Mechanistic investigations of multidentate organocatalyst-promoted counterion catalysis for fast and enantioselective aza-Morita-Baylis-Hillman reactions at ambient temperature

Article abstract of DOI:10.1039/c0ob00069h

Kinetic experiments were performed on the catalytic cycle of a trifunctional organocatalyst-promoted counterion catalysis of asymmetric aza-Morita-Baylis-Hillman reactions. The catalysis was found to be first order in the trifunctional catalyst with the Michael addition as the rate-limiting step. Temperature variation changed the rate of catalysis but not the enantioselectivity of the reaction.

Full text of DOI:10.1039/c0ob00069h

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Mechanistic investigations of multidentate organocatalyst-promoted
counterion catalysis for fast and enantioselective aza-Morita–Baylis–Hillman
reactions at ambient temperature
Christopher Anstiss, Jean-Marc Garnier and Fei Liu*
Received 30th April 2010, Accepted 22nd June 2010
DOI: 10.1039/c0ob00069h
Kinetic experiments were performed on the catalytic cycle of a trifunctional organocatalyst-promoted
counterion catalysis of asymmetric aza-Morita–Baylis–Hillman reactions. The catalysis was found to be
first order in the trifunctional catalyst with the Michael addition as the rate-limiting step. Temperature
variation changed the rate of catalysis but not the enantioselectivity of the reaction.
Introduction
Rapid advances in asymmetric organocatalysis have presented new
opportunities for the development of enantioselective multiden-
tate organocatalysts as potential enzyme mimics.1–3 The interest in
Scheme 1 A general catalytic cycle of the MBH or azaMBH reaction.
this approach is multifaceted and underpinned by the efficacy of
organocatalysts as green alternatives to traditional approaches.
In addition, these small, multifunctional systems can be used
as potential models for understanding more complex enzymatic
mechanisms and creating novel catalytic modes and motifs that
have no natural counterparts.
The recent emergence of trifunctional organocatalysts4 has
demonstrated that it is possible to instigate positive cooperativity
in systems with higher organization to approach the level of
complexity and proficiency seen in enzyme catalysis. On the other
hand, multifunctional systems, while potentially more versatile
and efficient, are inherently more prone to competition from
lower order or uncooperative catalysis, a problem that places
a much higher demand on the mechanistic understanding of
the catalyst. This challenge has led to careful definitions of the
catalytic role of trifunctional systems by functional studies4a–d
and, in some instances, mechanistic studies to confirm the likely
catalytic pathways in transesterification reactions using a racemic
trifunctional organocatalyst.4a
We have used the aza-Morita–Baylis–Hillman (azaMBH) as
a model system for exploring multifunctional or multidentate
organocatalysis that is not only rate enhancing but also capable
of a high level of asymmetric induction.4c–d The MBH reaction
is a multistep carbon–carbon bond-forming reaction that couples
an enone and an aldehyde (or an imine in the azaMBH reaction)
with the concurrent generation of one chiral alcohol or amine
(Scheme 1). This reaction is initiated by the conjugate addition
between a Lewis base catalyst with an enone to generate a
zwitterionic Michael adduct, which is followed by an aldol
addition between the Michael adduct and an electrophile (either
aldehyde or imine) to generate another zwitterionic aldol adduct.
The final proton transfer step, which is irreversible, eliminates the
Lewis base catalyst and furnishes the MBH or azaMBH adduct.
This atom-economic reaction, given its well recognized synthetic
utility, receives unabated attention but also has long standing
problems of slow reaction rates of days and capricious substrate
scope.5 This limits the process prospect of this important carbon–
carbon bond forming reaction, especially in the asymmetric sense.
Bifunctional asymmetric organocatalysts have been developed
for the MBH or azaMBH reaction.6 In the bifunctional approach,
a Lewis base and a Brønsted acid can be crafted onto one chiral
backbone to act cooperatively in the MBH reaction cycle. The
Lewis base functionality serves to initiate the Michael addition
step of the reaction, and the Brønsted acidity is thought to stabilize
the zwitterionic intermediates and promote the subsequent aldol
and proton-transfer-elimination step. The bifunctional strategy
has been very successful in addressing some of the scope issues,
while the solution to rate enhancement without loss of enantiose-
lectivity remains elusive.
