Synthesis of trans-3,4-dimethyl-4-(3-hydroxyphenyl)piperidine opioid antagonists: Application of the cis-thermal elimination of carbonates to alkaloid synthesis

Article abstract of DOI:10.1021/jo951403y

Improved syntheses of two trans-3,4-dimethyl-4-(3-hydroxyphenyl)piperidine opioid antagonists from 1,3-dimethyl-4-piperidinone are described. The 1,3-dimethyl-4-arylpiperidinol 23 was selectively dehydrated in a two step process to the 1,3-dimethyl-4-aryl-1,2,3,6-tetrahydropyridine 26 by the cis-thermal elimination of the corresponding alkyl carbonate derivative at 190°C. In the presence of a basic nitrogen, the success of the elimination was found to be critically dependent upon the nature of the carbonate alkyl group, with Et, i-Bu, and i-Pr being preferred (90% yield). Alkylation of the metalloenamine, formed by deprotonation of 26 with n-BuLi, proceeded regio- and stereospecifically to give the trans-3,4-dimethyl-4-aryl-1,2,3,4-tetrahydropyridine 27, which was converted in three steps to the common intermediate, (3R,4R)-3,4-dimethyl-4-(3-hydroxyphenyl)piperidine. LY255582, a centrally-active opioid antagonist, and LY246736-dihydrate, a peripherally-active opioid antagonist, were prepared from 1,3-dimethyl-4-piperidinone in 11.8% yield (8 steps) and 6.2% yield (12 steps), respectively.

Full text of DOI:10.1021/jo951403y

J . Org. Chem. 1996, 61, 587-597
587
Syn th esis of tr a n s-3,4-Dim eth yl-4-(3-h yd r oxyp h en yl)p ip er id in e
Op ioid An ta gon ists: Ap p lica tion of th e Cis-Th er m a l Elim in a tion of
Ca r bon a tes to Alk a loid Syn th esis
J ohn A. Werner,* Louis R. Cerbone, Scott A. Frank, J effrey A. Ward, Parviz Labib,
Roger W. Tharp-Taylor, and C. W. Ryan
Chemical Process Research and Development, Lilly Research Laboratories,
A Division of Eli Lilly and Company, Indianapolis, Indiana 46285-4813
Received J uly 31, 1995X
Improved syntheses of two trans-3,4-dimethyl-4-(3-hydroxyphenyl)piperidine opioid antagonists from
1,3-dimethyl-4-piperidinone are described. The 1,3-dimethyl-4-arylpiperidinol 23 was selectively
dehydrated in a two step process to the 1,3-dimethyl-4-aryl-1,2,3,6-tetrahydropyridine 26 by the
cis-thermal elimination of the corresponding alkyl carbonate derivative at 190 °C. In the presence
of a basic nitrogen, the success of the elimination was found to be critically dependent upon the
nature of the carbonate alkyl group, with Et, i-Bu, and i-Pr being preferred (90% yield). Alkylation
of the metalloenamine, formed by deprotonation of 26 with n-BuLi, proceeded regio- and
stereospecifically to give the trans-3,4-dimethyl-4-aryl-1,2,3,4-tetrahydropyridine 27, which was
converted in three steps to the common intermediate, (3R,4R)-3,4-dimethyl-4-(3-hydroxyphenyl)-
piperidine. LY255582, a centrally-active opioid antagonist, and LY246736-dihydrate, a peripherally-
active opioid antagonist, were prepared from 1,3-dimethyl-4-piperidinone in 11.8% yield (8 steps)
and 6.2% yield (12 steps), respectively.
In tr od u ction
linked to the selectivity of LY255582 for the κ2b-binding
site.5 LY246736-dihydrate is a peripherally-active an-
tagonist which has a high affinity for the µ-opioid
receptor in the lining of the gastrointestinal tract (Ki <
1 nM) and may prove useful in the treatment of GI
motility disorders.6
The trans-3,4-dimethyl-4-(3-hydroxyphenyl)piperidines
are an important class of compounds exhibiting pure
opioid antagonist activity as a result of the 3-methyl
substituent.1 Previously, all known opioid antagonists
were polycyclic analogs having a morphine-like structure,
such as naloxone and naltrexone. Endogenous opioids
are known to play a key role in moderating various
physiological functions, such as GI motility, appetite, pain
sensitivity, and emotional state.2 Thus, selective opioid
antagonists may find uses in a number of therapeutic
areas. An efficient synthetic approach to this class of
antagonists should provide an opportunity for the further
evaluation of these compounds.
LY255582 (1) and LY246736-dihydrate (2) are two
examples of this class of opioid antagonists. LY255582
is a centrally-active antagonist which has a high affinity
for both the µ- and κ-opioid receptors (Ki ) 0.4 and 2 nM,
respectively) and may prove useful in the treatment of
obesity and eating disorders.1c,3 This antagonist has been
shown to promote weight loss in obese Zucker rats over
an extended time without any tolerance observed.4 It has
recently been suggested that this anoretic effect may be
Our successful efforts in developing an efficient syn-
thesis of the (3R,4R)-3,4-dimethyl-4-(3-hydroxyphenyl)-
piperidine nucleus allowed for the preparation of multi-
kilogram quantities of both 1 and 2 to support clinical
investigations and provided a key intermediate for
further SAR studies. During the course of this work, the
use of an alkyl carbonate for a cis-thermal elimination
reaction in the presence of a basic nitrogen was demon-
strated for the first time. This method should find
application in alkaloid synthesis as a selective method
for dehydration of amino alcohols at reasonable temper-
atures under neutral conditions.
* Author to whom correspondence should be addressed at: Eli Lilly
and Company, Dropcode 4813, Indianapolis, IN 46285-4813; 317-276-
9385; 317-276-4507 (fax).
X Abstract published in Advance ACS Abstracts, December 15, 1995.
(1) (a) Zimmerman, D. M.; Nickander, R.; Horng, J . S.; Wong, D. T.
Nature 1978, 275, 332. (b) Zimmerman, D. M.; Leander, J . D.; Cantrell,
B. E.; Reel, J . K.; Snoddy, J .; Mendelsohn, L. G.; J ohnson, B. G.; Mitch,
C. H. J . Med. Chem. 1993, 36, 2833. (c) Mitch, C. H.; Leander, J . D.;
Mendelsohn, L. G.; Shaw, W. N.; Wong, D. T.; Cantrell, B. E.; J ohnson,
B. G.; Reel, J . K.; Snoddy, J . D.; Takemori, A. E.; Zimmerman, D. M.
J . Med. Chem. 1993, 36, 2842.
Resu lts a n d Discu ssion
Syn th esis of th e tr a n s-3,4-Dim eth yl-4-a r ylp ip er i-
d in e Nu cleu s. Establishing the trans-diaxial relation-
(2) (a) Zimmerman, D. M.; Leander, J . D. J . Med. Chem. 1990, 33,
895. (b) J affe, J . H.; Martin, W. R. In Goodman and Gilman’s the
Pharmacological Basis of Therapeutics, 7th ed.; Gilman, A. G., Good-
man, L. S., Rall, T. W., Murad, F., Eds.; Macmillan: New York, 1985;
p 491.
(3) For the synthesis of 1 see: Mitch, C. H.; Zimmerman, D. M.;
Snoddy, J . D.; Reel, J . K.; Cantrell, B. E. J . Org. Chem. 1991, 56, 1660.
(4) For a recent reference see: Shaw, W. N. Pharmacol. Biochem.
Behav. 1993, 46, 653.
(5) Rothman, R. B.; Xu, H.; Char, G. U.; Kim, A.; De Costa, B. R.;
Rice, K. C.; Zimmerman, D. M. Peptides 1993, 14, 17.
(6) (a) Zimmerman, D. M.; Gidda, J . S.; Cantrell, B. E.; Schoepp, D.
D.; J ohnson, B. G.; Leander, J . D. J . Med. Chem. 1994, 37, 2262. (b)
Zimmerman, D. M.; Gidda, J . S.; Cantrell, B. E.; Werner, J . A.; Parli,
C. J .; Franklin, R. B.; Francis, P. C.; Means, J . R.; Pohland, R. C.;
Leander, J . D. Drugs of the Future 1994, 19, 1078.
0022-3263/96/1961-0587$12.00/0 © 1996 American Chemical Society

588 J . Org. Chem., Vol. 61, No. 2, 1996
Werner et al.
Sch em e 1. Liter a tu r e Syn th esis of 13
However, an alternative synthetic strategy is required
to address the second and third issues mentioned above;
namely, the moderate selectivity in controlling the rela-
tive stereochemistry of the C3-methyl group and the late
resolution. Starting with 1,3-dimethyl-4-piperidinone (8)
would solve both of these problems (eq 1). Introduction
of the C3-methyl group from the start potentially allows
the resolution to be conducted after step 1 instead of step
6. Furthermore, it should also provide a control element
for establishing the relative stereochemistry between the
two methyl groups. Barnett7 has demonstrated that the
metalloenamine derived from deprotonation of 10 (Ar )
3-MeOC6H4) with n-BuLi can be alkylated with complete
regio- and stereocontrol with propyl bromide to give 11
(R ) Pr, Ar ) 3-MeOC6H4) in 93% yield.
ship between the two methyl groups has been a major
challenge in the synthesis of 7. Mitch and co-workers3
used a two-step reduction strategy to prepare piperidine
6 from the (dimethylamino)methyl enamine 47 (Scheme
1). Hydrogenolysis of the dimethylamino group in 4,
followed by reduction of enamine 5 with NaBH3CN in
MeOH, afforded racemic 6 in 94% yield as a 93:7 mixture
of diastereomers. After separation of the diastereomers
by recrystallization of the HBr salt (80%), (()-6 was
resolved with dibenzoyl-D-tartaric acid in 38% yield to
give the trans-3,4-dimethylpiperidine 6. The protect-
ing groups on 6 were removed in two steps to complete
the synthesis of the piperidine nucleus 7 (8.5% over-
all yield from 3), which was then carried on to complete
the synthesis of opioid antagonist 1 (1.6% overall yield
from 3).
This approach presents three major challenges for its
safe implementation on a large scale. First, one of the
synthetic intermediates, 1-methyl-4-(3-methoxyphenyl)-
1,2,3,6-tetrahydropyridine causes neurotoxicity in mice
similar to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
(MPTP)8 (which leads to Parkinson’s-like symptoms in
animals and humans). Second, only moderate selectivity
is achieved in controlling the relative stereochemistry of
the C3-methyl group. And finally, the resolution occurs
6 steps into a 10-step synthesis, which critically impacts
the efficient use of resources.
The neurotoxicity issue was solved by changing the
nitrogen and oxygen protecting groups to ethyl and
isopropyl, respectively. Substituents larger than methyl
on either the piperidine nitrogen9 or the aryl oxygen10
are known to prevent the MPTP-like neurotoxicity in
animals. C-Methyl substitution on the piperidine ring
also eliminates all neurotoxicity.11 This process was
safely implemented on a pilot plant scale by modification
of the literature procedures as summarized in the Sup-
porting Information and afforded the piperidine nucleus
7 in 5.5% overall yield from N-ethyl-4-piperidinone.
Implementation of this strategy requires the unprec-
edented regiospecific dehydration of 9 to 10. The acid-
promoted dehydration of 9 is well precedented and gives
a 2.3:1.0 mixture of olefins 10 and 12.7,12 Upon acid
equilibration, the more substituted alkene 12 predomi-
nates in a 1:9 mixture. Thus, acid-promoted dehydration
is not a viable solution because it gives low yields of 10
and would involve a difficult purification. Also, if 12 is
present as a contaminant, both strategic advantages from
starting with 8 are lost as deprotonation of 12 would give
a planar metalloenamine, which upon methylation would
produce enamine 5.
