An improved process for the N-alkylation of indoles using chiral N-protected 2-methylaziridines

Article abstract of DOI:10.1021/op020078v

An improved process for the N-alkylation of indoles using N-protected homochiral aziridines has been developed. This procedure allows reduced quantities of homochiral starting material to be used and leads to improved overall yields and operability.

Full text of DOI:10.1021/op020078v

Organic Process Research & Development 2003, 7, 22−24
An Improved Process for the N-Alkylation of Indoles Using Chiral N-Protected
2
-Methylaziridines
,†
‡
‡
†
Paul R. Giles,* Mark Rogers-Evans, Milan Soukup, and John Knight
Vernalis Research Ltd., Oakdene Court, 613 Reading Road, Winnersh, Wokingham RG41 5UA, UK, and
F. Hoffmann-La Roche Ltd., CH-4070 Basel, Switzerland
Abstract:
Scheme 1. Initial alkylation conditions for the synthesis of
indoles 3
An improved process for the N-alkylation of indoles using
N-protected homochiral aziridines has been developed. This
procedure allows reduced quantities of homochiral starting
material to be used and leads to improved overall yields and
operability.
Introduction
RO 60-0175 4a is a selective 5HT2C agonist with potential
therapeutic utility in the treatment of obsessive compulsive
1
disorder. Its synthesis includes a key step (Scheme 1)
involving the reaction of indole 1a with the N-protected
2
,3
alaninyl mesylate 2 to give the N-alkylated indole 3a.
In our hands this transformation proved to be less than
ideal for several reasons:
the amounts of indole 1b, alkylated indole 3b, and depro-
tected alkylated indole 4b. The HPLC method,4 using
1-methylnaphthalene as internal standard, gave baseline
separation of each of these entities.
(1) The reaction lacked generality, giving inconsistent
yields and incomplete conversion in the synthesis of close
analogues of indole 3a.
(
2) Incomplete reaction gave a mix of materials, which
A number of parameters were independently investigated
with the following results:
proved difficult to purify without recourse to chromatogra-
phy.
Temperature. Higher reaction temperatures increased the
loss of the BOC group to give amine 4b.
(
3) Alaninyl mesylate 2 required very slow addition to
the reaction mixture, typically 1-2 h.
4) The reaction conditions resulted in a partial loss of
the BOC protecting group to give indoles 4.
5) Large solvent volumes, typically 20-50 vols, and thus
t
Base. Replacement of KOH with KO Bu resulted in
(
increased loss of the BOC group to give amine 4b. The use
of powdered KOH prepared in a blender and KOH ground
with a mortar and pestle gave similar results.
(
large aqueous volumes, were required during the workup of
indoles 3.
Rate and Order of Addition of Compound 2. Slow addition
of 2 gave better initial alkylation although, by the end of
the reaction, yields of each entity were similar. It was also
found that the order of addition of reagents was critical. The
desired reaction was only observed when a solution of 2 was
added to a suspension of indole and powdered KOH. The
KOH could not be added to a solution of compound 2 and
the indole.
Reaction Times. Prolonged reaction times decreased the
yield of 3b, mainly by loss of the BOC group to give 4b.
The alkylation reaction was effectively over once the addition
of mesylate 2 was complete.
We sought a method which would circumvent these
problems and provide a scaleable, less capricious, and more
operable procedure. We therefore carried out a limited
amount of process development to improve this transforma-
tion.
Results and Discussion
The reaction illustrated in Scheme 1 using indole 1b and
2
literature conditions was monitored by HPLC to quantify
*
Corresponding author. E-mail p.giles@vernalis.com. Telephone: + 44 118
Water Content. The use of dry DMSO and standard grade
DMSO gave similar results.
9
77 3133. Fax: + 44 118 989 9300.
†
Vernalis Research Ltd.
F. Hoffmann-La Roche Ltd.
‡
(
(
(
1) Boes, M.; Jenck, F.; Martin, J. R.; Moreau, J.-L.; Sleight, A. J.; Wichmann,
J.; Widmer, U. J. Med. Chem. 1997, 40, 2762-2769.
2) Rogers-Evans, M.; Soukup, M. PCT Int. Appl., WO 9747598 A1, 1997;
Chem. Abstr. 1998, 128, 75296.
3) Adams, D. R.; Bentley, J. M.; Roffey, J. R. A.; Hamlyn, R. J.; Gaur, S.;
Duncton, M. A.; Bebbington, D.; Monck, N. J.; Dawson, C. E.; Pratt, R.
M.; George, A. R. PCT Int. Appl. WO 0012475, 2000; Chem. Abstr. 2000,
(4) The internal standard was weighed into the reaction vessel prior to addition
of the solvent and base. Samples (0.5 mL) of the reaction mixture were
taken for analysis and diluted with 10% aqueous acetic acid and methanol
prior to injection onto HPLC using the Vernalis conditions as described in
the Experimental Section. Typical retention times of the species are indole
1b, 3.76 min; alkylated product 3b, 5.78 min; internal standard 9.8 min;
deprotected amine 4b, 2.85 min.
1
32, 194289.
2
2
•
Vol. 7, No. 1, 2003 / Organic Process Research & Development
10.1021/op020078v CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/14/2002

