Thiourea-catalysed ring opening of episulfonium ions with indole derivatives by means of stabilizing non-covalent interactions

Article abstract of DOI:10.1038/nchem.1450

Small organic and metal-containing molecules (molecular mass <1,000) can catalyse synthetically useful reactions with the high levels of stereoselectivity typically associated with macromolecular enzymatic catalysts. Whereas enzymes are generally understood to accelerate reactions and impart selectivity as they stabilize specific transition structures through networks of cooperative interactions, enantioselectivity with chiral, small-molecule catalysts is rationalized typically by the steric destabilization of all but one dominant pathway. However, it is increasingly apparent that stabilizing effects also play an important role in small-molecule catalysis, although the mechanistic characterization of such systems is rare. Here, we show that arylpyrrolidino amido thiourea catalysts catalyse the enantioselective nucleophilic ring opening of episulfonium ions by indoles. Evidence is provided for the selective transition-state stabilization of the major pathway by the thiourea catalyst in the rate- and selectivity-determining step. Enantioselectivity is achieved through a network of attractive anion binding, cation-π and hydrogen-bond interactions between the catalyst and the reacting components in the transition-structure assembly.

Full text of DOI:10.1038/nchem.1450

ARTICLES
PUBLISHED ONLINE: 16 SEPTEMBER 2012 | DOI: 10.1038/NCHEM.1450
Thiourea-catalysed ring opening of episulfonium
ions with indole derivatives by means of stabilizing
non-covalent interactions
Song Lin and Eric N. Jacobsen
*
Small organic and metal-containing molecules (molecular mass <1,000) can catalyse synthetically useful reactions with
the high levels of stereoselectivity typically associated with macromolecular enzymatic catalysts. Whereas enzymes are
generally understood to accelerate reactions and impart selectivity as they stabilize specific transition structures through
networks of cooperative interactions, enantioselectivity with chiral, small-molecule catalysts is rationalized typically by the
steric destabilization of all but one dominant pathway. However, it is increasingly apparent that stabilizing effects also play
an important role in small-molecule catalysis, although the mechanistic characterization of such systems is rare. Here, we
show that arylpyrrolidino amido thiourea catalysts catalyse the enantioselective nucleophilic ring opening of episulfonium
ions by indoles. Evidence is provided for the selective transition-state stabilization of the major pathway by the thiourea
catalyst in the rate- and selectivity-determining step. Enantioselectivity is achieved through a network of attractive anion
binding, cation-p and hydrogen-bond interactions between the catalyst and the reacting components in the transition-
structure assembly.
ultifunctional urea and thiourea derivatives are known to background rate of 1a and acid alone. A broad screen of Brønsted
promote enantioselective reactions of cationic species in acids revealed a pronounced counterion effect. In conjunction
M
nonpolar media by binding the corresponding counter- with thiourea 3b, mineral acids with a nucleophilic counteranion,
anion through hydrogen bonding1–4, with selectivity imparted such as HCl, produced only trace amounts of the desired indole
through a combination of electrostatic association and additional addition product 2a (entry 1, Table 1), and the corresponding chlor-
non-covalent interactions that differentiate the diastereomeric tran- ide addition product predominated. In contrast, sulfonic acid
sition structures that lead to the product enantiomers5–7. We sought co-catalysts afforded 2a in useful yields and varying levels of enan-
to extend the anion-binding catalysis concept to episulfonium ions, tioselectivity, with the most promising result provided by 4-nitro-
which are highly reactive electrophilic species that readily undergo benzenesulfonic acid (4-NBSA) (entry 5, Table 1). The identity of
diastereospecific bond-forming reactions with nucleophiles8–11
.
the aromatic substituent on the pyrrolidino amide portion of the
Recently, Toste and co-workers demonstrated enantioselective cata- catalyst also exerted a profound effect on reaction enantioselectivity
lysis of the addition of alcohols to episulfonium ions using a chiral (entries 5, 7–12, Table 1). Catalyst 3a, which lacks an aryl
phosphoric acid catalyst12. On the basis of our previous work in group, induced little rate acceleration above that of the background
anion-binding catalysis, we envisioned that a urea or thiourea reaction (entry 6, Table 1) and afforded a nearly racemic product.
could serve as a suitable host for an episulfonium ion formed In contrast, catalysts 3c–3g, which bear more-extended aromatic
in situ through interactions with the counteranion (Fig. 1). High substituents, proved more enantioselective than catalyst 3b.
enantioselectivity might be achieved if additional interactions Correlation between enantiomeric excess (e.e.) and either the
between the catalyst and episulfonium intermediate could be incor- electronic properties of the aryl substituents or the polarizability
porated to stabilize differentially the diastereomeric pathways. This of the aromatic ring occurs in other reactions using this family
hypothesis was investigated in the context of a Friedel–Crafts-type of catalysts2,6,14. However, in the present case no such straight-
indole alkylation reaction13.
forward relationship was observed. Instead, e.e. improved on
expanding the aryl group from phenyl to phenanthryl (entries 5,
8–10, Table 1), and then to decrease slightly with more expansive
Results and discussion
Reaction methodology development. Preliminary efforts to
identify a suitable episulfonium-ion precursor revealed that a
relatively non-nucleophilic leaving group was required to achieve
the desired reactivity. Ultimately, we found that stable
trichloroacetimidates of type 1a (ref. 12) were particularly useful
substrates (equation (1), see Table 1) that undergo protonolysis
and substitution with a variety of strong Brønsted acids to form a
meso-episulfonium ion with a counteranion that can be varied
readily, based on the identity of the acid employed.
