N-Substituted 9beta-methyl-5-(3-hydroxyphenyl)morphans are opioid receptor pure antagonists.

Article abstract of DOI:10.1021/jm980290i

The inhibition of radioligand binding and [35S]GTPgammaS functional assay data for N-methyl- and N-phenethyl-9beta-methyl-5-(3-hydroxyphenyl)morphans (5b and 5c) show that these compounds are pure antagonists at the micro, delta, and kappa opioid receptors. Since 5b and 5c have the 5-(3-hydroxyphenyl) group locked in a conformation comparable to an equatorial group of a piperidine chair conformation, this information provides very strong evidence that opioid antagonists can interact with opioid receptors in this conformation. In addition, it suggests that the trans-3, 4-dimethyl-4-(3-hydroxyphenyl)piperidine class of antagonist operates via a phenyl equatorial piperidine chair conformation. Importantly, the close relationship between the 4-(3-hydroxyphenyl)piperidines and 5-(3-hydroxyphenyl)morphan antagonists shows that the latter class of compound provides a rigid platform on which to build a novel series of opioid antagonists.

Full text of DOI:10.1021/jm980290i

J . Med. Chem. 1998, 41, 4143-4149
4143
N-Su bstitu ted 9â-Meth yl-5-(3-h yd r oxyp h en yl)m or p h a n s Ar e Op ioid Recep tor
P u r e An ta gon ists
J ames B. Thomas,† Xiaoling Zheng,† S. Wayne Mascarella,† Richard B. Rothman,§ Christina M. Dersch,§
J ohn S. Partilla,§ J udith L. Flippen-Anderson,‡ Clifford F. George,‡ Buddy E. Cantrell,|
Dennis M. Zimmerman,| and F. Ivy Carroll*,†
Chemistry and Life Sciences, Research Triangle Institute, Research Triangle Park, North Carolina 27709,
Clinical Psychopharmacology Section, NIDA Addiction Research Section, P.O. Box 5180, Building C, 4980 Eastern Avenue,
Baltimore, Maryland 21224, Laboratory for the Structure of Matter, Naval Research Laboratory, Building 35, Room 455,
Overlook Avenue, Washington, D.C., 20375, and Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center,
Indianapolis, Indiana 46285
Received May 8, 1998
The inhibition of radioligand binding and [35S]GTPγS functional assay data for N-methyl- and
N-phenethyl-9â-methyl-5-(3-hydroxyphenyl)morphans (5b and 5c) show that these compounds
are pure antagonists at the µ, δ, and κ opioid receptors. Since 5b and 5c have the
5-(3-hydroxyphenyl) group locked in a conformation comparable to an equatorial group of a
piperidine chair conformation, this information provides very strong evidence that opioid
antagonists can interact with opioid receptors in this conformation. In addition, it suggests
that the trans-3,4-dimethyl-4-(3-hydroxyphenyl)piperidine class of antagonist operates via a
phenyl equatorial piperidine chair conformation. Importantly, the close relationship between
the 4-(3-hydroxyphenyl)piperidines and 5-(3-hydroxyphenyl)morphan antagonists shows that
the latter class of compound provides a rigid platform on which to build a novel series of opioid
antagonists.
The opioid receptor system has been extensively
investigated, and thousands of compounds have been
synthesized and evaluated in radioligand binding as-
says, tissue assays, and animal models.1 Numerous
structural types of opioid agonists have been discovered,
and several such as methadone, meperidine, fentanyl,
and pentazocine as well as others have become impor-
tant drugs for the treatment of pain.1 In contrast, there
are only a few structural types that show potent, opioid
pure antagonist activity.1,2 A resurgence in heroin use
in recent years coupled with the demonstrated effective-
ness of opioid antagonists for the treatment of other
substances of abuse has spurred new interest in the
development of novel antagonists for opioid receptors.3
The oxymorphone-related compounds such as naloxone
(1a ) and naltrexone (1b) (Chart 1), where the antagonist
activity is dependent upon the N-substituent, have
received considerable attention over the past few de-
cades.1 For example, pioneering studies by Portoghese
and co-workers led to the development of the proto-
typical κ and δ opioid receptor antagonists norbinal-
torphimine (2, nor-BNI) and naltrindole (3, NTI). In
contrast, the N-substituted trans-3,4-dimethyl(3-hy-
droxyphenyl)piperidine (4a , 4b, 4c, 4d ) class of pure
antagonist has received relatively little attention.
Studies with the N-methyl analogue 4a as well as
many other N-substituted analogues such as 4b, 4c
(LY255582), and 4d showed that the pure antagonist
activity was dependent on the 3-methyl substituent and
its trans relative relationship to the 4-methyl substitu-
ent on the piperidine ring and, unlike the oxymorphone
class, was independent of the nature of the N-substit-
uent.2,4-8 Interestingly, the 3,4-dimethyl cis isomer 4e
was found to be a mixed agonist-antagonist. May and
co-workers9 reported that 2,9R-dimethyl-5-(3-hydroxy-
phenyl)morphan (5a ), which has the 9-methyl group in
a configuration comparable to the cis-3,4-dimethyl-4-
(3-hydroxyphenyl)piperidine (4e) with the 5-(3-hydroxy-
phenyl) group locked in an equatorial conformation
relative to the piperidine ring in the morphan structure,
was a weak but pure antagonist. Since the trans-3,4-di-
methyl-4-(3-hydroxyphenyl)piperidine (4a ) is 40 times
more potent than its cis isomer 4e, one might predict
that 2,9â-dimethyl-5-(3-hydroxyphenyl)morphan (5b),
which has the 9-methyl group in a configuration com-
parable to 4a , would also be a more potent pure
antagonist than the 9R isomer 5a . In this article, we
describe the synthesis of N-methyl- and N-phenethyl-
9â-methyl-5-(3-hydroxyphenyl)morphans (5b and 5c,
respectively) and report that these compounds, which
have the 3-hydroxyphenyl group locked in an equatorial
position relative to the piperidine ring in the morphan
structure, are high-affinity, pure antagonists for opioid
receptors.
