Cyclization of zincated α-N-homoallylamino nitriles: A new entry to enantiopure 2,3-methanopyrrolidines

Article abstract of DOI:10.1002/chem.201001639

Stereoselective cyclization of zincated α-N-homoallylamino nitriles has been developed. Following treatment with lithium diisopropylamide (LDA) and transmetalation with zinc bromide, α-N-(1-phenylethyl)-N-homoallylamino nitriles lead to 2,3-methanopyrroli

Full text of DOI:10.1002/chem.201001639

DOI: 10.1002/chem.201001639
Cyclization of Zincated a-N-Homoallylamino Nitriles: A New Entry to
Enantiopure 2,3-Methanopyrrolidines
Sabrina Ouizem, Sandrine Cheramy, Candice Botuha, Fabrice Chemla,*
Franck Ferreira, and Alejandro Pꢀrez-Luna[a]
Dedicated to Professor Carmen Najera on the occasion of her 60th birthday
Abstract: Stereoselective cyclization of
zincated a-N-homoallylamino nitriles
has been developed. Following treat-
ment with lithium diisopropylamide
(LDA) and transmetalation with zinc
bromide, a-N-(1-phenylethyl)-N-homo-
allylamino nitriles lead to 2,3-methano-
pyrrolidines in moderate to good yields
(up to 66%) and excellent selectivities
(up to >98:2). With substrates derived
from a-branched homoallylic amines, a
stereospecific inversion of the homoal-
lylic stereogenic center was observed.
To account for this, a mechanistic ra-
tionale involving the formation of zin-
cioiminium ions from zincated a-amino
nitriles is put forward. 2,3-Methanopyr-
rolidines should then arise from a se-
quence involving an aza-Cope rear-
rangement providing a configurational-
ly stable (2-azoniaallyl)zinc species that
then undergoes a [3+2] cycloaddition
reaction.
Keywords: cyclization · enantiose-
lectivity · metalation · sigmatropic
rearrangement · ylides
Introduction
Inspired by literature reports on base-induced inter-[6] and
intramolecular[7] additions of metalated a-amino nitriles to
unactivated alkynes, and as part of our ongoing project on
the intramolecular carbometalation of unactivated alkenes
by zinc enolates,[8] we became interested in zincated a-N-ho-
moallylamino nitriles. Zinc enolates obtained from enantio-
pure a-N-homoallylamino esters such as 1 by lithium diiso-
propylamine (LDA)-mediated deprotonation/transmetala-
tion sequence have been shown to lead to the formation of
3-zinciomethyl prolines with excellent stereoselectivity
(Scheme 1).[9–11] We reasoned that if, in a similar way, a ste-
reoselective carbozincation took place starting from zincated
a-N-homoallylamino nitriles, such as 2a, it would provide
bifunctional 2-cyano-3-(zinciomethyl)pyrrolidines bearing
Metalated a-amino nitriles are well-known nucleophiles that
react with a range of electrophiles including alkyl or acyl
halides, epoxides, aldehydes, and C C multiple bonds.
[1]
ꢀ
Lithiated a-amino nitriles, which are easily obtained by de-
protonation of the parent a-amino nitriles, have been widely
used in organic synthesis and, more specifically, in asymmet-
ric synthesis.[2] Accordingly, their structure and reactivity
have been extensively studied.[3] To a lesser degree, sodium
and potassium derived metalated a-amino nitriles have also
been considered and have proven to be synthetically useful.
In contrast, very little is known about zincated a-amino ni-
triles which, to the best of our knowledge, have almost
never been used in organic synthesis.[4,5]
[a] S. Ouizem, Dr. S. Cheramy, Dr. C. Botuha, Prof. F. Chemla,
Dr. F. Ferreira, Dr. A. Pꢀrez-Luna
UPMC Univ Paris 06, UMR CNRS 7201
Institut Parisien de Chimie Molꢀculaire
Institut de Chimie Molꢀculaire (FR 2769)
Case 183, 4 place Jussieu, 75005 Paris (France)
Fax : (+33)144-27-75-67
E-mail: fabrice.chemla@upmc.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001639.
Scheme 1. Projected approach to methanopyrrolidines from a-N-homoal-
lylamino nitriles.
12668
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12668 – 12677

FULL PAPER
Table 1. Deprotonation of 2a with LDA.
both an electrophilic a-amino nitrile moiety and a nucleo-
philic alkylzinc moiety. As for previously disclosed analo-
gous cyclizations of w-stannyl appendages on N-acyloxyimi-
nium ions,[12] we then anticipated an intramolecular displace-
ment of the cyano group through the well-known Bruylants
reaction between a-amino nitriles and organometallic nucle-
ophiles, which has been reported, in particular, with organo-
zinc reagents.[13–16] If successful, our approach would thus
offer a new entry to enantiopure 2,3-methanopyrrolidines
(Scheme 1), which are interesting compounds[12,17] that are
still challenging to prepare in spite of recent elegant ap-
proaches based on the intramolecular Kulinkovich–de Mei-
jere reaction.[18]
Entry
LDA [equiv]
Solvent
T [8C]
t [h]
Ratio[a] 2a/2a-D/4
1
2
3
4
5
2.2
2.2
2.2
2.2
1.2
1
THF
Et2O
Et2O
Et2O
Et2O
ꢀ80
ꢀ60
ꢀ40
0
1
0:100:0
0:25:75
0:50:50
0:100:0
25:75:0
1.5
0.5
1
0
1
[a] Determined by H NMR spectroscopic analysis.
Results and Discussion
product 4, resulting from the addition of LDA to the nitrile
group, was the major product observed after hydrolysis. For-
tunately, the ratio of deprotonation over addition increased
with increasing temperature (Table 1, entries 2–4) and clean
metalation was obtained by carrying out the reaction at 08C
(Table 1, entry 4). Under these conditions, 2.2 equivalents of
LDA proved necessary to achieve complete deprotonation;
when 1.2 equivalents were used, 25% of unreacted (non-
deuterated) starting material 2a was recovered (Table 1,
entry 5). We thus established two lithiation procedures:
Method A (LDA (2.2 equiv), THF, ꢀ808C, 1 h), and Meth-
od B (LDA (2.2 equiv), Et2O, 0 8C, 1 h).
Deprotonation studies: We initiated our studies by looking
at unsubstituted a-N-homoallylamino nitriles 2a and 2b as
precursors. Starting from enantiopure (R)-1-phenylethyl-
ACHTUNGTRENNUNGamine or (R)-1-(4-methoxyphenyl)ethylamine, products 2a
and 2b were obtained in 72 and 50% yield, respectively,
after a two-step sequence involving two subsequent alkyl-
ACHTUNGTRENNUNGations, first with bromobutene and then with bromoacetoni-
trile (Scheme 2).
