NMR investigations on the proline-catalyzed aldehyde self-condensation: Mannich mechanism, dienamine detection, and erosion of the aldol addition selectivity

Article abstract of DOI:10.1021/jo200431v

The proline-catalyzed self-condensation of aliphatic aldehydes in DMSO with varying amounts of catalyst was studied by in situ NMR spectroscopy. The reaction profiles and intermediates observed as well as deuteration studies reveal that the proline-catalyzed aldol addition and condensation are competing, but not consecutive, reaction pathways. In addition, the rate-determining step of the condensation is suggested to be the C-C bond formation. Our findings indicate the involvement of two catalyst molecules in the C-C bond formation of the aldol condensation, presumably by the activation of both the aldol acceptor and donor in a Mannich-type pathway. This mechanism is shown to be operative also in the oligomerization of acetaldehyde with high proline amounts, for which the first in situ detection of a proline-derived dienamine was accomplished. In addition, the diastereoselectivity of the aldol addition is evidenced to be time-dependent since it is undermined by the retro-aldolization and the competing irreversible aldol condensation; here NMR reaction profiles can be used as a tool for reaction optimization.

Full text of DOI:10.1021/jo200431v

FEATURED ARTICLE
pubs.acs.org/joc
NMR Investigations on the Proline-Catalyzed Aldehyde
Self-Condensation: Mannich Mechanism, Dienamine Detection,
and Erosion of the Aldol Addition Selectivity
Markus B. Schmid, Kirsten Zeitler,* and Ruth M. Gschwind*
Institut f€ur Organische Chemie, Universit€at Regensburg, Universit€atsstr. 31, D-93053 Regensburg, Germany
S Supporting Information
b
ABSTRACT: The proline-catalyzed self-condensation of ali-
phatic aldehydes in DMSO with varying amounts of catalyst was
studied by in situ NMR spectroscopy. The reaction profiles and
intermediates observed as well as deuteration studies reveal that
the proline-catalyzed aldol addition and condensation are
competing, but not consecutive, reaction pathways. In addition,
the rate-determining step of the condensation is suggested to be
the CꢀC bond formation. Our findings indicate the involve-
ment of two catalyst molecules in the CꢀC bond formation of
the aldol condensation, presumably by the activation of both the
aldol acceptor and donor in a Mannich-type pathway. This mechanism is shown to be operative also in the oligomerization of
acetaldehyde with high proline amounts, for which the first in situ detection of a proline-derived dienamine was accomplished. In
addition, the diastereoselectivity of the aldol addition is evidenced to be time-dependent since it is undermined by the retro-
aldolization and the competing irreversible aldol condensation; here NMR reaction profiles can be used as a tool for reaction
optimization.
’ INTRODUCTION
The potential of this rediscovered route for the generation of versatile
R,β-unsaturated carbonyl compounds has been recognized and
successfully exploited recently.27,28 Concerning the mechanistic
understanding of the amine-catalyzed aldol condensation, it has
been shown that R,β-unsaturated ketones are not obtained from
the corresponding aldol products by treatment with proline in
acetone/chloroform.29 In addition, from a kinetic study on the
pyrrolidine-catalyzed aldehyde-aldehyde condensation in chloro-
form, evidence has been reported that the reaction is of second
order in the catalyst concentration.27 A second-order pathway for
the CꢀC bond-forming step was also suggested by kinetic
studies on the acetaldehyde condensation in aqueous and salt
solution with high catalyst concentrations.30 On the basis of these
experimental findings, Mannich-like mechanisms for the proline-
catalyzed aldol condensation have been proposed in contrast to
the presumably more intuitive pathway of aldol addition and
subsequent dehydration. Nonetheless, comprehensive mechan-
istic studies were not possible so far, in particular since more
detailed investigations have been impeded for a long time by the
inability to detect and characterize proline enamines in situ. Only
recently could this key intermediate of most of the mechanistic
proposals be snared in situ by means of NMR spectroscopy.20 It
should now allow access to more detailed insights into the
underlying reaction mechanism.
Detailed insights into the mechanisms of both intended
organic reactions and unwanted side reactions are of utmost
importance for better control and for the rational optimization of
reaction conditions in synthetic organic chemistry. Nevertheless,
such knowledge is often difficult to obtain since it is in many cases
intimately connected to the detection of elusive reaction inter-
mediates. Accordingly, in the growing field of asymmetric
organocatalysis,1ꢀ5 where the development of novel synthetic
applications has been predominant over mechanistic investiga-
tions, reaction optimization often appears to be rather empirical
than rational. For instance, the proline-catalyzed intermolecular
aldol reaction,6 archetypical for the concept of enamine catalysis
by secondary amines in general,7ꢀ12 has been reported right from
its beginnings to be accompanied by the aldol condensation
reaction;6 yet, despite proline being an important bifunctional
catalytic system of acknowledged model character, whose me-
chanistic features have been and are still a matter of intensive
theoretical13ꢀ16 and experimental17ꢀ23 studies, it has not been
clarified unambiguously up to now whether the aldol addition and
the aldol condensation must be regarded as competing or as
consecutive reactions.24 However, such knowledge should help to
rationally optimize reaction protocols for the suppression of the
aldol condensation as an unwanted side reaction in asymmetric
aldol addition reactions.25 On the other hand, based on a better
mechanistic understanding of the ongoing reactions, one may
improve the outcome of organocatalytic carbonyl condensations.26
Received: February 25, 2011
Published: March 29, 2011
r
2011 American Chemical Society
3005
dx.doi.org/10.1021/jo200431v J. Org. Chem. 2011, 76, 3005–3015
|

The Journal of Organic Chemistry
FEATURED ARTICLE
Here, we present our mechanistic studies on the proline-
catalyzed self-condensation of aliphatic aldehydes in DMSO.
Starting from the monomeric aldehyde or from the aldol dimer,
online-NMR spectroscopy was applied to monitor the aldol
addition, the aldol condensation, and the retro-aldolization,
which were shown to proceed in parallel within our samples.
This was achieved with the help of characteristic well-resolved
proton resonances of the different species (see Figure S2 in the
Supporting Information). Based on the reaction profiles ob-
tained and the intermediates observed, we can now thoroughly
interpret the ongoing events in the reaction mixtures with regard
to the formation pathway of the condensation product. Further-
more, a deuteration study by adding small amounts of D2O to the
reaction mixtures allowed us to follow the progress of the
R-deuteration via the reaction intermediates and to track the
accumulation of deuterium in the reaction products. The ob-
served reaction profiles and the degrees of deuteration rule out
the existence of an aldol additionꢀdehydration pathway for the
formation of the condensation product under proline catalysis in
DMSO. Instead, a Mannich-like mechanism, most probably via a
double aldehyde-activation, is evidenced for the proline-cata-
lyzed self-condensation of aliphatic aldehydes in DMSO. In
addition, the first detection of a proline-derived dienamine is
presented. In the context of parallel activation and reaction
pathways, using NMR reaction profiles we demonstrate that
the selectivity of the aldol addition reaction is dependent on the
reaction time because of the competition between the reversible
aldol addition and the irreversible aldol condensation pathway.