Ch em istr y
The synthesis of the N-methyl- and N-phenethyl-9â-
methyl-5-(3-hydroxyphenyl)morphans (5b and 5c, re-
spectively) was achieved as illustrated in Scheme 1.10
Treatment of 1,2,6-trihydro-1,3-dimethyl-4-(3-methoxy-
phenyl)pyridine (6) with sec-butyllithium followed by
quenching with allyl bromide provided the enamine
adduct 7 which was cyclized without isolation to give
2,9-dimethyl-5-(3-methoxyphenyl)-2-azabicyclo[3.3.1]-
† Research Triangle Institute.
§ NIDA Addiction Research Section.
‡ Naval Research Laboratory.
| Eli Lilly and Company.
S0022-2623(98)00290-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/16/1998

4144 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Thomas et al.
Ch a r t 1
Sch em e 1a
a
(a) s-BuLi, allyl-Br; (b) H3PO4, HCO2H; (c) NaHB(OAc)3; (d) HOAc, HBr; (e) PhOCOCl, then KOH, H2O; (f) (benzotriazol-1-yloxy)tris(di-
methylamino)phosphonium hexafluorophosphate, PhCH2CO2H; (g) borane-dimethyl sulfide, THF.
non-3-ene (8a , 8b) in a 3:1 9â:9R-methyl ratio, using
1:1 phosphoric and formic acid. Reduction of unpurified
8a and 8b using sodium borohydride triacetate followed
by separation of the major isomer gave 9. Subjection
of 9 to O-demethylation using hydrobromic acid in acetic
acid provided the desired phenylmorphan 5b. Single-
crystal X-ray analysis showed that 5b had the desired
9â-methyl relative configuration (Figure 1).
The N-phenethyl derivative 5c was prepared from
intermediate 9. Treatment of 9 with phenyl chlorofor-
mate followed by hydrolysis of the resulting urethane
with potassium hydroxide followed by O-demethylation
with hydrobromic acid in acetic acid gave 10. Com-
pound 10 was converted to 5c by coupling with phenyl-
acetic acid in the presence of (benzotriazol-1-yloxy)tris-
(dimethylamino)phosphonium hexafluorophosphate fol-
lowed by borane reduction of the resulting amide
intermediate.
Biologica l Resu lts
Table 1 lists the radioligand binding data for com-
pounds 5b and 5c along with data for naltrexone (1b).
While the binding of 5b to all three opioid receptors was
weak, it is particularly interesting to note that changing
the N-substituent from methyl to phenethyl (5c) pro-
vided a dramatic increase in binding affinity, a feature
shared by the corresponding 4-(3-hydroxyphenyl)pip-
eridine analogues 4a and 4b (Table 2).11 Furthermore,
the relative binding affinities displayed by 5b and 5c

Opioid Receptor Pure Antagonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 4145
Ta ble 3. Inhibition by Antagonists of [35S]GTPγS Binding in
Guinea Pig Caudate Stimulated by DAMGO (µ), SNC80 (δ),
and U69,593 (κ) Selective Opioid Agonistsa
Ki (nM ( SD)
compd
µ (DAMGO)a
δ (SNC80)b
κ (U69,593)c
5b
5c
21.2 ( 2.30
0.338 ( 0.028
0.930 ( 0.21
750 ( 85.9
12.6 ( 1.01
19.3 ( 2.25
105 (10.9
1.34 ( 0.084
2.05 ( 0.21
1b, naltrexone
a
DAMGO, (D-Ala2,MePhe4,Gly-ol5)enkephalin, agonist selective
b
for
µ
opioid receptor. SNC80, [(+)-4-[(RR)-R-(2S,5R)-4-allyl-
2,5-dimethyl-1-piperazinyl)-3-methoxybenzyl]-N,N-diethylbenz-
amide, agonist selective for δ opioid receptor. c U69,593, trans-
3,4-dich lor o-N-m et h yl[2-(1-pyr r olidin yl)cycloh exyl]ben zen e-
acetamide, agonist selective for κ opioid receptor.
tioned previously, retention of pure antagonist activity
regardless of the N-substituent structure is a key
feature that separates the 3,4-dimethyl-4-(3-hydroxy-
phenyl)piperidine class of antagonist from oxymor-
phone-based antagonists which display pure antago-
nism only for certain N-substituents such as the N-allyl
or N-cyclopropylmethyl derivatives. In their ability to
reverse agonist-stimulated GTP binding, compound 5c
displayed a higher potency than naltrexone. These
results are striking since agonist activity in several
opioid ligands is enhanced by N-substituents with two
methylene groups terminated by a phenyl group (N-
phenethyl). It is evident that the antagonist activity
of 5c is due to factors different from those of the
oxymorphone-type pure antagonists.
The data in Table 3 also demonstrate that the
N-methyl to N-phenethyl change, 5b to 5c, results in a
concomitant increase in antagonist potency. Thus, as
is the case for the 3,4-dimethyl-4-(3-hydroxyphenyl)-
piperidines, the antagonist potency and not the agonist/
antagonist behavior of the 9â-methyl-5-(3-hydroxy-
phenyl)morphans 5b and 5c is mediated by the N-
substituent.
F igu r e 1. X-ray structure of 5b drawn using the experimen-
tally determined coordinates.
Ta ble 1. Radioligand Binding Results at the µ, δ, and κ
Opioid Receptors for N-Methyl- and
N-Phenethyl-9â-methyl-5-(3-hydroxyphenyl)morphans
Ki (nM ( SD)
compd
µ [3H]DAMGOa δ [3H]DADLEb κ [3H]U69,593c
5b
5c
166 ( 15
3.11 ( 0.21
1.39 ( 0.40
>10000
272 ( 30
94.9 ( 6.6
816 ( 66
14.5 ( 0.99
4.71 ( 0.7
1b, naltrexone
a
[3H]DAMGO, (D-Ala2,MePhe4,Gly ol5)enkephalin, tritiated
ligand selective for µ opioid receptor. [3H]DADLE, (D-Ala2,D-
b
Leu5)enkephalin, tritiated ligand selective for δ opioid receptor.
c [3H]U69,593, [3H]-5R,7R,8â-(-)-N-methyl-N-[7-(1-pyrrolidinyl)-
1-oxaspiro[4,5]dec-8-yl]benzeneacetamide, tritiated ligand selective
for κ opioid receptor.
