Studies of intramolecular cyclizations of N-acyliminium ions derived from acyclic ketones: Unanticipated stereochemical and structural results

Article abstract of DOI:10.1021/ol035046m

(Matrix presented) Intramolecular cyclizations of a series of (E)- and (Z)-olefinic acyclic ketone-derived N-acyliminium ions have been studied. It has been found that both the course of the reaction and the stereochemistry of the products are critically

Full text of DOI:10.1021/ol035046m

ORGANIC
LETTERS
2003
Vol. 5, No. 16
2915-2918
Studies of Intramolecular Cyclizations of
N-Acyliminium Ions Derived from
Acyclic Ketones: Unanticipated
Stereochemical and Structural Results
Wenchun Chao, Jacob H. Waldman, and Steven M. Weinreb*
Department of Chemistry, The PennsylVania State UniVersity,
UniVersity Park, PennsylVania 16802
smw@chem.psu.edu
Received June 10, 2003
ABSTRACT
Intramolecular cyclizations of a series of (E)- and (Z)-olefinic acyclic ketone-derived N-acyliminium ions have been studied. It has been found
that both the course of the reaction and the stereochemistry of the products are critically dependent upon the tether length and olefin geometry
of the cyclization substrate.
N-Acylimines and N-acyliminium ions have become impor-
tant synthons in the construction of a wide array of nitrogen-
containing molecules, as well as nitrogen heterocycles, due
to their susceptibility to nucleophilic attack and their ability
to act as partners in several cycloaddition reactions.1 Much
of the early classical work on aromatic amidoalkylation
reactions made use of “linear” N-acylimines derived from
the direct condensation of highly electrophilic nonenolizable
aldehydes (e.g., formaldehyde, chloral, glyoxylate, etc.) with
amides.2 These kinds of acyclic N-acylimines have also been
used in other synthetic processes.1 However, to date there
has been minimal usage of linear N-acylimines derived from
enolizable aldehydes, perhaps due to a somewhat limited
amount of efficient methodologies for generating these
species,3 along with their propensity to hydrolyze readily.
Interestingly, even less has been done with acyclic acylimines
derived from ketones.
It has been known for a number of years that N-acylimines
and iminium salts can act as azadienes in hetero Diels-Alder
cycloadditions with olefins to produce 5,6-dihydro-1,3-
oxazines.4,5 Several years ago, we described the first
examples of intramolecular [4 + 2]-cycloadditions of this
type utilizing N-acyliminium salts derived from enolizable
aldehydes and showed that these reactions are completely
stereospecific.6 Thus, we found that it was possible to
generate N-acyliminium ions in situ from unsaturated alde-
(3) See, for example: (a) Johnson, A. P.; Luke, R. W. A.; Steele, R.
W.; Boa, A. N. J. Chem. Soc., Perkin Trans. 1 1996, 883. (b) Johnson, A.
P.; Luke, R. W. A.; Boa, A. N. J. Chem Soc., Perkin Trans. 1 1996, 895.
(c) Al-Talib, M.; Zaki, M.; Hehl, S.; Stumpf, R.; Fischer, H.; Jochims, J.
C. Synthesis 1996, 1115. (d) Hoffman, R. V.; Nayyar, N. K.; Shankweiler,
J. M.; Klinekole, B. W., III. Tetrahedron Lett. 1994, 35, 3231. (e) Chao,
W.; Weinreb, S. M. Tetrahedron Lett. 2000, 41, 9199.
(4) For reviews of uses of N-acylimines as heterodienes, see: (a) Boger,
D. L.; Weinreb, S. M. Hetero Diels-Alder Methodology in Organic
Synthesis; Academic Press: San Diego, 1987; Chapter 9. (b) Weinreb, S.
M.; Scola, P. M. Chem. ReV. 1989, 89, 1525.
(5) For more recent examples, see: (a) Gizecki, P.; Dhal, R.; Toupet,
L.; Dujardin, G. Org. Lett. 2000, 2, 585. (b) Shimizu, T.; Tanino, K.;
Kuwajima, I. Tetrahedron Lett. 2000, 41, 5715. (c) Suga, S.; Nagaki, A.;
Tsutsui, Y.; Yoshida, J.-I. Org. Lett. 2003, 5, 945. (d) Gizecki, P.; Dhal,
R.; Poulard, C.; Gosselin, P.; Dujardin, G. J. Org. Chem. 2003, 68, 4338.
(1) For extensive reviews of the chemistry of N-acylimines and iminium
ions, see: (a) Speckamp, W. N.; Hiemstra, H. Tetrahedron 1985, 41, 4367.
(b) Hiemstra, H.; Speckamp, W. N. In N-Acyliminium Ions as Intermediates
in Alkaloid Synthesis In The Alkaloids; Brossi, A., Ed.; Academic Press:
New York, 1988; Vol. 32, p 271. (c) Hiemstra, H.; Speckamp, W. N.
Additions to N-Acyl Iminium Ions. In ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 2, p 1047.
(d) Speckamp, W. N.; Moolenaar, M. J. Tetrahedron 2000, 56, 3817.
(2) (a) Zaugg, H. E.; Martin, W. B. Org. React. 1965, 14, 52. (b) Zaugg,
H. E. Synthesis 1984, 85, and 181.
10.1021/ol035046m CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/16/2003

