The Proline Enamine Formation Pathway Revisited in Dimethyl Sulfoxide: Rate Constants Determined via NMR

Article abstract of DOI:10.1021/jacs.5b03420

Enamine catalysis is a fundamental activation mode in organocatalysis and can be successfully combined with other catalytic methods, e.g., photocatalysis. Recently, the elusive enamine intermediates were detected, and their stabilization modes were revealed. However, the formation pathway of this central organocatalytic intermediate is still a matter of dispute, and several mechanisms involving iminium and/or oxazolidinone are proposed. Here, the first experimentally determined rate constants and rates of enamine formation are presented using 1D selective exchange spectroscopy (EXSY) buildup curves and initial rate approximation. The trends of the enamine formation rates from exo-oxazolidinones and endo-oxazolidinones upon variation of the proline and water concentrations as well as the nucelophilic/basic properties of additives are investigated together with isomerization rates of the oxazolidinones. These first kinetic data of enamine formations in combination with theoretical calculations reveal the deprotonation of iminium intermediates as the dominant pathway in dimethyl sulfoxide (DMSO). The dominant enamine formation pathway varies according to the experimental conditions, e.g., the presence and strength of basic additives. The enamine formation is zero-order in proline and oxazolidinones, which excludes the direct deprotonation of oxazolidinones via E2 mechanism. The nucleophilicity of the additives influences only the isomerization rates of the oxazolidinones and not the enamine formation rates, which excludes a nucleophile-assisted anti elimination of oxazolidinones as a major enamine formation pathway.

Full text of DOI:10.1021/jacs.5b03420

Subscriber access provided by CMU Libraries - http://library.cmich.edu
Article
The Proline-Enamine Formation Pathway Revisited
in DMSO – Rate Constants Determined via NMR
Michael H. Haindl, Johnny Hioe, and Ruth M. Gschwind
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b03420 • Publication Date (Web): 19 Sep 2015
Downloaded from http://pubs.acs.org on September 22, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
The Proline-Enamine Formation Pathway Revisited in DMSO – Rate
Constants Determined via NMR
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Michael H. Haindl, Johnny Hioe, and Ruth M. Gschwind*.
Institut für Organische Chemie, Universität Regensburg, Dꢀ93053 Regensburg
KEYWORDS: Enamine, proline, reaction mechanism, NMR, EXSY.
ABSTRACT: Enamine catalysis is a fundamental activation mode in organocatalysis and can be successfully combined with other
catalytic methods e.g. photocatalysis. Recently, the elusive enamine intermediates were detected and their stabilization modes were
revealed. However, the formation pathway of this central organocatalytic intermediate is still a matter of dispute and several mechꢀ
anisms involving iminium and/or oxazolidinone are proposed. Here, the first experimentally determined rate constants and rates of
enamine formation are presented using 1D selective EXSY buildup curves and initial rate approximation. The trends of the enamine
formation rates from exoꢀoxazolidinones and endoꢀoxazolidinones upon variation of the proline and water concentration as well as
the nucelophilic/basic properties of additives are investigated together with isomerization rates of the oxazolidinones. These first
kinetic data of enamine formations in combination with theoretical calculations reveal the deprotonation of iminium intermediates
as the dominant pathway in DMSO. The dominant enamine formation pathway varies according to the experimental conditions, e.g.
the presence and strength of basic additives. The enamine formation is zero order in proline and oxazolidinones, which excludes the
direct deprotonation of oxazolidinones via E2 mechanism. The nucleophilicity of the additives influences only the isomerization
rates of the oxazolidinones but not the enamine formation rates, which excludes a nucleophile assisted anti elimination of oxazoliꢀ
dinones as a major enamine formation pathway.
barrier for Zꢀiminium.35 In a second pathway they proposed a
INTRODUCTION
water assisted proton transfer with an amphoteric water moleꢀ
cule protonating the carboxylate and deprotonating the αꢀ
proton (see Fig. 1 II). In this water assisted pathway the calcuꢀ
lated barriers of Zꢀ and Eꢀiminium are similar, but both are
considerably higher than in the water free deprotonation from
Zꢀiminium.35 In a third pathway they calculated the participaꢀ
tion of an external base in the deprotonation step (see Fig. 1
III). However, with external base the resulting energy barriers
are again significantly higher.35 In the following, all enamine
formation pathways via deprotonation from iminium carboxꢀ
ylates herein are referred to as “iminium pathways”.
Enamine catalysis is one of the central activation modes in
organocatalysis and has proven to be a very powerful method
for the enantioselective αꢀfunctionalization of carbonyl comꢀ
pounds (e.g. in aldolꢀ or mannichꢀreactions, michaelꢀadditions,
αꢀhalogenations, αꢀoxygenations, αꢀaminations or domino
reactions).1,2 The primary synthetic applications using proline
enamine catalysis were followed by versatile and powerful
further developments e.g. prolinol3–5 and prolinolether cataꢀ
lysts,6–9 di /trienamine10–12 catalysis and combinations with
ꢀ
photocatalysis.13–15 Proline oxazolidinones16 and stabilized
enamines17 were characterized by Xꢀray crystallography.
However for decades exclusively oxazolidinones were detectꢀ
ed as intermediates by in situ NMR.18–22 Recently, we obꢀ
served the elusive proline enamines by in situ NMR23 and
elucidated the stabilization modes of enamines versus oxazoliꢀ
dinones.23,24 Further, the mechanism of aldol addition versus
aldol condensation was investigated25 and the formation pathꢀ
ways, conformational preferences and stereoinduction modes
of prolinol and prolinol ether enamines were revealed.26,27 The
formation of imines or iminium ions are generally accepted to
proceed via carbinolamines.28–30 NMR studies on two combiꢀ
nations of aldehydes/ketones with prolinol/prolinolether as
catalysts corroborated this intermediate.31,32 However, the
formation pathway of the central proline enamine intermediate
is still discussed controversially.
