The spirocyclopentadienyl derivative 2e (Table 1, entry
group by telluride ion to give the monoalkyltelluride ion 8
(which in the case of 2f leads to telluride 7). Then, 8
undergoes intramolecular nucleophilic attack on the cyclo-
propane ring to give the unstable epitelluride intermediate
9, which readily eliminates elemental tellurium to form
enolate 10, which is protonated during workup (Scheme 3).
5) underwent ring opening to give a mixture of 1- and 2-allyl-
1
1,3-cyclopentadienes 5a,b.26 The 600 MHz H NMR spec-
trum of the mixture allowed sufficient separation of the
chemical shifts of the two isomers for determination of the
isomer ratio of about 60:40 in favor of 2-allylcyclopenta-
diene.26 The facile ring-opening of the spirocyclopentadiene
is attributed to the stability of the cyclopentadienyl anion.
The reaction of (2-phenylcyclopropyl)methanol tosylate 2f
with telluride ion is different, yielding mainly unstable
telluride 7 (84%) with perhaps a trace of the ring-opened
product, 3-butenylbenzene 6 (Table 1, entry 6). The forma-
tion of a relatively less stable benzyl anion apparently is not
favored.
Scheme 3. Mechanism for the Telluride-Triggered
Nucleophilic Ring Opening of Cyclopropanes
The use of other reducing agents for tellurium gave lower
yields of 3a and often required longer reaction times: sodium
borohydride-water-benzene phase transfer catalyst (PTC),20b
2 h (17% plus isomers from migration of the double bond);
rongalite-NaOH-water-benzene PTC,19d 3 h (15-17%
plus products of double-bond migration); sodium hydride-
DMF,27 12 h (20%, complicated by reaction of Na2Te with
DMF27,28); sodium naphthalenide-THF,29 20 min (44%). We
have previously reported that the method of reduction of
tellurium could alter the outcome of some tellurium-triggered
reactions.19c Catalytic amounts (10 mol %) of tellurium with
stoichiometric LiEt3BH may be used, but the reaction is
slower (12 h) and the yield of 3a is lower (50%). The
presence of the Lewis acids, lithium ion, and triethylborane
(a byproduct in the reduction of tellurium) apparently is
helpful since the reaction seems to be somewhat inhibited
by the Lewis base, fluoride ion. The reaction of 2a with
lithium telluride in the presence of tetrabutylammonium
fluoride was slower and gave a lower yield of 3a (60% for
15 h vs 90% for 25 min, Table 1, entry 1). This result also
may be caused by some depletion of the telluride ion by
reaction with the quaternary salt.
The enolate from 2a can be trapped by reaction with
benzaldehyde to give syn- and anti-aldol products 11a14 and
11b,30 respectively, in 75-91% yield. Diastereoselectivity
of this aldol reaction varies depending on the temperature
and reaction time. The highly diastereoselective formation
of the syn-aldol 11a is kinetically controlled and observed
at low temperature, whereas the thermodynamically favored
anti-aldol 11b, resulting from a slow equilibration through
a retroaldol reaction, is formed at higher temperature, also
with a high degree of diastereoselectivity (Scheme 4). This
Scheme 4. Aldol Reaction of the Enolate Generated from 2a
with Telluride Ion
The reaction pathway for the nucleophilic ring opening
may be rationalized by initial substitution of the tosylate
(24) (a) Bergson, G. Acta Chem. Scand. 1957, 11, 571-572. (b) Engman,
L.; Cava, M. P. J. Org. Chem. 1982, 47, 3946-3949. (c) Clive, D. L. Y.;
Beaulieu, P. L. J. Org. Chem. 1982, 47, 1124-1126. (d) Osuka, A.; Suzuki,
H. Chem. Lett. 1983, 119-120. (e) Engman, L. Organometallics 1986, 5,
427-437. (f) Suzuki, H.; Inouye, M. Chem. Lett. 1986, 403-406. (g) Huang,
Z.; Xia, L.; Huang, X. Synth. Commun. 1988, 18, 1167-1170. (h) Matsuki,
T.; Hu, N. X.; Aso, Y.; Otsubo, T.; Ogura, F. Bull. Chem. Soc. Jpn. 1989,
62, 2105-2107. (i) Padmanabhan, S.; Ogawa, T.; Suzuki, H. Bull. Chem.
Soc. Jpn. 1989, 62, 2114-2116. (j) Li, C. J.; Harpp, D. N. Tetrahedron
Lett. 1990, 31, 6291-6294. (k) Huang, Z.-Z.; Zhou, X.-J. Synthesis 1990,
633-634. (l) Zhou, Z.-L.; Shi, L.-L.; Huang, Y.-Z. Synth. Commun. 1991,
21, 1027-1037. (m) Vasil’ev, A. A.; Engman, L. J. Org. Chem. 1998, 63,
3911-3917.
behavior closely resembles the aldol reaction of tert-butyl
n-butyl ketone with benzaldehyde reported by Heathcock and
Lampe.31 The aldol products 11a,b also have been obtained
by treatment of trimethylsilylmethylcyclopropyl tert-butyl
ketone and benzaldehyde with titanium tetrachloride at -78
°C (70%, anti/syn ) 1:11), with BF3‚Et2O (34%, anti/syn )
1.7:1), and with SnCl4 (40%, anti/syn ) 1:2).14
(25) For the use of the term “cyclopropanalogue”, see: Duffy, E. P. Ph.D.
Thesis, Tufts University, Boston, MA, 1987; Diss. Abstr. Int. B 1988, 48,
3568.
(26) 5-Allyl-1,3-cyclopentadiene was not detected. In addition, 1- and
5-allylcyclopentadienes are higher in energy than the 2-allylcyclopentadiene
by 0.545 and 3.859 kcal mol-1, respectively, according to the ab initio
calculations. A 60:40 ratio corresponds to a difference in energy of 0.240
kcal mol-1 at 298 K. The calculated energy difference of 0.545 kcal mol-1
is expected to afford a 72:28 mixture of isomers. Thermal 1,5-sigmatropic
rearrangements in the cyclopentadiene system can interconvert isomers. Our
reaction was carried out at room temperature. See Supporting Information.
(27) Suzuki, H.; Manabe, H.; Inouye, M. Chem. Lett. 1985, 1671-1674.
(28) Zingaro, R. A.; Herrera, C.; Meyers, E. A. J. Organomet. Chem.
1986, 306, C36-C40.
No reaction is observed under the conditions shown in
Scheme 2 if the acylcyclopropanemethanol 1b is used instead
(30) Anti-configuration of 11b was established on the basis of literature
analogies and by comparison with the known 11a:14 (a) Heathcock, C. H.;
Pirrung, M. C.; Sohn, J. E. J. Org. Chem. 1979, 44, 4294-4299. (b)
Ohtsuka, Y.; Koyasu, K.; Ikeno, T.; Yamada, T. Org. Lett. 2001, 3, 2543-
2546.
(29) Higa, K. T.; Harris, D. C. Organometallics 1989, 8, 1674-1678.
(31) Heathcock, C. H.; Lampe, J. J. Org. Chem. 1983, 48, 4330-4337.
Org. Lett., Vol. 6, No. 13, 2004
2227