Rhodium-catalyzed ring-opening reactions of N-boc-azabenzonorbornadienes with amine nucleophiles

Article abstract of DOI:10.1021/ja0577701

In the presence of a rhodium catalyst (5 mol %) generated in situ from [Rh(cod)Cl]₂ and (S, S′)-(R, R′)-C₂-ferriphos (4a), the asymmetric ring-opening reaction of azabenzonorbornadienes (1a-m) with various aliphatic and aromatic amin

Full text of DOI:10.1021/ja0577701

Published on Web 05/05/2006
Rhodium-Catalyzed Ring-Opening Reactions of
N-Boc-Azabenzonorbornadienes with Amine Nucleophiles
Yong-hwan Cho, Valentin Zunic, Hisanori Senboku, Madeline Olsen, and
Mark Lautens*
Contribution from the DaVenport Laboratories, Department of Chemistry, UniVersity of Toronto,
80 St. George Street, Toronto, Ontario, Canada M5H 3H6
Received November 15, 2005; E-mail: mlautens@chem.utoronto.ca
Abstract: In the presence of a rhodium catalyst (5 mol %) generated in situ from [Rh(cod)Cl]2 and (S,S′)-
(R,R′)-C2-ferriphos (4a), the asymmetric ring-opening reaction of azabenzonorbornadienes (1a-m) with
various aliphatic and aromatic amines (2a-l) proceeded with high enantioselectivity (up to >99% ee) to
give the corresponding 1,2-diamine derivatives 3 in high yields. In the specific case of pyrrolidine as
nucleophile, Et3NHCl was necessary as an additive for good reactivity and enantioselectivity. Additionally,
a practical protocol was developed for the ring-opening of 1a with volatile amines at elevated temperatures
and standard pressure, using R2NH2I and i-Pr2NEt. The experimental results showed that the nature of the
chiral ligand has the significant impact on the reactivity of the catalyst and the use of excess amount (2.2
eq to Rh) of the chiral ligand plays an important role to improve the enantioselectivity in the present
asymmetric reaction.
Introduction
substitution reaction. This class of reaction is particularly
attractive because more than one stereocenter can be generated
Transition metal-catalyzed asymmetric reactions are powerful
tools in modern organic synthesis.1 In particular, catalytic
asymmetric carbon-carbon bond and carbon-heteroatom bond
forming reactions have attracted a great deal of attention.
Although there are several other methods of providing optically
active compounds, for example optical resolution, catalytic
asymmetric synthesis is an ideal and practical method as far as
high enantioselectivity is obtained, because a large amount of
chiral product can be produced with a catalytic amount of chiral
material.2
The transition metal-catalyzed allylic substitution3 with
various nucleophiles represents a fundamentally important
reaction for the construction of useful chiral building blocks,
and many methods have been developed which enable a wide
range of allylic leaving groups to be used along with a variety
of carbon- and heteroatom nucleophiles.4 Ring-opening reactions
of heterobicyclic alkenes formally represent a type of allylic
by the desymmetrization of a meso-bicyclic alkene in a single
step.
We previously reported the rhodium-catalyzed asymmetric
ring-opening of oxabenzonorbornadiene with alcohols and
phenols, producing hydronaphthalenes in high yields and with
excellent enantioselectivity.5,6 Also, rhodium-catalyzed ring-
opening of oxabicyclic alkenes with amines,7 carboxylates,8 1,3-
dicarbonyl nucleophiles9 and sulfur nucleophiles10 have exten-
sively been studied.
We next focused our attention on expanding the scope of
the asymmetric ring-opening of azabicyclic alkenes so as to have
efficient access to the cyclohexyl-1,2-diamine moiety which has
been found to be an important class of compounds, particularly
for their utility as chiral ligands11 and for their biological
activity.12 Recently, we reported the catalytic asymmetric ring-
(4) For reviews of the catalytic asymmetric allylic substitution: (a) Trost, B.
M.; Crawley, M. L. Chem. ReV. 2003, 103, 2921. (b) Trost, B. M.; Chulbom,
L. In Catalytic Asymmetric Synthesis, 2nd ed.; Ojima, I., Ed.; VCH: New
York, 2000; p 593. (c) Pfaltz, A.; Lautens, M. In ComprehensiVe
Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.;
Springer: Berlin, 1999; Vol. 2, Chapter 24. (d) Trost, B. M.; Van Vranken,
D. L. Chem. ReV. 1996, 96, 395. (e) Hayashi, T. In Catalytic Asymmetric
Synthesis; Ojima, I., Ed.; VCH: New York, 1993; p 325. (f) Frost, C. G.;
Howarth, J.; Williams, J. M. J. Tetrahedron Asymmetry 1992, 3, 1089.
(5) (a) Lautens, M.; Fagnou, K.; Rovis, T. J. Am. Chem. Soc. 2000, 122, 5650.
(b) Lautens, M.; Fagnou, K.; Taylor, M. Org. Lett. 2000, 2, 1677.
(6) (a) Lautens, M.; Fagnou, K.; Taylor, M.; Rovis, T. J. Organomet. Chem.
2001, 624, 259. (b) Lautens, M.; Fagnou, K.; Hiebert, S. Acc. Chem. Res.
2003, 36, 48.
(1) (a) Morrison, J. D., Ed. Asymmetric Synthesis; Academic Press: London,
1983-1985; Vols. 1-5. (b) Ojima, I. Catalytic Asymmetric Synthesis, 2nd
ed.; VCH Publishers New York, 2000. (c) Jacobsen, E. N.; Pfaltz, A.;
Yamamoto, H. ComprehensiVe Asymmetric Catalysis, Supplement 1, 2;
Springer: Berlin, 2003, 2004. (d) Noyori, R. Asymmetric Catalysts in
Organic Synthesis; Wiley: New York, 1994.
(2) For a review: Kagan, H. B.; Fiaud, J. C. Top. Stereochem. 1978, 10, 175.
(3) For reviews, see: (a) Tsuji, J. Acc. Chem. Res. 1969, 2, 144. (b) Tsuji, J.
Pure Appl. Chem. 1982, 54, 197. (c) Trost, B. M.; Verhoeven, T. R. In
ComprehensiVe Organometallic Chemistry, Vol. 8; Wilkinson, G., Stone,
F. G. A., Abel, E. W., Eds.; Pergamon: Oxford, 1982; pp 799-938. (d)
Heck, R. F. Palladium Reagents in Organic Synthesis; Academic Press:
New York, 1985. (e) Tsuji, J. J. Organomet. Chem. 1986, 300, 281. (f)
Tsuji, J.; Minami, I. Acc. Chem. Res. 1987, 20, 140. (g) Tsuji, J. Organic
Synthesis with Palladium Compounds; Springer-Verlag: New York, 1990.
