Cyclization of nonterminal alkynic β-Keto esters catalyzed by gold(I) complex with a semihollow, end-capped triethynylphosphine ligand

Article abstract of DOI:10.1021/ol802079r

(Chemical Equation Presented) A cationic gold(I) complex with a semihollow-shaped trialkynylphosphine catalyzed 5-exo-dig and 6-endo-dig cyclizations of various internal alkynic β-keto esters, showing a marked advantage over a gold(I) - PPh₃ complex with respect to the rates of the reactions and the product yields. It is proposed that the gold-bound alkynic substrate in a catalytic pocket must be somewhat folded and that such a steric effect makes the carbon-carbon bond formation entropically more favorable.

Full text of DOI:10.1021/ol802079r

ORGANIC
LETTERS
2008
Vol. 10, No. 21
5051-5054
Cyclization of Nonterminal Alkynic
ꢀ-Keto Esters Catalyzed by Gold(I)
Complex with a Semihollow,
End-Capped Triethynylphosphine Ligand
Hideto Ito, Yusuke Makida, Atsuko Ochida, Hirohisa Ohmiya, and
Masaya Sawamura*
Department of Chemistry, Faculty of Science, Hokkaido UniVersity,
Sapporo 060-0810, Japan
sawamura@sci.hokudai.ac.jp
Received September 5, 2008
ABSTRACT
A cationic gold(I) complex with a semihollow-shaped trialkynylphosphine catalyzed 5-exo-dig and 6-endo-dig cyclizations of various internal
alkynic ꢀ-keto esters, showing a marked advantage over a gold(I)-PPh3 complex with respect to the rates of the reactions and the product
yields. It is proposed that the gold-bound alkynic substrate in a catalytic pocket must be somewhat folded and that such a steric effect makes
the carbon-carbon bond formation entropically more favorable.
Electrophilic activation of an alkyne by the coordination of
a π-Lewis acidic metal cation induces the attack of a
nucleophile that occurs at an appropriate position within the
molecule. For such alkyne cyclizations, gold is the most
active catalyst among various metal ions, and gold-phosphine
complexes have been applied to various types of alkyne
cyclizations.1,2 The ring-forming gold catalysts, however,
suffer from at least two serious problems. First, they are
limited to the cases where cyclization is entropically quite
favorable such as five-membered ring formations.2a Second,
the cyclization of internal alkynes is often hampered by steric
repulsion between the terminal substituent and a nucleophile
at the C-C bond-forming step.2b We previously reported
on the design and synthesis of semihollow-shaped trialky-
nylphosphine 1a, which consists of a triethynylphosphine
(1) For reviews on gold catalysis, see: (a) Hashmi, A. S. K. Chem. ReV.
2007, 107, 3180–3211. (b) Li, Z.; Brouwer, C.; He, C. Chem. ReV. 2008,
108, 3239–3265. (c) Arcadi, A. Chem. ReV. 2008, 108, 3266–3325. (d)
Jime´nez-Nu´n˜ez, E.; Echavarren, A. M. Chem. ReV. 2008, 108, 3326–3350.
(e) Gorin, D. J.; Sherry, B. D.; Toste, F. D. Chem. ReV. 2008, 108, 3351–
(2) For gold(I)-catalyzed cyclizations of acetylenic keto esters, see: (a)
Kennedy-Smith, J. J.; Staben, S. T.; Toste, F. D. J. Am. Chem. Soc. 2004,
126, 4526–4527. (b) Staben, S. T.; Kennedy-Smith, J. J.; Toste, F. D. Angew.
Chem., Int. Ed. 2004, 43, 5350–5352. (c) Me´zailles, N.; Ricard, L.; Gagosz,
3378
.
F. Org. Lett. 2005, 7, 4133–4136.
10.1021/ol802079r CCC: $40.75
Published on Web 10/16/2008
2008 American Chemical Society

