Synthesis of 6-substituted 7-bomoazabicyclo[2.2.1]heptanes via nucleophilic addition to 3-bromo-1-azoniatricyclo[2.2.1.0]-heptane bromide

Article abstract of DOI:10.1002/adsc.200404322

We describe herein an efficient method for the preparation of a functionalised bicyclic framework (6-substituted 7-bromo-aza-bicyclo[2.2.1] heptane) through the selective opening of the aziridium 2 with organocuprates in up to 90% yield. These interesting chiral building blocks were then utilised as novel ligands in the rearrangement of epoxides to afford chiral allylic alcohols.

Full text of DOI:10.1002/adsc.200404322

FULL PAPERS
Synthesis of 6-Substituted 7-Bomoazabicyclo[2.2.1]heptanes via
Nucleophilic Addition to 3-Bromo-1-azoniatricyclo[2.2.1.0]-
heptane Bromide
Arnaud Gayet, Pher G. Andersson*
Department of Organic Chemistry, Uppsala University, Box 531, 75121 Uppsala, Sweden
Fax: (þ46)-18-471-3818, e-mail: pher.andersson@kemi.uu.se
Received: October 20, 2004; Accepted: March 23, 2005
Supporting Information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
Abstract: We describe herein an efficient method for el ligands in the rearrangement of epoxides to afford
the preparation of a functionalised bicyclic frame- chiral allylic alcohols.
work (6-substituted 7-bromo-aza-bicyclo[2.2.1]hep-
tane) through the selective opening of the aziridium
2 with organocuprates in up to 90% yield. These inter- Keywords: asymmetric catalysis; aziridinium opening;
esting chiral building blocks were then utilised as nov- diamine ligand; enantiopure; organocuprates
Introduction
system occurs due to the fact that only the less sterically
hindered bond of the aziridinium bromide is available
Small, easily accessible enantiopure chiral molecules for SN2 nucleophilic attack (Scheme 1). A later patent[6]
are of great interest in both the development of new ef- describes the ring opening of the racemic aziridinium
ficient asymmetric catalysts and as building blocks in the bromide 3b with hydride as the nucleophile, leading to
synthesis of larger chiral molecules. In recent years our an unsubstituted bicyclic structure. In each case they
group has developed a number of low molecular weight employed this strategy to synthesise analogues of the
bicyclic compounds; with applications as chiral ligand muscarinic agonists arecoline and epibatidine.
for the transfer hydrogenation of ketones[1,2] (e.g., 1a)
We decided to extend this methodology by investigat-
and the rearrangement ofepoxides into allylic alcohols[3] ing the reactivity of the aziridinium ion 2 towards hard
(e.g., 1b). nucleophiles derived from organolithium- or
A limitation of these ligands is the difficulty in intro- Grignard-type reagents. This would allow a convenient
ducing diversity during the synthesis and the lack of a route to a large number of enantiopure bicyclic mole-
possible anchoring point on which to attach a solid sup- cules, bearing various substituents at the 6-position
port material, allowing a simpler purification and the and a bromine atom on the bridgehead position, allow-
catalyst to be easily recycled. We consequently investi- ing for further functionalisation and introduction of
gated the development of versatile methods to prepare more diversity to these molecules (Scheme 1).
such functionalised chiral building blocks.
Our attention was drawn to the work of Shapiro[4] and
Pombo-Villar[5] on the ring opening of enantiomerically
pure aziridinium bromide 3a with stabilised nucleo-
philes. Complete inversion of configuration of the ring
Figure 1. Low molecular weight bicyclic compounds 1a and
1b.
Figure 2. Aziridinium bromides 3a–c.
1242
ꢀ 2005WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/adsc.200404322
Adv. Synth. Catal. 2005, 347, 1242–1246

Synthesis of 6-Substituted 7-Bomoazabicyclo[2.2.1]heptanes
FULL PAPERS
When CuCN is exposed to two equivalents of an alkyl-
lithium, a diionic species of the general formula R2
Cu(CN)Li2 is produced. Such cuprates are far more re-
active and stable than the corresponding lower-order
cuprates.
