On the mechanism of bifunctional squaramide-catalyzed organocatalytic michael addition: A protonated catalyst as an oxyanion hole

Article abstract of DOI:10.1002/chem.201304553

A joint experimental-theoretical study of a bifunctional squaramide-amine-catalyzed Michael addition reaction between 1,3-dioxo nucleophiles and nitrostyrene has been undertaken to gain insight into the nature of bifunctional organocatalytic activation. F

Full text of DOI:10.1002/chem.201304553

DOI: 10.1002/chem.201304553
Full Paper
&
Reaction Mechanisms
On the Mechanism of Bifunctional Squaramide-Catalyzed
Organocatalytic Michael Addition; Protonated Catalyst as an
Oxyanion Hole
[
a]
[a]
[a]
[b]
[a]
Bianka Kꢀtai, Gyçrgy Kardos, Andrea Hamza, Viktor Farkas, Imre Pꢁpai,* and
Tibor Soꢀs*
[
a]
Abstract: A joint experimental–theoretical study of a bifunc-
tional squaramide-amine-catalyzed Michael addition reaction
between 1,3-dioxo nucleophiles and nitrostyrene has been
undertaken to gain insight into the nature of bifunctional or-
ganocatalytic activation. For this highly stereoselective reac-
tion, three previously proposed mechanistic scenarios for
the critical CÀC bond-formation step were examined. Ac-
cordingly, the formation of the major stereoisomeric prod-
ucts is most plausible by one of the bifunctional pathways
that involve electrophile activation by the protonated amine
group of the catalyst. However, some of the minor product
isomers are also accessible through alternative reaction
routes. Structural analysis of transition states points to the
structural invariance of certain fragments of the transition
state, such as the protonated catalyst and the anionic frag-
ment of approaching reactants. Our topological analysis pro-
vides deeper insight and a more general understanding of
bifunctional noncovalent organocatalysis.
Introduction
Bifunctional acid-base catalysis is a fundamental and enduring
concept that is invoked to increase reaction rates of organic re-
[
1]
actions and mimic, or explain, enzymatic catalysis. Although
its noncovalent organocatalytic version has been around for
[
2]
decades, the discovery of the chiral bifunctional 1,2 tertiary
[3]
amino-thiourea catalyst by Takemoto and co-workers revital-
ized the field. In their catalytic system, the double hydrogen-
bond donor thiourea was considered to position and activate
the electrophiles, while parallel activation of nucleophiles was
assumed to occur by the neighboring tertiary amine. This type
of catalyst has proved to be a highly efficient and versatile pro-
moter for a wide range of synthetically important 1,2 and 1,4
Figure 1. Bifunctional thiourea- and squaramide-based organocatalysts.
[
4,5]
nucleophile–electrophile additions.
In addition to stereose-
lective methodological developments, there has been a surge
of interest in bifunctional organocatalyst design, and this
heightened activity has resulted in the evolution of some
prominent catalyst variants, such as cinchona- or squaramide-
[6,7]
based bifunctional catalysts (Figure 1).
Despite a steady growth in the number and apparent
impact of noncovalent bifunctional organocatalysis in the last
decade, the mechanism of these processes is still far from
being well understood. Although kinetic experimental studies
in thiourea-amine catalysis are consistent with the dual activa-
tion mechanism, computational investigations of the rate-de-
termining transition states indicated that the bifunctional
mechanism is a blanket term covering at least three closely re-
lated mechanisms (Scheme 1).
[
a] B. Kꢀtai, G. Kardos, Dr. A. Hamza, Dr. I. Pꢁpai, Dr. T. Soꢀs
Institute of Organic Chemistry
Research Centre for Natural Sciences
of the Hungarian Academy of Sciences
Pusztaszeri fflt 59-67
H-1025 Budapest (Hungary)
Fax: (+36)1-438-1145
E-mail: papai.imre@ttk.mta.hu
soos.tibor@ttk.mta.hu
[b] Dr. V. Farkas
MTA-ELTE Protein Modelling Research Group
Pꢁzmꢁny PØter sØtꢁny 1A
H-1117 Budapest (Hungary)
Path A (Scheme 1), a mechanism postulated by Takemoto,
involves the dual activation of the electrophile (El) and nucleo-
phile (Nu) by the double hydrogen-bond donor thiourea and
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201304553.
[8]
the tertiary amine, respectively. Subsequent experimental
Chem. Eur. J. 2014, 20, 1 – 10
1
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Results and Discussion
Experimental background
To investigate the underlying principles of the bifunctional
squaramide-amine-catalyzed Michael reactions and compare
them with those of thiourea catalysis, we chose the reaction
between 1,3-dioxo nucleophiles and nitrostyrene. This is one
of the benchmark reactions in bifunctional thiourea- and
squaramide-based organocatalysis, and affords Michael ad-
ducts with high levels of enantio- and diastereoselectivity.
