The elusive enamine intermediate in proline-catalyzed aldol reactions: NMR detection, formation pathway, and stabilization trends

Article abstract of DOI:10.1002/anie.200906629

The missing link: The elusive enamine intermediate of nucleophilic proline catalysis was detected and stereochemically characterized by NMR analysis of the aldehyde self-aldolization reaction in dipolar aprotic solvents. NMR exchange spectroscopy (EXSY) w

Full text of DOI:10.1002/anie.200906629

Angewandte
Chemie
DOI: 10.1002/anie.200906629
Organocatalysis
The Elusive Enamine Intermediate in Proline-Catalyzed Aldol
Reactions: NMR Detection, Formation Pathway, and Stabilization
Trends**
Markus B. Schmid, Kirsten Zeitler, and Ruth M. Gschwind*
Dedicated to Professor Horst Kessler on the occasion of his 70th birthday
The detection and characterization of intermediates in
organic reactions is crucial for the understanding of mecha-
nisms and the rational optimization of reaction conditions.
However, especially in the rapidly expanding field of
asymmetric organocatalysis,[1–4] mechanistic studies are
scarce compared to new synthetic applications. Therefore,
organocatalysis was characterized as still being “in its
exploratory discovery phase before it can become contem-
plating”.[5] Among the different organocatalytic activation
modes and the wide range of identified general concepts,[6,7]
Brønsted acid[8,9] and Lewis base catalysis[10] have proven to
be broadly applicable. After the proline-catalyzed aldol
reactions (both origin[11,12] and prototype[13] for asymmetric
aminocatalysis), secondary amines[14–16] are preferentially
employed to activate substrates via iminium[17] or enamine
intermediates.[18] The generally accepted mechanism of
enamine catalysis[14,19] is based upon experimental[20] and
theoretical studies[21] that suggest a central enamine inter-
mediate in the proline-catalyzed reactions.
To the best of our knowledge, such enamine intermediates
have never been detected in situ; only product enamines[22,23]
or dienamines,[24] and dienamine intermediates[25] have been
reported—for different catalysts. In contrast, putative enam-
ine intermediates were synthesized, isolated, and character-
ized,[5,26–28] and recently an enamine intermediate was
observed in the crystal structure of an aldolase antibody.[29]
So far, in situ NMR spectroscopic approaches have only
resulted in the detection of the isomeric oxazolidi-
nones,[20,30–34] supposedly resulting from an “unwanted and
rate-diminishing parasitic equilibrium”,[20,35] which was
believed to be responsible for the inability to observe the
enamines. In fact, equilibria involving oxazolidinones have
been reported[20,32–34] and their energetic preference has been
calculated.[36] An alternate mechanistic model of proline-
catalyzed aldol reactions that attributes a pivotal role to the
predominant oxazolidinones has been proposed,[33] and
indeed such oxazolidinones have successfully been used as
“soluble proline catalysts”.[31–34]
The detection of enamine intermediates in proline-
catalyzed aldol reactions is the missing piece of evidence for
the commonly accepted mechanism of enamine catalysis.
Moreover, the structural characterization of key enamine
intermediates, the elucidation of their formation, and their
stabilization are important for a better understanding of and
the control of organocatalytic reactions, which could in turn
present new options in accelerating and controlling enamine-
catalyzed reactions.
Herein we present our real-time NMR studies that detail
the first detection and structural characterization of enamine
intermediates in proline-catalyzed aldol reactions. In addi-
tion, their direct formation from oxazolidinones is evidenced
in the solvent dimethylsulfoxide (DMSO). Moreover, the
influences of the carbonyl species, its substitution pattern, as
well as the effect of the solvent and its water content upon the
detectable enamine concentration are demonstrated.
The self-aldolization of propionaldehyde (1; c = 50 mm),
catalyzed by 20 mol% l-proline, in [D6]DMSO at 300 K
(Figure 1a) was used as a model reaction in our enamine
studies. The reaction was conducted within an NMR tube and
monitored by one-dimensional 1H NMR spectra (Figure 1b).
