Gosteli-claisen rearrangement of propargyl vinyl ethers: Cascading rearrangements molecular

Article abstract of DOI:10.1021/ol200558j

The ambivalent reactivity of 2-alkoxycarbonyl-substituted propargyl vinyl ethers has been explored. Depending on the conditions, the catalyzed and uncatalyzed Gosteli-Claisen rearrangement triggers downstream transformations that cascade from initially formed γ-allenyl α-keto esters to highly substituted furanes and cyclopentenes. In support of a mechanistic hypothesis, the results of a DFT study using the B1B95 and B3LYP functionals are revealed.

Full text of DOI:10.1021/ol200558j

ORGANIC
LETTERS
2011
Vol. 13, No. 8
2122–2125
GosteliꢀClaisen Rearrangement of
Propargyl Vinyl Ethers: Cascading
Molecular Rearrangements
Annika Gille, Julia Rehbein,† and Martin Hiersemann*
€
Fakultat Chemie, Technische Universitat Dortmund, 44227 Dortmund, Germany
€
martin.hiersemann@udo.edu
Received March 1, 2011
ABSTRACT
The ambivalent reactivity of 2-alkoxycarbonyl-substituted propargyl vinyl ethers has been explored. Depending on the conditions, the catalyzed
and uncatalyzed GosteliꢀClaisen rearrangement triggers downstream transformations that cascade from initially formed γ-allenyl R-keto esters
to highly substituted furanes and cyclopentenes. In support of a mechanistic hypothesis, the results of a DFT study using the B1B95 and B3LYP
functionals are revealed.
The Claisen rearrangement of allyl and propargyl vinyl
ethers has a long and successful scientific history.1 How-
ever, access to effectively stereodifferentiating chiral cata-
lysts for the aliphatic Claisen rearrangement was still
elusive at the end of the past century.2 During the past 10
years, intense research efforts have then led to the identi-
fication of chiral σ-Lewis acids and Brønsted acids as
catalysts for the GosteliꢀClaisen rearrangement of 2-alk-
oxycarbonyl-substituted allyl vinyl ethers providing δ,ε-
unsaturated R-keto esters as valuable building blocks.3ꢀ6
In light of this development, we became interested in the
GosteliꢀClaisen rearrangement of propargyl vinyl ethers
(“pve”, Figure 1, R = CO2R).7 As elegantly demonstrated
by Grissom,8 Toste,9 and Kirsch,10 the formal Claisen
rearrangement of pve with R = H or R = alkyl can be
catalyzed by π-Lewis acids such as Ag(I) or Au(I);11 this
type of rearrangement presumably proceeds by a cycliza-
tion-induced mechanism and provides opportunity for 1,3-
self-immolative chirality transfer.12,13 Toste and Kirsch
also ingeniously designed conditions under which the
† Physical Organic Chemistry Centre, School of Chemistry, Cardiff
University, Cardiff CF10 3AT, U.K.
(1) Hiersemann, M.; Nubbemeyer, U., Eds. The Claisen Rearrange-
ment; Wiley-VCH: Weinheim, Germany, 2007.
(2) Hiersemann, M.; Abraham, L. Eur. J. Org. Chem. 2002, 1461–
1471.
(3) Gosteli, J. Helv. Chim. Acta 1972, 55, 451–460.
(4) (a) Abraham, L.; Czerwonka, R.; Hiersemann, M. Angew. Chem.,
(7) For early work on the Claisen rearrangement of aliphatic pro-
pargyl vinyl ethers (R = H), see: (a) Black, D. K.; Landor, S. R. J. Chem.
Soc. 1965, 6784–6788.
(8) Grissom, J. W.; Klingberg, D.; Huang, D.; Slattery, B. J. J. Org.
Chem. 1997, 62, 603–626.
