A comparative study of the stereoselective addition of trimethylsilyl cyanide and diethylaluminum cyanide to chiral cyclic nitrones

Article abstract of DOI:10.1016/S0957-4166(02)00832-7

The mild cyanating agent trimethylsilyl cyanide adds with total stereoselectivity to α-alkoxy cyclic nitrones to afford the corresponding trans-hydroxyaminonitriles. The addition of Lewis acids to precomplexing the nitrones does not affect the stereoselectivity of these additions significantly. In all of the cases examined, excellent yields of diastereomerically homogeneous products were obtained. On the other hand, the use of diethylaluminum cyanide as cyanating agent leads to low diastereoselectivities. Both NMR studies and theoretical calculations show that whereas the addition of trimethylsilyl cyanide takes place through a concerted mechanism, in the addition of diethylaluminum cyanide, a complex is formed prior to the intramolecular delivery of the cyanide ion.

Full text of DOI:10.1016/S0957-4166(02)00832-7

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 14 (2003) 367–379
A comparative study of the stereoselective addition of
trimethylsilyl cyanide and diethylaluminum cyanide to chiral
cyclic nitrones
Pedro Merino,a,* Tomas Tejero,a Julia Revuelta,a Pilar Romero,a Stefano Cicchi,b Vanni Mannucci,b
Alberto Brandib and Andrea Gotib,*
aDepartamento de Qu´ımica Orga´nica, ICMA, Facultad de Ciencias, Universidad de Zaragoza-CSIC, E-50009 Zaragoza,
Aragon, Spain
bDipartimento di Chimica Organica ‘Ugo Schiff’, Universita` di Firenze, via della Lastruccia 13, I-50019 Sesto Fiorentino,
Firenze, Italy
Received 6 November 2002; accepted 5 December 2002
Abstract—The mild cyanating agent trimethylsilyl cyanide adds with total stereoselectivity to a-alkoxy cyclic nitrones to afford the
corresponding trans-hydroxyaminonitriles. The addition of Lewis acids to precomplexing the nitrones does not affect the
stereoselectivity of these additions significantly. In all of the cases examined, excellent yields of diastereomerically homogeneous
products were obtained. On the other hand, the use of diethylaluminum cyanide as cyanating agent leads to low diastereoselectiv-
ities. Both NMR studies and theoretical calculations show that whereas the addition of trimethylsilyl cyanide takes place through
a concerted mechanism, in the addition of diethylaluminum cyanide, a complex is formed prior to the intramolecular delivery of
the cyanide ion. © 2003 Elsevier Science Ltd. All rights reserved.
1. Introduction
form, the most stable configuration for aldonitrones
(with the exception of those having an electron-with-
drawing group at the nitrone carbon atom), which is
also assumed to be the reactive species. Consequently, it
is not possible to study the reactivity of the correspond-
ing E-isomers (Scheme 2).
In recent years, considerable progress has been made in
the stereocontrolled 1,2-addition of carbon nucleophiles
to chiral non-racemic nitrones under the influence of a
Lewis acid.1
In particular, nucleophilic additions to a-alkoxy
aldonitrones 1 take place with syn selectivity2 when the
reaction is carried out in the absence of any additive or
in the presence of Lewis acids such as ZnBr2 or MgBr2
(Scheme 1).
In order to carry out a comparative study on hydrocy-
anation of nitrones, we consider nitrones 24 as close
substrates to 1, but possessing two important structural
differences: (i) nitrones 2 are cyclic and (ii) they are
blocked into an E-configuration.
Although the hydrocyanation of cyclic imines5 and
iminium salts6 as well as other endocyclic CꢀN bonds7
has been reported before, only a handful of reports of
hydrocyanation of cyclic nitrones have appeared in the
literature.8 Moreover, only a couple of examples have
been reported for cyanide addition to chiral hydroxyl-
ated nitrones while this work was in progress, but a
detailed and unified study is lacking.8a,9
On the other hand, the use of Et2AlCl (DEAC) as a
precomplexing agent gives rise to anti adducts preferen-
tially, thus rendering the reaction stereodivergent.
A notable exception to this behavior is the addition of
cyanide: irrespective of the reagent used or the presence
of additives, the reaction invariably gives rise to syn
adducts.3 Nitrones 1 can only be obtained in the Z-
Herein, we disclose our results from cyanide additions
to the enantiopure nitrones 2a–d and racemic 2e (Fig.
* Corresponding authors. Tel.: +34-976-762075; fax: +34-976-762075
(P.M.); tel.: +39-055-4573505; fax: +39-055-4573531 (A.G.); e-mail:
pmerino@posta.unizar.es; andrea.goti@unifi.it
1). These nitrones can be easily prepared from L-malic
acid 2a,b,10 -tartaric acid 2d11 and arabinose 2e4a,12 as
L
0957-4166/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(02)00832-7

368
P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
Synthetically, the interest in this reaction lies in the
cyanide adducts being precursors of proline derivatives
and of chiral 1,2-diamines.
2. Results and discussion
2.1. Trimethylsilylcyanide additions
The addition of trimethylsilyl cyanide to nitrone 2a was
the first reaction to be examined. Thus, the nitrone was
treated with 1.0 equiv. of the reagent at ambient tem-
perature in dichloromethane as a solvent. The addition
product 3 was isolated in quantitative yield as a single
trans isomer13 after evaporation of the solvent no
Scheme 1.
1
purification being necessary (Table 1, entry 1). The H
NMR spectrum of 3 showed two sets of signals when
recorded at 22°C. When the solution was heated to
55°C they coalesced to a single, well-resolved set of
resonances, whereas cooling the sample restored the
spectrum to its original condition, thus suggesting the
existence of a dynamic equilibrium which could be
attributed to nitrogen inversion according to previous
theoretical and experimental investigations.14 However,
the presence of bulky groups linked to the hydroxyl-
amine oxygen means that an equilibrium between
rotamers around the NꢁO bond15 cannot be dis-
carded.16 Further treatment of compound 3 with 5%
methanolic citric acid gave rise quantitatively to the
free hydroxylamine 4 (Scheme 3). Since compound 4
was obtained as a single isomer, we can reasonably
assume that the observed ratio at the free hydroxyl-
amine stage reflects the diastereoselectivity of the addi-
tion step.
Scheme 2.
described before. Nitrone 2c has been appositely syn-
thesized for this study in analogy to 2a and 2b (see
Section 4). We undertook the investigation of the addi-
tion of two representative cyanating agents, i.e.
trimethylsilyl cyanide and diethylaluminum cyanide. A
parallel between these reagents has been established:
whereas TMSCN only acts as the cyanide-transfer
reagent and additives can be used, Et2AlCN can play
both the roles of additive and cyanide-transfer reagent
at the same time. Both NMR and computational stud-
ies have also been carried out to aid interpretation of
the experimental results.
Scheme 3. Reagents: (i) 5% citric acid, MeOH.
The trans relative stereochemistry has been established
for CN with respect to OtBu by 1H NMR analysis.
Indeed, in a NOESY spectrum, cross peaks H-2/H-5a,
Figure 1. a-Alkoxy nitrones 2.
Table 1. Trimethylsilyl cyanide addition to nitrone 2aa
Entry
Equiv.
Additive (equiv.)
Time
Yield (%)b
trans:cisc
1
2
3
4
5
6
1.0
3.0
1.0
1.0
1.0
1.0
None
None
TMSOTf (1.0)
Et2AlCl (1.0)
TMSOTf (0.2)
Et2AlCl (0.2)
16 h
2 h
5 min
5 min
5 min
5 min
100
100
100
100
100
100
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
a Further treatment with 5% methanolic citric acid of the crude mixtures afforded free hydroxylamines. No purification was necessary.
b All reactions showed quantitative yield.
c Only one isomer could be detected by NMR noise-to-signal ratio.