(h) Godleski, S. In ComprehensiVe Organic Synthesis, Vol. 4; Trost, B.
M., Fleming, I., Semmelhack, M. F., Eds.; Pergamon: Oxford, 1991; pp
585-659. (i) Tsuji, J. Palladium Reagents and Catalysts; John Wiley and
Sons: Chichester, 1995.
(7) Lautens, M.; Fagnou, K. J. Am. Chem. Soc. 2001, 123, 7170.
(8) Lautens, M.; Fagnou, K. Tetrahedron 2001, 57, 5067.
(9) Lautens, M.; Fagnou, K.; Yang, D. J. Am. Chem. Soc. 2003, 125, 14884.
(10) Leong, P.; Lautens, M. J. Org. Chem. 2004, 69, 2194.
(11) Fache, F.; Schulz, E.; Tommasino, M. L.; Lemaire, M. Chem. ReV. 2000,
100, 2159.
(12) Lucet, D.; Le Gall, T.; Mioskowski, C. Angew. Chem., Int. Ed. 1998, 37,
2580.
9
10.1021/ja0577701 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 6837-6846
6837

A R T I C L E S
Cho et al.
Table 1. Activating Group Effect on Enantioselectivity in
opening of azabenzonorbornadienes with aliphatic and cyclic
amines in the presence of a rhodium catalyst coordinated with
a chiral ferrocenyl phosphine ligand and the application of this
methodology in the total synthesis of an analgesic compound.13
In this type of reaction, the chiral 1,2-diamine is thought to be
generated via the enantioselective cleavage of a bridgehead
carbon-nitrogen bond in azanorbornadiene followed by SN2′
nucleophilic attack of amine.
Ring-Openinga
Here, we report a full description of the catalytic asymmetric
ring-opening reaction of azabenzonorbornadienes with aliphatic
and aromatic amines. The reaction proceeds under relatively
mild conditions and generates a variety of 1,2-diamino com-
pounds in high yield with excellent enantioselectivity.
Results and Discussion
entry
azabicycle
time (h)
product
yieldb (%)
% eec
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1f
24
24
24
24
24
24
24
8
3aa
3ba
3ca
3da
3ea
3fa
77
65
73
71
90
40
n. r.
96
86
63
68
96
99
25
The first example of the transition metal-catalyzed ring-
opening reaction of azabicyclic alkenes is the palladium-
catalyzed alkylative ring-opening of N-substituted azabenzonor-
bornadienes.14 Palladium-catalyzed ring-opening addition of
arylboronic acids to azabicyclic substrates has also been
reported.15 From our previous studies, azabicyclic alkenes,
including azabenzonorbornadienes, were found to be less
reactive than the corresponding oxabicyclic alkenes. For ex-
ample, the catalytic ring-opening of oxabenzonorbornadiene with
arylboronic acid proceeds at room temperature in the presence
of Pd(dppf)Cl2 whereas the reaction of N-Boc-azabenzonorbor-
nadiene requires heating (60 °C) to ensure complete conversion
under the same conditions.5,15 As expected, the activating group
on nitrogen has a strong effect on the reactivity of the substrate.
The results summarized in Table 1 show that the nature of
the activating group strongly influences both the reaction yield
and the enantioselectivity of the product. High enantioselectivity
was observed in the reaction of N-Boc-1a (entry 1). In the
1g
1h
3ha
10
a The reaction was carried out with azabicycle 1 (0.33 mmol) and 5.0
equiv. of pyrrolidine (2a) (1.64 mmol) in tetrahydropyran (1.6 mL) in the
presence of [Rh(cod)Cl]2 (2.5 mol %), 4a (10 mol %), and Et3NHCl (0.82
mmol). b Yield after silica gel chromatography. c Determined by HPLC
analysis of the corresponding ring-opening product 3 with chiral stationary
column (Chiralcel OD or Chiralcel OD-H).
most enantioselective substrate in the present reaction, 1a was
the more versatile substrate to find efficient conditions for ring-
opening with amine nucleophiles since deprotection of the Boc
group17 is easy, high-yielding, and more amendable to scale-
up with regards to starting material synthesis.
We then examined the asymmetric ring-opening of N-Boc
protected 1a with pyrrolidine. It was found that (S,S′)-(R,R′)-
C2-ferriphos (4a)18 gave the best enantioselectivity. We exam-
ined several chiral ligands such as BINAP, BPPFA,19 DIPOF,20
and PPF-P(t-Bu)2,21 all used successfully for other asymmetric
ring-opening reactions of bicyclic alkenes, but these failed to
give satisfactory results.6b,9 For example, under the optimized
conditions for the reaction of oxabenzonorbornadiene using
(S,R)-PPF-P(t-Bu)2 with NH4I, the reaction of 1a with pyrro-
lidine gave 3aa in 65% yield and only 25% ee.
According to our preliminary study on the catalytic ring-
opening reaction of oxabenzonorbornadiene with pyrrolidine,
various additives, containing a proton and a halide, were tested
to promote the reaction (Table 2).
The best overall results were obtained with Et3NHCl.7
Although the reaction of 1a with pyrrolidine in THP at 100 °C
gave the corresponding product 3aa in 50% yield and 59% ee
in the absence of an additive (entry 1), the use of Et3NHCl gave
75% yield and 78% ee (entry 6). Slightly higher levels of
16
presence of 2.5 mol % of [Rh(cod)Cl]2 and 10 mol % of
(S,S′)-(R,R′)-C2-ferriphos (4a), 1a with 2.5 equiv. of triethy-
lamine hydrochloride and 5.0 equiv. of pyrrolidine in tetrahy-
dropyran at 100 °C for 24 h gave the corresponding product
3aa in 77% yield with 86% ee. While other carbamates such
as methyl (1b) and benzyl (1c) showed good reactivity, the
product was obtained in lower enantioselectivity (entries 2 and
3). The reactions using N-benzamide (1d) and N-acetamide (1e)
also showed good reactivity (entries 4 and 5) and 1e was found
to be most enantioselective (99% ee). N-Phenyl (1f), the efficient
substrate for the palladium-catalyzed ring-opening with dim-
ethylzinc,14 showed poor reactivity and enantioselectivity (entry
6) and the reaction using N-methyl (1g) did not give any ring-
opening product (entry 7). The sulfonyl activating group showed
the best reactivity but produced very poor enantioselectivity
(entry 8), suggesting that electronic effects play a significant
role in the reaction outcome. Since it is postulated that the
rhodium-catalyzed ring-opening of strained allylic alkenes
proceeds by an oxidative insertion pathway, use of a more
electron-withdrawing tosyl group would be expected to facilitate
this pathway. Although N-acetamide (1e) was found to be the
(17) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic Synthesis,
3rd ed.; Willey: New York, 1999, and references therein.
