Catalytic asymmetric solid-phase cyclopropanation

Full text of DOI:10.1021/ja005776e

J. Am. Chem. Soc. 2001, 123, 2695-2696
2695
Scheme 1
Catalytic Asymmetric Solid-Phase Cyclopropanation
Tadamichi Nagashima and Huw M. L. Davies*
Department of Chemistry, UniVersity at Buffalo
State UniVersity of New York, Buffalo, New York 14260-3000
ReceiVed NoVember 9, 2000
Solid-phase synthesis of small molecules has been recognized
as an efficient tool to prepare chemical libraries.1 In recent years
a number of C-C bond forming reactions have been applied to
solid-phase reactions.2 A limited number of catalytic asymmetric
reactions have also been applied to substrates bound to solid
supports.3 In this communication, the first example of catalytic
asymmetric cyclopropanations of alkenes on solid support is
described. A major advantage of this process is that the high yield
of cyclopropanation can be achieved even when the alkene is
used as the limiting agent.
products would remain in the liquid phase and would be readily
removed by filtration with appropriate solvents.
The metal-catalyzed cyclopropanation of diazo compounds has
broad utility in organic synthesis.4 One of the major challenges
for efficient intermolecular cyclopropanations is the control of
the high reactivity of the carbenoid intermediate.4 Carbene
dimerization is a very prevalent side reaction, and typically an
excess of trapping agent and syringe pump techniques5 are
required to alleviate this problem. We have found that the
carbenoids derived from aryldiazoacetates are much less prone
to dimerization than the typical carbenoids derived from unsub-
stituted diazoacetates,6 and Rh2(S-DOSP)47 is an exceptional chiral
catalyst for aryldiazoacetates.8 In a recent study, however, directed
toward the asymmetric synthesis of cyclopropyl analogues of
tamoxifen,9 we found that carbene dimerization with an aryldia-
zoacetate could not be completely suppressed. The separation
difficulties associated with this reaction led us to consider a solid-
phase approach whereby an elaborate trapping agent could be
used as the limiting agent. Our studies to develop such an
approach are described here.
As shown in eq 1, a resin-bound olefin 3 was prepared from
the corresponding alcohol 1 and a polystyrene resin with a silicon
linker (PS-DES-SiH resin, Argonaut technologies). The silane
group in PS-DES-SiH resin was chlorinated,11 and resin 1 was
reacted with a 3-fold excess of olefin 2 to give 3. A small portion
of resin 3 was treated with HF-pyridine, and the olefin loading
in resin 3 was estimated based on the crude weight of olefin 2.
The range of the loading level was 0.83-1.0 mmol/g.
An earlier attempt has been made to limit the problems of
carbene dimerization by placing an unsubstituted diazoacetate on
a solid support.10 Rh2(OAc)4-catalyzed decomposition of the
diazoacetate on the solid support in the presence of 20 equiv of
alkyne resulted in the formation of a cyclopropene in 30% yield
after cleavage of the linker.10 In our case, we wished to have the
alkene trap as the limiting reagent and so it was placed on the
solid support (Scheme 1). If the carbenoid were sufficiently
selective, high conversion to the cyclopropane would be possible
by using an excess of the carbenoid source. Any carbene side
To optimize the stoichiometry of the phenyldiazoacetate relative
to the olefin in resin 3, Rh2(S-DOSP)4-catalyzed cyclopropanation
was conducted with 3 and 5 equiv of methyl phenyldiazoacetate
(4a) relative to the olefin in resin 3. Also, samples of resin 3
with different loading levels of alkene were examined to determine
the reproducibility of the reaction. When 3 equiv of the diazo
compound was used, the conversion varied from 75 to 97% (and
a significant amount of olefin 2 was recovered (3-14%)).