Two diastereomeric aldol dimers 2a and 2b, and the
condensation product 3 were observed, which is in accord
with previous studies.[37,38] In addition, three intermediate
species were detected, each of which disappeared at identical
rates (Figure 1c). By employing 100 mol% l-proline (which
led to an acceleration of the reaction and the predominant
formation of the condensation product 3, possibly through a
Mannich-like mechanism[39,40]), we successfully increased the
total amount of the intermediates from about 8% to 25–30%
without changing their relative ratios (see the Supporting
Information). This increased amount of the intermediates
allowed unambiguous identification and full characterization
of these species as the enamine 5 and the two diastereomeric
oxazolidinones 4a and 4b (in a ratio of about 3.8:1) by using
real-time homo- and heteronuclear two-dimensional NMR
spectroscopy during the reaction (see the Supporting Infor-
mation for the complete NMR assignments).
[*] M. B. Schmid, Dr. K. Zeitler, Prof. Dr. R. M. Gschwind
Institut fꢀr Organische Chemie, Universitꢁt Regensburg
Universitꢁtsstraße 31, 93053 Regensburg (Germany)
Fax: (+49)941-943-4617
E-mail: ruth.gschwind@chemie.uni-regensburg.de
[**] This work was supported by the DFG (SPP 1179). Scholarships by
the Cusanuswerk and the Studienstiftung des deutschen Volkes are
gratefully acknowledged. We thank the anonymous referees of this
manuscript for exceptionally helpful and constructive suggestions.
The detection of the ene unit of 5 is rather straightforward
due to characteristic 1H chemical shifts and multiplet patterns
(Figure 2b). The connection of the ene unit to the proline
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200906629.
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4997

Communications
Figure 2. NMR spectroscopic characterization of 5: a) nomenclature
1
and NOE pattern; b) sections of the H NMR spectra (top: experimen-
tal, bottom: simulated) of the ene spin system; c) section of a
1H,13C HMBC spectrum revealing through-bond correlations from H1
1
to Ca and Cd; d) sections of a H,1H NOESY spectrum showing
through-space correlations between proton H1 and Ha and Hd,Hd’.
(Figure 2d). This interpretation is supported by the slightly
more intensive HMBC cross-peak between H1 and Cd
compared to Ca (Figure 2c). The finding of an s-trans
E enamine is in agreement with reports on a crystal struc-
ture[5] and DFT calculations[28] of an aldehyde-derived
enamine of a prolinol ether-type catalyst, and the generally
accepted mechanism of enamine catalysis.[20]
Figure 1. a) l-Proline-catalyzed self-aldolization of propionaldehyde (1).
Reaction profile of substrate and products (b) and of intermediates (c)
as monitored by integration of selected NMR resonances in one-
1
dimensional H NMR spectra. Note: The total amount of C3 moieties
stemming from 1, detected in the first spectrum, was set to 100%.
Having both oxazolidinones and an enamine intermediate
at hand, we investigated the “parasitic equilibrium”[20] and the
associated reaction pathways by NMR exchange spectroscopy
(EXSY). By tracking the reaction of 1 and l-proline (see
EXSY trace for the resonance of 1 in Figure 3a), two
exchange cross-peaks, of almost equal intensity, for the
isomeric oxazolidinones 4a and 4b indicate similar formation
rates. Interestingly, in contrast, no exchange peak between 1
and 5, which would reveal the formation of 5 from 1 and
proline, was detected under these experimental conditions.
The EXSY trace for the resonance of 4a (example for both 4a
and 4b; Figure 3a) indicates a significantly faster exchange
between 4a and 4b than that between the other species, which
may account for the thermodynamic equilibration between
4a and 4b. A cross-peak to 1 shows the reverse reaction. In
addition and despite the low amounts of the intermediates,
exchange peaks between 4a and 5 as well as 4b and 5 are
residue is proven by through-bond correlations (1H,13C
HMBC spectrum; Figure 2c) of H1 to Ca and Cd, thereby
ruling out the existence of a potential enol that was recently
suggested for a different but related organocatalyst.[41,42] This
finding is additionally supported by through-space correla-
tions (1H,1H NOESY spectrum; Figure 2a and d) of H1 and
H2 to Ha and Hd.