Attractive
non-covalent
interactions
S
R3
R4
R1
R1
SR2
Thiourea
R1
R1
SR2
Nu
N
H
N
H
NuH
–
R1
R1
– HX
– thiourea
X
X
+
SR2
Enantio-
enriched
Anion-binding
Racemic
A wide variety of chiral urea and thiourea derivatives was evalu-
ated in the model reaction (Table 1). Only arylpyrrolidine-derived
thioureas of type 3 were found to induce reactivity above the
Figure 1 | Proposed thiourea-catalysed ring-opening of episulfonium ions
with indole via anion binding.
Department of Chemistry and Chemical Biology, Harvard University, Cambridge, Massachusetts 02138 USA. e-mail: jacobsen@chemistry.harvard.edu
*
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
© 2012 Macmillan Publishers Limited. All rights reserved.
817

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
nucleophilicity parameters N , 4) (ref. 19) proved unreactive.
Variation of the substituents on the carbon backbone of the electro-
phile revealed that aryl groups with meta- and ortho-functionalities
were compatible (entries 19–22, Table 2); substrates that carry
Table 1 | Reaction optimization.
Ph
SBn
Indole (2.0 equiv.)
Ph
Ph
SBn
O
Ph
Catalyst (10 mol%)
Acid (7 mol%)
a
para-substituent, regardless of its steric and electronic
(1)
properties, resulted in substantially lower enantioselectivity
(entries 23–26, Table 2). Ongoing computational studies suggest
that in the transition state leading to the major enantiomeric
product, one of the para-C–H bonds is engaged in an
attractive, electrostatic interaction with the thiourea-bound
sulfonate. We reasoned that the lack of this interaction in the
cases with para-substituted substrates might result in a less
well-organized transition structure and reduced selectivity. Finally,
three different acetimidate leaving groups displayed essentially the
same reactivity and enantioselectivity, which suggests that the
leaving group is not directly involved in the e.e.-determining step
(entries 28–30, Table 2).
Toluene (0.05 M)
N
HN
)-1a
CCl3
H
4 Å MS, –30 °C, 12 h
(
2a
CF3
Catalyst:
t-Bu
X
N
Ar = H
N
N
CF3
H
H
Ar
O
a
b
c
3 (X = S)
4 (X = O)
Elucidation of the rate- and enantio-determining step. Thiourea
derivatives of type 3 bearing specific arylpyrrolidino residues are
highly enantioselective catalysts for nucleophilic additions to a
remarkable variety of cationic electrophilic intermediates,
including oxocarbenium ions2, acyliminium ions6, acylpyridinium
ions20 and now episulfonium ions. Potentially, elucidation of the
mechanisms that underlie catalytic activity and stereoinduction
by these thioureas may serve as a foundation for the discovery
of new reactions and provide broader insights into cooperative
d
e
f
g
Entry
Catalyst
Acid
Yield (%)*
e.e. (%)†
1
2
3
4
5
6
7
8
9
10
11
12
13
3b
3b
3b
3b
3b
–
3a
3c
3d
3e
3f
HCl
HOTf
FSO3H
10
73
78
79
72
7
5
32
19
2,4-diNBSA
4-NBSA
4-NBSA
4-NBSA
4-NBSA
4-NBSA
4-NBSA
4-NBSA
4-NBSA
4-NBSA
63
73
n/a
12
84
85
93
91
non-covalent pathways. We therefore undertook
a detailed
experimental and computational investigation of the indole
addition to episulfonium ions promoted by 3e and related
catalysts. On the basis of the qualitative observations described
above and previous studies that involved pathways for binding the
thiourea anion, a simple catalytic cycle for the reaction of indoles
with 1 mediated by 3e can be advanced (Fig. 2). The absence of a
dependence of enantioselectivity on the identity of the acetimidate
group suggests that the reaction begins with protonation of the
trichloroacetimidate substrate, followed by the formation of
episulfonium ions in a step that may or may not be under the
influence of the thiourea catalyst. A series of 1H NMR studies
16
84
80
93
91
97
98
3g
4e
88
92
Optimization reactions were performed on a 0.05 mmol scale. *Isolated yields of material purified
chromatographically. †e.e. values determined by high-performance liquid chromatography analysis.
MS ¼ molecular sieves, 2,4-diNBSA ¼ 2,4-dinitrobenzenesulfonic acid.
aryl substituents (entries 11–12, Table 1). Urea catalyst 4e induced of preformed episulfonium-ion derivatives17 revealed that the
only a marginally lower e.e. than its thiourea counterpart (entries episulfonium sulfonate most probably exists as a covalent adduct,
10 versus 13, Table 1), which indicates that any mechanism both in the presence and the absence of the thiourea catalyst, so
wherein the thiourea sulfur is engaged productively as a Lewis an endothermic ionization to the episulfonium-ion complex
base catalyst is not operative15–18
.
presumably takes place before addition of the indole nucleophile.
A variety of substrate combinations was evaluated to define the Finally, re-aromatization to form the product and regenerate the
scope as well as gain insight into the mechanism of the reaction acid closes the catalytic cycle. In the thiourea-promoted pathway,
(equation (2), Table 2). Substrates bearing electronically and the chiral catalyst must therefore induce enantioselectivity
sterically diverse sulfur substituents (R2) underwent enantio- through association with the charged intermediates and transition
selective reactions (entries 1–9, Table 2), and S-benzyl-substituted structures, indicated in Fig. 2 as taking place via direct binding to
derivatives afforded the highest e.e. values (entries 1 and 7, the sulfonate counterion.