As a first step, we synthesized a bifunctional diamino-cyclo-
hexane-based squaramide catalyst (cat). The molecular size of
this catalyst allowed computational analysis at a reasonably
high level of theory. The synthesis of the catalyst was accom-
plished by following a reported synthetic sequence from
simple and available building blocks, such as squaric acid di-
Scheme 1. Transition-state variants of the bifunctional mechanisms.
and theoretical work provided evidence that reactions promot-
ed by bifunctional thiourea catalysts could proceed through
[
9]
this mechanism. The generality of this heuristic model was
challenged by our theoretical studies, which suggested an al-
[
10]
ternative scenario for the rate-determining step. In this path-
way (Scheme 1, Path B), the relative role of the catalytically
active sites in substrate binding and activation is different. The
thiourea, as an efficient anion receptor, binds the deprotonat-
ed nucleophile, while the protonated amine activates and posi-
tions the electrophile. The possibility of bifunctional catalysis
through mechanism B was supported by additional theoretical
[
7l]
ethyl ester and trans-1,2-diaminocyclohexane. In the model
reaction with acetylacetone 1 and nitrostyrene 2 (Scheme 2),
[
11]
and experimental results. Finally, Wang and co-workers re-
cently suggested a variant of the bifunctional mechanism in
which the distal acidic NH group of thiourea, and not the pro-
[12]
tonated amine, activates and positions the electrophile. In
Path C, the deprotonated nucleophile is involved in double hy-
drogen bond interactions with the protonated amine and the
available thiourea NÀH bond. This novel mechanistic proposal
was corroborated by DFT calculations on the vinylogous Mi-
Scheme 2. Enantio- and diastereoselective Michael addition to nitrostyrene.
[12]
chael addition of a,b-unsaturated butyrolactam to chalcone.
At first sight, the above mechanistic diversity in the rate-de-
termining step might seem irrelevant in terms of the outcome
of the organocatalytic process; however, an in-depth under-
standing of these elementary steps is essential for the rational
design of improved catalyst systems. Furthermore, the implicit
conclusion drawn from the mechanistic studies is that the
structural and electronic factors of the reacting substrates de-
termine which bifunctional pathway is followed. This, some-
what perplexing, “substrate mechanistic control” perception
also raises the question of whether the preference of a given
mechanistic scenario can be altered by changing the funda-
mental structural features of the catalysts, for instance, the dis-
tance between the catalytically essential H-bond donor units.
To gain more insight into these intriguing mechanistic
issues, we initiated a joint experimental–theoretical study of
a bifunctional squaramide-amine-catalyzed Michael addition re-
action. This particular choice of catalyst seemed interesting not
only because, to the best of our knowledge, no mechanistic in-
vestigations have previously been reported for squaramide-as-
this catalyst showed very similar catalytic performance to
those of previously reported catalysts. Both the yield and the
enantioinduction capacity of the catalyst cat were similar to
[6a,7l,14]
those of the related catalysts previously reported.
The possibility of an additional enantiodifferentiation capaci-
ty of cat was then probed by using ethyl 2-oxo-cyclopentane-
carboxylate (4) as a prochiral nucleophile. It was found that, in
addition to high levels of enantioselectivity, high level of dia-
stereoselectivity could also be achieved, affording the product
with a stereoarray including a quaternary all-carbon stereocen-
ter (91% ee, 26:1 d.r.). The absolute and relative stereochemis-
try of adduct 5 was determined by vibrational circular dichro-
[15]
ism (VCD) spectroscopic measurements.
The enormous
impact of the bifunctional catalyst on diastereoselectivity
could be estimated by comparing the outcomes with those of
the triethylamine (TEA) promoted reaction. Thus, the bifunc-
tional squaramide cat was able to shift the diastereomeric
ratio from 2:3 (TEA) to 26:1.
[
13]
sisted bifunctional organocatalytic reactions,
but also be-
Encouraged by the result that the catalyst cat behaved simi-
lar to previous squaramide catalysts in the above Michael addi-
tions, we conducted density functional theory (DFT) calcula-
tions to distinguish between potential pathways A–C operat-
ing in the rate-determining step.
cause Wang et al. prognosticated that squaramide-based bi-
functional catalysis may preferentially follow Path C due to the
[12]
longer distance between the NÀH protons in squaramide.
&
&
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
2
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
À1
Conformational analysis of the catalyst
found to be slightly more favored, lying only 1.1 kcalmol
above cat-a. Additional low-lying catalyst conformers with in-
ternal H-bonds and other types of weak noncovalent interac-
tions were identified computationally (cat-e and cat-f), but
they were not considered relevant to the bifunctional mecha-
We first explored computationally the conformational space of
the chiral squaramide cat. In these calculations, potential
energy scans were performed along dihedral angles describing
the relative orientation of different units of the catalyst. Struc-
tures corresponding to the located energy minima were then
fully optimized, and the relative stabilities were estimated by
using the protocol described in the Computational Approach
section. The most stable structures that emerged from this
analysis are depicted in Figure 2.
[16]
nism assumed for the present reactions.
