Organocatalysts Fold to Generate an Active Site Pocket for the Mannich Reaction

Article abstract of DOI:10.1021/acscatal.7b00336

Catalysts containing urea, thiourea, and tertiary amine groups fold into a three-dimensional organized structure in solution both in the absence as well as in the presence of substrates or substrate analogues, as indicated by solution NMR and computational studies. These foldamer catalysts promote Mannich reactions with both aliphatic and aromatic imines and malonate esters. Hammett plot and secondary kinetic isotope effects provide evidence for the C-C bond forming event as the turnover-limiting step of the Mannich reaction. Computational studies suggest two viable pathways for the C-C bond formation step, differing in the activation modes of the malonate and imine substrates. The results show that the foldamer catalysts may promote C-C bond formation with an aliphatic substrate bearing a cyclohexyl group by enhanced binding of the substrates by dispersion interactions, but these interactions are largely absent with a simpler catalyst. Additional control experiments demonstrate the ability of simple thiourea catalysts to promote competing side reactions with aliphatic substrates, such as reversible covalent binding of the thiourea sulfur to the imine which deactivates the catalyst, and imine to enamine isomerization reactions. In foldamer catalysts, the nucleophilicity of sulfur is reduced, which prevents catalyst deactivation. The results indicate that the improved catalytic performance of foldamer catalysts in Mannich reactions may not be due to cooperative effects of intramolecular hydrogen bonds but simply due to the presence of the folded structure that provides an active site pocket, accommodating the substrate and at the same time impeding undesirable side reactions.

Full text of DOI:10.1021/acscatal.7b00336

Subscriber access provided by Eastern Michigan University | Bruce T. Halle Library
Article
Organocatalysts Fold to Generate an Active
Site Pocket for the Mannich Reaction
Antti J. Neuvonen, Tamás Földes, Ádám Madarász, Imre Pápai, and Petri M. Pihko
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b00336 • Publication Date (Web): 30 Mar 2017
Downloaded from http://pubs.acs.org on March 30, 2017
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.
ACS Catalysis 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 13
ACS Catalysis
1
2
3
4
5
6
7
8
9
Organocatalysts Fold to Generate an Active Site Pocket for the
Mannich Reaction
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
Antti J. Neuvonen,† Tamás Földes,‡ Ádám Madarász,‡ Imre Pápai,*,‡ and Petri M. Pihko*,†
†Department of Chemistry and NanoScience Center, University of Jyväskylä, FIꢀ40014 JYU, Finland, and ‡Research Center
for Natural Sciences, Hungarian Academy of Sciences, Magyar tudósok körútja 2, Hꢀ1117, Budapest, Hungary
ABSTRACT: Catalysts containing urea, thiourea and tertiary amine groups fold into a threeꢀdimensional organized structure in
solution both in the absence as well as in the presence of substrates or substrate analogues, as indicated by solution NMR and comꢀ
putational studies. These foldamer catalysts promote Mannich reactions with both aliphatic and aromatic imines and malonate
esters. Hammett plot and secondary kinetic isotope effects provide evidence for the CꢀC bond forming event as the turnoverꢀ
limiting step of the Mannich reaction. Computational studies suggest two viable pathways for the CꢀC bond formation step, differꢀ
ing in the activation modes of the malonate and imine substrates. The results show that the foldamer catalysts may promote CꢀC
bond formation with an aliphatic substrate bearing a cyclohexyl group by enhanced binding of the substrates by dispersion interacꢀ
tions, but these interactions are largely absent with a simpler catalyst. Additional control experiments demonstrate the ability of
simple thiourea catalysts to promote competing side reactions with aliphatic substrates, such as reversible covalent binding of the
thiourea sulfur to the imine which deactivates the catalyst, and imineꢀtoꢀenamine isomerization reactions. In foldamer catalysts, the
nucleophilicity of sulfur is reduced, which prevents catalyst deactivation. The results indicate that the improved catalytic perforꢀ
mance of foldamer catalysts in Mannich reactions may not be due to cooperative effects of intramolecular hydrogen bonds, but
simply due to the presence of the folded structure that provides an active site pocket, accommodating the substrate and at the same
time impeding undesirable side reactions. KEYWORDS: organocatalysis, bifunctional, cooperativity, mechanism, kinetics, compuꢀ
tations, Mannich reaction
INTRODUCTION
and facilitates the turnoverꢀdetermining CꢀC bond formation
step.
Enzymes bind their substrates into a characteristic active
site cleft that contributes to binding and the remarkable degree
of selectivity obtained in enzymatic catalysis.1 In contrast,
many organocatalysts are small molecules that do not appear
to possess any folded active site cleft, yet they are often able
to exhibit remarkable degrees of selectivity as well.2
We have previously described a family of ureaꢀthioureaꢀ
tertiary amine catalysts (1aꢀc in Scheme 1) that appear to fold
into a conformation with an active site pocket. These catalysts
promote highly enantioselective Mannich reactions of maloꢀ
nates and βꢀketo esters with aromatic and aliphatic imines
(Scheme 1).3 In contrast to the prototype small molecule biꢀ
functional catalyst, Takemoto’s thioureaꢀtertiary amine cataꢀ
lysts (2a),4 these new catalysts readily accepted both aliphatic
and aromatic imines as electrophiles in the Mannich reaction.
The design of these catalysts was sparked by the seminal work
of Smith who showed that intramolecular hydrogen bonds
between urea and thiourea groups lead to enhanced reactivity.5
An alternative explanation for the higher activity of our cataꢀ
lysts is that the intramolecular hydrogen bond generates a
folded structure and an active site cleft in the catalyst. Such
folding is not possible with catalyst 2a.
Herein we show that these catalysts indeed form a wellꢀ
defined folded structure in solution, even in the presence of
substrates or substrate analogues. Furthermore, the most likely
reason for the superiority of the new catalyst family is that the
active site pocket, generated upon folding of the catalyst,
allows additional stabilizing interactions with the substrates
Scheme 1. Catalysts 1aꢀ1c and 2a and their contrasting reacꢀ
tivity
The dual activation modes of tertiary amineꢀthiourea bifuncꢀ
tional organocatalysts have previously been studied using both
computational and experimental methods.6 In addition, the Hꢀ
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 13
bond donor ability of thioureas and other hydrogenꢀbonding
a
NMR
CF3
b
NMR
catalysts have been studied by colorimetry by Kozlowski.7
However, the folded, more complex nature of catalysts 1aꢀc
suggests that mechanistic lessons learned in previous studies
should be complemented for these catalysts. To enable a more
rational design of future catalysts, and to expand the substrate
scope, we undertook a combined experimental and computaꢀ
tional mechanistic study of these new catalysts.
