Controlled synthesis of α-allenic ester and spiro ketone derivatives from tailored α-substituted cycloalkanones through cascade reactions: Exploring the possible reaction pathways by means of semiempirical MO calculations

Article abstract of DOI:10.1002/ejoc.200500339

Functionalized bicyclic electrophilic allenes and (or) spiro ketones have been obtained through a cascade reaction when acetylenic ω-keto esters were treated with tetra-n-butylammonium fluoride (TBAF). The isolated products can be modulated by using diffe

Full text of DOI:10.1002/ejoc.200500339

FULL PAPER
Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives from
Tailored α-Substituted Cycloalkanones Through Cascade Reactions: Exploring
the Possible Reaction Pathways by Means of Semiempirical MO Calculations
Antonio J. Mota,[a] Aurélie Klein,[b] Florence Wendling,[b] Alain Dedieu,[a] and
Michel Miesch*[b]
Keywords: Cascade reactions / Allenic esters / Spiro compounds / Semiempirical calculations / Transition states / HF
catalysis
Functionalized bicyclic electrophilic allenes and (or) spiro
ketones have been obtained through a cascade reaction
when acetylenic ω-keto esters were treated with tetra-n-bu-
tylammonium fluoride (TBAF). The isolated products can be
modulated by using different starting materials (i.e. varying
ring size, spacer length) and reaction times. The results indi-
cate that these reactions take place under kinetic control. In
line with this, a semiempirical study was carried out in order
to determine the preferred reaction pathways and the pre-
ferred cis/trans conformation of the bicyclic allenic deriva-
tives. An excellent correlation was found between theoretical
and experimental results, and the whole process dramatically
depends on the presence of the HF molecule.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
In this context, we have previously shown that acetylenic
ω-keto esters are readily available compounds that undergo
cascade reactions to give electrophilic allenes,[2] spiro
ketones, or oxetane derivatives when treated with different
bases.[3] However, in spite of the synthetic utility of these
systems, to the best of our knowledge, only a few reports
deal with the intramolecular reactivity of acetylenic ω-keto
esters. For example, Deslongchamps and co-workers re-
ported that polycyclic compounds could be readily ob-
tained through the intramolecular Michael addition of cy-
clic β-keto esters onto conjugated acetylenic ketones.[4]
More recently, Christoffers and Oertling achieved a seven-
membered ring annulation through an intramolecular
Michael reaction catalyzed by FeCl3.[5] Nevertheless, ole-
finic ω-keto esters have been reported to have different in-
tramolecular reactivities to those determined by our own
methodology. For example Fukumoto and co-workers, and
later Ihara and co-workers, have shown that a tandem
Michael–aldol reaction took place when olefinic ω-keto es-
ters were treated with either TBDMSOTf/NEt3 or TMSI/
(TMS)2NH.[6] On the other hand, Chiu et al. reported that
olefinic ω-keto esters undergo tandem conjugate reduction–
aldol cyclization reactions (using Stryker’s reagent) to read-
ily afford β-hydroxyketones in good yields.[7] We now wish
to report a detailed investigation of the reactivities of acety-
lenic ω-keto esters towards TBAF. We shall focus our atten-
tion on mechanistic considerations by means of semiempir-
ical MO calculations.
Introduction
Cascade reactions may be defined as multi-reaction,
“one-pot” sequences in which the first reaction creates the
functionality required to trigger the second one without a
need to change the reaction conditions. The performance
of an entire sequence of bond-forming reactions in one step
requires minimal effort for the isolation, purification, and
characterization of intermediate products, therefore making
the entire synthetic process economically and ecologically
more favorable. From the standpoint of efficiency, cascade
reactions are valuable, and are highly prized if they have
broad scope and occur under mild conditions. They offer a
very efficient approach to the formation of multiple bonds
and can therefore produce polycyclic structures in a con-
trolled fashion, thus leading to complex target molecules,
mainly natural products, that are difficult or impossible to
prepare by other means.[1] This has led to the idea that
chemical groups that are considered unreactive under nor-
mal conditions may react when a reactive intermediate is
generated in their vicinity.
[a] Laboratoire de Chimie Quantique et de Modélisation Molécu-
laire,
UMR 7551 CNRS/ULP, Université Louis Pasteur,
Strasbourg. 4, rue Blaise Pascal, 67000 Strasbourg, France
[b] Laboratoire de Chimie Organique Synthétique,
UMR 7123 CNRS/ULP, Université Louis Pasteur,
Strasbourg 1, rue Blaise Pascal, 67000 Strasbourg, France
Fax: +33-3-90241754
E-mail: miesch@chimie.u-strasbg.fr
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Semiempirical MO methods suffer certain well-known
problems, such as the description of the hydrogen bond (es-
4346
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200500339
Eur. J. Org. Chem. 2005, 4346–4358

Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives
FULL PAPER
pecially problematic in MNDO), typical organic hyperval- isomers) was formed as the main compound. Compound 6
ent atoms (i.e. sulfur or phosphorus), transition metals, was formed from 5 by increasing the reaction time
transition states, or molecules that contain atoms for which (Table 1).
good parametrization does not exist, or simply, for which
these parameters are not available.[8] Nevertheless, these
methods can quickly offer qualitative information about the
atomic organization and electronic distribution of experi-
mentally studied systems. For this reason, semiempirical
methods have become very useful tools in organic chemis-
try. In spite of the problems associated with the use of these
methods in theoretical calculations, they have been widely
employed,[9] even in the description of very complex reac-
tion mechanisms.[10]
Despite the high percentage of starting material 1 reco-
vered after 5 min (Entry 1), this was the optimal reaction
time for the production of spiro derivative 5. Nevertheless,
if our target is to obtain the polyfunctionalized ester 6, the
reaction time must be increased. Thus, after 48 h (Entry 4),
compound 6 was isolated as the main product (ca. 50%
yield) and results from the addition of a water molecule
onto the carbonyl group, followed by a ring-opening reac-
tion (Scheme 3).
When the length of the spacer between the cyclopenta-
none moiety and the electrophilic triple bond was increased
by one carbon–carbon bond, the spiro derivative 7 (a mix-
ture of the 7a, 7b, and 7c isomers) and the bicyclic allene 8
[a mixture of 8a and 8b isomers with a cis ring junction as
ascertained by NMR spectroscopy (NOESY experiments)]
[3a]
were formed as the major products. These compounds
were isolated in approximately the same ratio when the re-
action time was increased from 5 to 30 min (Table 2).
This means that there is no interconversion between
these spirocyclic and allenic forms. However, by increasing
the reaction time, allenic derivative 8 was completely trans-
formed into the β-keto ester 9. The latter probably results
from the intramolecular addition of the corresponding alk-
oxide onto the central carbon of the electrophilic allene,
affording a highly strained oxete ring, which evolves
through a ring-opening process to give the β-keto ester 9
(Scheme 4).
Scheme 1.
The theoretical contribution presented in this work con-
sists of a semiempirical MO study of the potential energy
space of an ensemble of propargylic (A) and enolate (B)
carbanions (Scheme 1). The evolution of allenic or spiro de-
rivatives, as well as the stereochemistry of the compounds
formed, will be discussed. For this purpose, the AM1 and
PM3 methods implemented in the Hyperchem 7.5 package
(see Expt. Sect. for computational details) were utilized.[11]
By replacing the cyclopentanone ring by a cyclohexa-
none ring in the starting acetylenic ω-keto ester (com-
pounds 3 and 4), we observed that either the allene deriva-
tive 10 (mixture of isomers 10a and 10b) was formed as the
major product along with the spiro derivatives 11, or the
unique compound 12, depending on the length of the
spacer (m value) (Table 3).
Note that the cis or trans diastereoselectivity observed in
the formation of the bicyclic allenes 10 and 12 depends on
the m value. The structures of these compounds were unam-
biguously secured by X-ray diffraction.[13]
Results and Discussion
The acetylenic ω-keto esters 1–4 were synthesized start-
ing from the corresponding N,N-dimethylhydrazones using
our previously developed reaction sequence (Scheme 2).[3b]
We decided to study in detail the reactivity of these com-
pounds towards tetra-n-butylammonium fluoride (TBAF),
whose behavior as a relatively strong base is well-documen-
ted.[12] First, we observed that no reaction took place when
TBAF was used in a catalytic amount. In order to promote
reaction, the use of stoichiometric quantities of TBAF
proved necessary. Reactions were performed at room tem-
In summary, the reactions of the acetylenic ω-keto esters
perature (ca. 25°C) and the reaction time was modified for 1–4 with TBAF generated two major products: spiro com-
each experiment in order to gain valuable information pounds 5, 7, and 11, and allene compounds 8, 10, and 12.
about the reaction kinetics.
Compounds 6 and 9 result from a subsequent transforma-
Starting from the acetylenic ω-keto ester 1, we noticed tion of the initial “primary” adducts, that is, of the spiro
that the spiro derivative 5 (as a mixture of 5a, 5b, and 5c and allene derivatives, respectively (Scheme 3 and
Scheme 2.
Eur. J. Org. Chem. 2005, 4346–4358
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4347

