Highly selective Diels-Alder reactions of directly connected enyne dienophiles

Article abstract of DOI:10.1021/ja065762u

The paper describes the course of cycloadditions of Diels-Alder dienophiles containing linked enyne sites, each substituted with activating groups. Consistently, it was found that in the enyne cases the Diels-Alder reaction occurred specifically at the ac

Full text of DOI:10.1021/ja065762u

Published on Web 12/29/2006
Highly Selective Diels-Alder Reactions of Directly Connected
Enyne Dienophiles
Mingji Dai,† David Sarlah,† Maolin Yu,† Samuel J. Danishefsky,*,†,‡
Gavin O. Jones,§ and K. N. Houk*,§
Contribution from the Department of Chemistry, Columbia UniVersity, HaVemeyer Hall,
New York, New York 10027, Laboratory for Bioorganic Chemistry, The Sloan-Kettering Institute
for Cancer Research Center,1275 York AVenue, Room RRL 1301, New York, New York 10021,
and Department of Chemistry and Biochemistry, UniVersity of California,
Los Angeles, California 90095-1569
Received August 8, 2006; E-mail: s-danishefsky@ski.mskcc.org; houk@chem.ucla.edu
Abstract: The paper describes the course of cycloadditions of Diels-Alder dienophiles containing linked
enyne sites, each substituted with activating groups. Consistently, it was found that in the enyne cases the
Diels-Alder reaction occurred specifically at the acetylenic center. Furthermore, it was found that the
regiochemical sense of the cycloaddition was apparently determined by the remote activating group bound
to the olefinic site. This remote acrylyl group totally dominated the course of the cycloaddition, relative to
the activating group bound directly on the acetylene site. Explanations for these findings at the computational
level are provided. The computations also rationalize the strong preference for cycloaddition to occur at
the acetylene linkage and encompass the otherwise surprising regiochemical dominance by the remote
ester on the olefinic site. The high selectivities available through such reactions provide important new
opportunities in the synthesis of orsenillate type substructures that are found in a variety of natural products
of contemporary interest.
Introduction
in sharp contrast to the situation with parent acetylenedicar-
boxylates, uncatalyzed Diels-Alder reactions of monoactivated
Since its discovery in 1928, the Diels-Alder reaction has
emerged as one of the most powerful and versatile methods
available for the construction of six-membered rings.1 This
reaction typically involves a cycloaddition between a 1,3-diene
and an alkene-based dienophile. For uncatalyzed normal electron
demand Diels-Alder reactions to work well in the absence of
the benefits of intramolecularity, it is necessary for the dieno-
philic double bond to be activated by an electron-withdrawing
group capable of resonance with the olefinic site. Rather less
attention has been given to the development of synthetically
viable Diels-Alder variants in which a similarly activated
alkyne serves as the dienophilic coupling partner. Of course,
the parent alkynes bearing two activating groups (cf. acetylene-
dicarboxylate esters) are well-known to be reactive dienophiles.
In the research reported herein, primary attention is focused on
alkyne dienophiles that are activated by only a single ester.
acetylenic dienophiles bearing nonactivating substitution at the
other acetylenic carbon have been found to be particularly slow.2
The problem of a highly concise orsenillate synthesis, which
was our goal, was partly ameliorated by the discovery of a more
reactive acetylenic Diels-Alder variant, i.e., an “ynolide” as
the dienophile (Scheme 1, eq 1). The ynolide concept has been
successfully applied to the total syntheses of cycloproparadicicol
and aigialomycin D.3 Cycloadditions of allene-activated dieno-
philes with activated dienes were also examined as possible
solutions to the problem, but the utility of the type of
construction implied in eq 2 is undercut by the unexpectedly
poor regioselectivity observed in the cycloaddition step
(Scheme 1).4
With various long-term total synthesis designs in mind, we
wondered about the relative dienophilicities of acetylenic
(ynoate) and ethylenic (enoate) dienophiles activated by identical
groups. We first studied a directly connected enyne system
wherein the olefinic and acetylenic dienophile sites would carry
identical activating groups, Ao (olefin activating group) and Aa
(acetylene activating group), respectively (cf. 2). In Scheme 2,
our expectations at the outset of the study are spelled out. At
Earlier, our laboratory had attempted to exploit Diels-Alder
cycloaddition reactions between monoactivated acetylenic di-
enophiles (cf. â-alkylated propiolic esters) and a variety of
dienes, with the goal of producing benzoate esters. However,
† Columbia University.
‡ The Sloan-Kettering Institute for Cancer Research Center.
§ University of California.
(3) (a) Yang, Z.-Q.; Danishefsky, S. J. J. Am. Chem. Soc. 2003, 125, 9602.
(b) Geng, X.; Danishefsky, S. J. Org. Lett. 2004, 6, 413. (c) Yang, Z.-Q.;
Geng, X.; Solit, D.; Pratilas, C. A.; Rosen, N.; Danishefsky, S. J. J. Am.
Chem. Soc. 2004, 126, 7881. (d) Geng, X.; Yang, Z.-Q.; Danishefsky,
S. J. Synlett 2004, 1325.
(1) (a) Diels, O; Alder, K. Justus Liebigs Ann. Chem. 1928, 460, 98. (b) Alder,
K.; Rickert, H. F. Justus Liebigs Ann. Chem. 1936, 524, 180.
(2) (a) Danishefsky, S. J.; Singh, R. K.; Gammill, R. B. J. Org. Chem. 1978,
43, 379. (b) Danishefsky, S. J.; Etheredge, S. J. J. Org. Chem. 1979, 44,
4716.
(4) Yoshino, T.; Ng, F.; Danishefsky, S. J. J. Am. Chem. Soc. 2006, 128, 14185.
9
10.1021/ja065762u CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 645-657
645

