Pronounced steric effects of substituents in the Nazarov cyclization of aryl dienyl ketones

Article abstract of DOI:10.1002/anie.200801542

(Chemical Presented) 1,3-'s a crowd: The facility of the AlCl ₃-catalyzed Nazarov cyclization of electron-rich aryl dienones was found to be dependent upon the substitution of the diene portion (see scheme). For α,γ-substituted systems, pronoun

Full text of DOI:10.1002/anie.200801542

Angewandte
Chemie
DOI: 10.1002/anie.200801542
Cyclizations
Pronounced Steric Effects of Substituents in the Nazarov Cyclization of
Aryl Dienyl Ketones**
Andrew P. Marcus, Amy S. Lee, Rebecca L. Davis, Dean J. Tantillo,* and Richmond Sarpong*
The acid-catalyzed 4 p-electrocyclic ring clo-
sure of a divinyl carbonyl substrate (i.e., the
Nazarov cyclization) is a reaction of substan-
tial synthetic utility for pentannulation.[1]
Because of its growing popularity in the
arena of complex molecule synthesis, the
reaction has evolved to include several var-
iants.[2,3] We became interested in the Nazarov
Scheme 1. Retrosynthetic approach to tetrapetalone A.
reaction as a means to access the A and B
rings of the natural product tetrapetalone A
(1, Scheme 1),[4] which is an attractive syn-
thetic target because of its biological rele-
vance and architectural complexity. To date,
there have been no reported syntheses of this
natural product.[5] It was envisioned that
indanone 2 would serve as a versatile frag-
ment to which rings C and D could be
appended. In turn, 2 could arise from a
Nazarov cyclization of aryl dienyl ketone 3.
Congruent with our plans to utilize inda-
Scheme 2. Unusual facilitation of a-alkyl dienone substrates.
none 2 as a synthetic starting point for the synthesis of 1,
initial investigations focused on the conversion of readily
prepared[6] aryl dienyl ketones (e.g., 3) into indanone
products such as 2. Surprisingly, whereas a screen of various
Lewis acids and solvents at numerous temperatures produced
either no reaction or the eventual decomposition of g-methyl-
substituted aryl dienone 3a (Scheme 2), the a,g-dimethyl
substituted aryl dienone 3b reacted smoothly (238C, 1 h) in
the presence of catalytic quantities of AlCl3 to yield the
indanones 2a and b as a 4:1 mixture of diastereomers in 71%
yield.[7]
Prompted by the unexpected reactivity differences
observed in the Nazarov cyclization studies of 3a and 3b,
we perused the literature for related transformations, and
found only two examples of using conjugated dienyl ketones
in the Nazarov reaction prior to our work.[8] The previous
reports both appear to be special cases, either involving the
use of extremely forcing conditions (formic acid/phosphoric
acid mixtures, 908C, 6 h)[9] to obtain low yields of pentannu-
lated products or an activated system that contained a
polarizing dihydropyran moiety.[10]
[*] R. L. Davis, Prof. D. J. Tantillo
Department of Chemistry
University of California, Davis
One Shields Avenue, Davis, CA 95616 (USA)
Fax: (+1)530-752-8995
E-mail: djtantillo@ucdavis.edu
Homepage: http://blueline.ucdavis.edu/
A. P. Marcus, A. S. Lee, Prof. R. Sarpong
Department of Chemistry
University of California, Berkeley
Berkeley, CA 94720 (USA)
Fax: (+1)510-642-8369
E-mail: rsarpong@berkeley.edu
Homepage: http://www.cchem.berkeley.edu/rsgrp/
More recently, during the preparation of this Communi-
cation, Frontier and co-workers reported that aryl dienyl
ketones bearing an electron-withdrawing group at the a po-
sition of the diene fragment and a phenyl group at the
terminus (e.g., 4; [Eq. (1)]) undergo the Nazarov cyclization
[**] A.P.M., A.S.L., and R.S. are grateful to UC Berkeley, the National
Science Foundation (CAREER: CHE-0643264), GlaxoSmithKline,
and Boehringer Ingelheim for generous financial support, and to
Dr. Kathleen Durkin (NSF CHE-0233882) for help with molecular
modeling, and Eric Bunnelle and Dr. William Bunnelle for helpful
discussions. R.L.D. and D.J.T. gratefully acknowledge support from
UC Davis and the National Science Foundation (CAREER: CHE-
0449845 and computer time from the Pittsburgh Supercomputer
Center).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200801542.
