.
Angewandte
Communications
Table 3: Benzannulation of triynes with different tethers in the presence
of various nucleophiles.
reaction at 908C in CH3CN with a small amount (0.1 mol%)
of the Grubbs second-generation complex (Grubbs II) sup-
pressed the formation of the unknown by-product,[4] thus
improving the yield of 3a to 63% (entry 3). Further improve-
ment was achieved when the triethylsilyl (TES) group of 1a
was replaced with tert-butyldimethylsilyl (TBS) group (1b)
under identical reaction conditions, thus leading to 3b in 98%
(entry 4). In contrast, the triisopropylsilyl (TIPS) group in 1c
or triphenylsilyl (TPS) in 1d had a marginal improvement,
thus providing the corresponding products 3c and 3d[5] in 68
and 52% yield, respectively (entries 5 and 6).
Having defined the optimal silyl functionality in the
substrate and assorted reaction conditions, we explored the
reaction of substrates containing more substituents in the
ester-tethered triyne platform (Table 2). The reaction of 1e,
Table 2: Benzannulation of ester-tethered 1,3,8-triynes.[a]
Numbers within parentheses represent yields of the isolated products.
[a] Grubbs II (0.1 mol%) was used in these reactions. [b] With a bromide
source described in Ref. [2c]. Ts=4-toluenesulfonyl.
1r significantly improved the reaction profile, thus providing
3r in 92% yield within 17 hours. However, only low yield of
3s (39%) was observed when AcOH was used as a trapping
agent. To our surprise, the substrate 1t (R’ = tBu) yielded 3t
which was devoid of the tert-butyl group. The reaction of 1u
with a bromide nucleophile afforded the aryl bromide 3u in
49% yield, and the same substrate in the presence of AcOH
as a nucleophile produced aryl acetate 3v in 53% yield.
Substrates having a ketone linkage can also undergo a benz-
annulation reaction only when a gem-dialkyl moiety is
present. Thus, the benzannulation products 3w and 3x were
obtained in the presence of MeOH and AcOH respectively,
albeit after prolonged heating.
To gain insight into the mechanism of this benzannulation,
we carried out DFT calculations[7] (M06-2X/6-31 + G*
level[8]) with the triyne 1aa as a model system (Scheme 2).
Calculations clearly indicate that the ene reaction leading to
the alkynyl enallene B[9] is kinetically more favorable by
4.6 kcalmolꢀ1 than the HDDA reaction leading to the aryne
A. Under the reaction conditions, B isomerizes to B’, from
which cyclization occurs to form C and D. The Saito–Myers
cyclization[1e,g,h,r] of B’ to form C via TSc (ꢀ6.3 kcalmolꢀ1)[10]
or its ionic version to form C’ via the slightly lower-energy
TSc’ (ꢀ6.6) is energetically reasonable but it does lead to
incorrect connectivity. However, the formation of the dirad-
ical D, bearing the correct connectivity of the observed
product via TSd (37.6 kcalmolꢀ1), does not seem feasible
from either a kinetic or thermodynamic aspect.[7,11] This
energetic consideration suggests an alternative mechanism
involving the Michael addition of AcOH to the allenoate
moiety of B,[12] where the barrier (TS1 = ꢀ9.9 kcalmolꢀ1)
leading to IN1, albeit slightly endergonic, is 3.3 kcalmolꢀ1
lower than even that of the Saito–Myers cyclization (B’!
C).[13] From IN1, all the remaining steps, involving proton-
shift-mediated relocation of p bonds to form IN2, its 6p elec-
trocyclization[14] to form IN3, and aromatization by a formal
[1,3]-H shift leading to 3aa, seem to be quite reasonable
energetically.
[a] Numbers within parentheses represent yield of the isolated products.
with an extra propyl substituent (compared to 1a), afforded
the expected product 3e in 66% yield. Although 1 f, with
a benzyloxy substituent at the propargylic carbon atom, did
not yield the product 3 f, its homologue 1g and silyloxy-
substituted substrates 1h and 1i produced the expected
products 3g–i in 66, 56, and 51% yield, respectively. The
slightly lower yields from 1h and 1i are most likely due to its
instability under the reaction conditions. Introducing a sub-
stituent at the propargylic site of the 1,3-diynyl moiety did not
change the reactivity of substrates 1j–m, thus the products
3k–m[5] were obtained in yields in the range of 72–89%, but
only 47% for 3j because of its labile TES group. Replacing
the silyl group with another alkyne in 1n led to the formation
of a mixture of 3n and 3n’ in 48% yield.
To broaden the scope of the reaction, we employed an
assortment of substrates of different tethers and trapping
agents (Table 3). Upon heating 1b at 908C in MeOH, the
benzannulation product 3o was obtained in 26% yield and
accompanied by the methanolysis product 1,3-diynyl prop-
argylic alcohol in 50% yield. The sterically hindered ester 1k
(R’’ = iPr), however, afforded 3p in 63% yield without
methanolysis.[6] Replacing the ester linkage with an amide
improved the yield, although a longer reaction time was
required. For example, the triyne 1q produced 3q in 85%
yield after heating for 72 hours. Replacing the phenyl group in
1q with a more-electron-withdrawing sulfonimide moiety in
2
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
These are not the final page numbers!