Table 2). Electron-potential maps calculated using density
To identify the rate- and enantioselectivity-determining step in
functional theory showed that the benzylic protons in the the catalytic mechanism, reaction-progress kinetic analysis21 of the
S-benzyl episulfonium ions bear a substantial amount of partial reaction that employed substrate 1a, indole, thiourea 3e and
positive charge, which may serve to enhance attractive interactions 4-NBSA in toluene at 0 8C was conducted with in situ infrared
between the cationic intermediate and an electronegative spectroscopy22,23. Using reaction rates measured over 10% to 60%
functionality on the catalyst (see below). Various indole derivatives conversion with ‘different excess’ experiments22, the empirical rate
with electron-donating and -withdrawing substituents at the 2-, 4-, equations were determined for both the racemic reaction catalysed
5- or 6-positions all underwent the addition reaction with high by only 4-NBSA (equation (3)) and the asymmetric reaction
levels of enantioselectivity (entries 10–16, Table 2). In sharp co-catalysed by 4-NBSA and thiourea (equation (4)).
contrast, N-methylindole provided the desired product 2p
rrac = d[1a]/dt = krac[4-NBSA]T[indole]
(3)
with a moderate yield and in almost racemic form (entry 17,
Table 2). This suggested that the indole N–H motif may be involved
in a key interaction during the e.e.-determining transition state (see
below). Benzotriazole also underwent an addition reaction to form a
rasym = d[1a]/dt = kasym[4-NBSA]T[3e]T[indole]
(4)
C–N bond with synthetically useful e.e. (entry 18, Table 2). Less Here, krac is the second-order rate constant for the racemic reaction
nucleophilic heterocycles (for example, p-nucleophiles with Mayr and kasym is the third-order rate constant for the asymmetric
818
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
Table 2 | Substrate scope of nucleophilic ring opening of episulfonium ions with indole derivatives.
R1
SR2
R3
(2.0 equiv.)
R1
N
H
R1
R1
SR2
LG
R3
Thiourea 3e (10 mol%), 4-NBSA (7 mol%)
(2)
4 Å MS, toluene (0.05 M)
–30 °C, 40–63 h
N
H
1
2
Entry
R1
R2
LG
R3
Product
Yield (%)
e.e. (%)
1
2
3*
4
5
6
7
8
9
Bn
Ph
Ph
2a
2b
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
2m
2n
2o
2p
2q
2r
99
83
84
73
76
90
.99
72
89
97
93
83
88
92
83
95
54
92
85
97
95
.99
91
93
85
80
81
87
88
94
84
87
91
93
92
95
85
91
NH
4-F-C6H4
Ph
4-Me-C6H4
2-naphthyl
PMB
Me
t-Bu
H
O
CCl3
(TCA)
10
11
5-Me
5-MeO
5-Br
5-F
6-F
12
13
14
15
16
17
18
19†
20
21
22
23
24
25†
26
27‡
28§
Ph
Bn
TCA
4-MeO
2-Me
N-Me
79
3
Benzotriazoleꢀ
80
93
95
93
79
45
60
6
3-MeO-C6H4
3-F-C6H4
3-Me-C6H4
2-Me-C6H4
4-F-C6H4
4-Me-C6H4
4-MeO-C6H4
4-CF3-C6H4
–(CH2)4–
2s
2t
2u
2v
2w
2x
2y
2z
2b
Bn
Ph
TCA
TCA
H
89
67
18
16
87
5
9
87
NPh
29§
Ph
H
2b
91
85
O
CF3
NH
30§
2b
85
85
O
CHCl2
†
Isolated yields of material purified chromatographically are reported. *With urea 4c. Yields determined by 1H NMR are reported. Product isolated via flash column chromatography usually contains a small
amount of 3e. ‡R2 ¼ Ph, with HOTf. §With 20 mol% 3e and 10 mol% 4-NBSA. ꢀBenzotriazole as nucleophile instead of indole. PMB, p-methoxy benzyl.
reaction, and [4-NBSA]T and [3e]T are the total initial concen- as the covalent adduct) is the resting state of the substrate in
trations of acid and thiourea, respectively.
the reaction. The kinetic data are consistent with a reaction
The rate of the asymmetric reaction was accelerated by the chiral with indole that involves either addition or reversible addition
thiourea catalyst relative to the reaction catalysed by 4-NBSA alone. followed by slow re-aromatization as the rate-determining
For instance, at 273 K, [3e]T ¼ 50 mmol l21, the rasym/rrac at step (Fig. 2).
10% conversion of 1a is 43.4+5.0. This rate acceleration corre-
To distinguish between these two possibilities, studies of the
sponds to a lowering of the free energy of activation of the reaction linear free-energy relationships and kinetic isotope effects were
by 3e of 2.0+0.1 kcal mol21
.
carried out. In the racemic reaction catalysed by 4-NBSA alone, a
The zeroth-order rate dependence on 1a and first-order rate linear correlation was observed between the reaction rate and
dependence on 4-NBSA indicate that quantitative protonation of Mayr’s nucleophilicity parameter N for five different 5-substituted
the substrate occurs under the reaction conditions prior to the indoles (log(krac) versus sNN, R2 ¼ 0.997, where N is the Mayr
rate-determining step (pKa(4-NBSA) ≈ 27, pKa(1a) ≈ 2) (ref. 24). In nucleophilicity parameter and sN is the nucleophile-specific
support of this conclusion, treatment of substrate 1a with 1 equiv. parameter (see the Supplementary Information), and R2 is the
of 4-NBSA resulted in an instantaneous complete consumption correlation coefficient), consistent with the indole addition step
of 1a and a concomitant, quantitative formation of trichloro- that results in nucleophilic ring opening of the episulfonium ion
acetamide (the by-product generated during episulfonium-ion being the rate-limiting step25. A correlation between indole nucleo-
formation) as determined by in situ infrared spectroscopy. philicity and rate was also obtained in the thiourea-catalysed
The first-order dependence on indole in both the reaction reaction (log(kasym) versus sNN, R2 ¼ 0.757). As discussed below, a
catalysed by 4-NBSA alone or by 4-NBSA þ 3e reveals that indole strict linear correlation is not observed in this case because the
is present in the rate-determining transition structure and that Brønsted acidity of the indole N–H group also influences
.
episulfonium 4-nitrobenzenesulfonate (existing predominantly the reaction rate in the thiourea-catalysed reaction. Evaluation of
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
© 2012 Macmillan Publishers Limited. All rights reserved.