In contrast, in the model of Seebach and Eschenmoser the
enamine intermediate is formed directly from the oxazoliꢀ
dinone species.20 An external base such as an additional proꢀ
line, oxazolidinone or product molecule is proposed to deproꢀ
tonate the oxazolidinone αꢀproton and to induce an anti E2
elimination (see Fig. 1 IV). Our own NMR investigations of
proline enamine intermediates seemed to corroborate a direct
formation of enamines from oxazolidinones providing only
EXSY cross peaks from oxazolidinones to enamines and none
from the aldehydes.23 In addition, initial studies hinted at a
nucleophile assisted enamine formation from oxazolidinones
(see Fig. 1 V).31 A nucleophilic ring opening would allow for a
water assisted anti elimination from oxazolidinones and avoid
the missing microreversibility of the Seebach Eschenmoser
pathway. Despite the fact that oxazolidinones are also generatꢀ
ed via iminium ions, all enamine formation pathways via
direct deprotonation from oxazolidinones are referred to hereꢀ
in as “oxazolidinone pathways”.
In the most generally accepted mechanism first proposed by
Houk and List33,34 the enamine is formed directly from the E/Z
zwitterionic iminium via intramolecular deprotonation of one
of its αꢀprotons by the carboxylate moiety (see Fig. 1 pathway
I). For this process Sunoj et al. calculated the lowest energy
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
three intermediates are ideal for the investigation of chemical
exchange between these three species.36
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Catalytic cycle of Lꢀproline catalyzed aldol reactions
and proposed enamine formation pathways. Enamine formations
via deprotonation of iminium carboxylates are shown in blue
(intramolecular I, water assisted II, external base induced III).
Deprotonations of oxazolidinones are shown in red (with external
base IV, nucleophile assisted V). For the sake of clarity the E/Zꢀ
iminium isomers, the diastereomeric endo/exoꢀoxazolidinones and
the sꢀcis/sꢀtrans enamine conformers are presented in condensed
forms using waved bonds.
To differentiate between the discussed enamine formation
pathways in this work, rates and rate constants of enamine
formation were determined experimentally using 1D selective
EXSY buildup curves and the initial rate approximation. The
presented large number of enamine formation rate constants
starting from both endo/exoꢀoxazolidinones, the influence of
various additives (Lꢀproline, water, bases and nucleophiles) on
these rate constants, the resulting trends and the comparison
with oxazolidinone isomerization rate constants allow for the
first time a detailed experimental insight into the so far hidden
mechanistic relation between enamines and oxazolidinones.
These extensive experimental data in combination with theoꢀ
retical calculations indicate that the enamine is neither formed
via E2 elimination nor via a nucleophile assisted pathway
from the oxazolidinone but rather via deprotonation of iminiꢀ
um ions.
Figure 2. A) LꢀProlineꢀcatalyzed selfꢀaldolization of
3ꢀmethylbutanal; B) section of the 1Hꢀspectrum showing well
separated H1 signals (circles) of the enamine and the two oxazoliꢀ
dinones (sample: Lꢀproline (saturated), 3ꢀmethylbutanal (50 mM),
DABCO (50 mM) in DMSOꢀd6 at 300 K); C) stack plot of 1D
selective EXSY spectra of this sample (irradiation on H1 of exoꢀ
oxazolidinone, mixing times 3ꢀ700 ms); D) 1D selective EXSY
buildꢀup curves (sample without additives); blue (P = endoꢀ
oxazolidinone) and black (P = enamine) lines represent the initial
slopes (Iτm(P)/τm) used for the initial rate approximation.
In principle, for such a slow exchange on the NMR time scale
magnetization transfer,37–44 usually called EXSY (exchange
spectroscopy), can be used to determine individual rate conꢀ
stants. In two dimensional EXSY experiments as previously
applied in our enamine studies,23 the complete exchange maꢀ
trix can be observed qualitatively within one spectrum. Howꢀ
ever, the quantitative interpretation of EXSY cross peak ratios
used in these studies to differentiate between the reaction
pathways requires an iminium isomerisation being fast comꢀ
pared to all other processes, which is not the case. In addition,
multistep transfers similar to spin diffusion in NOESY spectra
hampered a reliable quantitative interpretation (for details see
RESULTS AND DISCUSSION
Model system and methods. As model system for this
mechanistic study of the enamine formation pathway the
Lꢀproline catalyzed intermolecular aldol reaction of
3ꢀmethylbutanal in DMSOꢀd6 with 100 % Lꢀproline was choꢀ
sen (see Fig. 2a). In previous studies we observed for this
system the highest enamine concentration in combination with
1
a slow subsequent aldolisation rate. At 300 K the resulting H
spectra show well separated signals of the H1ꢀprotons of the
enamine as well as of the exoꢀ and endoꢀoxazolidinone (see
Fig. 2b). In addition, the quite similar signal intensities of all
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
SI page S2). Both problems are circumvented in the present
the formation pathway of enamines cannot be neglected, beꢀ
1
2
3
4
5
6
7
8
study. The determination of direct reaction rate constants in
combination with theoretical calculations avoids the mechanisꢀ
tic prerequisites for the application of the EXSY cross peak
ratios. The problem of mixed rates due to multistep transfers
can be solved by using the initial rate approximation, similar
to the NOE.
cause in all samples at least one equivalent of water is present
originating from the condensation reaction of aldehyde and
proline to oxazolidinones and enamines. In the following the
reaction rate constants from exoꢀoxazolidinone and en-
doꢀoxazolidinone to the enamine are discussed (for data reꢀ
garding the back reaction from enamine to oxazolidinones see
SI page S10). First, we investigated the influence of the
amount of Lꢀproline and water on the enamine formation rates.
