First experimental and theoretical evidence of a deactivating enone dienophile in the transannular Diels-Alder reaction

Article abstract of DOI:10.1021/jo0265129

A thorough study of the transannular Diels-Alder (TADA) reaction of trans-trans-trans macrocyclic trienes was carried out. It led to a better understanding of various parameters that govern the TADA reaction in particular and the Diels-Alder reaction in g

Full text of DOI:10.1021/jo0265129

F ir st Exp er im en ta l a n d Th eor etica l Evid en ce of a Dea ctiva tin g
En on e Dien op h ile in th e Tr a n sa n n u la r Diels-Ald er Rea ction
EÄ lyse Bourque,† Pierre Deslongchamps,*,† and Yves L. Dory*,‡
De´partement de chimie, Institut de Pharmacologie, Universite´ de Sherbooke, 3001, 12e avenue nord,
Sherbrooke, Que´bec J 1H 5N4, Canada
pierre.deslongchamps@usherbrooke.ca; yves.dory@usherbrooke.ca
Received October 2, 2002
A thorough study of the transannular Diels-Alder (TADA) reaction of trans-trans-trans macrocyclic
trienes was carried out. It led to a better understanding of various parameters that govern the
TADA reaction in particular and the Diels-Alder reaction in general. Thus, carbonyl activation of
the dienophile is thoroughly discussed in light of new experimental and theoretical data. An enone
dienophile is found to deactivate the reaction, although it remains planar at the transition state.
This unusual result was discussed in terms of tether substituents that provoke destabilization of
the transition state.
CHART 1. Selected Ta r gets for TADA Str a tegy
I. In tr od u ction
It is widely accepted that enones are activated dieno-
philes.1 Accordingly, they undergo fast Diels-Alder [4+2]
cycloaddition reactions, when compared to the corre-
sponding simple alkene dienophiles.2 There are no excep-
tions to this rule in intermolecular and very few in usual
intramolecular environments.3 However, the transannu-
lar Diels-Alder (TADA) strategy offers a much better
way to control the approach of the diene and dienophile.4,5
Indeed, the TADA reaction corresponds to a conforma-
tionally very restricted intramolecular Diels-Alder (IMDA)
reaction, since the two reacting partners are held to-
gether by means of two tethers instead of only one. With
this powerful probe as a tool, we have gathered experi-
mental and theoretical data that unequivocally prove
that enones can sometimes turn into deactivating dieno-
philes.
transition state appears to be asymmetric with the
following characteristics: (a) the short bond is located
at the end of the acrolein molecule (2.09 Å for 3-21G
calculations), (b) the long bond is next to the carbonyl
(2.35 Å), and (c) the acrolein skeleton remains flat to
ensure proper CdC/CdO conjugation and so to lower the
energy of the dienophile LUMO. These geometry require-
ments can be fought in two ways: (a) by preventing the
enone system from remaining flat or (b) by adding steric
congestion or strain at the short bond site. Since we are
currently interested in synthesizing molecules such as
Taibaihenryiin C,7 Odonicin,8 members of the Kaurane
family, like Leucamenin F,9 and other di- and triterpenes
in which the A ring is sterically congested (Chart 1), we
thought that we should take that opportunity to study
the aforementioned subject. Those compounds could
originate from TAC (trans-anti-cis) tricycles obtained by
TADA reaction of TTT (trans-trans-trans) 14-membered
macrocyclic trienes (Chart 2). Since another tricycle of
CAT (cis-anti-trans) geometry could be obtained as well
from the same triene precursor, there are therefore two
questions that we wanted to answer in model series.
First, is the desired TAC adduct going to be the major
product of the TADA reaction? If this is not so, then what
can be done in terms of substituents to get the right bias?
The geometry features of the transition state for the
acrolein-butadiene Diels-Alder have been calculated at
various RHF levels of ab initio theory.6 In all cases, the
† Laboratoire de synthe`se organique.
‡ Laboratoire de synthe`se supramole´culaire.
(1) Kistiakowsky, G. B.; Lacher, J . R. J . Am. Chem. Soc. 1936, 58,
123.
(2) Houk, K. N.; Li, Y.; Evanseck, J . D. Angew. Chem., Int. Ed. Engl.
1992, 31, 682 and references therein.
(3) (a) Craig, D. Chem. Soc. Rev. 1987, 16, 187. (b) J ung, M. E.;
Halweg, K. M. Tetrahedron Lett. 1981, 22, 3929. (c) Smith, D. A.;
Sakan, K.; Houk, K. N. Tetrahedron Lett. 1986, 27, 4877.
(4) (a) Lamothe, S.; Ndibwami, A.; Deslongchamps, P. Tetrahedron
Lett. 1988, 29, 1639. (b) Marsault, E.; Toro´, A.; Nowak, P.; Deslong-
champs, P. Tetrahedron 2001, 57, 4243.
(5) (a) Takahashi, T.; Katsuya, S.; Doi, T.; Tsuji, J . J . Am. Chem.
