Intramolecular 4+3 cycloadditions. A cyclohexenyl cation, its halogenated congener and a quasi-Favorskii rearrangement

Article abstract of DOI:10.1016/S0040-4039(02)00264-2

Treatment of alkoxycyclohexenols bearing a tethered diene substituent with a Lewis acid results in intramolecular 4+3 cycloaddition with complete endo selectivity. A cycloadduct bearing a bromo substituent at a bridgehead position undergoes a quasi-Favorskii rearrangement in near quantitative yield upon reaction with lithium aluminum hydride.

Full text of DOI:10.1016/S0040-4039(02)00264-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 2347–2349
Intramolecular 4+3 cycloadditions. A cyclohexenyl cation, its
halogenated congener and a quasi-Favorskii rearrangement
Michael Harmata,* Gary Bohnert, Laszlo Ku¨rti and Charles L. Barnes
Department of Chemistry, University of Missouri-Columbia, Columbia, MO 65211, USA
Received 16 January 2002; revised 2 February 2002; accepted 6 February 2002
Abstract—Treatment of alkoxycyclohexenols bearing a tethered diene substituent with a Lewis acid results in intramolecular 4+3
cycloaddition with complete endo selectivity. A cycloadduct bearing a bromo substituent at a bridgehead position undergoes a
quasi-Favorskii rearrangement in near quantitative yield upon reaction with lithium aluminum hydride. © 2002 Elsevier Science
Ltd. All rights reserved.
The intramolecular 4+3 cycloaddition reaction is a
powerful method for the construction of polycyclic ring
systems.1 In particular, when either the dienophilic
cation or the diene is contained in a ring, unique
structures can be obtained whose modification can give
rise to important ring systems.2 We demonstrated this
in our recent synthesis of (+)-dactylol.3 We have
reported that the intramolecular 4+3 cycloaddition is
possible when the cations are within rings of 5–8 as well
as 10 and 12 carbons, resulting in cycloadducts in very
good yields, in some cases with excellent stereoselectiv-
ity.4 In an attempt to broaden the scope of the chem-
istry, and to further study the chemistry of halogenated,
cyclic cations, we conducted the study described herein.
trichloride to the Grignard reagent was not advanta-
geous, the yield of 5 being only 42%.
The reaction of 5 with Lewis acids resulted in cycload-
dition to form 6 and other side products (Scheme 2). In
the presence of 1.2 equiv. of triflic anhydride and
2,6-lutidine, the yield of 6 varied between 37 and 48%
MeOH, CH(OMe)3
+
H+
O
O
O
OH
OMe
3, 23%
OMe
2, 44%
H+, C6H6
heat, 71%
MeO
1
Our goal was to study the generation of cyclohexenyl
allylic cations which were more reactive than oxyallylic
cations. We decided to pursue cation generation from
allylic alcohols, since there is good precedent that such
an approach would give cations capable of 4+3
cycloaddition.5
Scheme 1.
To that end, we prepared alkoxyenone 2 by treatment
of ketoenol
1 with trimethyl orthoformate and
methanol in the presence of acid (Scheme 1).6 On a 30
g scale, this resulted in the formation of a 44% yield of
2 along with a 23% yield of 3. The latter could be
cracked by heating in benzene in the presence of
pTsOH to afford 2 in 71% yield.
Treatment of 2 with Grignard reagent 4 produced
allylic alcohol 5 in 55% yield. Addition of cerium
* Corresponding author. Tel.: 573-882-1097; fax: 573-882-2754;
e-mail: harmatam@missouri.edu
Scheme 2.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00264-2