Table 1. Gold-Catalyzed Conia-Ene Reaction of 2a-da
convn
time
(h)
of 2b
(%)
3 + 4,
yield (%)
3/4
entry
2
ligand
1a
1b
1c
PPh3
PPh3
P(OPh)3
P(OPh)3
1a
PPh3
1a
PPh3
1a
ratiob
1
2
3
4
5
6
7
8
9
10
11
12
13
2a
2a
2a
2a
2a
2a
2a
2b
2b
2c
2c
2d
2d
3
24
24
3
24
3
24
12
24
20
24
24
24
100
0
12
4
12
6
13
100
21
100
12
100
23
80c
0b
80:20
N.A.d
85:15
N.D.e
89:11
N.D.e
85:15
91:9
95:5
92:8
100:0
45:55
54:46
12b
4b
Figure 1. Semihollow, end-capped triethynylphosphine 1a and
triethynylphosphines (1b,c) with smaller silicon end caps.
11b
6b
core and bulky end caps (Figure 1).3,4 This ligand (1a)
markedly accelerated ring-forming gold(I) catalysts with
various terminal alkynes, allowing for six- and seven-
membered ring formations that are difficult to achieve with
conventional ligands because of the flexibility of the sub-
strates.5 These results represented our approach to the first
issue. Here we report on our challenge toward the second
issue: The gold(I) complex with 1a catalyzed the cyclizations
of nonterminal alkynic ꢀ-keto esters that afford the sterically
congested five- and six-membered ring compounds.
13b
97c
12
94
12b
86c
13
PPh3
a Conditions: 2, 0.40 mmol; CH2Cl2, 1 mL. b Determined by NMR.
c Isolated yield. d Not applicable. e Not determined.
The semihollow triethynylphosphine (1a) showed a marked
advantage over conventional phosphine ligands such as PPh3
and P(OPh)3 when applied to the gold-catalyzed cyclization
of internal alkyne 2a with a keto ester functionality (Table
1, entry 1 vs entries 4-7).6 Thus, treatment of 2a with a
CH2Cl2 solution of cationic gold(I) complex [Au(NTf2)(1a)]
(1 mol %) resulted in smooth conversion at rt; the reaction
completed within 3 h to afford 5-exo-dig and 6-endo-dig
cyclization products 3a and 4a in a good combined yield,
with preference to the former (80:20), which seems to be
sterically more congested and hence energetically less
favorable but entropically more favorable than the latter
(entry 1). No E-isomer of exo-alkene 3a was formed,
suggesting anti-stereochemistry of a nucleophilic attack to
a gold-alkyne π complex (Scheme 1).2 To the best of our
knowledge, the gold-catalyzed 5-exo-dig/6-endo-dig cycliza-
tion of a nonterminal alkynic keto ester has not been
described in the literature.2b
Scheme 1. Proposed Mechanism for the Reaction of 2a-d
power is comparable with 1,3 was only as efficient as the
PPh3 complex (3 h, 6% conversion, entry 6; 24 h, 13%
conversion, entry 7). Furthermore, triethynylphosphines
(1b,c) bearing smaller silicon end caps such as Ph3Si and
(i-Pr)3Si groups were not effective (entries 2 and 3).
Even with the internal alkynes (2b-d) bearing bulkier
terminal substituents such as Bu, i-Pr, and Ph groups, the
reaction with the Au-1a catalyst proceeded smoothly and
afforded the corresponding five- and six-membered ring
compounds (3b-d/4b-d) in high yields (Table 1, entries
8, 10, and 12). With the Au-PPh3 catalyst, the reaction was
again much slower and gave the cyclization products in only
very low yields (entries 9, 11, and 13).
In contrast, there was almost no reaction with the corre-
sponding PPh3 complex with the same reaction time (3 h,
4% conversion, entry 4). Prolonging the time to 24 h caused
only a slight conversion (12% conversion, 24 h, entry 5).
The complex formed with the P(OPh)3 ligand, whose donor
(3) Ochida, A.; Sawamura, M. Chem.sAsian J. 2007, 2, 609–618
.
(4) For a review concerning the synthesis and applications of bowl-
shaped phosphine ligands, whose steric feature is related to our triethy-
nylphosphine ligands, see: Tsuji, Y.; Fujiwara, T. Chem. Lett. 2007, 36,
1296–1301
.
The effect of the terminal substituent (R) of 2 on the 5-exo/
6-endo selectivity may give an insight into the mechanism
of the gold-catalyzed Conia-ene reaction. As the terminal
alkyl substituents became bulkier (Me < Bu < i-Pr), the
(5) Ochida, A.; Ito, H.; Sawamura, M. J. Am. Chem. Soc. 2006, 128,
16486–16487.
(6) The cyclization products did not form in the presence of AuCl(1a),
Ag(NTf2), and HNTf2 (1 mol %) under othewise identical conditions (24
h). Slight decomposition of 2a (7%, 24 h) occurred with HNTf2.
5052
Org. Lett., Vol. 10, No. 21, 2008

