Polarizing the nazarov cyclization: The impact of dienone substitution pattern on reactivity and selectivity

Article abstract of DOI:10.1021/ja077162g

The impact of dienone substitution on the Nazarov cyclization has been examined in detail. Substrates bearing different substituents at each of four positions on the dienone backbone were systematically probed in order to identify trends leading to higher

Full text of DOI:10.1021/ja077162g

Published on Web 12/28/2007
Polarizing the Nazarov Cyclization: The Impact of Dienone
Substitution Pattern on Reactivity and Selectivity
Wei He, Ildiko R. Herrick, Tulay A. Atesin, Patrick A. Caruana,
Colleen A. Kellenberger, and Alison J. Frontier*
Department of Chemistry, UniVersity of Rochester, Rochester, New York 14627
Received September 15, 2007; E-mail: frontier@chem.rochester.edu
Abstract: The impact of dienone substitution on the Nazarov cyclization has been examined in detail.
Substrates bearing different substituents at each of four positions on the dienone backbone were
systematically probed in order to identify trends leading to higher reactivity and better selectivity.
Desymmetrization of the pentadienyl cation and oxyallyl cation intermediates through placement of polarizing
groups at both the C-2 and C-4 positions was found to be particularly effective. These modifications allowed
cyclizations to occur in the presence of catalytic amounts of mild Lewis acids. It was also found that
stereoconvergent cyclization of mixtures of E and Z isomers of alkylidene â-ketoesters occurred via an
efficient isomerization process that occurred under the reaction conditions.
the product expected from a conrotatory ring closure,3 and the
photochemical reaction gives the opposite diastereomer, as
Introduction
Electrocyclic reactions are powerful synthetic transformations
with the ability to create new carbon-carbon bonds stereospe-
cifically by simple orbital reorganization. One type of electro-
cyclic reaction is a 4π-electron process known as the Nazarov
cyclization, involving the conversion of divinyl ketones 1 to
cyclopentenones 5 by activation with a Lewis acid (eq 1).1
expected from disrotatory ring closure.2,4 In some cases,
however, isomerization of the dienone complicates analysis.5
The Nazarov reaction should be recognized as a valuable
synthetic transformation, since the stereospecific electrocycliza-
tion can convert achiral molecules into single stereoisomeric
products. However, the cyclization of simple dienones like those
depicted in eq 1 is often plagued with reactivity and selectivity
problems that seriously compromise synthetic utility. Specifi-
cally, (1) strong Lewis acids are often necessary to promote
cyclization; (2) one or more equivalents of promoter are required
in most cases; (3) regioselectivity of the elimination step can
be unselective (see 3 f 4); (4) elimination of the proton often
leads to loss of a stereocenter (see 4); and (5) the final enolate
protonation is often unselective (see 4 f 5). Two strategies
addressing some of the problems associated with synthetic utility
in Nazarov cyclization have been disclosed, and both involved
modification of the substitution pattern of the substrates 1.
Regioselective elimination became possible with the develop-
ment of Denmark’s silicon-directed Nazarov cyclization pro-
tocol, in which â-silyl divinyl ketones are employed in the
cyclization.6 West found that the intermediate cation (see 3, eq
Cyclization of pentadienyl cation 2 must proceed with
conservation of orbital symmetry, dictating conrotatory ring
closure to give a product with an anti relationship between R1
and R2 (see 3, eq 1). Since disrotatory closure is electronically
forbidden in the thermal reaction, stereospecificity is ensured
for the bond formation.2 Experimental data is consistent with
this prediction: thermal cyclization under acidic conditions gives
(3) Shoppee, C. W.; Cooke, B. J. A. J. Chem. Soc., Perkin Trans. 1 1972,
2271.
(4) Visser, C. P.; Cerfontain, H. Recl. TraV. Chim. Pays-Bas 1983, 102, 307.
(5) (a) Noyori, R.; Ohnishi, Y.; Kato, M. Tetrahedron Lett. 1971, 19, 1515.
(b) Noyori, R.; Ohnishi, Y.; Kato, M. Bull. Chem. Soc. Jpn. 1975, 48, 2881.
(c) Noyori, R.; Ohnishi, Y.; Kato, M. J. Am. Chem. Soc. 1975, 97, 928.
(6) (a) Denmark, S. E.; Jones, T. K. J. Am. Chem. Soc. 1982, 104, 2642. (b)
Jones, T. K.; Denmark, S. E. HelV. Chim. Acta 1983, 66, 2377.
(c) Jones, T. K.; Denmark, S. E. HelV. Chim. Acta 1983, 66, 2397. (d)
Denmark, S. E.; Habermas, K. L.; Hite, G. A.; Jones, T. K. Tetrahedron
1986, 42, 2821. (e) Denmark, S. E.; Klix, R. C. Tetrahedron 1988, 44,
4043. (f) Denmark, S. E.; Habermas, K. L.; Hite, G. A. HelV. Chim. Acta
1988, 71, 168. (g) Denmark, S. E.; Hite, G. A. HelV. Chim. Acta 1988, 71,
195. (h) Denmark, S. E.; Wallace, M. A.; Walker, C. B. J. Org. Chem.
1990, 55, 5543.
(1) (a) Habermas, K. L.; Denmark, S. E.; Jones, T. K. Org. React. (N.Y.) 1994,
45, 1. (b) Santelli-Rouvier, C.; Santelli, M. Synthesis 1983, 429. (c)
Denmark, S. E. In ComprehensiVe Organic Synthesis; Trost, B. M.; Fleming,
I., Eds.; Pergamon Press: Oxford, 1991; Vol. 5, p 751. (d) Pellissier, H.
