Asymmetric ion-pairing catalysis of the reversible cyclization of 2'-hydroxychalcone to flavanone: Asymmetric catalysis of an equilibrating reaction

Article abstract of DOI:10.1002/ejoc.201200838

The asymmetric catalytic cyclization of the simple 2'-hydroxychalcone (1) to flavanone (2), a model for the chalcone isomerase reaction, has been realized as a catalytic asymmetric ion-pairing process with chiral quaternary ammonium salts (e.g., 9-anthracenylmethlycinchoninium chloride; 9-Am-CN-Cl) and NaH as small-molecule co-catalyst. In toluene/CHCl₃ solution, the process reaches an intrinsic enantioselectivity of up to S = 14.4 (er = 93.5:6.5). The reversible reaction proceeds in two steps: A fast initial reaction approaches a quasi-equilibrium with K_R/S = 4.5, followed by a second, slow racemization phase approaching K_rac = 9. A simple mechanistic model featuring a living ion-pairing catalysis with full reversibility is proposed. Deuterium transfer from co-solvent CDCl₃ to product 2 and isolation of a Michael conjugate formed from 2 and 1 demonstrate the intermediacy of flavanone enolate ion pairs. A kinetic model shows good agreement with the experimentally observed, peculiar, time-dependent evolution of the species concentrations and the enantiomeric excess of 2. The reaction is a chemical model of the chalcone isomerase enzymatic reaction. Furthermore, it is an ideal model for studying the characteristic behavior of reversible asymmetric catalyses close to their equilibria.

Full text of DOI:10.1002/ejoc.201200838

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
FULL PAPER
DOI: 10.1002/ejoc.201200838
1
Asymmetric Ion-Pairing Catalysis of the Reversible Cyclization of
2Ј-Hydroxychalcone to Flavanone: Asymmetric Catalysis of an Equilibrating
Reaction[‡]
Lukas Hintermann*[a][‡‡] and Claudia Dittmer[a][‡‡]
Keywords: Asymmetric catalysis / Enzyme models / Natural products / Reaction mechanisms / Oxygen heterocycles
6
The asymmetric catalytic cyclization of the simple 2Ј-hy-
droxychalcone (1) to flavanone (2), a model for the chalcone
simple mechanistic model featuring a living ion-pairing ca-
talysis with full reversibility is proposed. Deuterium transfer
isomerase reaction, has been realized as a catalytic asymmet- from co-solvent CDCl3 to product 2 and isolation of a Michael
ric ion-pairing process with chiral quaternary ammonium
salts (e.g., 9-anthracenylmethlycinchoninium chloride; 9-
Am-CN-Cl) and NaH as small-molecule co-catalyst. In tolu-
conjugate formed from 2 and 1 demonstrate the intermediacy
of flavanone enolate ion pairs. A kinetic model shows good
agreement with the experimentally observed, peculiar, time-
21
26
11
16
ene/CHCl3 solution, the process reaches an intrinsic enantio- dependent evolution of the species concentrations and the
selectivity of up to S = 14.4 (er = 93.5:6.5). The reversible
reaction proceeds in two steps: A fast initial reaction ap-
proaches a quasi-equilibrium with KR/S = 4.5, followed by a
second, slow racemization phase approaching Krac = 9. A
enantiomeric excess of 2. The reaction is a chemical model
of the chalcone isomerase enzymatic reaction. Furthermore,
it is an ideal model for studying the characteristic behavior
of reversible asymmetric catalyses close to their equilibria.
not the simple substrate 1. Scheidt and co-workers demon-
strated that the introduction of an activating 2-alkoxycarb-
onyl group at C-2 of the chalcone substrate facilitates asym-
metric ring closure with chiral thiourea catalysts.[9] The
same model reaction has since been catalyzed by others
using chiral Lewis acids,[10] other bifunctional organocata-
lysts,[11] or chiral Brønsted acids,[12] providing indirect solu-
tions to the asymmetric synthesis of flavanones.[6a] The di-
rect asymmetric catalytic cyclization 1Ǟ2 has been at-
tempted without success in the past.[8,12–14] Several authors
have acknowledged the difficulty of realizing such a process,
Introduction
The reversible cyclization of 2Ј-hydroxychalcones to flav-
anones is an early step in flavonoid biosynthesis, which is
catalyzed by the enzyme chalcone isomerase (Scheme 1,
a).[2] Catalysis is effected by folding the chalconate anion
into a chiral conformation in the enzyme active site.[3] Ad-
ditional general acid and base interactions with amino acid
side-chains provide accurate structural organization of the
transition state of the intramolecular asymmetric conjugate
addition of phenolate (oxa-Michael reaction),[3,4] ensuring
high enantioselectivity (S ≈ 100000).[5] In spite of the revers-
ibility of the reaction, (2S)-flavanones are thus formed al-
most exclusively. Such high enantioselectivity cannot be
achieved with small-molecule catalysts (organocatalysts),
which have only recently been used successfully in asymmet-
ric oxa-Michael-type reactions[6,7] but not yet with any suc-
cess in the simple model reaction 1Ǟ2. We have previously
studied asymmetric catalytic cyclization reactions of hy-
droxychalcones with cinchona alkaloid catalysts[8] and were
able to cyclize specifically activated starting materials but
51
56
31
36
41
46
[‡] See ref.[1]
[a] Department Chemie, Technische Universität München,
Lichtenbergstr. 4, 85748 Garching bei München, Germany
Fax: +49-89-28913669
E-mail: lukas.hintermann@tum.de
Homepage: www.oca.ch.tum.de
[‡‡]Taken in part from the habilitation thesis of L. H. (RWTH
Aachen, Germany, 2008) and the thesis (Erste Staatsprüfung)
of C. D. (RWTH Aachen, Germany, 2004).
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201200838.
Scheme 1. a) Enzymatic cyclization of hydroxylated chalcones to
flavanones. b) Attempted asymmetric catalytic cyclization of 2Ј-hy-
droxychalcones to flavanones.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
L. Hintermann, C. Dittmer
FULL PAPER
ascribing it to its reversibility or the low reactivity of Table 1. Screening for asymmetric catalysts for the cyclization of 1
to 2.[a]
1.[6a,7,9,12,14–16] The topic of asymmetric catalysis of reac-
tions with a low thermodynamic driving force close to their
equilibrium is a problem of general interest that has been
little studied.[17] We now present a novel asymmetric ion-
pairing catalysis that achieves the difficult asymmetric cycli-
zation of 1 to 2 in a manner that is mechanistically compar-
able to the enzymatic reaction (Scheme 1, a). The reversibil-
ity of the reaction has specific consequences for the kinetic
evolution of the reaction mixture composition and the en-
antiomeric excess of flavanone (2). A mechanistic and ki-
netic model is presented that rationalizes the experimentally
determined reaction characteristics.
61
66
71
Entry Catalyst[b] (loading [mol-%])
t [h]
2 [%][c] ee [%][c]
1
2
3
4
5
6
Bn-CD-Cl/K2CO3 (10)
Bn-CD-CO3 (10)
28
28
6.5
52
9
65
41
54
19 (S)
21 (S)
6 (S)
Bn-CD-CO3 (10)[d]
9-Am-CD-Cl/K2CO3 (10/20)
9-Am-CD-CO3 (10)
9-Am-CD-Cl/NaH (10)
28
53 (S)
55 (S)
47 (S)
14
Results and Discussion
80
63[e]
Our earlier study indicated that the difficulty of catalyz-
ing the reaction 1Ǟ2 asymmetrically is connected to its
low driving force and reversibility (K ≈ 4 in nBuOH at
120 °C),[8] which creates the inherent risk of racemization
under the reaction conditions. The screening of a variety of
metal complexes and bifunctional organocatalysts, includ-
ing cinchona alkaloids, was not very successful (see
Table S1 in the Supporting Information for details);[18] only
(l)-proline displayed some activity with low enantio-
selectivity in the cyclization of 1 to 2 (Scheme 1, b) and
amino acid decarboxylation was a notable side-reaction.[19]
On the other hand, strongly basic catalysts like KOtBu
readily induced cyclization at ambient temperature. This led
to the idea of combining the successful chiral architecture
of cinchona alkaloids[20] with the potentially higher basicity
of inorganic bases in a two-component catalyst. When a
solution of hydroxychalcone 1 in toluene was stirred with
solid K2CO3 and the chiral phase-transfer catalyst N-benz-
ylcinchonidinium chloride,[21] an asymmetric catalytic cycli-
zation was indeed observed (Table 1, entry 1).
