (−)-53 was obtained from the tail-flick vs. morphine antagonist assay in mice (Table 4), although it was found to be less potent in vivo than would have been expected from the functional assay.

Table 2.

[¹²⁵I]Ioxy binding data for the oxalate salts of the racemic N-phenethyl substituted ortho- and para-e- and ortho- and para-f oxide-bridged phenylmorphans and their enantiomers. ^a

		Ki (nM)
Cmpd #	Compound	μ(cells)	δ(cells)	κ(cells)
43	1R*,4aR*,9aS*-ortho-e	2610 ± 160	2530 ± 100	353 ± 11
(−)-43	1S,4aS,9aR-(−)-ortho-e	2490 ± 140	2540 ± 74	156 ± 5
(+)-43	1R,4aR,9aS-(+)-ortho-e	3190 ± 160	2980 ± 90	2360 ± 68
20	1R*,4aR*,9aS*-para-e	133 ±11	377 ± 28	123 ± 4
(−)-20	1S,4aS,9aR-(−)-para-e	23 ± 1	156 ± 75	57 ± 1
(+)-20	1R,4aR,9aS-(+)-para-e	5130 ± 320	8000 ± 390	638 ± 41
53	1R*,4aR*,9aR*)-ortho-f	17 ± 0.96	1110 ± 41	32 ± 1
(−)-53	1R,4aR,9aR-(−)-ortho-f	7 ± 1	907 ± 32	25 ± 2
(+)-53	1S,4aS,9aS-(+)-ortho-f	201 ± 11	1190 ± 38	116 ± 3
26	1R*,4aR*,9aR*)-para-f	146 ± 8	920 ± 136	257 ± 9
(−)-26	1R,4aR,9aR-(−)-para-f	98 ± 13	1180 ± 60	758 ± 35
(+)-26	1S,4aS,9aS-(+)-para-f	195 ± 24	1150 ± 76	130 ± 9

^a[¹²⁵I]IOXY binding used membranes prepared from CHO cells that stably express the human mu, delta or kappa opioid receptors. Binding affinity is expressed as nM concentrations. All results are ± SD (n = 3). Assays were run as previously noted.¹

Table 3.

Functional data ([³⁵S]GTP-γ-S) ^a of N-phenethyl substituted oxide-bridged phenylmorphans.

		μagonism		δagonism		κagonism
		Emax	ED50 (nM)	Emax	ED50 (nM)	Emax	ED50 (nM)
20	rac-para-e	85 ± 4	770 ± 135	60 ± 4	1730 ± 400	-	NSb
(−)-20	(−)-para-e	75 ± 2	530 ± 41	65 ± 4	1040 ± 250	-	NSb
(−)-26	(−)-para-f	43 ± 3	630 ± 147	-	NSa	-	NSb
DAMGO		100	32 ± 2
		μ antagonism Ke (nM)		δ antagonism Ke (nM)		κ antagonism Ke (nM)
(−)-53	(−)-ortho-f	1.4				8.9
26	rac-para-f	300		NSa		275
(+)-26	(+)-para-f	310		NSa		120
(−)-26	(−)-para-f	244		NSa		NSa
Naloxone		6.4				10

^a[³⁵S]GTP-γ-S binding was performed using CHO cells that stably express the human mu, delta or kappa opioid receptors. All results are ± SD (n = 3). Emax values are expressed as a percent of the maximal stimulation, where 100% is defined as the stimulation produced by 1•M DAMGO (for mu receptors), 500 nM SNC80 (delta receptors) and 500 nM (−)-U50,488 (for kappa receptors). Assays were run as previously noted.¹

^bNS = Not significant activity.

Table 4. In vivo.

activity^a of the oxalate salts of the N-phenethyl substituted (−)-para-e ((−)-20) and (−)-ortho-f ((−)-53) oxide-bridged phenylmorphan enantiomers.

Compound	HP	PPQ	TF	TF vs M
(−)-20	6.4 (2.9–14.1)	1.4 (0.79–2.35)	1.8 (1.2–2.9)b	Inactive (30 mg/kg)
(−)-53	Inactive	Inactive (30 mg/kg)	Inactive 16% at 30 mg/kg	0.4 (0.24–0.54)
Morphine•SO4	0.9 (0.39–1.9)	0.4 (0.2–0.8)	1.9 (0.89–4.14)	Inactive (30 mg/kg)
Naloxone•HCl	Inactive (10 mg/kg)		Inactive (30 mg/kg)	0.04 (0.02–0.09)

^aED50, mg/kg, sc (95% confidence limits); HP = hot plate assay, PPQ = phenylquinone assay, TF = tail flick assay, in mice, TF vs M = tail flick vs. morphine (μ-antagonist assay) in mice, SDS = single dose suppression in monkeys. Vehicle 5% hydroxypropyl-β-cyclodextrin in H₂O for injection. Assays were run at Virginia Commonwealth University (Drs. M. Aceto and L. Harris) as previously noted.¹

^bMild Straub tail at 30 mg/kg in HP and 10 mg/kg in TF.

In order to gain insight into the structural features of the rigid N-phenethyl substituted ortho- and para-e and ortho- and para-f enantiomers that might be responsible for μ-receptor affinity and activity, their spatial relationships were investigated with the aim of providing a rationale for their observed agonist activity. For this purpose, the molecular structures of the compounds in Table 5 were energy minimized with density functional theory at the B3LYP/6-31G* level, and then overlayed onto the phenylmorphan compound (1R,5R,9S)-(−)-7 using the heavy atoms in the morphan moiety (i.e., C1–C9) as a common docking point. As noted, in the various Figures and Schemes, the C9a bonded to the O9 atom in the ortho- and para-e oxide-bridged phenylmorphan (1R*,4aR*,9aS*) has the 9aS*relative configuration, and the C9a bonded to the O9 atom in the ortho- and para-f oxide-bridged phenylmorphan (1R*,4aR*,9aR*) has the epimeric 9aR* relative configuration. The (−)-7, with a C9S configuration, as shown in Chart 2, is a very potent μ-agonist. It has a freely rotating phenolic moiety, and in our previous study¹ we hypothesized that a water molecule/chain H-bonded to the hydroxyl moiety at C9S of the compound (−)-7 facilitates proton transfer from the tertiary amine to an acceptor amino acid at the active site of the receptor as depicted in Fig. 2. We noted that the ease of protonation of the acceptor might be responsible for its activity as a potent agonist. In the present work, we have attempted to further extend the hypothesis of the proton transfer to ortho- and para- e and ortho- and para-f oxide-bridged phenylmorphan isomers. These rigid isomers have a fixed 3-dimensional angular position of the phenolic ring. The e and f oxide-bridged phenylmorphans were calculated to differ by 52° in the spatial position of their phenolic ring (C9a-C4a-C4b-C9).

Table 5.

Overlay and dihedral angles of energy-minimized structures.

Compound	Ki (nM)a	Ex vivo/in vivo	RMSD (Å)b	Dihedral Angle (°)c
(1R,5R,9S)-(−)-71	0.19	agonist	-	-
(1R,5R,9R)-(+)-81	54	ndd	0.03	-
(−)-43 ((−)-ortho e)	2488	ndd	0.06	26.1
(−)-53 ((−)-ortho f)	7	antagonist	0.10	−25.7
(−)-20 ((−)-para e)	23	agonist	0.06	25.7
(−)-26 ((−)-para f)	98	agonist	0.11	−25.5

^a[¹²⁵I]Ioxy binding using CHO cells stably transfected with human μ-DNA and express the human μ receptor.

