An improved process for the N-demethylation of opiate alkaloids using an iron(II) catalyst in acetate buffer

Article abstract of DOI:10.1002/adsc.200800632

An improved process to N-demethylate opiate alkaloids utilising a solution of the ferrous porphyrin, tetrasodium 5,10,15,20-tetra(4-sulfophenyl) porphyrinatoiron(II) [=Fe(II)-TPPS (8)], in acetate buffer is described. This method provided the corresponding N-demethylated opiates in good yield with high reproducibility.

Full text of DOI:10.1002/adsc.200800632

UPDATES
DOI: 10.1002/adsc.200800632
An Improved Process for the N-Demethylation of Opiate
Alkaloids using an Iron(II) Catalyst in Acetate Buffer
a
a
a,
Gaik Kok, Trent D. Ashton, and Peter J. Scammells *
a
Medicinal Chemistry and Drug Action, Monash Institute of Pharmaceutical Sciences, Monash University, 381 Royal
Parade, Parkville, Victoria, 3052, Australia
Fax : (+61)-9-903-9582; e-mail: peter.scammells@pharm.monash.edu.au
Received: October 15, 2008; Revised: December 14, 2008; Published online: January 2, 2009
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200800632.
Abstract: An improved process to N-demethylate
opiate alkaloids utilising a solution of the ferrous
porphyrin, tetrasodium 5,10,15,20-tetra(4-sulfophe-
nyl)porphyrinatoiron(II) [=Fe(II)-TPPS (8)], in
acetate buffer is described. This method provided
the corresponding N-demethylated opiates in good
yield with high reproducibility.
Keywords: alkaloids; amine N-oxides; iron(II) por-
phyrins; N-demethylation; Polonovski reaction
Naturally occurring opiates such as morphine and co-
deine possess an N-methyl group. Varying this sub-
stituent produces profound pharmacological effects Figure 1. Semi-synthetic opiates.
and there are a number of clinically relevant, semi-
synthetic opiates in which the N-methyl group has
been replaced by other alkyl moieties. Substituents chemistry and microorganisms. Our own laborato-
such as allyl, cyclopropylmethyl and cyclobutylmethyl ry has shown that the FeSO ·7H O-mediated non-
[
6]
[7]
4
2
typically endow antagonist properties, although ago- classical Polonovski reaction is also effective in the N-
[
8]
nist activity is restored with longer alkyl groups such demethylation of several opiate alkaloids. These and
as phenethyl. For example, nalorphine (1) (Figure 1), other methods for the N-demethylation of alkaloids
[
9]
which has an N-allyl group, has mixed agonist and an- have been the subject of a recent review.
[
10]
tagonist activity, while N-phenethylnormorphine (2)
We recently reported,
as summarised in
acts as a (m) agonist with 10-fold greater potency than Scheme 1, a new chemical process utilising a Fe(II)
[1]
morphine itself. The combination of an N-allyl or N- porphyrin-based complex for the N-demethylation of
cyclopropylmethyl group and some additional C-ring several opiate alkaloids such as dextromethorphan
modifications (6-keto and 14-hydroxy functionality) (5a), codeine methyl ether (CME) (6a), and thebaine
affords the potent antagonists, naloxone (3) and nal- (7a) (Figure 2). Thus, 0.3 molar equivalents of tetraso-
trexone (4), respectively. Accordingly, N-demethyla- dium
5,10,15,20-tetra(4-sulfophenyl)porphyrinato-
tion of natural opiates is a key chemical transforma-
tion in the synthesis of semi-synthetic opiates.
ACHTUNGTRENiNGNU ron(II) [Fe(II)-TPPS (8)] in MeOH readily effected
N-Demethylation of natural opiates has been
achieved in many ways including the use of reagents
[
2]
such as cyanogen bromide (von Braun reaction),
[3]
[4]
chloroformates and diethyl azodicarboxylate, as
well as procedures utilising photochemistry, electro- Scheme 1. Fe(II)-TPPS-mediated N-demethylation.
[
5]
Adv. Synth. Catal. 2009, 351, 283 – 286
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
283

Gaik Kok et al.
UPDATES
These findings prompted us to explore the use of 8 as
a solution in acetate buffer in N-demethylations.