[a] Conditions: 1 (0.3 mmol), toluene (2 mL), catalyst loading
(10 mol-%, as Q+), ambient temperature. [b] 9-Am = 9-an-
thrylmethyl; Bn = benzyl. CD = cinchonidine/cinchonidinium. [c]
Yield and ee of 2 determined by HPLC analysis. [d] Reaction in
CH2Cl2. [e] Isolated yield of 2. In addition, 1 (26%) and 3 (12%)
were also obtained.
76
81
86
91
96
catalysis mixtures displayed the characteristic dark-orange-
red color of the chalconate anion (for a photograph, see
Figure S2), which points to the presence of an ion pair [Q+/ 116
ChO–] (ChOH = 1) as the resting state of the catalysis. This
lipophilic ion pair is produced according to Equations (1)
and (2), the latter being an ion-exchange reaction driven
forward by the precipitation of the stable NaCl(s).
NaH + ChOH h ChONa + H2(g)
(1)
QCl + ChONa h Q+/ChO– + NaCl(s)
(2)
Alternatively, the catalyst components were combined to
a quaternary ammonium carbonate salt, which was ob-
tained from the chloride salt and silver carbonate, and used
as homogeneous catalysts (Table 1, entries 2 and 3). In
dichloromethane, the catalytic cyclization was faster but less
stereoselective (entry 3; see Table S2 in the Supporting In-
formation for more solvents). The superiority of N-(9-an-
thracenyl)methyl- over N-benzyl-quaternized cinchona al-
101 kaloids as chiral catalysts[22] was confirmed in this reaction
(entries 4–6). In view of the reversibility of the reaction,[23]
a catalytic experiment with 9-Am-CD-CO3 (Table 1, en-
try 5) was followed over time by removing samples for chi-
ral HPLC analysis (Figure 1). Both conversion (38%) and
106 ee (49%) values stabilized after 80 h. This reaction course
is not plausibly explained by thermodynamic or kinetic ar-
guments, but must be due to catalyst deactivation in air, a
conclusion justified by later experiments. The most conve-
nient catalyst system that emerged from the initial screening
Figure 1. Conversion of 1 and ee of 2 vs. time in the cyclization of
1 to 2 with 9-Am-CD-CO3 as catalyst (Table 1, entry 5); see Table 1
for conditions.
The practical and robust catalyst system composed of 2– 121
111 was a combination of the quaternary salt QCl with co-cata- 10 mol-% (typically 5 mol-%) of 9-anthrylmethylcinchonid-
lytic sodium hydride as base (entry 6); this resulted in a inium chloride (9-Am-CD-Cl) and co-catalytic NaH (5–
more active catalyst than the ill-defined carbonate salts. The 20 mol-%) was preferred in further studies. The base NaH
2
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
Asymmetric Catalysis of an Equilibrating Reaction
can be used in slight excess over QCl because the sodium
126 salt NaOCh is not soluble in the reaction medium and in-
duces no background reaction. Catalytic reactions were per-
formed under argon because this increased the catalyst life-
time. A fairly rapid cyclization of 1 to 2 ensued at 45 °C;
HPLC analysis of the reaction course revealed the charac-
131 teristic time-dependency of the conversion (Figure 2, a) and
the enantiomeric excess of the product (Figure 2, b).
Scheme 2. a) Generation of flavanone–chalcone conjugate 3 as a
side-product in the base-catalyzed condensation of 1. b) Asymmet-
ric catalytic synthesis of 3 from 1. The assignment of the absolute
configuration of 3 is based on that of 2 in the same reaction.
tive catalyst. The kinetic progress of the reaction with the
catalyst 9-Am-CN-Cl/NaH (9 mol-%) at 45 °C is shown in 151
Figure 3.
Figure 2. Course of the catalytic reaction 1Ǟ2 at 45 °C. Condi-
tions: catalyst 9-Am-CD-Cl/NaH (10%), toluene (11 mL/mmol).
Curves are interpolated to the measured values. a) Time-dependent
composition of the reaction mixture. The data for 2 and 3 represent
the sum of both enantiomers. Note the logarithmic timescale.
b) Evolution of the enantiomeric excess of flavanone (2) as a func-
tion of the conversion of 1.
Figure 3. Kinetics and time-dependent enantioselectivity of chalc-
one/flavanone dimer 3 generation from 2-hydroxychalcone (1). Ca-
talysis conditions: 9-Am-CN-Cl (9 mol-%), NaH (10 mol-%), tolu-
ene (4.5 mL/mmol), 45 °C. Internal standard for HPLC: 1,6-di-
bromo-2-isopropoxynaphthalene. Flavanone ([2] = [(S)-2] + [(R)-
2]) was completely racemized after 8 h. The inverted ee (S Ͼ R) of
2 after 10 h might be due to a kinetic resolution or a systematic
integration error.
The reaction was fast in the initial phase and reached a
quasi-equilibrium situation with Keq ≈ 4 (80% of 2, 20% of
1). The enantiomeric excess of product 2 remained relatively
136 stable up to this point, but then decreased in a slower race-
mization phase, and eventually was reduced to less than
10%. In this second phase, a side-product 3 was also
formed in considerable amounts; compound 3 is a conju-
gate of 1 and 2 and must result from the conjugate addition
The initial fast cyclization of chalcone to flavanone
141 (Michael reaction) of flavanone enolate to substrate 1 quickly reached quasi-equilibrium (≈70 mass-% 2). In the
(Scheme 2, a). second phase of the reaction, 2 was consumed at the ex-
The structural assignment of 3, a single diastereomer, pense of conjugate 3 and the remaining 2 had racemized 156
was particularly simple as it had earlier been obtained in after 7 h. Maximal conversion to 3 (50 mass-%) was
the base-induced dimerization of 1 under microwave heat- reached after 1 d. A preparative run was performed with
146 ing and had been fully characterized (NMR, X-ray).[14] In only 2.4 mol-% of 9-Am-CN-Cl at a higher substrate con-
our catalysis reaction, 3 was enantiomerically enriched. The centration (1.6 mL/mmol). After 40 h at 45 °C, conversion
direct asymmetric catalytic synthesis from 1 was improved to 3 had reached 49% (ee = 67%) and (–)-3 was isolated in 161
by using a higher substrate concentration and a more selec- 38% yield with 62.5% ee (Scheme 2, b).
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
L. Hintermann, C. Dittmer
FULL PAPER
Reaction Analysis
ted from Figure 4, b). When the catalytic reaction was per-
formed in a 3:1 mixture (v/v) of toluene and chloroform
(Figure 4, c), the initial enantioselectivity reached S/R = 5.0 201
(66.5% ee), which is higher than in either toluene (S/R =
3.9; ee = 59%) or CHCl3 (S/R = 2.3; ee = 39%) alone. The
reaction also proceeded faster in the mixed solvent than in
either of the pure solvents.
Generation of the side-product 3, although problematic 206
at high conversions, is suppressed to 2–3% at the onset of
the quasi-equilibrium after 4 h (conversion ≈80%). Flav-
anone enolates as fairly basic species (pKa = 17)[25,26] are
presumed reaction intermediates; the suppression of 3 in
chloroform could be due to the acidity of the solvent (pKa 211
= 13.6).[27] To test this hypothesis, an additional experiment
was performed in toluene/CDCl3 (3:1; Table 2). Samples
were removed and quenched after certain intervals of time,
and the incorporation of deuterium into the flavanone
product was analyzed by differential integration of 1H 216
NMR signals (accuracy: Ϯ2%). Table 2 shows that deute-
rium is indeed transferred from CDCl3 to flavanone (2),
and that this process becomes more important in the later
stages of the reaction, after the quasi-equilibrium has been
The composition of the catalytic reaction mixture was
analyzed by normal-phase chiral HPLC.[24] Chromato-
166 grams are shown in the Supporting Information (Fig-
ure S1). The signals for chalcone 1 and flavanone enantio-
mers (R)-2 and (S)-2 are clearly separated, as are those of
the enantiomers of conjugate 3, the broadened peaks of
which are better visible in the later stages of the reaction.
171 The concentrations of all reacting species could be deter-
mined from a single chromatogram. Quantification was
either performed by the summation of absorption-corrected
peak areas[18] and normalization to 100% or, for more accu-
rate work, by the use of an internal standard.