^bRoot mean square deviation from morphan backbone in (1R,5R,9S)-(−)-7.

^cC9a-C4a-C4b-C9 in the oxide-bridged compounds.

^dNot determined.

Fig. 2.

Schematic representation of the formation of a seven-membered ring via a water molecule that was hypothesized to facilitate proton transfer from the protonated nitrogen in compound 7¹ to an amino acid in the active site. Dotted lines indicate hydrogen bonding.

The rigid-fitting of the ortho-e ((−)-43), ortho-f ((−)-53), para-e ((−)-20) and para-f ((−)-26) compounds on (−)-7 gave ≤ 0.11 Å root mean square deviation (Table 5) indicating a good overlap of the morphan backbone of these isomers to that of (−)-7. Nonetheless, about only half of the phenolic ring of the e- and f-isomers spatially overlaps with the phenolic ring of (−)-7 since the five-membered oxide ring is forced to have a considerably smaller bond angle of C9a-C4a-C4b (e.g., 97.1° for the ortho-f, (−)-53) than the bond angle of C9-C5-C10 (111.3°) for (−)-7, and that results in the pulling of the phenolic ring of the oxide-bridged isomers away from the spatial area occupied by the phenolic ring of (−)-7 toward the C9a atom as shown in Fig. 3. Fig. 3 also illustrates that the ortho- and para-e isomers (−)-43 and (−)-20 differ mainly from the corresponding ortho- and para-f ((−)-53 and (−)-26) isomers in the spatial position of the bridged phenolic ring as shown; for example, the calculated dihedral angle of C9a-C4a-C4b-C9 of the ortho-e ((−)-43) and the ortho-f ((−)-53) are 26.1° and −25.7°, respectively (Table 5). In addition, the bridged oxygen atoms of the ortho-e and ortho-f isomers overlap well in space with the 9S-OH oxygen of (−)-7 and epimeric 9R-OH oxygen of (+)-8, respectively; these oxygen atoms are positioned to form a seven membered ring with the phenolic hydroxyl either via a water molecule as illustrated in Fig. 4 for the ortho-f or via the hydroxyl group of a polar residue such as Tyr in the binding pocket. This seven membered ring might alter the orientation of the protonated nitrogen of the ortho-f ((−)-53) with respect to that of (−)-7, making the proton transfer unfavorable. This may be one of the reasons why the ortho-f ((−)-53) compound does not show much agonist activity. Similarly, the ortho-e enantiomer could form a seven-membered ring using the two oxygen atoms, and this seven-membered ring would appear in the vicinity of the putative proton transfer pocket, and thus likely to disrupt the proton transfer process that was hypothesized to confer agonist activity to (−)-7.¹ In contrast, the phenolic hydroxyl of both the parae and para-f compounds ((−)-20 and (−)-26) is not positioned to form a seven-membered ring with the bridged oxygen atom (Fig. 4). However, the oxygen atom in the five-membered ring of the para-e isomer is only 0.35 Å away from the 9S-OH oxygen of (−)-7 suggesting that this particular oxygen might play a similar role to the 9S-OH oxygen by forming a seven-membered ring with the protonated nitrogen via a water molecule, and this may be a rationale for the morphine-like agonist potency of the para-e isomer (−)-20. Unlike the para-e isomer, the bridged oxygen of the para-f isomer (−)-26 is neither positioned to hydrogen-bond with the tertiary nitrogen via a water molecule (Fig. 4) nor to form a seven membered ring like ortho-f isomer that might alter the orientation of the protonated nitrogen, and this could be related to the much more modest agonist activity of the para-f isomer seen in the [³⁵S]GTP-γ-S assay (Table 3).

Fig. 3.

Two views of the oxide-bridged phenylmorphan ortho-e (1S,4aS,9aS-(−)-43) and ortho–f (1R,4aR,9aR-(−)-53) enantiomers overlapped with (1R,5R,9S)-(−)-7 (9S-OH) and (1R,5R,9R)-(+)-8 (9R-OH). Atoms are represented by colors as follows: white, hydrogen; green, carbon; blue, nitrogen; red, oxygen.

The binding affinity of (−)-7 (Ki = 0.19 nM) to the μ-receptor is ca. 300 times better than the affinity of its epimer, (+)-8 (Ki = 59 nM).¹ While with less affinity, a similar trend is observed for the para-e ((−)-20, Ki = 23 nM) and the para-f ((−)-26, Ki = 98 nM) isomers suggesting that the formation of the seven-membered ring between the bridged oxygen and the nitrogen atom, in a spatial area comparable to that occupied by the 9S-OH in (−)-7 (Fig. 3), enhances the binding energy to the μ-receptor via H-bonding interactions. The bridged oxygen of the ortho-e ((−)-43), however, is only 0.35 Å away from the 9S-OH oxygen of (−)-7 but its binding affinity (Ki = 2488 nM) to the μ-receptor is ca. 13,000 times less than (−)-7 suggesting that the presence of the ortho-oriented phenolic hydroxyl is detrimental to binding, possibly because of interference with the formation of the 7-membered ring between an oxygen and nitrogen atom, as noted above. Interestingly, the binding affinity of the ortho-f isomer ((−)-53, Ki = 7 nM) is about 8 times better than (+)-8 (Fig. 1) indicating that an ortho-oriented phenolic hydroxyl may enhance the interaction energy via H-bonding interaction possibly with a polar residue at the active site. This idea is consistent with the fact that the ortho-f isomer (−)-53 binds 14 times better to the μ-receptor than the para-f isomer-(−)-26.

While there are likely to be multiple mechanisms through which the types of ligands discussed in this and our previous paper (ligands with a phenylmorphan-like structure)¹ interact with the μ-receptor, we have proposed the idea of a proton transfer from the protonated nitrogen to a proton acceptor in the μ-opioid receptor via a water molecule(s) as one of the keys for imparting agonist activity. Also, the idea of the possible formation of a seven-membered ring with a putative water molecule may help relate the affinity and activity of the limited group of molecules in Table 5. Nonetheless, many more compounds still need to be examined before any definitive statement can be made about the plausibility of these ideas.

Conclusions

The N-phenethyl (−)- para-e enantiomer (1S,4aS,9aR-(−)-20) was found to be a μ-opioid agonist and was as potent as morphine as an antinociceptive. The N-phenethyl (−)-ortho-f enantiomer (1R,4aR,9aR-(−)-53) had good affinity for the μ-opioid receptor (Ki = 7 nM), and was found to be a μ-opioid antagonist in the [³⁵S]GTP-γ-S assay and in vivo. The epimeric difference between the rigid oxide -bridged e- and f-enantiomers gave them different fixed spatial three-dimensional patterns and this resulted in major differences in their pharmacological activity. Their spatial patterns are presumably relevant to those necessary for recognition by the μ-opioid receptor as agonists or antagonists. The superposition study suggests that i) the bridged oxygen of the para-e enantiomer and the 9S-OH of compound 7 are likely to play similar roles in forming a seven membered ring to facilitate proton transfer between the nitrogen atom and the oxygen atom of Asp that is hypothesized¹ to give rise to the μ-agonist activity and ii) while the hydroxyl of the ortho-f enantiomer enhances the binding affinity via H-bonding interactions with a polar residue at the binding site, it may affect the alignment of the protonated nitrogen, making the proton transfer process unfavorable. Even with the few molecules that have been examined thus far, insight gained from this mechanism can lead to new types of ligands and these might shed further light on the hypothesis. The synthesis of these ligands is currently in progress.