A “cis-dehydration” of alcohol 9 is required to achieve
the regiospecific elimination to 10. This has been ac-
complished by Portoghese13 through the use of a two-step
procedure: formation of the propionate ester followed by
thermal elimination at 300 °C in mineral oil for 1 h. This
afforded 10 (Ar ) Ph) in 62% yield. We chose to use a
similar strategy, the thermal cis-elimination of an acyl
derivative, but needed to identify a derivative that would
undergo elimination at a significantly lower temperature
and in higher yield.
Carbonate esters were selected as the acyl group to be
examined for the cis-thermal elimination. They are
known to eliminate at lower temperatures than carbox-
ylic esters, 150-300 vs 250-400 °C, and produce neutral
byproducts upon elimination (CO2 and alcohol).14 Xan-
thate esters were not examined because of our concern
about quaternization of the piperidine nitrogen during
their formation.
(7) Prepared in four steps using methodology adapted from Barnett’s
synthesis of picenadol, a cis-3,4-dialkyl-4-arylpiperidine: Barnett, C.
J .; Copley-Merriman, C. R.; Maki, J . J . Org. Chem. 1989, 54, 4795.
(8) For a review of the pharmacology of MPTP and related analogs
see: Markey, S. P.; Schmuff, N. R. Med. Res. Rev. 1986, 6, 389.
(9) Zimmerman, D. M.; Cantrell, B. E.; Reel, J . K.; Hemrick-Luecke,
S. K.; Fuller, R. W. J . Med. Chem. 1986, 29, 1517 and references cited
therein.
(12) Casy, A. F.; Beckett, A. H.; Iorio, M. A. Tetrahedron 1967, 23,
1405 and references therein.
(13) (a) Larson, D. L.; Portoghese, P. S. J . Med. Chem. 1973, 16,
195. (b) The analogous reaction in a seven-membered ring nitrogen
heterocycle (75%, 280-300 °C) has been reported: Diamond, J .; Bruce,
W. F.; Tyson, F. T. J . Med. Chem. 1964, 7, 57.
(10) Fuller, R. W., Eli Lilly and Company. Unpublished results, May
21, 1986.
(11) Fries, D. S.; de Vries, J .; Hazelhoff, B.; Horn, A. S. J . Med.
Chem. 1986, 29, 424.
(14) For a review of pyrolytic cis-eliminations see: DePuy, C. H.;
King, R. W. Chem. Rev. 1960, 431.

Synthesis of Piperidine Opioid Antagonists
J . Org. Chem., Vol. 61, No. 2, 1996 589
Ta ble 1. Th er m a l Elim in a tion of Ca r bon a tes in Deca lin
The thermal elimination of carbonates has been used
for some time in carbocyclic systems,14 but to the best of
our knowledge, it is unprecedented with alkaloids. The
dramatic effect that the basic nitrogen can have is
illustrated in eq 2. Thermal decomposition of the N-ethyl
analog 13 at 175 °C (with a slight vacuum to remove
phenol as it is formed) gives the desired olefin 14 in 21%
yield and 1.5% of the corresponding piperidinol (weight
percent analysis by capillary GC) along with a multitude
of other products (analysis by 1H NMR). Heating the
hydrochloride salt 16 under the same conditions gives a
mixture of the two olefin isomers, 17 (25%) and 18 (38%),
and 1.5% of the piperidinol. We suspect that the carbon-
ate is undergoing an acid-promoted solvolysis to give the
mixture of elimination products since the less substituted
product 17 does not undergo isomerization when subjec-
ted to the reaction conditions. Removal of the basic nitro-
gen has a dramatic impact on the success of the thermal
elimination process. Carbamate 19, which is prepared
in 87% yield by treating 13 with phenyl chloroformate
in refluxing toluene, gives olefin 20 in 92% yield upon
heating.15 Apparently the basic nitrogen is responsible
for the decomposition of phenyl carbonate 13, and at 175
°C, decomposition occurs more readily than does elimina-
tion. Unfortunately, we were unable to characterize the
decomposition products. We suspected that an alkyl car-
bonate derivative might be more stable than the phenyl
carbonate derivative in the presence of the basic amine,
particularly if it was somewhat sterically hindered.
entry
R
substrate
23 (%)a
24 (%)a
1
2
3
4
5
6
7
Me
Et
i-Bu
Bn
i-Pr
t-Bu
Ph
22a
22b
22c
22d
22e
22f
22g
13.7
0.5
0.1
4.4
0.6
68
91
92
75
90
13
41
86.4
0.5
a
GC yields using naphthalene as an internal standard. Condi-
tions: column; DB-wax (30 m × 0.32 i.d. × 0.25 µm film), cool
on-column injection, oven, 35 °C (1.5 min) to 225 °C (15 min) at
15 deg/min; detector, 260 °C; carrier, He (1.7 mL/min). tR ) 12.0
min (naphthalene), 18.5 min (24), 25.2 min (23).
nantly decomposition (poor mass balance and many
1
products by H NMR, as observed with 13 above).
Two factors are responsible for the successful elimina-
tion with the ethyl, isobutyl, and isopropyl carbonates.
First, those carbonates are sufficiently stable in the
presence of the basic nitrogen to permit elimination to
occur. Second, the R-groups are primary or secondary
alkyl groups, which are known to undergo elimination
at slower rates than tertiary alkyl groups (such as the
piperidinyl group).14 Thus, thermal elimination gives
exclusively tetrahydropyridine 24. When both alkyl
groups are tertiary (entry 6, R ) t-Bu), elimination occurs
at comparable rates, and a statistical mixture of elimina-
tion products is obtained, 24 and isobutene (with an equal
amount of alcohol 23).
Application of this methodology to the synthesis of the
piperidine nucleus 7 is shown in Scheme 2. Treatment
of 1,3-dimethyl-4-piperidinone (8) with the aryllithium
derived from 3-(isopropoxy)bromobenzene using litera-
ture conditions7 gives a 9:1 mixture of diastereomeric
alcohols. Recrystallization from heptane affords 23 as a
single diastereomer in 69% yield. Alcohol 23 can be
resolved with (+)-3-bromocamphor-8-sulfonic acid;18 how-
ever, upon scale up the results are highly variable and
the salt is difficult to filter.
Acylation of alcohol 23 with ethyl chloroformate in
EtOAc gives carbonate (()-25, which is efficiently re-
solved with (+)-di-p-toluoyl-D-tartaric acid (DTTA) in
ethanol as a 1:1 salt. Unlike the resolution of alcohol
23, the resolution of (()-25 is extremely reproducible on
all scales. Recrystallization of the salt from ethanol gives
the enriched salt as a 95:5 mixture in 41% yield.19
Furthermore, any unreacted 23 is also removed during
the process.
In screening a number of alkyl carbonates, we found
that thermal elimination of primary alkyl carbonates
with some steric hindrance and secondary alkyl carbon-
ates cleanly gave elimination, while a tertiary alkyl
carbonate gave predominately carbinol 23 (Table 1).16,17
The thermal eliminations were conducted under two sets
of conditions: (1) method A, 225 °C, 1 h, neat: (2) method
B, 190 °C, 24 h, refluxing Decalin. Similar results were
obtained with both procedures. The data from method
B are summarized in Table 1. The primary alkyl
carbonates (entries 1-4) give high yields only when there
is some additional steric hindrance around the carbonyl
group, as with the ethyl and isobutyl carbonates. The
methyl and benzyl carbonates provide 24, in moderate
yields of 68% and 75%, respectively. The lower yields
are due both to the formation of other byproducts (GC
analysis) and to the alcoholysis of the carbonate group
by the liberated ROH, as evidenced by the formation of
alcohol 23 in 4.4-13.7% yield. The secondary alkyl
carbonate (entry 5, R ) i-Pr) undergoes clean elimination,
while the phenyl carbonate (entry 7) undergoes predomi-
(16) Werner, J . A.; Cerbone, L. R. Poster presentation at the 203rd
National Meeting of the American Chemical Society, San Francisco,
CA, April 1992; paper ORGN 409.
(17) As expected, an ester gives only a low yield of the elimination
product at this temperature. Heating the propionate ester of 23 at 190
°C for 24 h (method B) affords tetrahydropyridine 24 in only 12% yield,
along with unreacted starting material (85%) and an uncharacterized
impurity (3%).
(15) Alkylation of the metalloenamine formed by deprotonation of
20 using the conditions described in Table 2 gave only a 30% yield of
the desired trans-γ-alkylation product along with significant amounts
of N-deprotected starting material. Therefore, a method for effecting
the elimination in the presence of the basic amine was required.
(18) Resolution results: 22% yield, 96% ee, 4 recrystallizations from
i-PrOH at 200 mg/mL. The diastereomeric enrichment of the initial
crystallization varies from 2:1 to 1:1.

590 J . Org. Chem., Vol. 61, No. 2, 1996
Werner et al.
Sch em e 2
1
Ta ble 2. Meta lloen a m in e Alk yla tion a n d En a m in e
Red u ction
clean alkylation (by H NMR), but a significant amount
of a byproduct is produced upon reduction of crude 27
with NaBH4. We were unable to characterize the byprod-
uct and could not prevent the formation of this compound
by modification of the MeBr stoichiometry. Methyl iodide
gives a number of impurities in the alkylation step.
Alkylation of the metalloenamine is regiospecific at the
γ-position and occurs exclusively trans to the C3-methyl
substituent, as anticipated from the Barnett7 precedent.
Apparently the C3-methyl group occupies a pseudoaxial
position in the flattened metalloenamine, thus blocking
alkylation from the â-face. The stereochemistry of the
alkylation was determined by preparing an authentic
sample of the cis-3,4-dimethyl analog of 2922 and compar-
ing the 1H NMR spectra of the crude alkylation/reduction
reaction mixtures (29, methyl doublet at 0.80 ppm; 3,4-
cis-isomer, methyl doublet at 0.61 ppm). None of the cis-
isomer could be detected although small quantities (3-
5%) of the C4-desmethyl analog, resulting from proto-
nation of the metalloenamine, were observed.
entry
MeX
MeI
MeBr
Me2SO4
Me2SO4
equiv
temp (°C)
29 (%)
purity (%)a
1
2
3
4
1.2
1.5
1.1
1.02
-50
-28
-50
-50
69b
56b
72b
80
71
87
97
99
a
b
HPLC area % after silica gel chromatography. Yields cor-
rected for purity of 26 (90% by HPLC).
Thermal elimination of the free base 25, as described
above, cleanly gives the desired olefin 26 in 92% yield
after 24 h in refluxing Decalin. None of the isomeric
elimination product can be detected by 1H NMR, and the
crude product is used without further purification.
Purification of 29 by crystallization of the (+)-DTTA
salt 28 proceeds in 65% overall yield from 25, reduces
all impurities to less than 1%, and increases the enan-
tiomeric purity to 99.8% ee. It also provides a stable
crystalline intermediate.
Methylation of the metalloenamine20 with dimethyl
sulfate (1.02 equiv) at -50 °C gives the desired trans-
3,4-dimethyl enamine 27 in excellent yield. The alkyla-
tion reaction was examined extensively, and a brief
summary is shown in Table 2 of our results with three
different methylating agents under optimized conditions
for each. Dimethyl sulfate is by far the superior methy-
lating agent giving 29 in 80% yield after chromatography.
However, due to the reactivity of this reagent, dimethyl
sulfate must be used in stoichiometric amounts, and the
reaction mixture must be quenched into aqueous am-
monium hydroxide to prevent N-methylation (compare
entries 3 and 4).21 Methyl bromide appears to give
Demethylation of the free base of 29 with phenyl
chloroformate in refluxing toluene followed by removal
of both the N- and O-protecting groups with HBr in
refluxing HOAc completes the synthesis of the key
intermediate 7 (85% yield from 28). Preparation of the
trans-3,4-dimethylpiperidine nucleus 7 from 8 in seven
steps and 14.4% overall yield represents a significant
improvement over the strategy shown in Scheme 1.