In addition, our spectroscopic results allow the stereo-
chemical characterization of enamine 5. The size of the scalar
coupling constant of H1 and H2 (3JH1,H2 = 13.7 Hz) is indica-
tive of an E configuration for the enamine double bond. The
Z isomer was not detected, even at higher E-enamine
concentrations (see below). Concerning the conformation of
ꢀ
the exocyclic N C bond, the NOESY cross-peak pattern
indicates an s-trans arrangement since the NOE between H1
and Ha is much larger than those between H1 and Hd,Hd’
4998 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003

Angewandte
Chemie
pathway would be expected to have different enamine
formation rates depending upon the starting oxazolidinone
(Figure 4b).
To rule out the possibility that the missing EXSY cross-
peak between I and V is simply below the detection limit,
additional positive evidence for the direct enamine–oxazoli-
dinone conversion was collected from our EXSY spectra: For
the iminium mechanism (Figure 4a), the aldehyde I, the
oxazolidinones IVa and IVb, and the enamine V should all
interconvert through the central iminium intermediate VI,
and each species should be formed from this iminium at a
distinct rate. By using ratios of these formation rates, for
example, the volume of the EXSY cross-peak IVa!I divided
by the volume of the EXSY cross-peak IVa!V, the
contribution of IVa is cancelled and a value of the relative
rates kal/ken is obtained. This value should be independent of
how the central intermediate was formed (e.g., starting from
either IVa or IVb). Therefore, in the case of a common
iminium intermediate VI, the resulting ratios of the formation
rate should be nearly identical [Eq. (1)].
IVa ! I
IVb ! I
kal
ð1Þ
ꢁ
ꢁ
IVa ! V IVb ! V ken
In contrast, significant differences in these ratios would
exclude a common central intermediate or a pool of rapidly
interconverting intermediates, and instead provide evidence
for the direct formation of the enamine V from the
oxazolidinones IVa,b (Figure 4b). To apply this key differ-
entiator to the two exchange pathways discussed, the ratios of
the crucial EXSY cross-peak volumes were calculated for
three different reaction mixtures each derived from a differ-
ent aldehyde (Figure 4d; representative EXSY traces are
shown for 3-methyl-butyraldehyde in Figure 4c). For all pairs
of EXSY cross-peak ratios significantly different volume
ratios, that is, formation-rate ratios (by factors of 2.5 to 5) are
obtained. These values indicate that the enamine V is indeed
formed directly from the oxazolidinones IVa,b (Figure 4b)
under the experimental conditions used, and excludes in this
process the presence of a common central intermediate or a
pool of rapidly interconverting intermediates such as the
commonly proposed iminium ion.
In view of these intermediate equilibria, we investigated
the influence of the amount of catalyst, the water content of
the sample, the solvent properties, and the substitution
pattern of the carbonyl compound upon the appearance and
the relative ratios of oxazolidinones and enamines. First, we
showed that for the reaction depicted in Figure 1a, increasing
the amount of catalyst from 20 mol% to 100 mol% leads to
higher concentrations of the intermediates, but does not alter
their equilibrium ratios (see the Supporting Information).
Therefore, all experiments discussed hereafter were per-
formed at substrate concentrations of 50 mm with 100 mol%
l-proline. Next, the influence of the amount of water upon the
overall intermediate concentration and the relative ratios of
the intermediates 4a, 4b, and 5 was investigated using 1 and
3-methyl-butyraldehyde because the latter had the highest
enamine amount (see Table 2 and discussion below). A
stepwise increase of the water content led to a reduction of
Figure 3. Experimentally detected exchange processes of a propion-
aldehyde/l-Pro reaction mixture by EXSY methods: a) rows from the
1H,1H EXSY spectrum for the protons of 1 (top) and 4a (bottom),
highlighted in gray, showing exchange peaks to the protons of 1
(d=9.69 ppm), 5 (d=6.06 ppm), 4b (d=5.30 ppm), and 4a
(d=5.08 ppm), and the missing exchange between 1 and 5 (circle;
diagonal peaks are omitted for clarity); b) schematic summary of the
exchange processes and the qualitative exchange intensities detected
by EXSY spectra.
detected. The observed exchange matrix is schematically
summarized in Figure 3b.
