WO2016075082A1

WO2016075082A1 - Stereoselective reductive amination of alpha-chiral aldehydes using omega-transaminases for the synthesis of precursors of pregabalin and brivaracetam

Info

Publication number: WO2016075082A1
Application number: PCT/EP2015/076077
Authority: WO
Inventors: Ferdinand Zepeck; Sven Nerdinger; Wolfgang Kroutil; Christine Fuchs; Robert Simon
Original assignee: Sandoz Ag
Priority date: 2014-11-10
Filing date: 2015-11-09
Publication date: 2016-05-19

Abstract

The present invention relates to processes comprising a combined racemization and stereoselective reductive amination step in which an aldehyde compound of formula (I) is contacted with an (R)-selective ω-transaminase or an (S)-selective ω-transaminase to racemize the compound of formula (I) and obtain an amine compound of formula (II). These processes are useful for the preparation of precursors of pharmaceutically active agents such as brivaracetam and pregabalin.

Description

Stereoselective reductive amination of a-chiral aldehydes using ω-transaminases for the synthesis of precursors of pregabaiin and brivaracetam

Field of the invention

The present invention relates to processes comprising a combined racemization and stereoselective reductive amination step in which an aldehyde compound of formula (I) is contacted either with an (f?)-selective ω-transaminase or an (S)-selective ω-transaminase to racemize the compound of formula (I) and obtain an amine compound of formula (II). These processes are useful for the preparation of precursors of pharmaceutically active agents, such as pregabaiin and brivaracetam.

Background of the invention

The compounds of formula (II) as described further below are key intermediates in the synthesis of optically active pharmaceuticals, such as the anticonvulsants pregabaiin and brivaracetam.

Pregabaiin Brivaracetam

A process for the synthesis of pregabaiin is described in EP-A-641330. However, the disclosed process is lengthy (requiring more than 10 steps), has a low efficiency, and uses pyrophoric or expensive reagents, such as butyl lithium and (+)-4-methyl-5-phenyl-2-oxazolidinone. respectively, which limits its use on an industrial scale.

Several further synthetic routes to pregabaiin are described in Hoekstra et al., 1997. Two processes of particular economic interest are disclosed in EP-A-828704 and EP-A-830338, respectively. In EP-A-828704, 3-isobutyl glutaric acid prepared from isovaleraldehyde and ethyl cyanoacetate serves as a key intermediate which is transformed via the corresponding cyclic anhydride to an amide that can be resolved in a classical manner with enantiopure phenylethylamine as the resolving agent. This amide is further subjected to a Hoffmann degradation leading to pregabalin. Improvements and variations of this process have been disclosed in WO 2006/122255, WO 2006/122258, WO 2006/122259, WO 2006/136087, WO 2007/035789, WO 2007/035790, and WO 2007/139933. Purification processes leading to pregabalin which is free of some process-related impurities are further described in WO 2006/121557.

In EP-A-830338 racemic 3-(aminomethyl)-5-methylhexanoic acid is prepared and the racemate is resolved by (S)-mandelic acid as a chiral resolution agent. The racemic starting material is prepared in five steps from isovaleraldehyde and diethylmalonate. The resolution of a racemate at the end makes the synthesis costly and inefficient because the undesired isomer has to be taken along the whole process. A variation of this process by performing the resolution prior to the reduction of the cyano group is disclosed in WO 2007/143152. Both processes suffer from disadvantages such as lengthy synthesis and low overall yield.

An asymmetric synthesis of an intermediate en route to pregabalin comprising a homogeneous catalytic hydrogenation with chiral phosphine-based ligands is disclosed in WO 2001/55090 and in WO 2005/087370. The starting material is prepared in three steps which include the use of carbon monoxide which is a hazardous reagent and Pd which is an expensive catalyst.

In WO 2006/110783 the conversion of chiral 2-(3-methyl-1 -nitromethyl-butyl)-malonic acid dialkyl ester to pregabalin using a reduction-decarboxylation strategy is described. The sequence follows a prior art reaction sequence which has been applied to the synthesis, e.g. of baclofen (Ooi et al., 2005).

WO 2009/141362 describes the synthesis of pregabalin via the enzymatic resolution of 5-methyl-3-nitromethyl-hexanoic acid ester as well as processes for the preparation of racemic 5-methyl-3-nitromethyl-hexanoic acid ester and its salts. The esterase EstC from Burkholderia gladioli preferentially hydrolyzes the (S)-enantiomer of 5-methyl-3-nitromethyl-hexanoic acid ester, leaving the (R)-enantiomer behind.

Enzymatic kinetic resolutions of nitrile-containing pregabalin precursors are disclosed in WO 2005/100580 and WO 2006/000904. These references describe syntheses of pregabalin which have the disadvantage of using potassium cyanide, the handling of which can be problematic at an industrial scale due to safety reasons. In WO 2007/1431 13 an enzymatic kinetic resolution via hydrolysis or esterification is described. Further enzymatic kinetic resolutions for the synthesis of pregabalin are disclosed in US 2009/0042262 and US 2011/0065168.