A. J. Mota, A. Klein, F. Wendling, A. Dedieu, M. Miesch
FULL PAPER
Scheme 3.
Scheme 4.
Table 1. Reaction of acetylenic ω-keto ester 1 with TBAF.
Entry
Reaction time
Yield of 5a, 5b, and 5c [%][a]
Yield of 6 [%]
Yield of 1 [%][b]
1
2
3
4
5 min
56
45
39
22
–
8
36
49
39
–
–
30 min
3 h 30 min
48 h
–
[a] The 5a, 5b, and 5c ratio (1:0.5:0.36) was only determined after 30 min. [b] Recovered starting material.
Table 2. Reaction of acetylenic ω-keto ester 2 with TBAF.
Entry
Reaction time
Yield of 7a, 7b, and 7c [%][a]
Yield of 8a and 8b [%]
Yield of other [%]
1
2
3
4
5 min
40
55
30
–
35
43
–
7[b]
–
30 min
3 h 30 min
48 h[d]
44[c]
–
–
[a] The 7a, 7b, and 7c ratio (1:0.14:0.06) was only determined after 30 min. [b] Recovered starting material (2). [c] Compound 9 (Scheme 4).
[d] Decomposition occurred.
4348
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4346–4358

Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives
Table 3. Reaction of acetylenic ω-keto esters 3 and 4 with TBAF.
FULL PAPER
m value
Reaction time
Yield of 10a + 10b [%]
Yield of 11a + 11b [%]
Yield of 12a + 12b [%]
Yield of other [%]
1
5 min
30 min
4 h[b]
5 min
30 min
4 h
45
59
–
–
–
15
15
–
–
–
–
–
–
50
70
45
25[a]
–
–
2
50[c]
20[c]
20[c]
–
–
[a] Recovered starting material 3. [b] Decomposition occurred. [c] Recovered starting material 4.
Scheme 5.
Scheme 4). Thus, TBAF can remove either one of two pro- atively wide range of new polyfunctionalized bicyclic com-
tons. 1) A propargylic proton leading to carbanion A which pounds (Scheme 6).
undergoes intramolecular addition onto the carbonyl car-
bon to give the alkoxide A1. The allene derivatives 8, 10,
and 12 were then isolated by prototropy A1 Ǟ A2 (upper
line in Scheme 5). 2) A hydrogen atom located in the posi-
tion α to the carbonyl group in order to give carbanion
B, which evolves through an intramolecular Michael-type
addition to afford the spiro derivatives 5, 7, and 11 (middle
line in Scheme 5).
Scheme 6.
Of course, an intramolecular process could also be in-
volved. For instance, carbanion B could result directly from
carbanion A by intramolecular hydrogen transfer or vice
versa (lower line in Scheme 5).
Moreover, due to the fact that the structures of interest
were formed within only five minutes, and taking into ac-
Thus, it seems quite clear that the structures of the pro- count the fact that only allenic or spiro structures were gen-
ducts obtained depend on the cycloalkanone ring size (n erated in situ, we can conclude that i. the reaction takes
value), the length of the spacer (m value), and the reaction place under kinetic control, ii. the two kinds of products
time. By combining these variables, we can synthesize a rel- (allenic or spiro derivatives) are not interconvertible, that
Eur. J. Org. Chem. 2005, 4346–4358
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4349

A. J. Mota, A. Klein, F. Wendling, A. Dedieu, M. Miesch
FULL PAPER
is, the A1 &rlhar2 B1 equilibrium does not exist, and iii. a disposition of the chain provides a greater number of nega-
relationship exists between the combined m and n values tive charge–hydrogen-stabilizing interactions (Table 4).
and the major product formed.
Table 4. AM1- and PM3-calculated relative energies of carbanions
In addition, it should be mentioned that we were unable
to isolate the acetylenic derivatives represented as carban-
ion A1 because they evolve towards the allenic form A2,
which is expected to be less stable.
A, B, and C for compounds 1 and 4.[a]
PM3 energy
Entry Anion
AM1 energy [kcalmol–1]
[kcalmol–1]
In order to clarify these different points, we carried out
an extensive semiempirical study.
1
2
3
4
5
6
7
1A
0
0
1B
+5.4
+12.5
0
+2.4
+6.1
+11.4
+2.2
+6.2
0
+1.0
+0.6
+11.1
1C
4Aax
4Aeq
4B
Semiempirical MO Study
4C
For this study, we focussed our attention on acetylenic
ω-keto esters 1 and 4 which generate a single class of deriva-
tive. We considered that three different carbanions (A, B,
and C) could be formed in each case, which will be referred
to as 1A–C and 4A–C, depending on the starting com-
pound and the type of carbanion formed. In the A-type
carbanions of compound 4 (cyclohexanone ring), the alky-
nyl chain may be present in either an equatorial or an axial
position,[14] corresponding, respectively, to isomers 4Aeq
and 4Aax (Scheme 7).
[a] See Scheme 1 within the Supporting Information for a graphical
representation of the energies with respect to the reaction coordi-
nate.
Note that carbanions A and B have very similar energies
according to the PM3 calculations. Although B anions can
be obtained directly by the reaction of TBAF with the start-
ing materials, the most stable conformations of anions A
and B are related by a hydrogen atom that can easily be
transferred between the two centres of negative charge. In-
deed, it can be seen that these hydrogen atoms are close
enough (2.646 and 2.847 Å, respectively) to allow this trans-
fer process (Figure 1).[15]
PM3 calculations show that this hydrogen-transfer pro-
cess occurs with a low activation energy, as well as there
being small energy differences between the A and B carban-
ions. This led us to presume that both carbanions are prob-
ably involved in a dynamic equilibrium (Scheme 6, Table 5).
Conversely, Tables 4 and 5 show that AM1 calculations give
particularly high energies for these systems.
Note that hydrogen transfer is much more likely for com-
pound 4. This means that a six-membered transition state
is preferred for this process. In addition, the chain in carb-
anion 4A can have either an equatorial or an axial confor-
mation. The results demonstrate, however, that the transfer
process in an axial conformation is particularly unstable rel-
ative to that in an equatorial position (ca. 3–4.5 kcalmol–1).
In fact, in the transition state the system changes to a boat
conformation in order to obtain an equatorial position for
the chain. Therefore, the chain must be in an equatorial
position in the starting carbanion (4A) in order to optimize
the hydrogen-transfer process (Figure 2). Note also that en-
ergy values obtained from AM1 calculations are signifi-
cantly higher than those obtained by PM3 calculations.[16]
Scheme 7.
In order to obtain energy minima for all the species, we
first carried out a systematic conformational analysis em-
ploying the AM1 method. Next, the geometries were opti-
mized by performing PM3 calculations. In all cases we in-
variably found that the A-type anions were the most stable.
There are two main reasons for this particularly effective
stabilization. First, A-type carbanions have the most ex-
tended charge delocalization. Secondly, the greater the
charge delocalization, the greater the number of possible
attractive intramolecular interactions with hydrogen atoms.
Similarly, the resonant structure of B-type carbanions,
which has a more substituted C=C double bond, is thermo-
Allenic versus Spiro Pathways
In the formation of allenic derivatives, the diastereoselec-
dynamically more stable than the C-type anions. The latter tivity of the bicycle formation must be considered. Indeed,
are relatively high in energy from a thermodynamic point- carbanion A could attack the carbonyl carbon atom either
of-view and we did not detect any compounds derived from on the α-face (cis addition) or on the β-face (trans addition)
them (i.e. functionalized bridged bicyclic systems). This evi- (Scheme 8).
dence led us to discard them from this study. In the same
Another difference should be highlighted. Formation of
way, the axial conformers of A-type carbanions are, in gene- the allenic derivative implies a two-step process: 1) forma-
ral, slightly more stable than the corresponding equatorial tion of the A1-type carbanion and 2) hydrogen transfer to
conformers. This is essentially due to the fact that an axial give the A2-type carbanion. For the spiro compounds only
4350
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4346–4358

Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives
FULL PAPER
Figure 1. PM3 representations of the most stable conformers of anions 1A and 1B, and the transition state for hydrogen transfer (1A-B
TS) between them.
Table 5. Activation energy for the prototropy between anions A and
B.[a]
From Table 6 the following conclusions can be deduced.
First, the allenic derivatives are undoubtedly formed
through a cis addition process (Entries 2 and 3). This is not
surprising since diquinanes are mostly characterized by a
cis ring junction. The trans ring junction leads to highly
strained compounds (Figure 3). However, if we take into
account the relative energies of 1A and 1B (Table 4), the
absolute energies of the corresponding transition states (En-
tries 2 and 9) determined by AM1 calculations are approxi-
mately the same. This is a consequence of the large differ-
ence in the energies of the parent carbanions (1A and 1B,
respectively) found by this semiempirical method and shows
that treatment by the AM1 method is inadequate for this
particular chemical problem. From PM3 calculations these
carbanions have a more similar energy and therefore the
spiro pathway is always favored, as found experimentally.
Nevertheless, even if the energy difference between both
transition states is similar, the rate conversion in each case
is considerably different. Thus, at the beginning, molecules
have different choices: 1A can be transformed into 1B, 1A
can be transformed into 1A1cis, 1B can be transformed
back to 1A, and 1B can be transformed into 1B1. Statistic-
ally it seems clear that for the majority of the time 1A is
transformed into 1B and vice versa because they have the
most favorable activation energy (8.9 kcalmol–1according
to PM3). In fact, 1B will be converted faster into 1A be-
cause the activation energy for the process 1B Ǟ 1A is only
6.7 kcalmol–1 (Table 4).
Activation energy for the A Ǟ
Species
Method
B hydrogen transfer[b]
[kcalmol–1]
1A-B TS
AM1
PM3
AM1
PM3
AM1
PM3
+18.6
+8.9
+13.8 (+11.4)
+4.8 (+3.8)
+15.8
4Aeq-B TS[c]
4Aax-B TS[d]
+7.0
[a] See Scheme 1 within the Supporting Information for a graphical
representation of the energies with respect to the reaction coordi-
nate. [b] The energies are given relative to those of compounds 1
and 4 (zero energy) in Table 4. [c] Activation energy for the equato-
rial conformer. The energy values given in brackets are given rela-
tive to the energy of 4Aeq, the zero reference. [d] Activation energy
for the axial conformer. In the transition state, the chain is no
longer in an axial position but the system changes to a boat confor-
mation.
Thus, owing to the low energy barrier for the 1A Ǟ 1B
transformation we find at any given moment an appreciable
population of 1B in solution. Moreover, one must remem-
ber that 1B can be formed in solution by direct reaction
with TBAF, as we have suggested before, and this is proba-
bly the main source of 1B. Therefore, both species are pres-
ent in solution and will be transformed independently into
1A1cis or 1B1, respectively, depending on their correspond-
ing activation energies (Table 6). In addition, the more reac-
tive species in solution (namely the 1B anion) has the most
favorable activation energy (by several kcalmol–1) to pro-
Figure 2. PM3 six-membered transition states for the 4Aeq and
4Aax Ǟ 4B hydrogen transfer.
one step is required: the formation of the B1-type anion (see
Scheme 5).
Compound 1 (n = m = 1)
Table 6 summarizes the energy values of the first step in ceed along the reaction coordinate. This vision is strongly
the synthesis of allenic and spiro derivatives from 1 (forma- supported if the reaction occurs under kinetic control, as
tion of A1 and B1 carbanions, respectively). In both cases supported by the experimental data (see Table 1), and also
this step corresponds to the generation of a new C–C bond. applies to compound 4, which is studied in the next section.
Eur. J. Org. Chem. 2005, 4346–4358
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4351

A. J. Mota, A. Klein, F. Wendling, A. Dedieu, M. Miesch
FULL PAPER
Scheme 8.
Table 6. Energies of the carbanions 1A, 1A1, and 1A2 (considering the cis/trans addition modes), 1B, 1B1, and the corresponding transition
states. Allenic versus spiro compounds.[a]
Series
Entry
Anion
AM1 energy [kcalmol–1]
PM3 energy [kcalmol–1]
Allenic
1
2
3
4
5
6
7
8
9
10
1A
0
0
1Acis TS
1Atrans TS
1A1cis
1A1trans
1A2cis
1A2trans
1B
+16.4
+34.2
+13.7
+30.6
–9.7
+8.3
0
+14.6
+30.9
+12.1
+28.3
–7.1
+10.4
0
Spiro
1B TS
1B1
+11.7
–12.5
+11.1
–15.1
[a] See Scheme 2 within the Supporting Information for a graphical representation of the energies with respect to the reaction coordinate.
Figure 3. PM3 transition states for the 1A Ǟ 1A1 and 1B Ǟ 1B1 transformations.
Compound 4 (n = m = 2)
plained by taking two factors into account: first, the
delocalization of the negative charge in the allene system
(4A2transeq) is much more important than the more
localized charge in the alkoxide (4A1transeq) and secondly,
the exocyclic allene derivative (sp2-hybridized carbon) is
less strained than the acetylenic side-chain (sp3-hybridized
carbon).[18] Moreover, activation energies indicate that
trans diastereoselectivity would be expected for decalin
skeleton formation (Entries 5 and 4), as found experimen-
tally.
The energy values of the first step (C–C bond formation)
in the synthesis of allenic and spiro derivatives from 4 are
summarized in Table 7.
In contrast to the previous case (m = n = 1), note
that PM3 calculations show that the formation of allenic
derivatives is kinetically preferred (ca. 5 kcalmol–1) over
that of the spiro compound (Entries 5 and 13). However,
the AM1 method gives approximately the same energies
for both transition states (A Ǟ A1 and B Ǟ B1) (Fig-
ure 4). The energy difference between 4A1transeq and
In order to explain the formation of the allenic deriva-
4A2transeq explains why the former compound was never tives it is also necessary to calculate the transition-state en-
isolated. The reaction always progresses toward allene for- ergy of the second step, that is, the 4A1transeq Ǟ 4A2transeq
mation. This significant energy difference can be ex- conversion.
4352
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4346–4358

Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives
FULL PAPER
Table 7. Energy values for carbanions 4A, 4A1, and 4A2 considering the cis/trans addition, 4B, 4B1, and the corresponding transition
states.[16] Allenic vs. spiro compounds.[a]
Series
Entry
Anion
AM1 energy [kcalmol–1]
PM3 energy [kcalmol–1]
Allenic
1
2
3
4
5
6
7
8
4Aax
0
0
4Aeq
+2.4
+19.1
+17.4
+14.8
+15.0
+12.7
+10.8
–8.6
–11.4
–12.2
0
+14.1
–13.3
+1.0
+19.3
+18.9
+15.3
+14.3
+12.4
+12.1
–3.8
–6.2
–5.6
0
+20.4
–13.2
4Acisax TS
4Aciseq TS
4Atranseq TS[b]
4A1cisax
4A1ciseq
4A1transeq
4A2cisax
4A2ciseq
4A2transeq
4B
4B TS
4B1
[b]
9
10
11
12
13
14
[b]
Spiro
[a] See Scheme 3 within the Supporting Information for a graphical representation of the energies with respect to the reaction coordinate.
[b] The trans-axial isomer (4A1transax) is geometrically impossible.
Table 8. Energies of the direct 4A1transeq Ǟ 4A2transeq hydrogen-
transfer process.[a]
Hydogen
transfer
AM1 energy
[kcalmol–1]
PM3 energy
[kcalmol–1]
Anion
Direct
4A1transeq
4A1transeq TS
4A2transeq
0
0
+33.8
–23.0
+27.0
–17.8
[a] See Scheme 3 within the Supporting Information for a graphical
representation of the energies with respect to the reaction coordi-
nate. This direct hydrogen transfer can take place only if there is a
cis relative configuration between the bridge-head oxo anion and
the propargylic hydrogen. This means that the side-chain must be
in an axial position.[17]
Figure 4. PM3 transition states for the 4A Ǟ 4A1 and 4B Ǟ 4B1
Therefore, according to these observations, the spiro
compounds should once again be the major products, in
contrast to the experimental results, which show the forma-
tion of only allenic derivatives.
transformations.
The 4A1transeq Ǟ 4A2transeq Transformation
One possible way to achieve this is to diminish the strain
From the results displayed in Tables 6 and 7 it can be energy by avoiding the formation of the four-membered
seen that the energy of A1 lies about 10 kcalmol–1 above transition state. Thus, it should be noted that the formation
that of the starting material (A carbanion). The “normal” of the A carbanions induced by TBAF is accompanied by
evolution of the A1 carbanion would be to undergo pro- the generation of a hydrogen fluoride (HF) molecule. By
totropy to give carbanion A2. This process can take place studying the role of this species, we have found that, indeed,
directly by transfer of the chain’s base hydrogen to the al- HF is able to catalyze this hydrogen transfer.
ready formed alkoxide, which implies the formation of a
four-membered transition state (4A1transeq TS) (Table 8).
The catalytic effect of the HF molecule can be seen from
the data in Table 9. The presence of HF dramatically lowers
Since this transition state connects two atoms with an the activation energy for the hydrogen-transfer process from
axial/equatorial relative configuration, a large deformation ca. 27 (Table 8) to ca. 15 kcalmol–1. The proposed mecha-
of the natural angles is found (132° for a sp3 carbon!), as nism for this process is depicted in Scheme 9.
depicted in Figure 5. This is the main reason for the high
activation energies calculated for this process.
We suggest that HF can interact with the alkoxide
4A1transeq through the formation of two hydrogen-like
Eur. J. Org. Chem. 2005, 4346–4358
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4353

A. J. Mota, A. Klein, F. Wendling, A. Dedieu, M. Miesch
FULL PAPER
bonds. According to PM3 calculations, hydrogen transfer
occurs by a two-step mechanism that involves the genera-
tion of two six-membered transition states, in contrast to
the “direct” transfer in which a four-membered transition
state was required. The first step of this mechanism
(4A1transeq-HF Ǟ 4A1transeq-HF inter) involves the proton-
ation of the alkoxide anion by breaking the HF bond.
Then, the generated fluoride anion picks up the propargylic
hydrogen atom (4A1transeq-HF inter Ǟ 4A2transeq-HF) to
finally afford the corresponding allene derivative.
Nevertheless, in order for the A Ǟ A1 and B Ǟ B1 trans-
formations to have comparable energies, it is necessary to
recalculate their energies in the presence of HF. These val-
ues are collected in Table 10.
Figure 5. Transition state for direct hydrogen transfer in compound
4.
Table 10. Energies of the anions 4Aax, 4A1transeq, 4B, 4B1, and the
corresponding transition states. Allenic versus spiro compounds in
the presence of HF.[a]
Table 9. Energies of the HF-catalyzed 4A1transeq Ǟ 4A2transeq hy-
drogen-transfer process.[a]
Series
Anion
PM3 energy [kcalmol–1]
Allenic
4Aax-HF
0
4Atranseq-HF TS
4A1transeq-HF
4B-HF
+16.3
+5.3
0
Spiro
4B-HF TS
4B1-HF
+22.2
–12.6
[a] See Schemes 4 and 5 within the Supporting Information for a
graphical representation of the energies with respect to the reaction
coordinate.
H-Transfer
Anion
PM3 energy [kcalmol–1]
HF-Catalyzed
4A1transeq-HF
4A1transeq-HF TS1
E1act
4A1transeq-HF inter
4A1transeq-HF TS2
E2act
0
+15.0
+15.0
+8.7
+9.1
+0.4
–7.9
Compound 4A1transeq-HF lies ca. 5 kcalmol–1 above the
starting material. This means that the presence of HF
greatly stabilizes the intermediate 4A1transeq, which was ca.
12 kcalmol–1 above 4A in the absence of HF (Table 7, En-
try 8). It has been shown (Table 9) that the activation en-
4A2transeq-HF
[a] See Scheme 4 within the Supporting Information for a graphical ergy for hydrogen transfer in this intermediate is ca.
15 kcalmol–1 (4Atranseq-HF TS), and therefore this transi-
representation of the energies with respect to the reaction coordi-
nate. For this case, only the PM3 method can properly treat the
tion state is ca. 20 kcalmol–1 above 4A. This process is the
presence of the fluorine atom.
most energy-demanding in the allenic pathway. However,
Scheme 9.
4354
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4346–4358

Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives
FULL PAPER
the transition state for the 4B-HF Ǟ 4B1-HF transforma- finite possibilities, but we think that the most probable is
tion is 22 kcalmol–1 above the starting material, showing the easiest, and once the fluoride anion picks up a hydrogen
that the presence of HF favors the allenic pathway, in ac- atom, the HF species then formed remains in the neighbor-
cord with the experimental results. Of course, 2 kcalmol–1 hood of the anion because of strong electrostatic interac-
in the correct sense (that is to say, in agreement with the tions between the anion and the HF molecule. Secondly,
experimental results) is not a big difference in energy, but it the fluoride atom in HF needs to be surrounded by the
is necessary to recognize that in the absence of the HF maximum number of hydrogen atoms in order to be stabi-
molecule this value was much higher in the opposite sense. lized. It seems normal that a very small molecule like HF,
Therefore, we can say that a drastic change in the reaction with strong interactions with the molecules under study and
preference has taken place, and 2 kcalmol–1 becomes a sig- in the proximity of the reaction centre, can participate more
nificant amount for this particular problem.
easily, catalyzing the process. Furthermore, treatment of
these molecules with other bases such as NaH, K2CO3, Na-
OEt, tBuOK, did not afford any allene derivatives.[3] This
probably means that the presence of HF is absolutely neces-
sary.
Compound 1 (n = m = 1) Revisited
Owing to the large effect of HF on the energies calcu-
lated for the transformations of compound 4, it is also nec-
essary to consider the effect of HF on the transformations
of carbanion 1. In this case, more significant differences
were obtained in the presence of HF, as can be seen in
Table 11.
Conclusions
The synthesis of allenic or spiro derivatives from acetyle-
nic ω-keto esters can be modulated by changing the ring
size (n value), the chain length (m value), or the reaction
time. Compounds 1 and 4 have been shown to be suitable
starting materials because only one type of product was
formed (spiro or allenic derivatives) from them. On the
other hand, with regard to the products observed, no evi-
dence for allene–spiro interconversion has been found.
Thus, it is clear that the reactions take place under kinetic
control. In order to justify the complex behavior of these
kinds of products, a semiempirical study was carried out
to account for the products formed in each case and the
stereochemistry of the bicyclic system generated in the case
of the allenic derivatives. In all cases (allenic or spiro), an
intramolecular reaction between the frontier orbitals of A
(for allenic derivatives) or B (for spiro compounds) occurs.
The activation energy of this process varies as a function of
both the size of the cycloalkanone moiety and the chain
length. A correct explanation for the whole process can be
obtained from PM3 calculations and involves an HF-cata-
lyzed reaction that is very important for compound 4. In
addition, from a qualitative point-of-view, an excellent cor-
relation between the experimental results and the theoreti-
cal calculations has been found. For compound 4, good
agreement with the experimental results is found only if we
consider that kinetic control (activation energies) governs
the reaction.
Table 11. Energies of the anions 1A, 1A1cis, 1B, 1B1, and the corre-
sponding transition states. Allenic versus spiro compounds in the
presence of HF.[a]
Series
Anion
PM3 energy [kcalmol–1]
Allenic
1A-HF
0
1Acis-HF TS
1A1cis-HF
1A1cis-HF TS1
E1act
+12.8
+4.3
+19.4
+15.1[b]
+13.5
+13.8
+0.2
–6.5
1A1cis-HF inter
1A1cis-HF TS2
E2act
1A2cis-HF
1B-HF
Spiro
0[c]
1B-HF TS
1B1-HF
+15.1
–18.7
[a] See Schemes 6 and 7 within the Supporting Information for a
graphical representation of the energies with respect to the reaction
coordinate. [b] The activation energy for the non-HF-catalyzed pro-
cess (1A1cis TS species) is 21.4 kcalmol–1 (PM3). See Scheme 2 in
the Supporting Information. [c] The 1B-HF structure is only
0.6 kcalmol–1 more stable than the 1A-HF isomer according to the
PM3 calculations.
In this case, compound 1A1cis-HF is only ca. 4 kcalmol–1
above the starting material, and the hydrogen transfer pro-
cess in this intermediate has an activation energy of ca.
15 kcalmol–1 (therefore the transitions state is ca.
19 kcalmol–1 above 1A). However, for the 1B-HF Ǟ 1B1-
HF transformation, the activation energy is ca.
15 kcalmol–1. Thus, the expected product is the spiro deriv-
ative, in accord with the experimental results, and in line
with previous calculations (Table 6).
We have analyzed proton transfer in other species. For
instance, it is known that commercial TBAF contains a
small amount of water. Thus, calculations carried out with
water as a catalyst (instead of HF) showed similar behavior,
but the activation energies are slightly higher. Other calcula-
tions have been carried out involving both water and HF,
but the best results were obtained when only HF molecules
Experimental Section
General Remarks: Reactions were carried out under argon with
magnetic stirring and by using degassed solvents. Et2O and THF
were distilled from Na/benzophenone. Thin-layer chromatography
(TLC) was carried out on silica gel plates and the spots were visual-
ized under a UV lamp (254 or 365 nm) and/or sprayed with a solu-
tion of vanillin (25 g) in EtOH/H2SO4 (98:2; 1 L) or with phos-
phomolybdic acid followed by heating on a hot plate. For column
chromatography, silica gel (40–60 μm) was used. Melting points
(m.p.) were measured on a hot stage apparatus. IR spectra were
1
were considered. Evidently, in solution, there are almost in- recorded as CCl4 solutions. H NMR spectra were recorded at 200
Eur. J. Org. Chem. 2005, 4346–4358
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4355