€
Int. Ed. 2001, 40, 4700–4703. (b) Abraham, L.; Korner, M.; Schwab, P.;
Hiersemann, M. Adv. Synth. Catal. 2004, 346, 1281–1294. (c) Abraham,
L.; Korner, M.; Hiersemann, M. Tetrahedron Lett. 2004, 45, 3647–3650.
For applications in target-oriented synthesis, see: (d) Pollex, A.; Hierse-
mann, M. Org. Lett. 2005, 7, 5705–5708. (e) Korner, M.; Hiersemann,
M. Org. Lett. 2007, 9, 4979–4982. (f) Gille, A.; Hiersemann, M. Org.
Lett. 2010, 12, 5258–5261.
€
€
(9) Sherry, B. D.; Toste, F. D. J. Am. Chem. Soc. 2004, 126, 15978–
15979.
(10) (a) Suhre, M. H.; Reif, M.; Kirsch, S. F. Org. Lett. 2005, 7, 3925–
3927. (b) Cao, H.; Jiang, H.; Mai, R.; Zhu, S.; Qi, C. Adv. Synth. Catal.
2010, 352, 143–152.
(11) For Rh(I)-catalysis, see: Tanaka, K.; Okazaki, E.; Shibata, Y.
J. Am. Chem. Soc. 2009, 131, 10822–10823.
(12) Overman, L. E. Angew. Chem., Int. Ed. Engl. 1984, 23, 579–587.
(13) Burgstahler, A. W.; Nordin, I. C. J. Am. Chem. Soc. 1961, 83,
198–206.
(5) Uyeda, C.; Jacobsen, E. N. J. Am. Chem. Soc. 2008, 130, 9228–
9229.
(6) For related work, see: (a) Marie, J.-C.; Xiong, Y.; Min, G. K.;
Yeager, A. R.; Taniguchi, T.; Berova, N.; Schaus, S. E.; Porco, J. A.
J. Org. Chem. 2010, 75, 4584–4590. (b) Wender, P. A.; D’Angelo, N.;
Elitzin, V. I.; Ernst, M.; Jackson-Ugueto, E. E.; Kowalski, J. A.;
McKendry, S.; Rehfeuter, M.; Sun, R.; Voigtlaender, D. Org. Lett.
2007, 9, 1829–1832.
r
10.1021/ol200558j
Published on Web 03/23/2011
2011 American Chemical Society

initially formed allenes are converted in situ into furans,10 2H-
pyrans, or dihydropyrans.14,15 However, the applicability of
the established Au(I) or Ag(I) catalysis to 2-alkoxycarbonyl-
substituted pve (R = CO2R) remained unexplored.
rate-accelerating effect, and because warming to 60 °C
for 23 h led to an incomplete conversion (Table 1, entry 3),
we extended the reaction time to 72 h (Table 1, entry 4).
The isolated product mixture then consisted of a residual
amount of 1a, the allene 2a, and, notably, the cycloalkenes
3a and 5a. In analogy, warming 1a in TFE for an extended
period of time also led to the formation of the correspond-
ing cyclization product 4a and 5a, albeit to a much lesser
extent (Table 1, entry 5).
In order to explore the scope of the rearrangement
chemistry delineated above, pve 1bꢀg were synthesized
and subjected to the conditions of the uncatalyzed Goste-
liꢀClaisen rearrangement (Table 2).17 Heating of 1aꢀf in
DCE for an optimized period of time led to the formation
of the unstable allenes 2aꢀf; extensive decomposition was
observedfor1g. Determinationof the conversion after 38h
in refluxing DCE indicated a slighty higher reactivity of the
aryl-substituted pve 1cꢀe compared to 1a,b,f.
Figure 1. GosteliꢀClaisen rearrangement (R = CO2R) of pro-
pargyl vinyl ethers.
Our study commenced with an investigation of the
uncatalyzed GosteliꢀClaisen rearrangement. In the event
of (Z)-1a (Table 1), heating in 1,2-dichloroethane (DCE)
to 80 °C for 23 h provided incomplete conversion to the
expected R-keto ester 2a (Table 1, Entry 1).