A R T I C L E S
Dai et al.
Scheme 1. Possible Syntheses of Aromatic Rings from Diels-Alder Reactions
Scheme 2
that time, the question of inherent chemoselectivity of “ene”
vs “yne” potential dienophilic sites had not been well aired in
the literature,5 and we had no solid basis to prognosticate as to
this matter. Specifically, we hoped to investigate two diene
types, 1 and 6, with respect to this question. With either of these
probe dienes, it was expected that, if the cycloaddition step were
to occur at the acetylenic dienophile site, an aromatic product
would ensue. In the case of diene type 1, cycloaddition followed
by fragmentation of isobutylene would lead directly to the
orsenillate type product 5. By contrast, if cycloaddition was to
occur at the olefinic center, producing an intermediate of the
type 3, retro-Diels-Alder fragmentation of isobutylene could
well be more problematic, since the coordinated bond breaking
steps (synchronous or sequential) would not directly accomplish
aromatization.
unravel was less clear. The formation of 8 as a primary product
from 7 would certainly find ample precedent.2a However, various
competing prototypically related courses would be open to this
enolizable, vinylogous, â-dicarbonyl system. It will have been
noted that, in drafting the prospectus summary shown in Scheme
2, we were operating under the tacit assumption that the
immediate activating group on each of the potentially dienophilic
linkages would control the regiochemical course of reaction of
its adjoining site. Actually, there was no data base to support
this expectation.
The issue of the relative reactivity of olefinic and acetylenic
dienophiles bearing an identical activating group had been taken
up with separate substrates.7-12 In our case, the two unsaturated
(6) (a) Oppolzer, W. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Paquette, L. A., Eds.; Pergamon Press: Oxford, U.K., 1991;
Vol. 5, p 315. (b) Nicolaou, K. C.; Snyder, S. A.; Montagnon, T;
Vassilikogiannakis, G. Angew. Chem., Int. Ed. 2002, 41, 1668. (c) Takao,
K; Munakata, R; Tadano, K. Chem. ReV. 2005, 105, 4779.
(7) (a) Sauer, J.; Mielert, A.; Wiest, H. Chem. Ber. 1964, 3183. (b) Sauer, J.
Angew. Chem., Int. Ed. Engl. 1967, 6, 16.
Related issues apply to the use of acyclic diene type 6.
Reaction at the ynoate site would be expected to yield 9.
Following expected de-alcoholysis (i.e., elimination of ROH),
the orsenillate 5 would be produced. Alternatively, cycloaddition
at the enoate site would lead to 7. The way in which 7 would
(8) (a) Konovalov, A. I.; Samuilov, Ya. D.; Slepova, L. F.; Breus, V. A. Zh.
Org. Khim. 1973, 9, 2519. (b) Samuilov, Ya. D.; Nurullina, R. L.;
Konovalov, A. I. Zh. Org. Khim. 1981, 17, 1494. (c) Samuilov, Ya. D.;
Nurullina, R. L.; Konovalov, A. I. Zh. Org. Khim. 1982, 18, 2253. (d)
Samuilov, Ya. D.; Nurullina, R. L.; Konovalov, A. I. Zh. Org. Khim. 1983,
19, 1431.
(5) (a) Kocien´ski, P. J.; Street, S. D. A.; Yeates, C.; Campbell, S. F. J. Chem.
Soc., Perkin Trans. 1 1987, 2171. (b) Barluenga, J.; Aznar, F.; Barluenga,
S.; Ferna´ndez, M.; Mart´ın, A.; Garc´ıa-Granda, S.; Pin˜era-Nicola´s, A.
Chem.sEur. J. 1998, 4, 2280-2298. (c) Barluenga, J.; Ferna´ndez-
Rodr´ıguez, M. A.; Aguilar, E. Org. Lett. 2002, 4, 3659.
(9) Froese, R. D. J.; Coxon, J. M.; West, S. C.; Morokuma, K. J. Org. Chem.
1997, 62, 6991.
9
646 J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007

