Stereoselective synthesis of N-alkylaziridines from N-chloroamines

Article abstract of DOI:10.1039/b608504k

We report the first racemic and stereoselective synthesis of cis- and trans-N-alkylaziridines via N-chloroamines; using this methodology an N-3,4,5-trimethoxybenzylaziridine was synthesised and efficiently cleaved, affording the corresponding NH aziridine

Full text of DOI:10.1039/b608504k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Stereoselective synthesis of N-alkylaziridines from N-chloroamines
Sean P. Bew,*a D. L. Hughes,a Nicholas J. Palmer,b Vladimir Savic,{b Katy M. Soapia and
Martin A. Wilsona
Received (in Cambridge, UK) 20th June 2006, Accepted 24th July 2006
First published as an Advance Article on the web 5th September 2006
DOI: 10.1039/b608504k
i.e. toxic/expensive metal salts or difficult to handle toxic Lewis
acids. Furthermore, generating carbenoids or nitrenes from diazo
species is, particularly on a large scale, potentially hazardous,12 as
is the synthesis/storage of the diazo precursors.13
We report the first racemic and stereoselective synthesis of cis-
and trans-N-alkylaziridines via N-chloroamines; using this
methodology an N-3,4,5-trimethoxybenzylaziridine was synthe-
sised and efficiently cleaved, affording the corresponding NH
aziridine in high yield.
It is apparent that alternative N-alkylaziridine syntheses are
required.
Refluxing an acetonitrile solution of 4-methoxybenzylamine and
methyl 3-bromopropionate afforded b-amino ester 1 in good yield
(Scheme 2). Critical for the success of the proposed aziridination
procedure, we undertook the N-chlorination of 1. Stirring an
ethereal solution of 1 and NCS afforded 2 in an excellent 94%
yield. With 2 in hand, synthesis of racemic N-(4-methoxybenzyl)
The development of protocols that afford difficult to synthesise
N-alkylaziridines is a challenge yet to be addressed.1 In this
Communication we report that N-chloroamines can be used for the
efficient synthesis of N-alkylaziridines. Our results demonstrate that,
contrarytoliteratureprecedent, N-chloroaminesarerelativelystable,
efficient, readily accessible starting materials that afford structurally
diverse N-alkylaziridines that would, using conventional protocols,
be available only via arduous, inefficient and laborious procedures.
Although N-chloroamines have been utilised for over 125 years,2
they have only been employed in a small number of reaction types
i.e. N-centered radicals,3 Ritter,4 Hoffman–Loffler reactions5 and
Stieglitz rearrangements.6 Their lack of widespread use can be
attributed to their reported instability,7 e.g. in protonating media
they decompose and the relatively weak N–Cl bond is readily
cleaved (photolysis and/or thermal processes).
aziridine-2-carboxylic acid methyl ester
3 was attempted.
Gratifyingly, following deprotonation of 2 racemic 3 was afforded
in an unoptimised 72% yield (Scheme 2).14 1H-NMR analysis of
the crude product 3 revealed that unreacted 2 comprised the
majority of the mass balance (15–20%). Contrary to literature
reports on different systems we could not find any evidence for the
formation of base induced 1,2-elimination products i.e. (E)- and
(Z)-alkanimines15 or of products resulting from nitrene or imine
formation/decomposition.
Confident that our protocol could be applied to the synthesis of
a series of racemic N-alkylaziridines, we investigated the generality
of the reaction. Secondary amines 4a–4h were synthesised in
excellent yields (84–99%) via aza-Michael reactions between
4-methoxybenzylamine, allylamine, tert-butylamine, cyclohexyl-
amine and tert-butyl/methyl acrylate, methyl crotonate/tiglate and
acrylonitrile (Table 1).16 Subsequent N-chlorination (NCS) of 4a–
4h produced 5a–5h in good to excellent yields (64–99%). In a
finding worthy of note and in contrast to reports on
N-chloroamine instability,17 we were surprised by the stability of
5a–5h. For example, no special precautions were required for their
synthesis although on exceptionally sunny days the flask was
wrapped in foil. Furthermore no special purification procedures or
techniques were required for their chromatography on silica. Due
to their ease of synthesis we did not attempt to synthesise and store
the N-chloroamines. It is also worth bearing in mind the potential
of N-chloroamines to have carcinogenic properties. Substituting
We considered that appending a chlorine onto a nitrogen induces
a net polarisation of the N–Cl bond towards the chlorine.