Table 2. Uncatalyzed GosteliꢀClaisen Rearrangement in 1,2-
Dichloroethane (DCE)a
time
(h)
yieldb
(%)
Table 1. Uncatalyzed GosteliꢀClaisen Rearrangement of (Z)-
1a to 2a and Formation of Unexpected Products 3aꢀ5a
entry
substrate
R1
product
1
2
3
4
5
6
7
1a
1b
1c
1d
1e
1f
Me
Bn
65
60
27
48
27
47
38
2a
2b
2c
2d
2e
2f
85c
99
Ph
99
p-MeOꢀC6H4
p-NO2ꢀC6H4
H2CdCH
MeCtC
99
99
99
99d
1g
2g
a Conditions: substrate, DCE, 80 °C. b Isolated yield of unpurified allene.
c Loss of mass due to volatility of 2a. d Incomplete conversion: 1g:2g =
41:59. Prolonged reaction times resulted in extensive decomposition.
Next, we subjected 1bꢀg to the conditions that had
promoted cycloalkene formation from 1a (Table 3).
NMR analysis of the crude product mixture indicated
the formation of 3 and 5, except for 1b. The extent and
ratio of the formation of 3 and 5 is markedly substrate-
dependent. In particular, we note the propensity of 1e
(R1 = p-NO2ꢀC6H4) and 1g (R1 = MeCtC) for the
formation of 3e and 3g in moderate yield (60%). Remark-
ably, using HFIP which contained small amounts (370
ppm) of water prevented cycloalkene formation and led
only to the allenes 2aꢀg. Control experiments in which
2aꢀg were warmed in dry HFIP (1 ppm H2O) also
provided mixtures of 3aꢀg and 5aꢀg suggesting that the
allenes are intermediates in a multistep mechanism
(vide infra). The structures of 3aꢀg were deduced from
NMR studies and corroborated by an X-ray crystal
structure analysis of 3g which also provided the basis
for the assignment of the relative configuration of
3aꢀg.18 The structural assignment of the very sensitive
temp
time
(h)
yieldb
(%)
ratioc
entry
solventa
(°C)
(1a:2a:3/4a:5a)
1
2
3
4
5
DCE
TFE
80
80
60
60
80
23
23
99
99
99
99
93
29:71:ꢀ:ꢀ
ꢀ:100:ꢀ:ꢀ
22:78:ꢀ:ꢀ
4:28:28:40
ꢀ:85:3:12
HFIP
HFIP
TFE
23
72
132
a DCE: 14 ppm H2O, HFIP: 1 ppm H2O, TFE: 14 ppm H2O. b Yield
of the unpurified product mixture. c Determined by 1H NMR from the
unpurified product mixture.
In order to accelerate the rearrangement, we exploited
the solvent-rate effect of 1,1,1-trifluoroethanol (TFE).16
As intended, full conversion to 2a was observed after 23 h
at 80 °C (Table 1, entry 2). The allene 2a proved to be very
sensitive, and purification by chromatography led to ex-
tensive decomposition. In search of milder conditions, we
tested 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) for its
(14) Menz, H.; Kirsch, S. F. Org. Lett. 2006, 8, 4795–4797.
(15) Sherry, B. D.; Maus, L.; Laforteza, B. N.; Toste, F. D. J. Am.
Chem. Soc. 2006, 128, 8132–8133.
(16) (a) Rehbein, J.; Leick, S.; Hiersemann, M. J. Org. Chem. 2009,
74, 1531–1540. (b) Brandes, E.; Grieco, P. A.; Gajewski, J. J. J. Org.
Chem. 1989, 54, 515–516.
(17) 1aꢀg were synthesized following our aldol condensation ap-
proach; see: Hiersemann, M. Synthesis 2000, 1279–1290. For experi-
mental procedures and characterization, see the Supporting
Information.
€
(18) Gille, A.; Schurmann, M.; Preut, H.; Hiersemann, M. Acta
Crystallogr. 2009, E65, o1660.