Scheme 2. (A) Possible Mechanistic Pathways for the
Process of Scheme 1A, Proposed in the Literature27,29,31ꢀ37
(Note: Species That Could Not Be Detected in This Study Are
Shown in Brackets). (B) Commonly Proposed
IminiumꢀOxazolidinone Equilibrium
’ RESULTS AND DISCUSSION
Model Reaction and General Mechanistic Considerations.
As a model system for our investigations on the proline-catalyzed
self-condensation of aliphatic aldehydes, propionaldehyde 1a
was selected as the starting material (c = 50 mM), 100 mol %
of L-proline as the amine catalyst and DMSO-d6 at 300 K as the
Scheme 1. (A) Model Reaction Investigated: Proline-
Catalyzed Self-Condensation of Propionaldehyde 1a. (B) and
(C) Aldol Condensation as a Process That Is Either Con-
secutive or Competing with the Aldol Addition, Respectively
solvent (Scheme 1A). This choice was based on our recent study
that had allowed us to detect in situ the elusive prolineꢀenamine
and further intermediate species in the intermolecular aldol
reaction for the first time.20 By thus presenting new options for
the interpretation of reaction profiles, these experimental con-
ditions were expected to also bring forward the mechanistic
understanding of the proline-catalyzed aldehydeꢀcondensation
reaction.
As a first simplistic issue, the relationship between the proline-
catalyzed aldol addition and the aldol condensation should be
addressed, i.e., whether the aldol addition and aldol condensation
are consecutive (Scheme 1B) or competing (Scheme 1C) reac-
tion steps. More detailed considerations on the potential path-
ways of the proline-catalyzed aldehyde self-condensation have
led to the three basic mechanistic pathways proposed in the
literature (Scheme 2A).27,29,31ꢀ37 They can be classified by the
nature of the species that are involved in the CꢀC bond-forming
steps into an aldol pathway (I), a Mannich pathway (II), and a
double-activation Mannich-type pathway (III).
3006
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
Pathway I initially constitutes the commonly accepted me-
chanism of the proline-catalyzed intermolecular aldol reaction; it
is characterized by the addition of the catalyst-derived enamine
4d of the donor aldehyde to the acceptor aldehyde molecule 1a
and is completed by hydrolysis and dehydration (or vice versa),
like the classical aldol condensation reaction.33ꢀ35
Pathway II corresponds to a textbook Mannich mechanism, in
which the enol tautomer 1b of the donor aldehyde adds to the
iminium ion 4a, formed from the acceptor aldehyde and the
amine catalyst, followed by elimination of the catalyst.29,31,32
A third mechanistic alternative, described by pathway III,27,36,37
has been pointed out recently: It resembles the Mannich pathway
IIbut comprises the simultaneous activation of both the donor and
the acceptor aldehyde as an enamine 4d and an iminium ion 4a,
respectively, and subsequent elimination and hydrolysis steps.38
Forward Aldol Addition and Condensation. To distinguish
between the three proposed mechanisms, we have employed our
recently gained knowledge on the observation of intermediates in
proline-catalyzed aldol reactions.20 On this basis, a very simple
kinetic consideration should help us to obtain first insights into
potential condensation reaction pathways: Whatever the exact
underlying mechanism of the formation of the enal 3 may be, a
certain correlation between the concentration of the intermedi-
ates involved in the rate-determining step and the formation rate
of 3, i.e., the slope of its NMR-derived concentrationꢀtime
curve, should be expected. To apply this criterion to the different
potential reaction pathways, we conducted the self-aldolization/
condensation of propionaldehyde with different amounts of
L-proline (100, 50, 20, and 10 mol %) in DMSO-d6 at 300 K
inside an NMR tube and monitored the progress of the reactions
by one-dimensional proton spectra. For all the different proline
concentrations applied as well as for butyraldehyde as the starting
material with 100 mol % of proline, very similar characteristic
features of the reaction progress were monitored. (Exemplary
results for 100 mol % of proline are shown in Figure 1; see
Figures S1 and S3 in the Supporting Information for the results of
the other experimental setups.) First, the consumption of the
starting material 1a and the steady decrease of the concentra-
tion of the proline-derived propionaldehyde intermediates
(the enamine 4d and the isomeric oxazolidinones 4b,c;
Figure 1A) are observed. In addition, the formation of two
product oxazolidinones 5b,c (among others) and of the diaster-
eomeric aldol dimers 2a,b is evidenced as well as their disap-
pearance (Figure 1B; maximum concentrations of 5b,c after
about 100 and 200 min and of 2a,b after about 70 and 100 min,
respectively).
Figure 1. Reaction profile of the proline-catalyzed self-aldolization/
condensation of propionaldehyde 1a with 100 mol % L-proline in
DMSO-d6 at 300 K: (A) starting material and intermediates derived
thereof, (B) aldol dimers and related oxazolidinones, and (C) condensa-
tion product and the slope (displayed in arbitrary units) of its buildup
curve, which corresponds to the formation rate of 3. (Those protons
whose resonances were used for the monitoring are highlighted in gray;
see Figure S2 in the Supporting Information. Note: The total amount of
C3 moieties stemming from propionaldehyde, detected in the first
spectrum, was set to 100%. Times of concentration maxima of 2a,b
and 5b,c are marked with vertical dotted lines.).
Finally, the aldehyde self-condensation product 3 is formed
irreversibly (Figure 1C) in about 90% yield after 12 h (The molar
ratio on the y axis refers to the initial concentration of the
aldehyde 1a; see the Supporting Information for observations
about aldehyde substitution effects on the reactivity and the enal
E/Z selectivity).