Equipped with the design insight from bifunctional systems,
we have made the initial demonstration of a trifunctional
organocatalyst-promoted counterion catalysis for this reaction
to address the rate issue without loss of enantioselectivity
(Scheme 2a).4c In this trifunctional system 1, a third functionality,
in the form of a nitrogen Brønsted base, is installed as an
activity switch in response to an additional Brønsted acid. This
activation elevates the catalysis from bifunctional to trifunctional
with significant rate enhancement. Unlike bifunctional catalysis,
where the reaction rate and enantioselectivity vary independently,
this trifunctional strategy has a unique catalytic profile in that
the rate of reaction and enantioselectivity arise in a coordinated
fashion when 1 becomes activated. Catalyst screening established
the optimal spatial orientation and electronic characteristics of
the three functionalities, and led to fast rates and good enan-
tioselectivity without the need to lower the reaction temperature
(Scheme 2b).4c–d Compared to the existing enantioselective bifunc-
tional counterparts that, in general, require half a day to days
for good conversion at lower temperatures, the trifunctional
Department of Chemistry & Biomolecular Sciences, Macquarie University,
Sydney, NSW, 2109, Australia. E-mail: fliu@alchemist.chem.mq.edu.au;
Fax: 61-2-9850-8313; Tel: 61-2-9850-8312
4400 | Org. Biomol. Chem., 2010, 8, 4400–4407
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Results and discussion
There are no precedents of substantial mechanistic studies on
bifunctional or trifunctional organocatalysts for the MBH or
azaMBH reactions. Several experimental mechanistic studies,
however, have been conducted for MBH or azaMBH reactions
with monofunctional Lewis base organocatalysts.7 Most of the
studies employed a nitrogen Lewis base for MBH reactions
between enones and aldehydes. One study utilized triphenyl phos-
phine as the monofunctional Lewis base for azaMBH reactions
between an enone nucleophile (methyl vinyl ketone) and an
imine electrophile. With the use of initial rate kinetics, this study
established that, with an acidic additive, the rate of catalysis was
first order with respect to the monofunctional phosphine catalyst,
the enone nucleophile, and the imine electrophile.7c For system
1 here, initial rate kinetic measurements were likewise performed
to examine the rate of catalysis with respect to the trifunctional
phosphine catalyst, methyl vinyl ketone, and the imine substrate.
1) Scalable synthesis of catalyst 1.
All of the existing mechanistic studies on the MBH or azaMBH
reactions utilized readily available catalysts (e.g. DABCO or
triphenylphosphine) from commercial sources. The trifunctional
system 1 is not commercially available and required optimization
for a gram-scale synthesis so that any one rate experiment
could be conducted without hindrance and with the same batch
of catalyst for each rate plot to maintain consistency in data
collection and interpretation. A scalable synthesis of 1 was
therefore developed from the cheap and commercially available
(S)-BINOL (Scheme 3).
Scheme 2 A trifunctional organocatalyst-promoted asymmetric counte-
rion-directed catalysis of azaMBH reactions.
catalysts 1 can reach completion within hours, and in some cases,
minutes, with no reduction in enantioselectivity. Comprehensive
bifunctional controls of the this system were also constructed
and tested, which showed that each of the three functionalities
of 1 was essential for switching to the trifunctional catalysis.4c
An investigation of the solvent effect found that the trifunctional
catalysis and asymmetric induction remained largely tolerant of
nonpolar aprotic solvents, but were both severely reduced in
polar or protic solvents, which suggested the requirement of a
tight ion-pair for this trifunctional catalysis.4c Various counterions
were also investigated to reveal that benzoic acid provided the
fastest rate with the highest enantioselectivity while, in general,
aryl carboxylates performed better in trifunctional catalysis than
acetates and phosphonates.4c
Due to the multifunctional nature of 1, several questions
emerged after the initial functional studies of this system. The
first and foremost is on the molecularity of the trifunctional
catalysis. While functionally we have established the positive
cooperativity of the phosphine Lewis base, the nitrogen Brønsted
base and the phenolic Brønsted acid, it remains untested if the
positive cooperativity is acting cis (intramolecularly) and not in
trans (intermolecularly). Another question relates to the rate-
limiting step of this trifunctional catalytic cycle. Also, the effect
of temperature on the enantioselectivity and rate of this system
remains unclear. Herein reported are experimental mechanistic
studies on azaMBH reactions catalyzed by 1. In this report,
kinetic experiments are described to address these questions with
additional substrate scope studies to shed light on the likely
operating features of 1. This will allow better design principles
for the next generation catalysts to be developed.
Scheme 3 Scalable synthesis of 1 from (S)-BINOL.
To access the key intermediate (S)-NOBIN, a modified pro-
cedure as described by Salinger and Bru¨ckner was followed.8a
Methoxymethyl protection of triflated (S)-BINOL for the
preparation of 2-methoxymethoxy-2¢-trifluoromethanesulfoxy-
1,1¢-binaphthyl intermediate 7 was found to be sluggish un-
der the reported conditions of phosphorous pentoxide and
dimethoxymethane. Alternatively, 7 was prepared using sodium
hydride and chloromethyl methyl ether as described by Maruoka
and co-workers.8b Furthermore, palladium-catalyzed debenzy-
lation, under the Salinger and Bru¨ckner conditions, to yield
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 4400–4407 | 4401
©

View Article Online
2-methoxymethoxy-2¢-amino-1,1¢-binaphthyl intermediate 8 was
found to suffer over-reduction. This yielded 14% of a side product
with saturated binaphthalene ring at the 6¢ and 8¢ position as
suggested by low resolution ESI mass spectrometry and 1H NMR
spectroscopy. The use of 10 mol% palladium on carbon under
1 atm of hydrogen at 45 ◦C eliminated the formation of this
unwanted side reaction to give the free amine 8 in excellent yield
without the need for column purification. Overall, 5.0 g of (S)-
BINOL was converted to 3.1 g of (S)-NOBIN over five steps in
62% yield using this modified route. The key amino phosphine
compound (S)-MAP was then synthesized on gram-scale in six
steps and 51% overall yield following the procedure reported
by Kocovsky.9 Catalysts 1 were then prepared from (S)-MAP as
previously described.4c–d The same batch of catalyst was used for
each set of experiments for a rate plot with varying concentrations
of one reactant.10
non-enantioselective pathway (without acid activation), which was
tested by the following delayed acid addition experiment.