Ta ble 2. Affinities of the N-Substituted
3,4-Dimethyl-(3′-hydroxyphenyl)piperidine Antagonists for the
µ and κ Opioid Receptorsa
Ki (nM)
Discu ssion
compd
µ [3H]Nalb
κ [3H]EKCc
In contrast to the naltrexone (1b) like antagonists,
the trans-3,4-dimethyl-4-(3-hydroxyphenyl)piperidines
do not rely on the N-substituent for expression of
antagonist activity.5 Instead, the effect of the N-
substituent has been shown to modulate the potency of
antagonism. Studies with a variety of 4-substituted
4-phenylpiperidines, each possessing both a 3-hydroxyl
substituent in their 4-phenyl ring and a 3â-methyl
substituent in the piperidine ring, displayed pure
antagonist activity.5,6 These studies also showed that
both opioid binding affinity and antagonist activity were
dependent on a trans relative relationship between the
3- and 4-methyl groups on the piperidine ring. This
information combined with the observation that the 3,4-
dimethyl cis isomer 4e displayed a peculiar pattern of
mixed agonist/antagonist activity led to the hypothesis
that a 3-hydroxyphenyl equatorial piperidine chair
conformation mediated the antagonists’ properties.
Further studies utilizing a variety of di- and trimethyl-
substituted 4-(3-hydroxyphenyl)piperidines7 supported
this hypothesis. We have demonstrated in this study
that N-methyl-9â-methyl-5-(3-hydroxyphenyl)morphan
(5b) is an opioid receptor pure antagonist. In addition,
we found that replacement of the N-methyl with an
N-phenethyl group to give 5c resulted in a 63-, 60-, and
4a
4b
80
1.5
0.56
833
52
3.9
1b, naltrexone
Data taken from ref 11. [3H]Naloxone (µ receptor assay).
a
b
c [3H]Ethylketocyclazocine (κ receptor assay).
for µ and κ opioid receptors are quite similar to that
observed for 4a and 4b. These results show that the
binding affinities of 5b and 5c are not adversely affected
by the 1,5-carbon bridge present in these structures. In
addition, it suggests a common binding mode for the two
types of structures.
The increase in binding of [35S]GTPγS stimulated by
opioid agonists is an assay able to distinguish com-
pounds of differing efficacy and intrinsic activity.8 The
antagonist properties of test compounds can be deter-
mined by measuring the inhibition of this stimulation.
To assess their potency as antagonists and to verify that
5b and 5c retain pure antagonist activity, the com-
pounds were analyzed for either stimulation or inhibi-
tion of agonist-stimulated GTP binding in comparison
with naltrexone (Table 3). In this functional assay,
neither 5b nor 5c stimulated GTP binding as measured
up to concentrations of 10 µM, showing that both
compounds were devoid of agonist activity.12 As men-

4146 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Thomas et al.
convincing evidence that opioid ligands of the phenyl-
piperidine class express potent opioid antagonist activity
with their 3-hydroxyphenyl group in an equatorial
position. Ong and co-workers15 reported that changing
the N-methyl group of the opioid agonist N-methyl-5-
(3-hydroxyphenyl)morphan (5d ) to an N-allyl- or N-
cyclopropyl group (5e and 5f) did not convert 5d to an
opioid antagonist. Thus, an equatorial 3-hydroxyphenyl
group even when combined with an N-allyl- or N-
cyclopropylmethyl substituent is not sufficient to ex-
press opioid antagonism. Since 5b and 5d differ only
by the presence of a 9â-methyl group in 5b, it is
apparent that this 9â-methyl substituent is essential
for pure opioid antagonist activity. Since both the
N-methyl and N-phenethyl analogues (5b and 5c,
respectively) are pure opioid antagonists, the antagonist
properties are not dependent on the N-substituent
structure which is the same behavior seen in the trans-
3,4-dimethyl-4-(3-hydroxyphenyl)piperidine class of an-
tagonists.
A comparison of the radioligand and [35S]GTPγS
binding properties of the N-substituted 9â-methyl-5-(3-
hydroxyphenyl)morphans 5b and 5c to those of the
N-substituted 3,4-dimethyl-4-(3-hydroxyphenyl)piperi-
dines 4a and 4b strongly suggests that these two types
of compounds are interacting with opioid receptors in a
similar mode. The pure antagonist activity of 5b, which
is increased when the N-methyl group is replaced by a
phenethyl group to give 5c, properties unique to the 3,4-
dimethyl-4-(3-hydroxyphenyl)piperidine class of antago-
nist, strongly supports the original hypothesis that this
class of opioid antagonist expresses pure antagonist
activity with the 4-(3-hydroxyphenyl) group in an
equatorial conformation.16
The difference in SARs required for the N-substitu-
ents of the oxymorphone class of antagonists, which
possess an axial 3-hydroxyphenyl group, and the 5-(3-
hydroxyphenyl)morphan class, which has an equatorial
3-hydroxyphenyl group, can be accounted for by an
extension of the model that Portoghese used to explain
the N-substituent requirement differences between the
benzomorphan and morphan classes of opioid ago-
nists.17,18 In a comparison of the opioid agonists of the
phenylmorphan series (phenyl equatorial) to those of the
benzomorphan series (phenyl axial), Portoghese pro-
posed that differences in behavior between the two
systems possessing the same N-substituents could be
explained by assuming a common binding site for the
phenyl rings in these compounds with the piperidine
nitrogen ends of the molecules sharing a common ionic
site of interaction but with their N-substituents situated
in different receptor environments.17,18 Viewed from
this point, the very different behaviors associated with
the N-substituents of the 3-hydroxyphenyl axial an-
tagonists compared with the 3-hydroxyphenyl equatorial
antagonists are easily understood. The N-substituent
structure in the phenyl axial series is the trigger for
antagonism, whereas the N-substituent structure in the
phenyl equatorial series only governs potency. It is the
axial 3-methyl substituent (phenylpiperidines) or the 9â-
methyl group (phenylmorphans) which triggers antago-
nist behavior.
F igu r e 2. Conformational representation of naltrexone (1b),
N-substituted 3,4-dimethyl-4-(3-hydroxyphenyl)piperidine, and
2-alkyl-9â-5-(3-hydroxyphenyl)morphan.