These results suggest that the calculated conformational dis-
tribution of catalyst cat is similar to that found previously for
[10]
the Takemoto catalyst in that the conformer that is consis-
tent with the dual substrate activation mechanism is an ener-
getically low-lying structure that is populated in the solution
phase.
Michael addition of acetylacetone to trans-b-nitrostyrene
In the next stage of our study, the elementary steps of the cat-
alytic Michael addition of acetylacetone (1) to trans-b-nitrostyr-
ene (2) were analyzed, focusing on possible mechanistic path-
ways regarding the CÀC bond formation and also on the
origin of the enantioselectivity.
Calculations carried out for binary catalyst–substrate systems
indicate that both reactants can form double H-bonds with the
squaramide unit (see Figure 3). These binding modes are
Figure 3. Double H-bonded catalyst–substrate adducts. Gibbs free energies
relative to the corresponding dissociation limits (cat+1 and cat+2) are
À1
shown in parenthesis (in kcalmol ).
slightly asymmetric, and both cat···1 and cat···2 complexes are
predicted to be thermodynamically unfavored (DG>0). Inter-
estingly, the enolic form of 1 displays another coordination
mode, with cat showing a single N-H···O type hydrogen bond
with the squaramide and also one O-H···N type bond (see
cat···1’ in Figure 4). This structure is slightly more favored and
it can be regarded as the reactant state for the protonation of
Figure 2. The most-stable conformers of catalyst cat. Relative free energies
À1
(in kcalmol ) are shown in parenthesis. Selected interatomic distances are
given in ꢃ.
The energetically lowest lying catalyst conformer (cat-a) fea-
the catalyst. The proton shift occurring via TS
is facile and
prot
+
À
tures both an intramolecular H-bond formed between the ter-
tiary amine and the neighboring NH group and a close contact
between the benzyl and dimethylamino groups. This latter in-
teraction seems to provide appreciable stabilization to this
structure because conformer cat-d, which differs only in the
leads to an ion-pair intermediate (catH ···1 ) lying only 1.3 kcal
À1
mol above the reactant level (cat+1). This species is stabi-
lized through multiple H-bonds involving one NH group from
the protonated amine and another from the squaramide. The
H-bonded network can be expanded through internal rear-
À1
+
À
position of the benzyl group, is predicted to be 1.9 kcalmol
rangement to structure catH ···1’ , however, this ion-pair con-
À1
+
À
higher in free energy. Conformers cat-b and cat-c represent
the active forms of the catalyst with the two N-H functionali-
ties aligned parallel and pointing towards the amine group,
forming a chiral catalytic pocket. Of the two structures, cat-b is
former is 4 kcalmol less stable than catH ···1 .
Three reaction pathways, corresponding to models A, B and
C, have been examined for the CÀC bond-formation process
between the activated nucleophile 1 and electrophile 2. Transi-
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
3
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
the squaramide NH units. The activation barrier represented by
À1
TS-A-R is 19.3 kcalmol with respect to the separated reac-
1
tants (cat+1+2). A slightly less stable transition state with sim-
ilar binding features, but different relative positioning of the H-
bond donor and acceptor sites was located on this pathway
(
see TS-A-R in Figure 5), pointing to a certain degree of struc-
2
tural flexibility of the stabilizing H-bonding network. The enan-
tiomeric adduct (S)-3 forms via TS-A-S and the computed acti-
1
À1
vation barrier is 20.8 kcalmol . Again, an additional transition
state (TS-A-S ) with a slightly higher barrier was found in our
2
calculations. It should be noted that in these latter two struc-
tures, only one of the -NO oxygen atoms engaged in H-bond-
2
ing with squaramide, thus the ideal double H-bonding pattern
is not established.
On reaction pathway B, the most favored transition state
À1
(
TS-B-R in Figure 6) provides a barrier of 15.5 kcalmol for
1
CÀC bond formation, which is significantly lower than that
found for Path A. This structure is characterized by a multiple
H-bonding pattern between the enolate and the squaramide,
wherein one of the carbonyl groups is bound to both NH units
and the second is involved in a single N-H···O bond shifted to-
wards the benzyl group of the catalyst. The nitro group of the
electrophile is bound to the protonated amine, which stabilizes
the negative charge developing on this unit upon CÀC bond
formation. Interestingly, an additional transition state having
a slightly different structure (the enolate shifted in the oppo-
site direction to the squaramide, that is, towards the cyclohexyl
Figure 4. Deprotonation of the nucleophilic acetylacetone (1) substrate.
Gibbs free energies relative to the cat+1 are shown in parenthesis (in kcal
mol ).
À1
tion states for pathways A and B could be easily located with
our standard computational protocol, however, all attempts to
identify C-type transition states failed. Our detailed analysis in-
dicates that the transition state of CÀC bond formation is not
[
15]
compatible with the H-bonding pattern defined by model C.