(18) (a) Hayashi, T.; Yamamoto, A.; Hojo, M.; Kishi, K.; Ito, Y.; Nishioka, E.;
Miura, H.; Yanagi, K. J. Organomet. Chem. 1989, 370, 129. (b) Schwink,
L.; Knochel, P. Chem. Eur. J. 1998, 4, 950.
(13) Lautens, M.; Fagnou, K.; Zunic, V. Org. Lett. 2002, 4, 3465.
(14) (a) Lautens, M.; Hiebert, S.; Renaud, J. Org. Lett. 2000, 2, 1971. (b) Cabrera,
S.; Arrayas, R. G.; Carretero, J. C. Angew. Chem., Int. Ed. 2004, 43, 3944.
(15) (a) Lautens, M.; Dockendorff, C. Org. Lett. 2003, 5, 3695. (b) Murakami,
M.; Igawa, H. Chem. Common. 2002, 390.
(16) The use of 2.5 mol % of [Rh(cod)Cl]2 is required to ensure high reaction
yield and enantioselectivity. Lower catalyst loading (1∼2 mol %) brought
lower enantioselectivity and slower reaction.
(19) Hayashi, T.; Mise, T.; Fukushima, M.; Kagotani, M.; Nagashima, N.;
Hamada, Y.; Matsumoto, A.; Kawakami, S.; Konishi, M.; Yamamoto, K.;
Kumada, M. Bull. Chem. Soc. Jpn. 1980, 53, 1138.
(20) (a) Richards, C. J.; Locke, A. J. Tetrahedron: Asymmetry 1998, 9, 2377.
(b) Nishibayashi, Y.; Uemura, S. Synlett 1995, 79. (c) Sammakia, T.;
Latham, H. A.; Schaad, D. R. J. Org. Chem. 1995, 62, 10.
(21) Togni, A.; Breutel, C.; Schnyder, A.; Spindler, F.; Landert, H.; Tijani, A.
J. Am. Chem. Soc. 1994, 116, 4062.
9
6838 J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006

Ring-Opening of Azabenzonorbornadienes with Amines
A R T I C L E S
Table 2. Catalytic Asymmetric Ring-Opening of
Table 3. Asymmetric Ring-Opening of 1A with Various Aliphatic
Amines (2A-e)a
N-Boc-Azabenzonorbornadiene (1A) with Pyrrolidine (2A)a
ligand
additive
(equiv.)
(equiv.
to Rh)
T
time
(h)
yieldb
(%)
entry
solvent
(
°
C)
% eec
ligand
yieldb
(%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
1.0
1.0
1.0
1.0
1.0
1.0
2.0
2.0
2.0
2.0
2.2
2.2
2.2
2.2
THP
THP
THP
THP
THP
THP
THP
DMF
CH3CN 100
toluene
THP
THP
THP
100
100
100
100
100
100
100
100
24
24
24
24
24
24
24
24
24
24
24
72
72
72
50
31
55
94
16
75
77
75
34
90
80
52
65
78
59
80
72
44
57
78
86
80
59
79
90
97
98
(equiv. to
Rh)
time
(h)
% eec
NH4Cl (2.5)
NH4Br (2.5)
NH4I (2.5)
entry
R2NH
product
1
2
3
4
2b
2c
2d
2e
4a (2.2)
4a (2.2)
4a (2.2)
4b (2.2)
60
24
48
48
3ab
3ac
3ad
3ae
83
92
85
98
97
93
97
97
nBu4NCl (2.5)
Et3NHCl (2.5)
Et3NHCl (2.5)
Et3NHCl (2.5)
Et3NHCl (2.5)
Et3NHCl (2.5)
Et3NHCl (2.5)
Et3NHCl (5.0)
Et3NHCl (2.5)
Et3NHCl (1.0)
a The reaction was carried out with 1a (0.33 mmol) and 5.0 equiv. of
amine 2 (1.64 mmol) in the presence of [Rh(cod)Cl]2 (2.5 mol %) and chiral
ligand (11.0 mol %). b Yield after silica gel chromatography. c Determined
by HPLC analysis of the product 3 with chiral stationary column: (Chiralcel
OD-H for 3aa and 3ab, Chiralcel AD for 3ac and 3ad, Chiralcel OD for
3ae).
100
100
80
80
80
THP
>99
a The reaction was carried out with 1a (0.33 mmol) and 5.0 equiv. of
pyrrolidine (2a) (1.64 mmol) in solvent (1.6 mL) in the presence of
[Rh(cod)Cl]2 (2.5 mol %), 4a, and additive. b Yield after silica gel
chromatography. c Determined by HPLC analysis of 3aa with Chiralcel OD-
H.
using toluene gave higher yield (90%) but lower ee (79%) (entry
10) than with THP (77%, 86% ee) (entry 7). The best results
were obtained by decreasing the reaction temperature. In the
presence of 2.5 mol % of [Rh(cod)Cl]2 and 11 mol % of 4a,
the reaction of 1a with 1.0 equiv. of Et3NHCl and 5.0 equiv. of
pyrrolidine in THP at 80 °C for 72 h gave 78% yield of
pyrrolidine ring-opening product 3aa in >99% ee (entry 14).
The use of 1.0 equiv. of the additive was found to be most
effective (entries 12, 13, and 14). We found that the ring-opening
reaction works better with basic amines which have six-
membered ring structures such as piperidine, morpholine, and
piperazine, in the absence of additives and solvents. The use of
additives was not beneficial in these cases and a significant
reduction of reactivity was observed when Et3NHCl was used.
As shown in Table 3, in the presence of 2.5 mol % of [Rh-
(cod)Cl]2 and 11 mol % of 4a, the asymmetric ring-opening of
1a with piperidine (2b) proceeded at 80 °C for 60 h without
additives and solvents to give 83% yield of the corresponding
product 3ab in 97% ee (entry 1). The reaction using morpholine
(2d) also gave excellent yield and enantioselectivity (entry 3).