70-fold increase in antagonist potency at the µ, δ, and κ
opioid systems. These results are particularly interest-
ing since changing an N-methyl to an N-phenethyl
substituent in all opioid systems which have the 3-hy-
droxyphenyl group in an axial relationship relative to
the piperidine ring results in an increase in opioid
agonist activity.1 This information strongly suggests
that 5b and 5c are acting as conformationally rigid
analogues of the trans-3,4-dimethyl-4-(3-hydroxy-
phenyl)piperidine class of opioid antagonists where the
3-hydroxyphenyl group is in an equatorial position
relative to the piperidine ring.
In opioid alkaloids such as naloxone (1a ) and nal-
trexone (1b), the 3-hydroxyphenyl ring is fixed in an
axial orientation relative to the piperidine ring by the
rigid framework of the structure (Figure 2). The 3-hy-
droxyphenyl ring in the 3,4-dimethyl-4-(3-hydroxy-
phenyl)piperidine analogues 4 can be in either axial or
equatorial positions (Figure 2). 1H and 13C NMR
studies13,14 as well as molecular modeling studies11
suggest a preference for the 3-hydroxyphenyl equatorial
conformation. 5-(3-Hydroxyphenyl)morphans such as
5a , 5b, and 5c are sterically constrained 4-(3-hydroxy-
phenyl)piperidines with the 3-hydroxyphenyl ring fixed
in the equatorial position (Figure 2). The pure antago-
nist activity of the morphans 5b and 5c provides
This assumed alignment for the two antagonist
classes also provides an explanation for other differences

Opioid Receptor Pure Antagonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 4147
1 H), 5.82 (m, 1 H), 5.13 (m, 2 H), 3.80 (s, 3 H), 2.68-2.40 (m,
3 H), 2.55 (s, 3 H), 2.22 (m, 1 H), 1.66 (m, 2 H), 1.52 (s, 3 H);
13C NMR (CDCl3) δ 159.2, 151.1, 136.7, 135.8, 128.7, 119.8,
117.4, 114.3, 110.1, 107.7, 55.1, 46.1, 43.1, 43.0, 41.7, 36.4, 17.3.
noted between the two classes of antagonists. For
example, it is interesting to note that the N-substituted
5-(3-hydroxyphenyl)morphans 5b and 5c were more
potent as inhibitors of agonist-stimulated [35S]GTPγS
binding than as inhibitors in the binding assay (compare
data from Tables 1 and 3). In contrast, the radioligand
binding and GTP binding data for naltrexone, particu-
larly at the µ and κ opioid receptors, are very similar.
In this regard, we recently reported that a series of
N-substituted (3R,4R)-dimethyl-4-(3-hydroxyphenyl)pi-
peridines such as 4c and 4d were also more potent as
inhibitors of agonist-stimulated [35S]GTPγS binding
than as inhibitors in the radioligand binding assay.8
While these differences could at least in part be linked
to changes in tissue and radioligands, this does not
account for the differences seen between the opioid
ligands 5b, 5c, and 4c with those of naltrexone (1b).
This difference could be due to the fact that the
N-substituents of the two classes of antagonists are
interacting with different parts of the opioid receptor.
In summary, we have demonstrated that 9â-methyl-
5-(3-hydroxyphenyl)morphans are a new structural type
of pure opioid antagonist. The data also strongly
support the proposed 4-(3-hydroxyphenyl) equatorial
piperidine chair mode of interaction for the trans-3,4-
dimethyl-(3-hydroxyphenyl)piperidine class of opioid
antagonist.5,16 This new antagonist class provides a
more suitable structure on which to attach receptor
subtype-selective recognition elements.
2,9-Dim eth yl-5-(3-m eth oxyp h en yl)-2-a za bicyclo[3.3.1]-
n on -3-en e (8a , 8b). A solution of 300 mg (1.17 mmol) of 7 in
6 mL of 85% H3PO4/HCO2H (1:1) was stirred at room temper-
ature for 72 h. The resulting dark-brown mixture was diluted
with water (6 mL) and cooled in an ice bath while NaOH (25%
w/w) was added until pH 8. The aqueous solution was
extracted with CHCl3 (3×). The combined organic layers were
washed with aqueous NaHCO3 and brine and dried over
Na2SO4. Evaporation of the solvent gave 270 mg (90%) of
crude products 8a and 8b in a ratio of 3:1. The crude products
were used directly in the next step without further purifica-
tion: 1H NMR (CDCl3) of the mixture δ 7.24-6.70 (m, 4 H),
6.16 (d, 1 H, J ) 9.2 Hz), 4.34 (d, 1 H, J ) 7.0 Hz), 4.13 (d, 1
H, J ) 9.1 Hz), 3.80 (s, 3 H), 2.80 (s, 3 H), 3.10-1.40 (m, 8 H),
0.74 (d, 3 H, J ) 8.6 Hz), 0.57 (d, 3 H, J ) 8.1 Hz).
2,9â-Dim eth yl-5-(3-m eth oxyp h en yl)-2-a za bicyclo[3.3.1]-
n on a n e (9). A solution of 270 mg (1.05 mmol) of 8a and 8b
mixture and acetic acid (1.05 mmol, 0.061 mL) in 5 mL of
dichloroethane was treated with NaBH(OAc)3 under N2 atmo-
sphere. The reaction was stirred at room temperature for 2
h. The reaction was quenched by adding 10% NaOH to pH ∼
10. The mixture was extracted with ether (3×) and washed
with water and brine. The organic phase was dried over
Na2SO4 and concentrated under reduced pressure. Isolation
of the major isomer by chromatography (1% Et3N/EtOAc) gave
135 mg (50%) of 9 as a colorless oil: 1H NMR (CDCl3) δ 7.26
(m, 1 H), 6.94 (m, 2 H), 6.70 (m, 1 H), 3.80 (s, 3 H), 3.05-2.90
(m, 2 H), 2.71 (m, 1 H), 2.43 (s, 3 H), 2.42-2.30 (m, 2 H), 2.28-
2.15 (m, 1 H), 2.00-1.35 (m, 6 H), 0.86 (d, 3 H, J ) 8.25 Hz);
13C NMR (CDCl3) 159.2, 152.0, 128.9, 118.0, 112.3, 109.6, 59.7,
55.1, 51.1, 43.1, 42.5, 40.0, 38.3, 29.1, 25.6, 23.4, 14.8. Anal.
(C17H25NO) C, H, N.
