Novel cleavage reaction of the C16-N17 bond in naltrexone derivatives

Article abstract of DOI:10.1016/j.tetlet.2008.08.037

A dealkylation reaction of tertiary amines using chloroformate was a useful method for synthesizing morphinan derivatives without 17-substituents; however, the reaction has been applied to only the 14-hydromorphinans. In the course of the investigation of the 17-dealkylation reaction in 14-hydroxymorphinan, the novel cleavage reaction of the C16-N17 bond in the naltrexone derivative was found. A plausible reaction mechanism based on the stereoelectronic effect is presented. The examinations of dealkylation reactions in general tertiary amines ranked the tendency of cleavage as follows: benzyl > cyclopropylmethyl (CPM) ≈ allyl > methyl, ethyl. The preferable cleavage of the CPM group may be explained by the polarization of CPM-N17 bond due to a postulation of extreme stability of the cyclopropylcarbinyl cation.

Full text of DOI:10.1016/j.tetlet.2008.08.037

Tetrahedron Letters 49 (2008) 6293–6296
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Novel cleavage reaction of the C16–N17 bond in naltrexone derivatives
*
Hideaki Fujii, Satomi Imaide, Akio Watanabe, Toru Nemoto, Hiroshi Nagase
School of Pharmacy, Kitasato University, 5-9-1, Shirokane, Minato-ku, Tokyo 108-8641, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 July 2008
Revised 8 August 2008
Accepted 11 August 2008
Available online 14 August 2008
A dealkylation reaction of tertiary amines using chloroformate was a useful method for synthesizing
morphinan derivatives without 17-substituents; however, the reaction has been applied to only the
14-hydromorphinans. In the course of the investigation of the 17-dealkylation reaction in 14-hydroxy-
morphinan, the novel cleavage reaction of the C16–N17 bond in the naltrexone derivative was found.
A plausible reaction mechanism based on the stereoelectronic effect is presented. The examinations of
dealkylation reactions in general tertiary amines ranked the tendency of cleavage as follows:
benzyl > cyclopropylmethyl (CPM) ꢀ allyl > methyl, ethyl. The preferable cleavage of the CPM group
may be explained by the polarization of CPM–N17 bond due to a postulation of extreme stability of
the cyclopropylcarbinyl cation.
Keywords:
Dealkylation
Chloroformate
Stereoelectronic effect
Morphinan
Ó 2008 Elsevier Ltd. All rights reserved.
Most therapeutically valuable opioids have a common struc-
ture, the morphinan scaffold. These morphinans include powerful
pain relieving agents such as naturally occurring alkaloids (e.g.,
morphine, codeine), semisynthetic derivatives (e.g., oxymorphone,
buprenorphine), and synthetic analogs (e.g., levorphanol, butor-
phanol). The 17-substituent of morphinans is believed to influence
the agonistic or antagonistic character for the opioid receptor. For
example, 17-methyl morphinans, morphine and oxymorphone are
agonists. On the other hand, 17-allyl and 17-cyclopropylmethyl
(CPM) morphinans, naloxone and naltrexone are antagonists.1
Therefore, a dealkylation, especially demethylation, reaction of
17-substituents has been widely used to synthesize morphinan
derivatives possessing various 17-substituents.2–8 However, the
reaction with chloroformates has been applied for only 14-hydro-
morphinans. To the best of our knowledge, the application of the
reaction to 14-hydroxymorphinan has not been reported. There-
fore, we were interested in the 17-dealkylation reaction in 14-
hydroxymorphinans such as naltrexone. In the course of the study,
we found a novel cleavage reaction of the C16–N17 bond in nal-
trexone derivative 1 that afforded oxazolidinone 2. This novel
cleavage reaction is interesting from the viewpoint not only of or-
ganic chemistry but also of medicinal chemistry, because the reac-
tion could open the door for the synthesis of the morphinan
derivatives without piperidine ring, which may be utilized to lead
we report the novel cleavage reaction of the C16–N17 bond in nal-
trexone derivative 1 to afford oxazolidinone derivative 2 and the
scission reaction that favors the CPM group over the methyl and
ethyl groups in compound 9.
Although we attempted to dealkylate the CPM group in naltrex-
one using 1-chloroethyl chloroformate (ACE-Cl) under the original
reaction conditions,8 the reaction hardly proceeded. After exten-
sive investigations, we found that the piperidine ring of naltrexone
derivative 1 was cleaved with ACE-Cl (9 equiv) in pyridine to give,
surprisingly, oxazolidinone 210 in 60% yield concomitantly with
starting material 1, but the objective carbamate derivative 3 was
not obtained (Scheme 1, condition (a), structure of compound 3
as shown in Fig. 1). We focused on the abnormal cleavage reaction
of the C16–N17 bond which afforded oxazolidinone 2, and at-
tempted to improve the yield of the compound 2 and also to exam-
ine the reaction mechanism.
