Angewandte
Chemie
our highest yields of 9, we delayed the addition of H2O as a 1:1
mixture in THF until the consumption of PIDAwas complete.
We believe that this sequential addition of H2O prevents the
over-oxidation of 9 by keeping the reaction at the stage of
para-quinone methide 12 until complete consumption of the
oxidant. Subsequent 5-exo-trig cyclization of the 18 alcohol
leads to (Æ)-tanegool (9) after hydration of 13. The anti
relationship between the aromatic ring and the hydroxy-
methyl substituent in 13 reflects a kinetic bias for the
cyclization of rotamer 12a over 12b, in which the para-
quinone methide adopts a conformation that minimizes allylic
strain.
Since the yield of 9 could not be improved beyond 40%
with IIII-based oxidants, we investigated iron trichloride
hexahydrate (FeCl3·6H2O) as an alternative one-electron
oxidant (Scheme 2b). Whereas PIDA likely proceeds through
the ring opening of oxonium ion 10,[26] FeCl3 has the potential
to promote a radical-based mechanism for ring opening via 14
or the corresponding arene radical cation.[10,12] Gratifyingly,
the addition of an aqueous solution of FeCl3·6H2O to
a solution of 8 in acetone at 08C dramatically improved the
mass balance of the reaction and afforded a 59% yield of (Æ)-
tanegool (9) along with its C7’ epimer 15 in 14% yield at 83%
conversion.[27] Selectivity for (Æ)-tanegool could be restored if
the oxidation was buffered with an amine base, and exposure
of pure 9 to aqueous hydrochloric acid (HCl) led to a 2:1
mixture with 15 after 20 h at RT. These results, along with the
complete absence of 15 when 8 is oxidized with PIDA, suggest
a kinetic bias for the formation of 9 during the cyclization/
hydration of quinone methide 12, which then isomerizes to 15
in the presence of acid. Among the four possible diastereo-
mers that can form from quinone methide 12, (Æ)-tanegool
(9) and diastereomer 15 are the only two that have been
reported as natural products.[25,28] Interestingly, both are
isolated in enantioenriched form, thus suggesting that enan-
tioselective cyclization of 12 may occur during biosynthesis.
Whether dirigent proteins can fulfill this function is an
intriguing question, since they have traditionally been asso-
ciated with the oxidative coupling of propenyl phenols[7,8] but
not with the subsequent cyclization of the para-quinone
methides.
While the tanegool diastereomers originate from the
desymmetrization of the S2-symmetric conformation of para-
quinone methide 12, pinoresinol (4) and many related lignans
possess a syn configuration at C2 and C3 of the THF ring
(Scheme 3a), which requires cyclization of a C2-symmetric
para-quinone methide 16. The distinction between isomeric
para-quinone methides 12 and 16 has been discussed in the
context of the direct oxidative coupling of coniferyl alcohol
(1), which favors 16 in the presence of a dirigent protein, but is
nonselective otherwise.[29] In principal, 16 should be accessible
from the oxidative opening of C1-symmetric cyclobutane diol
17 by following our approach. Thus, desymmetrization of
a suitably protected meso-cylobutane diester 18 by enoliza-
tion and protonation affords the trans-diester 19 (Scheme 3b),
which is readily converted to trans-cyclobutane diol 20
following reduction and deprotection. Subsequent oxidation
with FeCl3·6H2O under our previously optimized conditions
affords (Æ)-pinoresinol (4) in 48% yield (87% based on
Scheme 3. Synthesis of (Æ)-pinoresinol. a) Stereochemical analysis of
of tanegool and pinoresinol. b) Bio-inspired synthesis of pinoresinol
(4).
recovered starting material), wherein the relative configura-
tion at C2 and C3 of the cyclobutane is relayed to the cis-fused-
5,5-ring juncture of the natural product. As with the cycliza-
tion of 12a during the synthesis of (Æ)-tanegool (Scheme 2),
the 5-exo-trig cyclization of bis-para-quinone methide 21 is
selective for the rotamer that places the quinone methide in
a sterically favored syn conformation relative to the hydrogen
atom at C2. This gives rise to 22, which undergoes a second 5-
exo-trig cyclization with the expected stereoselectivity. This
completes the synthesis of (Æ)-pinoresinol by installing the
cis-fused 5,5-ring system and underscores the principal
distinction between the oxidation of cyclobutane diols 8 and
20. In the oxidation of 8, a second 5-exo-trig cyclization at the
stage of 13 is either precluded or thermodynamically disfa-
vored by the trans-configuration at C2 and C3, which would
produce a strained 5,5-trans-fused product.
In summary, we have developed a concise approach to the
furanolignans, which hinges on the oxidative opening of
diarylcyclobutane diols. The resulting para-quinone methides,
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!