β- and δ-Amino acids (2,3- and 3,4-trans-CHA) as catalysts in Knoevenagel condensation and asymmetric aldol reactions

The efficacy of the βand δ-amino acids (5S,6S)-6-amino-5-hydroxycyclohexa-1,3-dienecarboxylic acid (2,3-trans-CHA) and (3R,4R)-4-amino-3-hydroxycyclohexa-1,5-dienecarboxylic acid (3,4-trans-CHA) as catalysts in Knoevenagel condensation and aldol addition reactions is studied. Synthesis of the zinc(II) complexes of 2,3and 3,4-trans-CHA provided precipitated material of sufficient purity for use. The catalytic Knoevenagel reactions were carried out with the β-amino acid 2,3-trans-CHA and the δ-amino acid 3,4-trans-CHA, resulting in a product yield of up to 61%. The asymmetric aldol addition reactions were carried out with catalytic amounts of the zinc(II) complexes of 2,3and 3,4-trans-CHA. In this case, it was observed that the ee of the product (up to 90%) depends on the conversion and/or the reaction time.


Introduction
Organocatalysis has attracted the attention of many synthetic chemistry groups worldwide. Indeed, the beginnings of organocatalysis can be traced back to early in the nineteenth century. 1,2 In 1909, Dakin reported the organocatalytic behavior of some naturally occurring amino acids, peptides and proteins in aldol reactions. 3 Additionally, it was noted that primary amines catalyze the decarboxylation of α-ketocarboxylic acids to produce the corresponding aldehydes. 4,5 In 1953, Prout reported the catalytic effect of L-and D-amino acids as well as the catalytic behavior of β-amino acids in Knoevenagel condensations; good yields were achieved using β-alanine as a catalyst. 6 In the 1970s, L-proline was introduced as a catalyst in the asymmetric Robinson annulation reaction. 7,8 It is now established as a chiral organocatalyst and as a source of chirality in the syntheses of many products. 9,10 Several aspects were taken into consideration in the development of L-proline as a catalyst, including its small size, rigidity, inexpensiveness and ready availability. 9 This secondary amine has been used as a catalyst in intermolecular aldol reactions. [11][12][13][14][15] It has been proposed that proline catalyzes the aldol reaction according to a mechanism involving an enamine intermediate. 16 This provides hydrogen bonding between the carboxylic acid group of proline and the carbonyl group of the substrate. 7,8,[17][18][19] A stereoselective Mannich reaction has also been achieved using L-proline catalysis. [19][20][21] The use of primary amino acids as chiral catalysts has been thoroughly studied and improved within the past decade. [22][23][24][25] In 2005, Amedjkouh reported the L-valine-catalyzed intermolecular aldol addition of acetone with aromatic aldehydes; sufficient yields and enantiomeric excesses were achieved. 26 Other natural primary amino acids were also used. [27][28][29] Cordova and co-workers have reported that primary amino acids serve as powerful organocatalysts in intermolecular aldol reactions of cyclic ketones. [30][31][32] We were intrigued by the idea of using nonproteinogenic, yet natural, primary β-and δ-amino acids as catalysts. (5S,6S)-6-Amino-5-hydroxycyclohexa-1,3-dienecarboxylic acid (2,3-trans-CHA, 2) and (3R,4R)-4-amino-3-hydroxycyclohexa-1,5-dienecarboxylic acid (3,4-trans-CHA, 4) (Scheme 1) are nonproteinogenic primary amino acid metabolites of the shikimate pathway 33 and are accessible through metabolic engineering of recombinant E. coli strains. 34 The shikimate pathway is an instructive example of diversity-oriented biosynthesis of a wide spectrum of natural products (Scheme 1). 35,36 Shikimate-and chorismate-derived metabolites and their enzymatic transformations have been applied in synthesis, notably by Frost et al. 37 and by Müller et al. [38][39][40][41][42] These multifunctionalized and conformationally restricted natural products could serve as valuable chiral auxiliaries, chiral ligands, and organocatalysts in asymmetric transformations. Here, we focus on the production of the microbial nonproteinogenic amino acids 2,3-trans-CHA and 3,4trans-CHA (4), and derivatives thereof, and their use as potent (organo)catalysts.