We report herein an improved process which uti-
lises a solution of 8 in 1M acetate pH 4 buffer for the
N-demethylation of opiate alkaloids. A stock solution
of a 0.025M solution of 8 in 1M sodium acetate pH 4
buffer was prepared by heating FeSO ·7H O and
4
2
TPPS in buffer under reflux for 3.5 h, as described
above. The deep red-brown solution of Fe(II)-TPPS
(
8) in buffer was then cooled and used in N-demethy-
Figure 2. Dextromethorphan, codeine methyl ether and the-
baine.
lations without further purification.
The method was first illustrated with dextromethor-
phan (5a). Thus, oxidation of 5a with m-CPBA in
the N-demethylation of these N-methyl alkaloids, via CHCl and treatment with HCl provided the corre-
3
the corresponding N-oxide hydrochloride, in moder- sponding N-oxide hydrochloride in near quantitative
ate to high yields.
yield. The solution of Fe(II)-TPPS (8) in buffer was
Fe(II)-TPPS (8) then added to a stirred solution of the opiate N-oxide
[
10,11]
Using literature methodology,
was prepared via a two-step process: preparation of hydrochloride in MeOH. Using 20 mol% of catalyst,
the porphyrin TPPS and the subsequent incorporation the reaction was found to proceed effectively at room
of ferrous ion into the porphyrin. Employing micro- temperature and was complete, as monitored by TLC
wave-assisted irradiation at 2008C for 15 min, pyrrole analysis, after 144 h. Since only a substoichiometric
reacted with benzaldehyde to afford 5,10,15,20-tetra- amount of 7 was employed, when the reaction was
phenylporphyrin (TPP) in 25% yield. Subsequent sul- complete, the bulk of the catalyst was readily re-
fonation of TPP with concentrated sulfuric acid, fol- moved via precipitation after the addition of Et O.
2
lowed by work-up with sodium hydroxide (pH 8–10) The volatile organics were then removed and the
and lyophilisation, gave a quantitative yield of TPPS. crude product re-dissolved in CHCl . This was fol-
3
TPPS was then treated with FeSO ·7H O in a 1M lowed by a simple extractive work up and purification
4
2
acetate buffer (pH 4) under refluxing conditions for via column chromatography to remove a small
.5 h. The reaction mixture was cooled and filtered. amount of dextromethorphan (5a) (~5% by
3
1
The filtrate was poured into acetone and the resulting
H NMR). This provided an 87% yield of N-nordex-
precipitate isolated via centrifugation. The co-precipi- tromethorphan (5b) (Table 1, entry 1), which is com-
tation of the solution with acetone and the centrifuga- parable to that obtained in the case where isolated
[
10]
tion processes were repeated. This provided a total catalyst was employed.
isolated yield of 92% Fe(II)-TPPS (8). A number of experiments were then conducted to
(III)-porphyr- evaluate the scope of the reaction, varying the sub-
[
10]
In the earlier study, a number of Fe
ACHTUNGTRENNUNG
ins were evaluated as catalysts for N-demethylation. strate to catalyst loading ratio, the reaction time and
Using the N-oxide of CME (6a) as a substrate, stoi- the temperature. Reactions were slow at room tem-
chiometric or substoichiometric amounts of Fe ACHUTNGRNENGU( III)- perature, but proceeded at synthetically useful rates
TPPSCl failed to yield any of the desired N-demethy- at 50–1008C (Table 1). At these temperatures the re-
lated product. Therefore, in the preparation and isola- actions could be conducted with catalyst loadings as
tion of 8, much care was taken to minimise exposure low as 2 mol% (entry 6).
of the catalyst to air oxidation. Degassed buffer and
solvents were employed and all operations were con-
Table 1. N-Demethylation of dextromethorphan N-oxide hy-
ducted under an inert atmosphere. Despite these
measures we have subsequently observed variable ef-
ficiency from different batches of catalyst. Clearly a
simpler and more consistent operation would enhance
the synthetic utility of this catalyst.
drochloride with Fe(II)-TPPS in acetate buffer.
o
Entry Catalyst (equiv.) Temp. [ C] Time [h] Yield [%]
[a]
of 5b
1
2
3
4
5
6
0.2
0.1
0.1
0.05
0.05
0.02
20
20
50
50
100
50
144
168
7
24
7
87[
b]
Recently, Huszꢁnk and Horvꢁth reported that
Fe(II)-TPPS (8) was stable in an acetate buffer, even
nd
88
86
86
83
[
12]
after saturation with air or oxygen.
It was found
that when a solution of 8 in acetate buffer was ex-
posed to air, the ferrous species was stable even over
weeks, with no conversion to the ferric form. These
authors postulated that the acetate was able to trap
96
[a]
Isolated yield after column chromatography.