The use of these more reactive organocopper reagents
as nucleophiles allows the opening of 2 with bulkier al-
kyl groups. Compound 6 having an n-butyl group in
the 6-position can now be obtained in over 90% yield
(Table 2, entry 2), while compound 7 bearing the bulky
sec-butyl group can now be prepared in 57% yield (Ta-
ble 2, entry 3). We observed little change in the reactiv-
Scheme 1.
Results and Discussion
Following an analogous procedure[7] the Diels–Alder ity of the PhLi cuprate (Table 2 entry 5). However, the
adduct 4 was form by the in situ reaction of (S)-phenyl- reactivity of the MeLi cyanocuprate decreased com-
ethylamine hydrochloride with aqueous formaldehyde pared to its lower-order equivalent, the yield of 5 drop-
and cyclopentadiene (Scheme 2). Compound 4 was ping from 57% to 46% (Table 2, entry 1). This result em-
then converted to the aziridinium salt 2 by treatment phasises the unpredictable nature of organocopper re-
with bromine[8].
agents. Unfortunately, attempts to introduce the very
Initially we chose to use organocopper reagents bulky tert-butyl group were again unsuccessful.
known for their high reactivity and chemoselectivity as Since the number of readily available organolithium
nucleophiles to openthe aziridinium bromide 2. Our pri- reagents is limited, we decided to investigate the possi-
mary studies investigated the use of lower-order cup- ble use of Grignard species to form organocopper re-
rates as possible nucleophiles. Thus cuprates of the gen- agents suitable to open the aziridinium bromide 2. Ini-
eral formula R2CuLi were prepared by combining two tially we tried to use only a catalytic amount of copper
equivalents of an organolithium with one of copper(I) salt (Table 3, entry 1), but no open product was observ-
iodide.
able under these conditions. Subsequently, we tried to
The first homocuprate addition was achieved with use a stoichiometric amount of copper source compared
MeLi as the organolithium species, leading to com- to the Grignard reagent (Table 3, entry 2) forming a
pound 5 as a single product in 57% yield (Table 1, en- lower-order organocopper of the general formula
try 1). After this initial result we wanted to determine RCu·MgX2, but again no reaction occurred. This is
the scope of the reaction by testing other commercially probably due to the low reactivity of these organocop-
available organolithium reagents. The use of n-butyl- per species.
lithium resulted in a reduction in yield to 39% under
We then tested the more reactive magnesio-cuprates
the same conditions (Table 1, entry 2). Furthermore, of the type R2CuMgX and R2Cu(CN)MgX, formed by
the use of more bulky butyl organolithium did not pro-
duce anyopen product; with the corresponding sec-butyl
Table 1. Addition of lower order cuprate to aziridinium bro-
and tert-butyl organocopper reagents reacting with the
mide 2.
aziridinium (Table 1, entries 3 and 4). Surprisingly how-
ever, the use of phenyllithium produced the open prod-
uct 9 in good yield (60%) (Table 1, entry 5).
Due to the fact that bulky alkyllithium groups did not
successfully open 2using the above conditions, we decid-
ed to use higher-order cyanocuprates as nucleophiles.
Entry
R
Copper reagent Yield [%][a] Compound
1
2
3
4
Me
Me2CuLi
57
39
0
5
6
7
8
9
n-Bu n-Bu2CuLi
s-Bu s-Bu2CuLi
t-Bu
t-Bu2CuLi
0
5Ph
Ph
2CuLi
60
[a]
Yield of isolated product.
Table 2. Addition of higher-order cyanocuprate to aziridi-
nium bromide 2.
