Novel synthesis of 3,4-diaminobutanenitriles and 4-amino-2-butenenitriles from 2-(cyanomethyl)aziridines through intermediate aziridinium salts: An experimental and theoretical approach

Article abstract of DOI:10.1021/jo0704210

(Chemical Equation Presented) 1-Arylmethyl-2-(cyanomethyl)aziridines were transformed into 4-(N,N-bis(arylmethyl)amino)-3-(pyrrolidin-1-yl)butanenitriles and 4-(N,N-bis(arylmethyl)amino)-2-butenenitriles via 4-(N,N-bis(arylmethyl) amino)-3-bromobutanenitriles in high yields and purity. The key steps involve the unprecedented regiospecific ring opening of intermediate 2-(cyanomethyl)aziridinium salts by bromide and pyrrolidine in acetonitrile, exclusively at the substituted aziridine carbon atom. The results were rationalized on the basis of ab initio calculations.

Full text of DOI:10.1021/jo0704210

Novel Synthesis of 3,4-Diaminobutanenitriles and
4-Amino-2-butenenitriles from 2-(Cyanomethyl)aziridines through
Intermediate Aziridinium Salts: An Experimental and Theoretical
Approach
Matthias D’hooghe,† Veronique Van Speybroeck,‡ Andries Van Nieuwenhove,†
Michel Waroquier,‡ and Norbert De Kimpe*,†
Department of Organic Chemistry, Faculty of Bioscience Engineering, Ghent UniVersity,
Coupure Links 653, B-9000 Ghent, Belgium, and Center for Molecular Modeling, Laboratory of
Theoretical Physics, Ghent UniVersity, Proeftuinstraat 86, B-9000 Ghent, Belgium
norbert.dekimpe@UGent.be
ReceiVed March 1, 2007
1-Arylmethyl-2-(cyanomethyl)aziridines were transformed into 4-(N,N-bis(arylmethyl)amino)-3-(pyrrolidin-
1-yl)butanenitriles and 4-(N,N-bis(arylmethyl)amino)-2-butenenitriles via 4-(N,N-bis(arylmethyl)amino)-
3-bromobutanenitriles in high yields and purity. The key steps involve the unprecedented regiospecific
ring opening of intermediate 2-(cyanomethyl)aziridinium salts by bromide and pyrrolidine in acetonitrile,
exclusively at the substituted aziridine carbon atom. The results were rationalized on the basis of ab
initio calculations.
Introduction
(anticonvulsant)3 and 3-hydroxy derivatives (carnitine),4 have
received major interest in recent years because of their broad
pharmacological relevance. Furthermore, also â,γ-diamino-
butanoic acids comprise an attractive class of amino acids with
interesting applications as biologically active compounds. Emer-
icedin 1 has been isolated from the fungus Emericella quadri-
lineata and emeriamine 2 is a strong and specific inhibitor of
The search for new pathways toward amino nitriles as
precursors of the corresponding amino acids is an important
challenge in organic synthesis.1 The interesting pharmacological
properties of â- and γ-amino acids and their use in the synthesis
of the corresponding â- and γ-peptides have, especially in the
past decade, renewed the interest of organic chemists.2 3-Sub-
stituted 4-aminobutyric acids in particular, such as 3-alkyl
(2) (a) E. Juaristi, V. Soloshonok, EnantioselectiVe Synthesis of â-Amino
Acids, 2nd ed., John Wiley & Sons: Hoboken, 2005, (b) Abele, S.; Seebach,
D. Eur. J. Org. Chem. 2000, 1. (c) Juaristi, E.; Lu¨pez-Ruiz, H. Curr. Med.
Chem. 1999, 6, 983. (d) Cole, D. C. Tetrahedron 1994, 50, 9517. (e) Fu¨lo¨p,
F. Chem. ReV. 2001, 101, 2181. (f) Trabocchi, A.; Guarna, F.; Guarna, A.
Curr. Org. Chem. 2005, 9, 1127.
* Author to whom correspondence should be addressed. Phone: 32-9-264-
5951. Fax: 32-9-264-6243.
† Department of Organic Chemistry.
‡ Laboratory of Theoretical Physics.
(1) (a) Groger, H. Chem. ReV. 2003, 103, 2795. (b) Enders, D.; Shilvock,
J. P. Chem. Soc. ReV. 2000, 29, 359. (c) Winkler, M.; Martinkova, L.; Knall,
A. C.; Krahulec, S.; Klempier, N. Tetrahedron 2005, 61, 4249. (d) Preiml,
M.; Honig, H.; Klempier, N. J. Mol. Catal. B: Enzym. 2004, 29, 115.
(3) Yuen, P.-w.; Kanter, G. D.; Taylor, C. P.; Vartanian, M. G. Bioorg.
Med. Chem. Lett. 1994, 4, 823.
(4) Stryer, L. Biochemistry; Freeman W. H. and Company: San
Fransisco, 1975.
10.1021/jo0704210 CCC: $37.00 © 2007 American Chemical Society
Published on Web 05/26/2007
J. Org. Chem. 2007, 72, 4733-4740
4733

D’hooghe et al.
