Gold(I)-catalyzed synthesis of γ-hydroxyketones from 5-allyloxy-1-ynes

Article abstract of DOI:10.1021/jo102005x

The Au(I)-catalyzed reactions of 5-allyloxy-1-ynes gave various γ-hydroxyketones, via a hydration-terminated domino sequence involving sigmatropic allyl migration as the key event. Moreover, the scope of the sigmatropic allyl transfer was systematically d

Full text of DOI:10.1021/jo102005x

pubs.acs.org/joc
generate a mixture of the formal [1,3]-product 4 and [3,3]-
Gold(I)-Catalyzed Synthesis of γ-Hydroxyketones
from 5-Allyloxy-1-ynes
product 5 (path A). Alternatively, sigmatropic rearrange-
ment is feasible from the intermediate 2 (path B), which
will produce oxocarbenium ion intermediate 6 via con-
certed migration of the allylic moiety. Although the former
Jae Youp Cheong, Donghong Im, Miryeong Lee,
Wontaeck Lim, and Young Ho Rhee*
€
pathway has been reported by Furstner employing Pt
catalysts,2b,c the latter pathway remains unknown, to
our best knowledge.4 Moreover, development of the Claisen-
mediated domino pathways shown in path B would be highly
useful from a synthetic viewpoint, because the sigmatropic
rearrangement generates a structurally unique skeleton in a
highly controlled manner. Although formation of the formal
[3,3]-product 5 via demetalation is a plausible process, the
proposed domino sequence can be combined with external
nucleophiles to provide alternative structure 7, with or with-
out the involvement of cycloisomerization product 5 as an
intermediate.
Department of Chemistry, Pohang University of
Science and Technology, San-31 Hyojadong,
Pohang 790-784, Korea
yhrhee@postech.ac.kr
Received October 20, 2010
SCHEME 1. Mechanistic Consideration for the Gold(I)-Cata-
lyzed Synthesis of γ-Hydroxyketones from 5-Allyloxy-1-yne
The Au(I)-catalyzed reactions of 5-allyloxy-1-ynes gave
various γ-hydroxyketones, via a hydration-terminated
domino sequence involving sigmatropic allyl migration
as the key event. Moreover, the scope of the sigmatropic
allyl transfer was systematically determined.
Domino processes where multiple events are combined in
a single transformation have been considered an excellent
way to introduce molecular complexity into organic struc-
tures with high chemical efficiency.1 Metal-catalyzed dom-
ino reactions initiated by the addition of an X-R bond (X =
heteroatom, R = alkyl group) to alkynes offer the unique
possibility of installing various substitution patterns in het-
erocyclic compounds, via the subsequent rearrangement of
alkyl groups. Although this sequence has been employed
mostly in the carboalkoxylation of phenolic or benzylic
ethers,2 more challenging 5-alkoxy-1-ynes have rarely been
employed as a viable substrate.3 The reaction of 5-allyloxy-1-
ynes 1 is particularly interesting because two different types
of products can be formed depending upon the mechanism of
the reaction (Scheme 1). If the C-O bond scission of the
initial alkoxycyclization intermediate 2 is fast, the allylic
cation may be formed. The subsequent recombination would
Because of the importance of the tetrahydrofuran frame-
works in various natural products, we originally focused on
the formation of 5 via Claisen-mediated pathway. On the
basis of our recent study on the gold(I)-catalyzed cycloisome-
rization of 3-methoxy-1,6-enynes,3b,5 we envisioned that
highly electrophilic cationic gold(I) complexes would facilitate
this unprecedented tandem alkoxycyclization-sigmatropic
process.
A preliminary study using prenyl ether 8 showed an un-
expected result. Although the reaction was quickly completed
when complex 11a was employed in combination with AgSbF6
(5 mol %),6,7 the cycloisomerization product 9 could not be
obtained. Instead, we were able to identify a significant amount
(4) For the related study involving aza-Claisen rearrangement, see:
Istrate, F. M.; Gagosz, F. Org. Lett. 2007, 9, 3181.
(5) For reviews on the gold- and platinum-catalyzed reactions, see: (a)
(1) For a review on the domino reactions, see: Parsons, P. J.; Penkett,
C. S.; Shell, A. J. Chem. Rev. 1996, 96, 195.
€
Furstner, A.; Davies, P. W. Angew. Chem., Int. Ed. 2007, 26, 2. (b) Ma, S.;
(2) For some recent examples for transition metal-catalyzed carboalkox-
ylation generating benzylic or phenyl ethers, see: (a) Dube, P.; Toste, F. D. J.
Yu, S.; Gu, Z. Angew. Chem., Int. Ed. 2006, 45, 200. (c) Zhang, L.; Sun, J.;
Kozmin, S. A. Adv. Synth. Catal. 2006, 348, 2271. (d) Hashimi, A. S. K.;
€
Am. Chem. Soc. 2006, 128, 12062. (b) Furstner, A.; Daives, P. W. J. Am.
Chem. Soc. 2005, 127, 15024. (c) Furstner, A; Szillat, H.; Stelzer, F. J. Am.
Chem. Soc. 2001, 123, 11863. (d) Nakamura, I.; Sato, T.; Yamamoto, Y.
Angew. Chem., Int. Ed. 2006, 45, 4473.
(3) (a) Nakamura, I.; Chan, C. S.; Araki, T.; Terada, M.; Yamamoto, Y.
Org. Lett. 2008, 10, 309. (b) Bae, H. J.; Baskar, B.; An, S. E.; Cheong, J. Y.;
Thangadurai, D. T.; Hwang, I.-C.; Rhee, Y. H. Angew. Chem., Int. Ed. 2008,
47, 2263.
ꢀ
ꢀ~
Hutchings, G. J. Angew. Chem., Int. Ed. 2006, 45, 7896. (e) Jimenez-Nunez,
E.; Echavarren, M. Chem. Commun. 2007, 333. (f) Gorin, D. J.; Toste, F. D.
Nature 2007, 446, 395. (g) Hashimi, A. S. K. Chem. Rev. 2007, 107, 3180. (h)
€
ꢀ
Li, Z.; Brower, C.; He, C. Chem. Rev. 2008, 107, 3180. (i) Jimenez-Nunez, E.;
Echavarren, A. M. Chem. Rev. 2008, 108, 3266. (j) Furstner, A. Chem. Soc.
Rev. 2009, 38, 3208.
(6) For the preparation of complex 11a, see: Shin, S. Bull. Korean Chem.
Soc. 2005, 26, 1925.
ꢀ~
€
324 J. Org. Chem. 2011, 76, 324–327
Published on Web 12/14/2010
DOI: 10.1021/jo102005x
r
2010 American Chemical Society