Angew. Chem. Int. Ed. 2008, 47, 6379 –6383
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6379

Communications
to yield indanones (e.g., 5).[11] This recent study highlights an
quantum chemical methods.[15,16] Geometry optimizations and
frequency analyses (to verify that structures are minima or
transition-state structures) were carried out at the B3LYP/6-
31 + G(d,p) level.[17] Energy barriers for cyclization of all
Lewis acid bound intermediates were computed from the
s-trans conformation. Figure 1 shows AlCl3-bound 3a and 3b
(12 and 13, respectively), and the transition-state structures
for their cyclization (TS-12 and TS-13, respectively).
Consistent with our experimental results, the cyclization
barrier for 12 (DG° = 30.6 kcalmolÀ1) is predicted to be
considerably larger than that for 13 (DG° = 24.0 kcalmolÀ1),
even though the s-trans conformation of the reactant complex
was used for both. This barrier
important difference from our observations in that the 4 p
electrocyclizations reported by Frontier and co-workers are
electronically driven, whereas we present evidence for a
significant steric effect that facilitates the Nazarov cyclization
of conjugated dienes bearing a,g-substitution.
Our preliminary study of the scope of the observed steric
effect resulting from the diene substituents reveals it to be
quite general for a range of electron-rich aryl dienyl ketones
(Table 1).[12,13] For these aryl and heteroaryl substrates, the
Nazarov cyclization is only observed in those cases where
both a-Me and g-Me groups are present.
appears to be the result of an
unfavorable 1,3-allylic interaction
between the two methyl groups in
13 that is not present in 12. The
strain imparted to the AlCl3-bound
reactant by this interaction (see the
distorted
Table 1: Scope of the aryl dienyl ketone Nazarov reaction.[a]
Entry Substrate
1
t [h] Product
6a (R=H)
24
6
no reaction
60% yield (82:18 d.r.)
6b (R=Me)
C-C-C angle, 1328, in 13, Figure 1;
note also that the Me···Me distance
is significantly greater for TS-13
than for 13) is reduced as the
2
3
8a (R=H)
8b (R=Me)
16
6
no reaction
55% yield (75:25 d.r.)[b]
transition-state
structure
is
reached.[18,19] Replacement of the
g-methyl groups of 12 and 13 with
hydrogen atoms reduces the energy
difference (3.2 kcalmolÀ1) between
the corresponding barriers for cy-
clization (compare to the 6.6 kcal
molÀ1 difference between 12 and
13), additionally demonstrating the
importance of the 1,3-allylic steric
interaction.[14c]
10a (R=H) 18
10b (R=Me)
no reaction
84% yield (80:20 d.r.)
1
[a] All reactions were run at 708C with 25 mol% of AlCl3 as the catalyst in PhMe (0.1m). [b] 31% of the
starting material recovered.
In accord with our previous
It is well documented that Nazarov substrates bearing
alkyl groups at the a position undergo pentannulation with
high efficiency.[1a–c] This reactivity is attributable to a higher
population of the productive s-trans conformer (e.g., 14a;
[Eq. (2)]), which is preferred over the nonproductive s-cis
conformation (e.g., 14b).[14] This effect is certainly a signifi-
cant component of the observed reactivity for the
a,g-disubstituted substrates presented in Table 1,[14b–c] but
our results demonstrate that this is not the whole story (see
below).
To better understand the inherent differences that lead to
the observed facilitation of the 4 p electrocyclization with the
introduction of alkyl substituents at the a- and g-positions
(e.g., 3a versus 3b), a detailed analysis of the cyclizations of
Lewis acid bound dienyl ketones was undertaken by using
findings, d,d-dimethyl dienyl ketone 14, which does not
possess an a-Me or g-Me substituent, is unreactive under
the standard reaction conditions (25 mol% AlCl3, 0.1m in
PhMe, 238C) over 48 h. The computed DG° for this reaction
is 31.9 kcalmolÀ1 (based on AlCl3-complexed 14), which is
slightly higher than that for 12.
However, smooth conversion of aryl dienones 15 and 16
(both of which possess a Me but not g Me substituents) into
indanone products 17a and b and 18a and b, respectively, is
observed under identical reaction conditions after 2 hours.