Diels-Alder Reactions of Connected Enyne Dienophiles
A R T I C L E S
Figure 2. Frontier molecular orbitals for the reaction of dimethoxybuta-
diene (DMB) with dimethyl-hex-2-en-4-ynedioate. Orbital coefficients are
shown above the LUMO orbital of the dienophile.
Figure 1. Frontier molecular orbitals for the reactions of dimethoxybuta-
diene (DMB) with methyl acrylate and methyl propiolate.
sites (“o” and “a”) would be directly linked. We here note that
Kocienski et al. have reported that the reaction of an enyne,
activated by a single ester substituent on the alkyne moiety,
with a Danishefsky-type diene afforded the product formed from
addition to the double bond in about 40% yield.5a Two other
minor products were also formed from addition to the triple
bond, in yields of only about 9% each. Nevertheless, in light
of the massive body of Diels-Alder driven research,6 the
question we were implicitly asking, i.e., that of the chemose-
lectivity of reaction of a conjugated enyne, wherein each linkage
is equipped with an identical terminal activating group, had not
otherwise been posed in the literature.
Our reading of Sauer’s elegant investigations of Diels-Alder
reactions of symmetrical dienes, such as cyclopentadiene and
9,10-dimethylanthracene, with alkynes and alkenes suggested
that, at least in these cases, alkenes are usually more reactive
than alkynes in Diels-Alder reactions.7 Konovalov and co-
workers demonstrated that with symmetrical cyclopentadienones
as the dienic components, alkynes are more reactive than alkenes
when the accepting characteristics of the diene increases.8 In
general, it seems that, with symmetrical dienes, disubstituted
alkenes are somewhat more reactive than similarly substituted
alkynes.9-11
These trends are well accommodated by frontier molecular
orbital (FMO) theory,12 which teaches that reactions having
small HOMOdiene-LUMOdienophile gaps manifest faster rates.
Thus, regiochemistry of the Diels-Alder reaction is controlled
by the better overlap of the diene HOMO and dienophile LUMO
orbitals, both of which have coefficients controlled by the
substituent locations.12d,e
The FMO interactions in the reactions of 1,3-dimethoxyb-
utadiene (DMB) with the dienophiles, methyl acrylate and
methyl propiolate, are shown pictorially in Figure 1. HF
(Hartree-Fock) energies for the frontier orbitals of the diene
Scheme 3. Diels-Alder Reaction of Cyclic Diene 13 with the
Enyne Dienophile, 11
and dienophile components are shown; p-orbitals are sized in
proportion to the coefficient of the atomic orbitals in the
molecular orbitals.
The LUMO energy of methyl acrylate (3.16 eV) is lower
than that of methyl propiolate (3.29 eV). As a result, the
HOMOdiene-LUMOdienophile gap for the reaction involving
methyl acrylate (11.16 eV) is smaller than that for the corre-
sponding reaction involving methyl propiolate (11.29 eV). The
regiochemical outcome of each reaction is controlled by the
overlap of the largest lobes, on C4 of the diene and on the
â-carbon of the dienophile.
Similarly, FMO theory can also, in principle, be used to
predict the selectivities of Diels-Alder reactions involving
enynes. Figure 2 shows FMO interactions in the cycloaddition
reaction of DMB with the diactivated enyne, dimethyl-hex-2-
en-4-ynedioate. The HOMOdiene-LUMOenyne gap is smallest,
and both the chemoselectivity and regiochemistry of the reaction
should be controlled by overlap of C4 of the diene with C3 of
the dienophile. Thus, strictly on the basis of FMO consider-
ations, in keeping with the analogy to monoactivated dienophiles
described above, it would be predicted that reaction will occur
at the ene of the enyne, with the expectation that the directly
connected ester substituent will control the regiochemical
outcome (cf. Scheme 2).
A number of deficiencies of FMO theory have been identified
since the development of the theory, mainly due to the lack of
quantitative significance of FMO interactions.13,14 In connection
with these early findings, empirical findings as well as higher
(10) Singleton, D. A.; Schulmeier, B. E.; Hang, C.; Thomas, A. A.; Leung,
S.-W.; Merrigan, S. R. Tetrahedron 2001, 57, 5149.
(11) (a) Fuks, R.; Viehe, H. G. In Chemistry of Acetylenes; Viehe, H. G., Ed.;
Marcel Dekker: New York, 1969; p 478. (b) Veliev, M. G.; Guseinov,
M. M.; Yanovskaya, L. A.; Burstein, K. Y. Tetrahedron Lett. 1985, 41,
749. (c) Bondarev, G. N.; Ryzhov, V. A.; Chelpanova, L. F.; Petrov, A. A.
Zh. Org. Khim. 1967, 3, 816.
(12) (a) Sustmann, R. Tetrahedron Lett. 1971, 29, 2721. (b) Sustmann, R.;
Schubert, R. Angew. Chem., Int. Ed. Engl. 1972, 11, 840. (c) Sustmann,
R. Pure Appl. Chem. 1974, 40, 569. (d) Houk, K. N. Acc. Chem. Res.
1975, 8, 361. (e) Houk, K. N. J. Am. Chem. Soc. 1973, 95, 4092.
(13) (a) Alston, P. V.; Ottenbrite, R. M.; Shillady, D. D. J. Org. Chem. 1973,
38, 4075. (b) Alston, P. V.; Ottenbrite, R. M. J. Org. Chem. 1975, 40,
1111. (c) Alston, P. V.; Ottenbrite, R. M.; Cohen, T. J. Org. Chem. 1978,
43, 1864. (d) Cohen, T.; Ruffner, R. J.; Shull, D. W.; Daniewski, W. M.;
Ottenbrite, R. M.; Alston, P. V. J. Org. Chem. 1978, 43, 4052. For a review,
see (e) Ginsburg, D. Tetrahedron 1983, 39, 2095.
(14) Kahn, S. D.; Pau, C. F.; Overman, L. E.; Hehre, W. J. J. Am. Chem. Soc.
1986, 108, 7381.
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 647

