Synthesis and resolution of the benzomorphan-based kappa opioid agonist bremazocine

Article abstract of DOI:10.1055/s-2007-966026

An efficient synthesis of racemic bremazocine hydrochloride salt was developed. The racemic material was resolved using (+)- and (-)-3- bromocamphorsulfonic acid ammonium salt to give optically pure samples of (-)- and (+)-bremazocine. Georg Thieme Verlag

Full text of DOI:10.1055/s-2007-966026

PAPER
1481
Synthesis and Resolution of the Benzomorphan-Based Kappa Opioid Agonist
Bremazocine
B
remazoc
a
ine Synthes
m
is es B. Thomas, Lawrence E. Brieaddy, Karl G. Boldt, Carin Perretta, F. Ivy Carroll*
Center for Organic and Medicinal Chemistry, Research Triangle Institute, P. O. Box 12194, Research Triangle Park, NC 27709, USA
Fax +1(919)5418868; E-mail: fic@rti.org
Received 21 December 2006; revised 4 March 2007
OH
Abstract: An efficient synthesis of racemic bremazocine hydro-
chloride salt was developed. The racemic material was resolved us-
ing (+)- and (–)-3-bromocamphorsulfonic acid ammonium salt to
give optically pure samples of (–)- and (+)-bremazocine.
N
N
HO
O
Key words: bremazocine, kappa opioid agonist, synthesis, medici-
nal chemistry, receptors
OH
OH
(–)-2
1
N
Morphine (1) remains the treatment of choice for relief of
severe pain, but its side-effect profile, including respirato-
ry depression and addiction, has prompted researchers to
seek alternative remedies.1 One strategy employed to this
end has been the synthesis and testing of the various part
structures of morphine.2 Over the years, this has lead to
the discovery of numerous novel structures showing anal-
gesic activity. The benzomorphans, represented by brem-
azocine [(–)-2] and metazocine [(–)-3], are one class of
compound that arose through structural simplification of
morphine (Figure 1).3 Recently we required (–)-2 and its
enantiomer. Our previous experience working with brem-
azocine taught us that the resolution of the racemic mate-
rial was straightforward but the reported synthesis was
quite long (13 steps).4 For this reason, we sought a shorter
method of preparing racemic bremazocine. Much has
been written with regard to entries into the benzomorphan
system.5 Of particular interest to us, however, was the re-
cent synthesis of metazocine by Comins and Zhang.6 In
OH
(–)-3
Figure 1 Compounds 1–3
O
N
N
OMe
CO2Ph
5
(–)-3
OH
O
N
OMe
CO2Ph
4
Scheme 1
this work, the N-acyl-2-benzyl-dihydropyridone
4
(Scheme 1) was alkylated to give the methylated interme-
diate 5 that was ultimately converted to metazocine (–)-3.
In light of these findings, it seemed reasonable that di-
methylation of 4 might produce a comparable dimethyl in-
termediate that could then be converted to bremazocine
[(–)-2]. In practice, however, we were not able to identify
conditions that would provide such an intermediate start-
ing from 4. Nevertheless, we did find that the more syn-
thetically robust N-methyl compound 6 (Scheme 2) was
sufficiently stable to give the desired dialkyl intermediate
8 that was then converted to bremazocine as described
herein.
O
N
N
HO
OMe
OH
(–)-2
8
O
N
OMe
6
Scheme 2
We began this work with 2-(4-methoxybenzyl)-1-methyl-
2,3-dihydropyridin-4(1H)-one (6) that we obtained ac-
cording to the method previously described by Takeda
and May (Scheme 3).7 From an electronic perspective,
this intermediate is ideally suited to introduce the required
geminal dimethyl group found in (–)-2 via anion alkyla-
tion chemistry. The vinylogous amide character of the
carbonyl group controls the regiochemistry of alkylation
directing it solely to the saturated a-carbon. In attempting
SYNTHESIS 2007, No. 10, pp 1481–1484
x
x.
x
x
.