P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
369
H-3/H-4b, H-4a/H-5a and H-4b/H-5b were observed;
hence R configuration at the newly created stereogenic
center was assigned.
performed with cyanides, a complete stereoselectivity
was also observed,8a as it was by the authors of a recent
patent.9
Using an excess of trimethylsilyl cyanide the reaction
time shortened considerably (Table 1, entry 2). Since
trimethylsilyl triflate (TMSOTf) and diethylaluminum
chloride (DEAC) are considered to be effective activat-
ing agents for nitrones towards nucleophilic additions,17
we also examined the addition of trimethylsilyl cyanide
to 2a in the presence of such additives.18 Using 1.0
equiv. of either TMSOTf or DEAC considerably accel-
erates the reaction such that a 5 min reaction time was
only required for total conversion (Table 1, entries 3
and 4). The action of those activating agents was shown
to be catalytic since the same results were obtained
when the reaction was carried out in the presence of 20
mol% of additive (Table 1, entries 5 and 6). The
proposed catalytic cycle for the reaction in the presence
of TMSOTf is illustrated in Scheme 4; a similar cycle
may be proposed with DEAC.
Next, we took the best systems, catalytic TMSOTf or
DEAC, and surveyed nitrones 2b–e in the cyanosilyla-
tion reaction using these Lewis acids as additives, in
comparison with the reaction in absence of any additive
(Scheme 5). The results are shown in Table 2 and
establish the generality of the complete stereoselective
cyanation previously proved with nitrone 2a. The crude
mixture was treated in all instances with 5% methanolic
citric acid in order to obtain the free hydroxylamine
5–8.
Scheme 5.
Scheme 4.
It is of interest to compare the above results with that
reported for a structurally similar TBDMS protected
hydroxypyrroline N-oxide.8a In that case, the TMSOTf
catalyzed addition of TMSCN was reported to afford a
mixture of trans and cis N-hydroxypyrrolidines with a
poor 1.5:1 diastereoselectivity. It is unclear if this out-
come, which contrasts with the complete stereoselectiv-
ity observed by us, is a result of the different reaction
procedure employed. However, when the reaction was
It is noteworthy that the reactions were extremely
clean, the products were obtained essentially pure with-
out the need to resort to chromatographic purification.
In all cases, only a single isomer could be detected by
HPLC and NMR. The relative stereochemistries at
C-2/C-3 of the pyrrolidine ring could be inferred from
NOE experiments as before. In the case of hydroxy-
pyrrolidine (rac)-8 a single X-ray crystal structure
Table 2. Trimethylsilyl cyanide addition to nitrones 2b–ea,b
Entry
Nitrone
Additive (equiv.)
Time
Yield (%)c
trans:cisd
1
3
4
5
6
8
9
11
12
14
15
2b
None
16 h
100
100
100
100
100
100
100
100
100
100
100
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
\20:1
TMSOTf (1.0)
TMSOTf (0.2)
Et2AlCl (1.0)
Et2AlCl (0.2)
TMSOTf (0.2)
Et2AlCl (0.2)
TMSOTf (0.2)
Et2AlCl (0.2)
TMSOTf (0.2)
Et2AlCl (0.2)
5 min
5 min
5 min
5 min
5 min
5 min
5 min
5 min
5 min
5 min
2c
2d
2e
a All reactions were carried out in CH2Cl2 at 25°C with 1.0 equiv. of TMSCN.
b Further treatment with 5% methanolic citric acid of the crude mixtures afforded free hydroxylamines. No purification was necessary.
c All reactions showed quantitative yield.
d Only one isomer could be detected by NMR noise-to-signal ratio.

370
P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
analysis was accomplished, which confirmed the struc-
TMSOTf has no effect, the addition of a second equiv-
alent of DEAC effected additional deshielding of Ha
(7.90 ppm); two species were observed upon the addi-
tion of 1.5 equiv. DEAC. Thus, the formation of
complex C with a 2:1 stoichiometry DEAC:nitrone can
be postulated.
tural assignment (Fig. 2).19
The reaction of TMSOTf-activated nitrone with 1.0
equiv. of TMSCN proceeded with the immediate disap-
pearance of the complex. Upon injection of the cyanat-
ing agent and immediate recording of the spectrum
only silylated hydroxylamine 3 could be detected. A
similar result was observed when the nitrone was acti-
vated with DEAC. It proved impossible (even on cool-
ing the mixture to −60°C) to record the progress of the
reaction due to its extremely high rate.
At this stage, two major findings can be emphasized: (i)
only starting materials and final products can be
detected in the reaction of 2a with TMSCN either in
the absence of any additive or in the presence of Lewis
acids and (ii) the resonances of the azomethine proton
undergo a large downfield shift upon complexation of
the nitrone functionality. These observations will be of
importance for supporting the rationale for the cyanosi-
lylation that is presented in the mechanistic section.
2.2. Diethylaluminum cyanide additions
Figure 2. Perspective view (ORTEP) of (rac)-8. Non-hydro-
gen atoms are drawn as 50% thermal ellipsoids while hydro-
gens are drawn at an arbitrary size. Only the atoms refined
with anisotropic thermal parameters are drawn with the prin-
cipal axes indicated; the isotropic atoms are represented as
simple circles.
The results of the addition of Et2AlCN to nitrone 2a
are summarized in Table 3. We screened a variety of
conditions in the hope of obtaining information about
the reaction.
The stereoselective cyanosilylation of 2a was followed
using NMR spectroscopy. Upon injecting a CD2Cl2
solution of TMSCN (1.0 equiv.) into a CD2Cl2 solution
of nitrone 2a at 20°C, the 1H NMR spectrum was
immediately recorded, at which time only the signals
corresponding to nitrone 2a (without variation of dis-
placement) and TMSCN were observed. Upon ageing
the reaction mixture for suitable periods the reagents
afforded the corresponding product of the reaction.
In order to obtain information concerning the struc-
1
tures of the nitrone-additive complexes, H NMR anal-
ysis of the mixture of equimolar amounts of nitrone 2a
and the Lewis acid was carried out (Fig. 3). The
mixtures were prepared at −20°C by injecting CD2Cl2
solutions of the corresponding Lewis acid into an
1
NMR tube containing a CD2Cl2 solution of 2a. The H
NMR spectra were immediately recorded and they
showed, in both cases (TMSOTf and DEAC) consider-
able deshielding of several protons.20 Therefore the
structures of complexes A and B were formulated as
indicated in Fig. 3.
It is particularly revealing that the signal of the azome-
thine proton Ha of uncomplexed nitrone 2a shifts
downfield from 6.71 to 8.20 or 7.80 ppm upon complex-
ation with TMSOTf or DEAC, respectively. Moreover,
whereas the addition of a second equivalent of
Figure 3. Partial NMR spectra, showing the signal corre-
sponding to Ha, of 2a upon the addition of Lewis acids.

P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
Table 3. Diethylaluminum cyanide additions to nitrone 2aa
371
Entry
Additive (equiv.)
Solvent
Time
Yield (%)b
trans:cis
1
2
3
4
5
6
7
8
None
None
None
None
None
None
None
None
TMSOTf (1.0)
Et2AlCl (1.0)
TMSOTf (1.0)
Et2AlCl (1.0)
TMSOTf (1.0)
Et2AlCl (1.0)
TMSOTf (1.0)
Et2AlCl (1.0)
TMSOTf (1.0)
Et2AlCl (1.0)
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
THF
2 h
1 h
30 min
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
66:33
70:30
75:25
90:10
90:10
60:40
68:32c
63:37c
90:10
86:14
85:15
80:20
78:22
72:28
69:31
67:33
69:31
67:33
5 min
d
2 h
2 h
2 h
5 min
5 min
15 min
15 min
30 min
30 min
1 h
CH2Cl2
THF
9
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
10
11
12
13
14
15
16
17
18
1 h
2 h
2 h
a All reactions were carried out at 25°C with 1.0 equiv. of Et2AlCN.
b All reactions showed quantitative yield.
c 3.0 equiv. of Et2AlCN were used.
d The reaction was quenched prior to the immediate addition of the nitrone which was made to react for 10 min.
Et2AlCN or on changing the solvent of the reaction
Thus, a solution of nitrone 2a in CH2Cl2 was treated
with 1.0 equiv. of Et2AlCN and the mixture was stirred
at 0°C. Quenching of the reaction was made by addi-
tion of a saturated aqueous sodium bicarbonate solu-
tion. Surprisingly, rather different diastereoselectivities
were obtained depending on the time of the reaction
(Table 3, entries 1–4). This result might be due to the
reversibility of the reaction. However, in an indepen-
dent experiment a cis/trans 2:3 mixture of hydroxyl-
amines 4 was placed in CH2Cl2 in the presence of
aqueous NaHCO3 containing ‘quenched’ Et2AlCN.
After 2 days the spectroscopic (1H NMR) analysis of
the mixture showed the same 2:3 ratio. The same
experiment was repeated with pure trans and cis
hydroxylamines, no variation of the diastereomeric
purity being observed. Furthermore, these results show
that epimerization of the hydroxylamines did not occur
and, in consequence, the reversibility of the cyanide
addition can be discarded, at least under these
conditions.
(Table 3, entries 6–8).
The efficiency of TMSOTf and DEAC as activating
agents was lower than for the addition of TMSCN,
probably due to the necessity of formation of a reactive
complex in the case of Et2AlCN addition (Scheme 6).
As will be considered in more detail in the mechanistic
section, those results suggest that such a complex would
be less favored if an electron pair of the nitrone oxygen
is compromised with the Lewis acid.
Moreover, when a solution of Et2AlCN in CH2Cl2 was
treated with saturated aqueous sodium bicarbonate and
the resulting mixture was treated with a solution of 2a
in CH2Cl2 (Table 3, entry 5) the reaction was complete
after 15 min and a diastereoselectivity of 9:1 was
obtained. This experiment clearly indicates that ‘free’
cyanide anion, produced after quenching is actual
nucleophile when reaction of Et2AlCN is not complete.
These observations are in agreement with the previ-
ously reported addition of cyanides to a particular
nitrone 2 (R=TBS).8a,9
Scheme 6.
In order to estimate some rate enhancement of the
reaction it was quenched at different times. After 5 min
it was found a higher selectivity for the TMSOTf-pro-
moted reaction (Table 3, entry 9) than that activated
with Et2AlCl (Table 3, entry 10). The same trend was
From these experiments it can be concluded that at
least 2 h are needed for the Et2AlCN addition to reach
completion. Low selectivities were observed and no
better results were obtained on using an excess of