When the addition of benzoic acid was delayed by 30 min
(Table 1, entry 4), the ee is reduced significantly as the competing
pathway with lower enantioselectivity received a thirty minute
advantage in reaction time. Using the enantiomeric ratios (e.r.) for
1b, the enantioselective pathway with acid activation is estimated
to be about three-fold faster than the racemic pathway without
acid activation.
As a precaution to a potential dilution effect on catalysis in
the delayed-activation experiment, the test reaction from entry 2
was also repeated, with the reaction volume diluted by five-fold
after a few minutes of reaction (Table 1, entry 5). Although the
reaction rate was slightly reduced, the enantioselectivity remained
unchanged. In summary, in the “on” state with acid activation,
the trifunctional catalysis is not only enantioselective but also
kinetically dominating over competing pathways. With this clear
kinetic advantage, the subsequent rate measurements would be
reflective primarily of the trifunctional enantioselective pathway.
2) The role of the acid activation in the induction of asymmetry.
The trifunctional system 1 has two states of activity and switches
from the “off’ state to the “on” state after acid activation. We have
shown in our earlier functional studies that, in the “off” state of 1,
the catalysis is slow and with little enantioselectivity.4c Upon the
addition of an acid, the reaction improved markedly in rate, and
most importantly, proceeded with good to high enantioselectivity.
To ensure that the level of asymmetric induction is indeed coupled
to acid activation of the trifunctional catalyst, a delayed-activation
experiment was performed using catalyst 1b and a test azaMBH
reaction between imine 2a and MVK (Table 1 scheme).
As shown in Table 1, the trifunctional organocatalyst 1b could
confer little enantioselectivity in the “off’ state without acid
activation (Table 1, entry 1). Upon acid activation, catalyst
1b exhibited markedly improved enantioselectivity along with
significant rate enhancement (Table 1, entry 2). The test reaction
with 1b reached full conversion in under 30 min with good
enantioselectivity maintained (Table 1, entry 3). This suggests that
the catalytic pathway with acid activation (highly enantioselective)
is distinctly different from that without acid activation (much
less enantioselective). It was hypothesized that the enantiose-
lective pathway (with acid activation) would be faster than the
3) The reaction is first order with respect to the trifunctional
catalyst.
The multifunctional nature of the trifunctional system 1 presents
the possibility that the roles of the three functionalities may be
fulfilled by more than one catalyst molecule acting in concert.
This would not be discernable in functional studies but readily
determined via measurement of initial rates of the reaction with
various concentrations of the trifunctional catalyst. In order to
minimize error, a relatively slower azaMBH reaction between 4-
bromoaryl imine 2b and MVK was used with a slower catalyst 1c
to ensure rate measurement could be taken in time using NMR
spectroscopy with methyl benzoate as the integration standard for
all rate experiments described in sections 2 to 6 (Fig. 1).
The concentration of the catalyst was varied from 12 mM to
24 mM with the concentrations of the enone and imine substrates
kept constant at 36 mM. The linear relationship between the initial
rate of reaction and the concentration of catalyst 1c indicates that
the trifunctional enantioselective catalytic pathway is first order
in the catalyst (Fig. 2). The positive cooperativity of the three
functionalities is thus uni-molecular.
Table 1
A
delayed-activation experiment using N-(arylmethyl-
idene)arylsulfonamide 2a (1.0 equiv.), MVK (2.0 equiv.), 1b (10 mol%)
4) The reaction is zero order with respect to the imine substrate.
and benzoic acid (10 mol%) in dichloromethane
It has been shown that in azaMBH reactions catalyzed by triphenyl
phosphine, the reaction is first order with respect to the imine
substrate, provided that the proton-transfer-elimination step is
accelerated by acidic additives and no longer rate-limiting.7c The
aldol step between the electrophile and the Michael adduct is
the most typical rate-limiting step in either MBH or azaMBH
reactions for monofunctional organocatalysts such as DABCO or
triphenylphosphine.
Given that in this study only one trifunctional catalyst is
responsible for the activation of the azaMBH cascade and remains
covalently linked to all intermediates throughout the reaction
sequence until the release of the catalyst to complete the catalytic
cycle, the next question became which step in this catalytic cycle
would be rate-limiting. The reaction order in the imine substrate
was thus investigated at a constant concentration of 12 mM
Entry
Time/min
Conv. [%]a
ee [%]b
e.r.
1c
2
3
4
5
90
15
30
60d
30
90
91
16
88
88
48
88
58 : 42
94 : 6
94 : 6
74 : 26
94 : 6
>95
95
87e
a Calculated by 1H NMR spectroscopy. b Determined by chiral HPLC
analysis. c Without benzoic acid addition. d Benzoic acid added with a
30 minute delay after reaction has been initiated by MVK. e Reaction
mixture was diluted by the addition of 4 reaction volumes of CH2Cl2.