2348
M. Harmata et al. / Tetrahedron Letters 43 (2002) 2347–2349
with 16–20% recovered starting material and small
amounts of elimination product 7. Neat titanium tetra-
chloride added to a solution of 5 in dichloromethane at
−78°C destroyed the compound, but a 0.5 M solution
of this Lewis acid added in the same way afforded a
30% yield of the cycloadduct 6. At present, the best
Lewis acid for this process is scandium triflate, which
effects cycloaddition of 5 in catalytic amounts and at
room temperature to give 6 in 60% yield along with
small amounts (3–5%) of enol ether hydrolysis product
8.
Another approach to 15 involved the bromination/
dehydrobromination of 2. This afforded the target com-
pound in 60% yield (Scheme 5).
With 15 in hand, it was reacted with Grignard reagent
4. This resulted in only low yield of 16. Formation of
an organocerium reagent was helpful in this case. The
product 16 was then formed in 76% yield (16% recov-
ered starting material) (Scheme 6).
The cycloaddition of 16 was in general a low-yielding
process. Treatment with triflic anhydride and 2,6-
lutidine resulted in cycloaddition, but yields were only
as high as 45%. Elimination to form 18 was a problem
in these reactions. A catalytic amount of scandium
triflate in the presence of excess acetic anhydride gave
18 in 87% yield. The acetate derivative of 16 was either
destroyed (TiCl4) or underwent elimination (Sc(OTf)3,
80%). The best conditions for the cycloaddition so far
make use of 1.4 equiv. of triflic anhydride at −50°C (no
base!). This gave the cycloadduct 17 in 65% yield as a
single stereoisomer.10 The structure of 17 was estab-
lished chemically. Reaction of 17 with Bu3SnH and
AIBN gave 6 in 60% yield. We later confirmed the
structural assignment by obtaining a crystal structure of
17.
The structure of 6 was established by single-crystal
X-ray analysis of the corresponding oxime derivative. It
is important to note that 6 was formed as a single
diastereomer. In the context of related cycloadditions
involving cyclopentenyl cations, this is quite unusual.4
For example, we reported that the reaction of 9 with
titanium tetrachloride afforded the cycloadducts 10 and
11 in excellent yield, but as
a 2.4:1 ratio of
diastereomers (Scheme 3). On the other hand, West and
co-workers reported that the domino Nazarov/4+3
cycloaddition of 12 proceeded with complete exo
stereoselectivity (Scheme 4).7 It will be interesting to
establish the trends in simple diastereoselectivity for
other cyclic cations and tether lengths.
Our real interest in this area is to develop halogenated,
cyclic cations. We thus studied the preparation of 15.
Two routes were examined. The first involved the
preparation of the enol ether 14 by bromination of 1
followed by alkylation. This process worked well.
Treatment of 1 with 0.9 equiv. of bromine at −78°C
followed by reaction with 1 equiv. of triethylamine
afforded 14 in 90% yield on a l g scale.8 Methylation of
14 was attempted with TMSOMe, trimethyl orthofor-
mate and acid, potassium carbonate and methyl sulfate
and sodium hydride/methyl iodide. Of these, only the
latter worked, giving 15 in 42% yield. A better
approach involved the reaction of 14 with dia-
zomethane.9 Treatment of an ethereal solution of 14
with diazomethane at room temperature for 3 h
afforded 15 in 79% yield, along with recovered starting
material (15%) and a small amount of unidentified
byproducts.
We have recently been developing the quasi-Favorskii
rearrangement as a synthetic tool, since chemistry such
as that described herein allows the convenient prepara-
tion of bridgehead halides poised to undergo this reac-
tion.11 It was thus gratifying to find that the treatment
PhO2S
Scheme 5.
TiCl4
81%
C
O
C
O
+
OEt
H
HO
2.4:1
10
Tf2O
CH2Cl2, -50 oC
11
4, CeCl3
Br
9
Br
O
OMe
THF, 76%
OMe
65%
15
Scheme 3.
16
Me
Me
Br
FeCl3
-30 oC, 67%
Br
Me
Me
+
O
OMe
C
O
O
12
17
Scheme 6.
18
13
Scheme 4.