Exp er im en ta l Section
2,9â-Dim eth yl-5-(3-h yd r oxyp h en yl)-2-a za bicyclo[3.3.1]-
n on a n e (5b). Compound 9 was treated with 4 mL of glacial
acetic acid and 4 mL of 48% aqueous hydrobromic acid at
reflux temperature for 20 h. The reaction was cooled to room
temperature and diluted with 10 mL of water. The pH was
adjusted to 10 by using 50% NaOH with ice cooling. The
product was extracted into a mixture of 3:1 1-butanol/toluene,
dried over Na2SO4, and concentrated under reduced pressure.
Separation by chromatography [50% (80% CHCl3, 18% MeOH,
2% NH4OH) in chloroform] provided 199 mg (84%) of 5b as a
white solid: 1H NMR (CDCl3) δ 7.15 (m, 1 H), 6.87-6.75 (m,
2 H), 6.61 (m, 1 H), 3.10-2.90 (m, 2 H), 2.77 (m, 1 H), 2.44 (s,
3 H), 2.50-2.30 (m, 2 H), 2.25-2.10 (m, 1 H), 2.00-1.60 (m, 5
H), 1.60-1.40 (m, 1 H), 0.80 (d, 3 H, J ) 8.3 Hz); 13C NMR
(CDCl3) δ 155.9.
Melting points were determined on a Thomas-Hoover capil-
lary tube apparatus and are not corrected. Elemental analyses
were obtained by Atlantic Microlabs, Inc., and are within
(0.4% of the calculated values. 1H NMR spectra were
determined on a Bruker WM-250 spectrometer using tetra-
methylsilane as an internal standard. Silica gel 60 (230-400
mesh) was used for all column chromatography. All reactions
were followed by thin-layer chromatography using Whatman
silica gel 60 TLC plates and were visualized by UV or by
charring using 5% phosphomolybdic acid in ethanol. All
solvents were reagent grade. Tetrahydrofuran and diethyl
ether were dried over sodium benzophenone ketyl and distilled
prior to use.
The [3H]DAMGO, DAMGO, and [3H][D-Ala2,D-Leu5]enkeph-
alin were obtained via the Research Technology Branch, NIDA,
and were prepared by Multiple Peptide Systems (San Diego,
CA). The [3H]U69,593 and [35S]GTPγS (s.a. ) 1250 Ci/mmol)
were obtained from DuPont New England Nuclear (Boston,
MA). U69,593 was obtained from Research Biochemicals
International (Natick, MA). Levallorphan was a generous gift
from Kenner Rice, Ph.D., NIDDK, NIH (Bethesda, MD).
GTPγS and GDP were obtained from Sigma Chemical Co. (St.
Louis, MO). The sources of other reagents are published.19
1,2,3,4-Tet r a h yd r o-4-a llyl-1,5-d im et h yl-4-(3-m et h oxy-
p h en yl)p yr id in e (7). To a solution of 500 mg (2.3 mmol) of
1,2,6-trihydro-1,3-dimethyl-4-(3-methoxyphenyl)pyridine (6) in
15 mL of THF at -42 °C was added s-BuLi in cyclohexane
(1.3 M, 2.9 mmol). After 1 h, allyl bromide (2.3 mmol) was
added, and the color of the solution changed from dark red to
yellow. After been stirred for 1 h at -42 °C, the mixture was
allowed to warmed to 0 °C and then quenched with water (10
mL). Diethyl ether (10 mL) was added, and the aqueous layer
was extracted with ether (2×). The combined ether layers
were washed with water (10 mL), saturated NaHCO3, and
brine and dried over Na2SO4. Evaporation of solvent afforded
590 mg (∼100%) of crude 7. The crude product was used
directly in the next step without further purification: 1H NMR
(CDCl3) δ 7.26 (m, 1 H), 7.01 (m, 2 H), 6.74 (m, 1 H), 5.89 (s,
The hydrochloride salt was prepared and crystallized from
ether/methanol using 1 N HCl in ethyl ether: 152.0, 129.1,
117.5, 113.0, 112.4, 59.7, 51.0, 43.0, 42.0, 40.2, 38.0, 29.0, 25.6,
23.2, 14.6. The structure of this compound was determined
by single-crystal X-ray analysis. Anal. (C16H24ClNO) C, H,
N.
5-(3-H yd r oxyp h en yl)-9â-m et h yl-2-a za b icyclo[3.3.1]-
n on a n e (10). A solution of 200 mg (1.28 mmol) of phenyl
chloroformate was added dropwise to 300 mg (1.16 mmol) of 9
in 10 mL of dichloromethane at room temperature under a
nitrogen atmosphere. The reaction was heated to reflux for 6
h. Since the reaction was not complete by TLC, the solvent
was then changed to dichloroethane and the reflux was
continued for another 12 h. The mixture was cooled to room
temperature and concentrated under reduced pressure. The
resulting oil was treated with 10 mL of 1 N NaOH and stirred
with slight warming for 15 min. The product carbamate was
then extracted with ether, and the ether layer was washed
with 1 N HCl and water. The organic phase was dried over
Na2SO4 and concentrated under reduced pressure. The resi-
due was then treated with 5 mL of ethanol and 1.5 mL of 50%
aqueous KOH at reflux for 70 h. The mixture was cooled and
concentrated under reduced pressure. The resulting concen-

4148 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21
Thomas et al.