We developed the working hypothesis that at first a reaction
intermediate, oxooxazolidinium 4 (Figs. 1 and 2), would be formed
and then chloride ion would attack C16. If enough chloride ions
O
OH
14
O
17
N
N
O
O
a), b), c), or d)
a
l
opioid receptor partial agonist morphine9 to full agonistic
Cl
16
O
O
O
O
derivatives. We also observed a preferential cleavage of the CPM
group in general tertiary amine 9 under the dealkylation condi-
tions with 2,2,2-trichloroethyl chloroformate (Troc-Cl). Herein,
OMe
OMe
1
2
Scheme 1. Reagents and conditions: (a) ACE-Cl (9 equiv), pyridine, 100 °C, 6 h, 60%;
(b) ACE-Cl (22 equiv), pyridine, 100 °C, 1.5 h, 80%; (c) ACE-Cl (5.6 equiv), K2CO3,
1,1,2,2-tetrachloroethane, 150 °C, 21 h, 93%; (d) Troc-Cl (5.6 equiv), K2CO3, 1,1,2,2-
tetrachloroethane, 150 °C, 22 h, 70%.
* Corresponding author. Tel.: +81 3 5791 6372; fax: +81 3 3442 5707.
E-mail address: nagaseh@pharm.kitasato-u.ac.jp (H. Nagase).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.08.037

6294
H. Fujii et al. / Tetrahedron Letters 49 (2008) 6293–6296
the carbamate moiety of the intermediate 4 was fixed by the rigid
ring structure. Therefore, the -bond between the 17-nitrogen and
the CPM group in 4 would be fixed almost orthogonally to the
O
N
r
OH
O
O
N
p-
O
O
orbital of the carbonyl group in the oxazolidinone ring and, thus,
may not be conjugated with the carbonyl group of oxazolidinone
ring and not be activated by the carbonyl group. In contrast, the
Cl
O
O
O
O
O
OMe
OMe
3
4
C16–N17 bond would be parallel to the
group in the oxazolidinone ring (Fig. 2B). The
N17 bond could be conjugated with the carbonyl
p
-orbital of the carbonyl
-orbital of C16–
-orbital effec-
r
Figure 1.
p
tively to accelerate the cleavage reaction of the C16–N17 bond
with chloride ion. As a result, the C16–N17 bond would be stereo-
electronically preferred for cleavage as opposed to the bond be-
tween the 17-nitrogen and the CPM group to give oxazolidinone
2 (Fig. 2B). The Newman projection formula of intermediate 4 (C)
indicated the definitive difference of the positional relationship be-
would be present in the reaction system, the ion might attack C16
to easily cleave the C16–N17 bond. When sufficient free chloride
ion does not exist, the cleavage reaction could not proceed rapidly
enough, and the remaining intermediate 4 would be hydrolyzed to
give starting material 1 during workup. Although pyridine may
neutralize hydrochloric acid derived from the reaction of 1 with
ACE-Cl, the resulting pyridinium chloride may not give much free
chloride ion in the solution, because pyridine is a weak base and
pyridinium chloride cannot dissociate effectively into free chloride
ion and pyridine. On the basis of the above working hypothesis, we
used a large excess of ACE-Cl (22 equiv) in pyridine to access the
objective compound 2 in 80% yield (Scheme 1, condition (b)). We
next tried to use potassium carbonate as a base. Reaction of potas-
sium carbonate with hydrochloric acid gives potassium chloride
and carbonic acid, which will decompose to give carbon dioxide
and water. In fact, starting naltrexone derivative 1 was treated
with ACE-Cl (5.6 equiv) in the presence of potassium carbonate
(9 equiv) in 1,1,2,2-tetrachloroethane to give the objective oxazo-
lidinone 2 in 93% yield (Scheme 1, condition (c)). The reaction of
1 with Troc-Cl also afforded 2 in 70% yield (Scheme 1, condition
(d)).
tween the carbonyl
p-orbital and the CPM–N17 or the C16–N17
bond. On the other hand, in intermediate 5 of 14-hydromorphinan,
the CPM group would be preferentially cleaved because the N–C
bond of the carbamate moiety could rotate freely and the
r-bond
between the 17-nitrogen and the CPM group could be activated
(Fig. 2D).
