Unexpected electrophilic rearrangements of amides: A stereoselective entry to challenging substituted lactones

Article abstract of DOI:10.1002/anie.200906416

"Chemical Equation Presented" Surprise, surprise! An unexpected skeletal rearrangement was developed into a chemo- and stereoselective synthesis of aallyl and allenyl lactones with challenging substitution patterns (see scheme; EWC = electron-withdrawing group). The generality, unique features, and synthetic potential of this reaction were probed and a mechanism was proposed.

Full text of DOI:10.1002/anie.200906416

Angewandte
Chemie
DOI: 10.1002/anie.200906416
Rearrangements
Unexpected Electrophilic Rearrangements of Amides: A Stereoselec-
tive Entry to Challenging Substituted Lactones**
Claire Madelaine, Viviana Valerio, and Nuno Maulide*
Ever since the original preparation of ketenes and their
successful [2+2] cycloaddition with imines reported by
Staudinger in 1907,[1] heterocumulenes and their pericyclic
reactions have attracted considerable attention. In the latter
part of the 20th century, Ghosez and co-workers pioneered
the use of keteniminium salts as attractive alternatives to
ketenes for cycloadditions with alkenes to give cyclobuta-
nones.[2] Indeed, keteniminium salts are more electrophilic
than ketenes, show no tendency towards dimerization, and
can be prepared readily by treatment of an amide with
collidine and triflic anhydride.[3] In spite of their tremendous
potential as easily accessible, highly electrophilic reagents,
keteniminium salts generated from tertiary amides have
hardly been explored beyond the context of [2+2] cyclo-
additions.[4]
We became interested in the chemistry of keteniminium
salts as part of a research program aimed at the total synthesis
of a natural product. Having identified bicyclobutanone 3
(Scheme 1) as a pivotal intermediate in our plan, we
Scheme 1. Planned synthesis of compound 3 and unexpected observa-
tions. DCE=dichloroethane, Tf=trifluoromethanesulfonate, colli-
dine=2,4,6-trimethylpyridine.
envisioned its direct preparation from amide 1a through a
[2+2] cycloaddition of the keteniminium intermediate 2. In
Table 1: Optimization of the synthesis of 4a from 1a[a]
the event, treatment of g-allyloxyamide 1a with triflic
anhydride and collidine in dichloroethane did not lead to
the anticipated product 3. We were instead surprised to
observe the exclusive formation of a-allylbutyrolactone 4a,
after hydrolytic workup, in 40% yield (Scheme 1).
The unexpected skeletal reorganization and the complete
absence of any cyclobutanone product in the crude mixture
triggered our interest. Furthermore, the observed O-to-C allyl
transfer suggested an intermediary [3,3] sigmatropic rear-
rangement, and the inherent potential of such reactions in the
present context decisively warranted further inspection. We
present herein an unprecedented “Claisen[5,6]-like” rear-
rangement of keteniminium salts allowing direct and stereo-
selective access to challenging substituted lactones.
Entry
Solvent
Base
T [8C]
Yield of 4a [%]
1
2
3
4
5
6
7
8
THF
collidine
collidine
collidine
collidine
collidine
collidine
collidine
collidine
iPr2NEt
40
40
40
40
40
40
80
40
–
–
pyridine
toluene
MeCN
DCE
DCE
DCE
DCM
DCM
DCM
DCM
35
25
40[b]
[c]
–
51[b]
55[b]
10
9
10
11
40
DTBMP
collidine
40
23
120[d]
90[b]
[a] All reactions were conducted with 1.2 equivalents of base and
1.05 equivalents of Tf2O under the reported conditions. Yields were
estimated by H NMR analysis unless indicated otherwise. [b] Yield of
1
Systematic variation of the reaction parameters quickly
revealed that chlorinated solvents are optimal (Table 1,
isolated product. [c] Oxalyl chloride was used instead of Tf2O. [d] Micro-
wave irradiation for 5 min. DTMBP=2,6-di-tert-butyl-4-methylpyridine.
[*] Dr. C. Madelaine, V. Valerio, Dr. N. Maulide
Max-Planck-Institut fꢀr Kohlenforschung
Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim (Germany)
Fax: (+49)208-306-2999
entries 1–5 and 8). Among the amine bases tested, collidine
afforded the best results (Table 1, entries 8–10), whilst triflic
anhydride proved to be the most suitable activating agent
(Table 1, entries 5 and 6). In accordance with a high energy
barrier to the formation of the keteniminium intermediate,
higher yields were obtained at temperatures close to reflux
(Table 1, entries 5 and 7). Microwave irradiation also proved
beneficial; the optimal protocol thus entails microwave
irradiation of the reaction mixture in dichloromethane for
only 5 min, followed by simple hydrolytic workup with
aqueous bicarbonate solution. Under these conditions, a-
E-mail: maulide@mpi-muelheim.mpg.de
Homepage: http://www.kofo.mpg.de/maulide
[**] This work was supported financially by the Deutsche Forschungs-
gemeinschaft (grant MA4861/1-1). We are grateful to the Max-
Planck Society and the Max-Planck Institut fꢀr Kohlenforschung for
generous funding of our research programs. Dr. C. Wirtz is kindly
acknowledged for skillful NMR assistance.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200906416.
Angew. Chem. Int. Ed. 2010, 49, 1583 –1586
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1583