Tetrahedron 2005, 61, 6479. (e) Frontier, A. J.; Collison, C. Tetrahedron
2005, 6, 7577. (f) Harmata, M. Chemtracts: Org. Chem. 2004, 17, 416.
(g) Tius, M. A. Eur. J. Org. Chem. 2005, 11, 2193.
(2) (a) Woodward, R. B.; Hoffmann, R. Angew. Chem., Int. Ed. Engl. 1969,
8, 781. (b) Woodward, R. B.; Hoffmann, R. The ConserVation of Orbital
Symmetry; Verlag Chemie: Weinheim, 1970.
9
10.1021/ja077162g CCC: $40.75 © 2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 1003-1011
1003

A R T I C L E S
He et al.
Scheme 1. Polarized Substrates for Nazarov Cyclization
1) could be intercepted by various nucleophilic species.7,8 This
strategy is significant because trapping the intermediate cation
postpones elimination, such that both of the stereo-
centers created in the electrocyclic reaction are retained in the
product.
indicating multiple competitive reaction pathways. Better suc-
cess in the cyclization of these electron-poor substrates was
reported in two more recent studies: both Kerr17 and Aggarwal18
were able to achieve efficient cyclization of dienones bearing
electron-withdrawing groups using alternative Lewis acidic
promoters. However, Nazarov cyclizations of substrates bearing
both an electron-donating group at C-2 and an electron-
withdrawing group at C-4 had not been explored.
Our initial study19 and the closely related studies of Trauner,20
Tius,21 and Occhiato22 have established that divinyl ketones
bearing a heteroatom donor at one of the R-positions are indeed
more reactive and cyclize efficiently in the presence of catalytic
amounts of various Lewis acids. A full account of our study on
Nazarov cyclizations of polarized divinyl ketones of type 1P is
given in this article.
Inspired by the high reactivity engendered by the donor/
acceptor relationship of diene and dienophile in the Diels-Alder
reactions of Danishefsky9 and Rawal,10 we wondered whether
polarization or desymmetrization of the intermediate pentadienyl
cation 2 might improve the reactivity and selectivity of the
Nazarov cyclization. Synthesis and cyclization studies on a
divinyl ketone of type 1P, bearing an electron-donating group
at one R position (e.g., C-2, Scheme 1) and an electron-
withdrawing group at the other (e.g., C-4), were planned. It was
hoped that treatment of 1P with a Lewis acid would allow the
development of complementary orbital coefficients at the termini
of intermediate 2P, improving reactivity (Scheme 1). The
substituents at C-2 and C-4 would also desymmetrize the
oxyallyl cation (3P), which was expected to improve the
selectivity of the elimination step.
Before these studies were begun, little data concerning the
reactivity of dienones bearing electron-donating and electron-
withdrawing groups was available. It had been reported that
dienones bearing a C-2 heteroatom cyclized efficiently, although
like most other documented Nazarov cyclizations, a protic acid
medium or treatment with stoichiometic amounts of strong
Lewis acids was usually required to promote the reaction.11-13
The cyclization of C-4 carboalkoxy and carbamate-substituted
divinyl ketone substrates had been studied by Marino,14 Regan,15
and Takeda,16 who found that superstoichoimetric amounts of
the strong promoters SnCl4 and TMSI were required. The
reactions were typically slow (24 h) and low-yielding (<50%),
Results and Discussion
Development of the Catalytic Nazarov Cyclization of
Polarized Substrates. Initial screening was carried out on 2,4,6-
trimethoxyphenyl-substituted dihydropyran substrate 6. The
cyclization was very rapid (<1 min) even at room temperature
with stoichoimetric AlCl3, Al(OTf)3, Sc(OTf)3, and Cu(OTf)2,
as well as acetic acid, and the â-ketoester product 7 was isolated
in nearly quantitative yield (eq 2). Furthermore, highly efficient
cyclization of 6 was also observed upon treatment with 2 mol
% of either Cu(OTf)2, Sc(OTf)3, or Al(OTf)3 in CH2Cl2. The
reaction was fast (<5 min) and high yielding (>99%) in all
three cases. Since it was likely that triflic acid could also serve
as a catalyst for the cyclization, we carried out control
experiments to rule out the possibility of expeditious catalysis
by triflic acid, often present in the commercial Cu(OTf)2 reagent.
Thus, when a solution of Cu(OTf)2 was treated with excess
potassium carbonate prior to addition of substrate, the cyclization
behavior of 6 was unchanged.
(7) (a) Βender, J. Α.; Blize, A. E.; Browder, C. C.; Giese, S.; West, F. G. J.
Org. Chem. 1998, 63, 2430. (b) Giese, S.; West, F. G. Tetrahedron Lett.
1998, 39, 8393. (c) Bender, J. A.; Arif, A. M.; West, F. G. J. Am. Chem.
Soc. 1999, 121, 7443. (d) Browder, C. C.; West, F. G. Synlett 1999, 1363.
(e) Wang, Y.; Arif, A. M.; West, F. G. J. Am. Chem. Soc. 1999, 121, 876.
(f) Giese, S.; West, F. G. Tetrahedron 2000, 56, 10221. (g) Giese, S.;
Kastrup, L.; Stiens, D.; West, F. G. Angew. Chem., Int. Ed. Engl. 2000,
39, 1970. (h) Browder, C. C.; Marmsater, F. P.; West, F. G. Org. Lett.
2001, 3, 3033. (i) Wang, Y.; Schill, B. D.; Arif, A. M.; West, F. G. Org.
Lett. 2003, 5, 2747. (j) Browder, C. C.; Marmsater, F. P.; West, F. G. Can.
