Inner workings of a cinchona alkaloid catalyzed oxa-Michael cyclization: Evidence for a concerted hydrogen-bond-network mechanism

Article abstract of DOI:10.1002/chem.201203505

Cinchona alkaloids catalyze the oxa-Michael cyclization of 4-(2-hydroxyphenyl)-2-butenoates to benzo-2,3-dihydrofuran-2-yl acetates and related substrates in up to 99 % yield and 91 % ee (ee=enantiomeric excess). Catalyst and substrate variation studies reveal an important role of the alkaloid hydroxy group in the reaction mechanism, but not in the sense of a hydrogen-bonding activation of the carbonyl group of the substrate as assumed by the Hiemstra-Wynberg mechanism of bifunctional catalysis. Deuterium labeling at C-2 of the substrate shows that addition of RO-H to the alkenoate occurs with syn diastereoselectivity of ≥99:1, suggesting a mechanism-based specificity. A concerted hydrogen-bond network mechanism is proposed, in which the alkaloid hydroxy group acts as a general acid in the protonation of the α-carbanionic center of the product enolate. The importance of concerted hydrogen-bond network mechanisms in organocatalytic reactions is discussed. The relative stereochemistry of protonation is proposed as analytical tool for detecting concerted addition mechanisms, as opposed to ionic 1,4-additions. Secret of cyclization: The cinchona alkaloid catalyzed asymmetric oxa-Michael cyclization of 2′-hydroxyphenyl-2-butenoates to benzodihydrofurans proceeds by a highly enantio- and diastereoselective syn-specific addition mode (see scheme). Transition-state activation of the carbonyl group by hydrogen bonding to the catalyst is excluded. This represents a clear-cut demonstration of the importance of concerted hydrogen-bond network mechanisms in cinchona-based asymmetric organocatalysis. Copyright

Full text of DOI:10.1002/chem.201203505

FULL PAPER
DOI: 10.1002/chem.201203505
Inner Workings of a Cinchona Alkaloid Catalyzed Oxa-Michael Cyclization:
Evidence for a Concerted Hydrogen-Bond-Network Mechanism**
Lukas Hintermann,* Jens Ackerstaff, and Florian Boeck[a]
Abstract: Cinchona alkaloids catalyze
the oxa-Michael cyclization of 4-(2-hy-
droxyphenyl)-2-butenoates to benzo-
2,3-dihydrofuran-2-yl acetates and re-
lated substrates in up to 99% yield and
91% ee (ee=enantiomeric excess).
Catalyst and substrate variation studies
reveal an important role of the alkaloid
hydroxy group in the reaction mecha-
nism, but not in the sense of a hydro-
gen-bonding activation of the carbonyl
group of the substrate as assumed by
the Hiemstra–Wynberg mechanism of
bifunctional catalysis. Deuterium label-
ing at C-2 of the substrate shows that
network mechanism is proposed, in
which the alkaloid hydroxy group acts
as a general acid in the protonation of
the a-carbanionic center of the product
enolate. The importance of concerted
hydrogen-bond network mechanisms in
organocatalytic reactions is discussed.
The relative stereochemistry of proto-
nation is proposed as analytical tool for
detecting concerted addition mecha-
nisms, as opposed to ionic 1,4-addi-
tions.
ꢀ
addition of RO H to the alkenoate
occurs with syn diastereoselectivity of
ꢁ99:1, suggesting a mechanism-based
specificity. A concerted hydrogen-bond
Keywords: alkaloids · asymmetric
catalysis · conjugate additions · or-
ganocatalysis · reaction mechanisms
Introduction
The conjugate addition of hetero nucleophiles HX (X=N,
O, S) to Michael acceptors[1] has been studied early in asym-
metric catalysis.[2,3] Mechanistic ideas about this reaction
have been shaped by the study of Hiemstra and Wynberg in
1981 on the addition of thiophenols to cycloalkenons.[4]
They proposed a bifunctional catalysis mechanism with hy-
drogen-bond activation of the carbonyl group of the sub-
strate by the hydroxy group of the catalyst, and activation of
the hetero nucleophile by deprotonation with the tertiary
amine group of the catalyst to give an ammonium thiolate
ion pair (Figure 1). Activation of carbonyl groups through
hydrogen bonds is still a first guess for mechanisms of those
reactions, in which cinchona alkaloids are superior catalysts
to their O-protected derivative,[5] and binding of heteroa-
toms is the almost exclusive mode of action discussed in hy-
Figure 1. The Hiemstra–Wynberg-model of bifunctional hydrogen-bond-
ing catalysis.[4] a) Sketch of the bifunctional activation mode. b) Model of
a transition state.
drogen-bonding catalysis.[6] Jacobsen et al. have extended
this picture by using hydrogen-bond donors for activating
[7]
ꢀ
labile C X bonds (X=Cl, Br) by binding to halide anions.
We now find that heteroatom bonding and carbonyl group
activation is absent in a new cinchona alkaloid catalyzed
asymmetric oxa-Michael cyclization—even though the reac-
tion critically depends on the presence of a hydroxy group
in the catalyst. Evidence for a hydrogen-bond network relay
1,2-addition is presented, where the hydroxy group of the
catalyst acts as a general acid by protonating the enolate a-
carbon atom, in concert with phenoxide attack at the b-
carbon atom. This mechanistic principle may be of wider im-
portance in hydrogen-bonding organocatalysis.
[a] Prof. Dr. L. Hintermann, Dr. J. Ackerstaff, Dr. F. Boeck
Department Chemie, Technische Universitꢀt Mꢁnchen
Lichtenbergstrasse 4, 85747 Garching bei Mꢁnchen (Germany)
Fax : (+49)89-28913669
E-mail: lukas.hintermann@tum.de
[**] Part III of the series: “Mechanisms of conjugate additions of hetero
nucleophiles” (taken in part from the diploma thesis of J.A. (RWTH
Aachen, 2004) and the habilitation thesis of L.H. (RWTH Aachen,
2008)); for part I, see: L. Hintermann, C. Dittmer, Eur. J. Org. Chem.
2012, 5576; for Part II, see: L. Hintermann, A. Turockin, J. Org.
Chem. 2012, DOI: 10.1021/jo3021709.
Results and Discussion
Supporting information (containing full experimental details and
characterization of new compounds, additional catalyst screening re-
sults, and documentation of mechanistic experiments by spectra and
reaction curves) for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201203505.
Cinchona alkaloid catalyzed asymmetric 5-exo-trig oxa-Mi-
chael cyclizations: The addition of oxygen nucleophiles to
a,b-unsaturated carbonyl compounds (oxa-Michael reaction)
provides alternative access to aldol-type products.[8] For a
Chem. Eur. J. 2013, 19, 2311 – 2321
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2311

long time, the only example for an asymmetric catalytic oxa-
Michael addition had been a quinine-catalyzed synthesis of
2,3-dimethyl-chroman-4-ons through 6-endo-trig cycliza-
tion.[9] While the present work was in progress, a cinchona
alkaloid catalyzed 6-exo-trig cyclization to chroman-2-yl ace-
tate was reported.[10] Since then, progress has been achieved
by several groups[11–13] and the topic has been reviewed.[14]
Reaction development: Our approach to the topic started
with a study of the 5-exo-trig cyclization of 4-(2’-hydroxy-
phenyl) crotonates (1) to benzodihydro-2-furanyl acetates
(2) as model reaction for identifying new chiral catalysts for
the oxa-Michael-reaction (Scheme 1).[15] Similar substrates
Scheme 2. Synthesis of substrates 1 through lactols 4–6.
mol% and run to completion within 2 to 5 days. Hydrogen-
bond-accepting or -donating solvents display low activity
and selectivity (Table 1, entries 1–3), whereas chlorinated
arenes or chlorohydrocarbons are suitable solvents (for
more solvent screening results, see Table S-2 in the Support-
ing Information). In chloroform, the reactions were fastest,
reaching conversions in excess of 98% (HPLC) within 30 h.
