N-Demethylation of N-methyl alkaloids with ferrocene

Article abstract of DOI:10.1016/j.bmcl.2010.06.031

Under Polonovski-type conditions, ferrocene has been found to be a convenient and efficient catalyst for the N-demethylation of a number of N-methyl alkaloids such as opiates and tropanes. By judicious choice of solvent, good yields have been obtained for dextromethorphan, codeine methyl ether, and thebaine. The current methodology is also successful for the N-demethylation of morphine, oripavine, and tropane alkaloids, producing the corresponding N-nor compounds in reasonable yields. Key pharmaceutical intermediates such oxycodone and oxymorphone are also readily N-demethylated using this approach.

Full text of DOI:10.1016/j.bmcl.2010.06.031

Bioorganic & Medicinal Chemistry Letters 20 (2010) 4499–4502
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
N-Demethylation of N-methyl alkaloids with ferrocene
*
Gaik B. Kok, Peter J. Scammells
Medicinal Chemistry and Drug Action, Monash Institute of Pharmaceutical Sciences, Monash University, 381 Royal Parade, Parkville, Vic. 3052, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
Under Polonovski-type conditions, ferrocene has been found to be a convenient and efficient catalyst for
the N-demethylation of a number of N-methyl alkaloids such as opiates and tropanes. By judicious choice
of solvent, good yields have been obtained for dextromethorphan, codeine methyl ether, and thebaine.
The current methodology is also successful for the N-demethylation of morphine, oripavine, and tropane
alkaloids, producing the corresponding N-nor compounds in reasonable yields. Key pharmaceutical inter-
mediates such oxycodone and oxymorphone are also readily N-demethylated using this approach.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 14 April 2010
Revised 4 June 2010
Accepted 7 June 2010
Available online 10 June 2010
Keywords:
N-Demethylation
Polonovski reaction
Alkaloid
Ferrocene
N-Demethylation of opiate alkaloids is a key step in the semi-
synthesis of a number of medicinally important opiate deriva-
tives.1 For example, the mixed agonist antagonists, nalorphine (1)
and buprenorphine (2) as well as the potent antagonists, naloxone
(3) and naltrexone (4) (Fig. 1), are obtained from the corresponding
N-nor opiates.2–5 These N-nor opiates are, in turn, derived from
naturally occurring N-methyl opiates such as morphine, codeine,
thebaine and oripavine.6
N-Demethylation of alkaloids generally, and opiates in particu-
lar, has been a long-standing challenge to synthetic chemists.1
A wide variety of methods have been developed that attempt to
accommodate the often competing requirements of reagent costs,
toxicity, substrate specificity, yield, ease of operation and amena-
bility to scale up. More traditional methods include the use of
reagents such as cyanogen bromide (von Braun reaction),7 chloro-
formates,8 and dialkyl azodicarboxylates.9 Less generally applica-
ble methods include those that involve photochemical10 and
biochemical11 processes.
There is considerable scope for the further development of
alternative methods for the N-demethylation of opiate derivatives.
We have previously reported that, under Polonovski-type condi-
tions, iron(II) reagents such as FeSO4Á7H2O12 and the ferrous
porphyrin, tetrasodium 5,10,15,20-tetra(4-sulfophenyl)porphyri-
natoiron(II)13 are effective for N-demethylation of a number of opi-
ate and tropane alkaloids. Thus, the tertiary N-methylamine is first
converted into the corresponding N-oxide hydrochloride which,
following subsequent treatment with the ferrous reagent, has pro-
vided the N-nor compound in moderate to good yields. In most
cases, the only by-product obtained is the parent tertiary amine.
This result is consistent with a reaction pathway which involves
a number of intermediates. A plausible mechanism for the reaction
which implicates
a Fe(II)/Fe(III) redox couple, as shown in
Scheme 1, has previously been proposed.14
We now report a simple and mild two-step process employing
ferrocene to N-demethylated opiate and tropane alkaloids: forma-
tion and isolation of the corresponding N-methylamine N-oxide
hydrochloride, and subsequent reaction with ferrocene (Scheme 2).
HO
HO
O
O
N
N
HO
MeO
OH
Nalorphine (1)
HO
Buprenorphine (2)
HO
O
O
N
N
OH
OH
O
O
Naloxone (3)
Naltrexone (4)
* Corresponding author. Tel.: +61 3 9903 9542; fax: +61 9903 9582.
E-mail address: peter.scammells@pharm.monash.edu.au (P.J. Scammells).