176 Suppression of Side-Product 3 in Mixed Solvent Systems
The initial screening of solvents with the 9-Am-CD-Cl/
NaH precatalyst had used a single-point conversion and ee
analysis after 24 h. The results were not overly informative
because the ee depends on total conversion as much as on
181 the solvent used. Nevertheless, the screening showed that
the generation of side-product 3 also depends on the sol-
vent. Arenes produced as much as 10% of 3 after 24 h, but
the C–H acidic solvents chloroform or propionitrile sup-
pressed its generation to only 0.1 or 1%, respectively, in
186 spite of higher overall conversions (see Table S3 in the Sup-
porting Information). This observation provoked a closer
inspection of the time-dependent reaction course in selected
solvents. The catalysis in toluene (Figure 4, a) reached
maximum conversion to flavanone in 9 h. The ee of flav-
191 anone (2) was stable at 58% over the first 6 h, but then
decreased after the reaction had reached quasi-equilibrium
(Figure 4, a). Steady generation of side-product 3 occurred
throughout the reaction. The time-dependent analysis of
the reaction in chloroform (Figure 4, b) showed distinct dif-
196 ferences: Although the initial level of enantioselectivity was
lower (39% ee), generation of side-product 3 was almost
Table 2. Incorporation of deuterium from co-solvent CDCl3 into
flavanone.[a]
Entry Time [h]
2 [%][b]
ee [%][b]
Da [%][c] Db [%][c]
1
2
3
1
4
6.5
44
82
83
77 (R)
64 (R)
50 (R)
1
32
49
1
32
49
[a] Reaction conditions: 1 (100 mg), 9-Am-QD-Cl (10 mol-%),
NaH (10 mol-%), toluene (3 mL), CDCl3 (1 mL), room temp. [b]
Mol-% of 2, determined by HPLC. [c] Deuterium incorporation
1
completely suppressed in this solvent (Ͻ0.5%; curves omit- determined by H NMR spectroscopy.
Figure 4. Evolution of species in the catalytic cyclization of 1 to 2 in various solvents. General conditions: 1 (100 mg, 0.45 mmol), 9-Am-
CD-Cl/NaH (10%), room temp. a) Reaction performed in toluene (6 mL). b) Reaction performed in chloroform (5 mL). c) Reaction
performed in toluene/CHCl3 (3:1; 5 mL). See Tables S4–S6 in the Supporting Information for numerical values. Kinetic curves are fitted
to a model discussed later in the text and are intended as guidelines; high accuracy cannot be expected under the conditions used.
4
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
Asymmetric Catalysis of an Equilibrating Reaction
221 reached and the concentration of the acidic starting mate- vent, the quenching of B by acidic substrate 1 by proton
rial 1 (pKa ≈ 9.6)[28] has reached low levels. There was no transfer competes with intermolecular conjugate addition
site preference (Ha vs. Hb) for the deuteriation. This is in of B to either 1 or A to give 3. This irreversible side-reaction 256
line with the diastereoselectivity of deuterium incorporation is suppressed in buffered solvent containing CHCl3, in
observed in base-catalyzed cyclizations of 1 to 2 via flav- which the catalyst resting state is Q+/CCl3– (C) and the con-
226 anone enolates in protic, deuteriated solvents.[29]
centrations of A and B remain low. The catalytic principle
operating in the cyclization of 1 to 2 may be defined as
living[30] anionic catalysis. Each cyclization event AǞB 261
generates a strong base (enolate; pKa = 17) from a weak
base (phenolate; pKa = 9.6); hence the acidic substrate 1 is
not deprotonated by a chiral base catalyst, but by the conju-
gate base of the reaction product, and the chirality is associ-
ated with the counter-cation, just as in asymmetric phase- 266
transfer catalysis (APTC). In contrast to APTC, the current
reaction proceeds in homogeneous solution, and in contrast
to most APTCs, a stoichiometric quantity of base is not
required, only a co-catalytic amount.
Mechanistic Interpretation
The experimental data acquired so far give some hints
about the reaction mechanism of the catalytic cyclization of
2Ј-hydroxychalcone (1) to flavanone (2) with the QCl/NaH
231 precatalyst (Scheme 3). The acidic substrate 1 (pKa = 9.6)
[28]
is deprotonated by NaH to give insoluble ChONa and
hydrogen [Equation (1)]; ion exchange with Q+/Cl– [Equa-
tion (2)] gives the dissolved dark-orange ion pair A (for a
photograph of the reaction mixture, see Figure S2 in the
236 Supporting Information). The chalconate anion undergoes
intramolecular conjugate addition (oxa-Michael cycliza-
tion) in the presence of the chiral counter-cation Q+, the
latter inducing stereoselectivity in the attack of the phenol-
ate oxygen to the diastereotopic faces of the alkenone
241 double bond. In the resulting ion pair B, cation Q+ is com-
bined with a strongly basic flavanone enolate (pKa = 17),
which can deprotonate another molecule of 1 with regener-
ation of ion pair A and release of product 2. Alternatively,
B is protonated to product 2 by the acidic co-solvent chlo-
Kinetic Modeling of the Reversible Asymmetric Catalysis
271
The mechanistic model developed in the preceding sec-
tion can be translated into a kinetic reaction scheme. As a
test of the accuracy of the scheme and the underlying pos-
tulated mechanism, it remains to be shown if such a kinetic
scheme will correctly reproduce the peculiar time-depen- 276
dency of the reactant and product concentrations and of
the enantiomeric excess of the product (compare Figure 2).
The simplified kinetic model is based on Equations (3) and
(4), in which kR and kS are the pseudo-first-order rate con-
stants of the forward reactions and include the catalyst con- 281
centration, and k–R and k–S are the pseudo-first-order rate
constants of the backward reactions.
–
246 roform (pKa = 13.6),[27] creating the ion pair Q+/CCl3 (C);
if CDCl3 is the co-solvent, this process introduces deute-
rium at C-3 of 2 (see Table 2).
1 h (R)-2; Keq = KR = kR/k–R
(3)
1 h (S)-2; Keq = KS = kS/k–S
(4)
The enantiomers of 2 have equal free energy in a largely
achiral environment, translating to the additional restraint
shown in Equation (5).
286
KR = KS
(5)
There are two separate equilibria between the starting
material 1 and either enantiomer of 2, with equal equilib-
rium constants but different equilibration kinetics in the
case of asymmetric catalysis. For the kinetic fit, we per-
formed a catalytic reaction with 4.3 mol-% of the 9-Am- 291
CN-Cl/NaH catalyst in toluene/CHCl3 (3:1) at room tem-
perature. The reaction proceeded reversibly over a reason-
ably long timescale, and all the reaction components could
be quantified by a single chiral HPLC measurement against
an internal standard. The experimental values and fitted 296
curves are shown in Figure 5. The kinetic model correctly
Scheme 3. Mechanism of the QCl/NaH-catalyzed cyclization of hy-
droxychalcones to flavanones: An asymmetric ion-pairing catalysis.
Because ion pair C is stable towards the elimination of reproduces the time-dependent evolution of the reaction
Cl– in apolar media (no follow-up products from CCl2 have mixture (Figure 5) and the ee (Figure 6) of product 2. Some
251 been observed in our study), it must be the resting state of simplifications of the kinetic model need to be pointed out,
the catalyst that reinitiates the reaction by deprotonating 1 namely, concentrations of catalyst resting states (A, B, or C 301
to recreate ion pair A and CHCl3. In pure toluene as sol- in Scheme 3) are not accounted for and a steady catalyst
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
L. Hintermann, C. Dittmer
FULL PAPER
decomposition that eventually stops the reaction is not in-
cluded in the model.
A perfect fit of both the early and late reaction phase
with the same parameters cannot be expected. In view of 306
this, the good agreement of experimental datapoints and
modeled curves for 1, (R)-2, and (S)-2 with only three inde-
pendent parameters is satisfactory and supports the pro-
posed mechanism. Within the simplified kinetic model, the
irreversible side-reaction leading to 3 is satisfactorily de- 311
scribed by two additional kinetic parameters, Equations (6)
and (7).
1 + (R)-2 Ǟ (R)-3; kR3
(6)
1 + (S)-2 Ǟ (S)-3; kS3
(7)
In conclusion, three parameters are sufficient to charac-
terize this reversible asymmetric catalysis 1h2: 1) The in-
trinsic enantioselectivity S = kR/kS of the catalyst, which 316
defines the initial product ee, 2) the initial reaction rate,
which can be expressed as –d[1]/dt = (kR + kS)[1] and is
also dependent on the catalyst, and 3) the equilibrium con-
stant Krac (= KR + KS = 2KR), which is independent of the
Figure 5. Asymmetric catalytic cyclization of 1 to 2 showing experi-
mental datapoints and fitted curves. Conditions: 1 (2.23 mmol), 9-
AmCN-Cl/NaH (4.3/4.9 mol-%), solvent mixture toluene/CHCl3
(3:1; 20 mL), internal standard 1,6-dibromo-2-isopropoxynaph-
thalene (300 mg), ambient temperature (295 K). KR = KS = 4.5;
Krac = 9.0; S = 10.1; ktot = kR + kS = 0.33 h–1; kR3 = kS3
=
catalyst.