The solvent effects observed are different from those in the
6-endo-trig cyclizations described by Ishikawa et al.:[9b] their
preferred solvent, chlorobenzene, led to slower reactions
and increased levels of side product 7 in our reaction (see
Table S-4 in the Supporting Information for a dataset).[18]
Higher catalyst loading increases the reaction rate without
affecting the enantioselectivity. The substrate concentration
was an important parameter: higher ee (enantiomeric
excess) values were found in more dilute solutions (Table 1,
Scheme 1. 5-exo-trig Model reaction for investigating oxa-Michael cycli-
zations.
had been cyclized in the presence of a secondary amine[16]
or Na2CO3.[17] The required model substrates 1 were ob-
tained by ozonolysis of 2-allylphenol (3) to give compound 4
and a Wittig reaction to give (E)-1 (Scheme 2). Lactol 4 is
in equilibrium with the hydroxyaldehyde 4’ (4/4’=6:1) in
CDCl3. Other allylphenols analogously gave lactols 5 (5/5’=
10:1) and 6 (only lactol). Many potential catalysts were
tested in the cyclization of 1a
to 2a (see Tables S-1 and S-3 in
the Supporting Information).[18]
Table 1. Selected results for the asymmetric catalytic oxa-Michael cyclization of 4-(2’-hydroxyaryl)-2-bute-
noates 1 to benzodihydrofuranyl acetates 2.[a]
The initial catalyst screen
showed (Table S-1 in the Sup-
porting Information) that the
catalytic activity and the enan-
tioselectivity emerge for com-
Substrate
R
R’
Catalyst
Solvent
EtOH
tBuOMe 20
MeCN
PhCl
CHCl3
CHCl3
CHCl3
PhCl
c
G
t
Conv.
ACHTUNGTRENNUNG
(yield) [%][d]
Product ee of 2
[b]
[gLꢀ1
]
[h][c]
[%][e]
pounds with
a basic amine
group, a free hydroxy group in
proximity of the amine group,
and p-stacking groups in prox-
imity of the alcoholic function.
The use of cinchona alkaloids
as catalysts recommended itself
due to their success in catalyz-
ing conjugate additions,[4,19,20]
including 6-endo-trig[9,12] and 6-
exo-trig oxa-Michael cycliza-
tions.[10] In fact, they turned out
to be the most successful cata-
lysts for our 5-exo-trig cycliza-
tion (Table 1; see also Table S-1
in the Supporting Informa-
tion).[18] All reactions were per-
formed at room temperature
1
2
3
4
5
6
7
8
9
1a
1a
1a
1a
1a
1a
1a
1a
1a
Me
Me
Me
Me
Me
Me
Me
Me
Me
Bn
H
H
H
H
H
H
H
H
H
H
H
H
H
CD[f]
CD
CD
CD
CD
CN
CN
CN
20
215
215
215
200
120
120
40
120
90
108
96
40
30
>90
>95
>95
2a
2a
2a
2a
2a
2a
2a
2a
2a
2b
2b
2c
2c
2d
2d
20 (ꢀ)
50 (ꢀ)
64 (ꢀ)
62 (ꢀ)
76 (ꢀ)
86 (+)
89 (+)
86 (+)
90 (+)
83 (+)
79 (ꢀ)
77 (+)
75 (ꢀ)
86 (+)
78 (ꢀ)
20
15
20
20
8
(83)
99
>95
>95
>90 (87)
>90
>90
>90 (80)
>90
8
8
DHCN[g] CHCl3
10 1b
11 1b
12 1c
13 1c
14 1d
15 1d
CN
CD
CN
CD
CHCl3
CHCl3
CHCl3
CHCl3
CHCl3
CHCl3
20
20
20
20
20
20
Bn
tBu
tBu
Me
Me
144
144
96
OMe CN
OMe CD
96
>90
[a] Unless otherwise mentioned, conditions are: 15–20 mol% catalyst, RT. [b] Initial concentration of the start-
ing material. For 1a, 8 gLꢀ1 =0.042 molLꢀ1, 20 gLꢀ1 =0.1 molLꢀ1. [c] Time to complete conversion as deter-
mined by TLC. [d] Conversions are estimates by TLC, or determined by NMR spectroscopy/HPLC. Yields of
the isolated product are given in brackets. [e] The ee of 2 was determined by chiral HPLC. Sign of the optical
rotation of 2 (measured in tBuOMe) is given in brackets. The absolute configuration of (+)-2 is presumably
with a catalyst loading of 15–20 (2S).[18] [f] CD=cinchonidine. [g] DHCN=dihydrocinchonine.
2312
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2311 – 2321

Cinchona Alkaloid Catalyzed Oxa-Michael Cyclization
FULL PAPER
entry 6 vs. entry 7). Because the substrate concentration de-
creases with reaction progress, the ee of the product slowly
increases and is highest at the end of the reaction (see Fig-
ure S-2 in the Supporting Information).[18] An extended da-
taset for the initial and final ee values of 2a in dependence
of the initial concentration of 1a has been obtained in PhCl,
where the effect is stronger than in CHCl3 (Figure S-3 in the
Supporting Information), though the maximal ee (91%) of
2a at high dilution is the same in both solvents. An inverse
trend of the concentration versus the ee was observed in our
study of the cinchona alkaloid catalyzed chalcone/flavanone
cyclization (6-endo-trig).[12b] Here, we assume that the phe-
nolic substrate 1 competes with the hydroxyl group of the
catalyst as hydrogen-bond donor in an alternative transition
state producing 2a with lower selectivity. Substrate 1 was
varied in the ester or the arene moiety (1b–d). The steric
bulk of the ester group is of minor importance for both the
stereoselectivity and the rate of the catalysis (Table 1,
entry 6 vs. entries 10–15). Cinchonine (CN) gave the highest
enantioelectivities in most cases (see Tables S3–S10 in the
Supporting Information for detailed screening results with
other cinchona alkaloids).[18] A substitution in the arene por-
tion of the substrate has little effect at the C-5’ atom (1d;
Table 1, entries 14 and 15). However, a bulky substituent in
ortho position to the phenolic hydroxy group slows the reac-
tion and lowers the enantioselectivity (1e; Scheme 3). The
benzodihydrofuranyl ester 8 had been used by Kraus et al.
as intermediate in the total synthesis of the antibiotic (ꢂ)-
nanaomycin A (Scheme 4a).[21] They had obtained racemic 8
by combining acetyl-benzoquinone (9) with the diene 10a
and subsequent workup with tetrabutylammonium fluoride
(TBAF). A sequence of a Diels–Alder cycloaddition, a fluo-
ride-induced retro-Claisen reaction, and an oxa-Michael cyc-
lization was assumed.[21] In our hands, quenching of a reac-
tion mixture of compounds 9 and 10b[22] with KHF2 as
acidic fluoride source allowed the interception of (E)-1 f,
prior to its oxa-Michael cyclization (Scheme 4b).[23] The cat-
alytic cyclization of 1 f proceeded most satisfactorily with
only 10 mol% of cinchonine, generating 2 f in 90% ee
within a day (Scheme 4b; see Table S-9 in the Supporting
Information for reaction optimization). This particular reac-
tion provides a chiral switch for existing total syntheses of
pyrano-naphthoquinone antibiotics, with no need for addi-
tional steps.[24,25] The cyclization of hydroxyaryl ketones 11
was also shortly investigated (Scheme 5): being stronger Mi-
Scheme 5. a) Spontaneous cyclization of hydroxyalkenones 11 to benzodi-
hydrofurans 12. b) In situ generation of 11b and asymmetric cyclization
to 12b.
chael acceptors than esters 1, already their synthesis was ac-
companied by cyclization (Scheme 5a), which prevented iso-
lation of pure products. However, when the Wittig reaction
of lactol 4 with ylide 13 was carried out in the presence of
cinchonidine, the in-situ-generated 11b underwent alkaloid-
catalyzed cyclization to 12b, as shown by a moderate asym-
metric induction (Scheme 5b). Finally, we found that benzo-
dioxanyl acetate 15 can be obtained by catalytic asymmetric
oxa-Michael cyclization of ethyl ester 14[26] in a 6-exo-trig cy-
clization.[10,13a,h] The optimization of the catalysis is present-
ed in Table S-10 in the Supporting Information.[18] The best
result was again obtained with cinchonine as catalyst in 1,2-
dichlorobenzene (ODCB) solution (Scheme 6). The ee of
74% and the longer reaction time reflects the more difficult
6-exo-trig (as opposed to 5-exo-trig) ring closure.[27]
Scheme 3. The hindered substrate 1e only cyclizes with difficulty.
Scheme 4. Oxa-Michael cyclization in the total synthesis of nanaomy-
cin A. a) Total synthesis from a racemic benzodihydrofuranyl acetate as
described by Kraus et al.[21] b) Introducing a chiral switch by means of an
asymmetric catalytic oxa-Michael cyclization.