CF3
1
2
3
4
5
6
7
8
CF3
CF3
O
O
F
CF3
N
N
E
H
H
H
CF3
H
N
N
H
O
H
H
H
H
H
O
H
S
H
H
H
S
A
N
H
N
H
N
C
B
N
N
H
N
H
H
F3C
O
D
O
•
1a
free catalyst
1ahfacac
9
CF3
hfacac salt
RESULTS
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
Conformational Analysis of Catalyst 1a in Solution. We
had previously established that catalyst 1b, or its diastereomer
1c, could fold into at least two completely different conforꢀ
mations with different anions (Figure 1)8.
c
X-ray
CF3
CF3
O
CF3
N
N
H
H
O
cooperative mode
desired
anion receptor mode
undesired
S
(
)
(
)
N
H
N
N
H
O
H
N
H
H2N
•
•
1a
2 urea MeCN
urea complex
Figure 2. a) Structurally diagnostic 2D nOe cross peaks of 1a and
b) 1a·hfacac in benzeneꢀd6 at 30 °C. c) Xꢀray structure of 1a
complexed with urea and MeCN (MeCN molecule omitted for
clarity).
Our computational analysis9 provides further support for the
preference of the folded structure of catalyst 1a. We find that
the most stable form of 1a features an asymmetric double
hydrogen bonding pattern between the urea and thiourea moieꢀ
ties with an additional intramolecular hydrogen bond formed
between the tertiary amine and the adjacent NH group (see
Figure 3). The unfolded structures (i.e. those without NꢀHꢁꢁꢁS
hydrogen bonds) are all predicted to be at least 8 kcal/mol less
stable than the most favored conformer.10
X-ray of catalyst 1c . HCl
X-ray of catalyst 1c . hfacac
Figure 1. Comparison of Xꢀray structures of two different salts of
catalyst 1c, showing that at least two possible folds are readily
accessible.
The first conformation, corresponding to an intramolecularꢀ
ly hydrogen bonded structure, is observed when the catalyst
forms a salt with a malonate surrogate, hexafluoroacetyꢀ
lacetonate (hfacac). The hfacac salt might be viewed as an
analogue of the corresponding catalystꢀmalonate ion pair that
likely forms in the Mannich reaction. The second conforꢀ
mation was observed in the HCl salt of 1c.3a These results,
however, were obtained in the solid state.
To elucidate the major populated conformations in solution,
1
a benzeneꢀd6 solution of catalyst 1a was studied by H NMR
measurements. NOESY cross peaks in the free catalyst 1a in
benzene (Figure 2) are consistent with the obtained crystal
structure of the folded urea complex ((1a)2·urea·MeCN) of the
catalyst (see the Supporting Information). Importantly, the
same characteristic NOESY cross peaks were also observed in
the spectrum of the hfacac salt of 1a in benzene. These NMR
experiments thus established that the fold of the catalyst reꢀ
mains similar in both free state as well as the catalytically
relevant ion pair complex (assuming that the fold remains
similar with the malonate salt and the hfacac salt). The presꢀ
ence of several close contacts in the structure indicates that the
fold is tight, and the piperidine ring A as well as the indane
(CD) and one of the aryl rings (E) can form the walls of a
possible catalytic pocket (Figure 2).
Figure 3. The most stable conformer of catalyst 1a as predicted
by DFT calculations. Internal hydrogen bonds are indicated by
blue dotted lines (related distances are given in Å). Hydrogen
atoms are omitted for clarity, except those of NH groups.
Catalystꢀsubstrate binary complexes. In the folded strucꢀ
ture of catalyst 1a, the thiourea NH groups are easily accessiꢀ
ble by the substrate molecules, so the formation of binary
complexes is expected. Computations were carried out for
binary 1a∙substrate systems relevant to our present mechanisꢀ
tic studies (Scheme 2).
ACS Paragon Plus Environment

Page 3 of 13
ACS Catalysis
O
H
O
H
O
O
tBu
tBu
O
1
2
3
N
O
N
MeO
OMe
4a
: X = H
3a
4b: X = OCH3
4c
: X = CN
4d
X
4
5
6
7
Scheme 2. Substrates used in the computational study of binaꢀ
ry complexation.
8
9
Dimethylꢀmalonate 3a forms a multiple Hꢀbonded complex
with 1a as shown in Figure 4, however, the complexation is
predicted to be slightly endergonic. In complex 1aꢁ3a, one of
the CꢀH bonds of the methylene group is oriented towards the
basic center of the catalyst, so the malonate is structurally well
prepared for deprotonation. The deprotonation occurs via a
relatively low barrier (14.7 kcal/mol) and leads to an ionꢀpair
(1aH+ꢁ3a−) lying at +4.7 kcal/mol in free energy.11
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 5. The most stable form of complex 1aꢁ4a and the ion pair
formed upon substrate deprotonation. Relative stability (in
kcal/mol; with respect to 1a + 4a) is given in parenthesis.
Our attempts to identify complexes with malonate 3a and
catalyst 1a experimentally were not productive, which is in
line with the computed thermodynamics. However, with
acetylacetonate (acac), a more acidic βꢀdicarbonyl compound,
small shifts were indeed observed in NMR titration experiꢀ
ments (Figure 6). More pronounced shifts were observed in
NMR titrations with imine 4b, indicating that the imine interꢀ
acts with several key protons in the catalytic pocket of 1a (see
Figure 6).
CF3
CF3
acac 1a
Shift at 5:1 ratio of
HB1 +0.04 ppm
HC1 +0.02 ppm
HC2 +0.08 ppm
HN1 +0.32 ppm
HN2 +0.05 ppm
HN3 +0.04 ppm
:
O
F
E
CF3
N
N
2
3
H
H
O
H
2
H
S
A
1
C
B
N
N
N
1
1
H
H
D
H
Me
O
•
1a acac
O
Me
O
CF3
CF3
Figure 4. Binary complex 1aꢁ3a and the ionꢀpair formed upon
substrate deprotonation. Relative stabilities (in kcal/mol; with
respect to 1a + 3a) are given in parenthesis. For the sake of clariꢀ
ty, internal Hꢀbonds between urea and thiourea are not indicated.
Most of the hydrogen atoms are not shown either (except those of
NH groups and the CH2 of 3a in 1aꢁ3a). Carbon atoms of the
substrate are highlighted in dark green. Hydrogen involved in
proton shift is marked with blue H.
4b 1a
:
Shift at 5:1 ratio of
HB1 +0.14 ppm
F
E
CF3
N
N
2
3
HC1 +0.11 ppm
HC2 +0.19 ppm
H
H
O
H
2
H
S
A
1
HN1 +∼1.2 ppm
HN2 +0.15 ppm
HN3 +0.11 ppm
C
B
N
N
H
N
H
MeO
1
1
N
O
•
1a 4b
H
O
Various binding modes could be identified computationally
for imine 4a. In the most stable form, the imine binds via its N
atom to thiourea (see Figure 5) and this 1aꢁ4a complex is
practically isoenergetic with the dissociated state (1a + 4a).12
Paraꢀsubstituted aromatic imines 4b and 4c form very similar
binary complexes with 1a. The predicted stabilities (+0.3 and
+1.9 kcal/mol, respectively) do not fully reflect the trend exꢀ
pected from the nature of the substituents, which is likely
related to entropic effects.13 Analogous binary complex
formed with the aliphatic imine 4d is found to be notably less
stable (at +2.5 kcal/mol).