For reaction pathway A, the most stable transition state (TS-
A-R in Figure 5) is consistent with the formation of the (R) ste-
group) could also be identified (TS-B-R ), which was only
1
2
À1
reoisomer of adduct 3, which is the major product observed
experimentally. In this structure, the enolate ion is attached to
the protonated amine through N-H···O and C-H···O type hydro-
gen bonds, whereas nitrostyrene forms double H-bond with
1.5 kcalmol less stable than the former structure. In contrast
[10]
to bifunctional thioureas, bifunctional squaramides thus pro-
vide multiple reaction channels in this reaction even for
a given type of mechanistic pathway.
Along Path B, the computa-
tional analysis located two anal-
ogous transition states leading
to the minor enantiomeric prod-
uct. Those processes are predict-
ed to be significantly less fa-
vored in free energy (see TS-B-S1
and TS-B-S2 in Figure 6). It is
worth pointing out that the bar-
rier represented by TS-B-S is
1
À1
only 0.7 kcalmol
lower than
that arising from TS-A-S on the
1
other route, suggesting a scenar-
io in which a clear preference for
model B can be established for
the formation of the major prod-
uct, however, the minor product
species can be produced
through both pathways.
The relative free energies of
the entire set of transition states
are collected in Table 1 along
with their contributions to the
formation of (R) and (S) product
isomers. These results point to
Figure 5. Transition states identified on Path A for the reaction of 1 and 2. Predicted activation barriers (in kcal
mol ; computed as Gibbs free energies relative to cat+1+2) are given in parenthesis.
À1
&
&
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
4
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
2
(see Scheme 2). Based on our
experience with the previous re-
action, we considered only path-
ways A and B for computations.
Because substrate 4 is a prochiral
Michael donor, four different ste-
reoisomeric products can be
formed in this addition process,
depending on the facial ap-
proach of the reactants. A sche-
matic view of the corresponding
transition states on pathways A
and B is depicted in Figure 7
and Figure 8.
The conformational space of
these transition states is rather
complex. In addition to the
structural flexibility with respect
to the relative positions of the
H-bond donor and acceptor
sites, two puckered conforma-
tions of the five-membered ring
À
of the enolate (4 ) had to be
considered in the transition-state
search procedure. Consequently,
we identified four different tran-
Figure 6. Transition states identified on Path B in the reaction of 1 and 2. Predicted activation barriers (in kcal
mol ; computed as Gibbs free energies relative to cat+1+2) are given in parenthesis.
À1
Table 1. Relative Gibbs free energies of the located transition states and
[
a]
related product distribution for the reaction between 1 and 2.
À1
Path
A
Product
TS
DG [kcalmol
]
i
p [%]
TS-A-R
TS-A-R
TS-A-S
TS-A-S
TS-B-R
TS-B-R
1
19.3
20.1
20.8
22.4
15.5
17.0
20.1
21.3
0.2
0.0
0.0
0.0
92.6
7.2
(
(
(
(
R)
S)
R)
S)
2
1
2
1
2
B
TS-B-S
TS-B-S
1
0.0
0.0
2
[
a] Relative contribution of various pathways to the formation of 3 are
obtained from Maxwell–Boltzmann statistics: /RT)/
=100ꢄexp(ÀDG
(exp(ÀDG/RT)).
p
i
i
S
i
i
a very high degree of enantioselectivity for the investigated re-
action. However, the predicted enantiomeric ratio is much
higher than that observed experimentally. This might be par-
tially due to the inaccuracy of the present computational ap-
proach, but one cannot exclude that other unexplored reaction
pathways (e.g., simple base catalysis involving solely the dime-
thylamino group) may operate in addition to the bifunctional
Figure 7. Schematic view of transition states leading to various product iso-
mers along Path A in the reaction of 4 and 2.
[
17]
mechanism.
sition structures for each facial approach. These structures are
reported in the Supporting Information. Herein, we analyze
only the most stable forms of the transition states, which are
depicted in Figure 9 and Figure 10; the relative free energies
are collected in Table 2.
Michael addition of ethyl-2-oxocyclopentanecarboxylate to
trans-b-nitrostyrene
The CÀC bond formation transition states were also explored
for the reaction of ethyl 2-oxocyclopentanecarboxylate (4) with
The H-bonding patterns of the located transition states re-
semble those of the previous reaction and reflect the structural
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
5
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Table 2. Relative Gibbs free energies of the located transition states and
[
a]
related product distribution for the reaction between 4 and 2.
À1
Path
A
Product
TS
DG ([kcalmol
]
i
p [%]
(R,R)
(R,S)
(S,R)
(S,S)
(R,R)
(R,S)
(S,R)
(S,S)
TS-A-(R,R)
TS-A-(R,S)
TS-A-(S,R)
TS-A-(S,S)
TS-B-(R,R)
TS-B-(R,S)
TS-B-(S,R)
TS-B-(S,S)
18.0
15.4
18.4
16.2
17.2
12.0
19.1
15.0
0.0
0.3
0.0
0.1
0.0
92.0
0.0
0.5
B
[
a] Relative contribution of various pathways to the formation of product
isomers are obtained from Maxwell–Boltzmann statistics:
00ꢄexp(ÀDG/RT)/S(exp(ÀDG/RT)) using the entire set of transition
states (reported in Table S1 of the Supporting Information).
i
p =
1
i
i
i
by three N-H···O bonds and it is shifted from the symmetrical
arrangement to either the benzyl or cyclohexyl moieties of the
catalyst molecule.