Scheme 6. Asymmetric catalytic oxa-Michael cyclization of 14 to benzo-
dioxanyl acetate 15.
Chem. Eur. J. 2013, 19, 2311 – 2321
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2313

L. Hintermann et al.
Mechanism of the oxa-Michael cyclization: A large and
growing number of cinchona alkaloid catalyzed asymmetric
reactions in the literature[20] contrasts with relatively few de-
tailed studies of their mechanisms.[28] Having introduced a
new asymmetric cinchona alkaloid catalyzed oxa-Michael
cyclization and defined some of its characteristics, we
wished to study its mechanism in more detail.
Kinetics of the catalytic oxa-Michael cyclization: The kinetic
analysis of the model reaction 1a!2a was performed in the
less volatile solvent chlorobenzene with the more soluble
catalyst 2’-n-butylcinchonine[29] to ensure stable concentra-
tions over extended reaction times. The reaction progress
was analyzed by chiral HPLC. Plots of the initial reaction
rates versus c1 (ccat =constant) or alternatively versus ccat
(c1 =constant) were negatively curved (Figure S-4 in the
Supporting Information).[18] A logarithmic plot indicated re-
action orders of approximately 0.7 for the catalyst and 0.5
for substrate 1a (Figure S-4 in the Supporting Information).
The lower than unity reaction orders are presumably a con-
sequence of self-association of both the cinchona alka-
loids[30] and the phenols.[31] In a simplified analysis, associa-
tion equilibria reduce the concentration of the monomeric
substrate and the catalyst available for the catalytic reaction.
The kinetic measurements also illustrated that the ee of the
product was increasing with conversion (Figure S-2 in the
Supporting Information), and that there is a significant de-
pendency of the ee value of the product on the initial sub-
strate concentration c01 a (Figure S-3 in the Supporting Infor-
mation). Nevertheless, the kinetic analysis is in agreement
with a mechanism involving a single catalyst and a single
substrate molecule in the rate-limiting step.
Figure 2. The Hiemstra–Wynberg model of quinine-catalyzed thia-Mi-
chael addition of thiophenols to alkenones. Adapted from reference [4],
compare reference [32].
and reaction conditions. To further support this assumption,
we performed a series of asymmetric oxa-Michael cycliza-
tions with variably functionalized substrates and analyzed
the results in the light of the Hiemstra–Wynberg model
(Table 2): either methyl (1a), benzyl (1b), or tert-butyl
esters (1c) gave similar inductions in the catalytic cyclization
(Table 2, entries 1–4). A larger influence of the steric bulk
of the ester group on the enantioselectivity might have been
expected for a hydrogen-bonding activation mode requiring
close proximity of the carbonyl group of the substrate and
the catalyst.
Table 2. Structure–activity and –selectivity effects in the catalytic oxa-
Michael cyclization.[a]
Role of the hydrogen-bond donor in the catalysis: The key
role of a hydroxy group in the catalyst was evident from the
initial catalyst screening (Table S-1 in the Supporting Infor-
mation) and further reinforced by the loss of activity and se-
lectivity observed by moving from quinine (85% conversion
in 120 h, 72% ee) to O-acetyl quinine (1% conversion in
400 h, 8% ee; Table S-13 in the Supporting Information)[18]
as catalyst. This immediately pointed to the Hiemstra–Wyn-
berg mechanism[4] of cinchona alkaloid catalyzed conjugate
additions of thiophenols to cycloalkenones (Figure 2), where
the same observation had led to the key postulate of a hy-
drogen-bonding activation of the carbonyl group of the sub-
strate by the hydroxyl group of the catalyst. Together with
the primary directing interaction, namely the activation of
the thiol through deprotonation by the lone pair of the qui-
nuclidine, a three-dimensional arrangement of the starting
materials and the catalyst in the transition state could be de-
veloped (Figure 2, TS).[32] The above-described model as-
sumes a stepwise ionic reaction mechanism, where the
stereo-determining and rate-limiting conjugate addition step
generates an enolate, which undergoes fast subsequent pro-
tonation of unknown stereochemistry.[33] Transferring the
Hiemstra–Wynberg model to the cyclization 1!2 appeared
attractive in view of the similarity of the reaction, catalysts
Z (substrate)
Product
T
[8C]
t
Yield
[%]
ee
[h]
[%][b]
1
2
CO2Me (1a)
CO2Me (1a)
CO2Bn[d] (1b)
CO2tBu (1c)
CN (16)
2a
2a
1b
2c
17
RT
50
RT
RT
RT
50
40
99
99
87
80
85
81
62
89 (+)
85 (+)
83 (+)
77 (+)
72 (+)
75 (+)
61 (+)
24
3[c]
4[c]
5
96[e]
<144[e]
<96
25
6
CN (16)
17
7[f]
18[g]
19[g]
45
34
[a] Reaction conditions: CHCl3, 15–22 mol% catalyst, c=8 gLꢀ1. [b] Ee
determined by chiral HPLC,[18] the sign of the optical rotation in
tBuOMe is given in brackets. [c] Concentration 20 gLꢀ1. [d] Bn=benzyl.
[e] Reaction times not optimized. [f] Catalyst was quinine. [g] The struc-
tures of 18 and 19 are shown in Scheme 7.
Remarkably, nitrile 16 cyclized (to compound 17) with
comparable selectivity to the ester substrates (Table 2, en-
tries 5 and 6),[34] even though nitriles are less efficient hydro-
gen-bond acceptors with different stereoelectronic require-
ments than ester carbonyl groups.[35,36] Finally, even the fluo-
renylidene substrate 18, which is devoid of a heteroatomic
2314
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2311 – 2321

Cinchona Alkaloid Catalyzed Oxa-Michael Cyclization
FULL PAPER
adapted for substrates 1 in Figure 3c).[5g,10] This mechanism
is favorable in apolar reaction media or the gas phase, be-
cause it requires no separation of charges. A similar transi-
tion state has been considered theoretically by Merschaert
et al. besides a standard Hiemstra–Wynberg mechanism,[10]
but experimental evidence to discern the alternatives was
not available. In analyzing the stereochemical consequences
of the alternative mechanistic pathways, we notice a distinct
difference (Figure 4): the concerted mechanism (Figure 4a)
Scheme 7. Asymmetric catalytic hydroalkoxylation of the fluorenylene
substrate 18.
hydrogen acceptor group, cyclized to benzofurane 19 with
notable enantioselectivity (Scheme 7; see Table 2, entry 7).
In all of the above-mentioned reactions, the presence of a
free hydroxy group in the catalyst is pivotal, because O-
acetyl-quinine was no efficient catalyst for the cyclization of
either 1a or 18.[37] The conclusion from the experiments
shown in Table 2 is that, although the catalyst truly operates
by a bifunctional activation mode involving the amine base
of the catalyst and the alcoholic hydroxy group, the latter is
not involved in a hydrogen-bonding activation of the car-
bonyl group of the substrate. This follows from the observa-
tion that a carbonyl group (or any other heteroatomic hy-
drogen-bond acceptor group) in the substrate is no require-
ment for a successful catalysis, whereas the presence of a hy-
droxy group in the catalyst clearly is. This immediately sug-
gests that the role of the hydroxy group must be different
from that in the Hiemstra–Wynberg-model.
Figure 4. Stereochemical consequences of the addition mechanisms.
proceeds by a stereospecific, syn-facial 1,2-addition of the
heteroatom X and a hydrogen atom (Figure 4a), whereas
the ionic, stepwise 1,4-addition mechanism concludes with a
diastereoselective protonation of the enolate (Fig-
ure 4b).[41,42] Occurrence of a H X syn addition thus implies
ꢀ
a concerted addition, if the alternative ionic addition mode
does not plausibly display high syn diastereoselectivity. In
fact, ionic conjugate additions of oxygen (and other hetero)
nucleophiles occur either with low diastereoselectivity or a
preference for the anti addition[42,43] We initially determined
the stereoselectivity of protonation in the cyclization 1a!