819

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
Ph
Ph
SBn
O
Ph
Ph
S
SBn
R
Ar'
*
Episulfonium-
ion formation
N
N
HN
CCl3
H
H
N
H
+
O
O
O
Re- aromatization
S
Ar
OH
H2N
CCl3
Ar = p-nitrophenyl
CF3
S
S
t-Bu
S
S
S
R
Ar'
*
R
Ar'
*
S
N
H
N
H
N
R
Ar'
*
N
H
N
H
N
H
N
CF3
N
H
N
R
Ar'
H
*
N
N
=
O
H
+
H
Ph
H
H
+
SBn
Ph
SBn
N +
3e
O
O–
Ar
O
O–
Ar
S
Ph
Ph
H
O
Ph
O
O
SO2Ar
SBn
Ph
‡
S
R
Ar'
*
N
H
N
H
N
H
H
O
O–
Ar
δ+
N
H
S
H
O
Nucleophilic addition
δ+
Bn
S
A
δ+
Ph
Ph
Rate- and enantio-determining step
Figure 2 | Proposed catalytic cycle for thiourea-catalysed ring-opening of episulfonium ions with indole. R* denotes an unspecified chiral group.
3-deuterioindole in the thiourea-catalysed addition reaction revealed act as a hydrogen-bond donor, either by reduction of its acidity
a very small effect of isotopic substitution (kH/kD ¼ 0.93+0.12). through substituent effects or by excision of one of the donor
This rules out re-aromatization as the rate-determining step, groups, led to strong or complete decreases in reactivity
which would be expected to display a significant primary and enantioselectivity in the indole addition reaction (see
isotope effect (kH/kD . 2.5) (ref. 26), and is fully consistent Supplementary Section 4). These data, taken together with the
with a rate-determining indole addition. It can be concluded strong e.e. dependence on the sulfonate counterion of the
that indole addition is enantio-determining as well, because Brønsted acid (Table 1) suggest strongly that the thiourea–
the product’s stereogenic centres are generated in this rate- sulfonate interaction is a key element in the mechanism for
determining step27.
catalysis and stereoinduction.
Elucidation of the catalyst–substrate interactions in the enantio- Hydrogen-bond interaction with the indole N–H group. As noted
determining step. Although the kinetic studies served to define above, a striking difference in enantioselectivity was observed
the stoichiometry of the transition structure in the rate- and between the N–H indole and N–methylindole analogues in the
e.e.-determining addition of indoles to 1, they do not provide addition reaction (93% versus 3% e.e. using catalyst 3e in
any direct insight into the specific manner by which the thiourea additions to 1a). This suggests an important organizational role of
catalyst induces rate acceleration and enantioselectivity. This the indole N–H bond in the stereoinduction mechanism, probably
proved attainable, however, through the series of structure– through hydrogen bonding to a Lewis basic functionality on the
reactivity and structure–enantioselectivity studies detailed below.
catalyst. To define the catalyst–indole interactions, a structure–
reactivity and structure–enantioselectivity relationship study
Thiourea dual hydrogen-bond donation to the sulfonate anion. was undertaken with a series of p-nucleophiles (Table 3).
Anion-binding catalysis is recognized as the primary mode of In general, the absence of an N–H motif in a (1,3)-relationship
substrate activation in a variety of asymmetric reactions that with the reactive nucleophilic site led to very low levels of rate
involve hydrogen-bond donors1. In studies relevant to the system acceleration and enantioselectivity (entries 1 and 2 versus entries
described here, sulfonate-ion associations with urea or thiourea 3–5, Table 3).
derivatives via dual hydrogen-bond donation interactions were
To probe further into the role of hydrogen-bond interactions
identified in both binding and reactivity studies28,29. To test between the indole nucleophile and catalyst, various 5-substituted
whether such sulfonate binding operates in the current reaction, indoles were evaluated in competition experiments under both
model binding studies and catalyst structure–activity studies were racemic and thiourea-catalysed reaction conditions. The degree of
conducted. Titration of solutions of thiourea 3e in d8-toluene rate acceleration induced by thiourea 3e was found to be linked
with dibenzylmethylsulfonium triflate [(Bn2MeS)þ(OTF)2] (5) directly to the acidity of the indoles, and a linear correlation
revealed the formation of a 1:1 complex, as determined by 1H between log(kasym/krac) and the pKa of indoles was observed
NMR spectroscopy. Resonances assigned to each of the thiourea (Fig. 3). Therefore, in the presence of thiourea the rate of the
N–H protons sharpened and shifted downfield on complexation, reaction is correlated not only to the intrinsic nucleophilicity of
consistent with
a
dual hydrogen-bond interaction29,30
.
the indole, but also to its hydrogen-bond donor properties.