[b]
1
Crude product, as analysed by H NMR, showed incom-
plete conversion; N-oxide 16%, N-nordextromethorphan
81%, dextromethorphan 3%.
traces of Fe
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(III), responsible for conversion of Fe(II)-
porphyrins to the corresponding Fe AHCTUNGTRE(UNNNG III) species.
284
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 283 – 286

An Improved Process for the N-Demethylation of Opiate Alkaloids
UPDATES
Table 2. N-Demethylation of CME N-oxide hydrochloride
with Fe(II)-TPPS in acetate buffer.
A sample of the equatorial N-Me isomer was isolat-
ed via column chromatography (silica gel column
eluted with CHCl /MeOH/NH OH, 95:5:1–90:10:1).
o
3
4
Entry Catalyst (equiv.) Temp. [ C] Time [h] Yield [%]
[a]
The hydrochloride form was also prepared by washing
the free N-oxide in chloroform with 2M HCl in brine.
The N-demethylation of thebaine N-oxide was slower
than the corresponding reactions with the N-oxides of
dextromethorphan or CME. Accordingly, a higher
catalyst loading and reaction temperature were re-
quired to achieve full conversion in a reasonable
of 6b
1
2
3
0.05
0.05
0.05
20
50
80
408
43
6
61
90
84
[
a]
Isolated yield after column chromatography.
CME (6a) was also evaluated as a substrate for this timeframe (0.1 equiv. of 8 at 808C). The pure equato-
reaction. Oxidation of CME with m-CPBA in CHCl3 rial isomer of thebaine N-oxide gave a significantly
at 08C for 10 min followed by treatment with HCl higher yield of the desired N-demethylated product
provided the corresponding N-oxide hydrochloride in when the protonated form of the N-oxide was used
near quantitative yield. It is noteworthy that when (Table 4, cf. entries 1 and 2). A similar outcome was
[
5,10]
H O (30% w/v) was employed as the oxidant,
the observed when a mixture of thebaine N-oxide isomers
2
2
reaction took longer to complete (1–2 days at room was used (cf. entries 3 and 4).
temperature).
The influence of the isomeric ratio was further
Results for the N-demethylation of the CME N- probed through a comparison of the reaction of the-
oxide are summarised in Table 2. When a solution of baine N-oxide hydrochloride either as the pure equa-
CME N-oxide hydrochloride in MeOH was treated torial N-Me isomer or as a 65:35 or 75:25 mixture of
with 5 mol% of Fe(II)-TPPS (8) in buffer at room N-oxide isomers (Table 4, entries 1, 3 and 5). All of
temperature, the reaction took approximately 17 days these reactions afforded comparable isolated yields of
to complete and afforded the N-demethylated prod- N-northebaine (7b), which suggests that the isomeric
uct in modest yield (Table 2, entry 1). When the reac- configuration of the protonated N-oxide does not
tion was conducted at 508C, the reaction was com- affect the final outcome of the reaction. On the other
plete after 43 h and returned a 90% yield of the N- hand, the yield of 7b greatly diminished when a 65:35
norCME (6b) (entry 2). A further increase in temper-
ature reduced the reaction time to 6 h, but also gave a
slightly lower yield of 6b (entry 3).
Table 3. Oxidation of thebaine to thebaine N-oxide.
Entry
Oxidant
Solvent
Temp.
Time
[min]
Isomers
E:A
When thebaine N-oxide hydrochloride was reacted
with 0.1 equiv. of Fe(II)-TPPS in acetate buffer a 54%
yield of N-northebaine was obtained. It has previously
been noted that lower yields observed for the N-de-
methylation of thebaine may be related to the ratio
and relative stability of the N-oxide isomers formed.
Caldwell and co-workers have reported that the axial
N-Me isomer of thebaine N-oxide (in the non-proton-
ated form) completely decomposed upon standing in
solution for 5 days while the equatorial N-Me isomer
o
[a–c]
[
C]
A
H
U
G
R
N
U
G
1
2
3
4
5
m-CPBA
m-CPBA
m-CPBA
m-CPBA
CH Cl
0
10
10
10
20
960
65:35
71:29
75:25
75:25[
2
2
2
2
2
CH Cl
À18
À45
À78
0
2
CH Cl
2
CH
Cl
2
d]
H
O
MeOH
35:65
2
2
[a]
Oxidation with m-CPBA, in all cases, gave near quantita-
tive yields of N-oxide isomers.