A. J. Mota, A. Klein, F. Wendling, A. Dedieu, M. Miesch
FULL PAPER
or 300 MHz and 13C NMR spectra at 50 or 75 MHz using the
3.00 (m, 2 H), 4.10 (q, J = 7.1 Hz, 2 H), 5.85 (t, J = 1.9 Hz, 1
H) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ = 14.2, 20.3, 23.2,
36.3, 37.2, 37.4, 39.0, 59.6, 60.9, 112.71, 165.58, 168.82,
219.28 ppm.
signal of the residual nondeuteriated solvent as the internal refer-
1
ence. Significant H NMR spectroscopic data are tabulated in the
following order: chemical shift (δ) expressed in ppm, multiplicity (s,
singlet; d, doublet; t, triplet; q, quadruplet; m, multiplet), coupling
constants J in Hertz, number of protons. The ratios of compounds
indicated below were calculated from the NMR integrations.
1
Spiro Ketone 5c: Colorless oil. IR (CCl ): ν = 1712, 1650 cm–1. H
˜
4
NMR (300 MHz, CDCl3, 25 °C): δ = 1.23 (t, J = 7.1 Hz, 3 H),
1.35–2.15 (m, 7 H), 2.25–2.45 (m, 3 H), 2.94 (ABXYZ, JAB = 15.3,
JAX = JAY = JAZ = 1.4, JBX = JBY = JBZ = 2.1 Hz, δA = 2.88, δB
= 3.00, 2 H), 4.09 (q, J = 7.1 Hz, 2 H), 5.88 (m, 1 H) ppm. 13C
NMR (75 MHz, CDCl3, 25 °C): δ = 14.1, 19.9, 30.0, 34.2, 34.3,
36.9, 37.4, 61.0, 64.2, 132.9, 136.6, 171.0, 221.0 ppm. C13H18O3
(222.28): C 70.24, H 8.16, O 21.59; found C 70.18, H 8.12.
General Procedure for the Treatment of Compounds 1–4 with TBAF:
TBAF [1 m solution in THF (1.1 equiv.)] was added to a stirred
solution of acetylenic ω-keto esters (1 equiv., 0.4 mmol) in dry THF
(5 mL) at room temperature. The yellow solution was stirred at
the specified temperature for 30 min, hydrolyzed with a saturated
aqueous NH4Cl solution (10 mL), washed with water (3×10 mL),
and extracted with Et2O (2×10 mL). The organic layers were then
washed with a saturated NaCl solution (10 mL), dried with
MgSO4, filtered, and concentrated under reduced pressure (15 Torr,
25 °C). The residue was purified by chromatography on a silica gel
column (15 g SiO2; ethyl acetate/hexane, 1:99 to 15:85) to afford
the different products formed.
Compound 6: Colorless oil. IR (CCl ): ν = 3500, 2959, 1739,
˜
4
1652 cm–1. 1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.23 (t, J =
7.1 Hz, 3 H), 1.72 (q, J = 7.3 Hz, 2 H), 1.78 (quint., J = 7.1 Hz, 2
H), 1.85–2.05 (m, 2 H), 2.13 (t, J = 7.3 Hz, 2 H), 2.29 (t, J =
7.5 Hz, 2 H), 3.04 (s, 1 H), 4.10 (q, J = 7.1 Hz, 2 H) ppm. 13C
NMR (75 MHz, CDCl3, 25 °C): δ = 14.1, 22.8, 23.4, 27.6, 33.3,
34.4, 35.4, 36.3, 60.5, 129.1, 138.1, 171.6, 177.0 ppm. C13H20O4
(240.30): C 64.98, H 8.39, O 26.63; found C 65.08, H 8.38.
Compound 1: Colorless oil. IR (CCl ): ν = 2240, 1714 cm–1. 1H
˜
4
NMR (200 MHz, CDCl3, 25 °C): δ = 1.30 (t, J = 7.1 Hz, 3 H),
1.38–188 (m, 8 H), 2.00–2.33 (m, 3 H), 2.34 (t, J = 7.3 Hz, 2 H),
4.22 (q, J = 7.1 Hz, 2 H) ppm. 13C NMR (50 MHz, CDCl3, 25 °C):
δ = 14.1, 18.8, 20.8, 25.7, 29.1, 29.7, 38.1, 48.7, 61.8, 73.5, 88.7,
153.8, 220.7 ppm. C13H18O3 (222.28): C 70.24, H 8.16, O 21.59;
found C 70.01, H 8.22.
Spiro Ketone 7a: Colorless oil. IR (CCl ): ν = 1739, 1714,
˜
4
1635 cm–1. 1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.22 (t, J =
7.1 Hz, 3 H), 1.35–1.95 (m, 8 H), 2.00–2.40 (m, 5 H), 2.66 (dt, J =
9.7, 18.6 Hz, 1 H), 4.06 (ABX3, JAX = 7.1, JBX = 7.1, JAB
=
13.5 Hz, δA = 4.09, δB = 4.09, 2 H), 5.72 (s, 1 H) ppm. 13C NMR
(75 MHz, CDCl3, 25 °C): δ = 14.2, 19.2, 19.6, 23.8, 30.9, 34.2, 34.7,
37.1, 54.8, 59.9, 115.0, 159.1, 166.5, 218.3 ppm. C14H20O3 (236.31):
C 71.16, H 8.53, O 20.31; found C 71.22, H 8.42.
Compound 2: Colorless oil. IR (CCl ): ν = 2239, 1742, 1714 cm–1.
˜
4
1H NMR (200 MHz, CDCl3, 25 °C): δ = 1.30 (t, J = 7.1 Hz, 3 H),
1.15–2.30 (m, 11 H), 2.34 (t, J = 6.8 Hz, 2 H), 4.21 (q, J = 7.1 Hz,
2 H) ppm. 13C NMR (50 MHz, CDCl3, 25 °C): δ = 13.9, 18.4, 20.7,
26.8, 27.6, 29.3, 29.5, 37.7, 48.5, 61.4, 74.2, 88.6, 153.7, 217.5 ppm.
C14H20O3 (236.31): C 71.16, H 8.53, O 21.31; found C 71.17, H
8.66.
Spiro Ketone 7b: Colorless oil. IR (CCl ): ν = 1739, 1717,
˜
4
1635 cm–1. 1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.17 (t, J =
7.1 Hz, 3 H), 1.30–1.55 (m, 3 H), 1.60–1.90 (m, 6 H), 2.10–2.40 (m,
4 H), 3.51 (dt, J = 4.2, 14.0 Hz, 1 H), 4.04 (qd, J = 2.1, 7.1 Hz, 2
H), 5.34 (s, 1 H) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ =
14.1, 18.7, 21.8, 26.2, 27.1, 35.2, 35.7, 38.5, 57.0, 59.6, 113.8, 161.4,
166.3, 219.5 ppm. C14H20O3 (236.31): C 71.16, H 8.53, O 20.31;
found C 71.18, H 8.44.
Compound 3: Colorless oil. IR (CCl ): ν = 2239, 1712 cm-1. 1H
˜
4
NMR (300 MHz, CDCl3, 25 °C): δ = 1.24 (t, J = 7.1 Hz), 1.25–
1.40 (m, 3 H), 1.45–1.70 (m, 6 H), 1.71–1.90 (m, 2 H), 1.91–2.10
(m, 2 H), 2.28 (td, J = 1.5, 7.2 Hz, 2 H), 4.14 (q, J = 7.1 Hz, 2
H) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ = 14.1, 18.8, 24.9,
25.3, 27.9, 28.7, 33.9, 40.0, 50.1, 61.6, 73.2, 88.9, 153.7, 212.6 ppm.
C14H20O3 (236.31): C 71.16, H 8.53, O 21.31; found C 71.19, H
8.66.
1
Spiro Ketone 7c: Colorless oil. IR (CCl ): ν = 1750, 1648 cm–1. H
˜
4
NMR (300 MHz, CDCl3, 25 °C): δ = 1.25 (t, J = 7.1 Hz, 3 H),
1.50–2.50 (m, 12 H), 2.82 (AB, JAB = 15.1 Hz, δA = 2.96, δB = 2.68,
2 H), 4.10 (q, J = 7.1 Hz, 2 H), 5.89 (t, J = 3.7 Hz, 1 H) ppm. 13C
NMR (75 MHz, CDCl3, 25 °C): δ = 14.2, 19.2, 19.8, 23.8, 30.9,
34.7, 35.0, 37.1, 54.7, 59.9, 115.0, 159.1, 166.5, 218.3 ppm.
C14H20O3 (236.31): C 71.16, H 8.53, O 20.31; found C 71.30, H
8.62.
Compound 4: Colorless oil. IR (CCl ): ν = 2240, 1714 cm–1. 1H
˜
4
NMR (300 MHz, CDCl3, 25 °C): δ = 1.27 (t, J = 7.1 Hz), 1.35–
2.25 (m, 15 H), 2.32 (t, J = 7.1 Hz), 4.18 (q, J = 7.1 Hz, 2 H) ppm.
13C NMR (75 MHz, CDCl3, 25 °C): δ = 14.0, 18.5, 24.9, 26.4, 27.6,
27.99, 28.7, 33.9, 42.0, 50.5, 61.7, 73.2, 89.1, 153.8, 213.1 ppm.
C15H22O3 (250.34): C 71.97, H 8.86, O 19.17; found C 72.03, H
9.00.
Compound 9: Colorless oil. IR (CCl ): ν = 1746, 1682, 1614 cm–1.
˜
4
1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.24 (t, J = 7.1 Hz, 3 H),
1.35–1.65 (m, 3 H), 1.70–2.05 (m, 5 H), 2.10–2.40 (m, 3 H), 2.45–
2.90 (m, 2 H), 3.54 (AB, JAB = 15.7 Hz, δA = 3.57, δB = 3.51, 2
H), 4.17 (q, J = 7.1 Hz, 2 H) ppm. 13C NMR (75 MHz, CDCl3,
25 °C): δ = 14.1, 22.0, 23.7, 26.0, 27.2, 32.2, 33.1, 45.1, 47.8, 61.0,
127.1, 162.5, 168.1, 194.3 ppm. C14H20O3 (236.31): C 71.16, H 8.53,
O 20.31; found C 71.22, H 8.42.
Spiro Ketones 5: Spiro ketones 5a and 5b (mixture, as a colorless
oil) were not separated. IR (mixture 5a + 5b, CCl ): ν = 1742, 1714,
˜
4
1652 cm–1. C13H18O3 (222.28) (mixture 5a + 5b): C 70.24, H 8.16,
O 21.59; found C 70.20, H 8.32.
Allene 10a: Colorless crystals with m.p. 65–66 °C. IR (CCl ): ν =
˜
4
1
Spiro Ketone 5a: H NMR (300 MHz, CDCl3, 25 °C): δ = 1.22 (t,
3601, 1964, 1717 cm–1. 1H NMR (300 MHz, CDCl3, 25 °C): δ =
1.21 (t, J = 7.1 Hz, 3 H), 1.35–1.65 (m, 9 H), 1.80–1.95 (m, 2 H),
2.00–2.15 (m, 1 H), 2.60–2.70 (m, 2 H), 4.11 (ABX3, JAB = 10.8,
JBX = 7.1 Hz, δA = 4.07, δB = 4.15, 2 H), 5.59 (t, J = 4.1 Hz, 1
H) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ = 14.1, 23.1, 24.3,
26.4, 28.0, 28.6, 34.4, 47.3, 60.6, 83.3, 90.8, 112.3, 166.1,
J = 7.1 Hz, 3 H), 1.55–2.15 (m, 8 H), 2.20–2.45 (m, 2 H), 2.50–
2.60 (m, 1 H), 2.85–3.00 (m, 1 H), 4.03 (qd, J = 2.2, 7.1 Hz, 2 H),
5.50 (t, J = 2.5 Hz) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ =
14.2, 19.5, 23.5, 33.5, 36.8, 38.2, 38.3, 59.7, 62.5, 113.0, 166.4,
170.5, 219.6 ppm.
1
Spiro Ketone 5b: H NMR (300 MHz, CDCl3, 25 °C): δ = 1.18 (t, 207.0 ppm. C14H20O3 (236.31): C 71.16, H 8.53, O 20.31; found C
J = 7.1 Hz, 3 H), 1.55–2.15 (m, 8 H), 2.20–2.45 (m, 2 H), 2.85–
71.05, H 8.70.
4356
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4346–4358