In a previous study on a crossed-aldolization/-condensation of
aldehydes,27 the constant ratio of the aldol and the condensation
product throughout the reaction time could be interpreted as an
indication of the competition between the two reaction path-
ways. However, in the case of our self-aldolization/-condensation
this argument cannot be applied since 3 is the only product of the
reaction, while 2a,b vanish again, either because of dehydration
or because of retro-aldolization (Scheme 1B,C). Thus, the issue
of whether the aldol addition and the aldol condensation are
consecutive or competing reaction steps must be addressed by
other means. In view of the reaction profile of Figure 1, it is highly
striking that the maximum concentration of the aldol dimers 2a,b
(after 75 and 105 min, respectively, Figure 1B) does not at all
coincide with the maximum slope of the concentrationꢀtime
curve of the condensation product 3, i.e. with its rate of formation
(right at the beginning of the reaction, Figure 1C). According to
the criterion outlined above, reaction pathway Ia (the aldol
additionꢀhydrolysisꢀdehydration sequence in Scheme 2A) can
3007
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
hence clearly be ruled out as the major pathway for the formation
of enal 3. Accordingly, the decrease of the concentration of the
aldol dimers 2a,b must be due to retro-aldolization rather than
dehydration. In analogy, reaction pathway Ib (the aldol addi-
tionꢀdehydrationꢀhydrolysis sequence, Scheme 2A) can be
analyzed with regard to its plausibility. For the formation of enal
3 from the product iminium 5a via its dehydrated analogue 7a in
this pathway, the elimination of water should be rate determining
since we could show previously by cross-peak evaluation in EXSY
spectra20 that the hydrolysis of iminium ions/oxazolidinones to
the corresponding aldehydes is a fast process and that, in contrast,
the dehydration of aldol dimers is significantly slower (see also the
next paragraph). Accordingly, in case pathway Ib was operative,
the maximum formation rate of 3 would be expected to correlate
with the maximum concentration of the hypothetical iminium ion
5a (undetected in this study) and thus with the concentration of
the well-observed isomeric oxazolidinones 5b,c, as suggested by
the commonly proposed iminiumꢀoxazolidinone equilibrium
(Scheme 2B).20 Under these premises, the observed discrepancy
between the maximum of the concentrationꢀtime-curves of 5b,c
and the maximum rate of formation of 3 (cf. Figure 1B and 1C)
again indicates that pathway Ib is not the main route of the proline-
catalyzed aldehyde self-condensation, either. Instead, as the for-
mation of 3 is the faster the higher the concentration of the reactant
intermediates 4b,c,d is (cf. Figure 1A and 1C), processes involving
these intermediates in the rate determining step are suggested by the
monitored concentrationꢀtime curves. Altogether, as a first result
and in agreement with a previous study on the pyrrolidine-catalyzed
aldehyde self-condensation,27 the kinetic profile of an equimolar
reaction mixture of proline and propionaldehyde 1a in DMSO
evidences that the proline-catalyzed intermolecular aldol addition
and aldol condensation are not consecutive, but rather competing
reaction pathways (Scheme 1C).
Formal Dehydration of Aldoldimers via Retro-Aldoliza-
tion. To corroborate this finding, we investigated in more detail
whether the condensation product 3 could be generated from the
aldol products 2a,b under our experimental conditions. For that
purpose, the aldol addition dimers 2a,b were synthesized,39 isolated
as a mixture of diastereomers, dissolved in DMSO-d6 inside an
NMR tube, and exposed to benzoic acid or proline as potentially
dehydrating additives. The events in the reaction mixtures were then
again monitored by one-dimensional proton NMR spectra. First,
we studied the influence of Brønsted acid addition: However,
100 mol % of benzoic acid, a previously employed cocatalyst in
organocatalytic aldehyde self-condensations,27,32 virtually did
not effectuate any dehydration of the aldol dimers 2a,b over
15 h (data not shown). We conclude from this that the proline-
catalyzed aldehyde condensation is not a Brønsted acid-catalyzed
dehydration of aldol addition products. As a next approach, we
added 100 mol % of L-proline to 2a,b in DMSO-d6. The con-
centration curves of the observed species in the reaction mixture
are summarized in Figure 2.
In contrast to an early report on a similar approach in acetone/
chloroform starting from a β-hydroxy ketone,29 we could easily
detect the formal dehydration of the β-hydroxy aldehydes 2a,b in
DMSO-d6 under the influence of 100 mol % of proline: The
concentrations of the aldol dimers 2a,b decrease over time
(Figure 2A), while the amount of the condensation product 3
increases in turn (Figure 2C).
However, the monomeric propionaldehyde 1a, and—in agree-
ment with our previous EXSY findings20—the corresponding
oxazolidinones 4b,c and the enamine 4d (Figure 2B) are observed
Figure 2. Reaction profile of the proline-catalyzed formal dehydration
of the aldol addition dimers 2a,b in DMSO-d6 at 300 K under proline
catalysis (100 mol %): (A) aldol dimers and related oxazolidinones, (B)
propionaldehyde and intermediates derived thereof, and (C) condensa-
tion product and the slope (displayed in arbitrary units) of its buildup
curve, which corresponds to the formation rate of 3. (Notes: The total
amount of C3 moieties stemming from propionaldehyde, detected in the
first spectrum, was set to 100%. Times of concentration maxima of 1a
and 4b,c,d are marked with vertical dotted lines. The kinks in some of the
curves are due to the temporary removal of the sample from the NMR
spectrometer and the associated disturbances, resulting in a larger error
range for the maximum concentration of 4b,c,d.)
in the reaction mixture, too, being indicative of the readily
occurring retro-aldol reaction under our experimental condi-
tions. Thus, to clarify whether the condensation product 3 results
from 2a,b via a direct dehydration (Scheme 1B) or rather from a
sequence of retro-aldol reaction and Mannich-type condensation
3008
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
(Scheme 1C), the above-mentioned argumentation needs to be
applied once more: In concordance with the findings for the
forward aldol condensation (Figure 1), the maximum formation
rate of 3 (Figure 2C), starting from the aldol dimers 2a,b, does
not coincide with the highest concentration of 2a,b (at the very
beginning of the reaction, Figure 2A) or the product oxazolidi-
nones 5b,c. This confirms that the major formation pathway of 3
is not the direct dehydration of 2a,b (pathway Ia in Scheme 2A).
Instead, the formation rate of 3 shows an induction period; its
maximum is reached only after 4ꢀ5 h. Most interestingly, it is
correlated to the maximum concentration of the intermediates
4b,c,d. This supports the hypothesis that these species are
involved in the rate-determining step of the proline-catalyzed
aldehyde condensation.
Deuteration Experiments. Further evidence for the observed
competition between the proline-catalyzed aldol addition and
aldol condensation was collected from a deuteration study on the
above-mentioned formal dehydration of the aldol dimers 2a,b. In
the case of a consecutive aldol condensation (Figure 3A, left), the
degree of R0-deuteration of 3 is determined by the degree of
R0-deuteration of 2a,b (R0 and β0 refer to the R and β position of
the former aldol acceptor moiety within the dimers 2a,b and 3,
see Figure 3A), whereas in the case of a competing condensation
pathway (Figure 3A, right), it is determined by the degree of
R-deuteration of the aldol acceptor species (i.e., 1a or 4a in
Scheme 2A). The latter case implies that there is not necessarily a
causal relationship between the degrees of R0-deuteration of
2a,b and 3 so that the degree of R0-deuteration of 3 may well
exceed that of 2a,b. In contrast, for our experimental approach
starting from nondeuterated 2a,b, the consecutive condensation
pathway (Figure 3A, left) does not allow for a higher degree of
R0-deuteration of 3 compared to 2a,b because it can be assumed,
following well-known kinetic isotope effects,40 that the dehydra-
tion of R0-deuterated 2a,b is not faster than the one of non-
deuterated 2a,b. Hence, the observation of a higher degree of
R0-deuteration of 3 compared to 2a,b would provide further
strong evidence for the aldol condensation being in competition
with the aldol addition (Scheme 1C). For these deuteration
experiments, 2a,b and 100 mol % of proline were mixed in
DMSO-d6 to which 1 vol % of D2O had been added. Under
otherwise identical experimental conditions, but in the absence
of D2O, we had observed the readily occurring retro-aldolization
of 2a,b as well as the presence of the rapidly interconverting20
intermediates 4b,c,d (Figure 2). Since both the retro-aldol
reaction and the enamineꢀoxazolidinone exchange are asso-
ciated with protonation in the R-position of the carbonyl
compound, the R-deuteration of propionaldehyde 1a and the
derivatives thereof could be expected to proceed with ease in the
presence of D2O. Accordingly, an accumulation of deuterons in
R0-position of 2a,b and 3 should be observed over time
(Figure 3A) and the comparison of the deuteration of the
different species observed should provide insights into the actual
formation pathway of 3. To be able to connect this increasing
degree of R0-deuteration, i.e., the ratio of 2H in R0-position, to the
formation pathway of 3, the direct R0-deuteration of the reaction
products 2a,b and 3 must be ruled out first. For 2a,b, the lack of
CH-acidity generally prevents R0-deuteration. In the case of 3,
direct vinylogous R0-deuteration was excluded by treating non-
deuterated 3 with proline in DMSO/1 vol % of D2O for more
than 1 day: No R0-deuteration was observed (data not shown),
which proves at the same time the irreversibility of the con-
densation under our experimental conditions. Thus, the degree
Figure 3. (A) Possible relationships of the aldol condensation with the
aldol addition, with different implications for the relative degrees of
R0-deuteration of 2a,b and 3. (B) Exemplary deconvolution of over-
lapping NMR resonances of differently deuterated species of the
condensation product 3. Displayed are the experimental spectral section
(left top) as well as the underlying simulated spectra of the non-, mono-,
and dideuterated compounds (left rows 2ꢀ4). For an estimation of the
error range, the experimental spectrum and simulated spectra with
varying ratios of the differently deuterated species are presented (right).