Figure 6. Summary of 1H NMR shifts observed when catalyst 1a
was titrated with a) acetylacetone (acac) or b) imine 4b.
Competition Experiments. To be able to directly compare
the reactivity of aromatic and aliphatic NꢀBoc imines we conꢀ
ducted a series of competition experiments using two foldamer
bifunctional catalysts (1a and 1b), Takemoto catalyst (2a) and
ureaꢀTakemoto catalyst (2b) (Figure 7). In order to avoid
problems associated with the imine isomerization, we selected
the imine 4d as the aliphatic imine since 4d is less prone to
isomerization side reactions than other aliphatic imines.14 As
expected, arylimine 4a and aliphatic imine 4d exhibited clearꢀ
ly different reactivity when Takemoto catalyst (2a) was used.
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 13
Interestingly, the urea variant of 2a (2b) displayed increased moto catalyst, however, the reaction with aliphatic imine 4d is
1
2
3
4
activity but maintained the clear reactivity difference between
4a and 4d. These graphs indicate that for aromatic imine 4a,
the foldamer structure of the bifunctional catalysts 1a or 1b
provide only a slight rate acceleration as compared to Takeꢀ
catalyzed more efficiently by foldamer catalysts. Moreover,
the foldamer catalyst 1b gave similar initial rates for both
aromatic and aliphatic imine.
5
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
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
0
100
200
300
t / min
400
500
0
100
200
300
t / min
400
500
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0
0
100
200
300
t / min
400
500
0
100
200
300
t / min
400
500
Figure 7. Mannich reaction progress curves in competition experiments between imines 4a and 4d. The product concentrations are plotted
as a function of time relative to the internal standard. Reaction conditions: Catalyst (10 mol%), [3a]0 = 0.2 M, [4a]0 = 0.1 M, [4d]0 = 0.1
M, [dibenzyl ether] = 0.2 M, benzeneꢀd6 at 30 °C.
ACS Paragon Plus Environment

Page 5 of 13
ACS Catalysis
We had already demonstrated the superiority of catalyst 1a
poor aromatic imines. For both the electron donating (ρ =
3.22; R2 = 0.989) and the electron withdrawing (ρ = 0.882; R2
= 0.978) substituents excellent correlations were obtained.
This relatively small deviation from linearity does not support
any profound change in the reaction mechanism but instead
suggests a change in the energy profile of the reaction as the
electronic properties of the imine are changed. With different
imines, the relative stabilities of the complexes preceding the
turnover determining transition state may change, which
would influence the reaction rate.
over 1b in reactions with βꢀketo esters.3b Catalyst 1a was also
superior to catalyst 1b in preparative experiments with dimeꢀ
thyl malonate (Table 1). Both catalysts provide excellent isoꢀ
lated yields regardless of the imine used. For both imines,
catalyst 1a gave better selectivity than catalyst 1b. Especially
in the case of aromatic imine 4a the selectivity difference was
significant.
1
2
3
4
5
6
7
8
9
Table 1. Preparative Mannich Reaction Examples with
Malonate using Foldamer Catalysts 1a and 1b
Nevertheless, the Hammett plot clearly suggests that the
rateꢀdetermining step involves a transition state with increasꢀ
ing electron density at the imine carbon. Thus, initial binding
of the imine, or final protonation of the intermediate, are unꢀ
likely to be rateꢀlimiting. An internal rearrangement within the
catalyst structure after the C–C bond formation event might, in
theory, exhibit a positive ρ value. In order to rule out this
possibility, a kinetic isotope effect study with deuteriumꢀ
labeled imines was carried out.
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
O
O
catalyst (10 mol %)
O
O
NBoc
H
+
MeO
OMe
NHBoc
MeO
OMe
R
Na2SO4
toluene, 0 °C
18 h
R
3a (100 mol %)
4 (200 mol %)
5
Cataꢀ
lyst
R
Yield (%)
er
1a
1b
1a
1b
Ph (4a)
99
99
99
95
97.0:3.0
7.5:92.5
1
KIE measurements were conducted by H NMR reaction
monitoring by comparing the reaction rates16 with aromatic
imines 4aꢀc and 4aꢀcꢀd1 with dimethyl malonate in parallel
experiments (Table 2). In all cases, a notable inverse secondꢀ
ary KIE (0.90–0.93) was observed. These KIE’s are indicative
of a change in the bonding environment (hybridization) of the
imine carbon in the rateꢀdetermining step. Together with the
Hammett plot data, these experiments strongly indicate that
the C–C bond formation is rateꢀdetermining. An alternative
rationalization for the KIEs, an inverse 2° equilibrium isotope
effect17 arising from the binding event of the imine, is not
compatible with the results of the Hammett plot, which indiꢀ
cates that the rateꢀdetermining step involves an increase in the
electron density of the imine carbon.
Ph (4a)
cꢀhex (4d)
cꢀhex (4d)
>99.5:0.5
1.0:99.0
Kinetic measurements. We anticipated that a linear freeꢀ
energy relationship, such as a Hammett plot (Figure 8), could
clarify the rateꢀlimiting step of the catalytic Mannich reaction.
For example, if the rate is limited by the binding of the imine
to the catalyst, a negative reaction constant (ρ) should be obꢀ
tained. In contrast, with a rate limiting C−C bond formation,
the ρ value should be positive.
O
O
NBoc
H
1a (10 mol %)
dibenzyl ether (0.2 M)
O
O
MeO
OMe
NHBoc
Table 2. Kinetic isotope effect measurements
+
MeO
[3a
OMe
]0: 0.2 M
X
Na2SO4
benzene-d6
30 °C
H
O
O
NBoc
H/D
1a
(10 mol %)
X
O
O
4
[ ]0: 0.2 M
5
dibenzyl ether (0.2 M)
MeO
OMe
NHBoc
+
MeO
OMe
Na2SO4
benzene-d6
30 °C
X
H/D
X
[3a]0: 0.2 M
[4]0: 0.2 M
5
X
rate0H
(mM*minꢀ
rate0D
(mM*minꢀ
KIE
(kH/kD)
1
1
)
)
OMe (4b)
H (4a)
0.0779
0.499
1.82
0.0865
0.552
1.97
0.90 ± 0.04
0.90 ± 0.16
0.93 ± 0.07
CN (4c)
Further support for the C–C bond formation step as the rateꢀ
determining step could be obtained from a crossꢀover experiꢀ
ment with the Mannich reaction product 5a (0.2 M) and imine
4c (0.2 M) in the presence of 10 mol% catalyst 1a. In this
experiment, no crossꢀover product was observed after 24 h,
indicating that the C−C bond formation step is essentially
irreversible.