J. Chem. 2004, 82, 375. (k) Yungai, A.; West, F. G. Tetrahedron Lett.
2004, 45, 5445. (l) White, T. D.; West, F. G. Tetrahedron Lett. 2005, 46,
5629. (m) Rostami, A.; Wang, Y.; Arif. A. M. McDonald, R.; West, F. G.
Org. Lett. 2007, 9, 703.
(8) For other recent examples of interrupted Nazarov cyclization, see (a) Dhoro,
F.; Tius, M. A. J. Am. Chem. Soc. 2005, 127, 12472. and (b) Dhoro, F.;
Kristensen, T. E.; Stockmann, V.; Yap, G. P. A.; Tius, M. A. J. Am. Chem.
Soc. 2007, 129, 7256.
(9) (a) Danishefsky, S.; Kitahara, T. J. Am. Chem. Soc. 1974, 96, 7807. (b)
Danishefsky, S.; Kitahara, T.; Yan, C. F.; Morris, J. J. Am. Chem.
Soc. 1979, 101, 6996. (c) Danishefsky, S. J. Acc. Chem. Res. 1981, 14,
400.
Encouraged by the successful catalytic Nazarov cyclization
of substrate 6, a series of dihydropyran substrates were prepared
to explore the scope of the method (Table 1). While it was
possible to effect the cyclization using a catalytic amount of a
number of different Lewis acids, we chose to study the reaction
using Cu(OTf)2 because of the documented performance of
copper (II)-chiral ligand complexes in asymmetric reaction
(10) Kozmin, S. A.; Janey, J. M.; Rawal, V. H. J. Org. Chem. 1999, 64, 3039.
(11) Τius, Μ. Α.; Κwok, Ì.-Κ.; Gu, X.-q.; Zhao, C. Synth. Commun. 1994,
24, 871.
(17) Kerr, D. J.; Metje, C.; Flynn, B. L. Chem. Commun. 2003, 1380.
(18) Aggarwal, V. K.; Beffield, A. J. Org. Lett. 2003, 5, 5075.
(19) He, W.; Sun, X.; Frontier, A. J. J. Am. Chem. Soc. 2003, 125, 14278-
14279; addition/correction J. Am. Chem. Soc. 2004, 126, 10493.
(20) Liang, G. X.; Gradl, S. N.; Trauner, D. Org. Lett. 2003, 5, 4931.
(21) Bee, C.; Leclerc, E.; Tius, M. A. Org. Lett. 2003, 5, 4927.
(22) Occhiato, E. G.; Prandi, C.; Ferrali, A.; Guarna, A.; Venturello, P. J. Org.
Chem. 2003, 68, 9728.
(12) Casson, S.; Kocienski, P. J. Chem. Soc., Perkin Trans. 1 1994, 1187.
(13) Kim, S.-H.; Cha, J. K. Synthesis 2000, 2113.
(14) a) Marino, J. P.; Lindermann, R. J. J. Org. Chem. 1981, 46, 3696. (b)
Marino, J. P.; Lindermann, R. J. J. Org. Chem. 1983, 48, 4621.
(15) Andrews, J. F. P.; Regan, A. C. Tetrahedron Lett. 1991, 32, 7731.
(16) Sakai, T.; Miyata, K.; Takeda, A. Chem. Lett. 1985, 1137.
9
1004 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Polarizing the Nazarov Cyclization
A R T I C L E S
Table 1. Nazarov Cyclization of Dihydropyrans:a C-5 Variants
Table 2. Cyclization of Alkylidene â-Ketoesters:a C-1/C-2 Variants
a Reaction Conditions: 2 mol % Cu(OTf)2 in dichloromethane (rt) or
dichloroethane (higher temperatures). b E ) COOMe; TMP ) 2,4,6-
trimethoxyphenyl. c Even though shown in an E-configuration, substrates
sometimes contain Z-isomers. d Denotes position of double bond after
elimination.
Cyclization of 26 was particularly complex, resulting in a
mixture of regioisomers with 27 as the major component.
For most of the cyclizations examined in Tables 1 and 2,
elimination was regioselective and products were isolated as
single diastereoisomers. In contrast to the typically unselective
elimination pathways observed for unpolarized substrates (3 f
4, eq 1), in these polarized substrates elimination of protons
adjacent to the C-2 terminus of the oxyallyl cation was favored.
This selectivity can be attributed to the electron-donating, cation-
stabilizing group at the C-2 position and is further illustrated
by the outcome of the cyclization of 26. The electron-
withdrawing group at C-4 was also found to be essential to
diastereoselectivity in the final protonation step (4 f 5, eq 1).
In contrast to the typically poor facial selectivity observed in
the protonation of the enolate intermediate in substrates lacking
an electron-withdrawing group,6e,7f,20 the relative stereochemistry
at C-4 and C-5 of â-ketoester products in Table 2 was
predominantly trans, with minor (<10%) cis product detected
in a few cases. The trans relationship was indicated by the
coupling constant between the R-proton and the â-proton (1.4-
2.2 Hz), an assignment that was later confirmed by X-ray single
crystallography of 7 (Figure 1).
In these R, â-substituted cyclopentenones, it is expected that
fewer unfavorable steric interactions would exist in the trans
isomer relative to the cis. It has been assumed that a thermo-
dynamic equilibration at the C-4 carbon of the product â-ke-
toester occurs under the acidic reaction conditions of the
Nazarov cyclization, accounting for the diastereoselectivity of
the protonation in these cases.6a,15,17 Evidence for this has been
provided by NMR experiments: the kinetically favored cis
product is observed at low temperature and undergoes slow
conversion to the thermodynamically favored trans product
under the reaction conditions.25 No further epimerization at the
R-position of 7 was observed when the compound was treated
a Reaction Conditions: 2 mol % Cu(OTf)2, CH2Cl2 (0.06 M), rt. b TMP
) 2,4,6-trimethoxyphenyl, PMP ) p-methoxyphenyl, MMP ) m-methoxy-
phenyl. c Optimized conditions: Cu(OTf)2, 2 mol %, dichloroethane (0.5
M), 55 °C.
chemistry.23 Standard reaction conditions involving 2 mol %
of Cu(OTf)2 in chlorinated solvent led to efficient cyclization
of all substrates studied (Table 1), results that represented one
of the first examples of a general, efficient method for catalytic
Nazarov cyclization.