Soc. 1988, 110, 2674. (b) Wood, J . L.; Porco, J . A.; Taunton, J .; Lee, A.
Y.; Vlardy, J .; Schreiber, S. L. J . Am. Chem. Soc. 1992, 114, 5898. (c)
Marshall, J . A.; Wang, X.-J . J . Org. Chem. 1992, 57, 3387. (d) J ung,
S. H.; Lee, Y. S.; Park, H.; Kwon, D.-S. Tetrahedron Lett. 1995, 36,
1051. (e) Roush, W. R.; Works, A. B. Tetrahedron Lett. 1996, 37, 8065.
(f) J ones, P.; Li, W.-S.; Pattenden, G.; Thomson, N. M. Tetrahedron
Lett. 1997, 38, 9069.
(7) Li, B.-L.; Pan, Y.-J .; Yu, K.-B. Tetrahedron Lett. 2002, 43, 3845.
(8) Fujita, E.; Node, M. Prog. Chem. Org. Nat. Prod. 1984, 46, 77.
(9) (a) Yoshio, T.; Tetsuro, F.; Akira, U. Chem. Lett. 1981, 1229. (b)
Manes, L. V.; Crews, P.; Kernan, M. R.; Faulkner, D. J . J . Org. Chem.
1988, 53, 570.
(6) (a) Loncharich, R. J .; Brown, F. K.; Houk, K. N. J . Org. Chem.
1989, 54, 1129. (b) Dory, Y. L.; Hall, D. G.; Deslongchamps, P.
Tetrahedron 1998, 54, 12279.
10.1021/jo0265129 CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/21/2003
2390
J . Org. Chem. 2003, 68, 2390-2397

First TADA Deactivating Enone Dienophiles
CHART 2. Ma cr ocyclic Tr ien es a n d Th eir
Cor r esp on d in g TADA Tr a n sition Sta tes
SCHEME 1 a
CHART 3. “Activa ted ” a n d “In a ctiva ted ” Tr ien e
Mod els for th e Syn th esis of Ta iba ih en r yiin C
Second, is the gem-dimethyl group going to add too much
steric hindrance, so that much higher reaction temper-
atures will be required? This latter aspect is very
interesting, because this destabilizing steric effect could
be enhanced by adding a carbonyl (at position 11) next
to the dienophile, so that a normally very activated
dienophile could turn out to be deactivating. Indeed, the
methyl group of the dienophile and the axial methyl
group of the gem-dimethyl could collide very strongly at
the transition state, when the dienophile is conjugated
with a carbonyl group, because the corresponding transi-
tion state distance shortens from 2.21 Å (calculated 3-21G
distance for the ethylene/butadiene Diels-Alder transi-
tion state) to 2.09 Å.6 In the case of the conjugated
dienophile, the TADA reaction would then demand more
energy despite the “activation”.
Accordingly, we designed macrocycles 1, 2, and 3 to
study this phenomenon as well as to pursue the synthesis
of Taibaihenryiin C in particular, since an interesting key
TAC tricyclic intermediate 6 could be easily obtained
(Chart 3). The TADA experiments on the three macro-
cycles 1, 2, and 3 showed that the activation energy is
larger for the trienone 1 than it is for the triene 2. Ab
initio theoretical calculations were therefore carried out
to untangle all the different factors that could be respon-
sible for such an a priori unexpected result. We wish to
show these new results that complete our work on TT
dienes within 14-membered macrocycles in TADA reac-
tions (the first part was dealing with inactivated and
activated cis dienophiles)10 and that shed new light on
the mechanism of the Diels-Alder reaction.
a
Reagents and conditions: (a) KHMDS, THF, -78 °C, 84% (12:
13, 52:48). (b) TPAP, NMMO, MS 4 Å, CH2Cl2, rt, 61%. (c) NaBH4,
MnCl2, MeOH, 0 °C, 83% (12:13, 25:75). (d) MOMCl, DIEA,
CH2Cl2/hex (9:1), rt, 80%. (e) LiBH4, EtOH, Et2O, rt, 89%. (f) PivCl,
Py, DMAP, CH2Cl2, rt, 92%. (g) PPTS, MeOH, rt, 92%. (h) TPAP,
NMMO, MS 4 Å, CH2Cl2, rt, 86% (i) (i) 20, tBuLi, Et2O, -78°C;
(ii) 19, 82%. (j) AcCl, Py, DMAP, CH2Cl2, rt, 99%. (k) TBAF, THF,
rt, 87%. (l) TPAP, NMMO, MS 4 Å, CH2Cl2, rt, 89%.
