A simple divergence from asymmetric cyclopropane to lactone annulation

Full text of DOI:10.1021/ja963459v

J. Am. Chem. Soc. 1997, 119, 2735-2736
2735
Raising the temperature to reflux under otherwise identical
conditions allowed complete conversion of scalemic 3 to its
cyclopropane 4. The sluggishness of the cyclization of 3
stemmed from a slow ionization since the chirality of the ligand
and the remaining allylic benzoate represented a mismatch. Proof
for this contention arose by using racemic ligand 2 wherein
cyclization now proceeded at room temperature. Decreasing
the amount of catalyst to 1 mol % of [η3-C3H5PdCl]2 and 2.5
mol % 2 with 2 equiv of DBU at 0 °C to room temperature
overnight gave a 77% yield of 310 whose ee of 96% was
determined by HPLC (Chiralcel AD, 5% 2-propanol/heptane).
Better results were obtained by using sodium hydride as the
base at 0 °C to room temperature with the reaction requiring
only 2.5 h to give 3 in 81% yield and 99% ee.
Applying this last set of reaction conditions to the seven-
membered ring substrate 5a gave the monalkylated product 6a10
in 71% yield and >99% ee9 at -20 °C to room temperature
(eq 3). Initiating the reaction at 0 °C caused a diminishment
in ee to 86%. The five-membered ring substrate 4b showed
the greatest sensitivity to temperature. Initiating the reaction
at -20 °C under otherwise identical reaction conditions gave
6b10 in 75% yield and 92% ee.
A Simple Divergence from Asymmetric
Cyclopropane to Lactone Annulation
Barry M. Trost,* Shinji Tanimori, and Patrick T. Dunn
Department of Chemistry, Stanford UniVersity
Stanford, California 94305-5080
ReceiVed October 3, 1996
The widespread occurrence of lactones and their utility as
building blocks makes a simple protocol for their asymmetric
introduction desirable. The few existent protocols for an
annulation equivalent do not lend themselves readily for catalytic
asymmetric introduction.1 In addition, asymmetric syntheses
of cyclopropanecarboxylic acids almost invariably involve
transition metal catalyzed cyclopropanations of alkenes with
diazo compounds. While the copper-2 and rhodium-3catalyzed
asymmetric cyclopropanations have shown spectacular suc-
cesses, these reactions have limitations.4,5 An appealing alterna-
tive paradigm makes use of the ability to desymmetrize meso
diesters as shown in eq 1, path a, using an asymmetric palladium
catalyzed allylic alkylation.6 In the course of these studies, we
discovered that the asymmetric cyclopropanation7,8 was not
straightforward and that simple choice of malonate derivative
allowed either an asymmetric lactone annulation (eq 1, path b)
or an asymmetric cyclopropane annulation (eq 1, path a).
(3)
Cyclization to the cyclopropanes proved surprising. Using
triphenylphosphine as ligand, (η3-C3H5PdCl)2, and DBU as base,
cyclization proceeded moderately but significant racemization
accompanied ring closure. For example, 3 of 99% ee gave 4
of 42% ee in THF and 64% ee in CH2Cl2. On the other hand,
use of cesium carbonate with a catalyst formed from (η3-C3H5-
PdCl)2 and dppp (1,3-diphenylphosphinopropane) in THF at 0
°C to room temperature to reflux gave 410,11 in 90% yield and
94% ee.9 Control experiments demonstrated that the source of
the surprising racemization did not derive from racemization
of the cyclopropane. The improved conditions obtained for the
final cyclization with minimal racemization were applied to the
seven-membered (i.e., 6a) and five-membered (i.e., 6b) ring
substrates to give the cyclopropanes 7a10,11 and 7b10,12 in 95%
yield (96% ee9) and 58% yield (89% ee9), respectively.
(1)
Initial studies examined the reaction of the dibenzoate 1a with
dibenzyl malonate9 (eq 2). When 2.5 mol % [η3-C3H5PdCl]2
and 10 mol % 2 are utilized with DBU (1,8-diazabicyclo[5.4.0.]-
undec-7-ene) as base at 0 °C, the monoalkylation product 3 was
isolated in 63% yield and the cyclopropane in 22% yield.