Figure 8. Hammett plot. pꢀSMe imine was omitted from the line
fitting.
With 9 paraꢀsubstituted imines a clear positive Hammett
correlation was observed, however, there was significant deviꢀ
ation from linearity.15 In the Hammett plot, two subgroups
were clearly distinguishable, the electron rich and the electron
Inhibition and Complexation Experiments.
To clarify the importance of catalyst complexation with
imines we performed a kinetic experiment with pꢀCN benzalꢀ
dimine (4c) and substoichiometric amount of pꢀOMe benzalꢀ
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 13
dimine (4b) as a potent inhibitor (Scheme 3a). As the rate philic imine binds to the protonated amine unit via the carbonꢀ
1
2
3
4
5
6
7
8
difference of these imines in Mannich reaction is over 20ꢀfold,
4b can be considered purely as an inhibitor. The presence of
50 mol% of 4b retarded the Mannich reaction between pꢀCN
benzaldimine (4c) and dimethyl malonate by a factor of 1.88.
This inhibitory effect might result from a change in the turnoꢀ
verꢀdetermining intermediate (TDI) of the reaction.18 With
electronꢀrich imine, such as 4b, the complex with the catalyst
could become the TDI.
yl group of the Boc moiety.
In addition, 4b (50 mol%) inhibited the Mannich reaction of
4a by a factor of 2.91 (Scheme 3b). Although in this case the
reactivity difference of the imines is only 8ꢀfold, the Mannich
reaction rate of 4b was found to be negligible. These results
further support the notion that competitive substrates retard the
reaction by lowering the concentration of the productive comꢀ
plex, even if the interactions are weak (as evidenced by the
NMR titration experiments).
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
O
O
a)
NBoc
H
NBoc
H
dimethyl malonate (1 equiv)
1a (10 mol %)
dibenzyl ether (1 equiv)
MeO
OMe
NHBoc
+
NC
MeO
H
Na2SO4
benzene-d6
30 °C
[4c]0: 0.2M
[4b]0: 0.1M
NC
5c
inhib = 1.88
20-fold reactivity
difference
k/k
O
O
b)
NBoc
NBoc
H
dimethyl malonate (1 equiv)
1a (10 mol %)
dibenzyl ether (1 equiv)
MeO
OMe
NHBoc
H
+
MeO
H
Na2SO4
benzene-d6
30 °C
4a
[
4b
[ ]0: 0.1M
]0: 0.2M
5a
inhib = 2.91
8-fold reactivity
difference
k/k
Figure 9. CꢀC bond formation pathways explored computationalꢀ
ly for reaction 3a + 4a catalyzed by 1a. In the schematic view, the
carbon atoms involved in bond formation are labelled by black
dots. Relative stabilities (solution phase Gibbs free energies with
respect to reactant state 3a + 4a + 1a; in kcal/mol) are given in
parenthesis. Hydrogen bonds formed between the substrates and
the catalyst are indicated by blue dotted lines, whereas the develꢀ
oping CꢀC bond is illustrated by green dotted lines (related disꢀ
tances are given in Å). For clarity of images, hydrogen atoms are
omitted, except those of NH groups.
O
O
O
O
c)
dimethyl malonate (1 equiv)
NBoc
MeO
1a
(10 mol %)
OMe
NHBoc
MeO
OMe
NHBoc
dibenzyl ether (1 equiv)
H
+
H
H
Na2SO4
benzene-d6
30 °C
5a
k/kinhib = 2.26
4a
[
5a
[ ]0: 0.1M (90:10 er)
]0: 0.2M
Scheme 3. Inhibition experiments carried out with a) 4c and
4b, b) 4a and 4b, and c) 4a and 5a.
Finally, product inhibition with product 5a was also obꢀ
served (k/kinhib = 2.26, Scheme 3c), which justifies the need of
an excess of imine to drive the reactions to completion.
The CꢀC bond formation transition states identified compuꢀ
Ph
Ph
tationally on the two reaction pathways (TSF1 and TSF2 in
Figure 9) represent very similar barriers with respect to the 3a
+ 4a + 1a reactant state (17.7 and 17.9 kcal/mol, respectively)
suggesting that both addition mechanisms might be operative
in this particular reaction.19 This finding may seem surprising,
but it actually supports the view formulated recently in our
previous work that the application of a single reactivity model
might not always be sufficient to describe the mechanism of
bifunctional noncovalent organocatalysis.20
C–C bond formation pathways. The details of the C–C
bond formation process in present Mannich reactions were
explored computationally for the reaction 3a + 4a promoted
by foldamer catalyst 1a. We considered several possible
mechanistic scenarios for this step that involve different subꢀ
strate activation modes, and we found two feasible reaction
pathways as illustrated in Figure 9.
To test the two mechanistic scenarios against experimental
data, calculations were carried out for analogous reactions
with paraꢀsubstituted imines 4b and 4c as well. We find that
both CꢀC bond formation mechanisms account qualitatively
for the observed reactivity trend. As shown in Table 3, the
barriers predicted for the two pathways decrease gradually in
the 4bꢀ4aꢀ4c series, which is in line with the positive Hammett
correlation. It is also apparent that computations predict nearly
identical kinetic isotope effects for the two CꢀC bond forꢀ
mation pathways, all equal or very close to 0.9, which are in
reasonable agreement with the measured data (see Table 2).
On route 1, the deprotonated dimethylꢀmalonate is bound to
the protonated amine and to the proximal NH group of thiouꢀ
rea, whereas the imine is activated via the distal NH group of
the thiourea moiety of the catalyst. This type of substrate
activation has already been described in our previous work,3a
and it has also been identified as the most favored pathway for
organocatalytic vinylogous Michael reaction of α,βꢀ
unsaturatedꢀbutyrolactam to chalcone.6e On the other pathway,
on route 2, the deprotonated nucleophile is shifted to thiourea
displaying double Hꢀbonding interactions, and the electroꢀ
ACS Paragon Plus Environment

Page 7 of 13
ACS Catalysis
These results thus provide further support for the relevance of and they point to an enhanced reactivity of 4d with the foldaꢀ
1
2
3
4
5
6
7
8
parallel reaction channels in this Mannich reaction.