In addition, the effect of various additives with varying basic
and nucleophilic properties was tested (DABCO, TEA, sodiꢀ
um carbonate and sodium benzenesulfinate). Acidic additives
cannot be applied in this experimental setup, because proline
enamines are not detectable together with hydrogen bond
donors (e.g. in MeOH no enamine 1H signal can be
detected).23,24
Within the initial linear buildup of the exchange signal the
slope of this buildup (Iτm(P)/τm) (see. Fig. 2d) normalized by
the integral of the starting material (I0(A)) is directly correlatꢀ
ed to the formation rate of the product (r) divided by the equiꢀ
librium concentration of the starting material ([A], see Fig.
3).44,45 In case the rate determining step is an unimolecular
reaction the normalized formation rate is equivalent to the rate
constant (k). For bimolecular reactions the normalized forꢀ
mation rate is the rate constant multiplied with the equilibrium
concentration of the second reaction partner (k[B], see Fig. 3).
In the following, for the sake of readability all “normalized
rates” are denoted as “rates”.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Participation of proline, oxazolidinones and water. First,
the dependence of the enamine formation rate on the amount
of Lꢀproline and water was studied, because both are intrinsic
components of the reaction. Since the oxazolidinone pathway
according to Seebachs proposal (see Fig. 1 IV) includes the
participation of a second proline, oxazolidinone or product as
base (E2ꢀelimination), for this pathway it is expected that the
enamine formation rate increases with the proline concentraꢀ
tion (see Fig. 3). In Fig. 4a the experimentally determined
rates of enamine formation from both endoꢀoxazolidinone and
exoꢀoxazolidinone are presented. At any Lꢀproline or water
concentration investigated in the additive free case, the rate
starting from endoꢀoxazolidinone is smaller than that from
exoꢀoxazolidinone.
Figure 3. The normalized formation rate from EXSY spectra
using the initial rate approximation corresponds to rate constants
for unimolecular reactions and rate constants times concentration
of the reaction partner B in case of bimolecular reactions. r: rate;
k: rate constant; τm: mixing time; Iτm(P): EXSY integral of the
product signal (e.g. enamine) at τm; A: starting material (irradiated
species); P: product (buildup species); I0(A): EXSY integral of the
starting material signal (endoꢀ or exoꢀoxazolidinone) at τm =0 s;
[A], [B]: equilibrium concentrations of A and B.
2D EXSY buildup curves are extremely time consuming and
in reacting systems not applicable to one sample. Therefore, in
this mechanistic study 1D selective EXSY buildup curves
were measured, irradiating selectively the well separated H1
protons of endoꢀ/exoꢀoxazolidinone or enamine. Using very
short mixing times, exclusively the irradiated signal is obꢀ
served (see Fig. 2c). After a considerable mixing time the
signals of exchanging molecules appear. With the enamine
model system described above considerable amounts of both
intermediates oxazolidinones and enamines can be detected.
To the best of our knowledge this allows for the first time to
measure EXSYꢀbuildup curves between reaction intermediꢀ
ates. Thus, as an unique feature, this study provides rate conꢀ
stants and reaction orders of the rate limiting steps between
intermediate species and not for the whole reaction pathway.
In case several reaction steps are involved between the obꢀ
served intermediates, the rate limiting step has to be at least
one order of magnitude slower than all other steps, otherwise
mixed rates are observed.
Previous studies showed that the presence or addition of exꢀ
tra water does not affect the relative ratios of enamine to oxaꢀ
zolidinones
but
reduce
their
absolute
amounts
considerably.23,46,47 Therefore, dry solvents and starting mateꢀ
rials were used for all experiments to obtain optimal signal
intensities. Nevertheless a potential participation of water in
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
formation pathway (zero order in oxazolidinone). Thus, Seeꢀ
1
bach’s E2 elimination pathway from oxazolidinone (pathway
IV) with either proline or oxazolidinone as external base is not
supported by our data. Another possibility would be an aldol
product or oxazolidinone product as external base in pathway
IV. In this case an induction period of enamine and product
formation would be expected from an autocatalytic process.
However, in none of our enamine aldol studies23,25 such an
induction period was observed. An adduct of Lꢀproline and
oxazolidinone or of two oxazolidinones as reaction intermediꢀ
ates and a subsequent E1 elimination cannot be excluded by
these experimental data because the rate limiting step would
be again zero order in proline. However, theoretical calculaꢀ
tions of potential adduct complexes and subsequent E1 elimiꢀ
nations suggest that such processes are extremely unfavouraꢀ
ble under synthetic conditions (see SI page S23). Thus, the
experiments exclude all enamine formation pathways that are
first order in proline or oxazolidinones under synthetic condiꢀ
tions in DMSO and acetonitrile. Theoretical calculations sugꢀ
gest that pathways starting from proline/oxazolidinone adꢀ
ducts, which would be zero order in proline or oxazolidinone,
are highly unlikely.
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Next the influence of increasing amounts of water was invesꢀ
tigated. The addition of external water does not significantly
affect
the
rates
of
enamine
formation
from
exo/endoꢀoxazolidinone (see Fig. 4a), indicating again zero
order in water for this step.