A strong correlation between electron-donating ability of the
C-5 substituent and reaction rate was found. Among the C-5
aryl-substituted cases (entries 1-5), the relative reaction rates
were as follows: 2,4,6-trimethoxyphenyl (TMP) > p-methoxy-
phenyl (PMP) > 2-furyl > 3-methoxyphenyl (MMP) > phenyl
(Ph). An alkyl substitution at C-5 was much less reactive:
cyclization of 16 was very slow when carried out at room
temperature, and decomposition of the substrate started to
compete with the cyclization (entry 6). However, if the
cyclization was carried out at elevated temperature and higher
concentration, efficiency improved (entry 7).
The survey of C-1/C-2 substituent effects began with alky-
lidene â-ketoester substrates shown in Table 2. Five substrates
were studied, all bearing a carbomethoxy substituent on C-4
and the 2,4,6-trimethoxyphenyl group at C-5 but differing at
C-1 and C-2. All substrates cyclized efficiently with 2 mol %
or 5 mol % Cu(OTf)2 in chlorinated solvents. Cyclization of
substrate 18 gave a mixture of cyclohexene isomers (entry 1),
a result of unselective elimination in the final step (see eq 1).
The cyclization of aromatic substrate 22 required higher
temperatures, but cyclization was still efficient with 5 mol %
Cu(OTf)2 (entry 3). Cyclization of acyclic substrates 24 and 26
was slower and product yields were lower (entries 4 and 5).24
(24) Concurrent to our studies, Flynn and co-workers cyclized similar substrates
with a stoichiometric amount of methylsulfonic acid (MeSO3H): see ref
17.
(25) Carbon numbers in Figure 1 were assigned for the purposes of X-ray
structure analysis and are not consistent with the numbering scheme used
in the rest of the article.
(23) (a) Evans, D. A.; Rovis, T.; Johnson, J. S. Pure Appl. Chem. 1999, 71,
1407. (b) Johnson, J. S.; Evans, D. A. Acc. Chem. Res. 2000, 33, 325.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 1005

A R T I C L E S
He et al.
We have found that cyclization of a mixture of E and Z
alkylidene â-ketoesters of type I is stereoconvergent, giving a
single diastereomeric product II. This result contradicts the
commonly held belief that the olefin geometry of the substrate
is directly translated to the stereocenters of the product. The
Knoevenagel condensation of â-ketoesters with aldehydes, a
method often used to synthesize alkylidene â-ketoesters, is a
thermodynamically controlled process26 that gives a mixture of
E and Z isomers (see Table 3).27 The major component of the
mixture usually has E geometry, but in many cases the energy
difference between the isomers is not great. The isomers can
sometimes be isolated by chromatography, but reversion to the
thermodynamic equilibrium mixture is reported to occur upon
standing at or even below room temperature.17,28,29 This behavior
has complicated the study of the individual isomers.
Therefore, we were pleased to discover that when the mixture
of E and Z diastereomers was treated with catalytic copper
triflate, a high yield of a single Nazarov cyclization product
diastereomer was isolated. Interestingly, the stereochemistry of
the product corresponded to the expected conrotatory cyclization
of the minor Z isomer.30 Stereoconvergent behavior of this kind
was also observed in studies of reductive Nazarov cyclization:
both E-III and Z-III were converted to the same product mixture
IV (eq 3).7b,7f Similarly, in one of Denmark’s early studies, it
was found that the analogous cis-disubstituted â-silyl enone V
isomerized before it cyclized.6d However, it was not clear
whether this kind of isomerization could always be expected to
occur prior to cyclization, and in particular, no studies addressing
isomerization behavior in trisubstituted olefins had been con-
ducted. The efficiency and selectivity of the coupled isomer-
ization/cyclization prompted us to carry out a more thorough
study of stereoconvergence in the Nazarov cyclization of
alkylidene â-ketoesters.
Figure 1. X-ray crystal structure of 7.25
Scheme 2. Formation of Vicinal Stereocenters via Nazarov
Cyclization
with NaOH/MeOH in THF at 55 °C, indicating that the product
mixture already reflected the thermodynamic distribution of
diastereoisomers.
Stereoconvergent Cyclization of Alkylidene â-Ketoesters.
In many Nazarov cyclization applications, the elimination step
leads to the loss of a stereocenter during cyclization (Scheme
2; top). If the dienone substrate has a strategically placed
tetrasubstituted olefin, stereocenter loss does not occur and two
new stereocenters are created (I to II, Scheme 2).
One would expect that successful implementation of this
strategy would require synthesis of dienone I with geometric
fidelity, since a number of studies of the Nazarov cyclization
have indicated that the reaction proceeds with stereochemical
predictability according to the Woodward-Hoffmann rules.4,5
In most of these studies, substituents on the terminal carbons
of the pentadienyl cation system are fixed in one position (as a
member of a ring, for example), or the â-substituent is oriented
trans to the central carbonyl of the dienone (as the dienone R5
substituents are in Scheme 2). In these cases, olefin geometry
translates directly to the new sp3 stereocenters. However,
experiments have revealed that in cyclizations of acyclic Z
dienones,7f cyclic Z dienones with significant angle strain,5 and
acyclic alkylidene â-ketoesters (vide infra), olefin geometry did
not translate to the two new stereocenters of the product as one
would predict based on the conservation of orbital symmetry.