mer catalyst, the computed barriers are not fully consistent
with the results of competition experiments (Figure 2). The
reactivity difference between 4a and 4d is reasonably well
reproduced for the Takemoto catalyst, but computations seem
to overestimate the reactivity of 4d for the foldamer catalyst.21
The quantitative discrepancy between the computed barriers
and the observed rates could be related to the inaccuracy of the
applied computational method, which might be considerable
particularly for large molecular models involving the bulky
catalyst 1a.22
Table 3. Computed barriers (ꢀG‡; in kcal/mol) and kinetic
isotope effects (KIE) for reactions with paraꢀsubstituted
aromatic imines.a
route 1
route 2
ꢀG‡
ꢀG‡
Imine
4b
R
KIE
0.90
0.90
0.92
KIE
0.91
0.90
0.92
9
OMe
H
18.1
17.7
15.5
19.7
17.9
17.5
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
In addition to computational errors, the apparent discrepanꢀ
cy might be due to mechanistic events beyond the CꢀC bond
formation step, so we examined the mechanism of reprotonaꢀ
tion as well for the reaction 3a + 4a. This step involves a
proton shift from the protonated amine unit of the catalyst to
the N atom of the anionic adduct intermediate formed in the
CꢀC bond formation step. We find that proton transfer occurs
very easily on route 2 because the N atom of the adduct lies in
4a
4c
CN
a The barriers refer to reactions with nonꢀdeuterated imines.
Comparative analysis. To gain insight into the origin of the
unprecedented reactivity of foldamer catalysts in Mannich
reactions with aliphatic aldimines, we performed a comparaꢀ
tive computational analysis for reactions 3a + 4a and 3a + 4d
catalyzed by bifunctional organocatalysts 1a and 2a. The
computed barriers associated with the two CꢀC bond formation
mechanisms are collected in Figure 10.
Ph
close vicinity of the protonated amine (see TSF2 in Figure
9).23 On route 1, however, substantial structural rearrangement
in the Hꢀbonded network of this intermediate is required prior
to proton transfer (see TSF1Ph in Figure 9). This rearrangement
and subsequent reprotonation is still facile with the Takemoto
catalyst, however, it appears to be hindered with the bulky
foldamer catalyst. The difference found for the two catalysts is
clearly due to the presence of the urea sideꢀgroup in catalyst
1a, which imposes restriction on intramolecular structural
rearrangements needed for reprotonation. Potential energy
scan calculations suggest that transition states of this multistep
rearrangement process may lie above TSF1Ph, so the 3a + 4a
reaction may preferentially follow route 2 with catalyst 1a.24
In the 3a + 4d reaction, the reprotonation steps on route 1 is
likely hindered too, but this pathway could be still favored
over route 2,. It is therefore possible that the reactions with the
two imines (4a and 4d) take place via two different pathways
(route 2 and route 1, respectively) but with comparable rates.
We recall that experimentally, in the reaction with aromatic
imines (4aꢀ4c), the CꢀC bond formation is clearly rateꢀ
limiting, but the same may not hold true with aliphatic imine
4d. Unfortunately, reliable KIE analysis could not be perꢀ
formed for imine 4d due to side reactions.
ꢀG
22
20.1
4d
20
17.9
17.7
15.3
18
16
14
4a
4a
4d
route 1
route 2
ꢀG
22
20
18
16
14
19.7
17.6
4d
4a
17.2
15.1
4d
4a
route 1
route 2
On the beneficial role of the foldamer catalyst. To proꢀ
vide further understanding for the beneficial effect of catalyst
1a on the CꢀC bond formation barrier for imine 4d we anaꢀ
lyzed the structures and the nature of interactions in transition
states located on route 1 (see Figure 11).
Figure 10. Computed CꢀC bond formation barriers (in kcal/mol)
for reactions 3a + 4a (filled bars) and 3a + 4d (empty bars) proꢀ
moted by catalysts 1a and 2a.
Note first that the foldamer catalyst 1a offers a kinetically
favored pathway for the reaction with aliphatic imine 4d as
well, since the barrier computed on route 1 is only 15.3
kcal/mol, which is actually lower than that computed for 4a.
On the other hand, route 2 becomes less accessible in this
reaction as expected from the increased barrier (20.1
kcal/mol). The free energy data obtained for analogous reacꢀ
tions catalyzed by the Takemoto catalyst 2a reveal a different
pattern in the diagram. The barriers on route 1 are found to be
notably lower for both substrates, and the reaction with aroꢀ
matic imine 4a is predicted to be kinetically more favored on
both reaction pathways.
We note first that the intramolecular hydrogen bonds beꢀ
Ph
Cy
tween urea and thiourea in TSF1 and TSF1 are notably
strengthened as compared to the free foldamer catalyst (see
Figure 3), which is an evidence for the cooperativity, but these
changes alone do not account for the improved reactivity for
aliphatic imines. Cooperative effects amplify the acidity of the
thiourea unit in foldamer catalysts,25 which is likely beneficial
for both aromatic and aliphatic imines, however, this does not
translate to increased reactivity for aromatic imines, but only
for aliphatic imines.
Although these results corroborate the diverse catalytic efꢀ
fect of the two catalysts in the 3a + 4a and 3a + 4d reactions,
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 13
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 11. Bond distances (in Å) characteristic of Hꢀbonding interactions in CꢀC bond formation transition states on route 1 in reactions
3a + 4a and 3a + 4d promoted by catalysts 1a (upper structures) and 2a (lower structures). Catalystꢀsubstrate interactions are indicated in
blue, intramolecular ureaꢀthiourea interactions are in black. Relative stabilities (in kcal/mol; with respect to reactants) are given in parenꢀ
thesis.
Figure 12. Noncovalent contacts in C–C bond formation transition states on route 1 in reactions 3a + 4a and 3a + 4d promoted by cataꢀ
lysts 1a (upper structures) and 2a (lower structures). Protonated catalyst is represented via an isodensity surface (ρ = 0.01 au), imines 4a
ACS Paragon Plus Environment

Page 9 of 13
ACS Catalysis
and 4d are highlighted in red, malonate 3a in blue. Green regions represent weak noncovalent interactions as obtained from NCI analysis.
1
Applied cutoff for reduced density gradient is s = 0.3 au.
2
3
4
5
6
7
8
9
Ph
CꢀC bond formation transition states analogous to TSF1
Cy
Ph
and TSF1 were computed for each model. The barriers relatꢀ
In transition states involving the Takemoto catalyst (TST1
and TST1Cy) the binding of reacting partners is very similar, as
indicated by the distances characteristic of Hꢀbonding interacꢀ
tions (e.g. both imines form hydrogen bonds with a distance of
1.84 Å). In contrast, there are notable differences in these
ed to the identified transition states are listed in Table 4.