2a under conditions of the ionic enolate mechanism by per-
forming a H/D exchange reaction on product 2a in basic
MeOD, as described by Mohrig and coworkers.[42a,b] The ob-
served preferential exchange of H-anti over H-syn (ꢃ5:1;
Table 3, entry 1) corresponds to the established diastereose-
lectivity of protonation in intermolecular conjugate addi-
tions.[42a] The same anti preference is observed, albeit dimin-
ished, in a cesium carbonate catalyzed cyclization of the
deuterated substrate [2-D1]-1a in the non-protic solvent
MeCN (Table 3, entry 2). Sparteine is a tertiary amine cata-
lyst lacking a hydroxy group for bifunctional activation. It
catalyzes the reaction without induction (Table 3, entries 3
and 4), and with a low diastereoselectivity of protonation.[44]
In marked contrast, cinchonine catalyzes an enantioselective
cyclization that is almost perfectly diastereoselective
(Table 3, entries 5–8). Within the limits of accuracy of the
NMR integration (Figure 5), the newly entering hydrogen
atom is found at the H-2-syn position of 2a, proving an
almost perfect syn selectivity of the 1,2-addition of RO and
H across the double bond of [2-D1]-1a (Figure 5c). This ex-
cludes an enolate mechanism, but is the predicted result for
a concerted hydrogen-bonding mechanism (see Figures 3c
and 4a).
Relative stereochemistry of hydroalkoxylation: Conjugate
additions of hetero nucleophiles to Michael acceptors are
usually thought to proceed through resonance-stabilized
anionic intermediates. Theoretical and kinetic studies argue
for the possibility of alternative, concerted 1,2-addition
modes with modified transition states including an addition-
al molecule of ROH as mediator in a non-strained, cyclic
transition state (Figure 3a; proton-shuttle or proton-switch
mechanism).[38–40] For amino alcohol catalyzed additions, a
more rigid, extended transition state with a bifurcated hy-
drogen bond can be assumed (general scheme in Figure 3b;
Figure 3. Hydrogen-bond-network mechanisms (proton-shuttle or proton-
switch mechanisms)[38] for concerted additions to polarized double bonds.
a) Hydrogen-bond network mechanism including an R’OH catalyst.
b) Generalized model for amino alcohol catalysis in concerted additions
of hetero nucleophiles RXH to Michael acceptors. c) Plausible transition
state for the cyclization of 1, see reference [10] for a precedence.
Chem. Eur. J. 2013, 19, 2311 – 2321
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2315

L. Hintermann et al.
bifurcate (triangular) transition state;[48] 2) the non-synchro-
Table 3. Diastereoselectivity of the protonation in the cyclization 1a!
2a.[a]
ꢀ
ꢀ
nicity of C O (more advanced) with C H bond formation,
and 3) the presence of a reversible association between RO-
AHCTUNGTREG(NNUN H/D) and the catalyst. Associative pre-equilibria contribute
inversely (<1) to the overall KIE.[49] Overall, the measured
primary KIE is in agreement with a concerted, asynchro-
nous hydrogen-bond network mechanism, but would not ex-
clude a stepwise mechanism by itself.
Catalyst
KOMe[b]
Cs2CO3
sparteine[d]
sparteine[d]
CN
Solvent
t
Yield
[%]
ee
[%]
H-anti/H-syn
[h]
1
2
3
4
5
6
7
MeOD
MeCN
PhCl
CHCl3
PhCl
1.5
1
81[c]
66
44
47
90
90
89
–
–
83:17[c]
58:42
40:60
33:67
1:99
[d]
Control of two adjacent stereogenic centers: The clean syn
addition shown by [2-D1]-1a suggested testing the behavior
of substrates bearing groups other than hydrogen atoms at
C-2. Compound 20 (Scheme 8) bearing a methyl group at C-
28
25
34
34
34
3 (ꢀ)
4 (ꢀ)
88 (+)
88 (+)
90 (+)
CN
CN
CHCl3
CHCl3
1:99
1:99[e]
[a] Reaction conditions: RT, 20 mol% catalyst, concentration of [2-D1]-
1a (98% D at C-2)=8 gLꢀ1 (0.04 molLꢀ1). [b] 10 mol%. [c] Values are
derived from a H/D exchange of 2a in MeOD, yield corresponds to re-
isolated material, see the Supporting Information. [d] 30–40 mol%.
[e] Substrate concentration=4 gLꢀ1 (0.02 molLꢀ1).
Scheme 8. Stereoselective catalytic cyclization of a 2-methyl-substituted
4-(2-hydroxyphenyl)-2-butenoate.
2 was prepared.[18] Its catalytic cyclization with cinchonine
was successful in ODCB at 658C and gave as single product
diastereomer 21 (d.r.>20:1) with an ee of 90%. On the
other hand, the ionic catalyst Cs2CO3 displayed opposite
stereoselectivity in MeOH, producing a 1:5 mixture of dia-
stereomers 21 and 22 (Scheme 8). The analogy with the cyc-
lization of [2-D1]-1a suggests that 21 is the lk diastereomer
(+)-(2S,3S)-21, formed by enantioselective and stereospecif-
ic syn addition, whereas the 22/21 mixture results from dia-
stereoselective protonation of an intermediary enolate with
preference for anti protonation to give the ul diastereomer
(ꢂ)-(2R*,3S*)-22. The cyclization of 20 requires more forc-
ing conditions, because the electron-donating methyl group
opposes build-up of negative charge at C-2, which has carba-
nionic character in the transition state. The low yield of 21/
22 in both reactions is due to competitive base-induced E/Z
isomerization and anionic isomerization to 23 and 24, re-
spectively. Isomer (Z)-23 was independently submitted to
the cyclization conditions with cinchonine as catalyst. Ac-
cording to GC-MS, the reaction generated small amounts of
(E)-20, 21, 22, and 24 (<10% in 4 d at 708C). Presumably,
(Z)-23 cyclizes to 22 very slowly, but its concomitant isomer-
ization to (E)-20 starts a competing cyclization of 20 to 21.
The assignment of the diastereomeric 2-(benzodihydrofur-
an-2-yl)propionates (lk)-21/(ul)-22 is of immediate interest
for the chemistry of some natural benzodihydrofuranyl-pro-
pionate derivatives, such as the antibiotic salfredins A and C
1
Figure 5. Detail of the H NMR (500 MHz, CDCl3) spectrum of 2a, illus-
trating the diastereoselective deuterium incorporation. See Table 3 for
positional numbering and definition of H-syn/H-anti. The signal intensity
for H-3’ at d=2.93 ppm is the same in all spectra. a) Spectrum of all-[H]-
2a. b) Sample of [D1]-2a with H-2-anti/H-2-syn=40:60 (see Table 3,
entry 3). c) Sample of [D1]-2a from a cinchonine-catalyzed cyclization.
Note that the residual signal at d=2.7 ppm (H-2-anti) is largely derived
from residual hydrogen signal in the starting material [D1]-1a at H/D-2
(2% H).
Kinetic isotope effect (KIE): The primary kinetic isotope
effect of the reaction was measured with 1a and 2’-n-butyl-
cinchonine, in both of which the hydroxylic hydrogen atom
had been exchanged with D2O. The result[18] was kH/kD =1.4,
which may appear low for a reaction with a hydrogen trans-
fer in the rate-determining step. However, the value is in
the expected range for a concerted addition reaction, ap-
proaching a reverse E1cb-type mechanism,[43,45,46] or related
6-membered cyclic transition states involving two hydrogen-
bond donors.[39c] A similar primary deuterium KIE of 1.61
has been obtained for the crotonase-catalyzed (enoyl-CoA
hydratase) hydration, which proceeds by a concerted, ster-
ꢀ
eospecific syn addition of water to an unsaturated CoA S-
ester.[47] The low magnitude of the primary isotope effect in
such cases is explained by 1) the non-linearity of C-H-O in a
2316
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2311 – 2321

Cinchona Alkaloid Catalyzed Oxa-Michael Cyclization
FULL PAPER
Figure 6. Structures of naturally occurring 2-(benzodihydrofuran-2-yl)-
propionates.
Figure 7. Transition-state model and rationalization of the substituent ef-
fects. a) Transition-state model of the hydrogen-bond-network mecha-
nism. b) Substituent effects of the catalyst and the substrate on the enan-
tioselectivity. Variations: X=H versus OMe; Y=H versus NO2; RQ =H
versus RQ =alkyl or aryl; Ro =H versus Ro =tBu; Rm =H versus Rm =
acyl (Ac); Rp =OMe; Ra =H versus Ra =Me. In case of moderate or
strong effects, increasing bulk with substitution decreases the ee of the
product.