Org. Lett., Vol. 13, No. 8, 2011
2123

cyclobutenes 5, which could not be completely separated
from 3, is based on NMR investigations.
Table 3. Dependence of Product Ratio in Hexafluoroisopro-
panol (HFIP) on Substrate Structurea
ratiob
yield of 3c
entry
substrate
R1
(1:2:3:5)
(%)
1
2
3
4
5
6
1b
1c
1d
1e
1f
Bn
ꢀ:100:ꢀ:ꢀ
ꢀ:ꢀ:40:60
ꢀ:ꢀ:55:45
ꢀ:ꢀ:85:15
ꢀ:ꢀ:33:67
7:ꢀ:93:ꢀ
ꢀ
Ph
28
21
52
14
60
p-MeOꢀC6H4
p-NO2ꢀC6H4
H2CdCH
MeCtC
1g
a Conditions: substrate, dry HFIP (1 ppm H2O as determined by
KarlꢀFischer titration), 60 °C, 72 h. b Determined from the 1H NMR of
the crude product mixture. c Isolated yield after chromatography.
Figure 2. B1B95/6-31þG(d,p)-calculated (relative) electronic
plus zero-point energies (ΔE) at 298.15 K in kcal/mol. B3LYP/
6-31þG(d,p)-values are given in parentheses. 1a: R1 = Me, 1g:
R1 = MeCtC. HFIP = hexafluoroisopropanol.
A mechanistic hypothesis for the formation of 3a,g from
1a,g along with relative energies from a DFT study using
the meta hybrid functional B1B9519 and the hybrid func-
tional B3LYP20 is depicted in Figure 2.21 Without con-
sidering the explicit role of the solvent HFIP, we propose
that the mechanistic cascade is initiated by the concerted
GosteliꢀClaisen rearrangement of 1a,g via the transition-
state structures 6a,g to deliver the experimentally detect-
able allenes 2a,g. The computational study supports our
proposal by predicting a reasonable gas-phase barrier and
a significant driving force.22,23 We then suggest the forma-
tion of the dienones 8a,g from 2a,g by consecutive 1,3- and
1,5-sigmatropic hydrogen shifts.24 The results of our cal-
culation indicate that this process would traverse via
energetically accessible enols 7a,g to more stable dienones
8a,g;25 structurally, the calculations predict a nonplanar,
helical conformation of 8a,g as a consequence of steric
interference between the substituents at the tetrasubsti-
tuted double bond. Due to a favorable balance in π- to
σ-bond conversion, a very large driving force is predicted
for the formation of 3a,g from 8a,g.
stepwise mechanism via the cyclopentenyl cation 11 could
account for the cyclization of 8to 3(Figure3).27 Imitating the
strong hydrogen bond donor ability of HFIP28 by assuming
an initial protonation of the carbonyl group of 8a,g to afford
9a,g, our calculations predict a low barrier and a significant
driving force for the formation of 11a,gby nucleophilic attack
of the terminal double bond on the carbonyl group. Subse-
quent nucleophilic attack by HFIP would lead to 3a,g.
The formation of the methylidenecyclobutenes 5 from the
vinyl allenes 7 by electrocyclization appears plausible,26 and a
Figure 3. B1B95/6-31þG(d,p)-calculated ΔE at 298.15 K in
kcal/mol. B3LYP/6-31þG(d,p)-values are given in parentheses.
(19) Becke, A. D. J. Chem. Phys. 1996, 104, 1040–1046.
(20) (a) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–
789. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(21) For computational details, see Supporting Information. The
calculations were carried out with Gaussian03; see: Frisch, M. J. et al.
Gaussian 03, revision E.01; Gaussian, Inc.: Wallingford, CT, 2004. Full
reference given in the Supporting Information.