Slightly lower enantioselectivity (93% ee) was observed in the
reaction using 1-phenylpiperazine (2c) (entry 2). In the case of
dibenzylamine (2e), the use of (S,S′)-(R,R′)-4b, which has
p-tolyl groups on the phosphorus atom of the ligand instead of
phenyl groups, was found to be more efficient than 4a. The
chiral ligand 4b was prepared from (S,S)-1,1′-bis(R-N,N-
dimethylaminoethyl)ferrocene18 and di(4-methylphenyl)chloro-
phosphine. The reaction of 1a with 2e using a rhodium catalyst
generated from [Rh(cod)Cl]2 and (S,S′)-(R,R′)-4b gave 98%
yield of 3ae in 97% ee (entry 4).
enantioselectivity (80% ee) were observed with NH4Cl, but the
reaction was sluggish and gave a low yield (entry 2). With a
series of ammonium halides, an interesting trend was observed.
The reaction yields increased in order Cl < Br < I, but
enantioselectivities decreased (entries 2-4). This inverse rela-
tionship between reactivity and % ee is in contrast with the
reaction of oxabicyclic alkenes using a rhodium catalyst
prepared from [Rh(cod)Cl]2 and (S,R)-PPF-P(t-Bu)2, wherein
the iodide additive gave enhanced reactivity and enantioselec-
tivity.9 Tetrabutylammonium chloride was also tested and
showed low reactivity and enantioselectivity (entry 5). These
results indicated that both a proton and chloride of the additive
were required together to promote a faster reaction.22
The enantioselectivity of 3aa was influenced by the amount
of the chiral ligand. The reaction using 2.0 equiv. of (S,S′)-
(R,R′)-C2-ferriphos (4a) gave 77% yield of 3aa with 86% ee
(entry 7) whereas the use of 1.0 equiv. of chiral ligand gave
the product in 78% ee (entry 6). Higher enantioselectivity was
observed in the reaction using 2.2 equiv. of (S,S′)-(R,R′)-C2-
ferriphos which gave 80% yield of 3aa in 90% ee (entry 11).
Among the solvents examined for the asymmetric ring-
opening of azabenzonorbornadienes 1, tetrahydropyran (THP)
was found to be optimal to deliver the ring-opened product with
high enantioselectivity. DMF and CH3CN were tested and gave
80% ee and 59% ee, respectively (entries 8 and 9). The reaction
The absolute configuration of the ring-opening product 3ac
was determined to be (1S, 2S) by X-ray crystallography of a
derivative, 6. Crystals suitable for X-ray analysis were obtained
by deprotection of the Boc group of 3ac with TFA, followed
(22) In the presence of HCl, the regeneration of an active catalyst from the
poisoned rhodium complex by tight binding with amine to the metal center
was proposed. See: Vallarino, L. M.; Sheargold, S. W. Inorg. Chim. Acta
1979, 36, 243.
9
J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006 6839

A R T I C L E S
Scheme 1
Cho et al.
Table 4. Asymmetric Ring-Opening of 1A with Various Aromatic
Amines (2f-i)a
Scheme 2
Table 5. Rhodium-Catalyzed Asymmetric Ring-Opening of 1i and
1j with Various Amines (2b-f)a
entry
R2NH
time (h)
product
yieldb (%)
% eec
1
2
3
4
2f
48
48
30
30
3af
3ag
3ah
3ai
78
92
96
88
97
97
96
97
2g
2h
2i
a The reaction was carried out with azabicycle 1a (0.33 mmol) and 5.0
equiv. of amine 2 (1.64 mmol) in the presence of [Rh(cod)Cl]2 (2.5 mol
%) and 4a (11 mol %). b Yield after silica gel chromatography. c Determined
by HPLC analysis of the product 3 with chiral stationary column: (Chiralcel
AD for 3af and 3ag, Chiralcel OD for 3ah and 3ai).
entry
R2NH
R1 (azabicycle)
time (h)
product
yieldb (%)
% eec
1
2
2b
2c
2d
2e
2f
2c
2d
2e
H (1i)
H (1i)
H (1i)
H (1i)
120
96
96
96
96
96
96
144
3ib
3ic
3id
3ie
3if
3jc
3jd
3je
84
80
62
91
76
57
50
70
94
94
>99
99
96
89
94
99
3
by protection of 5 with 4-bromobenzenesulfonyl chloride
(Scheme 1). Additionally, 3ae was determined to be (1S, 2S)
by correlation with 7 whose absolute configuration was con-
firmed by X-ray analysis. Tartrate salt 7 was prepared from 3ae
by removal of the dibenzyl groups via catalytic hydrogenation23
and deprotection of the N-Boc group, followed by treatment of
the free 1,2-diamine with L-tartaric acid.
Aromatic amines also showed high reactivity and excellent
enantioselectivity in the asymmetric ring-opening under the new
conditions (Table 4). The reactions of 1a with N-methylaniline
(2f) or tetrahydroquinoline (2g) proceeded to give high enan-
tioselectivity of the corresponding products 3af (97% ee) and
3ag (97% ee), respectively (entries 1 and 2). Interestingly, the
reactions using primary aromatic amines such as aniline (2h)
(entry 3) and 1-aminonaphthalene (2i) (entry 4) proceeded at
80 °C in good yield and high enantioselectivity whereas the
reaction with the primary aliphatic amines such as benzylamine
did not produce the desired product.
4d
5
H (1i)
6
Me (1j)
Me (1j)
Me (1j)
7
8d
a The reaction was carried out with 1 (0.33 mmol) and 5.0 equiv. of
amine 2 (1.64 mmol) in the presence of [Rh(cod)Cl]2 (2.5 mol %) and 4a
(11 mol %). b Yield after silica gel chromatography. c Determined by HPLC
analysis of the product 3 with chiral stationary column: (Chiralcel OD-H
for 3ib and 3if, Chiralcel AD for 3ic, 3ie, 3jc, and 3je, Chiralpak AS for
3jd, Chiralcel OD for 3id). d The reaction was performed with C2-ferriphos-
tolyl (4b) (11 mol %).
2), rather than through N-1. The structure of the indole ring-
opening product 3aj was assigned by H-13C coupled NMR
1
experiments and identification of the N-H indole peak at 7.8
ppm. A similar pattern was also observed in the asymmetric
ring-opening reaction of oxabenzonorbornadiene with indole.7,9
The asymmetric ring-opening of 5,8-dimethyl-N-Boc-aza-
benzonorbornadiene (1i) and 5,6,7,8-tetramethyl-N-Boc-1j with
amine nucleophiles was investigated. The results are summarized
in Table 5 and show that the introduction of methyl groups on
the aromatic ring of azabenzonorbornadiene brought a decrease
in the reactivity probably due to steric effects. Although a longer
Indole (2j) was also found to be a good nucleophile and to
react exclusively at the C-3 position of the indole ring (Scheme
(23) Lautens, M.; Schmid, G. A.; Chau, A. J. Org. Chem. 2002, 67, 8043.