II. Exp er im en ta l a n d Th eor etica l Deta ils
A. Syn th esis. Synthesis of these macrocycles follows
the highly convergent general strategy developed in our
laboratory.11 The diene fragment 1112 was coupled with
the known ester 1013 via an aldol condensation to afford
a 48:52 mixture of the desired erythro adduct 13 and the
threo adduct 12 with a total yield of 84%. These com-
pounds could be easily separated by flash chromatogra-
phy (Scheme 1). The threo isomer 12 was oxidized to the
â-ketoester 14 with TPAP and N-methyl morpholine
oxide (NMMO)14 (61% yield) and recycled to the erythro
isomer 13 by reduction, using sodium borohydride-
manganese chloride complex (yield 83%, 75:25 mixture
(11) Deslongchamps, P. Pure Appl. Chem. 1992, 64, 1831.
(12) Phoenix, S.; Bourque, E.; Deslongchamps, P. Org. Lett. 2000,
2, 4149.
(13) Solladie´, G.; Almario, A. Tetrahedron: Asymmetry 1995, 6, 559.
(14) (a) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J .
Chem. Soc., Chem. Commun. 1987, 1625. (b) Griffith, W. P.; Ley, S.
V. Aldrichim. Acta 1990, 23, 13.
(10) Fortin, S.; Barriault, L.; Dory, Y. L.; Deslongchamps, P. J . Am.
Chem. Soc. 2001, 123, 8216.
J . Org. Chem, Vol. 68, No. 6, 2003 2391

Bourque et al.
SCHEME 2 a
SCHEME 3
a
Reagents and conditions: (a) MeOAc, LDA, THF, -78 °C, 87%.
tion precursor 28 following Magid’s procedure.18 28 was
then submitted to macrocyclization high-dilution condi-
tions using cesium carbonate in acetonitrile to yield an
inseparable mixture of the macrocyclic acetates 2 and 3
with a total yield of 60% when the reaction medium was
kept rigorously anhydrous; otherwise hydrolysis of the
acetate group occurred and further led to degradation
products.
The acetate group was removed by a treatment with
potassium carbonate to afford the alcohols 29 and 30 as
a 1:1 mixture (76% yield). Alcohols 30 and 29 were
separated and each transformed back into their corre-
sponding acetate 2 and 3 esters with the respective yields
of 65% and 71%. In addition, a mixture of both alcohols
was oxidized with the Dess-Martin reagent to afford the
macrocyclic trienone 1 with a yield of 84%. 1 was
crystallized and X-ray crystallography proved its struc-
ture.19
The thermal TADA reaction of macrocycle 1 was
carried out in toluene in a sealed Pyrex tube (Scheme
3). The best result was obtained when 1 was heated at
105 °C for 24 h and gave two tricyclic TAC adducts 4
and 5 in almost equal portions (48:52). Nothing happened
at 80 °C, but some decarboxymethylated product related
to tricycle 4 was observed at higher temperature. These
experiments suggest that the minimal activation tem-
perature stands between 100 and 105 °C. The structure
of 4 was established by X-ray crystallography;20 and that
of the other TADA adduct 5 was determined by NMR
correlation experiments and spectral comparisons with
(b) Dess-Martin periodinane, CH2Cl2, rt, 86%. (c) PPTS, MeOH,
rt, 88%. (d) Hexachloroacetone, Ph3P, THF, -30 °C, 99%. (e)
Cs2CO3, CsI, MeCN, 28 (2 × 10-3 M), 80 °C, 18 h, 60%. (f) K2CO3,
MeOH, 76%. (g) Dess-Martin periodinane, CH2Cl2, rt, 84%. (h)
AcCl, Py, CH2Cl2, rt, 65% and 71%.
of 13:12).15 The methoxymethyl ether 15 was obtained
by protection of the alcohol 13 with MOM chloride and
diisopropyl ethylamine (yield of 80%). The methyl ester
15 was reduced with lithium borohydride and with a yield
of 89%. The resulting alcohol 16 was protected as its
pivaloate ester 17 with a yield of 92%. The tert-butyldim-
ethylsilyl group was selectively removed by mild acidic
conditions with PPTS in methanol to yield the primary
alcohol 18 (yield 92%), which was subsequently oxidized
with TPAP to the aldehyde 19 with a yield of 86%. The
dienophile part was then added by nucleophilic attack
of the aldehyde 19 with the organolithium intermediate
issued from the known iodide 20.16 The alcohol 21 was
obtained in 82% yield and was protected as the acetate
ester 22 with the excellent yield of 99%. The tert-
butyldiphenylsilyl group was removed by a treatment
with tetrabutylammonium fluoride to afford the alcohol
23 with a yield of 87%. The alcohol group was then
oxidized to the aldehyde 24 (yield 89%) by means of
TPAP.
The â-ketoester moiety was further elaborated by aldol
condensation of methyl acetate with the aldehyde 24 to
give the alcohol 25 (yield 87%). That alcohol was im-
mediately oxidized with Dess-Martin periodinane17 to
give the corresponding â-ketoester 26 with a yield of 86%
(Scheme 2). The alcohol 27 was set free from its tetrahy-
dropyranyl protecting group with PPTS and with a yield
of 88%. Finally, the alcohol 25 was almost quantitatively
(99%) transformed into the allylic chloride macrocycliza-
(18) Magid, R. M.; Fruchey, O. S.; J ohnson, W. L.; Allen, T. G. J .
