Organic Letters
Letter
3
c
irradiation of visible light, which provided a facile solution for
the stereoselectivity of our previous work. Herein we
accomplished the asymmetric synthesis of fissistigmatins.
As depicted in Figure 1, we chose to set the construction of
the sesquiterpenoid moiety in the final stage and envisioned a
tributylstannyl-initiated radical cascade via the double 6-exo-
trig cyclization of dienyne 7 followed by the elimination of the
synergistic action of both organocatalysts. As shown in TS1
(Figure 1), the enamine derived from imidazolidinone A4
rendered the electrophile coming from the si face, and the ion
pairing of flavylium with the (R)-TRIP anion made the si face
less steric for the attack of enamine. Importantly, the reaction
provided satisfactory results (90% yield with 98% ee and >20:1
dr) on the gram scale, and two organocatalysts could be
With the reliable supply of flavonoid ent-10, we set out to
execute the construction of the cyclization precursor 7.
Conversion of the aldehyde moiety of ent-10 to nitrile by
successive reduction with DIBAL-H, mesylation of the
resulting alcohol, and substitution of the mesylate with
potassium cyanide produced nitrile ent-9 (82% yield in three
steps). To stereoselectively install the quaternary carbon center
C10′′, a sequential methylation followed by alkylation with 12
was intentionally executed, which gave access to ent-8 in 50%
4
thioether moiety. Despite the concern of regioselectivity and
stereoselectivity, this strategy was appealing because it would
enable forging of the trans-decalin scaffold and two terminal
olefins of fissistigmatin C via consecutive formation of the
C4′′−C5′′ and C6′′−C7′′ bonds. The alkyne and terminal
alkene moieties of 7 could, in turn, be derived from the TBS-
protected alcohol and nitrile moieties of 8 through functional
group manipulations. Next, we surmised that a sequential
alkylation of nitrile 9 would diastereoselectively set up the
quaternary carbon center induced by the stereogenic C1′′ with
a bulky flavonoid fragment. To this end, we came to a hybrid
flavonoid 10, which could be transferred to nitrile 9 by the
derivatization of the aldehyde moiety. In this context, an
enantioselective coupling of aldehyde 11 with 2-hydroxychal-
cone 4 promoted by cooperative organocatalysts driven by
6
yield (83% brsm) 4.5:1 dr. This outcome could be
rationalized by the transition state TS2, in which the smaller
ketenimine anion rested between the bulkiest flavonoid
5
fragment and the linear carbon chain. The electrophile
3c
came from the opposite direction to the flavonoid moiety to
give the desired diastereoisomer ent-8, which was confirmed by
the subsequent elaboration to ent-fissistigmatin C. (vide infra).
The subsequent removal of the TBS of ent-8 with HCl,
oxidation with Dess−Martin periodinane, followed by
visible light was utilized.
As shown in Scheme 1, the synthesis commenced with the
gram-scale synthesis of flavonoid ent-10 by using the
cooperative organocatalyst-promoted (R-TRIP and chiral
imidazolidinone A4) coupling of 2-hydroxylchalcone 4 with
a TBS-protected aliphatic aldehyde 11 driven by visible light.
The stereoselectivity of this reaction originated from the
7
3c
Seyferth−Gilbert homologation smoothly delivered the
alkyne 14. Fortunately, the minor stereoisomer could be
separated after the deprotection of TBS. Eventually, the
reduction of the nitrile to aldehyde mediated with DIABL-H
followed by the Wittig reaction furnished the cyclization
precursor 7.
Scheme 1. Synthesis of the Cyclization Precursor
The cascade radical cyclization of dienyne 7 proved to be
quite challenging. After extensive optimization, the radical
cascade of 7 was successively realized by employing AIBN/n-
Bu SnH via heating in dilute toluene (Table 1, entry 1). To
3
our delight, ent-fissistigmatin C could be isolated, albeit in 10%
yield, from reaction mixtures after HPLC purification, and the
other stereoisomers of ent-fissistigmatin C could not be
1
13
identified. The analytic data ( H and C NMR spectra and
optical rotation) of synthetic ent-fissistigmatin C were in good
products 15a, 15b, and 15c were isolated in 20, 23, and 8%
yield, respectively as the major products, which resembled
3
fissistigmatin D, a congener of fissistigmatins A−C. The
(
C5′′ and C7′′ originated from the boat conformation of I and
II via a double 6-exo-trig cyclization/elimination process (path
a). In this transition state, all substituents occupied the
pseudoequatorial position that ensured the correct stereo-
chemistry of C5′′ and C7′′. The 5/7 fused ring product 15 was
envisioned to be generated via 3-exo-dig cyclization of II (path
b), then the ring opening followed by the 5-exo-dig cyclization/
8
elimination process (path b). Although the cycloheptanyl
radical IV might be directly formed through the 7-endo-trig
cyclization of the vinyl radical to alkene (path c), this process
was kinetically unflavored based on previous observations and
8
mechanistic studies. The survey of reaction conditions for the
radical cascade was thus extensively performed to improve the
regioselectivity of the radical cascade. As shown in Table 1,
B
Org. Lett. XXXX, XXX, XXX−XXX