Experimental Section

Dry solvents were purchased from Aldrich and used without further purification. Compounds 12 and 15 were prepared according to published procedures.³ All melting points were determined on a Thomas Hoover apparatus and are uncorrected. NMR spectra were obtained on a Varian Gemini 300 spectrometer in CDCl₃, unless otherwise stated, with 0.1% v/v TMS as an internal standard (δ values in ppm, J (Hz) assignments of ¹H resonance coupling). HRMS were obtained on the racemates with a Waters/Micromass LCT ESI-TOF mass spectrometer at NIDDK. Dry distillation was carried out in a Buchi distillation oven (Kugelrohr). Thin layer chromatography (TLC) was performed on 250 μm Analtech GHLF silica gel plates using a CMA (CHCl₃:CH₃OH:concd NH₄OH (90:9:1)) solvent system. Enantiomeric purity was assessed on Shimadzu LC-6A HPLC with Shimadzu SPD-6AV UV detector (at 254 nm)) using a Daicel Chiralcel OD column (4.6 mm × 50 mm coupled to 4.6 mm × 250 mm). The mobile phase was hexanes:2-propanol:Et₂NH (90:10:0.3) at the constant flow rate of 1 mL/min. Preparative chiral separations were carried out on Daicel’s Chiralcel OD column (20 × 50 mm Guard plus 20 × 250 mm) using isocratic elution with hexanes:2-propanol:Et₂NH (80:20:0.2) at constant flow rate of 8 mL/min, detection at 254 nm. Racemates 43 and 53 were injected as solutions in 2-propanol (150 mg/1.5 mL, three runs, 500 μL each injection). Elemental analyses were performed on the racemic compounds by Atlantic Microlabs, Inc., Norcross, Ga.

Resolution of (1R*,4aR*,9aS*)-2-Methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propano-benzofuro[2,3-c]pyridine (15)

A boiling solution of [(1S)-(endo, anti)]-(−)-3-bromocamphor-8-sulfonic acid (5.60 g, 18 mmol, 1.05 equiv) in acetone (60 mL) was slowly added into a refluxing stirred solution of racemic amine 15 (4.70 g, 17 mmol) in acetone (100 mL). The reaction mixture was allowed to cool to ambient temperature and stirred overnight. After cooling in an ice bath, the yellow crystals of salt of (+)-15 that formed were filtered off and washed with cold acetone and air-dried (4.58 g, mp 269–270 °C). One recrystallization from the mixture of acetone-MeOH gave 2.48 g of yellow crystals (+)-15 salt (mp 275-276 °C), that were free based (concd NH₄OH/CHCl₃) to furnish (+)-15 base (1.06 g, [α]_D²⁶ +116.9 (c 1.0, CHCl₃)). A second crop was also obtained (0.98 g). The mother liquors were free-based to give the enriched (−)-enantiomer of 15, which was dissolved in boiling acetone (50 mL) and a boiling solution of [(1R)-(endo, anti)]-(+)-3-bromocamphor-8-sulfonic acid (2.73 g, 8.77 mmol, 1.05 equiv) in acetone (50 mL) was added portion-wise. The reaction mixture was allowed to cool to room temperature and stirred overnight. Crystals were collected after further cooling in an ice bath and washed with cold acetone to give the salt of (−)-15 (3.89 g). One recrystallization from an acetone-MeOH mixture gave 2.15 g of yellow crystals (mp 276–277 °C). This material was free based (concd NH₄OH/CHCl₃) to furnish (−)-15 base (0.984 g, [α]_D²³ -116.2 (c 1.0, CHCl₃)). A second crop was also obtained (0.97 g, mp 275–277 °C). The enantiomeric purity was assessed by NMR using S-(+)-1-phenyl-2,2,2-trifluoroethanol as a chiral shift reagent to be >99% e.e. The presence of the other enantiomer was not detected. The absolute stereochemistry of the (+)-enantiomer was determined to be 1R,4aR,9aS by X-ray crystallographic analysis of the salt of (+)-15.

(1R*,4aR*,9aS*)-2-(6-Nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]-pyridyl)-ethylcarboxylate (16)

In a flame-dried glassware apparatus cooled to room temperature under an argon atmosphere, 15 (1.00 g, 3.64 mmol) was dissolved in dry 1,2-dichloroethane (20 mL), and K₂CO₃ (1.01 g, 7.28 mmol, 2.0 equiv) was added followed by a slow addition of ethyl chloroformate (1.74 mL, 1.98 g, 18.24 mmol, 5.0 equiv). The reaction mixture was heated to reflux under an argon atmosphere for 1 day. Upon completion (TLC check) the reaction mixture was cooled to room temperature, diluted with H₂O (30 mL), and extracted with CHCl₃ (3 × 50 mL). The organic extracts were dried over anhydrous Na₂SO₄, filtered and evaporated. The resulting oil was redissolved in hot isopropanol (12 mL + 3 drops of H₂O), and the solution was boiled down to ca. 5 mL and cooled to room temperature and placed in the refrigerator overnight. The resulting off-white crystals were filtered and dried in vacuo (0.98 g, 81%). Mp ((±)-16) 143–145.0 °C (i-PrOH). (1S,4aS,9aR-(−)-16): [α]_D²⁵ -186.4 (c 1.0, CHCl₃), (1R,4aR,9aS-(+)-16): [α]_D²⁴ +187.0 (c 1.0, CHCl₃). ¹H-NMR: δ 8.14 (dd, 1H, J = 8.7, 2.4), 8.01 (d, 1H, J = 2.4), 6.95 (d, 1H, J = 8.7), 4.86, 4.75 (bs, 1H), 4.21 (m, 3H), 4.04 (m, 1H), 3.62 (m, 2H), 2.36-1.50 (m, 8H), 1.31 (m, 3H). ¹³C-NMR: δ 164.3, 156.3, 142.8, 140.1, 125.7, 118.1, 111.2, 90.9, 90.8, 61.6, 48.8, 40.9, 39.1, 38.8, 34.0, 33.9, 33.1, 30.4, 29.8, 18.9, 14.9, 14.8. HRMS: [M-H]⁺ calc. C₁₇H₂₁N₂O₅: 333.1450, found 333.1443. C₁₇H₂₀N₂O₅ requires: C, 61.44; H, 6.07; N, 8.43; found: C, 61.38; H, 6.22; N, 8.32%.

(1R*,4aR*,9aS*)-6-Nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (17)

Carbamate 16 (1.145 g, 3.45 mmol) was dissolved in 33% HBr/AcOH (20 mL) at room temperature, the resulting bright red solution was heated in an oil bath to 50 °C for 18 h to complete the reaction. After cooling to room temperature the reaction mixture was added dropwise into stirred mixture of concd NH₄OH (75 mL) and ice (50 mL), the yellow free base precipitated out. The precipitate was filtered off, washed with H₂O (2 × 20 mL) and air-dried to give a yellow amorphous solid. The aqueous layer was extracted with CHCl₃ (3 × 50 mL). Combined extracts were used to dissolve the precipitate and the solution was dried over Na₂SO₄. After filtration and the removal of solvent under reduced pressure, the residue was purified by column chromatography (silica, 2:1 v/v CHCl₃:CMA) to give 17 as a yellow solid (0.733 g, 82%). Optical rotations: (1S,4aS,9aR)-(−)-17): [α]_D²⁴ –149.3 (c 1.0, CHCl₃), (1R,4aR,9aS-(+)-17): [α]_D²² +151.2 (c 1.0, CHCl₃). ¹H-NMR: δ 8.14 (dd, 1H, J = 8.7, 2.7), 7.96 (d, 1H, J = 2.7), 6.96 (d, 1H, J = 8.7), 4.06 (d, 1H, J = 2.1), 3.69 (m, 1H), 3.30 (ddd, 1H, J = 13.2, 12.9, 5.1), 2.86 (dd, 1H, J = 13.5, 7.2), 2.43 (dd, 1H, J = 12.9, 5.7), 2.13 (m, 2H), 2.00-1.70 (m, 6H), 1.53-1.38 (m, 1H); ¹³C-NMR: δ 165.0, 142.4, 141.4, 125.6, 117.8, 111.0, 89.7, 48.8, 40.9, 40.6, 34.1, 33.1, 31.2, 21.8. HRMS: [MH]⁺ calcd C₁₄H₁₇N₂O₃: 261.1239, found 261.1246. C₁₄H₁₆N₂O₃•0.25H₂O requires: C, 63.50; H, 6.28; N, 10.56; found: C, 63.74; H, 6.17; N, 10.56%.