Entry
R
Copper reagent
Yield [%][a] Compound
1
2
3
4
Me
Me2Cu(CN)Li2
46
5
6
7
8
9
n-Bu n-Bu2Cu(CN)Li2 91
s-Bu s-Bu2Cu(CN)Li2 57
t-Bu t-Bu2Cu(CN)Li2
Ph 2Cu(CN)Li2
0
64
5Ph
Scheme 2. (a) Cyclopentadiene, CH2O, H2O, r.t., 16 h; (b)
CH2Cl2, Br2, 0 8C, 10 h, 75%.
[a]
Yield of isolated product.
Adv. Synth. Catal. 2005, 347, 1242–1246
asc.wiley-vch.de
ꢀ 2005WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1243

Arnaud Gayet, Pher G. Andersson
FULL PAPERS
Table 3. Addition of organocuprates derived from Grignard reagents to 2.
Entry
Cu(I) source
Equivs.
Grignard reagent
Equivs.
Yield [%][a]
Product
1
2
3
4
CuCN
CuCN
CuCN
CuI
0.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
EtMgBr
EtMgBr
EtMgBr
EtMgBr
EtMgBr
EtMgI
MeMgBr
VinylMgBr
VinylMgBr
n-PrMgBr
n-PrMgBr
n-BuMgBr
i-PrMgBr
i-PrMgBr
1.1
1.1
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
2.2
0
0
10
10
10
10
10
10
5
11
11
12
12
6
86
65
62
52
82
0
0
0
49
45
60
0
5CuBr·SMe
6
7
8
9
10
11
12
13
14
2
CuCN
CuCN
CuCN
CuBr·SMe2
CuCN
CuBr·SMe2
CuCN
CuCN
13
13
CuBr·SMe2
[a]
Yield of isolated product.
the addition of two equivalents of Grignard reagent to for a particular Grignard reagent. CuCN seems to be
one equivalent of copper(I) salt. The resulting organo- the first choice but if it is unsuccessful other cupper sour-
cuprates are very reactive and thermally unstable, effec- ces must always be tested prior to discounting the use of
tive stirring and cooling are essential for both reagent a particular Grignard reagent.
formation and subsequent introduction to the aziridini-
After developing successful methods for the synthesis
um bromide. We initially used CuCN as the copper of a variety of enantiopure 6-substituted 7-bromo-aza-
source and EtMgBr (Table 3, entry 3), which lead to bicyclo[2.2.1]heptanes we wanted to investigate the po-
the formation of 10 in excellent yield (86%). Variation tential use of these new bicyclic intermediates for chiral
of the copper source resulted in a lower activity with ligand synthesis.
CuI and CuBr·SMe2 forming 10 in 65and 62% yield re-
spectively (Table 3, entry 5and 6). Changing the
We decided to apply these building blocks to the syn-
thesis of chiral diamines which could be used as ligands
Grignard reagent from the bromide to the iodide (Ta- for the rearrangement of an epoxide into a chiral allylic
ble 3, entry 6) did not improve the reactivity of the cor- alcohol. In order to determine the effect of substitution
responding magnesio cuprate, with 10 formed in 52% on the 6-position, diamines featuring either a hydrogen
yield.
or a phenyl were prepared. The aziridinium ion 2 was
After the encouraging result obtained using EtMgBr ring opened using Red-Alꢁ as hydride donor to produce
we investigated the scope of the reaction by using a va- the unsubstituted bicycle 14 in 84% yield (Scheme 3).
riety of Grignard reagents. MeMgBr in combination
with CuCN afforded 5 in 83% yield (Table 3, entry 7).
Surprisingly the use of vinylmagnesium bromide with ei-
ther CuCN or CuBr·SMe2 as coppersource(Table 3, en-
try 8 and 9, respectively) could not open 2 to form the
desired product. When using larger Grignard reagents
yields decreased but stayed within an acceptable range,
compound 6 having a butyl group in the 6-position was
produced in 45% yield using CuCN as copper source
(Table 3, entry 12). While under the same conditions
compound 2 did not react with the corresponding prop-
yl-cuprate. However, it was found that the use of CuBr·
SMe2 enabled the addition of n-propyl with 12 produced
in 49% yield (Table 3, entry 11). Addition product 13,
bearing the bulky isopropyl substituent can be formed
in 60% yield when CuCN is employed as the copper
source. Interestingly, unlike with n-propyl the use of
CuBr·SMe2 yielded no product (Table 3, entry 13 and
14, respectively). These results highlight the difficulty
in predicting which copper salt provides the best yield Scheme 3. Synthesis of the diamines 17 and 18.