SCHEME 1
the computational viewpoint the system under study is chal-
lenging, as the solvent might be crucial for predicting the correct
reaction preference. To validate the importance of the incorpora-
tion of the molecular environment, the reaction route was
modeled at various levels of theory: in the gas phase, with
inclusion of explicit solvent molecules, and by embedding in a
continuum characterized by a dielectric constant which is
characteristic for the solvent. The latter was found to be im-
portant in other studies to capture the bulk electrostatic effects.11
carnitine palmitoyltransferase-2.5 Because of the medicinal
relevance of the latter amino acids, a variety of 3,4-diaminobu-
tanoic acid derivatives have been prepared, mostly starting from
aspartic acid derivatives,6 and evaluated as inhibitors of carnitine
palmitoyl transferase.7 However, very little effort has been made
toward the synthesis of their nitrile precursors, i.e., 3,4-diamino-
butanenitriles, as an alternative approach toward this valuable
class of compounds.8
Results and Discussion
The synthesis of 1-arylmethyl-2-(cyanomethyl)aziridines 6
as suitable synthons in organic chemistry has been reported
previously by treatment of 1-arylmethyl-2-(bromomethyl)-
aziridines,12 a peculiar yet promising class of â-halo amines,
with 1 equiv of potassium cyanide in DMSO and heating at
60-70 °C for 3 h.13 The combination of an aziridine moiety
and a cyano group in these unexplored substrates enables the
preparation of a variety of functionalized amino nitriles through
ring opening reactions of the constrained ring.
Only a limited number of reports are available on the
regioselective ring opening of 2-substituted aziridinium salts
(different from 2-aryl or 2-acyl substitution) at the more hindered
aziridine carbon atom.14 This unique feature offers a useful tool
in targeted organic synthesis and is considered as an important
synthetic advantage of nonactivated 2-substituted aziridines with
regard to their activated counterparts.
Although the preparation of 1,2-diamines from aziridines is
known in the literature,9 no reports are available on the synthesis
of 3,4-diaminobutanenitriles starting from suitably substituted
aziridine derivatives. From a retrosynthetic point of view, the
ring opening of 2-(cyanomethyl)aziridines 3 could offer an easy
access to 3,4-diaminobutanenitriles 4 and/or 5 depending on
the regioselectivity of this ring opening (Scheme 1).
In the present report, an elegant and straightforward approach
toward novel 4-(N,N-bis(arylmethyl)amino)-3-(pyrrolidin-1-yl)-
butanenitriles and 4-(N,N-bis(arylmethyl)amino)-2-butenenitriles
from 1-arylmethyl-2-(cyanomethyl)aziridines is described via
regiospecific ring opening of intermediate aziridinium salts. The
exclusive attack of bromide and pyrrolidine at the more hindered
aziridine carbon atom in the ring opening of 2-(cyanomethyl)-
aziridinium salts is rationalized on the basis of ab initio
calculations. At present, very few in-depth theoretical studies
on the ring opening of aziridinium salts are available.10 From
Treatment of aziridines 6 with 1 equiv of an arylmethyl
bromide in acetonitrile afforded novel 4-amino-3-bromobutane-
nenitriles 8 in good yields after several days at room temperature
(Scheme 2, Table 1). In this transformation, the arylmethyl
bromide is responsible for both the activation of the aziridine
ring toward an aziridinium intermediate and the delivery of the
nucleophilic bromide anion which induces ring opening of the
aziridinium ion. Since only one similar compound has been
reported in the literature, i.e., N-(2-bromo-3-cyanopropyl)-
succinimide,15 the presented methodology offers an elegant
approach toward this new class of versatile substrates 8 suitable
for further elaboration because of the presence of a γ-amino
nitrile moiety and a â-bromo amine in one structural entity.
In accordance with the previously observed reactivity of
2-(bromomethyl)-, 2-(aryloxymethyl)-, and 2-(alkanoyloxy-
methyl)aziridines toward arylmethyl bromides in acetonitrile,16
a regioselective ring opening of the intermediate 2-(cyanometh-
yl)aziridinium salts 7 by bromide occurred at the more hindered
(5) (a) Kanamara, T.; Shinagawa, S.; Asai, M.; Okazaki, H.; Sugiyama,
Y.; Fuijita, T.; Iwatsuka, H.; Yoneda, M. Life Sci. 1985, 37, 217. (b) Chiodi,
P.; Maccari, F.; Ramacci, M. T. Biochim. Biophys. Acta 1992, 1127, 81.
(6) (a) Thierry, J.; Servajean, V. Tetrahedron Lett. 2004, 45, 821. (b)
Calvisi, G.; Dell’Uomo, N.; De Angelis, F.; Dejas, R.; Giannessi, F.; Tinti,
M. O. Eur. J. Org. Chem. 2003, 4501. (c) Abdel-Aziz, M.; Yamasaki, T.;
Otsuka, M. Bioorg. Med. Chem. Lett. 2003, 13, 1041. (d) Bruckner, A. M.;
Schmitt, H. W.; Diederichsen, U. HelV. Chim. Acta 2002, 85, 3855. (e)
Markidis, T.; Kokotos, G. J. Org. Chem. 2001, 66, 1919. (f) Misiti, D.;
Santaniello, M.; Zappla, G. Synth. Commun. 1992, 22, 883.
(7) Giannessi, F.; Tassoni, E.; Tinti, M. O.; Conti, R.; Dell’uomo, N.;
Brunetti, T. PCT Int. Appl. WO 2006092204 A1, 2006; Chem. Abstr. 2006,
145, 293349.