Syn th esis of LY255582 (1). The N-alkyl substituent
of 1 is prepared in two steps from (S)-phenylpropane-
1,3-diol (30),23 Scheme 3. Reduction of the phenyl ring
proceeds in 92% yield. The primary alcohol in 31 is
readily derivatized to give the hydroxy brosylate 32 in
(19) (a) The “recrystallization” was conducted by heating a hetero-
geneous mixture at reflux for 2 h as described in the Experimental
Section. This process gives superior yields and equivalent enrichment
to those obtained by a true recrystallization from a homogeneous
solution. (b) The absolute stereochemistry was confirmed by comparing
the optical rotation of 7 with that reported previously: [R]25 +383°
(21) If excess dimethyl sulfate is present, the tertiary N-methyl
signal of 26 at 2.37 ppm disappears, and two quaternary N-methyl
signals at 3.55 and 3.59 ppm appear in the 1H NMR spectrum.
(22) The cis-3,4-dimethyl isomer of 29 was prepared from the cis-
dimethyl isomer recovered from the purification of 45b (see Scheme 5
in the Supporting Information). The N-ethyl substituent was converted
to the N-methyl substituent by dealkylation with phenyl chloroformate
followed by reduction of the phenyl carbamate with Red-Al.
365
(c 1.01, MeOH) (lit.3 [R]25 +361° (c 1.1, MeOH)). The (+)-DTTA salt
365
was the only crystalline salt obtained from the 12 chiral acids
examined.
(20) The introduction of an alkyl substituent at the 4-position of a
piperidine ring by alkylation of a metalated enamine was first
described by: Evans, D. A.; Mitch, C. H.; Thomas, R. C.; Zimmerman,
D. M.; Robey, R. L. J . Am. Chem. Soc. 1980, 102, 5955.

Synthesis of Piperidine Opioid Antagonists
J . Org. Chem., Vol. 61, No. 2, 1996 591
a
Sch em e 3. Syn th esis of LY255582 (1)
Sch em e 4
Syn th esis of LY246736-d ih yd r a te (2)
86% yield and greater than 90% ee24 after recrystalliza-
tion from toluene/cyclohexane.
Treatment of the piperidine nucleus 7 with brosylate
32 in refluxing 1,2-dimethoxyethane (DME) in the pres-
ence of NaHCO3 for 8 h affords 1 in 82% yield (98%
purity)25 after crystallization from EtOAc. The synthesis
of 1 from 1,3-dimethyl-4-piperidinone (8) in 8 steps and
11.8% yield (on a multikilo scale) offers an attractive
alternative to the previously reported synthesis3 (1.6%
yield, 10 steps from N-methyl-4-piperidinone).
a
Conditions: (a) methyl acrylate, THF, 40 °C, 3 h; (b) LDA (2
equiv), -30 °C; BnBr (2 equiv), -20 °C; HCl, MeOH; (c) NaOH,
H2O-MeOH (1:1), rt, 3 h; HCl, pH ) 6; (d) DCC, 1-hydroxyben-
zotriazole, Et3N, TsO-H3N+CHCH2CO2-i-Bu (36), THF, 48 h; rt;
(e) NaOH, EtOH/H2O (2:1), rt, 30 min; HCl, pH ) 6.
Syn th esis of LY246736-d ih yd r a te (2).26 LY246736-
dihydrate (2) was first prepared by the Michael addition
of 7 to 3-phenyl-2-(ethoxycarbonyl)-1-propene in metha-
nol.6a The reaction required 10 days at room temperature
and afforded a 1:1 mixture of diastereomeric esters (60:
40 mix of Me and Et esters) in 85% yield. The diaster-
eomeric mixture was carried on (hydrolysis and peptide
coupling) and was purified by a very difficult preparative
HPLC purification of isobutyl ester 37. Hydrolysis (6 N
HCl), extraction of the triethylammonium salt into bu-
tanol/toluene (3:1), and silica gel chromatography (EtOAc:
MeOH (9:1) to 100% MeOH) completed the synthesis of
2 with an overall yield of 4% from 7. We desired a
synthetic strategy that not only could be efficiently scaled
up but would also provide an opportunity for the dias-
tereoselective introduction of the benzyl substituent on
the side chain.27
zation of the hydrochloride salt 34 and also allows us to
examine diastereoselective alkylations by replacing the
methyl ester in 33 with a chiral auxiliary.
The Michael addition of 7 to methyl acrylate proceeds
smoothly in either methanol (15 min, rt) or THF (4 h, 40
°C), giving a 96% yield of 33. The difference in reaction
times is probably due to the increased solubility of 7 in
methanol.
Alkylation of the dianion of 33 proceeds in 97% yield,
giving a 47:53 mixture of the desired (3R,4R,RS)-isomer
and the undesired (3R,4R,RR)-isomer, respectively, which
are readily separated by recrystallization of their hydro-
chloride salts from methanol to give 34 in 34% yield.
Deprotonation of ester 33 with LDA (2.05 equiv,28 THF,
-30 °C) gives a slightly soluble dianion. The alkylation
with benzyl bromide must be conducted between -30 and
-15 °C. No reaction occurs below -30 °C, and the
dianion is unstable at temperatures above -15 °C (retro-
Michael). Due to the moderate reactivity of benzyl
bromide, 2 equiv must be used for complete conversion
of starting material. Use of less than 1.5 equiv gives only
80% conversion and does not change with the addition
of a catalytic amount of LiI (2.5 mol %). No reaction
occurs with benzyl chloride. Methyl iodide (1.05 equiv,
-25 °C) gives a 1:1 mixture of alkylation products in 97%
yield, thereby demonstrating that the low conversion is
likely a function of the reactivity of benzyl bromide. The
hydrochloride salt 34 must be formed prior to removing
the solvent to prevent N- or O-alkylation by the excess
benzyl bromide. Once the piperidine nitrogen is proto-
nated by the addition of anhydrous HCl, the mixture can
be concentrated without overalkylation. Upon the ad-
dition of methanol to the crude product, 34 precipitates
The alkylation-based approach shown in Scheme 4
meets both of these criteria. It permits the separation
of the diastereomeric alkylation products by recrystalli-
(23) Diol 30 is prepared in two steps from cinnamyl alcohol (Sharp-
less asymmetric epoxidation followed by Red-Al reduction): Gao, Y.;
Sharpless, K. B. J . Org. Chem. 1988, 53, 4081.
(24) The enantiomeric purity of 32 varies from 90 to 99% ee
depending on the reaction conditions of the hydrogenation step, the
scale of the reduction, and whether or not 32 is recrystallized.
Presumably the decrease in optical purity is due to the occurrence of
a small amount of benzylic oxidation, via a catalytic dehydrogenation
process, followed by a nonselective reduction back to diol 31.
(25) The largest impurity was the C4-desmethyl isomer (0.6%),
which was characterized by GC/MS: Murphy, A. T.; Breau, A. P.;
Werner, J . A.; Robbins, D. K. Structural Investigation of Impurities
in Synthesis of LY255582. Presented at the 41st American Society for
Mass Spectrometry Conference, San Francisco, CA, May 1993.
(26) This synthesis has been previously disclosed in communication
format. See ref 6b.
(27) (a) For a recent review of synthetic approaches to â-amino acids
see: Cole, D. C. Tetrahedron 1994, 50, 9517. (b) For a general review
of diastereoselective alkylation strategies see: Evans, D. A. In Asym-
metric Synthesis; Morrison, J . D., Ed.; Academic Press, Inc.: New York,
1984; Vol. 3, pp 2-110.
(28) Excess LDA must be avoided because it rapidly reacts with
benzyl bromide to give bromodiphenylethane (even at -70 °C). See:
Bazukis, P.; Dynak, J . N.; Wenkert, E. Syn. Commun. 1979, 9, 11.

592 J . Org. Chem., Vol. 61, No. 2, 1996
Werner et al.
as a crystalline solid. After a single recrystallization19a
in hot methanol, the (3R,4R,RS)-isomer 34 is isolated in
34% yield as a 97:3 mixture of diastereomers. Epimer-
ization (3 equiv LDA, -30 °C, THF; H2O) and recycling
of the (3R,4R,RR)-isomer increase the overall yield of the
benzylation step from 34% to 55%.
We have been unsuccessful to date in our attempts to
carry out a diastereoselective benzylation. Neither the
sodium nor the lithium enolates of various O-protected
oxazolidinone derivatives, 38, are sufficiently nucleophilic
to undergo alkylation with benzyl bromide. No reaction
occurs at -30 °C. Upon warming the reaction mixture
to 0 °C, acid 35 is isolated upon workup. Presumably
the acid is formed by hydrolysis of the ketene, which
results from the enolate undergoing elimination of the
chiral auxiliary. The lithium enolate of the sultam29
derivatives 39 are also unreactive toward benzyl bromide
(-78 to 0 °C, THF). Alternative strategies for the
diastereoselective introduction of the benzyl group are
under investigation and will be reported separately.
greater than 90% yield by a two-step procedure involving
formation and thermal elimination of an alkyl carbonate
derivative. The steric requirements of the alkyl group
are critically important to the success of the elimination
in the presence of a basic nitrogen, with Et, i-Bu, and i-Pr
being preferred. In the first application of this methodol-
ogy to the synthesis of alkaloids, the centrally-active
opioid antagonist LY255582 (1) was prepared in 8 steps
and 11.8% yield, and the peripherally-active opioid
antagonist LY246736-dihydrate (2) was prepared in 12
steps and 6.2% yield, from 1,3-dimethyl-4-piperidinone
(8) without any chromatographic purifications. This
methodology should find general application in the
synthesis of other alkaloids of this type, as it has been
used successfully on a pilot plant scale to prepare 50 kg
of the key intermediate 7 in 13.1% overall yield (cf. 14.4%
laboratory yield).
Exp er im en ta l Section
Gen er a l. All reactions were run under a nitrogen atmo-
sphere. Reagents were used as received unless otherwise
noted. Tetrahydrofuran (THF) was analyzed by Karl Fischer
titration and dried if the water content was greater than 0.03%
(either stored over 4 Å molecular sieves or distilled from a
sodium benzophenone ketyl solution). 1,3-Dimethyl-4-piperi-
dinone (8) was either prepared32 or purchased (Sumitomo or
Howard Hall International). (+)-Di-p-toluoyl-D-tartaric acid
monohydrate was purchased from Toray Industries, Inc.
Solutions of lithium diisopropylamide (LDA) were purchased
from either Aldrich or FMC Lithco. The term “3A ethanol”
refers to denatured ethanol containing 5% methanol. Proton
and carbon NMR spectra were obtained at 500 and 75.5 MHz,
respectively. NMR chemical shifts are reported in δ units
referenced either to TMS or to residual proton signals in the
deuterated solvents. All yields are corrected for chemical
purity of both the limiting reagent and the product (i.e. yield
) (weight of product × purity/MW of product)/(weight of
limiting reagent × purity/MW of limiting reagent) × 100). If
the purity of a product is not specified, it is greater than 99%.
1-Br om o-3-(1-m eth yleth oxy)ben zen e.33 A mixture of
ethanol (880 mL), potassium carbonate (448 g, 3.24 mol),
m-bromophenol (387 g, 2.24 mol), 2-bromopropane (400 g, 3.25
mol), and water (88 mL) was heated at reflux (78 °C) for 16 h.
