One-carbon homologation of N-sulfonylaziridines to allylic amines using dimethylsulfonium methylide

Article abstract of DOI:10.1021/ol051124p

(Chemical Equation Presented) Regio- and stereodefined allylic N-sulfonylamines are synthesized in high yields and under experimentally straightforward conditions by reaction of N-sulfonylaziridines with excess dimethylsulfonium methylide.

Full text of DOI:10.1021/ol051124p

ORGANIC
LETTERS
2005
Vol. 7, No. 15
3295-3298
One-Carbon Homologation of
N-Sulfonylaziridines to Allylic Amines
Using Dimethylsulfonium Methylide
David M. Hodgson,* Matthew J. Fleming, and Steven J. Stanway†
Department of Chemistry, UniVersity of Oxford, Chemistry Research Laboratory,
Mansfield Road, Oxford, OX1 3TA, UK
daVid.hodgson@chem.ox.ac.uk
Received May 14, 2005
ABSTRACT
Regio- and stereodefined allylic N-sulfonylamines are synthesized in high yields and under experimentally straightforward conditions by
reaction of N-sulfonylaziridines with excess dimethylsulfonium methylide.
Allylamines (e.g., 4) are important building blocks in organic
chemistry, and therefore methods for their synthesis are of
significance.1 Because of the increasing availability of
aziridines (e.g., 1), typically accessed from alkenes,2 the
eliminative ring-opening of aziridines constitutes an attractive
strategy to access allylamines. The latter has been achieved
from aziridines by, for example, cobalt(I)-catalyzed3 or base-
induced4 isomerization or by deoxygenation5 or deiodination6
of 2-hydroxymethyl or 2-iodomethyl derivatives, respec-
tively. The alkylative ring-opening of 2-(alkoxymethyl)-
aziridines to give 2-substituted 1-amino-2-alkenes has also
been recently reported.7 In seeking a straightforward method
to homologate aziridines to allylic amines, we were attracted
to a report in 1991 by Nadir and co-workers that focused on
the synthesis of azetidines from N-arylsulfonylaziridines
using dimethyloxosulfonium methylide.8
During that study, it was observed that reaction of excess
(4 equiv) dimethylsulfonium methylide 29 with trans-
aziridines 1 (R ) Ph; Ar ) Ph, p-ClC6H4) gave allylic amines
4 (Scheme 1); the corresponding cis-2,3-diphenyl-substituted
† GlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow,
Essex CM19 5AW, UK.
(1) For reviews, see: (a) Cheikh, R. B.; Chaabouni, R.; Laurent, A.;
Mison, P.; Nafti, A. Synthesis 1983, 685-700. (b) Johannsen, M.; Jørgensen,
K. A. Chem. ReV. 1998, 98, 1689-1708.
(2) (a) Jeong, J. U.; Tao, B.; Sagasser, I.; Henniges, H.; Sharpless, K.
B. J. Am. Chem. Soc. 1998, 120, 6844-6845. (b) Gontcharov, A. V.; Liu,
H.; Sharpless, K. B. Org. Lett. 1999, 1, 783-786. (c) Kim, S. K.; Jacobsen,
E. N. Angew. Chem., Int. Ed. 2004, 43, 3952-3954. (d) Bartoli, G.; Bosco,
M.; Carlone, A.; Locatelli, M.; Melchiorre, P.; Sambri, L. Org. Lett. 2004,
6, 3973-3975. For reviews, see: (e) Tanner, D. Angew. Chem., Int. Ed.
Engl. 1994, 33, 599-619. (f) Jacobsen, E. N. In ComprehensiVe Asymmetric
Catalysis; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-
Verlag: Berlin, 1999; Vol. 2, pp 607-618. (g) Mu¨ller, P.; Fruit, C. Chem.
ReV. 2003, 103, 2905-2919.
Scheme 1. Homologation of Aziridines to Allylic Amines
(3) da Zhang, Z.; Scheffold, R. HelV. Chim. Acta 1993, 76, 2602-2615.
(4) (a) Mordini, A.; Russo, F.; Valacchi, M.; Zani, L.; Degl’Innocenti,
A.; Reginato, G. Tetrahedron 2002, 58, 7153-7163. (b) Mu¨ller, P.; Riegert,
D.; Bernardinelli, G. HelV. Chim. Acta 2004, 87, 227-239.
(5) Dickinson, J. M.; Murphy, J. A. Tetrahedron 1992, 48, 1317-1326.
(6) Yadav, J. S.; Bandyopadhyay, A.; Reddy, B. V. S. Synlett 2001,
1608-1610.
and trans-2,3-dimethyl-substituted aziridines were found not
to be viable substrates, and the process was not examined
further. The reaction likely occurs by ring-opening of the
10.1021/ol051124p CCC: $30.25
© 2005 American Chemical Society
Published on Web 06/21/2005