9
6840 J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006

Ring-Opening of Azabenzonorbornadienes with Amines
A R T I C L E S
Table 6. Asymmetric Ring-Opening of
Table 7. Rhodium-Catalyzed Ring-Opening of 1A with Volatile
Amines (2j-k)a
6,7-Disubstituted-Azabicycles (1k-m) with Various Amines (2b-f)a
yield of
yield of
entry
amine (equiv.)
base (eq)
T (°
C)
time (h)
3 (%)b
8 (%)b
1c
2
3
4
5
6
7
Et2NH (3.0)
65
110
110
110
110
110
110
26
6
6
6
6
14
14
4d
trace
Et2NH2I (5.0)e
Et2NH2I (5.0)
Et2NH2I (3.0)
Et2NH2I (2.0)
Et2NH2I (2.0)
Me2NH2I (2.0)e
n. r.
38
59
76
90
89
0.3
0.3
3.0
3.0
3.0
55
30
10
<5
<5
R2
yieldb
(%)
entry
R2NH
(azabicycle)
time (h)
product
% eec
1
2c
2d
2e
2a
2d
2f
2a
2b
2f
Me (1k)
Me (1k)
Me (1k)
OMe (1l)
OMe (1l)
OMe (1l)
F (1m)
48
48
48
72
48
48
72
72
60
3kc
3kd
3ke
3la
3ld
3lf
3ma
3mb
3mf
97
98
95
80
86
91
70
67
95
94
>99
97
>99
94
99
>99
97
2
a The reaction was carried out with 1a (1.00 mmol), amine, i-Pr2NEt,
[Rh(cod)Cl]2 (1.0 mol %), and dppf (3 mol %) in 1,4-dioxane (5.0 mL).
b Yield after silica gel chromatography. c The reactions was carried out with
1.5 equiv. of NH4I in THF instead of 1,4-dioxane. d 89% of 1a was
recovered. e Dialkylammonium iodide was prepared by the halide exchange
of the commercial available corresponding ammonium chloride with NaI
in ethanol.
3d
4e
5
6
7e
8
F (1m)
F (1m)
9
98
a The reaction was carried out with 1 (0.33 mmol) and 5.0 equiv. of
amine 2 (1.64 mmol) in the presence of [Rh(cod)Cl]2 (2.5 mol %) and 4a
(11 mol %) at 80 °C (neat condition). b Yield after silica gel chromatography.
c Determined by HPLC analysis of the product 3 with chiral stationary
column: (Chiralcel AD for 3kc, 3ld, 3lf, and 3mf, Chiralcel OD for 3ke
and 3la, Chiralcel OD-H for 3ma and 3mb, Chiralpak AS for 3kd,). d The
reaction was performed with C2-ferriphos-tolyl (4b) (11 mol %). e The
reaction was carried out with 1 (0.33 mmol) and 2a (1.64 mmol) in THP
(1.6 mL) in the presence of [Rh(cod)Cl]2 (2.5 mol %), 4a (11 mol %), and
Et3NHCl (0.33 mmol).
1a with diethylamine (2k) and 1.5 equiv. of NH4I in THF at 65
°C for 24 h gave only 4% yield of the desired product 3ak and
89% of starting material 1a (Table 7, entry 1). It was previously
reported that the catalytic ring-opening of azabenzonorborna-
diene, which has a p-nitrobenzenesulfonyl group as the activat-
ing group instead of N-Boc group, proceeded with 10 equiv. of
diethylamine and NH4I as the additive in THP at 65 °C for 24
h to give the corresponding product in 91% yield.13 Since the
deprotection of the Boc group is easier and higher-yielding, 1a
was selected as the best substrate to find efficient conditions
for ring-opening with volatile amines.
Fortunately, it was found that the use of dialkylammonium
iodide dramatically increased the reactivity (Table 7). In the
presence of 1.0 mol % of [Rh(cod)Cl]2 and 3.0 mol % of dppf,
reaction of 1a with 2.0 equiv. of diethylammonium iodide and
3.0 equiv. of N,N-diisopropylethylamine (i-Pr2NEt) in 1,4-
dioxane at 110 °C for 14 h gave the corresponding product 3ak
in 90% yield (entry 6). The reaction using dimethylammonium
iodide gave 3al in 89% yield under the same conditions (entry
7). Addition of iodide led to an increase in the reactivity by
halide exchange between [Rh(cod)Cl]2 and iodide which gener-
ated the more reactive rhodium iodide species.9 The use of i-Pr2-
NEt is required to ensure the high yield. Without base, the ring-
opening reaction did not proceed using Et2NH2I (entry 2). The
use of 0.3 equiv. of base gave 38% yield of 3ak and a
considerable amount (55% yield) of the side product 8 (entry
3). Increasing the amount of i-Pr2NEt from 0.3 equiv. to 3.0
equiv. minimized the amount of 8 (entries 4 and 5). Dialkyl-
ammonium iodide was prepared from the commercially available
dialkylammonium chloride by exchange of the halide with NaI
at room temperature.24
reaction time (over 96 h) was required, generally all substrates
reacted in reasonable yields and high enantioselectivity. The
bulkier substrate 1j was found to be even less reactive than 1i.
The reactions of 1j with 1-phenylpiperazine (2c) and morpholine
(2d) at 80 °C for 96 h gave only 57% yield of 3jc and 50%
yield of 3jd, respectively (entries 6 and 7). No obvious trend
was observed in the relationship between substrate sterics and
the enantioselectivity of products. Some substrates showed slight
increases in enantioselectivity compared to 1a (entries 2, 3, 4,
and 8) but slight decreases were observed in other cases (entries
1, 5, 6, and 7).
The ring-opening of 6,7-disubstituted-N-Boc-azabenzonor-
bornadienes 1k-m were also examined (Table 6).
Except for the reaction of 1l with morpholine (2d) which
showed slightly decreased enantioselectivity (entry 5), all
nucleophiles gave the corresponding products in equal or better
enantioselectivity compared to 1a. A difference in reactivity
between 1k and the more electron-rich 1l was not observed.
