“This document is the Accepted Manuscript version of a Published Work that appeared in final form in ACS
Catalysis, copyright © American Chemical Society after peer review and technical editing by the publisher. To
access the final edited and published work see http://pubs.acs.org/doi/abs/10.1021/cs5006037 ”

Continuous Flow Enantioselective ThreeComponent anti-Mannich Reactions Catalyzed by a
Polymer-Supported Threonine Derivative
Carles Ayats,† Andrea H. Henseler,† Estefanía Dibello† and Miquel A. Pericàs†,‡,*
†

Institute of Chemical Research of Catalonia (ICIQ), Avda. Països Catalans, 16, E-43007,

Tarragona (Spain). Fax: (+34)977-920-244; Tel: (+34)977-920-243; E-mail: mapericas@iciq.es
‡

Departament de Química Orgànica, Universitat de Barcelona (UB), E-08028 Barcelona (Spain)
Keywords: β-amino-α-hydroxycabonyl compounds; asymmetric catalysis; continuous flow;
Mannich reaction; supported catalysts

Abstract: A series of primary amino acid-derived polystyrene-supported organocatalysts was
tested in anti-selective Mannich reactions. The polystyrene-immobilized threonine derivative
showed the best performance in three-component (hydroxyacetone, anilines and aldehydes)
Mannich reactions to provide anti-β-amino-α-hydroxycarbonyl compounds (11 examples; up to
95% ee) and its use could be extended to dihydroxyacetone and protected hydroxyacetones (7
examples; up to 90% ee). The high activity depicted by the catalyst has allowed its
implementation in continuous flow. Under this operation mode, the supported threonine catalyst
1

produces anti-Mannich adducts with generally higher diastereo- and enantioselectivity than in
batch. A family of five different enantioenriched anti-Mannich adducts has been sequentially
prepared in flow by passing different combinations of anilines and aromatic aldehydes over the
same sample of catalyst. This confirms the suitability of this methodology for the rapid access to
small libraries of enantioenriched compounds.

1. INTRODUCTION
The asymmetric Mannich reaction1 is one of the most versatile methodologies for the preparation
of valuable chiral amino carbonyl compounds and their derivatives.2 In 2000, List described the
utility of L-proline as an excellent organocatalyst3 for three-component intermolecular
asymmetric Mannich reactions,4 and L-proline became a routine catalyst for this transformation
affording syn-products with high stereoselectivity.5 While syn-Mannich products are readily
accessible, anti-Mannich products are more difficult to synthesize. Recently, different
organocatalysts have been found to catalyze Mannich reactions to selectively achieve either synor anti- Mannich products.2f,6 When α-hydroxyacetone and O-protected derivatives are used as
acyclic ketone donors in asymmetric Mannich-type transformations, adducts with up to four
adjacent functional carbons featuring a central, stereodefined 1,2-amino alcohol unit are
obtained. These synthons are common building blocks for the synthesis of complex α-amino
acid derivatives5b,5e as well as important precursors for different types of bioactive compounds
with high pharmaceutical value.7 The development of a simple and efficient way for the
preparation of these compounds in enantiomerically pure form is thus of utmost importance for
synthetic organic chemistry.
Natural primary amino acids and their derivatives have recently caught attention as successful
catalysts for different asymmetric transformations such as Mannich,8,9 aldol,8b,8e,8k,10 Michael,11 α2

amination12 and cyanosilylation reactions.13 While proline catalyzes Mannich reactions of
hydroxyacetone to form adducts with a syn-1,2-amino alcohol arrangement,5b,5e with primary
amino acids the corresponding anti adducts are formed.8b-l This is due to the fact that primary
amino acids mediate Mannich reactions via the (Z)-enamine intermediate, this configurational
preference results from an extra hydrogen bonding interaction between the NH group and the
oxygen atom of the hydroxy group of the enamine. Therefore, primary amino acid catalysis is
complementary to proline catalysis when hydroxyacetone and their O-protected derivatives are
used.8b,8h
The high polarity of organocatalysts (e.g., free amino acids) converts the isolation of similarly
polar reaction products (e.g., Mannich adducts) into wasteful and tedious processes. A successful
strategy to overtake this problem is catalyst immobilization onto solid supports, which allows
catalyst separation by simple filtration and the possibility of catalyst recycling.14 In addition,
properly designed, heterogenized catalysts often replicate the properties of their homogeneous
analogues, with the advantage of being suitable for continuous flow processes without co-elution
of the catalyst.5f,6d,15,16 Mannich reactions catalyzed by homogeneous primary amino acid
derivatives usually require high catalyst loadings (up to 30%);8a-b,8e-f,8i-k,9 thus, immobilization of
these species appears to be a very good alternative. In the literature, only two examples of
recyclable primary amino acid derivatives are reported to catalyze anti-Mannich reactions. One
of them involves a soluble catalyst and, therefore, extraction is required before the catalyst can
be reused.8m The other example reports the use of a cysteine-derivative anchored onto magnetic
nanoparticles, however non-enantioenriched Mannich products were obtained.17
We have recently reported on the use of polymer-supported prolines and pyrrolidines as highly
active and selective heterogenized catalysts for Mannich reactions under batch and continuous

3

flow conditions leading respectively to syn-products and anti-products, with excellent diastereoand enantioselectivities.5f,6d However, to the best of our knowledge there are no examples of
three-component

enantioselective

anti-Mannich

reactions

catalyzed

by

supported

organocatalysts. Herein, we report the use of a polymer-supported threonine derivative as a
highly active and stereoselective catalyst for three-component anti-Mannich reactions, and its
use for the implementation of a single pass, continuous flow process for the preparation of a
small library of anti-β-amino-α-hydroxycarbonyl compounds in high enantiomeric purity.