Cheong et al.
JOCNote
of γ-hydroxyketone 10 (∼30% yield) after carefull analysis of
TABLE 1. Optimization of the Reaction Conditions
spectral data.
entry
catalyst (mol %)
time
conversion (%) yield (%)a
1
2
3
4
5
6
11a (5)/ AgSbF6 (5) 2 min
11b (5)/ AgSbF6 (5) 5 h
11c (5)/ AgSbF6 (5) 5 h
11a (2)/ AgSbF6 (2) 30 min
100
95 (91b)
NDd
trace
trace
100
100
30c
NDd
98 (95b)
90
11a (5)/ AgOTf (5)
11a (5)/ AgBF4 (5)
5 min
5 h
NDd
We initially rationalized that 10 is formed by the hydration
of the cycloisomerization product 9. However, using anhy-
drous conditions failed to give 9, although formation of 10
could be minimized.8 This result strongly suggests that 10
may be formed via direct hydration of the oxacarbenium ion
intermediate (6 in Scheme 1), rather than the hydration of the
cycloisomerization product 9.
aNMR yield of compound 10 determined by using 1,3,5-trimethox-
ybenzene as an internal standard. bIsolated yield of compound 10.
cDetermined by the analysis of the 1H NMR of the crude reaction
mixture. dNot determined.
bearing an additional alkyl group at the bis-homopropargylic
position (entries 3-5). Thus, allyl ether 16 was only slowly
converted into the product 17 in moderate 46% yield (entry 3);
however, the crotyl ether 18 (entry 4) and the prenyl ether 20
(entry 5) significantly improved the yield of the reaction.11 As is
the case for the substrate 8, [1,3]-products analogous to 100 were
not observed. This result clearly shows that the sigmatropic
pathway is indeed dominating for the substrates possessing no
alkyl substituents at the allylic position. It should be emphasized
that the sigmatopic rearrangement shown here is unique in that
increasing the steric congestion does not slow the reaction.12-14
We then investigated the substrate possessing an alkyl
substituent at the allylic position (entries 6-10). On the basis
of the reactivity pattern of the previous examples, we antici-
pated that these substrates should be effective toward the
gold(I)-catalyzed tandem sequence. Indeed, substrate 22
produced the formal [3,3]-product 23 in reasonable 77%
yield with no indication of the formation of the [1,3]-product
(entry 6).15 Increasing the alkyl substitution at the terminal
position of the olefin moiety again facilitated the reaction
(entry 6 vs entries 7 and 8). Although the example shown in
entries 7 and 8 appears to significantly extend the scope of the
Claisen-mediated pathway, the cationic pathway cannot be
completely excluded because an identical structure would
arise by two mechanisms (path A vs path B in Scheme 1).
Thus, we tested 28 to rigorously examine the mechanistic
issue described in Scheme 1 (entry 9). Unlike the previous
examples, the gold(I)-catalyzed reaction of this compound
gave a chromatographically inseparable mixture (∼25:1 ratio)
Based on this preliminary study, we turned our attention
to the synthesis of γ-hydroxyketone 10 employing a hydra-
tion-terminated tandem sequence. A marked increase of the
yield occurred simply by adding exogenous water to the
reaction mixture (entry 1, Table 1). Use of an electrophilic
catalyst 11a was necessary, because employing less electro-
philic gold catalysts 11b and 11c significantly dropped the
conversion (entries 2 and 3). Reducing the catalyst loading of
11a to 2 mol % still maintained the catalytic activity with a
slight increase in the yield of 10 (entry 4). A significant
counteranion effect was also noted. For example, using
AgOTf showed significant catalytic activity (entry 5), while
employing AgBF4 slowed the reaction (entry 6). Notably, the
formal [1,3] product 100 was not observed even in trace
amounts in these optimization studies.9,10
Using the optimized conditions shown in entries 1 and 4 of
Table 1, various 5-allyloxy-1-ynes were converted into the corre-
sponding γ-hydroxyketones. As can be seen in Table 2, the sub-
stitution pattern of the allylic ether moiety had a crucial effect on
the rate and yield of the reaction. For example, the reaction of
crotyl ether 12 was significantly slower than that of the prenyl
ether 8 (entry 1). In this case, higher catalyst loading (5 mol %)
was required to give the desired product 13 in 90% yield. On the
other hand, the reaction of geranyl ether 14 was completed
within 2 min even at -15 °C, when 5 mol % catalyst was used
(entry 2). A similar reactivity pattern was seen with substrates
(7) For selected examples where the complex 11a is used, see: (a) Baskar,
B.; Bae, H. J.; An, S. E.; Cheong, J. Y.; Rhee, Y. H.; Duschek, A.; Kirsch,
S. F.; et al. Org. Lett. 2008, 10, 2605. (b) An, S. E.; Jeong, J.; Baskar, B.; Lee,
J.; Seo, J.; Rhee, Y. H. Chem.;Eur. J. 2009, 15, 11837. (c) Kim, C.; Bae,
H. J.; Lee, J. H.; Kim, H.; Sampath, V.; Rhee, Y. H. J. Am. Chem. Soc. 2009,
131, 14660. (d) Kang, J.-E.; Shin, S. Synlett 2006, 717. (e) Gung, B. W.; Craft,
D. T.; Bailey, L. N.; Kirschbaum, K. Chem.;Eur. J. 2010, 16, 639. (f) Kim,
C.; Lim, W.; Rhee, Y. H. Bull. Korean Chem. Soc. 2010, 31, 1465. (g) Lee,
P. H.; Kim, S.; Park, A.; Chary, B. C.; Kim, S. Angew. Chem., Int. Ed. 2010,
(11) All the γ-hydroxyketones obtained from studies exist as an open
form except for compound 17, which exists as an ca. 10:1 mixture of the open
and cyclized form in CDCl3 solution.
(12) For a review on the Claisen rearrangement, see: Castro, A. M. M.
Chem. Rev. 2004, 104, 2939.
(13) The sigmatropic nature of the allyl shift was further supported by a
crossover experiment using a mixture of 14 and 20, which provides only 15
and 21 without the apparent formation of the crossover products.
(14) A similar structural effect was reported for the aza-Claisen rearran-
gement, see ref 4.
(15) Exclusive formation of the trans isomer 23 strongly suggests involve-
ment of the chairlike transition state in the sigmatropic rearrangement. For a
related issue in the thermal Claisen rearrangement, see : Nguyen, N. N. M.;
Leclere, M.; Stogaitis, N.; Fallis, A. G. Org. Lett. 2010, 12, 168.
ꢀ
ꢀ~
122, 1. (h) Nieto-Oberhuber, C.; Munoz, M. P.; Lopez, S.; Jimenez-Nunez,
E.; Nevado, C.; Herrero-Gomez, E.; Raducan, M.; Echavarren, A. M.
Chem.;Eur. J. 2006, 12, 1677.
(8) In this case, an extensive amount of polymeric compounds was
obtained.
(9) Within the detection limit of 1H NMR analysis.
(10) (a) Using Pt complexes (PtCl2 and PtCl4) showed no conversion.
(b) Employing AgSbF6 also showed no conversion.
J. Org. Chem. Vol. 76, No. 1, 2011 325

