Double annulation route to fused bicyclic compounds with three contiguous quaternary centers

Article abstract of DOI:10.1021/ol990935o

(matrix presented). Fused bicyclic and tricyclic compounds featuring two or three new C-C σ bonds, two new rings, two or three new stereocenters, and three new contiguous quaternary centers can be prepared stereoselectively, atom-economically, and in one

Full text of DOI:10.1021/ol990935o

ORGANIC
LETTERS
1999
Vol. 1, No. 10
1583-1586
Double Annulation Route to Fused
Bicyclic Compounds with Three
Contiguous Quaternary Centers
Robert B. Grossman,* Aaron J. Skaggs, Arlene E. Kray, and Brian O. Patrick
Department of Chemistry, UniVersity of Kentucky, Lexington, Kentucky 40506-0055
rbgros1@pop.uky.edu
Received August 11, 1999
ABSTRACT
Fused bicyclic and tricyclic compounds featuring two or three new CsC σ bonds, two new rings, two or three new stereocenters, and three
new contiguous quaternary centers can be prepared stereoselectively, atom-economically, and in one synthetic operation via a cascade reaction
of two nonstereogenic, readily available starting materials: a tethered bis(malononitrile) and an internal alkynone. The course of the “double
annulation” changes drastically with the structure of the starting materials, alternately affording trans-decalins or cis-fused unsaturated lactones.
The efficient and stereoselective introduction of quaternary
centers into organic compounds represents a continuing
challenge to the synthetic organic chemist.1-3 The construc-
tion of three contiguous quaternary centers bearing only C
substituents is usually seriously impeded by steric encum-
brance, and methods for preparing such architectures in one
step are few and far between. The few methods reported so
far usually use cycloadditions such as the cycloaddition of
tetracyanoethylene with terminally disubstituted 1,3-dienes4-6
or electron-rich alkenes,7,8 the dimerization of tetrasubstituted
cyclopentadienones,9 and intramolecular [2 + 2] photocyclo-
additions,10,11 although double Michael reactions involving
heavily substituted cyclohexenone enolates and either iso-
propylidenemalononitrile12 or methyl R-chloro-R-cyclopro-
pylideneacetate13 have also been used, and Wender has
recently reported a spectacular arene-alkene meta-photo-
cycloaddition in which four new contiguous quaternary
centers are produced in a single step.14 The fact that to our
knowledge only two reactions other than cyclodditions (and
only a handful of cycloadditions, for that matter) have
previously been reported to make three new contiguous
quaternary centers is testament to the difficulty of this
transformation. We now report a novel “double annulation”
reaction that affords bicyclic products featuring three new σ
bonds (two or all three of which are CsC bonds), two new
rings, two or three new stereocenters, and three new
contiguous quaternary centers from two nonstereogenic,
readily available starting materials with excellent stereose-
lectivity and atom economy. We also report that slight
variations in the starting materials produce drastic changes
in the course of the double annulation reaction.
(1) Martin, S. F. Tetrahedron 1980, 36, 419.
(2) Fuji, K. Chem. ReV. 1993, 93, 2037.
(3) Corey, E. J.; Guzman-Perez, A. Angew. Chem., Int. Ed. 1998, 37,
388.
(4) Liebe, J.; Wolff, C.; Krieger, C.; Weiss, J.; Tochtermann, W. Chem.
Ber. 1985, 118, 4144.
(5) Brel’, V. K.; Abramkin, E. V.; Martynov, I. V. Bull. Acad. Sci. USSR,
DiV. Chem. Sci. (Engl. Transl.) 1989, 38, 2607.
(6) Thiemann, T.; Kohlstruck, S.; Schwaer, G.; de Meijere, A. Tetra-
hedron Lett. 1991, 32, 3483.
(7) Shipman, M.; Ross, T. M.; Slawin, A. M. Z. Tetrahedron Lett. 1999,
40, 6091.
(8) Cookson, R. C.; Halton, B.; Stevens, I. D. R.; Watts, C. T. J. Chem.
Soc. (C) 1967, 928.
(11) Winkler, J. D.; Doherty, E. M. J. Am. Chem. Soc. 1999, 121, 7425.
(12) Holton, R. A.; Williams, A. D.; Kennedy, R. M. J. Org. Chem.
1986, 51, 5480.
(9) Fuchs, B.; Pasternak, M.; Pazhenchevsky, B. Tetrahedron 1980, 36,
3443 and references therein.
(10) Kubo, Y.; Toda, R.; Araki, T. Bull. Chem. Soc. Jpn. 1987, 60, 429.
(13) Hadjiarapoglou, L.; Klein, I.; Spitzner, D.; de Meijere, A. Synthesis
1996, 525.
(14) Wender, P. A.; Dore, T. M. Tetrahedron Lett. 1998, 39, 8589.
10.1021/ol990935o CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/10/1999