However, 6,7-difluoro-N-Boc-1m was found to be a less reactive
than 1k and 1l in this reaction. The reactions of 1f with
pyrrolidine (2a) and piperidine (2b) gave lower yields (entries
7 and 8), except for the reaction using N-methylaniline (2f)
(entry 9).
We wanted to expand the scope of the reaction to include
sterically small amines because the resulting compounds were
highly desirable for ongoing medicinal chemistry studies.
However, the rhodium-catalyzed ring-opening of azabicyclic
alkenes with volatile amines such as diethylamine (bp 55 °C)
and dimethylamine (bp 7 °C) was a significant challenge because
of their poor reactivity. For example, in the presence of 1.0
mol % of [Rh(cod)Cl]2 and 3.0 mol % of dppf, the reaction of
The asymmetric ring-opening reactions of 1a with dialkyl-
ammonium iodides were also examined. Surprisingly, the
reaction of 1a with Et2NH2I using a rhodium catalyst generated
from [Rh(cod)Cl]2 and (S,S′)-(R,R′)-C2-ferriphos (4a) instead
of dppf did not give any desired product under the optimized
conditions. We also tested the use of Et2NH2Cl instead of Et2-
(24) See the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006 6841

A R T I C L E S
Scheme 3
Cho et al.
Scheme 6
Scheme 4
Scheme 7
Scheme 5
with other amine nucleophiles. For example, the reaction of 1a
with 1-phenylpiperazine using a rhodium catalyst generated in
situ from 2.5 mol % of [Rh(cod)Cl]2 and 11.0 mol % of 4a,
which contains two dimethylamino groups in its structure,
proceeded at 80 °C for 24 h to give 92% yield of 3ac whereas
the reaction using dppf gave only 6% yield after 48 h under
the same conditions (Scheme 6).
The reaction using PPF-P(t-Bu)2 did not produce the desired
product, and BPPFA, which has a dimethylamino group in its
structure provided a 28% yield of the product. Although a more
detailed explanation of the relationship between the reactivity
of the catalyst and the functionality present on the chiral ligand
is not possible at this point, experimental evidence has shown
that the two dimethylamino groups of 4a somehow participate
in the catalytic cycle and play an important role to promote the
ring-opening reaction.
The proposed catalytic pathway for the rhodium-catalyzed
ring-opening of azanorbornadienes follows closely that put
forward regarding oxabicyclic alkenes25 (Scheme 7). Beginning
with the chiral rhodium complex A, exo-binding can occur to
the nitrogen and the olefin on the substrate 1 to give the
intermediate B, as proposed for the oxabenzonorbornadienes.
Oxidative insertion of rhodium catalyst to C-N bond of 1 forms
C and an SN2′ displacement of the rhodium catalyst by the
nucleophile 2 gives product 3 and regenerates the catalyst.
NH2I according to previous studies on halide effects in the
asymmetric reaction of 1a with other amines (see Table 2). In
the presence of 2.5 mol % of [Rh(cod)Cl]2 and 11.0 mol % of
4a, the reaction of 1a with 5.0 equiv. of Et2NH2Cl and 7.5 equiv.
of Et3N in THP proceeded at 80 °C for 144 h to give 3ak with
85% ee but only in 9% yield (Scheme 3). Under the same
conditions, the reaction of 1a with Me2NH2Cl showed very slow
reaction but gave excellent enantioselectivity. The ring-opened
product 3al was isolated in 60% yield with 99% ee.
A competition experiment revealed that the poor nucleophi-
licity of diethylamine was not due to poisoning of the rhodium
catalyst (Scheme 4). For instance, pyrrolidine was added to the
reaction mixture containing Et2NH2I, yet the reaction proceeded
at 80 °C to give a 65% yield of the pyrrolidine ring-opening
product 3aa. The formation of 3al was not observed.
The nature of the chiral phosphine ligand strongly influenced
the reaction yield in the present reaction. As described above,
in the reaction of 1a with Et2NH2I and i-Pr2NEt, the use of
dppf gave 90% yield of the product while the reaction using 4a
did not proceed. Similar results were observed in the reaction
of 1a with Bn2NH (2e) and Et3NHI as the additive (Scheme 5).
In the presence of 1.0 mol % of [Rh(cod)Cl]2 and 3.0 mol % of
dppf, the reaction of 1a with 1.1 equiv. of Bn2NH and 1.1 equiv.
of iPr2NEt in 1,4-dioxane at 110 °C for 24 h gave the
corresponding product 3ae in 90% yield. However, the reaction
using 4a gave only 25% yield of 3ae under the same condition.
The reactivity trends between dppf and 4a were commonly
observed in the ring-opening reaction of azabenzonorbornadiene
(25) (a) Fagnou, K. In Modern Rhodium-Catalyzed Organic Reactions; Evans,
P. A., Ed.; Wiley-VCH: Weinheim, 2005; Chapter 9, pp 173-190. (b)
Lautens, M.; Fagnou, K. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5455.
9
6842 J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006

Ring-Opening of Azabenzonorbornadienes with Amines
A R T I C L E S
The oxidative insertion of the rhodium catalyst into the
strained carbon-nitrogen bond has been studied in the rhodium-
catalyzed carbonyl insertion of styrenyl aziridines and R-lac-
tams.30 Recently, Tam reported the ruthenium-catalyzed isomer-
ization of 7-oxa/azabenzonorbornadienes to the corresponding
vinyl epoxides and aziridines and a reaction pathway involving
oxidative insertion of the ruthenium catalyst into the strained
carbon-oxygen bond and carbon-nitrogen bond has also been
proposed.31 We attempted to insert carbon monoxide into the
carbon-nitrogen bond of 1a in order to support that an oxidative
pathway to give C in the catalytic cycle (Scheme 9).
Scheme 9
Figure 1. ORTEP drawing for 9.
Scheme 8
In the presence of 5.0 mol % of [Rh(CO)2Cl]2, the CO
insertion into 1a was examined in toluene at 100 °C for 20 h
under a CO atmosphere (1 atm). The isolation of 1-amino-
naphthalene (8) in quantitative yield suggests that insertion of
rhodium occurs but â-elimination is much faster than CO
insertion into the carbon-nitrogen bond.
The regioselective ring-opening of the unsymmetrical aza-
benzonorbornadiene (10) with N-methylaniline (2f) was also
examined (Scheme 10). In the presence of 3.0 mol % [Rh(cod)-
Cl]2 and 9.0 mol % of dppf, the reaction of 10 with 3.0 equiv.
of 2f proceeded in THF at 65 °C for 15 h to give the
corresponding product 11 in 45% yield and aminonaphthalene
12 in 30% yield, respectively. In this reaction, the regioisomer
11 was the sole product arising for selective carbon-nitrogen
bond cleavage at the more highly substituted position of 10.