For the enamine formation from aldehydes as well as from
oxazolidinones, mechanisms proceeding via iminium inter-
mediates have been proposed[14,33,43] wherein a central
iminium species would be responsible for the interconversion
between aldehydes, enamines, and oxazolidinones (Fig-
ure 4a). The iminium species VI is not detected by the
NMR methods herein, therefore its structure cannot be
identified and, in addition, the existence of various rapidly
interconverting iminium and/or carbinolamine species seems
possible. However, the fast exchange between the two
diastereomeric oxazolidinones IVa and IVb (Figure 3a and
4c) shows that the interconversion of different putative
intermediate species has to be fast compared to either
enamine or aldehyde formation. Therefore, for the sake of
clarity in Figure 4a only the iminium VI is depicted and
mentioned hereafter. Surprisingly and in contrast to this
commonly assumed iminium pathway, the missing cross-peak
between aldehyde I and enamine V indicates that V is formed
not via an iminium intermediate VI from aldehyde I, but
directly from the oxazolidinones IVab. For such a direct
formation of enamines from oxazolidinones a concerted E2
mechanism has recently been proposed,[33] and this reaction
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4999

Communications
mediates 4a, 4b, and 5 remain unaffected up to
about 2 vol% of water (limit of a reliable relative
integration, see the Supporting Information).
This suggests that the solvent properties respon-
sible for the relative stabilization of 4a, 4b, and 5
do not change significantly in these solvent
mixtures—at least on the level of microsolvation.
In addition, a severe line broadening of the
oxazolidinone signals and changes in the chem-
ical shifts of the oxazolidinones and the alde-
hydes are observed with increasing water con-
tent, whereas the line widths and the chemical
shifts of the enamine signals remain essentially
unaffected (see the Supporting Information).
This corroborates the equilibria presented in
Figures 3b and 4b showing the formation of the
oxazolidinones from aldehydes and proline, and
then a subsequent step for the formation of the
enamines under the experimental conditions
applied.
Secondly, we were interested in the extent to
which the amount of the enamine is influenced
by solvent properties. In addition to [D6]DMSO,
we repeated the self-aldolization of 1 in other
protic and dipolar aprotic organic solvents that
are successfully used in proline-catalyzed reac-
tions ([D3]MeCN, [D7]DMF, and [D4]MeOH).
The results as well as relevant solvent parameters
are summarized in Table 1. In DMF, just as in
DMSO, all three intermediates 4a, 4b, and 5
were detected, however, the amount of 5 was
reduced from 9% to 4% and the reaction is
considerably slower. In MeCN the reaction was
very slow and only 4a and 4b were detected,
whereas in [D4]MeOH the reaction proceeded
quickly, such that no intermediates could be
identified. A correlation of the amount of 5 to the
parameters of the respective solvents[46,47] reveals
that their hydrogen-bond donor (a) and acceptor
(b) ability, and not their dielectric constant (e),
are crucial for the relative enamine amount.
DMSO and DMF, having a = 0, allow for enam-
ine detection and the amount of enamine corre-
lates with the respective b values of DMSO and
DMF. In MeCN, which has a poor b value (and,
in addition, low a value), only oxazolidinones are
observed; this data is in good agreement with
reports on optimal oxazolidinone yields in
MeCN.[30] In contrast, for the dipolar protic
solvent MeOH that possesses a b value similar
to that of DMSO and DMF, and a high a value,
no enamines are detected. Therefore, both a high
Figure 4. Differentiation between alternative enamine formation pathways by analysis
of EXSY cross-peak ratios: a) previously proposed mechanism with a central iminium
intermediate or a pool of rapidly interconverting intermediates (see text); b) exper-
imentally confirmed direct enamine formation from oxazolidinones; c) relevant
sections of the 1H,1H EXSY spectrum of a 3-methylbutyraldehyde/l-Pro mixture for the
highlighted 1H resonances; d) quantitative evaluation of the 2D EXSY cross-peaks for
different aldehydes: the different ratios of the formation rates disprove the presence
of a central intermediate as depicted in (a). [a] Ratio determined by integration of the
peaks in the 1D EXSY trace.
