Steric and electronic effects on the Weiss reaction. Isolation of 1:1 adducts

Article abstract of DOI:10.1039/a701232b

The mechanism of the Weiss reaction has been studied with respect to the intermediacy of 4-hydroxycyclopent-2-en-1-ones (1:1 adducts) in this process. Analysis of these experiments provides additional evidence that 4-hydroxycyclopentenones are indeed key intermediates in the Weiss reaction. Based on the reaction of dimethyl 3-oxoglutarate with benzil, pyridil, thenil, furil and phenanthrenequinone, steric effects play the major role in the overall success of this condensation to provide substituted cis-bicyclo[3.3.0]octane-3,7-diones. Moreover a trihydroxyindene [5.6] system (see 26) has been isolated for the first time under the Weiss conditions which provides additional support for the existence of cyclopentenone intermediates in this process.

Full text of DOI:10.1039/a701232b

View Article Online / Journal Homepage / Table of Contents for this issue
Steric and electronic effects on the Weiss reaction. Isolation of 1:1
adducts
Scott G. Van Ornum, Jin Li, Greg G. Kubiak and James M. Cook*
Department of Chemistry, University of Wisconsin-Milwaukee, Milwaukee, WI 53201, USA
The mechanism of the Weiss reaction has been studied with respect to the intermediacy of
4-hydroxycyclopent-2-en-1-ones (1:1 adducts) in this process. Analysis of these experiments
provides additional evidence that 4-hydroxycyclopentenones are indeed key intermediates in the
Weiss reaction. Based on the reaction of dimethyl 3-oxoglutarate with benzil, pyridil, thenil, furil and
phenanthrenequinone, steric effects play the major role in the overall success of this condensation to
provide substituted cis-bicyclo[3.3.0]octane-3,7-diones. Moreover a trihydroxyindene [5.6] system (see 26)
has been isolated for the first time under the Weiss conditions which provides additional support for the
existence of cyclopentenone intermediates in this process.
MeO2C
CO2Me
O
Introduction
R1
R2
O
The majority of synthetic approaches to polyquinanes have
been directed toward the preparation of specific natural prod-
ucts or of ‘non-natural’ compounds of computational import-
ance. For a synthetic process or method to be termed general, it
must provide a common route capable of extension to a variety
of compounds of diverse structural type. The cis-bicyclo-
[3.3.0]octane unit is the basic component of many polyquinane
natural and non-natural products. At least one retrosynthetic
pathway will always be present in this series of compounds
which will terminate with the cis-bicyclo[3.3.0]octane system.
The Weiss reaction has been shown to occur in a general fash-
ion for a variety of 1,2-dicarbonyl compounds; 1,2-diketones,
α-keto aldehydes, and glyoxal all react with dimethyl 3-
oxoglutarate to yield the corresponding cis-bicyclo[3.3.0]octane
tetraesters (Scheme 1). In this regard Bertz reported an
approach to synthetic analysis1 which employs graph and
information theory in order to evaluate alternative synthetic
routes toward a common target. Analysis of the Weiss reaction
using this approach indicates that it ranks higher than the
Diels–Alder reaction for the rapid generation of molecular
complexity in a single step. Posner has referred to the Weiss
process as a three component (2 ϩ 3 ϩ 3) coupling reaction and
points out that an overall yield of 90% in this condensation can
be viewed as an average yield of 97.5% for each of the new
carbon–carbon bonds so formed.2 The ability to produce a
variety of monosubstituted and 1,5-disubstituted cis-bicyclo-
[3.3.0]octane-3,7-diones by simply varying the substituents on
the starting 1,2-dicarbonyl compound is a powerful feature of
this process.
The versatility of the Weiss reaction has been demonstrated
by the synthesis of many different cyclopentanoid systems
including those of ellacene,3 triquinacene,4 modhephene5 and
gymnomitrol.6 The mild conditions (room temp., pH 5.6–8.3)
under which this condensation can be carried out should permit
tolerance of a variety of functional groups (e.g. amides, halides,
acetals, esters) to be manipulated later as desired. The interest
in the reaction sequence 1 ϩ 2 → 6 (Scheme 1) for the
synthesis of polyquinane ring systems and natural products
has, therefore, prompted a study of the influence of electronic
and steric factors on the success of this condensation. The
preparation of a number of 1,2-diones 2 in which the electronic
and steric character of R1 and R2 have been varied is described
below and is followed by an evaluation of their reactivity
with 1.
O
+
+
O
MeO2C
CO2Me
1
2 R1 = R2 = H
3 R1 = Me, R2 = H
4 R1 = R2 = Me
1
alkaline pH
R1
E
E
E
CO2Me
O
O
1 equiv. 1 / NaOH
R1
R2
R2
O
E
6a R1 = R2 = H, E = CO2Me
6b R1 = Me, R2 = H, E = CO2Me
6c R1 = R2 = Me, E = CO2Me
OH
CO2Me
(proposed intermediate)
HCl–HOAc heat
5a R1 = R2 = H
5b R1 = Me, R2 = H
5c R1 = R2 = Me
R1
O
O
R2
6a' R1 = R2 = H
6b' R1 = Me, R2 = H
6c' R1 = R2 = Me
Scheme 1
1:1 Adducts
The mechanism of the Weiss reaction has been proposed7 and
can be subdivided into two principal processes; (i) the aldol
sequence, and (ii) the Michael sequence. The aldol condensa-
tion of one molecule of glyoxal 2 with one molecule of dimethyl
3-oxoglutarate 1 is proposed to generate a β-hydroxy ketone
which undergoes an intramolecular aldol reaction followed by
dehydration to form the 4-hydroxycyclopenten-2-one 5 (a 1:1
adduct). The Michael reaction of enone 5 with another mol-
ecule of ketoglutarate 1, which is followed by subsequent
dehydration of the β-hydroxy ketone that results, produces an
J. Chem. Soc., Perkin Trans. 1, 1997
3471

View Article Online
Table 1
O
R1
R2
E
E
E
E
MeO2C
O
R1
R2
O
MeO2C
CO2Me
R1
R2
NaOMe
CH3OH
O
O
O
+
NaOMe–CH3OH
O
E
MeO2C
OH
E
Dicarbonyl compound
R1, R2
1:1 Adduct (% yield)
1:2 Adduct (% yield)
1,2-Di(cyclopentyl)ethane-1,2-dione
1,2-Di(cyclohexyl)ethane-1,2-dione
1,2-Di(cyclopent-3-enyl)ethane-1,2-dione
Benzil
Phenanthrenequinone
1-Phenylpropane-1,2-dione
Phenylglyoxal
2,2Ј-Pyridil
2,2Ј-Furil
2,2Ј-Thenil
7
8
cyclopentyl
cyclohexyl
cyclopent-3-enyl
Ph
biphenyl-2,2Ј-diyl
Ph, Me
Ph, H
2-pyridyl
2-furyl
2-thienyl
7a (0)
7b (12–14)
8b (0)
9b (<10)
12b (0)
13b (54)
14b (50–68)
15b (>60)
16b (0)
17b (52)
19b (0)
8a (61)
9
9a (0)
12
13
14
15
16
17
19
12a (79)
13a (69)
14a (0)
15a (trace)
16a (67)
17a (56)
19a (77)
enone that is suitably set up for a second Michael addition. This
reaction proceeds in an intramolecular fashion to provide the
cis-bicyclo[3.3.0]octane-3,7-dione system 6 (1:2 adduct). In
reactions with 1,2-dicarbonyl compounds such as glyoxal 2,
pyruvaldehyde 3 and butane-2,3-dione 4, no 1:1 intermediates
have been observed or trapped.7 Consequently, the focus of the
present study was directed toward the use of dicarbonyl com-
pounds with suitable substituents (R1 and R2) to stabilize a 1:1
adduct which are capable of later addition of another molecule
of dimethyl 3-oxoglutarate to provide the 1:2 adduct. This
would furnish additional support for the involvement of a 4-
hydroxycyclopenten-2-one intermediate in the Weiss reaction.