Figure 1. Examples of semi-synthetic opiates.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.06.031

4500
G. B. Kok, P. J. Scammells / Bioorg. Med. Chem. Lett. 20 (2010) 4499–4502
H+
Fe(II)
+
+
O
CH3
+ OH
N
OH
CH3
Fe(II)
N
N
CH3
H+
Fe(II) Fe(III)
+
-H+
+
N CH2
N CH2
N
CH3 + Fe(III) + H2O
Fe(II)
hydrolysis
Fe(III)
N H
O
CH2
N
CH3
Scheme 1. Proposed mechanism14 for the non-classical Polonovski reaction.
Cl
OH
Me
R
R
R
R
R
R
ferrocene
(i) [O]
N
N H
(+ CH2O)
N Me
(ii) HCl
a
c
b
N-nor product
N-methylamine
N-oxide
hydrochloride
Scheme 2. N-Demethylation with ferrocene.
Structural effects have been studied using a range of opiates and tro-
panes and, for a number of these substrates, the effects of solvent on
the reaction outcome have been investigated.
of solvents: CHCl3, CH2Cl2, MeCN, MeOH, and i-PrOH. The solvents
were used as supplied and degassed with nitrogen prior to use.
Employing m-CPBA as oxidant, 5a was readily converted into
the corresponding N-oxide, which was isolated as the hydrochlo-
ride salt 5b. In a typical experiment,16 the N-oxide 5b and ferro-
cene were degassed prior to the addition of solvent. The resulting
solution was then stirred at the specified temperature until the
consumption of starting material 5b was complete (by TLC analy-
sis) or for the specified length of time. After standard workup,
the desired N-nor product 5c was isolated from a small amount
of the starting tertiary N-methylamine 5a via column chromatog-
Our initial efforts began with an examination of the conditions
that were effective for the N-demethylation of dextromethorphan
(5a) as a representative opiate substrate, and the optimized condi-
tions were subsequently applied to a larger test set of opiate alka-
loids (Fig. 2). As the ferrocene/ferrocenium redox couple has been
extensively studied and found to be solvent dependent,15 we were
interested in exploring the effects that different solvents have on
N-demethylation. Hence, experiments were conducted in a range
MeO
RO
RO
RO
O
O
O
NMe
NMe
NMe
NMe
OH
RO
MeO
O
6a
8a
10a
, R = Me)
Dextromethorphan (5a)
Morphine ( , R = H) Thebaine ( , R = Me) Oxycodeinone (
7a
9a
Oripavine ( , R = H) Oxymorphinone (
11a
, R = H)
CME ( , R = Me)
Me
N
Me
N
RO
O
MeO
O
OH
NMe
NMe
OH
O
OH
O
MeO
O
O
Oxycodone (12a, R = Me)
Oxymorphone (13a, R = H)
Thevinone (14a)
Tropine (15a)
Atropine (16a)
Figure 2. Alkaloid test set.

G. B. Kok, P. J. Scammells / Bioorg. Med. Chem. Lett. 20 (2010) 4499–4502
4501
Table 1
nating solvents such MeCN the reaction produced relatively more
tertiary amine 5a (47%, entry 5). The total percentage recovery of
product was also lower (84%). In MeOH, only a small amount of
the N-nor product 5c was produced after 96 h at 40 °C, with most
of the N-oxide being recovered unchanged (entry 6). We found that
the reaction was still incomplete even when the amount of ferro-
cene employed was increased fourfold (entry 7) and the reaction
conducted at a higher temperature (refluxing MeOH) for a pro-
longed period (336 h).
The results using either MeCN or MeOH as solvent suggest that
the ferrocene/ferrocenium one-electron oxidation reduction pro-
cess may be quasi-reversible in these solvents. Indeed, previous
studies on the electrochemical oxidation of ferrocene have sug-
gested that there is a very slow decomposition reaction of the oxi-
dized species in certain media such as MeCN, MeOH, and EtOH.17
The reaction conducted in i-PrOH proceeded to completion after
40 h at 40 °C, producing the N-nor product 5c in a modest yield
of 70% (entry 8). Finally, N-demethylation of dextromethorphan-
N-oxide with ferrocene using i-PrOH as solvent at 40 °C was re-
peated with the whole operation being conducted in air (entry
9). These conditions afforded effectively the same isolated yield
of N-nordextromethorphan (5c), although the reaction took longer
time to reach completion (68 h compared to 40 h).