Calculations predict the preferential formation of the (R,S)
product for route A, which is in line with experimental findings
Figure 8. Schematic view of transition states leading to various product iso-
mers along Path B in the reaction of 4 and 2.
(
the barriers represented by TS-A-(R,S) and TS-A-(S,R) differ by
À1
3.0 kcalmol ). However, the observed high diastereomeric
ratio (26:1) cannot be rationalized by this model, because the
S,S) diastereomeric product is calculated to be formed
(
through a relatively small barrier [TS-A-(S,S) is only 0.8 kcal
À1
mol less stable than TS-A-(R,S)]. On the other hand, model B
predicts high enantio- and diastereoselectivity for this reaction,
because the four transition states are significantly separated in
free energy. For the major (R,S) product, the barrier is found to
À1
be much lower on Route B (by 3.4 kcalmol ), whereas for the
other stereoisomers the computed barriers are quite similar on
the two reaction pathways. Consequently, these stereoisomers
are accessible through both reaction channels A and B. As
a matter of fact, the (S,R) enantiomeric product is predicted to
be formed preferentially through Route A.
Additional mechanistic insights from the structural analysis
The work described above reinforces the conclusion that the
underlying mechanistic principles in bifunctional squaramide
and bifunctional thiourea catalysis are the same. However,
there are some additional features for squaramide catalysis
that can be related to the elongated distance between the NÀ
H bonds of the squaramide. It was found that catalyst cat pro-
vides multiple transition states even within a given reaction
pathway. These energetically close-lying subvariant transition
structures differ slightly in the geometries and the pattern in
the hydrogen bonds, and inevitably increase the complexity of
the mechanism. Furthermore, the overall mechanistic picture
becomes even more complex by considering that some of the
stereoisomeric products seem to be formed in parallel through
two distinct pathways. Although, these findings can be inter-
preted as providing new insight, such mechanistic complexity
results in a much less general understanding and definitely
limits the potential of knowledge-based approaches in future
catalyst development.
Figure 9. Most-favored transition states identified on Path A for the reaction
of 4 and 2. Predicted activation barriers (in kcalmol ; computed as Gibbs
free energies relative to cat+4+2) are given in parenthesis.
À1
flexibility of these systems. On Route A, the nitroolefin is
bound through either one or both oxygen atoms to the squar-
amide, whereas on Route B, the enolate is attached to this unit
&
&
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
6
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
the mutual approach of reactants is also structurally well-de-
fined regardless of which pathway they follow.
We then compared the structures of transition states TS-B-
R and TS-B-S , which correspond to the most favored transi-
1
1
tion states leading to the formation of enantiomeric R or S
products, respectively. Similarly to the previous case, the posi-
tion and the orientation of the NH groups of the protonated
catalyst match remarkably well in the two structures (see
Figure 12). Surprisingly, the H-bonding patterns show close re-
semblance in the two transition states. Of the four oxygen
1 1
Figure 12. Overlays of transition states TS-B-R (blue) and TS-B-S (red) su-
perimposed via the cyclohexyl group of the protonated catalyst (the super-
imposed atoms are indicated by asterisks). H atoms are omitted for clarity
Figure 10. Most-favored transition states identified on Path B for the reac-
tion of 4 and 2. Predicted activation barriers (in kcalmol ; computed as
Gibbs free energies relative to cat+4+2) are given in parenthesis.
À1
(
except those involved in H-bonding).
atoms of the reacting species, three are involved in hydrogen
bonding and their positions relative to the chiral H-bond
donor environment are structurally invariant.
To develop a more general view on the investigated mecha-
nism, we carried out a detailed topological analysis of some of
the transition states identified for the reaction of 1 and 2. In
this analysis, we superimposed various transition states and ex-
amined the structural variation of two fragments, namely the
protonated catalyst and the remaining anionic adduct-like
species.
Based on the above fragment analysis, a more general
mechanistic picture begins to emerge. The results suggests
that the protonated catalyst formed upon the activation of
acetylacetone acts as a chiral oxyanion hole in this reaction
and stabilizes the anionic transition state of the approaching
reactants. Although the positions of the H-bond donor sites of
the oxyanion hole and the acceptor sites of the anionic species
are well-defined, several H-bonding patterns are available,
giving rise to different reaction pathways for the CÀC bond-
formation process and even to similar parallel reaction chan-
nels. The feasibility of these pathways and the stereoisomeric
preference are determined by the strength of the H-bonding
network formed between the two structurally invariant
fragments.
We first generated an overlay of the two most stable transi-
tion states obtained for routes A and B (TS-A-R and TS-B-R ).