HO
O
1, Na, xylene, heat
CO2Et
2, H +
10
11
[O]
O
O
Steric interactions
9
Clearly the steric bulk of the substituent attached to the 1,2-
dicarbonyl system plays an important role in the success of the
Weiss reaction. Consequently, it was decided to investigate
this process with 1,2-diones which are fully substituted with
alicyclic groups. 1,2-Di(cyclopentyl)ethane-1,2-dione 7 and
1,2-di(cyclohexyl)ethane-1,2-dione 8 were synthesized and
reacted (individually) with dimethyl 3-oxoglutarate 1 (Table 1).
When the reactions were carried out under aqueous conditions
(pH 5.6 or 8.3), only starting materials were recovered. When 7
was heated with dimethyl 3-oxoglutarate 1 in a solution of meth-
anolic sodium hydroxide,8,23 a 12–14% yield of the bicyclic 1:2
adduct 7b was obtained. Under the same conditions, the di(cy-
clohexyl) analog 8 failed to produce any 1:2 adduct 8b, how-
ever, when allowed to stir at room temp. for 9 days a 61% yield
of the 1:1 adduct, 4-hydroxycyclopentenone 8a, was observed.
Apparently, the di(cyclohexyl)ethane-1,2-dione 8 reacted with
one molecule of 1 to form the intermediate 1:1 adduct 8a,
which was too crowded sterically to allow addition of a second
molecule of 1 to the enone system. The 1,2-di(cyclopent-3-
enyl)ethane-1,2-dione 9 was prepared in two steps starting from
ethyl cyclopent-3-enecarboxylate 10 by the method of Stocker
(Scheme 2).9 The oxidation of 11 with copper acetate10 pro-
vided the desired diketone 9, albeit in poor yield (27%). Rigby11
had reported that bismuth trioxide served as a mild oxidant for
acyloins in cases where copper reagents failed. When Bi2O3 was
used for the oxidation of 11, the yield of 9 was increased to
52%. When 9 was stirred with dimethyl 3-oxoglutarate, under a
variety of conditions, the formation of a low yield of the
bicyclic 1:2 adduct 9b and a material with the spectral charac-
teristics of the 1:1 adduct 9a was observed. The CI mass spec-
trum of this material contained the molecular ion of 9b at m/z
503 and an ion for 9a at m/z 347. Repeated attempts at column
chromatography provided a material enriched in 9a, but this
1:1 adduct proved too unstable for isolation and complete
characterization. Formation of 9b was apparently retarded by
steric interactions since all attempts to improve the yield of this
1:2 adduct were unsuccessful.
CO2Me
O /NaOH
CO2Me
1 equiv.
CO2Me
CO2Me
MeO2C
O
+
O
O
MeO2C
CO2Me
OH
9a
CO2Me
9b
Scheme 2
From these results, it appeared that substituents which occupy
a ‘molecular volume’ smaller than a cyclopentane unit (as in 7)
would serve as useful substrates in this condensation, while
those which contain the cyclohexyl rings would not react in the
Weiss reaction to provide 1:2 adducts. The formation of both
the 1:1 and 1:2 adducts from di(cyclopentenyl)ethane-1,2-
dione provides additional support for the involvement of a 4-
hydroxycyclopentenone intermediate in the mechanism of the
Weiss reaction. Unfortunately, the instability of this 1:1 adduct
9a did not permit investigation of this hypothesis.
Electronic effects
Earlier work had demonstrated that reaction of dimethyl 3-
oxoglutarate with benzil 12 furnished a stable 4-hydroxy-
cyclopentenone 1:1 adduct 12a. The failure of benzil to provide
a 1:2 adduct was felt to arise from electronic as well as steric
factors since phenanthrenequinone, under analogous condi-
tions, gave a similar result.7 In phenanthrenequinone 13, the
3472
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
phenyl substituents of the 1,2-dicarbonyl system are con-
strained in a six-membered ring, consequently, steric factors in
this case were considered minimal since the aromatic substitu-
ents are tied back in the ring system. These two experiments with
benzil and phenanthrenequinone indicated that electronic fac-
tors (stabilization of the intermediate enone by the aromatic π
system) may play a role in the lack of reactivity of 1 with these
substrates; however, in the case of benzil, steric factors may also
be involved. Phenylglyoxal 15, on the other hand, gave a 1:2
adduct with 1 accompanied by a 1:1 adduct observed by spec-
troscopy but too unstable to isolate.6 Furthermore, 1-phenyl-
propane-1,2-dione 14 reacted in a similar fashion with 1 to
provide the 1:2 adduct 14b. When 14 was stirred with 1 in either
aqueous sodium hydrogen carbonate or in methanolic sodium
methoxide the formation of the 1:2 adduct 14b was observed in
68 and 50% yields, respectively.
cis-bicyclo[3.3.0]octane-3,7-dione tetraester 17b in 52% yield
from 17.12 When 2,2Ј-furil and dimethyl 3-oxoglutarate were
reacted in a 1:1 stoichiometry, under the same conditions, a
new component was observed by TLC, accompanied by a con-
siderable amount of baseline material. In order to avoid
decomposition, the crude material was eluted rapidly under
nitrogen through a short silica gel column providing 17a in pure
form as a light yellow solid (1:1 adduct) in 56% yield. The
isolation and characterization of a stable 1:1 adduct 17a
accompanied by the cis-bicyclo[3.3.0]octane-3,7-dione tetra-
ester 17b from the same dicarbonyl compound provides
strong evidence for the intermediacy of a 4-hydroxycyclopent-
enone on the reaction coordinate of the Weiss reaction.
The ability of 17a to add a second molecule of dimethyl 3-
oxoglutarate could be attributed to either differences in size or
electronic character of the aromatic substituents on the 1,2-
dicarbonyl compound. In order to better differentiate the effect
of size or electronic character of the substituents on the Weiss
reaction, it was decided to investigate the reaction of 2,2Ј-thenil
with dimethyl 3-oxoglutarate. This would allow a comparison
of the effect of thiophene rings on the condensation relative to
that of benzene and furil. The 1,2-dione, 2,2Ј-thenil 19, was
prepared in two steps from commercially available thiophene-2-
carbaldehyde.13–15 The 1:1 adduct 19a was prepared under the
conditions analogous to those employed earlier for preparation
of the furil analog 17a. Sodium dimethyl 3-oxoglutarate was
prepared in methanol and 2,2Ј-thenil was added at room temp.
resulting in precipitation of the sodium salt of the 1:1 adduct.
Treatment of this material with aqueous HCl provided the
di(thienyl) substituted 1:1 adduct 19a in 77% yield (Scheme 3).
This material proved to be more stable than the furil 1:1 adduct
and could be easily purified by gravity column chromatography.