Table 4. Computed Barriers and their Differences Obꢀ
tained for Model Catalysts.a
distances in transition states involving the foldamer catalyst 1a
(TSF1 and TSF1Cy), and they show a more compact binding
Ph
ꢀGPh
17.7
18.5
17.9
19.9
11.8
ꢀGCy
15.3
18.6
18.3
21.1
14.9
ꢀꢀG = ꢀGCy – ꢀGPh
catalyst
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
for aliphatic imine 4d. For instance, the N−H∙∙∙N hydrogen
bond associated with the activation of imine 4d is significantly
1a
m1
ꢀ2.4
+0.1
+0.4
+1.2
+3.1b
Ph
Cy
elongated in TSF1 and it is shorter in TSF1 (1.94 and 1.80
Å, respectively). The longer N−H∙∙∙N hydrogen bond in
TSF1Ph is consistent with the trend obtained for relative stabiliꢀ
ties of the transition states and suggests that the foldamer
catalyst cannot accommodate aromatic imines as efficiently as
aliphatic imines into the catalytic pocket.
m2
m3
no cat
ꢀGPh and ꢀGCy refer to barriers computed for aromatic and
aliphatic imines (4a and 4d). In the absence of catalyst, the
barriers are computed relative to the 3a− + imine state.
a
b
Figure 12 shows another representation of the four transition
states, and highlights the differences in noncovalent interacꢀ
tions between the catalyst and the evolving adduct species. It
is apparent that the Ph/Cy groups of the imines and the Me
substituents of dimethylꢀmalonate have practically no contact
with catalyst 2a, but these groups do interact with various
parts of catalyst 1a. The presence of the urea side group in
catalyst 1a provides a wellꢀdefined binding pocket for the
reacting substrates, which induces selectivity with respect to
imine substitution. The cyclohexyl moiety of 4d seems to
accommodate more favorably into this pocket than the flat and
more rigid phenyl group of 4a, presumably because the more
flexible cyclohexyl ring can bend and fit in the binding pocket
generated by the rings C, D and E of the catalyst. This is also
visible from the more extended contact surface in TSF1Cy. The
As reported above, the barriers computed for imines 4a and
4d with the original catalyst (1a) are 17.7 and 15.3 kcal/mol
giving ꢀꢀG = ꢀ2.4 kcal/mol as a difference. The presence of
the indene ring and the neighbouring CF3ꢀsubstituted phenyl
group appears to be important in getting enhanced reactivity
with 4d, because without these groups (models m1 and m2) the
computed ꢀꢀG is close to zero (similar barriers for 4a and
4d). The elimination of the entire urea sideꢀgroup (model m3)
gives ꢀꢀG = +1.2 kcal/mol, so the catalysis becomes less
efficient for the aliphatic imine similarly to that with Takemoꢀ
to catalyst 2a (for which ꢀꢀG = +2.1 kcal/mol). We note that
in the absence of catalyst, the CꢀC bond formation is clearly
favored for the aromatic imine (ꢀꢀG = +3.1 kcal/mol).26
Ph
Cy
longer N−H∙∙∙N distance in TSF1 compared to TSF1 (and
also compared to TST1Ph) implies destabilizing steric hinꢀ
drance.
Our computational analysis thus suggest that the improved
catalytic performance of foldamer catalysts in Mannich reacꢀ
tions with aliphatic imines cannot be explained solely by the
cooperative effects of intramolecular Hꢀbonding interactions
between urea and thiourea, although cooperativity is clearly
evident in the shortening of the intramolecular Hꢀbonds in the
transition states. The structural fit of the substrates to the bindꢀ
ing pocket of the folded structure appears to be better with the
aliphatic imines, and this also contributes to their higher reacꢀ
tivity.
The role of the urea sideꢀgroup in catalytic selectivity was
further analyzed by calculations carried out for model catalysts
derived from 1a (Scheme 4).
Catalyst deactivation and substrate isomerization. To
further rationalize the differences between the catalysts 1a and
2a, we examined whether the aliphatic imines might suffer
additional disadvantages with catalyst 2a that are not present
with 1a. We were especially concerned about potential side
reactions, which were investigated by using a lessꢀhindered
and more reactive imine 4e (Scheme 5).
Indeed, we observed isomerization of the imine 4e to the
enamine (6) form with catalyst 2a (Scheme 5). This isomerizaꢀ
tion leads to low conversion as the isomerization is irreversiꢀ
ble in the reaction conditions. Previously, Takemoto and
coworkers have noted that problems with the bifunctional
thiourea catalysts27 can be avoided if the highly nucleophilic
thiourea moiety is replaced with a less nucleophilic Hꢀbond
donor.28
Scheme 4. Model catalysts derived from 1a. Units altered with
respect to the original catalyst are highlighted in blue.
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 13
The effect of the catalysts on the isomerization reaction was
studied in more detail by H NMR monitoring (Scheme 5a).
enamine. With the bulkier piperidine catalyst (1a) the rate of
1
1
2
3
4
5
6
7
8
isomerization was much slower than with the Nꢀ
dimethylamine foldamer catalyst (1b) and the E:Z selectivity
was also close to those obtained with 2a and 2b. These results
demonstrate that blocking the sulfur atom of the catalyst by
the foldamer structure does not solve the isomerization probꢀ
lem.
Mixing an aliphatic imine 4e and catalyst 2a in 1:1 ratio in
benzene led to immediate disappearance of the imine related
peaks. In addition, most of the catalyst had formed a new
1
species 7 that was stable enough to be detected on the H
NMR timescale. The degradation of this new species was
directly related to formation of the enamines 6 and also to an
increase in the catalyst peak integrals.
Overall, the deactivation and isomerization experiments
suggest that the reactivity difference between the foldamer
catalysts 1a and 1b and Takemoto catalyst 2a could at least
partially be attributed to the nucleophilicity of the sulfur atom
in 2a, resulting in fast but reversible deactivation of the cataꢀ
lyst by the aliphatic imine such as 4e. In catalysts 1a and 1b,
the intramolecular hydrogen bonds and the folding may comꢀ
pletely block the nucleophilic attack of the thiourea sulfur and
thus prevents deactivation of the catalyst. Isomerization of the
imine appears to be a side reaction that is common to all cataꢀ
lysts. A control experiment with triethylamine showed that
aliphatic imines are relatively stable under moderately basic
nonꢀnucleophilic conditions. The imineꢀtoꢀenamine isomerizaꢀ
tion appears to require the presence of Hꢀbond donors in the
catalyst.
Mixing imine 4e with 2ꢀmercaptopyridine resulted in a staꢀ
ble product (8) that could be fully characterized with NMR
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
1
(Scheme 5b, see Supporting Information). H, 13C, COSY and
HMQC measurements were all consistent with a 1,2ꢀaddition
product of 2ꢀmercaptopyridine to imine 4e. The stability of
this product implies that the amine base in Takemoto catalyst
is necessary for the adduct to decompose.
Conclusions. The folded structure of the foldamer ureaꢀ
thioureaꢀtertiary amine catalyst 1a was confirmed by solution,
solid state NMR and computational analyses, both in the free
state as well as when complexed with the imine substrate or in
an intramolecular salt form with substrate analogues (acac or
hfacac). In all cases, evidence for the folded structure could be
readily inferred from the structural data.