Cyclization of a-N-homoallylamino nitriles: With the appro-
priate metalation procedures in hand, we studied the cycli-
zation reactions (Table 2). No formation of 2,3-methanopyr-
rolidine 5 was observed following treatment of amino nitrile
2a with 2.2 equivalents of LDA in Et2O at 08C and further
stirring at room temperature. In contrast, and to our delight,
addition of 4 equivalents ZnBr2 following the deprotonation
led to the formation of methanopyrrolidine 5 in 32% yield,
essentially as a single diastereoisomer (95:5 d.r.) (Table 2,
Scheme 2. Synthesis of a-N-homoallylamino nitriles 2a and 2b.
Building on our previous experience with a-amino esters,
we intended to access the zincated amino nitriles by lithia-
tion of the amino nitriles with LDA, followed by transmeta-
lation with a zinc salt. However, because difficulties in the
deprotonation of hindered a-amino nitriles either due to a
lack of reactivity or to competi-
1
entry 1). H NMR spectroscopic analysis of the crude prod-
uct revealed the presence of several side products, however,
only amine 3a, which formed in 30%, could be identified.
tive addition of the base to the
Table 2. Cyclization reactions from 2a following deprotonation with LDA.
nitrile moiety have been report-
ed,[4] we decided to first secure
the metalation procedure by
studying the product distribu-
tion obtained by deprotonation
of 2a followed by performing a
Entry
Method
ZnX2 [equiv]
T [8C]
t [h]
5 (d.r.][a] [%][b]
3a [%][c]
deuterium
Table 1).
quench
(D2O;
1
2
3
4
5
6
A[d]
B[e]
B[e]
B[e]
A[d]
ZnBr2 (4)
ZnBr2 (4)
0 to RT
1.5
1.5
2
2.5
18
4
32 (95:5)
45 (77:23)
40 (70:30)
<10 (nd)
<10
30
17
20
ꢀ80 to RT
ꢀ80 to RT
ꢀ80 to RT
0 to RT
ZnI2 (4)
Whereas complete metalation
was achieved after 1 h at
ꢀ808C in THF using 2.2 equiv-
BuZnI.LiI[f] (4)
Zn(CN)2 (4)
60
<10
[g]
–
Zn
N
0 to RT
14 (40:60)
16[h]
alents
of
LDA
(Table 1,
[a] Determined by 1H NMR spectroscopic analysis of the crude material. [b] Combined yield of isolated diaste-
reoisomers. [c] Isolated yield. [d] Method A: LDA (2.2 equiv), Et2O, 0 8C, 1 h. [e] Method B: LDA (2.2 equiv),
THF, ꢀ808C, 1 h. [f] Prepared by mixing BuLi with ZnI2. [g] Reaction carried out in THF with a 2:1 mixture
entry 1), the use of diethyl
ether as solvent was problemat-
ic. At ꢀ608C (Table 1, entry 2), of LiNiPr2/ZnBr2. [h] 24% of starting material was recovered.
Chem. Eur. J. 2010, 16, 12668 – 12677
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12669

F. Chemla et al.
In particular, no starting material was recovered. It is impor-
tant to note that, to obtain 1H NMR spectra of sufficient
clarity, the reaction mixture had to be stirred in the presence
of ethanolamine for 1 h during workup to remove zinc salts
that, otherwise, led to broadening and shifting of the signals.
Encouraged by these results and seeking improved yields,
we studied the influence of several reaction parameters. A
standard amount of zinc salt (4 equiv) was chosen because it
was noticed that lower quantities led to lower diastereose-
lectivities and had little impact on the yield and product dis-
Figure 1. ORTEP view of 7.
1
tribution. In all cases, H NMR spectroscopic analysis of the
crude reaction mixture revealed the presence of amine 3a
and several other unidentified side products. Unless other-
wise stated in Table 2, no starting material was recovered.
Conducting the reaction in THF afforded a slight increase in
yield (45%), although a significant decrease in selectivity
was observed (77:23 d.r.) (Table 2, entry 2). Whereas similar
results were observed using ZnI2 (40% yield and 70:30 d.r.)
(Table 2, entry 3), the reaction failed to produce the desired
cyclization compound when BuZnI and Zn(CN)2 were used
(Table 2, entries 4–5). Interestingly, formation of methano-
pyrrolidine 5 was observed, albeit in only 14% yield, when
Although rather disappointing in terms of yield, we never-
theless decided to explore the scope of this new cyclization
procedure. We first considered substrates bearing a substitu-
ent a to the nitrile group. Compounds 8 and 9 were readily
accessed through the Strecker reaction between amine 3a,
trimethylsilyl cyanide (TMSCN) and either ethanal or cyclo-
hexylcarboxaldehyde, respectively (Scheme 5).[20] Compound
10, which is both an a-amino ester and an a-amino nitrile,
was prepared in 72% yield by alkylation of amino nitrile 2a
with methylchloroformate (Scheme 6). In all cases, the prod-
ucts were isolated and used as mixtures of two diastereoiso-
mers.
Zn
entry 6).
Changing the stereodirecting group from a phenyl-ethyl-
ACHTUNGTRENNUNGamino to a 4-methoxyphenyl-ethylamino group had little in-
ACHTUNGTRENNUNG(NiPr2)2 was used without any additional base (Table 2,
fluence on the reaction outcome starting from 2b; methano-
pyrrolidine 6 was obtained in 28% yield and excellent selec-
tivity (d.r. >95:5) (Scheme 3).
Scheme 5. Preparation of substrates 8 and 9.
Scheme 3. Cyclization of substrate 2b.
Scheme 6. Preparation of substrate 10.
The absolute configuration of the newly created centers in
6 was determined by X-ray crystallographic analysis of its
hydrochloride salt 7 (Scheme 4, Figure 1).[19] The absolute
configuration of 5 was inferred from this result. It is worth
noting that the sense of induction brought about by the ste-
reodirecting phenylethylamino group was found to be oppo-
site to the sense observed previously for the carbocyclization
of a-N-homoallylamino esters.[9–11]
Treatment of alkyl-substituted aminonitriles 8 and 9
under the previously developed conditions resulted in the
formation of methanopyrrolidines 11 and 12 with a signifi-
cant increase in yield (75–80%) with respect to unsubstitut-
ed aminonitriles (Table 3, entries 1 and 2). Interestingly, for-
mation of the free amine 3a was nearly completely sup-
pressed. Unfortunately, the diastereoselectivity of the reac-
tion dropped, and diastereoisomeric ratios of 75:25 were ob-
served. LDA did not deprotonate substrate 10, presumably
because the acidic proton was too sterically hindered. Metal-
ation was finally effected using 2.2 equivalents of sodium hy-
dride. Following deprotonation, addition of ZnBr2 and usual
workup led to methanoproline 13 in 67% isolated yield but,
very disappointingly, only as a 55:45 diastereoisomeric mix-
Scheme 4. Hydrochloride salt formation from 6 for X-ray analysis.