Switching to Meldrum’s acid as the pronucleophile, as shown
in eq 4, gave similar results wherein the monoalkylation product
8a,10 mp 154 °C, was isolated in 60% yield and 98% ee.9
Resubjection of 8a to the same conditions, with the exception
being the use of a racemic mixture of ligand 2, gave none of
the cyclopropane 9 and only lactone 10a13 in 75% yield. Use
of achiral ligands, such as triphenylphosphine and triisopropyl
(2)
(1) For lactone annulations, see: Corey, E. J.; Gross, A. W. Tetrahedron
Lett. 1985, 26, 4291. Larock, R. C.; Stinn, D. E. Tetrahedron Lett. 1989,
30, 2767. Jones, G. B.; Huber, R. S.; Chau, S. Tetrahedron 1993, 49, 369.
Pearson, A. J.; Khan, M. N. I. J. Org. Chem. 1985, 50, 5276. For a protocol
using a chiral auxiliary via a cyclobutanone, see: Genicot, C.; Ghosez, L.
Tetrahedron Lett. 1992, 33, 7357 and references cited therein.
(2) (a) Aratani, T. Pure Appl. Chem. 1985, 57, 1839. (b) Pfaltz, A. Acc.
Chem. Res. 1993, 26, 339. (c) Evans, D. A.; Woerpel, K. A.; Hinman, M.
M.; Fual, M. M. J. Am. Chem. Soc. 1991, 113, 726. (d) Mu¨ller, D.; Umbricht,
G.; Weber, B.; Pfaltz, A. HelV. Chim. Acta 1991, 74, 232. (e) Lowenthal,
R. E.; Masamune, S. Tetrahedron Lett. 1991, 32, 7373. (f) Brunner, H.;
Singh, V. P.; Boeck, T.; Altmann, S.; Scheck, T.; Wrackmeyer, B. J.
Organomet. Chem. 1993, 443, C16. (g) Ito, K.; Katsuki, T. Tetrahedron
Lett. 1993, 34, 2661.
(3) Doyle, M. P. In Catalytic Asymmetric Synthesis; Ojima, I., Ed.;
VCH: New York, 1993; Chapter 3, pp 63-99. Adams, J.; Pero, D. M.
Tetrahedron 1991, 47, 1765. O’Malley, S.; Kodadek, T. Organometallics
1992, 11, 2299. Doyle, M. P.; Winchester, W. R.; Hoorn, J. A. A.; Lynch,
V.; Simonsen, S. H.; Ghosh, R. J. Am. Chem. Soc. 1993, 115, 9968. Martin,
S. F.; Spaller, M. R.; Liras, S.; Hartmann, B. J. Am. Chem. Soc. 1994, 116,
4493. Corey, E. J.; Gant, T. G. Tetrahedron Lett. 1994, 35, 5373.
(4) For a Ru catalyst, see: Nishiyama, H.; Itoh, Y.; Matsumoto, H.; Park,
S.-B.; Itoh, K. J. Am. Chem. Soc. 1994, 116, 2223.
(5) For asymmetric Simmons-Smith type reactions, see: Kitajima, H.;
Aoki, Y.; Ito, K.; Katsuki, T. Chem. Lett. 1995, 1113. Charette, A. B.;
Prescott, S.; Brochu, C. J. Org. Chem. 1995, 60, 1081. Imai, N.; Sakamoto,
K.; Takahashi, H.; Kobayshi, S. Tetrahedron Lett. 1994, 35, 7045. For an
asymmetric Kulinkovich reaction, see: Corey, E. J.; Rao, S. A.; Noe, M.
C. J. Am. Chem. Soc. 1994, 116, 9345.
(6) For a review, see: Trost, B. M. Acc. Chem. Res. 1996, 29, 355. Trost,
B. M.; Van Vranken, D. L.; Bingel, C. J. Am. Chem. Soc. 1992, 114, 9327.
Trost, B. M.; Li, L.; Guile, S. D. J. Am. Chem. Soc. 1995, 116, 8746.