Perturbations to the chiral catalyst that diminished its ability to No correlation was observed between reaction enantioselectivity
820
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
both rate- and e.e.-determining. The rate constant that corresponds
to the major pathway (kasym,major) could be deduced for catalysts
3a–3g from in situ infrared-based kinetic measurements combined
with enantiomeric ratio (e.r.) determinations. A strong correlation
between this rate and reaction enantioselectivity was observed: a
good linear fit was obtained in plots of ln(kasym,major) versus
ln(e.r.) (Fig. 4a). This provides unambiguous evidence that
enantioselectivity increases because variations of the aryl component
of the catalyst 3 are, indeed, tied to stabilization of the major
transition structure14. The rate of the pathway that leads to the
minor enantiomer also displays a linear, positive correlation with the
reaction e.r., which indicates that the minor transition structure is
also stabilized selectively by the more enantioselective catalysts,
albeit to a substantially lesser extent.
Insight into the nature of the transition-state stabilizing inter-
actions that may be at play was provided through spectroscopic ana-
lyses of thiourea derivatives 3a and 3e complexed to a sulfonium-
ion model system. The dibenzylmethylsulfonium ion triflate 5 was
selected for these studies because episulfonium sulfonate could
not be examined directly (see above) and the cation in 5 was
found by computational methods to have a very similar charge dis-
tribution to that of the episulfonium ion (Fig. 4c). With thiourea 3e,
the resonances of the benzylic and methyl protons of 5 underwent a
significant upfield shift (0.6–0.8 ppm) on formation of the 1:1
complex (Fig. 4b)37. In contrast, the chemical shift of these
protons was barely perturbed to any measurable extent (Dd ,
0.05 ppm) on complexation of 5 with the thiourea derivative 3a,
which lacks an aryl substituent. This points to an attractive
cation-p interaction between the p-face of the arene in catalyst 3e
and the sulfonium ion in the model system38–42. Such interaction
in the transition structure of the indole addition to episulfonium
ions could underlie the observed enantioselectivity effects, as the
cation-p interactions are expected to increase in magnitude with
Table 3 | Structure reactivity and enantioselectivity
relationship of p-nucleophiles.
Entry
Nucleophile
Yield (%)*
e.e. (%)*
r
asym/r†rac
1
93
93
43.4
n.d.
3.2
N
H
N
N
92
68
80
32
2
3
N
H
4
5
NH
67
57
9
1
3.2
3.8
N
Me
*Yields and enantiomeric excesses were obtained under reaction conditions described in equation
†
(2). The initial reaction rates with 4-NBSA alone (rrac) and with 4-NBSA plus 3e (rasym) were
determined directly by in situ infrared spectroscopy. The rasym/rrac value was not determined
(n.d.) for benzotriazole (entry 2) because the kinetic analysis was complicated by the poor
solubility of the nucleophile in the reaction medium.
and the acidity of the indoles, which indicates that this interaction more extended aromatic substituents43,44. Based on the kinetic and
probably exists to a similar degree in both the major and the enantioselectivity data, we therefore propose that the difference in
minor pathways. The kinetic data are thus consistent with a
general base activation of indole through an indole N–H–catalyst
4.3
hydrogen-bond interaction in the rate-determining addition to epi-
Cl
sulfonium ions31–34
. Catalyst 3e possesses few Lewis-base
F
functionalities (namely, the thiourea, the amide and perhaps the
extended arene substituent) and therefore the possible catalyst
hydrogen-bond acceptor sites are limited in number. The similar
reactivity and enantioselectivity displayed with urea 4e and thiourea
3e appear to rule out a direct role of the thiourea sulfur atom.
As discussed below, the extended aromatic plays a key role that
can be tied to interactions with the episulfonium ion. Therefore,
by this simple process of elimination we propose that the amide
oxygen is the most likely hydrogen-bond acceptor site for the
activation of the indole35,36. Consistent with this hypothesis, catalyst
3a was found to be more reactive than the corresponding thioamide
analogue in the model reaction (equation (2)). In addition, the
reaction catalysed by 3a is about four times as fast as the reaction
with Schreiner’s thiourea (1,3-bis(3,5-bis(trifluoromethyl)phe-
nyl)thiourea), which is a stronger H-bond donor, but lacks an
amide appendage.
Br
4.2
4.1
4.0
3.9
3.8
3.7
MeO
H
R2 = 0.9674
Me
15.5
16.0
16.5
pKa (indole N–H)
17.0
17.5
Stabilization of the cationic transition state through cation-p
interactions. The strong correlation between reaction
enantioselectivity and the identity of the arene on the catalyst Figure 3 | Rate acceleration by 3e and acidity of the N–H motif of the
suggests a direct role of the extended p-system in the mechanism 5-indole derivatives. Correlation between the degree of rate acceleration by
3e over the background racemic reaction (kasym/krac) and the acidity of the
of stereoinduction. In principle, this arene affect may be caused
primarily either by an acceleration of the major pathway through N–H motif of 5-substituted indole derivatives (pKa). The rate data were
transition-state stabilization or by inhibition of pathways that lead obtained by in situ infrared and 1H NMR spectroscopy (see Supplementary
to the minor enantiomer through destabilizing interactions. This Section 14 for detailed experimental procedures and data analyses). Error
question was addressed through a kinetic analysis of the reaction, bars reflect the range of experimental data from 2–3 individual
measurements, and the line represents the least-squares fit.
which takes advantage of the fact that the indole addition step is
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
821
© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
a
6
5
4
3
2
1
0
c
3f
Major pathway
Slope = 1.2 0.1
R2 = 0.95
3e
3d
3g
3c
3b
3b
3f
3d
3c
3a
3a
H
3e
H
H
3g
+S
H
Minor pathway
Slope = 0.2 0.1
R2 = 0.45
Ph
Ph
0.0
0.4
0.8
1.2
1.6
ln(e.r.)