A R T I C L E S
Scheme 4
Dai et al.
level theoretical analyses have shown that FMO interactions
alone can fail to predict the relative reactivities of alkenes and
alkynes in Diels-Alder reactions.15 The necessity to go beyond
FMO theory to explain the relative reactivities of alkenes and
alkynes in nucleophilic additions was noted some time ago.16
Because of the nucleophilic character of the dienes to be
employed in our projected synthetic ventures, the alkyne might
be expected to be more reactive than the alkene. As shown in
an early theoretical treatment, nucleophilic (“one-bond”) dienes
may exhibit differing reactivity patterns relative to “two-bond”
dienes such as cyclopentadiene and 9,10-dimethylanthracene.17
In connection with these findings, Singleton and co-workers
have recently noted the shortcomings of FMO theory in and of
itself to predict the regiochemical sense of Diels-Alder
cycloaddition reactions.18
As will be seen via the empirical results described herein,
remarkably efficient and selective syntheses of conjugated arenes
were accomplished via Diels-Alder reactions involving enyne
dienophiles. In our case, exclusive attack on the yne is observed,
and stereoselectivity is controlled by the remote vinyl ester
substituent. Theoretical explorations prompted by these results
have provided explanations of these phenomena as well as new
insights into structural factors which impact upon the stability
of transition states of Diels-Alder reactions.
eVen in the presence of the attached ester, in contrast to the
expectation, shown in Scheme 2.
It was necessary to establish whether product 18 arises from
a strong kinetic preference, or whether it reflects a subsequent
transformation (via retro Diels-Alder reaction followed by a
Diels-Alder reaction in the observed sense). Put differently,
we wondered whether or not 18 actually reflects a thermody-
namic sink of the various Diels-Alder, retro Diels-Alder, and
aromatization progressions (Scheme 4). This matter was studied
with careful monitoring of the course of the events by proton
NMR spectroscopy.
As shown in Figure 3, cycloaddition happened slowly at room
temperature (15% conversion after 24 h). At a slightly higher
temperature (50 °C) the cycloaddition was faster (58% conver-
sion after 24 h). The fragmentation of isobutylene leading to
aromatic product occurred very slowly at approximately 80 °C
wherein trace quantities of 18a could be detected. Aromatization
was completed at approximately 120 °C in approximately 12
h, giving 18a as the only product. Its bis-desilylated derivative,
18, was obtained after preparative TLC purification. We
emphasize that during this whole monitoring exercise, there
appeared no other obserVable intermediates. We conclude that
under our conditions, 18 is arising under strictly kinetic control.
This conclusion is further confirmed by the isolation and
characterization 4aa from silyl enol ether 4a.
Results and Discussion
This result prompted further exploration. Accordingly, four
other dienes (14,21 15, 16,4 and 1722) were prepared and studied.
Each was subjected to thermolysis with enyne 11. Diels-Alder
adducts (18-21) were obtained in good to excellent yields
(Table 1). In each case, the cycloadduct was the product of
addition of the diene to the triple bond of the enyne. The
regioselectivity of each transformation was apparently solely
dictated by the activating group on the vinyl moiety. In fact,
only one product was also observed arising from reaction of
11 with cyclohexadiene (15), which lacks additional “nucleo-
philic” activation. We do note, however, that, in the case of
15, the yield of the cycloaddition reaction with 11 was notably
lower than that with the “nucleophilic” dienes.
Diels-Alder Cycloadditions of Enynes. 2-Hexen-4-ynedioic
acid dimethyl ester, 11, easily synthesized via the dimerization
of methyl propiolate,19 was coheated with dimedone-derived
diene 13 at 120 °C for 12 h.20 Of four possible products which
might have been anticipated from this reaction, only the styrenyl
ester, 18, was observed (98% yield) (Scheme 3). The structure
assigned to 18 was confirmed through a crystal structure of its
bismethylated derivative (i.e., the corresponding resorcinol
dimethyl ether). Clearly, the diene had reacted with the triple
bond of 11. However, surprisingly to us at the time, the remote
Vinyl ester had totally dictated the regioselectiVity of the reaction
(15) (a) Jones, G. O.; Ess, D. H.; Houk, K. N. HelV. Chim. Acta 2005, 88, 1702.
(b) Jones, G. O.; Robbins, J. A.; Theofanis, P. L.; Dai, M. J.; Houk, K. N.;
Danishefsky, S. J., manuscript in preparation.
(16) Strozier, R. W.; Caramella, P.; Houk, K. N. J. Am. Chem. Soc. 1979, 101,
1340.
(17) Houk, K. N.; Munchausen, L. L. J. Am. Chem. Soc. 1976, 98, 937.
(18) Ussing, B. R.; Hang, C.; Singleton, D. A. J. Am Chem. Soc. 2006, 128,
7594.
(19) (a) Wenkert, E.; Leicht, C. L.; Adams, K. A. H. Can. J. Chem. 1963, 41,
1844. (b) Ramachandran, P. V.; Rudd, M. T.; Reddy, M. V. R. Tetrahedron
Lett. 2005, 46, 2547.
(20) Rubottom, G. M.; Krueger, D. S. Tetrahedron Lett. 1977, 18, 611 and
references therein.
In an effort to gain a further grasp of the governing issues,
we next explored reactions between diene 13 and four other
enyne bis-dienophiles (Table 2). Two adducts were obtained
when the ester group on the alkene moiety was deleted and a
bromine was inserted on the dienophile (cf. 24). We also
(21) Kra¨geloh, K.; Simchen, G.; Schweiker, K. Liebigs Ann. Chem. 1985, 2352.
(22) Rubottom, G. M.; Gruber, J. M. J. Org. Chem. 1978, 43, 1599.
9
648 J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007