(7) For Pd-catalyzed formation of cyclopropanes, see: (a) Michelet, V.;
Besnier, I.; Geneˆt, J. P. Synlett 1996, 215 and earlier references cited therein.
(b) Ba¨ckvall, J. E.; Vagberg, J. O.; Zercher, C.; Geneˆt, J. P.; Denis, A. J.
Org. Chem. 1987, 52, 5430. (c) Hanzawa, Y.; Ishizawa, S.; Kobayashi, Y.
Chem. Pharm. Bull. 1988, 36, 4209. (d) Trost, B. M.; Tometzki, G. B.;
Hung, M. H. J. Am. Chem. Soc. 1987, 109, 2176.
(8) Cf.: Hayashi, T.; Yamamoto, A.; Ito, Y. Tetrahedron Lett. 1988,
29, 669.
(9) All ee values were determined by HPLC (Daicel Chiracel AD or
OD). The dibenzyl esters were utilized prefentially because of ease of
analysis using a chiral HPLC column for determination of ee.
(10) This compound has been fully characterized spectrally and elemental
composition established by high-resolution mass spectrometry and/or
combustion analysis.
(11) Cf. the dimethyl ester in ref 7b.
(12) For methyl ester, see: Burgess, K. J. Org. Chem. 1987, 52, 2046.
(13) Larock, R. C.; Hightower, T. R. J. Org. Chem. 1993, 58, 5298.
Nicolaou, K. C.; Seitz, S. P.; Sipio, W. J.; Blount, J. F. J. Am. Chem. Soc.
1979, 101, 3884.
S0002-7863(96)03459-2 CCC: $14.00 © 1997 American Chemical Society

2736 J. Am. Chem. Soc., Vol. 119, No. 11, 1997
Communications to the Editor
phosphite, in the same reaction also gave 10a in comparable
yields. Adding tert-butyl alcohol to the reaction mixture
produces the tert-butyl ester 10b instead in 52% yield.
(6)
(4)
effectively hydrolyze it. Formation of the O-alkylated product
13 may derive from an isomerization of the initial C-alkylated
product 16. The known reactivity of these vinylcyclopropanes
toward Pd(0)12,16 and the special activation toward ring opening
of cyclopropanes provided by Meldrum’s acid17 make this
scenario quite reasonable.
The racemization accompanying formation of the cyclopro-
pane indicates this seemingly straightforward reaction is much
more complex. The simplest explanation invokes reversible
cyclization to a symmetrically bridged bicyclic system 16 (eq
7, path b) by attack at Cb competitive with formation of
cyclopropane 17 (path a) by attack at Ca. For n ) 2, the
Switching to the dicarbonate 1b in the absence of base led
directly to the lactone 10c but only in 30% yield (83% ee). Use
of (η3-C3H5PdCl)2 as catalyst precursor gave a slightly improved
ee to 90%, but still only in 43% yield. The reaction consists
of two stagessasymmetric alkylation and cyclization. For the
first step, use of potassium carbonate as base proved beneficial
to maximize ee; however, it did not appear to be satisfactory
for the cyclization step. The latter was improved by using a
tertiary amine base in the presence of methanol. The best
conditions involved performing the alkylation with 1 mol %
(η3-C3H5PdCl)2, 2.5 mol % 2, and 0.5 equiv of potassium
carbonate with 3 equiv of methanol in THF at 0 °C. After 1 h,
0.5 equiv of (diisopropylethyl)amine was added, the reaction
was allowed to warm to room temperature and subsequently
heated to 55 °C. By this method, crystalline lactone 10a,10,14
mp 98-99 °C was obtained in 70% yield and >99% ee.9
The sluggishness of the reaction for the seven-membered
substrate 11b required the alkylation to be performed at 50 °C
(eq 5). At this temperature, use of 0.5 equiv of potassium
carbonate as base even in the absence of methanol generated
the lactone 13a10 in satisfactory yield (70%) and ee (96%9).