Scheme 2. Unexpected reaction
Scheme 3. Improved process for the alkylation of indoles 1
Table 1. Comparison of improved yields and reaction
conditions for indole alkylations
Stirring Efficiency. The use of either a mechanical stirrer
or a magnetic follower to mix the resulting suspension gave
similar results on a small scale (up to 5 g).
Protecting Group in Compound 2. Changing the N-BOC
protecting group to N-tosyl, N-benzylidine imine, or N-CBz
resulted in no improvement in the overall alkylation.
LeaVing Group in Compound 2. Similarly, the use of
O-tosyl, O-nosyl, and iodo gave no improvement in the
overall alkylation process.
% purity
% area
previous
yieldc
entry
indole 1
yield %a NMRb purity HPLCb
1
2
3
4
5
6
indole
80
79
86
89
85
82
85
85
92
85
90
90
85
82
93
88
75
88
not reported
56
29
11
not reported
59
6-bromoindole
5-chloroindole
6-trifluoromethylindole
5-methoxyindole
5-fluoro-6-chloroindole
a
b
All figures are for crude reaction products. The major impurity in each
case was the unreacted starting indole. c Yields previously obtained in-house
9
after chromatography using initial general procedure.
The identification by proton NMR of the chiral N-BOC
aziridine 5 as a byproduct in the crude reaction product
mixtures from many of the above investigations led to a
suspicion that N-BOC alaninyl mesylate 2 may be unstable
under the reaction conditions. Indeed, compound 2 is
also gave the same outcome, as determined by HPLC
analysis.
Observations during our initial investigations also led us
to suspect that the BOC protected indole 3b may also be
completely and rapidly (less than 5 min) converted to
unstable to the reaction conditions. Examination of a sample
t
aziridine 5 on treatment with KO Bu in d
6
-DMSO at ambient
t
of indole 3b treated with K OBu in d
6
-DMSO by proton
temperatures. The use of powdered KOH resulted in a slower
transformation unless D O was added to the solution, in
NMR demonstrated a slow conversion to the deprotected
indole 4b (1:1 mixture of 3b and 4b after 24 h). We therefore
investigated the reduction or omission of base in this
procedure.
2
which case conversion was more rapid presumably as a result
of homogenising the reaction mixture. Aziridine 5 was
t
prepared by treatment of mesylate 2 (5 g) with KO Bu (1.5
Treatment of indole 3b with aziridine 5 in DMSO at 40
g) in THF (83% yield, 95% purity by NMR).5
°
C in the absence of base resulted in no reaction. However,
These observations led us to suspect that aziridine 5 might
be useful as an alternative, more stable alkylating agent in
this reaction. Various N-protected aziridines have been
reacted with N-lithiated indoles to afford N-alkylated and
alkylation was observed after adding a small quantity of
powdered KOH to the aziridine-indole solution. This
demonstrated that not only could substoichiometric quantities
of base be used but, more importantly, the order of the
addition of reagents in this modified procedure was no longer
key to the outcome (Scheme 3).
Further investigation led to an optimised procedure (see
Experimental Section) with only 0.2 equiv of KOH being
necessary. The generality of the procedure was demonstrated
by the conversion of a number of substituted indoles, some
of which had previously proved problematic using the
literature procedure, to the corresponding alkylated indoles
3. The yields and product purities were superior to those
obtained by the initial general procedure (see Table 1, entries
1-6).
3
-alkylated products, the exact ratios depending on the
6
reaction solvent and the nature of the N-protecting group.
Indoles and N-alkyl indoles afford tryptamine derivatives on
reaction of aziridines under Lewis acid catalysis.
7
This evidence was further supported by the observation
that alkylation with a regioisomer of 2 did not give the
desired alkylation product 6, but the unexpected regioisomer
3
(Scheme 2), derived from the ring-opening of the inter-
8
mediate aziridine.
Instead of direct displacement of the mesylate group, the
replacement of mesylate 2 with aziridine 5 using the literature
conditions (2.4 equiv of alkylating agent added to indole-
KOH slurry in DMSO) resulted in efficient alkylation of
indole 1b. It was noted that using only 1.2 equiv of aziridine
Conclusions
We have developed an improved technical process for
the efficient N-alkylation of indoles using the N-protected
homochiral aziridine 5. The optimised procedure minimises
the required number of equivalents of homochiral reagent,
the volumes of solvent and thus the quantity of waste
products. The reduced quantity of base used also reduces
unwanted side reactions. The new procedure also has
improved operability, yields, and conversions over a range
of indole substrates.
(5) Conversion of (S)-BOC-alaninyl mesylate 2 to the corresponding aziridine
5
has been reported using 4 equiv of KOH and 1.2 equiv TsCl in ether; see
Wessig, P.; Schwarz, J. Synlett 1997, 893.
(6) Onistschenko, A.; Stamm, H. Chem. Ber. 1989, 122, 2397.
3 2
(7) For an example using BF .Et O: Shima, I.; Shimazaki, N.; Imai, K.; Hemmi,
K.; Hasimoto, M. Chem. Pharm. Bull. 1990, 38, 564; using scandium
triflate: Bennani, Y. L.; Zhu, G.-D.; Freeman, J. C. Synlett 1998, 754.
8) Mansell, H. L. Unpublished results.
9) Roffey, J. R. A.; Duncton, M. A.; Dawson, C. E.; Hamlyn, R. J. Unpublished
results.
(
(
Vol. 7, No. 1, 2003 / Organic Process Research & Development
•
23