(1R*,4aR*,9aS*)-6-Nitro-2-(2-phenethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (18)

N-nor amine 17 (660 mg, 2.54 mmol) was dissolved in dry acetonitrile (25 mL), finely ground K₂CO₃ (651 mg, 5.08 mmol, 2.0 equiv) and KI (105 mg, 0.64 mmol, 0.25 equiv) were added followed by (2-bromoethyl)benzene (429 μL, 3.17 mmol, 1.25 equiv, 587 mg). The reaction mixture was stirred and heated to reflux under an argon atmosphere for three days. After cooling to room temperature, the reaction mixture was diluted with H₂O (50 mL) and extracted with CHCl₃ (3 × 50 mL). The organic extracts were dried over Na₂SO₄, filtered and evaporated with small amount of silica. The crude mixture was then purified by column chromatography (silica, 85:15 hexanes:EtOAc). The product 18 was obtained as a yellow solid (0.860 g, 93%). Optical rotations: (1S,4aS,9aR)-(−)-18: [α]_D²⁴ -133.4 (c 1.0, CHCl₃), (mp 90–92 °C (2-propanol:H₂O), (1R,4aR,9aS)-(+)-18): [α]_D²⁴ +133.9 (c 1.0, CHCl₃) (mp 88–90 °C (2-propanol:H₂O). ¹H-NMR: δ 8.13 (dd, 1H, J = 8.7, 2.4), 7.99 (d, 1H, J = 2.4), 7.26 (m, 5H), 6.97 (d, 1H, J = 9.0), 4.23 (d, 1H, J = 3.0), 3.67 (m, 1H), 2.88 (m, 6H), 2.41-2.28 (m, 2H), 2.10-1.70 (m, 5 H), 1.50 (m, 1H). ¹³C-NMR: δ 164.9, 142.4, 141.0, 140.4, 128.9, 128.6, 126.3, 125.6, 118.0, 111.4, 91.6, 58.1, 54.0, 47.9, 40.6, 34.6, 34.0, 33.3, 24.9, 22.0. HRMS: [MH]⁺ calcd C₂₂H₂₅N₂O₃: 365.1865, found 365.1863. C₂₂H₂₄N₂O₃ requires: C, 72.50; H, 6.647; N, 7.69; found: C, 72.20; H, 6.58; N, 7.68%.

(1R*,4aR*,9aS*)-6-Amino-2-(2-phenethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (19)

10% Pd-C (0.100 g, 5 mol%) was added to a solution of nitro derivative 18 (700 mg, 1.92 mmol) in abs EtOH (120 mL). The reaction mixture was shaken in a H₂ atmosphere (45 psi) for 7 h. Catalyst was filtered on Celite pad and washed with EtOH (3 × 30 mL). The filtrate was evaporated and purified by column chromatography (silica, 7:1 CHCl₃:CMA) to give product 19 as yellow solid (0.606 g, 94%). Mp ((± )-19) 120–123 °C (2-propanol:H₂O). Optical rotations: (1S,4aS,9aR)-(−)-19: [α]_D²⁵ –59.7 (c 1.0, CHCl₃), (1R,4aR,9aS)-(+)-19: [α]_D²⁵ +58.9 (c 1.0, CHCl₃). ¹H-NMR: δ 7.26 (m, 5H), 6.73 (d, 1H, J = 8.4), 6.50 (d, 1H, J = 2.4), 6.45 (dd, 1H, J = 8.1, 2.4), 4.01 (d, 1H, J = 3.0), 3.61 (m, 1H), 3.41 (bs, 2H), 2.87 (m, 6H), 2.26 (m, 2H), 2.10-1.70 (m, 5 H), 1.50 (m, 1H). ¹³C-NMR: δ 152.0, 140.7, 140.4, 128.9, 128.5, 126.1, 114.0, 111.5, 109.6, 89.5, 58.3, 54.2, 48.2, 34.6, 34.0, 33.5, 25.1, 22.0. HRMS: [MH]⁺ calc. C₂₂H₂₇N₂O: 335.2123, found 335.2146. C₂₂H₂₆N₂O•0.25H₂O requires: C, 77.96; H, 7.88; N, 8.26; found: C, 77.58; H, 7.83; N, 8.07%.

(1R*,4aR*,9aS*)-6-Hydroxy-2-(2-phenethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (20)

The aniline derivative 19 (250 mg, 0.747 mmol) was dissolved in cold 35% wt H₂SO₄ (12 mL) and the resulting red solution was cooled to −3 °C. A cold solution of NaNO₂ (57 mg, 0.823 mmol, 1.1 equiv) in H₂O (1 mL) was added under the surface of the stirred reaction mixture at a rate that kept the temperature below 0 °C. The reaction mixture was stirred at less than 0 °C for 1 h. Cold diazonium salt solution was then added dropwise into a vigorously stirred boiling H₂SO₄ solution (prepared by mixing concd H₂SO₄ with H₂O (20 mL each), bp ca. 135 °C) under argon atmosphere. The resulting red solution was refluxed for another ten min, cooled to room temperature, and the solution was added dropwise into a cold NH₄OH (50 mL of concd NH₄OH/ca. 75 mL of ice) with vigorous stirring keeping the temperature below 30 °C at all times. The resulting alkaline suspension was extracted with CHCl₃ (5 × 100 mL), and the combined organic extracts were dried over Na₂SO₄. The drying agent was removed by filtration and the filtrate was evaporated with a small amount of silica. Purification of the preadsorbed material by chromatography (silica, 4:1 CHCl₃:CMA) gave 20 as a pink foam (0.205 g, 82%). The oxalate salt of 20 was prepared. Mp ((±)-20•oxalate) 205–207 °C (acetone). Optical rotations: (1S,4aS,9aR)-(−)-20•oxalate: [α]_D²⁴ –64.1 (c 1.0, 90% EtOH), (1R,4aR,9aS)-(+)-20•oxalate: [α]_D²⁴ +65.3 (c 1.0, 90% EtOH). Optical rotations (free bases - oils): (1S,4aS,9aR)-(−)-20: [α]_D²⁵ –51.7 (c 1.0, CHCl₃), (1R,4aR,9aS) )-(+)-20: [α]_D²⁵ +40.3 (c 1.0, CHCl₃). ¹H-NMR: δ 7.31-7.17 (m, 5H), 6.74 (d, 1H, J = 8.1), 6.62 (d, 1H, J = 2.7), 6.57 (dd, 1H, J = 8.1, 2.7), 4.04 (d, 1H, J = 3.0), 3.62 (m, 1H), 2.86 (m, 6H), 2.27 (m, 2H), 2.00-1.70 (m, 5 H), 1.47 (m, 1H). ¹³C-NMR: δ 152.9, 150.3, 141.0, 140.7, 129.0, 128.6, 126.2, 113.9, 111.5, 109.5, 89.9, 58.4, 54.3, 48.2, 41.0, 34.5, 34.0, 33.5, 25.2, 22.0. HRMS: [MH]⁺ calcd C₂₂H₂₆NO₂: 336.1963, found 336.1969. C₂₂H₂₅NO₂•C₂H₂O₄ requires: C, 67.75; H, 6.40; N, 3.29; found: C, 67.46; H, 6.30; N, 3.31%.