12670
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12668 – 12677

Enantiopure 2,3-Methanopyrrolidines
FULL PAPER
Table 3. Cyclization from substrates 8–10 bearing a substituent a to the
CN group.
yields by the diastereoselective addition of allyltitanium to
imines 15 and 16, respectively (Scheme 8). Alkylation with
bromoacetonitrile afforded the corresponding a-amino ni-
triles, which were isolated in nearly diastereomerically pure
form (d.r. >98:2).
Entry Method Substrate
R
Product Yield (d.r.)[a] 3a
[%][b]
[%][c]
1
2
3
A[d]
A[d]
C[e]
8
9
10
Me
c-Hex
CO2Me 13
11
12
75 (75:25)
80 (75:25)
67 (55:45)[f]
10
10
<5
[a] Determined by 1H NMR spectroscopic analysis of the crude material.
[b] Combined yield of isolated diastereoisomers. [c] Isolated yield.
[d] Method A: LDA (2.2 equiv), Et2O, 0 8C, 1 h. [e] Method C: NaH
(2.2 equiv), Et2O, RT, 1 h. [f] Stereochemistry of the major product was
not established.
Scheme 8. Preparation of substrates 19 and 20.
Initially, when the deprotonation–transmetalation–cycliza-
tion procedure was carried out in Et2O starting from 19,
methanopyrrolidine 21 was obtained in excellent selectivity
(95:5 d.r.) but, again, in quite poor yield (30%; Table 4,
ture (Table 3, entry 3). Again, only small amounts of free
amine 3a were formed.
The absolute configuration of the newly created centers
of the major products were also determined by X-ray crys-
tallographic analysis of the hydrochloride salt 14 obtained
from 12 (Scheme 7, Figure 2).[21] The same sense of induc-
tion brought by the phenylethylamino group was also ob-
served in the unsubstituted case.
Table 4. Cyclization from substrates 19 and 20.
Entry
Method
Substrate
R
Product
Yield (d.r.)[a] [%]
1
2
3
A[b]
B[d]
B[d]
19
19
20
Et
Et
iPr
21
21
22
30[c] (95:5)
66[c] (96:4)
71[e] (95:5)
[a] Determined by 1H NMR spectroscopic analysis of the crude material.
[b] Method A: LDA (2.2 equiv), Et2O, 0 8C, 1 h. [c] Combined yield of
isolated isomers. [d] Method B: LDA (1.2 equiv), THF, ꢀ808C, 1 h.
[e] Yield determined by 1H NMR spectroscopy based on analysis of the
crude mixture with biphenyl as internal standard. Problematic separation
from side products resulted in 43% isolated yield.
Scheme 7. Hydrochloride salt formation from 12 for X-ray analysis.
entry 1). Fortunately, in this case, changing the solvent to
THF proved to be highly beneficial; under these conditions
the yield was increased to 66% without loss of diastereose-
lectivity (Table 4, entry 2). With substrate 20, which bears an
isopropyl group, a similar 71% yield was obtained, again
with a 95:5 diastereomeric ratio (Table 4, entry 3).
As described previously, to determine the absolute config-
uration of the newly created centers of 21, X-ray analysis of
its hydrochloride salt 23 was performed (Scheme 9,
Figure 3).[23] The same sense of induction observed for the
unsubstituted case was also found with regard to the (R)-1-
phenylethylamino stereodirecting group but, to our surprise,
Figure 2. ORTEP view of 14.
Cyclization of a-N-homoallylamino nitriles bearing a sub-
stituent in the homoallylic position: We next investigated
the influence of a substituent in the homoallylic position.
Following a procedure reported by Gao and Sato,[22] enan-
tiomerically pure amines 17 and 18 were prepared in good
Scheme 9. Hydrochloride salt formation from 21 for X-ray analysis.
Chem. Eur. J. 2010, 16, 12668 – 12677
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12671

F. Chemla et al.
methanopyrrolidine 25 showed a trans relative configuration
between the ethyl and the cyclopropyl groups, implying a
relative configuration between the methyl group of the (R)-
1-phenylethylamino moiety and the cyclopropyl group oppo-
site to that observed for 21. The structure of 25 was further
secured by removal of the nitrogen protecting group, be-
cause, in so doing, two enantiomeric compounds (26 and
ent-26) were obtained starting, respectively, from 21 and 25
(Scheme 12). Thus, in conclusion, the reaction of phenyl-
Figure 3. ORTEP view of 23.
an inversion of the homoallylic stereogenic center was de-
tected. Furthermore, when these results are compared with
the analogous carbocyclization of N-homoallyl a-amino
esters,[11] it can be seen that not only is the sense of induc-
tion of the stereodirecting group opposite, but the diastereo-
selectivity is also much higher (96:4 here versus 40:60, in the
case of proline formation) and the relative configuration be-
tween the substituents of the pyrrolidine ring is opposite
(trans here versus cis for the major carbocyclization prod-
uct).
Scheme 12. Stereochemical correlation for products 21 and 25.
So that the relative influence of each stereogenic center
during the cyclization could be determined, we decided to
prepare a-amino nitrile 24, having a relative configuration
opposite to that of 19 (Scheme 10). Allylation of imine 15
AHCTUNGTREGeNNUN thylamino-derived N-homoallylamino nitriles substituted at
the homoallylic position is stereospecific; it proceeds with
inversion of configuration of the homoallylic center and af-
fords methanopyrrolidines with a trans relationship between
the cyclopropyl and the homoallylic substituents.
Mechanistic considerations: The unexpected stereochemical
outcome of the cyclization reactions involving a-N-homoal-
lylamino nitriles substituted at the homoallylic position led
us to rule out a mechanistic rationale involving a carbocycli-
zation/intramolecular Bruylants reaction (see path A,
Scheme 14 below), because this route could not account for
the stereospecific inversion of the homoallylic center.
Scheme 10. Preparation of substrate 24.
with allyl Grignard followed by alkylation of the resulting
mixture of amines with bromoacetonitrile led to a 41:59
mixture of amino nitriles 19 and 24. Careful separation by
silica-gel flash chromatography allowed compound 24 to be
isolated in nearly diastereomerically pure form (94:6 d.r.).
With compound 24 in hand, the cyclization procedure was
carried out and methanopyrrolidine 25 was obtained in 36%
yield in practically diastereomerically pure form (98:02 d.r.;
Scheme 11). As evidenced by their different NMR spectra,
two diastereomeric methanopyrrolidines were obtained
starting from 19 and 24. NOE experiments carried out with
Because a-amino nitriles are potential iminium ion pre-
cursors in the presence of Lewis acids,[24] we decided to
make sure that metalation was, as anticipated, the first step
of the cyclization sequence leading to methanopyrrolidine
formation. We thus carried out a test reaction in which zinc
bromide was added prior to addition of LDA (Scheme 13).