In addition, we determined the enamine formation and the
oxazolidinone isomerization rate constants for the linear aldeꢀ
hyde 3ꢀphenylpropanal in the additive free case in DMSO
(approx. 0.5 eq. water). The results are very similar and thereꢀ
fore the influence of the aldehyde structure is assumed to be
minor (for more details also concerning additional results in
acetonitrile please refer to the SI on pages S13 and S3).
Theoretical calculation of the enamine formation. In orꢀ
der to translate these experimental data into mechanistic
enamine formation pathways, various theoretical calculations
of potential enamine formation pathways were performed and
compared to the experimentally determined rate constants. The
low stability of iminium and enamine has been reported by
Houk and Blackmond as the major problem in predicting the
Figure 4. A) Enamine formation rate constants from en-
do/exoꢀoxazolidinone dependent on Lꢀproline and external water
amount (sample: 50 mM 3ꢀmethylbutanal in DMSOꢀd6 at 300 K);
experimental
reactant/product
distribution
(oxazoliꢀ
dinone/enamine).48 In this case the inclusion of an explicit
solvent molecule in the model is becoming compulsory to
cover the missing interaction between solvent and solute. This
is also confirmed by our present study. At least one solvent
molecule is needed to predict the enamine/oxazolidinoneꢀratio
and, as shown later, the barrier heights correctly (for the data
regarding the discussion of solvent molecule and the simulaꢀ
tion without explicit solvent refer to the SI on page S19).
Therefore, calculations with a cluster continuum model49–51
were performed.
B) Summary of calculated free energy barriers ꢁG‡ kJ [molꢀ1]
298
(ring opening and proton transfer) and thermodynamics stability
(in orange; referenced to exoꢀoxazolidinone) using a cluster conꢀ
tinuum model at CCSD(T)/CBS level of theory in DMSO. For the
CBS extrapolation procedure, please refer to SI page S24. The
experimental values are shown in brackets.
Upon increasing Lꢀproline concentration the enamine forꢀ
mation is not accelerated but remains constant within the
experimental error range with a slightly decreasing trend.
Similar results were obtained in acetonitrile (see SI page S3).
This indicates that the rate determining step of the enamine
formation is zero order in Lꢀproline (unimolecular; r/[A]=k).
Our current and previous23 NMR studies showed that at higher
Lꢀproline concentrations also the amount of the two oxazoliꢀ
dinones increases directly proportional (see SI page S12).
Therefore, the rather constant enamine formation rates exclude
also the involvement of a second oxazolidinone in the enamine
The contribution of additional Lꢀproline or oxazolidinone in
bimolecular mechanisms was excluded experimentally and
therefore omitted in the theoretical calculations. Our previous
EXSY studies suggested a nucleophile assisted enamine forꢀ
mation from oxazolidinone (Fig. 1: Path V).31 Therefore, we
tried to simulate the ring opening of oxazolidinone by
DABCO as nucleophile/base together with a simultaneous αꢀ
proton transfer to the carboxylate group and a subsequent
dissociation of the nucleophile. Furthermore, we also tried to
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
The experimental conditions being closest to those of our
abstract the αꢀproton by DABCO and to locate an E2 eliminaꢀ
1
2
3
4
5
6
7
8
theoretical calculations (one Lꢀproline, one aldehyde and one
water molecule) are those at saturation with Lꢀproline and
without additional water. For this sample an activation barrier
of +79.4 kJ molꢀ1 is experimentally determined starting from
Zꢀiminium and of +80.9 kJ molꢀ1 starting from Eꢀiminium.
This is qualitatively and quantitatively in good agreement with
our theoretical data for the internal deprotonation from Zꢀ
iminium as well as the water assisted deprotonation from both
E- and Zꢀiminium (see numbers in Fig. 4b). Considering the
absence and the presence of water in the three pathways of the
lowest transition states, at first glance one would expect that
increasing amounts of water would affect these pathways
differently. However, the experimental data show that the rate
determining steps of all pathways are zero order in water (see
Fig. 4a and discussion above). That is directly obvious for the
water free pathway starting from Zꢀiminium. For the water
assisted pathways starting from Eꢀ and Zꢀiminium the theoretiꢀ
cal data suggest the formation of an intermediate consisting of
iminium and water, followed by a deprotonation step which is
zero order in water, because the reference iminium intermediꢀ
ate already includes water. Thus, from the experimental data
the assistance of water cannot be directly deduced but the
measured rate constants are in agreement with the most preꢀ
ferred pathways from theoretical calculations.
tion transition state as proposed by Seebach (Fig. 1: Path IV).
However, all attempts to locate any transition state lead to
either iminiumꢀnucleophile adducts, or oxazolidinones. The
only transition state found stems from the addition of a nucleꢀ
ophile to the iminium ion after the ring opening of the oxazolꢀ
idinone facilitating the E/Zꢀiminium isomerization. As shown
later, this is corroborated by the experimental data, which
show that nucleophiles accelerate the exoꢀ/endoꢀoxazolidinone
exchange rate but not the enamine formation rates (see below
and SI on page S5). Therefore, the Houk List pathway was
recalculated on a refined level of theory (see Fig. 4b) and
compared to the experimental rate constants (see Fig. 4a).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Our calculations showed that in the Houk List pathway the
endoꢀoxazolidinone is exclusively connected with the Eꢀ
iminium and the exoꢀoxazolidinone with the Zꢀiminium (Figꢀ
ure 4b), which is consistent with previous calculations of
Sharma and Sunoj.35 Without water the activation barrier of
the subsequent internal deprotonation step from Zꢀiminium to
enamine is calculated to be +74.4 kJ molꢀ1 compared to
+114 kJ molꢀ1 when starting from Eꢀiminium. This higher
activation barrier of approx. 40 kJ molꢀ1 from Eꢀiminium is
rather due to the unfavourable geometry of the transition state.