Cheong et al.
JOCNote
TABLE 2. Scope of the Gold(I)-Catalyzed Formation of γ-Hydroxyketones
aMixture of olefin isomers was used (E:Z = 5:1). bThe product was obtained as a diastereomeric mixture (1:1). cThe product was obtained as an
inseparable mixture with a small amount of the isomeric [1,3]-product.
of [3,3]- and [1,3]-product (compound 29 and 31) in 85% yield
(entry 9). Substrate 30, in which the R1 and R3 groups are
interchanged from 28, showed a similar pattern in erms of both
yield and the product. Thus, a mixture of the [3,3]-product 31
and the [1,3]-product 29 (ca. 25:1 ratio) was obtained in 83%
yield. This result indeed shows that increasing the alkyl substitu-
tion promotes the cationic mechanism (path A in Scheme 1). As
depicted in eq 3, the effect of alkyl substitution promoting the
cationic mechanism is more clearly seen with the substrate 32. In
this case, a nearly equimolar mixture of the [3,3]-product 33 and
[1,3]-product 34 was obtained in 72% overall yield.16
Finally, it should be noted that another external nucleophile
also can be easily incorporated with the domino alkoxycycliza-
tion-Claisen sequence. Thus, 2-alkoxytetrahydrofuan 35 could
be obtained in 71% yield by the gold(I)-catalyzed reaction of
substrate 22 in the presence of methanol (eq 4).
In summary, we discovered gold(I)-catalyzed access to synthe-
tically useful γ-hydroxyketones from 5-allyloxy-1-ynes via a
hydration-terminated domino sequence. Notably, this study
systematically determines the scope of the Claisen pathway in
allyl migration. Extrapolation of this tandem sequence to other
catalytic cycles, as well as the application to the synthesis of
bioactive natural products synthesis, is currently under investi-
gation.
Experimental Section
(16) In contrast to the primary allylic ether 8, isomeric tertiary allylic ether
generated from isoprenol showed no reactivity. Presumably, the initial
cyclization is significantly slow because of the steric effect of the alkyl groups.
General Procedure for the Gold(I)-Catalyzed Reaction. To a
solution of AgSbF6 (2.5 mg, 0.0073 mmol) in dry CH2Cl2 (1 mL)
326 J. Org. Chem. Vol. 76, No. 1, 2011