(8) (a) Jones, D. S.; Srinivasan, A.; Kasina, S.; Fritzberg, A. R.;
Wilkening, D. W. J. Org. Chem. 1989, 54, 1940. (b) Gmeiner, P.; Hummel,
E.; Haubmann, C.; Ho¨fner, G. Arch. Pharm. (Weinheim) 1995, 328, 265.
(c) Einsiedel, J.; Hubner, H.; Gmeiner, P. Bioorg. Med. Chem. Lett. 2003,
13, 851. (d) Baeza, A.; Casas, J.; Na´jera, C.; Sansano, J. M. J. Org. Chem.
2006, 71, 3837.
(9) From nonactivated aziridines: (a) Anand, R. V.; Pandey, G.; Singh,
V. K. Tetrahedron Lett. 2002, 43, 3975. (b) Katagiri, T.; Takahashi, M.;
Fujiwara, Y.; Ihara, H.; Uneyama, K. J. Org. Chem. 1999, 64, 7323. (c)
Sekar, G.; Singh, V. K. J. Org. Chem. 1999, 64, 2537. (d) Padwa, A.; Dent,
W. J. Org. Chem. 1987, 52, 235.
(10) (a) Bouyacoub, A.; Jean, Y.; Volatron, F. J. Mol. Struct.
(THEOCHEM) 1996, 371, 51. (b) Silva, M. A.; Goodman, J. M. Tetrahedron
Lett. 2005, 46, 2067.
(11) (a) Van Speybroeck, V.; Moonen, K.; Hemelsoet, K.; Stevens, C.
V.; Waroquier, M. J. Am. Chem. Soc. 2006, 128, 8468. (b) Ando, K. J.
Am. Chem. Soc. 2005, 127, 3964. (c) Pratt, L. M.; Streitwieser, A. J. Org.
Chem. 2003, 68, 2830.
(12) (a) De Kimpe, N.; Jolie, R.; De Smaele, D. J. Chem. Soc., Chem.
Commun. 1994, 1221. (b) De Kimpe, N.; De Smaele, D.; Szakonyi, Z. J.
Org. Chem. 1997, 62, 2448. (c) D’hooghe, M.; Waterinckx, A.; De Kimpe,
N. J. Org. Chem. 2005, 70, 227. (d) D’hooghe, M.; Rottiers, M.; Jolie, R.;
De Kimpe, N. Synlett 2005, 931.
(13) (a) D’hooghe, M.; Mangelinckx, S.; Persyn, E.; Van Brabandt, W.;
De Kimpe, N. J. Org. Chem. 2006, 71, 4232. (b) D’hooghe, M.; Vervisch,
K.; Van Nieuwenhove, A.; De Kimpe, N. Tetrahedron Lett. 2007, 48, 1771.
(14) For a recent example, see: Savoia, D.; Alvaro, G.; Di Fabio, R.;
Gualandi, A.; Fiorelli, C. J. Org. Chem. 2006, 71, 9373.
(15) Bailey, W. J.; Bello, J. J. Org. Chem. 1955, 20, 525.
(16) (a) D’hooghe, M.; Van Brabandt, W.; De Kimpe, N. J. Org. Chem.
2004, 69, 2703. (b) D’hooghe, M.; Waterinckx, A.; Vanlangendonck, T.;
De Kimpe, N. Tetrahedron 2006, 62, 2295. (c) D’hooghe, M.; Van
Speybroeck, V.; Waroquier, M.; De Kimpe, N. Chem. Commun. 2006, 1554.
(d) D’hooghe, M.; De Kimpe, N. Synlett 2006, 2089.
4734 J. Org. Chem., Vol. 72, No. 13, 2007

Synthesis of 3,4-Diaminobutanenitriles
SCHEME 2
TABLE 1. Synthesis of 4-Amino-3-bromobutanenitriles 8 Starting
from 1-Arylmethyl-2-(cyanomethyl)aziridines 6
the neurotransmitter γ-aminobutyric acid (GABA), and the
second the manifold efforts reported in the literature for the
production of L-carnitine from crotonobetaine.17
entry
R1
R2
time (d)
compd
yield (%)a
Another useful elaboration of 4-amino-3-bromobutanenitriles
8 comprises their transformation into â,γ-diaminobutanenitriles
as precursors for the corresponding amino acid derivatives.
Treatment of â-bromoamines 8 with 2 equiv of pyrrolidine in
refluxing acetonitrile afforded 3,4-diaminobutanenitriles 10 as
the sole reaction products in good yields after 3 h (Scheme 4).