However, 5 equiv of the diazo was found to be sufficient to give
quantitative conversion (>99%) of the olefin (Table 1). None of
the olefin was detected in the 1H NMR of the crude mixture after
treating the resin with HF-pyridine. The same reaction conditions,
with 5 equiv of ethyl diaozacetate (EDA), gave a poor conversion
of olefin (35%). Even with 50 equiv of EDA the conversion of
olefin was only 54% and the yield of cyclopropanation was only
8%. The contrasting results obtained with phenyldiazoacetate and
EDA demonstrate the clear advantages of using the more
chemoselective carbenoid.12 The solid-phase cyclopropanation
reaction between phenyldiazoacetate and resin-bound olefin 3
dispalys almost identical stereoselectivities (diastereo- and enan-
tioselectivities) to those obtained in the corresponding reaction
in solution phase. The diastereomer ratio of the cyclopropane 5a
(1) (a) Dolle, R. E.; Nelson, K. H., Jr. J. Comb. Chem. 1999, 1, 235. (b)
Thompson, L. A.; Ellman, J. A. Chem. ReV. 1996, 96, 555.
(2) (a) Burgess, K., Ed. Solid-Phase Organic Synthesis; Wiley-Inter-
Science: New York, 2000. (b) Lorsbach, B. A.; Kurth, M. J. Chem. ReV.
1999, 99, 1549. (c) Franze´n, R. G. J. Comb. Chem. 2000, 2, 195.
(3) For other catalytic asymmetric reactions with the substrate bound to a
solid-support see: Hydrogenation: (a) Ojima, I.; Tsai, C.-Y.; Zhang, Z.
Tetrahedron Lett. 1994, 35, 5785. Dihydroxylation: (b) Han, H.; Janda, K.
D. Angew. Chem., Int. Ed. Engl. 1997, 36, 1731. (c) Xia, Y.; Yang, Z.-Y.;
Brossi, A.; Lee, K.-H. Org. Lett. 1999, 1, 2113. 1,3-Dipolar cycloaddition:
(d) Zou, N.; Jiang, B. J. Comb. Chem. 2000, 2, 6.
(4) Doyle, M. P.; McKervey, M. A.; Ye, T. In Modern Catalytic Methods
for Organic Synthesis with Diazo Compounds; Wiley-Interscience: New York,
1998.
(5) Doyle, M. P.; van Leusen, D.; Tamblyn, W. H. Synthesis 1981, 787.
(6) Davies, H. M. L.; Hodges, L. M.; Matasi, J. J.; Hansen, T.; Stafford,
D. G. Tetrahedron Lett. 1998, 39, 4417.
(7) Rh2(S-DOSP)4 is tetrakis[N-[4-dodecylphenyl)sulfonyl]-(S)-prolinato]-
dirhodium.
(8) (a) Davies, H. M. L. Eur. J. Org. Chem. 1999, 2459, 9. (b) Davies, H.
M. L. Aldrichim. Acta 1997, 30, 105.
(11) Hu, Y.; Porco, J. A., Jr.; Labadie, J. W.; Gooding, O. W.; Trost, B.
M. J. Org. Chem. 1998, 63, 4518.
(12) Davies, H. M. L.; Panaro, S. A. Tetrahedron 2000, 56, 4871.
(9) Davies, H. M. L.; Nagashima, T.; Klino, J. L. Org. Lett. 2000, 2, 823.
(10) Cano, M.; Camps, F.; Joglar, J. Tetrahedron Lett. 1998, 39, 9819.
10.1021/ja005776e CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/24/2001

2696 J. Am. Chem. Soc., Vol. 123, No. 11, 2001
Table 1. Asymmetric Cyclopropanation on Solid Phase
Communications to the Editor
1
1
1
a Olefin was not detected in H NMR of the crude mixture. b Determined by H NMR with DMAP as an internal standard. c Determined by H
NMR of the crude mixture. d 3 equiv of methyl phenydiazoacetate was used.
prepared under solid-phase conditions was 84:16 favoring the trans
isomer. The enantiomeric excess of the major diastereomer was
91% ee.13
The diastereo- and enantioselectivities of the cyclopropanation
reactions are also summarized in Table 1. Except for the results
obtained with 4-CF3-phenyldiazoacetate (entry 5), narrow ranges
of diastereomer ratios were obtained. The diastereomer ratios were
85:15 to 88:12 with Rh2(S-DOSP)4 and were 80:20 to 84:16 with
Rh2(TPA)4. The enantiomeric excess was also in a narrow range
(86-93% ee). Interestingly, an electron-rich diazo (4-MeO-
phenyldiazoacetate, entry 4) and an electron-deficient diazo (4-
CF3-phenyldiazoacetate, entry 5) gave similar enantiomeric excess.
