The lithium-ene cyclization and the purported carbanion-induced ene cyclization [16]

Full text of DOI:10.1021/ja992583m

J. Am. Chem. Soc. 1999, 121, 10241-10242
10241
Scheme 1
The Lithium-Ene Cyclization and the Purported
Carbanion-Induced Ene Cyclization
Dai Cheng, Shirong Zhu, Xiaojun Liu, Steven H. Norton, and
Theodore Cohen*,1
Scheme 2
Department of Chemistry, UniVersity of Pittsburgh
Pittsburgh, PennsylVania 15260
ReceiVed July 21, 1999
We provide evidence herein that the previously unstudied
lithium-ene cyclization (eq 1; Met ) Li) is more facile than the
Scheme 3
widely used Mg analogue;2-4 that the products, however, are
unstable, accepting protons from various donors and possibly
reverting to uncyclized organolithiums; that the only example of
such a cyclization cited in reviews of this reaction involved a
crucial misassignment of the structure of the product; and that
another subsequently discovered cyclization, thought to be a
carbanion-induced ene reaction, is actually a lithium-ene cy-
clization followed by an intramolecular proton transfer.
The “surprisingly smooth” lithium-ene cyclization cited in the
Oppolzer reviews2 refers to the addition of n-butyllithium to triene
1 and isolation of the products of electrophilic trapping of 3
(Scheme 1).5,6 However, as shown below, the actual trapped
intermediate was 4.
proximate cyclization products 10 via the 1,5-proton transfer
shown (Scheme 2).9
Indeed, the alkyllithium trans-10 (n ) 1) rearranged in good
yield to trans-11 (n ) 1) at ambient (Scheme 3).
After completion of this work, we became aware of a similar
cyclization to an allyllithium; the report10 (see below) did not
cite the cyclization in Scheme 1 and was not referenced in the
reviews of metallo-ene cyclizations,2 possibly because a different
mechanism was invoked. The question then arose as to why 2
cyclized to 3 rather than to the allyllithium 4. In fact, a critical
reading of the paper5 revealed that the reported 60 MHz NMR
spectrum of the product obtained upon quenching with paraform-
aldehyde was inadequate to distinguish between attack of
formaldehyde on 3 or 4. Our repetition of the procedure,5 however
quenching with D2O, deliVered the deuteration product not of 3
but of 4.
As to our failure to observe cyclized product 8 from 6 (Scheme
2), the low yield of recovered product and the methallyllithium
results made us suspect that cyclization had occurred but that the
cyclized alkyllithium 7 removed a proton from THF11 more
rapidly than the less basic allyllithiums such as 11 and that the
resulting volatile hydrocarbon was lost upon workup. This
hypothesis was tested by reductive lithiation of 15 in THF. The
cyclized organolithium, which was not trapped by added paraform-
aldehyde, removed a proton from solvent to deliver an excellent
yield of 16. The important conclusion can be drawn that the
lithium-ene cyclization occurs at a considerably lower temper-
ature than the magnesium analogue2 but that because of the
greater basicity of organolithium than organomagnesium com-
pounds, the cyclized organolithium abstracts a proton readily,
either intramolecularly to produce an allyllithium if an allylic
methyl group is nearby or intermolecularly from THF.
It appears likely that this type of facile cyclization, possibly
followed by 1,5-allylic proton transfer as demonstrated above, is
Our suspicions concerning the reported5 structure (3) of the
cyclic organolithium arose when we were unable to detect the
thioether derivative 8 of the lithium-ene cyclization product 7
of 6 (readily prepared by reductive lithiation7 of 5 using lithium
4,4′-di-tert-butylbiphenylide (LDBB)) when 6 was warmed from
-78 to 20 °C and the reaction mixture was treated with bis(p-
methoxyphenyl) disulfide. The acyclic allyllithium 6 could be
recovered as its thioether derivative in 40% yield (Scheme 2).
On the other hand, when a methyl group was present at the
2-position of the allyllithium 9, as was the case in Scheme 1,8
cyclization to five- and six-membered rings occurred partially
during 1 h at 0 °C and completely at ambient temperature.
However, the cyclic thioethers, isolated in yields of 77% and 64%,
respectively, for n ) 1 and 2, were not those from electrophilic
capture of 10, as expected from ref 5, but those from allyllithiums
11. It seemed likely that these products were formed from the
(1) Taken in part from the M.S. Theses of Shirong Zhu (1997) and Xiaojun
Liu (1999) and the Ph.D. Thesis of Steven H. Norton (1999) at the University
of Pittsburgh.
(2) Reviews of metallo-ene reactions: Oppolzer, W. Angew. Chem., Int.