4402 | Org. Biomol. Chem., 2010, 8, 4400–4407
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Fig. 1 Representative NMR traces of the reaction between 2b and MVK.
Fig. 3 The catalytic cycle promoted by 1c is zero order with respect to
the imine substrate.
Fig. 2 The reaction cycle promoted by the trifunctional organocatalyst
1c is first order with respect to the catalyst.
interactions between the catalyst and the imine substrate may
reduce slightly the effective concentration of the catalyst at higher
imine concentrations and render the reaction slower. Another
possibility is that the acid additive could also be sequestered
by excess imine, which would lead to less prominence of the
enantioselective trifunctional pathway.
The role of the acid additive in this interference was then
examined with 60 mM imine at 20 mol% acid additive and 10 mol%
loading of catalyst 1c. Excess acid additive did not rescue the rate of
catalysis but again resulted in a reduction of the reaction rate (data
not shown). This suggested the possibility of interference from
unwanted protonation of the Lewis base of the catalyst by excess
acidity. Taken together, the results point to the sensitive nature of
the reaction rate to factors early in the catalytic cycle prior to the
involvement of the imine substrate, indicating that in this catalytic
cycle the Michael addition may be the rate-limiting step. Another
catalyst 1c and 36 mM MVK with varying concentrations of the
imine from 25 mM to 75 mM. The initial rate did not increase but
rather decreased slightly at higher imine concentrations (Fig. 3).
The zero order in the imine substrate suggests first that the aldol
step involving the imine is not rate-limiting in this trifunctional
catalytic cycle. Since the trifunctional catalyst contains a protic
component and is therefore capable of catalyzing the proton
transfer-elimination step, the hypothesized mode of action of the
trifunctional catalyst in promoting the proton transfer step, as
suggested earlier,4c is consistent with the observed zero order in
the imine substrate. It is possible that the aldol step in this catalytic
cycle could also be promoted by the action of catalyst 1c.
Secondly, a slightly faster reaction rate at lower imine concentra-
tions seems to indicate an inhibitory role of excess imine substrate
in the rate-determining step of this catalytic cycle. Noncovalent
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 4400–4407 | 4403
©

View Article Online
Table 2 Parameters from rate plots from Fig. 3 indicating the reaction is
point was that, while required to switch on the catalyst, the amount
of the acid additive should be controlled carefully to preserve the
activity of the Lewis base on the trifunctional catalyst.
likely to be first order
Entry
T/K
Rate/hr-1
R2
ee [%]a
ee [%]b
5) The reaction is first order with respect to MVK.
1
2
3
4
313
303
293
273
0.23
0.10
0.041
0.0051
0.9905
0.9972
0.9397
0.9954
93
94
94
94
90
90
90
91
In the studies of monofunctional organocatalysis of MBH or
azaMBH reactions, the cycle was invariably found to be first
order with respect to the enone nucleophile.7 The reaction order
in the imine or aldehyde electrophile, as discussed earlier, can be
either first or second, with the aldol addition step typically as
rate-limiting.7a,b,d The proton-transfer-elimination step is usually
not rate-limiting when acidic additives are used, or the reaction
has proceeded long enough to exhibit autocatalysis by the action
of the protic MBH adduct.7d In the trifunctional case here, with
constant catalyst concentration of 12 mM and imine concentration
of 36 mM, the reaction order with respect to MVK, varied from
45 mM to 75 mM, was again found to be first order (Fig. 4). Given
that the reaction is zero order with respect to the imine substrate,
the Michael, not the aldol, addition step is now rate-limiting in the
case here. As the Michael step is reversible, the use of the Michael
acceptor in large excess can be employed to saturate the rate of
Michael addition at the expense of efficiency. The implication here
for catalyst design, under economical and practical conditions, is
that the nucleophilicity of the phosphine could be enhanced to
accelerate further the catalysis to attenuate the use of MVK in
large excess.
a Measured at 5–10% conversion by chiral HPLC analysis. b Measured at
>90% conversion by chiral HPLC analysis.
respect to time, and ee analysis was performed at the starting and
finishing stages of the reaction for all temperatures. As the reaction
is first-order with respect to the catalyst 1c, which is kept constant,
the rate of reaction should follow a first order mechanism with
respect to MVK. This is supported by the apparent first-order
kinetic analysis at all four temperatures (Fig. 5 and Table 2).
Fig. 5 The first order plot of the reaction at various temperatures from
0
◦C to 40 ◦C.
While the rate of reaction was lower at lower temperatures
as expected, the enantioselectivity of the reaction intriguingly
was found to be completely insensitive to temperature variation
(Table 2). There was, however, a slight reduction in enantiose-
lectivity at higher conversions of the reaction (Table 2, last two
columns). This leakage at higher conversions could be due to a
larger presence of the MBH product that is itself a Brønsted acid.