(1R*,4aR*,9aR*)-2-Methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (21)

A solution of 29 (2.00 g, 6.8 mmol) in dry THF (50 mL) was slowly added into refluxing suspension of NaH (0.408 g, 10.2 mmol, 1.5 equiv, 60% oil suspension) in dry THF (20 mL). The reaction mixture was refluxed under an argon atmosphere for 4 h. After cooling to room temperature, the reaction mixture was slowly quenched with ice and poured into cold H₂O (100 mL). Most of the THF was removed under reduced pressure and the aqueous solution was extracted with CHCl₃ (5 × 50 mL). The organic extracts were dried over Na₂SO₄, filtered and evaporated and purified by column chromatography (silica, 4:1 CHCl₃:CMA) to give 1.43 g (77 %) of 21. Mp 173-174.5°C (2-propanol:H₂O). Optical rotations: (1S,4aS,9aS)-(+)-21: [α]_D²⁴ +114.8 (c 1.0, CHCl₃) (from the (−)alcohol), (1R,4aR,9aR)-(−)-21: [α]_D²⁴ –111.9 (c 1.0, CHCl₃) (from the (+)-alcohol). ¹H-NMR: 8.14 (dd, 1H, J = 9.0, 2.7), 7.97 (d, 1H, J = 2.7), 6.95 (d, 1H, J = 9.0), 4.28 (d, 1H, J = 3.9), 3.40 (m, 1H), 3.11 (ddd, 1H, J = 18.0, 12.6, 6.0), 2.92 (dd, 1H, J = 12.6, 6.9), 2.55 (s, 3H), 2.30 (dd, 1H, J = 12.9, 6.7), 2.24-2.13 (m, 2H), 1.78-1.43 (m, 5H). ¹³C-NMR: 165.0, 142.4, 144.4, 125.4, 117.8, 111.0, 90.5, 54.8, 50.5, 42.4, 40.8, 33.7, 32.2, 20.7, 20.3. HRMS: [MH]⁺ calcd C₁₅H₁₉N₂O₃: 275.1383, found 275.1396. C₁₅H₁₈N₂O₃ requires: C, 65.68; H, 6.61; N, 10.21; found: C, 65.46; H, 6.84; N, 10.03%.

Experimental data for compounds 22 through 30 can be found in the Supplemental section.

(1R*,4aR*,9aR*)-5-(2-Fluoro-5-nitrophenyl)-2-methyl-2-azabicyclo[3.3.1]nonan-9α-ol (29)

Alcohol 31 (13.50 g, 54.2 mmol) was dissolved in cold 70% HNO₃ (135 mL). The stirred solution was warmed to room temperature and heated at 58–60 °C for 4 h. After cooling to room temperature the reaction mixture was added dropwise into a stirred cold NH₄OH solution (300 mL of concd NH₄OH/100 mL of ice, cooled in ice bath). The free base 29 precipitated out as a light yellow solid. The precipitate was collected and washed with cold H₂O (ca. 50 mL) and recrystallized from 2-propanol:H₂O (300 mL/40 mL, boiled down to ca. 100 mL), to give two crops of pure 29 (total 12.29 g, 77%). Note: Reaction progress must be carefully monitored (mini work-up, ¹H-NMR is needed, since the starting material co-elutes with the product). Overly long heating results in appreciable formation of various by-products and substantially lowers yield. Mp ((± )-29) 175–177 °C (EtOAc). Resolved using O,O′-di-p-toluoyltartaric acids (acetone-MeOH 1:1). The absolute stereochemistry of the (+)-enantiomer was found by single crystal X-ray analysis of the diastereomeric salt of (+)-29 to be 1R,4aR,9aR. Optical rotations: (1R,4aR,9aR) )-(+)-29: [α]_D²⁴ +41.8 (c 1.0, CHCl₃) (from the salt with (−)-L-acid), (1S,4aS,9aS)-(−)-29: [α]_D²⁴ –42.0 (c 1.0, CHCl₃) (from the salt with (+)-D-acid). ¹H-NMR: 8.49 (dd, 1H, J = 6.6, 2.7), 8.09 (m, 1H), 7.12 (dd, 1H, J = 12.0, 9.0), 4.52 (d, 1H, J = 3.9), 3.02 (ddd, 1H, J = 12.0, 11.9, 5.1), 2.93 (bs, 1H), 2.75 (m, 1H), 2.49 (s, 3H), 2.21-1.73 (m, 8H). ¹³C-NMR: 167.4, 164.0, 144.5, 137.0, 136.8, 125.5, 125.4, 124.0, 123.9, 117.9, 117.6, 71.1, 60.0, 50.4, 43.2, 41.0, 40.9, 35.3, 35.2, 28.3, 28.2, 21.7, 18.1. HRMS: [MH]⁺ calcd C₁₅H₂₀N₂O₃F: 295.1458, found 295.1441. C₁₅H₁₉FN₂O₃ requires: C, 61.21; H, 6.51; N, 9.52; found: C, 0.99; H, 6.55; N, 9.53%.

Resolution protocol

A hot MeOH solution of (+)-O,O′-di-p-toluoyl-D-tartaric acid (16.30 g, 42 mmol, 1.05 equiv in 150 mL) was slowly added to a hot stirred solution of racemic amine 29 (11.82 g, 40 mmol) in acetone (150 mL). The reaction mixture was allowed to cool to ambient temperature and then chilled in an ice bath. The white crystals of the salt of (−)-29 that formed were filtered and washed with cold acetone (10 mL) and air-dried (10.86 g, mp 153–155 °C). One recrystallization from acetone:MeOH:H₂O (155:155:15 mL, boiled down to ca. 300 mL) gave 8.69 g of white crystals (mp 157–158 °C). This salt was converted to the free base (concd NH₄OH/CHCl₃) of (−)-29 (3.14 g, mp 205–207 °C (2-propanol:H₂O), [α]_D²⁴ – 39.8 (c 1.0, CHCl₃)). The second crop of the tartrate salt and mother liquor were pooled, converted to the free base and purified by column chromatography (silica, 1:1 CHCl₃:CMA) to give an additional 0.660 g of of (−)-29 as a white solid ([α]_D²³ –42.0 (c 1.0, CHCl₃). The mother liquor was converted to the free base to give the enriched (+) enantiomer (7.07 g of brown solid), which was dissolved in hot acetone (70 mL) and a hot solution of (−)-O,O′-di-p-toluoyl-L-tartaric acid (8.15g, 21 mmol, 1.05 equiv) in MeOH (70 mL) was added portionwise. The salt crystallized rapidly, and after cooling to room temperature and then in an ice bath, the crystals were collected and washed with cold acetone to give the salt of (+)-29 (10.67 g, mp 153–156 °C). One recrystallization from a mixture of acetone:MeOH:H₂O (150:150:30 mL, boiled down to ca. 200 mL) gave 9.58 g of a white crystalline salt (mp 157.5–159 °C). This salt was converted to the free base (concd NH₄OH/CHCl₃) to give (+)-29 (3.46 g, mp 205–207 °C (2-propanol:H₂O), [α]_D²⁴ +40.6 (c 1.0, CHCl₃)). The second crop of (+)-29 salt and mother liquor from recrystallization were pooled, converted to the free base and purified by column chromatography (silica, 1:1 CHCl₃:CMA) to give 0.382 g of white solid (+)-29 ([α]_D²³ +41.8 (c 1.0, CHCl₃). The retention times of (+)-25 and (−)-29 were 6.58 and 10.44 min, respectively. The presence of any other enantiomer in the final free base products was not detected under those conditions.