C) Experimentally determined degrees of R0-deuteration in the course
of the formal dehydration of 2a,b to 3 under catalysis of 100 mol % of
proline in DMSO-d6 with 1 vol % of D2O at 300 K, supporting the
general reaction scheme on the right-hand side of A). (Note: Data points
for the first 3 h are not depicted because the amounts of 3 are so small, cf.
Figure 2C, that an evaluation of the degree of deuteration is not
reasonable.)
of R0-deuteration of 3 must be really caused by the degree of
deuteration of the reaction intermediates involved.
To distinguish between the two fundamental reaction schemes
(Figure 3A), the degrees of deuteration of the reaction products
2a,b and 3 were evaluated. For 2a,b, the loss of signal intensities
of the R0-protons over time, compared to the aldehyde proton
resonance, was ascribed to the deuteration and was hence used as
3009
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
a measure for the degree of R0-deuteration. In the case of 3, the
β0-methyl proton resonance of the aldol acceptor moiety was used
as a probe for the deuteration in the neighboring R0-position.
On the basis of isotope-induced chemical shift changes of the
β0 resonance, the amounts of h,h-3, h,d-3, and d,d-3 and, hence,
the degree of R0-deuteration can be calculated after direct
integration of the different branches of the β0 resonance (see
Figure 3B for an exemplary deconvolution of the experimental
spectrum). The resulting degree of deuteration agrees very well
with an exemplary simulation of the peak structure (Figure 3B,
right top). The accuracy of this direct integration approach was
furthermore supported by peak simulation with deviating shares
of h,h-3, h,d-3, and d,d-3, respectively (see Figure 3B, right
rows 2ꢀ4): Variations of the individual shares on the order of
5 percentage points would be easily detected despite the multiple
resonance overlap. Accordingly, the error range of the degrees of
deuterationof 3, determined by thisapproach, can be estimated to
be below 5 percentage points. On the basis of this good
correspondence of the integration and the simulation, the more
convenient integration method was used for the determination of
the degrees of deuteration of 3. The results of the evaluation of the
degrees of deuteration of 2a,b and 3 are shown in Figure 3C.
The comparison of the degrees of R0-deuteration of 2a,b and 3
reveals substantially higher values for 3 than for 2a,b throughout
the reaction time observed. According to our above-mentioned
argument, pathway I (Scheme 2A) can definitely be ruled out as
the formation pathway of 3. From this observation we can
conclude that the proline-catalyzed condensation of aliphatic
aldehydes does not proceed via the aldol addition and a
consecutive dehydration step but rather represents a reaction
pathway that competes with the aldol addition.
For the differentiation between pathways II and III, the rate-
determining step of the aldol condensation should be clarified
first. For the proline-catalyzed aldol addition, the CꢀC bond
formation has been shown to be rate determining.19,41ꢀ43 Our
experimental findings suggest the same for the aldol condensa-
tion: On the one hand, EXSY analyses20 revealed that the
formation of the enamine from the aldehyde via the oxazolidi-
nones is a fast process (opposite to a study on the amino acid-
catalyzed acetaldehyde self-condensation in aqueous and salt
solution30). In contrast, neither the aldol addition nor the aldol
condensation product formation could be traced beyond the
enamine by EXSY. On the other hand, the inability to detect
species of type 6, 7, and 8 may be taken as an additional hint that
the final elimination and hydrolysis steps are relatively rapid in
comparison to the CꢀC bond formation. For the hydrolysis, this
is covered experimentally by our EXSY findings20 and also by the
fact that the elimination is unlikely to be rate limiting.27 Together
with the above-mentioned correlation of the maximum forma-
tion rates of 3 with the maximum concentrations of 4b,c,d, these
experimental findings suggest that the CꢀC bond formation is
rate determining in pathways II and III. Since both of these
pathways of the aldol condensation share the same aldol acceptor
species, a correlation of the concentration of 4a to the formation
rate of 3 is expected. To check this and hence to further support
the assumption that the CꢀC bond formation is the rate-
determining step in the proline-catalyzed aldol condensation,
the progress of the aldol condensation was investigated for a
sample of propionaldehyde 1a and 100 mol % of proline in
DMSO-d6 with 0.5 vol % of D2O at 300 K (Figure 4A). The
concentration of the oxazolidinone 4b was employed once more
as a measure for the concentration of the aldol acceptor iminium
Figure 4. (A) Hypothetical formation of differently deuterated species
of 3 via differently deuterated aldol acceptors (The oxazolidinone 4b is
shown as a representative for the typically postulated, yet undetected,
iminium ion 4a.) (B)ꢀ(D) Concentration curves of selected species in a
reaction mixture of propionaldehyde with 100 mol % of L-proline in
DMSO-d6 with 0.5 vol % of D2O at 300 K: (B) nondeuterated 4b and 3,
(C) monodeuterated 4b and 3, and (D) dideuterated 4b and 3 and, in
each single case, the slope (displayed in arbitrary units) of the buildup
curve of 3, which corresponds to its formation rate. (Note: The total
amount of C3 moieties stemming from propionaldehyde, detected in the
first spectrum, was set to 100%.)
ion 4a (see above and Scheme 2B), and the slope of the
concentrationꢀtime curve of 3 was again used as a measure for
the formation rate of 3. The beneficial effect of the addition of
D2O was an increase of the information density of our
3010
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
investigation, in that three different aldol acceptors (h,h-4b, h,
d-4b, d,d-4b) and three different condensation products (h,h-3,
h,d-3, d,d-3) could be studied in parallel by using only one
sample. The results of this experiment are depicted in Figure 4.