1
1
In Figure 11a, the protonated catalyst fragments of the two
transition states are superimposed at the cyclohexyl groups.
Surprisingly, the two structures appear to be rather similar;
only a slight variation in the orientation of the squaramide unit
can be observed. The overlay of the anionic fragments reveal
almost identical geometries (see Figure 11b), illustrating that
Conclusion
We have conducted combined experimental and theoretical in-
vestigations on Michael addition reactions promoted by the bi-
functional squaramide tertiary amine organocatalyst cat. Based
on the presented results, the following conclusions can be
drawn. The mechanism of the present squaramide-catalyzed
reaction is analogous to that reported for thiourea catalysis in
terms of the basic elementary steps and the relative impor-
tance of previously proposed reactivity models. We found that
reaction pathway B, corresponding to electrophile activation
Figure 11. Overlays of transition states TS-A-R1 (blue) and TS-B-R1 (red).
a) The protonated catalyst superimposed via the cyclohexyl group, and
b) anionic adduct fragments (the superimposed atoms are indicated by as-
terisks). The complementary fragments are omitted for clarity.
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
7
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
via the protonated amine unit of the catalyst, is the most plau-
sible mechanism, which accounts reasonably well for the ob-
served stereoselectivity. An interesting new feature of the
squaramide-catalyzed reaction is that some of the reaction
channels described as pathway A also become accessible, im-
plying that the application of a single reactivity model might
not always be sufficient to rationalize the stereoselectivity out-
come of Michael addition reactions. The previously envisioned
Path C could not be identified as a viable mechanism for the
present reaction.
The application of the B3LYP functional for geometry optimizations
(
and subsequent single-point calculations at higher level) was also
[21]
justified in previous work.
The nature of the stationary points was characterized by using vi-
brational analysis. The initial structures for transition-state calcula-
tions were determined from the potential energy surface scan cal-
culations with respect to selected internal coordinates, which were
then followed by transition-state optimizations. The harmonic fre-
quencies were computed at the B3LYP/6–31G(d) level. These data
were also utilized to estimate the zero-point energies as well as
the thermal and entropic contributions to the Gibbs free energies.
The thermochemical data were obtained within the ideal-gas-rigid-
rotor-harmonic-oscillator approximation for T=298.15 K and p=
An additional feature of squaramide catalysis is the existence
of energetically close-lying multiple transition states within the
same pathway, differing only in the relative positions of H-
bond acceptor sites with respect to the squaramide NH bonds.
This is likely related to the increased H···H distance in squara-
mide (2.7 vs. 2.1 ꢃ in thiourea), which offers alternative binding
modes for the acceptor sites.
1
atm, however, concentration correction (0.00302 a.u.) was ap-
3
plied to Gibbs free energies corresponding to c=1 mol/dm condi-
tions in solvent phase. Solvent effects were also taken into account
at the B3LYP/6–31G(d) level by estimating the solvation free ener-
gies (solvent=dichloromethane) by using the integral equation
[22]
formalism variant of the polarizable continuum model (IEFPCM).
The overlay analysis of transition-state fragments revealed
that the protonated catalyst creates a geometrically invariant
chiral oxyanion hole. The topology of the acceptor oxygen
atoms of the anionic transition state fragment of reacting spe-
cies is also well-defined, therefore, the competing mechanistic
pathways can be unified under the concept of transition-state
tethering to the chiral oxyanion hole. This theoretical view sim-
plifies and explains the mechanism, suggesting that it is the
best fitting to the structurally rigid chiral oxyanion hole that
determines not only which pathway operates, but also which
enantiomer will be delivered in the catalytic cycle.
The atomic radii and nonelectrostatic terms in the IEFPCM calcula-
tions were those introduced recently by Truhlar and co-workers
[23]
(SMD solvation model).
The energy values reported in the paper correspond to solution-
phase Gibbs free energies, which are based on M06–2X/6–
3
11++G(d,p) electronic energies, and all additional terms were ob-
tained at the B3LYP/6–31G(d) level. All DFT calculations were car-
[24]
ried out with the Gaussian 09 software.
Acknowledgements
Due to the loose nature of hydrogen-bonding interactions,
the optimal H-bonding pattern cannot be easily predicted
even when the structures of the protonated catalyst and the
anionic fragment are available. However, the present findings
may help in the development of more efficient computational
tools for transition-state analysis, which may aid the design of
new catalysts. The development of force-field-based transition-
state screening methods is in progress in our laboratory.
We are grateful for the financial support of the Hungarian Sci-
entific Research Fund (OTKA, grants K-69086 and K-81927) and
the Lendꢅlet Program E-13/11/2010.
Keywords: density functional calculations · Michael addition ·
organocatalysis · reaction mechanisms · reaction intermediates
[
1] a) T. M. Lowry, I. J. Falkner, J. Chem. Soc., Trans. 1925, 127, 2883; b) T. M.