(1S*,4aS*)-5-(2-Fluorophenyl)-2-methyl-2-azabicyclo[3.3.1]nonan-9α-ol (31)

Ketone 30 (12.16 g, 49.2 mmol) was dissolved in MeOH (120 mL) and PtO₂ (335 mg, 1.48 mmol, 0.03 equiv) was added. The reaction mixture was hydrogenated in a Parr-Shaker apparatus at ca. 45 psi for 20 h at ambient temperature. The reaction mixture was filtered through a Celite column and the catalyst was washed with additional MeOH (ca. 3 × 40 mL). The filtrate was evaporated to dryness and further dried in vacuo to give 31 (11.50 g, 94%). ¹H-NMR: 7.45 (td, 1H, J = 8.1, 2.1), 7.19 (m, 1H), 7.21 (m, 1H), 7.10 (bt, 1H, J = 7.5), 7.02 (ddd, 1H, J = 17.1, 7.8, 1.2), 4.54 (d, 1H, J = 3.9), 2.98 (m, 2H), 2.75 (m, 1H), 2.49 (s, 3H), 2.15 (m, 4H), 1.98-1.70 (m, 6H); ¹³C-NMR: 163.7, 160.4, 134.8, 134.7, 128.5, 128.4, 128.1, 127.9, 124.3, 124.2, 117.0, 116.7, 71.18, 71.1, 59.7, 50.6, 43.2, 40.5, 40.4, 35.7, 35.6, 28.5, 28.4, 21.8, 18.2. HRMS: [MH]⁺ calcd C₁₅H₂₁NOF: 250.1607, found 250.1627. C₁₅H₂₀FNO requires: C, 72.26; H, 8.09; N, 5.62; found: C, 72.00; H, 8.06; N, 5.62%.

(1R*,4aR*,9aR*)-6-Amino-2-methyl-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (32)

Nitro derivative 21 (2.85 g, 10.4 mmol) was dissolved in mixture of MeOH:2-propanol (100 mL:100 mL)), and the solution was purged with a gentle argon stream. The catalyst (10% Pd-C, 0.553 g) was then added portion-wise. The reaction mixture was hydrogenated in a Parr-Shaker apparatus at room temperature and 50 psi of hydrogen pressure for 3 h. The catalyst was filtered on a Celite bed, which was washed with MeOH (3 × 30 mL). The combined filtrates were evaporated and the residue was recrystallized from 2-propanol:H₂O to give 32 (2.16 g, 85%) as pink crystals. Mp ((± )-32) 203–204 °C. ¹H-NMR: 6.71 (d, 1H, J = 8.1), 6.49 (d, 1H, J = 2.7), 6.46 (dd, 1H, J = 8.4, 2.7), 4.07 (d, 1H, J = 3.3), 3.42 (bs, 2H), 3.33 (m, 1H), 3.06 (ddd, 1H, J = 11.7, 11.7, 6.3), 2.87 (ddd, 1H, J = 12.0, 6.6, 1.5), 2.53 (s, 3H), 2.21-2.07 (m, 3H), 1.77-1.45 (m, 5H). ¹³C-NMR: 152.3, 141.3, 140.5, 113.9, 111.2, 109.7, 88.7, 55.4, 50.9, 42.5, 41.1, 33.8, 32.6, 21.1, 20.7. HRMS: [MH]⁺ calcd C₁₅H₂₁N₂O: 245.1654, found 245.1666. C₁₅H₂₀N₂O requires: C, 73.74; H, 8.25; N, 11.47; found: C, 73.80; H, 8.31; N, 11.44%.

Experimental data for compounds 30 and 33 through 34 can be found in the Supplemental section.

(1R*,4aR*,9aS*)-8-Bromo-2-methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (35)

Bromide 33 (1.10 g, 3.1 mmol) was dissolved in 70% HNO₃ (10 mL) and the solution was heated to reflux for 2 h. After cooling to room temperature the reaction mixture was poured on ice and the pH was adjusted to ca. 9 with concd NH₄OH, followed by extraction with CHCl₃. The combined extracts were dried over Na₂SO₄, filtered and evaporated. ¹H-NMR spectra of the crude mixture revealed the presence of two products, i.e. 35 and 15 in ca. 6:4 ratio. Compound 35 was isolated by column chromatography to give 0.55 g (44 %) of solid. Recrystallization from 2-propanol gave light yellow crystals, mp 128–129 °C. ¹H-NMR: 8.31 (d, 1H, J = 2.2), 7.91 (d, 1H, J = 2.2), 4.30 (d, 1H, J = 2.7), 3.55 (m, 1H), 2.89-2.70 (m, 2H), 2.53 (s, 3H), 2.45-2.31 (m, 2H), 1.99-1.88 (m, 2H), 1.86-1.78 (m, 3H), 1.49-1.40 (m, 1H). HRMS (FAB): [MH]⁺ calcd C₁₅H₁₈BrN₂O₃: 353.0501, found 353.0493, and 355.0480, found 355.0492. C₁₅H₁₇BrN₂O₃ requires: C, 51.01; H, 4.85; N, 7.75; found: C, 50.93; H, 4.94; N, 7.73%.

Experimental data for compounds 36 through 46 can be found in the Supplemental section.

(1R*,4aR*,9aR*)-6-Chloro-2-methyl-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (47)

The aniline derivative 32 (650 mg, 2.66 mmol) was dissolved in cold concd HCl (15 mL). The resulting yellow solution was cooled to −5 °C. A cold solution of NaNO₂ (208 mg, 2.93 mmol, 1.10 equiv) was then added dropwise under the surface of the stirred reaction mixture at a rate that maintained the temperature of the reaction mixture below 0 °C. The purple reaction mixture was stirred at less than 0 °C for 75 min, while a solution of copper (II) sulfate (99%) (489 mg, 3.06 mmol, 1.15 equiv), and NaCl (649 mg, 11.10 mmol, 4.17 equiv) in H₂O (10 mL) was heated to 70–75 °C. Addition of a solution of sodium bisulfite (167 mg, 0.88 mmol, 0.33 mmol) and NaOH (112 mg, 2.79 mmol, 1.05 equiv) in H₂O (10 mL) to the copper (II) sulfate solution and subsequent short heating (15 min) generated copper (I) chloride. Cold diazonium salt solution was then added dropwise and the heating continued for 1 h. Upon cessation of nitrogen evolution the reaction mixture was stirred at room temperature overnight. The resulting suspension was added dropwise into a mixture of concd NH₄OH and ice (50 mL and ca. 20 g, respectively). The free base precipitated and was extracted with CHCl₃ (4 x100 mL). The combined organic extracts were dried over Na₂SO₄, filtered, and the solvent removed under reduced pressure. The residue was purified by column chromatography (silica, 5:1 CHCl₃:CMA) to give the product 47 as a white solid (0.497 g, 71%), mp ((±)-47) 96–98 °C (2-propanol:H₂O). ¹H-NMR: 7.08 (dd, 1H, J = 8.4, 2.4), 7.03 (d, 1H, J = 2.7), 6.81 (d, 1H, J = 8.4), 4.14 (d, 1H, J = 3.9), 3.35 (m, 1H), 3.07 (ddd, 1H, J = 12.3, 12.0, 2.7), 2.88 (dd, 1H, J = 12.6, 6.9), 2.54 (s, 3H), 2.34-2.08 (m, 3H), 1.75-1.42 (m, 5H). ¹³C-NMR: 158.1, 142.1, 127.5, 125.9, 112.1, 89.4, 60.1, 55.8, 50.8, 42.5, 41.4, 33.8, 32.5, 21.0, 20.6. HRMS: [MH]⁺ calcd C₁₅H₁₉ClNO: 264.1155, found 264.1156. C₁₅H₁₉ClNO requires: C, 68.30; H, 6.88; N, 5.31; found: C, 68.50; H, 6.94; N, 5.30%.