This reaction can be explained considering the formation of an
intermediate aziridinium salt 7, followed by nucleophilic ring
opening of the latter by pyrrolidine at the more hindered aziri-
dine carbon atom (Scheme 4). The structural identity of 4-(N,N-
bis(arylmethyl)amino)-3-(pyrrolidin-1-yl)butanenitriles 10 (and
not the regioisomers) was determined by detailed spectroscopic
analysis and confirmed by comparison with literature data of
3-(N,N-dibenzylamino)-4-(pyrrolidin-1-yl)butanenitrile.8b
From a synthetic point of view, the ring opening of aziri-
dinium salts by nitrogen nucleophiles constitute a powerful
method for the preparation of 1,2-diamino derivatives,18 although
very little reports on this type of chemistry are available in the
literature. Recently, the synthesis of methyl R,â-diaminocar-
boxylates by treatment of methyl 3-phenyl-2-(trifluoromethane-
sulfonyloxy)propenoate with secondary amines has been re-
ported, in which ring opening of the intermediate aziridinium
salts occurred at the benzylic position.19 However, elimination
toward the corresponding enamines appeared to be a common
side-reaction in this type of transformations. This side-effect
was avoided by means of a N,N-dibenzyl-protected serine methyl
ester, leading to chiral R,â-diamino esters via ring opening of
intermediate aziridinium salts by amines.20 Other rare examples
of the ring opening of aziridinium salts by amines include the
reaction of a 2-(trifluoromethyl)aziridinium salt with N-benzyl-
N-methylamine and butylamine, affording the corresponding 1,2-
diamines through ring opening at the less hindered aziridine
carbon atom.21 In another report, the ring opening of 2-substi-
tuted 1,1-dibenzylaziridinium salts by methylamine has been
evaluated, showing a preferential (yet never exclusive) attack
at the less hindered carbon atom for 2-methyl-, 2-benzyl-, and
2-isopropylaziridinium salts, and opening at the benzylic position
1
2
3
4
5
6
H
H
Me
Me
Me
Cl
H
Br
H
Br
Me
Me
10
10
3
5
12
4
8a
8b
8c
8d
8e
8f
66
74
80
87
77
74
a Yields after column chromatography
SCHEME 3
aziridine carbon atom, affording â-bromo amines 8 as the sole
reaction products in high purity. Detailed spectral analysis
confirmed the structural identity of these novel N-(2-bromo-3-
cyanopropyl)amines 8, excluding the formation of the corre-
sponding regioisomers. The regiospecific ring opening of
2-(cyanomethyl)aziridinium salts 7 by bromide in acetonitrile
at the more hindered position was further validated by perform-
ing ab initio calculations (see section Ab Initio Calculation
Results).
It should be remarked that upon heating of aziridines 6 in
the presence of an arylmethyl bromide under reflux for 5 h, the
reaction mixtures were contaminated with reasonable amounts
of the corresponding 4-amino-2-butenenenitriles 9 (15-50%)
because of the acidity of the R-proton with respect to the cyano
group, whereas no dehydrobromination occurred at room
temperate (although longer reaction times of 3-12 days were
required to drive the reaction to completion). The selective
preparation of 4-amino-2-butenenenitriles 9 as mixtures of E/Z-
isomers could be accomplished by a simple base treatment of
4-amino-3-bromobutanenitriles 8 with 1.1 equiv of Et3N in
refluxing CH2Cl2 for 3 h (Scheme 3). Other base/solvent
combinations, such as KOtBu/Et2O and NaH/THF, appeared
to be less successful, resulting in complex reaction mixtures.
Two main reasons can be adduced for the relevance of
γ-aminobutenenitrile derivatives such as 9, the first being the
general interest in structural (in casu unsaturated) analogues of
(17) (a) Canovas, D. M.; Iborra Pastor, J. L. PCT Int. Appl. WO
2004104207 A1, 2004; Chem. Abstr. 2004, 142, 22341. (b) Lee, B. K.;
Kim, M. S.; Hahm, Kim, D. S.; Lee, W. K.; Ha, H.-J. Tetrahedron 2006,
62, 8393. (c) Kim, Y.; Ha, H-J.; Han, K.; Ko, S. W.; Yun, H.; Yoon, H. J.;
Kim, M. S.; Lee, W. K. Tetrahedron Lett. 2005, 46, 4407.
(18) Kotti, S. R. S. S.; Timmons, C.; Li, G. Chem. Biol. Drug Des. 2006,
67, 101.
(19) Tranchant, M.-J.; Dalla, V. Tetrahedron 2006, 62, 10255.
(20) Couturier, C.; Blanchet, J.; Schlama, T.; Zhu, J. Org. Lett. 2006, 8,
2183.
(21) Katagiri, T.; Takahashi, M.; Fujwara, Y.; Ihara, H.; Uneyama, K.
J. Org. Chem. 1999, 64, 7323.
J. Org. Chem, Vol. 72, No. 13, 2007 4735

D’hooghe et al.
SCHEME 4
SCHEME 5
in the case of 2-phenyl substitution.22 The present work
comprises the first report of the regiospecific ring opening of
2-substituted aziridinium salts 7 by a secondary amine (pyrro-
lidine) at the more hindered aziridine carbon atom in an elegant
and straightforward approach. This new and valuable synthetic
methodology was further supported on the basis of ab initio
calculations (see section Ab Initio Calculation Results).
It is important to note that the Michael addition of pyrrolidine
onto 4-amino-2-butenenenitriles 9 does not offer a suitable
alternative for the synthesis of 3,4-diaminobutanenitriles 10,
since treatment of butenenitriles 9 with 2-16 equiv of pyrro-
lidine resulted in all cases in mixtures of 3,4-diaminobutane-
nitriles 10 (22-48%), N,N-bis(arylmethyl)amines 11 (15-73%),
and sometimes unreacted starting material 9 (0-43%) (Scheme
5). On the contrary, the methodology involving the preparation
of 4-amino-3-bromobutanenenitriles 8 followed by treatment
with pyrrolidine offers an elegant route toward 3,4-diaminobu-
tanenitriles 10 in good yields. The formation of N,N-bis-
(arylmethyl)amines 11 can be rationalized considering a base-
assisted migration of the double bond in butenenitriles 9 toward
intermediate iminium salts, which are then hydrolyzed affording
bis(arylmethyl)amines 11.
known to give reliable estimates of these parameters.24 More-
over, a variety of other electronic levels were tested and the
selected methods have been found most suitable. Caution is
needed in interpreting the quantitative values for the barriers
resulting from DFT calculations, as the absolute values for the
energies might vary largely depending on the specific functional
that is used.25 Our main objective in this study is the prediction
of qualitative trends that might contribute to a better understand-
ing of the experimental results, and therefore the proposed
functional is well suited.
