Trimethylaluminum-triflimide complexes for the catalysis of highly hindered diels-alder reactions

Article abstract of DOI:10.1021/ol302172y

Two catalysts, Me₂AlNTf₂ and MeAl(NTf ₂)₂, derived from the mixing of trimethylaluminum with triflimide, proved to be highly effective catalysts in hindered Diels-Alder reactions, generating the desired Diels-Al

Full text of DOI:10.1021/ol302172y

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
TrimethylaluminumÀTriflimide
Complexes for the Catalysis of Highly
Hindered DielsÀAlder Reactions
Michael E. Jung* and Mikhail Guzaev
Department of Chemistry and Biochemistry, University of California, Los Angeles,
California 90095, United States
jung@chem.ucla.edu
Received August 4, 2012
ABSTRACT
Two catalysts, Me2AlNTf2 and MeAl(NTf2)2, derived from the mixing of trimethylaluminum with triflimide, proved to be highly effective catalysts in
hindered DielsÀAlder reactions, generating the desired DielsÀAlder cycloadducts from both hindered 2-silyloxydienes and hindered
dienophiles. Thus reaction of 1 with 2 afforded the hindered cycloadduct 4 in excellent yield in 0.5À1.5 h at À40 °C.
The DielsÀAlder reaction is ubiquitous in organic
synthesis.1 In particular substituted silyloxydienes have
proved very useful in this reaction.2 We have previously
Scheme 1. Triflimide-Catalyzed DielsÀAlder Reaction
found two very effective catalyst systems, namely a 5:1
mixture of aluminum tribromide (AlBr3) and trimethyla-
luminum (AlMe3) and tert-butyldimethylsilyl triflamide
(TBSNTf2), prepared in situ from triflimide (Tf2NH) and a
silyl enol ether. These have proven to be valuable and
potent catalysts for more challenging transformations of
this kind.3 However, in our studies of a novel synthetic
approach toward Rhodexin A, neither of these methods
provided satisfactory results.4 Yamamoto and co-workers
recently reported the generation of highly active Lewis acid
species through the formation of triflimide complexes of
(1) For reviews of DielsÀAlder reactions, see: (a) Nicolaou, K. C.;
Snyder, S. A.; Montagnon, T.; Vassilikogiannakis, G. Angew. Chem.,
trimethylaluminum.5 However, to the best of our knowl-
edge these species have never been applied to the
Int. Ed. 2002, 41, 1668–98. (b) Hayashi, Y. Cycloaddit. React. Org.
Synth. 2002, 5–55. (c) Whiting, A. Adv. Asymmetric Synth. 1996, 126–
145. (d) Oppolzer, W. Intermolecular DielsÀAlder Reactions. In Com-
prehensive Organic Synthesis; Trost, B. M., Ed.; Pergamon Press: Oxford,
DielsÀAlder reaction. We therefore wanted to explore
U.K., 1991; Vol. 5, Chapter 4.1, pp 315À99.
their synthetic viability in this reaction.
(2) Danishefsky, S. Acc. Chem. Res. 1981, 14, 400–406.
(3) (a) Jung, M. E.; Ho, D.; Chu, H. V. Org. Lett. 2005, 7, 1649–51.
(b) Jung, M. E.; Ho, D. Org. Lett. 2007, 9, 375–78. (c) Jung, M. E.; Chu,
H. V. Org. Lett. 2008, 10, 3647–49.
(4) For the synthesis of Rhodexin A, see: Jung, M. E.; Yoo, D. Org.
Lett. 2011, 13, 2698–01 and references therein.
(5) (a) Marx, A.; Yamamoto, H. Angew. Chem., Int. Ed. 2000, 39,
178–81. (b) Boxer, M. B.; Yamamoto, H. Org. Lett. 2005, 7, 3127–29.
(c) Brady, P. B.; Yamamoto, H. Angew. Chem., Int. Ed. 2012, 51, 1942–46.
We recently reported a triflimide promoted stepwise
MukaiyamaÀMichael vinylogous aldol addition of hin-
dered silyloxy dienes 1 to hindered enones 2 to give formal
DielsÀAlder cycloadducts (Scheme 1).3 This method al-
lows easy synthetic access to highly hindered systems that
are difficult to obtain via alternative Lewis acid catalysis or
r
10.1021/ol302172y
XXXX American Chemical Society