(1R*,4aR*,9aR*)-6-Chloro-2-methyl-8-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (48)

Chloride 47 (1.43 g, 5.4 mmol) was dissolved in trifluoroacetic acid (50 mL) and NaNO₂ (2.63 g, 38 mmol, 7.0 equiv) was added in two portions (NO_x evolves vigorously). The dark brown reaction mixture was then stirred at room temperature for 2 h and the progress of the reaction was monitored by HPLC of worked up aliquots. The resulting clear orange solution was slowly added dropwise into a vigorously stirred cold NH₄OH (150 mL of concd NH₄OH with ca. 50 g of ice). The free base precipitated as a yellow powder, and after cooling it was collected by filtration and washed with cold H₂O (2 × 10 mL). The filter cake was dried overnight in vacuum oven (25 psi/55 °C) to give 48 as a yellow solid (1.62 g, 98%). Two recrystallizations from 2-propanol:H₂O were necessary to remove the traces of a by-product, (presumably the 7-nitro isomer). Pure compound was obtained as light yellow crystals (1.07 g, 74%). Mp ((±)-48) 126–128 °C (2-propanol:H₂O). ¹H-NMR: 7.92 (d, 1H, J = 1.8), 7.26 (d, 1H, J = 1.8), 4.35 (d, 1H, J = 3.6), 3.49 (m, 1H), 3.10 (ddd, 1H, J = 12.3, 12.2, 6.0), 2.90 (dd, 1H, J = 12.6, 7.5), 2.55 (s, 3H), 2.29-2.10 (m, 3H), 1.79-1.45 (m, 5H). ¹³C-NMR: 153.4, 146.4, 133.9, 127.5, 126.1, 122.9, 91.2, 54.6, 50.3, 42.4, 41.1, 33.6, 32.2, 20.7, 20.1. HRMS: [MH]⁺ calcd C₁₅H₁₈ClN₂O₃: 309.1006, found 309.0990. C₁₅H₁₇ClN₂O₅ requires: C, 58.35; H, 5.55; N, 9.07; found: C, 58.47; H, 5.67; N, 9.05%.

Experimental data for compounds 49 through 51 can be found in the Supplemental section.

(1R*,4aR*,9aR*)-8-Amino-2-(2-phenylethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (52)

Compound 51 (500 mg, 1.25 mmol) and NaOAc•3H₂O (1.656 g, 0.0125 mol, 10 equiv) were dissolved in hot EtOH (100 mL). The solution was cooled to room temperature and then placed in an ice bath and purged with a stream of argon. The catalyst (10% Pd-C, Engelhardt C3645, 0.200 g, 15 mol% of Pd) was carefully added and the reaction mixture was hydrogenated in a Parr-Shaker apparatus at 45–50 psig and ca. 45–50 °C for 3 h, then at room temperature overnight. Reaction progress was monitored by HPLC. The reaction mixture was filtered through a Celite bed, which was washed with EtOH (3 × 10 mL). The filtrate was evaporated to dryness and the residue was converted to the free base by addition of concd NH₄OH (50 mL) and extracted with CHCl₃ (5 × 50 mL). The combined extracts were dried over Na₂SO₄, filtered, evaporated with a small amount of silica and purified by column chromatography (silica, 5:1 CHCl₃:CMA to give 52 as a colorless viscous oil (0.318 g, 76%). ¹H-NMR: 7.32-7.20 (m, 4H), 6.73 (t, 1H, J = 7.5), 6.60-6.53 (m, 2H), 4.15 (d, 1H, J = 3.6), 3.60 (bs, 2H), 3.52 (m, 1H), 3.15-2.80 (m, 6H), 2.25-2.00 (m, 3H), 1.80-1.45 (m, 5H). ¹³C-NMR: 146.4, 140.5, 140.2, 131.7, 128.8, 128.3, 126.0, 121.6, 115.0, 111.3, 88.7, 56.8, 53.2, 49.2, 41.5, 34.7, 33.7, 32.7, 21.1, 21.0. HRMS: [MH]⁺ calcd C₂₂H₂₇N₂O: 335.2123, found 335.2104. C₂₂H₂₆N₂O requires: C, 79.00; H, 7.84; N, 8.28; found: C, 79.15; H, 7.91; N, 8.28%.

(1R*,4aR*,9aR*)-8-Hydroxy-2-(2-phenylethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine (53)

Amine 52 (112 mg, 0.335 mmol) was dissolved in 35% wt H₂SO₄ (14 mL) and the stirred solution was cooled to 3 °C. An aqueous solution of NaNO₂ (0.5 M, 1.0 mL, 0.502 mmol, 1.5 equiv) was added dropwise over 5 min. The reaction mixture was stirred at 3 °C for 4 h then urea (0.010 g, 0.5 equiv) was added. After an additional 5 min of stirring, a solution of copper (II) nitrate (45 mL, 1.5 M, 67.0 mmol, 200 equiv) was added and the reaction mixture was stirred for another 5 min in a cooling bath. Finally, copper (I) oxide (60 mg, 0.419 mmol, 1.25 equiv) was added and the reaction mixture was vigorously stirred at room temperature for 1 h, (as nitrogen gas evolved some foaming occurs). The pH was adjusted to 9 by addition of concd NH₄OH (40 mL) to the chilled reaction mixture. The resulting dark blue mixture was extracted with CHCl₃ (4 × 50 mL). The combined organic extracts were dried over Na₂SO₄, filtered and evaporated with a small amount of silica. The mixture was purified by column chromatography (silica, 4:1 CHCl₃:CMA to give 53 as a dark yellow oil (0.096 g, 86%). Further chromatographic purification gave a yellow oil (0.068g, 61%). ¹H-NMR: 7.33-7.18 (m, 5H), 6.80-6.74 (m, 2H), 6.63 (dd, 1H, J = 6.3, 2.4), 4.11 (d, 1H, J = 3.3), 3.72 (m, 1H), 3.07 (dd, 2H, J = 9.9, 3.3), 2.92 (m, 4H), 2.25-2.05 (m, 3H), 1.80-1.50 (m, 4H). ¹³C-NMR: 146.3, 142.2, 141.4, 140.3, 129.0, 128.6, 126.3, 122.3, 116.8, 113.0, 89.2, 56.7, 52.3, 50.0, 41.3, 33.9, 33.2, 32.9, 21.1, 20.7. HRMS: [MH]⁺ calcd C₂₂H₂₆NO₂: 336.1964, found 336.1945. C₂₂H₂₅NO₂•C₂H₂O₄ requires: C, 67.75; H, 6.40; N, 3.29; found: C, 67.93; H, 6.53; N, 3.49%.