In case of a water assisted deprotonation step the situation
changes. Now the activation barrier starting from an Eꢀ
iminium water complex is considerably lower (+83.5 kJ molꢀ1)
than in the water free pathway. In the case of deprotonation
from Zꢀiminium via water, the barrier is only marginally highꢀ
er (+77.2 kJ molꢀ1) than in the internal deprotonation. Qualitaꢀ
tively these two preferred pathways are in agreement with the
previous calculations of Sunoj and Sharma using a smaller
model in the gasꢀphase, and a lower level of theory (ꢀG298
B3LYP/6ꢀ31G**).35 As expected, the change of phase (gasꢀ
phase to condensed phase) and theoretical model in our preꢀ
sent work led to three significant differences. First, the ring
opening barriers in the study of Sunoj and Sharma are rather
high (+53.6 kJ molꢀ1 (endo), +73.6 kJ molꢀ1 (exo)) compared to
ours (+38.7 kJ molꢀ1 (endo), +47.4 kJ molꢀ1 (exo)), most probꢀ
ably due to the fact that they did not include solvent effects in
their calculation (see SI page S20 for further discussion).
Second, the energy barrier for the water assisted deprotonation
from Eꢀ and Zꢀiminium are very similar in the gasꢀphase at
DFT level of theory (differing by 2.5 kJ molꢀ1) and quite low
(+53.6 kJ molꢀ1 for Eꢀiminium and +51.1 kJ molꢀ1 for Zꢀ
iminium). In our calculation the difference amounts to 6.3 kJ
molꢀ1 and the barriers are significantly higher (Fig. 4b). Again,
this is an artefact from the gasꢀphase calculation. As stated in
their SI, the solvent correction in acetonitrile leads to an inꢀ
crease of barrier height by ~40 kJ molꢀ1 for both E/Zꢀiminiumꢀ
enamine transition states, which would approach our predicted
values. Third, although it does not have a significant impact,
we also notice here that the energy barrier difference between
Zꢀiminium and Eꢀiminium for the water free pathway is slightꢀ
ly higher (+54.0 kJ molꢀ1) in their calculation than in our preꢀ
sent data (+39.6 kJ molꢀ1).
Figure 5. Summary of enamine formation without additives for
R=C3H7, phenyl in DMSO. The experimentally measured interꢀ
mediates and rate constants are highlighted in green, the calculatꢀ
ed activation barriers are highlighted in yellow.
A summary of the active enamine formation pathways without
additives is shown in Fig. 5. Without additives the dominant
process is the proton transfer from Zꢀiminium, which is conꢀ
nected with the exo-oxazolidinone. The water assisted deproꢀ
tonation of Eꢀiminium is also observed but at a slower rate.
Influence of basic additives. Next the influence of basic
additives (DABCO, TEA, sodium carbonate and sodium benꢀ
zenesulfinate) was measured to elucidate a potential correlaꢀ
tion between enamine formation rates and basicity. However,
not all pKaH values of the additives are known in DMSO.
Therefore, an internal basicity scale was created. Our previous
enamine study with basic additives showed that the amount of
enamine increases at the expense of oxazolidinone with inꢀ
creasing basicity of the additive.24 Therefore, the ratio of
enamine to total intermediate concentration was used as a
Next, the experimentally determined rate constants were conꢀ
verted into activation barriers (for details see SI page S24) and
compared to our theoretical calculation. The activation barriers
for the ring opening processes are at least 20 kJ molꢀ1 lower
than those of the deprotonation processes from iminium ions
to enamines. This translates to several orders of magnitudes in
rate constants. Hence, the experimental rate constants correꢀ
spond exclusively to the barrier of the deprotonation process.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 9
measure for the internal basicity in DMSO. The results DMSO). For the tertiary amine bases DABCO and TEA only
1
2
3
4
5
6
7
8
showed an increasing basicity from sodium benzenesulfinate,
over TEA and DABCO to sodium carbonate (for data and
details see SI page S4).
N values in acetonitrile are available, which are considerably
lower (18.8 and 17.1, respectively). The experimental enamine
formation rates from both oxazolidinones are significantly
lower for benzenesulfinate than for DABCO (see Fig. 6b),
which supports the discussion above, that not the nucleophilicꢀ
ity but the basicity of the additive is crucial for the enamine
formation rate constants. These results exclude pathway V, i.e.
a nucleophile assisted antiꢀelimination of oxazolidinones
leading to enamines as a major formation pathway in DMSO.
In contrast to the measurements with varying water and proꢀ
line concentrations the rate of enamine formation is depending
on the concentration of the additives. This indicates that the
rate limiting deprotonation step is bimolecular (participation
of starting molecule and base) (see Fig. 3 bimolecular and SI
page S16 for details). In Fig. 6a the rate constants for the addiꢀ
tive free sample and those with TEA are given. Upon addition
of TEA the rate constant of enamine formation raises drastiꢀ
cally and the relative order of enamine formation rate conꢀ
stants starting from endoꢀ and exoꢀoxazolidinones are invertꢀ
ed. In Fig. 6b the normalized rates for all additives (50 mM) as
well as their correlation with the internal basicity scale are
presented. All additives show high enamine formation rates
starting from endoꢀoxazolidinone. In contrast the rates starting
from exoꢀoxazolidinone are much lower. Overall, both trends
follow the relative internal basicity of the additives, but with
different slopes. One exception is sodium carbonate which
shows a considerably lower enamine formation rate from
endoꢀoxazolidinone than DABCO52 .Another striking change
is the general switch of the relative rate constants and rates
starting from endoꢀoxazolidinones and exoꢀoxazolidinones,
respectively. Without basic/nucleophilic additives the rate
constants from endo-oxazolidinone are smaller than those
from exoꢀoxazolidinone (see Figure 4a and Fig. 6a first entry).