2. RESULTS AND DISCUSSION
We have recently reported the preparation of a series of polymer-supported primary amino acid
derivatives 1-6 (Figure 1) and their use as heterogeneous catalysts for aldol reactions.10j Among
them, catalyst 6 showed the best catalytic performance regarding activity and stereoselectivity,
yielding aldol products of high enantiopurity.
N
N

N

N
S

1
N

N
O

NH2

2

CO2H
N

CO2H

N
O

O

3
N

N

NH2

N N
N

NH2

N
N
O

N
NH2

5
N
H

CO2H

4

CO2H

N N

NH2
O

6

NH2

CO2H

CO2H

Figure 1. Primary amino acid derived catalysts.
With these catalysts in hand, we proceeded to evaluate them in Mannich reactions, using
hydroxyacetone and preformed N-(p-methoxyphenyl) (N-PMP) ethyl glyoxylate imine 7 as the

4

benchmark process (Table 1). DMF was used as the solvent for this transformation due to both
its good swelling ability for polymer-supported organocatalysts and previous experience in the
laboratory.5f,6d

Table 1. Evaluation of PS-supported primary amino acids as catalysts for the anti-selective
Mannich reaction.a
OMe

OMe
O

catalyst (20 mol%)
+
OH

HN

O

CO2Et

DMF, RT

N

OH

CO2Et
7

8

	
  

entry

catalyst

time (h)

conv.b (%)

syn:antib

eec (%)

1

1

17

>99

58:42

43

2

2

12

>99

50:50

69

3

3

6

>99

55:45

58

4

4

3

>99

40:60

36

5

5

6

>99

50:50

78

6

6

6

>99

35:65

85

7

6d

8

>99

40:60

73

8

6e

8

>99

45:55

80

a

Reactions were performed with catalyst 1-6 (20 mol%), preformed imine 7 (0.125 mmol) and
hydroxyacetone (6 equiv.) in DMF (0.25 mL). bBy 1H NMR of the crude mixture. canti-Isomer;
by chiral HPLC after purification. d10 mol% of 6 was used. e15 mol% of 6 was used.

Polystyrene-supported cysteine and serine derivatives 1 and 2 catalyzed the formation of
compound 8 but showed only low activity (entries 1 and 2). Moreover, in the case of cysteine
derivative 1, the syn-Mannich product was obtained as the major diastereomer.18 Catalyst 4 was

5

synthesized as a promising potential catalyst due to its close structural similarity with
immobilized trans-4-azidoproline.5f Although this species did not show catalytic activity for
aldol reactions,10j it mediated the formation of anti-Mannich product 8 in a remarkable short
reaction time (3 h), albeit with disappointingly low enantioselectivity (entry 4). In turn, catalyst 5
derived from 5-hydroxytryptophan afforded product 8 with high levels of enantioselectivity but
in poor diastereoselectivity (entry 5).
Threonine derivatives have been reported to display very good catalytic activity for
enantioselective anti-Mannich reactions.8b-f,8i-k Very gratifyingly, the supported threonine 6 also
displayed this behavior and showed the best catalytic performance (entry 6) in terms of both
diastereo- and enantioselectivity.
Once 6 was selected as the optimal catalyst, the effect of the solvent nature on the performance
of the reaction was studied (Table 2). With DMF, no difference was observed regarding either
catalytic activity or selectivity when reagent grade or anhydrous solvent was used (entries 1 and
2). In both cases, reaction rates were low, although good stereoselectivities were achieved. Using
THF stereoselectivity was not improved, but a shorter reaction time was required (entry 3).
Despite the fact that some examples of anti-selective Mannich reactions catalyzed by primary
amino acid derivatives using NMP (N-methyl-2-pyrrolidone) have been described,8b,8f,8i-k in the
presence of catalyst 6 a slow reaction took place and the product was obtained with low
enantioselectivity (entry 4). Conversely, the reaction showed to be very fast when CH2Cl2,
toluene or solvent free conditions were tested; unfortunately, only moderate to very poor
enantioselectivities were achieved in these cases (entries 5, 6 and 7). On the other hand, the
presence of water as a co-solvent exclusively led to the decomposition of imine 7 (entry 8).5f In
the light of these results, we tested different mixtures of DMF (better results regarding

6

stereoselectivity) and CH2Cl2 (better results regarding reaction rate and swelling ability for
Merrifield-derived resins) (entries 9, 10 and 11). Gratifyingly, a 1:1 mixture of DMF and CH2Cl2
led to a significantly improved yield as well as to excellent enantioselectivity (Table 2, entry 10).
As a consequence of the hydrolytic instability of imine 7, optimal reaction conditions involved
performing the reaction in the glovebox with anhydrous solvents (entry 12) and in the presence
of molecular sieves (entry 13). In this way, the Mannich adduct 8 could be obtained in very short
reaction time (1 h) and with good diastereoselectivity (syn:anti = 21:79) and excellent
enantioselectivity (96% ee).

