Journal of the American Chemical Society
Communication
Hydrolysis of the enol ether (hydrochloric acid), removal of
the acetate (potassium carbonate, methanol), and silylene
ether formation (di-tert-butylsilyl ditrifluoromethanesulfo-
nate) provided the tricycle 22 (60% over three steps).
Semihydrogenation of the alkyne (palladium−barium sulfate,
dihydrogen) followed by ozonolysis of the resulting alkene
formed the keto aldehyde 6 (85% over two steps).
To obtain the lactone 4, the keto aldehyde 6 was first
exposed to freshly prepared sodium ethoxide in ethanol, which
generated the expected intramolecular aldol addition product
Unfortunately, all attempts to achieve an intermolecular or
directed oxidation of C8 of 4 (and various derivatives) were
20
ultimately unsuccessful. Many of these attempted oxidations
1
9
returned unreacted starting material 4 or proceeded to
(see Table S4). We envisioned employing the primary alcohol
derived from the lactone in a directed oxidation; however, we
were unable to successfully reduce 4. We also attempted an
intramolecular dioxirane insertion using the C9 alcohol as a
21
directing group.
(
not shown) as an uncharacterized mixture of diastereomers
Scheme 4). Activation of the alcohol with methanesulfonyl
In light of these difficulties, an alternative sequence was
devised (Scheme 5). Removal of the silylene protecting group
(tetra-n-butylammonium fluoride, TBAF) followed by site-
selective oxidation of the resulting 1,3-diol (DMP) and silyl
ether formation (tert-butyldimethylsilyl trifluoromethanesulfo-
nate, triethylamine) provided the ketone 28 (67% from 4).
Oxidation of 28 (lead tetraacetate, 88%) generated an α-
(
chloride followed by elimination of the resulting mesylate
afforded the α,β-unsaturated ketone 26 (74% over two steps).
Scheme 4. Synthesis of the Fully Protected Lactone 4
acetoxy ketone (not shown) as a single detectable diastereomer
1
(
H NMR analysis). Cleavage of the resulting acetate
protecting group (potassium carbonate, methanol) formed
the α-hydroxy ketone 29 (88% over two steps). Attempts to
reduce the C9 ketone of 29 using sodium borohydride,
diisobutylaluminum hydride, or Superhydride provided the
undesired trans-vicinal diol predominantly. Consequently, we
employed an α-ketol rearrangement mediated by trimethyla-
9a,b
luminum (as first reported by White and co-workers
for a
similar substrate) to access the isomeric α-hydroxy ketone 30
1
(
90%, single diastereomer, H NMR analysis). Reduction of
the ketone (sodium borohydride, cerium chloride) proceeded
with 4:1 diastereoselectivity at C8. We found that the major
diastereomer in the reduction step underwent partial trans-
lactonization upon purification. Consequently, the unpurified
reduction product was protected (p-toluenesulfonic acid, 2,2-
dimethoxypropane) to provide the stable bis(acetonide) 31
13
(
61% over two steps). Removal of the silyl ether (TBAF)
followed by exhaustive reduction of the lactone (lithium
aluminum hydride, LAH) afforded the triol 32 (66% over two
steps). All attempts to reduce the lactone within the silyl ether
3
1 resulted in recovery of starting material, presumably
because of the steric encumbrance introduced by the silicon
substituents.
Finally, removal of the protecting groups (aqueous acetic
acid, 85 °C) afforded euonyminol (3). Because of the polarity
of 3 and the lack of a chromophore to guide purification, the
unpurified sample was subjected to exhaustive acetylation
2
2
(
acetic anhydride), to provide, following normal-phase
chromatographic purification, euonyminol octaacetate (33)
in analytically pure form (60% over two steps). Spectroscopic
data for synthetic 33 obtained in this way were identical in all
acetate residues (sodium methoxide, methanol, >99%) then
provided 3. To the best of our knowledge, spectroscopic data
for natural euonyminol (3) have not been published. White
and co-workers reported proton chemical shifts for synthetic 3
1
,2-Addition of methyllithium proceeded with 9:1 diaster-
1
eoselectivity ( H NMR analysis) to provide the alcohol 27 in
0% yield. Oxidation with dimethyldioxirane proceeded
smoothly to afford the epoxide 5 as a single detectable
diastereomer ( H NMR analysis). Nucleophilic cleavage of the
1
3
9
1
9a,b
methyl ester (lithium chloride, 130 °C) proceeded with
opening of the epoxide in situ. Protection of the resulting
vicinal diol (p-toluenesulfonic acid, 2,2-dimethoxypropane)
generated the lactone 4 (68% over two steps). X-ray
crystallographic analysis of 4 served to confirm its structure
as well as the stereoselectivity in the epoxidation step (27 →
discrepancy is not known but the polar nature of 3 may render
its chemical shifts sensitive to concentration and impurities.
We eluted our sample over an ion-exchange resin as in White’s
9a,b
report,
but the NMR spectroscopic data were unchanged.
1
3
Unfortunately, White and co-workers did not disclose
C
5).
NMR data or graphical reproductions of their proton NMR
7
02
J. Am. Chem. Soc. 2021, 143, 699−704