The computed free energy barriers for these systems are
27.4 kcalmolÀ1 and 27.3 kcalmolÀ1, respectively—higher than
that for 13, which points to the importance of a g-Me
substituent in lowering the activation barrier, but significantly
lower than those associated with 12 and AlCl3-complexed 14.
The mixture of diastereomers 17a and b was readily
equilibrated to a 4:1 mixture favoring 17b[20] upon
treatment with 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) over 10 hours at 238C.[21]
The facile electrocyclization of 15 and 16 may be
rationalized by a preference for the productive s-
trans conformation for these substrates, which is
dictated by the a-Me substituent. Additionally,
6380 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6379 –6383

Angewandte
Chemie
serve any electronic effects while removing the steric
effect that we suggest is dominating the reactivity.
Exposure of 19 to the reaction conditions at 238C over
48 hours returns the starting substrate (19, 82% recov-
ery), in line with our prediction. The computed free
energy barrier for Lewis acid complexed 19 is 29.2 kcal
molÀ1, again higher than those for the Lewis acid
complexes of 3b, 15, and 16, and not much higher than
those for the complexes of 15 and 16. Consequently, we
hypothesized that 19 might be prone to cyclization if
heated. In fact, increasing the temperature to 708C led
to a 37% conversion of 19 into a 1.8:1 mixture of double
bond isomers of tricycle 20.[22]
On the basis of the experimentally observed reac-
tivity and the quantum calculations described above, it
would appear that systems for which free energy
barriers for cyclization of s-trans Lewis acid bound
reactants are computed to be less than 28 kcalmolÀ1 are
likely to undergo Nazarov cyclization under our stan-
dard conditions.[23] Conversely, those with computed
barriers greater than 29 kcalmolÀ1 are unlikely to
Figure 1. Computed structures and relative energies (B3LYP/6-31+G(d,p) [17]
for cyclization of 12 and 13 (energies in kcalmolÀ1; blacknumbers corre-
)
= À =
spond to distances in Š; blue numbers correspond to X C C C dihedral
angles; red numbers correspond to C-C-C bond angles).
however, we believe that compounded steric
effects (i.e. pseudo 1,3-allylic interactions;
[Eq. (3)]) of the methyl substituents in 16,
which are manifested in a variety of small
geometric distortions,[6] also contribute to effi-
cient cyclization for this substrate. These effects
are borne out in the computed barriers for these reactions,
which are calculated by using the s-trans reactants.
cyclize under these conditions, even if they possess an a-
alkyl substituent. This result should be taken as a rough
guideline for predicting reactivity, especially
for systems with computed barriers near to
the 28–29 kcalmolÀ1 range. For systems with
computed barriers below, but near to this
range, cyclization may be effected by increas-
ing the temperature.[24]
Finally, to access indanone products that
do not possess an a-alkyl substituent, a
removable activating group, similar to the
method of Frontier and co-workers,[11] may be
applied. For example, the cyclization of 21
(Scheme 3) proceeds in 59% yield at 238C to
afford b-ketoester 22. Saponification of 22
and subsequent decarboxylation upon workup affords inda-
none 2c.
In conclusion, we have found the Nazarov cyclization of
aryl dienyl ketones to be dependent on steric interactions
between diene substituents at the a- and g-positions.[25] The
proposed models are highly predictive for a range of aryl
To distinguish the effect of g-alkyl substituents from the
previously established facilitation by a-alkyl substituents for
this subset of Nazarov substrates, cyclic aryl dienyl ketone 19
[Eq. (4)] was designed. We hypothesized that this substrate
would be unlikely to undergo a Nazarov cyclization under our
standard conditions at room temperature, despite the fact that
the s-trans conformation should be favored
by virtue of the a-alkyl substitution (our
calculations suggest that the s-trans con-
formation is favored by ꢀ 2 kcalmolÀ1 over
the s-cis conformation[14]). The lack of
reactivity of 19 was anticipated because
1,3-steric clashes for the a and g substitu-
ents on the diene fragment are replaced by
a covalent tether. This design should pre-
Scheme 3. Synthesis of indanones without a substituents.