The regioselective ring-opening of 10 suggests that ionization
of the carbon-nitrogen bond is a key step in the catalytic cycle.
If oxidative insertion were occurring, then ionization of the
tertiary carbon-nitrogen bond should be preferred from an
electronic point of view since the tertiary carbocation 13 will
be more stabilized.
Nucleophilic attack with inversion from the endo-direction
provides the 1,2-trans-diamino product,26 in an SN2′ fashion
relative to the rhodium metal.
On the basis of steric arguments, the coordination of the
rhodium catalyst to the azabenzonorbornadienes should be
favored at the more accessible exo face.27 The exo-coordination
of the aryl rhodium (I) complex to the olefin can be explained
by the stereochemical outcome in the ring-opening of oxabi-
cyclic alkenes with arylboronic acid.28 We were fortunate to
obtain evidence that exo-coordination of the rhodium catalyst
with azabenzonorbornadiene is a possible binding mode (Scheme
8) though we cannot say if this is an intermediate on the reaction
pathway.
Treatment of N-Boc-azabenzonorbornadiene (1a) with [Rh-
(CO)2Cl]2 in CH2Cl2 at room temperature for 12 h produced
the complex 9 whose structure was characterized by X-ray
analysis29 (Figure 1).
Interestingly, extrusion of one CO ligand is observed and the
rhodium complex is isolated in its dimeric form. The lone pair
on nitrogen of 1 facilitates this bidentate binding of the metal
on the exo face of the substrate.
The oxidative insertion of the catalyst into a bridgehead
carbon-nitrogen bond is considered as the enantiodiscriminating
step in the catalytic cycle. As shown in the regioselective ring-
opening of 10 (Scheme 10), the formation of 12 supports
oxidative insertion occurs prior to nucleophilic attack.
Scheme 10
9
J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006 6843

A R T I C L E S
Cho et al.
Figure 2. 31P spectra (at 12×e2P MHz in toluene at 25 °C) of rhodium-phosphine complexes. (a) [Rh(cod)Cl]2 and 4a (1:2) in toluene at 80 °C for 1.5 h.
Complexes 14 and 15 were observed. (b) [Rh(cod)Cl]2 and 4a (1:2) in toluene at 80 °C for 10 h. Complex 14 was observed. (c) [Rh(cod)Cl]2 and 4a (1:4)
in toluene at 80 °C for 12 h. Complex 14 and free 4a were observed.
Scheme 11
If the ring-opening step was coincident with nucleophilic
attack, then the side product 12 should have not been obtained.
Thus, the nature of nucleophile should not influence the
stereochemical outcome of the reaction. However, the stereo-
selectivity is influenced by the amine nucleophile in some cases.
For example, of 2.5 mol % of [Rh(cod)Cl]2 and 11.0 mol % of
4a, the reaction of 1a with piperidine (2b) proceeded at 80 °C
for 60 h to give 83% yield of 3ab in 97% ee while the reaction
using pyrrolidine (2a) gave 53% yield of 3aa in 58% ee under
the same conditions (Scheme 11).
The nucleophile obviously interacts9 with the catalyst in a
manner which affects the stereoselectivity of the reaction. The
specifics of this interaction are unknown at this point. It is well-
known that amines bind to the metal center of the catalyst and
cause deactivation or poisoning.32 However, pyrrolidine did not
cause catalytic poisoning but rather influenced the catalytic
efficiency resulting in low reactivity and enantioselectivity. As
(26) The trans-stereochemistry was also observed in the catalytic ring-opening
reactions of oxabicyclic alkenes. See: refs 5 and 7.
(27) (a) Sun, C. H.; Chow, T. J. J. Chem. Soc., Chem. Commun. 1988, 535. (b)
Sun, C. H.; Chow, T. J.; Liu, L. K. Organometallics 1990, 9, 560. (c) Liu,
L. K.; Sun, C. H.; Yang, C. Z.; Shih, S. Y.; Lin, K. S.; Wen, Y. S.; Wu,
C. F. Organometallics 1992, 11, 972. (d) Chow, T.; Hwang, J. J.; Sun, C.
H.; Ding, M. H. Organometallics 1993, 12, 3762.
(29) The crystals suitable for X-ray were obtained by the addition of hexanes
to a solution of 9 in CH2Cl2 until the solution changed slightly cloudy,
followed by slow evaporation of the solvent mixture.
(28) (a) Lautens, M.; Dockendorff, C.; Fagnou, K.; Malicki, A. Org. Lett. 2002,
4, 1311. (b) Fagnou, K.; Lautens, M. Chem. ReV. 2003, 103, 169.
9
6844 J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006

Ring-Opening of Azabenzonorbornadienes with Amines
A R T I C L E S
Scheme 12
Scheme 13
shown in Table 2, the use of Et3NHCl prevents deactivation of
the catalyst and leads to high yield and excellent enantioselec-
tivity (see Table 2, entry 14). From the experimental evidence
shown in Tables 3 and 4, the six-membered ring amines 2b-
d, dibenzylamine (2e), and the aromatic amines 2f-i may bind
to the catalyst less tightly than pyrrolidine and may be the reason
these reactions showed the best results under the conditions
without the additive and solvent.
Scheme 14
In the present reaction, the experimental results revealed that
the stereochemical outcome is also strongly dependent on the
ratio of the chiral ligand 4a to the Rh (Scheme 12).
In the reaction of 1a with dibenzylamine (2e) at 110 °C for
24 h, the use of 1.5 equiv. of 4a to rhodium (7.5 mol %) gave
78% ee, whereas the use of 2.0 equiv. (10 mol %) gave 90%
ee. Higher enantioselectivity was observed in the reaction using
3.0 equiv. of 4a (15 mol %) which gave 92% ee. Finally, the
use of 2.2 equiv. of 4a (11 mol %) was found to be most
effective (92% ee) for the present asymmetric reaction. Similar
results were observed in the ring-opening of 1a with pyrrolidine
(2a) in Table 2 (entries 6, 7, and 11).
To try to learn more about the beneficial effort of the excess
ligand on the enantioselectivity of the reaction, we undertook
some simple 31P NMR studies to characterize the rhodium-
phosphine complexes in solution (Figure 2 and Scheme 13).