Also the effect of solvent is taken into account, explicitly by
surrounding the reactive center of the complexes with two or
four acetonitrile molecules and implicitly by embedding all
structures in a polarizable medium characterized by a fixed
dielectrical constant which is typical for acetonitrile (i.e., ꢀ )
36.64). The solvation energy originates from two contribu-
tions: the explicit coordination with acetonitrile molecules and
the electrostatic effect of the solvent dielectricum. The first effect
might be expected to involve the bare bromide ion, as it is an
important part of our model system, and it is expected to
coordinate explicitly with surrounding solvent molecules. The
solvation due to the bulk electrostatic effects is obtained by
enclosing the reactive intermediate in a cavity within a continu-
ous dielectricum medium. The CPCM option in Gaussion was
applied. A recent study by Kelly, Cramer, and Truhlar on the
accurate determination of solvation free energies might be
interesting for a more detailed reading on this subject.26
Ab Initio Calculation Results. Effect of Solvation on the
Intermediate Aziridinium Salts 7. From a molecular modeling
viewpoint, the system under study is challenging, as it concerns
Computational Details. All ab initio calculations were
performed with the Gaussian 03 software package.23 Density
Functional Theory (DFT) methods have been shown to be more
efficient than wave function based procedures such as highly
correlated post-Hatree-Fock methods because of their excellent
cost to performance ratio. The MPW1B95/6-31+g(d) level of
theory was used for both geometries and energies, which is
(22) O’ Brien, P.; Towers, T. D. J. Org. Chem. 2002, 67, 304.
(23) Gaussian 03, Revision D, Frisch, M. J.; Trucks, G. W.; Schlegel,
H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A.,
Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.;
Tomasi, J.; Barone, V. et al. Gaussian, Inc., Pittsburgh, PA, 2003.
(24) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2004, 108, 6908.
(25) Kormos, B. L.; Cramer, C. J. J. Phys. Org. Chem. 2002, 15, 712.
(26) Kelly, C. P.; Cramer, C. J.; Truhlar, D. G. J. Chem. Theory Comput.
2005, 1, 1133.
4736 J. Org. Chem., Vol. 72, No. 13, 2007

Synthesis of 3,4-Diaminobutanenitriles
The bromide ion prefers a position located above the aziridinium
moiety at a distance of about 3.6 and 3.7 Å from the unhindered
and hindered carbon atom, respectively. The C-N bond lengths
of the intermediate are different, the one involving the sub-
stituted carbon atom (C2) being longer. The optimized structure
with four acetonitrile molecules is shown in Figure 2. In the
aziridinium intermediate 7, additionally surrounded by a pyr-
rolidine molecule (Scheme 4), the bromide has a similar posi-
tion with respect to the aziridinium moiety, but it makes
an additional contact with the hydrogen of pyrrolidine
(Br-Hpyrrolidine ) 2.573 Å).
Some important geometrical parameters of 7 in terms of a
varying coordination number are given in Table 2. As for the
CSE values discussed previously, also the distances of the
bromide with respect to the aziridinium moiety are only
converged when taking four explicit acetonitrile molecules into
account. However, the C-N distances seem to be less sensitive
to the coordination number.
FIGURE 1. The coordination solvation energy (CSE) in terms of
a varying number of acetonitrile solvent molecules for the inter-
mediate of the ring opening of the aziridinium salt 7 by bromide (dotted
line) and the ring opening of aziridinium salt 7 by pyrrolidine (full
line).
Atomic Charges of the Intermediates 7. The nature of the
intermediates can be further understood by studying the atomic
charges on the optimized geometries. The Mulliken atomic
charges with the hydrogens summed into the heavy atoms at
the MPW1B96/6-31+g(d) level of theory are listed in Table 3
for varying coordination numbers. In the gas-phase calculation
without explicit account of acetonitrile molecules, both the
hindered and unhindered carbon atom of the aziridine moiety
carries a partial positive charge. The latter charges are sensitive
to the number of coordinating solvent molecules. With two, four,
or five acetonitrile molecules the hindered carbon atom carries
a partial positive charge, whereas the unhindered position is
partially negatively charged. These observations give a good
indication that the hindered position is preferred for the further
reaction with the bromide anion. The partial charges give further
evidence that the correct reaction preference will be sensitive
to the account of the first solvation shell of the bromide anion.
For the intermediate 7, the largest partial positive charge is also
located at the hindered carbon atom, indicating reaction prefer-
ence at this site.
an ion pair which is further solvated in acetonitrile. In our
previous communication on the ring opening of 1-alkyl-2-
(aryloxymethyl)aziridinium salts by bromide,16c preliminary
indications were found that the explicit account of solvent mole-
cules together with further embedding in a continuous dielec-
tricum was crucial for the prediction of the correct reactivity
pattern. The importance of explicit solute-solvent interactions
and the bulk electrostatic effects is properly investigated on the
intermediate aziridinium salts. A first point of interest concerns
the degree of coordination of the ion pair under consideration.