Because these substrate dienones were not fixed in an E or Z
configuration, they underwent isomerization under the reaction
conditions, which explains the apparent violation of the Wood-
ward-Hoffmann rules.
Stereoconvergent cyclization behavior was observed in di-
enones with both aryl and alkyl substituents, as shown in Tables
3 and 4. Both [5,6]- and [5,7]-fused ring systems were formed
in good yield. However, cyclization of substrates with an internal
methyl substituent at C-1 and R ) alkyl substitution was
sluggish (34E/Z) or completely inhibited (36E/Z). Cyclopen-
tenyl enone 41E/Z did not cyclize.31
Creation of vicinal tertiary centers was also possible but
compromised by competing internal elimination of a proton (and
loss of a stereocenter) to give a mixture of products of types A
(26) Janka, M.; He, W.; Frontier, A.; Flaschenriem, C.; Eisenberg, R. Tetra-
hedron 2005, 61, 6193.
(27) (a) Tietze, L. F.; Beifuss, U. In ComprehensiVe Organic Synthesis; Trost,
B. M.; Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 2, p 341.
(b) Tanikaga, R.; Konya, N.; Hamamura, K.; Kaji, A. Bull. Chem. Soc.
Jpn. 1988, 61, 3211.
3
(28) Assignment of E and Z isomers was accomplished by analysis of JC,H
coupling constants: see Kingsbury, C. A.; Draney, D.; Sopchik, A.; Rissler,
W.; Durham, D. J. Org. Chem. 1976, 41, 3863.
(29) We observed slow reversion of a neat sample of a single isomer (isolated
by chromatography) to the thermodynamic mixture in days at 25°C.
(30) The same result was obtained with catalytic Sc(OTf)3, AlCl3, or [IrMe-
(CO)(dppe)(DIB)](BARF)2.
(31) NOE data were used to assign stereochemistry: see Supporting Information.
9
1006 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Polarizing the Nazarov Cyclization
A R T I C L E S
Table 3. Nazarov Cyclization of Alkylidene â-Ketoesters I:a C-5 Variants; Internal C-1 Substituent ) CH3
a Cu(OTf)2 (5 mol %), Cl(CH2)2Cl (0.2 M). b TMP)2,4,6-trimethoxyphenyl; PMP ) p-methoxyphenyl; PNP ) p-nitrophenyl. c Dienones 36E/Z were
recovered unchanged. d Run at 0.5 M in Cl(CH2)2Cl. e Significant decomposition occurred; no desired product was isolated.
and B (Table 4). Only one diastereomer of product type A was
observed (again corresponding to Z isomer cyclization), even
though the starting alkylidene was greater than 95% E geometry.
Cyclization occurred efficiently for 42E/Z and 44E/Z, in
contrast to the poor reactivity observed in the analogous
tetrasubstituted systems (Table 3, 34E/Z and 36E/Z). Appar-
ently, if the substituents at the internal â-positions of the dienone
(C-1 and C-5) experience too much steric congestion, cyclization
efficiency is compromised.32
Observed cyclization times for pure E and Z isomers are
presented in Table 5. For each pair, cyclization of the individual
isomers led to the same diastereomeric product. 20E and 20Z
cyclized at the same rate and so did 28E and 28Z (entries
1-4).33 It was found that 42Z isomer cyclized faster and slightly
more efficiently than the 42E isomer (entry 7 vs entry 8).
Similarly, although 30E and 30Z were inseparable, 30Z was
consumed more quickly than 30E (entry 5). After 1 h, it was
possible to isolate 31 and unreacted 30E, which allowed us to
carry out cyclization of pure 30E (entry 6).
substitution. When R is an electron-rich aromatic (entries 1-4),
observed reaction time is the same for both isomers, so
isomerization must be faster than ring closure.34 In contrast,
when R ) phenyl or n-propyl (entries 5-8), the Z isomer
cyclizes faster than the E isomer, indicating that Ecation-Zcation
isomerization must be slow relative to ring closure of Zcation
.
Quaternary center formation was found to be most efficient
for substrates with aromatic substitution, giving products with
limited synthetic utility (Table 3). Fortunately, cleavage of the
styrenyl double bond of 38 could be accomplished via
ozonolysis, to give versatile aldehyde 48 in 61% yield (Scheme 3).
A proposed mechanistic pathway is shown in Scheme 4.
Lewis acid activation of the E or Z dienone provides the
corresponding pentadienyl cation intermediates. While multiple
rotational conformers can exist, only two are properly aligned
for cyclization (Ecation and Zcation; Scheme 4). The stereochem-
istry of the product observed indicates that Ecation does not
(33) Harmata, M.; Schreiner, P. R.; Lee, D. R.; Kirchhoefer, P. L. J. Am. Chem.
Soc. 2004, 126, 10954.
According to the results in Table 5, isomerization can be
either faster or slower than ring closure, depending on dienone
1
(34) Further evidence: the E/Z isomer ratio (as observed by H NMR) did not
change over the course of cyclization with either copper triflate or [IrMe-
(CO)(dppe)(DIB)](BARF)2 as catalyst. When pure E/Z 20 or 28 was
subjected to copper triflate, the interconversion was observed within 5 min
of mixing by NMR at rt. The isomers had reached equilibrium before any
cyclization took place.
(32) The poor reactivity of cyclopentenyl-substituted â-ketoester alkylidenes has
been documented: see ref 16.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 1007

A R T I C L E S
He et al.