Incorporating this principle we speculated on the ability of
N-chloroamines to act as efficient nitrogen centered electrophiles
forheterocycleformationviaintramolecularcyclisations(Scheme1).
Apart from a report in 19618 on racemic N-tert-butyl-a-amino
acids synthesis there are no reports of any intramolecular
cyclisations that utilise N-chloroamines for aziridine synthesis.
In comparison to N-activated aziridines there are few procedures
that afford N-alkylaziridines. These can be grouped into: addition
of carbenes or ylids to imines;9 addition of nitrenes to alkenes;10
and nucleophile mediated 3-exo-tet cyclisations (Scheme 1).11 The
first two have significant environmental/experimental drawbacks
Scheme 1 Proposed pathway to N-alkylaziridine synthesis.
aSchool of Chemical Sciences & Pharmacy, UEA, Norwich, UK
NR4 7TJ. E-mail: s.bew@uea.ac.uk; Fax: 44 (0)1603 592003;
Tel: 44 (0)1603 593142
bBiofocus Discovery Ltd, Chesterford Research Park, Saffron Waldon,
Essex, UK CB10 1XL. E-mail: vladimir.savic@pharmacy.bg.ac.yu
{ Present address: Department of Chemistry, Vojvode Stepe 450, 11000
Belgrade, Serbia and Montenegro. Tel: 00381 113 970 379 ext 643.
Scheme 2 Synthesis of racemic 3 via N-chloroamine 2.
4338 | Chem. Commun., 2006, 4338–4340
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Table 1 Yields for the individual reaction steps towards 6a–g
R1
R2 R3 R4
4
5
6
4-MeOC6H4CH2
4-MeOC6H4CH2
4-MeOC6H4CH2
H
H
H
H
H
H
H
H
H
H
–CO2Me 4a 89% 5a 94% 6a 72%
–CN 4b 84% 5b 93% 6b 70%
t
–CO2 Bu 4c 87% 5c 65% 6c 39%
–CO2Me 4d 87% 5d 64% 6d 79%
–CO2Me 4e 92% 5e 78% 6e 58%
4-MeOC6H4CH2 Me
H
Allyl
tert-Bu
t
H
Me
H
–CO2 Bu 4f 94% 5f 82% 6f 90%
–CO2 Bu 4g 97% 5g 64% 6g 63%
t
4-MeOBn
4-MeOBn
Me –CO2Me 4h 99% 5h 99%
—
Fig. 1 X-ray crystal structure of (R,R)-8.
NCS for NIS or NBS failed to return N-haloamines that were
suitable for cyclisation. Gratifyingly, when 5a–5g were deproto-
nated the desired 3-exo-tet cyclisation reactions afforded unopti-
mised moderate (39%) to excellent yields (90%) of racemic
N-alkylaziridines 6a–6g. Of note, the cyclisation of 5g (the
secondary amine starting material for 5g was synthesised via an
aza-Michael reaction between para-methoxybenzylamine and
trans-methyl crotonate), a b-methyl substituted ester was tolerated;
racemic trans-6g (J2,3 2.8 Hz) was produced in a 63% yield.
However attempted cyclisation of 5h afforded the corresponding
imine derived from a 1,2-elimination process. Although further
studies are essential, these preliminary results suggest that: the
protocol is substrate tolerant; the process is amenable to nitrogen
substituent variation; the cyclisation procedure accommodates
structurally diverse alkyl esters and that different electron-
withdrawing moieties i.e. nitrile and ester groups are tolerated.
With these racemic results in hand we sought to extend our
methodology to the asymmetric synthesis of N-alkylaziridines.