The coordination number is estimated by coordinating the
intermediate 2-(cyanomethyl)aziridinium salts 7 by an increasing
number of explicit acetonitrile molecules and calculating the
coordination solvation energy (CSE). These values are shown
in Figure 1.
Transition State Structures. The transition state structures
for ring opening at the more or less hindered carbon atom
starting from intermediate 7 (Scheme 2) are shown in Figure 3.
A specific geometrical feature of the transition states in
aziridinium salt 7 surrounded by an additional pyrrolidine
molecule (Scheme 4) is the length of the breaking carbon-
nitrogen bond, which amounts to 1.77 and 1.82 Å for attack at
C3 and C2, respectively. In the latter case, the bond is already
far elongated in the transition state, and this is an indication
for a “late transition state” which is in most cases an indication
of a lower reaction barrier compared to “early transition states”.27
For the transition states following from the intermediate 7
(Scheme 2), the differences in geometrical characteristics be-
tween the two possible ring openings is less pronounced, and
this is a preliminary indication that the correct prediction of
the reaction preference is subjected to a lot of factors, such as
a proper account of the solvent environment. Moreover, the
geometrical characteristics of forming and breaking bonds in
the transition state vary largely on the number of explicit solvent
molecules that is accounted. The underlying reason for this dif-
ferent behavior of the two intermediates is probably the coor-
dination of the bromide anion with pyrrolidine in intermediates
7 as shown in Scheme 4. Therefore, an already large part of
The coordination of the intermediate aziridinium salt 7
(Scheme 2) is highly exothermic, giving 95, 176, and 192
kJ/mol for two, four, and five acetonitrile molecules. These
results are not too surprising, as it concerns a bromide anion
that is coordinated with the aziridinium ion but that is further
largely stabilized by surrounding solvent molecules. From these
numbers it can already be expected that the correct reaction
preference for attack at the less or more hindered carbon atom
will be sensitive to the number of acetonitrile molecules that is
explicitly taken into account. For the intermediate aziridinium
salt 7 which is additionally surrounded by a pyrrolidine molecule
(Scheme 4), the degree of coordination was also determined.
The coordination of the first two acetonitrile molecules is still
very exothermic, but the CSE value is smaller in absolute value,
as a large amount of the stabilization of the bromide anion is
already captured by the surrounding pyrrolidine ring (i.e., -444
kJ/mol). The Gibbs free energy change is expected to converge
more rapidly than the CSE because of the entropic effect of the
solvation.
Geometries of the Intermediates 7. Before studying the
transition-state structures, it is interesting to study some geo-
metrical properties of the intermediate aziridinium salts. The
approach of the aziridinium ion 7 with bromide (Scheme 2)
leads to a stable intermediate on the potential energy surface.
(27) Fischer, H.; Radom, L. Angew. Chem., Int. Ed. 2001, 40, 1340.
J. Org. Chem, Vol. 72, No. 13, 2007 4737

D’hooghe et al.
FIGURE 2. Geometries of intermediates 7 which are further coordinated by four acetonitrile molecules.
TABLE 2. Most Important Geometrical Characteristics of the
Intermediates 7 at the MPW1B96/6-31+g(d) Level of Theory (Å).
C2 and C3 Are the Hindered and Unhindered Carbon Atoms of the
Aziridinium Salts
TABLE 3. Mulliken Atomic Charges at the MPW1B96/6-31+g(d)
Level of Theory and with Hydrogens Summed into Heavy Atoms for
the Intermediates 7
gas
phase
two acetonitrile
molecules
four acetonitrile
molecules
gas
two acetonitrile
molecules
four acetonitrile
molecules
phase
Intermediate 7 (Scheme 2)
Intermediate 7 (Scheme 2)
N
-0.61
-0.60
-0.22
+0.96
-0.70
-0.54
-0.44
+0.69
-0.86
C3-N
1.483
1.485
1.493
3.447
3.394
3.728
3.619
1.486
1.496
3.641
3.703
3.729
3.629
3.726
3.676
C3
C2
Br
+0.30
+1.11
-0.55
C2-N
1.498
3.407
3.313
C3-Br
C2-Br
Intermediate 7 + pyrrolidine (Scheme 4)
Br-Caceto1
Br-Caceto2
Br-Cateto3
Br-Cacteo4
N
C3
C2
Br
-0.53
+0.76
+0.53
-0.61
+0.39
-0.45
+0.19
+0.86
-0.71
+0.44
-0.30
+0.29
+0.46
-1.21
Intermediate 7 + Pyrrolidine (Scheme 4)
Hpyrrolidine
C3-N
1.483
1.500
3.452
3.374
2.525
1.486
1.501
3.557
3.449
2.584
3.719
3.687
1.486
1.498
3.645
3.601
2.573
3.764
3.752
3.556
3.511
C2-N
C3-Br
C2-Br
BrvHpyrrolidine
Br-Caceto1
Br-Caceto2
Br-Cateto3
Br-Cacteo4
are highly dependent on the explicit account of acetonitrile
molecules. Taking into account two explicit solvent molecules
which are further embedded in a continuum dielectricum gives
a preference for attack at the hindered carbon atom with an
energy barrier advantage of 26 kJ/mol. The results for the ring
opening of the intermediate 7 by bromide (Scheme 2) are less
transparent. The barriers obtained without inclusion of bulk
solvent effects do not succeed in predicting the correct reaction
preference, even with inclusion of two or four acetonitrile
molecules. Similar observations have been reported earlier.11a
Once bulk solvent effects are accounted for, the correct reaction
preference is predicted, provided that at least two acetonitrile
molecules are explicitly taken into account. From previous
calculations it follows that the intermediate 7 with a naked
bromide anion attacking the aziridinium moiety is challenging
in order to get correct quantitative values for the barriers, and
further investigation would be appropriate to design a reliable
cluster model that enables prediction of correct quantitative
values for the barriers. Moreover, it seems that a variety of stable
minima exists on the potential energy surface that differ in
orientation of the surrounding solvent molecules. The results
given here are obtained by taking the energetically most stable
configuration.