(25),[50–52] or the ester 26 found in the roots of an analgesic
tropical climbing shrub (Figure 6).[53] The absolute configu-
ration of 26 has been assigned as 2’S with regard to the ben-
zodihydrofuranyl unit, but its relative configuration at C-2
1
was not known.[53] Comparing the H NMR data of (+)-26
with those of 21/22,[18] we find a very close match with the
anti-addition product 22 (see Table S-23 in the Supporting
Information). Therefore, (+)-26 can be assignment the abso-
lute (ul)-(2S,2’R) configuration shown in Figure 6. The same
relative ul configuration has earlier been found in the salfre-
dins A (25) by X-ray crystal structure analysis, though their
absolute configuration is currently unknown.[51]
amino alcohol fragment, rather than being involved in steric
interactions with the substrate.[56]
Generalization—cyclic-hydrogen-bond-network mechanisms
in asymmetric organocatalysis: Concerted hydrogen-bonding
network mechanisms present an alternative to the ionic con-
jugate addition mechanism. It is possible that similar con-
certed hydrogen-bonding network mechanisms are at work
in other organocatalytic reactions. In the following, we
would like to discuss some examples from the literature,
where evidence points to the occurrence of concerted hydro-
gen-bond network mechanisms. For some cases, the evi-
dence is limited, and additional mechanistic work will be re-
quired to discern between ionic stepwise or concerted mech-
anisms, for example, by determination of the relative stereo-
chemistry (diastereoselectivity) of the protonation.[55] A rel-
atively clear-cut case appears to be that of the Cope
hydroamination (“retro-Cope elimination”), which is a stoi-
chiometric addition of N-alkyl hydroxylamines to alkenes
(Scheme 9).[57] Its concerted mechanism is supported by the
syn specificity for either intramolecular (Scheme 9a)[58] or
intermolecular additions to Michael acceptor alkenes (Sche-
me 9b),[59] or by mechanistic calculations.[57b] Particularly rel-
evant to organocatalysis is the work of Pꢃte et al. on amino
Discussion
Cyclic, concerted transition states in hydrogen-bonding cat-
alysis: We have introduced a new asymmetric oxa-Michael
cyclization catalyzed by the cinchona alkaloids. Besides the
synthetic potential of this methodology, important stereo-
chemical details of this process have been elucidated, which
place restrictions on potential transition-state models.
Strong evidence for a concerted, asynchronous hydrogen-
bond network mechanism with limited degrees of freedom
in the transition state has been provided. The key arguments
are 1) the observation of an almost perfect syn selectivity,
which points to a stereospecific reaction mechanism and
2) evidence for the absence of a hydrogen-bonding activa-
tion of the carbonyl group of the substrate, which would be
required in an enolate-type mechanism. Both observations
are opposed to predictions based on the Hiemstra–Wynberg
reaction mechanism,[4,54] which consequently cannot apply
for the present reaction.[55] A model of our proposed transi-
tion state is shown in Figure 7. This model is consistent with
the substituent effects on the enantioselectivity observed in
either the substrates or the catalyst (Figure 7b): groups that
interfere with the central hydrogen-bond network display a
negative effect on the enantioselectivity. For example,
whereas Me-C-2 and Z are perpendicular to the central hy-
drogen-bond network and thus show little effect, the reac-
tion is sensitive to the presence of a large tBu group next to
the phenolic hydroxy group (see 1e) that opposes the ap-
proach of substrate 1e and the catalyst. The key stereo-de-
termining interaction appears to be the proper placement of
the aryloxy fragment of the substrate in a chiral cleft created
by the quinuclidinyl substituents RN, RN’, and R2 (Figure 7a).
The group R1 serves as a conformational anchor for the
Scheme 9. a) Stereospecific intramolecular “retro-Cope elimination”.[58]
b) Cope-type hydroamination of an acceptor alkene.[59]
Chem. Eur. J. 2013, 19, 2311 – 2321
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2317

L. Hintermann et al.
alcohol catalyzed asymmetric tautomerization of photo-
chemically generated enols,[60] which has been discussed in
the framework of asymmetric protonation.[61] The reaction is
more selective in apolar media and is believed to proceed
by
a hydrogen-bond network mechanism rather than
through ionic intermediates.[60a] A cyclic transition state with
a 9-membered ring has been favored (Figure 8a). The au-
Scheme 10. Catalytic reactions displaying syn selectivity. a) Catalytic syn-
specific hydroacylation and a concerted transition state proposed by
Rovis et al.[65] b) Remarkable syn selectivity observed in a metal-cata-
lyzed hydroalkoxylation.[11]
for related metal-catalyzed hydroaminations have been pro-
posed,[66] which essentially correspond to the proton-switch
mechanism in Scheme 3a, if H of the first RXH is replaced
by a metal (M). In a final example, Luo and coworkers have
speculated that an asymmetric organocatalytic Michael addi-
tion of indoles to unsaturated aldehydes involves a hydro-
gen-bond network including a single water molecule in the
rate-limiting proton transfer step.[67] Apparently, the small
amounts of water present in regular organic solvents were
often sufficient to provide this water molecule. To test the
potential role of water in our catalysis, we have carried out
the cyclization 1a!2a (CN as catalyst) in dry (1 ppm H2O
by Karl Fischer titration) and water-saturated chloroform,
but have not found a significant difference of the kinetics,
and only a slightly higher enantioselectivity in the dry sol-
vent (89.5% ee vs. 87% ee in the wet solvent), excluding a
key role of water in our mechanism.
Figure 8. Various hydrogen-bond network transition states. a) 9-Mem-
bered cyclic transition state proposed for the tautomerization of photoen-
ols.[60a] b) Alternative 8+5 bicyclic transition state considered in referen-
ce [60a]. c) Formal, generalized transition state for 1,3-H-shifts.
d) Formal transition state for amino alcohol catalyzed 1,2-additions of
hetero nucleophiles R1XH to carbonyl groups.
thors also considered a bifurcated 8+5-membered transition
state (Figure 8b),[60c,62] but in any case this arrangement
cannot be realized with tertiary amino alcohols as catalysts.
In view of our findings, a bifurcate 6+5-membered transi-
tion state (Figure 8c; Y=O) could also be considered. The
bifurcate 5+6 bicyclic hydrogen-bond network mechanism
can be formally generalized to describe amino alcohol medi-
ꢀ
ꢀ
ated 1,3-H-shifts in HY CR=CHR systems to Y=C(R)
CHR, including allylic shifts in all-carbon systems (Y=
CR’R’’; see Figure 8c).[63] Likewise, asymmetric amino alco-
hol mediated additions of hetero nucleophiles R1XH to car-
bonyl groups can principally follow a similar mechanism
(Figure 8d), for example, in the asymmetric alcoholysis of
meso anhydrides.[64] Although formally possible, it is not
clear to what extent such bifurcate transition states are com-
petitive with the currently preferred monocyclic, ionic
ones;[64e] nevertheless, they are mentioned here, because it
will be desirable to include and evaluate them in future in
silico studies. Returning to experimental evidence, some
syn-stereospecific, catalytic alkene addition reactions have
been reported (Scheme 10): Rovis et al. have found stereo-
specific hydroacylations catalyzed by N-heterocyclic car-
benes.[65] The syn specificity was ascribed to either a two-
step ionic mechanism with fast proton transfer to a carbon
atom, or by a concerted transition state similar to the Cope
hydroamination (Scheme 10a). An interesting observation
of a syn hydration of a Michael acceptor was reported by
Roelfes and coworkers:[11] conjugate addition of D2O to a
Michael acceptor, if catalyzed by a copper/ligand/DNA com-
plex, led to a D/OD syn-addition product (Scheme 10b).