trate was extracted with ether (2×), and the ether layers were
concentrated in vacuo. The resulting oil was dissolved into
10 mL of 1 N HCl and washed with ether. The aqueous layer
was then made strongly basic (pH > 12) with 50% NaOH with
ice cooling. The desired amine was extracted into ether (2×),
and the ether extracts were washed, dried over Na2SO4, and
concentrated under reduced pressure to give 207 mg (70%) of
a light-yellow oil. This was treated with 4 mL of glacial acetic
acid and 4 mL of 48% aqueous hydrobromic acid at reflux
temperature for 20 h. The reaction was cooled to room
temperature and diluted with 10 mL of water. The pH was
adjusted to 10 by using 50% NaOH with ice cooling. The
product was extracted into a mixture of 3:1 1-butanol/toluene,
dried over Na2SO4, and concentrated under reduced pressure
to yield 100 mg (51%) of 10 as a semisolid. The crude product
10 was used directly in the next step without further purifica-
tion: 1H NMR (CD3OD) δ 7.15 (m, 1 H), 6.79-6.75 (m, 2 H),
6.65 (m, 1 H), 3.70-3.30 (m, 3 H), 2.70 (m, 1H), 2.45-1.70 (m,
8 H), 0.87 (d, 3 H, J ) 8.3 Hz).
5-(3-H yd r oxyp h en yl)-9â-m et h yl-2-(2′-p h en ylet h yl)-2-
a za bicyclo[3.3.1]n on a n e (5c). To a solution of 100 mg (0.43
mmol) of 10, 190 mg (0.43 mmol) of BOP reagent, and 0.19
mL (1.38 mmol) of triethylamine in 15 mL of THF was added
phenylacetic acid (70.25 mg, 0.52 mmol). The mixture was
stirred at room temperature for 1 h. The reaction was diluted
with 45 mL of water and ether (45 mL). The aqueous layer
was extracted with ether (2×). The combined ether layers
were washed with NaHCO3 and brine and dried over Na2SO4.
Evaporation of solvent provided the crude product as a
colorless oil. The crude amide was dissolved in THF (8 mL).
The solution was cooled to 0 °C, and borane-methyl sulfide
complex (0.4 mL, 0.8 mmol) was added dropwise. After
vigorous reaction ceased, the resulting mixture was slowly
heated to reflux and maintained at that temperature for 4 h.
The reaction mixture was cooled to 0 °C, 6 mL of methanol
was added, and the mixture was stirred for 1 h. Anhydrous
hydrogen chloride in ether (1 mL) was added to attain a pH <
2, and the resulting mixture was gently refluxed for 1 h. After
the mixture was cooled to room temperature, methanol was
added, and the solvents were removed on a rotovap. The
residue obtained was made basic (pH > 12) by adding 25%
NaOH and extracted with ether (3×). The combined ether
layers were dried over Na2SO4 and concentrated under reduced
pressure. Separation by chromatography (1% Et3N/50% EtOAc/
hexanes) gave 38 mg (71%) of amine 5c as a colorless oil: 1H
NMR (CDCl3) δ 7.30-7.14 (m, 6 H), 6.85 (m, 2 H), 6.63 (m, 1
H), 4.71 (br s, 1 H), 3.05 (m, 2 H), 2.88 (m, 1 H), 2.79 (s, 4 H),
2.43-2.15 (m, 3 H), 1.94-1.65 (m, 5 H), 1.65-1.45 (m, 1 H),
0.83 (d, 3 H, J ) 8.2 Hz); 13C NMR (CDCl3) δ 155.7, 152.5,
140.9, 129.1, 128.8, 128.3, 125.9.
The hydrochloride salt was prepared and crystallized from
ether/methanol using 1 N HCl in ethyl ether: 117.7, 113.0,
112.4, 57.4, 57.2, 49.5, 42.4, 40.0, 38.7, 34.1, 29.1, 26.2, 23.4,
14.7. Anal. (C23H30ClNO) C, H, N.
Op ioid Bin d in g Assa ys. µ Binding sites were labeled
using [3H][D-Ala2-MePhe4,Gly-ol5]enkephalin ([3H]DAMGO)
(2.0 nM, s.a. ) 45.5 Ci/mmol), and δ binding sites were labeled
using [3H][D-Ala2,D-Leu5]enkephalin (2.0 nM, s.a. ) 47.5 Ci/
mmol) using rat brain membranes prepared as described.20 κ-1
binding sites were labeled using [3H]U69,593 (2.0 nM, s.a. )
45.5 Ci/mmol) and guinea pig membranes pretreated with BIT
and FIT to deplete the µ and δ binding sites.19
[3H]DAMGO binding proceeded as follows: 12- × 75-mm
polystyrene test tubes were prefilled with 100 µL of the test
drug which was diluted in binding buffer (BB: 10 mM Tris-
HCl, pH 7.4, containing 1 mg/mL BSA), followed by 50 µL of
BB and 100 µL of [3H]DAMGO in a protease inhibitor cocktail
(10 mM Tris-HCl, pH 7.4), which contained bacitracin (1 mg/
mL), bestatin (100 µg/mL), leupeptin (40 µg/mL), and chymo-
statin (20 µg/mL). Incubations were initiated by the addition
of 750 µL of the prepared membrane preparation containing
0.2 mg/mL protein and proceeded for 4-6 h at 25 °C. The
ligand was displaced by 10 concentrations of test drug, in
triplicate, 2×. Nonspecific binding was determined using 20
µM levallorphan. Under these conditions, the Kd of [3H]-
DAMGO binding was 4.35 nM. Brandel cell harvesters were
used to filter the samples over Whatman GF/B filters, which
were presoaked in wash buffer (ice-cold 10 mM Tris-HCl, pH
7.4).
[3H][D-Ala2,D-Leu5]enkephalin binding proceeded as fol-
lows: 12- × 75-mm polystyrene test tubes were prefilled with
100 µL of the test drug which was diluted in BB, followed by
100 µL of a salt solution containing choline chloride (1 M, final
concentration of 100 mM), MnCl2 (30 mM, final concentration
of 3.0 mM), and, to block µ sites, DAMGO (1000 nM, final
concentration of 100 nM), followed by 50 µL of [3H][D-Ala2,D-
Leu5]enkephalin in the protease inhibitor cocktail. Incuba-
tions were initiated by the addition of 750 µL of the prepared
membrane preparation containing 0.41 mg/mL protein and
proceeded for 4-6 h at 25 °C. The ligand was displaced by 10
concentrations of test drug, in triplicate, 2×. Nonspecific
binding was determined using 20 µM levallorphan. Under
these conditions the Kd of [3H][D-Ala2,D-Leu5]enkephalin bind-
ing was 2.95 nM. Brandel cell harvesters were used to filter
the samples over Whatman GF/B filters, which were presoaked
in wash buffer (ice-cold 10 mM Tris-HCl, pH 7.4).