Ed. Engl. 1989, 28, 38-52. Oppolzer, W. In ComprehensiVe Organic
Synthesis; Trost, B. M., Ed.; Pergamon Press: Oxford, England, 1991; Vol.
5, pp 29-61.
(3) The Zn-ene cyclization, used less extensively in synthesis, has recently
come into prominence. Recent references: Nakamura, E.; Kubota, K. J. Org.
Chem. 1997, 62, 792-793. Lorthiois, E.; Marek, I.; Normant, J.-F. J. Org.
Chem. 1998, 63, 2442-50.
(4) Cyclization of nonallylic unsaturated organolithiums: Bailey, W. F.;
Ovaska, T. V. In AdVances in Detailed Reaction Mechanisms; Coxon, J. M.,
Ed.; JAI Press: Greenwich, CT, 1994; Vol. 3, pp 251-273.
(5) Josey, A. D. J. Org. Chem. 1974, 39, 139-145.
(9) A reviewer observed that the mainly cis configuration of 11 supports
the metallo-ene cyclization since a trans configuration is generated when a
nonallylic secondary organolithium cyclizes in a similar fashion: Bailey, W.
F.; Nurmi, T. T.; Patricia, J. J.; Wang, W. J. Am. Chem. Soc. 1987, 109,
2442-48.
(10) Edwards, J. H.; McQuillin, F. J. J. Chem. Soc., Chem. Commun. 1977,
838-839. See also: Cuvigny, T.; Julia, M.; Rolando, C. J. Organomet. Chem.
1988, 344, 9-28.
(11) Bates, R. B.; Kroposki, L. M.; Potter, D. E. J. Org. Chem. 1972, 37,
560-562.
(6) Two lithium-ene cyclizations, not included in the reviews,2 may be
special cases: Klumpp, G. W.; Schmitz, R. F. Tetrahedron Lett. 1974, 2911-
14 (strained enophile). Krief, A.; Derouane, D.; Dumont, W. Synlett 1992,
907-908 (1,1,3,3-tetrasubstituted allyllithium).
(7) Cohen, T.; Bhupathy, M. Acc. Chem. Res. 1989, 22, 152-161.
(8) Allyllithiums are depicted here as covalent (η1) only for simplicity and
analogy with allylmagnesiums. They are known to be π (η3) complexes.
Fraenkel, G.; Cabral, J. A. J. Am. Chem. Soc. 1993, 115, 1551-1557.
10.1021/ja992583m CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/13/1999

10242 J. Am. Chem. Soc., Vol. 121, No. 43, 1999
Communications to the Editor
Scheme 4
∆Eq for the lithium-ene cyclization 6 f 7 (19.4-25.8 kcal/mol,18
depending on stereochemistry) is slightly higher than that for the
intermolecular version in Scheme 6, probably partly due to the
enophile being substituted. When continuum models of solvation,
included in Gaussian, were applied to the transition states,
reactants, and products for both the Li-ene and CI-ene reactions,
the calculated energy difference between the two processes in
Scheme 6 was even greater. Self-Consistent Reaction Field
(SCRF) models used included the Polarized Continuum Model1919
(PCM) and Isodensity Surface PCM20 (IPCM) models. It is thus
likely that this process is actually a lithium-ene cyclization as
originally surmised.
The original putative result which is quoted in the reviews
indicates not only that the lithium-ene cyclization occurs at -25
°C, which it does not (the original paper5 is less than clear on
this point), but that the cyclization is thermodynamically favorable.
Now that we know that the product structure was misassigned,
the possibility exists that this is not so and that the ring closure
is driven to completion by the subsequent intramolecular or
intermolecular proton transfer. To test this reversibility, 7 was
prepared by reductive lithiation of the corresponding phenyl
thioether and an attempt was made to trap it or its retrocyclization
product with paraformaldehyde. Apparently, only protonated
product was generated since no trapped product could be isolated
unless paraformaldeyde was added prior to warming to 20 °C, in
which case the cyclic alcohol was isolated in excellent yield.