This may interfere with hydrogen bonding interactions with the
catalyst that are important in conferring enantioselectivity. Alter-
natively, the enantioselectively formed product may be racemized
by other pathways as previously suggested.7c
Fig. 4 The catalytic cycle promoted by 1c is first order with respect to the
enone substrate MVK.
6) The effect of temperature on the rate and enantioselectivity of
the catalysis.
7) The role of the azaMBH adduct in catalysis.
Bifunctional organocatalysts for the MBH reaction typically
require the use of lower temperatures to access higher levels
of enantioselectivity. The effects of temperature on catalysis are
difficult to resolve, as temperature variations can change confor-
mational arrangement, apparent acidities, and kinetic preferences
for intermediates. Generally, however, enantioselectivity tends to
improve at lower temperatures. The disadvantage of this is that
the reaction is less facile as the rate of reaction is compromised for
enantioselectivity, which limits the utility of the catalyst.
One of the mechanistic intricacies of the MBH reaction is the
autocatalysis from the protic MBH product, which becomes more
evident as the reaction proceeds.7d In order to investigate the
possible role of autocatalysis in the trifunctional organocatalyst-
promoted azaMBH reaction, a test azaMBH reaction was per-
formed using 2a and MVK with 1c as the catalyst at 10 mol% load-
ing (Scheme 4a). This test reaction is one of the fastest azaMBH
reactions, with completion under one hour, and therefore used to
minimize effects from other potential racemization pathways that
may be more pronounced at longer reaction times. The reaction
was spiked with 20 mol% racemic azaMBH adduct 3a at the
In the trifunctional system here, the dependence of the enan-
tioselectivity and rate of reaction on temperature was investigated
from 0 ◦C to 40 ◦C. The reaction conversion was monitored with
4404 | Org. Biomol. Chem., 2010, 8, 4400–4407
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 3 Generic or azaMBH reactions of electrophiles 4 (1.0 equiv.) with
MVK (2.0 equiv.) in dichloromethane with catalyst 1c (10 mol%) and
benzoic acid (10 mol%)
Entry
R
Time/hr
Conv. [%]a
ee [%]b
1
2
p-bromophenyl
3
16
95
44
94
77
c
3
4
16
16
—
—
—
c
—
Scheme 4 The effect the azaMBH adduct on the trifunctional catalytic
cycle.
start to examine if this would result in reduced enantioselectivity
due to the protic product interference from the start. If product
interference, from either enantiomer, were significant, then a
considerable drop in enantioselectivity would manifest, with a
possible change in reaction rate as well. The rate of reaction in
this case was found to be unaffected, and the final apparent ee
at completion was 70%. Given the initial spike of 20 mol% of
racemic product, the expected apparent ee would be 71% if the
enantioselective catalytic pathway was immune to the spike. This
indicates that product interference is an unlikely or negligible event
in this catalytic cycle.
As the azaMBH adduct contains an enone moiety, there exists
the possibility of the trifunctional catalyst reacting with the
product and initiating a reversal of the azaMBH reaction. The
reversibility of this catalytic cycle of the azaMBH reaction was
examined using the same azaMBH adduct 3a (Scheme 4b). The
racemic adduct 3a was treated with only catalyst 1c and benzoic
acid. In order to maximize any potential effect for observation,
the catalyst and benzoic acid were used in excess at 2 equiv. of the
adduct 3a. The reaction was allowed to stand for one week and
found to remain completely unchanged in composition. Enantio-
enriched 3a (88% ee) was also incubated under catalysis conditions
(without the imine substrate to stop conversion of MVK to the
MBH product) for one week and found to retain its ee (Scheme 4c).
The absence of any observable effect of the azaMBH product
during the catalytic time course suggests that the mild reduction
in enantioselectivity over the course of the reaction may be due to
racemization pathways but not due to product interference of the
catalytic cycle itself.
a Calculated by 1H NMR spectroscopy. b Determined by chiral HPLC
analysis. c Starting material recovered.
(Table 3, entry 1),4d these substrates are less reactive and sterically
more demanding (Table 3, entries 2 to 4).
In the case of a less reactive substrate that still contains the a-sp2
carbon characteristic with comparable steric bulk, the catalyst was
able to produce the product at a slower rate and reduced level of
enantioselectivity (Table 3, entry 2). With alkyl imine substrates
that are weaker electrophiles and also bulkier, the catalyst was
not able to convert the substrate to the desired product (Table 3,
entries 3 and 4). Interestingly, these alkyl imines remained intact
during the reaction, which is contrary to the reported instability
of these alkyl imines in bifunctional catalysis of the azaMBH
reaction.12 In general, alkyl imines are known to be less reactive
substrates compared to the aryl imines in the azaMBH pathway,
and this low activity can render these imines more susceptible to
the competing hydrolysis pathway, especially in the presence of
basic reactive species. However, in this trifunctional catalysis, the
low activity of the imine substrate was not accompanied by the
usual hydrolysis bleed-off. The absence of the unwanted hydrolysis
of the imine substrates in this system suggests that, unlike the
bifunctional counterparts, the activity of the trifunctional catalyst
may be restricted to the MBH reaction pathway. The higher level
of catalytic chemospecificity here in turn may also hint that the
activity of the trifunctional catalyst in the MBH pathway may
depend on more stringent molecular recognition of the imine
substrate by the catalyst.