N-chlorination of the precursor amine afforded (R)-(para-meth-
oxy-a-methylbenzyl)-N-chloroamine 7; its deprotonation and
subsequent cyclisation gave a diastereomeric mixture (83 : 17,
72% yield) of aziridines which were assigned by 1H-NMR as
(R,R)-8 (major) and (R,S)-9 (minor, Scheme 3). Purification of the
major product resulted in a crystal suitable for X-ray analysis.{
The structure (Fig. 1) clearly shows the favorable trans-disposition
between the N-appended group and the ester on the aziridine ring
and confirms the absolute stereochemistry as (R,R)-8.
Using (R)-para-methoxy-a-methylbenzylamine, phenyl vinyl
sulfone and benzyl acrylate as Michael acceptors the correspond-
ing amines were N-chlorinated and cyclised affording diastereo-
meric mixtures of aziridines 9 and 10 in 65% and 87% yields and
75 : 25 diastereomeric ratios respectively (Fig. 2, tentative major
diastereomers shown).
Further studies, employing 7, investigated the effect on the
cyclisation reaction when alternative bases were employed.
Substituting LHMDS (Scheme 3) with NaH, KH, BEMP,
DBU, Et3N or NaOMe returned starting material 7. However
LDA, tert-BuOK, NaHMDS or KHMDS afforded poor to good
yields (20–85%) and diastereoselectivities (50 : 50–81 : 19) of (R,R)-
8 and (R,S)-9 compared with LHMDS. Similarly, raising the
reaction temperature (Scheme 3) from 278 uC to RT had a
relatively small effect on the yields (73 ¡ 8%) and diastereoselec-
tivities (83 : 17–77 : 23) of the resulting (R,R)-8 and (R,S)-9. In
contrast to this, changing the solvent had a significant impact on
both the yield and diastereoselectivity (Table 2). Acyclic and cyclic
ethers (entries 2, 4 and 5), aromatics (entries 6 and 7) and the aryl
ether anisole (entry 3) afforded substantially higher yields and/or
diastereoselectivities than chlorinated solvents. Of note, either
petrol or TBDME (entries 1 and 2 respectively) independently
produced excellent diastereoselectivities and yields of (R,R)-8 and
(R,S)-9 (93 : 7 and 89 : 11; 88% and 82% respectively).
Fig. 2 Asymmetric synthesis of 9 and 10.
Scheme 3 Asymmetric synthesis of N-alkylaziridines.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 4338–4340 | 4339

View Article Online
needle-plates. Intensity data were measured on a Nonius CAD4
Table 2 Probing for solvent effects during the cyclisation of 7
diffractometer (with monochromated radiation); 1400 reflections to
hmax = 20u, the limit of useful diffraction; 1240 unique reflections (Rint
0.034), 685 ‘observed’ with I . 2sI. Corrections were applied for Lorentz-
polarisation effects, slight crystal deterioration, and to eliminate negative
net intensities (by Bayesian statistical methods). Structure determined by
direct methods in SHELXS20 and refined by full-matrix least-squares, on
Entry Solvent
Yield 8 : 9
Entry Solvent
Yield 8 : 9
1
2
3
4
5
6
petrol 88% 93 : 7
TBDME 82% 89 : 11
anisole
THF
ether
7
8
9
benzene 67% 78 : 22
DMF 64% 72 : 28
dioxane 57% 80 : 20
neat 57% 57 : 43
1,2-DCE 15% 77 : 23
DCM 10% 71 : 29
73% 79 : 21
F2’s, in SHELXL.20 At convergence, wR2 = 0.139 and R1 = 0.100 (A2) for
82% 83 : 17 10
40% 88 : 12 11
70% 85 : 15 12
all 1240 reflections weighted w = [s2(Fo2) + (0.0542P)2]21 with P = (Fo
+
2
2Fc2)/3; for the ‘observed’ data only, R1 = 0.056. The Flack parameter, x =
23(5), allows no valid conclusions to be drawn about the absolute
configuration, but the enantiomer shown corresponds with that prepared
from known (R) material. CCDC 611802. For crystallographic data in CIF
or other electronic format see DOI: 10.1039/b608504k
toluene
1 S. P. Bew, D. L. Hughes, V. Savic, K. M. Soapi and M. A. Wilson,
Chem. Commun., 2006, 3513.