Both the CI and EI mass spectra of the 1:1 adduct 19a con-
tained fragment ions which were similar to those of 17a, fur-
thermore the 1H and 13C NMR spectra were consistent with the
proposed structure of 19a. The methoxy groups appeared as
two singlets in the proton spectrum at δ 3.72 and 3.84, respect-
ively. The signal which represented the hydroxy proton at δ 5.42
disappeared upon treatment with D2O. The carbon spectrum
contained 17 lines and the cyclopentanone carbonyl carbon
atom appeared at δC 190.0, consistent with the signal from this
group observed in related molecules. Examination of both the
1H and 13C NMR spectra of this solid indicated the presence of
a second isomer (<10%). The 1:1 adduct 19a could potentially
exist in two epimeric forms, represented by 20a and 20b. The
2,2Ј-thenil 19 was then heated with 2 equiv. of dimethyl 3-
oxoglutarate in the presence of NaOMe–MeOH to determine
if a 1:2 adduct would result. Although the presence of 19a was
observed by TLC, heating the mixture for 16 h in refluxing
methanol did not provide any indication of the presence of a
Pyridil 16, a heterocyclic analog of benzil, behaved in a simi-
lar fashion. When 16 was stirred with excess 1 in a methanolic
solution of sodium methoxide a yellow precipitate (sodium salt)
was formed. After the usual work-up (aqueous HCl) a 69%
yield of the di(pyridyl) substituted 1:1 adduct 16a was obtained
(Scheme 3). No material which could be identified as the 1:2
X
X1
O
CO2Me
O
O
MeO2C
CO2Me
5
8
14
O
NaOH, heat
OH
CO2Me
X
X
12
16 X = -C N-(pyridine)
17 X = O (furan)
19 X = S (thiophene)
16a X= -C N-(pyridine)
17a X = O (furan)
19a X = S (thiophene)
S
S
CO2Me
CO2Me
O
O
OH
S
CO2Me
OH
CO2Me
S
20a
20b
Scheme 3
adduct 16b was isolated from this reaction or from the same
reaction when repeated at a higher temperature (60 ЊC) in the
presence of additional glutarate 1. In order to determine the
involvement of the 4-hydroxycyclopentenone intermediate in
the Weiss reaction, it was necessary to find a 1,2-dicarbonyl
compound which formed both isolable 1:1 and 1:2 adducts.
Furthermore, for proof that the 1:1 adduct was indeed an
intermediate it was necessary to demonstrate that addition of
one equivalent of dimethyl 3-oxoglutarate to the monocyclic
1:1 adduct would result in the formation of the cis-
bicyclo[3.3.0]octane-3,7-dione system. Consideration of both
steric and electronic effects suggested that a 1,2-dicarbonyl
compound which carried substituents with less steric bulk than
phenyl or cyclohexyl groups and possessed the capability to
stabilize the reactive π system of the 4-hydroxypentenone
intermediate would be required. The 2,2Ј-furil 17 appeared to
be an ideal candidate. It is substituted with two aromatic
groups, the π-systems of which would stabilize the 1:1 adduct,
and the furan rings of furil are smaller than benzene, pyridine
and cyclohexane and might permit the addition of a second
equivalent of dimethyl 3-oxoglutarate 1 to the 1:1 adduct.
When 2,2Ј-furil 17 was heated with 1 in a solution of NaOH
1
1:2 adduct 19b. The IR, H NMR and mass spectra of the
reaction mixture were consistent with the presence of the 1:1
adduct 19a and dimethyl 3-oxoglutarate. An attempt to add
dimethyl 3-oxoglutarate to the 1:1 adduct 19a directly was
likewise unsuccessful.
Evaluation of steric and electronic influences on the success of the
Weiss reaction
The failure of 2,2Ј-thenil to form a 1:2 adduct could result
from either steric effects, electronic effects, or a combination of
both. It was difficult to resolve whether these observations were
due to steric congestion in the intermediate 1:1 adduct [as in
the case of the di(cyclohexyl) analog 8], or to electron release to
the enone system decreasing the reactivity of the 1:1 adduct
toward Michael addition of a second molecule of dimethyl 3-
oxoglutarate. Previously it was felt that both electronic and
steric effects were important in the success of the Weiss conden-
sation, since dimethyl 3-oxoglutarate and phenanthrenequin-
one yielded only a 1:1 adduct upon reaction with excess di-
methyl 3-oxoglutarate.7 In the case of phenanthrenequinone,
1
and methanol, a precipitate formed. Examination of the H
NMR spectrum and mass spectrum of this material indicated
that the 1:2 adduct 17b had been obtained. Treatment of this
sodium salt with aqueous HCl provided the di(furyl) substituted
J. Chem. Soc., Perkin Trans. 1, 1997
3473

View Article Online
Table 2
Substituent R1 = R2
1:1 adduct
* δC of β carbon
in 1:1 adduct
Aldehyde (control)
δC of carbonyl carbon
8a Cyclohexyl
186.5
cyclohexanecarbaldehyde
acetaldehyde
benzaldehyde
2-furaldehyde
2-thienaldehyde
202.1 (ref. 24a)
200.5 (ref. 24a)
190.7 (ref. 24a)
178.2 (ref. 22)
182.9 (ref. 24b)
12a Phenyl
17a 2-Furyl
19a 2-Thienyl
175.0
164.1
153.0
the effect of steric congestion should be less important. This is
due to the unique position of the aromatic substituents which
are tied back in a ring which occupies a molecular volume simi-
lar to that in cyclohexane-1,2-dione at the site of reaction, as
mentioned earlier.
shifts are listed in Table 2. While it is true that 13C chemical
shifts do not depend entirely on electron density, the correlation
between the two is often good in a similar series of com-
pounds.21 For example, the decreased electron density at the
β carbon of α,β-unsaturated ketones is responsible for the
downfield shift of this signal (δ 165.2 for 2-cyclopentenone vs.
δ 130.8 for cyclopentene) relative to that of cyclopentene.22
The assignment of chemical shifts for the β-carbon of the 1:1
adducts could not be made simply on the basis of the broad-
band (BB) decoupled 13C NMR spectrum. Although this signal
would be expected to occur around δ 160, the spectra of the 1:1
adducts contained many peaks in the region from δ 150–170.
To determine the assignments, the quaternary carbons were first
defined through the NORD (noise off-resonance decoupled)
13C NMR spectrum. The 1H-coupled, 13C NMR spectrum was
then recorded and examined. It was possible to make assign-
ments for all the carbon atoms of the cyclopentenone system.
The values for the chemical shifts of the β-carbons of the enone
system in the 1:1 adducts were compared to those for the car-
bonyl carbons of the corresponding substituted aldehydes. In
these aldehydes, it is anticipated that the chemical shift should
be influenced by the electronic interaction of the substituent
with the carbonyl function. A good correlation can be seen
between the chemical shifts of the β-carbons of the 1:1 adducts
and the corresponding shifts of the aldehydic carbonyl carbon
atoms. In addition there is a reasonable correlation between the
chemical shifts of these carbon atoms and their σϩ values.
These pieces of evidence, taken together, would tend to rank the
electron density of the β-carbon atom of the enone system of
the 1:1 adduct in the order of 2-furyl 17 > 2-thienyl 19 > phenyl
12 > cyclohexyl 8. This order should be opposite to their ability
to act as Michael acceptors, for the higher electron density on
the β-carbon of the enone system would be expected to render
this site less reactive toward Michael addition. On this basis, the
1:1 adduct of benzil should be better suited, electronically, to
add a second molecule of dimethyl 3-oxoglutarate than the 1:1
adduct of furil. This is again exactly opposite to the experi-
mental observations! This study suggests that the predominant
effect in the conversion of the 1:1 adduct into the 1:2 adduct
in the series of aromatic substituted 1,2-dicarbonyl compounds
rests on steric factors, analogous to the results observed for 8.
The ability of furil to proceed to the 1:2 adduct in contrast to
the results with thenil, benzil and pyridil was important.