solvation is captured by the presence of pyrrolidine, which coor-
dinates strongly with Br- (cfr. CSE indicated in Figure 1). In
case of the intermediate 7 without the presence of pyrrolidine
(Scheme 2), the bromide anion behaves as an unscreened ion
which must fully be solvated by solvent acetonitrile molecules,
and proper embedding will be essential.
Reaction Barriers. The reaction barriers for ring opening at
the hindered and unhindered carbon atom of the aziridine moiety
are given in Table 4. Both the values with and without inclusion
of bulk solvent effects are given. The reaction preference for
the nucleophilic ring opening of the intermediate 7 by pyrro-
lidine (Scheme 4) is always correctly predicted at all levels of
our cluster model. Even in the gas phase without explicit account
of solvent molecules and without accounting for bulk solvation
effects, a slight preference for ring opening at the more hindered
carbon aziridine atom is found. However, the quantitative values
4738 J. Org. Chem., Vol. 72, No. 13, 2007

Synthesis of 3,4-Diaminobutanenitriles
FIGURE 3. The transition state structures for ring opening at the unhindered (u) or hindered (h) carbon atom starting from intermediate 7 with
or without an additional pyrrolidine.
TABLE 4. Reaction Barriers for Ring Opening at the Hindered and Unhindered Carbon Atom of the Aziridine Moiety MPW1B96/6-31+g(d)
Level of Theory Starting from the Intermediates 7 (Energies Are in kJ/mol)a
without inclusio n of bulk solvent effects
with inclusio n of bulk solvent effects
∆Eu
∆Eh
∆(∆Eu - ∆Eh)
7 (Scheme 2)
∆Gu
∆Gh
∆(∆Gu - ∆Gh)
gas phase
two acetonitrile molecules
four acetonitrile molecules
51.24
37.66
51.70
57.07
70.79
56.48
-5.83
-33.13
-4.78
41.42
49.92
52.93
45.38
44.97
47.23
-3.96
4.95
5.70
7 + pyrrolidine (Scheme 4)
gas phase
two acetonitrile molecules
40.93
53.95
38.68
17.34
2.25
33.70
54.71
27.76
28.61
5.94
26.1
36.61
a ∆Euand ∆Eh are the reaction barriers without inclusion of zero point vibrational energy for ring opening at the less hindered and more hindered aziridine
carbon atom, and ∆(∆Eu - ∆Eh) is the difference between the aforementioned barriers. ∆Gu and ∆Gh are the free energy reaction ring barriers for ring
opening at the less hindered and more hindered aziridine carbon atom.
Conclusions
pyrrolidine at the more hindered aziridine carbon atom, which
was validated by means of high level ab initio calculations. For
the ring opening by pyrrolidine, the correct reaction preference
was always predicted irrespective of the molecular environment
that was taken into account. For the reaction with bromide, the
result is prone to a lot of effects such as the number of explicit
acetonitrile molecules by which the intermediate is surrounded.
In conclusion, an efficient and straightforward synthesis of
4-(N,N-bis(arylmethyl)amino)-3-(pyrrolidin-1-yl)butaneni-
triles and 4-(N,N-bis(arylmethyl)amino)-2-butenenitriles is pre-
sented starting from 1-arylmethyl-2-(cyanomethyl)aziridines.
The key steps comprise the novel regiospecific ring opening of
intermediate 2-(cyanomethyl)aziridinium salts by bromide and
J. Org. Chem, Vol. 72, No. 13, 2007 4739

D’hooghe et al.
obtain an analytically pure sample. (E/Z)-4-(N,N-Dibenzylamino)-
2-butenenitrile 9a. Yellow oil. Yield 77%. Rf ) 0.34 (hexane/
CH2Cl2 9/1). 1H NMR (270 MHz, CDCl3): δ 3.19 (2H, d×d, JE )
5.3, 2.0 Hz); 3.40 (2H, d×d, JZ ) 6.0, 1.7 Hz); 3.59 (8H, s); 5.34
(1H, d×t, JZ ) 11.1, 1.7 Hz); 5.66 (1H, d×t, JE ) 16.5, 2.0 Hz);
6.55 (1H, d×t, JZ ) 11.1, 6.0 Hz); 6.73 (1H, d×t, JE ) 16.5, 5.3
Hz); 7.21-7.40 (20H, m). 13C NMR (68 MHz, CDCl3): δ 54.1,
54.5, 58.4, 58.7, 100.7, 100.8, 115.5, 117.3, 127.3, 127.3, 128.4,
Beside the energetic results, several other parameters also give
valuable information about the reaction preference, such as
geometries, charges, etc., which all point toward a selective ring
opening at the hindered carbon atom of the aziridine moiety.