1244
ꢀ 2005WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1242–1246

Synthesis of 6-Substituted 7-Bomoazabicyclo[2.2.1]heptanes
FULL PAPERS
larimetry was performed using a Perkin-Elmer 241 polarime-
ter. 1Hand 13C NMR were obtained onaVarian Unity 400 using
the residual peak from CDCl3 d¼7.26 for 1H or CDCl3 d¼77.0
for 13C as reference. Unless otherwise noted, materials were
obtained from commercial suppliers and used without further
purification. THF was distilled from a sodium-benzophenone
solution.
Nucleophilic substitution of the bromide in the 7-posi-
tion of compound 14 and 9 by pyrrolidine produced
the protected diamines 15 and 16 in excellent yield. Re-
moval of the benzylic protecting group by hydrogena-
tion at atmospheric pressure led to both the non-substi-
tuted diamine 17 and the 6-phenyl substituted diamine
18 in quantitative yield.
The two diamines were tested as a catalytic chiral base
in the rearrangement of cyclohexene oxide using LDA
as stoichiometric base (Scheme 4). The corresponding
allylic alcohol was obtained in 60% ee after 12 hours us-
(1R,4S)2-[(S)1-Phenylethyl]-2-azabicyclo[2.2.1]hept-
5-ene (4)
ing the non-substituted diamine 17 and 61% ee after 14 Formaldehyde (37%, 15.5 mL, 190 mmol) was added to a solu-
tion of (ꢀ)-(S)-1-phenylethylamine hydrochloride (21 g,
hours using the 6-phenyl substituted diamine. This result
133 mmol) in water (60 mL) and the solution was stirred at
is very promising since only very few catalytic systems
exist for this reaction.[3,9] Substitution at the 6-position
08C for 15min then freshly distilled cyclopentadiene (25mL,
380 mmol) was added. The mixture was vigorously stirred at
does not affect the selectivity or the activity of the cata-
08C for 4 hours before being diluted with water (250 mL)
lyst offering us an ideal anchoring point for a polymer or
and washed (hexane/Et2O, 1:1, 3ꢁ100 mL). The aqueous
other solid support.
layer was made alkaline by addition of KOH pellets and ex-
tracted with Et2O (3ꢁ100 mL). The organic phases were dried
(MgSO4) and evaporated. The crude product was subjected to
purification by column chromatography (EtOAc/pentane,
70:30) to offer pure 4; yield: 51%. All spectroscopic and phys-
ical data were in accordance with those published.[7]
(1R,2R,3R,4R,6R)-3-Bromo-1-[(S)-1-phenylethyl]-1-
azoniatricyclo[2.2.1.0]heptane Bromide (2)
A solution of 4 (15g, 75mmol) in CH 2Cl2 (50 mL) was added
dropwise over 1 hour to a cooled solution of Br2 (3.6 mL,
67 mmol) in CH2Cl2 (50 mL) and the solution was kept at
08C overnight. The yellow crystals that formed were filtered
and recrystallised from CH2Cl2/Et2O to afford 2 as colourless
crystals; yield: 19 g (72%). All spectroscopic and physical
data were in accordance with those published.[8]
Scheme 4.