Water (880 mL) was added to the reaction mixture, and 900
mL of solvent was removed by distillation at atmospheric
pressure over 4 h. Heptane (880 mL) was added to the
reaction mixture, and the layers were separated. The aqueous
layer was extracted with heptane (80 mL), and the combined
organic fractions were washed sequentially with 1 N HCl (300
mL), water (300 mL), 1 N NaOH (300 mL), and water (300
mL). Removal of the solvent by rotary evaporation afforded
453.6 g of crude product. Distillation through a short-path
column (81-85 °C, 2 Torr) afforded 431.5 g of a colorless liquid
(94% purity, 84% yield). Spectral data: 1H NMR (CDCl3) δ
1.32 (d, J ) 6.0 Hz, 6H), 4.51 (septet, J ) 6.1 Hz, 1H), 6.80
(ddd, J ) 8.2, 2.4, 0.9 Hz, 1H), 7.03 (d, J ) 2.1 Hz, 1H), 7.05
(m, 1H), 7.11 (t, J ) 8.2 Hz, 1H); 13C NMR (CDCl3) δ 21.1 (2C),
69.5, 113.9, 118.3, 122.0, 122.7, 129.7, 158.0; IR (CHCl3) 3030,
2982, 1588, 1572, 1474, 1115, 959, 872 cm-1. Anal. Calcd for
C9H11BrO: C, 50.26; H, 5.15; Br, 37.15. Found: C, 50.49; H,
5.26; Br, 37.38.
Saponification of ester 34 proceeds in 94% yield with
1 N NaOH without epimerization. Amino acid 35 can
be isolated in one of two ways: precipitation of the
zwitterion (isoelectric point, pH ) 6) from MeOH/H2O (1:
1)30 in 94% yield as the monohydrate or extraction of the
zwitterion into chloroform/2-propanol (3:1) in 95% yield
as the anhydrate. X-ray crystal structures were obtained
for the monohydrate of 35 and the ester hydrochloride
34 and showed the R-carbon to have the (S)-configuration.
Condensation of amino acid 35 (as the O-acyl deriva-
tive of 1-hydroxybenzotriazole) with the isobutyl ester31
of glycine gives ester 37 in 92% yield, which can either
be carried on crude or purified as a salt. Two crystalline
salts have been identified to date: the sesquimalate salt
using L-malic acid (90% yield, EtOAc/acetone (12.5:1))
and the HCl‚(acetone) salt (85% yield, acetone). Saponi-
fication of crude 37 with 1 N NaOH is very rapid in 2:1
EtOH/H2O (15 min, 25 °C), and the crystalline dihydrate
2 is isolated directly upon neutralization with hydrochlo-
ric acid in 93% yield. Opioid antagonist 2 was success-
fully prepared in 5 steps and 42% yield from the
piperidine nucleus 7 and in 12 steps and 6.2% yield from
1,3-dimethyl-4-piperidinone (8).
Con clu sion s
Alkylation of the metalloenamine derived from a C3-
substituted 1,2,3,6-tetrahydropyridine is a very efficient
method for the stereospecific preparation of the trans-
3,4-dimethyl-4-arylpiperidine family of opioid antago-
nists. Preparation of the required tetrahydropyridine
from the 4-aryl-4-piperidinol is readily accomplished in
cis-(()-1,3-Dim et h yl-4-[3-(1-m et h ylet h oxy)p h en yl]-4-
p ip er id in ol (23). n-BuLi (565 mL, 1.47 M in heptane, 0.831
mol) was added dropwise to a solution of 1-bromo-3-(1-
methylethoxy)benzene (200 g, 94% purity, 0.870 mol) in THF
(540 mL) over 1 h at -75 °C, and the resulting solution was
stirred for 1 h. 1,3-Dimethyl-4-piperidinone (106.7 g, 0.8389
mol) was then added over 1 h while maintaining the reaction
(29) Oppolzer, W.; Moretti, R.; Thome, S. Tetrahedron Lett. 1989,
30, 6009.
(30) The MeOH/H2O ratio is critical for a successful crystallization.
At lower MeOH levels a gum is formed.
(31) The isobutyl ester was used because multigram quantities of
ester 37 were also needed for preclinical evaluation.
(32) Howton, D. R. J . Org. Chem. 1945, 10, 277.
(33) Procedure developed by Dr. Mary Peters, Eli Lilly and Com-
pany, unpublished results, 1987.

Synthesis of Piperidine Opioid Antagonists
J . Org. Chem., Vol. 61, No. 2, 1996 593
temperature below -70 °C, and the resulting mixture was kept
below -60 °C for 1.5 h. The reaction was quenched by adding
the cold reaction mixture to 6 N HCl (280 mL) while keeping
the temperature at 25 °C, and then the pH was adjusted to 1
with 12 N HCl (ca. 8 mL). After phase separation, heptane
(320 mL) and 50% aqueous NaOH (48 mL) were added to the
aqueous layer (pH ) 13-14), and the resulting mixture was
allowed to stand at room temperature overnight. The layers
were separated at 45 °C, and the warm aqueous layer was ex-
tracted with heptane (320 mL). The combined organic frac-
tions were washed with water (120 mL) at 45 °C. The resul-
ting organic layer was concentrated to a volume of 235 mL by
vacuum distillation (55 °C, 100 Torr). Heptane (380 mL) was
added to the crude product (ca. 9:1 ratio of diastereomers), and
the mixture was reheated to 55 °C. The reaction mixture was
seeded at 45 °C and allowed to cool slowly overnight to room
temperature. The slurry was then cooled to 0 °C for about 2
h and filtered, and the cake was washed with cold heptane
(135 mL). The white solid was dried (in vacuo at 50 °C) to
give 151.8 g of 23 (69% yield), mp 75.0-76.0 °C. Spectral data
for 23: 1H NMR (CDCl3) δ 0.63 (d, J ) 6.9 Hz, 3H), 1.32 (d, J
) 6.2 Hz, 6H), 1.69 (dd, J ) 14.0, 1.4 Hz, 1H), 1.85 (s, 1H),
2.10-2.22 (m, 2H), 2.25-2.31 (m, 1H), 2.33 (s, 3H), 2.38 (t, J
) 12.1 Hz, 1H), 2.66 (br d, J ) 11.1 Hz, 1H), 2.74 (br d, J )
9.35 Hz, 1H), 4.54 (septet, J ) 6.0 Hz, 1H), 6.75 (d, J ) 8.2
Hz, 1H), 7.00 (d, J ) 7.6 Hz, 1H), 7.03 (m, 1H), 7.22 (t, J )
7.9 Hz, 1H); 13C NMR (CDCl3) δ 12.9, 22.7 (2C), 39.9, 41.2,
46.8, 52.2, 59.5, 70.4, 74.0, 113.4, 114.5, 117.6, 129.8, 149.7,
158.5; IR (CHCl3) 3610, 2978, 2941, 2805, 1606, 1582, 1483,
1468, 1116, 952 cm-1. Anal. Calcd for C16H25NO2: C, 72.96;
H, 9.57; N, 5.32. Found: C, 72.82; 9.46, N, 5.11.
(m, 9H), 1.86-2.00 (m, 1H), 2.19 (td, J ) 12.1, 2.0 Hz, 1H),
2.25 (t, J ) 11.4 Hz, 1H), 2.33 (s, 3H), 2.33-2.40 (m, 1H), 2.65
(dd, J ) 11.5, 3.4 Hz, 1H), 2.79 (br d, J ) 11.6 Hz, 1H), 2.98
(dt, J ) 14.4, 2.5 Hz, 1H), 4.11-4.21 (m, 2H), 4.50 (septet, J
) 6.1 Hz, 1H), 6.75-6.79 (m, 3H), 7.21 (t, J ) 7.9 Hz, 1H); 13
C
NMR (CDCl3) δ 12.6, 14.4, 22.0, 22.1, 32.8, 42.6, 45.9, 51.1,
58.9, 63.5, 69.9, 84.3, 113.2, 114.3, 117.3, 129.0, 143.4, 153.2,
157.7; IR (CHCl3) 2980, 2942, 2805, 1743, 1604, 1584, 1278,
1260, 1235 cm-1
. Anal. Calcd for C19H29NO4: C, 68.03; H,
8.71; N, 4.18. Found: C, 67.74; H, 8.66; N, 4.13.
(R )-1,2,3,6-Te t r a h yd r o-1,3-d im e t h yl-4-[3-(1-m e t h yl-
eth oxy)p h en yl]p yr id in e (26). A 1-L three-neck round-
bottom flask containing a magnetic stir bar and equipped with
a thermometer, a 12-in. Vigreux column fitted with a short-
path distillation condenser, and a glass stopper was charged
with 25 (50.0 g, 0.149 mol) and Decalin (250 mL). After the
system was evacuated and purged with nitrogen three times,
the solution was heated at reflux (190-195 °C) for 24 h. The
ethanol produced was removed by distillation. The yellow-
orange solution was cooled to room temperature under nitro-
gen, and 1 N HCl (155 mL) was added with stirring. The
aqueous layer was extracted with heptane (2 × 30 mL) to
remove residual Decalin. NaOH (10 mL, 50% aq) was added
(pH ) 13), and the free base was extracted into heptane (130
mL). The organic fraction was dried over Na2SO4. Filtration,
followed by removal of the solvent by rotary evaporation,
afforded 36.5 g of 26 as a yellow-orange liquid (92% purity,34
92% yield). This material was used without further purifica-
tion. An analytical sample was obtained by Kugelrohr distil-
lation (160 °C/0.2 Torr). Optical purity: ca. 90% ee based on
25; [R]25 -67.2° (c 1.01, MeOH). Spectral data: 1H NMR
D
Ca r b on ic Acid , E t h yl (3S,4R)-1,3-Dim et h yl-4-[3-(1-
m eth yleth oxy)-p h en yl]-4-p ip er id in yl Ester Com p ou n d
with (+)-D-2,3-Bis[(4-m eth ylben zoyl)oxy]bu tan edioic Acid
(1:1) (25‚(+)-DTTA). Ethyl chloroformate (205 mL, 2.14 mol)
was added over 60 min to a solution of 23 (463 g, 1.76 mol) in
ethyl acetate (2.28 L) while maintaining the temperature
below 15 °C. When the addition was complete, the reaction
mixture was allowed to warm to 25 °C and was stirred for an
additional 3 h. The reaction was quenched by pouring into a
mixture of 5 N NaOH (750 mL) and ethyl acetate (100 mL).
The organic fraction was washed with water (240 mL).
Removal of the solvent by rotary evaporation afforded 591 g
(93% purity,34 93% yield) of a viscous oil.
(CDCl3) δ 1.00 (d, J ) 7.0 Hz, 3H), 1.33 (d, J ) 5.8 Hz, 6H),
2.37 (s, 3H), 2.40 (d, J ) 5.2 Hz, 1H), 2.68 (dd, J ) 11.2, 4.9
Hz, 1H), 2.85-2.94 (m, 1H), 2.97 (dt, J ) 16.7, 2.8 Hz, 1H),
3.09 (dt, J ) 16.7, 2.8 Hz, 1H), 4.53 (septet, J ) 6.1 Hz, 1H),
5.83 (t, J ) 3.1 Hz, 1H), 6.76 (dd, J ) 8.1, 2.3 Hz, 1H), 6.83 (t,
J ) 1.9 Hz, 1H), 6.87 (d, J ) 7.8 Hz, 1H), 7.19 (t, J ) 7.9 Hz,
1H); 13C NMR (CDCl3) δ 18.8, 22.11, 22.16, 32.2, 46.0, 55.5,
60.3, 69.8, 114.06, 114.13, 118.6, 122.6, 129.2, 141.1, 142.6,
157.9; IR (CHCl3) 2979, 2940, 2789, 1603, 1576, 1194, 1117
cm-1. Anal. Calcd for C16H23NO: C, 78.32; H, 9.45; N, 5.71.
Found: C, 78.05; H, 9.42; N, 5.81.