Lowry, I. J. Falkner, J. Chem. Soc. 1927, 2554; c) H. D. Paull, C. J. Abra-
ham, M. T. Scerba, E. Alden-Danforth, T. Lectka, Acc. Chem. Res. 2008, 41,
Experimental Section
6
55; d) T. Setoyama, Catal. Today 2006, 116, 250; e) M. Shibasaki, N.
Yoshikawa, Chem. Rev. 2002, 102, 2187; f) Catalysis by Acid and Bases
Eds.: K. Tanabe, B. Imelik, C. Naccache, G. Coudurier, Y. Ben Taarit, J. C.
General synthetic procedure
(
Catalyst cat (5 mol%, 0.05 mmol) was added to a mixture of nitro-
styrene 2 (149.2 mg, 1.0 mmol) and the appropriate nucleophile
Vedrine), Elsevier, Amsterdam, 1985; g) H. Steinhagen, G. Helmchen,
Angew. Chem. 1996, 108, 2489; Angew. Chem. Int. Ed. Engl. 1996, 35,
2339; h) J. D. Wuest, Acc. Chem. Res. 1999, 32, 81; i) J.-M. Ma, D. Cahard,
(
2.0 mmol) in dichloromethane (2 mL) and the mixture was stirred
Angew. Chem. 2004, 116, 4666; Angew. Chem. Int. Ed. 2004, 43, 4566.
2] a) C. G. Swain, J. F. Brown, J. Am. Chem. Soc. 1952, 74, 2534; b) W. P.
Jencks, Chem. Rev. 1972, 72, 705; c) H. Hiemstra, H. Wynberg, J. Am.
Chem. Soc. 1981, 103, 417.
3] T. Okino, Y. Hoashi, Y. Takemoto, J. Am. Chem. Soc. 2003, 125, 12672.
4] For recent books on organocatalysis, see: a) Comprehensive Enantiose-
lective Organocatalysis: Catalysts, Reactions, and Applications (Ed.: P. I.
Dalko), Wiley-VCH, Weinheim, 2013; b) Science of Synthesis: Asymmetric
Organocatalysis Volume 2 (Ed.: K. Maruoka), Thieme, Stuttgart, 2012.
for 5 h at RT until the reaction was complete (followed by TLC
analysis). The reaction mixture was then concentrated in vacuo
and the crude product was subjected to column chromatography
[
(
hexane/ethyl acetate, 5:1 to 1:1).
[
[
Computational approach
The geometries of the stationary points were optimized by using
[18]
[5] For recent reviews, see: a) J. Alemꢁn, A. Parra, H. Jiang, K. A. Jørgensen,
density functional theory (DFT) with the B3LYP functional and
the 6–31G(d) basis set. To provide accurate energetics, additional
single-point energy calculations were carried out for each structure
Chem. Eur. J. 2011, 17, 6890; b) W.-Y. Siau, J. Wang, Catal. Sci. Technol.
2
011, 1, 1298; c) S. J. Connon, Chem. Commun. 2008, 2499.
[
6] For seminal works on catalyst development, see: a) J. P. Malerich, K. Ha-
gihara, V. H. Rawal, J. Am. Chem. Soc. 2008, 130, 14416; b) B. Vakulya, S.
Varga, A. Csꢁmpai, T. Soꢀs, Org. Lett. 2005, 7, 1967; c) S. H. McCooey,
S. J. Connon, Angew. Chem. 2005, 117, 6525; Angew. Chem. Int. Ed.
2005, 44, 6367; d) J. Ye, D. J. Dixon, P. S. Hynes, Chem. Commun. 2005,
[19]
at the M06–2X/6–311++G(d,p) level of theory. In these calcula-
tions, the ultrafine integration grid was employed to ensure accu-
racy of numerical integration. The M06–2X functional was previous-
[20]
ly shown to perform reasonably for Michael addition reactions.
&
&
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
8
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
4
2
481; e) B.-J. Li, L. Jiang, M. Liu, Y.-C. Chen, L.-S. Ding, Y. Wu, Synlett
005, 603; f) see ref. [3].
[16] For a related NMR study, see: G. Tꢁrkꢁnyi, P. Kirꢁly, T. Soꢀs, S. Varga,
Chem. Eur. J. 2012, 18, 1918.
[
7] For recent selected examples using bifunctional squaramides, see:
a) K. S. Yang, A. E. Nibbs, Y. E. Turkmen, V. H. Rawal, J. Am. Chem. Soc.
[17] As noted by one of the referees, transition states derived from the E
À
isomer of enolate 1 may also be competitive in this reaction. Our pre-
2
2
2
013, 135, 16050; b) Y. Fukata, K. Asano, S. Matsubara, J. Am. Chem. Soc.
013, 135, 12160; c) Y. Liu, Y. Wang, H. Song, Z. Zhou, Adv. Synth. Catal.
013, 355, 2544; d) Q. Guo, J. C.-G. Zhao, Org. Lett. 2013, 15, 508; e) X.
liminary calculations indicate that some of these alternative pathways
are indeed feasible; however, the large energy difference between the
R- and S-type transition states (i.e., the overestimation of enantioselec-
tivity) is maintained. For a related computational study highlighting the
relevance of E-enolates in organocatalytic addition reactions, see: L.