Diels-Alder Reactions of Connected Enyne Dienophiles
A R T I C L E S
Figure 3. Diels-Alder reaction of diene 13 and dienophile 11 at different temperatures. Dienophile 11 (1.0 equiv, 6.5 mg) and diene 13 (1.2 equiv, 13.2
mg) were dissolved in 1 mL of toluene (d8) in an NMR tube. The temperature was elevated from rt to 120 °C. Proton NMR was taken every 12 h.
examined the case of dienophile 25,23 which carries an olefinic
gem-dimethyl group. Even in these cases, the major product of
each cycloaddition reaction had been formed under the apparent
control of the vinyl moiety, though yields of the Diels-Alder
reaction were somewhat reduced. Remarkably, despite the
similarity of these cycloadditions to those first investigated by
Kocienski et al., the chemoselectivity of the reaction proceeded
only by addition to the triple bond of the monoactivated enynes.
These issues will be explored in the theoretical section.
Application of the chemistry to reactions with the enyne,
methyl pent-4-en-2-ynoate (see 22, R1 ) CO2Me, R2, R3 ) H)
was attempted. However, this dienophile is unstable, even below
the temperatures required for cycloaddition. By contrast, cy-
cloaddition reactions of the more stable 26 and 27 (available
via one-flask tandem manganese dioxide oxidation/Horner-
Wadsworth-Emmons olefination)24 with 13 at 160 °C did afford
adducts 32 and 33, respectively, in excellent yields and with
complete selectivity. It will be noted that the reaction of 13
with 27 points the way to an efficient synthesis of biaryl
compounds.
product stereochemistry when the olefin undergoes the cycload-
dition. We were also interested to learn which dienophile
substituent (ester or vinyl ester) would dictate the stereochemical
outcome of the Diels-Alder reaction in the bis-enophile case.
In the event, treatment of bis-dienophile 3525 with diene 40
led to a 1:1 mixture of endo/exo type cycloadducts (Scheme
6). Since the ene linkages of 35 are both E, each Diels-Alder
adduct per force contains one endo and one exo group. For
purposes of this discussion, we define the “endo product” as
the one in which the apparent regiocontrolling (i.e., “ortho”)
group emerges endo. If there is a dichotomy between the regio-
and stereo-determinants, we classify this as the “exo product.”
In both cases presented below, a single regioisomer was
observed, as shown. The use of the more reactive Rawal diene
(44)26 provided Diels-Alder adducts with the same sense of
regioselectivity and with an improved diastereomeric ratio
(endo(45)/exo(46) ) 4:1). The observed sense of regiochemistry
is consistent with the regioselectivities obtained in our previously
described enynyl-dienophile cases, where cycloaddition had
occurred at the ynoate site.
By contrast, reaction of Rawal’s diene (44)26 with enyne 11
at room temperature provided a single regioisomeric product
in moderate yield (Scheme 7). Reaction of diene 40 in refluxing
benzene yielded a similar result. Once again, the regiochemistry
was controlled by the vinyl ester moiety. Neither of these dienes
produced adducts resulting from reaction across the enoate
sector. These phenomena were reproduced by computations. In
The unexpected sense of regiocontrol observed in the Diels-
Alder reactions of the enyne-dienophiles described above led
us to speculate on the regiochemical outcome of Diels-Alder
reactions of dienophiles possessing â,â′-connected bis-dienyl
ester functionality (cf. 35) (Scheme 5). Furthermore, with this
type of dienophile, it would be possible to investigate issues of
(23) (a) Negishi, E.-I.; Van, Horn, D. E.; Yoshida, T. J. Am. Chem. Soc. 1985,
107, 6639. (b) Negishi, E.-I.; Qian, M.; Zeng, F.; Anastasia, L.; Babinski,
D. Org. Lett. 2003, 5, 1597.
(25) Liang, J, J; Foote, C. S. Tetrahedron Lett. 1982, 23, 3039.
(26) (a) Kozmin, S. A.; Rawal, V. H. J. Org. Chem. 1997, 62, 5252. (b) Kozmin,
S. A.; Green, M. T.; Rawal, V. H. J. Org. Chem. 1999, 64, 8045.
(24) Blackburn, L.; Pei, C.; Taylor, R. J. K. Synlett 2002, 215.
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 649

A R T I C L E S
Dai et al.
Table 1. Diels-Alder Reactions of Cyclic Dienes with the Enyne Dienophile, 11
a The mixture of diene (2.5 equiv) and dienophile (1.0 equiv) were heated at 120 °C for 12 to 36 h. b Isolated yields after silica gel chromatography.
this context, a new phenomenon controlling the chemoselectivity
and regioselectivity observed in these reactions was found (vide
infra).
from addition of 5,5-dimethyl-1,3-dimethoxycyclohexa-1,3-
diene, 49, the dimethoxy analogue of dienes 13 and 16, to the
triple bond (TS1, TS2) or the double bond (TS3, TS4) of 11
are shown in Table 3. The vinyl ester controlled yne Diels-
Alder adduct is formed from TS1 which has an activation
enthalpy of 14 kcal/mol (∆Gq ) 27 kcal/mol). The ester-
controlled yne Diels-Alder adduct is formed from TS2 and
has an activation enthalpy of 20 kcal/mol (∆Gq ) 33 kcal/mol).
Both intermediates, formed from reactions involving the enyne
triple bond, extrude isobutylene through a retro-Diels-Alder
reaction to form styrenyl products.
Theoretical Studies. In order to determine the chemo- and
regioselectivities in these reactions, the transition states for many
of the systems that had been explored experimentally were also
studied computationally. Lowest energy conformations of transi-
tion states are reported in each case. The transition states for
the reactions were obtained using density functional theory,
which has been found to give geometries and energies of
transition states in accord with experimental measures such as
kinetic isotope effects and reaction rates. In the case at hand,
the calculations give excellent agreement with experiment and
also led to the discovery of a phenomenon that causes acetylenes
activated by two electron-releasing groups to be more reactive
toward nucleophilic dienes, or nucleophiles in general, than
ethylenes activated by two electron-withdrawing groups.
Computational Methods. Geometry optimizations for all
reactants and transition structures were performed using the
B3LYP27,28 density functional theory (DFT) method as imple-
mented in the Gaussian 0329 suite of programs, with the 6-31G-
(d) basis set.
The reaction enthalpies of -33 and -31 kcal/mol (∆Gint
)
-18 and -17 kcal/mol) for formation of the vinyl ester
controlled and ester controlled intermediates, respectively,
indicate that these cycloadducts might be experimentally
observable under short reaction times. With reference to the
intermediates, the free energy of activation for isobutylene loss
from the vinyl ester controlled DA adduct is 27 kcal/mol; this
is approximately equal to the initial cycloaddition step. In
contrast, the free energy of activation for isobutylene loss from
the ester controlled DA adduct is 28 kcal/mol, which is about
5 kcal/mol less than the corresponding cycloaddition step.
Transition States of Enyne Cycloadditions. The activation
enthalpies for formation of the four possible adducts resulting
The activation enthalpies for formation of adducts resulting
from addition of the diene to the enyne double bond are both
equal to 23 kcal/mol; these are 9 kcal/mol greater than the
activation barrier for TS1, involving vinyl ester controlled
addition to the triple bond of the enyne. The reaction enthalpies
for these reactions are about -10 kcal/mol; these cycloadducts
are about 21 kcal/mol less stable than those produced from
addition to the triple bond. What is more, the reaction free
energies of these reactions are about +5 kcal/mol, which
(27) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang, W.;
Parr, R. G. Phys. ReV. B 1988, 37, 785.(c) Becke, A. D. J. Chem. Phys.
1993, 98, 1372.
(28) (a) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J. J. Phys.
Chem. 1994, 98, 11623. (b) Stephens, P. J.; Devlin, F. J.; Ashvar, C. S.;
Bak, K. L.; Taylor, P. R.; Frisch, M. J. ACS Symp. Ser. 1996, 629, 105. (c)
Hehre, W. J.; Radom, L.; Schleyer, P. V.; Pople, J. A. Ab Initio Molecular
Orbital Theory; Wiley: New York, 1986.
(29) Frisch, M. J., et al. Gaussian 03, revision C.02; Gaussian, Inc.: Wallingford,
CT, 2004.
9
650 J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007

