Regio- and diastereoselective tandem rhodium-catalyzed allylic alkylation/Pauson-Khand annulation reactions [5]

Full text of DOI:10.1021/ja015531h

J. Am. Chem. Soc. 2001, 123, 4609-4610
4609
Table 1. Development of the Sequential Rhodium-Catalyzed
Allylic Alkylation/Pauson-Khand Annulation
Regio- and Diastereoselective Tandem
Rhodium-Catalyzed Allylic Alkylation/
Pauson-Khand Annulation Reactions
P. Andrew Evans*,† and John. E. Robinson†
Department of Chemistry, Indiana UniVersity
Bloomington, Indiana 47405
ReceiVed January 10, 2001
The construction of complex polycyclic systems using transition
metal-catalyzed annulation reactions provides a powerful strategy
for target directed synthesis.1 The Pauson-Khand (PK) annulation
is representative of this class of transformations in which a
tethered eneyne undergoes a formal [2 + 2 + 1] reaction to
furnish a bicyclic cyclopentenone.2,3 A significant limitation with
this process is the necessity for intramolecularity in order to
suppress competitive intermolecular metal-catalyzed reactions.
However, a recent study provided a new approach to this problem,
which utilized a dual catalytic system to facilitate a one-pot
palladium-catalyzed allylic alkylation followed by a rhodium-
catalyzed PK annulation reaction.4
We envisioned an alternative approach to this problem that
utilized a single metal-catalyst to facilitate both transformations
in a tandem sequence, using only the reaction temperature to
modulate the catalytic activity. The obvious advantage of this
strategy was the ability to significantly increase the molecular
complexity of the bicyclic adduct through the introduction of a
stereogenic center at C-2, which we anticipated would control
diastereoselectivity in the PK annulation.5 Herein, we now
describe the development of the regio- and diastereoselective
rhodium-catalyzed tandem allylic alkylation/Pauson-Khand an-
nulation reaction using stabilized carbon and heteroatom nucleo-
philes (eq 1).
2°:1° yield ds yield
entry
catalysta
additive solvent 4:5b,c (%)d 2:3e (%)d
1
2
3
4
5
6
7
8
9
[RhCl(CO)dppp]2 None
THF
toluene 12:1 47
AgOTf THF 16:1 58
AgOTf toluene 15:1 59
31:1 86
3:1 40
5:1 80
5:1 63
5:1 35
1:1 57
"
"
"
"
"
"
"
None
DMSO 18:1 57
DME
MeCN 37:1 88
"
"
"
"
"
17:1 36 10:1 34
7:1 87
3:1 12
4:1 18
5:1 20
RhCl(CO)(PPh3)2
RhCl(CO)dppe
"
"
"
3:1 72
8:1 36
8:1 62
10 [RhCl(CO)dppb]2
a All reactions were carried out on a 0.25 mmol reaction scale using
10 mol% rhodium and 1.2 equiv of the propargyllic nucleophile.
b Ratios of regioisomers were determined by capillary GLC on aliquots
of the crude reaction mixture. c The primary product 5a was prepared
independently Via Pd(0) catalysis.8 d GLC yields. e Ratios of diaste-
reoisomers were determined by capillary GLC.
modified Wilkinson’s catalyst (Rh(PPh3)3Cl) that had proven
effective for the allylic substitution reactions6 furnished only trace
amounts of the annulation adducts 2a/3a. Hence, we decided to
reexamine the allylic alkylation with rhodium catalysts that had
proven effective for the PK reaction. Interestingly, although the
PK catalysts were significantly different, we reasoned the strong
π-acidity of a carbon monoxide ligand may be sufficient to attain
high regioselectivity in the allylic substitution.6a Treatment of the
allylic carbonate 1 with the sodium salt of the R-branched
malonate and [RhCl(CO)dppp]2, in tetrahydrofuran and toluene
respectively, furnished alkylation products 4a/5a with g12:1
regioselectivity (entries 1-2). The corresponding PK reaction with
4a furnished the bicyclic adducts 2a/3a with modest diastereo-
selectivity, favoring the cis-adduct 2a.
Preliminary studies examined the feasibility of the metal-
catalyzed multicomponent annulation reaction (Table 1). We
anticipated that the tandem process would evolve from the
combination of the rhodium-catalyzed allylic substitution6 with
the PK annulation.3c,4 Initial efforts with the trimethyl phosphite
Initial efforts to improve catalyst turnover and selectivity,
through an in situ counterion exchange with AgOTf, resulted in
reduced turnover and similar selectivity irrespective of the solvent
(entries 3-4). The disparity between the solvents prompted the
examination of alternatives that could potentially facilitate both
transformations with improved turnover and selectivity. This study
provided acetonitrile as the potential medium to facilitate the
tandem reaction (entries 5-7). Finally, a series of electronically
similar rhodium catalysts, in which the steric environment of
phosphine was systematically altered, were examined to determine
the optimum catalyst (entries 8-10).
The ability to catalyze both transformations with [RhCl(CO)-
dppp]2 in acetonitrile, albeit at different reaction temperatures,
allowed the development of the tandem process (Table 2).