Table 4. Nazarov Cyclization of Alkylidene â-Ketoesters II: C-5 Variants; Internal C-1 Substituent ) H
a Cu(OTf)2 (5 mol %), Cl(CH2)2Cl (0.2 M). b TMP ) 2,4,6-trimethoxyphenyl.
Table 5. Comparison of Z versus E Alkylidene â-Ketoester
tion, to give a single product with stereochemistry corresponding
to Z isomer cyclization. It is likely that unfavorable steric
interactions in the E-configured pentadienyl cation intermediate
prevent cyclization. This stereoconvergent cyclization behavior
is significant for the future planning of asymmetric cyclization
strategies: a nonracemic chiral catalyst could allow enantiose-
lective synthesis of both types of targets shown in Scheme 4,
from a mixture of E and Z stereoisomers.
Cyclizationa
time
(h)
temp
C)
yield
(%)
entry
dienone
(
°
product
1
2
3
4
5
6
7
8
20Z
20E
28Z
28E
30Z
30E
42Z
42E
4
4
25
25
45
45
65
65
65
65
∼99
∼99
97
21
21
29
29
31
31
0.75
0.75
1b
10
2
96
71c
60
Effect of C-4 Electron-Withdrawing Groups on Reaction
Rate. Divinyl ketone substrates bearing different C-4 substit-
uents were also studied. In order to a make meaningful
assessment, identical reaction conditions were used for all the
cyclizations: 2 mol % Cu(OTf)2, in dichloromethane or
dichloroethane, at 0.5 M concentration. Substrates bearing 2,4,6-
trimethoxyphenyl substitution at C-5 were examined so that
isomerization would be faster than cyclization, and cyclization
rates could be more accurately compared (Table 6).
75
68
43A/Bd
43A/Bd
10
a Reaction conditions: Cu(OTf)2 (5 mol %), Cl(CH2)2Cl (0.2 M). b Time
observed for consumption of 30Z during cyclization of the 30E/30Z mixture.
c Estimated yield based on recovered 30E. d Ratio of 43A/B was 1:2.
Scheme 3. Ozonolysis of Cinnamyl-Substituted Nazarov
Cyclization Products
Most of the substrates cyclized with 2 mol % Cu(OTf)2 in
good to excellent yields. Significant decomposition was ob-
served in the reaction of 51, and product yield was low (entry
3). Since this substrate is not “polarized,” it was expected to be
less reactive. When the C-4 substituent was not an ester, the
stereoselectivity of the enolate protonation (see 4 f 5, eq 1)
was poor (entries 7 and 8). Overall, the reaction rates did
correlate with C-4 electron-withdrawing ability (CO2R > Cl >
H > CH3, entries 5-8), but the rate differences observed were
very modest.
cyclize, probably because of the significant steric interactions
between R1 and R2. However, Ecation can isomerize via σ bond
rotation to Zcation, which is able to cyclize.35-37
In summary, our findings indicate that alkylidene â-ketoesters
with E geometry undergo Nazarov cyclization to give products
that do not bear the stereocenters expected from conrotatory
cyclization. Instead, E/Z isomerization occurs prior to cycliza-
The versatile Knoevenagel condensation was the standard
protocol used to synthesize alkylidene â-ketoesters for Nazarov
cyclization studies. However, the relatively harsh conditions of
this procedure, such as high temperatures, lengthy reaction times,
and difficulties in making starting material, did not allow
(35) Ecation is most efficiently delocalized in these substrates, accounting for the
facility of the isomerization.
(36) Since selectivity arises from the differential reactivity of two isomers in
equilibrium (Ecation and Zcation), the process could be viewed as an example
of dynamic kinetic diastereoselection.
(37) It is likely that a similar process is responsible for the stereoselectivity
observed in the rearrangement of furfuryl carbinols to 3-hydroxycyclopen-
tenones: (a) Piancatelli, G.; Scettri, A.; Barbadoro, S. Tetrahedron Lett.
1976, 17, 3555. For a theoretical analysis of this reaction with conclusions
consistent with our experimental findings, see (b) Faza, O. N.; Lopez, C.
S.; Alvarez, R.; de Lera, A. R. Chem. Eur. J. 2004, 10, 4324.
(38) (a) Padwa, A.; Meske, M.; Ni, Z. Tetrahedron 1995, 51, 89. (b) Padwa,
A.; Chiacchio, U.; Kline, D. N.; Perumattam, J. J. Org. Chem. 1988, 53,
2238.
(39) (a) Canterbury, D. P.; Frontier, A. J.; Um, J. M.; Cheong, P. H.-Y.; Goldfeld,
D. A.; Huhn, R. A.; Houk, K. N. Unpublished results.
9
1008 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Polarizing the Nazarov Cyclization
A R T I C L E S
Scheme 4. Stereoconvergent Behavior of E/Z Mixtures in Nazarov Cyclization: Proposed Mechanism
Table 6. Cyclization of C-5 Trimethoxyphenyl-Substituted Dienones: C-4 Variants
a The olefin geometry was determined by coupling constants. b General reaction conditions: 2 mol % Cu(OTf)2, dichloroethane, 0.5 M. c Ratio of cis/
1
trans isomers at C-4.. d A complicated mixture was obtained. H NMR revealed the possible presence of other regioisomeric products. However, only the
product shown was isolated and characterized.
synthesis of certain dienone targets. In order to access the desired
alkylidene â-ketoesters, a reaction sequence reported by Padwa38
was adopted as a mild alternative (Scheme 5).39 The one-pot
procedure commences with a (3 + 2) cyclization of an alkyne
and a nitrone to form an isoxazoline (VII), followed by oxidative
extrusion of nitrosomethane by m-CPBA to form the appropriate
divinyl ketone (VIII).