We recently disclosed that nucleophiles containing two
tethered carbon acids (1) underwent two sequential Michael
reactions to 3-butyn-2-one to give functionalized cycloal-
kanes (2) containing two new quaternary centers (Scheme
1).15-17 The reactions proceeded when at least one of the
Scheme 2
Scheme 1
determined by X-ray crystallography. The reaction of 1d with
4-hexyn-3-one proceeds similarly to afford trans-decalin 3b
as a single stereoisomer in 49% yield.20 These remarkably
efficient reactions create three new CsC σ bonds, two new
rings, two or three new stereocenters, and three new
contiguous quaternary centers in one step from two non-
stereogenic, readily available starting materials with excellent
stereoselectivity and complete atom economy, albeit in
moderate yield.21
The double annulation takes a very different course when
the electrophile has an Et group attached to the triple bond.
When 1c is combined with 3-hexyn-2-one in THF with
catalytic NaH, the double Michael reaction is followed by
attack of the nascent enolate oxygen on a CN group to give
cis-hexahydroisochromanone 4a as a single stereoisomer in
79% yield after aqueous acidic workup (Scheme 3). The
groups attached to the acidic carbons was a CN group.
Compound 2 could then be induced to undergo a Dieckmann
reaction to give a trans-decalin (X ) CH2) or a trans-
hydrindan (X ) nothing).
Although dicyano diesters 1a and 1b smoothly underwent
double Michael reactions with the terminal alkynone 3-butyn-
2-one, multiple attempts to promote their double Michael
reaction with internal alkynones failed to provide any
identifiable cyclic products. In our earlier work we had taken
advantage of the very low steric bulk of the cyano group,15
and so we decided to see whether tethered carbon acids
containing only cyano groups were sufficiently unhindered
to react with internal alkynones in a double Michael reaction.
The requisite tetranitriles 1c-e (Scheme 1) were easily
prepared in good yield using the protecting group methodol-
ogy we have reported;18 that is, (3-pentylidene)malononitrile
was alkylated with an R,ω-dibromide, and the product was
deblocked by ozonolysis and reflux in acidic EtOH. Unfor-
tunately, we were unable also to prepare tetranitrile 1f in
this way, as it rapidly underwent a Thorpe-Ziegler reaction
under the conditions of the deblocking. Both 1c and 1e
smoothly underwent the double Michael reaction with
terminal alkynone 3-butyn-2-one to give monocyclic products
2c and 2e in 67% and 66% yields, respectively (Scheme 1).
Scheme 3
structure of 4a has been established by X-ray crystallography.
The reaction of 1d with 3-hexyn-2-one proceeds similarly
to afford 4b as a single stereoisomer in 29% yield, and 1c
also reacts stereoselectively with 4-heptyn-3-one to give 4c
in 45% yield.21 Either complex mixtures are formed or no
reaction occurs when 1c is combined with the more sterically
hindered electrophiles 4-phenyl- and 4-trimethylsilyl-3-butyn-
2-one.
When 1c is combined with internal alkynone 4-hexyn-3-
one in THF with 1 equiv of NaH, not only does the double
Michael reaction proceed smoothly, but it is followed
immediately by a Thorpe-Ziegler reaction19 to give trans-
decalin 3a as a single stereoisomer in 48% isolated yield
(Scheme 2). The stereochemistry at the ring junction has been
(20) The crystal structure of 3a reveals NsH‚‚‚OdC hydrogen bonds
(1.99 and 2.15 Å) involving both O lone pairs and both NH hydrogens,
and its IR spectrum (KBr) shows two strong, broad absorbances at 3353
and 3195 cm-1. By contrast, the IR spectrum of 3b shows five strong, sharp
absorbances at 3440, 3411, 3350, 3275, and 3242 cm-1, perhaps indicating
a lesser degree of hydrogen bonding in this compound. The carbonyl regions
of the IR spectra of both of these compounds are also unusual: 3a shows
a strong absorbance at 1551 cm-1 and a weaker one at 1666 cm-1, whereas
3b shows a strong absorbance at 1548 cm-1 and weaker ones at 1654, 1623,
(15) Grossman, R. B.; Varner, M. A.; Skaggs, A. J. J. Org. Chem. 1999,
64, 340.
(16) Grossman, R. B.; Rasne, R. M.; Patrick, B. O. J. Org. Chem. 1999,
64, 7173.
and 1599 cm-1
.
(17) Grossman, R. B.; Pendharkar, D. S.; Patrick, B. O. J. Org. Chem.
1999, 64, 7178.
(18) Grossman, R. B.; Varner, M. A. J. Org. Chem. 1997, 62, 5235.
(19) Schaefer, J. P.; Bloomfield, J. J. Org. React. (N. Y.) 1967, 15, 1.
(21) Although the yields are low or moderate in some cases, in no case
does GC-MS analysis of the crude reaction mixture show any evidence for
products with molecular weights identical or close to those of the isolated
products.
1584
Org. Lett., Vol. 1, No. 10, 1999