Reaction of acyclic amine 6 with Troc-Cl afforded oxazolidinone
derivative 711 without CPM group in 65% yield (Scheme 2). In the
case of compound 6, the methyl group is less hindered than the
CPM group and was predicted to be preferentially attacked by a
O
Me
O
OH
N
N
O
O
Me
O
O
O
O
We examined why the C16–N17 bond cleavage occurred with-
out removal of the CPM group. Figure 2 shows respective interme-
diate 4 ((A) and (B)) or 5 (D) of 14-hydroxy or 14-hydromorphinan
and their Newman projection formulas ((C) and (E)), and includes a
postulated reaction mechanism of the bond cleavage. In Figure 2A,
OMe
OMe
7
6
Scheme 2. Reagents and conditions: Troc-Cl (10 equiv), K2CO3, 1,1,2,2-tetrachloro-
ethane, 6 h, 150 °C, 65%.
(A)
(B)
(C)
O
O
O
O
CPM
O
N
O
14
14
Cl
17
N
N
9
9
16O
O
17
C16
C9
OMe
OMe
Cl
4
4
H
(D)
(E)
O
Cl
14
C16
O
C9
O
O
17N
9
16
O
N
CPM
OMe
5
Cl
Figure 2. Respective intermediate 4 ((A), (B)) or 5 (D) of 14-hydroxy or 14-hydromorphinan and their Newman projection formulas ((C) and (E)), and postulated reaction
mechanism of the bond cleavage. Dioxolane moieties were omitted for clarity. In the intermediate 4, the CPM–N17 bond seems to be nearly orthogonal to the -orbital
(green) of the carbonyl group ((A) and (C)); however, the C16–N17 bond seems to be parallel to the -orbital (green) ((B) and (C)). The CPM–N17 bond in intermediate 5 could
be parallel to the -orbital (green) ((D) and (E)).
p
p
p

H. Fujii et al. / Tetrahedron Letters 49 (2008) 6293–6296
6295
O
N
O
O
N
O
OH
Me
N
O
O
O
a)
Cl
O
O
O
O
O
O
OMe
OMe
OMe
8
2
1
O
N
O
N
Figure 3.
O
O
b)
O
nucleophile in a bimolecular nucleophilic substitution reaction.
Nevertheless, the CPM group was preferentially removed over
the methyl group to afford oxazolidinone 7. In the reaction inter-
mediate 8 (Fig. 3), because the conformation of 8 was more flexible
than that of 4, both the nitrogen–methyl and nitrogen–CPM bonds
Cl
Cl
O
O
OMe
OMe
11
12
O
N
O
could be oriented parallel to the carbonyl
p-orbital and thus be
O
O
activated. Therefore, the stereoelectronic effect could not explain
the preferential cleavage of the CPM group. It may be explained
by postulating the extreme stability of the cyclopropylcarbinyl cat-
ion.12–15 Generally speaking, the CPM tosylate is solvolyzed 106
faster than isobutyl tosylate because the tosylate ion could be
solvolitically eliminated from the CPM tosylate to afford stable
cyclopropylcarbinyl cation.14 The polarization of the CPM–N17
bond probably due to the stability of tentatively formed cyclopro-
pylcarbinyl cation may give the carbon of the CPM group a partial
positive charge,16,17 and as a result a nucleophile could easily
attack on the carbon of the CPM group. We examined the dealkyl-
ation reaction of tertiary amines 9 with Troc-Cl to confirm the pref-
erential elimination of CPM group. A similar investigation has
already been reported using the compounds shown in Figure 4.18
However, no example of the CPM group was found. The order of re-
moval was reported as: benzyl > allyl > methyl. So, we examined
the order of reactivity among CPM, benzyl, allyl, ethyl, and methyl
groups using model compounds 9 (Table 1). The stabilizing effect
of a cyclopropyl group on a carbocation center was known to be
larger than that of a vinyl group.19 Although controversy exists
as to a relative stabilizing effects of cyclopropyl and phenyl
N
O
O
c), d)
HO
O
O
O
O
OMe
OMe
13
14
Scheme 3. Reagents and conditions: (a) NaOH, DMSO, 115 °C, 50%; (b) mCPBA,
CH2Cl2, rt, 67% as a mixture of
ꢁ78 °C to rt; (d) NaBH4, MeOH, rt, 77% (in 2 steps).
a
and b isomer; (c) O3, CH2Cl2, ꢁ78 °C, then Me2S,
groups,15 it was recently reported that a high-level ab initio calcu-
lation showed the cyclopropyl group to have the largest stabilizing
effect on a carbocation center among the cyclopropyl, phenyl, and
vinyl groups.20 According to this information, we expected that the
CPM group would be more easily removed than the benzyl and
allyl groups. However, the benzyl group was preferentially
removed over the CPM group (Table 1, entry 1). The allyl and
CPM groups were eliminated at almost equivalent rates (entry 2).