hyde bis-amides such as 1 using boron trifluoride etherate
as a Lewis acid catalyst at room temperature. These
intermediates then cyclized to afford, with both (E)- and (Z)-
olefinic dienophiles, exclusively trans-fused bicyclic dihy-
drooxazines such as 3 (Scheme 1). We rationalized this
Scheme 2
Scheme 1
stereochemical outcome on the basis of the assumption that
due to unfavorable nonbonded interactions, the N-acylimin-
ium ion transition state conformer 4, leading to the cis-fused
product 5, is of higher energy than conformer 2, which would
afford the observed trans-fused product 3. A similar prefer-
ence for the trans product was also observed in systems
leading to 6,5-fused dihydrooxazines.6 In addition, since
synthesis of bis-amides such as 1 from aldehydes often
suffers from low yields, we have developed a new and more
efficient approach for generation of the N-acyliminium ion
heterodienes using our radical transposition methodology.3e
More recently, we considered the possibility of extending
this intramolecular hetero Diels-Alder chemistry to ketone-
derived N-acyliminium ions7 since the [4 + 2]-cycloadducts
from this variation of the methodology would have a
quaternary center adjacent to nitrogen (vide infra), and we
believed these heterocycles might be attractive intermediates
for alkaloid synthesis. It was our expectation a priori that
the stereochemistry of the ketone-derived systems should
parallel that of the aldehyde N-acyliminium ion substrates.
However, as described in this communication, we have found
that there are significant and unexpected differences in the
chemistry of these two types of systems.
The substrates for the cycloadditions were readily prepared
by the route shown in Scheme 2. Thus, alkylation of tert-
butyl acetoacetate (6) with some (E)- and (Z)-olefinic primary
iodides afforded products 7a-d (see Supporting Informa-
tion). Decarboxylation of these compounds produced ketones
8a-d, which could be converted in two steps to the desired
enamides 9a-d (mixtures of isomers).
To effect the desired Diels-Alder reaction, substrate 9a
was exposed to trifluoroacetic acid in 1,2-dichloroethane at
room-temperature overnight (Scheme 3).3e A single cycload-
duct was produced in 65% yield which, to our surprise,
proved to be the cis-fused bicyclic dihydro-1,3-oxazine 11.
The structure of adduct 11 was confirmed by its conversion
into a crystalline cyclic carbamate 12 by hydrogenolysis of
Scheme 3
The strategy was to generate the requisite ketone-derived
N-acyliminium salt by protonation of an enamide (cf. 9,
Scheme 2). Since we knew from previous experience that
enamides bearing an alkyl group on nitrogen are more stable
to hydrolysis by adventitious moisture than unsubstituted
(NH) enamides and also more easily prepared, we opted to
investigate N-p-methoxybenzyl (PMB)-substituted cases. In
addition, on the basis of observations in our previous work,
we anticipated that this substituent would be lost under the
conditions of the cycloaddition.3e
(6) (a) Scola, P. M.; Weinreb, S. M. J. Org. Chem. 1986, 51, 3248. (b)
Scola, P. M. Ph.D. Thesis, The Pennsylvania State University, University
Park, Pennsylvania, 1993.
(7) A few cases exist of intermolecular hetero Diels-Alder cycloadditions
involving N-acylimines derived from nonenolizable ketones. See for
example: (a) Hall, H. K., Jr.; Miniutti, D. L. Tetrahedron Lett. 1984, 25,
943. (b) Safronova, Z. V.; Simonyan, L. A.; Zeifman, Y. V.; Gambaryan,
N. P. Bull. Acad. Sci USSR, DiV. Chem. Sci. (Engl. Transl.) 1979, 1688.
2916
Org. Lett., Vol. 5, No. 16, 2003