Communications
allylbutyrolactone 4a was generated in an excellent 90%
yield. It is worth mentioning that during this optimization we
never detected even traces of the [2+2] cycloadduct 3 or
related isomers.
Applying the optimized conditions for this skeletal
rearrangement, we then proceeded to examine its scope and
functional-group tolerance (see Table 2). A wide variety of
Table 2: Scope of the rearrangement reaction.[a]
Yield[%][b]
Scheme 2. Results with differently substituted allyl moieties.
Entry
Substrate
Product
as the exclusive reaction products—a consequence of the allyl
inversion expected from a [3,3] sigmatropic shift.[8,9] More
importantly, high diastereoselectivities were observed in most
cases.[10] For instance, our system is able to effectively
differentiate between ethyl and vinyl substituents as in
product 6c, and easily generate nearly diastereopure lactone
6a. Both of these products would be almost impossible to
prepare by direct alkylation of the corresponding allylic
halides. It is therefore not surprising that such simple
compounds are, to the best of our knowledge, unknown in
the chemical literature.
1
2
3
4
5
6
7
8
R=H (1a)
R=Me (1b)
R=CO2Et (1c)
R=CH2OMe (1d)
R=CH2OAc (1e)
R=CN (1 f)
4a
4b
4c
4d
4e
4 f
90
61[c]
78
50
54
65
R=Br (1g)
R=CH2Cl (1h)
4g
4h
33
75[d]
At this juncture, we became intrigued with the possibility
of extending the substrate scope to propargyl ethers. The
preparation of allenes through the thermal, metal-free [3,3]
sigmatropic rearrangement of alkynes is known to be
difficult.[11,12] We thus awaited the outcome of the reaction
of propargyl ether 7a with some trepidation (Scheme 3). In
the event, alkyne 7a smoothly rearranged to a-allenyl lactone
8a in an excellent yield for such a challenging transformation.
Neither the isomeric propargylated butyrolactone nor [2+2]
cycloadducts were detected in the reaction mixture. As can be
further inferred from Scheme 3, this reaction also appears
9
R=H (1i)
4i
57
10
11
R=H (1j)
R=CO2Et (1k)
4j
4k
50
49
[a] All reactions were conducted with 1.2 equivalents of base and
1.05 equivalents of Tf2O unless otherwise noted. [b] Yield of isolated
product. [c] 3 equivalents of base used. [d] 1.5 equivalents of Tf2O used.
allyloxyamides could be converted into the corresponding
lactones in good to excellent yields. Ester, nitrile, and halide
functionalities are well tolerated, and oxygenated substrates
were also amenable to this reaction (Table 2, entries 3–8).
Isomeric [2+2] cycloaddition products were never detected in
the crude mixtures. This suggests that the electronic proper-
ties of the allyl moiety do not affect the reaction. It is
interesting to note that the length of the tether could be
increased with only a slight decrease in yield (Table 2,
entry 9). Moreover, this reaction has the potential to generate
products containing all-carbon-substituted quaternary centers
(Table 2, entries 10 and 11).
Having realized that this unusual reaction is indeed
general, we proceeded to probe substrates containing more
challenging allyl moieties. Since the starting allyloxyamides
are accessible by simple allylation of the corresponding
hydroxy derivatives, a range of substrates could be synthe-
sized readily (Scheme 2).[7] In the event, lactones bearing
branched allyl substitutents in the a position were generated
Scheme 3. Synthesis of a-allenyl lactones.
1584
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1583 –1586