No cyclization was detected. Along with 47% of unreacted
starting material, the only products isolated were compound
4, resulting from the addition of LDA to 2a, and small
amounts of amine 3a. From a mechanistic standpoint, this
Scheme 11. Cyclization from substrate 24.
Scheme 13. Reaction with 2a, reversing the reagent addition order.
12672
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12668 – 12677

Enantiopure 2,3-Methanopyrrolidines
FULL PAPER
result suggests that methanopyrrolidine formation only
arises from the metalated amino nitrile, whereas formation
of amine 3a can occur by reaction of the starting amino ni-
trile with the zinc salt (iminium formation and hydrolysis).
Accordingly, starting from lithiated amino nitriles 27
(Scheme 14; for the sake of clarity only the formation of
cloaddition[36] would then afford the observed methanopyr-
rolidines. Formation of an azomethine ylide from 30, which
would then undergo cycloaddition,[37] cannot be ruled out
completely in presence of halides (or cyanide).[38] However,
this possibility is considered unlikely because intramolecular
dipolar cycloadditions of azomethine ylides bearing such a
short tether have not previously been reported, to the best
of our knowledge.[37] Thus, we prefer to consider a more
plausible sequence involving the addition of the pendant
alkene onto the iminium moiety[34a,39] of 30 to afford 31, fol-
lowed by anionic annulation[40] by reaction between the or-
ganozinc and the formed cationic center (path D,
Scheme 14).
The stereochemical outcome of the cyclization reactions
can be rationalized on the basis of this mechanism. The ste-
reospecific inversion of the homoallylic center observed in
the cyclization procedure can only be explained if a chirality
transfer takes place from 29 to the zincated azomethine
imine ylide 30, and if the latter is configurationally stable
during the time scale of the reaction.[41] The configuration of
the (2-azoniaallyl)zinc intermediate depends both on the
configuration of the zincioiminium ion and on the preferred
six-center chair-like transition state for the aza-Cope rear-
rangement.[42,43] Furthermore, according to Polniaszek and
co-workersꢂ studies[44] on additions to N-acyliminum ions
bearing an a-methylbenzyl group,[45] in the most reactive
conformation the phenyl group is perpendicular to the plane
of the C=N double bond to maximize the interaction be-
ꢀ
tween the s* of the C Ph bond and the p* of the p system.
Scheme 14. Mechanistic rationales for methanopyrrolidine formation
from lithiated a-N-homoallylamino nitriles.
As a consequence, two possible chair-like diastereomeric
transition states for the aza-Cope rearrangement are to be
considered for a given configuration of the reacting zincio-
methanopyrrolidines 5, 21, and 22 are depicted), one can en-
vision that addition of the zinc salt will first result in a trans-
metalation to afford the zincated derivative 28 (probably C-
metalated as observed for other zincated nitriles).[25,26] The
Lewis acidic salt can then assist the departure of the cyano
group to lead to the metalated iminium salt 29, which is a
species that has precedents in literature.[27–29] From this
point, several possibilities can be considered. First, the
postulated zincioiminium ion 29 (or the parent zincated
a-amino nitrile) is a carbenoid species. Because both in-
tra-[27e,30] and intermolecular[27a–d,31] cyclopropanations of a-
amino carbenes or carbenoids have been reported, trapping
of a carbenoid or carbene intermediate by the pendant ho-
moallylic chain is a possible pathway that may account for
the formation of methanopyrrolidines (paths B or C,
Scheme 14). However, as in the case of the carbocyclization
mechanism, this possibility can be excluded because it
cannot explain the stereospecific inversion of the homoallyl-
ic center. Nevertheless, this highly reactive intermediate
might be involved in degradation pathways and side reac-
tions and, hence, could lead to the observed moderate
yields.
ACHTUNGTRENNUNG
lylic substituent, the observed product arises from a path in-
volving the aza-Cope rearrangement of the (E)-zincioimini-
um ion via TS1a. A probable explanation is that formation
of the (E)-zincioiminium ion is favored over that of the Z
isomer to minimize steric repulsive interactions between the
zinc moiety and the bulkier a-methylbenzyl group.[33d] Tran-
sition-state TS1a should be lower in energy than TS1b as a
consequence of a more favorable allylic interaction between
the iminium proton and the hydrogen (in TS1a) compared
with the methyl group of the a-methylbenzyl moiety (for
TS1b).[46]
Through TS1a, 3,3-sigmatropic rearrangement leads to
ylide 30a (Scheme 15). Although no previous reports on the
reactivity of (2-azoniaallyl)zinc species have been disclosed,
Alternatively, the zincioiminium intermediate could un-
dergo a 2-aza-Cope[32–34] rearrangement and thus lead to the
(2-azoniaallyl)zinc species 30.[35] A [3+2] intramolecular cy-
Chem. Eur. J. 2010, 16, 12668 – 12677
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12673

F. Chemla et al.
Scheme 15. Proposed rationale for the stereochemical outcome of the
synthesis of 5.
we hypothesize that the subsequent alkene addition follows
Lindermanꢂs model[47] for nucleophilic additions to a-silyl-
and stannyl oxocarbenium ions, which has also been used to
rationalize the behavior of [3+2] cycloadditions involving
(2-azoniaallyl)stannanes.[38] Hence, cyclization to 31a should
take place via transition-state TS2, wherein the iminium
adopts a favored conformation involving a hyperconjugative
ꢀ
interaction between the Zn C bond and the neighboring
electron-deficient iminium p system.[48] From 31a, methano-
pyrrolidine 5 can then be produced by anionic cyclization,
resulting in cyclopropane formation. For reaction of the car-
Scheme 16. Proposed rationale for the stereochemical outcome of the
synthesis of 21 and 25.
ꢀ
banionic center with the carbocation to occur, the Zn C
bond and the p orbital of the cation must adopt a W confor-
mation to enable back-lobe orbital overlap (Scheme 15).[49]
This requirement accounts for the stereoselective formation
of the cyclopropyl group, provided that the organozinc inter-
mediate 31a is configurationally stable on the reaction time
scale.
the timescale of the reaction. Intramolecular cyclopropana-
tion involving a W conformation then stereoselectively af-
fords methanopyrrolidine 21 or 25 (Scheme 16).
It should be noted that the stereochemical outcome of the
reactions starting from substrates 8 and 9, bearing a sub-
stituent a to the nitrile, is also consistent with this picture.
Indeed, the formation of the E isomer of the zincioiminium
should be less selective as a consequence of a lower differ-
ence in bulk between the zinc atom and the alkyl group.
This might be one of the reasons (though perhaps not the
only one) for the overall loss of stereoselectivity noted.