7

Table 2. Solvent screening for the Mannich reaction between hydroxyacetone and preformed
imine 7 catalyzed by resin 6.a
OMe

OMe
O

catalyst 6 (20 mol%)
+
OH

HN

O

CO2Et

solvent, RT

N

OH

CO2Et
7

8

entry

solvent

time (h)

yieldb (%)

syn:antic

eed (%)

1

DMFe

6

35

35:65

85

2

anh. DMF

5

35

37:63

84

3

anh. THF

3

37

50:50

74

4

NMP

6

16

59:41

33

5

anh. CH2Cl2

1,5

53

64:36

66

6

Neat

0,5

46

65:35

0

7

anh. Toluene

1

51

63:37

60

8

DMF/H20 (80:20)

7

-

-

-

9

DMF/CH2Cl2 (80:20)e

3

42

36:64

85

10

DMF/CH2Cl2 (50:50)e

2

52

26:74

93

11

DMF/CH2Cl2 (20:80)e

2

45

41:59

88

12f

anh. DMF/ anh. CH2Cl2 (50:50)

1,5

56

26:74

94

13f,g

anh. DMF/ anh. CH2Cl2 (50:50)

1

58

21:79

96

a

Unless otherwise stated, the reactions were performed with catalyst 6 (20 mol%), preformed
imine 7 (0.125 mmol) and hydroxyacetone (6 equiv.) in 0.25 mL solvent. bIsolated product. cBy
1
H NMR of the crude mixture. danti-Isomer; by chiral HPLC after purification. eSynthesis grade
solvents were used. fReaction performed in the glovebox. g4Å molecular sieves were used.

8

Once the reaction conditions were optimized, the next step was to explore the scope of the
anti-Mannich reactions of hydroxyacetone with different amines and aldehydes (Table 3). To
avoid possible problems arising from hydrolytic instability of the intermediate imines, the
possibility of performing the reactions as "one pot", three component processes was explored.
Interestingly, the reactions took place under these conditions without the need for special
precautions.
Table 3. Scope of the reaction of hydroxyacetone with different amines and aldehydes catalyzed
by resin 6.a
R1
NH2

O
+

CHO
catalyst 6 (20 mol%)

+

OH
R1

R2

DMF/CH2Cl2 (50:50)
RT

O

HN

OH

R2

9-19

entry

R1

R2

product

time (h)

yieldb (%)

syn:antic

eed (%)

1

H

H

9

6

96

8:92

93

2

H	
  

p-­‐NO2	
  

10

5

88

9:91

92

3	
  

p-­‐CH3O

H

11

7

90

28:72

82

4	
  

p-­‐CH3O	
  

p-­‐NO2	
  

12

4

89

13:87

85

5	
  

p-­‐CH3O	
  

p-­‐CH3O	
  

13

8

62

52:48

49

6	
  

p-­‐CH3O	
  

p-­‐Br	
  

14

4

84

22:78

81

7	
  

p-­‐CH3O	
  

p-­‐CN	
  

15

5

86

19:81

83

8

p-­‐CH3O	
  

p-­‐Cl	
  

16

8

72

37:63

67

9

p-­‐CH3O	
  

o-­‐Cl	
  

17

8

70

15:85

83

10	
  

p-­‐CH3	
  

p-­‐NO2	
  

18

5

87

11:89

87

11	
  

p-­‐Cl	
  

p-­‐NO2	
  

19

3

88

12:88

95

9

a

Reactions were performed with catalyst 6 (20 mol%), amine (0.2 mmol), aldehyde (1.1 equiv.)
and hydroxyacetone (6 equiv.) in 1:1 DMF/CH2Cl2 (0.4 mL). bIsolated product. cBy 1H NMR of
the crude mixture. danti-Isomer, by chiral HPLC after purification.

The results show that three-component anti-Mannich reactions mediated by 6 can be
completed in 3-8 hours using unsubstituted aniline (entries 1-2) and p-substituted (methoxy,
methyl, chloro) derivatives (entries 3-11), yielding anti-adducts 9-19 in moderate to good
diastereo- (up to 8:92) and enantioselectivities (up to 95% ee), with yields up to 96%. When the
reaction was carried out with an aldehyde bearing an electron-donating substituent in para
position (entry 5) poor enantioselectivity was observed as previously reported in the literature.8b
It is worth noting that the present procedure affords anti-Mannich products in good
enantioselectivities when performing the reactions at room temperature, while most of the
reported anti-Mannich reactions with hydroxyacetone catalyzed by primary amino acid
derivatives require low temperatures for the achievement of high enantioselectivities.8b,8i-k
The scope of the asymmetric, anti-selective Mannich reaction mediated by 6 could be further
expanded (Table 4) with the use of additional ketones, including 1,3-dihydroxyacetone (entries 1
and 2), protected hydroxyacetones (entries 3-7), aromatic and aliphatic aldehydes, and panisidine.