2
0
0
7
Advanced online publication: 02.05.2007
DOI: 10.1055/s-2007-966026; Art ID: M07706SS
© Georg Thieme Verlag Stuttgart · New York

1482
J. B. Thomas et al.
PAPER
the dialkylation, we found that we were unable to intro- alcohol solvent, preparation of the potassium acetylide,
duce both of the methyl groups in a single-pot reaction in and drying and purification of acetylene. Treatment of 9
good yield. Indeed while the first methylation went with ethylmagnesium chloride gave none of the desired
smoothly to give 7, using sodium hydride as the base in product but instead gave reduction of the ketone. Treat-
THF, the second methylation required lithium diisopropyl- ment of 9 with ethyllithium gave the desired adduct 10,
amide or an equivalently strong base to give a good yield. but the reaction could not be forced to completion. The re-
Lithium pyrrolidide worked equally well. It was this re- action gave product 10 and recovered starting material 9
quirement for a strong base that precluded the use of com- in nearly equal portions. The acetylide apparently acts like
pound 5 in the synthesis since the strong base inevitably a more compact nucleophile than ethyl an effect that is
reacted with the N-protecting group to give a urea by- probably magnified in this instance due to the adjacent
product. Protecting the nitrogen with a methyl group in- gem-dimethyl group. Use of cerium chloride did not im-
stead worked well since this side reaction was precluded. prove the yield. Overall, the use of ethyllithium is more
Despite the requirement for the stronger base, quenching convenient than the acetylide method and would be the re-
the anion derived from 7 with methyl iodide gave 8 in agent type of choice if one were preparing a library of
good yield.
compounds requiring introduction of diversity at position
5. However, for the preparation of bremazocine (2), the
acetylide procedure is preferred.
Reduction of the vinylogous amide in 8 was accomplished
using lithium aluminum hydride to give the piperidone in-
termediate 9. The conversion of 8 to give 9 intercepts the Heating the adduct 10 with phosphoric acid gave the ben-
original synthesis of bremazocine [(–)-2] described by zomorphan scaffold 11 in 88% yield as reported previous-
Akkerman and Janssen.4 Use of LiAlH4 provided the de- ly. In our hands, this reaction is easily reproducible and
sired ketone 9 as well as some of the alcohol correspond- under these conditions, the methyl phenyl ether was
ing to over-reduction. In the original work, this cleaved to the phenol. In other work with opioid ligands,
intermediate provided the functionality to introduce the 5- we have found that removing N-methyl groups in the pres-
ethyl substituent and concomitantly produce the Grewe ence of phenols can be problematic. In the present work,
cyclization precursor 10. In that work, the ethyl group was the use of cyanogen bromide described in the original syn-
introduced by a two-step procedure beginning with the ad- thesis failed to give good yields of the N-demethylated
dition of potassium acetylide to the ketone 9 followed compound 12. However, we found that we could remove
thereafter by reduction to an ethyl group using hydrogen the N-methyl substituent in good yield using 10 equiva-
and palladium on carbon. We reproduced the original lents of 1-chloroethylchloroformate (ACE-Cl) in chloro-
work to obtain an 84% yield of 10 over the two steps. form with sodium bicarbonate as the scavenger base.
However, we also examined other methods that did not re- Intermediate 12 was converted to racemic bremazocine
quire the use of potassium metal, drying of the tert-butyl ( )-2 using the much-improved approach recently report-
O
O
N
O
N
OMe
OMe
OMe
a
b
N
79%
68%
6
7
8
c
66%
OMe
O
N
Et
OH
N
OMe
g
d,e or f
84% or 42%
N
88%
11
OH
10
9
88%
N
h
N
HO
H
i,j
53%
k
(+)- and (–)-2
OH
OH
12
( )-2
Scheme 3 Reagents and conditions: (a) NaH, THF, MeI; (b) lithium diisopropylamide, THF, MeI; (c) LiAlH4, THF; (d) potassium acetylide;
(e) H2, Pd/C; (f) EtLi, THF; (g) 85% H3PO4; (h) ACE-Cl, MeOH; (i) 1-hydroxy-1-cyclopropane carboxylic acid, BOP, Et3N; (j) LiAlH4; (k)
resolve with (+)- or (–)-3-bromocamphorsulfonic acid.
