Communication
ion with the diene could deliver a gold carbene intermediate
that would then ring-close with the unreacted alkene unit to
produce a fused polycyclic structure featuring a bridgehead
alkene moiety (Scheme 2, path b). This approach would not
only allow rapid and unprecedented access to highly strained
polycyclic bridgehead olefins, but would also further highlight
the potential of gold catalysis for the generation of structural
complexity from simple starting materials.[9]
present transformation (Table 1, entries 7–9). Lowering the re-
action temperature led to a lower yield of 3a without improv-
ing the diastereoisomeric ratio (Table 1, entry 10). Finally, the
loading of AuCl·SMe2 (11) could be reduced to 2 mol% and
diene 5a used in lower excess (3 equiv). Under these opti-
mized conditions, bridgehead olefin 3a could be isolated in
87% yield as a 9:1 mixture of diastereoisomers (Table 1,
entry 11).
o-Alkynylbenzylaldehyde 4a[10] and conjugated diene 5a[11]
were initially chosen as model substrates for our study
(Scheme 3). We were aware that their reaction in the presence
of a gold catalyst might lead to some selectivity issues. Indeed,
we found that the desired reaction pathway leading ultimately
to the polycyclic bridgehead alkene 3a could be in competi-
tion with: 1) the frequently encountered [4+2] cycloaddition
reaction 4a+5a!6 and 2) the intra- or intramolecular trap-
ping of the postulated gold carbene intermediate with an
alkene leading to cyclopropanes 7 and 8.[12] However, by
taking into account the reactivity of conjugated diene struc-
tures in such a type of transformation,[13] we considered that
the introduction of methyl groups at positions C2 and C3
could potentially favor the desired ring-closing reaction lead-
ing to 3a.
Table 1. Optimization of reaction conditions.[a]
Entry
Catalyst[b]
(4 mol%)
T
[8C]
t
[h]
Yield
[%][c]
1
2
3
4
5
6
7
8
9
AuCl3 (9)
23
23
23
40
23
23
23
23
40
0
2
4
1
73 (9:1)
61 (8:1)
88 (9:1)
0
18 (>20:1)
31 (>20:1)
0
NaAuCl4·2H2O (10)
AuCl·SMe2(11)
PPh3AuNTf2 (12)
IPrAuNTf2 (13)
SIPrAuNTf2 (14)
AgSbF6
AgNTf2
PtCl2
11
11
6
12
12
6
6
6
4
1
0
0
10
11[d]
60 (9:1)
87 (9:1)
23
[a] Unless otherwise noted, all reactions were carried out with 4a
(0.2 mmol), 5a (1.0 mmol, 5 equiv), and catalyst (4 mol%) in CH2Cl2
(2.0 mL). [b] Tf=trifluoromethanesulfonyl; IPr=1,3-bis(2,6-diisopropylphe-
nyl)imidazol-2-ylidene; SIPr=1,3-bis(2,6-diisopropylphenyl)-4,5-dihydroi-
midazol-2-ylidene. [c] Yield of isolated products, the d.r. values were de-
termined by 1H NMR spectroscopy and given within parentheses.
[d] 2 mol% of AuCl·SMe2 (11) and 3 equivalents of 5a were used.
Scheme 3. Potential competitive pathways in the reaction of 4a with 5a.
A primary experiment was performed by treating at room
temperature a solution of o-alkynylbenzylaldehyde 4a and 2,3-
dimethylbutadiene (5a, 5 equiv) in dichloromethane with
4 mol% of AuCl3 (9). We were delighted to observe the rapid
formation (2 h) of the fused tetracyclic product 3a, which
could be isolated in 73% yield as a 9:1 mixture of diastereoiso-
mers (Table 1, entry 1).[14] Notably the reaction was surprisingly
highly selective; no trace of compound 6 or cyclopropanes 7
We then performed the scope of the reaction and first stud-
ied the possibility to use various ortho-alkynylarylaldehydes
(Table 2). It was found that the procedure could be successfully
applied to a large panel of enynals (4a–u). The corresponding
strained tetracyclic bridgehead olefins (3a–u) were obtained in
generally high yield (51-99%) and diastereoselectivity (9:1 to
>20:1). While enynals 4b and 4d–h, possessing substituents
of various electronic nature at the 4-, 5-, and 6-positions of the
aromatic ring, smoothly gave the desired products, the reac-
tion of substrate 4c, bearing a strong electron-donating
group, was almost inefficient (Table 2, entries 1–8). With regard
to the substitution pattern of the R2 group attached to the
alkyne moiety, it was shown that the reaction could be per-
formed with a similar efficiency whatever the electronic nature
of the substituent at the para position of the aromatic ring
(73–99% yield, 18:1 to 1:0 diastereoselectivity, Table 2, en-
tries 9–11 and 13–18). meta-Substituted enynals (4s and 4t)
were also suitable substrates (Table 2, entries 19 and 20). A
limit in reactivity was observed when the sterically hindered
ortho-substituted substrate 4u was employed. In this case, an
elevated temperature and a longer reaction time were re-
quired to produce 3u in acceptable yield and good selectivity
1
and 8 could be detected by H NMR spectroscopy analysis of
the reaction mixture (Scheme 3). We then focused our atten-
tion on the optimization of the experimental reaction condi-
tions. Results of this study are reported in Table 1. While
NaAuCl4·2H2O (10) could also promote the formation of 3a
(Table 1, entry 2), the best result in terms of reaction rate and
yield was obtained when AuCl·SMe2 (11) was used as the cata-
lyst (Table 1, entry 3). [Ph3PAuNTf2] gold complex (12), which
possesses a phosphine ligand, was completely inefficient even
at the elevated temperature (Table 1, entry 4); gold(I) com-
plexes 13 and 14 bearing an N-heterocyclic carbene ligand
were also far less reactive than 11 (Table 1, entries 5 and 6). It
was also shown that other transition-metal salts, such as
AgSbF6, AgNTf2 and PtCl2, were not suitable catalysts for the
Chem. Eur. J. 2016, 22, 9125 – 9129
9126
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim