$$\rightleftharpoonup{xx}$$
$$\longleftharp{xx}$$,
$$\longrightharp{xx}$$,
Despite the benefits offered by biocatalysis, the integration of biocatalytic steps in synthetic pathways or total biocatalytic routes remains mostly limited to enzymatic kinetic resolutions. These routes have been widely used as a first step in asymmetric chemo-enzymatic synthesis, but biocatalysis offers many more possibilities in functional group interconversions with high stereoselectivity1,2,3. Moreover, as biocatalytic reactions are conducted in similar conditions, it is therefore feasible to perform cascade reactions in a one-pot fashion4,5.
Chiral amino alcohols are versatile molecules for use as auxiliaries or scaffolds in organic synthesis6. The amino alcohol moiety is frequently found in secondary metabolites and in active pharmaceutical ingredients (API). Primary β-amino alcohols are readily available from the corresponding α-amino acids by conventional chemical synthesis, but access to chiral γ-amino alcohols or secondary amino alcohols often requires tedious synthetic pathways together with sensitive control of the stereochemistry7,8,9,10. Due to its high stereoselectivity, biocatalysis may provide a superior synthetic route to these chiral building blocks11,12,13,14.
We previously reported the synthesis of mono- and di-hydroxy-L-lysines by diastereoselective enzymatic hydroxylation catalyzed by dioxygenases of the iron(II)/α-ketoacid-dependent oxygenase family (αKAO) (Figure 1)15. In particular, starting from L-lysine, the KDO1 dioxygenase catalyzes the formation of the (3S)-hydroxy derivative (1), while the (4R)-derivative (2) is formed by the reaction with KDO2 dioxygenase. Successive regiodivergent hydroxylations by KDO1 and KDO2 lead to the formation of the (3R,4R)-dihydroxy-L-lysine (3) in optically pure form. However, the limited substrate range of these enzymes impedes their large utilization in chemical synthesis, especially in the hydroxylation of simple amines, as a carboxylic acid moiety in the α-position of the amino group is essential for activity16.

Figure 1: Biocatalytic conversions of L-lysine. Conversion into (3S)-hydroxy-L-lysine (1) catalyzed by KDO1 dioxygenase; (4R)-hydroxy-L-lysine (2) catalyzed by KDO2 dioxygenase; and (3R,4R)-dihydroxy-L-lysine (3) by cascade reaction catalyzed successively by KDO1 and KDO2 dioxygenases. Please click here to view a larger version of this figure.
Decarboxylation is a common reaction in metabolism17. In particular, amino acid DCs (EC 4.1.1) are cofactor-free (pyruvoyl-dependent) or PLP-dependent enzymes, and catalyze the decarboxylation of amino acids into the corresponding polyamines in bacteria and higher organisms18,19,20,21,22. The mono- and dihydroxy compounds (Figure 3) 4-7, 10-11 correspond to hydroxylated cadaverine, the diamine obtained by decarboxylation of L-lysine. Cadaverine is a key building block for the chemical industry, specifically it is a component of polyamide and polyurethane polymers. Therefore, bio-based production of this diamine from renewable resources has attracted attention as an alternative to the petroleum-based route, and various microorganisms have been engineered for this purpose. In these metabolic pathways, lysine DC (LDC) is the key enzyme. LDC is a PLP-dependent enzyme belonging to the alanine racemase (AR) structural family23. The PLP-dependent DCs (PLP-DCs) are known to be highly substrate-specific. However, a few enzymes own the capability of slight promiscuity, being active towards both L-ornithine and L-lysine amino acids, as for example the LDC from Selenomonas rumirantium (LDCSrum), which has similar kinetic constants for lysine and ornithine decarboxylation24,25. This extended substrate specificity makes this enzyme a good candidate for the decarboxylation of mono- and di-hydroxy-L-lysine. In addition, to find DCs active towards the hydroxyl derivatives of lysine, we examined the genomic context of the genes encoding the αKAO enzymes. Indeed, in prokaryotic genomes the genes encoding enzymes involved in the same biosynthetic pathway are generally co-localized in gene clusters. The KDO2 (from Chitinophaga pinensis) gene was found co-localized with a gene encoding putative PLP-DC (Figure 2). In contrast, no gene encoding for DC has been found when analyzing the genomic context of the KDO1 dioxygenase. The PLP-DC protein from C. pinensis (DCCpin) was therefore selected as a promising candidate to catalyze the decarboxylation step of the cascade reaction.

Figure 2: Genomic context of KDO2 gene in C. pinensis. Please click here to view a larger version of this figure.
Consequently, we designed enzymatic cascade reactions involving dioxygenases and DCs to achieve the synthesis of aliphatic chiral β- and γ-amino alcohols from amino acids (Figure 3). As previously reported, the C-H oxidation catalyzed by the αKAO introduces the hydroxy-substituted stereogenic center with total diastereoselectivity; the Cβ/γ chirality will be preserved in the decarboxylative step, which only affects the Cα carbon of the amino acid moiety16.

Figure 3: Retrosynthetic analysis. (A) Retrosynthesis of β- and γ-amino alcohols (R)-1,5- diaminopentan-2-ol (4) from (5R)-hydroxy-L-Lysine, and (S)-1,5-diaminopentan-2-ol (5) and 1,5-diaminopentan-3-ol (6) from L-lysine. (B) Retrosynthesis of β,γ- and β,δ-amino diols (2S,3S)-1,5-diaminopentane-2,3-diol (10) and (2R,4S)-1,5-diaminopentane-2,4-diol (11) starting from (5R)-hydroxy-L-lysine, and (2R,3R)-1,5-diaminopentane-2,3-diol (7) starting from L-lysine. Please click here to view a larger version of this figure.
Starting from L-lysine and its (5R)-hydroxy derivative, we herein report a two/three step, one pot, enzymatic procedure combining dioxygenases and PLP-DCs to obtain the target amino alcohols. Prior the synthesis at the laboratory scale of the target molecules, the method was developed at the analytical scale to adjust the reaction conditions, e.g., the enzyme concentrations, required to allow full conversion of the starting materials; we present this procedure as well.