Synthesis 2007, No. 10, 1481–1484 © Thieme Stuttgart · New York

PAPER
Bremazocine Synthesis
1483
ed by Greiner and Rice.8 Their two-step procedure elimi-
nated the need to use the volatile and difficult to prepare
reagent 1-oxaspiro[2.2]pentane. Overall, the yield for the
two-step procedure was 58%. Obtaining optically pure
(+)- or (–)-2 via the resolution of ( )-2 was straightfor-
ward and accomplished by crystallization of racemic 2 as
its 3-bromocamphorsulfonic acid salt.
2-(4-Methoxybenzyl)-1,3,3-trimethyl-2,3-dihydropyridin-
4(1H)-one (8)
BuLi (65 mL, 0.161 mol, 2.5 M in hexane) was added to a solution
of (i-Pr)2NH (8.6 g, 0.085 mol) in THF (200 mL) at –78 °C. The
mixture was stirred for 30 min and compound 7 (18.1 g, 0.074 mol)
in THF (200 mL) was added. The mixture was stirred at –78 °C for
1 h and MeI (41.9 g, 0.29 mol) in THF (100 mL) was added. The
mixture was stirred at –78 °C for 1 h, and then at r.t. for 1 h. It was
then was quenched with aq sat. NH4Cl (100 mL) and diluted with
brine (200 mL). The organic layer was separated, dried (Na2SO4),
and concentrated to give 21.2 g of a dark red oil. The oil was chro-
matographed on silica gel using EtOAc as the eluent to yield 13.0 g
of 8 (68%) as an orange-yellow solid.
In summary, a new synthesis of (+)- and (–)-bremazocine
(2) was developed that involved ten steps starting with 2-
(4-methoxybenzyl)-1-methyl-2,3-dihydropyridin-4(1H)-
one (6).7 This route eliminates four chemical steps of the
originally reported synthesis. An important feature of the
new route was the dimethylation of 6 to provide the key
1H NMR (300 MHz, CDCl3): d = 1.17 (d, J = 6.0 Hz, 3 H), 2.45 (s,
3 H), 2.78 (m, 1 H), 3.05 (t, J = 12.0 Hz, 1 H), 3.13 (m, 1 H), 3.80
intermediate 2-(4-methoxybenzyl)-1,3,3-trimethyl-2,3-di- (s, 3 H), 4.87 (d, J = 9.0 Hz, 1 H), 6.83 (m, 3 H), 7.08 (d, J = 9.0 Hz,
2 H).
hydropyridin-4(1H)-one (8). In addition ACE-Cl demeth-
13C NMR MHz, CDCl3): d = 20.4, 26.9, 30.5, 43.8, 44.6, 55.6, 71.8,
ylation is superior to the previously described cyanogen
bromide for removing the N-methyl group in intermediate
11. Both enantiomers of bremazocine are obtained on
crystallization of the (+)- and (–)-3-bromocamphorsulfon-
ic acid salt of the racemic bremazocine.
95.3, 114.3, 130.6, 131.3, 150.6, 158.7, 197.5.
Anal. Calcd for C16H21NO2: C, 74.10; H, 8.16; N, 5.40. Found: C,
73.82; H, 8.22; N, 5.30.
2-(4-Methoxybenzyl)-1,3,3-trimethylpiperidin-4-one (9)
Compound 8 (5.0 g, 0.019 mol) was added to LiAH4 (0.36 g, 0.009
mol) in THF (80 mL), chilled in an ice/salt bath. The mixture was
stirred for 30 min and quenched with aq sat. NH4Cl (50 mL) and fil-
tered. The solids were washed with Et2O (3 × 50 mL) and the filtrate
was combined with the THF solution and concentrated. The residue
(5.1 g) was chromatographed on silica gel using 50% EtOAc–hex-
anes as the eluent to yield 3.3 g of 9 (66%) as an orange oil. A more
polar by-product (1.27 g, 25%) was also isolated that had the mass
expected for the reduced ketone.