Results and Discussion
The aim of this study was to prove the efficiency of the βand δ-amino acids 2,3-trans-CHA (2) and 3,4-trans-CHA (4) as catalysts in Knoevenagel condensation and asymmetric aldol addition reactions. The reference compound in this work was the proteinogenic amino acid L-proline (6), which was used as an organocatalyst in its pure form and as the zinc(II) complex in Knoevenagel and aldol reactions, respectively. L-Proline (6) is a secondary α-amino acid, while 2 and 4 are primary βand δ-amino acids, respectively. L-Proline and both the amino alcohols 2 and 4 are chiral, and all three possess conformational restrictions with somewhat rigid structures. Scheme 1. Amino acids and 1,2-diol metabolites derived from chorismic acid (1) as a branching point. 33 Compounds 2 and 4 are metabolites of the shikimate biosynthetic pathway (Scheme 1). Microbial access to 2 was realized by heterologous expression of phzDE in E. coli cells, which allowed the production of up to 12 g·L -1 of 2. 33,43 By using a 300-L fed-batch culture approach, 2 was produced on a kg scale. 44 Isolation of 2 was performed by concentration of the cell-free fermentation broth and crystallization from water at 4 °C.
A concentration of 1.7 g·L -1 of 3,4-trans-CHA (4) was obtained by combining the pabAB gene and phzD gene in a recombinant E. coli strain. 33  The zinc complexes 8-10, which precipitated from the reaction mixtures and were sufficiently pure for use without the need for further purification steps, were obtained in 52-84% yield. These complexes were tested in the asymmetric aldol addition reaction of acetone and pnitrobenzaldehyde (see below).
The Knoevenagel condensation of dimethyl malonate (11) and 3-methylbutyraldehyde (12) was tested using L-proline (6), as well as the two amino acids 2 and 4. In the case of L-proline (6), the reaction was carried out using 13 mol% of the catalyst, and the product 13 was obtained in 97% chemical yield (Scheme 3).
The transformation using 2,3-trans-CHA (2) and 3,4-trans-CHA (4) was achieved by employing similar reaction conditions. In this case, the product 13 was obtained in 61% and 46% yield, respectively, clearly showing that the βand δ-amino acids can be used as catalysts in the Knoevenagel condensation.
Intermolecular aldol condensation of acetone (15) and p-nitrobenzaldehyde (14) was achieved in the presence of a catalytic amount (5 mol%) of Zn(L-proline)2 complex (8) ( Table 1, entry 1). After 4 hours, there was 83% conversion, and (S)-16 was obtained in 11% ee (as determined by chiral-phase HPLC). The reaction of 14 and 15 was carried out under similar conditions with complexes 9 and 10. We observed that the ee of the product 16 depends on two factors: the conversion and the reaction time. In the presence of Zn(2,3-trans-CHA)2 complex (9), the transformation showed a low yield of 22% with a high ee of 88% after 43 hours; after 70 hours, the conversion increased to 88% and the ee decreased to 26% (Table 1, entries 2, 3). The conversion continued to increase and the ee to decrease with increasing reaction time. We observed similar results in the case of Zn(3,4-trans-CHA)2 complex (10). A conversion of 24% with 90% ee was observed after 42 hours; after 70 hours, the conversion increased to 43% and the ee decreased to 84% (Table 1, entries 4, 5). For both catalysts 9 and 10, the (R)-enantiomer of the aldol product 16 was obtained predominantly, which nicely complements the S-selectivity of the L-proline-derived complex 8. A similar dependency of the enantiomeric excess on reaction time (or catalyst loading) has recently been described by Gröger, Berkessel and co-workers. 47 No product formation was detected when the aldol reaction was carried out using 2,3-trans-CHA (2) or 3,4-trans-CHA (4) in the pure forms. Also, the condensation product (elimination of water with formation of the ,β-unsaturated ketone) was not observed.
Our results confirm that amino acids, irrespective of the position of the amino group (α-, βand δ-amino acids were tested in parallel), are valuable chiral (organo)catalysts. While the two amino acids 2,3-trans-CHA (2) and 3,4-trans-CHA (4) proved to be just as good as the L-proline catalyst with respect to enantioselectivity, they intrinsically possess several advantages. As we and others have shown for 2 and 4, and for the respective diols 2,3-trans-CHD (3) and 3,4-trans-CHD (5), 39,41,48,49 the carboxylate moiety and the vicinal diol or amino alcohol moieties can be protected selectively. The 1,3-diene system of these compounds can be modified highly specifically through oxidative or reductive processes (e.g., epoxidation, dihydroxylation, selective hydrogenation), as well as via pericyclic reactions such as Diels-Alder and hetero-Diels-Alder additions (Scheme 5). In the case of proline (6), modification of the amino and carboxylate groups, as well as hydroxylation at C3 and C4, is restricted. Recently, Al-Momani and Lataifeh have studied the efficiency of some proline modifications and the resulting catalytic behavior in aldol reactions. 50 Hence, the catalysts 2 and 4 and derivatives thereof should be able to be tailored specifically and optimized with respect to selectivity and reactivity through the adjustment of steric or electronic effects. Moreover, these amino acids are natural products, easily accessible on a large scale starting from renewable resources. Thus, the diversity-oriented aspects of chorismate in biosynthesis have been successfully transferred to the diversity-oriented application of its derivatives in synthesis and catalysis.