In the beginning of the reaction, h,h-4b is the only oxazolidi-
none present (Figure 4B). Then, the deuterated isotopologues
h,d-4b and d,d-4b appear quite rapidly, reaching their maximum
concentrations after about 20 min and 1 h, respectively
(Figure 4C,D). Likewise, h,h-3 is the only condensation product
formed in the beginning (Figure 4B), whereas the formation of
h,d-3 and d,d-3 is retarded by an induction period (Figure 4C,D).
In view of the convoluted proton spectra of the differently deuterated
species and the associated deconvolution for the quantitative
evaluation (see Figure 3B and Figure S5 in the Supporting
Information), it is highly striking how well the maxima of the
concentration of 4b and of the slope of 3 coincide in the cases of
all three isotopologues (Figure 4BꢀD). This suggests according
to the above-mentioned reasoning that the hypothetical iminium
ion 4a (represented in its concentrationꢀtime curve by 4b) is in
fact involved in the rate-determining step of the formation of 3 in
our experimental system. Thus, the CꢀC bond formation should
be the crucial and rate-determining step in the proline-catalyzed
aldol condensation of aldehydes in DMSO.
Variation of the Catalyst Amount. With the knowledge that
the proline-catalyzed aldol addition and aldol condensation are
competing reactions and based on the experimentally supported
assumption that the CꢀC bond formation is rate determin-
ing both in the condensation and in the addition pathway,19,41,42
there is now a possibility to differentiate between pathway II and
pathway III: This option is due to the fact that both the aldol
addition reaction and pathway II of the aldol condensation comprise
one catalyst molecule in the rate determining step, either as the
enamine 4d or as the iminium ion 4a. But in contrast, following
pathway III, two catalyst molecules (4a and 4d) are involved in the
rate-determining step of the aldol condensation.27,36,37 Thus,
changes in the amount of catalyst should not have a significant
impact on the competition between the aldol addition and the
aldol condensation if the aldol condensation proceeds via pathway
II. In contrast, if pathway III is operative for the aldol condensation
reaction, lower catalyst amounts should initially favor the aldol
addition, whereas higher catalyst amounts should increase the
extent of the aldol condensation relative to the aldol addition.
Revealing such different trends for the addition/condensation
rate dependencies on the catalyst amount would thereby qualita-
tively confirm the detailed kinetic studies that have already pro-
ven the aldol condensation to be second order in the catalyst
concentration.27,30
To check this influence of the catalyst amount on the relative
reaction rates of the aldol addition and condensation, different
amounts of proline (10, 20, 50, and 100 mol %) were offered to
propionaldehyde 1a in DMSO-d6 at 300 K, and the progress of
the reactions was followed by one-dimensional proton spectra.
The interpolated yield of 2a,b and 3 after 30 min was used as a
measure for the initial formation rates of 2a,b and 3, respectively.
(Note: Retro-aldolization is expected to be negligible at early
stages of the reaction and plotted against the varying amounts of
L-Pro as the catalyst which are represented by the detected
enamine concentration as a measure for the active amount of
catalyst available to the system.47) For all three dimer species, an
increase in the initial formation rate is found with increasing
amounts of proline. Without a doubt, another tendency becomes
obvious from Figure 5 concerning the relative rates of the aldol
Figure 5. (A) Initial formation rates of the aldol products 2a,b and of
the condensation product 3 in DMSO-d6 at 300 K as a function of
enamine concentration which represents the available, “active” amount
of catalysts as offered to the system. (B) Direct comparison of
representative 1H NMR spectra of reaction mixtures subjected to
100 mol % (upper trace) and 10 mol % of L-Pro (lower trace) as catalyst
taken at the same reaction time (10 min). The two regions of interest
providing characteristic signals of the competing products 2 and 3 are
highlighted within the spectra.
addition and condensation: While the condensation product 3 is
formed more slowly than the addition products 2a,b at low
catalyst concentration (10 mol %, enamine concentration ca.
0.7%), its formation is the fastest at high catalyst loadings (50 and
100 mol %, enamine concentration >2%). In addition, it is
confirmed by the comparison of the initially formed quantities
of 2a,b and 3 (cf. Figure S1C and S1D in the Supporting
Information): While 2a,b exceeds 3 in the early stages of the
aldol addition/condensation with low catalyst loadings (10 and
20 mol %), the formation of 3 is preferred over the one of 2a,b
right from the beginning if higher amounts of catalyst (50 and
100 mol %) are used. This trend is exemplarily illustrated in
Figure 5B. These qualitatively different acceleration tendencies
of the aldol addition and aldol condensation indicate that
different amounts of catalyst-derived species are involved in
the rate determining steps of these reactions. Therefore, as
elaborated above, pathway II can be ruled out as the pathway
for the aldol condensation since, just like the aldol addition, it
3011
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
requires one catalyst molecule in the CꢀC bond-forming step. In
contrast, pathway III of the aldol condensation is associated with
two catalyst-derived molecules in the rate-determining step.
Thus, this pathway can well explain the stronger acceleration
effect of the aldol condensation in comparison to the aldol
addition upon increasing amounts of proline.
Hence we can provide further evidence that the proline-
catalyzed aldol condensation of aldehydes proceeds via a double
activation of both the aldol donor and the aldol acceptor
molecule (pathway III in Scheme 2A). Since the enamine 4d is
readily detected in our study, there can be little doubt that it is
involved as the aldol donor species in the formation of the
condensation product 3. On the other hand, the iminium ion 4a
has been typically proposed as the electrophilic aldol acceptor
species. However, since the existence of this intermediate
iminium zwitterion in solution has not been proven experimen-
tally in proline-catalyzed reactions under our experimental con-
ditions, one may well speculate about different electrophilically
activated species acting as aldol acceptors, for instance, the
carbinolamines (likewise undetected in our study) or the readily
observed oxazolidinones. Further experimental and theoretical
approaches will be necessary to shed more light on this unsolved
issue of the proline-catalyzed aldol reaction. Likewise, more
investigations will be necessary to clarify whether the aldol
condensation might proceed via pathway IIIa or IIIb. We cannot
further address this issue here since we have not been able to
detect any of the intermediates of type 6, 7, or 8.
Detection of an Acetaldehyde-Derived ProlineꢀDienamine.
Nevertheless, the observation that high proline loadings favor the
Mannich-type aldehyde self-condensation over the aldehyde self-
aldolization is supported by our findings on the proline-catalyzed
self-aldolization/-condensation of acetaldehyde 9. As reported
previously,20 the detection of the acetaldehyde-derived prolineꢀ
enamine was not feasible because of the rapid progress of the
proline-catalyzed self-aldolization reaction of acetaldehyde in
DMSO. While this had largely impeded the use of acetaldehyde
9 as a nucleophilic reaction partner in amine-catalyzed crossed-
aldol, Mannich or Michael reactions until recently,48ꢀ53 the pro-
line-catalyzed asymmetric self-aldolization of 9 could be exploited
synthetically to generate the triketide 5-hydroxy-2-hexenal 12.33
For this acetaldehyde trimerization, an aldol-Mannich cascade
(Figure 6A, top)33 as well as a Mannichꢀaldol sequence
(Figure 6A, middle) are possible. In contrast, 13, the condensation
product of the trimerization of 9, must be formed by two subsequent
Mannich-type condensations (Figure 6A, bottom). Accordingly,
while the formation of the triketide 12 was predominant at proline
loadings on the order of 1 mol %,33 the condensation product 13
should be favored at higher proline amounts. To check this hypo-
thesis, a reaction mixture of acetaldehyde 9 (c = 50 mM) and 100 mol
% L-proline was prepared in DMSO-d6 at 300 K and the progress of
the reaction was monitored by 1D 1H NMR spectra (Figure 6B).