aziridine 1 by ylide 2 acting as a nucleophile to give betaine
3, from which elimination of Me2S takes place by either (or
both) of two pathways (A and B).10 In the present paper, we
demonstrate that this transformation, in fact, holds consider-
able promise for the one-carbon homologation of aziridines
to allylic amines.
One important requirement for successful homologation
of an aziridine by the above process is that the rate of
nucleophilic ring-opening of the aziridine is faster than (or
competitive with) decomposition11 of the labile ylide 2.
Therefore, we initially investigated the reaction of dimeth-
ylsulfonium methylide 2 with terminal aziridines 5 possessing
different protecting groups on nitrogen (Bn, Boc, and Ts,
Table 1). An excess (3 equiv) of ylide 2 was generated under
unreactive, while the N-Boc-protected aziridine 5b underwent
rapid decomposition, giving no identifiable products (Table
1, entries 1 and 2). The N-tosylaziridine 5c gave, after 3 h,
the desired allylic amine 6c (38% yield; 84% yield based
on recovered starting material, entry 3). Extended reaction
times did not improve the yield of 6c, likely due to
decomposition of the ylide 2. However, it was found that
maintaining the reaction at -15 °C for 18 h led to an
improvement in the yield of 6c (80%, entry 4), but some
starting material was still observed.
Increasing the amount of ylide 2 to 5 equiv and carrying
out the reaction from -10 to 5 °C over 3 h improved the
yield of 6c to 91% (entry 5). At this point, we investigated
whether alternative ways of generating the ylide 2 (3 equiv,
using BuMgCl or NaH as the base), changing the solvent
(Et2O, t-BuOMe, or DMSO), or maintaining the reaction at
lower temperatures would improve the yield of allylic amine
6c; however, all these changes were detrimental (entries
6-9). Given the initial success of the N-tosylaziridine 5c,
other sulfonyl protecting groups were investigated. N-
Nosylaziridines are reported to be comparatively more
susceptible to ring-opening by nucleophiles.12 However, in
the present chemistry, only 15% yield of the corresponding
allylic amine 6d (entry 10) was obtained using the best
reaction conditions we had identified for 5c. The yield of
6d could be improved to 45% by maintaining the reaction
temperature at -15 °C (entry 11). An N-Bus (Bus ) tert-
butylsulfonyl)-protected aziridine 5e13 was next examined.
Pleasingly, this gave allylic amine 6e in excellent yield (99%)
using 3 equiv of ylide 2 (entry 12). Reducing the quantity
of ylide used gave a lower yield of 6e together with unreacted
aziridine 5e (entries 13 and 14), supporting the idea that the
excess of ylide 2 acts as the base [rather than the sulfonamido
nitrogen in the putative intermediate betaine (cf. 3, Scheme
1)] to eliminate Me2S. Lowering the temperature and
increasing the reaction time in an attempt to maintain high
yields, while reducing the amount of ylide used, was
unsuccessful (entry 15).
Table 1. Addition of Ylide 2 to Various N-Protected Terminal
Aziridines 5a
dSMe2
(equiv)
temp
(°C)
time
(h)
yield
(%)b
entry
R
1
2
3
4
5
6c
7d
8c,e
9f
10
11
12
13
14
15
Bn
Boc
Ts
Ts
Ts
Ts
Ts
Ts
Ts
Ns
Ns
5a
5b
5c
5c
5c
5c
5c
5c
5c
5d
5d
5
3
3
3
5
3
3
3
3
5
3
3
2.1
1.1
2.1
-10 to 5
-10
-10 to 5
-15
-10 to 5
-10 to 5
-10 to 5
-10 to 5
20
-10 to 5
-15
-10 to 5
-10 to 5
-10 to 5
-15
3
1
3
18
3
3
3
3
5
3
18
3
3
3
18
6a
6b
6c
6c
6c
6c
6c
6c
6c
6d 15
6d 45
6e
6e
6e
6e
0 (99)
0
38 (55)
80 (15)
91
31
28
0
40
Bus 5e
Bus 5e
Bus 5e
Bus 5e
99
63 (35)
40 (56)
78 (20)
The methodology developed above was applied to a range
a Ylide 2 generated from Me3SI and n-BuLi unless indicated otherwise.
b Recovered 5 (%) in parentheses. c Et2O as solvent. d t-BuOMe as solvent.
e BuMgCl used as base. f NaH used as base and DMSO as solvent.
of N-Bus-protected aziridines 7 (Table 2).14
The aziridine 7a, which is readily available in enantiopure
form from (R)-tritylglycidyl ether, gave allylic amine 8a in
90% yield and >99% ee (entry 1).15 This example illustrates
that the homologation process occurs with no loss of
stereochemical integrity. Terminal alkene, (unprotected)
terminal alkyne, and primary chloride functionalities were
tolerated in the reaction (Table 2, entries 2-4). Interestingly,
standard conditions (Me3SI, n-BuLi, THF, -10 °C),11 and
then the aziridine was added and the reaction allowed to
warm slowly to 5 °C over a certain time period. Under these
conditions, the benzyl-protected aziridine 5a proved to be
(7) (a) Hodgson, D. M.; Stefane, B.; Miles, T. J.; Witherington, J. Chem.
Comm. 2004, 2234-2235. (b) Rosser, C. M.; Coote, S. C.; Kirby, J. P.;
O’Brien, P.; Caine, D. Org. Lett. 2004, 6, 4817-4819.
(8) Nadir, U. K.; Sharma, R. L.; Koul, V. K. J. Chem. Soc., Perkin Trans.
1 1991, 2015-2019.
(12) Maligres, P. E.; See, M. M.; Askin, D.; Reider, P. J. Tetrahedron
Lett. 1997, 38, 5253-5256.
(13) (a) Sun, P.; Weinreb, S. M.; Shang, M. J. Org. Chem. 1997, 62,
8604-8608. (b) Hodgson, D. M.; Humphreys, P. G.; Ward, J. G. Org. Lett.
2005, 7, 1153-1156.
(9) For a recent review of dimethylsulfonium methylide 2, see: Li, J. J.
In Name Reactions in Heterocyclic Chemistry; Li, J. J., Ed.; John Wiley &
Sons: Hoboken, 2005; pp 2-14.
(14) Representative Procedure. n-BuLi (1.6 M in hexanes, 0.64 mL,
1.02 mmol) was added dropwise to a stirred suspension of trimethylsulfo-
nium iodide (208 mg, 1.02 mmol) in THF (3 mL) at -10 °C and left stirring
for 15 min. Aziridine 5e (80 mg, 0.34 mmol) in THF (0.5 mL) was added
dropwise, and the reaction was allowed to warm to 5 °C over 3 h. After
quenching with brine solution, the layers were separated. The aqueous layer
was extracted with Et2O; the combined organic layers were dried (MgSO4),
and solvent was evaporated in vacuo. The residue was purified by column
chromatography (SiO2, 30% Et2O in petroleum ether) to give allylic
sulfonamide 6e (83 mg, 99%) as a colorless oil.
(10) For related studies with epoxides, see: (a) Alcaraz, L.; Harnett, J.
J.; Mioskowski, C.; Martel, J. P.; Le Gall, T.; Shin, D.-S.; Falck, J. R.
Tetrahedron Lett. 1994, 35, 5449-5452. (b) Alcaraz, L.; Cridland, A.;
Kinchin, E. Org. Lett. 2001, 3, 4051-4053. (c) Alcaraz, L.; Cox, K.;
Cridland, A. P.; Kinchin, E.; Morris, J.; Thompson, S. P. Org. Lett. 2005,
7, 1399-1401.
(11) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353-
1364.
3296
Org. Lett., Vol. 7, No. 15, 2005