Cheong et al.
JOCNote
was added a solution of 11a (5.6 mg, 0.0073 mmol) in dry
CH2Cl2 (1 mL). The solution was stirred for 10 min. The
resulting solution was filtered though a pad of Celite and
concentrated. The residue was dried over high vacuum for 2 h
and then cooled to 0 °C. To this residue was added the solution
of 8 (56 mg, 0.37 mmol) in CH2Cl2:H2O = 10:1 (7.4 mL,
precooled to 0 °C). The resulting green solution was stirred for
30 min. Triethylamine (1 mL) was added and the solution was
stirred for 5 min. The resulting solution was filtered through a
pad of silica and concentrated. The crude oil was purified by
flash chromatography on silica gel (eluted with hexane:ethyl
acetate = 70:30) to give the compound 10 as a colorless oil
(60 mg, 0.35 mmol, 95% yield). Rf 0.08 (hexane:ethyl acetate =
80:20); 1H NMR (300 MHz, CDCl3) δ 1.10 (s, 6H), 1.77 (quint,
J = 6.3 Hz, 2H), 2.03 (br s, 1H), 2.42 (s, 2H), 3.60 (t, J = 6.1 Hz,
2H), 4.90-4.96 (m, 2H), 5.89 (dd, J = 17.6 Hz, J = 10.5 Hz,
1H); 13C NMR (75 MHz, CDCl3) δ 26.5, 27.2, 36.6, 41.8, 54.5,
62.4, 111.1, 147.3, 210.8; IR (cm-1) ν 3397, 2966, 1640, 1456,
1371; HRMS calcd for C10H18O2 170.1307, found 170. 1308.
Acknowledgment. This project was supported by the Na-
tional Research Foundation (NRF-2010-0009458 and NRF-
2008-0061957). J.Y.C. thanks the BK21 program for finan-
cial support.
Supporting Information Available: Experimental proce-
1
dures for catalytic formation of 10, and copies of H and 13C
NMR spectra of all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 76, No. 1, 2011 327