2
2.4
2.8
CH3
+S
b
Ph
Ph
–OTf
H
H H H
5
H
H
3e + 5
3a + 5
H
H
H
+ S
H
Ph
H
Ph
5
5.0
4.5
4.0
3.5
2.5
2.0
1.5
1.0
Chemical shift (ppm)
Figure 4 | Enantio-induction with thiourea catalysts through a selective, attractive cation-p interaction between the extended aromatic residue on the
catalyst and the acidic a-protons in the episulfonium ion. a, Correlation between rate and enantioselectivity of reactions catalysed by thioureas 3a–3g.
Each data point represents the average rate determined from two individual kinetic experiments, and the error bar shows the range of the measurements.
The rate constants for the major and the minor pathways (kasym,major and kasym,minor, respectively) are calculated on the basis of the rate equations
(equations (3), (4), rasym ¼ rasym,major þ rasym,minor and e.r. ¼ rasym,major/rasym,minor). The lines represent least-square fits. b, A 1H NMR binding study of
thiourea and 5 in d8-toluene showed attractive interactions between the aromatic group in 3e and the a-protons in 5. The resonances of the benzylic protons
and the methyl protons in 5 are labelled with blue and green dots, respectively. In the bottom two spectra, the methyl resonance overlaps with the solvent
peak, but can be identified in expanded displays. c, Electrostatic potential maps for fully optimized structures (B3LYP/6-31G(d)) of the episulfonium ion
derived from 1a and the sulfonium ion in 5 reveal a similar distribution of positive charge over the benzylic protons. Negative potentials are shown in red and
positive potentials in blue.
the strength of the cation-p interaction between the major and
Taken together, the data presented above allow the construction of
minor pathways lies at the origin of the high reaction enantioselec- a detailed mechanistic model for the enantioselective reaction,
tivity observed6,14,45. On the basis of polarizability effects alone, wherein rate acceleration and enantioselectivity are induced by
which are known to correlate directly with cation-p binding the thiourea catalyst through a network of attractive non-covalent
ability46, the most expansive aryl-substituted catalysts 3f and 3g interactions. In particular, we propose that the transition structure
might be expected to be the most enantioselective. However, these for the rate-determining addition of indole to the episulfonium ion
two thiourea derivatives afford slightly lower e.e. values (see is stabilized by a combination of anion binding of the thiourea to
Table 1) in the model reaction than the smaller phenanthryl catalyst the sulfonate, general base activation of the indole via a catalyst
3e. The reason for the lack of an exact correlation between the polar- amide–indole N–H interaction and a cation-p interaction between
izability of the aromatic substituent and reaction enantioselectivity the arene of the catalyst and the benzylic protons of the episulfonium
is not yet established. However, it seems reasonable to expect that ion (Fig. 5). We anticipate that characterization of these enzyme-like
any number of minor steric or conformational factors could attenu- non-covalent stabilizing elements with small-molecule catalysts such
ate the ability of the largest substituents to engage fully in the stabi- as 3e may enable the future design and application of such biomimetic
lizing cation-p interaction.
strategies in organic asymmetric synthesis.
822
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
9. Smit, W. A., Caple, R. & Smoliakova, I. P. Stepwise electrophilic addition. Some
novel synthetic ramifications of an old concept. Chem. Rev. 94, 2359–2382 (1994).
10. Fachini, M., Lucchini, V., Modena, G., Pasi, M. & Pasquato, L. Nucleophilic
reactions at the sulfur of thiiranium and thiirenium ions. New insight in the
electrophilic additions to alkenes and alkynes. Evidence for an episulfurane
intermediate. J. Am. Chem. Soc. 121, 3944–3950 (1999).
‡
F3C
t-Bu
S
N
11. Toshimitsu, A., Hirosawa, C. & Tamao, K. Retention of configuration in the
Ritter-type substitution reaction of chiral/3-arylthio alcohols through the
anchimeric assistance of the arylthio group. Tetrahedron 50, 8997–9008 (1994).
12. Hamilton, G. L., Kanai, T. & Toste, F. D. Chiral anion-mediated asymmetric ring
opening of meso-aziridinium and episulfonium ions. J. Am. Chem. Soc. 130,
14984–14986 (2008).
N
H
N
H
CF3
O
H
O–
R
δ+
N
H
O
13. Wu, Y. J. in Heterocyclic Scaffolds II: Topics in Heterocyclic Chemistry Vol. 26
(ed. Gribble, G. W.) 1–29 (Springer, 2011).
H
S
H
14. Uyeda, C. & Jacobsen, E. N. Transition-state charge stabilization through
multiple non-covalent interactions in the guanidinium-catalyzed enantioselective
Claisen rearrangement. J. Am. Chem. Soc. 133, 5062–5075 (2011).
15. Denmark, S. E., Kornfilt, D. J. P. & Vogler, T. Catalytic asymmetric
thiofunctionalization of unactivated alkenes. J. Am. Chem. Soc. 133,
15308–15311 (2011).
H
O
Ph
Ph
S
δ+
Ph
16. Denmark, S. E., Collins, W. R. & Cullen, M. D. Observation of direct sulfenium
and selenenium group transfer from thiiranium and seleniranium ions to
alkenes. J. Am. Chem. Soc. 131, 3490–3492 (2009).
17. Denmark, S. E. & Vogler, T. Synthesis and reactivity of enantiomerically
enriched thiiranium ions. Chem. Eur. J. 15, 11737–11745 (2009).
18. Lucchini, V., Modena, G. & Pasquato, L. Enantiopure thiosulfonium salts in
asymmetric synthesis. Face selectivity in electrophilic additions to
unfunctionalised olefins. J. Chem. Soc. Chem. Commun. 1565–1566 (1994).