We again turned to computations. Although the intermolecular
Li-ene process shown in Scheme 6 was found to have ∆E )
-3.1 kcal/mol, the intramolecular process 6 f 7 was found to
be endergonic (∆E ) +11.2 to +11.9 kcal/mol, depending on
stereochemistry; the six-membered ring formation is only slightly
endergonic). It thus appears that the incorrect assignment of 3 to
the product of cyclization of 2 led to considerably incorrect
conclusions about the lithium-ene cyclization. While it is a
process that occurs more readily than the far more common
magnesium analogue, unlike the latter it is probably thermody-
namically unfavorable and observed only when the cyclization
product undergoes a subsequent irreversible reaction.21 This
concept has been utilized synthetically in a remarkably efficient,
stereoselective, tandem cyclization.22 The lack of an isotope effect
in 9 f 11 is consistent with the proton transfer being far more
rapid than the retrocyclization, i.e., the ring closure is the rate
determining step. This conclusion is also suggested by Scheme
3. If ring opening had preceded proton transfer, the resulting
product would be cis as in Scheme 2.23
Scheme 5
Scheme 6
responsible for the previously inadequately explained high degree
of carbanionic cyclopolymerization of butadiene under certain
conditions.12
For the cyclization of a terminally unsaturated methallyllithium
to an allyllithium analogous to 4 and 11, Edwards and McQuillin10
suggested a very interesting carbanion-induced ene reaction
(Scheme 5), which differs fundamentally from our suggestion of
a lithium-ene cyclization followed by a 1,5-proton transfer. In
their concept, one conjugated organolithium is converted directly
to another, no nonconjugated alkyllithium intervening. Whereas
the Alder ene reaction usually occurs at temperatures above 200
°C,13 it was suggested that the carbanionic site greatly facilitates
the reaction.
We examined the two mechanisms both by determining the
deuterium isotope effect for the reaction 9 f 11 (Scheme 2),
where 9 contains CH3 or CD3, and by ab initio computations.
The absence of a primary isotope effect would essentially disprove
the mechanism in Scheme 5 as well as that involving a reversible
cyclization followed by a rate-determining proton transfer.14 An
isotope effect would be revealed if allowing 9 (n ) 1; R ) CH3
or CD3) to cyclize for a given length of time resulted in
significantly more cyclization when R ) CH3 than when it is
CD3. In the event, allowing each cyclization to occur at 0 °C for
3 h led to an 86/14 ratio of cyclized to uncyclized thioether when
R ) CH3 and a ratio of 81/19 when R ) CD3; this leads to an
approximate isotope effect of 1.09.15 An independent experiment
involving about 90% cyclization of an equimolar mixture of
labeled and unlabeled 9 and determination of the deuterium
content of uncyclized 9 by 13C NMR revealed no isotope effect
within experimental error.15
Acknowledgment. We thank the National Science Foundation for
financial support and Professors Fu-Tyan Lin and Kenneth Jordan
(University of Pittsburgh) and William Saunders (University of Rochester)
for valuable advice.
The computed activation energy of the Li-ene reaction of an
allyllithium/ethylene complex was compared with the carbanion-
induced ene (CI-ene) reaction of a methallyllithium/ethylene
complex (Scheme 6), using the Gaussian program.16,17 The results
indicate that the lithium-ene addition has a considerably lower
activation energy in the gas phase than the CI-ene process. The
Supporting Information Available: Experimental procedures, details
of compound characterization, and computational details (PDF). This
material is available free of charge via the Internet at http://pubs.acs.org.
JA992583M
(12) McElroy, B. J.; Merkley, J. H., U.S. Patent 3,678,121, July 18, 1972.
Halasa, A. F., U.S. Patent 3,966,691, June 29, 1976. Quack, G.; Fetters, L. J.
Macromolecules 1978, 11, 369-373.
(13) Snider, B. B. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Paquette, L. A., Eds.; Pergamon: Oxford, 1991; Vol. 5, pp 1-25.
(14) Melander, L.; Saunders, W. H., Jr. Reaction Rates of Isotopic
Molecules; John Wiley & Sons: New York, 1980; Chapter 5.
(15) See Supporting Information for details.
(16) Gaussian 94, Revision B.2, Frisch, M. J.; Trucks, G. W.; Schlegel, H.
B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.;
Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.;
Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B.
B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A.; Gaussian, Inc.: Pittsburgh, PA, 1995.
(17) Full computations concerning these and related reactions will be
published separately.
(18) This and subsequent energy values are from Becke3-LYP/6-31G*
calculations.
(19) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117-129.
Miertus, S.; Tomasi, J. Chem. Phys. 1982, 65, 239-245. Cossi, M.; Barone,
V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett. 1996, 255, 327-335.
(20) Foresman, J. B.; Keith, T. A.; Wiberg, K. B.; Snoonian, J.; Frisch,
M. J. J. Phys. Chem. 1996, 100, 16098-16104.
(21) The energy of the intramolecular proton transfer can be approximated
by that for the transfer of an allylic proton from propene to ethyllithium, -25.7
kcal/mol at the same level of theory. Thus, ∆E = 11.5-25.7 ) -14.2 kcal/
mol for the cyclization + proton transfer.17
(22) Cheng, D.; Knox, K. R.; Cohen, T. Manuscript in preparation.
(23) In a recent report, an intramolecular displacement of an allylic chloride
by an allyllithium is represented as starting as a lithium-ene cyclization and
merging into an SN′ reaction. Dieters, A.; Hoppe, D. Angew. Chem., Int. Ed.
Engl. 1999, 38, 546-548.