Although thiophene and furan are five-membered heterocycles,
thiophene is larger than furan as a consequence of the size of the
sulfur atom. The atomic radius of oxygen is reported to be 0.74
Å, and its van der Waals radius is 1.40 Å. In contrast, the
atomic radius for sulfur is 1.04 Å and its van der Waals radius is
1.85 Å.16 Examination of models of these substituents indi-
cated that benzene and thiophene are similar in size, and are
both larger than furan. The electronic nature of these substitu-
tions also varies. The electronic effects deemed important here
are those that would alter the electron density at the β-position
of the enone system of the 1:1 adduct (Scheme 4). Substituents
CO2Me
CO2Me
R
R
*
δ+
O –
O
*
R
R
OH
CO2Me
OH
CO2Me
Scheme 4
capable of donating electrons to this site (designated as *)
would be expected to render the enone system less electrophilic,
thereby decreasing the rate of the Michael addition of dimethyl
3-oxoglutarate to the enone in favor of competing side reac-
tions such as the formation of cyclopentadienones. A number
of criteria can be used to judge the electronic nature of these
aromatic groups. The resonance energy of these substituents
can serve to gauge how well their π-electrons are delocalized,
and therefore, their ability to donate electron density to the
enone system of the 1:1 adducts. Resonance energies for
these compounds follow the order: benzene > pyridine > thio-
phene > furan.17–19 On the basis of resonance energies, the furan
substituent should be better able to donate electron density to
the enone system of the 1:1 adduct than either thiophene or
benzene rendering this Michael acceptor the least electrophilic.
The acid strength of the corresponding carboxylic acids follows
the order: furan-2-carboxylic acid > thiophene-2-carboxylic
acid > benzoic acid. This order of acidities indicates that the 2-
furyl and 2-thienyl substituents are able to withdraw electrons
through an inductive mechanism to a greater extent than
benzene. Another measure of the electronic nature of these
substituents is the experimental Hammett σϩ values. These
values have been employed to gauge electronic effects in cases
where direct conjugation is possible and should be appropriate
for the substituted 1:1 adducts. The σϩ values for this series of
compounds were reported by Hill et al. to be: phenyl (1.0),
2-thienyl (Ϫ0.85) and 2-furyl (Ϫ0.95).20 These data suggested
that the di(furyl) 1:1 adduct should be less reactive than either
the benzil or thenil 1:1 adducts based entirely on electronic
factors. This is contrary to the observed results!
Reactions of phenanthrenequinone with dimethyl 3-oxoglutarate
The above results prompted the reinvestigation of the reaction of
phenanthrenequinone 13 with dimethyl 3-oxoglutarate. If elec-
tronic effects are less important in the formation of 1:2 adducts
from aromatic substituted 1,2-diketones than steric effects, then
phenanthrenequinone should be capable of adding two mol-
ecules of dimethyl 3-oxoglutarate to form the bicyclic 1:2 adduct
13b (Scheme 5). When the reaction was repeated under the condi-
tions reported earlier25 (NaOH, MeOH) at room temperature,
the sodium salt of 13a was again found to precipitate from the
reaction mixture. The earlier inability to convert 13a into 13b7
appeared to result from problems of solubility, followed by
decomposition of 13a. To circumvent this difficulty, phenan-
threnequinone was reacted with dimethyl 3-oxoglutarate at
higher dilution and higher temperature (NaOH, MeOH, 68 ЊC)
which resulted in the formation of a white solid. Examination of
the 1H and 13C NMR spectra of the material indicated the pres-
ence of a single isomer with two-fold symmetry consistent with
To evaluate the effect of the substituent on the electron dens-
ity of the β-carbon atom of the enone system in various 1:1
adducts, 13C NMR spectroscopy was employed. The chemical
3474
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
period of 6 h, both 2,5,7-trimethoxyindene 27 and 2,4,6-
trimethoxy-1-methylindene 28 were obtained (Scheme 7). Fur-
CO2Me
O
O
1 equiv. 1
OH
MeO2C
CO2Me
NaOH–MeOH
O
HO
OH
CO2Me
OH
MeO2C
13
CO2Me
13a
2 equiv. 1
NaOH–MeOH
heat
26
1 equiv. 1
CH2N2
NaOH–MeOH
heat
OMe
OMe
MeO2C
MeO2C
MeO
MeO2C
H
CO2Me
OMe
CO2Me
OMe
CO2Me
CO2Me
MeO
MeO2C Me
+
HCl–HOAc
O
OH
OH
O
heat
CO2Me
CO2Me
27
(62.4%)
28
(23.7%)
CO2Me
CO2Me
22
13b
CH2N2
Scheme 5
structure 13b. Hydrolysis of 13b and decarboxylation provided a
93% yield of the biphenylene fused cis-bicyclo[3.3.0]octane-3,7-
dione system 22. In addition, the bicyclic 1:2 adduct was also
formed by adding dimethyl 3-oxoglutarate to the 1:1 adduct,
heating 1:1 adduct 13a and 1 in methanolic sodium hydroxide
for 5 h resulted in the formation of 13b in 53% yield. Dimethyl
3-oxoglutarate 1 was also added to 13a at room temperature,
although only in 29% yield. The formation of 13b from 13a
serves as additional proof for a 4-hydroxycyclopentenone inter-
mediate involved in the Weiss reaction.
Recently the reaction of 1 with ethyl 2,3-dioxopropionate 23
was examined under the Weiss reaction conditions and pro-
vided the first case of the formation of a [5.6] indene ring sys-
tem 25. As shown in Scheme 6, 2 equiv. of 1 and 1 equiv. of 23
(85%)
Scheme 7
thermore when 27 was exposed overnight to CH2N2, the
tetramethylated material 28 was isolated in 85% yield. Outlined
in Scheme 8 is a proposed mechanism for the formation of
dihydroxyindene 25 through an intermediate 1:1 adduct 29 and
cyclopentenone 31. It is believed 4-hydroxycyclopentenone 29
underwent a Michael addition with an additional equivalent of
oxoglutarate 1 and this was followed by subsequent loss of
water to furnish cyclopentenone 31 which can undergo cycliz-
ation by path a (Weiss path) or path b. Path a provides the 1:2
adduct 24 characteristically derived from the Weiss process;
however, reaction by path b provides indene 25 presumably
through keto tautomer 32. Both products which were isolated
and characterized provide additional support for the existence
of cyclopentenones such as 29 and 31 in the Weiss reaction.
CO2Et
CO2Me
MeO2C
O
O
O
+
+
O
CO2Me
MeO2C
H
Experimental
1
23
The experimental details are analogous to those reported earl-
ier.7,11 All materials were commercially available and used as
received, except where indicated. Ether refers to diethyl ether.
Petroleum refers to light petroleum boiling in the range 40–
60 ЊC. Analytical TLC plates used were E. Merck Brinkman
UV-active silica gel or alumina on plastic. Compounds were
visualized by UV light, 2,4-DNP and FeCl3 spray reagents or
iodine vapor. The 2,4-DNP spray reagent was composed of 2,4-
dinitrophenylhydrazine, ethanol and sulfuric acid. The ferric
chloride spray reagent was prepared by dissolving ferric chlor-
ide (2.0 g) in 100 cm3 of water. In the NMR spectral data J
values are given in Hz.