372
P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
Table 4. Diethylaluminum cyanide additions to nitrone 2b–ea
Entry
R
Additive (equiv.)
Time (h)
Yield (%)b
trans:cis
1
2
3
4
5
6
7
8
2b
None
2
1
1
2
1
1
2
1
1
2
1
1
100
100
100
100
100
100
100
100
100
100
100
100
66:33
70:30
72:28
70:30
68:32
67:33
65:35
60:40
62:38
78:22
75:25
72:28
TMSOTf (1.0)
Et2AlCl (1.0)
None
TMSOTf (1.0)
Et2AlCl (1.0)
None
TMSOTf (1.0)
Et2AlCl (1.0)
None
TMSOTf (1.0)
Et2AlCl (1.0)
2c
2d
2e
9
10
11
12
a All reactions were carried out in CH2Cl2 at 25°C with 1.0 equiv. of Et2AlCN.
b All reactions showed quantitative yield.
complex D with concomitant formation of new species
to which structures 9 were tentatively assigned. After 2
h the spectrum did not show any additional change,
therefore it is experimentally proved that the use of
only 1.0 equiv. of Et2AlCN consumes completely the
nitrone, giving a quantitative yield of the product. For
shorter reaction times the reaction is complete after
quenching as a consequence of the free cyanide liber-
ated upon aqueous treatment of the reaction mixture.
maintained after 15 min, 30 min, 1 and 2 h (Table 3,
entries 11–18) at which time one can estimate that the
reaction is finished. These results confirmed a slight
Lewis acid activation for the Et2AlCN addition, the
TMSOTf-promoted reaction being slower than the
Et2AlCl promoted one. It can be concluded that a higher
reaction rate leads to lower diastereoselectivity.
The hydrocyanation with Et2AlCN was extended to
nitrones 2b–e; the results are summarized in Table 4.
The observed diastereoselectivities were similar to those
observed for 2a and similar slight enhancement of the
reaction rates were observed upon activation with
Lewis acids. Thus, the presence of TMSOTf or DEAC
speeds up the reaction, although diastereoselectivities
remain practically the same. Moreover, the selectivity is
mainly inferred by the vicinal alkoxy group, remaining
almost the same for all nitrone substrates and only
slightly increasing for the cis configured diol 2e and
decreasing for the trans 2d, as expected.
On the assumption that in the presence of additives the
formation of the final adducts occurs by nucleophilic
attack to the coordinated nitrone, we hoped that addi-
tion of Et2AlCN in the presence of DEAC would allow
us to detect an intermediate containing a nitrone bridg-
ing ligand between two aluminum atoms (in a similar
way to that found when 2 equiv. of DEAC were added
to 2a to form complex C—see above). Indeed, it was
possible to observe that the signal for the azomethine
proton of Et2AlCl complexed nitrone B shifts upfield
from 7.80 to 7.86 ppm upon addition of Et2AlCN at
−60°C, thus indicating the formation of complex E
(Scheme 7). This observation could not be confirmed in
the case of the TMSOTf activated nitrone because the
otherwise sharp lines of complex A became broad even
at low temperatures.
Unambiguous evidence for the formation of an initial
complex between nitrone 2a and Et2AlCN was obtained
1
by carrying out a H NMR study of the complexation
in Scheme 6 in CD2Cl2. In all cases, freshly prepared
Et2AlCN was used, obtained by the reaction of
trimethylsilyl cyanide and diethylaluminum chloride in
dry CH2Cl2.21
Upon injecting a CD2Cl2 solution of Et2AlCN (1
equiv.) into a CD2Cl2 solution of nitrone 2a at −60°C
1
the H NMR spectrum was immediately recorded. The
signal of the azomethine proton shifted downfield from
6.71 ppm (nitrone) to 7.70 ppm, thus indicating the
formation of complex D in a similar way to the addi-
tion of Et2AlCl. The complex showed broad signals,
probably due to either nitrogen inversion or presence of
rotamers around the NꢁO bond (see discussion
above).14–16 To this respect, it proved impossible to
record the spectrum at temperatures above 0°C,
because under these conditions irreversible CꢁC bond
formation set in simultaneously. At −60°C there is no
CꢁC bond formation for at least 1 h. Aging the reaction
by warming to 0°C showed gradual disappearance of
Scheme 7.
Both TMSOTf- and DEAC-promoted reactions pro-
ceeded more rapidly than those carried out in the
absence of any additive, the former being slower than
the latter. Moreover, relative rates between activated
and non-activated reactions with TMSCN and
Et2AlCN indicated a higher efficiency of the activation

P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
373
in the case of TMSCN addition. These observations
provided considerable insight into the two-step (com-
plex formation and cyanide transfer) mechanism of the
addition of Et2AlCN to nitrones 2 illustrated in Scheme
6.
Calculations were carried out with the most stable
conformer of the nitrone. The calculated reaction paths
are depicted in Scheme 8. For the addition of TMSCN
two transition structures, TS1-Re and TS1-Si, were
located. The geometries of these stationary points and
the nitrone are depicted in Fig. 4.28 The total and
relative energies as well as selected geometrical parame-
ters are summarized in Table 5.
Reaction retardation for the TMSOTf-promoted
Et2AlCN addition with respect to the Et2AlCl-pro-
moted reaction is consistent with lower availability of
the remaining electron-pair of the nitrone oxygen for
the crucial complexation with the aluminum atom of
the cyanating reagent. Thus, it can be postulated an
internal delivery of CN from complexes D or E for
Et2AlCN in contrast to the external delivery occurring
in reactions with TMSCN.
The transition structure corresponding to a Re attack,
TS1-Re, is more stable (3.74 kcal/mol) than TS1-Si
(corresponding to the Si attack), the former having a
relatively high energy barrier (37.25 kcal/mol), which is
in agreement with a slow reaction even at ambient
temperature. The IRC calculations confirm that transi-
tion structures connect 2 and the corresponding prod-
ucts and no intermediate stable complexes (minima of
energy) having a pentacoordinate silicon atom were
found following the reaction path. Such a pentacoordi-
nated silicon atom was found only in the transition
structures in agreement with previous calculations for
nucleophilic additions of silyl ketene acetals to
nitrones.29
2.3. Computational studies
To investigate the origin of the stereoselectivity of the
cyanation reactions we carried out a theoretical study
using semiempirical and DFT methods. Semiempirical
calculations were performed with MOPAC 200022 and
DFT calculations were carried out with Gaussian98.23
All stationary points were optimized first at the
semiempirical AM1 method,24 then moved to Gaus-
sian98 and optimized without any restriction at the
B3LYP/6-31G(d)25 level of theory. Vibrational frequen-
cies were calculated (1 atm, 298.15 K) for all B3LYP/6-
31G(d) optimized structures and used, unscaled, to
compute both ZPVE and activation energies. Intrinsic
reaction coordinate (IRC) analysis26 was performed to
verify that the transition structures connect reactants
and products. Graphical animations of the imaginary
frequencies (transition vectors) for all transition struc-
tures and IRC calculations ensured that the motions
were appropriate for converting reactants to products
and for verifying the reaction path.27 The only simplifi-
cation of the calculations consisted of replacement of
the tert-butyl group of the nitrone and the ethyl groups
of the Et2AlCN by methyl groups in all cases. So, the
calculated reactions were those between nitrone 2 (R=
Me) and Me2AlCN or TMSCN. A molecule of water
was also included in the study of the reaction with
Me2AlCN in order to simulate the presence of the
solvent.
The lengths of the forming CꢁC bonds in TS1-Re and
,
TS1-Si are 2.26 and 2.31 A, respectively. The lengths of
the SiꢁO distances in TS1-Re and TS1-Si are 1.84 and
,
1.83 A, respectively, thus showing that the SiꢁO bond
formation is more advanced than the CꢁC bond forma-
tion (even when the forming bond lengths are com-
pared with typical values of the stated bonds, i.e. CꢁC
and SiꢁO).
In the case of the addition of Me2AlCN the reaction
proceeds through a two-step mechanism, in agreement
with experimental observations and as illustrated in
Scheme 6. The first step corresponds to the formation
of a stable complex D1 (8.45 kcal/mol below the reac-
tants for DFT calculations) without energy barrier,
through the displacement of the molecule of solvent
(water) by the nitrone. The conformational analysis of
D1 (by rotation of NꢁO and OꢁAl bonds) give us clues
to understand the experimentally observed 1H NMR
spectra which presented broad signals at −60°C. Very
Scheme 8.