Diels-Alder Reactions of Connected Enyne Dienophiles
A R T I C L E S
Table 2. Diels-Alder Reactions of Cyclic Dienes with Enyne Dienophiles
a The mixture of diene (2.5 equiv) and dienophile (1.0 equiv) were heated at the designated temperature for 12-36 h. b Isolated yield after silica gel
chromatography.
Scheme 5a
a Note: Endo and exo are defined here based on the relationship between the regio-directing group and the inside proton of the diene. Thus, a trans
relationship between these moieties would be considered endo and a cis relationship is considered exo.
indicates that the equilibrium of the reaction is shifted toward
reactants. Therefore, in comparison to the fast, irreversible
cycloadditions of dienes with the triple bonds of diactivated
enynes, cycloaddition involving the double bond is slow and
reversible.
The barriers for these cycloadditions are rationalized by the
transition structures for each of these cycloadditions, which show
the contributions of the ester and vinyl ester moieties. Four
transition structures corresponding to both diene CtCenyne (TS1,
TS2) and both diene CdCenyne (TS3, TS4) cycloadditions are
shown in Figure 4. Like the methyl acrylate and methyl
propiolate cycloadditions, these transition structures are all
highly asynchronous, with the difference between partial bonds
ranging from 0.8 to 1.3 Å.30
Overall, these results are in agreement with the experimental
results for Diels-Alder reactions of diester-substituted enynes
outlined above: kinetics and thermodynamics favor the addition
of dienes to the triple bonds of diester-substituted enynes over
addition to double bonds. In addition, they suggest that the
regioselectivity is controlled by the vinyl ester, not the ester
group, which is also in agreement with experiment.
The p-orbitals of the ester and vinyl ester substituents are
orthogonal in the transition structures for diene CtCenyne
cycloadditions. Because of this arrangement of substituents in
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 651

A R T I C L E S
Scheme 6
Dai et al.
Scheme 7
calculations on the isodesmic reactions of the vinyl anion with
methyl acrylate and methyl-2,4-pentadienoate to form ethylene
and the ester and vinyl ester stabilized vinyl anions (Figure 5).
The isodesmic reaction preserves the number and type of bonds
on each side of the equation. This model provides a measure
of stabilization of a fully formed vinyl anion when the group is
directly conjugated with the carbanion (direct stabilization), or
when there is what we call indirect stabilization. These equations
measure the stabilization of a vinyl anion by the substituents.
The relevance to the cycloadditions is that, in the transition states
in Figure 4, the highly nucleophilic diene is forming a bond to
one terminus of the acetylene, and there is considerable vinyl
anion character built up at the other terminus. The isodesmic
equations assess how the substituents would stabilize the vinyl
anion were it to be fully formed as an intermediate in these
cycloadditions.
Returning to the stabilizations assessed in Figure 5, the direct
stabilization of the vinyl anion is largest. An ester provides 32
kcal/mol stabilization, and an acrylate group gives an even larger
48 kcal/mol. The amazing fact is that indirect stabilization by
the ester or the acrylate is also very large, 21 and 27 kcal/mol.
In a resonance picture, indirect stabilization can be thought of
as the stabilization of the carbene resonance form of the vinyl
anion:
TS1 and TS2, the partial negative charge that develops in the
rather zwitterionic transition states (a charge transfer from diene
to dienophile is 0.3e in TS1) can be stabilized by both
substituents at once. However, in reactions involving the double
bond (TS3, TS4), the partial negative charge can be stabilized
mostly by the â substituent only. Highly asymmetric transition
structures such as these were first discovered by Domingo et
al.31 and later investigated by Morokuma et al.9 and Singleton
and co-workers10 for related cycloadditions of diactivated
alkynes.
Regioselectivity in Enyne Cycloadditions. In highly asyn-
chronous TS1, the acrylate group is in a position to stabilize
the negative charge directly, while the immediately bound ester
can do so only indirectly. It is clearly advantageous for the
acrylate to be able to stabilize the developing anion.
B3LYP predicts that the vinyl anion is stabilized by -32 kcal/
mol by direct interaction with the ester group and by -21 kcal/
mol from indirect interaction with the ester group. In contrast,
the acrylate directly stabilizes the negative charge by -48 kcal/
mol and indirectly by -27 kcal/mol; these acrylate stabilization
energies are 16 and 6 kcal/mol, respectively, greater than the
stabilization energies obtained with the ester substituent.
The preference for the formation of the vinyl ester controlled
adduct in the reactions of 11 with dienes is primarily the result
of an electronic effect, in which the negative charge that
develops at the transition state is best stabilized by the
combination of direct conjugation to the vinyl ester substituent
and indirect stabilization by the ester group. This indirect
stabilization is a new aspect of cycloaddition reactivity that will
have important effects on the relative reactivities of alkenes and
alkynes.
To model the effect of the ester and vinyl ester substituents
on the regioselectivities of these reactions, we have performed
(30) IRC calculations performed on the transition state for the formation of the
vinyl ester-controlled adduct uncovered the zwitterionic intermediate
(S2 ) 0) shown below, corresponding to the adduct formed from
nucleophilic addition of the diene. The energy of this intermediate is 12.5
kcal/mol, which is 1.5 kcal/mol lower than the transition state. However,
the ring-closing transition structure proceeding from this intermediate could
not be found, most likely because of the near-zero activation barrier for
conversion of this zwitterions to product.
(31) (a) Domingo, L. R.; Jones, R. A.; Picher, M. T.; Sepu´lveda-Arques, J.
Tetrahedron 1995, 51, 8739. (b) Domingo, L. R.; Arno´, M.; Andre´s, J.
Tetrahedron Lett. 1996, 37, 7573. (c) Domingo, L. R.; Picher, M. T.; Aurell,
M. J. J. Phys. Chem. A 1999, 103, 11425.
Reactions of Monoactivated Enynes. Two examples of
cycloadditions involving monoester-substituted enynes are
9
652 J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007