the overall amount of the intermediates present until their
detection limit was reached at around 15 vol% of water in
DMSO (for details see the Supporting Information). This
result is in agreement with previous reports on the reduction
of the amount of oxazolidinone upon the addition of
water,[44,45] as well as with their reversible formation by
condensation. Interestingly, the relative ratios of the inter-
b value and a low a value together are favorable for enamine
stabilization.
Our experimental observations suggest that the stabiliza-
tion of the carboxylic proton in the proline-enamine species
by solvents having high b values decreases the collapse of the
enamine into the oxazolidinone and is therefore crucial for
the detectability of enamines. The exquisite capability of
5000 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003

Angewandte
Chemie
Table 1: Selected solvent properties and their influence upon the relative
ratios of the intermediates 4a, 4b, and 5 (total amount of intermediates
was set to 100%).[a]
and by the preference of ketone-derived oxazolidinones as
anticipated by the Thorpe–Ingold effect.[51,52]
Next, starting from 1, we investigated the influence of
a- and b-alkyl substituents on the aldehyde upon the relative
ratios of oxazolidinones and enamines. For acetaldehyde
(Table 2, entry 3), self-aldolization was so fast that among a
variety of intermediates and products only the dienamine
could be identified. In contrast, for isobutyraldehyde (Table 2,
entry 4) no product formation was observed and only tiny
amounts of enamine were detected, thereby suggesting—as
has already been reported[30]—a-substitution on the aldehyde
favors oxazolidinones instead of enamines, probably because
of the presence of the unfavorable allylic strain in the
enamine which is more predominant than the stabilization of
the double-bond by higher alkyl substitution. For butyralde-
hyde (Table 2, entry 6) higher enamine ratios than those for
propionaldehyde (1; Table 2, entry 5) were detected, and can
be ascribed to the stronger + I- and hyperconjugation effects
of the ethyl group. Additional alkyl substitution in the
b position, as shown for 3-methylbutyraldehyde (Table 2,
entry 7), leads to an additional increase of the enamine ratio
which might be accounted for by further + I-effects and a
reduced reactivity resulting from steric crowding.
On the basis of the NMR detection and characterization
of reaction intermediates, and the observation of the reaction
pathways by EXSY methods (see Figure 3 and 4 and the
Supporting Information), we propose a slightly modified
mechanism of proline-catalyzed aldol reactions in dipolar
aprotic solvents exclusively having hydrogen-bond acceptor
properties (Figure 5). First, oxazolidinones 4 are formed from
1 and l-proline, probably via the commonly proposed
iminium intermediate 6, which also allows the rapid isomer-
ization of 4a and 4b. The extremely short lifetime of 6
resulting from its immediate collapse into 4a,b may account
for the inability to detect iminium ion 6 in our study, and can
be rationalized by the poor solvation capability of DMSO
towards anionic species,[53] that is, the carboxylate moiety.
Next, the s-trans E-enamine 5 is formed directly from 4a and
4b, as evidenced by EXSY analysis. For such a direct
transformation of oxazolidinones into enamines, Seebach
et al. recently proposed a concerted E2 mechanism.[33] The
latter steps of the reaction are mostly in agreement with the
generally accepted mechanism of enamine catalysis.[14,19] We
speculate that two very small intermediate peaks, slightly
upfield shifted from those of 2a,b, belong to the iminium
Solvent
er
a
b
4a [%]
4b [%]
5 [%]
[D6]DMSO
[D7]DMF
[D3]MeCN
[D4]MeOH
46.5
37.8
35.9
32.7
0.00
0.00
0.19
0.98
0.76
0.69
0.40
0.66
72
77
80
n.d.
19
19
20
n.d.
9
4
n.d.
n.d.