(7)
considerably higher strain of 17 compared to that of 16,
according to molecular mechanics, would suggest the reverse
scenario (i.e., equilibration will favor 16 not 17). Vinylcyclo-
propanes such as 17 are known to be substrates for Pd(0).16
Furthermore, these calculations suggest that bridged bicycles
like 16 (n ) 2 or 3) should be observable, but they have not
been detected. Thus, a much more complicated mechanism
must be involved. The speculative nature of any such sugges-
tions leads us to postpone any further comment at this time.
Thus, we have established a very simple protocol for facile
formation of either a cyclopropane or a lactone both with
exceptionally high ee. This sequence constitutes a new
paradigm for asymmetric cyclopropanation that complements
the traditional carbenoid additions to alkenes. Lactone annu-
lations are usually multistep processes and require chiral
scalemic starting materials to be asymmetric. Thus, this one-
pot protocol from readily available achiral starting materials in
a catalytic asymmetric reaction makes these valuable building
blocks much more accessible. The widespread presence of
lactones in target molecules and their importance as building
blocks (e.g., recently for carbanucleoside synthesis)18 imparts
special significance to this new reaction.
(5)
The five-membered ring substrate 11b (eq 5) proved to be
exceptionally reactive. When the reaction was performed as
for 1b, 11b produced the lactone 12b having only 76% ee.
Lowering the temperature improved the ee. The insolubility
of potassium carbonate made it impractical to employ it at low
temperature. Switching to methanolic sodium methoxide al-
lowed the reaction to be performed at -20 °C. After warming
to room temperature, (diisopropylethyl)amine was added, and
the reaction was allowed to proceed to completion. The bicyclic
lactone 12b10 was obtained in 54% yield and 96% ee.9
The formation of the lactone in the reactions with Meldrum’s
acid is unexpected.15 It presumably arose from an unprec-
edented O-alkylation as shown in eq 6. Formation of the
O-alkylated product 13 may trigger a retro-Diels-Alder reaction
to form the acyl ketene 14 which may then be trapped by alcohol
to give the observed lactone 15 (path a). Alternatively, the
O-alkylated product 13 may react directly with the alcohol to
give 15 (path b). The mildness of the reaction conditions and
precedent with related systems suggest path a to be less likely.
Either pathway suggests that any R1 group may be introduced.
Indeed, as shown in eq 4, adding tert-butyl alcohol to the
cyclization of benzoate 8a gave the tert-butyl ester. Addition
of no alcohol led to trapping by water (R1OH ) H2O) followed
by decarboxylation to give the parent lactone. The sensitivity
of 13 toward alcoholysis was illustrated by the fact that methanol
released from the carbonate leaving group was sufficient to
Acknowledgment. Dedicated to the memory of Patrick T. Dunn,
deceased 5/19/95. We thank the National Institutes of Health, General
Medical Sciences Institute, for their generous support of our research.
Mass spectra were kindly provided by the Mass Spectrometry Regional
Center of University of California-San Francisco, supported by the
NIH Division of Research Resources.
Supporting Information Available: Sample experimental proce-
dures and characterization data for 3, 4, 6a,b, 7a,b, 10c, 12a,b (4 pages).
See any current masthead for ordering and Intenet access instructions.
JA963459V
(15) For a review, see: Chen, B. C. Heterocycles 1991, 32, 529.
(16) Morizawa, Y.; Oshima, K.; Nozaki, H. Tetrahedron Lett. 1982, 23,
2871. Shimizu, I.; Aida, F. Chem. Lett. 1988, 601. Shimizu, I.; Ohashi, Y.;
Tsuji, J. Tetrahedron Lett. 1985, 26, 3825.
(17) Danishefsky, S. Acc. Chem. Res. 1979, 12, 66.
(18) Marschner, C.; Gaumgartner, J.; Griengl, H. J. Org. Chem. 1995,
60, 5224. Akella, L. B.; Vince, R. Tetrahedron 1996, 52, 2789.
(14) Pearson, A. J.; Khan, M. N. I.; Clardy, J. C.; Cun-heng, H. J. Am.
Chem. Soc. 1985, 107, 2748.