10

Table 4. Scope of the reaction using different ketone donors.a
OMe
NH2

O
+
R1

+

catalyst 6 (20 mol%)

O
R3

OR2
OMe

DMF/CH2Cl2 (50:50) R
1
RT

O

HN
R3
OR2
20-26

R1

entry

R2

R3

product

time (h)

yieldb (%)

syn:antic

eed (%)

1

e

OH

H

p-­‐BrC6H4	
  

20

8

74

34:66

67

2

e

OH

H

p-­‐CNC6H4	
  

21

8

76

29:71

82

3

H

Bn

p-­‐NO2C6H4	
  

22

38

94

23:77

80

4

H

Bn

p-­‐CNC6H4	
  

23

36

80

25:75

90

5

H

Bn

2,4-­‐Cl2C6H3	
  

24

38

84

12:88

77

6

H

Bn

i-Bu

25

22

52

47:53

82

7

H

TBS

p-­‐NO2C6H4	
  

26

24

77

67:33

51f

a

Reactions were performed with 6 (20 mol%), p-anisidine (0.2 mmol), aldehyde (1.1 equiv.)
and ketone (6 equiv.) in a 1:1 mixture of DMF/CH2Cl2 (0.4 mL). bIsolated product. cBy 1H NMR
of the crude mixture. danti-Isomer; by chiral HPLC after purification. eDihydroxyacetone dimer
(3 equiv.) and acetic acid (10 mol%) were used. fanti-Isomer; by chiral HPLC of the desilylated
product.

Notably, resin 6 catalyzed the three-component anti-Mannich reactions of dihydroxyacetone19
affording anti-dihydroxy-β-aminoketone 20-21 (entries 1 and 2). These densely functionalized
compounds are of high synthetic importance for the preparation of aminosugars.20 Despite only
moderate stereoselectivities were recorded (and to 82% ee), these results are among the best
previously reported for the same transformation.8b In the case of O-benzylhydroxyacetone, we
studied this transformation with both electron-deficient aromatic aldehydes (entries 3, 4 and 5)
and aliphatic aldehydes (entry 6), leading to the formation of the orthogonally protected β11

amino-α-hydroxyketones 22-25. In these cases, the reactions were slower but took place with
high enantioselectivity. Notably, when the reaction was performed with TBS-protected αhydroxyacetone, the diastereoselectivity of the process could be reversed, the syn-Mannich
adduct being obtained as the major product, and poor enantiomeric excess of the anti-isomer was
recorded (Table 4, entry 7).8d
To exploit one of the advantages offered by catalyst immobilization, we explored the
recyclability of PS-supported threonine 6 in the Mannich reaction between α-hydroxyacetone,
aniline and p-nitrobenzaldehyde to afford anti-β-amino-α-hydroxyketone 10 (Table 5). After
each run, the catalyst was recovered by simple filtration and directly used in the next cycle.
Working under the specified reaction conditions, a decrease in the activity and selectivity of the
catalyst was observed after the third cycle in a highly reproducible manner. Starting from the
preformed imine (see Supporting Information) or performing the recycling in the glovebox
entailed no improvement of these results. Attempts to reactivate the resin by washing with AcOH
also proved unsuccessful (entry 5). To understand the origin of the deactivation process, the
initial and the deactivated resins were studied by elemental analysis and by IR spectroscopy (see
Supporting Information). No change in the functionalization level (%N) was observed,
suggesting that no cleavage of the monomer has taken place. On the other hand, the IR spectrum
shows an almost complete disappearance of the carboxylate bands and the appearance of a weak
absorption at 1715 cm-1, suggesting that the excess of hydroxyacetone required for the reaction
to proceed, slowly reacts with the supported amino acid leading to some catalytically less active
and less stereoselective derivative. In any case, the recycling process allows achieving an
accumulated TON >20, which compares very favorably with the values (3-5) achieved with the
reference homogeneous catalysts.8b

12

Table 5. Recycling experiments for the reaction between the α-hydroxyacetone, aniline and pnitrobenzaldehyde.a

NH2

O
+

CHO
catalyst 6 (20 mol%)

+

OH
NO2

DMF/CH2Cl2 (50:50)
RT, 5 h

O

HN

OH

NO2

10

run

conv.b (%)

syn:antib

eec (%)

1

>99

9:91

92

2

>99

10:90

92

3

>99

10:90

89

4

82

17:83

55

5d

25

28:72

67

a

Reactions were performed with catalyst 6 (20 mol%), aniline (0.2 mmol), p-nitrobenzaldehyde
(1.1 equiv.) and hydroxyacetone (6 equiv.) in a 1:1 mixture of DMF/CH2Cl2 (0.4 mL). bBy 1H
NMR of the crude mixture. canti-Isomer; by chiral HPLC after purification. dAfter washing with
acetic acid (10%) in a 1:1 mixture of DMF/CH2Cl2.

In view of the results obtained with immobilized catalyst 6 under batch conditions, we
envisaged the possibility of performing the asymmetric, three component Mannich reaction in
continuous flow. The experimental set-up for the flow synthesis of compound 10 is represented
schematically in Figure 2 (see Supporting Information for more details). It consisted of a
vertically mounted and fritted, volume adjustable low-pressure glass chromatography column
loaded with the polymer-supported catalyst 6. The reactor inlet was connected to a T-shaped
adaptor, which allows switching between two channels, connected to two syringe pumps. One
channel was connected to a solution with the three reactants (no reaction occurs in absence of
13

catalyst) in a 1:1 mixture of DMF/CH2Cl2, and the other channel was connected to a flask
containing the same solvent mixture, to rinse the system. Continuous inline IR analysis was used
to determine the optimal flow rate for complete conversion.21

IR
N

CHO

NH2

O
+
OH

N N

NH2
O

CO2H
6
(300 mg, 0.31 mmol)

+
O

DMF/CH2Cl2 (50:50)

HN

NO2
T-shaped
piece

OH

NO2

10
Mixture of
DMF/CH2Cl2

Figure 2. Experimental set-up for the continuous flow experiments.