[a] n.d.=not detected; er =dielectric constant; a,b=Kamlet–Taft
parameters characterizing hydrogen-bond donor and acceptor ability,
respectively.
DMSO to form complexes with hydroxy groups was recently
explored by means of a
3-hydroxyflavone (3HF) which, upon undergoing intermo-
lecular proton transfer, forms a 3HFꢀ·DMSOH+ complex.[48]
Experimental findings suggest that DMF follows this trend
better than MeCN.[48–50] To verify such an interaction between
5 and DMSO, we used diffusion-ordered spectroscopy
(DOSY) and found significantly slower diffusion of 5
compared to that of 4a and 4b (see the Supporting
Information); this result cannot be simply explained by the
different shapes of the molecules, but instead we believe this
to be indicative of “strong interactions” between the carboxy
group of 5 and solvent molecules.
Furthermore, we studied the influence of different
substitution patterns of the carbonyl compound upon the
amount of enamine observed (Table 2). First, enamine
detection was successful only for aldehydes, and not for
ketones: Butanone (Table 2, entry 1) did not form any adduct
with l-proline under our experimental conditions, whereas
for acetone (Table 2, entry 2), only the oxazolidinone was
observed, which is in agreement with previous NMR stud-
ies.[20] In both cases, neither an enamine intermediate nor a
reaction was observed. This result may be rationalized by the
lower carbonyl activity of ketones compared to aldehydes,
Table 2: Influence of the carbonyl species and its substitution pattern
upon the preference of the equilibrium between the enamine and the
oxazolidinones in [D6]DMSO.[a]
ꢀ
species 7 (see the Supporting Information) formed by C C
bond formation between 5 and 1. Subsequently, product
oxazolidinones 8 are observed and in the final step the
catalyst is released by hydrolysis to 2a,b.
In summary, the first in situ detection of enamine
intermediates in proline-catalyzed aldol reactions, accom-
plished by using NMR methods, is reported and new insights
into the stabilization and the formation pathway of enamines
in dipolar aprotic solvents are presented. Exclusively
E-configured s-trans enamines are detected; in DMSO,
these enamines are formed directly from oxazolidinones
and not via central iminium or iminium-like intermediates as
evidenced by EXSY analyses. The position of these oxazoli-
dinone-enamine equilibria is not affected by additional water
Entry
R1
R2
R3
IVa [%]
IVb [%]
V [%]
1
2
3
4
5
6
7
Me
H
H
Me
Me
Et
H
H
H
Me
H
Me
Me
H
H
H
n.d.
n.d.
n.d.
n.d.
n.d.
1
9
13
a 100% a
n.d.
90
72
64
57
n.d.
9
19
23
23
H
H
H
H
iPr
20
[a] n.d.=not detected.
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5001

Communications
[4] “Asymmetric Organocatalysis” (Ed.: B. List): Top. Curr. Chem.
2009, 291.
[5] D. Seebach, U. Groselj, D. M. Badine, W. B. Schweizer, A. K.
Beck, Helv. Chim. Acta 2008, 91, 1999 – 2034.
[6] J. Seayad, B. List, Org. Biomol. Chem. 2005, 3, 719 – 724.
[7] D. W. C. MacMillan, Nature 2008, 455, 304 – 308.
[8] T. Akiyama, Chem. Rev. 2007, 107, 5744 – 5758.
[9] Hydrogen Bonding in Organic Synthesis (Ed.: P. M. Pihko),
Wiley-VCH, Weinheim, 2009.
[10] S. E. Denmark, G. L. Beutner, Angew. Chem. 2008, 120, 1584 –
1663; Angew. Chem. Int. Ed. 2008, 47, 1560 – 1638.
[11] U. Eder, G. Sauer, R. Wiechert, Angew. Chem. 1971, 83, 492 –
493; Angew. Chem. Int. Ed. Engl. 1971, 10, 496 – 497.
[12] Z. G. Hajos, D. R. Parrish, J. Org. Chem. 2002, 39, 1615 – 1621.
[13] B. List, R. A. Lerner, C. F. Barbas III, J. Am. Chem. Soc. 2000,
122, 2395 – 2396.