Treatment of the allylic carbonate 1 with the sodium salt of the
R-branched dimethyl malonate derivative and [RhCl(CO)dppp]2
† This work was initiated at the Department of Chemistry and Biochemistry,
University of Delaware, Newark, DE 19716.
(1) Ojima, I.; Tzamarioudaki, M.; Li, Z.; Donavan, R. J. Chem. ReV. 1996,
96, 635.
(2) For recent reviews on the Pauson-Khand reaction, see: (a) Schore, N.
E. In ComprehensiVe Organometallic Chemistry II; Hegedus, L. S., Ed.;
Pergamon: Oxford, 1995, p 703. (b) Geis, O.; Schmalz, H. G.; Angew. Chem.,
Int. Ed. 1998, 37, 911. (c) Jeong, N. In Transition Metals in Organic Synthesis;
Beller, M., Bolm, C., Eds.; Wiley-VCH: Weinheim, 1998; Vol. 1, p 560. (d)
Buchwald, S. L.; Hicks, F. A. In ComprehensiVe Asymmetric Catalysis;
Jacobsen, E. N., Pfalz, A., Yamamoto, H., Eds.; Springer: Berlin, 1999; Vol.
2, p 491. (e) Brummond, K. M.; Kent, J. L. Tetrahedron 2000, 56, 3263.
(3) (a) Co: Pagenkopf, B. L.; Livinghouse, T. J. Am. Chem. Soc. 1998,
120, 2285. (b) Ir: Shibata, T.; Takagi, K. J. Am. Chem. Soc. 2000, 122, 9852.
(c) Rh: Koga, Y.; Kobayashi, T.; Narasaka, K. Chem. Lett. 1998, 249. Jeong,
N.; Lee, S.; Sung, B. K. Organometallics 1998, 17, 3642. Jeong, N.; Sung,
B. K.; Choi, Y. K. J. Am. Chem. Soc. 2000, 122, 6771. (d) Ru: Morimoto,
T.; Chatani, N.; Fukumoto, Y.; Murai, S. J. Org. Chem. 1997, 62, 3762. Kondo,
T.; Suzuki, N.; Okada, T.; Mitsudo, T. J. Am. Chem. Soc. 1997, 119, 6187.
(e) Ti: Hicks, F. A.; Buchwald, S. L. J. Am. Chem. Soc. 1996, 118, 11688
and references therein.
(4) Jeong, N.; Seo, S.-D.; Shin, J.-Y. J. Am. Chem. Soc. 2000, 122, 10220.
(5) For a related diastereoselective Pauson-Khand reaction using cobalt,
see: Breczinski, P. M.; Stumpf, A.; Hope, H.; Krafft, M. E.; Casalnuovo, J.
A.; Shore, N. E. Tetrahedron 1999, 55, 6797.
(6) (a) Evans, P. A.; Nelson, J. D. J. Am. Chem. Soc. 1998, 120, 5581. (b)
Evans, P. A.; Robinson, J. E.; Nelson, J. D. J. Am. Chem. Soc. 1999, 121,
6761. (c) Evans, P. A.; Leahy, D. K. J. Am. Chem. Soc. 2000, 122, 5012. (d)
Evans, P. A.; Kennedy, L. J. J. Am. Chem. Soc. 2001, 123, 1234.
10.1021/ja015531h CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/19/2001

4610 J. Am. Chem. Soc., Vol. 123, No. 19, 2001
Communications to the Editor
Table 2. Scope of the Regio- and Diastereoselective Tandem
Rhodium-Catalyzed Allylic Alkylation/Pauson-Khand Annulation
Reaction (eq 1)a
the substituent at C-2. Insertion of the metal then leads to the
irreversible formation of the diastereomeric metallacycles iii/iv,
which are poised to undergo migratory insertion of metal bound
carbon monoxide, followed by reductive elimination to afford the
bicyclic adducts 2/3. Hence, provided the initial coordination is
reversible, increasing the size of the substituent at C-2 should
significantly improve the diastereoselectivity.
entry
X
M
R
2°:1° 4:5b ds 2:3c yield (%)d
1
2
3
4
5
6
7
8
9
(MeO2C)2C Na
H
a
b
c
d
e
f
g
h
i
27:1
19:1
11:1
20:1
32:1
57:1
5:1
5:1
6:1
9:1
3:1
6:1
82
80
78
79
84
81
63
73
81
"
"
"
"
Li
"
"
Li
"
Me
Ph
H
Me
Ph
H
TsN
O
"
"
7:1
g19:1
g19:1
g19:1
"
"
Me
Ph
7:1
8:1
"
a All reactions were carried out on a 0.5 mmol reaction scale using
5-6 mol % of [RhCl(CO)dppp]2 and 1.2 equiv. of nucleophile. b Ratios
of regioisomers were determined by HPLC and capillary GLC. c Ratios
of diastereoisomers were determined by 400 MHz 1H NMR. d Isolated
yields.