buffer
pH = 8.3
CO2Et
OH
MeO2C
O
MeO2C
HO
CO2Me
CO2Me
OH
1
+
OH
4
5
CO2Me
H
MeO2C
MeO2C
CO2Me
24
25
Scheme 6
1,2-Di(cyclopentyl)ethane-1,2 dione 7
were dissolved in NaHCO3 buffer (pH = 8.3) and stirred at
room temp. for 3 days to provide a mixture of 24 and 25. The
normal Weiss condensation product 24 and the trihydroxy-
indene [5.6] system 25 were obtained in 26 and 35% yields,
respectively. The [5.6] system 25 was isolated as a mixture of
isomers including some of the enol form 26. This is the first
report of the origin of a [5.6] indene system from the Weiss
reaction and may lead to extension of this process to other [5.6]
related systems. The structure of the trihydroxyindene [5.6] sys-
tem 25 was confirmed on examination of the IR, NMR and MS
data as well as X-ray analysis of the corresponding trimethoxy
derivative 27.26 When indene 26 was treated with CH2N2 over a
To a 500 cm3 round bottom flask equipped with a magnetic stir-
rer, heating mantle, Dean–Stark trap and reflux condenser, was
added cyclopentanecarboxylic acid (32.1 g, 0.28 mol), dry ben-
zene (250 cm3), absolute ethanol (70 cm3) and sulfuric acid
(conc., 1 cm3). The mixture was heated at reflux for 30 h. The
excess ethanol and benzene were removed by distillation at
atmospheric pressure. The crude product was washed with aque-
ous NaHCO3 solution (10%, 100 cm3) and brine (100 cm3). The
crude product was purified by distillation under reduced pressure
to provide 29.5 g (74%) of ethyl cyclopentanecarboxylate, bp 80–
85 ЊC (water aspirator); νmax/cmϪ1 2950, 1725 and 1150; δH(60
MHz; CDCl3) 1.26 (3 H, t, J 7, CH3CH2), 1.43–2.20 (8 H, m),
J. Chem. Soc., Perkin Trans. 1, 1997
3475

View Article Online
CO2Et
OH
CO2Me
CO2Et
OH
MeO2C
O
MeO2C
CO2Me
O
EtO2C
H
O
pH = 8.3
buffer
O
O
1
+
O
CO2Me
MeO2C
1
MeO2C
MeO2C
H
23
29
30
– H2O
CO2Et
O
O
OEt
b
OH
MeO2C
MeO2C
O
MeO2C
O
MeO2C
HO
CO2Me
OH
CO2Me
CO2Me
O
CO2Me
OH
path a
path b
a
O
O
CO2Me
CO2Me
MeO2C
H
MeO2C
H
MeO2C
MeO2C
CO2Me
CO2Me
32
31
24
25 mixture of isomers
including enols
Scheme 8
2.20–3.13 (1 H, m), 4.13 (2 H, q, J 7, CH3CH2); m/z (CI) 143
(Mϩ ϩ 1, 100). This material was used directly in the next step.
In a three-neck 500 cm3 round bottom flask fitted with a
reflux condenser, a fine dispersion of sodium metal was pre-
pared (10.0 g, 0.43 mol). The sodium metal was washed by
withdrawing the solvent through a sintered filter stick under
vacuum followed by addition of fresh xylene (50 cm3) and dry
ether. Dry ether (200 cm3) was added and the solution rapidly
stirred. Ethyl cyclopentanecarboxylate (26.7 g, 0.20 mol) was
dissolved in dry ether (50 cm3) and added dropwise over one
hour. The solution was heated to reflux and allowed to stir an
additional 2 h. The mixture was cooled (ice bath) and aqueous
sulfuric acid (50 cm3, H2SO4–H2O, 1:2) was added dropwise to
the stirred mixture under a blanket of nitrogen. Diethyl ether
(100 cm3) was added and the ethereal solution decanted from
the solid salts. The salts were washed with ether (3 × 100 cm3).
The combined fractions were washed with aqueous NaHCO3
(200 cm3, 5%) and dried over MgSO4. The ethereal solution was
filtered and concentrated by rotary evaporation and provided
16.6 g (0.085 mol, 90%) of crude ketol as a bright yellow oil.
The crude ketol was oxidized with cupric acetate as follows: the
crude ketol (16.0 g, 81.2 mmol) was added to a solution com-
posed of glacial acid (125 cm3), methanol (13 cm3), water (14
cm3) and cupric acetate monohydrate (16.2, 82 mmol). The mix-
ture was refluxed for 3 h and then allowed to stand at room
temperature for 12 h. The solution was then filtered through
Celite and the residue was washed several times with ether. The
organic layer was washed with aqueous sodium hydrogen car-
bonate (5%) until the aqueous layer remained basic to pH
paper. The organic layer was then washed with brine (100 cm3),
dried (Na2SO4) and the solvent removed under reduced pres-
sure. The dark oil which resulted was purified by passing it
through a column of silica gel (petroleum–benzene, eluent)
which provided a viscous yellow oil (7.8 g). This material was
purified further by vacuum distillation to yield 3.4 g of 7
(21.5%) which was employed directly in the next step. 7: bp 65–
72 ЊC at 0.3–0.5 mmHg; νmax/cmϪ1 2955, 1730, 1705, 1450;
δH(60 MHz; CDCl3) 1.10–2.23 (18 H, m); m/z (CI) 195 (Mϩ ϩ 1,
25.9), 97 (100) (Found: Mϩ, 194.1295. C12H18O2 requires M,
194.1307).
with cold aqueous HCl (10%), followed by extraction with
chloroform. The organic layer was dried (Na2SO4) and the solv-
ent removed under reduced pressure to provide a viscous oil
(3.29 g). Chromatography on silica gel (benzene–EtOAc eluent)
of this material yielded a white solid 7b (0.30 g, 12%). This
material was recrystallized from methanol to provide an ana-
lytical sample of 7b, mp 177–179 ЊC; νmax/cmϪ1 (KBr) 3480,
1750, 1712, 1655; δH(60 MHz; CDCl3) 0.8–2.4 (18 H, m), 3.0–
3.6 (4 H, br), 3.6–3.8 (12 H); m/z 507 (Mϩ ϩ 1, 100%) (Found:
Mϩ, 506.2132. C26H34O10 requires M, 506.2151). When this
reaction was repeated under conditions analogous to those
described above, with the exception of the period of time
required for heating (2 h reflux in this experiment), a 14% yield
of the 1:2 adduct 7b was realized, as well as the recovery of
starting dione (12%). At no time was any of the 1:1 adduct 7a
isolated from the reaction or observed by TLC or spectroscopy,
although several attempts to accomplish this were executed.
Preparation of 1,2-di(cyclopent-3-enyl)ethane-1,2 dione 9
(oxidation with Bi2O3)
To a three-neck round bottomed flask equipped with a magnetic
stirrer, condenser, thermometer and heating mantle was added
the di(cyclopentene) ketol 11 (4.06 g, 0.021 mol) and glacial
acetic acid (60 cm3). Bismuth oxide (4.4 g, 0.009 mol) was added
and the solution heated to 90 ЊC. The progress of the reaction
was monitored by TLC (silica gel; 25% EtOAc–benzene). With-
in 20 min, a new spot of higher Rf was observed. The reaction
was allowed to stir for 12 h at 90–100 ЊC. The reaction mixture
was cooled and then diluted with ether (40 cm3) and hexane (40
cm3). The bismuth salts which precipitated were removed by
filtration through Celite. The volume of the filtrate was reduced
under vacuum, diluted with ether (300 cm3) and washed with
10% NaHCO3 until the pH of the aqueous layer remained
slightly basic. The ether layer was dried over MgSO4 and the
solvents removed under reduced pressure to provide 2.59 g
(64%) of crude diketone 9. The crude product was purified by
column chromatography (silica gel; 10% EtOAc–hexane) to
provide 2.0 g (52%) of pure 9: bp 75–79 ЊC (0.5–0.7 mmHg);
νmax/cmϪ1 (neat) 3064, 1708; δH(60 MHz; CDCl3) 2.6 (8 H, d, J 7,
CH᎐CHCH ), 4.0 (2 H, quintet, J 7, CH CHCH ), 5.6 (4 H, s);
᎐
2
2
2
m/z 191 (Mϩ ϩ 1, 100%). This material was employed directly
in the next step.
Tetramethyl 1,5-(dicyclopentyl)-3,7-dioxo-cis-bicyclo[3.3.0]-
octane-2,4,6,8-tetracarboxylate 7b
2,5-Bis(methoxycarbonyl)-3,4-dicyclohexyl-4-hydroxycyclopent-
2-enone 8a
To a solution of dimethyl 3-oxoglutarate 1 (1.74 g, 10 mmol) in
methanol (15 cm3), solid sodium hydroxide (0.40 g) was added.