By contrast with the behavior of 1c,d, combination of 1e
with either 4-hexyn-3-one or 3-hexyn-2-one affords only the
corresponding cis-lactones 5a,b in the indicated yields
(Scheme 4). The stereochemistry of the ring junction of 5b
has been determined by X-ray crystallography.22,23
closely,24 and 6-Req is calculated to be about 2-12 kcal/mol
lower in energy than 6-Rax (when R ) Et and R′ ) H; MM2,
MOPAC), so the rate of formation of 6-Req is probably much
greater than that for 6-Rax.25 When R ) Et, the steric
compression in 6-Req is so great that it lactonizes irreversibly
before it has an opportunity to flip to 6-Rax or to tautomerize,
and thus cis-4 is formed exclusively. On the other hand, when
R ) Me, the smaller steric compression leads to a much
slower or a reversible lactonization, an equilibrium is
established between 6-Req, 6-Rax, and their tautomeric
enolates, and the exclusive formation of trans-3 reflects the
lower energy of the TS leading to this thermodynamically
preferred isomer (Curtin-Hammett principle). Finally, when
R ) H, the steric compression is so small that a second
cyclization does not occur at all, and 2 is obtained. In support
of this hypothesis, when 1d is combined with 4-hexyn-3-
one (R ) Me) and the reaction mixture is quenched after 2
h, some isochromanone can be detected in the GC-MS of
the crude reaction mixture (m/z ) 283), but when the reaction
mixture is allowed to stir overnight before quenching, only
decalin is detected (m/z ) 282).26 A similar analysis can be
carried out with 1e, except here the greater steric compression
attendant upon formation of the seven-membered ring causes
the lactonization to occur rapidly and irreversibly even when
R ) Me.
Scheme 4
Why does the course of the double annulation change so
drastically with small changes in the reagents? A plausible
explanation can be formulated upon examination of the
nascent double Michael adduct, 6, that is produced upon
combination of 1c and an internal alkynone (Scheme 5).
We must comment on the experimental simplicity of the
double annulation reactions. The nucleophile 1 is dissolved
in THF at 0 °C, NaH (1.2 equiv for 3, 0.2 equiv for 4 or 5)
is added, and after 15 min, a solution of the electrophile in
THF is added via addition funnel. When the reaction is
judged to be complete (TLC or GC, usually a few hours),
the reaction mixture is diluted with 1 N HCl, the organic
solvent is evaporated, and the aqueous suspension is diluted
with EtOH. Less soluble products are isolated in pure form
simply by filtration, and more soluble products are isolated
by extraction and recrystallization.
Scheme 5
In conclusion, we have discovered an efficient, stereose-
lective route to bi- and tricyclic compounds featuring three
new contiguous quaternary centers. Experiments that expand
the scope of the double annulation further will be reported
in due course.
Acknowledgment. The support of this work by the
National Science Foundation (CAREER award to R.B.G.,
Grant CHE-9733201, and REU program, Grant CHE-
(22) The large R value was due to rather large thermal motion, but nothing
was found which would have cast doubt on the stereochemistry.
(23) A subsequent X-ray study of this compound revealed the possibility
of a second crystal form. Details will be presented in a future publication.
Brock, C. P.; Patrick, B. O.; Grossman, R. B.; Skaggs, A. J. Unpublished
results.
(24) The Michael reaction in question converts a very stabilized enolate
[pKa(conj. acid) ≈ 11] into a moderately stabilized one [pKa(conj. acid) ≈
20] . Such an elementary step is not expected to be very exothermic, and
hence the Hammond postulate suggests that the TS should resemble the
product.
(25) The difference in energy between 6-Eteq and 6-Etax seems to be
due to dipolar interactions between the axial CN groups and the enolates,
not steric interactions.
(26) However, the imidate precursors to lactones 4 could not be
equilibrated to the corresponding trans-decalins, even after the reaction
mixtures were allowed to stir with excess base for extended periods of time.
Enolate 6 can exist as two interconvertible chair conformers
(6-Req and 6-Rax). The transition states for the formation of
6-Req and 6-Rax should resemble these enolates rather
Org. Lett., Vol. 1, No. 10, 1999
1585

9531406) and the University of Kentucky is gratefully
acknowledged. The authors thank Professor Carolyn Brock
for preliminary work on the X-ray structure of 4a.
5a,b, and 2-amino-2-cyclopentene-1,1,3-tricarbonitrile, and
X-ray data for 3a, 4a, and 5b. This material is available free
of charge via the Internet at http://pubs.acs.org.
Supporting Information Available: Experimental pro-
cedures and full characterization of 1c-e, 2c,e, 3a,b, 4a-c,
OL990935O
1586
Org. Lett., Vol. 1, No. 10, 1999