To test this hypothesis, the C-2 methyl group was replaced
with a 2-naphthyl substituent. This study also provided an
opportunity to examine the enantiospecificity of the allylic
amination with this particular catalyst and thus ascertain whether
the allylic substitution was consistent with our previous studies.6
Treatment of the chiral nonracemic allylic carbonate (S)-6 (g
99% ee) under the optimized reaction conditions furnished the
azabicycles 7/8 in 82% yield, as a 43:1 mixture of diastereo-
isomers favoring 7 (eq 2).9 Furthermore, the allylic amination is
highly enantiospecific (98% cee) and proceeds with retention of
absolute configuration.10
In conclusion, we have developed a new regio- and diastereo-
selective tandem rhodium-catalyzed allylic substitution/Pauson-
Khand annulation reaction. This study demonstrates that excellent
regioselectivity may be obtained with additional rhodium catalysts,
provided there is a strong π-acidic ligand present on the metal
center. The annulation reaction was also applied to chiral
nonracemic allylic carbonate, which is likely to have significant
utility for asymmetric synthesis.
Figure 1.
Acknowledgment. We sincerely thank the National Institutes of
Health (GM58877) for generous financial support. We also thank Zeneca
Pharmaceuticals for an Excellence in Chemistry Award, Eli Lilly for a
Young Faculty Grantee Award, GlaxoWellcome for a Chemistry Scholar
Award, and Novartis Pharmaceuticals for an Academic AchieVement
Award. The Camille and Henry Dreyfus Foundation is also thanked for
a Camille Dreyfus Teacher-Scholar Award (P.A.E.). Professor Paul A.
Wender is acknowledged for stimulating discussions.
in acetonitrile at 30 °C, under an atmosphere of carbon monoxide,
furnished the enynes 4a/5a favoring the secondary product 4a
(2°:1° ) 27:1 by GLC). The reaction mixture was then heated at
reflux for ∼24 h, resulting in the formation of the bicyclic
cyclopenteneones 2a/3a in 82% yield, as a 5:1 mixture of
diastereoisomers favoring 2a (entry 1). The relative configuration
of the major isomer was established by NMR (nOe), which
confirmed the cis-relationship of the C-1 bridgehead proton with
the C-2 methyl group. This represents the first regio- and
diastereoselectiVe rhodium-catalyzed tandem allylic alkylation/
Pauson-Khand annulation reaction. Table 2 summarizes the
application of this method to substituted acetylenic stabilized
carbon and heteroatom nucleophiles. This study indicates that the
nitrogen nucleophiles provide optimum secondary substitution
(TsNRLi > E2CRNa > ROLi), while the diastereoselectivity, with
the exception of the oxygen derivatives 2g-i,8 is relatively
consistent throughout the series.
Supporting Information Available: Spectral data for 2a-i and 7
(PDF). This material is available free of charge via the Internet at
http://pubs.acs.org.
JA015531H
(9) RepresentatiVe Experimental Procedure: Lithium hexamethyldisilyl
azide (575 µL, 0.575 mmol, 1.0 M solution in THF) was added dropwise to
a solution of p-toluenesulfonyl propargylic amine (126 mg, 0.6 mmol) in
anhydrous MeCN (3.0 mL) at 30 °C under a CO atmosphere. The anion was
allowed to form over ca. 20 min, then added Via Teflon cannula to a solution
of [RhCl(CO)dppp]2 (31.4 mg, 0.03 mmol) in anhydrous MeCN (1.0 mL) at
30 °C, rinsing with anhydrous MeCN (2 × 0.5 mL). The optically active
allylic carbonate 6 (119 mg, 0.5 mmol; g99% ee by capillary GLC) was then
added from a tared vial and the reaction mixture heated at 30 °C for ∼5 h
(TLC control; 2°:1° g99:1 by HPLC). The reaction mixture was heated at
reflux under a CO atmosphere for ∼24 h (TLC control). The reaction mixture
was concentrated in Vacuo and purified by flash chromatography (eluting with
60% ethyl acetate/hexane) to furnish the azabicycles 7/8 (0.162 g, 82%) as a
white crystalline solid, with 43:1 diastereoselectivity (by crude HPLC analysis)
favoring 7 (98% cee).6b
The diastereoselectivity in the rhodium-catalyzed Pauson-
Khand annulation is consistent with the mechanistic hypothesis
outlined in Figure 1. Initial complexation of the enyne 4
presumably results in a mixture of diastereomeric complexes
(i/ii), where the relative population is influenced by the size of
(10) The absolute configuration for the rhodium-catalyzed allylic amination
was assigned by comparison of (S)-II prepared from (S)-6 with (R)-II prepared
Via Mitsunobu inversion of the allylic alcohol (S)-I.
(7) Tsuji, J. In Palladium Reagents and Catalysts; Wiley: New York, 1996;
Chapter 4, pp 290-404. For a recent review on the transition metal-catalyzed
allylic alkylation, see: Trost, B. M.; Van Vranken, D. L. Chem. ReV. 1996,
96, 395.
(8) The excellent diastereoselectivity observed in this particular series is
presumably due to reduced A1,3-strain between the C-2 alkyl substituent and
ring oxygen (x ) 0).