Diels-Alder Reactions of Connected Enyne Dienophiles
A R T I C L E S
Figure 4. B3LYP transition structures for the four modes of cycloaddition of 49 with enyne 11. Forming bond distances are in angstroms.
Table 3. B3LYP Relative Enthalpies of Transition States (above Arrows) and Products of Diels-Alder Reactions of 49 with Enyne 11 (Free
Energies of Transition States and Products Are Shown in Parentheses)
represented by eqs 1 and 2 in Table 2. As shown, addition to
the double bond does not occur. Notably, however, both enynes
react to give vinyl-controlled adducts (cf. 28 and 30) in an ∼3:1
preference over ester-controlled adducts (cf. 29 and 31).
Geometries and energies for model reactions have been
computed in an attempt to understand the observed chemo- and
regioselectivities. Relative energies for the reaction of one of
enyne, 50, a model for enyne 25 used in experimental studies,
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 653

A R T I C L E S
Dai et al.
Figure 5. B3LYP stabilization energies (kcal/mol) by ester and acrylate substituents from isodesmic reactions of methyl acrylate and methyl-2,4-pentadienoate
with the vinyl anion. Direct and indirect stabilization is assessed.
Table 4. B3LYP Relative Enthalpies of Transition States (above Arrows) and Products of Diels-Alder Reactions of 49 with Enyne 50 (Free
Energies of Transition States and Products Are Shown in Parentheses)
with 49 are shown in Table 4; transition state geometries and
relative energies are also shown in Figure 6. Enynes 24 and 25
are similar enough that conclusions derived from the study of
one of these reactants should be readily applicable to the other.
The activation enthalpies for vinyl-controlled and ester-
controlled cycloadditions to the triple bond of the enyne are 24
and 25 kcal/mol, respectively. These are about 12 kcal/mol
smaller than the activation enthalpy for addition to the double
bond of the enyne (36 kcal/mol). These results are similar to
the predicted preference for vinyl-controlled addition to the triple
bond of diactivated enynes; more significantly, the final products
observed in experiment result from cycloadditions of the diene
with the triple bond of monoactivated enynes.
The reaction enthalpies for formation of the yne, vinyl-
controlled and yne, ester-controlled adducts are -26 and -24
kcal/mol, respectively; these are about 30 kcal/mol more stable
than the reaction enthalpy for formation of the ene-controlled
cycloadduct (+4 kcal/mol). Moreover, cycloaddition of the diene
9
654 J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007

Diels-Alder Reactions of Connected Enyne Dienophiles
A R T I C L E S
Figure 6. B3LYP transition state geometries and activation enthalpies, in kcal/mol, for the reaction of methyl pent-4-en-2-ynoate, 50, with 49.
Figure 7. B3LYP reaction enthalpies, in kcal/mol, for the isodesmic reactions of methyl acrylate and trans-1,3-butadiene with the vinyl anion forming the
directly, and indirectly, stabilized vinyl anions and ethylene.
Table 5. B3LYP Activation Enthalpies for Diels-Alder Reactions
49 with Substituted Propiolates to Give Ester-Controlled (I) and
Non-Ester-Controlled (II) Diels-Alder Adducts
unstable norbornene adducts, kinetics control the course of the
reaction. Therefore, initial fast cycloaddition of the diene with
the triple bond of the enyne results in the formation of stable
norbornadiene cycloadducts, which subsequently extrude isobu-
tylene to form the styrenyl final products.
The product differentiation arises in the initial rate-determin-
ing cycloaddition step. The predicted 1 kcal/mol preference for
formation of the vinyl-controlled over the ester-controlled
regioisomer in this model system is consistent with the 5:2
regioisomeric ratio of styrenyl products obtained in experimental
studies. To account for this preference we have evaluated the
ability of the vinyl and ester groups to stabilize the negative
charges that develop in these transition states.
∆
Hq
R
I
II
H
Me
16.0
22.7
24.5
24.8
27.3
23.9
CHdCMe2
Figure 7 depicts the energies of anion stabilization by butylene
and ester groups. Thus, direct interaction of the butylene group
with the negative charge is preferable to indirect interaction.
The ester group stabilizes the anion more than the vinyl group
does, both in a direct (15 kcal/mol) and indirect (12 kcal/mol)
sense, through interaction with the negative charge; the opposite
was true for the vinyl ester substituent.
If these substituent effects were additive, then the ester-
controlled adducts would be expected to be formed preferentially
in the cases of dienophiles 26 and 27. However, both experi-
mental results and transition state calculations indicate that the
vinyl controlled adduct is formed in preference to the ester-
controlled adduct.
with the double bond of the monoactivated enyne is very
endergonic (+20 kcal/mol), more so than for reactions involving
addition to the double bond of the diactivated enyne. Presum-
ably, the large endergonicity of the reactions involving these
monoactivated enynes is the result of significant reactant
stabilization by the gem-dimethyl substituents on the vinyl
moiety. With reference to the norbornadiene intermediates,
extrusion of isobutylene requires about 30 and 28 kcal/mol (+4.4
and +4.2 kcal/mol with reference to the enyne and diene
reactants) for formation of the vinyl-controlled and ester-
controlled adducts, respectively. The formation of the styrenyl
final products is exothermic by about -43 kcal/mol.
These results suggest that while the thermodynamics greatly
favor formation of the stable norbornadiene cycloadducts over
The influence of steric repulsion in the transition states for
the reaction of 49 with 50 was evaluated by comparison with
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 655