321
0.10 Lmol–1h–1.
Reversibility and Time-Dependent Evolution of the
Enantiomeric Excess
The reversibility of the reaction renders enantio-
selectivity a function of time or conversion. The plot of
product ee versus conversion has a characteristic shape 326
(Figure 6). In the initial reaction phase, the backward reac-
tion is negligible and the product ee reflects the intrinsic
selectivity (S = kR/kS) of the chiral catalyst. Only after the
conversion has reached quasi-equilibrium (ca. 80% 2 and
20% 1 for KR = 4), the product ee decreases quickly and 331
reaches 0% at the equilibrium conversion value (88.9% for
Krac = 8). With very selective catalysts it is possible to ap-
proach quasi-equilibrium concentrations of the product
with little loss of product ee (Figure 6, b, curve S = 100).
The modeling ignores the generation of side-product 3, 336
which is responsible for the deviations at the higher conver-
sions seen in Figure 6 (a).
Effects of Catalyst and Substrate Structure on
Enantioselectivity
A specific catalyst in the chalcone/flavanone cyclization 341
is fully characterized by the initial enantioselectivity and
reaction rate. Single-point ee determinations of isolated re-
action products, as is common analytical practice in
irreversible asymmetric catalysis, are not suitable in revers-
ible systems. In practice, three to five samples were removed 346
from a single catalytic run in the initial reaction phase (10–
60% conversion) for analysis by chiral HPLC. Use of an
internal standard for HPLC was not necessary because
side-products were absent in the initial reaction phase.
Quantification of 1 and 2 (in mol-%) was possible from the 351
Figure 6. Conversion dependency of product ee in reversible asym-
metric catalysis. a) Datapoints for catalysis with 9-Am-CN-Cl/NaH
(see Figure 5) are compared with a modeled curve using a simpli-
fied kinetic model that neglects the generation of 3. The data for
9-Am-CD-Cl from Figure 2 are included for comparison. b) Calcu-
lated ee vs. conversion profiles for the reversible asymmetric cataly-
sis 1 = (S)-2/(R)-2 with KR/S = 4 (Krac = 8) at different levels of
catalyst enantioselectivity S.
6
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
Asymmetric Catalysis of an Equilibrating Reaction
relative HPLC peak areas corrected for UV absorption fac- Influence of the Alkaloid Structure
381
tors. The reaction rate is expressed in simplified form as the
The N-(9-anthrylmethyl) salts of all major cinchona al-
kaloids were tested in the reaction (Table 4). Cinchonine
turnover frequency (TOF [h–1]) and was calculated from
early datapoints according to Equation (8).
Table 4. Variation of the catalyst in the asymmetric cyclization of
1Ǟ2.[a]
(8)
356
The relative molar fraction of the catalyst corresponds to
the loading in mol-%. Because the TOF of a homogeneous
catalysis depends on the initial reactant concentration, the
latter was kept constant in the catalyst screening experi-
ments.
361 Influence of the N-Alkyl Group in the Quaternized Alkaloid
The effect of the N-alkyl group of the quaternized cin-
chonidine on the catalytic reaction is shown in Table 3.
These experiments from an early stage of the project were
performed in toluene. Enantioselectivities are somewhat
366 lower than in the optimized solvent systems toluene/CHCl3
or cymene/CHCl3. Starting from N-benzylcinchonidinium
chloride (Table 3, entry 1), an increase in the size of the N-
alkyl group to 1-naphthylmethyl and 9-anthracenylmethyl
(9-Am) increased both the selectivity and activity of the re-
371 action (entries 2 and 3). However, bulkiness is not the only
factor in play because groups like 9-fluorenyl or a silyloxy-
substituted benzyl displayed lower selectivity with inversion
of induction sense (entries 4 and 5). The extended aromatic
π system of the 9-Am substituent might be required for
376 stacking interactions in the deprotonated substrate.
Table 3. Effect of quinuclidine N-substitution on catalytic perform-
ance.[a]
Entry
Quaternary ammonium salt S[b] ee[b]
TOF[b]
[h–1]
(QX)
R
[%]
R1
X
1
2
Bn
H
H
H
Cl 1.7 26 (S)
Cl 2.2 37 (S)
0.3
0.8
1-naphthylmethyl
9-anthracenylme-
thyl
9-fluorenyl
3,5-bis(TBPSO)
benzyl
Cl
3
4
5
6
3.9 59 (S)
1.0
0.06
0.4
H
H
Br 1.6 23 (R)
Br
1.4 18 (R)
Bn
Bn Br 1.4 16 (S)
0.05
[a] Conditions: cf. scheme, and 1 (0.45 mmol), toluene (3 mL).
TBPS = tert-butyldiphenylsilyl. [b] Determined by HPLC.
The presence of the alcoholic hydroxy group is also criti-
cal for achieving selective catalysis: An experiment with O-
benzylated N-benzylcinchonidinium bromide[31] displayed
both low selectivity and activity (entry 6).
[a] Conditions: 1 (0.45 mmol), QCl (5 mol-%), NaH (6 mol-%), tol-
uene (3 mL), CHCl3 (1 mL). [b] Cymene (3 mL) was used in place
9
of toluene. [c] Anthryl = 9-anthracenyl.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
L. Hintermann, C. Dittmer
FULL PAPER
(CN) produced the most selective catalyst (entries 1 and 2), Am-CD-Cl/NaH precatalyst (Scheme 4, b). This result was
which was also more active than the pseudo-enantiomeric contrary to our expectation: Generation of the ion pair A
386 cinchonidine (CD) derivative (entries 3 and 4). Variations and thus initiation of the catalysis was predicted as 2,4-di- 426
in the cinchonine structure, such as hydrogenation, isomer- tert-butylphenolate (pKa = 11.6)[34] is sufficiently basic to
ization,[32] or 2Ј-alkylation[33] (entries 3–7) led to modest deprotonate 1 (pKa = 9.6). Future studies on homogeneous
changes in the catalytic activity with retention of selectivity. ion-pairing catalysis (and also on APTC) must address the
The methoxy-substituted alkaloids gave irregular results: role of the counter-ion in chiral quaternary salts QX more
391 Although the quinidine (QD) derivative was only slightly closely.
less selective than its non-methoxylated analogue (entry 8
vs. 1), the quinine (QN) derivative was considerably less se-
lective and active than its demethoxylated CD analogue
431
Variation of the Substrate Structure
(entry 11 vs. 10). In general, use of cymene instead of tolu-
396 ene as co-solvent led to somewhat higher enantioselectivi-
ties.
The asymmetric cyclization of substituted 2Ј-hydroxy-
chalcones to flavanones was investigated with the catalyst
systems 9-Am-CD-Cl/NaH (series I; reaction in toluene/
CHCl3) and 9-Am-CN-Cl/NaH (series II; reaction in cy- 436
mene/CHCl3). Figure 7 gives the intrinsic enantioselectivi-
ties of flavanone generation for both catalyst systems. The
induction with precatalyst 9-Am-CD-Cl is around 60% ee
for many flavanones with monosubstitution patterns (Fig-
ure 7, a–k). Likewise, catalyst 9-Am-CN-Cl induces consis- 441
tent but higher selectivities of 80–85% ee for these flav-
anones. In all cases (a–p), the order of elution of the major
and minor enantiomer is reversed in the chiral HPLC
analysis for the two precatalysts. By analogy to 2, the flav-
anones of series I probably display excess of the 2S configu- 446
ration, and those of series II the 2R configuration. Devia-
tions from the above induction pattern occur particularly
with benzo-annulated flavanones (l–p).