(22) For a computational study on the Au(I)-catalyzed Claisen
rearrangement of propargyl vinyl ethers, see: Mauleon, P.; Krinsky,
J. L.; Toste, F. D. J. Am. Chem. Soc. 2009, 131, 4513–4520.
(23) For a computational study on the GosteliꢀClaisen rearrange-
ment of 2-alkoxycarbonyl-substituted allyl vinyl ethers, see: Rehbein, J.;
Hiersemann, M. J. Org. Chem. 2009, 74, 4336–4342.
(24) (a) Otter, B. A.; Saluja, S. S.; Fox, J. J. J. Org. Chem. 1972, 37,
2858–2863. (b) Bhat, L.; Ila, H.; Junjappa, H. J. Chem. Soc., Perkin
Trans. 1 1994, 1749–1752.
(25) The attempted catalytic asymmetric GosteliꢀClaisen rearrange-
ment of 1a using [Cu{(S,S)-t-Bu-box}(H2O)2](SbF6)2 (20 mol%) in
CH2Cl2 at ambient temperature provided (()-2a. We assume a rapid
racemization of 2a via the enol 7a in the presence of the Lewis acid.
(26) (a) Gil-Av, E.; Herling, J. Tetrahedron Lett. 1967, 8, 1–4. (b)
Lopez, S.; Rodriguez, J.; Rey, J. G.; de Lera, A. R. J. Am. Chem. Soc.
1996, 118, 1881–1891.
It seemed obvious at that point that the presence of a
Brønsted acid could further promote the cyclization.
Therefore, we treated 1aꢀg with HOSO2CF3 (HOTf, 0.5
equiv) in HFIP containing 370 ppm of H2O. In the event,
however, formation of tetrasubstituted furans was trig-
gered (Table 4). In detail, conversion of the methyl-sub-
stituted pve 1a proceeded sluggishly and delivered the
furan 12a in a mediocre yield (Table 4, entry 1); interest-
ingly, a much slower reaction and the formation of the
allene 2a were observed in dry HFIP (Table 4, entry 2).
Replacement of HFIP by DCE or CH2Cl2 led to enol ether
hydrolysis and formation of isopropyl 2-oxo-butyrate; no
(27) For a related cyclization, see: Iglesias, B.; de Lera, A. R.;
Rodriguez-Otero, J.; Lopez, S. Chem.;Eur. J. 2000, 6, 4021–4033.
€
€
(28) Berkessel, A.; Adrio, J. A.; Huttenhain, D.; Neudorfl, J. M.
J. Am. Chem. Soc. 2006, 128, 8421–8426.
2124
Org. Lett., Vol. 13, No. 8, 2011

reaction was observed in wet i-PrOH. The conversion
of the Bn-substituted pve 1b was low yielding, and a sig-
nificant amount of the allene 2b was detected (Table 4,
entry 3).29 The phenyl-substituted derivative 1c underwent
furan formation in good yield (Table 4, entry 4); notably,
the presence of a p-MeO-Ph group resulted in the forma-
tion of large amounts of the “abnormal” furan 13 (Table 4,
entry 5). Replacement of the p-MeO by a p-NO2 group
retarded furan formation (Table 4, entry 6). The vinyl- and
the 1-propinyl-substituted pve 1f and 1g showed no pro-
pensity at all for the formation of furans under Brønsted
acid catalysis (Table 4, entries 7,8).30
ambient temperature, as reported by Kirsch,10 catalyzed
the formation of the furans 12aꢀf in moderate to excellent
yield (48ꢀ96%),31 even with the capricious 1d (48%) or
previously unreactive 1f (57%). No furan formation was
detected in control experiments using only Ph3PAuCl or
AgSbF6. Alternatively, AuCl (0.1 equiv) in CH2Cl2 or the
combination of Ph3PAuCl (0.1 equiv) and Ag(BF4) (0.1
equiv) in CH2Cl2 also catalyzed furan formation but, in
our hands, appeared to be less active. Notably, for 1b and
1c, Cu(OTf)2 (0.1 equiv) in wet HFIP at ambient tempera-
ture catalyzed the formation of the furans 12b (75%) and
12c (84%) in good yield. As before, the 1-propinyl-sub-
stituted pve 1g opposed the participation in the cascade
process under various conditions and currently represents
the limiting case for the established methodology.