We next examined whether the chemistry could be
broadened to 2,3-disubstituted aziridines other than trans-
2,3-diphenyl-substituted N-arylsulfonylaziridines.8
A range of unsymmetrical aziridines 9 possessing unsat-
uration adjacent to the three-membered ring were found to
undergo homologation with complete (Table 3, entries 1,
3-5) or high (entry 2) selectivity for initial reaction at the
benzylic (or allylic) positions. The synthesis of diene 10e
(75% yield, entry 5) from vinyl aziridine 9e illustrates a way
to access potentially useful substrates for cycloaddition
chemistry.
Table 2. Allylic N-Bus-Amines 8 from Terminal and
2,2-Disubstituted N-Bus-Aziridines 7
Table 3. Allylic N-Sulfonylamines 10 from Benzylic or Allylic
N-Sulfonylaziridines 9
a Regioisomeric allylic amine also isolated (12%).15 b Performed with
6 equiv of dSMe2.
a Performed with 5 equiv of dSMe2. b Performed with 6 equiv of dSMe2.
c Ratio 82:18.
Both N-Bus and N-Ts cyclopentyl- and cyclohexyl-fused
aziridines gave the corresponding allylic amines in good
yields (Table 4).
primary alkyl bromide-containing aziridine 7f and bisaziri-
dine 7g underwent clean double homologations to give dienes
8f16 and 8g (entries 6 and 7). The styrene-derived aziridine
7h gave a regioisomeric mixture of allylic amines (95%
combined yield, 82:18 in favor of initial attack at the terminal
position, entry 8). The chemistry was also successfully
extended to 2,2-disubstituted aziridines to give the corre-
sponding tertiary allylic amines 8i,j in excellent yields
(entries 9 and 10). In these latter cases 5 equiv of ylide 2
was found to be required for the reactions to go to
completion.
Table 4. Allylic N-Sulfonylamines 12 from Cyclopentyl- and
Cyclohexyl-Fused N-Sulfonylaziridines 11
entry
R
n
yield (%)
1
2
3
4
11a
11b
11c
11d
Ts
Bus
Ts
1
1
2
2
12a
12b
12c
12d
82
92
86
94
(15) See Supporting Information for details.
(16) Alcaraz, L.; Harnett, J. J.; Mioskowski, C.; Martel, J. P.; Le Gall,
T.; Shin, D. S.; Falck, J. R. Tetrahedron Lett. 1994, 35, 5453-5456. In
our system, it was not necessary to add LiI to avoid substantial competing
E2 elimination of the bromide.
Bus
Org. Lett., Vol. 7, No. 15, 2005
3297