[
[
[
b]
c]
d]
A, axial N-Me isomer; E, equatorial N-Me isomer.
[
13]
1
remained stable over this period. As a result we in-
vestigated the ratio of N-oxide isomers formed by
varying the oxidation conditions and the effect of this
ratio on subsequent N-demethylation reactions.
Isomer ratio as analysed by H NMR.
[13]
Ref.
Table 4. N-Demethylation of thebaine N-oxide with Fe(II)-
Oxidation of thebaine (7a) with m-CPBA at 08C for TPPS (0.1 equiv.) in acetate buffer.
1
0 min provided a 65:35 mixture of two N-oxide iso-
Entry
N-Oxide
Isomer ratio
Yield [%] of
7b (7a)
mers (Table 3, entry 1), with the N-oxide having the N-
Me group equatorial being the major isomer. Isomer
ratios were determined by H NMR, according to peak
assignments previously reported.
mers obtained was influenced by temperature. For ex-
ample, when the oxidation was conducted at À458C,
the ratio increased to 75:25 in favour of the equatorial
isomer (cf. entries 1 and 3). In contrast, when a differ-
ent oxidant and solvent were employed (30% H O in
MeOH),
[a]
[b,c]
A
H
U
G
R
N
U
G
1
1
2
3
4
5
HCl salt
free base
HCl salt
free base
HCl salt
100:0
100:0
65:35
65:35
75.25
71 (28)
49 (22)
69 (28)
21 (44)
72 (23)
[
13]
The ratio of iso-
[a]
A, axial N-Me isomer; E, equatorial N-Me isomer.
Isolated yield after column chromatography.
Reactions were conducted at 808C for 24 h with
0.1 equiv. of Fe(II)-TPPS.
[
[
b]
c]
2
2
[10,13]
the isomers were produced in a ratio of
3
5:65, with the axial isomer predominating (entry 5).
Adv. Synth. Catal. 2009, 351, 283 – 286
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
285

Gaik Kok et al.
UPDATES
mixture of the free N-oxide isomers was employed, stock solution in 1M acetate pH 4 buffer was added. The
total volume of buffer was then adjusted to 2.5 mL with ad-
ditional buffer. The deep red-brown solution was then
with the major product being the tertiary amine 7a
(
entry 4). These outcomes are consistent with the ob-
stirred for the specified time and temperature. When the re-
action was complete, as analysed by TLC analysis, Et
15 mL) was added and the precipitate was filtered over
Celite. The filter pad was washed with Et O/MeOH (2:1,
served inherent instability of the free base form of
[
13]
O
2
the axial N-Me N-oxide isomer. On the other hand,
(
the protonated forms of both isomers were found to
2
be stable in CHCl for at least 10 days at room tem-
3
5
mLꢂ3) and the combined filtrate was concentrated. The
perature.
resultant residue was taken up in brine (5 mL) and adjusted
Tropane alkaloids have also been N-demethylated to pH 10 using NH OH. The mixture was extracted with
4
using the iron sulfate-mediated Polonovski reaction, CHCl (3ꢂ5 mL), dried (Na SO ), filtered, and evaporated.
3
2
4
[
8b]
albeit in modest yield. This Fe(II)-TTPS procedure Subsequent column chromatography (CHCl /MeOH/
3
NH OH; 90:10:0.5 to 85:15:1) gave N-nordextromethorphan
effected the N-demethylation of atropine in 42% iso-
lated yield, which is slightly lower than the best yield
previously obtained with iron sulfate (51%).
In conclusion, N-methyl alkaloids can be efficiently
N-demethylated via conversion to the corresponding
N-oxide and treatment with a solution of Fe(II)-TPPS
in acetate buffer. This method proved to be highly re-
4
1
as a pale yellow oil. H NMR: d=7.05 (1H, d, J=8.4 Hz),
6
(
.80 (1H, d, J=2.4 Hz), 6.72 (1H, dd, J=8.4, 2.4 Hz), 3.78
3H, s), 3.20–3.08 (2H, m), 3.00 (1H, br, NH), 2.81 (1H, d,
J=18.0 Hz), 2.78–2.71 (1H, m), 2.63 (1H, ddd, J=12.6,
2.6, 3.0 Hz), 2.32 (1H, d, J=12.6 Hz), 1.82 (1H, m), 1.72–
.58 (2H, m), 1.57–1.47 (1H, m), 1.43–1.25 (5H, m), 1.07
1
1
(
+
1H, m); HR-MS: m/z=258.1851, calcd. for [M+H]
producible and obviated the tedious processes associ- C H NO: 258.1852.