Melting points were determined on a Thomas-Hoover capillary tube
apparatus and are not corrected. Elemental analyses were obtained
by Atlantic Microlabs, Inc. and are within 0.4% of the calculated
values. All optical rotations were determined at the sodium D line
using a Rudolph Research Autopol III polarimeter (1-dm cell).
Mass spectral data were obtained using a Finnegan LCQ electro-
spray mass spectrometer in positive ion mode at atmospheric pres-
sure. 1H NMR spectra were determined on a Bruker 300
spectrometer using tetramethylsilane as an internal standard. Silica
gel 60 (230–400 mesh) was used for all column chromatography.
All reactions were followed by TLC using Whatman silica gel 60
TLC plates and were visualized by UV, charring using 5% phospho-
molybdic acid in ethanol, or iodine stain. All solvents were reagent
grade. Anhyd THF and Et2O were purchase from Aldrich Chemical
Co.
1H NMR (300 MHz, CDCl3): d = 1.04 (s, 3 H), 1.32 (s, 3 H), 2.33
(s, 3 H), 2.47–2.77 (m, 6 H), 2.91–2.97 (m, 1 H), 3.78 (s, 3 H), 6.82
(d, J = 8.7 Hz, 2 H), 7.10 (d, J = 8.4 Hz, 2 H).
LCMS (APCI): m/z = 262.2 [M + H]+.
4-Ethyl-2-(4-methoxybenzyl)-1,3,3-trimethylpiperidin-4-ol (10)
Compound 9 (8.15 g, 0.031 mol) in THF (90 mL) was added to an
ice-chilled solution of EtLi (120 mL, 0.060 mol, 0.5 M). The bath
was allowed to come to r.t. and the mixture was stirred at r.t. over-
night. The reaction was quenched with brine and extracted with
EtOAc (2 × 150 mL). The organic layers were separated, dried
(Na2CO3), and concentrated to give 8.6 g of an orange oil. The oil
was chromatographed on silica gel, using 50% (80% CHCl3, 18%
MeOH, 2% NH4OH) in EtOAc, to yield 10 as two diastereomeric
alcohols 3.05 g (34%) and 0.7 g (8%). Starting material 9 (4.4 g,
54%) was also isolated.
1H NMR (300 MHz, CDCl3): d = 0.90 (t, J = 4.5, 9.0 Hz, 3 H), 1.00
(d, J = 6.0 Hz, 6 H), 1.45–1.48 (m, 1 H), 1.54 (br s, 1 H), 1.57 (br
s, 1 H), 1.64–1.71 (m, 2 H), 2.27 (t, J = 6.0, 15.0 Hz, 1 H), 2.75–
2.78 (m, 2 H), 2.96 (dd, J = 3.0, 6.0 Hz, 1 H), 3.78 (s, 3 H), 6.84 (d,
J = 6.0 Hz, 2 H), 7.14 (d, J = 5.1 Hz, 2 H).
2-(4-Methoxybenzyl)-1,3-dimethyl-2,3-dihydropyridin-4(1H)-
one (7)
To NaH (9.2 g, 0.23 mol, 60%) in THF (200 mL), in an ice bath, was
added 2-(4-methoxybenzyl)-1-methyl-2,3-dihydropyridin-4(1H)-
one (6; 21.0 g, 0.091 mol) in THF (200 mL). The mixture was
stirred at bath temperature for 45 min and then MeI (42.6 g, 0.30
mol) in THF (50 mL) was added. The bath was removed and the
mixture was stirred overnight under N2. Brine (200 mL) and EtOAc
(100 mL) were added to the mixture. The organic layers were sepa-
rated, dried (Na2SO4) and concentrated. The residue was purified by
column chromatography on silica gel, using 95% EtOAc–MeOH as
the eluent, to afford 17.7 g of 7 (79%) as an orange oil. The dimeth-
ylated ketone 8 (1.8 g, 7.6%) was also isolated from the column.
1H NMR (300 MHz, CDCl3): d = 1.13 (d, J = 6.0 Hz, 3 H), 2.18 (m,
1 H), 2.76 (m, 1 H), 2.91 (s, 3 H), 2.96 (m, 1 H), 3.18 (t, J = 7.5 Hz,
1 H), 3.79 (s, 3 H), 4.87 (d, J = 6.0 Hz, 1 H), 6.85 (m, 3 H), 7.08 (d,
J = 9.0 Hz, 2 H).