The cause of the syn selectivity is not yet known. It is inter-
esting to note, though, that concerted mechanistic models
Conclusion
We have introduced a new cinchona alkaloid catalyzed
asymmetric oxa-Michael cyclization that gives benzodihy-
drofuranyl acetates in up to 91% ee. Through deuterium-la-
beling and substrate variation studies, we have experimen-
tally distinguished two mechanisms for cinchona alkaloid
catalyzed conjugate additions of hetero nucleophiles to Mi-
chael acceptors. Our 5-exo-trig oxa-Michael cyclization does
not follow the Hiemstra–Wynberg model of bifunctional cat-
alysis, but displays an alternative syn-facial addition mecha-
nism without carbonyl group participation. This is most
plausibly explained by a concerted, but not necessarily syn-
chronous, proton-switch mechanism. The hydroxy group of
the cinchona alkaloid acts as a general acid in the protona-
tion of the evolving product carbanion within a hydrogen-
bond network. Anslyn and co-workers have recently studied
the influence of hydrogen-bonding activation of carbonyl
groups in carbonyl compounds on the rate of enolization
and have concluded: “.. that catalysts of enolate formation
2318
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2311 – 2321

Cinchona Alkaloid Catalyzed Oxa-Michael Cyclization
FULL PAPER
[11] a) R. P. Megens, G. Roelfes, Chem. Commun. 2012, 48, 6366; b) A. J.
Boersma, D. Coquiꢃre, D. Geerdink, F. Rosati, B. L. Feringa, G.
Roelfes, Nat. Chem. 2010, 2, 991.
will be most effective if the binding groups are focused on
stabilizing negative charge that is forming on the enolate
carbon rather than on the enolate oxygen”.[68] A related
case could apply for the hydrogen-bond-catalyzed genera-
tion of enolates by conjugate addition. In any case, the role
of hydroxy groups in bifunctional catalysts should not be
considered to be limited to carbonyl group activation. The
relative stereochemistry of protonation in catalytic Michael
additions contains important mechanistic information. Addi-
tional mechanistic and in particular theoretical studies are
required to define the nature and scope of 1,2-syn-addition
mechanisms in amino alcohol catalyzed reactions.
[12] a) L. Hintermann, C. Dittmer, Eur. J. Org. Chem. 2012, 5573; b) C.
Dittmer, G. Raabe, L. Hintermann, Eur. J. Org. Chem. 2007, 5886.
[13] a) T. Okamura, K. Asano, S. Matsubara, Chem. Commun. 2012, 48,
5076; b) K. Asano, S. Matsubara, J. Am. Chem. Soc. 2011, 133,
16711; c) P. G. McGarraugh, S. E. Brenner-Moyer, Org. Lett. 2011,
13, 6460; d) Q. Gu, Z.-Q. Rong, C. Zheng, S.-L. You, J. Am. Chem.
Soc. 2010, 132, 4056; e) A. Pohjakallio, P. M. Pihko, U. M. Laitinen,
Chem. Eur. J. 2010, 16, 11325; f) D. Dꢄez, M. G. NfflÇez, A. BenØitez,
R. F. Moro, I. S. Marcos, P. Basabe, H. B. Broughton, J. G. Urones,
Synlett 2009, 390; g) N. R. Andersen, S. G. Hansen, S. Bertelsen,
K. A. Jørgensen, Adv. Synth. Catal. 2009, 351, 3193; h) N. Saito, A.
Ryoda, W. Nakanishi, T. Kumamoto, T. Ishikawa, Eur. J. Org.
Chem. 2008, 2759; i) D.-R. Li, A. Murugan, J. R. Falck, J. Am.
Chem. Soc. 2008, 130, 46; j) S. Bertelsen, P. DinØr, R. L. Johansen,
K. A. Jørgensen, J. Am. Chem. Soc. 2007, 129, 1536.
Acknowledgements
[14] Reviews: a) C. F. Nising, S. Brꢀse, Chem. Soc. Rev. 2012, 41, 988;
b) C. F. Nising, S. Brꢀse, Chem. Soc. Rev. 2008, 37, 1218; c) M. G.
NfflÇez, P. Garcꢄa, R. F. Moro, D. Dꢄez, Tetrahedron 2010, 66, 2089.
[15] J. Ackerstaff, Diploma Thesis, RWTH Aachen (Germany), 2004.
[16] K. M. Rupprecht, J. Boger, K. Hoogsteen, R. B. Nachbar, J. P.
Springer, J. Org. Chem. 1991, 56, 6180.
L.H. is grateful to the DFG (Emmy-Noether-Programm) for funding,
and to Prof. Dr. Carsten Bolm, RWTH Aachen, for generous support.
We thank Martial Aime Wankeu (research student, 2004) and Tanja
Kremer (lab trainee, 2004), for performing supportive experiments.
[17] T. Kometani, Y. Takeuchi, E. Yoshii, J. Org. Chem. 1983, 48, 2630.
[18] See the Supporting Information for additional information.
[19] S. France, D. J. Guerin, S. J. Miller, T. Lectka, Chem. Rev. 2003, 103,
2985.
[1] M. E. Jung in Comprehensive Organic Synthesis Vol. 4 (Eds.: B. M.
Trost, M. F. Semmelhack, I. Fleming), Pergamon Press, New York,
1991, pp. 30–41.
[2] J.-X. Ji, A. S. C. Chan in Catalytic Asymmetric Synthesis (Ed.: I.
Ojima), Wiley, New York, 2010, p. 439.
[3] a) S. C. Jha, N. N. Joshi, ARKIVOC 2002, vii, 167; b) N. Krause, A.
Hoffmann-Rçder, Synthesis 2001, 171.
[20] a) T. Marcelli, H. Hiemstra, Synthesis 2010, 1229; b) K. Kacprzak, J.
Gawronski, Synthesis 2001, 961; c) H. Wynberg, Top. Stereochem.
1986, 16, 87; d) S. Tian, Y. Chen, J. Hang, L. Tang, P. McDaid, L.
Deng, Acc. Chem. Res. 2004, 37, 621.
[4] H. Hiemstra, H. Wynberg, J. Am. Chem. Soc. 1981, 103, 417.
[5] Some examples where the presence of OH or NH donors was found
crucial for selective catalysis with cinchona alkaloids: a) T. Marcelli,
R. N. S. van der Haas, J. H. van Maarseveen, H. Hiemstra, Angew.
Chem. 2006, 118, 943; Angew. Chem. Int. Ed. 2006, 45, 929; b) F.
Wu, H. Li, R. Hong, L. Deng, Angew. Chem. 2006, 118, 961; Angew.
Chem. Int. Ed. 2006, 45, 947; c) B. Tçrçk, M. Abid, G. London, J.
Esquibel, M. Tçrçk, S. C. Mhadgut, P. Yan, G. K. S. Prakash, Angew.
Chem. 2005, 117, 3146; Angew. Chem. Int. Ed. 2005, 44, 3086;
d) M. M. H. Verstappen, G. J. A. Ariaans, B. Zwanenburg, J. Am.
Chem. Soc. 1996, 118, 8491; e) O. Riant, H. B. Kagan, L. Ricard, Tet-
rahedron 1994, 50, 4543; f) A. Kumar, R. V. Salunkhe, R. A. Rane,
S. Y. Dike, J. Chem. Soc. Chem. Commun. 1991, 485; g) O. Piva, R.
Mortezaei, F. Henin, J. Muzart, J.-P. Pete, J. Am. Chem. Soc. 1990,
112, 9263; h) A. Sera, K. Takagi, H. Katayama, H. Yamada, J. Org.
Chem. 1988, 53, 1157; i) M. Aune, A. Gogoll, O. Matsson, J. Org.
Chem. 1995, 60, 1356.
[21] a) G. A. Kraus, J. Shi, J. Org. Chem. 1990, 55, 1105; b) G. A. Kraus,
M. T. Molina, J. A. Walling, J. Org. Chem. 1987, 52, 1273; c) G. A.
Kraus, M. T. Mollin, J. A. Walling, J. Chem. Soc. Chem. Commun.
1986, 1568.
[22] K. Krohn, G. Schꢀfer, Liebigs Ann. 1996, 265.
[23] The isolation of (E)-1 f rather than the Z isomer could hint towards
a non-cyclic transition state in the reaction of compounds 9 and 10.
[24] M. A. Brimble, M. R. Nairn, H. Prabaharan, Tetrahedron 2000, 56,
1937.
[25] Our synthesis is not strictly a formal total synthesis of nanaomycin,
because Kraus started with methyl ester 8, whereas we prepared the
tBu ester for practical reasons.
[26] Available from catechol in three simple steps, see the Supporting In-
formation.
[27] The highest ee reported by Merschaert et al. was 52% for an E sub-
strate, or 80% (with only 20% conversion) for a Z substrate lacking
the ether oxygen atom.[10]
[6] a) Hydrogen Bonding in Organic Synthesis (Ed.: P. Pihko), Wiley-
VCH, Weinheim, 2009; b) M. S. Taylor, E. N. Jacobsen, Angew.
