tuted allenoate 1b was responsive to cyclization delivering
compound 2b in 70% yield (Table 2, entry 2). The reaction
proceeded rapidly (1.5 h) with allenoate 1c, which has elec-
tron-donating groups on the phenyl ring and led to 2c in
84% yield (Table 2, entry 3). The trifluoromethyl-substitut-
ed allenoate 1d cyclized successfully, but this substrate re-
quired a longer reaction time (16 h) and 2d was isolated in
much lower chemical yield (Table 2, entry 4). Under our op-
timized reaction conditions, substrate 1e with the benzyl
group geminal to the tert-butyl ester gave access to an alter-
native structural motif, the tricyclic 3,8-dihydro-1H-indeno-
AHCTUNGTREG[NNUN 2,1-c]furan-1-one (5e) (Table 2, entry 5). With allenoate 1 f
presenting two benzyl groups that could compete for the
Scheme 2. Gold(I)-catalyzed cyclization of 1a with and without Select-
fluor.
À
C C coupling event, the dihydroindenofuranones 2 f and 5 f
were obtained in 70% overall yield (Table 2, entry 6). The
successful cyclization of 1a–1h
prompted us to investigate the
efficiency of the transfer of ste-
reochemical information. Pleas-
Table 1. Optimization studies for oxidative cyclization of 1a.
Entry
Catalyst
Conditions[a]
Product ratio[b]
ingly, for (2S,5S)-1i, which was
prepared enantiopure (enantio-
meric excess (ee) >97%) and
as a single diastereomer, com-
plete axis-to-center chirality
2a/3a[c]/4a/1a (Yield [%])[d]
[e]
1
2
3
4
5
6
7
8
AuCl3
Selectfluor, MeCN, 5 d
Selectfluor, MeCN, 5 d
Selectfluor, MeCN, 24 h
Selectfluor, MeCN, H2O, 24 h
Selectfluor, MeCN, H2O, 4 h
NFSI, MeCN, H2O, 72 h
NFSI, CH2Cl2, 16 h
A, MeCN, H2O, 5 d
tBuOOH, MeCN, H2O, 6 d
Ph2SO, MeCN, H2O, 6 d
26:74:0:0 (2a=20) (3a=41)
no reaction
[e]
SIPrAuNTf2
AuCl
57:43:0:0 (2a=56)
67:33:0:0 (2a=58)
80:20:0:0 (2a=62)
38:62:0:0 (2a=20) (3a=45)
0:0:100:0 (4a=57)
no reaction
AuCl
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
Ph3PAuNTf2
À
transfer for the C O bond
forming event followed by ary-
lation delivered the single dia-
stereomer (8R,8aS)-2i in 80%
yield and >97% ee (Table 2,
entry 9).
9
0:0:50:50
0:0:69:31
10
11
12
13
PhI
N
0:100:0:0[f] (3a=23)
12:88:0:0 (2a=10) (3a=46)
100:0:0:0 (2a=95)
Oxone, MeCN, H2O, 6 d
Selectfluor, MeCN (0.01m), H2O, 4.5 h
Mechanistically, we envisage
initial AuI coordination to the
allene and addition of the pend-
ent tert-butyl ester to afford in-
termediate A, two elementary
steps for which experimental
evidence has been provided by
Hammond and co-workers.[6b,c]
[a] All reactions at RT; catalyst (10 mol%) unless otherwise stated; [1a]=0.15m unless otherwise stated, H2O
(10 equiv), oxidant (2.5 equiv); A=1-fluoro-2,4,6-trimethylpyridinium tetrafluoroborate. [b] Determined by
1H NMR spectroscopy of the crude reaction mixture. [c] (Æ)-(2S,2’S)-3a and (2R,2’S)-3a. [d] Yield of the iso-
lated product where appropriate. [e] 5 mol%. [f] Significant decomposition of 1a was observed.
reaction performed in dichloromethane afforded 57% of
the butenolide 4a as the only product (Table 1, entry 7).[7] 1-
Fluoro-2,4,6-trimethylpyridinium tetrafluoroborate led to re-
covery of the starting material (Table 1, entry 8). Alternative
In one case, oxidative fluorination to the AuIII species B fol-
lowed by Friedel–Crafts-type arylation with fluoride dis-
placement would lead to the auracycle C, an intermediate
susceptible to AuIII/AuI reductive elimination (Path I). A
conjugated addition–reductive elimination mechanism from
B leading to 2a is also plausible (Path II). Alternatively, the
oxidative fluorination may lead to the b-fluorinated g-bute-
nolide D either directly by fluorodeauration of A or by re-
ductive elimination of B. This fluorinated intermediate
could subsequently undergo intramolecular arylation upon
addition–elimination (Path III). g-Butenolide 4a was un-
reactive under the standard oxidative cyclization conditions,
ruling out a mechanism involving the protodeaurated prod-
uct (Scheme 3).
oxidants, such as tBuOOH, Ph2SO, and PhIACTHNURGTNE(GNU OAc)2, did not
lead to 2a, but typically led to 4a or 3a (Table 1, entries 9–
11). With Oxone, compound 2a was isolated in 10% yield
(Table 1, entry 12).[10] Pleasingly, by using Ph3PAuNTf2 with
Selectfluor, the formation of the dimeric butenolides 3a was
completely suppressed upon lowering the reaction concen-
tration from 0.15 to 0.01m. Under these optimized condi-
tions, compound 2a was formed within 4.5 h and isolated in
95% yield (Table 1, entry 13).[12] The structural core of 2a is
surprisingly uncommon,[13] but presents similarities with ca-
dinane sesquiterpenes isolated from Heritiera littoralis.[14]
Experiments to probe the scope of the allenoate sub-
strates are summarized in Table 2. Access to all of the sub-
strates 1a–1i was achieved by the Wittig olefination of ke-
tenes by following literature procedures.[6a,10] The trisubsti-
To support a mechanistic scenario involving the vinyl gold
complex A, this intermediate (A; L=PPh3) was prepared
independently[6b,c] and treated with Selectfluor in CD3CN/
D2O. Pleasingly, the conversion of this complex into 2a and
3a was complete within 2 h with no detectable formation of
4740
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 4739 – 4743