Despite the complexity of the in situ NMR spectra of the
reaction mixture, various rapidly vanishing species could be
identified (because of the short lifetimes of all theses species
more detailed investigations could not be performed and their
Figure 6. (A) Potential formation mechanisms of 12 via an al-
dolꢀMannich33 or via a Mannichꢀaldol cascade and of 13 via a
MannichꢀMannich sequence. (B) Reliably assignable intermediates in
the reaction mixture and the development of their concentrations over
time. (Those protons whose resonances were used for the monitoring
are highlighted in gray; see Figure S7 in the Supporting Information.
Note: The total amount of C2 units stemming from 9, detected in the
first spectrum, was set to 100%. However, because of the incomplete
characterization of the whole of the reaction mixture, the actual ratios
may be substantially lower.)
neither aldol dimers of 9 nor the triketide 12 were detected.
Hence, thepresence ofthe condensation dimer10and theabsence
of aldol dimers of 9 can be interpreted such that, for monomeric 9,
the Mannich-type condensation is favored over the aldol addition
reaction when 100 mol % of proline is applied. The same
conclusion can be drawn for the formation of trimers starting
from 10: The dienamine 11 should, as the common intermediate,
in principle allow for the aldol addition to the triketide 12 and for
the condensation to 13. But only 13 is detected experimentally.
Altogether, this preliminary study of the proline-catalyzed self-
aldolization/-condensation in DMSO corroborates our finding
that high proline loadings favor the aldol condensation over the
competing aldol addition and that hence a Mannich-type mechan-
ism with a 2-fold substrate activation underlies the condensation
reaction. In addition, these investigations unearth the first in situ
prolineꢀdienamine. Thus, they should moreover guide the way
toward the elucidation of mechanistic issues of the growing field of
dienamine catalysis.55ꢀ62
1
assignments are based on H NMR data, see Figure S7 in the
Supporting Information). Among them, two condensation pro-
ducts, the dimeric enal 10 and the trimeric 13 (Figure 6B), could
1
be assigned reliably owing to their characteristic H chemical
shifts and multiplet patterns and in comparison with literature
data.54 Additionally, a butadienyl unit was recognized that can be
assigned to 11, the proline dienamine derived from 9. In contrast,
Time-Dependence of the Diastereoselectivity of the Aldol
Addition. Another interesting finding of our studies, closely
3012
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
for the monomeric species: The condensation withdraws the
monomeric aldehyde(1a)-derived species of type 4 irreversibly
from the equilibrium between aldolization and retro-aldolization.
In addition, the retro-aldol reaction is significantly faster for the
main aldolization product 2a than for the minor product 2b (see
Figure 2A). Thus, upon retro-aldolization the competing irre-
versible condensation reaction causes the faster disappearance of
the mainly formed 2a, which in turn leads to a decrease in the
diastereomeric ratio of the two aldol products. Since the rates of
both the aldol addition (thus most probably also of the retro-
aldol reaction) and the aldol condensation are enhanced by
higher catalyst amounts but the rate of the aldol condensation
significantly more (see Figure 5), the steeper decays in the
selectivityꢀtime curves with 50 and 100 mol % (Figure 7A)
are readily rationalized.
Thus, highest diastereoselectivities can only be achieved if
short reaction times are applied. However, in terms of a valuable
and useful synthetic procedure the yield of the process is also
crucial and might be affected in a detrimental way by such an
experimental setup. Accordingly, only the combined optimum of
both (diastereo)selectivity (Figure 7A) and yield (Figures 1B
and S1, Supporting Information, for other catalyst loadings) can
provide ideal reaction conditions. Use of the NMR reaction
profiles points to other possible options for reaction optimization
of the aldol addition investigated here. Again, for all catalyst loadings
a decrease in diastereoselectivity for higher yields (conversion) is
visible. In addition, depending on the amount of catalyst, different
maximum yields can be reached and at a constant diastereoselec-
tivity the yield increases using lower catalyst loadings.
Interestingly, the two plots presented in Figure 7A and 1B
(product curves) correspond to direct information available from
the reaction profiles mapping the products without any informa-
tion about the reaction intermediates or the mechanism of the
reaction. Accordingly, even in complex reactions with competing
reaction pathways useful information for the optimization of the
synthetic reaction conditions can be gained from simple NMR-
reaction profiles of the products (for a further example using
DMF-d7 catalyst loading of mol %, see Figure S8, Supporting
Information) that could serve as a valuable tool for the determi-
nation of optimal reaction parameters.
Figure 7. (A) Dependence of the product ratio 2a:2b of the proline-
catalyzed self-aldolization of 1a on the reaction time and on the amount
of proline offered. (B) Graphical summary of the parameters and
reaction pathways influencing the diastereoselectivity of the proline-
catalyzed aldol addition of 1a.
associated with the competitive nature of aldol addition and aldol
condensation, concerns an issue that has not been addressed so
far to our knowledge: the time-dependence of the diastereo-
selectivity of proline-catalyzed intermolecular aldol reaction of
aldehydes. As can be seen from Figure 1B and 2A, the time-
dependent appearance and vanishing is qualitatively different for
the diastereomeric aldol products 2a and 2b in DMSO and in
DMF alike (Figure S8 in the Supporting Information). In
addition, Figure S1 (Supporting Information) reveals that these
differences depend on the catalyst amount, too. As a synopsis of
the experimental approaches discussed above, the ratio of the two
aldol products 2a and 2b, hence the diastereoselectivity of the
reaction, is depicted in Figure 7A as a function of reaction time.
For all catalyst amounts sampled, ranging from 10 to 100 mol %,
a decrease of the ratio of 2a:2b over time is observed. The
more catalyst is employed, the steeper is this decay, for instance,
with 100 mol % of proline, the diastereomeric ratio drops within
3 h from 4:1 to 2:1, with 50 mol % of proline, still a reduction
from 3.5:1 to 1.75:1 within 5 h is observed. Unlike, at lower
catalyst loadings (20 mol % and 10 mol %) the slope of the
decreasing diastereoselectivity is very similar and only a slight
offset of 0.2 to higher diastereomeric ratios is found for lower
catalyst loading of 10 mol %. This time dependent loss of
diastereoselectivity can be attributed to the different rates of
the retro-aldolization of the aldol dimers 2a and 2b together with
the competition of the aldol addition and the aldol condensation
’ CONCLUSION
We have presented our NMR spectroscopic study on the
proline-catalyzed self-condensation of aliphatic aldehydes in
DMSO. By a detailed interpretation of the NMR-monitored
reaction profiles of the condensation, starting from the mono-
meric aldehyde and from the aldol dimers, respectively, we have
provided evidence that the aldol addition and the aldol con-
densation are competing, but not consecutive reaction pathways.