Simꢀn, J. M. Goodman, J. Am. Chem. Soc. 2012, 134, 16869.
Fang, J. Li, C.-J. Wang, Org. Lett. 2013, 15, 3448; f) X. Fang, Q.-H. Li, H.-Y.
Tao, C.-J. Wang, Adv. Synth. Catal. 2013, 355, 327; g) M.-X. Zhao, H.-L. Bi,
H. Zhou, H. Yang, M. Shi, J. Org. Chem. 2013, 78, 9377; h) H. Wang, Y.
Wang, H. Song, Z. Zhou, C. Tang, Eur. J. Org. Chem. 2013, 4844; i) W.
Yang, Y. Yang, D.-M. Du, Org. Lett. 2013, 15, 1190; j) C. C. J. Loh, D. Hack,
D. Enders, Chem. Commun. 2013, 49, 10230; k) C. Jarava-Barrera, F. Este-
ban, C. Navarro-Ranninger, A. Parra, J. Alemꢁn, Chem. Commun. 2013,
[
[
[
[
[
[
18] A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
19] Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215.
20] T. A. Rokob, A. Hamza, I. Pꢁpai, Org. Lett. 2007, 9, 4279.
21] L. Simꢀn, J. M. Goodman, Org. Biomol. Chem. 2011, 9, 689.
22] J. Tomasi, B. Mennucci, E. Cancꢆs, J. Mol. Struct. 1999, 464, 211.
23] A. V. Marenich, C. J. Cramer, D. G. Truhlar, J. Phys. Chem. B 2009, 113,
49, 2001; l) G. Kardos, T. Soꢀs, Eur. J. Org. Chem. 2013, 4490.
[
[
8] T. Okino, Y. Hoashi, Y. Takemoto, J. Am. Chem. Soc. 2005, 127, 119.
9] a) S. J. Zuend, E. N. Jacobsen, J. Am. Chem. Soc. 2007, 129, 15872; b) P.
Hammar, T. Marcelli, H. Hiemstra, F. Himo, Adv. Synth. Catal. 2007, 349,
6378.
[
24] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, J. E. Peralta, F. Ogliaro, M.
Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobaya-
shi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J.
Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V.
Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Moro-
kuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dap-
prich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski,
D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
2
537; c) S. J. Zuend, E. N. Jacobsen, J. Am. Chem. Soc. 2009, 131, 15358.
10] A. Hamza, G. Schubert, T. Soꢀs, I. Pꢁpai, J. Am. Chem. Soc. 2006, 128,
3151.
11] a) D. Almasi, D. A. Alonso, E. Gꢀmez-Bengoa, C. Nꢁjera, J. Org. Chem.
[
[
1
2
2
009, 74, 6163; b) B. Tan, Y. Lu, X. Zeng, P. J. Chua, G. Zhong, Org. Lett.
010, 12, 2682; c) X. Han, R. Lee, T. Chen, J. Luo, Y. Lu, K. W. Huang, Sci.
Rep. 2013, 3, 2557; d) T. Azuma, Y. Kobayashi, K. Sakata, T. Sasamori, N.
Tokitoh, Y. Takemoto, J. Org. Chem. 2014, 79, 1805.
[
[
[
12] J.-L. Zhu, Y. Zhang, C. Liu, A.-M. Zheng, W. Wang, J. Org. Chem. 2012, 77,
9813.
13] For recent theoretical work on bifunctional squaramide-enamine cataly-
sis, see: T. Lu, S. E. Wheeler, Chem. Eur. J. 2013, 19, 15141.
14] a) H. Y. Bae, S. Some, J. S. Oh, Y. S. Lee, C. E. Song, Chem. Commun. 2011,
47, 9621; b) C. Min, X. Han, Z. Liao, X. Wu, H.-B. Zhou, C. Dong, Adv.
Synth. Catal. 2011, 353, 2715.
Received: November 20, 2013
[15] For details see the Supporting Information.
Published online on && &&, 0000
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
9
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
Finding the path: The mechanism of bi-
functional squaramide-promoted Mi-
chael addition of prochiral 1,3-dioxo nu-
cleophiles and nitroolefin has been
studied on the basis of DFT calculations.
Among the investigated mechanistic
scenarios, the pathway corresponding
to electrophile activation via the proton-
ated amine unit is found to be the most
feasible (see figure). For some of the
minor stereoisomeric products, alterna-
tive pathways are also accessible.
&
Reaction Mechanisms
B. Kꢀtai, G. Kardos, A. Hamza, V. Farkas,
I. Pꢁpai,* T. Soꢀs*
&
& – &&
On the Mechanism of Bifunctional
Squaramide-Catalyzed Organocatalytic
Michael Addition; Protonated Catalyst
as an Oxyanion Hole
&
&
Chem. Eur. J. 2014, 20, 1 – 10
www.chemeurj.org
10
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!