The mixture was heated to reflux, followed by addition of 1,2-
di(cyclopentyl)ethane-1,2-dione 7 in one portion. The heat was
removed and the reaction mixture was allowed to stir at room
temperature for 70 h. The methanol was then removed under
reduced pressure, and the light brown residue was dissolved in
water. This aqueous solution was brought to slightly acidic pH
To a solution of sodium hydroxide (0.8 g, 20 mmol) in meth-
anol (16 cm3), dimethyl 3-oxoglutarate 1 (7.08 g, 40 mmol) was
added. The mixture was heated to reflux for 10 min, after which
1,2-di(cyclohexyl)ethane-1,2-dione 8 (4.45 g, 20 mmol), pre-
pared by the method of Stocker,9 was added in one portion. The
heat was removed and the reaction mixture allowed to stir at
room temperature for 9 d. The methanol was subsequently
3476
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
removed under reduced pressure, and the remaining residue dis-
solved in water, acidified with cold 1  HCl and extracted with
ether. The organic layer was then dried over MgSO4 and con-
centrated under reduced pressure to provide crude 8a. This
material was purified by column chromatography on silica gel
(CHCl3) to provide the pure 1:1 adduct 8a (3.15 g, 61.1%) and
unreacted dione 8 (1.2 g). An analytically pure sample of 8a
was obtained by recrystallization from petroleum, mp 58–61 ЊC
(Found: C, 66.90; H, 8.25. Calc. for C21H30O6: C, 66.68; H,
7.99%); νmax/cmϪ1 3510, 2945, 1755, 1735, 1700, 1620; δC(62.86
MHz; CDCl3) 25.59, 25.80, 26.01, 26.09, 26.23, 26.27, 27.00,
27.22, 30.02, 32.03, 38.28, 44.13, 52.14, 52.97, 57.89, 83.12,
133.39, 164.20, 168.91, 186.28, 193.5; m/z (CI) 396 (Mϩ ϩ 18,
100%), 379 (M ϩ 1, 95%).
2,5-Bis(methoxycarbonyl)-3,4-di(2-pyridyl)-4-hydroxycyclopent-
2-enone 16a
A solution of sodium methoxide was prepared by dissolving
sodium metal (0.24 g, 10.4 mmol) in dry methanol (50 cm3).
Dimethyl 3-oxoglutarate 1 (1.82 g, 10.5 mmol) was added and
the solution allowed to stir for 10 min at room temperature.
The 2,2Ј-pyridil analog 16 was added in one portion, and
immediately dissolved to provide a dark yellow–brown colored
solution. Within 20 min, a green–yellow precipitate began to
form. The solution was stirred for 12 h, and the resulting solid
was filtered from the medium and dried to provide 16a (2.79
g) as its sodium salt. This material decomposed at 230 ЊC
without melting. The salt 16a was then dissolved in water (50
cm3), the pH adjusted to 7.0, and the neutral solution
extracted with CHCl3 (3 × 50 cm3). The combined organic
fractions were dried over MgSO4 and concentrated under
reduced pressure to give the 1:1 adduct 16a (2.4 g, 67%). An
analytical sample was obtained through recrystallization from
methanol, mp 127–130 ЊC (Found: C, 61.95; H, 4.27; N, 7.46.
2,5-Bis(methoxycarbonyl)-3,4-di(2-thienyl)-4-hydroxycyclopent-
2-enone 19a
Sodium hydroxide (0.470 g, 11.7 mmol) was dissolved in
methanol (30 cm3). This solution was filtered and dimethyl 3-
oxoglutarate 1 (1.20 g, 5.86 mmol) was added and the solution
allowed to stir at room temperature for 10 min. 2,2Ј-Thenil 19
(1.5 g, 5.86 mmol) was added to the mixture in one portion
and the mixture allowed to stir at room temperature for a total
of 8 h. The resulting precipitate was filtered and dried in vac-
uum to provide 1.95 g of a bright orange salt (mp >250 ЊC,
darkened at 190 ЊC); νmax/cmϪ1 (KBr) 3531, 3441, 2899, 1711,
1672, 1567, 1543, 1451, 1440. This salt was suspended in a
separatory funnel in H2O (50 cm3) and CH2Cl2 (100 cm3).
Aqueous HCl (10%, 50 cm3) was added and the mixture
shaken until the solid dissolved. The layers were separated and
the aqueous phase was extracted with CH2Cl2 (50 cm3). The
organic phases were combined, dried over MgSO4 and concen-
trated under reduced pressure to provide 19a (1.7 g, 76.7%) as
a yellow solid, mp 110–112 ЊC. An analytical sample of 19a
was obtained through recrystallization from EtOAc–hexane;
mp 112–113 ЊC (Found: C, 53.49; H, 3.81. Calc. for
C17H14O6S2: C, 53.96; H, 3.73%); νmax/cmϪ1 (KBr) 3340, 3101,
1751, 1705, 1696, 1686, 1596, 1497; δH(250 MHz; CDCl3) 3.72
(3 H, s), 3.82 (1 H, s), 3.84 (3 H, s), 5.42 (1 H, s), 6.71 (2 H, m),
6.98 (1 H, d, J 3.5, 9-H), 7.17 (1 H, dd, J 1.5, 5, 4-H), 7.53 (1
H, dd, J 1.3, 4, 11-H), 7.65 (1 H, dd, J 1.3, 4, 2-H); m/z (EI)
378 (Mϩ, 24.5), 360 (24.5), 346 (26.5), 320 (36.7), 319 (100),
289 (22.4), 288 (57.1), 287 (67.3), 267 (26.5), 235 (20), 195
(38.8), 166 (32.7), 111 (51.1) (Found: Mϩ, 378.0217. C17H14O6S2
requires M, 378.0232).
Calc. for C19H16N2O6: C, 61.95; H, 4.39; N, 7.61%); νmax
/
cmϪ1 (KBr) 3400–3000, 1760–1745, 1715; m/z (CI) 369
(Mϩ ϩ 1, 55.7), 353 (29.7), 351 (11.0), 337 (17.3), 321 (34.0),
311 (100) (Found: Mϩ, 368.1008. C19H16N2O6 requires M,
368.1008).
2,4,6,8-Tetramethyl 3,7-dioxo-1,5-(biphenyl-2,2Ј-diyl)-cis-
bicyclo[3.3.0]octane-2,4,6,8-tetracarboxylate 13b
Sodium hydroxide (3.8 g, 96 mmol) was dissolved in dry meth-
anol (300 cm3) and dimethyl 3-oxoglutarate 1 (8.36 g, 48 mmol)
was added. The solution was heated to 60 ЊC (oil bath) and
phenanthrenequinone 13 (4.0 g, 19.0 mmol) was added. The
mixture was then stirred at 60 ЊC for 2 d. The solution was
cooled to room temperature and the precipitate which formed
was filtered from the medium, dissolved in water (100 cm3) and
acidified to pH 3 with aqueous HCl (10%). The precipitate was
filtered and dried in vacuo to yield 13b (1.7 g) as a cream colored
solid. The original filtrate was concentrated under reduced
pressure, dissolved in water (100 cm3), brought to pH 3 and
filtered to provide an additional 3.7 g of 13b. The combined
yield (5.4 g) of 13b was 54%. An analytical sample of 13b was
obtained by recrystallization from CHCl3–hexane, mp 234–
235 ЊC; νmax/cmϪ1 (KBr) 3428, 3071, 1746, 1652, 1619; δH(250
MHz; CDCl3) 2.90 (6 H, s), 4.00 (6 H, s), 4.70 (2 H, s), 7.24–
7.27 (4 H, t), 7.63–7.67 (2 H, t), 7.73–7.77 (2 H, t), 11.3 (2 H, s,
broad enol absorptions); δC(62.86 MHz; CDCl3) 51.68, 52.20,
59.18, 108.51, 122.92, 127.84, 128.14, 128.34, 131.40, 135.52,
167.96, 170.69, 175.84; m/z (CI, CH4) 549 (Mϩ ϩ 29, 5.6), 521
(Mϩ ϩ 1, 26.6), 489 (67.2), 485 (16.9), 457 (100), 431 (10.9), 425
(16.3), 347 (10.7) (Found: Mϩ, 520.1366. C28H24O10 requires M,
520.1369).