(3R ,4S )-1,2,3,4-T e t r a h y d r o -1,3,4-t r im e t h y l-4-[3-(1-
m eth yleth oxy)p h en yl]p yr id in e (27). n-BuLi (70.0 mL, 1.6
M in hexane, 112 mmol) was added with stirring over 30 min
to a solution of 26 (19.6 g, 92% purity, 73.5 mmol) in THF
(175 mL) while maintaining the temperature at -10 to -20
°C. When the addition was complete, the deep red reaction
mixture was stirred for an additional 30 min at -15 °C. The
solution was then cooled to -50 °C, and dimethyl sulfate (7.7
mL, 81 mmol) was added dropwise over 30 min at -50 °C.
The reaction is very exothermic! When the addition was
complete, the yellow-brown mixture was stirred for an ad-
ditional 30 min at -50 °C and then slowly transferred via
cannula into a dilute solution of aqueous NH4OH (15.5 mL of
aqueous NH4OH/55 mL of water) and heptane (70 mL) at 0
°C. The mixture was warmed to 25 °C and stirred an addition-
al 2 h. The phases were separated, and the organic layer was
washed with water (40 mL). Removal of the solvent by rotary
evaporation afforded 21.4 g of 27 (86% purity,34 96% yield) as
an orange liquid, which was used immediately in the next step
without purification. An analytical sample was obtained by
Resolu tion . (+)-Di-p-toluoyl-D-tartaric acid monohydrate
(343 g, 0.850 mol) was added to a solution of 25 (285 g, 93%
purity, 0.790 mol) in 3A ethanol (2.85 L) at 55 °C. The solution
was heated to reflux and slowly cooled to room temperature
with stirring (overnight). The mixture was cooled to 0 °C for
2 h and filtered, and the cake washed with cold ethanol (170
mL). The product (323.4 g, 81:19 ratio of diastereomers) was
obtained as a white powder after drying (6 h at 50 °C).
Recrystallization of the crude salt from 3A ethanol (10 mL/g
of salt) afforded 232.8 g of 25‚(+)-DTTA (41% yield, 90% de,35
38% yield from 23) as white powder: mp 153.5-155 °C; [R]25
D
+64.2° (c 1.01, MeOH). Spectral data for 25‚(+)-DTTA: 1H
NMR (CDCl3) δ 0.62 (d, J ) 6.8 Hz, 3H), 1.25-1.30 (m, 9H),
2.24-2.32 (m, 1H), 2.33 (s, 6H), 2.71 (s, 3H), 2.65-2.85 (m,
3H), 2.97 (d, J ) 15.2 Hz, 1H), 3.28 (d, J ) 9.8 Hz, 1H), 3.55
(d, J ) 11.0 Hz, 1H), 4.05-4.20 (m, 2H), 4.52 (septet, J ) 6.1
Hz, 1H), 5.82 (s, 2H), 6.69 (d, J ) 7.9 Hz, 1H), 6.73 (br s, 1H),
6.77 (dd, J ) 8.3, 2.0 Hz, 1H), 7.12 (d, J ) 8.2 Hz, 4H), 7.14-
7.17 (m, 1H), 7.95 (d, J ) 8.1 Hz, 4H); 13C NMR (CDCl3) δ
11.8, 14.2, 21.6 (2C), 21.98, 22.02, 30.1, 40.0, 43.5, 50.0, 55.9,
64.1, 70.0, 73.1 (2C), 82.3, 113.0, 115.1, 117.2, 127.0 (2C), 129.0
(4C), 129.6, 130.1 (4C), 140.7, 143.6 (2C), 152.8, 157.7, 165.8
(2C), 170.6 (2C); IR (CHCl3) 2982, 1748, 1724, 1612, 1266,
1248, 1179, 1109 cm-1. Anal. Calcd for C39H47NO12: C, 64.90;
H, 6.56; N, 1.94. Found: C, 65.20; H, 6.65; N, 1.91.
Kugelrohr distillation (165 °C/0.25 Torr). Optical purity: [R]25
D
-57.8° (c 1.01, MeOH, ca. 90% ee based on 25). Spectral
data: 1H NMR (CDCl3) δ 0.59 (d, J ) 7.0 Hz, 3H), 1.33 (d, J
) 6.1 Hz, 6H), 1.43 (s, 3H), 1.87-1.95 (m, 1H), 2.47 (t, J )
10.9 Hz, 1H), 2.67 (s, 3H), 2.66-2.69 (m, 1H), 4.34 (d, J ) 7.9
Hz, 1H), 4.52 (septet, J ) 6.1 Hz, 1H), 5.97 (d, J ) 7.8 Hz,
1H), 6.71 (dd, J ) 7.9, 2.4 Hz, 1H), 6.95-6.96 (m, 2H), 7.17 (t,
J ) 8.2 Hz, 1H); 13C NMR (CDCl3) δ 15.0, 22.14, 22.16, 28.0,
38.8, 40.1, 42.5, 52.8, 69.7, 106.2, 112.7, 117.3, 121.2, 127.8,
1
An a lysis of 25. Optical purity: 90% ee by H NMR using
chiral shift reagent;35 [R]25 -10.2° (c 1.02, MeOH). Spectral
D
data: 1H NMR (CDCl3) δ 0.73 (d, J ) 6.8 Hz, 3H), 1.30-1.33
(35) Analysis of the free base by 1H NMR using S-(+)-2,2,2-trifluoro-
1-(9-anthryl)-ethanol as a chiral shift reagent showed the product to
be a 95:5 ratio of enantiomers (by integration of the methyl doublets
at 0.40 and 0.53 ppm).
(34) HPLC conditions for analysis of 25, 26, 27, 28: Zorbax RX-C8
(25 cm), 75% 0.1% trifluroracetic acid/25% acetonitrile, 2 mL/min, 273
nm.

594 J . Org. Chem., Vol. 61, No. 2, 1996
Werner et al.
135.2, 148.0, 157.0; IR (CHCl3) 3009, 2977, 1643, 1604, 1577,
1483, 1119, 975 cm-1. Anal. Calcd for C17H25NO: C, 78.72;
H, 9.71; N, 5.40. Found: C, 78.71; H, 9.77; N, 5.52.
Hz, 1H), 7.05-7.15 (m, 1H), 7.17 (t, J ) 7.5 Hz, 1H), 7.23 (t,
J ) 8.0 Hz, 1H), 7.27 (d, J ) 8.4 Hz, 1H), 7.34 (br t, J ) 7.4
Hz, 1H), 7.41 (t, J ) 4.3 Hz, 1H); 13C NMR (CDCl3, additional
signals due to amide rotamers) δ 14.3, 22.0 (2C), 26.4, 29.1,
29.6, 38.3, 38.4, 38.7, 40.3, 40.7, 45.9, 46.5, 69.6, 112.1, 112.2,
113.9, 117.5, 120.8, 121.6, 125.0, 126.2, 129.1, 129.4, 150.9,
157.8; IR (CHCl3) 3012, 2979, 1780, 1706, 1606, 1595, 1581,
1432, 1245, 983 cm-1. Anal. Calcd for C23H29NO3: C, 75.17;
H, 7.95; N, 3.81. Found: C, 74.95; H, 7.72; N, 3.54.
(3R,4R)-1,3,4-Tr im eth yl-4-[3-(1-m eth yleth oxy)p h en yl]-
p ip er id in e Com p ou n d w ith (+)-D-2,3-Bis[(4-m eth ylben -
zoyl)oxy]bu ta n ed ioic Acid (1:1) (28). To a solution of 27
(21.2 g, 86% purity, 70.3 mmol) in methanol (195 mL) at 0 °C
was added sodium borohydride (4.20 g, 111 mmol) while
keeping the temperature below 15 °C. When the addition was
complete, the reaction mixture was stirred for 3 h at 25 °C.
The reaction was quenched by the addition of acetone (21 mL)
and a saturated solution of NaHCO3 (25 mL). The solvent was
removed by rotary evaporation, and the residue was dissolved
in EtOAc (95 mL) and water (95 mL). The phases were
separated, the aqueous phase was extracted with EtOAc (20
mL), and the combined organic fractions were washed with
water (95 mL). Removal of the solvent by rotary evaporation
afforded 20.5 g (82% purity,34 92% yield) of a yellow liquid.
The crude product was purified by crystallization of the (+)-
di-p-toluoyl-D-tartaric acid salt from 3A ethanol (ca. 5 mL/g of
salt). The salt was further purified by heating a heterogeneous
mixture of the crude solid at reflux in 3A ethanol (5 mL/g of
salt) for 2 h, cooling to 0 °C, and filtering. This “hot reslurry”
procedure gave equivalent purity (greater than 99%)34 and
superior yields (33.4 g, 80% yield) to those of a typical recrys-
tallization process: mp 150.0-151.5 °C; [R]25D +113.6° (c 1.02,
MeOH). Spectral data: 1H NMR (DMSO-d6) δ 0.65 (d, J )
7.1 Hz, 3H), 1.26 (d, J ) 5.9 Hz, 6H), 1.31 (s, 3H), 1.72 (d, J
) 14.1 Hz, 1H), 2.15-2.28 (m, 2H), 2.37 (s, 6H), 2.65 (s, 3H),
2.95-3.14 (m, 2H), 3.15-3.25 (m, 2H), 4.60 (septet, J ) 6.1
Hz, 1H), 5.66 (s, 2H), 6.75 (br s, 1H), 6.76 (d, J ) 7.1 Hz, 1H),
6.78 (d, J ) 7.7 Hz, 1H), 7.21 (t, J ) 8.1 Hz, 1H), 7.30 (d, J )
8.1 Hz, 4H), 7.83 (d, J ) 8.1 Hz, 4H); 13C NMR (DMSO-d6) δ
14.6, 21.2 (2C), 21.9 (2C), 26.8, 36.5, 37.0, one aliphatic carbon
under residual DMSO signal, 43.0, 49.6, 55.0, 68.9, 72.5 (2C),
112.5, 113.5, 117.4, 128.0 (2C), 129.2-129.4 (10C), 143.7 (2C),
157.5, 165.0 (2C), 168.4 (2C); IR (CHCl3) 2979, 1725, 1612,
1579, 1268, 1179, 1109 cm-1. Anal. Calcd for C37H45NO9: C,
68.61; H, 7.00; N, 2.16. Found: C, 68.76; H, 7.03; N, 2.30.
An a lysis for F r ee Ba se of 28. Optical purity: 99.9% ee
(3R,4R)-3-(3,4-Dim eth yl-4-p iper idin yl)ph en ol (7). Com-
pound 28b (14.0 g, 38.0 mmol), 48% HBr (17.1 mL, 151 mmol),
and glacial acetic acid (17.1 mL) were combined and heated
at reflux for 18 h. When the reaction was complete, the solu-
tion was allowed to cool to room temperature, and 50 mL of
water was added. The solution was extracted with methyl tert-
butyl ether (3 × 30 mL) to remove the phenol byproduct. The
aqueous phase was titrated with a solution of 15% NaOH to a
pH of 8.5-8.8. Methanol (15 mL) was added to solubilize any
gummy solids, and the pH was adjusted to 10.3-10.5 with 15%
NaOH to precipitate the product. The mixture was stirred for
1.5 h at 25 °C, cooled to 0 °C, filtered, and washed with cold
water (10 mL). Drying (70 °C/5 Torr) afforded 6.86 g of 7 (88%
yield) as a tan crystalline solid: mp 179.4 °C; [R]25 +114.2°
D
(c 1.01, MeOH); [R]25 +383° (c 1.01, MeOH). Spectral data:
365
1H NMR (DMSO-d6) δ 0.67 (d, J ) 7.0 Hz, 3H), 1.29 (s, 3H),
1.40 (d, J ) 12.9 Hz, 1H), 1.79-1.89 (m, 1H), 2.00 (td, J )
12.4, 5.0 Hz, 1H), 2.56 (d, J ) 11.9 Hz, 1H), 2.82 (td, J ) 12.3,
2.6 Hz, 1H), 2.89 (br d, J ) 11.5 Hz, 1H), 3.09 (dd, J ) 12.6,
3.1 Hz, 1H), 6.57 (dd, J ) 7.9, 1.9 Hz, 1H), 6.69 (br s, 1H),
6.72 (d, J ) 7.9 Hz, 1H), 7.11 (t, J ) 7.9 Hz, 1H); 13C NMR
(DMSO-d6) δ 14.9, 27.2, 30.4, 37.7, 38.4, 42.0, 47.8, 112.1,
112.3, 115.8, 128.8, 152.3, 157.2; IR (KBr) 3291, 2939, 2665
(br), 2566 (br), 1583, 705 cm-1
. Anal. Calcd for C13H19NO:
C, 76.06; H, 9.33; N, 6.82. Found: C, 76.27; H, 9.19; N, 6.88.