Org. Chem. 1979, 44, 359.
(19) Cambridge Data Bank reference for X-ray of macrocycle 1:
CCDC 190392. Formula: C27H40O8. Unit cell parameters: a ) 5.993-
(5) Å, b ) 19.071(5) Å, c ) 12.063(5) Å, â ) 96.130(5)°. Space group
P21.
(15) Fujii, H.; Oshima, K.; Utimoto, K. Tetrahedron Lett. 1991, 43,
6147.
(16) Liu, F.; Negishi, E. I. J . Org. Chem. 1997, 62, 8591.
(17) Dess, D. B.; Martin, J . C. J . Org. Chem. 1983, 48, 4155.
(20) Cambridge Data Bank reference for X-ray of tricycle 5: CCDC
190393. Formula:
C27H40O8. Unit cell parameters: a ) 15.8491(11)
Å, b ) 9.845(17) Å, c ) 17.9449(12) Å, â ) 106.430(6)°. Space group
P21/c.
2392 J . Org. Chem., Vol. 68, No. 6, 2003

First TADA Deactivating Enone Dienophiles
SCHEME 4
B. Th eor et ica l P r oced u r es. Although most triene
systems undergoing TADA reactions could be adequately
handled by low-level calculations (semiempirical AM1
and PM3 methods),23 it quickly appeared that the TTT
case could not be treated as easily. In fact, reproducing
the TAC:CAT adduct ratio experimentally observed
proved impossible without the addition of empirical
corrections.24 We looked for a better way to model the
different effects arising in the competing transition states
(TSs) leading to the TAC and CAT TADA adducts, to put
sufficiently accurate figures on all the different interac-
tions involved. We calculated the different transition
states (TSs) corresponding to the TADA reactions at the
RHF/3-21G theory level25 by means of GAMESS.26 The
zero-point energy corrections were not applied due to the
large size of the systems, which always prevented such
calculations. However, these corrections are known not
to affect the relative energies between TSs in a significant
way.27 To estimate the activation energy corresponding
to each case, it was necessary to locate the lowest energy
macrocyclic conformers. We first performed a systematic
conformational search of each a the macrocycles using
the maximin2 force field within SYBYL.28 All conformers
having a relative energy below 4 kcal/mol were then
individually calculated at the RHF/3-21G ab initio level.
No conformers having an s-cis diene were found. Only
the lowest energy conformers were selected to calculate
the activation energy, because all other conformers were
less stable by at least 2 kcal/mol (apart from two
macrocycles where competing conformers were found
with relative energies of 0.44 and 1.11 kcal/mol in one
case and 1.28 kcal/mol in the other case). Although the
RHF/3-21G method is not sophisticated by current stan-
dards, we have already proven that it was very efficient
for our transannular systems.10 Moreover, numerous
calculations have to be carried out on rather large
systems, so that the RHF/3-21G theory level appears to
be the ideal compromise. We first validated its ability to
deal with our TTT systems by modeling the prototype
reaction from the triene 37 that yields a 67:33 mixture
of TAC and CAT adducts 38 and 39 (Chart 4).29 AM1
calculations give a 29:71 ratio, whereas a ratio of 72:28
is obtained by means of 3-21G calculations.30 This latter
ratio is therefore in excellent agreement with the experi-
mental results. Furthermore, these calculations allow us
to characterize the geometry of the favored transition
states. Thus, both adducts 38 and 39 must have issued
from chair-boat-chair (cbc) transition structures (46% and
28% respectively), but a fair share of the TAC adduct
should form through a chair-boat-boat (cbb) transition
a similar tricycle 32 obtained by TADA reaction of the
macrocycle 31 at an optimal temperature slightly above
115 °C (Scheme 4).12 Thus, the carbonyl function at C-11
of the trienone 1 lowers the TADA temperature by only
10 °C by comparison with the nonactivated series 31.
This almost nonexistent activation is further demon-
strated by the inability of Lewis acids to catalyze the
reaction:3c,21 When the macrocycle 1 was treated for 3 h
with tin tetrachloride in toluene at -20 °C, no trace of
tricycle was observed, and at room temperature, only
degradation products were obtained. The two other
“inactivated” trienes 2 and 3 were also submitted to the
TADA reaction conditions. The 11-R-acetoxy-isomer 2
starts reacting at 80 °C and yields only one adduct, the
TAC tricycle 6. The 11-â-isomer 3 reacts less readily,
since no significant amounts of TADA adducts 7, 8, and
9 are obtained below 105 °C.