the dihydrooxazine,8 followed by treatment of the resulting
amino alcohol with 1,1′-carbonyldiimidazole and subsequent
X-ray analysis.9 It seems likely that dihydrooxazinium salt
10 is an intermediate in this process but is not observed since
it apparently undergoes facile dealkylation in situ. We cannot,
however, rule out the possibility that the PMB group comes
off prior to the cyclization. Interestingly, it was subsequently
found that the isolated yield of cyclization product 11
improves to 88% if, after ∼20 h at room temperature, boron
trifluoride etherate and anisole are added to the reaction
mixture, which was then heated at reflux for approximately
18 additional hours (see Supporting Information). Although
we are not sure as to the exact role of these reagents at
present, it is possible that they help promote cleavage of the
PMB group of any remaining oxazinium salt 10 and/or help
in product isolation by destroying PMB-containing byprod-
ucts. This same yield improvement was also observed in
other cyclizations (vide infra).
Another cycloaddition was conducted with the related
enamide substrate 13 using the optimum reaction conditions
described in eq 1 (Scheme 3). Once again the cis-fused
dihydrooxazine 14 was produced here in good yield as the
only isolable product. In this example, the terminal olefin
functionality proved to be compatible with the reaction.
We next turned to the substrate 9b containing a (Z)-olefin
dienophile (Scheme 4). In this case, exposure of enamide
Scheme 5
seems probable that 20 arises by stepwise cyclization of
N-acyliminium ion 17 via intermediate carbocation 18 to
oxazinium salt 19, which then loses the PMB group.
The cyclization of (Z)-olefin analogue 9d was investigated
next (Scheme 6). This transformation yielded a 1.1:1 mixture
Scheme 6
Scheme 4
9b to the identical reaction conditions developed for the (E)-
olefins now afforded a 1:3 mixture of the cis-fused cycload-
duct 15 and the trans-fused system 16 in good overall yield.
The structure of the minor isomer 15 was confirmed by X-ray
crystallography.9
Another system that was investigated was enamide 9c,
bearing an (E)-olefinic dienophile, which we anticipated
would afford a 6,5-fused dihydro-1,3-oxazine Diels-Alder
product. Again to our surprise, exposure of substrate 9c to
the standard cyclization conditions led cleanly and in 93%
yield to the bridged dihydrooxazine 20 as a single stereo-
isomer (Scheme 5). If the boron trifluoride/anisole treatment
is omitted, the product is obtained in lower yield and
contaminated by an inseparable impurity. The structure of
product 20 was again secured by X-ray crystallography.9 It
of the trans- and cis-fused [4 + 2]-cycloadducts 21 and 22.
The stereochemistry of these Diels-Alder products was
determined by X-ray analysis of the crystalline derivatives
23 and 24.9
Since it seemed possible that these major differences in
the reactions of aldehyde- vs ketone-derived systems were
due to the method of generation of the N-acyliminium salts
and/or the presence of the N-PMB group, we conducted the
experiments shown in Scheme 7 to probe these issues. Thus,
(E)-olefin aldehyde 25 was converted to N-PMB enamide
26 in two steps. When treated under the same conditions as
for the ketone-derived cases, substrate 26 afforded a single
product 28, which could be isolated in good yield. Amide
alcohol 28 likely derives via an initial hetero Diels-Alder
process leading to oxazinium salt 27, which does not lose
the PMB group but rather is hydrolyzed on aqueous workup.
The structure and stereochemistry of 28 were confirmed by
X-ray analysis of the dealkylated derivative 29.9 Signifi-
(8) Cf.: Fulop, F.; Simon, L.; Simon-Talpas, G.; Bernath, G. Synth.
Commun. 1998, 28, 2303.
(9) We thank Dr. Louis Todaro (Hunter College) for the X-ray deter-
mination of carbamate 12 and Dr. Hemant Yennawar (Penn State Small
Molecule X-ray Crystallographic Facility) for the structures of compounds
15, 20, 23, 24, and 29.
Org. Lett., Vol. 5, No. 16, 2003
2917

tether length as well as the geometry of the olefin dienophile.
In general, hetero Diels-Alder reactions of electrophilic
N-acyliminium ions with olefins are considered to be “polar”
cycloadditions which proceed via a concerted but nonsyn-
chronous mechanism.4b,10 Such a mechanistic pathway prob-
ably holds for the aldehyde N-acyliminium ion cases, but
with the keto systems described here there may in fact be a
change of mechanism to one that is stepwise (cf. Scheme
5). At this stage, it is difficult to fully rationalize all of the
diverse results we have observed. We are, however, continu-
ing to explore this chemistry to better understand the
parameters controlling the reactions and also currently
utilizing cis-fused cycloadducts such as 11 and 14 in
synthesis of some marine alkaloids.
Scheme 7
Acknowledgment. We are grateful to the National
Institutes of Health (GM-32299) for financial support of this
research and for NSF Grant CHE-0131112 for the purchase
of an X-ray diffractometer.
Supporting Information Available: Experimental pro-
cedures for preparation of new compounds, including spectral
data, along with X-ray data for compounds 12, 15, 20, 23,
24, and 29 in CIF format. This material is available free of
charge via the Internet at http://pubs.acs.org.
cantly, the trans-stereospecificity of this variation of the
cycloaddition is in accord with that shown in Scheme 1,
which utilizes a bis-amide precursor 1 for generating the
N-acyliminium ion.6
OL035046M
The work described above has demonstrated that in
contrast to aldehyde-derived systems, the course of the
intramolecular hetero Diels-Alder cyclizations with ketone-
derived N-acyliminium ions is highly dependent upon the
(10) (a) Schmidt, R. R. Angew. Chem., Int. Ed. Engl. 1973, 12, 212. (b)
Schmidt, R. R. Angew. Chem., Int. Ed. Engl. 1969, 8, 602. Schmidt, R. R.
Angew. Chem., Int. Ed. Engl. 1970, 9, 31. (c) Schmidt, R. R. Chem. Ber.
1970, 103, 3242. Schmidt, R. R.; Hoffmann, A. R. Chem. Ber. 1974, 107,
78.
2918
Org. Lett., Vol. 5, No. 16, 2003