(S)-1-Cycloh exyl-1,3-p r op a n ed iol (31). The (S)-phenyl-
propane-1,3-diol (30) (4.0 g, 26.3 mmol) was dissolved in 12
mL of 2-propanol. The 5% Rh/Al2O3 (1.0 g) was slurried in 8
mL of 2-propanol and added to the diol solution in a Parr
bottle. The catalyst was then rinsed with 4 mL of 2-propanol.
The mixture was shaken at 50 psi of hydrogen for 9 h at 25
°C. The catalyst was removed by filtration through Celite.
Removal of the solvent by rotary evaporation afforded 3.93 g
of 31 (97% pure by capillary GC,24 92% yield) as a colorless
by chiral HPLC;36 [R]25 +76.2° (c 1.01, MeOH). Spectral
D
data: 1H NMR (CDCl3) δ 0.80 (d, J ) 7.0 Hz, 3H), 1.30 (s,
3H), 1.32 (d, J ) 6.1 Hz, 3H), 1.33 (d, J ) 6.0 Hz, 3H), 1.55-
1.62 (m, 1H), 1.95-2.02 (m, 1H), 2.27 (s, 3H), 2.29-2.33 (m,
2H), 2.48-2.54 (m, 2H), 2.76-2.80 (m, 1H), 4.53 (septet, J )
6.1 Hz, 1H), 6.70 (dd, J ) 8.1, 2.3 Hz, 1H), 6.82 (t, J ) 2.0 Hz,
1H), 6.85 (d, J ) 7.8 Hz, 1H), 7.20 (t, J ) 8.0 Hz, 1H); 13C
NMR (CDCl3) δ 16.5, 22.09, 22.13, 27.6, 30.7, 38.0, 38.8, 46.8,
52.3, 58.6, 69.7, 112.0, 114.4, 118.1, 128.9, 151.9, 157.8; IR
(CHCl3) 3009, 2979, 2939, 2805, 1606, 1580, 1486, 1385, 1257,
1118, 986 cm-1. Anal. Calcd for C17H27NO: C, 78.11; H, 10.41;
N, 5.36. Found: C, 78.05; H, 10.41; N, 5.38.
oil: [R]25 -23.4° (c 1.03, MeOH). Spectral data: 1H NMR
D
(CDCl3) δ 0.95-1.30 (m, 5H), 1.31-1.38 (m, 1H), 1.60-2.00
(m, 7H), 2.66 (br s, 1H), 2.88 (br s, 1H), 3.59 (br s, 1H), 3.75-
3.90 (m, 2H); 13C NMR (CDCl3) δ 26.2, 26.3, 26.5, 28.2, 28.9,
35.2, 44.0, 61.8, 76.2; IR (CHCl3) 3622, 3600-3020, 3015, 2931,
2856, 1059, 1044 cm-1. Anal. Calcd for C9H18O2: C, 68.31;
H, 11.47. Found: C, 68.24; H, 11.37.
(S)-3-Cycloh exyl-3-h yd r oxyp r op yl (4-Br om oben zen e)-
su lfon a te (32). Diol 31 (7.70 g, 97% purity, 47.2 mmol),
triethylamine (10.2 mL, 73.0 mmol), and 4-(dimethylamino)-
pyridine (0.59 g, 4.9 mmol) were dissolved in CH2Cl2 (100 mL),
and the solution was cooled to 0 °C. A solution of 4-bromoben-
zenesulfonyl chloride (13.1 g, 51.2 mmol) in CH2Cl2 (50 mL)
was added over 30 min, and the mixture was stirred for 1 h
at 0 °C. The reaction mixture was washed successively with
0.5 N HCl (120 mL), water (100 mL), and a saturated solution
of NaHCO3 (100 mL) and was dried over anhydrous MgSO4.
Removal of the solvent by rotary evaporation at 60 °C followed
by recrystallization of crude product from toluene (20 mL) and
cyclohexane (160 mL) afforded 15.9 g of 32 (96% purity,37 86%
yield) as a white crystalline solid: mp 57.0-58.5 °C. Optical
(3R,4R)-3,4-Dim eth yl-4-[3-(1-m eth yleth oxy)p h en yl]-1-
p ip er id in eca r boxylic Acid P h en yl Ester (28b). Com-
pound 28 (61.9 g, 95.6 mmol) was neutralized with 2 N NaOH
(114.7 mL, 229.4 mmol, 2.4 equiv) in toluene (265 mL). The
free base was isolated, redissolved in dry toluene (160 mL),
and heated to 85 °C. Phenyl chloroformate (17.2 g, 110 mmol)
was added slowly, and the solution was heated at reflux for 2
h. After the solution was cooled to 45 °C, aqueous NaOH (5
mL of 50% aqueous NaOH in 40 mL of water) was added, and
the mixture was allowed to cool to room temperature with
stirring. The organic fraction was washed with 1:1 MeOH/1
N HCl (3 × 50 mL), 1:1 MeOH/1 N NaOH (50 mL), and water
(50 mL). Removal of the solvent by rotary evaporation
afforded 33.9 g of 28b (97% yield) as an oil which solidified
purity: 94.2% ee;24 [R]25 -18.8° (c 1.0, MeOH). Spectral
D
data: 1H NMR (CDCl3) δ 0.90-1.05 (m, 2H), 1.05-1.30 (m,
4H), 1.54 (s, 1H), 1.59-1.71 (m, 3H), 1.72-1.79 (m, 3H), 1.86-
1.92 (m, 1H), 3.44-3.48 (m, 1H), 4.18-4.22 (m, 1H), 4.27-
4.32 (m, 1H), 7.70 (d, J ) 8.6 Hz, 2H), 7.77 (d, J ) 8.6 Hz,
2H); 13C NMR (CDCl3) δ 26.0, 26.2, 26.4, 27.8, 28.9, 33.3, 43.8,
upon standing: mp 63 °C; [R]25 +61.0° (c 1.00, MeOH).
D
Spectral data: 1H NMR (CDCl3) δ 0.70-0.80 (m, 3H), 1.34 (d,
J ) 6.0 Hz, 6H), 1.41 (s, 3H), 1.63 (d, J ) 13.1 Hz, 1H), 2.00-
2.10 (m, 1H), 2.29-2.32 (m, 1H), 3.10-3.56 (m, 2H), 3.97-
4.05 (m, 1H), 4.25-4.38 (m, 1H), 4.54 (septet, J ) 6.1 Hz, 1H),
6.73 (dd, J ) 8.1, 2.3 Hz, 1H), 6.81 (br s, 1H), 6.84 (d, J ) 7.9
(37) HPLC conditions for analysis of 28, 28b, 7, 32, 1: Zorbax RX
(25 cm); 40 °C; gradient with acetonitrile and 0.1 M NaH2PO4 (pH )
3), 15% acetonitrile for 2 min then program to 70% acetonitrile in 18
min and hold for 10 min; 1.5 mL/min, 262 nm. tR ) 4.5 min (7), 13.6
min (free base of 28), 14.3 min (1), 22.3 min (32), 25.7 min (28b).
(36) HPLC conditions: column, Chiralcel OD (25 cm, Daicel Ind.);
mobile phase, 1% i-PrOH and 0.2% Et2NH in hexane (1.0 mL/min);
detector, 273 nm. tR ) 5.5 min (free base of 28), 6.5 min (ent-28).

Synthesis of Piperidine Opioid Antagonists
J . Org. Chem., Vol. 61, No. 2, 1996 595
69.0, 71.9, 129.0, 129.4 (2C), 132.6 (2C), 135.3; IR (CHCl3)
3650, 3050, 3010, 2931, 2856, 1579, 1365, 1187, 1178, 923, 825
. Anal. Calcd for C15H21BrO4S: C, 47.75; H, 5.61; Br,
21.18; S, 8.50. Found: C, 48.04; H, 5.65; Br, 20.94; S, 8.60.
r id e (34). A dry 250-mL three-neck round-bottom flask
equipped with a mechanical stirrer, a thermometer, and a
rubber septum was purged with nitrogen and charged with
THF (100 mL) and a 2.0 M solution of LDA (17.6 mL, 35.2
mmol). The yellow solution was cooled to -30 °C, and the
solution of 33 (15.2 g, 32.8% by weight in THF, 17.2 mmol)
was added over 20 min at -28 °C. The dianion precipitated
during the addition. After 15 min benzyl bromide (5.81 g, 34.3
mmol) was added over 5 min while maintaining the temper-
ature between -15 and -20 °C. Most of the solid dissolved
during the addition. The reaction mixture was stirred for 3 h
at this temperature.
The reaction mixture was quenched by adding 1 N HCl (22
mL, 22 mmol) and adjusting the pH to 10.6-9.5 with 12 N
HCl (ca. 2.3 mL) while keeping the temperature below -10
°C. Heptane (50 mL) was added, and the layers were
separated. Methanol (25 mL) was added to the organic layer,
and the solution was cooled to -5 °C. The hydrochloride salt
was formed by sparging anhydrous HCl (ca. 1.3 g) into the
solution while maintaining the temperature below 5 °C until
the mixture was acidic (pH ) 1, moist litmus paper). The
mixture was concentrated via rotary evaporation to a net
weight of 32.6 g (20% of initial weight) prior to the addition of
methanol (36 mL), which resulted in the precipitation of a
white solid after a few minutes. The mixture was stirred
overnight at room temperature. After cooling to 0 °C for 1.25
h the precipitate was filtered, the cake was washed with cold
methanol (10 mL), and the product was dried (50 °C/5 Torr)
to give 2.93 g (35% yield, 72% de41) of a white powder. The
crude product (2.75 g, 72% de, 5.66 mmol) was added to
methanol (14 mL), and the slurry was heated at reflux for 2
h. The mixture was cooled to room temperature with stirring
and then cooled to 0 °C for 1 h. The precipitate was filtered,
washed with cold methanol (1.5 mL), and dried (50 °C/5 Torr)
to afford 2.32 g of 34 (95% yield, 94% de41) as a white solid.