Thus, a solution of aniline, hydroxyacetone and p-nitrobenzaldehyde was pumped through a
column loaded with 300 mg of catalyst 6 (f = 1.02 mmol·g-1) with a flow rate of 30 µL·min-1
(corresponding to 17 min residence time) (Scheme 1a).22 The system was operated for 6h
producing 3.13 mmol of 10, with diastereo- and enantioselectivity very close to those recorded
under batch conditions (Table 3, entry 2).23 Likewise, after pumping for 4 hours a solution of pchloroaniline, hydroxyacetone and p-nitrobenzaldehyde,24 2.78 mmol of highly enantioenriched
anti-Mannich adduct 19 were obtained (Scheme 1b).25 In this case, slightly better diastereo- and
enantioselectivity was recorded with respect to the corresponding batch process (Table 3, entry

14

11). Remarkably, both continuous flow processes allow for a two-fold reduction of the catalyst
amount compared to the batch processes mediated by resin 6.26

Scheme 1. Continuous production of anti-Mannich products 10 and 19.

a)

NH2

O
+

CHO

flow = 30 µL·min-1
residence time = 17 min

O

HN

+

OH

OH

NO2
10
3.13 mmol product in 6h
dr = 88:12
89% ee

NO2

6 (300 mg)
Cl

b)
NH2

O
+

CHO

flow = 50 µL·min-1
residence time = 10 min

O

HN

+

OH

OH
Cl

NO2

NO2

19
2.78 mmol product in 4h
dr = 92:8
96% ee

The successful use of resin 6 in continuous flow conditions, together with the importance of
the resulting enantioenriched Mannich adducts as advanced synthons for the preparation of
biologically active compounds,7 prompted us to explore the possibility of preparing a small
library of enantioenriched anti-Mannich adducts in a sequential manner. This is a rather unique
application of flow processes with immobilized catalytic systems, and allows the preparation of
focused libraries of drug candidates or drug intermediates with important advantages. Thus, the
products are obtained free of contamination by the catalyst, and the scale of production can be
simply controlled through the operation time.6c,16c,f-g Not less importantly, these sequential
processes are easily amenable to automation. Thus, by operating the same continuous flow setup

15

described above, a single sample of catalytic resin 6 could be used to prepare five different antiMannich compounds in a sequential process. In this event, the flow preparation of every adduct
was run for 1 h (flow rate of 30 µL·min-1) and the operation only involved washing the resin by
circulation of CH2Cl2 for 30 min between two different substrates. The results obtained in the
preparation of this sequential library are summarized in Scheme 2.
The five target anti-Mannich adducts 9, 12, 15, 18 and 19 were obtained in good yields and in
high diastereo- and enantiomeric purities. Remarkably 9, 12, and 19 were obtained in this
manner with slightly better stereoselectivities than under batch conditions (see Table 3, entries 1,
4 and 11). High productivities were recorded with all tested substrates, ranging from 1.6−2.1
mmolproduct·mmolresin-1·h-1.

16

R1
O

HN

OH

R2
R1

R2

1h

15, 74% yield
81:19, 80% ee, (1.7)

-OCH3

-CN

1h

19, 80% yield
92:8, 96% ee, (1.8)

-Cl

-NO2

1h

12, 69% yield
91:9, 88% ee, (1.6)

-OCH3

-NO2

1h

18, 86% yield
89:11, 92% ee, (1.9)

-CH3

-NO2

1h

9, 94% yield
90:10, 92% ee, (2.1)

-H

-H

6
(300 mg, 0.31 mmol)
NH2

O
+

CHO
+

OH
R1

R2

Scheme 2. Continuous flow production of a library of enantioenriched anti-Mannich adducts.
Productivities in mmolproduct·mmolresin-1·h-1are shown in parentheses.

3. CONCLUSIONS
In summary, we have identified a threonine derivative immobilized onto polystyrene through a
1,4-disubstituted 1,2,3-triazole linker (6) as an efficient organocatalyst for the three-component
anti-Mannich reaction. The immobilized catalyst depicts high activity and stereoselectivity in
dichloromethane/N,N-dimethylformamide mixtures and allows recycling and reuse (3 cycles).
The good overall performance of 6 allowed adapting its use to single-pass, continuous flow

17

processes. We have also shown that this flow system can be applied to the diastereo- and
enantioselective medium-scale preparation of a diverse library of anti-Mannich adducts in a
sequential manner. Noteworthy, these are the first examples of anti-selective, three component
asymmetric Mannich reactions performed in continuous flow.

4. EXPERIMENTAL DETAILS
General Procedure for the Asymmetric Multicomponent Mannich reaction
Polymer-supported catalyst 6 (20 mol%) was swollen in a vial with a 1:1 mixture of
DMF/CH2Cl2 (0.4 mL). The amine (0.2 mmol), aldehyde (1.1 equiv.) and ketone (6.0 equiv. or 3
equiv. of dihydroxyacetone dimer plus 10 mol% acetic acid) reactants were added and the
reaction mixture was shaken at room temperature for the times indicated in Tables 3 and 4. Then,
the resin was filtered off, washed with AcOEt (3 x 1 mL) and dried under vacuum. The
combined liquid phases were concentrated under reduced pressure. Purification by column
chromatography (eluting with cyclohexane/AcOEt) gave the corresponding Mannich products 926. Conversion and diastereomeric ratio were determined by 1H NMR of the crude samples after
removal of the resin. The enantiomeric excess was determined by HPLC on a chiral stationary
phase after purification by column chromatography. In the recycling experiments, the dried resin
was used in the next run without further treatment.