The CPM group was removed preferentially as compared to the
ethyl and methyl groups (entries 3 and 4). The order of removal
of the five groups was established as: benzyl > CPM ꢀ allyl > ethyl,
methyl. Contrary to our expectation, the elimination rate of CPM
group was not so high. This discrepancy may be attributed to steric
hindrance of the CPM group and/or polarity of the reaction solvent.
Hydrolysis of oxazolidinone 2 with sodium hydroxide in DMSO
provided original morphinan 1 in 50% yield. This result means that
the oxazolidinone group can protect the 17-nitrogen group from
oxidation. Naltrexone derivative 1 was oxidized with mCPBA to
give a complex mixture owing to oxidation of nitrogen. On the
other hand, oxazolidinone 11 was oxidized with mCPBA to afford
R
Bn
N
R
N
N
R'
Me
Me
Bn N
Bn N
R, R' = alkyl
R'
Me
Figure 4.
objective epoxide 12 in 67% yield as a mixture of the
a and b iso-
mers. Oxazolidinone 13 was also oxidized with ozone followed by
the reduction with NaBH4 to give alcohol 14 in 77% yield (Scheme
3).
Table 1
Examination of order of reactivity among cyclopropylmethyl, benzyl, allyl, ethyl, and
methyl groupsa
In summary, we found the novel cleavage reaction of C16–N17
bond of naltrexone derivative 1 to give oxazolidinone 2. The mech-
anism of the cleavage reaction was proposed. The CPM groups in
tertiary amines 9c, d were found to be eliminated faster than the
ethyl and methyl groups, although the benzyl group in 9a was
more readily removed than the CPM group. The allyl and CPM
groups in 9b were removed at almost equivalent rates. The oxazo-
lidinone group in compounds 11 and 13 was also shown to be use-
ful in protecting the 17-nitrogen in naltrexone derivatives from
oxidation.
R'
N
N
R
Troc
10a -d
9a - d
Entry
Substrate 9
Product 10
R
R0
Yield (%)
1
2
Bn (a)
allyl (b)
CPMb (a)
CPMb (b)
allyl (b0)
Et (c)
Quant
54
45
Quant
Quant
3
4
Et (c)
Me (d)
Me (d)
Acknowledgements
a
To the solution of 9 (0.11 mmol) in 1,2-dichloroethane were added Troc-Cl
We acknowledge the financial supports from Grant-in-Aid
for Scientific Research (C) (19590105). We also acknowledge the
(0.22 mmol) and proton sponge (0.22 mmol), and stirred at rt for 3–4 h.
b
CPM: cyclopropylmethyl.

6296
H. Fujii et al. / Tetrahedron Letters 49 (2008) 6293–6296
m), 1.53–1.64 (2H,m), 1.71–1.85 (1H, m), 1.89–2.08 (2H, m), 2.37 (1H, ddd,
J = 4.8, 11.7, 14.1 Hz), 2.66 (1H, ddd, J = 0.9, 9.6, 14.7 Hz), 2.86 (1H, dt, J = 5.1,
11.4 Hz), 2.97 (1H, dd, J = 7.2, 14.7 Hz), 3.21 (1H, dd, J = 6.9, 14.7 Hz), 3.52 (1H,
dd, J = 7.2, 14.4 Hz), 3.58 (1H, dt, J = 5.1, 11.1 Hz), 3.80–3.90 (2H, m), 3.88 (3H,
s), 3.92 (1H, dd, J = 6.6, 13.2 Hz), 4.02–4.11 (1H, m), 4.21–4.29 (1H, m), 4.76
(1H, s), 6.71 (1H, dd, J = 0.9, 7.8 Hz), 6.80 (1H, d, J = 7.8 Hz). HRMS (FAB) calcd
for C24H29ClNO6 [M+H]+: 462.1683; found 462.1675.
Institute of Instrumental Analysis of Kitasato University, School of
Pharmacy for its facilities.
References and notes
1. Casy, A. F.; Parfitt, R. T. In Opioid Analgesics; Plenum Press: New York, 1986.
2. Hagemen, H. A. Org. React. 1953, 7, 198.
3. von Braun, J. Ber. 1914, 47, 2312.