Overall yield was 34% from 33. Recycling the undesired
diastereomer (free base, dianion formation with LDA (3 equiv),
protonation, and HCl salt formation) increased the overall
yield from 34% to 55%.
cm-1
(rS,3R,4R)-r-Cycloh exyl-4-(3-h yd r oxyp h en yl)-3,4-d i-
m eth yl-p ip er id in ep r op a n ol (1).38 A mixture of 7 (36.0 g,
0.175 mol), 32 (69.3 g, 0.183 mol), NaHCO3 (20.2 g, 0.240 mol),
and 1,2-dimethoxyethane (DME, 400 mL) was heated at reflux
(85 °C) for 8 h. After the reaction was complete, the reaction
mixture was cooled to 25 °C and diluted with THF (512 mL),
and Celite (27.0 g) was added. After being stirred for 30 min,
the mixture was cooled to 0 °C and stirred for 1 h. The mixture
was filtered to remove the sodium brosylate/celite, and the cake
was washed with THF (2 × 164 mL). The filtrate and washes
were combined and concentrated to 140 g by rotary evapora-
tion at 85 °C. The concentrate was then diluted with EtOAc
(600 mL) and concentrated to 250 g. The resulting solution
was filtered, and the filter was washed with hot EtOAc (2 ×
75 mL). The filtrate was reconcentrated to 250 g. The slurry
was cooled to 0 °C and stirred for 2 h. The solid was then
collected and washed with cold EtOAc (2 × 120 mL). After
the solid was dried at 40 °C in vacuo, 50.5 g of 1 (98% purity,37
82% yield) was obtained: mp 156.0-159.6 °C; [R]25D +73.1° (c
1.0, MeOH). Spectral data: 1H NMR (CDCl3) δ 0.54 (d, J )
7.1 Hz, 3H), 0.95-1.26 (m, 5H), 1.28 (s, 3H), 1.33-1.40 (m,
1H), 1.50-1.55 (m, 2H), 1.61-1.73 (m, 5H), 1.91-1.96 (m, 2H),
2.27 (td, J ) 12.9, 4.3 Hz, 1H), 2.37 (dd, J ) 11.5, 2.9 Hz, 1H),
2.45 (td, J ) 12.0, 1.9 Hz, 1H), 2.59 (dt, J ) 12.8, 3.5 Hz, 1H),
2.67 (td, J ) 12.2, 2.7 Hz, 1H), 2.83 (d, J ) 11.4 Hz, 1H), 2.88
(d, J ) 11.4 Hz, 1H), 3.58 (br t, J ) 8.2 Hz, 1H), 6.60 (dd, J )
8.0, 2.1 Hz, 1H), 6.67 (br s, 1H), 6.70 (d, J ) 8.0 Hz, 1H), 7.08
(t, J ) 7.9 Hz, 1H), 7.30-7.50 (br s, 2H, D2O exchange); 13C
NMR (DMSO-d6) δ 16.0, 25.8, 26.0, 26.3, 27.2, 27.7, 28.7, 30.0,
30.2, 37.95, 38.03, 43.5, 50.2, 55.2, 56.4, 74.3, 112.2, 112.5,
116.0, 128.8, 151.6, 157.1; IR (KBr) 3112, 2925, 1604, 1443,
1236, 705 cm-1; UV (EtOH) λmax 201, ꢀ 18 851; 274, ꢀ 2068.
Anal. Calcd for C22H35NO2: C, 76.47; H, 10.21; N, 4.05.
Found: C, 76.45; H, 10.21; N, 4.08.
Meth yl (3R,4R)-4-(3-Hyd r oxyp h en yl)-3,4-d im eth yl-1-
p ip er id in ep r op a n oa te (33). Methyl acrylate (46.4 mL,
0.515 mol) was added over 3 min to a suspension of 7 (70.5 g,
0.343 mol) in THF (1 L) at 45 °C. After 4 h the reaction
mixture was cooled to room temperature and filtered through
Celite to give a clear amber solution. The solvent and excess
methyl acrylate were removed by concentration of the solution
via rotary evaporation at 40 °C to a net weight of 120 g. The
crude product was redissolved in THF (180 g) to give a 33.3%
solution (by weight)39 of 33 for use in the benzylation step (96%
Analytical data for both the salt and the free base are
recorded. The solubility of the HCl salt is very low in most
solvents. In DMSO 34 exists as an 85:15 mixture of diaster-
eomeric salts due to the axial and equatorial protonation of
1
the piperidine nitrogen, giving rise to a complicated H NMR
spectrum.
An a lytica l Da ta (34). Mp 230.0-232.0 °C dec. Spectral
data: 1H NMR (DMSO-d6) δ (0.78 (d, J ) 7.2 Hz, 0.85 × 3H)
and 1.02 (d, J ) 7.2 Hz, 0.15 × 3H), diastereomeric salts), (1.28
(s, 0.15 × 3H), 1.34 (s, 0.85 × 3H), diastereomeric salts), 1.76
(br d, 1H), 2.10-2.48 (m, 2H), 2.75-3.65 (m, 12H), 6.60-6.90
(m, 3H), 7.11 (t, J ) 7.8 Hz, 1H), 7.15-7.35 (m, 5H), 9.43 (br
s, 1H), 9.75 (br s, 1H); IR (KBr) 3174, 1732, 1620, 1586, 1276,
785, 749, 706 cm-1. Anal. Calcd for C24H32ClNO3: C, 68.97;
H, 7.72; N, 3.35; Cl, 8.48. Found: C, 69.27; H, 7.84; N, 3.42;
Cl, 8.38.
An a lytica l Da ta for F r ee Ba se of 34. Mp 81.0-84.0 °C;
[R]25D +67.1° (c 1.02, MeOH). Spectral data: 1H NMR (CDCl3)
δ 0.73 (d, J ) 7.0 Hz, 3H), 1.27 (s, 3H), 1.51 (br d, J ) 12.8
Hz, 1H), 1.89-1.96 (m, 1H), 2.25 (td, J ) 12.5, 4.5 Hz, 1H),
2.36-2.42 (m, 2H), 2.49 (dd, J ) 11.2, 2.9 Hz, 1H), 2.62 (br d,
J ) 10.9 Hz, 1H), 2.70 (dd, J ) 12.1, 9.2 Hz, 1H), 2.75 (d, J )
11.3 Hz, 1H), 2.80 (dd, J ) 13.2, 5.4 Hz, 1H), 2.88-2.98 (m,
2H), 3.55 (s, 3H), 5.04 (br s, 1H), 6.62 (dd, J ) 8.0, 2.3 Hz,
1H), 6.74 (br s, 1H), 6.81 (d, J ) 8.1 Hz, 1H), 7.13-7.20 (m,
4H), 7.24-7.27 (m, 2H); 13C NMR (CDCl3) δ 16.0, 27.4, 30.6,
36.7, 38.5, 38.9, 46.1, 50.9, 51.5, 55.3, 60.4, 112.4, 112.9, 117.9,
126.4, 128.5 (2C), 128.8 (2C), 129.2, 139.2, 152.5, 155.8, 176.0;
IR (CHCl3) 3597, 3382 (br), 2953, 2812, 1729, 919 cm-1. Anal.
Calcd for C24H31NO3: C, 75.56; H, 8.19; N, 3.67. Found: C,
75.63; H, 8.29; N, 3.66.
yield by HPLC);40 [R]25 +75.3° (c 1.01, MeOH). Spectral
D
data: 1H NMR (CDCl3) δ 0.73 (d, J ) 7.0 Hz, 3H), 1.29 (s,
3H), 1.57 (br d, J ) 12.1 Hz, 1H), 1.92-2.00 (m, 1H), 2.29 (td,
J ) 12.4, 4.3 Hz, 1H), 2.40 (td, J ) 11.6, 2.7 Hz, 1H), 2.52 (t,
J ) 7.3 Hz, 2H), 2.56-2.60 (m, 2H), 2.64-2.76 (m, 2H), 2.81-
2.85 (m, 1H), 3.66 (s, 3H), 5.50-6.20 (br s, 1H, D2O exch), 6.62
(dd, J ) 8.1, 2.3 Hz, 1H), 6.74 (br s, 1H), 6.82 (d, J ) 8.1 Hz,
1H), 7.15 (t, J ) 7.9 Hz, 1H); 13C NMR (CDCl3) δ 16.1, 27.5,
30.8, 32.0, 38.4, 38.9, 49.9, 51.7, 53.9, 55.8, 112.6, 113.2, 117.7,
129.2, 151.6, 156.1, 173.4; IR (CHCl3) 3600, 3600-3100, 1732,
1440 cm-1. Anal. Calcd for C17H25NO3: C, 70.07; H, 8.65; N,
4.81. Found: C, 67.80; H, 8.26; N, 4.73 (viscous oil which
contains 2.5 wt % THF).
Meth yl (rS,3R,4R)-4-(3-Hyd r oxyp h en yl)-3,4-d im eth yl-
r-(p h en ylm et h yl)-1-p ip er id in ep r op a n oa t e H yd r och lo-
(38) The coupling procedure with a nonaqueous workup allows for
the potential recovery and recycling of sodium brosylate. J ohn Qua-
troche, Eli Lilly and Company, unpublished results, 1987.
(39) The concentration was determined by the amount of THF added
to the crude product (assuming 100% yield). The concentration can be
determined by HPLC using the amino acid of 33 as the external
standard (preferred method) or by integration of the signals from THF
and 33 in the 1H NMR spectrum.
(40) HPLC conditions for 33-35: column, Zorbax SB-Phenyl (4.6
mm × 250 mm); mobile phase; solvent A (CH3CN), solvent B (0.1%
TFA), gradient, 30% A:70% B (5 min), increase over 10 min to 70%
A:30% B; detector, 273 nm. tR ) 5.5 min (33), 14.6 min (35), and 16.7
min (34).
(41) HPLC conditions for analysis of the diastereomers: column,
Chiralcel OD (4.6 mm x 250 mm, Regis); Mobile phase, 8% n-PrOH
and 0.5% Et2NH in hexane (1.0 mL/min); column temperature, 40 °C;
detector, 273 nm. tR ) 11.5 min ((3R,4R,RS)-isomer) and 13.3 min
((3R,4R,RR)-isomer)).

596 J . Org. Chem., Vol. 61, No. 2, 1996
Werner et al.
(rS ,3R ,4R )-4-(3-H y d r o x y p h e n y l)-3,4-d i m e t h y l-r-
(p h en ylm eth yl)-1-p ip er id in ep r op a n oic Acid Mon oh y-
d r a te (35). Compound 34 (25.0 g, 94% de, 58.0 mmol) was
added to a solution of 50% aqueous NaOH (20.0 g, 250 mmol)
in water (230 mL), and the mixture was stirred at room
temperature for 4 h. The mixture was then filtered through
No. 1 Whatman paper. Methanol (240 mL) was charged to
the filtrate,30 and the pH was adjusted to 6.0 with concentrated
hydrochloric acid (32.1 g). After the product had precipitated,
most of the methanol was removed by rotary evaporation (50
°C, 100 Torr). The slurry was stirred for 4 h, the pH was
readjusted to 6.0, and the slurry was stirred at 0 °C for 1.5 h.
The desired product was filtered and washed with water (3 ×
50 mL). After drying in an air dryer at 25 °C overnight, the
desired monohydrate product 35 was isolated as a white
granular solid (21.3 g, 94% yield on an anhydrous basis, 96%
de, 4.07% H2O by Karl Fischer analysis (calcd for monohy-
of the solvent by rotary evaporation afforded to 25.0 g of 37
(91% purity, 96% de, 92% yield) as an amorphous, tacky solid
which was used without purification: [R]25 +55.0° (c 1.01,
D
MeOH). Spectral data: 1H NMR (CDCl3) δ 0.71 (d, J ) 7.0
Hz, 3H), 0.91 (d, J ) 6.6 Hz, 6H), 1.25 (s, 3H), 1.57 (d, J )
12.7 Hz, 1H), 1.91 (septet, J ) 6.6 Hz, 1H), 1.92-2.00 (m, 1H),
2.25-2.35 (m, 2H), 2.37 (dd, J ) 12.7, 4.0 Hz, 1H), 2.45 (t, J
) 11.6 Hz, 1H), 2.57-2.71 (m, 4H), 2.83 (d, J ) 10.9 Hz, 1H),
3.34 (dd, J ) 13.9, 4.5 Hz, 1H), 3.83 (dd, J ) 18.3, 3.3 Hz,
1H), 3.90 (d, J ) 6.7 Hz, 2H), 4.19 (dd, J ) 18.3, 5.9 Hz, 1H),
5.75 (br s, 1H), 6.65 (dd, J ) 8.0, 2.3 Hz, 1H), 6.75 (t, J ) 1.7
Hz, 1H), 6.79 (d, J ) 6.1 Hz, 1H), 7.14 (t, J ) 7.9 Hz, 1H),
7.18-7.21 (m, 3H), 7.26-7.29 (m, 2H), 9.0 (br s, 1H); 13C NMR
(DMSO-d6) δ 15.8, 18.7 (2C), 27.1, 27.2, 29.9, 35.6, 37.8, 38.0,
40.6, 44.3, 49.9, 55.0, 59.7, 70.0, 112.1, 112.4, 115.9, 125.8,
128.1 (2C), 128.7 (2C), 128.8, 140.1, 151.7, 157.1, 169.8, 174.1;
IR (CHCl3) 3600, 3010, 2967, 2936, 1741, 1711, 1658 cm-1
.
drate: 4.70%)): mp 178.0-180.0 °C dec; [R]25 +95.7° (c 1.01,
Anal. Calcd for C29H40N2O4: C, 72.47; H, 8.39; N, 5.83.