Catalyst Recuperation and Ion-Pairing Effects
After workup of the catalysis reaction mixtures, the cata-
lyst salt remains in the acidic water phase. Catalyst recuper-
401 ation was performed for 9-Am-CD-Cl by careful neutraliza-
tion of the water phase and extraction with CH2Cl2 to give
the catalyst salt in yields of 50–92%, depending on the reac-
tion scale and solvent volumes. Addition of KPF6 to the
neutralized water phase simplified the extraction of the cat-
406 alyst, which was then obtained as its more lipophilic 9-Am-
CD-PF6 salt (82% yield). Surprisingly, this salt displayed
essentially no catalytic activity in the chalcone/flavanone
cyclization reaction; this was confirmed with a pure sample
of the independently prepared (from 9-Am-CD-Cl and
411 NH4PF6) analytically pure salt. The reduced catalytic ac-
tivity of 9-Am-CD-PF6 could be related to the position of
the ion-exchange equilibrium between sodium chalconate
(ChONa), QX, and the ion pair A and NaX [Equation (2),
Scheme 4, a]. For small ions like Cl–, the equilibrium might
416 be on the right due to the high stability of NaCl(s), whereas
anions like PF6– form lipophilic ion pairs QX that are read-
ily soluble in organic medium and have little tendency to
exchange counter-ions with the insoluble NaOCh. The in-
terference of ion-exchange equilibria in the asymmetric
421 catalytic 1Ǟ2 cyclization reaction is supported by the ob-
servation that the salt 9-Am-CD-(2,4-di-tert-butylphenol-
ate) is around 100 times less active than the standard 9-
Figure 7. Intrinsic enantioselectivities (% ee) for the generation of
substituted flavanones with two catalysts. Series I: 9-Am-CD-Cl/
NaH (10 mol-%) in toluene/CHCl3 (3:1). Series II: 9-Am-CN-Cl
(5 mol-%) in cymene/CHCl3 (3:1). Reaction conditions: hydroxy-
chalcone (0.4 mmol) in solvent (4 mL), room temp. Analysis by chi-
ral HPLC sampling of the reaction mixture; products were not iso-
lated.
Scheme 4. a) The position of the ion-exchange equilibrium produc-
ing the catalyst resting state may depend on the nature of the
counter-ion X–. b) The catalytic activity of 9-Am-CD-X salts in the
1Ǟ2 cyclization varies with the nature of X.
8
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
Asymmetric Catalysis of an Equilibrating Reaction
Preparative Applications of the Reaction
Conclusions
We have developed a model for the asymmetric chalcone
isomerase reaction that catalyzes the fundamental reaction 481
1Ǟ2, for which no asymmetric process was previously
known. Past literature reports have mentioned the difficulty
of catalyzing this particular reaction because of its revers-
ibility. Our new process is an asymmetric ion-pairing cataly-
sis in homogeneous solution. The key species is ion pair A, 486
formed from the achiral chalconate anion ChO– and the
chiral quaternary ammonium counter-cation Q+. The cycli-
zation of chalconate within this ion pair produces a strongly
basic flavanone enolate ion pair B with asymmetric induc-
tion. Important structural elements of the chiral cation cat- 491
alyst are a hydrogen-bond-donating hydroxy group, a posi-
tively charged ammonium unit, and an extended aromatic
unit (9-anthracenyl) connected to the quaternary nitrogen.
We have as yet refrained from drawing a detailed structural
model of the transition state. However, it is clear that the 496
above structural elements are responsible for noncovalent
interactions within the substrate ion pair, presumably by
arene and CH-to-arene stacking, hydrogen bonding, and
coulombic attraction. Similar interactions, namely shape
complementarity of the binding site[3a] and hydrogen bond- 501
ing,[3c] have been postulated for the transition state of the
enzymatic chalcone isomerase reaction.[3,4] Furthermore,
kinetic studies of the chalcone isomerase reaction show that
the natural hydroxychalcone substrate is deprotonated at
the 2Ј-hydroxy group in bulk solution prior to binding to 506
the active site.[3b] Therefore, the asymmetric ion-pairing
cyclization reported herein has many parallels and can be
considered a small-molecule model for the enzymatic pro-
cess.
451
To date, the asymmetric catalytic chalcone/flavanone cy-
clization has been studied analytically and mechanistically.
By taking into account the peculiarities imposed by the re-
versibility of the process, it should be possible to develop
preparative applications. A careful analysis of the conver-
456 sion and enantioselectivity as a function of reaction time is
required to determine the optimal reaction conditions and
times. For example, an analytical run with 4ЈЈ-fluoro-2Ј-hy-
droxychalcone using only 2 mol-% of 9-Am-CN-Cl/NaH
indicated that quenching of the reaction after 9 h gives a
461 decent yield of 4Ј-fluoroflavanone with 85% ee, whereas
doubling of the reaction time increased the yield at the cost
of a slightly lower enantiomeric excess (Scheme 5, a). The
trade-off between yield and enantioselectivity is a character-
istic of asymmetric catalyses of reversible processes. A cata-
466 lytic synthesis of enantioenriched 2 was performed on a
gram scale (Scheme 5, b) and showed excellent scalability.
HPLC analysis of the reaction mixture indicated a conver-
sion of 75% and a product enantioselectivity of 78% after
4 h (Figure S5 in the Supporting Information). Quenching
471 and isolation of flavanone (R)-2 gave a 71% yield with 80%
ee. Although this level of enantioselectivity is not satisfac-
tory for synthetic applications,[35] we expect that the reac-
tion can profit from ongoing improvements in asymmetric
phase-transfer catalysts.[36] In addition, this new asymmet-
476 ric catalytic reaction is significant as it is currently the only
process known for the direct asymmetric cyclization of 2Ј-
hydroxychalcone (1) to flavanone (2).
This work has also shed light on issues of asymmetric 511
catalysis of reversible reactions close to equilibrium. Few
such examples have been described in the literature, and the
practical consequences of their reversibility are not widely
appreciated.[17] The simplistic assumption that a reversible
catalysis cannot be catalyzed asymmetrically because of the 516
inherent danger of racemization of the reaction product
must be rejected. It is necessary to perform time-dependent
analyses of conversion and enantiomeric excess in asym-
metric catalyses in order to recognize reversible processes.
Single-point ee determinations of reaction products after 521
workup may conceal the true efficiency of an asymmetric
catalyst. The characteristic course of a reversible asymmet-
ric catalysis involves an initial quasi-equilibration of the
substrate and major enantiomer (characterized by KR/S) fol-
lowed by a slower equilibration of the substrate with race- 526
mic product (characterized by Krac = 2KR/S). This racemiza-
tion is necessarily accompanied by a further increase of
conversion. We have been able to reproduce the kinetic
course of the asymmetric catalytic reaction 1Ǟ2, including
its characteristic ee versus conversion plot, by means of a 531
simple model using only the initial reaction rate, the equilib-
rium constant Krac, and the intrinsic enantioselectivity S of
the catalyst as variables. The structurally simple cyclization
of 2Ј-hydroxychalcone (1) to flavanone (2) has proven to be
Scheme 5. Establishing optimal conditions for preparative applica-
tions of the reversible asymmetric catalytic 2Ј-hydroxychalcone to
flavanone cyclization reaction. a) Analytical testing of a reaction
for optimal yield vs. ee values by chiral HPLC. b) Preparative run
of the reaction with 9-Am-CN-Cl and 1 for obtaining enantioen-
riched flavanone (+)-(2R)-2.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
9

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
L. Hintermann, C. Dittmer
FULL PAPER
N-[3,5-Bis(di-tert-butylphenylsilyloxy)phenyl]methylcinchonidinium
Bromide: Synthesis according to the general procedure GP1 from
cinchonidine and 3,5-bis(tert-butyldiphenylsilyloxy)benzyl brom-
536 an interesting model reaction of general interest for asym-
596
metric catalysis.
1
ide, yyield 60%. H NMR (300 MHz, [D6]DMSO): δ = 1.05 (s, 18
H), 1.10–1.23 (m, 2 H), 1.38–1.54 (m, 1 H), 1.79–1.93 (m, 2 H),
2.00–2.15 (m, 2 H), 2.24–2.37 (m, 1 H), 2.51–2.82 (m, 3 H), 3.34–
Experimental Section
3.46 (m, 1 H), 4.40 (d, J = 12.2 Hz, 1 H), 4.40–4.51 (m, 1 H), 4.90– 601
5.03 (m, 2 H), 5.24–5.39 (m, 1 H), 5.66 (d, J = 12.2 Hz, 1 H), 6.44
(t, J = 2.1 Hz, 1 H), 6.51–6.61 (m, 2 H), 6.63 (d, J = 2.1 Hz, 2 H),
7.23–7.63 (m, 20 H), 7.75–7.82 (m, 2 H), 8.06 (d, J = 8.4 Hz, 1 H),
8.87 (d, J = 4.5 Hz, 1 H) ppm. 13C NMR (76 MHz, CDCl3): δ =
19.4 (C), 21.6 (CH2), 24.9 (CH2), 26.6 (CH3), 38.0 (CH), 51.1 606
(CH2), 61.4 (CH2), 63.6 (CH), 64.2 (CH2), 69.4 (CH), 114.3 (CH),
117.9 (CH2), 118.3 (CH), 120.5 (CH), 122.4 (CH), 124.6 (C), 127.4
(CH), 127.9 (CH), 129.1 (CH), 130.0 (CH), 130.1 (CH), 130.6
(CH), 132.3 (C), 135.6 (CH), 136.5 (CH), 144.8 (C), 148.0 (C),
General: 9-Am-CD-Cl, Bn-CD-Cl, and Bn-QN-Cl were obtained
commercially. The following salts were prepared according to lit-
541
546
erature procedures: 9-Am-QN-Cl[37] and other 9-Am salts,[38,39] 1-
Npm-CD-Cl,[40] and N,O-dibenzylcinchonidinium bromide.[31] The
2Ј-hydroxychalcones were obtained by known methods, of which
aldol condensation of 2-hydroxyacetophenones with aryl aldehydes
and finely powdered Ba(OH)2·8H2O in EtOH proved to be the
most general;[41] see the Supporting Information for detailed pro-
cedures. CC is flash column chromatography. Kinetic fits were per-
formed by using the Dynafit program (BioKin).[42]
150.2 (CH), 156.9 (C) ppm;
1
signal (C) not detected.
