Conjugated Dienes of Variable Substitution Patterns
A R T I C L E S
Scheme 3. Possible Transition States in Ireland-Claisen Step
Scheme 2. Chirality Transfer in Claisen Rearrangement and
Assignment of Absolute Stereochemistry
formed under kinetic control in >95:5 ratio. This assignment
was made based on proton NMR determination observing the
quartet at δ 3.78 ppm, similar to that observed by Ireland et al.
in their comprehensive study.5b Having verified high diastereo-
selectivity in the enolization step, we considered that the slight
erosion of chirality transfer may be due to destabilizing 1,3-
diaxial interaction in the dominant chairlike transition state F.
Ireland et al.5c suggest that the energy differences between
favored chair- and boat-like transition states can be very small
depending on substitution and the electronic nature of the alkene.
As a result, a competing boat transition state G may become
energetically accessible (Scheme 3). Leakage through the boat-
like transition state would result in formation of the enantiomer
R-21, thereby eroding the enantiomeric excess of S-21 arising
through the chairlike transition state. This explains the modest
chirality transfer, which is unusual for the Ireland-Claisen
rearrangement.
Application. Chiral cyclohexadiene diol derivatives are
versatile intermediates used in the synthesis of many natural
products.11 Enantioselective oxidation of monosubstituted ben-
zenes using P. putidia are very powerful, providing access to
3-substituted-3,5-cyclohexadiene-1,2-diols (see J).16 While highly
useful to produce the 3-substitution pattern, the enzymatic
method is not well-suited to the direct synthesis of 4-substituted-
3,5-cyclohexadiene-1,2-diols. The 4-substitution pattern can be
found in a number of natural products including the ep-
oxyquinoids.17 Because of our interest in synthesizing ambe-
welamide A, we sought a metathesis route to 4-substituted
cyclohexadiene diols. Initial experiments focused on cross-
metathesis to make the cyclohexadiene ring similar to the
methylene-free metathesis, but the allylic oxygen atoms inhibited
this process. For example, direct enyne metathesis between 3,4-
dialkoxy-1,5-hexadienes H and alkynes failed to give the desired
1,3-cyclohexadienes. The tandem sequence of cross-enyne
metathesis and Ireland-Claisen reaction was used tactically to
assemble the unsaturated elements to set the stage for a final
ring-closing metathesis. The intramolecular nature of the me-
tathesis step was expected to overcome the difficulties encoun-
tered in cross-metathesis with allylic ethers present in the alkene
substrate.
20 in 94% ee. In propionates 20, silyl ketene acetal formation
under standard conditions (LiHMDS, TBSCl THF/HMPA, -78
°C) was problematic and resulted in poor conversions. We
overcame these difficulties using the conditions described by
McIntosh (KHMDS/TMSCl, toluene, -78 °C)11 to give more
efficient silyl ketene acetal generation, thereby obtaining the
Claisen product in 82% yield. Esterification using TMSCHN2
gave the methyl ester S-21, which was used for analysis.
Enantiomeric excess determination was performed on 21, which
showed 73% ee (hplc), indicating 78% chirality transfer.12
The absolute stereochemistry of the Claisen product S-21 was
determined on the basis of a Mosher method. Reduction of the
ester to primary alcohol 22 (LiAlH4, Et2O, 96% yield) was
followed by Mosher ester synthesis. The alcohol 22 was reacted
separately with the R*-MTPA and S*-MTPA to provide the
2S,R*-diastereomer 23A and the 2S,S*-diastereomer 23B,
respectively (Scheme 2).13 Using the method devised by
Kobayashi,14 the diastereotopic C1 methylene protons (adjacent
to the C2 stereocenter under scrutiny) are influenced differen-
tially by the Mosher ester chiral center. The empirical model
predicts that if the compound has S stereochemistry, then the
2S,S* diastereomer will show a larger separation of the chemical
shift (∆δ) of the protons Ha and Hb comprising the AB system
whereas the 2S,R* diastereomer will not show a major difference
in the chemical shift of Ha,Hb. This was observed: the AB
quartet for S*,S-22B solved for Ha at δ 4.16 and Hb at δ 4.00
with ∆δ ) 0.16 ppm, whereas the diastereomer R*,S-22A
showed Ha at δ 4.09 and Hb at δ 4.05 (JAB) 10.7 Hz) with ∆δ
) 0.03 ppm.15
The origin of the modest chirality transfer in the Ireland-
Claisen step was examined next. First, high diastereoselectivity
of the enolization step was established in a model ester enolate
using menthyl propionate. Under similar enolization/trapping
conditions to that used in eq 3, the E-silyl ketene acetal was
The tandem metathesis/Claisen sequence was used to access
chiral cyclohexadiene diol 27 (Scheme 4). In this case, the
tandem sequence was followed by a ring-closing metathesis to
form the cyclohexadiene ring system. The starting material,
erythronolactol 24, is readily available in two steps from
isoascorbic acid.18 A series of routine transformations were
(11) (a) Hong, S.; Lindsay, H. A.; Yaramatsu, T.; Zhang, X.; McIntosh, M. C.
J. Org. Chem. 2002, 67, 2042-2055. (b) Hutchinson, J. M.; Hong, S.;
McIntosh, M. C. J. Org. Chem. 2004, 69, 4185-4191.
(12) Similar to that observed: Ho¨ck, S.; Koch, F.; Borschberg, H. J. Tetrahe-
dron: Asymmetry 2004, 15, 1801-1808.
(16) (a) Boyd, D. R.; Bugg, T. D. H. Org. Biomol. Chem. 2006, 4, 181-192.
(b) Johnson, R. A. Org. React. (New York) 2004, 63, 117-264. (c)
Hudlicky, T.; Gonzalez, D.; Gibson, D. T. Aldrichim. Acta. 1999, 32, 35-
62. (d) Hudlicky, T.; Thorpe, A. J. Chem. Commun. 1996, 17, 1993-2000.
(17) (a) Kakeya, H.; Onose, R.; Yoshida, A.; Koshino, H.; Osada, H. J. Antibiot.
2002, 55, 829-831. (b) Williams, D. E.; Bombuwala, K.; Lobkovsky, E.;
De Silva, E. D.; Veranja, K.; Allen, T. M.; Clardy, J.; Andersen, R. J.
Tetrahedron Lett. 1998, 39, 9579-9582. (c) Ernst-Russel, M. A.; Chai, C.
L. L.; Hurne, A. M.; Waring, P.; Hockless, D. C. R.; Elix, J. A. Aust. J.
Chem. 1999, 52, 279.
(13) In these stereochemical descriptors, the asterisk denotes the known chirality
from the Mosher ester.
(14) Tsuda, M.; Toriyabe, Y.; Endo, T.; Kobayashi, J. Chem. Pharm. Bull. 2003,
51, 448-451.
(15) In the few instances where the Mosher method is unsuccessful (see ref
14), poor separation (∆δ ) 0.00-0.05 ppm) in both diastereomers was
observed. In the case of 23A,B, this was not the case as there was found
to be greater chemical shift dispersion in one diastereomeric Mosher ester
(expected range is ∆δ ) 0.14-0.20 ppm) and almost none in the other.
9
J. AM. CHEM. SOC. VOL. 128, NO. 49, 2006 15635