The N-demethylation of morphine-N-oxide (6b), CME-N-oxide
(7b), thebaine-N-oxide (8b), and oripavine-N-oxide (9b) were next
investigated and the results are summarized in Table 2. Both CHCl3
and i-PrOH were evaluated as solvents for these reactions. Chloro-
form was chosen as it gave the best results for the N-demethyla-
tion of dextromethorphan, while isopropanol was included as a
less toxic, non-halogenated solvent that may be more suitable for
industrial applications. N-NorCME and N-northebaine were ob-
tained in good yield (92% and 84%, respectively) when CHCl3 was
used as the reaction solvent. With oripavine-N-oxide (9b), using
a catalyst loading of 0.5 molar equiv, no reaction was observed
after 24 h in CHCl3 at 50 °C, and it was thought that solubility
may be the limiting factor. Likewise, N-demethylation was not ob-
N-Demethylation of dextromethorphan-N-oxide hydrochloride (5b)a with ferrocene
Entry
Ferrocene
(equiv)
Solvent
Temperature
(°C)
Time (h)
% Isolated yieldb
5c
5a
1
2
3
4
5
6
7
8
9f
0.25
0.25
0.05
0.25
0.25
0.25
1.00
0.25
0.25
CHCl3
CHCl3
CHCl3
40
60
60
Reflux
40
40
Reflux
40
72
16
22
16
96
96
336
40
92
92
89
82
37
13
64
70
71
8
7
10
16
47
0d
23e
26
26
c
CH2Cl2
MeCN
MeOH
MeOH
i-PrOH
i-PrOH
40
68
a
Unless otherwise indicated, concentration: 5 mL solvent per 100 mg
(0.309 mmol) of substrate 5b.
b
Isolated yield after column chromatography.
Concentration: 10 mL solvent per 100 mg of substrate 5b.
No dextromethorphan was isolated; however, 86% of N-oxide was recovered.
12% of N-oxide was also recovered.
c
d
e
f
Reaction conducted in air.
raphy on silica gel, using a CHCl3/MeOH/NH4OH gradient. Most of
the ferrocene, if desired, could readily be recovered either via
extraction with hexane at an appropriate stage of the workup, or
via column chromatography. For example, in the reaction using
CH2Cl2 as solvent, a 97% recovery of ferrocene was obtained via col-
umn chromatography; in the case of MeCN, the recovery was
approximately 90%.
Table 1 compiles the key results for the N-demethylation of
dextromethorphan-N-oxide under a variety of conditions. The
reaction in CHCl3 proceeded efficiently and provided N-nordextro-
methorphan (5c) in high isolated yields of ꢀ90% (entries 1 and 2).
Although catalyst loadings as low as 5 mol % were effective (entry
3), 25 mol % of ferrocene was typically used for subsequent exper-
iments to minimize the reaction time. Dichloromethane also
proved to be a reasonable solvent, returning an 82% yield of
N-nordextromethorphan (entry 4). However, with higher coordi-
Table 2
N-Demethylation of other opiate and tropane alkaloid N-oxides
Entry
Substratea (N-oxide)
Ferrocene
(equiv)
Solvent
Temp (°C)
Time (h)
Isolation
methodb
% Isolated yield
#c
#a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Morphine (6b)
CME (7b)
0.2
i-PrOH
CHCl3
i-PrOH
CHCl3
i-PrOH
i-PrOH
CHCl3
i-PrOH
i-PrOH
i-PrOH
CHCl3
i-PrOH
i-PrOH
i-PrOH
CHCl3
i-PrOH
CHCl3
i-PrOH
CHCl3
i-PrOH
40
50
50
50
50
70
40
40
40
40
40
40
40
40
35
60
120
22
48
48
48
20
5
18
16
17
20
20
20
48
24
3
B
A
A
A
A
D
C
B
C
B
A
B
C
B
A
A
A
A
A
A
63
92
75
84
83
38
50
45
51
34
32
47
58
59
37
16
74
50
55
59
35
8
13
7
0.25
0.25
0.5
0.5
2.0
0.5
0.5
0.2
0.2
0.2
0.2
0.2
0.2
0.25
0.25
0.25
0.25
0.1
CME (7b)
Thebaine (8b)c
Thebaine (8b)c
Oripavine (9b)
11
29
40
32d
30
32
23
22
23
32
62
83
8
Oxycodeinone (10b)
Oxycodeinone (10b)
Oxycodeinone (10b)
Oxymorphinone (11b)
Oxycodone (12b)
Oxycodone (12b)
Oxycodone (12b)
Oxymorphone (13b)
Thevinone (14b)
Thevinone (14b)
Tropine (15b)
Reflux
80
Reflux
80
24
24
24
24
Tropine (15b)
Atropine (16b)
Atropine (16b)
46
6
21
0.1
a
Unless otherwise specified, concentration: 10 mL solvent per 100 mg of substrate.