This was deduced from the observation that the maximum
formation rate of the condensation product does not correspond
to the maximum concentration of the aldol addition products,
but to the one of the aldehydeꢀproline adducts. This result was
confirmed by the formal dehydration of the aldol dimers in the
presence of proline and D2O for which a higher degree of
R0-deuteration of the condensation product was found in com-
parison to the aldol dimers. Additionally, it was observed that the
acceleration of the aldol condensation by increasing the catalyst
loading is more pronounced than for the competing aldol
addition. Thus, it is concluded that two catalyst molecules are
involved in the rate-determining step of the aldol condensation
3013
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
which is suggested by our experimental findings to be the CꢀC
bond formation. Accordingly, a Mannich-type reaction pathway,
including the activation of both the aldol acceptor and the aldol
donor, is indicated for the proline-catalyzed self-condensation of
aldehydes. This can also rationalize the preference of the aldol
condensation over the aldol addition when high amounts of
catalyst are applied in the proline-catalyzed oligomerization of
acetaldehyde. In this reaction mixture, the first in situ detection of
a proline dienamine was accomplished. In addition, we could
demonstrate the time-dependence of the stereoselectivity of the
aldol addition, resulting from the competitive nature of the aldol
addition and condensation. Extracting information from easily
accessible NMR reaction profiles reveals that the yield (anti-aldol
dimer 2a) can be considerably improved by using low catalyst
loadings with only a slight loss in diastereoselectivity during
increased reaction times.
(2) Dalko, P. I., Ed. Enantioselective Organocatalysis: Reactions and
Experimental Procedures; Wiley-VCH: Weinheim, 2007.
(3) Chem. Rev. 2007, 107, 12: special issue on organocatalysis.
(4) Reetz, M. T.; List, B.; Jaroch, S.; Weinmann, H.; Eds.; Ernst
Schering Foundation Symposium Proceedings “Organocatalysis”; Springer:
Berlin, 2008.
(5) List, B., Ed. Top. Curr. Chem. 2010, 291: Asymmetric
Organocatalysis.
(6) List, B.; Lerner, R. A.; Barbas, C. F., III. J. Am. Chem. Soc. 2000,
122, 2395–2396.
(7) List, B. Chem. Commun. 2006, 819–824.
(8) Mukherjee, S.; Yang, J. W.; Hoffmann, S.; List, B. Chem. Rev.
2007, 107, 5471–5569.
(9) Melchiorre, P.; Marigo, M.; Carlone, A.; Bartoli, G. Angew.
Chem., Int. Ed. 2008, 47, 6138–6171.
(10) Bertelsen, S.; Jørgensen, K. A. Chem. Soc. Rev. 2009, 38, 2178–2189.
(11) Pihko, P. M.; Majander, I.; Erkkil€a, A. Top. Curr. Chem. 2010,
291, 29–75.
(12) Nielsen, M.; Worgull, D.; Zweifel, T.; Gschwend, B.; Bertelsen,
S.; Jørgensen, K. A. Chem. Commun. 2011, 47, 632–649.
(13) Bahmanyar, S.; Houk, K. N. J. Am. Chem. Soc. 2001,
123, 11273–11283.
(14) Bahmanyar, S.; Houk, K. N.; Martin, H. J.; List, B. J. Am. Chem.
Soc. 2003, 125, 2475–2479.
(15) Hoang, L.; Bahmanyar, S.; Houk, K. N.; List, B. J. Am. Chem.
Soc. 2003, 125, 16–17.
(16) Sharma, A.; Sunoj, R. Angew. Chem., Int. Ed. 2010, 49,
6373–6377.
’ EXPERIMENTAL SECTION
All monitored reactions were conducted inside standard 5 mm NMR
tubes by adding freshly distilled aldehyde 1a or a mixture of 2a,b or 3
(30 μmol, each) to a suspension of L-proline (10, 20, 50, or 100 mol %)
in 0.6 mL of DMSO-d6 (optionally with 0.5 or 1 vol % of D2O). The
NMR tube was transferred to the spectrometer immediately after the
mixing of all reacting components.
NMR measurements were performed at 300 K on two different NMR
spectrometers (600.13 and 600.25 MHz), the latter equipped with a
cryoprobe. NMR data were processed and evaluated with TOPSPIN 2.1,
and the included DAISY software was used for the simulation of NMR
spectra. The concentrationꢀtime curves of the observed aldol addition
and condensation products were fit and its slope was calculated with the
help of Origin 6.0.
(17) List, B.; Hoang, L.; Martin, H. J. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 5839–5842.
(18) Seebach, D.; Beck, A. K.; Badine, D. M.; Limbach, M.;
Eschenmoser, A.; Treasurywala, A. M.; Hobi, R.; Prikoszovich, W.;
Linder, B. Helv. Chim. Acta 2007, 90, 425–471.
(19) Zotova, N.; Broadbelt, L. J.; Armstrong, A.; Blackmond, D. G.
Bioorg. Med. Chem. Lett. 2009, 19, 3934–3937.
(20) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. Angew. Chem., Int.
Ed. 2010, 49, 4997–5003.
The aldol dimers 2a,b were synthesized according to a slightly
modified literature procedure of MacMillan and co-workers.39
(21) Bock, D. A.; Lehmann, C. W.; List, B. Proc. Natl. Acad. Sci. U.S.
A. 2010, 107, 20636–20641.
’ ASSOCIATED CONTENT
(22) Kanzian, T.; Lakhdar, S.; Mayr, H. Angew. Chem., Int. Ed. 2010,
49, 9526–9529.
(23) Blackmond, D. G.; Moran, A.; Hughes, M.; Armstrong, A.
S
Supporting Information. Reaction profiles with varying
b
amounts of catalyst as well as with butyraldehyde (including
NMR spectra and assignments), observations on enal E/Z-selectivity,
spectral deconvolution for partially deuterated 4b, NMR spec-
trum and assignment of the reaction mixture with acetaldehyde,
and time-dependent diastereoselectivity of the propionaldehyde
aldol dimers in DMF. This material is available free of charge via
the Internet at http://pubs.acs.org.
J. Am. Chem. Soc. 2010, 132, 7598–7599.
(24) For a recent GC-based investigation of aldol/Mannich ratios in
the condensation of isobutyraldehyde with acetone, see: Domínguez de
María, P.; Bracco, P.; Castelhano, L. F.; Bargeman, G. ACS Catal. 2011,
1, 70–75.
(25) Trost, B. M.; Brindle, C. S. Chem. Soc. Rev. 2010, 39, 1600–
1632.
(26) Recently, the organocatalytic asymmetric Knoevenagel con-
densation has been introduced: Lee, A.; Michrowska, A.; Sulzer-Mosse,
S.; List, B. Angew. Chem., Int. Ed. 2011, 50, 1707–1710.
(27) Erkkil€a, A.; Pihko, P. M. Eur. J. Org. Chem. 2007, 4205–4216.