In the case of substrates bearing a substituent in the ho-
moallylic position, such as 19 and 24, there is no clear differ-
ence of bulk between the nitrogen substituents (a-(ethyl)ho-
moallyl versus a-methylbenzyl), so one would also expect
iminium ion formation to be less selective. The excellent se-
lectivity nevertheless obtained must then arise from the ste-
reoselectivity of the aza-Cope rearrangement. Indeed, the
observed outcome is consistent with preferred transition
states TS3 and TS4 wherein both the homoallylic substituent
and the zinc moiety adopt less hindered pseudo-equatorial
positions. Based on Lindermanꢂs model, 5-endo cyclization
of 30b or 30c then occurs through TS5 or TS6, respectively,
to produce 31b or 31c, which is configurationally stable on
Conclusion
We have developed a new stereoselective cyclization reac-
tion that provides enantiopure 2,3-methanopyrrolidines
from zincated a-N-homoallylamino nitriles in moderate to
good yields. In addition to the synthetic potential offered by
this new approach, the analysis of the stereochemical out-
come of the reaction has provided evidence for an unexpect-
ed mechanistic behavior. To account for the stereospecific
inversion of the homoallylic stereogenic center for sub-
strates bearing a substituted homoallylic side chain, we pro-
12674
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12668 – 12677

Enantiopure 2,3-Methanopyrrolidines
FULL PAPER
lution was cooled to ꢀ808C. A solution of 19 (242 mg, 1 mmol) in anhy-
drous THF (1.5 mL) was added and, after 1 h of stirring at ꢀ808C, ZnBr2
(1.0m in anhydrous THF, 4.0 mL, 4.0 mmol) was added dropwise, and the
cooling bath was removed and replaced by a water bath. Once the solu-
tion reached RT, the bath was removed and the mixture was stirred at
RT for 2 h. An aqueous solution of NH4Cl/NH3 (2:1; 10 mL) and etha-
nolamine (1 mL) were added. The biphasic mixture was stirred vigorous-
ly for 30 min before the aqueous layer was extracted with ethyl acetate
(3ꢃ10 mL). The combined organic layers were washed with brine, dried
over MgSO4, filtered, and concentrated. Purification by flash chromatog-
raphy afforded 21 as a yellow oil (142 mg, 66%, d.r.=96:4). Further pu-
rification by flash chromatography afforded a stereochemically pure
sample. [a]2D0 =ꢀ100.7 (c=1.03 in chloroform); 1H NMR (400 MHz,
CDCl3, 258C, TMS): d=7.41 (d, J=7.0 Hz, 2H), 7.32–7.21 (m, 3H), 3.64
(q, J=6.6 Hz, 1H), 2.75 (dt, J=5.9, 2.8 Hz, 1H), 2.36–2.29 (m, 1H),
2.11–2.07 (m, 1H), 1.70 (ddd, J=12.5, 9.4, 5.3 Hz, 1H), 1.52 (d, J=
6.8 Hz, 3H), 1.33–1.27 (m, 1H), 1.11–1.01 (m, 1H), 0.96–0.85 (m, 1H),
0.67 (t, J=7.4 Hz, 3H), 0.69–0.65 (m, 1H), 0.21 ppm (dt, J=5.5, 8.4 Hz,
1H); 13C NMR (100 MHz, CDCl3, 258C, TMS): d=145.9, 128.2, 127.7,
126.8, 61.1, 60.5, 39.5, 34.5, 27.8, 19.0, 12.7, 10.6, 5.5 ppm; IR (neat): n˜ =
2958, 1638, 1492, 1454, 1112, 952, 701 cmꢀ1; HRMS (ESI): m/z calcd for
C15H22N: 216.17468 [M+ꢀH]; found: 216.17420.
pose that zincated a-amino nitriles are new zincioiminium
ion precursors. We believe that this approach will foster new
developments in the chemistry of zincated iminiums. In the
particular case of a-N-homoallylamino nitriles, a subsequent
aza-Cope rearrangement takes place to afford previously
unreported (2-azoniaallyl)zinc intermediates,[50] which lead
to 2,3-methanopyrrolidines after a [3+2] cycloaddition pro-
cess. Remarkable features of this mechanism include a chir-
ality transfer through an asymmetric aza-Cope rearrange-
ment, which is rather uncommon,[51] and the formation of a
metalated azomethine-ylide that is configurationally stable
on the timescale of the subsequent reaction. Additional
mechanistic studies are underway in our group to fully un-
derstand the [3+2] process and will be reported in due
course.
Experimental Section
General information: Experiments involving organometallic compounds
were carried out in dried glassware under a positive pressure of dry N2.
Liquid nitrogen was used as a cryoscopic fluid. A three-necked, round-
bottomed flask equipped with an internal thermometer, a septum cap
and a nitrogen inlet was used. Anhydrous solvents were distilled to
remove stabilizers and dried with a double column purification system.
Zinc bromide (98%) was melted under dry N2 and, immediately after
cooling to RT, was dissolved in anhydrous Et2O or THF. All other re-
agents and solvents were of commercial quality and were used without
further purification. 1H and 13C NMR spectra were recorded with a
Bruker AVANCE 400 spectrometer fitted with a BBFO probe or with a
Bruker ARX 200 spectrometer fitted with a dual probe (13C/1H). Chemi-
cal shifts are reported in d units relative to an internal standard of residu-
al chloroform (d=7.27 ppm for 1H NMR spectra and d=77.16 ppm for
13C NMR spectra). IR spectra were recorded with a diamond ATR spec-
trometer. High-resolution mass spectra (HRMS) were obtained with a
Finnigan MAT 95 instrument. Elemental analyses were performed at the
Service de Microanalyses de lꢂUniversitꢀ Pierre et Marie Curie - Bat F -
case 55–4 place Jussieu - 75252 Paris Cedex 05.
Acknowledgements
The authors wish to thank Patrick Herson and Kamal Boubekeur for X-
ray structure determination and the MESR for grants (S.O. and S.C.).
[1] T. Opatz, Synthesis 2009, 1941–1960.
[2] Selected reviews : a) J. D. Albright, Tetrahedron 1983, 39, 3207–
3233; b) H.-P. Husson, J. Royer, Chem. Soc. Rev. 1999, 28, 383–394;
c) D. Enders, J. P. Shilvock, Chem. Soc. Rev. 2000, 29, 359–373.
[3] For leading references, see: D. Enders, J. Kirchhoff, P. Gerdes, D.
Mannes, G. Raabe, J. Runsink, G. Boche, M. Marsch, H. Ahlbrecht,
H. Sommer, Eur. J. Org. Chem. 1998, 63–72, and references therein.
[4] An example of addition of a zincated dibenzylaminoacetonitrile to
an aldehyde has been reported, see: E. Leclerc, E. Vrancken, P.
Mangeney, J. Org. Chem. 2002, 67, 8928–8937.
[5] Alteration of the reactivity of lithiated b,g-unsaturated a-dialkylami-
nonitriles by addition of ZnCl2 has been reported, see: R. M. Jacob-
son, J. W. Clader, Tetrahedron Lett. 1980, 21, 1205–1208.