Chem. 2006, 118, 1550; Angew. Chem. Int. Ed. 2006, 45, 1520; c) T.
Akiyama, J. Itoh, K. Fuchibe, Adv. Synth. Catal. 2006, 348, 999;
d) P. M. Pihko, Angew. Chem. 2004, 116, 2110; Angew. Chem. Int.
Ed. 2004, 43, 2062; e) P. Schreiner, Chem. Soc. Rev. 2003, 32, 289.
[7] a) R. R. Knowles, S. Lin, E. N. Jacobsen, J. Am. Chem. Soc. 2010,
132, 5030; b) T. Mita, E. N. Jacobsen, Synlett 2009, 1680; c) S. E. Re-
isman, A. G. Doyle, E. N. Jacobsen, J. Am. Chem. Soc. 2008, 130,
7198; d) I. T. Raheem, P. S. Thiara, E. A. Peterson, E. N. Jacobsen, J.
Am. Chem. Soc. 2007, 129, 13404.
[28] a) T. Marcelli, WIREs Comput. Mol. Sci. 2011, 1, 142; b) N. Celebi-
OlÅꢁm, V. Aviyente, K. N. Houk, J. Org. Chem. 2009, 74, 6944;
c) C. S. Cucinotta, M. Kosa, P. Melchiorre, A. Cavalli, F. L. Gervasio,
Chem. Eur. J. 2009, 15, 7913; d) H. Li, X. Liu, F. Wu, L. Tang, L.
Deng, Proc. Natl. Acad. Sci. USA 2010, 107, 20625; e) P. Hammar,
T. Marcelli, H. Hiemstra, F. Himo, Adv. Synth. Catal. 2007, 349,
2537.
[29] L. Hintermann, M. Schmitz, U. Englert, Angew. Chem. 2007, 119,
5256; Angew. Chem. Int. Ed. 2007, 46, 5164.
[30] a) G. Uccello-Barretta, L. D. Bari, P. Salvadori, Magn. Reson. Chem.
1992, 30, 1054; b) T. Williams, R. G. Pitcher, P. Bommer, J. Gutzwil-
[8] A. Berkessel in Methods of Organic Chemistry (Houben–Weyl),
Vol. E21, 4th ed., Thieme, Stuttgart, 1995, p. 4818.
ler, M. Uskokovic, J. Am. Chem. Soc. 1969, 91, 1871; c) Ref.[4]
.
[31] C. Laurence, M. Berthelot, J. Graton, Hydrogen-bonded Complexes
of Phenols in The Chemistry of Phenols (Ed.: Z. Rappoport), Wiley,
New York, 2003; p. 529.
[32] Note, however, that the widely referenced transition-state structure
TS[4] was actually rejected in favor of a modified one in a subse-
quent paper: G. D. H. Dijkstra, R. M. Kellog, H. Wynberg, Rec.
Trav. Chim. Pays-Bas 1989, 108, 195.
[9] a) E. Sekino, T. Kumamoto, T. Tanaka, T. Ikeda, T. Ishikawa, J. Org.
Chem. 2004, 69, 2760; b) T. Tanaka, T. Kumamoto, T. Ishikawa, Tet-
rahedron: Asymmetry 2000, 11, 4633; c) T. Tanaka, T. Kumamoto, T.
Ishikawa, Tetrahedron Lett. 2000, 41, 10229; d) T. Ishikawa, Y. Oku,
T. Tanaka, T. Kumamoto, Tetrahedron Lett. 1999, 40, 3777.
[10] A. Merschaert, P. Delbeke, D. Daloze, G. Dive, Tetrahedron Lett.
2004, 45, 4697.
Chem. Eur. J. 2013, 19, 2311 – 2321
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2319

L. Hintermann et al.
[33] The relative stereochemistry of the ArSH addition to cyclohexe-
none, namely the stereoselectivity of protonation, was not deter-
mined or discussed in the work of Wynberg.
[34] The absolute configurations of (+)-17 and (+)-2a are the same by
chemical correlation, see the Supporting Information.
[35] Hydrogen bonding is preferred in the direction of the available lone
pairs. The hydrogen-bonding basicity pKHB is 1.09 for ethyl acetate
and 0.90 for acetonitrile. For hydrogen-bonding properties of ni-
triles, see: J.-Y. Le Questel, M. Berthelot, C. Laurence, J. Phys. Org.
Chem. 2000, 13, 347.
[36] For hydrogen-bonding properties of ester carbonyl groups, see:
a) J. P. Lommerse, S. L. Price, R. Taylor, J. Comput. Chem. 1997, 18,
757; b) F. Besseau, C. Laurence, M. Berthelot, J. Chem. Soc. Perkin
Trans. 2 1994, 485.
b) D. Leow, S. Lin, S. K. Chittimalla, X. Fu, C.-H. Tan, Angew.
Chem. 2008, 120, 5723; Angew. Chem. Int. Ed. 2008, 47, 5641.
[47] a) B. J. Bahnson, V. E. Anderson, Biochemistry 1989, 28, 4173;
b) B. J. Bahnson, V. E. Anderson, Biochemistry 1991, 30, 5894;
c) W.-J. Wu, Y. Feng, X. He, H. A. Hofstein, D. P. Raleigh, P. J.
Tonge, J. Am. Chem. Soc. 2000, 122, 3987.
[48] F. Westheimer, Chem. Rev. 1961, 61, 265.
[49] B. Dmuchovsky, B. D. Vineyard, F. B. Zienty, J. Am. Chem. Soc.
1964, 86, 2874.
[50] K. Matsumoto, K. Nagashima, T. Kamigauchi, Y. Kawamura, Y.
Yasuda, K. Ishii, N. Uotani, T. Sato, H. Nakai, Y. Terui, J. Kikuchi,
Y. Ikenisi, T. Yoshida, T. Kato, H. Itazaki, J. Antibiot. 1995, 48, 439.
[51] T. Yoshida, T. Kato, Y. Kawamura, K. Matsumoto, H. Itazaki,
EP459449A1, 1991.
[37] The catalytic activity of O-acetylquinine in the cyclization of 1a can
be increased by adding trifluoroethanol as external hydrogen-bond
donor to the catalysis. See Table S-13 in the Supporting Information.
[38] a) L. Pardo, R. Osman, H. Weinstein, J. R. Rabinowitz, J. Am.
Chem. Soc. 1993, 115, 8263; b) C. F. Bernasconi, D. E. Fairchild, C. J.
Murray, J. Am. Chem. Soc. 1987, 109, 3409; c) C. F. Bernasconi, Tet-
rahedron 1989, 45, 4017; d) L. Simón, F. M. MuÇiz, S. Sµez, C.
Raposo, J. R. Morµn, ARKIVOC 2007, iv, 47; e) J. Hine, J. Am.
Chem. Soc. 1972, 94, 5766.
[39] 1,2-Additions to strained C=C or C=X bonds in ketenes, cumulenes,
silenes (Si=C), or (Z)-cyclohexenones by proton-shuttle mechanisms
have also been investigated: a) G. Raspoet, M. T. Nguyen, S. Kelly,
A. F. Hegarty, J. Org. Chem. 1998, 63, 9669; b) K. Sung, T. T. Tid-
well, J. Am. Chem. Soc. 1998, 120, 3043; c) K. B. Wiberg, Y. Wang,
S. J. Miller, A. L. A. Puchlopek, W. F. Bailey, J. D. Fair, J. Org.
Chem. 2009, 74, 3659; d) W. J. Leigh, T. R. Owens, M. Bendikov,
S. S. Zade, Y. Apeloig, J. Am. Chem. Soc. 2006, 128, 10772; e) R.
Kowalewski, P. Margaretha, Helv. Chim. Acta 1992, 75, 1925.
[40] Theoretical studies of the related thia-Michael addition have consid-
ered ionic or concerted additions, including a 1,2-addition of RSH
to alkenes for unsaturated esters. The lower stability of the ester
enol (relative to the aldehyde and ketone enol) could favor a con-
certed 1,2-addition: a) D. Mulliner, D. Wondrousch, G. Schꢁꢁrmann,
Org. Biomol. Chem. 2011, 9, 8400; b) A. Paasche, M. Schiller, T.
Schirmeister, B. Engels, ChemMedChem 2010, 5, 869.
[52] T. Neumann, B. Schlegel, P. Hoffmann, S. Heinze, U. Grꢀfe, J. Basic
Microbiol. 1999, 39, 357.