Controlled Synthesis of α-Allenic Ester and Spiro Ketone Derivatives
FULL PAPER
46–126; i) A. J. McCarroll, J. C. Walton, J. Chem. Soc., Perkin
Trans. 1 2001, 3215–3229; j) A. J. McCarroll, J. C. Walton, An-
gew. Chem. Int. Ed. 2001, 40, 2225–2248; Angew. Chem. 2001,
113, 2282–2307; k) R. Schobert, G. J. Gordon, Curr. Org.
Chem. 2002, 6, 1181–1196.
M. Miesch, Synthesis 2004, 746–752.
a) F. Wendling, M. Miesch, Org. Lett. 2001, 3, 2689–2691; b)
A. Klein, M. Miesch, Tetrahedron Lett. 2003, 44, 4483–4485.
a) B. L. Roy, P. Deslongchamps, Can. J. Chem. 1986, 64, 2068–
2075; b) J. F. Lavallée, G. Berthiaume, P. Deslongchamps, Tet-
rahedron Lett. 1986, 27, 5455–5458; c) J. F. Lavallée, P. De-
slongchamps, Tetrahedron Lett. 1987, 28, 3457–3458.
Allene 10b: Colorless oil. IR (CCl ): ν = 3500, 3392, 1963,
˜
4
1716 cm–1. 1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.26 (t, J =
7.1 Hz, 3 H), 1.35–1.70 (m, 10 H), 1.80–2.00 (m, 1 H), 2.05–2.25
(m, 1 H), 2.50–2.70 (m, 1 H), 2.71–2.85 (m, 1 H), 4.16 (q, J =
7.1 Hz, 2 H), 5.69 (t, J = 4 Hz, 1 H) ppm. 13C NMR (75 MHz,
CDCl3, 25 °C): δ = 14.2, 23.6, 24.0, 26.4, 28.1, 28.3, 34.1, 46.8,
60.8, 83.0, 91.0, 113.6, 166.72, 206.78 ppm. C14H20O3 (236.31): C
71.16, H 8.53, O 20.31; found C 71.32, H 8.69.
[2]
[3]
[4]
Spiro Ketone 11a: Colorless oil. IR (CCl ): ν = 1709, 1657 cm–1.
˜
4
1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.26 (t, J = 7.1 Hz, 3 H),
1.50–1.90 (m, 8 H), 1.95–2.15 (m, 2 H), 2.25–2.65 (m, 2 H), 2.90
(ddd, J = 2.5, 6.7, 7.7 Hz, 2 H), 4.14 (q, J = 7.3 Hz, 2 H), 5.67 (t,
J = 2.5 Hz, 1 H) ppm. 13C NMR (75 MHz, CDCl3, 25 °C): δ =
14.2, 22.1, 22.8, 26.8, 33.0, 37.1, 38.1, 39.1, 59.6, 63.4, 114.3, 166.7,
168.6, 211.0 ppm. C14H20O3 (236.31): C 71.16, H 8.53, O 20.31;
found C 71.13, H 8.48.
[5]
[6]
J. Christoffers, H. Oertling, Tetrahedron 2000, 56, 1339–1344.
a) M. Ihara, T. Taniguchi, K. Makita, M. Takano, M. Ohnishi,
N. Taniguchi, K. Fukumoto, C. Kabuto, J. Am. Chem. Soc.
1993, 115, 8107–8115 and references cited therein; b) M. Ihara,
K. Fukumoto, Angew. Chem. Int. Ed. Engl. 1993, 32, 1010–
1022; Angew. Chem. 1993, 105, 1059–1071; c) K. Takasu, M.
Ueno, M. Ihara, J. Org. Chem. 2001, 66, 4667–4672 and refer-
ences cited therein.
a) P. Chiu, C. P. Szeto, Z. Geng, K. F. Cheng, Org. Lett. 2001,
3, 1901–1903; b) P. Chiu, B. Chen, K. F. Cheng, Tetrahedron
Lett. 1998, 39, 9229–9232.
Spiro Ketone 11b: Colorless oil. IR (CCl ): ν = 1709, 1739 cm–1.
˜
4
1H NMR (300 MHz, CDCl3, 25 °C): δ = 1.24 (t, J = 7.1 Hz, 3 H),
1.50–2.15 (m, 8 H), 2.20–2.50 (m, 4 H), 3.04 (ABXYZ, JAX = JAY
= JAZ = 2.0 Hz, 2 H), 4.12 (q, J = 7.1 Hz, 2 H), 5.77 (dddd, J =
4.4, 1.3, 2.3, 3.6 Hz, 1 H) ppm. 13C NMR (75 MHz, CDCl3, 25 °C):
δ = 14.1, 22.1, 26.4, 29.7, 34.3, 35.6, 35.8, 39.6, 60.5, 64.6, 130.3,
138.2, 171.7, 213.2 ppm. C14H20O3 (236.31): C 71.16, H 8.53, O
20.31; found C 71.21, H 8.65.
[7]
[8]
a) J. B. Foresman, Æ. Frisch, Exploring Chemistry with Elec-
tronic Structure Methods, 2nd ed., Gaussian, Inc., Pittsburgh
PA, 1996; b) F. Jensen, Introduction to Computational Chemis-
try, John Wiley & Sons Ltd., Chichester, 1999.
[9]
For examples, see: a) J. W. McIver Jr., A. Komornicki, J. Am.
Chem. Soc. 1972, 94, 2625–2633; b) C. S. Wu, W. C. Neely,
S. D. Worley, J. Comput. Chem. 1991, 12, 862–867; c) Y. L.
Dory, P. Soucy, M. Drouin, P. Deslongchamps, J. Am. Chem.
Soc. 1995, 117, 518–529; d) W. M. F. Fabian, G. Kollenz, J.
Chem. Soc., Perkin Trans. 2 1995, 515–518; e) A. B. J. Parusel,
R. Schamschule, D. Piorun, K. Rechthaler, A. Puchala, D. Ras-
ala, K. Rotkiewicz, G. Köhler, J. Mol. Struct. 1997, 419, 63–
75; f) Z. Markovic, S. Konstantinovic, I. Juranic, L. Dosen-
Micovic, Gazz. Chim. Ital. 1997, 127, 429–434; g) L. Lorenc,
V. Pavlovich, I. Juranich, M. L. Mihailovich, L. Bondarenko-
Gheorghiu, N. Krstich, M. Dabovich, Molecules 1999, 4, 272–
278; h) M. Grotjahn, N. Jäger, H.-J. Drexler, H.-J. Holdt, E.
Kleinpeter, J. Mol. Model. 1999, 5, 72–77; i) S. Bohm, R. Pohl,
J. Kuthan, Collec. Czech. Chem. Commun. 1999, 64, 1761–1769;
j) P. Merino, J. A. Mates, Arkivoc 2001, xi, 12–30; k) A. J.
Mota, R. Robles, L. Álvarez de Cienfuegos, A. Lamenca, Tet-
rahedron Lett. 2004, 45, 3349–3353; l) L. Álvarez de Cien-
fuegos, A. J. Mota, C. Rodríguez, R. Robles, Tetrahedron Lett.
2005, 46, 469–473.
A. J. Mota, E. Castellanos, L. Álvarez de Cienfuegos, R.
Robles, Tetrahedron: Asymmetry 2005, 16, 1615–1629.
Hyperchem™, Hypercube, Inc., 1115 NW 4th Street, Gaines-
ville, Florida 32601, USA.
a) H. B. Bürgi, J. D. Dunitz, J.-M. Lehn, G. Wipff, J. Am.
Chem. Soc. 1974, 96, 1956–1957; b) H. B. Bürgi, J. M. Lehn,
Tetrahedron 1974, 30, 1563–1572.
CCDC-271434 (for 10a), and -271435 (for 12a) contain the
supplementary crystallographic data for this paper. These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
The spectroscopic data for the allene derivatives 8a, 8b and 12a,
12b have already been reported.[3a]
Computational Details: The computational study was performed by
means of AM1 and PM3 semiempirical MO methods implemented
in the HyperchemTM 7.5 package. Close-shell type calculations
(RHF) were carried out, unless otherwise indicated, using the Po-
lak–Ribiere optimization algorithm, except for the transition states
for which an eigenvector-following algorithm was employed. Con-
vergence limits for geometry optimization were fixed at
0.01 kcalÅ–1 mol–1 (RMS gradient). The limit for the iterative SCF
calculations was fixed at 0.001 kcalmol–1. Transition states were
evaluated from estimated initial geometries using a transition-state
search implemented in the HyperchemTM 7.5 package by means of
a trial-and-error sequence. This transition state is characterized by
its unique negative frequency. Geometries for all the starting carb-
anions were obtained after a systematic conformational search
using the semiempirical methods mentioned above by changing all
the dihedral angles in 30° steps for the acyclic torsion variations.
The energies and Cartesian coordinates of all the structures studied
are included in the Supporting Information.
[10]
[11]
[12]
Acknowledgments
The authors thank the Junta de Andalucía and the Universidad de
Granada for the research contract to A.J.M., and the CNRS and
ULP for financial support. We also thank Dr. Jennifer Wytko for
her help in preparing the manuscript.
[13]
[14]
[15]
Related to this see: a) K. L. Servis, D. J. Bowler, C. Ishii, J.
Am. Chem. Soc. 1975, 97, 73–80; b) E. A. Basso, C. Kaiser, R.
Rittner, J. B. Lambert, J. Org. Chem. 1993, 58, 7865–7869.
[1] a) L. F. Tietze, U. Beifuss, Angew. Chem. Int. Ed. Engl. 1993,
32, 131–163; Angew. Chem. 1993, 105, 137–170; b) L. F. Tietze,
Chem. Rev. 1996, 96, 115–136; c) P. J. Parsons, C. S. Penkett,
A. J. Shell, Chem. Rev. 1996, 96, 195–206; d) A. Padwa, M. D.
Weingarten, Chem. Rev. 1996, 96, 223–269; e) M. Malacria,
Chem. Rev. 1996, 96, 289–306; f) E.-I. Negishi, C. Copéret, S.
Ma, S.-Y. Liou, F. Liu, Chem. Rev. 1996, 96, 365–393; g) K.
Neuschütz, J. Velker, R. Neier, Synthesis 1998, 227–255; h)
K. C. Nicolaou, D. Vourloumis, N. Winssinger, P. S. Baran, An-
gew. Chem. Int. Ed. 2000, 39, 44–122; Angew. Chem. 2000, 112,
a) B. P. Roberts, J. Chem. Soc., Perkin Trans. 2 1996, 2719–
2725; b) M. Ramos, I. Alkorta, J. Elguero, Tetrahedron 1997,
53, 1403–1410; c) H. F. Dos Santos, L. F. C. De Oliveira, S. O.
Dantas, P. S. Santos, W. B. De Almeida, Int. J. Quantum Chem.
2000, 80, 1076–1086; d) P. E. Allegretti, M. S. Cortizo, C.
Guzmán, E. A. Castro, J. J. P. Furlong, Arkivoc 2003, x, 24–31;
e) C. Ögretir, S. Yarhgan, H. Berber, T. Arslan, S. Topal, J.
˘
Mol. Model. 2003, 9, 390–394.
Eur. J. Org. Chem. 2005, 4346–4358
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4357