The structure of compound 6 was confirmed by crystal
X-ray diffraction analysis22 and that of compound 8 was
determined by NMR correlation with the known tricycles
32 and 34 (Scheme 4).12 Compounds 7 and 8 could not
be separated because they have the same polarity in
various solvents. In the same way 9 could not be
separated from the starting material 3. According to the
1
shape of the H NMR CHOAc signals for the tricycles, 7
has an equatorial acetate (wide multiplet signal), whereas
8 and 9 have an axial acetate (narrow multiplet signal).
The respective relative geometry of the three tricycles
was then easily attributed according to these observa-
tions and to a previous experiment carried out on related
macrocycle 33. In that experiment, 36, the only CAT
adduct obtained, has the same geometry as 9 and the
TAC adduct 34 corresponding to 8 is also the most
abundant product (Scheme 4).
(23) Dory, Y. L.; Soucy, P.; Drouin, M.; Deslongchamps, P. J . Am.
Chem. Soc. 1995, 117, 518.
(24) Takahashi, T.; Sakamoto, Y.; Doi, T. Tetrahedron Lett. 1992,
33, 3519.
(25) Binkley, J . S.; Pople, J . A.; Hehre, W. J . J . Am. Chem. Soc. 1980,
102, 939.
(26) Schmidt, M. W.; Baldridge, K. K.; Boatz, J . A.; Elbert, S. T.;
Gordon, M. S.; J ensen, J . H.; Koseki, S.; Matsuraga, N.; Nguyen, K.
A.; Su, S. J .; Windus, T. L.; Dupuis, M.; Montgomery, J . A. J . Comput.
Chem. 1993, 14, 1347.
(27) Garc´ıa, J . I.; Mart´ınez-Merino, V.; Mayoral, J . A.; Salvatella,
L. J . Am. Chem. Soc. 1998, 120, 2415.
(28) Tripos Inc.: 1699 South Hanley Road, St. Louis, MO 63144-
2913.
(21) (a) Yates, P.; Eaton, P. J . Am. Chem. Soc. 1960, 82, 4436. (b)
Inukai, T.; Kojima, T. J . Org. Chem. 1971, 36, 924. (c) Fleming, I.
Frontier Orbitals and Organic Chemical Reactions; Wiley: New York,
1976; Chapter 4. (d) Yamabe, S.; Dai, T.; Minato. T. J . Am. Chem. Soc.
1995, 117, 10994.
(22) Cambridge Data Bank reference for X-ray of tricycle 6: CCDC
190394. Formula:
Å, b ) 16.9509(16) Å, c ) 13.899(3) Å, â ) 113.728(19)°. Space group
P21/n.
C
116H176O36. Unit cell parameters: a ) 13.676(5)
(29) Ndibwami, A.; Lamothe, S.; Soucy, P.; Goldstein, S.; Deslong-
champs, P. Can. J . Chem. 1993, 71, 714.
(30) Zefirov, N. S. Tetrahedron 1977, 33, 2719.
J . Org. Chem, Vol. 68, No. 6, 2003 2393

Bourque et al.
CHART 4. TTT Mod el Stu d y Resu lts
adducts. Nevertheless, it is impossible to identify the
exact nature of that ketone (either ketone or enol) at the
transition state. We made the assumption that the ketone
form would still give a fair representation of the systems
at the transition state. The removal of all other groups
is less dramatic, so that the overall discussion should be
very meaningful. The aim of this study is to discover the
trends of three selected groups, separately at first, in
regards to their effect on the TADA selectivity, then to
see the results of their combinations, leading to either
synergism or antagonism. The three groups of particular
interest are (i) a carbonyl at position 3, (ii) a gem-
dimethyl at position 4, and (iii) a carbonyl at position 11
that could activate the dienophile.
For the forthcoming discussion, the bond that is formed
between rings A and B, in front of the methyl group of
the dienophile, will be called a , and the one that forms
at the BC ring junction will be designed as b. Let us now
analyze carefully each of the eight systems A-H.
A: The macrocycle A32 corresponds to the parent
system; its TADA reaction at 100 °C would theoretically
lead to 56% (55 + 1) of CAT adduct and 44% (38 + 6) of
TAC adduct. In this reference case, the TS cbc geometry
is always favored, although the TAC cbb geometry is not
negligible. This again confirms that a boat tether ring is
possible when it connects a cis ring junction. The mean
a bond value is 2.25 Å and that of bond b is 2.22 Å.
B: By adding a gem-dimethyl group at position 4 in
the macrocycle A,33 the cbc TSs are still favored but there
is even more TAC adduct arising from a cbb geometry.
The gem-dimethyl now favors the TAC adduct (95% )
63% + 30% + 2%) over the CAT adduct (5%) and the
activation energy has increased by 3.48 kcal/mol; it can
be assumed that this value corresponds roughly to the
destabilizing effect of the axial methyls facing each other
across the forming bond a . Indeed, this interaction looks
like a 1,3 diaxial steric interaction, although it is slightly
less severe than that found in 1,3 diaxial dimethyl
cyclohexane (3.7 kcal/mol).35 Addition of the gem-dimethyl
group leads also to concomitant lengthening of bond a
by 0.05 Å (now 2.30 Å) and shortening of bond b by 0.04
Å (2.18 Å). The presence of some TAC bbc adduct (2%)
demonstrates that there could exist some steric hin-
drance in the cbc and cbb TSs because of the colliding
axial methyl groups at positions 4 and 10 (see Chart 2).