Experimental Section
Synthesis of Aziridine (5). To a solution of alaninyl
mesylate 2 (5.2 g, 20.5 mmol) in THF (100 mL) was added
solid potassium tert-butoxide (2.5 g, 22.3 mmol) portionwise
over 5 min. The reaction mixture was stirred for 10 min,
checked by TLC to ensure complete reaction (condition a;
Reagents and solvents were used as received from
commercial suppliers. All equipment was inspected visually
for cleanliness and integrity before use.
Analytical HPLC was performed on a Hewlett-Packard
1
2
2
050 or 1100 system with UV detection at a wavelength of
10 nm using a LiChrospher 100 RP18 endcapped (5 µm)
50 mm × 4 mm column (Merck) and gradient elution with
R
f
product 0.7, starting material, R
water (100 mL), washed with diethyl ether (1 × 100 mL, 1
), and evaporated to afford
f
0.2) then diluted with
× 50 mL) and brine, dried (MgSO
4
H
2
O-MeCN (Roche) or a Perkin-Elmer series 200 system
the aziridine 5 (2.7 g, 83% crude yield).
with UV detection at a wavelength of 210 nm using a
Improved General Procedure starting from 5-Fluoro-
6-chloroindole (1a). To a solution of aziridine 5 (0.566 g,
3.54 mmol) and indole 1a (0.5 g, 2.90 mmol) in DMSO (5
mL) was added ground potassium hydroxide (34 mg, 0.2
equiv) and the resulting solution was stirred at 25-40 °C
and monitored by either TLC (condition b) or HPLC. When
judged to be complete, typically 1-8 h, the reaction mixture
was poured into ice-water (20 vols) and allowed to solidify;
the solid was filtered off and dried to give alkylated indole
3a, crude yield 82%, 88% purity by HPLC.
+
Supercosil ABZ 150 mm × 6 mm column and isocratic
elution with MeOH-ammonium acetate buffer (70/30 ratio)
(Vernalis). TLCs were perfomed either with (a) isohexanes-
ethyl acetate (1:1) and developed using KMnO
in DIPE and developed using Ninhydrin spray.
4
dip or (b)
Initial General Procedure Using 5-Fluoro-6-chloroin-
dole (1a). To a suspension of ground potassium hydroxide
(
1
0.58 g, 10.3 mmol) in DMSO (20 mL) was added indole
a (0.44 g, 2.6 mmol) and the resulting suspension was
stirred at 40 °C. An inert atmosphere was not used. After
0 min, a solution of alaninyl mesylate 2 (1.64 g, 6.5 mmol)
Acknowledgment
3
We thank Dr. Ken Heatherington (Vernalis) for his help
during this investigation and Antoine Ritter and Gerd
Schaffner (Roche) for skillful synthetic assistance.
in DMSO (20 mL) was added slowly over a period of 2 h,
and the resulting mixture was stirred overnight at 40 °C.
The viscous gel-like suspension was poured into ice-water,
extracted with diethyl ether (2 × 20 mL), washed with water
Received for review September 6, 2002.
OP020078V
4
and brine, dried (MgSO ), and evaporated to afford the
alkylated indole 3a, 0.5 g, 59% yield after purification).
24
•
Vol. 7, No. 1, 2003 / Organic Process Research & Development