This methodology was used to synthesize a series of aromatic
Nazarov substrates bearing different electron-withdrawing groups
(Table 7). This series of dienones was chosen for study because
(40) (a) Boger, D. L.; Lerner, R. A.; Cravatt, B. F. J. Org. Chem. 1994, 59,
5078. (b) Yamamura, K.; Watarai, S.; Kinugasa, T. B. Chem. Soc. Jpn.
1971, 44, 2440. (c) Dornow, A.; Menzel, H. Ann. 1954, 588, 40.
(41) Gilmartin, B.; Eisenberg, R. Unpublished results.
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 1009

A R T I C L E S
He et al.
Scheme 5. Sequential (3 + 2)Dipolar Cycloaddition/Oxidative Extrusion Protocol for the Synthesis of Alkylidene â-Ketoesters
Table 7. Cyclization of Aryl Enones: C-4 ) Electron-Withdrawing
Group Variants
relative rates: ester > amide > sulfone > phosphonate (entries
1-4). The ester (59) and amide (61) aryl enones cyclized at
45 °C, while the sulfone (63) and phosphonate (65) enones did
not begin to cyclize until the temperature was elevated to
80 °C. In all cases of successful cyclization, the product was
isolated in high yield as a single diastereomer with a trans
relationship between the R- and â-substituents on the product
indanone. Two substrates tested (R-cyanoenone 67; entry 5 and
R-nitroenone 70; entry 7) did not cyclize at 45 °C and then
decomposed at elevated temperatures. No cyclization products
were detected in the reaction mixtures, and higher catalyst
loadings did not improve the results. It is likely that the cyano
group (entry 5) does not cyclize because the Lewis acid binds
preferentially at the terminal nitrogen.41 In this linear arrange-
ment, the Lewis acid sits too far away from the ketone carbonyl
to allow formation of the pentadienyl cation. In the case of the
R-nitroketone, rapid decomposition occurs under the cyclization
conditions.
According to Hammett constants, the expected order of
reactivity based on electron-withdrawing character for the groups
studied should be nitro > sulfone > cyano > phosphonate >
ester > amide.42 Since phosphonate 65 and sulfone 63 were
both less reactive than ester 59 and amide 61 (Table 7),
cyclization rates are not dependent on the electron-withdrawing
ability of the C-4 substituent alone. However, reactivity will
also depend upon how well the substrate (acting as a bidentate
ligand) and the Lewis acid catalyst are able to coordinate to
produce an intermediate with strong cationic character at the
C-5 position. Facility of coordination will depend upon the
σ-donating ability of the C-4 substituent as well as the geometry
the substrate adopts as a bidentate ligand. Bond lengths increase
as heteroatoms are introduced at the C-4 position, and in the
case of sulfone 63 and phosphonate 65, this could lead to
significant skewing of the angle at which the oxygen of the
C-4 substituent sits, thus distorting the binding site for the
catalyst. This could explain the necessary increase in reaction
temperature, as well as the slower relative reaction rates. It is
also possible that with stronger electron-withdrawing groups,
the Lewis acid binds more tightly to the product indanones
(compare 60 and 62), and the slower reaction rates are a
reflection of slower rates of catalyst turnover rather than slower
cyclization.43
a Reaction conditions: 5 mol % Cu(ClO4)2, 0.1 M dichloroethane. b All
dienones were prepared by nitrone sequence shown in Scheme 5 unless
otherwise noted. c No cyclization was observed before decomposition of
the starting ketone. d Prepared by Knoevenagel condensation. e Prepared by
a modified Knoevenagel condensation.40
the nitrone synthetic strategy (Scheme 5) made them available
as single isomers with Z (out) geometry, rather than as the
thermodynamic mixture of E/Z isomers expected from synthesis
via Knoevenagel condensation. Studying the cyclization of
substrates with exclusively Z geometry avoids complications
caused by the isomerization processes noted for the E isomers
(see Table 1), allowing access to true relative reaction rates. In
fact, a pair of closely related Z and E isomers in this series
provides another example of this pattern: cyclization of the Z
isomer is complete in 8 h (entry 1), while the E isomer requires
26 h for conversion to product (entry 6), presumably because
isomerization is slow and must occur before cyclization does.
The results of selective 1H-decoupled 13C NMR and NOE
experiments indicate Z stereochemistry for R-nitroenone 70,
although this geometry is opposite to the results typically
obtained from Knoevenagel condensation (vide supra) and from
the results reported for synthesis of analogous alkylidene
R-nitroketones.40a,b
Summary
A series of Nazarov substrates bearing electron-donating
substituents at C-2 and electron-withdrawing substituents at C-4
were synthesized (1P). Treatment with catalytic amounts of a
mild Lewis acid (2 mol % Cu(OTf)2) gave high yields of the
Nazarov product for these substrates, which were expected to
cyclize via “polarized” pentadienyl cation intermediates. De-
(42) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165.
(43) Experiments conducted with high loadings of catalyst were complicated
by poor solubility of the copper(II) salts in chlorinated solvents.