In competition experiments, aliphatic imine 4d and aromatic
imine 4a reacted at comparable rates with dimethyl malonate
when the foldamer catalysts 1a or 1b were used. In contrast,
with 2a and its urea variant 2b, 4d reacted very slowly as
compared to 4a.
Scheme 5. Deactivation of catalyst 2a: a) isomerization of
imine 4e, and b) addition of 2ꢀmercaptopyridine to 4e.
When the isomerization experiment was conducted with 20
mol% of 2a the isomerization process was sluggish as only
73% conversion was obtained after 2h (Table 5). However, a
urea version of the Takemoto catalyst (2b) was more effective
in the isomerization process while no covalently bound species
could be observed in the 1H NMR spectra.
The Hammett plot and the secondary kinetic isotope effects
measured for the Mannich reaction with aromatic imines supꢀ
ported a mechanism where the CꢀC bond forming event is the
turnoverꢀlimiting step. Computational studies revealed two
viable CꢀC bond formation pathways in these reactions, route
1 and route 2, that differ by the alternate activation modes of
the malonate and imine substrates. For aliphatic imine 4d
bearing a cyclohexyl group, route 1 allows a kinetically faꢀ
vored CꢀC bond formation process as well, which is found to
be a unique feature of foldamer catalyst 1a. For reactions with
Takemoto catalyst, computations predict significantly reduced
reactivity of 4d as compared to aromatic imines 4aꢀc, which is
in line with experimental observations. Our computational
analysis suggests that in addition to Hꢀbonding interactions,
the foldamer catalyst can further facilitate the CꢀC bond forꢀ
mation via dispersion forces provided by the catalyst’s binding
pocket. These stabilizing noncovalent interactions are scarcely
present in CꢀC bond formation transition states with simpler
thioureaꢀtertiary amine catalysts. These differences may exꢀ
plain the improved performance of the foldamer catalyst with
aliphatic imines.
Table 5. Isomerization of Aliphatic Imine with Bifunctionꢀ
al Organocatalysts
Catalyst
1a
Conversion (%)
Time (min)
E:Z
1:1.5
80
100
73
97
6
120
<8
1b
3.8:1
1:1.2
1:1.8
1:2.1
2a
120
40
2b
Et3N
120
Catalyst 2b also produced the enamine in higher Z selecꢀ
tivity than Takemoto catalyst. Surprisingly, the foldamer cataꢀ
lyst 1b was found to be the most effective in the isomerization
of the imine to the enamine, implying that the isomerization
process is favored by stronger Hꢀbond donors. In addition, the
selectivity for the double bond was inversed to favor the E
The folded structure of catalysts 1a and 1b also helps to
block the nucleophilicity of the thiourea sulfur atom, preventꢀ
ing catalyst deactivation via nucleophilic attack to imines.
This provides an additional reason for the improved catalytic
performance of the foldamer catalysts in the Mannich reacꢀ
tions with aliphatic imines.
ACS Paragon Plus Environment

Page 11 of 13
ACS Catalysis
7. (a) Huynh, P. N. H.; Walvoord, R. R.; Kozlowski, M. C. J. Am.
In summary, the results obtained herein point to the imꢀ
Chem. Soc. 2012, 134, 15621–15623. (b) Walvoord, R. R.; Hyunh, P.
N. H.; Kozlowski, M. C. J. Am. Chem. Soc. 2014, 136, 16055–16065.
8. Xꢀray structure of 1b and its hfacac salt have been published in
ref 3a.
9. In our computational analysis, most of the DFT calculations
(geometry optimizations, vibrational analysis, estimation of solvent
effects) were carried out at the M06ꢀ2X/6ꢀ311G(d,p) level of theory.
For each located structure, we carried out additional singleꢀpoint
energy calculations using the same functional along with the larger 6ꢀ
311++G(3df,3pd) basis set. The reported energetics refers to relative
solutionꢀphase Gibbs free energies (with benzene as a solvent). For
further details, see Supporting Information.
10. For details of the conformational analysis, see Supporting Inꢀ
formation.
11. For the corresponding transition state, see Supporting Inforꢀ
mation. The experimentally observed ionꢀpair complex formed beꢀ
tween catalyst 1a and hfacac is predicted to 10.4 kcal/mol more stable
than the reactant state.
12. Two different Oꢀcoordinated structures of complex 1aꢁ4a were
also identified computationally both lying about 2 kcal/mol higher
than the Nꢀcoordinated form.
13. Entropic loss appears to be more important upon the coordinaꢀ
tion of the methoxy substituted imine 4b (rotational degrees of freeꢀ
dom of the OMe group are constrained in the 1aꢁ4b complex. The
expected stability trend is well reproduced in terms of the binding
energies computed from the electronic energies (−20.9, −19.5 and
−18.5 kcal/mol, for the 4b, 4a and 4c series.
1
2
3
4
5
6
7
8
portance of the folded structure with an active site cleft, in
contrast to cooperative effects associated with the intramoꢀ
lecular hydrogen bond, as the explanation for the enhanced
reactivity of foldamer catalyst 1a and 1b with aliphatic imines.
AUTHOR INFORMATION
Corresponding Author
papai.imre@ttk.mta.hu, petri.pihko@jyu.fi
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
ASSOCIATED CONTENT
Supporting Information. Experimental procedures, additional
experiments pertaining to the mechanism, characterization data,
computational details, and copies of NMR spectra and GC chroꢀ
matograms.
ACKNOWLEDGMENT
Financial support from the Academy of Finland (project
#259532), University of Jyväskylä (postgraduate fellowship to A.
J. N.), and the Hungarian Scientific Research Fund (OTKA, grant
Kꢀ112028) are gratefully acknowledged. Computer facilities
provided by NIIF HPC Hungary (project 85708 kataproc) is also
acknowledged. We thank Dr. Elina Kalenius and Ms. Johanna
Lind for assistance with mass spectrometry and Mr. Esa
Haapaniemi for NMR assistance.
14. Formation of exocyclic double bonds to a 6ꢀmembered ring is
known to be less favorable than acyclic double bonds and as such
isomerization to form the enamine is disfavoured with 4d compared
to other aliphatic imines. For a discussion see: Brown, H. C.; Brewꢀ
ster, J. H.; Shechter, H. J. Am. Chem. Soc. 1954, 76, 467–474.
15. See the Supporting Information for details about rate measureꢀ
ments and initial rate determination.
REFERENCES
1. For related discussions, see: (a) Frey, P. A.; Hegeman, A. D.
Enzymatic Reaction Mechanisms, Oxford University Press, Oxford,
2007. (b) Petsko, G. A.; Ringe, D. In: Protein Structure and Function,
New Science Press, London, 2004.