Table 2. Gold-Catalyzed Conia-Ene Reaction of Nonterminal Alkynic, Cyclic Keto Esters 2f-ja
a Conditions: 2, 0.40 mmol; Au(NTf2) (ligand), 0.004 mmol (1 mol %); CH2Cl2, 1 mL; 25 °C unless otherwise noted. b Determined by NMR. c Isolated
yield.
selectivity for 5-exo product 3 increased: 80%, 91%, and
Scheme 2. Gold-Catalyzed Conia-Ene Reaction of 2e
92% with 1a (Table 1, entries 1, 8 and 10); 89%, 95%, and
100% with PPh3 (entries 5, 9 and 11). In the proposed C-C
bond-forming transition states (Scheme 1), the terminal
substituent should encounter steric repulsions not only with
the incoming nucleophile but also with the Au–PR′3 moiety.
The former should be larger in the 5-exo-transition state
(geminal), while the latter in the 6-endo-transition state (cis-
vicinal). Accordingly, the increase of the exo selectivity upon
the increase of the size of the terminal substituent R is likely
due to the steric repulsion between R and the Au-PR′3
moiety.
fragment encounters more steric repulsion with the geminally
substituted inner substituent than with the terminal Me
substituent.
The catalysis of gold-1a appeared to be effective even
when the acyclic alkyne (2e) has a quaternary carbon center
at the internal R-position (Scheme 2). While no reaction
occurred at rt, the reaction at 80 °C in DCE converted the
hindered alkyne (2e) cleanly into a mixture of 5-exo-dig (3e)
and 6-endo-dig (4e) cyclization products with a preference
to the 6-endo isomer.7 The observed regioselectivity is
consistent with our rationale for the 5-exo selectivity in the
reaction of 2a-d: In the reaction of 2e, the Au-PR′3
Albeit to a reduced extent, the advantage of 1a over PPh3
appeared to be significant in the synthesis of sterically
congested bicyclic compounds by the annulation of cyclic
compounds (Table 2). When [Au(NTf2)(1a)] (1 mol %) was
employed, the 5-exo/6-endo annulation of an alkyne pendant
onto a cyclopentanone (Table 2, entry 1) and a cyclohex-
anone (entry 4) completed within 24 h to afford 3f/4f (49:
51) and 3g/4g (67:33), respectively, in high yields and in
complete diastereoselectivity. The same reactions also pro-
ceeded with [Au(NTf2)(PPh3)], but they were slower and
formed unidentified side products, resulting in lower yields
(7) The reaction of 2e with the Au-PPh3 catalyst: 24 h; 41% conversion;
35% yield; 3e/4e, 12:88. 48 h; 44% conversion; 44% yield; 3e/4e, 12:88.
Org. Lett., Vol. 10, No. 21, 2008
5053

of the desired products even after complete comsumption
of the alkynes (entries 2, 3, 5, and 6). Although the insertion
of an aromatic ring between an alkyne and a nucleophile
tends to make the cyclization easier, the reaction of 2h,
having a keto ester attached to the terminal of a side chain,
appeared to be less feasible under the conventional conditions
with the PPh3 ligand (entry 8 and 9). On the contrary, the
gold-1a catalyst achieved quatitative conversion for this
transformation (entry 7).
Notably, the reaction of 2i proceeded with complete
regioselectivity to afford bicyclo[3.3.1]nonenone 4i irrespec-
tive of the ligand used; however, 1a showed much higher
efficiency in terms of both the rate of the reaction and the
yield of 4i (Table 2, entries 10-12). Additionally, similar
results were obtained in the reaction of 2j, which provided
highly substituted bicyclo[2.2.2]octenone derivative 4j (en-
tries 13-15).
carbon-carbon bond formation entropically more favorable
so that the gold-catalyzed cyclization of internal alkynes
becomes feasible against steric repulsions (see Scheme 1)
at the transition state.
In summary, the scope of the gold-catalyzed Conia-ene
reaction has substantially been expanded so as to involve
5-exo-dig and 6-endo-dig cyclizations of internal alkynic
ꢀ-keto esters by employing the semihollow-shaped end-
capped triethynylphosphine as a ligand of a cationic gold(I)
catalyst. Studies to explore the applicability of the semihol-
low-shaped ligand toward various transition-metal catalysts
is underway.
Acknowledgment. This work was supported by Grant-
in-Aid for Scientific Research (B) (No. 18350047) from JSPS
and Grant-in-Aid for Scientific Research on Priority Area
(No. 20037003, “Chemistry of Concerto Catalysis”) from
MEXT. A.O. thanks JSPS for a fellowship.
Our molecular modeling simulations indicated that the
gold-bound alkynic substrate8 in a catalytic pocket created
by the semihollow ligand must be somewhat folded. Ac-
cordingly, we assume that such a steric effect makes the
Supporting Information Available: Experimental pro-
cedures and NMR spectra for new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
(8) For the structure of an isolated gold(I)-alkyne complex, see: Shapiro,
N. D.; Toste, F. D. Proc. Nat. Acad. Sci. U.S.A. 2008, 105, 2779–2782.
OL802079R
5054
Org. Lett., Vol. 10, No. 21, 2008