[6] a) A. Jonczyk, D. Lipiak, T. Zdrojewski, Tetrahedron 1990, 46,
1025–1038; b) T. Zdrojewski, A. Jonczyk, Synthesis 1990, 224–233;
c) T. Zdrojewski, A. Jonczyk, Liebigs Ann. Chem. 1993, 375–378;
d) A. Jonczyk, Z. Pakulski, Tetrahedron Lett. 1996, 37, 8909–8912.
[7] J.-L. Wang, C.-H. Ueng, M.-C. P. Yeh, Tetrahedron Lett. 1995, 36,
2823–2826.
[8] A. Pꢀrez-Luna, C. Botuha, F. Ferreira, F. Chemla, New J. Chem.
2008, 32, 594–606.
[9] E. Lorthiois, I. Marek, J.-F. Normant, Tetrahedron Lett. 1997, 38,
89–92.
[10] P. Karoyan, G. Chassaing, Tetrahedron Lett. 1997, 38, 85–88.
[11] E. Lorthiois, I. Marek, J.-F. Normant, J. Org. Chem. 1998, 63, 2442–
2450.
[12] S. Hanessian, L. Azzas, Acc. Chem. Res. 2008, 41, 1241–1251, and
references therein.
Cyclization of a-aminonitriles (Method A): Typical procedure (Com-
pound 5; Table 2, entry 1): Diisopropylamine (0.31 mL, 2.2 mmol) was
added at RT to nBuLi (0.94 mL, 2.2 mmol). Once a gummy mixture
formed, anhydrous Et2O (0.5 mL) was added and the solution was cooled
to 08C. A solution of 2a (214 mg, 1.0 mmol) in anhydrous Et2O (4 mL)
was added dropwise. After 1 h of stirring at 08C, ZnBr2 (1.0m in Et2O,
4.0 mL, 4.0 mmol) was added in one portion. The reaction mixture was
allowed to warm to RT and stirred for 2 h. An aqueous solution of
NH4Cl/NH3 (2:1; 10 mL) and ethanolamine (1 mL) were added. The bi-
phasic mixture was stirred vigorously for 30 min before the aqueous layer
was extracted with Et2O (3ꢃ10 mL). The combined organic layers were
washed with brine, dried over MgSO4, filtered, and concentrated in
vacuo. Purification by flash chromatography afforded 5 as a yellow oil
(60 mg, 32%, d.r.=95:05). Further purification by flash chromatography
afforded a stereochemically pure sample. [a]2D0 =+6.9 (c=0.94 in chloro-
form); 1H NMR (200 MHz, CDCl3, 258C, TMS): d=7.32–7.20 (m, 5H),
3.24 (q, J=6.6 Hz, 1H), 2.91 (dt, J=5.9, 2.9 Hz, 1H), 2.58–2.48 (m, 1H),
2.00–1.84 (m, 1H), 1.81–1.62 (m, 2H), 1.49 (d, J=6.4 Hz, 3H), 1.46–1.32
(m, 1H), 0.75–0.69 (m, 1H), 0.13 ppm (dt, J=8.4, 5.9 Hz, 1H); 13C NMR
(50 MHz, CDCl3, 258C, TMS): d=144.3, 126.8, 125.7, 125.4, 62.8, 46.7,
37.7, 25.3, 22.3, 13.9, 0.0 ppm; IR (neat): n˜ =2931, 1947 cmꢀ1; HRMS
(ESI): m/z calcd for C13H18N: 188.14338 [M+ꢀH]; found: 188.14320.
[13] E. Le Gall, C. Gosmini, M. Troupel, Tetrahedron Lett. 2006, 47,
455–458.
[14] L. Bernardi, B. F. Bonini, E. Capitꢄ, G. Dessole, M. Fochi, M.
Comes-Franchini, A. Ricci, Synlett 2003, 1778–1782.
[15] C. Agami, F. Couty, G. Evano, Org. Lett. 2000, 2, 2085–2088.
[16] C. Schneider, C. Bçrner, Synlett 1998, 652–654.
[17] F. Brackmann, A. de Meijere, Chem. Rev. 2007, 107, 4538–4583.
[18] Selected references: a) J. Lee, J. K. Cha, J. Org. Chem. 1997, 62,
1584–1585; b) G.-D. Tebben, K. Rauch, C. Stratmann, C. M. Wil-
Cyclization of a-aminonitriles (Method B): Typical procedure (Com-
pound 21; Table 4, entry 2): Diisopropylamine (0.17 mL, 1.2 mmol) was
added at RT to nBuLi (2.4m in hexane, 0.50 mL, 1.2 mmol). Once a
gummy mixture formed, anhydrous THF (1.5 mL) was added and the so-
Chem. Eur. J. 2010, 16, 12668 – 12677
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12675

F. Chemla et al.
liams, A. de Meijere, Org. Lett. 2003, 5, 483–485; c) N. Ouhamou,
Y. Six, Org. Biomol. Chem. 2003, 1, 3007–3009; d) C. Madelaine, N.
Ouhamou, A. Chiaroni, E. Vedrenne, L. Grimaud, Y. Six, Tetrahe-
dron 2008, 64, 8878–8898; e) C. Madelaine, A. K. Buzas, J. Kowal-
ska-Six, Y. Six, B. Crousse, Tetrahedron Lett. 2009, 50, 5367–5371.
[19] X-ray crystal data of compound 7: C14H20ClNO; Mw =253.77; color-
less parallelepiped crystals; crystal size 0.30ꢃ0.35ꢃ0.45 mm; ortho-
rhombic; space group P212121; Z=4; a=9.857(4), b=11.006(7), c=
12.942(5) ꢅ, a=90.0008, b=90.000, g=90.0008; V=1404(1) ꢅ3;
[28] Lithioiminiums: M. Buswell, I. Fleming, U. Ghosh, S. Mack, M. Rus-
sell, B. P. Clark, Org. Biomol. Chem. 2004, 2, 3006–3017.
[29] For selected reviews on related a-azametalo-carbenoids, see: a) L. S.
Hegedus, Tetrahedron 1997, 53, 4105–4128; b) T. S. Powers, W. D.
Wulff, J. Quinn, Y. Shi, W. Q. Jiang, R. P. Hsung, M, Parisi, A
Rahm, X. W. Jiang, G. A. Yap, A. L. Rheingold, J. Organomet.
Chem. 2001, 617, 182–208.
[30] Selected references: a) C. P. Casey, A. J. Shusterman, N. W. Vollen-
dorf, K. J. Haller, J. Am. Chem. Soc. 1982, 104, 2417–2423; b) C. P.