374
P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
Figure 4. B3LYP/6-31G(d) optimized structures of important stationary points of the TMSCN and Me2AlCN additions to 2
(R=Me). Some hydrogen atoms have been omitted for clarity. For selected geometrical data see Table 5.
,
Table 5. Total free energies, relative free energies and selected bond lengths (in A), for the stationary points of the
nucleophilic additions of TMSCN and Me2AlCN to 2 (R=Me)
AM1
B3LYP/6-31G(d)
Geometric parametersa
CꢁCd CꢁNe NꢁOe CꢁXf OꢁXf
Total energyb
Relative
Total energyc
Relative
energyb
energyb
2 (R=Me)
Me3SiCN
Me2AlCN·H2O −70.94
−23.21
−27.20
–
–
–
−400.946434
−502.039177
−491.513985
−902.926255
−902.920290
−903.002184
−903.000895
−892.474878
−892.438060
−892.429810
−892.488932
−892.487750
–
–
–
–
–
–
1.31
–
–
1.258
–
–
–
–
ꢁ
2.06
1.84
1.83
1.71
1.71
1.94
1.91 (2.18)
1.92 (2.20)
1.79 (2.03)
1.80 (2.02)
1.88
1.96
2.49
2.53
–
TS1-Re
TS1-Si
P1-trans
P1-cis
−7.04
−2.60
43.37g
47.81g
−26.96g
−24.43g
−3.65h
26.92i
33.01i
−28.05h
−21.94h
37.25g
40.99g
−10.40g
−9.59g
−8.45h
22.48i
27.65i
−17.89h
−17.15h
2.26
2.31
1.47
1.47
–
1.97
2.00
1.47
1.47
1.30
1.30
1.49
1.48
1.29
1.34
1.33
1.47
1.48
1.34
1.34
1.46
1.46
1.31
1.35
1.35
1.43
1.43
−77.37
−74.84
−97.80
−70.88
−64.79
−122.20
−119.74
−
D1j
2.01
2.32
2.28
–
TS2-Re
TS2-Si
P2-trans
P2-cis
–
a Only those corresponding to B3LYP/6-31G(d) optimized structures are given.
b kcal/mol.
c Hartrees.
d Referred to the forming CꢁC bond in transition structures.
e Referred to nitrone/hydroxylamine.
f X=Si for TS1 and P1; X=Al for D1, TS2 and P2; referred to nitrone oxygen; referred to coordinating water in brackets.
g Referred to 2+Me3SiCN.
h Referred to 2+Me2AlCN·H2O.
i Referred to complex D1.
j The energy corresponding to a molecule of water (−59.25 kcal/mol (AM1) and −76.406097 au (B3LYP/6-31G(d)) has been included for
comparison of relative energies.
low energy differences (1–2 kcal/mol) are found when
the PES defined by the two dihedral angles mentioned
above was analyzed30 (the most stable conformer is
depicted in Fig. 5). The second step is the intramolecu-
lar delivery of cyanide via transition structures TS2-Re
and TS2-Si to give trans and cis adducts, respectively.
The energy barriers are 22.48 and 27.65 kcal/mol for
TS2-Re and TS2-Si, respectively and the trans adduct is
more stable than the cis adduct. The final products are
lower in energy than complex D1, indicating an
exothermic transformation.
,
The lengths of the OꢁAl bonds change from 1.94 A in
,
the starting complex to 1.91 and 1.92 A in the transi-
tion structures, showing a continuance of the OꢁAl
bond during the rate-limiting step. The lengths of the
CꢁC forming bonds in TS2-Re and TS2-Si are 1.97 and
,
2.00 A, respectively.

P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
375
3. Conclusions
A complete study of the hydrocyanation of a-alkoxy
chiral non-racemic cyclic nitrones has been carried out.
Additionally, a rationale for the cyanation of cyclic
nitrones using TMSCN and Et2AlCN has been given.
According to both theoretical and experimental obser-
vations, whereas TMSCN reacts through a concerted
transition state, Et2AlCN initially forms a complex,
which evolves to the final product via intramolecular
carbon-carbon bond formation. On the basis of intrin-
sic reactivity data, there is no doubt that Et2AlCN
behaves as an intramolecular cyanating agent.
A comparison of the results reported in the Tables for
the two cyanating reagents employed under a variety of
conditions clearly show that the stereoselectivity fur-
nished by TMSCN is always considerably higher than
that seen with Et2AlCN. A study of the reaction profile
1
by H NMR and theoretical calculations allows us to
assign this difference to the switch from an intermolecu-
lar to an intramolecular mode of addition on passing
from TMSCN to Et2AlCN.
Careful examination of the reaction conditions
employed in this study allows us to raise a general
warning when Et2AlCN is used as a cyanating agent, to
ensure that the addition is complete before quenching,
the rate of addition being much slower than that of free
cyanide.
From a practical point of view, virtually quantitative
yields of adducts were obtained and complete stereose-
lectivity has been achieved in the addition of TMSCN
with all the nitrones under study. Moreover, very short
reaction times are attainable with the use of catalytic
amounts of TMSOTf or DEAC and no separation
technique is needed for recovery of hydroxylamine
products. This study suggests that the same behavior
will be displayed by at least all 3-oxy substituted pyrro-
line N-oxides, which makes their cyanation a general
and valuable method for the synthesis of adducts of the
type 3. In turn, compounds 3 can be envisaged as useful
intermediates for obtaining substituted prolines, as
demonstrated by a recently patented work,9 or chiral
diamines usable as chelating ligands for metal complex-
ation and related to those already employed in clinical
trials with Pt-based anticancer agents.31
Figure 5. Energy profiles of the TMSCN (blue) and
Me2AlCN (red) additions to 2 (R=Me) at the AM1 and
B3LYP/6-31G(d) level.
For the purpose of comparison, the energy profiles of
the addition reactions of the two cyanating reagents are
given in Fig. 5. These energetic differences are in agree-
ment with experimental results concerning both selec-
tivity (Re attack is preferred to Si attack) and reactivity
(Me2AlCN is more reactive than TMSCN). However,
the selectivity differences observed between the two
cyanating reagents are not well-described. It should be
noted that for
a quantitative evaluation of the
4. Experimental
diastereomeric ratio the B3LYP/6-31G(d) level of the-
ory might not be sufficiently flexible to produce reliable
energetics across the entire potential energy surface.
The reaction flasks and other glass equipment were
heated in an oven at 130°C overnight and assembled in
a stream of Ar. All reactions were monitored by TLC
on silica gel 60 F254; the position of the spots were
detected with 254 nm UV light or by spraying with one
of the following staining systems: 50% methanolic sul-
furic acid, 5% ethanolic phosphomolybdic acid and
iodine. Preparative centrifugally accelerated radial thin-
layer chromatography (PCAR-TLC) was performed
with a Chromatotron® Model 7924 T (Harrison
Research, Palo Alto, CA, USA) and with solvents that
In conclusion, based upon these calculations we pro-
pose that TMSCN is added to nitrone 2 in a one-step
process without formation of any stable intermediate,
via transition states TS1-Re and TS1-Si. In contrast,
the addition of Et2AlCN to 2 takes place via transition
structures TS2-Re and TS2-Si formed from the inter-
mediate complex D1, which has been detected by
NMR.