Attempted preparation of tetramethyl 3,7-dioxo-1,5-di(2-
thienyl)-cis-bicyclo[3.3.0]octane-2,4,6,8-tetracarboxylate 19b
Dimethyl 3-oxoglutarate 1 (0.78 g, 4.5 mmol) was added to a
solution of methanol (20 cm3) and sodium hydroxide (0.224 g,
5.6 mmol). The mixture was heated to 60 ЊC (oil bath) and 2,2Ј-
thenil 19 (0.5 g, 2.25 mmol) was added. The mixture was held at
60 ЊC for 12 h and then allowed to cool to room temperature.
The methanol was removed under reduced pressure and the
dark residue was dissolved in CHCl3 (100 cm3) in a separatory
funnel and shaken with aqueous HCl (10%, 10 cm3). The
organic layer was dried (MgSO4) and concentrated under
reduced pressure leaving a dark viscous material. Examination
of the reaction mixture by TLC (silica gel, 60% EtOAc–hexane)
indicated the presence of dimethyl 3-oxoglutarate 1, the 1:1
adduct 19a of 2,2Ј-thienyl, and baseline material. The 1H NMR
(60 MHz, CDCl3) spectrum of this material consisted of
absorptions in the aromatic region of the spectrum (δ 6.7–7.8)
and overlapping singlets in the region of ester methyl hydrogens
(δ 3.5–4.1); m/z (CI) 379 (6.7), 361 (31.1), 347 (8.9), 331 (20.0),
303 (11.1), 289 (6.7), 175 (4.4), 143 (100). The molecular ion of
the desired 1:2 adduct (m/z 353) was not present nor were the
characteristic fragment ions which would result from the suc-
cessive losses of methanol from the parent ion observed on
mass spectroscopy of this material.
Conversion of the 1:1 adduct 13a from phenanthrenequinone 13
into the 1:2 adduct 13b with 1 at 60 ЊC in methanol
Sodium hydroxide (1.9 g, 4.8 mmol) was dissolved in dry
methanol and the solution was heated to 60 ЊC. The mixture
was stirred at this temperature for 5 min after which 13a (2.0 g,
6 mmol) was added. Within 10 min a precipitate began to form
and this solution was refluxed for 4.5 h and then allowed to cool
to room temperature. The solution was filtered, the filtrate
reduced in volume (50 cm3) and brought to pH 3 with cold
aqueous HCl (10%). The suspension was filtered and dried in
vacuo to provide 13b (1.81 g, 63%) as a tan solid which was
recrystallized from CHCl3–hexane to yield 1.65 g (53%) of
product as an off white solid, mp 234–235 ЊC. This material
proved to be identical to an authentic sample of 13b by com-
parison of mass and IR spectra, as well as TLC.25
1,5-(Biphenyl-2,2Ј-diyl)-cis-bicyclo[3.3.0]octane-3,7-dione 22
The tetraester 13b (0.454 g, 0.87 mmol) was added to a mixture
J. Chem. Soc., Perkin Trans. 1, 1997
3477

View Article Online
of acetic acid (30 cm3) and aqueous HCl (10%, 30 cm3). The
mixture was heated to reflux and within 1 h the solid dissolved
to provide a homogeneous solution which was refluxed for 18 h
then cooled to room temperature. Upon cooling, crystals pre-
cipitated from the light yellow solution. This solution was
cooled to 0 ЊC (ice bath) and was filtered to provide 204 mg of
crystalline 22. The filtrate was extracted with CHCl3 (3 × 20
cm3), and the combined organic layers dried over MgSO4 and
concentrated under reduced pressure to give an additional 30
mg of 22 (93%). A sample of 22 was obtained by recrystalliz-
ation from EtOAc–hexane; mp 220–222 ЊC; νmax/cmϪ1 (KBr)
1725; m/z (CI, CH4) 328 (Mϩ ϩ 41, 7.8), 317 (Mϩ ϩ 29, 14.9),
289 (Mϩ ϩ 1, 75.7), 231 (100); δH(250 MHz; CDCl3) 2.62 (4 H,
d, J 19.4, CH2 endo), 2.93 (4 H, d, J 19.4, CH2 exo), 7.92 (2 H,
m), 7.35 (6 H, m); δC(62.86 MHz; CDCl3) 50.00, 50.64, 124.56,
127.80, 128.30, 128.97, 131.09, 136.2, 214.04 (Found: Mϩ,
288.1129. C20H16O2 requires Mϩ, 228.1150).
18.64, 51.90, 52.47, 58.72, 59.66, 62.61, 63.98, 104.49, 118.89,
123.04, 131.32, 142.59, 154.42, 163.66, 165.31, 166.06, 167.00,
170.16; m/z (CI) 453 (46, Mϩ ϩ 1).
Reaction of indene 27 with CH2N2 to provide indene 28
Indene 27 (20 mg, 0.0457 mmol) was added to 8 cm3 of an ether
solution of CH2N2 (0.67 ) at 0 ЊC. The mixture was stirred at
0 ЊC for 2 h and was then stirred at room temp. for 4 h. The
solvent was removed under reduced pressure. The residue
was purified by flash chromatography (silica gel; EtOAc–
hexane = 2:3) to give 28 (17.5 mg, 85%) which was identical
with 28 obtained in the previous experiment.
Acknowledgements
We thank Mr Frank Laib for providing mass spectroscopic data
and Mr Keith Krumnow for elemental analysis. We also would
like to thank NSF for financial support.
Reaction of ethyl 2,3-dioxopropionate with dimethyl 3-
oxoglutarate to provide tetramethyl 5,7-dihydroxy-2-oxo-2,3-
dihydro-1H-indene-1,3,4,6-tetracarboxylate 25 and tetramethyl
1-ethoxycarbonyl-3,7-dihydroxybicyclo[3.3.0]octa-2,6-diene-
2,4,6,8-tetracarboxylate 24
References
1 (a) S. H. Bertz, J. Am. Chem. Soc., 1981, 103, 3599; (b) S. H. Bertz,
J. Am. Chem. Soc., 1982, 104, 5801.
2 G. H. Posner, Chem. Rev., 1986, 86, 831.
Dimethyl 3-oxoglutarate 1 (5.22 g, 0.03 mol) was dissolved in
aqueous sodium hydrogen carbonate buffer (1.4 g NaHCO3 in
100 cm3 H2O) and ethyl 2,3-dioxopropionate27 23 (2.22 g, 0.015
mol) was then added in one portion. This mixture was allowed
to stir for 4 d and then acidified to pH 1 with ice-cold 6 
aqueous HCl. The water layer was extracted with CHCl3 (3 × 50
cm3). The combined extracts were dried (Na2SO4) and the solv-
ent removed under reduced pressure to provide an oil which was
dissolved in ethanol and allowed to crystallize. The crystals were
filtered from the solution and dried to furnish 25 (2.08 g, 35%)
as a white crystalline solid, mp 173–175 ЊC (Found: C, 51.33; H,
4.01. Calc. for C17H16O11: C, 51.52; H, 4.04%); νmax/cmϪ1 (KBr)
3373, 2955, 1772, 1739, 1665; δH(250 MHz; CDCl3) major iso-
mer 3.72 (3 H, s), 3.74 (3 H, s), 3.85 (3 H, s), 4.01 (3 H, s), 4.46 (1
H, s), 4.78 (1 H, s), 12.63 (1 H, s), 12.93 (1 H, s); δC(62.86 MHz;
CDCl3) 52.10, 53.01, 53.09, 56.48, 62.98, 63.11, 102.05, 102.42,
118.08, 118.38, 144.34, 144.50, 164.45, 165.47, 166.01, 166.52,
166.74, 169.92, 171.10, 199.59, 199.89; m/z 397 (18.6, Mϩ ϩ 1).