[14] B. List, Chem. Commun. 2006, 819 – 824.
[15] P. Melchiorre, M. Marigo, A. Carlone, G. Bartoli, Angew. Chem.
2008, 120, 6232 – 6265; Angew. Chem. Int. Ed. 2008, 47, 6138 –
6171.
[16] S. Bertelsen, K. A. Jørgensen, Chem. Soc. Rev. 2009, 38, 2178 –
2189.
[17] A. Erkkilꢁ, I. Majander, P. M. Pihko, Chem. Rev. 2007, 107,
5416 – 5470.
[18] S. Mukherjee, J. W. Yang, S. Hoffmann, B. List, Chem. Rev. 2007,
107, 5471 – 5569.
[19] B. List, Acc. Chem. Res. 2004, 37, 548 – 557.
[20] B. List, L. Hoang, H. J. Martin, Proc. Natl. Acad. Sci. USA 2004,
101, 5839 – 5842.
Figure 5. Proposal for the catalytic cycle of proline-catalyzed aldol
reactions in dipolar aprotic solvents based on NMR-detected inter-
mediates and exchange pathways (the equilibrium arrows are omitted
for clarity).
[21] C. Allemann, R. Gordillo, F. R. Clemente, P. H.-Y. Cheong,
K. N. Houk, Acc. Chem. Res. 2004, 37, 558 – 569.
[22] S. Lakhdar, T. Tokuyasu, H. Mayr, Angew. Chem. 2008, 120,
8851 – 8854; Angew. Chem. Int. Ed. 2008, 47, 8723 – 8726.
[23] S. Lakhdar, R. Appel, H. Mayr, Angew. Chem. 2009, 121, 5134 –
5137; Angew. Chem. Int. Ed. 2009, 48, 5034 – 5037.
[24] A. L. Fuentes de Arriba, L. Simon, C. Raposo, V. Alcazar, J. R.
Moran, Tetrahedron 2009, 65, 4841 – 4845.
[25] S. Bertelsen, M. Marigo, S. Brandes, P. Diner, K. A. Jørgensen, J.
Am. Chem. Soc. 2006, 128, 12973 – 12980.
[26] S.-I. Yamada, G. Otani, Tetrahedron Lett. 1969, 10, 4237 – 4240.
[27] T. J. Peelen, Y. Chi, S. H. Gellman, J. Am. Chem. Soc. 2005, 127,
11598 – 11599.
[28] U. Groselj, D. Seebach, D. M. Badine, W. B. Schweizer, A. K.
Beck, I. Krossing, P. Klose, Y. Hayashi, T. Uchimaru, Helv. Chim.
Acta 2009, 92, 1225 – 1259.
[29] X. Zhu, F. Tanaka, R. A. Lerner, C. F. Barbas III, I. A. Wilson, J.
Am. Chem. Soc. 2009, 131, 18206 – 18207.
[30] F. Orsini, F. Pelizzoni, M. Forte, M. Sisti, G. Bombieri, F.
Benetollo, J. Heterocycl. Chem. 1989, 26, 837 – 841.
[31] H. Iwamura, S. P. Mathew, D. G. Blackmond, J. Am. Chem. Soc.
2004, 126, 11770 – 11771.
or the amount of the catalyst. The presence of strong
hydrogen-bond acceptor properties together with the absence
of hydrogen-bond donor properties of the solvent increase the
amount of enamine present. Ketones and a-branched alde-
hydes show no or very low enamine concentrations, whereas
b-alkyl substituents of propionaldehydes cause enamine
stabilization. Our results corroborate not only the central
role of enamines in proline-catalyzed aldol reactions, but also
elucidate a new role of the oxazolidinones as a bridge
between aldehydes and enamines in dipolar aprotic solvents.
In conclusion, the first detailed insights into the oxazolidi-
none-enamine conversion processes presented herein should
allow rational and directed optimizations of proline-catalyzed
reactions.
Received: November 24, 2009
Revised: April 10, 2010
[32] H. Iwamura, D. H. Wells, Jr., S. P. Mathew, M. Klussmann, A.