19. Mayr, H., Kempf, B. & Ofial, A. R. p-nucleophilicity in carbon–carbon bond-
forming reactions. Acc. Chem. Res. 36, 66–77 (2003).
20. Birrell, J. A., Desrosiers, J-N. & Jacobsen, E. N. Enantioselective acylation of
silyl ketene acetals through fluoride anion-binding catalysis. J. Am. Chem. Soc.
133, 13872–13875 (2011).
21. Blackmond, D. G. Reaction progress kinetic analysis: a powerful methodology
for mechanistic studies of complex catalytic reactions. Angew. Chem. Int. Ed. 44,
4302–4320 (2005).
Figure 5 | Proposed transition-structure model. The transition structure for
the rate-determining addition of indole to the episulfonium ion is stabilized
by a combination of attractive, non-covalent interactions, which include
anion binding of the thiourea to the sulfonate, general base activation of the
indole via a catalyst amide–indole N–H interaction and a cation-p interaction
between the arene of the catalyst and the benzylic protons of the
episulfonium ion.
Methods
General procedure for thiourea 3e-catalysed nucleophilic ring opening of
episulfonium ions with indole derivatives. An oven-dried 1.5 dram vial was
charged with substrate 1 (0.05 mmol, 1.0 equiv.), 3e (3.2 mg, 0.005 mmol, 0.1
equiv.), indole (11.7 mg, 0.1 mmol, 2.0 equiv.) and 4 Å molecular sieves (25 mg,
powder, activated) under an atmosphere of N2. The vial was cooled to –30 8C, and
toluene (1 ml) was added with stirring. Once the reactants and catalyst had dissolved
fully, the mixture was cooled to –78 8C and solid 4-NBSA (0.7 mg, 0.0035 mmol,
0.07 equiv.) was added at once against a counterflow of N2. The resulting solution
was stirred at –30 8C and the progress of the reaction was monitored by thin layer
chromatography (see Supplementary Section 3). When the progress of the reaction
was determined to be complete, triethylamine (ꢁ10 ml) was added at –30 8C. The
resulting mixture was applied directly to a pipette column that contained 4–5 cm of
silica gel, and the product was isolated by eluting hexanes/ethyl acetate (20:1 to 10:1)
and solvent removal.
22. Blackmond, D. G., Ropic, M. & Stefinovic, M. Kinetic studies of the asymmetric
transfer hydrogenation of imines with formic acid catalyzed by Rh-diamine
catalysts. Org. Process Res. Dev. 10, 457–463 (2006).
23. Denmark, S. E. & Burk, M. T. Lewis base catalysis of bromo- and
iodolactonization, and cycloetherification. Proc. Natl Acad. Sci. USA 107,
20655–20660 (2010).
24. French, D. C. & Crumrine, D. S. Improved correlation of 33S chemical shifts with
pKas of arenesulfonic acids: use of 33S NMR for pKa determination. J. Org. Chem.
55, 5494–5496 (1990).
Detailed experimental procedures, syntheses of the substrates and catalysts,
characterization data for all the new compounds, procedures and data for
mechanistic investigations (including reaction-progress kinetic analysis, linear free-
energy relationship studies with Mayr’s reactivity parameters, kinetic isotope-effect
studies, model binding studies by 1H NMR spectroscopy and other experimental
kinetic studies) are given in the Supplementary Information. Crystallographic
information for compounds 2b, 2g and 2q are deposited at the Cambridge
Crystallographic Data Centre and allocated the deposition numbers CCDC 862750,
CCDC 862751 and CCDC 862752, respectively).
25. Phan, T. B., Breugst, M. & Mayr, H. Towards a general scale of nucleophilicity?
Angew. Chem. Int. Ed. 45, 3869–3874 (2006).
26. Maresh, J. J. et al. Strictosidine synthase mechanism of a Pictet–Spengler
catalyzing enzyme. J. Am. Chem. Soc. 130, 710–723 (2008).
27. Bosnich, B. Asymmetric Catalysis 14 (Marinus Nijhoff, 1986).
28. Kelly, T. R. & Kim, M. H. Relative binding affinity of carboxylate and its isosteres:
nitro, phosphate, phosphonate, sulfonate, and d-lactone. J. Am. Chem. Soc.
116, 7072–7080 (1994).
29. Xu, H., Zuend, S. J., Woll, M. G., Tao, Y. & Jacobsen, E. N. Asymmetric
cooperative catalysis of strong Brønsted acid-promoted reactions using chiral
ureas. Science 327, 986–990 (2010).
Received 19 January 2012; accepted 31 July 2012;
published online 16 September 2012
30. Scheerder, J., Engbersen, J. F. J., Casnati, A., Ungaro, R. & Reinhoudt, D. N.
Complexation of halide anions and tricarboxylate anions by neutral urea-
derivatized p-tert-butylcalix[6]arenes. J. Org. Chem. 60, 6448–6454 (1995).
31. Lan, T. & McLaughlin, L. W. The energetic contribution of a bifurcated
hydrogen bond to the binding of DAPI to dA-dT rich sequences of DNA.
J. Am. Chem. Soc. 123, 2064–2065 (2001).
References
1. Zhang, Z. G. & Schreiner, P. R. (Thio)urea organocatalysis – what can be learnt
from anion recognition? Chem. Soc. Rev. 38, 1187–1198 (2009).
2. Reisman, S. E., Doyle, A. G. & Jacobsen, E. N. Enantioselective thiourea-
catalyzed additions to oxocarbenium ions. J. Am. Chem. Soc. 130,
7198–7199 (2008).