376
P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
were distilled prior to use; the rotors (1 or 2 mm layer
thickness) were coated with silica gel Merck grade type
7749, TLC grade, with binder and fluorescence indica-
tor (Aldrich 34,644-6) and the eluting solvents were
delivered by the pump at a flow-rate of 0.5–1.5 mL
min−1. Melting points were uncorrected. 1H and 13C
NMR spectra were recorded on a Varian Unity or on a
Bruker 300 instrument in CDCl3. Chemical shifts are
reported in ppm (l) relative to CHCl3 (l=7.26) in
CDCl3. Optical rotations were taken at 25°C on a
Perkin–Elmer 241 polarimeter. Elemental analysis were
performed on a Perkin Elmer 240B microanalyzer.
Nitrones 2a,10a 2b,10b 2d11 and 2e4a were prepared as
described. Diethylaluminum cyanide was prepared as
described.21
solid was washed thoroughly with ether (3×50 mL). The
ethereal washings were concentrated and the residue
was purified by flash column chromatography
(petroleum ether:ethyl acetate, 3:1) to afford pure 12
(1.0 g, 66%) as a pale yellow oil. Rf=0.51; [h]2D5=+17 (c
1
2.2, CHCl3); H NMR (CDCl3) l 1.56 (bs, 1H), 1.98,
(bs, 1H), 2.65–3.05 (m, 5H), 3.09 (s, 3H), 4.1 (bs, 1H),
4.36 (AB system, 2H). 13C NMR (CDCl3) l 29.8, 54.9,
57.1, 64.3, 75.3, 95.4. Anal. calcd for C6H13NO3: C,
48.97; H, 8.90; N, 9.52. Found: C, 48.90; H, 8.80; N,
9.13%.
4.1.3. 4-Methoxymethoxy-3,4-dihydro-2H-pyrrole 1-
oxide. A solution of 12 (700 mg, 4.76 mmol) in CH2Cl2
(25 mL), cooled into an ice bath, was added with MnO2
(472 mg, 5.26 mmol). The suspension was stirred at 0°C
for 1 h, then at rt overnight. The suspension was
filtered on a short pad of Celite and Na2SO4 and the
solution concentrated. The crude reaction mixture con-
tained the two regioisomeric nitrones 2c and 13 in a 9:1
ratio. Separation by flash column chromatography
(ethyl acetate:ethanol, 1:5) afforded pure nitrones 2c
(590 mg, 85%) as a colourless oil and 13 (100 mg, 14%)
as a yellowish solid.
4.1. Synthesis of 4-methoxymethoxy-3,4-dihydro-2H-
pyrrole 1-oxide
The nitrone 2c was synthesized following the procedure
reported in Scheme 9.
4.1.1. Methanesulfonic acid 4-methanesulfonyloxy-3-
methoxymethoxy-butyl ester, 11. A solution of methane-
sulfonic acid 2-hydroxy-4-methanesulfonyloxy-butyl
ester 1032 (1.2 g, 4.57 mmol) in CHCl3 (60 mL) was
mixed with dimethoxymethane (9 mL, 101.7 mmol) and
phosphorus pentoxide (4.5 g, 31.71 mmol) and stirred
at room temperature for 45 min. The suspension was
then cooled with ice and quenched by slow addition of
a saturated Na2CO3 solution. The aqueous phase was
washed with ether (3×30 mL) and the washings were
added to the CHCl3 solution. The organic phase was
washed with brine, dried with Na2SO4, and concen-
trated. The crude reaction mixture was purified by
passage on a short pad of silica gel (petroleum
ether:ethyl acetate, 1:1) to afford pure 11 (1.41 g, 100%)
as a colorless oil; Rf=0.37; [h]2D5=−40 (c 0.46, CHCl3);
1H NMR (CDCl3) l 1.98–2.08 (s, 2H), 3.03 (s, 3H),
3.05 (s, 3H), 3.40 (s, 3H), 3.98 (dq, 1H, J=6.6, 4.4 Hz),
4.20–4.32 (m, 2H), 4.37 (t, 2H, J=6.6 Hz), 4.67–4.76
(AB system, 2H). 13C NMR (CDCl3) l 31.5, 37.5, 37.6,
56.1, 65.6, 70.3, 71.9, 96.8. Anal. calcd for C8H18S2O8:
C, 31.37; H, 5.93. Found: C, 31.20; H, 5.80%.
4.1.3.1. 4-Methoxymethoxy-3,4-dihydro-2H-pyrrole 1-
oxide, 2c. Rf=0.29; [h]2D5=−77 (c 1.28, CHCl3); 1H
NMR (CDCl3) l 1.97–2.12 (m, 1H), 2.37–2.56 (m, 1H),
3.20 (s, 3H), 3.71 (ddd, 1H, J=9.5, 5.2, 5.1 Hz),
3.90–4.05 (m, 1H), 4.51 (AB system, 2H), 4.67 (d, 1H,
J=7.3 Hz), 6.85, (s, 1H). 13C NMR (CDCl3) l 27.8,
55.5, 60.9, 77.2, 99.0, 134.5. Anal. calcd for C6H11NO3:
C, 49.65; H, 7.64; N, 9.65. Found: C, 49.84; H, 7.36; N,
9.83%.
4.1.3.2. 3-Methoxymethoxy-3,4-dihydro-2H-pyrrole 1-
oxide, 13. Mp 35–37°C. Rf=0.11; [h]2D5=33 (c 2.63,
1
CHCl3); H NMR (CDCl3) l 2.75 (dd, 1H, J=19.4,
1.5Hz), 3.03 (ddd, 1H, J=19.4, 6.6, 1.5 Hz), 3.35 (s,
3H), 3.92 (d, 1H, J=14.6 Hz), 4.15 (dd, 1H, J=14.6,
6.6 Hz), 4.51 (dq, 1H, J=9.6, 2.9 Hz), 4.67 (AB system,
2H), 6.85, (s, 1H). 13C NMR (CDCl3) l 36.8, 55.2, 68.1,
70.6, 95.7, 132.8. Anal. calcd for C6H11NO3: C, 49.65;
H, 7.64; N, 9.65. Found: C, 49.55; H, 7.74; N, 9.45%.
4.1.2. 3-Methoxymethoxy-pyrrolidin-1-ol, 12. A suspen-
sion of 11 (3.0 g, 9.43 mmol) and hydroxylamine hydro-
chloride (3.14 g, 45.24 mmol) in triethylamine (60 mL)
was heated under reflux with vigorous stirring for 4 h.
The suspension was concentrated and the remaining
4.2. (2R,3S)-3-tert-Butoxy-1-(trimethylsiloxy)-pyrro-
lidine-2-carbonitrile, 3
To a solution of nitrone 2 (0.157 g, 1 mmol) in CH2Cl2
(10 mL) was added trimethylsilyl cyanide (0.100 g, 1
mmol). The resulting solution was stirred at ambient
temperature for 16 h, at which time the solvent was
rotatory evaporated to afford pure 3 (0.256 g, 100%) as
1
a colourless oil; [h]2D5=−62 (c 0.56, CHCl3); H NMR
(CDCl3, 55°C) l 0.14 (s, 9H), 1.18 (s, 9H), 1.65 (ddd,
1H, J=3.9, 8.3, 12.7 Hz), 2.19 (dt, 1H, J=8.3, 12.7
Hz), 2.86–3.12 (m, 2H), 3.50 (d, 1H, J=5.4 Hz), 4.30–
4.36 (m, 1H). 13C NMR (CDCl3, 55°C) l −0.7, 28.2,
31.8, 57.1, 66.4, 73.9, 74.6, 114.4 Anal. calcd for
C12H24N2O2Si: C, 56.21; H, 9.43; N, 10.92. Found: C,
56.49; H, 9.11; N, 11.25%.
Scheme 9. Reagents: (i) CH3OCH2OCH3, P2O5. (ii) NH2OH.
(iii) MnO2.