13C NMR (75 MHz, CDCl3): d = 17.6, 33.0, 42.6, 42.8, 55.2, 67.2,
94.9, 114.1, 129.3, 130.4, 150.8, 158.5, 195.0.
LCMS (APCI): m/z = 292.3 [M + H]+.
6-Ethyl-3,11,11-trimethyl-1,2,3,4,5,6-hexahydro-2,6-methano-
3-benzazocin-8-ol (11)
Compound 10 (4.3 g, 0.0147 mol) and H3PO4 (40 mL, 85%) were
stirred at reflux for 21 h. The reaction mixture was cooled to r.t.,
added to 150 mL of 5 N NaOH and 150 mL of saturated NaHCO3.
The mixture was extracted with Et2O, dried (Na2CO3), and concen-
trated to give 3.2 g (88%) of 11 as a tan solid.
Anal. Calcd for C15H19NO2: C, 73.44; H, 7.81; N, 5.71. Found: C,
73.20; H, 7.83; N, 5.53.
1H NMR (300 MHz, CDCl3): d = 0.96 (s, 3 H), 1.03 (t, J = 9.0, 15.0
Hz, 3 H), 1.29 (s, 3 H), 1.88 (m, 2 H), 2.03 (m, 2 H), 2.31 (s, 3 H),
Synthesis 2007, No. 10, 1481–1484 © Thieme Stuttgart · New York

1484
J. B. Thomas et al.
PAPER
2.37 (m, 2 H), 2.67 (m, 2 H), 2.97 (d, J = 18.0 Hz, 1 H), 6.60 (dd,
J = 3.0, 9.0 Hz, 2 H), 6.73 (s, 3 H), 6.94 (d, J = 7.5 Hz, 1 H).
recrystallized from MeOH–Et2O twice to yield 0.345 g (–)-brema-
zocine [(–)-2]; mp 244–245 °C; [a]D –105.9 (c = 0.37, MeOH).
LCMS (APCI): m/z = 260.6 [M + H]+.
Anal. Calcd for C20H29NO2⋅HCl⋅0.25H2O: C, 67.39; H, 8.63; N,
3.93. Found: C, 67.37; H, 8.47; N, 3.95.
6-Ethyl-11,11-dimethyl-1,2,3,4,5,6-hexahydro-2,6-methano-3-
benzazocin-8-ol (12)
(+)-Bremazocine [(+)-2]
1-Chloroethyl chloroformate (24.0 g, 0.165 mol) was added to 11
(5.34 g, 0.021 mol) and NaHCO3 (13.9 g, 0.17 mol) in CHCl3 (200
mL). The mixture was stirred at reflux overnight and concentrated
to afford 8.4 g of a foam. MeOH (250 mL) and concd HCl (3 mL)
were added, and the mixture was stirred at reflux overnight. The so-
lution was concentrated and chromatographed on silica gel using
50% (80% CHCl3, 18% MeOH, 2% NH4OH) in EtOAc to afford 5.1
g (88%) of 12 as a foam.
1H NMR (300 MHz, CDCl3): d = 0.93 (s, 3 H), 1.02 (t, J = 6.0, 15.0
Hz, 3 H), 1.25 (s, 3 H), 1.80–1.91 (m, 1 H), 2.59 (td, J = 3.0, 9.0 Hz,
1 H), 2.71 (d, J = 3.0 Hz, 2 H), 2.76 (s, 1 H), 2.82 (s, 1 H), 3.21 (dd,
J = 9.0, 18.0 Hz, 1 H), 4.0 (br s, 1 H), 6.58 (dd, J = 6.0, 9.0 Hz, 1
H), 6.69 (s, 1 H), 6.93 (d, J = 9.0 Hz, 1 H).