The reaction of [Rh(cod)Cl]2 and 4a (1.0 equiv. to Rh) in toluene
monomer complex based on the larger JRh-P value. The trends
in the coupling constants between monomer and dimer of the
[RhL2Cl]2 type complexes have been investigated by Van
Gaal.34 A mixture of [Rh(cod)Cl]2 and 4a (2.0 equiv to Rh)
was also reacted at 80 °C for 12 h and showed only the doublets
of 14 at 42.3 ppm (JRh-P ) 205.1 Hz) and a peak for free 4a.
Monomeric [RhL2]+X- complexes 16, known to have a smaller
coupling constant of ca. 132 Hz, were not observed.35 For
-
at 80 °C for 1.5 h showed two doublets, 42.3 ppm (JRh-P
)
example, [Rh(dppe)2]+Cl- and [Rh(dppf)2]+BF4 have JRh-P
)
205.1 Hz) for the complex 14 and 43.1 ppm (JRh-P ) 209.4
Hz) for the complex 15 in a ratio of 1:1. After heating the
mixture of 14 and 15 at 80 °C for 10 h, the peak of attributed
to complex 15 disappeared and only the doublet of 14 was
observed. According to previous reports,33 [RhL2Cl]2 type of
dimers showed a large coupling constant of 197 Hz. For
132 and 139 Hz, respectively. Thus, the rhodium catalyst 14 is
generated from [Rh(cod)Cl]2 and 4a in a ratio 1 to 2.
The results indicate that after the formation of the dimer 14
consumed 2.0 equiv. of ligand 4a, 2.0 equiv. of free 4a still
remained. It was shown by 31P NMR that excess ligand does
not lead to a detectable amount of a new complex such as 16,
however excess ligand is still envisioned to play an important
role by interacting with the catalytic intermediates to minimize
catalyst-nucleophile interactions which presumably cause lower
enantioselectivity at the enantiodiscriminating step (Scheme 14).
example, [Rh(binap)Cl]233a and [Rh(dppe)Cl]233b have JRh-P
)
197.0 and 198.5 Hz, respectively. On the basis of the analogy
of the coupling constant with other rhodium complexes already
reported and the stoichiometry, complex 14 should be [Rh(C2-
ferriphos)Cl]2. Also, complex 15 is proposed to be a 14-electron
Conclusion
(30) (a) Calet, S.; Urso, F.; Alper, H. J. Am. Chem. Soc. 1989, 111, 931. (b)
Alper, H.; Urso, F.; Smith, D. J. H. J. Am. Chem. Soc. 1983, 105, 6737.
(c) Murahashi, S.; Horiie, S. J. Am. Chem. Soc. 1956, 78, 4816.
We have developed rhodium-catalyzed asymmetric ring-
(31) Villeneuve, K.; Tam, W. J. Am. Chem. Soc. 2006, 128, 3514.
(32) (a) Wilkinson, G., Ed. ComprehensiVe Organometallic Chemistry, Vol. 5;
Pergamon Press: Elmsford, New York, 1982. (b) Fu, G. C.; Nguyen, S.
T.; Grubbs, R. H. J. Am. Chem. Soc. 1993, 115, 9856. (c) Bazan, G. C.;
Oskam, J. H.; Cho, H.-N.; Park, L. Y.; Schrock, R. R. J. Am. Chem. Soc.
1991, 113, 6899. (d) Schrock, R. R.; Murdzek, J. S.; Bazan, G. C.; Robbins,
J.; DiMare, M.; O’Regan, M. J. Am. Chem. Soc. 1990, 112, 3875.
(33) (a) Hayashi, T.; Takahashi, M.; Takaya, Y.; Ogasawara, M. J. Am. Chem.
Soc. 2002, 124, 5052. (b) Ball, G. E.; Cullen, W. R.; Fry, M. D.; Henderson,
W. J.; James B. R.; MacFarlane, K. S. Inorg. Chem. 1994, 33, 1464. (c)
Kim, T.-J.; Lee, K.-C. Bull. Korean Chem. Soc. 1989, 10, 279.
opening reaction of azabenzonorbornadienes 1 with various
(34) Van Gaal, H. L. M.; van den Bekerom, F. L. A. J. Organomet. Chem.
1977, 134, 237.
(35) (a) Longato, B, Coppo, R.; Pilloni, G.; Corvaja, C.; Toffoletti, A.; Bandoli,
G. J. Organomet. Chem. 2001, 637-639, 710. (b) Di Noto, V.; Valle, G.;
Zarli, B.; Longato, B.; Pilloni, G.; Corain, B. Inorg. Chim. Acta 1995, 233,
165. (c) Casellato, U.; Corain, B.; Graziani, R.; Longato, B.; Pilloni, G.
Inorg. Chem. 1990, 29, 1193. (d) Kunin, A. J.; Nanni, E. J.; Eisenberg, R.
Inorg. Chem. 1985, 24, 1852.
9
J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006 6845

A R T I C L E S
Cho et al.
aliphatic, aromatic, and volatile amines and an efficient route
to the optically active 1,2-diamine derivatives 3 that can be
obtained in high yields and with excellent enantioselectivities
(up to >99% ee). The nature of the activating group on nitrogen
influenced the reactivity of azabicyclic alkenes and the stereo-
chemical outcome of the ring-opening. Good reactivity and
enantioselectivity was generally observed with a variety of
substituted azabenzonorbornadienes. The asymmetric ring-
opening with small amines such as pyrrolidine and dimethyl-
amine give the best results with ammonium hydrochloride
additives, but generally no additives are needed when 4a is used
as the ligand. Also, the combination of R2NH2I and i-Pr2NEt
(optimal ratio 2:3) led to ring-opening reaction of 1a with
volatile amines such as Et2NH and Me2NH, with relatively low
catalyst loadings (1.0 mol % of [Rh(cod)Cl]2. Experiments
revealed that the nature of the chiral ligand has the significant
impact on the reactivity of the catalyst and the use of an excess
amount (2.2 equiv. to Rh) of the chiral ligand plays an important
role to increase the enantioselectivity in the ring-opening
reactions of azabenzonorbornadienes with amine nucleophiles.
Acknowledgment. We thank NSERC, the University of
Toronto, AstraZeneca Research Centre in Montre´al, and Solvias
AG (Basel) for financial support and for various ligands. We
thank Professor Keith Fagnou for early experiments that
confirmed the feasibility of ring-opening with azabicyclic
alkenes. We also thank Mr. Chris Dockendorff for valuable
discussions and useful suggestions during the writing of this
manuscript.
Supporting Information Available: Experimental procedures
and spectral, analytical data for all reaction products (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA0577701
9
6846 J. AM. CHEM. SOC. VOL. 128, NO. 21, 2006