4. Bentley, K. W.; Hardy, D. G. J. Am. Chem. Soc. 1967, 89, 3281.
5. Cooley, J. H.; Evain, E. J. Synthesis 1989, 1.
6. Montzka, T. A.; Matiskella, J. D.; Partyka, R. A. Tetrahedron Lett. 1974, 15, 1325.
7. Olofson, A. R.; Schnur, R. C.; Bunes, L.; Pepe, J. P. Tetrahedron Lett. 1977, 18,
1567.
11. Spectral data for compound 7. IR (neat) cmꢁ1: 3482, 2966, 1747, 1508, 1433,
1280, 1168, 1046. 1H NMR (CDCl3, 300 MHz): d 0.60 (3H, t, J = 7.5 Hz), 1.40–
1.59 (2H, m), 1.66–2.02 (4H, m), 2.59 (1H, ddd, J = 0.9, 9.6, 14.4 Hz), 2.92 (3H,
s), 3.12 (1H, dd, J = 7.2, 14.4 Hz), 3.55 (1H, dd, J = 7.2, 9.6 Hz), 3.79–3.93 (2H,
m), 3.84 (3H, s), 4.01–4.10 (1H, m), 4.18–4.25 (1H, m), 4.67 (1H, s), 6.65 (1H,
dd, J = 0.8, 8.1 Hz), 6.75 (1H, d, J = 8.1 Hz). HRMS (FAB) calcd for C21H26NO6
[M+H]+: 388.1760; found 388.1809.
12. Gould, E. S. In Mechanism and Structure in Organic Chemistry; Henry Holt and
Company: New York, 1959; pp 561–617.
13. Richey, H. G., Jr. In Carbonium Ions; Olah, G. A., Schleyer, P. R., Eds.; John Wiley &
Sons: Canada, 1972; Vol. III, pp 1201–1294.
14. Lowry, T. H.; Richardson, K. S. In Mechanism and Theory in Organic Chemistry,
3rd ed.; Harper & Row Publishers: New York, 1987. pp 425–515.
15. Olah, G. A.; Prakash Reddy, V.; Surya Prakash, G. K. Chem. Rev. 1992, 92. 69.
16. Nagase, H.; Yamamoto, N.; Nemoto, T.; Yoza, K.; Kamiya, K.; Hirono, S.;
Momen, S.; Izumimoto, N.; Hasebe, K.; Mochizuki, H.; Fujii, H. J. Org. Chem.,
in press.
17. Fujii, H.; Osa, Y.; Ishihara, M.; Hanamura, S.; Nemoto, T.; Nakajima, M.; Hasebe,
K.; Mochizuki, H.; Nagase, H. Bioorg. Med. Chem. Lett., in press. doi:10.1016/
j.bmcl.2008.08.019.
8. Olofson, R. A.; Martz, J. T.; Senet, J. P.; Piteau, M.; Malfroot, T. J. Org. Chem. 1984,
49, 2081.
9. Although morphine is a classical
is known to be lower than that of a representative
agonist, DAMGO. Moreover, it was recently reported that morphine was a
partial agonist for opioid receptor in some pharmacological assays. See the
l
opioid receptor agonist, its intrinsic activity
l
opioid receptor full
l
following references: (a) Selley, D. E.; Sim, L. J.; Xiao, R.; Liu, Q.; Childers, S. R.
Mol. Pharmacol. 1997, 51, 87; (b) Brown, G. P.; Yang, K.; Ouerfelli, O.; Standifer,
K. M.; Byrd, D.; Pasternak, G. W. J. Pharmacol. Exp. Ther. 1997, 282, 1291; (c)
Traynor, J. R.; Clark, M. J.; Remmers, A. E. J. Pharmacol. Exp. Ther. 2002, 300, 157;
(d) Schulz, S.; Mayer, D.; Pfeiffer, M.; Stumm, R.; Koch, T.; Höllt, V. EMBO J.
2004, 23, 3282; (e) Johnson, E. A.; Oldfield, S.; Braksator, E.; Gonzalez-Cuello,
A.; Couch, D.; Hall, K. J.; Mundell, S. J.; Bailey, C. P.; Kelly, E.; Henderson, G. Mol.
Pharmacol. 2006, 70, 676.
18. Kapnang, H.; Charles, G. Tetrahedron Lett. 1983, 24, 3233.
19. Mayr, H.; Olah, G. A. J. Am. Chem. Soc. 1977, 99, 510.
20. van Alem, K.; Lodder, G.; Zuilhof, H. J. Phys. Chem. A 2002, 106, 10681.
10. Spectral data for compound 2. IR (KBr) cmꢁ1: 2938, 1742, 1439, 1070, 1033. 1H
NMR (CDCl3, 300 MHz): d 0.23–0.40 (2H, m), 0.47–0.68 (2H, m), 0.94–1.09 (1H,