(28) Zumbansen, K.; D€ohring, A.; List, B. Adv. Synth. Catal. 2010,
352, 1135–1138.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: kirsten.zeitler@chemie.uni-regensburg.de; ruth.gschwind@
chemie.uni-regensburg.de
(29) List, B.; Pojarliev, P.; Castello, C. Org. Lett. 2001, 3, 573–575.
(30) Noziꢀere, B.; Cꢁordova, A. J. Phys. Chem.
112, 2827–2837.
A 2008,
’ ACKNOWLEDGMENT
We are grateful to Johannes Franz for the synthesis of the aldol
dimers 2a,b. This work was supported by the DFG (SPP 1179).
Scholarships by the Cusanuswerk and the Studienstiftung des
deutschen Volkes are gratefully acknowledged.
(31) Arend, M.; Westermann, B.; Risch, N. Angew. Chem., Int. Ed.
1998, 37, 1044–1070.
(32) Ishikawa, T.; Uedo, E.; Okada, S.; Saito, S. Synlett 1999, 450–452.
(33) Cꢁordova, A.; Notz, W.; Barbas, C. F., III. J. Org. Chem. 2002,
67, 301–303.
(34) Casas, J.; Sundꢁen, H.; Cꢁordova, A. Tetrahedron Lett. 2004,
45, 6117–6119.
’ REFERENCES
(1) Berkessel, A.; Gr€oger, H. Asymmetric Organocatalysis - From
Biomimetic Concepts to Applications in Asymmetric Synthesis; Wiley-
VCH: Weinheim, 2005.
(35) Smith, M. B.; March, J. March’s Advanced Organic Chemistry:
Reactions, Mechanisms, and Structure, 6th ed.; John Wiley & Sons, Inc.:
Hoboken, 2006.
3014
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015

The Journal of Organic Chemistry
FEATURED ARTICLE
(36) Hong, B.-C.; Wu, M.-F.; Tseng, H.-C.; Liao, J.-H. Org. Lett.
2006, 8, 2217–2220.
(37) Erkkil€a, A.; Pihko, P. M. J. Org. Chem. 2006, 71, 2538–2541.
(38) A similar dual activation of carbonyl species by proline has
recently been proposed for the organocatalytic homodimerization of
enones: Guidi, V.; Sandoval, S.; McGregor, M. A.; Rosen, W. Tetra-
hedron Lett. 2010, 51, 5086–5090.
(39) Northrup, A. B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002,
124, 6798–6799.
(40) Miyano, M. J. Org. Chem. 1981, 46, 1854–1857.
(41) Zotova, N.; Franzke, A.; Armstrong, A.; Blackmond, D. G.
J. Am. Chem. Soc. 2007, 129, 15100–15101.
(42) Zotova, N.; Moran, A.; Armstrong, A.; Blackmond, D. Adv.
Synth. Catal. 2009, 351, 27652769.
(43) Interestingly, the rate-limiting step of the proline-catalyzed
intramolecular HajosꢀParrishꢀEderꢀSauerꢀWiechert44,45 aldol reac-
tion has been shown to precede the CꢀC bond formation.46 However,
this may be attributed to the acceleration of the CꢀC bond formation by
the “apparent concentration effect in the intramolecular case”.19
(44) Eder, U.; Sauer, G.; Wiechert, R. Angew. Chem., Int. Ed. 1971,
10, 496–497.
(45) Hajos, Z. G.; Parrish, D. R. J. Org. Chem. 1974, 39, 1615–1621.
(46) Zhu, H.; Clemente, F. R.; Houk, K. N.; Meyer, M. P. J. Am.
Chem. Soc. 2009, 131, 1632–1633.
(47) This type of graphical representation is used to show the clear
trend of a strongly increasing formation rate for the condensation product
3 upon higher catalyst loadings in contrast to the nearly unchanging
formation rates of the aldol products 2a,b which are by far less affected by
increasing catalyst amounts and only exceed the formation rate of the
condensation product at the lowest catalyst loading (approximately 0.7%
enamine concentration ≈ 10 mol% of proline). Because of the limited
solubility of proline in DMSO at amounts exceeding 30 mol % of catalyst
loading and the influence of water on the reaction rates (enamine
formation) as noticed by us and others,19 we found the relation of the
formation rates against the actual present catalyst amount represented by
the enamine concentration to be more accurate, especially in case of short
reaction times. (Note: enamine formation is fast and reversible according
to our earlier EXSY analyses.20)
(48) Alcaide, B.; Almendros, P. Angew. Chem., Int. Ed. 2008,
47, 4632–4634.
(49) García-García, P.; Ladꢁep^eche, A.; Halder, R.; List, B. Angew.
Chem., Int. Ed. 2008, 47, 4719–4721.
(50) Hayashi, Y.; Itoh, T.; Aratake, S.; Ishikawa, H. Angew. Chem., Int.
Ed. 2008, 47, 2082–2084.
(51) Hayashi, Y.; Itoh, T.; Ohkubo, M.; Ishikawa, H. Angew. Chem.,
Int. Ed. 2008, 47, 4722–4724.
(52) Hayashi, Y.; Okano, T.; Itoh, T.; Urushima, T.; Ishikawa, H.;
Uchimaru, T. Angew. Chem., Int. Ed. 2008, 47, 9053–9058.
(53) Hayashi, Y.; Samanta, S.; Itoh, T.; Ishikawa, H. Org. Lett. 2008,
10, 5581–5583.
(54) http://riodb01.ibase.aist.go.jp/sdbs/, accessed March 11, 2011.
(55) Bertelsen, S.; Marigo, M.; Brandes, S.; Diner, P.; Jørgensen,
K. A. J. Am. Chem. Soc. 2006, 128, 12973–12980.
(56) Hong, B.-C.; Wu, M.-F.; Tseng, H.-C.; Huang, G.-F.; Su, C.-F.;
Liao, J.-H. J. Org. Chem. 2007, 72, 8459–8471.
(57) Utsumi, N.; Zhang, H.; Tanaka, F.; Barbas, C. F., III. Angew.
Chem., Int. Ed. 2007, 46, 1878–1880.
(58) de Figueiredo, R. M.; Fr€ohlich, R.; Christmann, M. Angew.
Chem., Int. Ed. 2008, 47, 1450–1453.
(59) Liu, K.; Chougnet, A.; Woggon, W.-D. Angew. Chem., Int. Ed.
2008, 47, 5827–5829.
(60) Han, B.; Xiao, Y.-C.; He, Z.-Q.; Chen, Y.-C. Org. Lett. 2009,
11, 4660–4663.
(61) Marquꢁes-Lꢁopez, E.; Herrera, R. P.; Marks, T.; Jacobs, W. C.;
K€onning, D.; de Figueiredo, R. M.; Christmann, M. Org. Lett. 2009,
11, 4116–4119.
(62) Bencivenni, G.; Galzerano, P.; Mazzanti, A.; Bartoli, G.;
Melchiorre, P. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 20642–20647.
3015
dx.doi.org/10.1021/jo200431v |J. Org. Chem. 2011, 76, 3005–3015