Method A: column chromatography on SiO2, eluting with a gradient of CHCl3/MeOH/NH4OH (90:10:1–85:15:1) or ethyl acetate/MeOH/NH4OH (70:30:1–60:40:1). Method
b
B: extraction of an aqueous solution of the crude at pH 2–10 with a suitable solvent. Method C: 10% aqueous HCl was added to the crude and the solution was heated at 50 °C
for 1–48 h. The pH of the resultant solution was adjusted to 2–10 before extractions with a suitable solvent/solvent system. Method D: column chromatography on SiO2,
eluting with a gradient of CHCl3/MeOH (97:1–9:1).
c
Concentration: 20 mL solvent per 100 mg of substrate.
d
Tertiary amine was approximately 90% pure by 1H NMR.

4502
G. B. Kok, P. J. Scammells / Bioorg. Med. Chem. Lett. 20 (2010) 4499–4502
served for morphine-N-oxide (6b) using 0.2 molar equiv of catalyst
in refluxing CHCl3 for 48 h. However, when these reactions were
conducted in i-PrOH, modest amounts of N-nororipavine (38%)
and N-normorphine (63%) were obtained (Table 2).
The use of ferrocene also has a number of advantages over
Fe(II)TPPS, as the latter is not commercially available and only gave
the best results when it was retained in an acetate buffer to pre-
vent decomposition.
Our current methodology is also applicable to the N-demethyla-
tion of the N-oxides of various 14-hydroxy semi-synthetic opiates:
oxycodeinone-N-oxide (10b), oxymorphinone-N-oxide (11b), oxy-
codone-N-oxide (12b), and oxymorphone-N-oxide (13b). For 10b
and 12b, reactions were conducted in both CHCl3 and i-PrOH. How-
ever, due to solubility issues, reactions for 11b and 13b were per-
formed only in i-PrOH. Results are summarized in Table 2. In the
Acknowledgment
The authors thank the ARC Centre for Free Radical Chemistry
and Biotechnology for financial support.
Supplementary data
product isolation process, we have found that the a,b-unsaturated
ketones such as oxycodeinone and oxymorphinone are not stable
to chromatography over silica gel. After some experimentation, we
have found that the desired N-nor products 10c and 11c could be iso-
lated from the respective tertiary N-methylamine 10a and 11a via
simple extractions with a suitable solvent or solvent mixture. The
extractive workup protocol was also successful in the separation
of oxycodone (12a) from N-noroxycodone (12c), and oxymorphone
(13a) from N-noroxymorphone (13c). N-Demethylation of thevi-
none-N-oxide (14b) was also investigated using both CHCl3 and i-
PrOH. Interestingly, for this substrate, relatively more of the tertiary
N-methylamine 14a was produced compared to the N-demethylat-
ed product 14c (entries 15 and 16).
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.06.031.
References and notes
1. Thavaneswaran, S.; McCamley, K.; Scammells, P. J. Nat. Prod. Commun. 2006, 1,
885.
2. McCawley, E. L.; Hart, E. R.; Marsh, D. F. J. Am. Chem. Soc. 1941, 63, 314.
3. Bentley, K. W. Br. Patent 1136214, 1968.
4. Lewenstein, M. J. Br. Patent 955493, 1964.
5. Blumberg, H.; Pachter, I. J.; Matossian, Z. U.S. Patent 3332950, 1967.
6. Berényi, S.; Csutorás, C.; Sipos, A. Curr. Med. Chem. 2009, 16, 3215.
7. von Braun, J. Chem. Ber. 1909, 42, 2035.
Finally, N-demethylation of the N-oxides of tropane alkaloids
(15b and 16b) was also investigated. Experiments were conducted
in both CHCl3 and i-PrOH. For 15b, the reaction in CHCl3 delivered a
better outcome; for 16b, comparable yields of N-nor product 16c
were obtained for both solvents (Table 2).
8. (a) Cooley, J. H.; Evian, E. J. Synthesis 1989, 1; (b) Rice, K. C. J. Org. Chem. 1975,
40, 1850; (c) Rice, K. C.; May, E. L. J. Heterocycl. Chem. 1977, 14, 665.
9. (a) Schwab, L. S. J. Med. Chem. 1980, 23, 698; (b) Merz, H.; Pook, K. H.
Tetrahedron 1970, 26, 1727.
10. Ripper, J. A.; Tiekink, E. R. T.; Scammells, P. J. Bioorg. Med. Chem. Lett. 2001, 11,
443.