611
C58H65BrN2O3Si2·3H2O (1028.29): calcd. C 67.75, H 6.96, N 2.72;
found C 67.80, H 6.70, N 2.53.
General Procedure for the Quaternization of Cinchona Alkaloids
(GP1):[22c] A mixture of the alkaloid (1 equiv.) and alkylating agent
(1 equiv.) was heated at reflux for 3 h in toluene (5 mL/mmol). Af-
ter cooling to room temp., the mixture was diluted with 10 times
its volume of diethyl ether. The precipitated salt was filtered,
washed with diethyl ether, and dried in high vacuum. This general
procedure gave satisfactory results, but the reaction products con-
tained variable amounts of the alkaloid hydrochloride (or hydro-
bromide). Suitable recrystallization procedures must be found for
each individual case. Because alkaloid hydrohalides do not interfere
with the catalytic cyclization of 1 to 2, crude products can in prin-
ciple be used when screening new quaternary salts for catalysis.
551
556
9-Anthracenylmethylcinchonidinium Hexafluorophosphate (9-Am-
CD-PF6):
N-(9-Anthracenylmethyl)cinchonidinium
chloride
(178.6 mg, 90% purity, 0.308 mmol) was dissolved in CH2Cl2
(10 mL) and shaken for several minutes with a solution of NH4PF6
(500 mg) in water (20 mL). The aqueous phase was re-extracted
with CH2Cl2 (5 mL). This procedure (aq. NH4PF6, then re-extrac-
tion) was repeated twice. The organic phase was dried with MgSO4,
filtered, and the solvents evaporated. The residue was dissolved in
CH2Cl2 (2 mL) and overlayed with Et2O (6 mL) and set aside for
crystallization at 4 °C. Filtration and washing with pentane gave a
bright-yellow powder (149 mg, 77%). 1H NMR (300 MHz,
CDCl3): δ = 1.19 (m, 1 H), 1.72–2.89 (several m, 7 H), 3.44–3.48
(m, 1 H), 4.06–4.21 (m, 1 H), 4.48–4.64 (m, 1 H), 4.84–4.97 (m, 2
H), 5.19–5.34 (m, 2 H), 5.37–5.52 (m, 1 H), 5.97 (d, J = 13.6 Hz,
1 H), 6.71 (s, 1 H), 7.22–7.32 (m, 1 H), 7.43 (dd, J = 8.4, 6.7 Hz,
1 H), 7.52–7.73 (m, 4 H), 7.79 (d, J = 7.9 Hz, 1 H), 7.83 (d, J =
4.5 Hz, 1 H), 7.89 (d, J = 7.9 Hz, 1 H), 8.02 (dd, J = 8.4, 1.0 Hz,
1 H), 8.13 (d, J = 8.3 Hz, 1 H), 8.21 (d, J = 9.1 Hz, 1 H), 8.28 (s,
1 H), 8.32 (d, J = 9.0 Hz, 1 H), 8.86 (d, J = 4.5 Hz, 1 H) ppm. 31P
NMR (121 MHz, CDCl3): δ = –144 (JPF = 715 Hz) ppm. 19F NMR
616
621
626
631
636
561
566
571
576
581
586
591
N-(9-Anthracenylmethyl)cinchoninium Chloride (9-Am-CN-Cl): This
procedure includes a purification protocol and gives crystalline
product. Cinchonine (2.94 g, 10.0 mmol; dried in an oven at
100 °C) and 9-chloromethylanthracene (2.30 g, 10.1 mmol) in tolu-
ene (30 mL) were heated at reflux (oil-bath 130 °C) with stirring
using a large magnetic stirring bar. After 1 h, the resulting thick
yellow suspension was diluted with more toluene (20 mL) and
heated at reflux for another 2 h with efficient magnetic stirring. The
reaction mixture was diluted with tBuOMe (50 mL), filtered, and
the solids washed with tBuOMe. After drying at 100 °C, 4.049 g
(85%) of a yellow powder was obtained, which consisted (accord-
ing to 1H NMR) of a 3.75:1 ratio of 9-Am-CN-Cl and CN·HCl.
Purification: The solid was suspended in EtOAc (20 mL) and dis-
solved by the addition of sufficient CHCl3 (ca. 50 mL). The organic
phase was extracted twice with water (50 mL) containing aq. HCl
(0.5 mL of a 2.4 m solution; 2ϫ 1.2 mmol) and once with water
(30 mL), dried with MgSO4, and evaporated to a small volume.
The remaining viscous liquid was diluted with toluene (50 mL) and
set aside for crystallization. [If crystallization started during the
evaporation, the resulting suspension was alternatively diluted with
toluene (50 mL) and slowly evaporated to a small volume prior to
filtration.] Filtration and washing with toluene gave yellow crystals
which were dried in an oven at 100 °C to give the product as a
toluene solvate QCl·1.3C7H8 (4.015 g, 63%). Notes: This salt is
markedly less soluble in CH2Cl2 than in CHCl3. The washing se-
quence with dilute aq. HCl extracts CN·HCl into the water phase
(as the highly soluble CN·2HCl). The amount of aq. HCl needed
for this selective extraction is estimated from the mass and compo-
(282 MHz, CDCl3):
δ
=
–68.4 (JPF
=
715 Hz) ppm.
C34H33F6N2OP·0.3H2O: calcd. C 64.21, H 5.32, N 4.40; found C
64.24, H 5.44, N 4.21.
General Procedure for the Catalytic Cyclization of 2Ј-Hydroxychalc-
one (1) to Flavanone (2) (GP2): Solid NaH (60% in oil; 5–10 mol-
%; see note a below) was added to a degassed suspension/solution
of 2Ј-hydroxychalcone (100 mg, 0.45 mmol) in toluene (or cymene,
3 mL; see note b below). The mixture was stirred for 5 min to give
an orange suspension. The quaternary ammonium salt (2–10 mol-
%) was then added in a counter-stream of argon either as a solid
(single solvent reaction) or as a solution in CHCl3 (0.7–1.0 mL; see
note b below). The reaction mixture was stirred at room temp. for
the desired time. Sampling for HPLC analysis: A small quantity of
the stirred reaction mixture was removed with a Pasteur pipette
(ca. 0.05 mL), diluted with tBuOMe (1 mL), and shaken with aq.
HCl (2.4 m, 0.5 mL). The organic phase was directly used for
HPLC analysis. Workup: The reaction was quenched by addition
of aq. HCl (2.4 m, 5 mL) and tBuOMe (10 mL). The organic phase
was washed with water (3ϫ 20 mL) and dried with MgSO4. After
filtration and evaporation, the residue was purified by CC (SiO2,
tBuOMe/hexanes, 1:10). Notes: a) This small quantity of NaH is
best removed with a spatula from a tared vessel on an analytical
balance in a single operation. The spatula is then dipped into the
reaction mixture until the NaH is completely washed into the reac-
641
646
651
656
1
sition (by H NMR) of the crude product. Use of excess acid will
lead to loss of 9-Am-CN-Cl and the formation of a yellow precipi-
tate (probably 9-Am-CN-Cl·HCl) in the aqueous phase. Commer-
cial cinchonine may contain several % of dihydrocinchonine; this
is also quaternized and the resulting 9-Am-DHCN-Cl can be recog-
nized by a peak at δH = 0.46 (t, J = 7.3 Hz, 3 H) ppm.