The filtrate B was free based (with NaHCO3), then converted to
1.14 g of the HCl salt. This salt and (–)-3-bromocamphorsulfonic
acid ammonium salt (1.06 g, 0.0032 mol) were heated in H2O (75
mL). Initially, the solution turned clear but within minutes crystals
appeared after cooling to r.t.; filtration gave 1.01 g of salt B. Solid
B was triturated in warm H2O (75 mL), cooled, and filtered to yield
0.71 g of a solid (mp 244–245 °C). This solid was free-based (with
NaHCO3) to yield 0.36 g of a white foam. The base was dissolved
in Et2O and acidified with ethereal HCl to afford 0.40 g of a white
solid. The salt was recrystallized from MeOH–Et2O twice to give
0.29 g of (+)-bremazocine [(+)-2]; mp 240–241 °C; [a]D +100.0
(c = 0.325, MeOH).
The spectral properties were identical to those of (–)-bremazocine.
LCMS (APCI): m/z = 246.4 [M + H]+.
Anal. Calcd for C20H29NO2⋅HCl: C, 68.26; H, 8.59; N, 3.98. Found:
C, 67.97; H, 8.65; N, 3.95.
( )-Bremazocine [( )-2]
1H NMR (300 MHz, CD3OD): d = 0.86 (m, 1 H), 1.07–1.10 (m, 3
H), 1.12 (d, J = 3.0 Hz, 1 H), 1.33 (t, J = 6.0, 12.0 Hz, 6 H), 1.57
(dd, J = 6.0, 15.0 Hz, 1 H), 1.68 (s, 2 H), 2.22 (m, 2 H), 2.53 (td,
J = 6.0, 12.0 Hz, 1 H), 3.30 (dd, J = 3.0, 6.0 Hz, 1 H), 3.40 (m, 2 H),
3.51 (d, J = 15.0 Hz, 1 H), 3.69 (d, J = 12.0 Hz, 1 H), 3.96 (d, J =
6.0 Hz, 1 H), 6.91 (dd, J = 3.0, 6.0 Hz, 1 H), 7.01 (d, J = 3.0 Hz, 1
H), 7.27 (d, J = 6.0 Hz, 1 H). The 1H NMR spectrum was identical
to that of an authentic sample obtained from Sigma/RBI (Natick,
MA).
Acknowledgment
The work was supported by the National Institute on Drug Abuse on
project DA-3-7736.
References
(1) Aldrich, J. V. In Burger’s Medicinal Chemistry and Drug
Discovery, 6th ed, Vol. 6; Abraham, D. J., Ed.; Wiley: New
York, 2003, Chap. 7.
(2) Eddy, N. B.; May, E. L. Synthetic Analgesics; Pergamon:
Oxford, 1966.
(3) Grewe, R.; Mondon, A. Chem. Ber. 1948, 81, 279.
(4) Akkerman, A. M.; Janssen, P. A. J. German Patent 2027077,
1970; Chem. Abstr. 1971, 74, 125486.
(5) Palmer, D. C.; Strauss, M. J. Chem. Rev. 1977, 77, 1.
(6) Comins, D. L.; Zhang, Y.; Joseph, S. P. Org. Lett. 1999, 1,
657.
Resolution of Bremazocine
(–)-Bremazocine [(–)-2]
A mixture of (+)-3-bromocamphorsulfonic acid ammonium salt
(1.86 g, 0.0057 mol) and ( )-bremazocine HCl (2.0 g, 0.00568 mol)
in H2O (135 mL) was heated and stirred until a clear solution was
obtained. The flask was cooled, and then transferred to the refriger-
ator for 4 h. The solids were filtered and dried to obtain 1.25 g (A)
and filtrate (B). Solid A was recrystallized from 125 mL of hot H2O
to get 1.00 g of salt A (mp 244–256 °C). Salt A was free-based (with
NaHCO3) to get 0.40 g of a foam. The base was dissolved in Et2O
and acidified with ethereal HCl to give a white solid. The salt was
(7) Takeda, M.; Jacobson, A.; Kanematsu, K.; May, E. L. J.
Org. Chem. 1969, 34, 4154.
(8) Greiner, E.; Folk, J. E.; Jacobson, A.; Rice, K. C. Bioorg.
Med. Chem. 2004, 12, 233.
Synthesis 2007, No. 10, 1481–1484 © Thieme Stuttgart · New York