This effect is not present in a bbc transition state, where
the methyl group at 4 is no longer axial.
structure (26%). The cbb TS structure leading to the CAT
adduct, on the other hand, is much less stable (0%), and
confirms the rule that states that a boat tether ring (ring
C in this case) is not favored when the corresponding ring
junction bond undergoing formation (bond b) is trans
(ring junction BC here).10,23,31 The other competing bbc
and bbb TS structures are even less stable.
Therefore, this particular difficult case was successfully
handled, and all of our previous work indicates that the
RHF/3-21G level of ab initio theory is well suited for the
current problem. All possible transition states dealing
with the TTT systems under scrutiny have been calcu-
lated. The results of these calculations have been drawn
up at the eight corners of a cube to ease the analysis and
the comparisons necessary to fully understand all the
various factors at work (Chart 5). Thus, the four back
transition structures A-D correspond to systems without
carbonyl at position 3, whereas the four structures E-H
display a carbonyl at that position. The four right
structures A, C, E, and F have no gem-dimethyl at
position 4, contrary to the four structures B, D, G, and
H on the left side. Finally the difference between the four
bottom transition structures A, B, E, and G and the four
top structures C, D, F , and H resides in the absence of
a carbonyl at position 11 in the former case and in its
presence in the latter. The transition structure G matches
rather closely to the experimental TADA reactions car-
ried out from the macrocyclic trienes 2 and 3, and H
corresponds to the conjugated case 1. The major differ-
ence between the theoretical calculation and the true
experiments consists of the removal of some pendant
groups that would lead to unreasonably long computer
times: (i) the methyl ester at position 2, (ii) the acetate
at position 11 (cases corresponding to 2 and 3), (iii) the
CH2OPiv group at position 13, and (iv) the MOM ether
at position 14. We are well aware of the importance of
the methyl ester at position 2, because the ketone at
position 3 stands truly as a ketone in the macrocycles
(as shown in the crystal structure of 1 and in the NMR
data) but it prefers to exist in its enol form in the tricyclic
C: From the reference system A, a carbonyl group is
added at position 11 to see its sole effect;34 the resulting
dienone C requires less energy (-1.56 kcal/mol). The
bond lengths are inverted; a is now the shorter at 2.11 Å
(bond shortened by 0.14 Å) and b the longer at 2.34 Å
(increase of 0.12 Å). The selectivity effect is the same as
that of a gem-dimethyl at position 4 (B), since the TAC
adduct is also favored (79% ) 12% + 67%) over the CAT
adduct (21% ) 19% + 2%). It is worth noting that the
most favored TAC transition state (67%) assumes a chair-
(32) The third calculated conformation with a relative energy of 2.78
kcal/mol corresponds to the crystal structure of macrocycle 1.
(33) The third calculated conformation with a relative energy of 2.45
kcal/mol corresponds to the crystal structure of macrocycle 1.
(34) The lowest calculated conformation corresponds to the crystal
structure of macrocycle 1.
(31) (a) Coe, J . W.; Roush, W. R. J . Org. Chem. 1989, 54, 915. (b)
Marshall, J . A.; Grote, J .; Audia, J . E. J . Am. Chem. Soc. 1987, 109,
1186. (c) Dory, Y. L.; Ouellet, C.; Berthiaume, S.; Favre, A.; Deslong-
champs, P. Bull. Soc. Chim. Fr. 1994, 131, 121.
(35) Allinger, N. L.; Miller, M. A. J . Am. Chem. Soc. 1961, 83, 2145.
2394 J . Org. Chem., Vol. 68, No. 6, 2003

First TADA Deactivating Enone Dienophiles
CHART 5. Resu lts of th e Ca lcu la tion sa
a
Each corner of the cube corresponds to a particular macrocyclic system among A-H. For each system the population of the expected
products at 100 °C is indicated based on the calculated relative energies of the TSs.30 The figures in brackets are the average lengths for
the TS bonds a and b (Å). The bold figures at the corners are the activation energies (∆∆E) for the most favored TSs (kcal/mol). The
figures along the vertices of the cube are differences of activation energies. For example, in the vertex connecting TSs A and B, the arrow
goes from A to B, the difference of activation energy is therefore equal to 40.72 - 37.24 ) +3.48 kcal/mol; this positive energy indicates
that the introduction of a gem dimethyl at position 4 led to an increase of the activation energy as expected. On the contrary, the differences
of activation energies that are underlined are unexpected from currently accepted rules.
boat-boat conformation. These calculations that are
directly supported by experimental data obtained from
the same macrocycle C24,36 add further credit to our
molecular modeling results.