[3H]U69,593 binding proceeded as follows: 12- × 75-mm
polystyrene test tubes were prefilled with 100 µL of the test
drug which was diluted in BB, followed by 50 µL of BB,
followed by 100 µL of [3H]U69,593 in the standard protease
inhibitor cocktail with the addition of captopril (1 mg/mL in
0.1 N acetic acid containing 10 mM 2-mercaptoethanol to give
a final concentration of 1 µg/mL). Incubations were initiated
by the addition of 750 µL of the prepared membrane prepara-
tion containing 0.4 mg/mL protein and proceeded for 4-6 h
at 25 °C. The ligand was displaced by 10 concentrations of
test drug, in triplicate, 2×. Nonspecific binding was deter-
mined using 1 µM U69,593. Under these conditions the Kd of
[3H]U69,593 binding was 3.75 nM. Brandel cell harvesters
were used to filter the samples over Whatman GF/B filters,
which were presoaked in wash buffer (ice-cold 10 mM Tris-
HCl, pH 7.4) containing 1% PEI.
For all three assays, the filtration step proceeded as fol-
lows: 4 mL of the wash buffer was added to the tubes, was
rapidly filtered, and was followed by two additional wash
cycles. The tritium retained on the filters was counted, after
an overnight extraction into ICN Cytoscint cocktail, in a
Taurus beta counter at 44% efficiency.
[
35S]GTP γS Bin d in g Assa y. Ten frozen guinea pig brains
(Harlan Bioproducts for Science, Inc., Indianapolis, IN) were
thawed, and the caudate putamen were dissected and homog-
enized in buffer A (3 mL/caudate) (buffer A: 10 mM Tris-HCl,
pH 7.4 at 4 °C, containing 4 µg/mL leupeptin, 2 µg/mL
chymostatin, 10 µg/mL bestatin, and 100 µg/mL bacitracin)
using a polytron (Brinkman) at setting 6 until a uniform
suspension was achieved. The homogenate was centrifuged
at 30000g for 10 min at 4 °C and the supernatant discarded.
The membrane pellets were washed by resuspension and
centrifugation twice more with fresh buffer A, aliquotted into
microfuge tubes, and centrifuged in a Tomy refrigerated
microfuge (model MTX 150) at maximum speed for 10 min.
The supernatants were discarded, and the pellets were stored
at -80 °C until assayed.
For the [35S]GTPγS binding assay, all drug dilutions were
made up in buffer B (50 mM Tris-HCl, pH 7.7/0.1% BSA).
Briefly, 12- × 75-mm polystyrene test tubes received the
following additions: (a) 50 µL of buffer B with or without an
agonist, (b) 50 µL of buffer B with or without 60 µM GTPγS
for nonspecific binding, (c) 50 µL of buffer B with or without
an antagonist, (d) 50 µL of salt solution which contained in
buffer B 0.3 nM [35S]GTPγS, 600 mM NaCl, 600 µM GDP, 6
mM dithiothreitol, 30 mM MgCl2, and 6 mM EDTA, and (e)
100 µL of membranes in buffer B to give a final concentration
of 10 µg/tube. The final concentrations of the reagents were
100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM dithiothreitol,
100 µM GDP, 0.1% BSA, 0.05-0.1 nM [35S]GTPγS, 500 nM or
10 µM agonists, and varying concentrations (at least 10
different concentrations) of antagonists. The reaction was

Opioid Receptor Pure Antagonists
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 21 4149
(6) Zimmerman, D. M.; Smits, S.; Nickander, R. Further investiga-
tion of novel 3-methyl-4-phenylpiperidine narcotic antagonists.
In Proceedings of the 40th Annual Scientific Meeting of the
Committee on Problems of Drug Dependence, 1978; pp 237-
247.
(7) Zimmerman, D. M.; Smits, S. E.; Hynes, M. D.; Cantrell, B. E.;
Leander, J . D.; Mendelsohn, L. G.; Nickander, R. Picenadol. Drug
Alcohol Depend. 1985, 14 (3-4), 381-401.
(8) Thomas, J . B.; Mascarella, S. W.; Rothman, R. B.; Partilla, J .
S.; Xu, H.; McCullough, K. B.; Dersch, C. M.; Cantrell, B. E.;
Zimmerman, D. M.; Carroll, F. I. Investigation of the N-
substituent conformation governing potency and µ receptor
subtype-selectivity in (+)-(3R,4R)-dimethyl-4-(3-hydroxyphenyl)-
piperidine opioid antagonists. J . Med. Chem. 1998, 41, 1980-
1990.
(9) Awaya, H.; May, E. L.; Aceto, M. D.; Merz, H.; Rogers, M. E.;
Harris, L. S. Racemic and optically active 2,9-dimethyl-5-(m-
hydroxyphenyl)morphans and pharmacological comparison with
the 9-demethyl homologues. J . Med. Chem. 1984, 27, 536-539.
(10) Evans, D. A.; Mitch, C. H.; Thomas, R. C.; Zimmerman, D. M.;
Robey, R. L. Application of metalated enamines to alkaloid
synthesis. An expedient approach to the synthesis of morphine-
based analgesics. J . Am. Chem. Soc. 1980, 102, 5955-5956.
WARNING: Read the background information relating to
analogues of MPTP including refferences for Zimmerman et al.
(J . Med. Chem. 1986, 29, 1517-1520) and references cited in
ref 11.
(11) Zimmerman, D. M.; Leander, J . D.; Cantrell, B. E.; Reel, J . K.;
Snoddy, J .; Mendelsohn, L. G.; J ohnson, B. G.; Mitch, C. H.
Structure-activity relationships of the trans-3,4-dimethyl-4-(3-
hydroxyphenyl)piperidine antagonists for µ and κ opioid recep-
tors. J . Med. Chem. 1993, 36, 2833-2841.
(12) Xu, H.; Lu, Y.-F.; Partilla, J . S.; Brine, G. A.; Carroll, F. I.; Rice,
K. C.; Lai, J .; Porreca, F.; Rothman, R. B. Opioid peptide receptor
studies. 6. The 3-methylfentanyl congeners RTI-4614-4 and its
enantiomers differ in efficacy, potency, and intrinsic efficacy as
measured by stimulation of [35S]GTP-γ-S binding using cloned
µ-opioid receptors. Analgesia 1997, 3, 35-42.
initiated by the addition of membranes and terminated after
4 h by addition of 3 mL of ice-cold (4 °C) purified water (Milli-Q
UV-Plus, Millipore) followed by rapid vacuum filtration through
Whatman GF/B filters presoaked in purified water. The filters
were then washed once with 5 mL of ice-cold water. Bound
radioactivity was counted by liquid scintillation spectroscopy
using a Taurus (Micromedic) liquid scintillation counter at 98%
efficiency after an overnight extraction in 5 mL of Cytoscint
scintillation fluid. Nonspecific binding was determined in the
presence of 10 µM GTPγS. Assays were performed in tripli-
cate, and each experiment was performed at least 3 times.