10
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
Asymmetric Catalysis of an Equilibrating Reaction
1
tion mixture. The quantity of base was adapted as follows: For 5–
10 mol-% of QCl, around 1.8 mg of NaH (60% in oil, 0.045 mmol,
10 mol-%) are applied, for 2–5 mol-% of QCl, around 0.8–1 mg of
NaH (4–5 mol-%). The presence of a small excess of base is not
problematic. b) The reaction solvents were filtered through neutral
Al2O3 prior to use; this removes acid traces from CHCl3 and perox-
ide traces from aromatic solvents (particularly for cymene).
evaporation, the residue was analyzed by H NMR and, if neces-
sary, precipitated from CH2Cl2/tBuOMe or CH2Cl2/pentane.
Yields: 50–92%, depending on reaction scale and catalyst loading.
721
661
Catalyst Recuperation as the Hexafluorophosphate Salt: A catalytic
reaction according to GP2 (2.2 mmol scale, 10 mol-% catalyst) was
quenched and worked up as described above. The aqueous phase
collected from workup/extraction/washing was washed with tBu-
OMe (2ϫ 20 mL) and neutralized by the addition of satd.
NaHCO3. A solution of NH4PF6 (200 mg, 1.2 mmol) in H2O
(2 mL) was added and the turbid mixture extracted with CH2Cl2
(2ϫ 10 mL). The organic phase was washed with a solution of
NH4PF6 (200 mg in 5 mL H2O), dried with MgSO4, filtered, and
the solvents evaporated to dryness. The residue was analyzed di-
rectly by 1H NMR or after crystallization (precipitation) from
CH2Cl2/tBuOMe; yield 82%.
726
731
Synthesis of Enantioenriched Flavanone (2): In a Schlenk vessel un-
666 der argon, 2Ј-hydroxychalcone (1; 1.000 g, 4.46 mmol) and NaH
(60% in oil; 10.6 mg, 0.265 mmol, 6 mol-%) were stirred in toluene
(20 mL, degassed) to give a yellow solution/suspension. A solution
of N-(9-anthracenylmethyl)cinchoninium chloride (108.3 mg,
0.208 mmol, 4.7 mol-%) in CHCl3 (7 mL; filtered through Al2O3)
671
676
681
was added and the reaction mixture stirred for 4 h at room temp.
(22 °C). The reaction was quenched by the addition of aq. HCl
(2.4 m, 4 mL) and the organic phase washed with water. After evap-
oration, the residue was purified by CC (SiO2, tBuOMe/hexanes =
1:10) to give a colorless solid (709 mg, 71%). [α]D = +50.2Ϯ2.0 (c
= 0.75; CHCl3); lit.:[43] [α]D = +62.8 (c = 1 in CHCl3) for a sample
Supporting Information (see footnote on the first page of this arti-
cle): Detailed catalyst screening results, analytical data, procedures
for chalcone synthesis, and selected NMR spectra of reaction prod-
ucts.
736
1
of (+)-(2R)-2 with 95% ee. H NMR (400 MHz, CDCl3): δ = 2.90
(dd, J = 16.8, 2.9 Hz, 1 H), 3.10 (dd, J = 16.8, 13.4 Hz, 1 H), 5.50
(dd, J = 13.3, 2.9 Hz, 1 H), 7.04–7.09 (m, 2 H), 7.39–7.55 (m, 6
H), 7.94 (dd, J = 8.1, 1.8 Hz, 1 H) ppm. 13C NMR (100 MHz,
CDCl3): δ = 44.7 (CH2), 79.6 (CH), 118.2 (CH), 120.5 (C), 121.6
(CH), 126.2 (CH), 127.1 (CH), 128.8 (CH), 128.9 (CH), 136.2
(CH), 138.8 (C), 162.0 (C), 192.0 (C) ppm. C15H12O2 (224.26):
calcd. C 80.34, H 5.39; found C 80.14, H 5.52.
Acknowledgments
This project was supported by the Deutsche Forschungsgemein-
schaft (DFG) through the Emmy Noether Programm (L. H.) and 741
within the DFG project SPP 1179.
(2R,3S)-3-[(S)-3-(2-Hydroxyphenyl)-3-oxo-1-phenylpropyl]-2-phen-
[1] Part I of the series “Mechanisms of conjugate additions of het-
eronucleophiles”; for Part II, see: L. Hintermann, A. Turockin,
manuscript in preparation.
[2] a) K. Springob, J. Nakajima, M. Yamazaki, K. Saito, Nat.
Prod. Rep. 2003, 20, 288; b) E. Moustafa, E. Wong, Phytochem-
istry 1967, 6, 625.
686 ylchroman-4-one (Flavanone/Chalcone Conjugate 3): In a Schlenk
vessel under argon, 2Ј-hydroxychalcone (1.003 g, 4.47 mmol) and
N-(9-anthracenylmethyl)cinchoninium chloride (57 mg, 0.11 mmol,
2.4 mol-%) were stirred in toluene (7 mL, degassed). NaH (60% in
oil; 6 mg, 0.15 mmol, 3.3 mol-%) was added to give a yellow solu-
746
751
756
761
766
771
776
781
[3] a) J. M. Jez, M. E. Bowman, R. A. Dixon, J. P. Noel, Nat.
Struct. Biol. 2000, 7, 786; b) J. M. Jez, J. P. Noel, J. Biol. Chem.
2002, 277, 1361; c) J. M. Jez, M. E. Bowman, J. P. Noel, Bio-
chemistry 2002, 41, 5168.
[4] S. Hur, Z. E. R. Newby, T. C. Bruice, Proc. Natl. Acad. Sci.
USA 2004, 101, 2730.
[5] R. A. Bednar, J. R. Hadcock, J. Biol. Chem. 1988, 263, 9572.
[6] a) A. E. Nibbs, K. A. Scheidt, Eur. J. Org. Chem. 2012, 449; b)
M. G. Núñez, P. García, R. F. Moro, D. Díez, Tetrahedron
2010, 66, 2089; c) C. F. Nising, S. Bräse, Chem. Soc. Rev. 2012,
41, 988; d) C. F. Nising, S. Bräse, Chem. Soc. Rev. 2008, 37,
1218.
[7] For a low-yielding direct asymmetric catalytic synthesis of 6-
fluoroflavanone from benzaldehyde and 5-fluoro-2-hydroxy-
acetophenone, see: S. Zhou, Y. Zhou, Y. Xing, N. Wang, L.
Cao, Chirality 2011, 23, 504.
691
tion/suspension. The reaction mixture was stirred for 40 h at 45 °C
in an oil-bath. The reaction was quenched by the addition of aq.
HCl (2.4 m, 4 mL) and tBuOMe (50 mL) and the organic phase
was washed with water. After evaporation, the residue was purified
by CC (SiO2, tBuOMe/hexanes = 1:5–1:3) to give a colorless solid
696 (378.4 mg, 38%); m.p. 140–142 °C. [α]D = –107.8 (c = 0.5, EtOH).
HPLC (Chiralcel-OJ, n-heptane/iPrOH = 80:20, 0.8 mL/min, λ =
1
254 nm): 62.5% ee. H NMR (400 MHz, CDCl3): δ = 3.31 (dd, J
= 10.7, 2.5 Hz, 1 H), 3.47 (dd, J = 17.6, 4.8 Hz, 1 H), 3.58 (dd, J
= 17.6, 8.9 Hz, 1 H), 3.98 (ddd, J = 10.7, 8.9, 4.8 Hz, 1 H), 5.29
701
706
(d, J = 2.3 Hz, 1 H), 6.79–7.78 (m, 18 H), 11.89 (s, OH) ppm. 13C
NMR (100 MHz, CDCl3): δ = 39.7 (CH), 42.1 (CH2), 56.0 (CH),
79.0 (CH), 118.1 (CH), 118.4 (CH), 118.8 (CH), 119.3 (C), 120.6
(C), 121.5 (CH), 126.3 (CH), 127.0 (CH), 127.4 (CH), 128.0 (CH),
128.2 (CH), 128.6 (CH), 129.0 (CH), 129.7 (CH), 136.2 (CH), 136.7
(CH), 137.7 (C), 141.1 (C), 159.3 (C), 162.1 (C), 193.4 (C), 203.3
[8] C. Dittmer, G. Raabe, L. Hintermann, Eur. J. Org. Chem. 2007,
5886.
(C) ppm. IR (KBr): ν = 3451 (s), 3061 (w), 2925 (s), 1682 (s), 1284 [9] a) M. M. Biddle, M. Lin, K. A. Scheidt, J. Am. Chem. Soc.
˜
(m) cm–1. MS (EI): m/z (%) = 448 (1) [M]+, 312 (16), 224 (66), 223
(100), 147 (16), 121 (64). C30H24O4 (448.51): calcd. C 80.34, H 5.39;
found C 80.21, H 5.66. Assignment of the 2R,1ЈS predominating
2007, 129, 3830; b) R. L. Farmer, M. M. Biddle, A. E. Nibbs,
X. Huang, R. C. Bergan, K. A. Scheidt, ACS Med. Chem. Lett.
2010, 1, 400.