P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
377
4.3. Addition of trimethylsilyl cyanide to nitrones in the
absence of Lewis acids. General procedure
(CDCl3) l 24.3, 25.8, 60.5, 62.9, 76.1, 80.1, 112.6, 114.8.
Anal. calcd for C8H12N2O3: C, 52.17; H, 6.57; N, 15.21.
Found: C, 52.21; H, 6.53; N, 15.18%.
To a solution of the corresponding nitrone (1 mmol) in
CH2Cl2 (10 mL) was added trimethylsilyl cyanide
(0.100 g, 1 mmol). The resulting solution was stirred at
ambient temperature for 16 h, at which time the solvent
was rotatory evaporated. The residue was taken up into
5% methanolic citric acid (10 mL) and the resulting
mixture was stirred for 1 h, at which time saturated
aqueous sodium bicarbonate (25 mL) was added. The
reaction mixture was extracted with dichloromethane
(3×15 mL) and the combined organic extracts were
dried (MgSO4) and evaporated to give the free hydroxyl-
amines which did not need further purification.
4.4. Addition of trimethylsilyl cyanide to nitrones in the
presence of Lewis acids. General procedure
To a solution of the corresponding nitrone (1 mmol) in
CH2Cl2 (10 mL) was added the corresponding Lewis
acid (1 mmol for stoichiometric reactions or 0.2 mmol
for catalytic reactions). The resulting mixture was
stirred for 5 min at ambient temperature, at which time
trimethylsilyl cyanide (0.100 g, 1 mmol) was added.
After stirring for 5 min, the solvent was rotatory evap-
orated. The residue was taken up into 5% methanolic
citric acid (10 mL) and the resulting mixture was stirred
for 1 h, at which time saturated aqueous sodium bicar-
bonate (25 mL) was added. The reaction mixture was
extracted with dichloromethane (3×15 mL) and the
combined organic extracts were dried (MgSO4) and
evaporated to give the free hydroxylamines which did
not need further purification.
4.3.1.
(2R,3S)-3-tert-Butoxy-1-hydroxy-pyrrolidine-2-
carbonitrile, 4. (0.184 g, 100%); oil; [h]2D5=+20 (c 1.00,
1
CHCl3); H NMR (CDCl3) l 1.14 (s, 9H), 1.69 (ddt,
1H, J=13.4, 8.6, 3.7 Hz), 2.22 (dq, 1H, J=13.4, 8.6
Hz), 3.06 (dt, 1H, J=11.7, 8.6 Hz), 3.18 (ddd, 1H,
J=11.7, 8.1, 3.7 Hz), 3.58 (d, 1H, J=5.9 Hz), 4.33
(ddd, 1H, J=8.9, 5.9, 3.7 Hz), 7.41 (s, 1H, ex. D2O).
13C NMR (CDCl3) l 28.1, 31.6, 56.2, 67.7, 73.3, 74.9,
117.8. Anal. calcd for C9H16N2O2: C, 58.67; H, 8.75; N,
15.21. Found: C, 58.77; H, 8.63; N, 15.31%.
4.5. Addition of diethylaluminium cyanide to nitrones in
the absence of Lewis acids. General procedure
To a solution of the corresponding nitrone (1 mmol) in
CH2Cl2 (10 mL) was added diethylaluminium cyanide
(0.111 g, 1 mmol). The resulting solution was stirred at
ambient temperature (for time see Tables 3 and 4), at
which time the reaction mixture was quenched by the
addition of saturated aqueous sodium bicarbonate (15
mL). The resulting mixture was stirred for additional 5
min at ambient temperature and then extracted with
CH2Cl2 (3×15 mL). The combined organic extracts
were dried (MgSO4) and evaporated to give the crude
mixture which was taken up into 5% methanolic citric
acid (10 mL) and the resulting mixture was stirred for 1
h, at which time saturated aqueous sodium bicarbonate
(25 mL) was added. The reaction mixture was extracted
with dichloromethane (3×15 mL) and the combined
organic extracts were dried (MgSO4) and evaporated to
give the pure mixture of free hydroxylamines. Separa-
tion of the diastereomers was performed by preparative
centrifugally accelerated thin layer chromatography
with a Chromatotron®.
4.3.2. (2R,3S)-1-Hydroxy-3-(triisopropylsiloxy)-pyrro-
lidine-2-carbonitrile, 5. (0.285 g, 100%); oil; [h]2D5=−37
1
(c 0.27, CHCl3); H NMR (CDCl3) l 1.05 (s, 21H),
1.68–1.72 (m, 1H), 2.28–2.31 (m, 1H), 3.20–3.24 (m,
2H), 3.73 (s, 1H), 4.63–4.66 (m, 1H), 6.48 (bs, 1H, ex.
D2O). 13C NMR (CDCl3) l 6.8, 12.3, 27.8, 50.5, 62.1,
69.2, 118.2. Anal. calcd for C14H28N2O2Si: C, 59.11; H,
9.92; N, 9.85. Found: C, 59.22; H, 9.86; N, 9.97%.
4.3.3. (2R,3S)-1-Hydroxy-3-(methoxymethoxy)-pyrro-
lidine-2-carbonitrile 6. (0.173 g, 100%); oil; [h]2D5=+9 (c
1
0.22, CHCl3); H NMR (CDCl3) l 1.79–2.01 (m, 1H),
2.20–2.44 (m, 1H), 3.07–3.34 (m, 2H), 3.43 (s, 3H), 3.85
(d, 1H, J=4.4 Hz), 4.37–4.50 (m, 1H), 4.69 (pseudo d,
2H, J=1.5 Hz), 5.62 (bs, 1H, ex. D2O). 13C NMR
(CDCl3) l 28.7, 55.0, 55.2, 64.1, 77.7, 95.6, 117.7. Anal.
calcd for C7H12N2O3: C, 48.83; H, 7.02; N, 16.27.
Found: C, 48.88; H, 6.90; N, 16.15%.
4.3.4. (2S,3S,4S)-3,4-Di-tert-butoxy-1-hydroxy-pyrro-
lidine-2-carbonitrile, 7. (0.256 g, 100%); oil; [h]2D5=−16
4.6. Addition of diethylaluminium cyanide to nitrones in
the presence of Lewis acids. General procedure
1
(c 0.80, CHCl3); H NMR (CDCl3) l 1.15 (s, 9H), 1.22
(s, 9H), 3.14 (dd, 1H, J=10.3, 4.4 Hz), 3.18 (dd, 1H,
J=10.3, 6.4 Hz), 3.60 (d, 1H, J=5.0 Hz), 3.89 (ddd,
1H, J=6.4, 4.4, 3.4 Hz), 4.12 (dd, 1H, J=5.0, 3.4 Hz),
5.80 (bs, 1H, ex. D2O). 13C NMR (CDCl3) l 28.2, 28.6,
31.5, 56.4, 69.3, 74.0, 75.8, 76.2, 116.4. Anal. calcd for
C13H24N2O3: C, 60.91; H, 9.44; N, 10.93. Found: C,
61.06; H, 9.38; N, 11.00%.
To a solution of the corresponding nitrone (1 mmol) in
CH2Cl2 (10 mL) was added the corresponding Lewis
acid (1 mmol for stoichiometric reactions or 0.2 mmol
for catalytic reactions). The resulting mixture was
stirred for 5 min at ambient temperature, at which time
diethylaluminium cyanide (0.111 g, 1 mmol) was added.
The reaction mixture was stirred at ambient tempera-
ture (for time see Tables 3 and 4), and then it was
quenched by the addition of saturated aqueous sodium
bicarbonate (15 mL). The resulting mixture was stirred
for additional 5 min at ambient temperature and then
extracted with CH2Cl2 (3×15 mL). The combined
organic extracts were dried (MgSO4) and evaporated to
4.3.5.
(2R*,3S*,4R*)-5-Hydroxy-2,2-dimethyl-tetra-
hydro-[1,3]dioxolo[4,5-c]pyrrole-4-carbonitrile, 8. (0.184
g, 100%); white solid; mp 110–112°C; 1H NMR
(CDCl3) l 1.27 (s, 3H), 1.44 (s, 3H), 3.15 (dd, 1H,
J=11.4, 4.8 Hz), 3.42 (d, 1H, J=11.4 Hz), 4.20 (s, 1H),
4.70–4.80 (m, 2H), 6.18 (bs, 1H, ex. D2O). 13C NMR