General Procedure for Nucleophilic Addition of
Lower-Order Cuprates (R2CuLi) to 2
Conclusion
A solution of organolithium (1.54 mmol) was added dropwise
at ꢀ308C to a slurry of cuprous iodide (145mg, 0.77 mmol) in
THF (2 mL) under an argon atmosphere. The solution was stir-
redfor30 minat08C.Thereactionmixturewascooledagainto
ꢀ308C, and aziridinium bromide 2 was added portionwise. Af-
ter 2 hours the reaction was allowed to slowly reach room tem-
perature (over several hours). The black/brown reaction was
poured into a solution of saturated NH4Cl (5mL), the aqueous
phase was extracted with CH2Cl2 (3ꢁ10 mL), and the com-
bined organic layers dried over MgSO4. The solvent was evapo-
rated and the resulting oil subjected to a short column chroma-
tography of silica gel (pentane/diethyl ether, 90:10).
We have developed successful methods to prepare a va-
riety of enantiopure 6-substituted 7-bromoazabicy-
clo[2.2.1]heptane, using organocopper reagents as nu-
cleophiles for the regioselective opening of the tricyclic
aziridinium bromide 2. The usefulness of these 6-substi-
tuted 7-bromoazabicyclo[2.2.1]heptane compounds as
intermediate in the preparation of successful chiral li-
gands for catalytic asymmetric synthesis has been high-
lighted.
Experimental Section
General Procedure for Nucleophilic Addition of
Higher-Order Cuprates [R2Cu(CN)Li] to 2
General Remarks
Flash chromatography was performed on silica gel (Matrex
60A, 37–70 mm). Analytical TLC was carried out on precoated
plates SIL G-60 UV254, purchased from Macherey-Nagel. Po-
A solution of organolithium (1.54 mmol) was added dropwise
at ꢀ308C to a slurry of CuCN (145mg, 0.77 mmol) in THF
(2 mL) under an argon atmosphere. The solution was stirred
Adv. Synth. Catal. 2005, 347, 1242–1246
asc.wiley-vch.de
ꢀ 2005WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1245

Arnaud Gayet, Pher G. Andersson
FULL PAPERS
for 30 min at 08C. The reaction mixture was cooled again to
ꢀ308C, followed by the portionwise addition of aziridinium
bromide 2. After 2 hours the reaction mixture was allowed to
slowly reach room temperature. The black/brown reaction
was poured into a solution of saturated NH4Cl (5mL), the
aqueous phase was extracted with CH2Cl2 (3ꢁ10 mL), and
the combined organic layers dried over MgSO4. The solvent
was evaporated and the resulting oil subjected to a short col-
umn chromatography of silica gel (pentane/diethyl ether,
90:10).
(1R,4R,6S,7S)-2-[(S)-1-Phenylethyl]-7-pyrrolidino-6-
phenyl-2-azabicyclo[2.2.1]heptane (16)
Following the procedure described for compound 15 using 9
(250 mg, 0.7 mmol), compound 16 was obtained in 88% yield
(213 mg).
(1R,4R,7R)-7-Pyrrolidino-2-azabicyclo[2.2.1]heptane
(17)
Pd(OH)2 on carbon (80 mg, 20 wt %) was dried under vacuum
for 4 hours, and then added to a solution of the benzyl-protect-
ed amine 15 (400 mg, 1.48 mmol) in MeOH (15mL). The reac-
tion mixture was stirred overnight at room temperature under
a hydrogen atmosphere. The solution was filtered over celite
and the solvent removed under vacuum to afford 17 as a yellow
oil in quantitative yield.
General Procedure for Nucleophilic Addition of
Organocopper-Derived Species from Grignard
Reagents to 2
A solution of Grignard reagent was added dropwise at ꢀ508C
to a slurry of copper salt in THF (2 mL) under argon atmos-
phere. The solution was stirred for 30 min at ꢀ208C, then
the reaction mixture was cooled again to ꢀ508C, and aziridini-
um bromide 2 was added portionwise. After 2 hours the reac-
tion mixture was allowed to slowly reach room temperature.
The black/brown reaction was poured into a solution of satu-
rated NH4Cl (5mL), the aqueous phase was extracted with
CH2Cl2 (3ꢁ10 mL), and the combined organic layers dried
over MgSO4. The solvent was evaporated and the resulting
oil subjected to a short column chromatography of silica gel
(pentane/diethyl ether, 90:10).
