Atomic Fukui indices16 at the carbon atoms of the reacting
polyenes 6ꢀ8 (see figures on the structures in Scheme 4)
were in line with the data obtained by List, foreseeing a
preferential electrophilic attack of the aldehydes at the
terminal carbon atom of the silyloxy nucleophiles. As for
simple furan 1 and the related bisvinylogous furan candi-
date 3a above disclosed, for which exclusive attack at the
most distant point has been validated, the calculations for
higher homologues 6ꢀ8 also point to the terminal positions
as the most reactive sites, though the values of the compet-
ing nucleophilic positions are more leveled as compared to
3a. Interestingly, as the polyene chain grows in length, the
Fukui indices of the terminal methylenes decrease in the
series, a possible foreshadowing of a decreased reactivity of
these very remote sites. However, other issues could impact
the practicality and selectivity of the reactions, including the
diverse steric congestion of the competing nucleophilic sites
of the donors, the nature of the acceptors, and the mutual,
often unpredictable disposition of the reactants and catalyst
in the transition states.
Encouraged by the computational results, we turned our
attention toward finding the experimental proof of the
concept. In practice, under our optimized conditions,
HVMAR between tetraene 6 and brominated aldehyde
2b was shown to work well and, as was our hope and as
predicted by DFT calculations, θ-hydroxylated butenolide
9 formed in acceptable yield and with complete η-site
selectivity, as a mixture of geometrical isomers in which
the 5Z,20E isomer highly predominated (>91:9 dr). Re-
markably, the enantioselectivity proved very good (>98:2 er)
as demonstrated by HPLC and 1H NMR analyses of the
corresponding (ꢀ)-menthyloxycarbonyl derivative;17 this
highlights once more the superior ability of Denmark’s
catalyst system even with these elongated vinyl nucleo-
philes. Similarly, chiral catalyst (R,R)-4/SiCl4 was success-
fully tried out with 30-substituted tetraene 7 and the
hyperextended pentaene 8, in asymmetric HVMAR addi-
tions to 2b. To our delight, the same reactivity and
remarkable enantiocontrol were observed for methyl-sub-
stituted furan 7, affording carbinol 10 in 73% isolated
yield, which was obtained as a 56:44 mixture of 5Z/20Z and
5Z/20E isomers (>98:2 er for both).
inspection of their high resolution 1D and 2D NMR
1
1
spectra, including detailed Hꢀ H NOESY experiments
and proton coupling constant measurements.18 As for the
assignment of the absolute configuration of the sole stereo-
genic carbon in the butenolide products, we assumed that
all compounds were invariably R-configured as indicated,
based on the experienced stereoinduction trend dictated by
the chiral ligand in the catalyst, featuring a preferential
attack of the incoming nucleophiles at the Re-face of the
aldehyde carbonyls.10ꢀ12 To validate this assumption,
derivatization of butenolide 5db to the corresponding
crystalline menthyl derivative allowed us to definitely
confirm its three-dimensional disposition via single crystal
X-ray analysis (Figure S1 in the Supporting Information
(SI)). By assuming that the sense of stereocontrol exerted
by the catalyst was the same for all the substrates, the R-
absolute configuration for the remaining butenolide can-
didates in this study was assigned by analogy.19
To conclude, we have successfully developed a reliable
catalytic, asymmetric bisvinylogous and hypervinylogous
Mukaiyama-type aldol methodology using easily available
extended furan-based silyloxy polyenes. We have demon-
strated, for the first time, a perfect relay of the enolate
reactivity over a distance of up to five conjugated double
bonds and named this phenomenon “hypervinylogy”.
These novel extended and overextended aldolizations dis-
played excellent enantiocontrol (up to >99:1 er), complete
selectivity at the most remote nucleophilic site of the
substrates, and good to moderate control of the product
geometries. Our findings contrast with the results obtained
by List et al.5 with open-chain bisvinylogous silyl nucleo-
philes, where variable mixtures of ε- and R-substituted
adducts were formed. The extended enantioenriched poly-
ene alcohols formed are rich in functionality and varied in
shape, with a butenolide skeleton flanked by diverse con-
jugated double bonds, a chiral secondary hydroxyl, and
several prochiral centers. This is a prelude for a number of
significant skeletal transformations and challenging op-
portunity in asymmetric synthesis. Work is planned to
address this goal.
ꢀ
Acknowledgment. Funding from Universita degli Studi
Noteworthy, unsubstituted all-trans furan pentaene 8
also proved to be a pertinent partner, providing the
expected hypervinylogous aldol 11 cleanly with complete
ι-site selectivity, while maintaining the excellent enantio-
control exerted by its lower counterparts (>98:2 er).
However, as predicated by our DFT studies, the reduced
reactivity of extended substrate 8 required a higher reac-
tion temperature and a prolonged reaction time for the
process to be acceptably productive.
di Parma is gratefully acknowledged. The authors thank
the Centro Interdipartimentale Misure “G. Casnati”
ꢀ
(Universita di Parma) for instrumental facilities and Dr.
Filippo Romiti (Universita di Parma) for preliminary
ꢀ
experiments.
Supporting Information Available. Experimental pro-
cedures, characterization data, chiral HPLC traces,
NMR spectra for new compounds, crystallographic data.
This material is available free of charge via the Internet at
The geometry of all γ-alkylidene butenolide adducts in
this study was firmly established as shown by direct
(17) For extended butenolides 9ꢀ11, for which no conditions for
(18) See the SI for details.
(19) A detailed 1H NMR investigation on (ꢀ)-menthyloxycarbonyl
derivatives of (E)-5cb and (Z)-5cb allowed us to argue that both
candidates possess the same 30R-configuration; see the SI for details.
Thanks are due to a reviewer for bringing this issue to our attention.
direct measurement of the enantiomeric purity via chiral HPLC analysis
were found, the er values were calculated indirectly via H NMR and
1
HPLC analyses of the corresponding (ꢀ)-menthyloxycarbonyl deriva-
tives; see the SI for details.
Org. Lett., Vol. 13, No. 17, 2011
4741