2 A. W. Hoffman, Berichte, 1881, 14, 1497.
3 G. Heuger, S. Kalsow and R. Gottlich, Eur. J. Org. Chem., 2002, 1848.
4 K. I. Brooker-Milburn, D. J. Guly, B. Cox and P. A. Procopiou, Org.
Lett., 2003, 5, 3313.
5 R. S. Neale and M. R. Walsh, J. Am. Chem. Soc., 1965, 87, 1255.
6 P. A. Grieco and Y. Dai, J. Am. Chem. Soc., 1998, 120, 5128.
7 L. Stella, Angew. Chem., Int. Ed. Engl., 1983, 22, 337; J. M. Antelo,
F. Arce, J. Crugeiras, J. Franco, F. Lopez, P. Rodriguez and A. Varela,
An. Quim., 1991, 87, 195; S. Wawzonek and J. D. Nordstrom, J. Org.
Chem., 1962, 27, 3726.
8 H. E. Baumgarten, R. L. Zey and U. Krolls, J. Am. Chem. Soc., 1961,
83, 4469.
9 S. Taylor, J. Gullick, N. Galea, P. McMorn, D. Bethell, P. C. B. Page,
F. E. Hancock, F. King, D. J. Willock and G. J. Hutchings, Top. Catal.,
2003, 25, 81.
Scheme 4 Cleavage of the N-3,4,5-trimethoxybenzyl group off trans-12.
The cleavage of N-appended activating groups and even more
so non-activating groups off aziridines requires harsh reagents
and/or reaction conditions that often result in the partial or
complete destruction of the heterocycle. We considered the
synthesis of an oxidatively cleavable electron-rich N-benzyl
substituted aziridine using the N-chlorination–cyclisation metho-
dology reported here to offer a convenient solution to the problem.
Reductive amination of the imine derived from 3,4,5-trimethoxy-
benzaldehyde and tert-butyl (3S)-3-amino-3-phenylpropionate
followed by N-chlorination afforded 11 in an excellent 92% yield
(Scheme 4). Disappointingly, all attempted cyclisations using 11
failed (1,2-elimination products resulted).15 However the corre-
sponding ethyl ester (R = Et, 11), enolate generation and presumed
SN2 cyclisation afforded trans-12 in a 70% yield. Critically, when
trans-12 was reacted with DDQ the N-(3,4,5-trimethoxybenzyl)
substituent cleaved, returning trans-13 in a 75% yield. Attempted
oxidative cleavage using DDQ of the corresponding 4-methoxy-
benzyl group off (R,R)-13 was slow and incomplete.
10 S. Fioravanti, A. Morreale, L. Pellacani and P. A. Tardella, Synlett,
2004, 1083.
11 G. Cardillo, L. Gentilluci and A. Tolomelli, Aldrichimica Acta, 2003, 36,
39; H. M. I. Osborn and J. B. Sweeney, Tetrahedron: Asymmetry, 1997,
8, 1693; P. Garner, O. Dogan, W. J. Youngs, V. O. Kennedy,
J. Protasiewicz and R. Zaniewski, Tetrahedron, 2001, 57, 71.
12 J. R. Fulton, V. K. Aggarwal and J. de Vicente, Eur. J. Org. Chem.,
2005, 1479.
13 Organic Syntheses, Coll. Vol. 3, John Wiley, Chichester, p. 56.
14 A flame dried round bottomed flask was charged (under argon) with 1
(210 mg, 0.97 mmol) in anhydrous diethyl ether (10 mL) followed by
NCS (251 mg, 1.88 mmol). The resulting suspension was stirred for
1 hour at ambient temperature after which time the white precipitate
was removed via filtration. The resulting ether filtrate was concentrated
in vacuo affording a clear oil which was purified via flash chromato-
graphy (silica, hexane : ether, 2 : 1) returning 2 (234 mg, 94%). A
solution of 2 (150 mg, 0.58 mmol) in anhydrous THF (4 mL) was cooled
to 278 uC (under argon); to this was added LHMDS (1.06 M in THF,
0.713 mL, 0.757 mmol). The reaction was stirred for 30 minutes at
278 uC and subsequently quenched with saturated NH4Cl solution.