Found: C, 72.21; H, 8.53; N, 6.12.
D
MeOH). Spectral data: 1H NMR (DMSO-d6) δ 0.65 (d, J )
7.0 Hz, 3H), 1.21 (s, 3H), 1.51 (d, J ) 13.2 Hz, 1H), 1.90-2.00
(m, 1H), 2.12 (td, J ) 12.8, 4.3 Hz, 1H), 2.37 (dd, J ) 10.9, 5.6
Hz, 1H), 2.43-2.53 (m, 2H), 2.60 (t, J ) 11.2 Hz, 1H), 2.65-
2.75 (m, 2H), 2.75-2.85 (m, 2H), 2.91 (dd, J ) 13.4, 7.3 Hz,
1H), 6.54 (dd, J ) 7.9, 2.0 Hz, 1H), 6.65 (br s, 1H), 6.68 (d, J
) 7.9 Hz, 1H), 7.07 (t, J ) 7.9 Hz, 1H), 7.16-7.21 (m, 3H),
7.24-7.27 (m, 2H), 9.14 (br s, 1H); 13C NMR (DMSO-d6) δ 15.5,
26.9, 29.6, 35.2, 37.5, 37.7, 42.7, 49.7, 53.7, 58.8, 112.2, 112.3,
115.9, 126.0, 128.2 (2C), 128.7 (2C), 128.9, 139.4, 151.2, 157.1,
175.1; IR (KBr) 3600, 3360, 3272, 2967, 1622, 1585, 1363, 844
cm-1. Anal. Calcd for C23H31NO4: C, 71.66; H, 8.10; N, 3.63.
Found: C, 72.29; H, 8.10; N, 3.71. Analysis is off due to
incomplete hydration (vide supra).
[[2(S)-[[4(R)-(3-Hyd r oxyp h en yl)-3(R),4-d im eth yl-1-p ip -
er din yl]m eth yl]-1-oxo-3-ph en ylpr opyl]am in o]acetic Acid
Dih yd r a te (2). One molar NaOH (77 mL, 77 mmol, 3.0 equiv)
was added slowly to a solution of 37 (12.5 g, 91% purity, 96%
de, 23.7 mmol) in 3A ethanol (315 mL) and water (74 mL) over
15 min at 25 °C (pH ) 12-13). After 45 min the solution was
neutralized (pH ) 6.0) by addition of concentrated hydrochloric
acid and seeded. The mixture was stirred at 25 °C for 2 h
and filtered, and the cake was washed with water (30 mL).
The crystals were dried overnight (25 °C, 33% relative humid-
ity) by pulling air through the product in the filter funnel
under slight suction to give 10.2 g of the dihydrate 2 (99% de,43
7.95% H2O by Karl Fischer analysis (theoretical: 7.4% H2O),
2-Meth ylp r op yl Glycin e, p-Tolu en esu lfon ic Acid Sa lt
(36). A mixture of toluene (600 mL), glycine (22.5 g, 300
mmol), p-toluenesulfonic acid monohydrate (62.8 g, 330 mmol),
and isobutyl alcohol (60 mL, 650 mmol) was heated at reflux,
and the water was removed as it was formed via a Dean-Stark
trap. After 2 h the reaction mixture was homogeneous (109
°C), and no additional water was being formed. After an
additional 1.5 h, the reaction mixture was cooled to 50 °C and
concentrated via rotary evaporation at 60 °C to a net weight
of 135 g. The residue was added to hot EtOAc (450 mL) and
hexane (450 mL), and the solution was heated to reflux and
then allowed to cool slowly (seeded at 38 °C to initiate
crystallization). After cooling at 5 °C for 1 h, the product was
filtered and dried overnight (40 °C, 5 Torr). A total of 89.1 g
of 36 (98% yield) was obtained as a white crystalline solid:
mp 77.2-79.6 °C; pKa (67% aqueous DMF) ) 7.68. Spectral
data: 1H NMR (CDCl3) δ 0.82 (d, J ) 6.9 Hz, 6H), 1.79 (septet,
J ) 6.8 Hz, 1H), 2.33 (s, 3H), 3.66 (br s, 2H), 3.78 (d, J ) 6.6
Hz, 2H), 7.10 (d, J ) 8.1 Hz, 2H), 7.72 (d, J ) 8.2 Hz, 2H),
8.03 (br s, 3H); 13C NMR (CDCl3) δ 18.9 (2C), 21.3, 27.4, 40.3,
72.0, 126.1 (2C), 128.9 (2C), 140.3, 141.4, 167.5; IR (CHCl3)
3300-2600, 3018, 2970, 1752, 1125, 1034, 1011 cm -1. Anal.
Calcd for C13H21NO5S: C, 51.47; H, 6.98; N, 4.62; S, 10.57.
Found: C, 51.74; H, 6.77; N, 4.76; S, 10.73.
[[2(S)-[[4(R)-(3-Hyd r oxyp h en yl)-3(R),4-d im eth yl-1-p ip -
er din yl]m eth yl]-1-oxo-3-ph en ylpr opyl]am in o]acetic Acid
2-Meth ylp r op yl Ester (37). Amino acid 35 (20.1 g, 96% de,
4.1% H2O, 51.4 mmol), amino ester 36 (17.6 g, 58.0 mmol),
and 1-hydroxybenzotriazole monohydrate (7.83 g, 58.0 mmol)
were added to THF (144 mL). To this solution was added Et3N
(8.08 mL, 58.0 mmol), followed by dicyclohexylcarbodiimide
(12.0 g, 58.0 mmol) dissolved in THF (60 mL). The mixture
was stirred at 25 °C under nitrogen42 for 2 days. The slurry
was cooled at 0 °C for 2 h and then filtered to remove around
98% of the 1,3-dicyclohexylurea (DCU). The filtrate was then
evaporated to near dryness (40 °C/10 Torr). The oil was taken
up in EtOAc (250 mL), and the organic layer was washed with
a pH 10 buffer solution (250 mL of a 0.5 M Na2CO3-NaHCO3
solution), followed by brine (250 mL). The organic layer was
dried over Na2SO4, cooled with stirring to -20 °C, and allowed
to set unstirred at -20 °C overnight (16 h). The precipitated
DCU and drying agent were removed by filtration. Removal
93% yield on an anhydrous basis): mp 210-213 °C; [R]25
D
+51.8° (c 1.00, DMSO). Spectral data: 1H NMR (DMSO-d6)
δ 0.73 (d, J ) 6.9 Hz, 3H), 1.26 (s, 3H), 1.53 (d, J ) 12.9 Hz,
1H), 1.98 (br s, 1H), 2.22 (td, J ) 12.5, 4.0 Hz, 1H), 2.33 (dd,
J ) 12.3, 5.5 Hz, 1H), 2.40 (t, J ) 10.7 Hz, 1H), 2.49 (d, J )
11.0 Hz, 1H), 2.62 (dd, J ) 12.3, 8.9 Hz, 1H), 2.68-2.71 (m,
2H), 2.81 (d, J ) 11.3 Hz, 1H), 2.85-2.98 (m, 2H), 3.73 (dd, J
) 17.7, 5.3 Hz, 1H), 3.76 (dd, J ) 17.7, 5.6 Hz, 1H), 6.60 (dd,
J ) 8.0, 2.0 Hz, 1H), 6.71 (br s, 1H), 6.75 (d, J ) 7.9 Hz, 1H),
7.13 (t, J ) 7.9 Hz, 1H), 7.22 (t, J ) 7.2 Hz, 1H), 7.23-7.32
(m, 4H), 8.38 (t, J ) 5.6 Hz, 1H), 9.18 (br s, 1H); 13C NMR
(DMSO-d6) δ 15.7, 27.0, 29.8, 35.6, 37.9, 38.0, 41.2, 43.9, 50.1,
54.8, 59.8, 112.3, 112.5, 116.1, 126.0, 128.2 (2C), 128.9 (2C),
129.0, 140.0, 151.5, 157.2, 171.4, 173.7; IR (KBr) 3419, 3204,
3028, 1684, 1591 cm-1. Anal. Calcd for the anhydrate of 2,
which was obtained by block drying at 120 °C prior to analysis,
C25H32N2O4: C, 70.73; H, 7.60; N, 6.60. Found: C, 70.97; H,
7.64; N, 6.36.
Ack n ow led gm en t. The authors thank Drs. Charles
Barnett, Mary Peters, Charles Mitch, and Dennis Zim-
merman for their helpful discussions on this work and
for sharing their expertise on the synthesis of related
alkaloids and Professor K. Barry Sharpless for sugges-
tions on the preparation of diol 30. We also acknowl-
edge the analytical support of Dr. Sally Anliker, Lynn
Burks, Paul Dodson, Don Hodges, Dr. Eric J ensen, J oe
Kennedy, Mike Kuhfeld, Stacy Linzie, Tom Lobdell, and
Dave Robbins; the engineering support of Mark Kamer,
J ohn Quatroche, and David Staehler; and the technical
support of Richard Hoying, Doug Prather, David Bar-
low, Henry Giddens, Allen Glasson, Kay Stephenson,
and Tony Winningham. Finally, we acknowledge J ack
Deeter and J ennifer Olkowski of the Technology Core
division for obtaining the X-ray crystal structures of 34
and 35.
Su p p or tin g In for m a tion Ava ila ble: A summary of the
results of our pilot plant scale synthesis of 7 from N-ethyl-4-
(43) HPLC chiral assay conditions for analysis of diastereomer:
column, ES-OVM; mobile phase, 10% acetonitrile in 0.02 M phosphate
buffer (0.5 mL/min); column temperature, 30 ˚C; detector, 273 nm. tR
) 8.3 min (2) and 11.3 min (RR,3R,4R isomer).
(42) An orange impurity is formed if any oxygen is present.

Synthesis of Piperidine Opioid Antagonists
J . Org. Chem., Vol. 61, No. 2, 1996 597
piperidinone along with a brief description of the modifications
made to the literature procedures; experimental procedures
for carbonates 13, 19, 22a , and 22c-g and the propionate ester
ordered from the ACS; see any current masthead page for
ordering information. The authors have deposited atomic coor-
dinates for compounds 34 and 35 with the Cambridge Crystal-
lographic Data Centre. The coordinates can be obtained, on
request, from the Director, Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, U.K.
1
of alcohol 23; and the H NMR spectra for compounds 1, 2, 7,
13, 19, 22a , 22c-g, 23-28b, 30-37, and the (3R,4R,RR)-
isomers of 34, 35, 37, and 2 (42 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
J O951403Y