These experimental results comprise a useful contribution to
the chemistry of 2-substituted aziridinium salts and offer
perspectives for prospective research in this area.
128.4, 128.6, 128.8, 138.4, 138.6, 153.2, 153.2. IR (NaCl): νCN
)
2223. MS (70 eV): m/z (%): 263 (M+ + 1, 100); 238 (20); 210
Experimental Section
(14); 91 (11).
Synthesis of 4-(N,N-Di(arylmethyl)amino)-3-bromobutane-
nitriles 8. As a representative example, the synthesis of 4-(N,N-
dibenzylamino)-3-bromobutanenitrile 8a is described. To a stirred
solution of 1-benzyl-2-(cyanomethyl)aziridine 6a (1.72 g, 10 mmol)
in acetonitrile (15 mL) was added benzyl bromide (1.71 g, 1.0
equiv), and the resulting mixture was further stirred for 10 days at
room temperature. Evaporation of the solvent afforded 4-(N,N-
dibenzylamino)-3-bromobutanenitrile 8a, which was purified by
means of column chromatography (hexane/ethyl acetate 9/1) on
silica gel in order to obtain an analytically pure sample. 4-(N,N-
Dibenzylamino)-3-bromobutanenitrile 8a. Yellow oil. Yield 66%.
Rf ) 0.26 (hexane/ethyl acetate 9/1). 1H NMR (300 MHz,
CDCl3): δ 2.67 and 3.01 (2H, 2×d×d, J ) 17.3, 7.4, 3.9 Hz);
2.89 and 2.96 (2H, 2×d×d, J ) 13.6, 9.8, 5.2 Hz); 3.57 and 3.71
(4H, 2×d, J ) 13.5 Hz); 3.71-3.81 (1H, m); 7.27-7.40 (10H,
m). 13C NMR (75 MHz, CDCl3): δ 25.5, 43.6, 59.9, 60.5, 117.1,
127.8, 128.8, 129.2, 138.4. IR (NaCl): νCN ) 2252. MS (70 eV):
m/z (%): 343/5 (M+ + 1, 30); 263 (100).
Synthesis of 4-(N,N-Di(arylmethyl)amino)-2-butenenitriles 9.
As a representative example, the synthesis of 4-(N,N-dibenzyl-
amino)-2-butenenitrile 9a is described. To a stirred solution of
4-(N,N-dibenzylamino)-3-bromobutanenitrile 8a (1.72 g, 5 mmol)
in dichloromethane (10 mL) was added triethylamine (0.56 g, 1.1
equiv), and the resulting solution was heated under reflux for 3 h.
The reaction mixture was poured into water (10 mL) and extracted
with dichloromethane (3 × 10 mL). Drying (MgSO4), filtration of
the drying agent, and removal of the solvent in vacuo afforded
4-(N,N-dibenzylamino)-2-butenenitrile 9a as a mixture of E/Z
isomers (54/46), which was purified by means of column chroma-
tography (hexane/dichloromethane 9/1) on silica gel in order to
Synthesis of 4-(N,N-Di(arylmethyl)amino)-3-(pyrrolidin-1-yl)-
butanenitriles 10. As a representative example, the synthesis of
4-(N,N-dibenzylamino)-3-(pyrrolidin-1-yl)butanenitrile 10a is de-
scribed. To a stirred solution of 4-(N,N-dibenzylamino)-3-bromobu-
tanenitrile 8a (1.72 g, 5 mmol) in acetonitrile (10 mL) was added
pyrrolidine (0.71 g, 2.0 equiv), and the resulting solution was heated
under reflux for 3 h. The reaction mixture was poured into water
(10 mL) and extracted with diethyl ether (3 × 10 mL). Drying
(MgSO4), filtration of the drying agent, and removal of the sol-
vent in vacuo afforded 4-(N,N-dibenzylamino)-3-(pyrrolidin-1-yl)-
butanenitrile 10a, which was purified by means of column
chromatography (hexane/ethyl acetate 2/1) on silica gel in order to
obtain an analytically pure sample. 4-(N,N-Dibenzylamino)-3-
(pyrrolidin-1-yl)butanenitrile 10a. Yellow oil. Yield 80%. Rf ) 0.23
1
(hexane/EtOAc 2/1). H NMR (300 MHz, CDCl3): δ 1.68-1.78
(4H, m); 2.43-2.69 (9H, m); 3.42 and 3.75 (4H, 2×d, J ) 13.3
Hz); 7.30-7.34 (10H, m). 13C NMR (75 MHz, CDCl3): δ 19.6,
23.4, 51.0, 56.0, 58.8, 59.6, 119.3, 127.4, 128.5, 129.2, 138.8. IR
(NaCl): νCN ) 2246. MS (70 eV): m/z (%): 334 (M+ + 1, 100).
Acknowledgment. The authors are indebted to the “Fund
for Scientific Research - Flanders (Belgium)” (F.W.O.-Vlaan-
deren) and to Ghent University (GOA) for financial support.
Supporting Information Available: Spectroscopic data of
compounds 8-f, 9b-d,f, and 10b-f. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO0704210
4740 J. Org. Chem., Vol. 72, No. 13, 2007