With basic/nucleophilic additives the situation is inverted, in
all cases the rate constants (see Fig. 6a second entry) and rates
(see Fig. 6b) from endoꢀoxazolidinone are significantly larger
than those from exoꢀoxazolidinone. These experimental data
fit to the theoretical calculations shown in Figure 4b with
additional pathways of Eꢀ and Zꢀiminium deprotonation by
external bases (see Figure 6c). The rapid increase of the rate
constant and rates from endoꢀoxazolidinone indicates that
external bases can easily deprotonate the Eꢀiminium. For
exoꢀoxazolidinone the slope with increasing basicity is signifiꢀ
cantly lower, which can be explained by steric hindrance and
electrostatic repulsion between the base and the carboxylate
moiety in the case of the Zꢀiminium.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
External bases potentially could also deprotonate the oxaꢀ
zolidinones directly as previously proposed by Seebach et
al..20 Endoꢀoxazolidinones are well known to be sterically
more congested than the corresponding exoꢀoxazolidinones.
However, for the proposed E2 antiꢀelimination not the internal
steric but the steric hindrance of one of the two αꢀprotons in
anti-conformation to the oxazolidinone CO bond has to be
considered. Nevertheless, the structures of endo- and exoꢀ
oxazolidinones do not exhibit significant differences in the
steric of these αꢀprotons (see SI page S12). Thus, steric arguꢀ
ments of the oxazolidinones cannot explain the very different
behaviour of enamine formation rates from endo- and exoꢀ
oxazolidinones with increasing basicity. This excludes the
direct deprotonation of oxazolidinones (pathway IV) as a
major pathway in the presence of bases.
Figure 6. a) The enamine formation rate constants drastically
increase with the addition of the basic additive TEA, while the
relative rate constants from endoꢀ and exoꢀoxazolidinone invert.
b) The enamine formation rates increase with the basicity of the
additives. Additives invert the relative rates from endoꢀ and exoꢀ
oxazolidinones and accelerate especially those from endoꢀ
oxazolidinone (sample: Lꢀproline (saturated), 3ꢀmethylbutanal (50
mM), additive (50 mM, for sodium carbonate: saturated solution)
in DMSOꢀd6 at 300 K). c) Summary of enamine formation pathꢀ
ways in the presence of additives. Basic additives accelerate
strongly the deprotonation from Eꢀiminium, but only slightly that
from Zꢀiminium. Nucleophiles accelerate the E/Zꢀiminium isomꢀ
erization.
Influence of nucleophilic additives. Next, a correlation of the
enamine formation rates towards Mayr’s nucleophilicity
scale53 was considered. Benzenesulfinate shows hardly any
basicity but is a very strong Sꢀnucleophile (N value of 19.6 in
However, another effect of additional nucleophiles was obꢀ
served experimentally, the accelerated exchange between exoꢀ
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
Supporting Information. Details about the experimental setup,
oxazolidinones and endoꢀoxazolidinones (for details see SI
page S5).
1
2
3
the theoretical calculations and the calculation of rate constants as
well as additional NMR and theoretical data. This material is
available free of charge via the Internet at http://pubs.acs.org.
CONCLUSION
4
5
6
7
8
9
In summary, detailed experimental enamine formation rate
constants and rates from both endoꢀoxazolidinone and exoꢀ
oxazolidinone are presented. Their dependence on the addition
of Lꢀproline, water as well as additives with various basic and
nucleophilic properties was investigated. The mechanistic
interpretation of these data is confirmed by higher level theoꢀ
retical calculations of the enamine formation pathway. First of
all the enamine formation is zero order in proline and oxazoliꢀ
dinones, which excludes the Seebach pathway V proposing an
E2 deprotonation of oxazolidinones. Prereacting oxazoliꢀ
dinoneꢀproline complexes undergoing E1 elimination cannot
be excluded experimentally but are highly unlikely according
to theoretical calculations. Without additives the fastest proꢀ
cess is the proton transfer from Zꢀiminium (pathways I + II
from exoꢀoxazolidinone), while the deprotonation of Eꢀ
iminium follows the water assisted pathway II (from endoꢀ
oxazolidinone) at a slower rate. Basic additives change the
situation considerably. Now the deprotonation via an external
base is preferred (pathway III). The acceleration of the
enamine formation is significantly more pronounced for Eꢀ
iminium (from endoꢀoxazolidinone) than for Zꢀiminium (from
exoꢀoxazolidinone) and correlates to the basicity of the addiꢀ
tive. These trends are in agreement with the steric and elecꢀ
tronic properties of the iminium intermediates (pathways I, II,
III) but not with those of the oxazolidinones (exclusion of
pathway IV with basic additives). The nucleophilicity of the
additives influences only the isomerization rates of the two
oxazolidinones but not the enamine formation rates. This
excludes a nucleophile assisted antiꢀelimination of oxazoliꢀ
dinones as a major enamine formation pathway (pathway V).