The mother liquor was evaporated slowly to provide 24 as a
3 (a) X. Fu and J. M. Cook, Tetrahedron Lett., 1990, 31, 3409;
(b) X. Fu and J. M. Cook, J. Org. Chem., 1992, 57, 5121.
4 S. H. Bertz, G. Lannoye and J. M. Cook, Tetrahedron Lett., 1985, 26,
4695.
5 J. Wrobel, K. Takahashi, V. Honkan, S. H. Bertz, G. Lannoye and
J. M. Cook, J. Org. Chem., 1983, 48, 139.
6 (a) R. M. Coates, S. K. Shah and R. W. Mason, J. Am. Chem. Soc.,
1979, 101, 6765; (b) Y. K. Han and L. A. Paquette, J. Org. Chem.,
1979, 44, 3731; (c) Y. K. Han and L. A. Paquette, J. Am. Chem. Soc.,
1981, 103, 1831.
7 S. Yang-Lan, M. Mueller-Johnson, J. Oehldrich, D. Wichman,
J. M. Cook and U. Weiss, J. Org. Chem., 1976, 41, 4053.
8 S. H. Bertz, PhD Thesis, Harvard University, Cambridge, MA, 1978.
9 J. H. Stocker, J. Org. Chem., 1964, 29, 3593.
10 A. T. Blomquist and A. Goldstein, Org. Synth., 1948, Coll. Vol. 4,
838.
11 W. Rigby, J. Chem. Soc., 1951, 793.
12 G. Kubiak and J. M. Cook, J. Org. Chem., 1984, 49, 561.
13 W. S. Ide and J. S. Buck, Org. React., Wiley, NY, 1948, 4, 269.
14 S. Z. Cardon and H. P. Lankelma, J. Am. Chem. Soc., 1948, 70, 4248.
15 I. Deschamps, W. J. King and F. F. Nord, J. Org. Chem., 1949, 14, 184.
16 CRC Handbook of Chemistry and Physics, ed. R. C. West, CRC
Press, West Palm Beach, FLA, p. D-230.
17 J. March, Advanced Organic Chemistry, McGraw-Hill, New York,
2nd edn., 1977, p. 32.
18 A. F. Bedford, A. E. Beezer and C. T. Mortimer, J. Chem. Soc., 1963,
2039.
19 R. M. Acheson, in An Introduction to the Chemistry of Heterocyclic
Compounds, Wiley, New York, 1980, 3rd edn., p. 124 (thiophene),
p. 153 (furan).
20 E. A. Hill, M. L. Gross, M. Stasiewicz and M. Manion, J. Am.
Chem. Soc., 1969, 91, 7381.
21 R. L. Levin and L. Weingarten, Tetrahedron Lett., 1975, 611.
22 R. M. Silverstein, G. C. Bassler and T. C. Morril, in Spectrometric
Identification of Organic Compounds, Wiley, New York, 4th edn.,
1974, p. 277.
23 K. Avasthi, M. N. Deshpande, W. C. Han, J. M. Cook and U. Weiss,
Tetrahedron Lett., 1981, 22, 3475.
24 (a) G. C. Levy, R. L. Lichter and G. L. Nelson, in Carbon-13
Nuclear Magnetic Resonance Spectroscopy, Wiley, New York, NY,
1980, p. 145; (b) E. Breitmaier and W. Voelter, 13C NMR
Spectroscopy, in vol. 5 of Monographs in Modern Chemistry, ed.
H. F. Ebel, Verlag Chemie, Weinhein/Bergst., 1974, p. 185.
25 G. Kubiak, J. M. Cook and U. Weiss, Tetrahedron Lett., 1985, 26,
2163.
white crystalline solid (1.72 g, 26%), mp 121–123 ЊC; νmax
/
cmϪ1 (KBr) 3286, 2956, 1735, 1677; δH(250 MHz; CDCl3) 1.21 (3
H, t, J 9, CH3CH2), 3.61 (3 H, s), 3.71 (3 H, s), 3.72 (1 H, d, J 7,
5-H), 3.77 (3 H, s), 3.84 (3 H, s), 3.91 (1 H, d, J 7, 4-H), 4.22 (2
H, q, J 9, CH3CH2), 4.62 (1 H, s), 10.12 (1 H, s), 10.55 (1 H, s);
m/z (CI, CH4) 443 (Mϩ ϩ 1, 78.4%).
Reaction of 25 with CH2N2 28 to provide 2,5,7-trimethoxy-
1,3,4,6-tetramethoxycarbonylindene 27 and 1-methyl-2,4,6-
trimethoxy-1,3,5,7-tetramethoxycarbonylindene 28
The indene 25 (1 g, 2.53 mmol) (enol form, see 26) was added to
a 50 cm3 (0.67 ) ether solution of CH2N2 at 0 ЊC. The mixture
was stirred at 0 ЊC for 2 h and then stirred at room temperature
for 4 h. The solvent was removed under reduced pressure and
the residue was purified by flash chromatography (silica gel;
hexane–EtOAc = 1:2) to furnish 27 (0.69 g, 62.4%) and 28 (0.27
g, 23.7%). Indene 27, mp 154–155 ЊC (Found: C, 54.78; H, 5.01.
Calc. for C20H22O11: C, 54.79; H, 5.02%); νmax/cmϪ1 (KBr) 2952,
1735; δH(250 MHz; CDCl3) 3.72 (3 H, s), 3.83 (3 H, s), 3.87 (3
H, s), 3.88 (3 H, s), 3.89 (3 H, s), 3.92 (3 H, s), 4.03 (3 H, s), 4.59
(1 H, s); δC(62.86 MHz; CDCl3) 51.58, 51.84, 52.14, 52.43,
52.90, 59.79, 60.67, 64.25, 111.66, 112.36, 116.24, 120.67,
145.18, 153.62, 157.90, 163.64, 165.93, 165.99, 167.71, 167.90;
m/z (CI) 439 (46, Mϩ ϩ 1). Indene 28, mp 136–137 ЊC
(Found: C, 55.66; H, 5.27. Calc. for C21H24O11: C, 55.75; H,
5.31%); νmax/cmϪ1 (KBr) 2950, 1735; δH(250 MHz; CDCl3) 1.50
(3 H, s), 3.62 (3 H, s), 3.64 (3 H, s), 3.67 (3 H, s), 3.77 (3 H, s),
3.91 (3 H, s), 3.92 (3 H, s), 3.94 (3 H, s); δC(62.86 MHz; CDCl3)
26 The X-ray coordinates of 27 and 28 will be published elsewhere:
J. Li, D. Bennett, D. Grubisha and J. M. Cook, unpublished work.
27 H. Ihmels, M. Maggini, M. Prato and G. Scorrano, Tetrahedron
Lett., 1991, 32, 6215.
28 M. Hudlicky, J. Org. Chem., 1980, 45, 5377.
Paper 7/01232B
Received 21st February 1997
Accepted 11th July 1997
3478
J. Chem. Soc., Perkin Trans. 1, 1997