11. See for example: (a) Madyastha, K. M. Proc. Indian Acad. Sci. 1994, 106, 1203; (b)
Madyastha, K. M.; Reddy, G. V. B. J. Chem. Soc., Perkin Trans. 1 1994, 911; (c)
Chaudhary, V.; Leisch, H.; Moudra, A.; Allen, B.; De Luca, V.; Cox, D. P.;
Hudlicky, T. Coll. Czech. Chem. Commun. 2009, 74, 1179.
12. (a) McCamley, K.; Ripper, J. A.; Singer, R. D.; Scammells, P. J. J. Org. Chem. 2003,
68, 9847; (b) Thavaneswaran, S.; Scammells, P. J. Bioorg. Med. Chem. Lett. 2006,
16, 2868.
13. (a) Dong, Z.; Scammells, P. J. J. Org. Chem. 2007, 72, 9881; (b) Kok, G.; Ashton, T.
D.; Scammells, P. J. Adv. Synth. Catal. 2009, 351, 283.
14. Grierson, D. Org. React. 1990, 39, 85.
15. Noviandri, I.; Brown, K. N.; Fleming, D. S.; Gulyas, P. T.; Lay, P. A.; Masters, A. F.;
Philips, L. J. Phys. Chem. B 1999, 103, 6713.
16. General procedure for N-demethylation of the tertiary N-methylamine N-oxide
In summary, a number of opiate and tropane alkaloids, includ-
ing key pharmaceutical intermediates such as oxycodone and oxy-
morphone, are readily N-demethylated in a two-step Polonovski-
type process using a sub-stoichiometric amount of ferrocene. To
the best of our knowledge, this is the first time ferrocene has been
shown to effect a Polonovski reaction. This method offers a number
of advantages with the ferrocene catalyst being inexpensive and
readily available, as well as being air and thermally stable. If de-
sired, most of the catalyst could readily be recovered from the
reaction via a simple extraction with hexane or column chroma-
tography. The reaction is mild and, as demonstrated for substrates
such as oripavine, morphine, and oxymorphone, does not require
protection of functional groups such as hydroxyl. As ferrocene is
both stable and soluble in most organic solvents, we have been
able to demonstrate that the reaction medium greatly influences
the outcome of Fe(II)-mediated N-demethylation of opiate and tro-
pane alkaloids, both with respect to completion time as well as to
the product yield ratio of starting tertiary amine and N-nor
product.
In general, the N-demethylation of tertiary amine-N-oxides
with ferrocene proceeds in a comparable or superior yield to the
previously reported Polonovski reactions that employed FeS-
O4Á7H2O12 or Fe(II)TPPS [tetrasodium 5,10,15,20-tetra(4-sulfophe-
nyl)porphyrinatoiron(II)]13 in cases where a direct comparison
can be made. Furthermore, ferrocene is required in amounts as
low as 0.05 molar equiv, whereas the best results with FeSO4Á7H2O
required an excess of the Fe(II) species (typically 2 molar equiv).
hydrochlorides with ferrocene:
A solution or slurry of the tertiary N-
methylamine N-oxide hydrochloride (approx. 0.27 mmol), ferrocene (0.05–
2.0 molar equiv) and CHCl3, CH2Cl2, MeCN, MeOH or i-PrOH (10 mL) was
stirred at 40–80 °C until reaction was complete or for the specified length of
time. The reaction mixture was concentrated to dryness to give a crude
mixture of the hydrochloride salt of the N-nor compound and the starting
tertiary N-methylamine. Pure N-nor compound was isolated via one of the
following methods: Method A: The crude mixture was dissolved in CHCl3 or
CHCl3/i-PrOH (3:1) and the resulting solution was washed with 10% aqueous
NaOH, dried (Na2SO4), filtered and concentrated. The remaining residue was
subjected to column chromatography on SiO2, eluting with a gradient of CHCl3/
MeOH/NH4OH (90:10:1–85:15:1) or, in the case of 15a/15c, using ethyl
acetate/MeOH/NH4OH (70:30:1–60:40:1), which provided first the starting
tertiary amine followed by the N-nor product. Method B: extraction of an
aqueous solution of the crude at pH 2–10 (adjusted with concd NH4OH) with a
suitable solvent or solvent system. Method C: 10% aqueous HCl was added and
the solution was heated at 50 °C for 1–48 h. The pH of the resultant solution
was adjusted to 2–10 (concd NH4OH) before extractions with a suitable solvent
or solvent system. Method D: as per Method A, with column chromatography
using a CHCl3/MeOH gradient (97:3–9:1).
17. Zara, A. J.; Machado, S. S.; Bulhões, L. O. S. J. Electroanal. Chem. 1987, 221, 165.
and references cited therein.