[10] L. J. Wang, X. H. Liu, Z. H. Dong, X. Fu, X. M. Feng, Angew.
Chem. 2008, 120, 8798; Angew. Chem. Int. Ed. 2008, 47, 8670.
[11] H.-F. Wang, H. Xiao, X.-W. Wang, G. Zhao, Tetrahedron 2011,
67, 5389.
711 configuration is based on the generation of (2R)-flavanone with the
same catalyst.
Catalyst Recuperation as the Chloride Salt: A catalytic reaction ac-
cording to GP2 was quenched and worked up as described in the
general procedure. The aqueous phases were collected and washed
with tBuOMe (2ϫ 20 mL) and then neutralized by the addition of
satd. NaHCO3. An equal volume of satd. NaCl was added and the
aqueous phase extracted with CH2Cl2 (3ϫ 20 mL). The organic
phase was washed with satd. NaCl and dried with MgSO4. After
[12] Z. Feng, M. Zeng, Q.-L. Xu, S.-L. You, Chin. Sci. Bull. 2010,
55, 1723.
[13] K. Maruyama, K. Tamanaka, A. Nishinaga, A. Inada, T. Nak-
anishi, Tetrahedron Lett. 1989, 30, 4145.
[14] a) T. Patonay, R. S. Varma, A. Vass, A. Lévai, J. Dudás, Tetra-
hedron Lett. 2001, 42, 1403; b) C. J. Adams, L. Main, B. K.
Nicholson, Acta Crystallogr., Sect. C 1990, 46, 2424.
716
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
11

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
L. Hintermann, C. Dittmer
FULL PAPER
[15] F. Han, G. Chen, X. Zhang, J. Liao, Eur. J. Org. Chem. 2011,
2928.
[16] K. J. Hodgetts, ARKIVOC 2001, 74.
performed the cyclization of 1 in [D4]methanol with sodium
benzoate as a weakly basic catalyst and have similarly found
syn/anti = 48:52; a similar result was obtained in the base-cata-
lyzed H/D exchange of 3-H in 2 (see the Supporting Infor-
mation).
[17] a) L. Hintermann in Asymmetric Synthesis More Methods and
Applications (Eds.: M. Christmann, S. Bräse), Wiley-VCH,
Weinheim, 2012; b) for an asymmetric organocatalytic aldol
reaction in which either kinetic control (product isolated with
high ee) or thermodynamic control (racemization) has been
achieved under suitable conditions, see: G. Rulli, N. Duangdee,
K. Baer, W. Hummel, A. Berkessel, H. Gröger, Angew. Chem.
2011, 123, 8092; Angew. Chem. Int. Ed. 2011, 50, 7944.
[18] Additional data is provided in the Supporting Information.
[19] The decarboxylation is presumably catalyzed by chalcone 1.
For the decaboxylation of amino acids by an unsaturated
ketone as catalyst, see: M. Hashimoto, Y. Eda, Y. Osanai, T.
Iwai, S. Aoki, Chem. Lett. 1986, 893.
826
831
786
791
796
801
806
811
816
821
[30] The term “living” is used in analogy to “living anionic polyme-
rization”, in which the reactive carbanionic center resides on
the growing polymer chain, and each monomer addition event
recreates a new reactive basic centre.
[31] M. J. O’Donnell, S. Wu, J. C. Huffman, Tetrahedron 1994, 50,
4507.
[32] C. Exner, A. Pfaltz, M. Studer, H.-U. Blaser, Adv. Synth. Catal.
2003, 345, 1253.
[33] L. Hintermann, M. Schmitz, U. Englert, Angew. Chem. 2007, 836
119, 5256; Angew. Chem. Int. Ed. 2007, 46, 5164.
[34] I. Leupold, H. Musso, Justus Liebigs Ann. Chem. 1971, 746,
134.
[20] a) T. Marcelli, H. Hiemstra, Synthesis 2010, 1229; b) K.
Kacprzak, J. Gawronski, Synthesis 2001, 961.
[35] Crystallization from EtOH did not change the ee of 2.
[21] a) J. J. O’Donnell, “Asymmetric Phase-Transfer Reactions” in
Catalytic Asymmetric Synthesis (Ed.: I. Ojima), Wiley-VCH,
Weinheim, 2000, p. 727; b) A. Nelson, Angew. Chem. 1999, 111,
1685; c) J. Vachon, J. Lacour, Chimia 2006, 60, 266.
[22] a) B. Lygo, B. I. Andrews, Acc. Chem. Res. 2004, 37, 518; b)
B. Lygo, P. G. Wainwright, Tetrahedron Lett. 1997, 38, 8595; c)
E. J. Corey, F. Xu, M. C. Noe, J. Am. Chem. Soc. 1997, 119,
12414.
[36] Since our study was carried out (2004–2007), new classes of
asymmetric phase-transfer catalysts have been developed that
might prove superior in the current reaction: a) T. Hashimoto,
K. Maruoka, Chem. Rev. 2007, 107, 5656; b) T. Ooi, D. Ohara,
K. Fukumoto, K. Maruoka, Org. Lett. 2005, 7, 3195.
841
[37] a) T. Perrard, J.-C. Plaquevent, J.-R. Desmurs, D. Hébrault, 846
Org. Lett. 2000, 2, 2959; b) C. Hofstetter, P. S. Wilkinson, T. C.
Pochapsky, J. Org. Chem. 1999, 64, 8794.
[23] A. Cisak, C. Mielczarek, J. Chem. Soc. Perkin Trans. 2 1992,
[38] a) B. Lygo, P. G. Wainwright, Tetrahedron 1999, 55, 6289; b)
1603, and references cited therein.
B. Lygo, J. Crosby, T. R. Lowdon, P. G. Wainwright, Tetrahe-
[24] For the peculiarities of the analytical separation of 1 and (R/
dron 2001, 57, 2391.
851
856
861
S)-2 by HPLC on chiral stationary phases, see: L. Hintermann, [39] P. Schanen, Asymmetric Cyclopentannulation Reactions: Scope
J. Org. Chem. 2007, 72, 9790.
[25] R. G. Button, P. J. Taylor, J. Chem. Soc. Perkin Trans. 2 1992,
1571.
[26] The pKa values are for aqueous solutions and thus only general
guidelines.
[27] P. Scharlin, Acta Chem. Scand., Ser. A 1986, 40, 207.
[28] K. B. Old, L. Main, J. Chem. Soc. Perkin Trans. 2 1982, 1309.
[29] The relative stereochemistry of the addition of RO–H to the
C=C double bond in the base-catalyzed 1/2 cyclization reaction
in aqueous/alcoholic solution is syn/anti = 55:45: C. O. Miles,
L. Main, J. Chem. Soc. Perkin Trans. 2 1989, 1623. We have
and Limitation, Doctoral Thesis, Universite Catholique de
Louvain, Louvain, 2003.
[40] S. Caron, N. M. Do, P. Arpin, A. Larivee, Synthesis 2003, 1693.
[41] a) S. Sogawa, Y. Nihro, H. Ueda, A. Izumi, T. Miki, H. Matsu-
moto, T. Satoh, J. Med. Chem. 1993, 36, 3904; b) A. R. Alcan-
tara, J. M. Marinas, J. V. Sinisterra, Tetrahedron Lett. 1987, 28,
1515.
[42] P. Kuzmic, Anal. Biochem. 1996, 237, 260.
[43] Y. Noda, M. Watanabe, Helv. Chim. Acta 2002, 85, 3473.
Received: June 22, 2012
Published Online: ᭿
12
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20838
/KAP1
Date: 21-08-12 17:29:15
Pages: 13
Asymmetric Catalysis of an Equilibrating Reaction
Asymmetric Catalysis
L. Hintermann,* C. Dittmer ........... 1–13
Asymmetric Ion-Pairing Catalysis of the
Reversible Cyclization of 2Ј-Hydroxychalc-
one to Flavanone: Asymmetric Catalysis of
an Equilibrating Reaction
866
Keywords: Asymmetric catalysis / Enzyme
The cyclization of nonactivated 2Ј-
hydroxychalcones to flavanones has been
realized for the first time with asymmetric
induction. In spite of the reversibility of the
reactions, this ion-pairing catalysis reaches
intrinsic enantioselectivities of up to S =
14.4 (er = 93.5:6.5) and illustrates the
characteristic behavior of reversible asym-
metric catalyses close to their equilibria.
871
876
models
/ Natural products / Reaction
mechanisms / Oxygen heterocycles
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
13