Angew. Chem. Int. Ed. 2008, 47, 6379 –6383
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6381

Communications
[14] a) For a detailed discussion, see the recent reviews on this
subject (references [1a–c]); b) Our quantum chemical calcula-
tions also bear this out. For example, the s-trans form of 12 is ca.
2 kcalmolÀ1 higher in energy than the s-cis form, whereas the
s-trans form of 13 is ca. 1 kcalmolÀ1 lower in energy than the
s-cis form; c) See the Supporting Information for details and
additional related examples.
[15] Cyclization is expected to be the rate-determining step of the
entire transformation. For a discussion, see reference [1].
[16] For computational analyses of other variants of the classic
Nazarov reaction, see: a) M. Harmata, P. R. Schreiner, D. R.
Lee, P. L. Kirchhoefer, J. Am. Chem. Soc. 2004, 126, 10954 –
10957; b) A. Cavalli, M. Masetti, M. Recanatini, C. Prandi, A.
Guarna, E. G. Occhiato, Chem. Eur. J. 2006, 12, 2836 – 2845;
c) O. Nieto Faza, C. S. López, R. lvarez, A. R. de Lera, Chem.
Eur. J. 2004, 10, 4324 – 4333.
[17] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648 – 5652; b) A. D.
Becke, J. Chem. Phys. 1993, 98, 1372 – 1377; c) C. Lee, W. Yang,
R. G. Parr, Phys. Rev. B 1988, 37, 785 – 789; d) P. J. Stephens, F. J.
Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys. Chem. 1994, 98,
11623 – 11627; e) These calculations were run by using Gaus-
sian03 (full citation can be found in the Supporting Information,
along with additional details on these calculations, including test
calculations in toluene). Free energy barriers include entropy
corrections computed for 258C.
[18] Initially we expected this strain to manifest itself by twisting the
diene out of conjugation, which would thereby prepare it for the
cyclization reaction, during which the conjugation is necessarily
reduced as the transition-structure is reached and the cyclic
delocalization typical for a pericyclic reaction develops in its
place. However, only small deviations from planarity are
observed in our computed structures for AlCl3-complexed
reactants (see Figure 1 and the Supporting Information). Twist-
ing of 2,4-dialkyl 2-trans-4-trans pentadienals has been studied:
H. Ogawa, Y. Taketugu, T. Imoto, Y. Taniguchi, H. Kato,
Tetrahedron Lett. 1979, 20, 3457 – 3460.
dienone compounds possessing alkyl substituents, and pro-
vide a more complete picture of the reactivity of this subset of
Nazarov substrates.[26] Our current efforts are focused on
applying these reactions to a total synthesis of tetrapetal-
one A.
Received: April 2, 2008
Revised: May 18, 2008
Published online: July 11, 2008
Keywords: annulation · density functional calculations ·
.
dienones · electrocyclic reactions · Nazarov cyclization
[1] For pertinent reviews, see: a) A. J. Frontier, C. Collison,
Tetrahedron 2005, 61, 7577 – 7606; b) M. A. Tius, Eur. J. Org.
Chem. 2005, 2193 – 2206; c) H. Pellissier, Tetrahedron 2005, 61,
6479 – 6517; d) M. Harmata, Chemtracts 2004, 17, 416 – 435.
[2] For a recent account featuring allenyl vinyl ketones, see: M. A.
Tius, Acc. Chem. Res. 2003, 36, 284 – 290.
[3] For examples of tandem reactions that involve capture of the
Nazarov intermediate, see: a) M. Janka, W. He, I. E. Haedicke,
F. R. Fronczek, A. J. Frontier, R. Eisenberg, J. Am. Chem. Soc.
2006, 128, 5312 – 5313; b) Y. Wang, B. D. Schill, A. M. Arif, F. G.
West, Org. Lett. 2003, 5, 2747 – 2750; c) G. Liang, Y. Xu, I. B.
Seiple, D. Trauner, J. Am. Chem. Soc. 2006, 128, 11022 – 11023.
[4] a) T. Komoda, Y. Sugiyama, N. Abe, M. Imachi, H. Hirota, H.
Koshino, A. Hirota, Tetrahedron Lett. 2003, 44, 7417 – 7419; b) T.
Komoda, M. Kishi, N. Abe, Y. Sugiyama, A. Hirota, Biosci.
Biotechnol. Biochem. 2004, 68, 903 – 908.