Casey, N. W. Vollendorf, K. J. Haller, J. Am. Chem. Soc. 1984, 106,
3754–3764; c) M. J. McGeary, T. L. Tonker, J. L. Templeton, Orga-
nometallics 1985, 4, 2102–2106; d) M. Brookhart, W. B. Studabaker,
Chem. Rev. 1987, 87, 411–432; e) A. Ogawa, N. Takami, M. Sekigu-
chi, I. Ryu, N. Kambe, N. Sonoda, J. Am. Chem. Soc. 1992, 114,
8729–8730; f) S. Couty, C. Meyer, J. Cossy, Synlett 2007, 2819;
g) Z. F. Al-Rashid, R. P. Hsung, Org. Lett. 2008, 10, 661–663.
[31] Selected references: a) U. Schçllkopf, M. Hauptreif, J. Dippel, M.
Nieger, E. Egert, Angew. Chem. 1986, 98, 187–189; Angew. Chem.
Int. Ed. Engl. 1986, 25, 192–193; b) A. Ogawa, N. Takami, T.
Nanke, S. Ohya, T. Hirao, Tetrahedron 1997, 53, 12895–12902; c) I.
Merino, L. S. Hegedus, Organometallics 1995, 14, 2522–2531; d) J.
Barluenga, F. Aznar, I. Gutiꢀrrez, S. Garcꢆa-Granda, M. A. Llorca-
BaragaÇo, Org. Lett. 2002, 4, 4273–4276; e) R. P. Wurz, A. B. Charr-
ette, J. Org. Chem. 2004, 69, 1262–1269.
1calcd =1.20 gcmꢀ3
; ; Enraf
l=0.710690 ꢅ (MoKa); m=2.57 cmꢀ1
Nonius KAPPA CCD diffractometer; temperature 295 K; q range
1.0–25.08; min/max h=0/11, min/max k=0/13, min/max l=0/15;
1453 collected reflections, 1432 independent, 1225 used, 156 parame-
ters; R=0.0458 (R=SjjF0 jꢀjFc jj/SjF0 j), Rw=0.0499 (Rw=
[Sw(jjF0jꢀjFc jj)2/SwF02]1/2), goodness of fit=0.990, max/min D1=
0.38/ꢀ0.22 eꢅꢀ3. CCDC-777553 contains the supplementary crystal-
lographic data for this paper. These data can be obtained free of
charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[20] K. Mai, G. Patil, Tetrahedron Lett. 1984, 25, 4583–4586.
[21] X-ray crystal data of compound 14: C19H28ClN; Mw =305.89; color-
less blocks; crystal size 0.30ꢃ0.35ꢃ0.45 mm; orthorhombic; space
group P212121; Z=4; a=9.870(5), b=10.888(7), c=15.929(8) ꢅ, a=
90.000, b=90.000, g=90.0008; V=1712(2) ꢅ3; 1calcd =1.19 gcmꢀ3
;
l=0.710690 ꢅ (MoKa); m=2.18 cmꢀ1; Enraf–Nonius Cad-4 diffrac-
tometer; temperature 295 K; q range 1.0–25.08; min/max h=0/11,
min/max k=0/12, min/max l=0/18; 1755 collected reflections, 1734
independent, 1321 used, 192 parameters; R=0.0485 (R=SjjF0 jꢀ
jFc jj/SjF0 j), Rw=0.0552 (Rw=[Sw(jjF0jꢀjFc jj)2/SwF02]1/2), good-
ness of fit=1.042, max/min D1=0.31/ꢀ0.19 eꢅꢀ3. CCDC-777551
contains the supplementary crystallographic data for this paper.
These data can be obtained free of charge from The Cambridge
Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/
cif.
[32] For the first 2-azonia-Cope rearrangement, see: R. M. Horowitz,
T. A. Geissman, J. Am. Chem. Soc. 1950, 72, 1518.
[33] For selected recent examples, see: a) M. Rueping, A. Antonchick,
Angew. Chem. Int. Ed. 2008, 47, 10090–10093; b) M. Sugiura, C.
Mori, S. Kobayashi, J. Am. Chem. Soc. 2006, 128, 11038–11039; c) S.
Stas, K. A. Tehrani, G. Laus, Tetrahedron 2008, 64, 3457–3463;
d) B. F. Johnson, E. L. Marrero, W. A. Turley, H. A. Lindsay, Synlett
2007, 893–896.
[34] For reviews of the aza-Cope–Mannich rearrangement cyclization
and related reactions, see: a) L. E. Overman, Aldrichimica Acta
1995, 28, 107–120; b) L. Overman, Acc. Chem. Res. 1992, 25, 352–
359; c) L. Overman, D. Ricca in Comprehensive Organic Synthesis
Vol. 8 (Ed.: B. M. Trost), Pergamon, New York, 1991, pp. 1007–
1046; d) J. Royer, M. Bonin, L. Micouin, Chem. Rev. 2004, 104,
2311–2352.
[35] For an example of isomerization of a carbene to a nonstabilized azo-
methine ylide, see: X. Cattoꢇn, S. Solꢀ, C. Pradel, H. Gornitzka, K.
Miqueu, D. Bourrissou, G. Bertrand, J. Org. Chem. 2003, 68, 911–
914.
[36] Pyrrolidine ring formation by inter- and intramolecular [3+2] cyclo-
additions between (2-azaallyl)lithium and nonactivated alkenes is
well-documented; for selected reviews, see: a) W. H. Pearson, Pure
Appl. Chem. 2002, 74, 1339–1347; b) W. H. Pearson, P. Stoy, Synlett
2003, 0903–0921.
[37] Select reviews: a) L. M. Harwood, R. J. Vickers in Synthetic Applica-
tions of 1,3-Dipolar Cycloaddition Chemistry Toward Heterocycles
and Natural Products (Eds.: A. Padwa, W. H. Pearson), Wiley, New
York, 2003, Chapter 3, p. 169; b) C. Najera, J. M. Sansano, Curr.
Org. Chem. 2003, 7, 1105–1150; c) I. Coldham, R. Hufton, Chem.
Rev. 2005, 105, 2765–2809.
[38] For selected references on azomethine ylide formation from (2-azo-
niaallyl)stannanes or (2-azoniaallyl)silanes, see: a) M. T. Epperson,
D. Y. Gin, Angew. Chem. 2002, 114, 1856–1858; Angew. Chem. Int.
Ed. 2002, 41, 1778–1780; b) W. H. Pearson, P. Stoy, Y. Mi, J. Org.
Chem. 2004, 69, 1919–1939; c) G. Bꢀlanger, V. Darsigny, M. Dorꢀ,
F. Lꢀvesque, Org. Lett. 2010, 12, 1396–1399, and references therein.
[39] For a recent example of a 6-endo cyclization of an iminium onto an
alkene of an organometallic species, see: C. Kim, H. J. Bae, J. H.
Lee, W. Jeong, H. Kim, V. Sampath, Y. H. Rhee, J. Am. Chem. Soc.
2009, 131, 14660–14661.
[22] Y. Gao, F. Sato, J. Org. Chem. 1995, 60, 8136–8137.