Conclusions
8) Substrate scope profile.
This trifunctional catalytic system has been found to be partic-
ularly effective for a wide range of aryl imines, regardless of the
electronic nature or position of the substituent on the aromatic
ring of the imine. Given the kinetic profile of the catalytic cycle
implicating the role of the catalyst in promoting the proton
transfer step, it would be important to assess the sensitivity of the
catalytic outcome with respect to the substrate. Three additional
imine substrates 4a–c were prepared as previously described.11
Compared to the control reaction with an aryl imine substrate
The complex nature of the MBH reaction with competing auto-
catalysis has made the rate enhancement and enantioselectivity
difficult to attain concurrently for this reaction with generality
in substrate scope. The trifunctional organocatalytic approach
has shown some success in achieving more proficient catalysis
with good enantioselectivity for subclasses of substrates, from
which new generations of catalysts can be developed to expand
its substrate scope. Mechanistic investigations here attribute
the positive cooperativity of the three functionalities to only
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 4400–4407 | 4405
©

View Article Online
one catalyst molecule. With the molecularity of the catalysis
ascertained, future structural and computational studies will be
able to shed light on the likely transition structures for an improved
design of this new catalytic system. The need for a rational
approach is evident as the possibilities for catalyst modification
are vast with multiple functionalities. This study also suggests
that, for some substrates, the rate of the Michael addition could
be enhanced to shunt more of the imine substrate into the faster
enantioselective catalytic cycle, provided that the subsequent steps
after the Michael addition are not rate limiting by the action
of the catalyst. For less reactive imine substrates that are not
recognised by the current generation of trifunctional catalysts,
the limitation may also be due to an uncatalyzed proton transfer
step. This suggests that modification of the Brønsted acid moiety,
and its associated counterion, may be important to enable proper
molecular recognition at the aldol and proton transfer stages for
future generations of catalysts.
preparation of 2-methoxymethoxy-2¢-trifluoromethanesulfoxy-
1,1¢-binaphthyl intermediate 7 is as follows: to a suspension of
sodium hydride (455 mg, 18.9 mmol) in THF (80 mL) at 0 ◦C was
slowly added the precursor 2-hydroxy-2¢-trifluoromethanesulfoxy-
1,1¢-binaphthyl (7.15 g, 17.1 mmol) under nitrogen and stirred
for 30 min. To this yellow mixture chloromethyl methyl ether
(1.43 mL, 18.9 mmol) was added dropwise. The reaction was
stirred at room temperature for 2 h, quenched carefully with water
and extracted with ethyl acetate. The extracts were washed once
with brine, dried over sodium sulfate and concentrated to produce
a viscous oil purified by column chromatography on silica gel with
petroleum ether–ethyl acetate (9 : 1) to give 7 (7.46 g, 16.1 mmol,
94%). The procedure for the preparation of 2-methoxymethoxy-
2¢-amino-1,1¢-binaphthyl intermediate 8 is as follows: to a solution
of 2-N-benzyl-2¢-trifluoromethanesulfoxy-1,1¢-binaphthyl (5.82 g,
13.8 mmol) in ethyl acetate (30 mL) was added palladium (10%
on carbon, 1.41 g, 1.38 mmol). The reaction was heated at 45 ◦C
under H2 (1 atm) for 3 h. After filtration through Celite the solvent
was removed to produce 8 (4.3 g, 12.9 mmol; 94%), which was
used without further purification. The preparation of (S)-MAP
and catalysts 1 followed the procedures as previously described
without change.4c
Experimental
General
Unless otherwise stated, all chemicals and reagents were received
from Sigma Aldrich, Castle Hill, NSW Australia and used
without further purification. Reactions were performed under
a nitrogen atmosphere. Reactions were magnetically stirred and
monitored by thin-layer chromatography (TLC) using silica gel
60 F254 aluminium pre-coated plates from Merck (0.25 mm).
Reaction procedures
azaMBH reactions. An imine (0.015 mmol), phosphine cat-
alyst (10 mol%, 0.0015 mmol) and benzoic acid (10 mol%,
0.0015 mmol) were combined under N2. Dichloromethane (0.1 mL
per mg of catalyst) was added, followed by distilled methyl
vinyl ketone (2 equiv.) dropwise. The reaction was stirred at
room temperature until completion. The reaction mixture was
evaporated to dryness, and the crude mixture was immediately
subjected to chiral HPLC analysis.
˚
Flash column chromatography was performed on silica gel (60 A
0.06–0.2 mm, 400 mesh from Scharlau). 1H NMR and 13C NMR
spectra were obtained in dry CDCl3 on a Bruker Avance DPX
400 MHz spectrometer. Chemical shifts were reported in ppm
using chloroform or dichloromethane as the internal reference
(1H, 7.26 ppm, 13C, 77.09 ppm for CDCl3 and 1H, 5.25 ppm, 13C,
53.8 ppm for CD2Cl2). All spectra were processed using Bruker
TOPSPIN (1.4) and MestRec 4.9.9.6. Kinetics experiments were
performed under dry conditions in a sealed NMR tube on a
Bruker Avance DPX400 or DRX600 NMR spectrometers without
mechanical agitation. Infrared spectra were taken on a Perkin
Elmer paragon 1000PC FTIR spectrometer. Optical rotations
were measured at 23 ◦C on a P1010 digital polarimeter (Jasco,
Japan). High resolution mass analysis was provided by University
of Illinois, Urban-Champaign USA.