Angewandte
Chemie
quite general with respect to the acetylenic component. In
each case, the crude reaction mixtures are extremely clean
and the allenes are the only products detected; only their
pronounced volatility precluded higher yields of isolated
products. To our knowledge, all of these products were
hitherto unknown in the chemical literature. Indeed, and
despite their apparent structural simplicity, straightforward
routes to such compounds are not particularly easy to
conceive.
Our mechanistic rationale for this transformation is
depicted in Scheme 4. After initial activation of the amide
Scheme 5. Unambiguous distinction between allyl transfer pathways in
the rearrangement of 1a.
mechanistic scenario. Further evidence in support of the
direct [3,3] shift was obtained in experiments with the 13C-
labeled allyl ether 1a* (Scheme 5). Upon rearrangement, the
lactone 4a* bearing the label exclusively at the unsaturated
terminus of the allyl group was obtained as the only reaction
product (NMR analysis of the crude reaction mixture).[7] This
unambiguous result lends further support to a direct “Claisen-
like” sigmatropic rearrangement (Scheme 5, path a). Based
on such a mechanistic rationale, we propose that the highly
stereoselective reorganization of substrates 5a and 5c pro-
ceeds through
a
boatlike transition-state structure
(Scheme 6), in order to minimize eclipsing interactions
between the allyl substituent(s) and the dihydrofuran ring.[10]
Scheme 4. Proposed reaction mechanism.
carbonyl by the electrophilic reagent, elimination to form
keteniminium 2 probably takes place. The enhanced electro-
philicity of this intermediate then triggers an unusual
nucleophilic attack of the ether, which may be reversible.
This addition, however, generates a vinyl allyl oxonium
intermediate B. Such a species should be ideally poised to
undergo a “Claisen-like” [3,3]-sigmatropic rearrangement,
leading to the stabilized carbenium ion C, hydrolysis of which
then accounts for the formation of lactone 4a. From a
mechanistic point of view, this reaction is reminiscent of the
Scheme 6. Mechanistic proposal for diastereoselectivity.
Finally, we evaluated a possible asymmetric version
(Scheme 7). Our initial results were promising: the reaction
of the readily available chiral amide 9 led to lactone 10 in high
yield and 80:20 e.r.[7,18] This preliminary result opens up
exciting vistas for this chemistry.
Claisen–Eschenmoser–Ficini–Bellus
rearrangements.[13–16]
However, and in stark contrast to the latter, the reactive
intermediates herein are generated from stable, easily han-
dled amides under essentially neutral conditions.
From the outset, we were aware of possible, alternative
mechanisms to account for the transfer of the allyl moiety
(Scheme 5). Two such mechanisms in particular appeared
plausible: a direct or stepwise 1,3-allyl migration (not shown)
and a nitrogen-assisted allyl transfer proceeding via a
In summary, we have developed an efficient synthesis of
challenging substituted lactones of different ring sizes. This
quaternary allylammonium species
D
(path b in
Scheme 5).[17] For geometric reasons, the latter process
should take place with double allylic inversion.
Our results with substrates having more complex allyl
(and propargyl) residues already strongly argue against such a
Scheme 7. Preliminary experiments towards an asymmetric variant.
Angew. Chem. Int. Ed. 2010, 49, 1583 –1586
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1585