These results may suggest that the electronic property of the
aryldiazoacetate had no significant effect on the enantioselectivity.
In summary, we have developed an effective solid-phase
catalytic asymmetric cyclopropanation in which a trapping agent
(olefin) can be used as the limiting agent.15 Excellent conversions
of olefin on a solid support were achieved with 5 equiv of
electron-rich aryldiazoacetates and a catalytic amount of Rh2(S-
DOSP)4 or Rh2(TPA)4. High enantiomeric excesses of the
cyclopropanes were obtained in these solid-phase reactions, and
the stereoselectivities are comparable to those obtained in solution-
phase reactions. Because it is possible to prepare a variety of
olefins such as 3 on the solid phase, our method would be a useful
tool for constructing a library of aryl-substituted cyclopropane
derivatives. Further studies are in progress to develop a parallel
synthesis of tamoxifen analogues with this solid-phase cyclopro-
panation as a key step.
To determine the range of the solid-phase chemistry, the effect
of substituents on the aryldiazoacetate was examined. Two
7
dirhodium catalysts were used, Rh2(S-DOSP)4 and Rh2-
(triphenylacetate)4 [Rh2(TPA)4].14 The conversion of olefin 3 and
the yield of cyclopropanes 4a-g are shown in Table 1. With Rh2-
(S-DOSP)4, aryldiazoacetates with an electron-rich aryl group gave
quantitative conversion of the olefin (entries 1-4), and the
cyclopropanes were obtained in good to high yields (83-96%).
In contrast to the electron-rich aryldiazoacetates, electron-deficient
aryldiazoacetates gave significantly lower conversions of the
olefin with Rh2(S-DOSP)4 (entries 5 and 6). 4-CF3-phenyldiazo-
acetate gave only 52% conversion of the olefin, and the yield of
cyclopropanation product 4e was poor (16%). With 4-chlorophe-
nyldiazoacetate, the conversion of the olefin was 91% and the
yield of the cyclopropane 4f was moderate (59%). The low
conversion of the olefin with 4-CF3- and 4-Cl-phenyldiazoacetates
may suggest that competing side reactions (e.g. carbene dimer-
ization) are more prevalent when the carbenoid is more electron-
deficient. The cyclopropanations with Rh2(TPA)4 gave consis-
tently good to high conversions of the olefin (89-98%). Similar
to the results obtained with Rh2(S-DOSP)4, high conversions of
olefin were achieved with electron-rich aryldiazoacetates (entries
1-4). Quite interestingly, even with 4-CF3-phenyldiazoacetate,
the conversion is still good (89%), and with 4-Cl-phenyldiazo-
acetate the conversion was nearly quantitative. These results
suggest that the carbenoids derived from the Rh2(TPA)4-catalyzed
reactions are more chemoselective than those derived from the
Rh2(S-DOSP)4-catalyzed reactions.
Acknowledgment. This work was supported by the National Institutes
of Health (CA85641). Partial support of this work by a Johnson & Johnson
Focused Giving Grant is also gratefully acknowledged. We thank
Professor Hiroaki Suga for some helpful discussions.
Supporting Information Available: Experimental data for 1, 3 and
5 (PDF). This material is available free of charge via the Internet at
http://pubs.acs.org.
(13) In the corresponding cyclopropanation of the chloro derivative of 1
with phenyldiazoacetate 4a in solution phase (CH2Cl2 as solvent), the
diastereomer ratio for 7 was 83:17, and the enantiomeric excess of the major
diastereomer was 90% ee.
(14) Rh2(TPA)4 was used as an achiral catalyst instead of the traditional
Rh2(OAc)4 because the equivalent solution-phase cyclopropanations were
cleaner when catalyzed with Rh2(TPA)4 than with Rh2(OAc)4.
JA005776E
(15) In solution-phase competition studies it was found that 1,1-diphenyl-
ethylene was 2.5 times less reactive than styrene in reaction with 3a. Thus, it
is expected that this chemistry can be extended at least to various electron-
rich alkenes.