ASSOCIATED CONTENT
Supporting Information.
Experimental details and characterization data. This material is available free of charge via the
Internet at http://pubs.acs.org.
18

AUTHOR INFORMATION
Corresponding Author
*E-mail: mapericas@iciq.es
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENT
This work was funded by MINECO (grant CTQ2012-38594-C02-01), DEC (grant
2014SGR827), and ICIQ Foundation. We also thank MINECO for support through Severo
Ochoa Excellence Accreditation 2014-2018 (SEV-2013-0319). C. A. thanks MICINN for a Juan
de la Cierva postdoctoral fellowship. A. H. H. thanks the MECD for an FPU fellowship. E. D.
thanks to ANII, CSIC Uruguay and PEDECIBA for financial support.
REFERENCES
(1) Mannich, C.; Krösche, W. Arch. Pharm. 1912, 250, 647−667.
(2) For reviews see: (a) Arend, M.; Westermann, B.; Risch, N. Angew. Chem. Int. Ed. 1998,
37, 1044−1070. (b) Denmark, S. E.; Nicaise, O. J.-C. In Comprehensive Asymmetric
Catalysis, Vol. II; Jacobsen, E. N.; Pfaltz, A.; Yamamoto, H., Eds.; Springer, Heidelberg,
1999; pp 923−961. (c) Kobayashi, S.; Ishitani, H. Chem. Rev. 1999, 99, 1069−1094. (d)
Córdova, A. Acc. Chem. Res. 2004, 37, 102−112. e) Arrayás, R. G.; Carretero, J. C. Chem.
Soc. Rev. 2009, 38, 1940−1948. f) Cai, X.-H.; Xie, B. Arkivoc 2013, 264−293, and
references herein.

19

(3) (a) de Meijere, A.; Diederich, F. In Asymmetric Organocatalysis; Berkessel, A.; Gröger, H.
Eds.; Wiley-VCH, Weinheim, 2005; pp 104-106. (b) Tanaka, F.; Barbas III, C. F. In
Enantioselective Organocatalysis; Dalko, P. I., Ed.; Wiley-VCH, Weinheim, 2007; pp 3849. (c) Jacobsen, E. N.; MacMillan, D. W. C. Proc. Natl. Acad. Sci. USA 2010, 107,
20618−20619.
(4) List, B. J. Am. Chem. Soc. 2000, 122, 9336−9337.
(5) For selected examples using L-proline for syn-Mannich reactions, see: (a) Córdova, A.;
Watanabe, S.; Tanaka, F.; Notz, W.; Barbas III, C. F. J. Am. Chem. Soc. 2002, 124,
1866−1867. (b) List, B.; Pojarliev, P.; Biller, W. T.; Martin, H. J. J. Am. Chem. Soc. 2002,
124, 827−833 (c) Hayashi, Y.; Tsuboi, W.; Ashimine, I.; Urushima, T.; Shoji, M.; Sakai,
K. Angew. Chem. Int. Ed. 2003, 42, 3677−3680. (d) Notz, W.; Tanaka, F.; Watanabe, S.;
Chowdari, N. S.; Turner, J. M.; Thayumanavan, R.; Barbas III, C. F. J. Org. Chem. 2003,
68, 9624−9634. (e) Pojarliev, P.; Biller, W. T.; Martin, H. J.; List, B. Synlett 2003,
1903−1905. (f) Alza, E.; Rodríguez-Escrich, C.; Sayalero, S.; Bastero, A.; Pericàs, M. A.
Chem.−Eur. J. 2009, 15, 10167−10172.
(6) See for example: (a) Zhang, H.; Mitsumori, S.; Utsumi, N.; Imai, M.; García-Delgado, N.;
Mifsud, M.; Albertshofer, K.; Cheong, P. H.-Y.; Houk, K. N.; Tanaka, F.; Barbas III, C. F.
J. Am. Chem. Soc. 2008, 130, 875−886. (b) Pouliquen, M.; Blanchet, J.; Lasne, M.-C.;
Rouden, J. Org. Lett. 2008, 10, 1029−1032. (c) Martín-Rapún, R.; Fan, X.; Sayalero, S.;
Bahramnejad, M.; Cuevas, F.; Pericàs, M. A. Chem.−Eur. J. 2011, 17, 8780−8783. (d)
Martín-Rapún, R.; Sayalero, S.; Pericàs, M. A. Green Chem. 2013, 15, 3295−3301.