378
P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
give the crude mixture which was taken up into 5%
methanolic citric acid (10 mL) and the resulting mixture
was stirred for 1 h, at which time saturated aqueous
sodium bicarbonate (25 mL) was added. The reaction
mixture was extracted with dichloromethane (3×15 mL)
and the combined organic extracts were dried (MgSO4)
and evaporated to give the pure mixture of free hydroxyl-
amines. Separation of the diastereomers was performed
by preparative centrifugally accelerated thin layer chro-
matography with a Chromatotron®.
Cofin 2002); MCYT and MIUR for a bilateral
exchange project (2000-2001). J.R. and V.M. acknowl-
edge a fellowship within an Erasmus/Socrates student
exchange program. Dr. Cristina Faggi (University of
Firenze) is acknowledged for performing the X-ray
analysis.
References
1. Merino, P.; Merchan, F. L.; Franco, S.; Tejero, T. Synlett
2000, 442–454 and references cited therein.
4.6.1.
(2S,3S)-3-tert-Butoxy-1-hydroxy-pyrrolidine-2-
carbonitrile, cis-4. (0.061 g, 33%); oil; [h]2D5=+21 (c
2. In this work we use the syn/anti nomenclature according
to Masamune convention. See: (a) Masamune, S.; Ali, S.
A.; Snitman, D. L.; Garvey, D. S. Angew. Chem., Int. Ed.
Engl. 1980, 19, 557–558; (b) Masamune, S.; Kaiho, T.;
Garvey, D. S. J. Am. Chem. Soc. 1982, 104, 5521–5523.
3. (a) Merino, P.; Merchan, F. L.; Lanaspa, A.; Tejero, T.
J. Org. Chem. 1996, 64, 9028–9032; (b) Merchan, F. L.;
Merino, P.; Tejero, T. Tetrahedron Lett. 1995, 36, 6949–
6952.
1
0.40, CHCl3); H NMR (CDCl3) l 1.22 (s, 9H), 1.74–
1.97 (m, 1H), 2.07–2.34 (m, 1H), 2.84 (dt, 1H, J=10.3,
8.8 Hz), 3.36 (ddd, 1H, J=10.3, 8.8, 4.4 Hz), 3.85 (d,
1H, J=6.6 Hz), 4.28 (ddd, 1H, J=8.9, 5.9, 3.7 Hz),
6.00 (s, 1H, ex. D2O).13C NMR (CDCl3) l 27.4, 31.3,
54.7, 64.4, 68.3, 74.5, 115.6. Anal. calcd for C9H16N2O2:
C, 58.67; H, 8.75; N, 15.21. Found: C, 58.57; H, 8.80;
N, 15.15%.
4. For leading references on the synthetic utility of nitrones
2, see: (a) Goti, A.; Cicchi, S.; Cacciarini, M.; Cardona,
F.; Fedi, V.; Brandi, A. Eur. J. Org. Chem. 2000, 3633–
3645 and references cited therein; (b) Goti, A.; Cacciarini,
M.; Cardona, F.; Cordero, F. M.; Brandi, A. Org. Lett.
2001, 3, 1367–1369; (c) Socha, D.; Jurczak, M.; Frelek, J.;
Klimek, A.; Rabiczko, J.; Urbanczyk-Lipkowska, Z.;
Suwinska, K.; Chmielewski, M.; Cardona, F.; Goti, A.;
Brandi, A. Tetrahedron: Asymmetry 2001, 12, 3163–3172;
(d) Cordero, F. M.; Pisaneschi, F.; Gensini, M.; Goti, A.;
Brandi, A. Eur. J. Org. Chem. 2002, 1941–1951; (e)
Giovannini, R.; Marcantoni, E.; Petrini, M. J. Org.
Chem. 1995, 60, 5706–5707; (f) McCaig, A. E.; Meldrum,
K. P.; Wightman, R. H. Tetrahedron 1998, 54, 9429–
9446; (g) Murahashi, S.-I.; Imada, Y.; Ohtake, H. J. Org.
Chem. 1994, 59, 6170–6172; (h) Ishikawa, T.; Tajima, Y.;
Fukui, M.; Saito, S. Angew. Chem., Int. Ed. Engl. 1996,
35, 1863–1864.
5. (a) Xiang, Y.-B.; Drenkard, S.; Baumann, K.; Hickey,
D.; Eschenmosher, A. Helv. Chim. Acta 1994, 77, 2209–
2250; (b) Wong, C.-H.; Provencher, L.; Porco, J. A.;
Jung, S.-H.; Wang, Y.-F.; Chen, L.; Wang, R.; Steensma,
D. H. J. Org. Chem. 1995, 60, 1492–1501; (c) Yasuda,
M.; Shigehisa, S.; Arakawa, Y.; Yoshifuji, S. Chem.
Pharm. Bull. 1995, 43, 1318–1324; (d) Arakawa, Y.;
Yoshifuji, S. Chem. Pharm. Bull. 1991, 39, 2219–2224; (e)
Mulzer, J.; Meier, A.; Buschmann, J.; Luger, P. Synthesis
1996, 123–132; (f) Halab, L.; Belec, L.; Lubell, W. D.
Tetrahedron 2001, 57, 6439–6446; (g) Sigman, M. S.;
Vachal, P.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl.
2000, 39, 1279–1281.
4.6.2. (2S,3S)-1-Hydroxy-3-(triisopropylsiloxy)-pyrro-
lidine-2-carbonitrile, cis-5. (0.094 g, 33%); oil; [h]2D5=
1
+15 (c 0.10, CHCl3); H NMR (CDCl3) l 1.05 (s, 21H),
1.83–1.86 (m, 1H), 2.26–2.28 (m, 1H), 2.95 (q, 1H,
J=8.3 Hz), 3.92 (d, 1H, J=6.8 Hz), 4.60 (pseudo q,
1H, J=6.5 Hz), 6.12 (bs, 1H, ex. D2O). 13C NMR
(CDCl3) l 6.8, 12.0, 32.8, 55.5, 65.7, 69.8, 115.7. Anal.
calcd for C14H28N2O2Si: C, 59.11; H, 9.92; N, 9.85.
Found: C, 59.34; H, 9.69; N, 9.91%.
4.6.3.
(2S,3S)-1-Hydroxy-3-(methoxymethoxy)-pyrro-
lidine-2-carbonitrile, cis-6. (0.052 g, 30%); sticky oil;
1
[h]2D5=−50 (c 0.10, CHCl3); H NMR (CDCl3) l 1.93–
2.17 (m, 1H), 2.24–2.47 (m, 1H), 2.87–3.05 (m, 1H),
3.35–3.45 (m, 1H), 3.46 (s, 3H), 3.96 (d, 1H, J=5.9
Hz), 4.31–4.47 (m, 1H), 4.74 (s, 2H), 5.34 (bs, 1H, ex.
D2O). 13C NMR (CDCl3) l 30.1, 55.2, 56.2, 63.9, 73.9,
96.1, CN carbon undetected. Anal. calcd for
C7H12N2O3: C, 48.83; H, 7.02; N, 16.27. Found: C,
48.80; H, 6.95; N, 16.10%.
4.6.4.
(2S*,3S*,4R*)-5-Hydroxy-2,2-dimethyl-tetra-
cis-8.
hydro-[1,3]dioxolo[4,5-c]pyrrole-4-carbonitrile,
1
(0.040 g, 22%); oil; H NMR (CDCl3) l 1.31 (s, 3H),
1.54 (s, 3H), 2.80 (d, 1H, J=11.4 Hz), 3.47 (s, 1H), 3.55
(d, 1H, J=11.4 Hz), 3.60–3.62 (m, 1H), 4.73–4.76 (m,
2H, ex. 1H D2O). 13C NMR (CDCl3) l 24.7, 25.8, 61.7,
63.6, 76.6, 76.7, 112.8, 115.5. Anal. calcd for
C8H12N2O3: C, 52.17; H, 6.57; N, 15.21. Found: C,
52.26; H, 6.68; N, 15.03%.
6. (a) Koskinen, A. M. P.; Rapoport, H. J. Med. Chem.
1985, 28, 1301–1309; (b) Shiosaki, K.; Rapoport, H. J.
Org. Chem. 1985, 50, 1229–1239; (c) Niopas, I.; Smail, G.
A. J. Chem. Soc., Perkin Trans. 1 1991, 119–125 see also
Ref. 5f.
Acknowledgements
We thank for their support of our programs: the
Spanish Ministry of Science and Technology (MCYT)
and FEDER Program (Project CASANDRA,
BQU2001-2428) and the Government of Aragon (Pro-
ject P116-2001); the Ministry of Instruction, University
and Scientific Research (MIUR, Italy, Cofin 2000 and
7. Guerra, F. M.; Mish, M.; Carreira, E. M. Org. Lett.
2000, 2, 4265–4267.
8. (a) Ohtake, H.; Imada, Y.; Murahashi, S.-I. Bull. Chem.
Soc. Jpn. 1999, 72, 2737–2754; (b) Peer, A.; Vasella, A.
Helv. Chim. Acta 1999, 82, 1044–1065; (c) Lombardo,
M.; Fabbroni, S.; Trombini, C. J. Org. Chem. 2001, 66,

