ring can therefore dramatically affect the pharmacodynamic
and pharmacokinetic profile of flavanones, justifying a
systematic scrutiny of this issue also in flavones, by far the
most common type of prenylflavonoids and also their more
diverse group in terms of bioactivity.9
respectively) are available by Robinson synthesis from their
corresponding diprotected, meta-substituted phloracetophe-
none precursors (5 and 7, respectively). This operation
translates differences in terms of relative orientation of the
free hydroxyl and the substituent (ortho in 5 and para in 7)
at the acetophenone stage into a C-8- or a C-6-flavone
substitution pattern. We reasoned that the functional asym-
metry of 5 and 7 could be interchanged by a protection swap
between the ortho-hydroxyls and that this could be achieved
through the agency of the orthogonally triprotected inter-
mediate 6 (Scheme 2).
The shortage of systematic studies on prenylflavones is
mainly due to their limited availability by synthesis compared
to other classes of prenylated flavonoids. In general, the direct
prenylation of flavonoids is difficult to control in terms of
chemoselectivity (C- vs O-alkylation), regiochemistry, and
number of prenyls introduced, and the issue is better
addressed at the stage of their acetophenone precursors.
However, the harsh conditions required for the elaboration
of acetophenones into flavones (Robinson synthesis)10 greatly
limits the application of this strategy to prenylated deriva-
tives.11 The development of a straightforward and tunable
entry into C-6 and C-8 prenylflavones from a common
precursor would therefore represent a significant addition to
the field, paving the way to systematic structure-activity
studies for this class of compounds. To combine proof-of-
concept and relevance, we have investigated the regiodiver-
gent synthesis of cannflavin B (1d), the COX-inhibiting
principle of marijuana,3 and its unnatural C-8 regioisomer
(isocannflavin B, 1e) from a common acetophenone precur-
sor. Legal issues aside, cannflavin B is very difficult to obtain
by isolation because of its low concentration in marijuana
and the occurrence of related phenolics that complicate its
purification.3 Its biomedical potential is therefore still un-
tapped.
Scheme 2. Wessely-Moser Rearrangement of Flavones and
Regiodivergent Retrosynthesis of 8- and 6-Substituted Flavones
(P and P′ ) Orthogonal Protecting Groups)
To address the problem of the complementary synthesis
of C-6 and C-8 prenylflavones, we took inspiration from the
Wessely-Moser rearrangement (Scheme 2).12 This venerable
reaction formally interconverts C-8- and C-6-substituted
flavones (2 and 4, respectively) through a diaroylmethane
intermediate that can freely rotate around the C-4a/C-7 axis
(3).13 The C-8- and C-6-substituted flavones (2 and 4,
(8) Rad, M.; Huemperl, M.; Schaefer, O.; Schoemaker, R. C.; Schleu-
niing, W. D.; Cohen, A. F.; Burggraaf, J. Br. J. Clin. Pharmacol. 2006, 62,
288–296.
(9) For a review, see: Barron, D.; Ibrahim, R. K. Phytochemistry 1996,
43, 921–982.
To put this line of thinking into practice, we had to address
two synthetic issues, namely, (a) the design of a 3-pre-
nylphloracetophenone orthogonally protected at the ortho-
hydroxyls (6), and (b) the development of a milder version
of the Robinson synthesis, compatible with the presence of
acid-labile prenyl groups and the regiodirecting ortho-
protection of the acetophenone precursors.
Within the various combinations of ether (MEM, MOM),
ester (acetate, pivalate), carbamate (BOC, Cbz), and silyl
(TES, TBDMS) protecting groups assayed, the 2,4-TBDMS-
6-pivaloyl combination (10) proved optimal in terms of
introduction, stability during the prenylation step, and
orthogonality. Thus, the TBDMS group was superior to all
(10) There is considerable confusion in the literature regarding the
paternity of the preparation of 4H-chromones from o-hydroxyacetophenones
by reaction with an activated carboxylic acid, a transformation referred to
asRobinson,Allan-Robinson,Baker-Venkataraman,orKostanecki-Robinson
synthesis. The original Robinson reaction involved a thermal, sodium
benzoate catalyzed process, involving o-hydroxyacetophenone and aroyl
anhydrides: Allan, J.; Robinson, R. J. Chem. Soc. 1924, 125, 2192–2195.
A more general two-step protocol based on the base-catalyzed isomerization
of o-acyloxyacetophenones to o-hydroxybenzoylacetophenones followed by
cyclization under acidic conditions was later developed and became known
as the Baker-Venkataraman synthesis: Baker, W. J. Chem. Soc. 1933,
1381–1389. Mahal, H. S.; Venkataraman, K. J. Chem. Soc. 1934, 1767–
1769. While Kostanecki had reported prior to Robinson the formation of
flavones by reaction of acetophenones with alkylated salycilic acids in the
presence of sodium: von Kostanecki, S.; Rozycki, A. Ber. 1901, 349, 102–
109. For sake of clarity, we refer to the transformation of o-hydroxyac-
etophenones to flavones simply as the “Robinson reaction”. For a discussion,
see: Finar, I. L. Organic Chemistry; Longman: London, 1977; Vol. 2, pp
782-784.
(14) MOM-Cl has been reported to selectively diprotect phloracetophe-
none (Bu, X. Y.; Zhao, L. Y.; Li, Y. L. Synthesis 1997, 1246–1248), but,
in our hand, this procedure gave a chromatographically difficult to separate
2.5:1 mixture of di- and triprotected derivatives, in analogy with what
observed with ester and carbamate protecting groups. An improved
procedure has been recently reported: Khupse, R. S.; Erhardt, P. W. J. Nat.
Prod. 2007, 70, 1507–1509.
(11) Alternative strategies are the dehydrogenation of prenylflavanones
to prenylflavones by treatment with iodine in pyridine (Dong, X.; Fan, Y.;
Yu, L.; Hu, Y. Arch. Pharm. Chem. Life Sci. 2007, 340, 372–376. ) or the
use of robust acid-sensitive, but difficult to remove, protecting groups.
(12) Wessely, F.; Moser, G. H. Monatsch. Chem. 1930, 56, 97–105.
(13) Flavone numbering.
2268
Org. Lett., Vol. 10, No. 11, 2008