Communications
[5] L. Claisen, Chem. Ber. 1912, 45, 3157.
unprecedented “Claisen-like” rearrangement draws from the
[6] For reviews, see: a) G. B. Bennett, Synthesis 1977, 589; b) F. E.
Ziegler, Chem. Rev. 1988, 88, 1423; c) S. Blechert, Synthesis 1989,
71; d) U. Nubbemeyer, Synthesis 2003, 961; e) A. M. Martin
Castro, Chem. Rev. 2004, 104, 2939. See also: f) The Claisen
Rearrangement: Methods and Applications (Eds.: M. Hierse-
mann, U. Nubbemeyer), Wiley-VCH, Weinheim, 2007, and
references therein.
high electrophilicity of keteniminium derivatives and com-
pletely disrupts their well-known propensity to undergo [2+2]
cycloaddition reactions with olefins. The reaction proceeds
under essentially neutral conditions and tolerates a variety of
useful functionalities. Furthermore, it allows the generation of
all-carbon quaternary centers, as well as the stereoselective
[7] See the Supporting Information for details.
À
formation of conjugate C C bonds and (perhaps most
[8] In contrast to other related reactions: a) A. Fꢁrstner, H. Szillat,
F. Stelzer, J. Am. Chem. Soc. 2000, 122, 6785; b) A. Fꢁrstner, F.
Stelzer, H. Szillat, J. Am. Chem. Soc. 2001, 123, 11863.
[9] Y.-J. Li, Y.-K. Chang, G.-M. Ho, H.-M. Huang, Org. Lett. 2009,
11, 4224.
[10] For further discussion of the diastereoselectivity of this reaction
and additional experiments, see the Supporting Information.
[11] For a metal-catalyzed propargyl-Claisen rearrangement, see:
B. D. Sherry, F. D. Toste, J. Am. Chem. Soc. 2004, 126, 15978.
[12] M. O. Frederick, R. P. Hsung, R. H. Lambeth, J. A. Mulder,
M. R. Tracey, Org. Lett. 2003, 5, 2663.
strikingly) a-allenyl lactones; most of the compounds
reported herein would be very difficult to access by other
routes. Evidence in support of a [3,3]-sigmatropic mechanism
has been garnered. Additionally, promising results indicate
the potential for the development of an asymmetric version.
Efforts aimed at broadening the scope of this methodology,
elucidating mechanistic intricacies, and developing highly
enantioselective asymmetric versions are currently underway
in our laboratories.
[13] a) J. Ficini, C. Barbara, Tetrahedron Lett. 1966, 7, 6425; b) J.
Ficini, N. Lumbroso-Bader, J. Pouliquen, Tetrahedron Lett. 1968,
9, 4139; c) P. A. Bartlett, W. F. Hahne, J. Org. Chem. 1979, 44,
882.
[14] A. E. Wick, D. Felix, K. Steen, A. Eschenmoser, Helv. Chim.
Acta 1964, 47, 2425.
Received: November 13, 2009
Published online: February 2, 2010
Keywords: Claisen rearrangement · keteniminium ions ·
.
lactones · rearrangement · stereoselective synthesis
[15] J. A. Mulder, R. P. Hsung, M. O. Frederick, M. R. Tracey, C. A.
Zificsak, Org. Lett. 2002, 4, 1383.
[16] a) R. Malherbe, G. Rist, D. Bellus, J. Org. Chem. 1983, 48, 860;
b) J. Gonda, Angew. Chem. 2004, 116, 3600; Angew. Chem. Int.
Ed. 2004, 43, 3516, and references therein. For other mechanis-
tically related reactions, see e.g.: c) D. J. Pollart, H. W. Moore, J.
Org. Chem. 1989, 54, 5444.
[17] A reviewer suggested an alternative possible scenario: the
formation of a triflyl-enamine and its subsequent intramolecular
attack onto the allyl (or propargyl) moiety in SN2’ fashion,
followed by cyclization of the resulting alcohol. We believe that
the implied seven- or eight-membered-ring transition-state
structure for the SN2’ attack might be considerably higher in
energy than the rationale presented herein.
[1] H. Staudinger, Justus Liebigs Ann. Chem. 1907, 356, 51.
[2] For leading references from the Ghosez group, see: a) L.
Ghosez, J. Marchand-Brynaert, in Iminium Salts in Organic
Chemistry, Part 1 (Eds.: H. Bꢀhme, H. G. Viehe), Wiley, New
York, NY, 1976, pp. 421; b) J.-B. Falmagne, J. Escudero, S. Taleb-
Sahraoui, L. Ghosez, Angew. Chem. 1981, 93, 926; Angew.
Chem. Int. Ed. Engl. 1981, 20, 879; c) L. Ghosez, I. Marko, A.-M.
Hesbain-Frisque, Tetrahedron Lett. 1986, 27, 5211; d) C. Schmit,
J. B. Falmagne, J. Escudero, H. Vanlierde, L. Ghosez, Org. Synth.
1990, 69, 199; e) C. Genicot, B. Gobeaux, L. Ghosez, Tetrahe-
dron Lett. 1991, 32, 3827.
[3] For a review on keteniminium chemistry, see: a) B. B. Snider,
Chem. Rev. 1988, 88, 793. For further examples, see: b) B. B.
Snider, R. A. H. F. Hui, J. Org. Chem. 1985, 50, 5167.
[4] For a Pictet–Spengler cyclization of a keteniminium salt, see: Y.
Zhang, R. P. Hsung, X. Zhang, J. Huang, B. W. Slafer, A. Davis,
Org. Lett. 2005, 7, 1047.
[18] To our knowledge, the only alternative reported synthesis of
enantioenriched (ca. 85:15 e.r.) 10 entails five steps from
commercially available substances: A. I. Meyers, E. D. Mihelich,
J. Org. Chem. 1975, 40, 1186.
1586
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1583 –1586