Table 4. Brønsted Acid Promoted Furan Formation in Wet
HFIP (370 ppm H2O)
Table 5. Au(I)-Catalyzed GosteliꢀClaisen Rearrangement/
Cycloisomerization of 1aꢀf to the Furans 12aꢀfa
entry substrate
R1
product t (min) yield (%) b
1
2
3
4
5
6
7
1a
1b
1c
1d
1e
1f
Me
Bn
Ph
12a
12b
12c
12d
12e
12f
20
78
89
90
48c
96
57
20
yielda
(%)
ratiob
20
entry
substrate
R1
1:2:12(:13)
p-MeOꢀC6H4
p-NO2ꢀC6H4
H2CdCH
150
20
1
2c
3
4
5
6
7
8
1a
1a
1b
1c
1d
1e
1f
Me
Me
44
70
34
84
99
86
47
ꢀ
ꢀ:ꢀ:100
69:16:15
ꢀ:43:57
20
d
1g
MeCtC
12g
24 h
Bn
a Conditions: substrate, dry CH2Cl2, Ph3PAuCl (0.05 equiv), Ag-
Ph
ꢀ:1:99
(SbF6) (0.05 equiv), rt. b Isolated yield after chromatographic purifica-
p-MeOꢀC6H4
p-NO2ꢀC6H4
H2CdCH
MeCtC
ꢀ:ꢀ:27(:73)
49:22:29
tion. c Along with 8% of 13. d 77% 1g reisolated.
100:ꢀ:ꢀ
d
1g
In conclusion, during the development of a catalyzed
cascade reaction that transforms 2-alkoxycarbonyl-substi-
tuted propargyl vinyl ethers into tetrasubstituted furans,
we observed an unprecedented carbacyclization involving
fluorinated alcohols as nucleophiles; we are currently
exploring the synthetic opportunities arising from the
discovery of this cascade reaction.
a Isolated yield after chromatography. b Determined from the 1H
NMR of the isolated product mixture. c Run in dry HFIP (1 ppm H2O).
d Small amounts of an unindentified byproduct were detected.
In light of these results, we surmised that the lack of
efficiency of the Brønsted acid catalyzed cascade reaction,
which culminates in furan formation, was caused by the
ineffective catalysis of the initiating GosteliꢀClaisen re-
arrangement. Consequently, and guided by the ground-
breaking contributions of Toste and Kirsch (vide supra),
we turned our attention to Au(I) catalysis (Table 5). Much
toour delight, a precatalystsystem consisting of Ph3PAuCl
(0.05 equiv) and Ag(SbF6) (0.05 equiv) in CH2Cl2 at
Acknowledgment. Financial support by the Technische
€
Universitat Dortmund is gratefully acknowledged. Com-
putational resources were provided by the ITMC of the
TU Dortmund. A.G. thanks Claudia God (TUDortmund)
for technical assistance.
Supporting Information Available. Text, tables, and
figures giving computational details, experimental proce-
dures, spectral and analytical data, and 1H and 13C NMR
spectra for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(29) The formation of allenes at ambient temperature under the
reaction conditions indicates that the GosteliꢀClaisen rearrangement
proceeds under specific-acid catalysis by the hydronium ion.
(30) 1aꢀg were also treated with PTSA•H2O (0.1 equiv) in HFIP (370
ppm H2O) at ambient temperature for 23 h. In the event, furan
formation was much slower for 1aꢀe; neither 1f nor 1g was converted
into the corresponding furan. Interestingly, however, subjecting the
allenes 2a and 2c to the identical conditions led to the high-yielding
formation of 12a (99%) and 12c (83%).
(31) We did not attempt a rigorous optimization of the catalyst
loading.
Org. Lett., Vol. 13, No. 8, 2011
2125