20

(7) (a) Ager, D. J.; Prakash, I.; Schaad, D. R. Chem. Rev. 1996, 96, 835-876. (b) Bergmeier, S.
C. Tetrahedron 2000, 56, 2561−2576.
(8) For primary α-amino acid derived catalysts, see: (a) Ibrahem, I.; Zou, W.; Engqvist, M.;
Xu, Y.; Córdova, A. Chem.−Eur. J. 2005, 11, 7024−7029. (b) Ramasastry, S. S. V.; Zhang,
H.; Tanaka, F.; Barbas III, C. F. J. Am. Chem. Soc. 2007, 129, 288−289. (c) Cheng, L.; Wu,
X.; Lu, Y. Org. Biomol. Chem. 2007, 5, 1018−1020. (d) Cheng, L.; Han, X.; Huang, H.;
Wong, M. W.; Lu, Y. Chem. Commun. 2007, 4143−4145. (e) Xu, L.-W.; Lu, Y. Org.
Biomol. Chem. 2008, 6, 2047−2053. (f) Zhang, H.; Ramasastry, S. S. V.; Tanaka, F.;
Barbas III, C. F. Adv. Synth. Catal. 2008, 350, 791−796. (g) Teo, Y.-C.; Lau, J.-J.; Wu, M.C. Tetrahedron: Asymmetry 2008, 19, 186−190. (h) Fu, A.; Li, H.; Si, H.; Yuan, S.; Duan,
Y. Tetrahedron: Asymmetry 2008, 19, 2285−2292. (i) Dziedzic, P.; Ibrahem, I.; Córdova,
A. Tetrahedron Lett. 2008, 49, 803−807. (j) Wu, C.; Fu, X.; Ma, X.; Li, S.; Li, C.
Tetrahedron Lett. 2010, 51, 5775−5777. (k) Wu, C.; Fu, X.; Li, S. Tetrahedron: Asymmetry
2011, 22, 1063−1073. (l) An, Y.; Qin, Q.; Wang, C.; Tao, J. Chin. J. Chem. 2011, 29,
1511−1517. (m) Yong, F.-F.; Teo, Y.-C. Synth. Commun. 2011, 41, 1293−1300. (n)
Nugent, T. C.; Sadiq, A.; Bibi, A.; Heine, T.; Zeonjuk, L. L.; Vankova, N.; Bassil, B. S.
Chem.−Eur. J. 2012, 18, 4088−4098.
(9) For primary β-amino acid derived catalysts, see: Dziedzic, P.; Córdova, A. Tetrahedron:
Asymmetry 2007, 18, 1033−1037.
(10) For selected examples, see: (a) Bassan, A.; Zou, W.; Reyes, E.; Himo, F.; Córdova, A.
Angew. Chem. Int. Ed. 2005, 44, 7028−7032. (b) Córdova, A.; Zou, W.; Ibrahem, I.; Reyes,
21

E.; Engqvist, M.; Liao, W.-W. Chem. Commun. 2005, 3586−3588. (c) Jiang, Z. Q.; Liang,
Z.; Wu, X. Y.; Lu, Y. X. Chem. Commun. 2006, 2801−2803. (d) Ramasastry, S. S. V.;
Albertshofer, K.; Utsumi, N.; Tanaka, F.; Barbas III, C. F. Angew. Chem. Int. Ed. 2007, 46,
5572−5575. (e) Utsumi, N.; Imai, M.; Tanaka, F.; Ramasastry, S. S. V.; Barbas III, C. F.
Org. Lett. 2007, 9, 3445−3448. (f) Khan, S. S.; Shah, J.; Liebscher, J. Tetrahedron 2010,
66, 5082−5088. (g) Jiang, Z. Q.; Yang, H.; Han, X.; Luo, J.; Wong, M. W.; Lu, Y. X. Org.
Biomol. Chem. 2010, 8, 1368−1377. (h) Wu, C.; Fu, X.; Li, S. Eur. J. Org. Chem. 2011,
1291−1299. (i) Wu, C.; Long, X.; Li, S.; Fu, X. Tetrahedron: Asymmetry 2012, 23,
315−328. (j) Henseler, A. H.; Ayats, C.; Pericàs, M. A. Adv. Synth. Catal. 2014, 356, 17951802.
(11) (a) Sato, A.; Yoshida, M.; Hara, S. Chem. Commun. 2008, 6242−6244. (b) Yoshida, M.;
Narita, M.; Hirama, K.; Hara, S. Tetrahedron Lett. 2009, 50, 7297−7299. (c) Yoshida, M.;
Sato, A.; Hara, S. Org. Biomol. Chem. 2010, 8, 3031−3036. (d) Yoshida, M.; Kitamikado,
N.; Ikehara, H.; Hara, S. J. Org. Chem. 2011, 76, 2305−2309.
(12) Fu, J.-Y.; Yang, Q.-C.; Wang, Q.-L.; Ming, J.-N.; Wang, F.-Y.; Xu, X.-Y.; Wang, L.-X.
J. Org. Chem. 2011, 76, 4661−4664.
(13) Liu, X.; Qin, B.; Zhou, X.; He, B.; Feng, X. J. Am. Chem. Soc. 2005, 127, 12224−12225.
(14) (a) Blaser, H.-U., Pugin, B.; Studer, M. In Chiral Catalyst Immobilization and
Recycling; De Vos, D. E.; Vankelecom, I. F. J.; Jacobs, P. A., Eds.; Wiley-VCH,
Weinheim, 2000; pp 1-17. (b) Cozzi, F. Adv. Synth. Catal. 2006, 348, 1367−1390. (c)
Wang, Z.; Ding, K., Uozumi, Y. In Handbook of Asymmetric Heterogeneous Catalysis;
22