Armstrong, D. G. Blackmond, J. Am. Chem. Soc. 2004, 126,
16312 – 16313.
[33] D. Seebach, A. K. Beck, D. M. Badine, M. Limbach, A.
Eschenmoser, A. M. Treasurywala, R. Hobi, W. Prikoszovich,
B. Linder, Helv. Chim. Acta 2007, 90, 425 – 471.
[34] C. Isart, J. Bures, J. Vilarrasa, Tetrahedron Lett. 2008, 49, 5414 –
5418.
Published online: June 8, 2010
Keywords: aldol reaction · enamines · NMR spectroscopy ·
.
organocatalysis · proline catalysis
[35] In contrast, in an ESI-MS study, a cationic enamine species was
detected, but the missing observation of the isomeric oxazoli-
dinones suggests a method-induced effect: C. Marquez, J. O.
Metzger, Chem. Commun. 2006, 1539 – 1541.
[36] Y.-F. Hu, X. Lu, Chin. J. Struct. Chem. 2008, 27, 547 – 552.
[37] R. Mahrwald, B. Costisella, B. Gꢂndogan, Synthesis 1998, 262 –
264.
[1] A. Berkessel, H. Grꢀger, Asymmetric Organocatalysis—From
Biomimetic Concepts to Applications in Asymmetric Synthesis,
Wiley-VCH, Weinheim, 2005.
[2] special issue on organocatalysis: Chem. Rev. 2007, 107(12).
[3] Ernst Schering Foundation Symposium Proceedings “Organo-
catalysis” (Eds.: M. T. Reetz, B. List, S. Jaroch, H. Weinmann),
Springer, Berlin, 2008.
5002 www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003

Angewandte
Chemie
[38] A. B. Northrup, D. W. C. MacMillan, J. Am. Chem. Soc. 2002,
124, 6798 – 6799.
[46] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
2nd ed., Wiley-VCH, Weinheim, 1988.
[39] B. List, P. Pojarliev, C. Castello, Org. Lett. 2001, 3, 573 – 575.
[40] A. Erkkilꢁ, P. M. Pihko, Eur. J. Org. Chem. 2007, 4205 – 4216.
[41] D. A. Yalalov, S. B. Tsogoeva, T. E. Shubina, I. M. Martynova, T.
Clark, Angew. Chem. 2008, 120, 6726 – 6730; Angew. Chem. Int.
Ed. 2008, 47, 6624 – 6628.
[47] Y. Marcus, Chem. Soc. Rev. 1993, 22, 409 – 416.
[48] S. Protti, A. Mezzetti, J.-P. Cornard, C. Lapouge, M. Fagnoni,
Chem. Phys. Lett. 2008, 467, 88 – 93.
[49] A. Douhal, M. Sanz, L. Tormo, J. A. Organero, ChemPhysChem
2005, 6, 419 – 423.
[42] C. T. Wong, Tetrahedron 2009, 65, 7491 – 7497.
[43] K. N. Rankin, J. W. Gauld, R. J. Boyd, J. Phys. Chem. A 2002,
106, 5155 – 5159.
[50] V. I. Tomin, R. Javorski, Opt. Spectrosc. 2007, 103, 952 – 957.
[51] R. M. Beesley, C. K. Ingold, J. F. Thorpe, J. Chem. Soc. Trans.
1915, 107, 1080 – 1106.
[44] A. I. Nyberg, A. Usano, P. M. Pihko, Synlett 2004, 1891 – 1896.
[45] P. M. Pihko, K. M. Laurikainen, A. Usano, A. I. Nyberg, J. A.
Kaavi, Tetrahedron 2006, 62, 317 – 328.
[52] S. M. Bachrach, J. Org. Chem. 2008, 73, 2466 – 2468.
[53] T. F. Magnera, G. Caldwell, J. Sunner, S. Ikuta, P. Kebarle, J. Am.
Chem. Soc. 2002, 124, 6140 – 6146.
Angew. Chem. Int. Ed. 2010, 49, 4997 –5003
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 5003