Communication
bulky phosphine ligands such as Sphos (2-dicyclohexylphos-
phino-2’,6’-dimethoxybiphenyl, catalyst C) catalyzed this reac-
tion less efficiently, leading to 2a in 42% yield (entry 6). Inter-
estingly, the use of (Me)4tBuXphos as the ligand (catalyst D) re-
sulted in the formation of a diastereomeric mixture of [4+2]
cycloadducts 4a and 4a’ in high yield (entry 7). These results
indicated that the steric and electronic properties of phosphine
ligands on gold catalysts had a significant impact on the reac-
tion outcome. When the reaction was run at room tempera-
ture using catalyst B, the cycloadduct 4 could also be ob-
served, along with 27% of 2a (entry 9). This suggests that 4
might be the intermediate leading to product 2. Decreasing
the amount of catalyst B to 2 mol% gave 2a in a lower yield
of 59% (entry 10). Changing the solvent to toluene or THF af-
forded 2a also in lower yields of 49 or 33% (entries 11,12, re-
spectively). We then investigated the effect of the protecting
groups on this reaction. Acetyl- (Ac), pivaloyl- (Piv), or benzoyl-
protected (Bz) substrates 5a–c, respectively, resulted in remark-
able decreases in product yields (24–46%), along with the for-
mation of several byproducts (entries 13–15). These results
reveal that the nature of the protecting groups on the alcohol
is also crucial for successful transformation. In this reaction,
a carbonate substrate is converted into 2a more efficiently
than those of esters, possibly due to the enhanced stability of
the carboxyallene intermediate for the carbonate cases.
Scheme 1. Metal-catalyzed [4+3] cycloaddition of furan-allenes.
Scheme 2. Gold-catalyzed [4+2] cycloaddition of furans with in situ
generated allenes.
nish the anthracenones containing
(Scheme 2, eq 1).
a quaternary carbon
Recently, we developed various gold-catalyzed intramolecu-
lar cyclizations of benzene-tethered furan-ynes, initiated by
attack of the furan on the gold-activated alkyne, followed by
furan ring-opening.[13] These initial results make the current
target 3,3-rearrangement/IMDAF reactions with high selectivity
a challenging task. To initiate an allene formation, 3,3-rear-
rangement of the propargyl carbonate moiety in designed
substrates 1 should occur more rapidly than the furan-yne cyc-
lization.[14] Here, we expected that the desired transformation
might be achieved through variation of the different ligands
on gold. To test the feasibility of our hypothesis, we initially ex-
amined the gold-catalyzed reactions of 1,5-furan-yne 1a bear-
ing a phenyl substituent at the alkyne terminus; the results are
shown in Table 1. Treatment of 1a in DCE (1,2-dichloroethane)
with cationic gold(I) complex PPh3AuSbF6, generated in situ, af-
forded only a complex reaction mixture (Table 1, entry 1). To
our delight, the use of a gold catalyst with an N-heterocyclic
carbene (NHC) ligand, after stirring the reaction mixture in DCE
at 508C for 2 h, provided the desired anthracen-1-(2H)-one 2a
in 56% yield, along with 7% of byproduct anthracene 3a
(entry 2). The structure of 2a was unambiguously confirmed
by X-ray crystallography.[15] The results indicated that an inter-
esting 1,2-migration of the phenyl group occurred during the
reaction process, resulting in the assembly of a quaternary
carbon center adjacent to a carbonyl functionality in 2a. De-
composition of the substrate was observed only in the pres-
ence of AgSbF6 (entry 3). When [JohnphosAu(MeCN)]SbF6 (cat-
alyst A) was used (Johnphos=(2-biphenyl)di-tert-butylphos-
phine), the yield of 2a increased to 67%, along with 9% of 3a
(entry 4). Gratifyingly, it was found that gold catalyst B with
tBuXphos (2-di-tert-butylphosphino-2’,4’,6’-triisopropylbiphenyl)
as the ligand was more efficient for this transformation, lead-
ing to 84% yield of 2a, whereas the formation of 3a was com-
pletely inhibited (entry 5). Gold complexes possessing other
Next, the scope of the gold-catalyzed cascade reaction was
evaluated under the conditions shown in Table 1, entry 5 (con-
ditions a); the results are shown in Table 2. We first investigat-
ed the electronic effect of the aryl substituents on the alkyne
terminus. We were pleased to find that a series of electron-
withdrawing groups substituted on the aryl alkynes such as p-
F, p-Cl, p-CF3, and p-CO2Et underwent the cyclization smoothly
to provide the corresponding anthracenones 2b,c,e,f in 64–
86% yields in short reaction times. However, when aryl alkynes
bearing substituents such as p-Br or p-OMe were employed,
the desired products were obtained only in low yields under
conditions a, possibly due to the rapid decomposition of the
starting materials. We then made efforts to re-optimize the re-
action conditions for those particular alkynes. Gratifyingly, we
À
found that simply changing the counterion from SbF6 to OTfÀ
(Tf=trifluoromethanesulfonyl) in the gold catalyst led to 70%
yield of 2g bearing a p-OMe-substituted aryl group (condi-
tions b, 5 mol% tBuXphosAuCl/AgOTf). Under these conditions,
p-Br-substituted 1d afforded 2d in a high yield of 88%. Elec-
tron-rich aryl alkynes with p-Me, p-tBu, 3-OMe, or 3,4,5-(OMe)3
substituents were also well suited under these conditions, fur-
nishing 2h–k in 42–78% yields. A thienyl group was well toler-
ated under conditions b, leading to the desired product 2l in
73% yield. The use of alkynes with an alkyl substituent such as
nBu resulted in a complex mixture of products, in DCE. In con-
trast, when the reaction was carried out in THF, this substrate
isomerized rapidly within 30 minutes at room temperature to
give allenyl carbonate 6m in 69% yield. A cyclopropyl-substi-
tuted alkyne also afforded allene product 6n in 79% yield in
THF. However, extending the reaction time formed a complex
reaction mixture, and the desired anthracen-1(2H)-ones were
not observed in these cases. Cyclization of substrate 1o with
&
&
Chem. Eur. J. 2016, 22, 1 – 7
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!