Thus, the first kinetic data of enamine formation together with
theoretical calculations reveal that enamines are formed most
likely via deprotonation of iminium intermediates (HoukꢀList
pathway) in DMSO. The dominant pathway varies according
to the experimental conditions, e.g. the presence and strength
of a basic additive.
AUTHOR INFORMATION
Corresponding Author
ruth.gschwind@ur.de
Author Contributions
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the manuꢀ
script.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACKNOWLEDGMENT
We gratefully acknowledge financial support from the DFG, grant
GS 13/4ꢀ1. We thank Dr. Markus Schmid for the initial NMR
investigations leading to the proposal of the nucleophile assisted
deprotonation pathway from oxazolidinones. We also thank the
Leibniz Rechenzentrum for providing computational facility.
ABBREVIATIONS
DABCO, 1,4ꢀdiazabicyclo[2.2.2]octane; TEA, triethylamine,
CBS, complete basis set; CCSD(T), coupled cluster singles douꢀ
bles with triples from perturbation theory; DMSO, dimethylꢀ
sulfoxide; NMR, nuclear magnetic resonance; EXSY, exchange
spectroscopy; NOE, nuclear overhauser effect.
REFERENCES
(1) Pihko, P. M.; Majander, I.; Erkkilä, A. Topics in Cur-
rent Chemistry, 2010, 291, 29.
(2) List, B. Synlett 2001, 2001, 1675.
(3) Zhong, G.; Fan, J.; Barbas, C. F. Tetrahedron Lett.
2004, 45, 5681.
(4) Hayashi, Y.; Itoh, T.; Aratake, S.; Ishikawa, H. An-
gew. Chem. Int. Ed. 2008, 47, 2082.
(5) Enders, D.; Chow, S. European J. Org. Chem. 2006,
2006, 4578.
(6) Hayashi, Y.; Gotoh, H.; Hayashi, T.; Shoji, M. An-
gew. Chem. Int. Ed. 2005, 44, 4212.
(7) Chi, Y.; Gellman, S. H. Org. Lett. 2005, 7, 4253.
(8) Marigo, M.; Wabnitz, T. C.; Fielenbach, D.;
Jørgensen, K. A. Angew. Chem. Int. Ed. 2005, 44, 794.
(9) Palomo, C.; Mielgo, A. Angew. Chem. Int. Ed. 2006,
45, 7876.
(10) Ramachary, D. B.; Reddy, Y. V. European J. Org.
Chem. 2012, 2012, 865.
(11) Kumar, I.; Ramaraju, P.; Mir, N. A. Org. Biomol.
Chem. 2013, 11, 709.
(12) Dieckmann, A.; Breugst, M.; Houk, K. N. J. Am.
Chem. Soc. 2013, 135, 3237.
(13) Nicewicz, D. A.; MacMillan, D. W. C. Science
2008, 322, 77.
(14) Neumann, M.; Füldner, S.; König, B.; Zeitler, K.
Angew. Chem. Int. Ed. 2011, 50, 951.
(15) Pirnot, M. T.; Rankic, D. A.; Martin, D. B. C.;
MacMillan, D. W. C. Science 2013, 339, 1593.
(16) Seebach, D.; Schweizer, W. B. J. Am. Chem. Soc.
1983, 105, 5390.
COMPUTATIONAL DETAILS
The geometry of all systems was optimized in the gas phase at
TPSSꢀD3/augꢀSVP level of theory (Ahlrichs SVP54,55 augꢀ
mented with diffuse function from augꢀccꢀVDZ56,57) corrected
with empirical dispersion from Grimme.58,59 Harmonic vibraꢀ
tional frequency and thermochemical correction were perꢀ
formed at the geometry optimization level. Single point calcuꢀ
lations were done at DLPNOꢀCCSD(T)/CBS60,61 using density
fitting (SplitꢀRIꢀJ) and semi numerical approximation for the
exchange term (RIJCOSX) for the reference wavefunction62–64
and domain based local pair natural orbital (DLPNO) approxꢀ
imation for the postꢀHF part. Two points extrapolation techꢀ
nique to approach CBS was used using def2ꢀQZVPP and def2ꢀ
TZVPP basis sets as implemented in ORCA.65,66 The solvent
correction ꢁGsolv was calculated at TPSSꢀD3/augꢀTZVP (Ahlꢀ
richs TZVP54,55 augmented with diffuse function from augꢀccꢀ
VDZ56,57) using COSMO model in DMSO and subsequently
added to the single point energy.67,68 The software used was
ORCAꢀ3.0.3. The extrapolation procedure, energies, and
structures are provided in the supplemental information.
(17) Bock, D. A.; Lehmann, C. W.; List, B. Proc. Natl.
Acad. Sci. U. S. A. 2010, 107, 20636.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
(18) List, B.; Hoang, L.; Martin, H. J. Proc. Natl. Acad.
Sci. U. S. A. 2004, 101, 5839.
(46) Nyberg, A. I.; Usano, A.; Pihko, P. M. Synlett 2004,
11, 1891.
1
2
3
4
5
6
7
8
(19) Iwamura, H.; Wells, D. H.; Mathew, S. P.;
Klussmann, M.; Armstrong, A.; Blackmond, D. G. J. Am.
Chem. Soc. 2004, 126, 16312.
(20) 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.
(47) Pihko, P. M.; Laurikainen, K. M.; Usano, A.;
Nyberg, A. I.; Kaavi, J. A. Tetrahedron 2006, 62, 317.
(48) Hein, J. E.; Burés, J.; Lam, Y.; Hughes, M.; Houk,
K. N.; Armstrong, A.; Blackmond, D. G. Org. Lett. 2011,
13, 5644.