Separation of the enantiomers

The racemic compound was prepared in larger quantity (ca. 0.150 g) and the enantiomers were separated on a chiral phase (Daicel’s Chiralcel OD column (20×50 mm Guard plus 20×250 mm) using isocratic elution with a hexane:2-propanol:Et₂NH (80:20:0.2) mixture (8 mL/min, detection at 254 nm). Racemate was injected as a solution in 2-propanol (150 mg/1.5 mL, three runs, 500 μL each). Retention times of the enantiomers were around 19.5 and 31 min, (+) and (−) respectively. Corresponding fractions were pooled together, evaporated and analyzed on an analytical column of the same type. The oily products were converted into corresponding oxalate salts in acetone. Optical rotations: (1S,4aS,9aS)-(+)-53 (base): [α]_D²³ +21.9 (c 1.0, CHCl₃), oxalate: [α]_D²³ +37.0 (c 1.0, 90% EtOH), (1R,4aR,9aR)-(−)-53 (base): [α]_D²³ -26.3 (c 1.0, CHCl₃), oxalate: [α]_D²³ –37.4 (c 1.0, 90% EtOH).

X-ray Crystal Structure of (1R,4aR,9aS)-2-Methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propano-benzofuro[2,3-c]pyridine ((+)-15, (1R*,4aR*,9aR*)-5-(2-fluoro-5-nitrophenyl)-2-methyl-2-azabicyclo[3.3.1]nonan-9α-ol (21), (1R,4aR,9aR)-5-(2-fluoro-5-nitrophenyl)-2-methyl-2-azabicyclo[3.3.1]nonan-9α-ol ((+)-29), and (1S,4aS,9aS)-1,3,4,9a-tetrahydro-8-hydroxy-2-(2-phenylethyl)-2H-1,4a-propanobenzofuro[2,3-c]pyridine•HBr ((+)-53•HBr)

Single-crystal X-ray diffraction data on compounds (+)-15, (+)-29, 21, and (+)-53•HBr (Fig. 2) were collected using MoKα radiation and a Bruker APEX 2 CCD area detector. The structures were solved by direct methods and refined by full-matrix least squares on F² values using the programs found in the SHELXTL suite (Bruker, SHELXTL v6.10, 2000, Bruker AXS Inc., Madison, WI). Parameters refined included atomic coordinates and anisotropic thermal parameters for all non-hydrogen atoms. Hydrogen atoms on carbons were included using a riding model [coordinate shifts of C applied to H atoms] with C-H distance set at 0.96 Å. Atomic coordinates for these compounds have been deposited with the Cambridge Crystallographic Data Centre (deposition numbers 656371, 656369, 656370, and 656372 for compounds (+)-15, 21, (+)-29, and (+)-53, respectively). Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK [fax: +44(0)-1223-336033 or e-mail: deposit@ccdc.cam.ac.uk

(1R,4aR,9aS)-2-Methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propano-benzofuro[2,3-c]pyridine ((+)-15•[(1S)-(endo,anti)]-(−)-3-bromocamphor-8-sulfonate

A 0.65 × 0.35 × 0.14 mm³ crystal of (+)-15 was prepared for data collection coating with high viscosity microscope oil (Paratone-N, Hampton Research). The oil-coated crystal was mounted on a glass rod and transferred immediately to the cold stream (93 °K) on the diffractometer. The crystal was orthorhombic in space group P2₁2₁2₁ with unit cell dimensions a = 7.719(3) Å, b = 13.922(5) Å, c = 24.899(12) Å. Corrections were applied for Lorentz, polarization, and absorption effects. Data were 91.8% complete to 28.51° θ (approximately 0.74 Å) with an average redundancy of 4.7. The absolute configuration was determined from the x-ray data and confirmed using a co-crystallized reference molecule ([(1S)-(endo,anti)]-(−)-3-bromocamphor-8-sulfonate).

(1R*,4aR*,9aR*)-5-(2-Fluoro-5-nitrophenyl)-2-methyl-2-azabicyclo[3.3.1]nonan-9α-ol (21)

A 0.64 × 0.36 × 0.28 mm³ crystal of 21 was prepared for data collection coating with high viscosity microscope oil (Paratone-N, Hampton Research). The oil-coated crystal was mounted on a glass rod and transferred immediately to the cold stream (93 °K) on the diffractometer. The crystal was orthorhombic in space group P2₁2₁2₁ with unit cell dimensions a = 7.4315(16) Å, b = 8.4430(18) Å, c = 21.576(5) Å. Corrections were applied for Lorentz, polarization, and absorption effects. Data were 97.7% complete to 28.05° θ (approximately 0.75 Å) with an average redundancy of 4.8. The relative stereochemistry (1R,4aR,9aR) was determined from the x-ray data.

(1R,4aR,9aR)-2-Methyl-6-nitro-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine ((+)-29)

A 0.50 × 0.46 × 0.14 mm³ crystal of (+)-29 was prepared for data collection coating with high viscosity microscope oil (Paratone-N, Hampton Research). The oil-coated crystal was mounted on a glass rod and transferred immediately to the cold stream (93 °K) on the diffractometer. The crystal was orthorhombic in space group P2₁2₁2₁ with unit cell dimensions a = 8.7028(11) Å, b = 13.9149(15) Å, c = 30.493(3) Å. Corrections were applied for Lorentz, polarization, and absorption effects. Data were 97.9% complete to 28.34° θ (approximately 0.75 Å) with an average redundancy of 5.7. The absolute configuration was set using a co-crystallized reference molecule (R,R-(−)-di-O,O′-p-toluoyl-D-tartaric acid).

(1S,4aS,9aS)-8-Hydroxy-2-(2-phenylethyl)-1,3,4,9a-tetrahydro-2H-1,4a-propanobenzofuro[2,3-c]pyridine•HBr ((+)-53•HBr)

A 0.84 × 0.54 × 0.30 mm³ crystal of (+)-53•HBr was mounted on a glass rod and transferred to the diffractometer and data collected at room temperature (298 °K) on the. The crystal was orthorhombic in space group P2₁2₁2₁ with unit cell dimensions a = 7.5366(2) Å, b = 11.9922(4) Å, c = 21.9804(11) Å. Corrections were applied for Lorentz, polarization, and absorption effects. Data were 97.3% complete to 30.32° θ (approximately 0.72 Å) with an average redundancy of 7.0. The absolute configuration was determined from the X-ray data.

Quantum Chemical Methods

Geometry optimization and energetic calculations for all the compounds in Table 5 have been conducted in the gaseous phase with the density functional theory at the level of B3LYP/6-31G*.²⁹ Superposition of these geometry-optimized structures was carried out using the rigid-fit of Quanta 2005 (Accelyrs).

Supplementary Material

arrays

Supporting Information Available: Atomic coordinates for 1R,4aR,9aS)-(+)-15, (1S*,4aS*,9aS-21, (1R,4aR,9aR)-(+)-29, and (1S,4aS,9aS)-(+)-53•HBr have been deposited with the Cambridge Crystallographic Data Centre (deposition numbers 656371, 656369, 656370, and 656372, respectively). Copies of the data can be obtained, free of charge, on application to CCDC, 12 Union Road, Cambridge, CB2 1EZ, UK [fax: +44(0)-1223-336033 or e-mail: deposit@ccdc.cam.ac.uk. This material is also available free of charge via the Internet at http://pubs.acs.org.