Extraction and purification of the product (silica, hexane : ethyl acetate,
1 : 1) afforded racemic-3 as a clear oil (93 mg, 72%). Rf 0.2 (ether–
hexane 1 : 2). 1H-NMR (400 MHz, CDCl3) d 7.23 (2H, d, J 8.7), 6.85
(2H, d, J 8.7), 3.78 (3H, s, OCH3), 3.70 (3H, s, OCH3), 3.47 (2H, m),
2.22 (1H, dd, J 3.2, 1.2), 2.18 (1H, dd, J 6.5, 3.2), 1.78 (1H, dd, J 6.5,
1.2); 13C-NMR (125 MHz) d 171.5, 159.2, 129.9, 129.7, 114.1, 63.5, 55.5,
52.5, 37.4, 34.6; IR (nmax/cm21) 1742, 1611.
15 T. M. Chapman, S. Courtney, P. Hay and B. G. Davis, Chem.–Eur. J.,
2003, 9, 3397; B. Han, Z. Wang, B. Jaun, R. Krishnamurthy and
A. Eschenmoser, Org. Lett., 2003, 5, 2071.
16 Attempted aza-Michael addition between ethyl cinnamate, methyl
tiglate and 4-methoxybenzylamine failed.
17 The chemistry of amino, nitroso and nitro compounds and their derivatives,
ed. S. Patai, Wiley, Chichester, 1982, Part 2, p. 1095; Comprehensive
Organic Synthesis, ed. B. Trost, Pergamon, Oxford, 1991, Vol. 7, p. 741.
18 S. Farooq, W. E. Swain, Jr, R. Daeppen and G. Rihs, Tetrahedron:
Asymmetry, 1992, 3, 51–63.
19 K.-D. Lee, J.-M. Suh, J.-H. Park, H.-J. Ha, H. G. Choi, C. S. Park,
J. W. Chang and W. L. Lee, Tetrahedron, 2001, 42, 8267–8276.
20 G. M. Sheldrick, SHELX-97 – Programs for crystal structure
determination (SHELXS) and refinement (SHELXL), University of
Go¨ttingen, Germany, 1997.
In conclusion, we have demonstrated that N-alkyl-b-amino
esters and nitriles are efficiently transformed into the corresponding
N-chloroamines and that these are convenient starting materials for
the synthesis of structurally diverse racemic and optically active
N-alkylaziridines. The regioselective ring opening of 2-substituted-
N-alkylaziridines using a variety of reagents, e.g. triazole,18
hydrogenation,19 azide,19 and halide anions,19 corroborates their
importance as valuable synthetic intermediates. Uniquely, we have
also ascertained that electron-rich N-benzyl substituents can be
efficiently installed/cleaved via exceptionally mild conditions,
affording the corresponding NH aziridine in good yield.
The authors would like to acknowledge UEA and Biofocus for
financial support. The EPSRC Mass Spectrometry Centre at
Swansea is gratefully acknowledged.
Notes and references
{ Crystal structure analysis of (R,R)-1-[1-(4-methoxyphenyl)-ethyl]-aziri-
dine-2-carboxylic acid tert-butyl ester.
Crystal data: C16H23NO3, M = 277.4. Orthorhombic, space group
˚
P212121 (no. 19), a = 5.614(2), b = 11.689(3), c = 25.371(6) A, V =
1665.1(7) A . Z = 4, Dc = 1.106 g cm23, F(000) = 600, T = 293(2) K,
3
˚
m(Mo-Ka) = 0.76 cm21, l(Mo-Ka) = 0.71069 A. Crystals are colourless
˚
4340 | Chem. Commun., 2006, 4338–4340
This journal is ß The Royal Society of Chemistry 2006