(49) Bryantsev, V. S.; Diallo, M. S.; Iii, W. a G. Water
2008, 9709.
9
(21) Isart, C.; Burés, J.; Vilarrasa, J. Tetrahedron Lett.
2008, 49, 5414.
(50) Pliego, J. R.; Riveros, J. M. J. Phys. Chem. A 2001,
105, 7241.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(22) F. Orsini, F. Pelizzoni, M. Forte, M. Sisti, G.
Bombieri, F. B. J. Heterocycl. Chem. 1989, 26, 837.
(23) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. Angew.
Chemie Int. Ed. 2010, 49, 4997.
(24) Schmid, M. B.; Zeitler, K.; Gschwind, R. M.
Chemistry 2012, 18, 3362.
(25) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. J. Org.
Chem. 2011, 76, 3005.
(26) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. J. Am.
Chem. Soc. 2011, 133, 7065.
(27) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. Chem.
Sci. 2011, 2, 1793.
(28) Bruice, P. Y.; Bruice, T. C. J. Am. Chem. Soc. 1978,
100, 4793.
(29) Bruice, P. Y. J. Am. Chem. Soc. 1989, 111, 962.
(30) Bruice, P. Y. J. Am. Chem. Soc. 1990, 112, 7361.
(31) Schmid, M. B. Ph. D. Thesis, University of Regensꢀ
burg, RegensburgꢀGermany, January 2011.
(32) Seegerer, A. Master Thesis, University of Regensꢀ
burg, RegensburgꢀGermany, September 2014.
(33) Bahmanyar, S.; Houk, K. N.; Martin, H. J.; List, B.
J. Am. Chem. Soc. 2003, 125, 2475.
(34) List, B.; Lerner, R. A.; Barbas, C. F. J. Am. Chem.
Soc. 2000, 122, 2395.
(35) Sharma, A. K.; Sunoj, R. B. Angew. Chemie Int. Ed.
2010, 49, 6373.
(36) So far no iminium can be detected in our system.
Therefore only oxazolinones and enamine can be directly
investigated.
(37) Orrell, K. G.; Šik, V.; Stephenson, D. Prog. Nucl.
Magn. Reson. Spectrosc. 1990, 22, 141.
(38) Campbell, I. D.; Dobson, C. M.; Ratcliffe, R. G.;
Williams, R. J. P. J. Magn. Reson. 1978, 29, 397.
(39) Forsén, S.; Hoffman, R. A. J. Chem. Phys. 1963, 39,
2892.
(40) Perrin, C. L.; Johnston, E. R. J. Magn. Reson. 1979,
33, 619.
(41) Günther, H. NMR Spectroscopy, 3rd ed.; Wileyꢀ
VCH: Weinheim, 2013.
(42) Bain, A. D. Prog. Nucl. Magn. Reson. Spectrosc.
2003, 43, 63.
(43) Aski, S. N.; Takacs, Z.; Kowalewski, J. Magn.
Reson. Chem. 2008, 46, 1135.
(51) Thar, J.; Zahn, S.; Kirchner, B. J. Phys. Chem. B
2008, 112, 1456.
(52) Most probably this is due to the bad solubility of soꢀ
dium carbonate in DMSO. The amount of sodium carꢀ
bonate dissolved is much lower than in the other additive
samples. Therefore, the enamine formation rates are not as
high as expected from its internal basicity.
(53) Baidya, M.; Kobayashi, S.; Mayr, H. J. Am. Chem.
Soc. 2010, 132, 4796.
(54) Schäfer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys.
1994, 100, 5829.
(55) Schäfer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys.
1992, 97, 2571.
(56) Kendall, R. A.; Jr., Dunning, T. H.; Harrison, R. J.
Chem. Phys. 1992, 96, 6796.
(57) Woon, D. E.; Dunning, T. H. J. Chem. Phys. 1993,
98, 1358.
(58) Tao, J.; Perdew, J. P.; Staroverov, V. N.; Scuseria,
G. E. Phys. Rev. Lett. 2003, 91, 146401.
(59) Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput.
Chem. 2011, 32, 1456.
(60) Riplinger, C.; Sandhoefer, B.; Hansen, A.; Neese, F.
J. Chem. Phys. 2013, 139.
(61) Riplinger, C.; Neese, F. J. Chem. Phys. 2013, 138.
(62) Neese, F.; Wennmohs, F.; Hansen, A.; Becker, U.
Chem. Phys. 2009, 356, 98.
(63) Neese, F. J. Comput. Chem. 2003, 24, 1740.
(64) Izsák, R.; Neese, F. J. Chem. Phys. 2011, 135.
(65) Weigend, F.; Furche, F.; Ahlrichs, R. J. Chem. Phys.
2003, 119, 12753.
(66) Neese, F. Wiley Interdiscip. Rev. Comput. Mol. Sci.
2012, 2, 73.
(67) Klamt, A.; Schüürmann, G. COSMO: a new apꢀ
proach to dielectric screening in solvents with explicit
expressions for the screening energy and its gradient. Jour-
nal of the Chemical Society, Perkin Transactions 2, 1993,
799.
(68) Klamt, A. The COSMO and COSMOꢀRS solvation
models. Wiley Interdisciplinary Reviews: Computational
Molecular Science, 2011, 1, 699.
(44) Perrin, C. L.; Dwyer, T. J. Chem. Rev. 1990, 90,
935.
(45) Kumar, A.; Wagner, G.; Ernst, R. R.; Wuethrich, K.
J. Am. Chem. Soc. 1981, 103, 3654.
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table of Contents
9
ACS Paragon Plus Environment