General procedure for the experiments on the catalyst order.
Catalyst 1c (4.0–8.0 mg, 6.1–12.1 mmol, 12.1–24.2 mM), benzoic
acid (1 equiv. to the catalyst), methyl benzoate (0.64 mg, 5.6
mmol, 11.2 mM) and 4-bromo-N-tosyl imine (6.2 mg, 18.3 mmol,
36.6 mM) were dissolved in dichloromethane-d2 to afford a total
reaction volume of 0.5 mL. Freshly distilled methyl vinyl ketone
(1.27 mg, 18.1 mmol, 36.3 mM) was added and the solution was
transferred to an NMR tube. The tube was sealed and the time
course of the reaction was followed by 1H NMR spectroscopy.
General procedure for the experiments on the imine order. 4-
bromo-N-tosyl imine 2b (4.2–12.7 mg, 12.4–37.6 mmol, 24.8–
75.1 mM), catalyst 1c (4.0 mg, 6.1 mmol, 12.1 mM), benzoic acid
(1 equiv. to the catalyst) and methyl benzoate (0.61 mg, 5.3 mmol,
10.6 mM) were combined and dissolved in dichloromethane-d2 to
afford a total reaction volume of 0.5 mL Freshly distilled methyl
vinyl ketone (1.27 mg, 18.1 mmol, 36.3 mM) was added and the
solution was transferred to an NMR tube. The tube was sealed
General information for the kinetics experiments.
Unless otherwise stated, all kinetics experiments were performed
under dry conditions in a sealed NMR tube on a Bruker Avance
DRX600 Cryoprobe NMR spectrometer without mechanical
agitation at 301 K. The receiver gain and relaxation time were
set to 128 and 5 s, respectively. Dichloromethane-d2 was pre-
treated by filtering through a plug of activated basic alumina.
Freshly distilled methyl benzoate was used as a standard, with
its integration normalized internally against the concentration of
imine and MVK. Liquid transfer technique from stock solutions
was employed for quantities less than 10 mg.
1
and the time course of the reaction was followed by H NMR
spectroscopy.
General procedure for the experiments on the methyl vinyl
ketone order. Catalyst 1c (4.0 mg, 6.1 mmol, 12.1 mM), benzoic
acid (1 equiv. to the catalyst), methyl benzoate (0.61 mg, 5.3
mmol, 10.6 mM) and 4-bromo-N-tosyl imine (6.2 mg, 18.3 mmol,
36.6 mM) were combined and dissolved in dichloromethane-d2 to
afford a total reaction volume of 0.5 mL. Freshly distilled methyl
Synthesis of catalysts 1
The synthesis of (S)-NOBIN followed literature procedures with
modifications for synthesizing 7 and 8. The procedure for the
4406 | Org. Biomol. Chem., 2010, 8, 4400–4407
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
vinyl ketone (1.69–2.64 mg, 24.1–37.6 mmol, 48.2–75.2 mM was
added and the solution was transferred to an NMR tube. The tube
was sealed and the time course of the reaction was followed by 1H
NMR spectroscopy.
2 (a) For recent reviews of asymmetric organocatalysis, see: D. Enders
and A. Narine, J. Org. Chem., 2008, 73, 7857–7870; (b) A. Dondoni
and A. Massi, Angew. Chem., Int. Ed., 2008, 47, 4638–4660.
3 For a proper definition of bifunctional organocatalysis, see: J. Seayad
and B. List, Org. Biomol. Chem., 2005, 3, 719–724.
4 (a) For recent examples of trifunctional organocatalysts: T. Ema, D.
Tanida, T. Matsukawa and T. Sakai, Chem. Commun., 2008, 957–958;
(b) W. Chen, W. Du, Y-Z. Duan, Y. Wu, S-Y Yang and Y-C Chen,
Angew. Chem., Int. Ed., 2007, 46, 7667–7670; (c) J-M. Garnier, C.
Anstiss and F. Liu, Adv. Synth. Catal., 2009, 351, 331–338; (d) J-M.
Garnier and F. Liu, Org. Biomol. Chem., 2009, 7, 1272–1275; (e) D-Q.
Xu, L-P. Wang, S-P. Luo, Y-F. Wang, S. Zhang and Z-Y. Xu, Eur. J. Org.
Chem., 2008, 1049–1053; (f) C-J. Wang, Z-H. Zhang, X-Q. Dong and
X-J. Wu, Chem. Commun., 2008, 1431–1433.
5 (a) Recent representative reviews on the MBH reaction: D. Basavaiah,
A. J. Rao and T. Satyanarayana, Chem. Rev., 2003, 103, 811–891; (b) D.
Basavaiah, K. V. Rao and R. J. Reddy, Chem. Soc. Rev., 2007, 36, 1581–
1588; (c) V. Declerck, J. Martinez and F. Lamaty, Chem. Rev., 2009,
109, 1–48.