[53] T. J. Schmidt, M. R. Hildebrand, G. Willuhn, Planta Med. 2003, 69,
258.
[54] Two referees have pointed to the argument of Hiemstra and Wyn-
berg that the presence of an intramolecular hydrogen bond (OH to
N of quinuclidine) in the 9-epi cinchona alkaloids causes inferior
catalytic activity in the thia-Michael addition.[4] One referee propos-
es that, because an intramolecular hydrogen bond does not interfere
with the hydrogen-bond-network mechanism, the epi-alkaloids could
be suitable catalysts for the oxa-Michael cyclization. The other refer-
ee suggests that the lack of an intramolecular hydrogen bond in the
ground state of the regular alkaloids should correlate with a loose-
ness of the hydrogen-bond-network in the cyclic mechanism, which
might be a critical aspect of that reaction mechanism. We plan to
follow those interesting suggestions in future work, including a com-
putational study.
[55] This statement immediately opens the question whether the Hiem-
stra–Wynberg mechanism[4] is truly operating in the cinchona alka-
loid catalyzed conjugate addition of thiophenols to cycloalkenones,
for which it was originally proposed. This question has been investi-
gated by analyzing the relative stereochemistry of protonation. Re-
sults will be reported in a separate study: L. Hintermann, unpublish-
ed results.
[56] The 1H NMR spectrum of a mixture of equimolar amounts of cin-
chonidine and 1a in CDCl3 shows only notable change of the shift
[41] Stereoselectivity of protonation of enolates: a) H. E. Zimmerman,
Acc. Chem. Res. 1987, 20, 263; b) S. Hꢁnig, Protonation of Carban-
ions and Polar Double Bonds in Houben–Weyl, Methods of Organic
Chemistry, Vol. E21D7, Thieme, Stuttgart, 1996, p. 3851.
of
dACHTNGUTERN(UNG OMe) of 1a from d=3.72 ppm (pure substance) to d=
3.55 ppm in the mixture. This implies that already the complexation
of 1a and the alkaloid in the ground state places the OMe group of
1a into the vicinity of the quinolinyl group of the alkaloid. Signals
for the OH groups are very broad (4–6 ppm) in the mixture.
[57] a) N. J. Cooper, D. W. Knight, Tetrahedron 2004, 60, 243; b) J.
Moran, S. I. Gorelsky, E. Dimitrijevic, M.-E. Lebrun, A.-C. BØdard,
C. SØguin, A. M. Beauchemin, J. Am. Chem. Soc. 2008, 130, 17893.
[58] E. Ciganek, J. M. Read, J. C. Calabrese, J. Org. Chem. 1995, 60,
5795.
[59] D. Niu, K. Zhao, J. Am. Chem. Soc. 1999, 121, 2456.
[60] Photoenolization/tautomerization: a) reference [5g]; b) J. Muzart, F.
HØnin, J.-P. Pꢃte, A. M’Boungou-M’Passi, Tetrahedron: Asymmetry
1993, 4, 2531; c) F. HØnin, S. LØtinois, J. Muzart, Tetrahedron: Asym-
metry 2000, 11, 2037.
[42] a) J. R. Mohrig, R. E. Rosenberg, J. W. Apostol, M. Bastienaansen,
J. W. Evans, S. J. Franklin, C. D. Frisbie, S. S. Fu, M. L. Hamm, C. B.
Hirose, D. A. Hunstad, T. L. James, R. W. King, C. J. Larson, H. A.
Latham, D. A. Owen, K. A. Stein, R. Warnet, J. Am. Chem. Soc.
1997, 119, 479; b) J. R. Mohrig, S. S. Fu, R. W. King, R. Warnet, G.
Gustafson, J. Am. Chem. Soc. 1990, 112, 3665; c) J. R. Mohrig, K. A.
Moerke, D. L. Cloutier, B. D. Lane, E. C. Person, T. B. Onasch, Sci-
ence 1995, 269, 527; d) J. R. Mohrig, P. K. Lee, K. A. Stein, M. J.
Mitton, R. E. Rosenberg, J. Org. Chem. 1995, 60, 3529; e) H. Hart,
E. Dunkelblum, J. Am. Chem. Soc. 1978, 100, 5141.
[43] Variable syn/anti ratios have been observed in a similar hydroxyal-
kenoate cyclization in protic solution: T. L. Amyes, A. J. Kirby, J.
Am. Chem. Soc. 1988, 110, 6505.
[44] Relatively higher amounts of syn-addition products can be expected
in oxa-Michael additions under conditions of ion pairing in non-
protic media.[42] An example with very high syn selectivity (ꢁ90%;
large error margin) is reported for a buffer-catalyzed cyclization of a
[61] C. Fehr, Angew. Chem. 1996, 108, 2726; Angew. Chem. Int. Ed.
Engl. 1996, 35, 2566.
[62] For yet another inverted nine-membered transition state in an ion
pair, see reference [60c].
[63] Cinchona alkaloid catalyzed allylic double-bond shifts: a) referen-
ce [5i]; b) L. Hintermann, M. Schmitz, Adv. Synth. Catal. 2008, 350,
1469; c) Y. Wu, R. P. Singh, L. Deng, J. Am. Chem. Soc. 2011, 133,
12458.
[64] a) J. Hiratake, M. Inagaki, Y. Yamamoto, J. Oda, J. Chem. Soc.
Perkin Trans. 1 1987, 1053; b) C. Bolm, I. Schiffers, C. L. Dinter, A.
Gerlach, J. Org. Chem. 2000, 65, 6984; c) C. Bolm, I. Atodiresei, I.
Schiffers, Org. Synth. 2005, 82, 120; d) Y. Chen, P. McDaid, L. Deng,
Chem. Rev. 2003, 103, 2965; e) B. Dedeoglu, S. Catak, K. N. Houk,
V. Aviyente, ChemCatChem 2010, 2, 1122.
substituted 4-(2’-hydroxyphenyl)-2-butenoic acid at
a buffer/sub-
strate ratio of approximately 25:1.[43]
[45] Small KIEs of approximately 2 have been obtained for E1cb or
E1cb-like E2 eliminations: a) M. Schlosser, V. Ladenberger, Chem.
Ber. 1971, 104, 2873; b) M. P. Cooke, J. L. Coke, J. Am. Chem. Soc.
1968, 90, 5556; c) J. Sicher, Angew. Chem. 1972, 84, 177; Angew.
Chem. Int. Ed. Engl. 1972, 11, 200.
[46] KIE for an aza-Michael addition: 1.25–2.01; for a thia-Michael reac-
tion: 1.50: a) H. K. Oh, Bull. Korean Chem. Soc. 2008, 29, 1195;
2320
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 2311 – 2321

Cinchona Alkaloid Catalyzed Oxa-Michael Cyclization
FULL PAPER
[65] a) J. Read de Alaniz, T. Rovis, J. Am. Chem. Soc. 2005, 127, 6284;
b) Q. Liu, T. Rovis, Org. Lett. 2009, 11, 2856; c) J. Read de Alaniz,
T. Rovis, Synlett 2009, 1189.
[66] a) K. Manna, S. Xu, A. D. Sadow, Angew. Chem. 2011, 123, 1905;
Angew. Chem. Int. Ed. 2011, 50, 1865; b) J. F. Dunne, D. B. Fulton,
A. Ellern, A. D. Sadow, J. Am. Chem. Soc. 2010, 132, 17680; c) S.
Tobisch, Chem. Eur. J. 2011, 17, 14974; d) M. Arrowsmith, M. R.
Crimmin, A. G. M. Barrett, M. S. Hill, G. Kociok-Kçhn, P. A. Proco-
piou, Organometallics 2011, 30, 1493; e) M. R. Gagne, C. L. Stern,
T. J. Marks, J. Am. Chem. Soc. 1992, 114, 275.
[67] N. Fu, L. Zhang, J. Li, S. Luo, J.-P. Cheng, Angew. Chem. 2011, 123,
11653; Angew. Chem. Int. Ed. 2011, 50, 11451.
[68] Z. Zhong, T. S. Snowden, M. D. Best, E. V. Anslyn, J. Am. Chem.
Soc. 2004, 126, 3488.
Received: October 2, 2012
Published online: January 4, 2013
Chem. Eur. J. 2013, 19, 2311 – 2321
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2321