(1R,4R,6S,7S)-7-Pyrrolidino-6-phenyl-2-
azabicyclo[2.2.1]heptane (18)
Following the procedure described for compound 17 using 6
(150 mg, 0.43 mmol), compound 18 was obtained as a yellow
oil in quantitative yield.
Acknowledgements
Characterization data for products 5, 6, 7, 9, 10, 12, 13, 14, 15,
16, 17, and 18 are available in the Supporting Information.
This work was supported by grants from The Swedish Research
Council and COST Chemical Action D24.
References
(1R,4R,7R)-2-[(S)1-Phenylethyl]-7-bromo 2-
azabicyclo[2.2.1] heptane (14)
[1] a) D. A. Alonso, S. J. M. Nordin, P. Roth, T. Tarnai, M.
Thommen, U. Pittelkow, P. G. Andersson, J. Org. Chem.
2000, 65, 3116–3122; b) S. J. M. Nordin, P. Roth, T. Tarnai,
D. A. Alonso, P. Brant, P. G. Andersson, Chem. Eur. J.
2001, 7, 1431–1436.
Red-Alꢁ (65wt %, 8.5mL, 28.9 mmol) was added to a solution
of 2 (10.3 g, 28.9 mmol) in tetrahydrofuran (250 mL) cooled to
ꢀ108C. After stirring for 2 hours, the reaction mixture was
quenched with a saturated aqueous sodium bicarbonate solu-
tion (125mL) followed by brine (125mL), and then extracted
with ethyl acetate (3ꢁ200 mL). The combined organic ex-
tracts were dried over magnesium sulfate and then concentrat-
ed on a rotary evaporator to afford 14 as a colourless oil; yield:
7.6 g (27.5mmol).
[2] A. Gayet ,C. Bolea, P. G. Andersson, Org. Biomol. Chem.
2004, 2, 1887–1893.
[3] a) M. J. Sçdergren, S. K. Bertilsson, P. G. Andersson, J.
Am. Chem. Soc. 2000, 122, 6610–6618; b) A. Gayet, S.
Bertilsson, P. G. Andersson, Org. Lett. 2002, 4, 3777–3779.
[4] P. Floersheim, E. Pombo-Villar, G. Shapiro, Chimia 1992,
46, 323–334.
[5] E. Pombo-Villar, P. Supavilai, H. P. Weber, H. W. G. M.
Noddeke, Bioorg. Med. Chem. Lett. 1992, 2, 501–504.
[6] C. H. Mitch, S. J. Quinby, World Patent WO 00/75140 A1,
2000.
[7] S. D. Larsen, P. A. Grieco, J. Am. Chem. Soc. 1985, 107,
1768.
[8] E. Pombo-Villar, J. Boelsterli, M. Magdalena, J. France, B.
Fuchs, M. Walkinshaw, H. O. Weber, Helv. Chem. Acta
1993, 76, 1203–1215.
[9] a) M. Asami, t. Ishizaki, S. Inoue, Tetrahedron Lett. 1994,
35, 793–796; b) J. P. Tierney, A. Alexakis, P. Mangeney,
Tetrahedron: Asymmetry 1997, 8, 1019–1022.
(1R,4S,7R)-2-[(S)-1-Phenylethyl]-7-pyrrolidino-2-
azabicyclo[2.2.1]heptane (15)
Compound 14 (0.5g, 1.8 mmol) was dissolved in pyrrolidine
(5mL), and heated under reflux for 18 hours. Volatiles were re-
moved under vacuum and the resulting oil washed with a satu-
rated aqueous solution of NaHCO3 (5mL), extracted with
CH2Cl2 (3ꢁ10 mL) and dried over MgSO4. Removal of the sol-
vent in vacuum afforded 15 as a colourless oil; yield: 430 mg
(90%).
1246
ꢀ 2005WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1242–1246