General procedure for the experiments on the rate and enan-
tioselectivity dependence on temperature. Catalyst 1c (1 mg, 1.5
mmol, 3 mM), benzoic acid (1 equiv. to the catalyst), methyl
benzoate (0.64 mg, 5.6 mmol, 11.2 mM) and 4-bromo-N-tosyl
imine (5.1 mg, 15.1 mmol, 110 mM) were combined and dissolved
in 0.5 mL dichloromethane-d2. The solution was transferred to
an NMR tube and freshly distilled methyl vinyl ketone (2.65 mg,
37.8 mmol, 275 mM) was added. The tube was sealed and the time
course of the reaction was followed by 1H NMR spectroscopy at
273–303 K. Sample aliquots at each temperature were taken at
approximately 10% and >90% conversions and subjected to chiral
HPLC for enantioselectivity analysis.
6 (a) Y-L. Shi and M. Shi, Eur. J. Org. Chem., 2007, 18, 2905–2916;
(b) G. Masson, C. Housseman and J. Zhu, Angew. Chem., Int. Ed.,
2007, 46, 4614–4628; (c) C. Menozzi and P. I. Dalko, Enantioselective
Organocatalysis, Wiley-VCH, Weinheim, 2007, 151–187.
Spectroscopic data
7 (a) Recent experimental mechanistic studies on the MBH or azaMBH
reaction: K. E. Price, S. J. Broadwater, H. M. Jung and D. T. McQuade,
Org. Lett., 2005, 7, 147–150; (b) K. E. Price, S. J. Broadwater, B. J.
Walker and D. T. McQuade, J. Org. Chem., 2005, 70, 3980–3987; (c) P.
Buskens, J. Klankermayer and W. Leitner, J. Am. Chem. Soc., 2005, 127,
16762–16763; (d) V. K. Aggarwal, S. Y. Fulford and G. C. Lloyd-Jones,
Angew. Chem., Int. Ed., 2005, 44, 1706–1708; (e) I. T. Raheem and E. N.
Jacobsen, Adv. Synth. Catal., 2005, 347, 1701–1708; (f) M. E. Krafft,
T. F. N. Haxell, K. A. Seibert and K. A. Abboud, J. Am. Chem. Soc.,
2006, 128, 4174–4175; (g) G. W. Amarante, H. M. S. Milagre, B. G. Vaz,
B. R. Vilacha Ferreira, M. N. Eberlin and F. Coelho, J. Org. Chem.,
2009, 74, 3031–3037; (h) L. S. Santos, C. H. Pavam, W. P. Almeida, F.
Coelho and M. N. Eberlin, Angew. Chem., Int. Ed., 2004, 43, 4330–
4333.
(E)-4-Methyl-N-(4-methylene-5-oxo-1-phenylhex-1-en-3-yl)ben-
zenesulfonamide (5a). The spectroscopic data of the compound
was found to match those in a previous report.13 1H NMR
(400 MHz, CDCl3) d 2.17 (s, 3H); 2.31 (s, 3H); 4.80 (dd, J1 =
9.0 Hz, J2 = 7.8 Hz, 1H); 5.70 (d, J = 9.0 Hz, 1H); 6.98 (m, 3H);
6.28 (d, J = 15.9 Hz, 1H); 7.14–7.27 (m, 7H), 7.69 (d, J = 8.1 Hz,
2H). ESI [M+Na+]: found 378.1138, calcd for (C20H21NO3SNa)+:
378.1140. Chiral HPLC analysis: eluent: hexane–isopropanol =
80 : 20; flow rate: 0.7 mL min-1; tmajor = 12.8 min, tminor = 14.6 min.
8 (a) D. Salinger and R. Bru¨ckner, Synlett, 2009, 109–111; (b) T. Ooi, K.
Ohmatsu and K. Maruoka, J. Am. Chem. Soc., 2007, 129, 2410–2411.
9 S. Vyskocil, M. Smrcina, V. Hanus, M. Polasek and P. Kocovsky, J. Org.
Chem., 1998, 63, 7738–7748.
Acknowledgements
C. A. is supported by an Australian Postgraduate Award.
10 Overall two batches of catalysts 1c were used. Both batches of the
catalyst were found to be identical in activity in control reactions.
11 A. Solladie´-Cavallo, M. Roje, R. Welter and V. Sunjic, J. Org. Chem.,
2004, 69, 1409–1412.
12 M. Shi, Y-M. Xu and Y-L. Shi, Chem.–Eur. J., 2005, 11, 1794–1802.
13 M. Shi, L-H. Chen and C-Q Li, J. Am. Chem. Soc., 2005, 127, 3790–
3800.
References
1 For an overview of asymmetric organocatalysis, see A. Berkessel and
H. Gro¨ger, Asymmetric Organocatalysis: from Biomimetic Concepts to
Applications in Asymmetric Synthesis, Wiley-VCH, Weinheim, 2005.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 4400–4407 | 4407
©