These latter results demonstrate that unsaturation adjacent
to the three-membered ring is not a requirement for effective
homologation at substituted positions. However, the chem-
istry could not be usefully extended to aziridines bearing
acyclic alkyl substituents (Table 5). Both cis and trans
aziridines 13a and 13b gave either poor or no yield of the
desired allylic amine with either high or quantitative recovery
of starting material. Lowering the reaction temperature and
increasing the reaction time did not have a beneficial effect.
In summary, we have established a process of useful
generality for the conversion of N-sulfonylaziridines to
regiodefined one-carbon-homologated allylic N-sulfonyl-
amines. The method uses readily available reagents and
occurs under experimentally straightforward conditions.
Additional studies in the area of aziridines and sulfonium
ylides are currently underway.
Table 5. Attempted Homologation of 2,3-Dialkyl-Substituted
N-Sulfonylaziridines 13
Acknowledgment. We thank the EPSRC and Glaxo-
SmithKline for a CASE award and the EPSRC National Mass
Spectrometry Service Centre for mass spectra.
Supporting Information Available: Experimental pro-
entry
R
yield (%)
cedures and characterization data for all new aziridines and
1
allylic amines and H and 13C NMR spectra for all allylic
1
2
3
4
trans-13a
trans-13b
cis-13a
Ts
Bus
Ts
14a
14b
14a
14b
5
0
31
0
amines. This material is available free of charge via the
Internet at http://pubs.acs.org.
cis-13b
Bus
OL051124P
3298
Org. Lett., Vol. 7, No. 15, 2005