A (109.5°). The effect of a large angle inside the tether
ring A (particularly at positions 2 or 3) could be to hold
the diene and the dienophile at larger distances so that
closing of the bond a (its formation) becomes less easy.
D: In system D,34 the gem-dimethyl group at 4 and the
carbonyl at 11 are both present. Since they both indi-
vidually favor the TAC adduct, their effects are now
synergetic and no CAT should be obtained with this
macrocycle. The preferred TS geometry is a cbb as in B.
On going from C to D the activation energy increases by
3.97 kcal/mol, virtually the same energy as that found
between A and B. From B to D, the activation of the
dienophile by a carbonyl is evidenced by the decrease of
activation energy of 1.07 kcal/mol. This activation gap is
not very different from that observed between A and C
(1.56 kcal/mol).
E: In the macrocycle E,34 we can probe the “pure” effect
of a carbonyl at position 3. The tendency is now to favor
the CAT TS (90% ) 89% +1%) mostly via the cbc
geometry (89%); this has no effect on the a and b bond
distances (see A and E). However, the activation energy
increases by 1.23 kcal/mol. This could come from the C2-
C3-C4 angle, which is larger in E (around 120°) than in
F : This explanation seems to gain credit in the case of
F ,34 in which the carbonyl at position 11 shortens the
bond a (E, 2.26 Å; F , 2.12 Å), while the effect of the
carbonyl at position 3 is precisely to prevent that bond
from shortening. A direct consequence of both opposing
effects is to drastically increase the energy of activation
by 5.64 kcal/mol from C to F and by 2.85 kcal/mol from
E to F . The latter increase is very impressive because,
from E to F , one normally expects the activation energy
to decrease, as it did on going from A to C, the dienophile
being more reactive. This constitutes therefore the first
evidence of an enone behaving as a deactivating dieno-
phile, and this theoretical assertion is strongly supported
by our experimental data. In terms of adduct selection,
the effects of both carbonyls are antagonistic so that there
is more CAT adduct in the case of F (70% ) 64% + 6%)
than in the case of C (21%) and less in F than in E (90%).
The geometrical features of E and F indicate that the
bonds a and b are only influenced by the groups at
positions 4 (none or gem dimethyl) and 11 (none or
carbonyl). Consequently, E and F are geometrically
(36) Roush, W. R.; Warmus, J . S.; Works, A. B. Tetrahedron Lett.
1993, 34, 4427.
J . Org. Chem, Vol. 68, No. 6, 2003 2395

Bourque et al.
CHART 6. Ca lcu la tion Resu lts for System s I a n d J
would be too heavy for 3-21G ab initio calculations and
the number of rotamers would unduly complicate the
problem.
There are 16 competing TSs in each case, but the eight
that bear an axial methyl group at position 13 have been
discarded because their energies are on the whole ex-
ceedingly high. For system I34 that resembles closely the
case of macrocycle 2, the calculations predict that only
one TAC adduct would be obtained over a total of four
possibilities. This TAC adduct would mostly be formed
via a bbc TS geometry, and it corresponds exactly to the
one experimentally observed (6). In the case of J ,37 the
calculations indicate that three adducts should be equally
formed: the two possible TAC adducts and one CAT
adduct over two possibilities. Noteworthy, this CAT
geometry theoretically found had been experimentally
observed in another TADA experiment involving a closely
related macrocyclic triene (compound 36 in Scheme 4).19
System J mimics closely the case of macrocycle 3, it yields
theoretical predictions that reasonably match the experi-
mental results (7/8/9 ratio at 105 °C calcd 34/34/32, found
10/60/30). Moreover, the relative reactivity of both sys-
tems is correctly modeled, since the lowest activation
energy for I is 41.74 kcal/mol and that for J is 43.50 kcal/
mol. These figures match the fact that the R acetate
macrocycle 2 already reacts at 80 °C whereas its â epimer
3 does not produce significant amounts of products before
105 °C.
similar to A and C, respectively, and the same rule can
be applied to the two remaining systems G and H that
look like B and D.
G: For G,34 there are also two opposing effects: the
carbonyl at position 3 that gives more CAT adduct (10%
) 2% + 8%) than in B (5%), where that carbonyl is not
present, and the gem-dimethyl group at position 4,
yielding more TAC adduct than in the case of E (10%),
which bears no gem-dimethyl. The 90:10 TAC:CAT
theoretical ratio matches well with the experimental
results of 2 and 3 where no or little CAT adducts were
isolated. In this case again, the additive effect of the
factors is further evidenced by the differences of activa-
tion energies which are equivalent for the couples AE/
BG (1.23 kcal/mol/1.39 kcal/mol) and AB/EG (3.48 kcal/
mol/3.64 kcal/mol).
III. Con clu sion s
An enone is usually known to behave as an activated
dienophile, but we could prove that in special environ-
ments, the results of the activation could be surmounted
by other factors so that deactivation would in fact ensue.