[5] The earlier report from Wang and Porco of an advanced
tetracyclic intermediate has recently been retracted, see: a) X.
Wang, J. A. Porco, Angew. Chem. 2006, 118, 6759; Angew. Chem.
Int. Ed. 2006, 45, 6607; b) X. Wang, J. A. Porco, Angew. Chem.
2005, 117, 3127 – 3131; Angew. Chem. Int. Ed. 2005, 44, 3067 –
3071.
[19] The relative stabilities of the uncomplexed aryl dienyl ketones
(e.g., 3a and 3b) may also be an important component.
However, in our studies, consideration of the reactive Lewis
acid bound structures were found to be a more significant
predictor of reaction facility since no sizable differences in the
lmax values were found for the aryl dienone substrates.[6]
[20] The syn and anti determinations were made on the basis of
coupling constants by using the Karplus correlation and NOESY
analysis.[6]
[6] For details, see the Supporting Information.
[7] Similarly, i was unreactive under the standard conditions,
whereas ii provided the corresponding indanone in 58% yield
as a 92:8 mixture of diastereomers.
[21] This equilibration protocol to enrich the product in one
diastereomer was found to be general for all the indanones
(see Table 1) obtained and in all other cases gave diastereomeric
ratios ꢁ 9:1.[6]
[22] Presumably, the exocyclic double bond isomerizes to an
endocyclic position under the reaction conditions.
[23] At this level of theory, and for systems bearing 3,5-dimethoxy-
phenyl groups (additional calculations are required to assess the
limitations of this energetic criterion). Note also that these
barriers should not be compared directly—in terms of their
absolute magnitudes—with experimental barriers, since they are
based on complexed reactants and do not include the effects of
solvation.[19]
[24] Barriers for several other systems have been calculated ahead of
experimental testing. See the Supporting Information for details.
[25] Although we have shown that the propensity of aryl dienyl
ketone substrates is to undergo Nazarov cyclization, those
substrates not expected to undergo the cyclization may partic-
ipate in alternate transformations. For example, upon exposure
of 23a [Eq. (5)] to catalytic AlCl3 at 238C, a 50% yield of Diels–
Alder dimer 24 is formed. On the other hand, related substrate
23b [Eq. (6)] bearing an a-methyl substituent undergoes the
[8] The work detailed in this manuscript was presented by RS and
APM in poster sessions at a Gordon Research Conference (July
16–21, 2006, Organic Reactions and Processes, Bryant Univer-
sity) and 232nd ACS National Meeting (San Francisco, CA,
United States, Sept. 10–14, 2006, ORGN-523), respectively.
[9] E. A. Braude, W. F. Forbes, J. Chem. Soc. 1953, 2208 – 2216.
[10] G. Liang, S. N. Gradl, D. Trauner, Org. Lett. 2003, 5, 4931 – 4934.
[11] W. He, I. R. Herrick, T. A. Atesin, P. A. Caruana, C. A.
Kellenberger, A. J. Frontier, J. Am. Chem. Soc. 2008, 130,
1003 – 1011.
[12] An electron-rich aryl fragment in the aryl dienyl ketone
substrate is important and consistent with recent findings, see:
J. A. Malona, J. M. Colbourne, A. J. Frontier, Org. Lett. 2006, 8,
5661 – 5664.
[13] This initial study is restricted to alkyl substituents.
6382
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6379 –6383

Angewandte
Chemie
Nazarov cyclization as expected. Calculations on
Nazarov cyclizations for these substrates can be
found in the Supporting Information. Aryl
dienyl ketone substrates possessing non-elec-
tron-rich aryl groups do not undergo the Naz-
arov reaction even at elevated temperatures. For
example, 26, bearing a tolyl group undergoes an
“anomalous” Nazarov cyclization[27] at 708C to
provide cyclopentenone 28 (via 27) in 61% yield
[Eq. (7)]. Efforts to understand the underlying
selectivity principles for these latter transforma-
tions are underway.
[26] Interestingly, problems with Cu-promoted Naz-
arov cyclization of an a-substituted (NO2),
g-unsubstituted aryl dienyl ketone were also
described in reference [11] (see Table 7 therein).
[27] S. E. Denmark, G. A. Hite, Helv. Chim. Acta
1988, 71, 195 – 208.
Angew. Chem. Int. Ed. 2008, 47, 6379 –6383
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6383