A. J. Mota, A. Klein, F. Wendling, A. Dedieu, M. Miesch
FULL PAPER
[16] Higher level DFT (B3LYP/6-31+G*) calculations were carried
out for the 1A-B hydrogen transfer to test both the accuracy
of semiempirical calculations and the treatment of anions. We
found that this transfer occurs with an activation energy of
+15.3 kcalmol–1, compared with values of +18.6 (AM1) and
+8.9 kcalmol–1 (PM3) found by semiempirical calculations.
[17] The formation of 4A1 derivatives implies the generation of two
diastereomers depending on the axial or equatorial position of
the electrophilic acetylene side-chain after the cyclisation pro-
cess. The energies given in Table 7 for these derivatives corre-
spond to the most stable isomer in the transition state. Accord-
ing to PM3 calculations the axial diastereomer is preferred by
ca. 2 kcalmol–1, while AM1 calculations showed a preference
of only ca. 0.6 kcalmol–1 for this conformation.
[18] a) K. B. Wiberg, Angew. Chem. Int. Ed. Engl. 1986, 25, 312–
322; Angew. Chem. 1986, 98, 312–322; b) T. Dudev, C. Lim, J.
Am. Chem. Soc. 1998, 120, 4450–4458.
Received: May 11, 2005
Published Online: September 1, 2005
4358
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 4346–4358