...
...
32. Braun, J., Neusser, H. J. & Hobza, P. N–H p interactions in indole benzene-
...
h6,d6 and indole benzene-h6,d6 radical cation complexes. Mass analyzed
threshold ionization experiments and correlated ab initio quantum chemical
calculations. J. Phys. Chem. A 107, 3918–3924 (2003).
33. Biswal, H. S. & Wategaonkar, S. Nature of the N–H S hydrogen bond. J. Phys.
3. Brown, A. R., Kuo, W-H. & Jacobsen, E. N. Enantioselective catalytic a-alkylation
of aldehydes via an SN1 pathway. J. Am. Chem. Soc. 132, 9286–9288 (2010).
4. De, C. K., Klauber, E. G. & Seidel, D. Merging nucleophilic and hydrogen
bonding catalysis: an anion binding approach to the kinetic resolution of amines.
J. Am. Chem. Soc. 131, 17060–17061 (2009).
5. Knowles, R. R. & Jacobsen, E. N. Attractive noncovalent interactions in
asymmetric catalysis: links between enzymes and small molecule catalysts.
Proc. Natl Acad. Sci. USA 107, 20678–20685 (2010).
6. Knowles, R. R., Lin, S. & Jacobsen, E. N. Enantioselective thiourea-catalyzed
cationic polycyclizations. J. Am. Chem. Soc. 132, 5030–5032 (2010).
7. Zuend, S. J. & Jacobsen, E. N. Mechanism of amido-thiourea catalyzed
enantioselective imine hydrocyanation: transition state stabilization via multiple
non-covalent interactions. J. Am. Chem. Soc. 131, 15358–15374 (2009).
8. Fox, D. J., House, D. & Warren, S. Mechanisms of sulfanyl (RS) migrations:
synthesis of heterocycles. Angew. Chem. Int. Ed. 41, 2462–2482 (2002).
...
Chem. A 113, 12763–12773 (2009).
34. Herrera, R. P., Sgarzani, V., Bernardi, L. & Ricci, A. Catalytic enantioselective
Friedel–Crafts alkylation of indoles with nitroalkenes by using a simple thiourea
organocatalyst. Angew. Chem. Int. Ed. 44, 6576–6579 (2005).
35. James, W. H. III et. al. Evolution of amide stacking in larger g-peptides: triamide
H-bonded cycles. J. Phys. Chem. A 115, 13783–13798 (2011).
36. Gustafson, J., Lim, D. & Miller, S. J. Dynamic kinetic resolution of biaryl
atropisomers via peptide-catalyzed asymmetric bromination. Science 328,
1251–1255 (2010).
37. Shepodd, T. J., Petti, M. A. & Dougherty, D. A. Tight, oriented binding of an
aliphatic guest by a new class of water-soluble molecules with hydrophobic
binding sites. J. Am. Chem. Soc. 108, 6085–6087 (1986).
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
823
© 2012 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1450
ARTICLES
38. Kearney, P. C. et al. Molecular recognition in aqueous media. New binding
studies provide further insights into the cation-p interaction and related
phenomena. J. Am. Chem. Soc. 115, 9907–9919 (1993).
46. Ngola, S. M. & Dougherty, D. A. Evidence for the importance of polarizability
in biomimetic catalysis involving cyclophane receptors. J. Org. Chem. 61,
4355–4360 (1996).
39. Ma, J. C. & Dougherty, D. A. The cation-p interaction. Chem. Rev. 97,
1303–1324 (1997).
40. Dougherty, D. A. Cation-p interactions in chemistry and biology: a new view of
benzene, Phe, Tyr, and Trp. Science 271, 163–168 (1996).
41. Raju, R. K., Bloom, J. W. G., An, Y. & Wheeler, S. E. Substituent effects on
non-covalent interactions with aromatic rings: insights from computational
chemistry. ChemPhysChem 12, 3116–3130 (2011).
42. Xiu, X., Puskar, N. L., Shanata, J. A. P., Lester, H. A. & Dougherty, D. A. Nicotine
binding to brain receptors requires a strong cation-p interaction. Nature 458,
534–538 (2009).
Acknowledgements
This work was supported by the National Institute of General Medical Sciences (NIGMS)
(P50 GM-69721 and RO1 GM-43214) and by a predoctoral fellowship to S.L. from Eli Lilly.
We thank R. Knowles, K. Brak and H. Mayr for helpful discussions, A. Brown, D. Lehnherr
and A. Hyde for the use of catalysts, and S-L. Zheng for crystal structure determinations.
Author contributions
S.L. conducted the experiments, S.L. and E.N.J. co-wrote the manuscript and E.N.J. guided
the research.
43. Vijay, D. & Sastry, G. N. Exploring the size dependence of cyclic and acyclic
p-systems on cation-p binding. Phys. Chem. Chem. Phys. 10, 582–590 (2008).
44. Gal, J-F. et al. Lithium-cation/p complexes of aromatic systems. The effect
of increasing the number of fused rings. J. Am. Chem. Soc. 125,
10394–10401 (2003).
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permission information is available online at http://
www.nature.com/reprints. Correspondence and requests for materials should be addressed
to E.N.J.
45. Li, X., Liu, P., Houk, K. N. & Birman, V. B. Origin of enantioselectivity in CF3–
PIP-catalyzed kinetic resolution of secondary benzylic alcohols. J. Am. Chem.
Soc. 130, 13836–13837 (2008).
Competing financial interests
The authors declare no competing financial interests.
824
NATURE CHEMISTRY | VOL 4 | OCTOBER 2012 | www.nature.com/naturechemistry
© 2012 Macmillan Publishers Limited. All rights reserved.