Organic Letters
Letter
Scheme 2. Retrosynthetic Modes of the Aquatolide Core
Scheme 4. Synthesis of the Phosphonate
To construct the requisite eight-membered ring, alkene 16
was first subjected to hydroboration. This addition reaction
proved to be very slow, indicating considerable steric hindrance.
It appeared that 6 equiv of the BH3·THF complex was required
to obtain a reasonable yield (36%) of a pure product after
oxidation of the borane adduct. Much to our satisfaction, NOE
measurements proved the desired relative configuration of 17,
as strong NOE effects were observed between one of the
methyl groups at 1.21 ppm and the newly introduced proton in
the hydroboration at 2.49 ppm (the other methyl group at 1.05
ppm did not show this NOE effect). One reason for the
disappointing yield of 17 was probably competitive reduction of
the lactone carbonyl group.13
The final part of the total synthesis was determined by the
choice for the intramolecular Mukaiyama-type aldol reaction to
cyclize the eight-membered ring as pioneered by Kocienski.14
Alcohol 17 was oxidized to the aldehyde by using Dess−Martin
periodinane (DMP). The aldehyde was then treated with
ethylmagnesium bromide to give a 58:42 mixture of secondary
alcohols, which was not purified but directly subjected to
hydrogenolysis. The crude product was oxidized with excess
DMP to finally produce ketoaldehyde 18 as a single product in
49% yield over four steps.
Ti(OiPr)4 effected the desired aldol coupling, but the product
aldehyde isolated after acidic workup appeared rather unstable.
It was therefore immediately converted, by using trimethyl
orthoformate and catalytic p-TsOH, into its methyl acetal 11,
which was stable enough for chromatographic purification.9
Alcohol 11 was then esterified with acid 10 (X = H),10 but all
attempts to cyclize the resulting ester under basic conditions
failed. We then decided to examine a Horner−Wadsworth−
Emmons-type cyclization to arrive at the pentenolide 8. Thus,
alcohol 11 was esterified with the phosphonate containing acid
12b, which was readily synthesized as shown in Scheme 4. The
resulting ester 13 was then subjected to excess sodium hydride
under dilute conditions which caused the desired cyclization to
pentenolide 14 in satisfactory yield. Finally, the acetylene was
readily converted into the allene 15 using the Crabbe
homologation.11 This product showed characteristic double
́
1
bond protons in the H NMR spectrum at 6.29 ppm (ring C4
proton), at 5.22 ppm for the internal allene proton, and at
4.83−4.96 ppm for the two terminal allene protons. In the IR
spectrum the allene appeared at 1958 cm−1 and the unsaturated
lactone carbonyl at 1719 cm−1.
The aldehyde 18 was then chemoselectively converted into
methyl acetal 19 in 81% yield under the influence of
cerium(III) chloride as developed by Luche and Gemal.15
Regioselective enolate silylation proceeded cleanly, and the
crude silyl enol ether was directly subjected to cationic
cyclization caused by BF3·OEt2 to provide a mixture of
stereoisomers. This crude mixture was refluxed in toluene in
the presence of p-toluenesulfonic acid, which furnished a single
crystalline product (mp 148−149 °C) in 59% yield over the last
With the substrate 15 in hand, the key photochemical [2 +
2]-cycloaddition was investigated. Irradiation of 15 in a
degassed mixture of acetonitrile/acetone 9:1 with 300 nm
light led to full conversion into a new product within 3 h (see
Scheme 5). In neat acetone, the reaction was complete in 1 h,
and a pure product was obtained in 77% yield after
chromatographic purification. The lactone carbonyl now
resonated at 1771 cm−1, indicative of a saturated γ-lactone.12
The two remaining double-bond protons now appeared in the
NMR spectrum as singlets at 4.71 and 4.70 ppm. These spectral
data were sufficient evidence that the photochemical reaction
provided the desired product 16.
1
three steps. Both the H and the 13C NMR spectra of our
racemic material were identical to those published by Tantillo2
for natural aquatolide (see the full comparison of chemical
shifts in the Supporting Information). The racemate was
resolved on preparative HPLC to provide the pure
enantiomers, which showed [α]D values of +27.7 and −29.0
(c = 1, CHCl3) and mp 180−182 °C, identical for both
Scheme 3. Synthesis of the Substrate for the Photochemical [2 + 2]-Cycloaddition
B
Org. Lett. XXXX, XXX, XXX−XXX