A R T I C L E S
Scheme 8
Dai et al.
analogous reactions involving methyl propiolate and methyl
butynoate. The expectation is that replacement of the terminal
acetylenic hydrogen atom of methyl propiolate with a methyl
group or a butylene group will increase the activation enthalpy
for formation of the ester-controlled adducts, presumably due
to steric repulsion in the transition state. This is indeed the case.
B3LYP activation enthalpies for the formation of ester-
controlled (I) and non-ester-controlled (II) adducts of these
reactions are shown in Table 5. The activation enthalpy for
formation of the ester-controlled reaction of 49 with methyl
propiolate (R ) H) is 16 kcal/mol. In comparison, the activation
enthalpies for the reactions involving methyl butynoate (R )
Me) and 50 (R ) butylene) are 23 and 24 kcal/mol, respectively.
The approximately 7 kcal/mol increase in the activation enthal-
pies for these reactions in comparison with the reaction involving
methyl propiolate is evidently due to steric repulsion between
the olefinic proton of the diene and the methyl and butylene
groups of both dienophiles.
Indeed, calculations support the experimental results reported
by Kocienski. The formation of the ene-propargyl ester cy-
cloadduct is favored over formation of the yne, vinyl controlled
cycloadduct, but by only 0.2 kcal/mol (∆Hq ) 21.5 and 21.7
kcal/mol, respectively). This suggests that both adducts should
be formed in almost equal amounts, whereas the experiment
gives an ∼1.3 kcal/mol lower activation enthalpy for formation
of the ene, propargyl ester controlled cycloadduct. The transition
state corresponding to formation of the yne, ester controlled
cycloadduct could not be located, presumably because of the
steric interactions between the methyl group on the cis-dimethyl
vinyl substituents and the incoming diene.
The reaction enthalpies for formation of the intermediates
resulting from addition to the triple bond are about -40 kcal/
mol; these are more exothermic reactions than reactions with
the double bond (-18 kcal/mol). Significantly, ene-controlled
reactions are exergonic by -1 kcal/mol, in contrast to reactions
involving 24 and 25, which are endergonic by as much as +20
kcal/mol. The exothermicity of this variant of the reaction
reported by Kocienski et al. is due to the absence of significant
steric strain in the monocyclic product. Although calculations
do not fully account for the large kinetic preference for
formation of ene-controlled cycloadducts, they bolster the notion
that these cycloadducts, once formed, are thermodynamically
stable.
The activation enthalpy for the formation of the non-ester-
controlled adducts of methyl propiolate and methyl butynoate
are 25 and 27 kcal/mol, respectively. By contrast, the activation
enthalpy for the formation of the vinyl-controlled adduct of 50
is only 24 kcal/mol. These differences are consistent with the
inability of the terminal hydrogen atom and methyl group of
methyl propiolate and methyl butynoate, respectively, to stabilize
the negative charge that develops at the transition state of these
reactions. As discussed in relation to Figure 7, the butylene
group is, in fact, able to stabilize the developing negative charge
appreciably.
Conclusions
In summary, we have described highly efficient chemo- and
regioselective Diels-Alder reactions for the syntheses of
polysubstituted aromatic compounds and biaryl compounds from
unsymmetrical dienes and enynes in good to excellent yield. A
dienophile based on a â,â′-connected dienyl ester was also
explored and gave adducts with the same sense of regioselec-
tivity. Theoretical considerations of these reactions account for
experimentally observed chemoselectivities and regioselectivi-
ties. The importance of indirect stabilization was discovered;
this causes a deactivated alkyne to be substantially more reactive
than a deactivated alkene in cycloadditions with strongly
asymmetric nucleophilic dienes. The consequences of this effect
in other cycloadditions are being actively explored.
These results signify that, while the transition states for
formation of ester-controlled adducts from reactions involving
monoactivated enynes are stabilized by interaction of the
developing negative charge with the ester group, steric repulsion
from the butylene group overwhelms this interaction. These
combined effects result in the formation of vinyl-controlled
adducts from transition states which are stabilized by direct
interaction of the vinyl group and indirect interaction of the
ester group with the developing vinyl anion.
Finally, to complement these findings we have also investi-
gated the cycloaddition reactions first reported by Kocienski,
shown in Scheme 8, involving addition of an activated diene
with a monoactivated enyne, in which the ene, propargyl ester
controlled adduct, 53, is produced in preference to either of the
isomers resulting from addition to the triple bond (54 and 55).
Substrates 56 and 57 were used to model the substrates used in
experimental studies.
Acknowledgment. We are grateful to the National Science
Foundation (CHE-0548209 to K.N.H.) and the National Insti-
tutes of Health (GM36700 to K.N.H. and HL25848 to S.J.D.)
for financial support of this research. This research was
facilitated through the Partnerships for Advanced Computational
Infrastructure (PACI) through the support of the National
9
656 J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007

Diels-Alder Reactions of Connected Enyne Dienophiles
A R T I C L E S
Science Foundation. The computations were performed on the
National Science Foundation Terascale Computing System at
the Pittsburgh Supercomputing Center (PSC) and on the UCLA
Academic Technology Services (ATS) Hoffman Beowulf
cluster. We thank Ms. Daniela Buccella for crystal structure
analysis and Ms. Rebecca Wilson for valuable help in editing
the manuscript.
Supporting Information Available: Experimental procedures,
data on new compounds, and B3LYP coordinates and energies
of reactants, transition structures, intermediates, and products.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA065762U
9
J. AM. CHEM. SOC. VOL. 129, NO. 3, 2007 657