Ding, K.; Uozumi, Y., Eds.; Willey-VCH, Weinheim, 2008; pp 1-23. (d) Gruttadauria, M.;
Giacalone, F.; Noto, R. Chem. Soc. Rev. 2008, 37, 1666−1688. (e) Lu, J.; Toy, P. H. Chem.
Rev. 2009, 109, 815−838. (f) Zhao, G.; Chai, Z. In Recoverable and Recyclable Catalysts;
Benaglia, M., Ed.; Wiley, Chichester, 2009; pp 49-75. (g) Kristensen, T. E.; Hansen, T.
Eur. J. Org. Chem. 2010, 3179−3204.
(15) For recent examples, see: (a) Kirschning, A.; Jas, G. Applications of Immobilized
Catalysts in Continuous Flow Processes. In Immobilized Catalysts; Kirschning, A., Ed.;
Topics in Current Chemistry; Springer: Berlin, Heidelberg, 2004; Vol 242, pp 210-239. (b)
Baxendale, I. R.; Deeley, J.; Griffiths-Jones, C. M.; Ley, S. V.; Saaby, S.; Tranmer, G. K.
Chem. Commun. 2006, 24, 2566−2568. (c) Baxendale, I. R.; Ley, S. V. Solid Supported
Reagents in Multi-Step Synthesis. In New Avenues to Efficient Chemical Synthesis:
Emerging Technologies; Seeberger, P. H.; Blume, T., Eds.; Springer, Berlin, 2007; pp 151185. (d) Mason, B. P.; Price, K. E.; Steinbacher, J. L.; Bogdan, A. R.; McQuade, D. T.
Chem. Rev. 2007, 107, 2300−2318. (e) Bedore, M. W.; Zaborenko, N.; Jensen, K. F.;
Jamison, T. F. Org. Process Res. Dev. 2010, 14, 432−440. (f) Noel, T.; Buchwald, S. L.
Chem. Soc. Rev. 2011, 40, 5010−5029. (g) Zhao, D.; Ding, K. ACS Catal. 2013, 3,
928−944. (h) Puglisi, A.; Benaglia, M.; Chiroli, V. Green Chem. 2013, 15, 1790−1813. (i)
Tsubogo, T.; Ishiwata, T.; Kobayashi, S. Angew. Chem. Int. Ed. 2013, 52, 6590−6604. (j)
Pastre, J. C.; Browne, D. L.; Ley, S. V. Chem. Soc. Rev. 2013, 42, 8849−8869. (k) Wiles,
C.; Watts, P. Green Chem. 2014, 16, 55−62.
(16) For some examples of asymmetric continuous flow processes with immobilized catalysts,
see: (a) Alza, E.; Sayalero, S.; Cambeiro, X. C.; Martín-Rapún, R.; Miranda, P. O.; Pericàs,
23

M. A. Synlett 2011, 464−468. (b) Cambeiro, X. C.; Martín-Rapún, R.; Miranda, P. O.;
Sayalero, S.; Alza, E.; Llanes, P.; Pericàs, M. A. Beilstein J. Org. Chem. 2011, 7,
1486−1493. (c) Ayats, C.; Henseler, H. A.; Pericàs, M. A. ChemSusChem 2012, 5,
320−325. (d) Osorio-Planes, L.; Rodríguez-Escrich, C.; Pericàs, M. A. Org. Lett. 2012, 14,
1816−1819. (e) Fan, X.; Sayalero, S; Pericàs, M. A. Adv. Synth. Catal. 2012, 354,
2971−2976. (f) Kasaplar, P.; Rodríguez-Escrich, C.; Pericàs, M. A. Org. Lett. 2013, 15,
3498−3501. (g) Osorio-Planes, L.; Rodríguez-Escrich, C.; Pericàs, M. A. Chem.−Eur. J.
2014, 20, 2367−2372.
(17) Gawande, M. B.; Velhinho, A.; Nogueira, I. D.; Ghumman, C. A. A.; Teodoro, O. M. N.
D.; Branco, P. S. RSC Adv. 2012, 2, 6144−6149.
(18) Switches regarding the diastereoselectivity using different primary amino acid
derivatives have been observed previously. See ref. 8c.
(19) When 1,3-dihydroxyacetone is used as a donor in the reaction, the commercially
available dimer is employed in combination with a catalytic amount of acetic acid
(10 mol%). The role of the acetic acid is to catalyze the in situ conversion of the dimer into
the reacting monomer. See ref. 8i.
(20) (a) Westermann, B.; Neuhaus, C. Angew. Chem. Int. Ed. 2005, 44, 4077−4079. (b)
Enders, D.; Grondal, C.; Vrettou, M.; Raabe, G. Angew. Chem. Int. Ed. 2005, 44,
4079−4083.
(21) Rueping, M.; Bootwicha, T.; Sugiono, E. Beilstein J. Org. Chem. 2012, 8, 300−307.

24

(22) The residence time was determined by pumping a solution of methyl red through the
system and measuring the time elapsed between the first contact of the dye with the resin
and the moment when red color appeared at the column output.
(23) The use of continuous flow conditions for six hours results in a turnover number (TON)
of 10 (referred to the product formed) and a productivity of 1.7 mmolproduct·mmolresin-1·h-1.
(24) A flow rate of 50 µL·min-1 was used (equivalent to 10 min residence time).
(25) The use of continuous flow conditions results a turnover number (TON) of 9 (referred to
the product formed) and a productivity of 2.2 mmolproduct·mmolresin-1·h-1.
(26) The turnover number (TON) for the synthesis of compounds 10 and 19 under batch
conditions was 4.5 (referred to the product formed).

TOC
O
N
OH
NH 2

R1

N N

NH 2
O

CO2H

CHO

R2

anti-Mannich adducts
Wide scope
Up to 96% ee

R1
O HN

OH

R2

25