1
7
23
ated with isolation of the Fe(II)-TPPS catalyst. The
improved stability of the catalyst in the acetate buffer
allowed significantly lower catalyst loading to be em-
ployed without any deleterious effect on the reaction
yield. Furthermore, it was found that mild heating of
Acknowledgements
this reaction system could also significantly reduce re- The authors thank the ARC Centre for Free Radical Chemis-
action times (from over 100 h to less than 10 h).
The N-demethylation of thebaine was also studied
in greater detail. It had previously been thought that
try and Biotechnology for financial support.
the lower yields obtained from this substrate may References
have resulted from the instability of the axial isomer
of the N-oxide. However, the hydrochloride salts of
both thebaine N-oxide isomers were found to be
[1] D. S. Fries, in: Foyeꢀs Principles of Medicinal Chemistry,
th
5
edn., (Eds.: D. A. Williams, T. L. Lemke), Lippin-
cott William & Wilkins, Philadelphia, 2002.
stable in CHCl for at least 10 days at room tempera-
3
[
[
2] J. von Braun, Ber. dtsch. chem. Ges. 1909, 42, 2035.
3] a) J. H. Cooley, E. J. Evian, Synthesis 1989, 1; b) K. C.
Rice, J. Org. Chem. 1975, 40, 1850; c) K. C. Rice, E. L.
May, J. Heterocycl. Chem. 1977, 14, 665.
ture. Furthermore, N-demethylation of the pure equa-
torial isomer of thebaine N-oxide hydrochloride pro-
ceeded in similar yield to 65:35 and 75:25 mixtures of
isomers, suggesting that the isomeric configuration of
the N-oxide hydrochloride does not affect the final
outcome of the reaction.
[
4] a) L. S. Schwab, J. Med. Chem. 1980, 23, 698; b) H.
Merz, K. H. Pook, Tetrahedron 1970, 26, 1727.
[5] J. A. Ripper, E. R. T. Tiekink, P. J. Scammells, Bioorg.
Med. Chem. Lett. 2001, 11, 443.
[
6] a) P. J. Smith, C. K. Mann, J. Org. Chem. 1969, 34,
821; b) J. E. Barry, M. Finkelstein, E. A. Mayeda, S. D.
1
Experimental Section
Ross, J. Org. Chem. 1974, 39, 3488.
[
7] a) K. M. Madyastha, Proc. Indian Acad. Sci. 1994, 106,
General N-Demethylation Procedure
1
203; b) K. M. Madyastha, G. Vijay Bhasker Reddy, J.
Dextromethorphan·HBr·H O (3.00 g, 8.10 mmol) was dis-
Chem. Soc. Perkin Trans. 1 1994, 911.
2
solved in CHCl (60 mL) and extracted using brine/NH OH
[8] a) K. McCamley, J. A. Ripper, R. D. Singer, P. J. Scam-
mells, J. Org. Chem. 2003, 68, 9847; b) S. Thavaneswar-
an, P. J. Scammells, Bioorg. Med. Chem. Lett. 2006, 16,
2868.
[9] S. Thavaneswaran, K. McCamley, P. J. Scammells, Nat.
Prod. Commun. 2006, 1, 885.
3
4
(
pH 10), dried (Na SO ), filtered and cooled to 08C. m-
2 4
CPBA (2.31 g of ~73% reagent, 9.77 mmol) was added in
one portion and after 10 min, the solution was successively
extracted with 10% NaOH/brine (pH 10) (2ꢂ10 mL), brine
(
2ꢂ5 mL) and 2M HCl/brine (pH 2). The organic phase was
dried (Na SO ), filtered, and concentrated to afford dextro-
methorphan N-oxide hydrochloride as a white foam; yield:
[10] Z. Dong, P. J. Scammells, J. Org. Chem. 2007, 72, 9881.
[11] P. Rothemund, A. R. Menotti, J. Am. Chem. Soc. 1941,
63, 267.
2
4
2.49 g (94%).
The crude dextromethorphan N-oxide hydrochloride
100 mg, 0.309 mmol) was dissolved in MeOH (7 mL) and
the specified equivalent of Fe(II)-TPPS from a 0.025M
[12] R. Huszꢁnk, O. Horvꢁth, Chem. Commun. 2005, 224.
[13] G. W. Caldwell, A. D. Gauthier, J. E. Mills, Magn. Res.
Chem. 1996, 34, 505.
(
286
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 283 – 286