indeed by any other synthetic route. Although good yields
could be obtained in many systems, transformations were
complicated by considerably prolonged reaction times
under AlBr3ÀAlMe3 catalysis and the tendency of trifli-
mide catalysis to stop at the MukaiyamaÀMichael inter-
mediates, such as the enone derived by desilylation of 3,
without proceeding to the cyclized product 4. In an attempt
to improve upon these conditions, we explored the use of the
potentially more highly active Lewis acid systems described
by Yamamoto and co-workers and here report the results.
The reaction of the silyl enol ether 1 with 3-methyl-
2-cyclohexen-1-one 2 has previously been shown to yield
the desired cycloadduct 4 in 92% yield under Tf2NTBS
catalysis (prepared in situ from reaction of 1 with trifli-
mide) (Table 1, entry 1).3 We observed that Me2AlNTf2
(prepared by mixing Me3Al with equimolar amounts
of Tf2NH) resulted in a marked rate acceleration of this
reaction, while affording comparable yields of the cycload-
duct 4, 96% in 1.5 h (Table 1, entry 2). Use of an additional
equivalent of triflimide (to generate MeAl(NTf2)2) pro-
vided further rate acceleration with a similar yield, 94%
in 0.5 h (Table 1, entry 3). Based on these findings, we
have identified Me2AlNTf2 and MeAl(NTf2)2 as poten-
tially useful catalysts for highly hindered DielsÀAlder
cycloadditions. We found that a Me2AlNTf2 catalyst
loading of 3À10% gave excellent yields. Interestingly, the
Table 2. Me2AlNTf2 Catalyzed DielsÀAlder Cycloadditions
Table 1. AluminumÀTriflimide Catalysts for the Preparation of 4
entry
catalyst
Tf2NH
time (h)
yield (%)a
1
2
3
4
4
92
Me2AlNTf2
MeAl(NTf2)2
Al(NTf2)3
1.5
0.5
0.1
96
94
N/A
a Isolated yield based on 3-methyl-2-cyclohexen-1-one.
tris(triflimido)aluminum species, Al(NTf2)3,6 did not re-
sult in constructive catalysis, creating instead a complex
mixture of products (Table 1, entry 4). In addition, the use
of solvents other than dichloromethane, including toluene,
benzene, diethyl ether, and tetrahydrofuran, gave poorer
results.7
Based on these results, we then proceeded to screen
Me2AlNTf2 as a catalyst in a series of DielsÀAlder
cycloadditions with hindered dienes and/or dienophiles
(Table 2). Initially we focused on improving the extensive
reaction times and/or difficult conditions of a number
of substrates that have been reacted with the mixed
Lewis acid catalyst (AlBr3ÀMe3Al) or Tf2NTBS. Using
Me2AlNTf2 as the catalyst, we reacted the silyl enol ether 1
with the enone 2 to produce the desired cycloadduct 4 in
excellent yield under mild conditions (0 °C, 0.5 h) (Table 2,
entry 1).8 Similarly, reaction of 1 with the enone 5 afforded
the previously obtained product 6 in 71% yield, with the
reaction time drastically reduced from the reported 48 h.3b
As we had seen earlier, the double bond of the silyl enol
(6) Mikami, K.; Kotera, O.; Motoyama, Y.; Sakaguchi, H.; Maruta,
M. Synlett 1996, 171–2.
(7) The catalyst system was poorly soluble in toluene and benzene,
the yields decreased dramatically in diethyl ether, and THF was highly
prone to polymerization under these conditions.
(8) In all cases, 1.3 equiv of the silyloxydiene were used.
B
Org. Lett., Vol. XX, No. XX, XXXX

Interestingly, when silyloxycyclohexadiene 13 and the
enone 5 were left under the same reaction conditions, that
is, Me2AlNTf2 in DCM at 0 °C for 14 h, we isolated, in
addition to the expected exo adduct 16, the novel twista-
none 18 (Scheme 2). It is likely that this compound is
formed via enolization of the cyclohexanone 15 to give
an enol derivative 19 (A = H, Tf) followed by protonation
of the silyl enol ether from the less hindered exo face to give
the stabilized carbocation 20 and final CÀC bond forma-
tion via an aldol-like reaction (Scheme 3). This finding
proved helpful in elucidating the diastereoselectivity of the
reaction, as this structure is geometrically accessible only
via the endo DielsÀAlder product 15.
Scheme 2. Formation of Twistanone 18 from 13 and 5
ether migrated easily to the allylic position and yielded
exclusively the migrated product 6 (Table 2, entry 2).
Likewise the reaction of 1 with the hindered enone 7 gave
the desired cycloadduct 8 in good yield under much milder
conditions than those previously used (Table 2, entry 2).3b
The hindered cyclopentenyl diene system 9 also reacted
well with the dienophiles 2, 5, and 7 under Me2AlNTf2
catalysis to afford the cycloadducts 10, 11, and 12 in fair to
excellent yields (41À99%). In the first two cases, a mixture
of diastereomers was obtained, 5:2 for 10nx and 1:1 for
11nx,9 while in the last case the pure diastereomer was
able to be separated out10 (Table 2, entries 4À6). In order
to investigate the formation of bridged polycyclic com-
pounds, we prepared the cyclic diene, 2-trialkylsilyloxy-
3-methyl-1,3-cyclohexadiene 13, and reacted it with the same
series of dienophiles. Withthe dienophile 2, weobtained an
approximate 5:4 mixture of the endo and exo adducts
14nx in 48% yield, whereas the dienophile 5 gave a 3:2
mixture of endo and exo diastereomers 15 and 16 in
92% yield9 (Table 2, entries 7À8). Finally the very
hindered dienophile 7 afforded the endo adduct 17 in
fair yield (Table 2, entry 9). All of the cycloadditions
shown in Table 2 occurred under very mild conditions
even though the products were quite hindered, thus
showing the efficacy of this new catalyst system.
Scheme 3. Possible Mechanism for Formation of Twistanone 18
In summary, we have demonstrated the efficacy of
complexes of Me3Al in combination with Tf2NH in cata-
lyzing very hindered DielsÀAlder reactions. The reactivity
of Me2AlNTf2 has permitted isolation of products that are
either difficult to obtain or inaccessible through conven-
tional Tf2NH catalysis. In addition, MeAl(NTf2)2 shows
promise as a more powerful Lewis acid with even greater
rate acceleration. Further studies on the use of these com-
plexes in natural product synthesis are in progress.
Acknowledgment. M.G. gratefully acknowledges the
support of a UCLA Graduate Division Dissertation Year
Fellowship.
Supporting Information Available. Experimental pro-
cedures and proton and carbon NMR for all new com-
pounds. This material is free of charge via the Internet at
http://pubs.acs.org.
(9) The structures of the major and minor diastereomers have not
been rigorously proven and are assigned based on mechanistic reasons.
(10) Interestingly, the reaction of 5 and 9 at 0 °C for only 1 h gave
mainly the simple MukaiyamaÀMichael product.3b
The authors declare no competing financial interest.
Org. Lett., Vol. XX, No. XX, XXXX
C