P. Merino et al. / Tetrahedron: Asymmetry 14 (2003) 367–379
379
1264–1268; (d) Goessinger, E. Monatsh. Chem. 1982,
495–508.
9. Takahashi, H.; Yamaura, M.; Yamanobe, A.;
Yoshimura, T. Jpn. Kokai Tokkyo Koho 2002, JP
2002088059.
10. (a) Cicchi, S.; Goti, A.; Brandi, A. J. Org. Chem. 1995,
60, 4743–4748; (b) Goti, A.; Cacciarini, M.; Cardona, F.;
Brandi, A. Tetrahedron Lett. 1999, 40, 2853–2856.
11. Cicchi, S.; Ho¨ld, I.; Brandi, A. J. Org. Chem. 1993, 58,
5274–5275.
12. Enantiomerically pure (+)- and (−)-2e are also available:
(a) Cicchi, S.; Corsi, M.; Brandi, A.; Goti, A. J. Org.
Chem. 2002, 67, 1678–1681; (b) Closa, M.; Wightman, R.
H. Synth. Commun. 1998, 28, 3443–3450.
13. The diastereoselectivity was sufficiently high that the
minor diastereomer was not observed. In such instances
the minimal diastereomer ratio is given as >20:1, the limit
of detectability by 13C NMR spectroscopy. This is the
signal-to-noise limit of detection. See: Hiemstra, H.;
Wynberg, H. Tetrahedron Lett. 1977, 18, 2183–2186.
14. (a) Riddell, F. G.; Lehn, J.-M.; Wagner, J. Tetrahedron
1970, 26, 4227–4240; (b) Kost, D.; Raban, M. J. Org.
Chem. 1976, 41, 1748–1751; (c) Perrin, C. L.; Thoburn, J.
D.; Elsheimer, S. J. Org. Chem. 1991, 56, 7034–7038.
15. For a general study on the internal rotation of hydroxyl-
amines, see: Radom, L.; Hehre, W. J.; Pople, J. A. J. Am.
Chem. Soc. 1972, 94, 2371–2381.
16. For studies with O-substituted hydroxylamines, see: (a)
Raban, M.; Kost, D. J. Org. Chem. 1972, 37, 499–501;
(b) Walker, S.; Yang, D.; Kahne, D. J. Am. Chem. Soc.
1991, 113, 4716–4717; (c) Walker, S.; Gange, D.; Gupta,
V.; Kahne, D. J. Am. Chem. Soc. 1994, 116, 3197–3206.
17. Silyltriflate activation: (a) Wuts, P. G. M.; Jung, Y.-W. J.
Org. Chem. 1988, 53, 1957–1965; (b) Camiletti, C.;
Dhavale, D. D.; Donati, F.; Trombini, C. Tetrahedron
Lett. 1995, 36, 7293–7296; (c) Camiletti, C.; Dhavale, D.
D.; Gentilucci, L.; Trombini, C. J. Chem. Soc., Perkin
Trans. 1 1993, 3157–3165; (d) Bojin, M. L.; Evans, S. A.,
Jr. Phosphorus, Sulfur & Silicon 1996, 111, 789; (e)
DeRisi, C.; Dondoni, A.; Perrone, D.; Pollini, G. P.
Tetrahedron Lett. 2001, 42, 3033–3036; (f) Merino, P.;
Del Alamo, E.; Bona, M.; Franco, S.; Merchan, F. L.;
Tejero, T.; Vieceli, O. Tetrahedron Lett. 2000, 41, 9239–
9243. Aluminum chloride activation: (g) Merino, P.;
Franco, S.; Merchan, F. L.; Tejero, T. Tetrahedron:
Asymmetry 1997, 8, 3489–3496; (h) Dondoni, A.; Franco,
S.; Junquera, F.; Merchan, F. L.; Merino, P.; Tejero, T.;
Bertolasi, V. Chem. Eur. J. 1995, 1, 505–520. See also
Ref. 1.
The coordinates can be obtained on request from the
Director, Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB2 1EZ, UK.
20. For previous reports on 1H NMR data of hydroxylam-
monium ions formed from nitrones and Lewis acids, see
Refs. 17c and 17h.
21. Imi, K.; Yanagihara, N.; Utimoto, K. J. Org. Chem.
1987, 52, 1013–1016.
22. MOPAC 2000 was used as implemented in the
ChemOffice Ultra 2002 suite of programs. CSC Cam-
bridge Software, Cambridge, 2002.
23. All DFT calculations were performed using the Gaussian
98 Revision A.9 suite of programs. Frisch, J.; Trucks, G.
W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.;
Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J.
A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.;
Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M.
C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.;
Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clif-
ford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.;
Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;
Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.;
Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,
A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.;
Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople,
J. A. Gaussian, Inc., Pittsburgh PA, 1999.
24. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart,
J. J. P. J. Am. Chem. Soc. 1985, 107, 3902–3909.
25. (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652; (b)
Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100; (c) Lee,
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–
789.
26. (a) Fukui, K. Acc. Chem. Res. 1981, 14, 363–368; (b)
Head-Gordon, M.; Pople, J. A. J. Chem. Phys. 1988, 89,
5777–5786.
27. We used MOLDEN 3.7 software for graphical anima-
tions. See: Schaftenaar, G.; Noordik, J. M. J. Comput.
Aided Mol. Design 2000, 14, 123–134 or visit: http://
www.cmbi.kun.nl/ꢀschaft/molden/molden.html.
28. The graphic views showed in Fig. 3 were made with
PovChem software. Copyright by Thiessen, P.A., 1999–
2000.
29. Jost, S.; Gimbert, Y.; Greene, A. E.; Fotiadu, F. J. Org.
Chem. 1997, 62, 6672–6677.
30. The optimized geometry of D1 was obtained by generat-
ing 72×72 (5184) different conformations by stepwise
rotation of 5° around both NꢁO and OꢁAl bonds. From
the corresponding contour map in which the heat of
formation (DHf) of all these conformations was plotted as
a function of the two dihedral angles, the minimum
energy conformation was gained and then it was reopti-
mized.
31. Kim, D.-K.; Kim, Y.-W.; Kim, H.-T.; Kim, K. H.
Bioorg. Med. Chem. Lett. 1996, 6, 643–646.
32. Goti, A.; Cicchi, S.; Fedi, V.; Nannelli, L.; Brandi, A. J.
Org. Chem. 1997, 62, 3119–3125.
18. When the reaction was carried out in the presence of
additives, the nitrone was previously treated at ambient
temperature with the additive for 5 min. Therefore, it can
be assumed that complexes between the nitrone and the
additive were formed (for experimental evidence see
below).
19. The graphic view shown in Fig. 4 was made with
ORTEP3 software. Copyright by Farrugia, L. J. Univer-
sity of Glasgow, 1997–2000. The authors have deposited
the atomic coordinates for this structure with the Cam-
bridge Crystallographic Data Centre (CCDC 200710).