[23] X-ray crystal data of compound 23·H2O: C15H24ClNO; Mw =269.82;
colorless sticks; crystal size 0.20ꢃ0.20ꢃ0.40 mm; orthorhombic;
space group P212121; Z=4; a=7.470(6), b=8.590(7), c=
25.586(16) ꢅ, a=90.000, b=90.000, g=90.0008; V=1642(2) ꢅ3;
1calcd =1.09 gcmꢀ3
;
l=0.710690 ꢅ (MoKa); m=2.24 cmꢀ1
;
KAPPA
type diffractometer; temperature 295 K; q range 1.0–25.08; min/max
h=0/8, min/max k=0/10, min/max l=0/30; 1719 collected reflec-
tions, 1693 independent, 1110 used, 173 parameters; R=0.0729 (R=
SjjF0 jꢀjFc jj/SjF0 j), Rw=0.0782 (Rw=[Sw(jjF0jꢀjFc jj)2/SwF02]1/2),
goodness of fit=0.888, max/min D1=0.60/ꢀ0.31 eꢅꢀ3
. CCDC-
777552 contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif.
[24] For an example of iminium formation in presence of ZnBr2, see:
J. E. Baldwin, D. R. Spring, R. C. Whitehead, Tetrahedron Lett.
1998, 39, 5417–5420.
[25] a) H. Brombacher, H. Vahrenkamp, Inorg. Chem. 2004, 43, 6054–
6060; b) F. Orsini, Synthesis 1985, 500–503; c) N. Goasdouꢀ, M.
Gaudemar, J. Organomet. Chem. 1972, 39, 17–27.
[26] For a discussion on the structure of metalated nitriles as a function
of the metal, see: F. F. Fleming, S. Gudipati, Eur. J. Org. Chem.
2008, 5365–5374.
[27] Zincioiminiums: a) G. Bꢀgis, D. E. Cladingboel, W. B. Motherwell,
Chem. Commun. 2003, 2656–2657; b) G. Bꢀgis, D. E. Cladingboel,
W. B. Motherwell, T. D. Sheppard, D. A. Tocher, Synthesis 2005,
3186–3188; c) W. B. Motherwell, G. Bꢀgis, D. E. Cladingboel, L.
Jerome, T. D. Sheppard, Tetrahedron 2007, 63, 6462–6476; d) G.
Bꢀgis, D. E. Cladingboel, L. Jerome, W. B. Motherwell, T. D. Shep-
pard, Eur. J. Org. Chem. 2009, 1532–1538; e) L. Jerome, T. D. Shep-
pard, A. E. Aliev, W. B. Motherwell, Tetrahedron Lett. 2009, 50,
3709–3712; f) A. Serieys, C. Botuha, F. Chemla, F. Ferreira, A.
Pꢀrez-Luna, Tetrahedron Lett. 2008, 49, 5322–5323.
[40] Ring contraction of lithio N-Boc-4-chloropiperidine to 2-azabicylo-
AHCTUNGTREG[NNUN 3.1.0]hexane has precedent, see: a) P. Beak, S. Wu, E. K. Yum,
Y. M. Jun, J. Org. Chem. 1994, 59, 276–277; b) Y. S. Park, P. Beak,
Tetrahedron 1996, 52, 12333–12350.
12676
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 12668 – 12677

Enantiopure 2,3-Methanopyrrolidines
FULL PAPER
ꢀ
[41] C C single bond rotation in acyclic N-homoallyl iminium ions was
2453; c) J. Cossrow, S. D. Rychnovsky, Org. Lett. 2002, 4, 147–150;
d) S. D. Rychnovsky, J. Cossrow, Org. Lett. 2003, 5, 2367–2370.
[48] Moreover this stereochemical outcome is also consistent with that of
intramolecular cyclizations of vinylsilanes onto a-cyanoenamines
leading to 2,6-disubstituted piperidines, see: a) P. Castro, L. E. Over-
man , X. Zhang, P. S. Mariano, Tetrahedron Lett. 1993, 34, 5243–
5246; b) T.-K. Yang, T.-F. Teng, J.-H. Lin, Y.-Y. Lay, Tetrahedron
Lett. 1994, 35, 3581–3582.
shown to be even faster than intramolecular Mannich cyclization be-
tween two oriented reactive donor and acceptor, see: E. J. Jacobsen,
J. Levin, L. E. Overman, J. Am. Chem. Soc. 1988, 110, 4329–4336.
[42] a) R. J. Doedens, G. P. Meier, L. E. Overman, J. Org. Chem. 1988,
53, 685–690; b) H. Hiemstra, W. N. Speckamp in Comprehensive
Organic Synthesis, Vol. 4 (Eds.: B. M. Trost, I. Fleming), Pergamon,
Oxford, 1991, pp. 1047–1082.
[49] For a discussion on the conformational requirements for cyclopro-
pane formation by internal nucleophilic substitution, see: a) D. Ber-
uben, I. Marek, J. F. Normant, N. Platzer, J. Org. Chem. 1995, 60,
2488–2501; b) F. Ferreira, C. Herse, E. Riguet, J. F. Normant, Tetra-
hedron Lett. 2000, 41, 1733–1736.
[50] In contrast, (2-azaallyl)zinc intermediates have a precedent, see: V.
Lyaskovskyy, R. Frçlich, E.-U. Wꢉrthwein, Chem. Eur. J. 2007, 13,
3113–3119.
[51] For other selected examples of 2-aza-Cope rearrangements with no
epimerization, see: a) P. G. M. Wuts, Y. W. Jung, J. Org. Chem. 1988,
53, 1957–1965; b) H.-S. Cheng, A.-H. Seow, T.-P. Loh, Org. Lett.
2008, 10, 2805–2807, and reference [33b].
[43] M. Lukowski, K. Jacobs, P. Hsueh, H. A. Lindsay, M. C. Milletti, Tet-
rahedron 2009, 65, 10311–10316.
[44] R. P. Polniaszek, S. E. Belmont, R. Alvarez, J. Org. Chem. 1990, 55,
215–223.
[45] For a general review on the use of a-phenylethylamine as a chiral
reagent in the stereodifferentiation of prochiral substrates, see: E.
Juaristi, J. L. Leꢈn-Romo, A. Reyes, J. Escalante, Tetrahedron:
Asymmetry 1999, 10, 2441–2495.
[46] The sense of asymmetric induction is the sense generally observed
for 1,3-stereoinduction for chiral imines derived from phenylethyl-
amine, see reference [45].
[47] a) R. J. Linderman, T. V. Anklekar, J. Org. Chem. 1992, 57, 5078–
5080; b) R. J. Linderman, K. Chen, J. Org. Chem. 1996, 61, 2441–
Received: June 10, 2010
Published online: September 21, 2010
Chem. Eur. J. 2010, 16, 12668 – 12677
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12677