M. Harmata et al. / Tetrahedron Letters 43 (2002) 2347–2349
2349
(27 mL). pTsOH (3 g, 5 mol%) was added and the
reaction mixture was refluxed overnight. The pale yellow
reaction mixture was cooled and diluted with H2O (50
mL) and extracted with CH2Cl2 (5×30 mL). The organic
layers were washed with 5% NaHCO3 (3×20 mL) and
dried over MgSO4. The concentrated crude product was
purified by flash chromatography (hexanes:EtOAc, 2:1)
to afford 2 (14.8 g, 44%) and 3 (10.2g, 23%).
7. Wang, Y.; Arif, A. M.; West, F. G. J. Am. Chem. Soc.
1999, 121, 876–877.
8. Sato, K.; Suzuki, S.; Kojima, Y. J. Org. Chem. 1967, 32,
339–341.
9. Compound 14 (2.0 g, 10.5 mmol) was dissolved in 100
mL of freshly distilled ether. A solution of diazomethane
(CAUTION) was slowly added over 3 h to the stirred
solution of 14. After each addition, the subsequent addi-
tion was delayed until the yellow color of diazomethane
dissipated. When the yellow color remained the solution
was allowed to stir for and additional 30 min and the
solvent was removed in vacuo, after the addition of a
small amount of acetic acid to remove remaining dia-
zomethane. The crude product was purified by flash
chromatography (hexanes:EtOAc, 8:1) to afford 1.68 g of
15 (79%). Starting material was also recovered (10%).
Scheme 7.
of 17 with LAH afforded not a simple reduction
product but one in which a quasi-Favorskii rearrange-
ment had taken place; and in nearly quantitative yield
(Scheme 7). The implications of this process for the
synthesis of angular triquinanes are obvious and are
being pursued.
1
Data on 15: H NMR (CDCl3, 500 MHz) l 3.75 (s, 3H),
In summary, we have shown two examples of an
intramolecular 4+3 cycloaddition reaction of cyclohex-
enyl cations which proceed with complete endo
2.92 (t, 2H, J=6.1), 2.54–2.51 (m, 2H), 2.04 (quintet, 2H,
J=6.3); 13C NMR (CDCl3, 125.8 MHz) l 192.1, 151.0,
135.2, 59.8, 38.3, 35.8, 22.6; IR (neat) 1686s, 1288s, 1204s
cm−1. Anal. calcd for C7H9BrO2: C, 41.00; H, 4.42.
Found: C, 40.87; H, 4.60.
diastereoselectivity.
The
straightforward
quasi-
Favorskii rearrangement of 17 as well as its facile
preparation suggest that good use of such halides can
be made in synthesis. Further progress in this area will
be reported in due course.12
10. Compound 16 (301 mg, 1 mmol) was dissolved in 10 mL
of freshly distilled dichloromethane and was cooled to
−50°C in a cold bath. After 30 min, Tf2O (233 mL, 1.4
equiv.) was added in 20 min. The reaction mixture
became pink after 2 h. The color slowly turned to deep
brown. After 11 h at −50°C, TLC showed 17 as the only
product. The reaction mixture was quenched with satu-
rated NaHCO3. The organic layer was extracted with
water and dried over MgSO4. After removal of solvent,
the residue was purified by flash chromatography (hex-
anes:EtOAc, 7:1) to afford 175 mg (65%) of 17. Data on
Acknowledgements
This work was supported by the National Science
Foundation to whom we are grateful. Preliminary
experiments by Mr. Asitha Abeywardane are gratefully
acknowledged. Support of the Elmer O. Schlemper
X-ray diffraction facility at the University of Missouri-
Columbia by the Department of Chemistry is gratefully
acknowledged.
1
17: mp 68°C; H NMR (CDCl3, 500 MHz) l 5.71 (dtd,
1H, J=2.0, 6.2, 10.8 Hz), 5.64 (ddd, 1H, J=1.4, 5.1,
10.8), 3.40 (ddd, 1H, J=1.0, 6.5, 16.2), 2.92 (dd, 1H,
J=5.57, 16.3), 2.85 (q, 1H, J=5.8), 2.69 (ddt, 1H, J=
2.70, 4.16, 13.5), 2.6 (ddd, 1H, J=13.3, 8.1, 5.3), 2.54 (td,
1H, J=4.5, 13.5), 2.37 (qt, 1H, J=4.3, 14.1), 1.98–1.88
(m, 2H), 1.78 (td, 1H, J=4.9, 13.7), 1.71–1.57 (m, 3H),
1.54 (septet, 1H, J=6.2), 1.20 (dt, 1H, J=7.9, 13.2); 13C
NMR (CDCl3, 125.8 MHz) l 204.5, 137.7, 126.0, 76.8,
62.0, 46.8, 46.1, 40.9, 40.4, 36.3, 33.3, 23.5, 21.9; IR
(neat) 1723s, cm−1. Anal. calcd for C13H17BrO: C, 58.01;
H, 6.37. Found: C, 57.89; H, 6.54.
References
1. (a) Harmata, M. Acc. Chem. Res. 2001, 34, 595–605; (b)
Harmata, M. Tetrahedron 1997, 53, 6235–6280.
2. Cha, J. K.; Oh, J. Curr. Org. Chem. 1998, 2, 217–232.
3. Harmata, M.; Rashatasakhon, P. Org. Lett. 2000, 2,
2913–2915.
4. Harmata, M.; Elomari, S.; Barnes, C. L. J. Am. Chem.
Soc. 1996, 118, 2860–2871.
5. For examples and leading references, see: (a) Harmata,
M.; Bohnert, G.; Barnes, C. L. Tetrahedron Lett. 2001,
42, 149–151; (b) Harmata, M.; Carter, K. W. Tetrahedron
Lett. 1997, 38, 7985–7988.
6. Preparation of 2: Compound 1 (30 g, 268 mmol) was
dissolved in a mixture of MeOH (52 mL) and HC(OMe)3
11. (a) Harmata, M.; Rashatasakhon, P. Org. Lett. 2001, 3,
2533–2535; (b) Harmata, M.; Rashatasakhon, P. Tetra-
hedron Lett. 2001, 42, 5593–5595; (c) Harmata, M.; Shao,
L. Synthesis 1999, 1534–1540; (d) Harmata, M.; Shao, L.;
Ku¨rti, L.; Abeywardane, A. Tetrahedron Lett. 1999, 40,
1075–1078.
12. All new compounds exhibited satisfactory 1H and 13C
NMR data as well as satisfactory combustion analysis or
high resolution exact mass data.