2. For reviews, see: (a) Pihko, P. M.; Rahaman, H., Bifunctional
Acid-Base Catalysts. In: Enantioselective Organocatalyzed Reactions
1, R. Mahrwald Ed., Springer, Heidelberg, 2011, pp. 185–207. (b)
Takemoto, Y.; Inokuma, T. Bifunctional Thiourea Catalysts. In:
Asymmetric Synthesis II: More Methods and Applications, Christꢀ
mann, M.; Bräse, S. Ed., WileyꢀVCH, Weinheim, 2012, pp. 233–237.
(c) Claraz, A.; Siitonen, J. H..; Pihko, P. M. Iminium catalysis. In:
Lewis Base Catalysis in Organic Synthesis; Vedejs, E.; Denmark S.
Ed., WileyꢀVCH, Weinheim, 2016, Vol 2, pp. 805–856.
3. (a) Probst, N.; Madarász, Á.; Valkonen, A.; Pápai, I.; Rissanen,
K.; Neuvonen, A.; Pihko, P. M. Angew. Chem. Int. Ed. 2012, 51,
8495–8499. (b) Neuvonen, A. J.; Pihko, P. M. Org. Lett. 2014, 16,
5152–5155.
4. (a) Okino, T.; Hoashi, Y.; Takemoto, Y. J. Am. Chem. Soc.
2003, 125, 12672–12673. (b) Okino, T.; Hoashi, Y.; Furukawa, T.;
Xu, X.; Takemoto, Y. J. Am. Chem. Soc. 2005, 127, 119–125.
5. (a) Jones, C. R.; Pantoş, G.D.; Morrison, A. J.; Smith, M. D.
Angew. Chem. Int. Ed. 2009, 48, 7391–7394. (b) So, S. S.; Burkett, J.
A.; Mattson, A. E. Org. Lett. 2011, 13, 716–719. (c) A recent review
on the topic: Auvil, T. J.; Schafer, A. G.; Mattson, A. E. Eur. J. Org.
Chem. 2014, 2633–2646.
6. (a) Hamza, A.; Schubert, G.; Soós, T.; Pápai, I. J. Am. Chem.
Soc. 2006, 128, 13151–13160. (b) Hammar, P.; Marcelli, T.; Hiꢀ
emstra, H.; Himo, F. Adv. Synth. Catal. 2007, 349, 2537–2548. (c)
Almasi, D.; Alonso, D. A.; GómezꢀBengoa, E.; Nájera, C. J. Org.
Chem. 2009, 74, 6163–6168. (d) Tan, B.; Lu, Y.; Zeng, X.; Chua, P.
J.; Zhong, G. Org. Lett. 2010, 12, 2682–2685. (e) Zhu, J.ꢀL.; Zhang,
Y.; Liu, C.; Zheng, A.ꢀM.; Wang, W. J. Org. Chem. 2012, 77, 9813–
9825. (f) Han, X.; Lee, R.; Chen, T.; Luo, J.; Lu, Y.; Huang, K. W.
Sci. Rep. 2013, 3, 2557. (g) Azuma, T.; Kobayashi, Y.; Sakata, K.;
Sasamori, T.; Tokitoh, N.; Takemoto, Y. J. Org. Chem. 2014, 79,
1805–1817. (h) Žabka, M.; Šebesta, R. Molecules 2015, 20, 15500. (i)
Grayson, M. N.; Houk, K. N. J. Am. Chem. Soc. 2016, 138, 1170–
1173.
1
16. The reaction progress was followed by H NMR and the data
obtained were analyzed by linear regression of first measurement
points, i.e. the method of initial rates. See the Supporting Information
for details.
17. Gajewski, J. J.; Ngernmeesri, P. Org. Lett. 2000, 2, 2813–2815.
18. For the basic concepts of the energetic span model, see: Kozuch,
S.; Shaik, D. Acc. Chem. Res. 2011, 44, 101–110.
19. Although the analysis of stereoselectivity was not in the main
focus of the present study, we examined possible pathways towards
the minor enantiomeric product as well. The most favored transition
state corresponds to activation mode of route 1 lying at 25.8 kcal/mol
in free energy implying that the sense and the high degree of enantiꢀ
oselectivity is reproduced by computations (for details, see Supportꢀ
ing Information).
20. B. Kótai, G. Kardos, A. Hamza, V. Farkas, I. Pápai, T. Soós,
Chem. Eur. J., 2014, 20, 5631–5639.
21. Based on the results of competition experiments (see Figure 2),
one expects similar or slightly lower barrier for reaction with aromatic
imine 4a.
22. Test calculations carried out with the ωB97XꢀD functional give
similar trends for the CꢀC bond formation barriers suggesting that the
discrepancy is probably not related to the choice of the approximated
energy functional (for details, see Supporting Information).
23. CꢀC bond formation transition states and subsequent adduct
intermediates show a close structural resemblance (for details, see
Supporting Information).
24. The adequate description of the full pathway from the ionꢀpair
intermediate formed on route 1 to the reprotonation transition state is
challenging for large model used herein, because the structural rearꢀ
rangement likely involves several steps. In our present work, we
performed potential energy scan calculations using a single structural
parameter, namely the distance between the adduct N atom and the H
atom of the protonated amine. This simple approach can provide only
an upper limit for the barrier of structural rearrangement.
25. Computations predict very similar acidities of thiourea in cataꢀ
lysts 1a and 2a. Despite the aliphatic nature of the indene linker in 1a,
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 13
the pKa of the NꢀH group that binds the imine on route 1 is computed
28. (a) Inokuma, T.; Furukawa, M.; Uno, T.; Suzuki, Y.; Yoshida,
to be only 1.1 unit higher than that in 2a. For acidity enhancement via
internal activation, see: Couch, E. D.; Auvil, T. J.; Mattson, A. E.
Chem. Eur. J., 2014, 20, 8283–8287.
26. The barriers computed in the absence of catalyst are low because
they do not involve the deprotonation and catalyst binding processes,
which are both endergonic.
K.; Yano, Y.; Matsuzaki, K.; Takemoto, Y. Chem. Eur. J. 2011, 17,
10470–10477. (b) Inokuma, T.; Furukawa, M.; Suzuki, Y.; Kimachi,
T.; Kobayashi, Y.; Takemoto, Y. ChemCatChem 2012, 4, 983−985.
(c) Xiao, H.; Kobayashi, Y.; Takemoto, Y.; Morokuma, K. ACS
Catal. 2016, 6, 2988–2996.
1
2
3
4
5
27. Xu, X.; Yabuta, T.; Yuan, P.; Takemoto, Y. Synlett, 2006, 137–
140.
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
ACS Paragon Plus Environment

Page 13 of 13
ACS Catalysis
TOC graphic:
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
13
ACS Paragon Plus Environment