Da ta An a lysis. The data of the two separate experiments
(opioid binding assays) or three experiments ([35S]GTPγS
assay) were pooled and fit, using the nonlinear least-squares
curve-fitting language MLAB-PC (Civilized Software, Be-
thesda, MD), to the two-parameter logistic equation21 for the
best-fit estimates of the IC50 and slope factor. The Ki values
were then determined using the equation: IC50/(1 + ([L]/Kd)).
Sin gle-Cr ysta l X-r a y An a lysis of 5b. Crystals of 5b were
grown from ethyl ether/methanol. Data were collected on a
computer-controlled automatic diffractometer, Siemens P4,
with a graphite monochromator on the incident beam. Data
were corrected for Lorentz and polarization effects, and a face-
indexed absorption correction was applied. The structure was
solved by direct methods with the aid of the program SHELXS22
and refined by full-matrix least-squares on F2 values using
the program SHELXL.22 The parameters refined included the
coordinates and anisotropic thermal parameters for all non-
hydrogen atoms. Hydrogen atoms on carbons were included
using a riding model in which the coordinate shifts of their
covalently bonded atoms were applied to the attached hydro-
gens with C-H ) 0.96 Å. H angles were idealized and Uiso(H)
values set at fixed ratios of Uiso values of bonded atoms.
Coordinates were refined for H atoms bonded to nitrogen and
oxygen. Additional experimental and structural analysis
including an ORTEP figure and tables of atomic coordinates,
bond lengths, and band angles are available as Supporting
Information. Atomic coordinates are also available from the
Cambridge Crystallographic Data Centre (Cambridge Univer-
sity Chemical Laboratory, Cambridge CB2 1EW, U.K.).
(13) Casy, A. F.; Dewar, G. H.; Al-Deeb, O. A. A. Stereochemical
influences upon the opioid ligand activities of 4-alkyl-4-aryl-
piperidine derivatives. Chirality 1989, 1, 202-208.
(14) Casy, A. F.; Dewar, G. H.; Al-Deeb, O. A. A. Stereochemical
studies of the 4-alkyl-4-arylpiperidine class of opioid ligand.
Magn. Reson. Chem. 1989, 27, 964-972.
(15) Ong, H. H.; Oh-ishi, T.; May, E. L. Phenylmorphan agonists-
antagonists. J . Med. Chem. 1974, 17, 133-134.
(16) Zimmerman, D. M.; Smits, S. E.; Hynes, M. D.; Cantrell, B. E.;
Reamer, M.; Nickander, R. Structural requirements for affinity
and intrinsic activity at the opiate receptor defined in 4-phen-
ylpiperidine and related series. In Problems of Drug Dependence,
1981, Proceedings of the 43rd Annual Scientific Meeting, The
Committee on Problems of Drug Dependence, Inc.; NIDA Re-
search Monograph 41; Harris, L. S., Ed.; Committee on Problems
of Drug Dependence, Inc., 1982; pp 112-118.
(17) Portoghese, P. S. A new concept on the mode of interaction of
narcotic analgesics with receptors. J . Med. Chem. 1965, 8, 609-
616.
(18) Portoghese, P. S. The role of concepts in structure-activity
relationship studies of opioid ligands. J . Med. Chem. 1992, 35,
1927-1937.
Ack n ow led gm en t. This research was supported by
the National Institute on Drug Abuse, Grant DA09045.
Su p p or tin g In for m a tion Ava ila ble: Crystal data, struc-
tural refinement analysis, atomic coordinates, bond lengths,
bond angles, anisotropic displacement parameters, hydrogen
coordinates, and isotropic displacement parameters of 5b (6
pages). Ordering information is given on any current mast-
head page.
Refer en ces
(19) Rothman, R. B.; Bykov, V.; de Costa, B. R.; J acobson, A. E.; Rice,
K. C.; Brady, L. S. Interaction of endogenous opioid peptides and
other drugs with four kappa opioid binding sites in guinea pig
brain. Peptides 1990, 11, 311-331.
(20) Rothman, R. B.; Xu, H.; Seggel, M.; J acobson, A. E.; Rice, K. C.;
Brine, G. A.; Carroll, F. I. RTI-4614-4: an analogue of (+)-cis-
3-methylfentanyl with a 27, 000-fold binding selectivity for mu
versus delta opioid binding sites. Life Sci. 1991, 48, PL111-
PL116.
(21) Rodbard, D.; Lenox, R. H.; Wray, H. L.; Ramseth, D. Statistical
characterization of the random errors in the radioimmunoassay
dose-response variable. Clin. Chem. 1976, 22, 350-358.
(22) Sheldrick, G. M. SHELXTL-Plus, release 5.03; Siemens Analyti-
cal X-ray Instruments, Inc.: Madison, WI, 1995.
(1) Aldrich, J . V. Analgesics. In Burger’s Medicinal Chemistry and
Drug Discovery; Wolff, M. E., Eds.; J ohn Wiley & Sons: New
York, 1996; Vol. 3: Therapeutic Agents.
(2) Zimmerman, D. M.; Leander, J . D. Selective opioid receptor
agonists and antagonists: Research tools and potential thera-
peutic agents. J . Med. Chem. 1990, 33, 895-902.
(3) Kreek, M. J . Opiates, opioids and addiction. Mol. Psychiatry
1996, 1 (3), 232-254.
(4) Zimmerman, D. M.; Gidda, J . S.; Cantrell, B. E.; Schoepp, D.
D.; J ohnson, B. G.; Leander, J . D. Discovery of a potent,
peripherally selective trans-3,4-dimethyl-4-(3-hydroxyphenyl)-
piperidine opioid antagonist for the treatment of gastrointestinal
motility disorders. J . Med. Chem. 1994, 37, 7 (15), 2262-2265.
(5) Zimmerman, D. M.; Nickander, R.; Horng, J . S.; Wong, D. T.
New structural concepts for narcotic antagonists defined in a
4-phenylpiperidine series. Nature 1978, 275, 332-334.
J M980290I