To achieve this reverse activation effect, the enone
dienophile and the diene were rigidly held in 14-
membered macrocycles by means of two four-carbon
tethers. On the side of the short forming bond, away from
the ketone, a methyl group was placed on the dienophile
and a gem-dimethyl group was added next to diene as
part of the tether. Theoretical calculations showed that
the resulting steric interaction, almost equivalent to 1-3
interaction, was not as strong as expected. In fact, these
calculations suggest two important geometrical features
in regard to the introduction of these three methyl
groups: (a) a methyl on the dienophile acts as a donor
and the bond that forms away from it becomes naturally
shorter (2.22 Å) with concomitant lengthening of the bond
directly linked to it (2.25 Å) and (b) a gem-dimethyl group
next to the diene acts in exactly the same way by
shortening the bond that forms further away from it. In
the systems studied here, both effects tend to open the
same forming bond, the one that is close to these methyls
(2.30 Å). However, when a carbonyl becomes conjugated
to the dienophile, it shortens this forming bond (2.15 Å)
in such a way that the hydrogen atoms of the methyl
groups give rise to steric repulsion (shortest H-H
distance calculated at 1.91 Å instead of 2.02 Å without
carbonyl) for chair-shaped tethers. Nevertheless, these
H: Finally, in H34 all three factors are at work, two of
which favor TAC selectivity. As a consequence, there is
more TAC adduct in H (97% ) 4% + 73% + 15% + 5%)
than there was in G (90% ) 6% + 4% + 79% + 1%). The
deactivating effect of a carbonyl at position 11 becomes
very severe in this case, since the activation energy
increases by 3.62 kcal/mol from G to H.
Systems G and H are obviously good mimics of the
TADA reaction for the macrocyclic trienes 1-3. However,
2 and 3 behave very differently from each other, in terms
of temperature of reaction and selectivity, although they
both have been simplified by the same macrocycle G for
the calculations. Clearly, the acetate at position 11, which
is either R in 2 or â in 3, has a drastic influence on the
outcome of their TADA reactions. The methyl pivaloate
substituent at position 13 and the methoxymethyl ether
at position 14 must also play a significant role in the
matter. Thus, to gain further confidence in the validity
of our theoretical approach, we wanted to check if
calculations could properly model both systems; so we
examined macrocycles I and J (Chart 6) and calculated
all their corresponding TSs. In I and J , the acetate at
position 11 is explicitly there, but the methyl pivaloate
at 13 and the MOM ether at 14 are replaced by a methyl
and a methyl ether, respectively. These simplifications
become necessary because the real macrocycles 2 and 3
(37) The second calculated conformation with a relative energy of
2.02 kcal/mol corresponds to the crystal structure of macrocycle 1 (this
comes as no surprise since the pseudoaxial acetate experiences strong
transannular interaction).
2396 J . Org. Chem., Vol. 68, No. 6, 2003

First TADA Deactivating Enone Dienophiles
calculations indicate that methyl-methyl 1-3 steric
interactions across a TS forming bond remain usually
moderate, precisely because those partially formed bonds
are sufficiently long to accommodate rather bulky groups.
Moreover, this steric interaction can be avoided when the
tether (short bond side) assumes a boat conformation
(shortest H-H distance calculated at 2.29 Å). On the
other hand, a carbonyl at position 3 proved excessively
efficient at antagonizing the activating effect of the enone
dionophile. This effect went as far as turning the enone
into a deactivating dienophile. The tether ketone is more
effective than steric interactions because it affects the
geometry of the tether, contrary to weaker steric non-
bonding interactions. The resulting tether wants to hold
the reactants further apart because the ketone internal
angle of 120° is more open than a CH2 having a angle of
109.5°. It is very likely that such unexpected results could
not be reproduced in intermolecular cases, where the
reactants can easily adjust their trajectories to avoid all
detrimental steric and electronic interactions.
a carbonyl at position 3. The required TAC tricycle is the
sole TADA adduct issued from macrocycle 2 having an
R-acetoxy group at position 11. Theoretical calculations
indicate that all three additional competing adducts
would be obtained through much more energetic TS,
contrary to macrocycles 1 and 3 that give mixtures of
tricycles.
Ack n ow led gm en t. Financial help from NSERC-
Canada and FCAR-Quebec (fellowship to Elyse Bourque)
is highly appreciated. We are also grateful to Marc
Drouin for the X-ray analysis.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and spectral data of compounds 1-9, 12-19, and 21-
30 and crystal structure of macrocycle 1 and tricycles 4 and
6; Cartesian coordinates, energies, and populations of the
macrocycles and TSs calculated at the 3.21G level of ab initio
theory (Charts 5 and 6). CIF file corresponding to the three
crystal structures 1, 5, and 6. This material is available free
of charge via the Internet at http://pubs.acs.org.
This work also puts us closer to the synthesis of
Kaurane or diterpines such as Taibaihenryiin C having
J O0265129
J . Org. Chem, Vol. 68, No. 6, 2003 2397