Finally, varying the electron-withdrawing group at C-4 in the
aryl enones with Z olefin geometry provided the following
9
1010 J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008

Polarizing the Nazarov Cyclization
A R T I C L E S
symmetrization of the pentadienyl cation through substrate
engineering is thought to improve reactivity, by creating an
electron-donating terminus and an electron-accepting terminus
(2P). Similarly, the asymmetry of the resulting oxallyl cation
is thought to allow greater selectivity in the final elimination
step, which occurs adjacent to the terminus with the highest
population of positive charge (i.e., C-2, 3P). The electron-
withdrawing group at C-4 also allows thermodynamic equilibra-
tion to occur under the reaction conditions at that stereocenter,
delivering the more stable trans diastereomer with high dias-
tereoselectivity (5P). Little or no cis diastereomer was observed.
necessary to assess the impact of C-4 substituents. Two studies
were conducted: one employed substrates with E geometry
bearing the 2,4,6-trimethoxyphenyl substituent at C-5, which
isomerize faster than they cyclize, and the second focused on
substrates with exclusively Z geometry. These studies showed
that the role of the C-4 substituent in catalytic Nazarov
cyclization is complex, and that reaction rate is probably affected
by a combination of factors. The electron-withdrawing ability
of the C-4 substituent, the efficiency and geometry of substrate/
catalyst binding, and the facility of catalyst turnover are all
variables that could impact reaction rate, depending on the
substrate. It was found that substrates with a C-4 ester substituent
cyclized most readily, suggesting that this substituent has the
combination of electron-withdrawing ability and catalyst binding
profile most able to facilitate the Nazarov cyclization.
In conclusion, the reactivity and selectivity of the Nazarov
cyclization can be controlled by careful positioning of substit-
uents on the five-carbon dienone backbone. Desymmetrization
of the intermediates 2 and 3 through placement of polarizing
groups at both the C-2 and C-4 positions was found to be
particularly effective. These modifications allowed cyclizations
to occur in the presence of catalytic amounts of mild Lewis
acids, which could lead to the eventual development of a general
method for catalytic asymmetric Nazarov cyclization.44 During
the course of these studies, an efficient isomerization process
leading to the stereoconvergent cyclization of mixtures of E
and Z isomers of alkylidene â-ketoesters was fully characterized,
demonstrating that geometric fidelity is not guaranteed in
substrates with internal substituents at the C-5 position.
Systematic variation of the substituents on the C-1, C-2, C-4,
and C-5 revealed the steric and electronic impact of substitution
on each position of the polarized Nazarov substrate. The findings
can be summarized as follows:
Substitution at C-2. Substituents at this position had the most
profound impact on reaction rate. If the results in Tables 1 and
2 are compared, it is clear that dihydropyran substrates in Table
1 (oxygen substituent at C-2) cyclize much faster than the
cyclohexenyl counterparts in Table 2 (alkyl substituent at C-2).
Internal substitution at C-1 was found to slow cyclization
because of steric interactions with internal substituents at C-5:
the larger the internal C-1 substituent, the slower the cyclization.
The same steric limitations would be expected for internal
substituents at C-5, except that in alkylidene â-ketoesters,
isomerization can occur, meaning that the identity of the internal
substituent is not fixed.
Acknowledgment. We are grateful to the Research Corpora-
tion, the Petroleum Research Fund, NSF (CAREER: CHE-
0349045), and the University of Rochester for generous financial
support. We thank Prof. Patrick Holland (University of Roch-
ester) for solving the structure of â-ketoester 7 by X-ray
crystallography and Dr. Alice Bergmann (University of Buffalo)
for carrying out high-resolution mass spectroscopy. We also
acknowledge Dr. Sandip Sur (University of Rochester) for
valuable assistance with NMR spectroscopy.
Substitution at C-5. An unusual reactivity trend was
observed: 2,4,6-trimethoxyphenyl > 4-methoxyphenyl > 2-fu-
ryl > 3-methoxyphenyl > phenyl > cyclohexyl (entries 1-7,
Table 1). This trend was initially difficult to rationalize, since
it did not correlate with steric bulk or with the expected
electronic impact of substituents at this position on the reactivity
of reaction intermediates. It was eventually determined that in
the alkylidene â-ketoester substrates studied, the C-5 substituent
has a significant impact on E/Z isomerization rates of alkylidene
â-ketoester substrates. When the C-5 substituent was 2,4,6-
trimethoxyphenyl or 4-methoxyphenyl, isomerization was faster
than cyclization, so the reaction times for entries 1 and 2 in
Table 1 reflect cyclization rate. However, entries 3-6 in Table
1 do not represent rates of cyclization but rather rates of
isomerization prior to cyclization because isomerization is slow.
Supporting Information Available: Experimental procedures
and characterization data for all new compounds, X-ray crystal
structure coordinates, and files for compound 7 in CIF format.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA077162G
(44) Asymmetric Nazarov cyclization has been accomplished with the use of
chiral auxiliaries; see ref 17 and (a) Tius, M. A.; Harrington, P. E. Org.
Lett. 2000, 2, 2447. (b) Harrington, P. E.; Murai, T.; Chu, C.; Tius, M. A.
J. Am. Chem. Soc. 2002, 124, 10091. (c) Pridgen, L. N.; Huang, K.; Shilcrat,
S.; Tickner-Eldridge, A.; DeBrosse, C.; Haltiwanger, R. C. Synlett 1999,
1612. (d) delos Santos, D. B.; Banaag, A. R.; Tius, M. A. Org. Lett. 2006,
8, 2579. (e) Banaag, A. R.; Tius, M. A. J. Am. Chem. Soc. 2007, 129,
5328. For the use of chiral Lewis acids, (f) see ref 18 and Liang, G.; Trauner,
D. J. Am. Chem. Soc. 2004, 126, 9544. Recently, catalytic asymmetric
Nazarov cyclization has been achieved using chiral Bronsted acids: (g)
Rueping, M.; Ieawsuwan, W.; Antonchick, A. P.; Nachtsheim, B. J. Angew.
Chem., Int. Ed. 2007, 46, 2097.
Substitution at C-4. Since the rate of E to Z isomerization
complicated the measurement of relative cyclization rates for
many of the substrates studied, careful design of substrates was
9
J. AM. CHEM. SOC. VOL. 130, NO. 3, 2008 1011