The synthesis of racemic pregabalin is described in Andruszkiewicz et al., 1989. The synthesis starts from (£)-5-methyl-hex-2-enoic ac d ethyl ester which is converted into 5-methyl-3- nitromethyl-hexanoic acid ethyl ester by a conjugate addition of nitromethane. This compound is converted into racemic pregabalin by catalytic hydrogenation followed by saponification.

WO 2014/155291 furthermore describes the synthesis of 5-hydroxy-4-(2-methylpropyl)-3,4- dihydro-5H-2-furanone and derivatives thereof, and their use as intermediates in the preparation of pregabalin.

Brivaracetam can be prepared by different methods as described, e.g., in WO 2001/062726, WO 2003/014080, WO 2005/028435, WO 2007/031263 and US 2008/0009638. All these methods, however, face major drawbacks as they require either a preparative separation on a chiral column or a separation of diastereomers via preparative HPLC. ω-Transaminases ("ω-TAs") are enzymes of the class of transaminases (EC 2.6.1.x) that catalyze the reductive amination of a carbonyl compound into an amine compound and vice versa, and are distinguished from a-transaminases which transform exclusively a-amino acids and a-keto acids, respectively.

The use of ω-transaminases for the preparation of enantioenriched or enantiopure compounds has been described, e.g., in Malik et al., 2012, Koszelewski et al., 2010, Shin et al., 2001 , and Shin et al., 1999. A stereoselective reductive amination catalyzed by ω-transaminases leading to 4-phenylpyrrolidin-2-one via dynamic kinetic resolution is furthermore described in Koszelewski et al., 2009. Moreover, Fuchs et al. described the stereoselective reductive amination of certain a-chiral 2-phenylpropanal derivatives by ω-transaminases, leading to the corresponding enantioenriched β-chiral 2-phenylpropylamine derivatives (Fuchs et al., 2013; Fuchs et al., 2014); these a-chiral 2-phenylpropanal derivatives were found to racemize spontaneously in buffer at pH 7 and could therefore be aminated via dynamic kinetic resolution. On the basis of these references, however, it could not have been expected that the specific aldehyde compounds of formula (I) to be employed in the processes of the present invention, which are structurally different from the carbonyl compounds used in these references, could be transformed by ω-transaminases into the corresponding amine compounds of formula (II) in a highly efficient and stereoselective manner and with high yield. WO 2008/127646 relates to processes for the preparation of racemic and enantiomerically enriched precursors of pregabalin and processes for the synthesis of pregabalin via the chemical or enzymatic amination of an aldehyde precursor. While the use of transaminases is mentioned in WO 2008/127646, this document fails to provide any corresponding experimental details or experimental examples. Moreover, when it was attempted to reproduce the teaching of WO 2008/127646 using various transaminases mentioned in this document, it was found that most of the tested transaminases did not convert the substrate at all or only gave racemic product. In fact, only one of the tested transaminases gave the (S)-isomer of the product, and only at a very low enantiomeric excess (see Example 1 1 ). The teaching of WO 2008/127646 is thus defective and not workable without undue effort. Furthermore, WO 2008/127646 does not provide any indication that ω-transaminases could be used for the racemization of the specific aldehyde compounds of formula (I) that are employed in the processes of the present invention. As explained further below, the racemization of the compounds of formula (I) is highly advantageous as it allows the stereoselective conversion of these compounds via dynamic kinetic resolution, leading to high yields of the desired stereoisomer even if the compound of formula (I) is employed as a racemate.

Although some processes for the synthesis of pregabalin and brivaracetam as well as their precursors are available, further improvements, particularly in terms of increasing the overall yield and increasing the degree of enantiopurity of the desired product would be highly desirable. It is thus an object of the present invention to provide improved processes for the preparation of optically active pharmaceuticals such as brivaracetam and pregabalin and precursors thereof. Summary of the invention

In the context of the present invention, it has surprisingly been found that an aldehyde compound of formula (I) can be racemized and stereoselectively aminated using an (R)-selective ω-transaminase or an (S)-selective ω-transaminase to obtain either the (f?)-isomer or the (S)-isomer of the corresponding amine compound of formula (II) in high enantiomeric excess and in high yield.

Accordingly, in a first aspect, the present invention provides a process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (I)

(I) wherein R¹ is a carboxy protecting group or hydrogen, and R² is alkyl or arylalkyi, is contacted either with an (R)-selective ω-transaminase or with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (I) and obtain a compound of the following formula (II)

(II) wherein R¹ and R² are the same as in formula (I).

It is to be understood that, throughout the present specification and independently from the choice of the group R², the (R)-isomer and the (S)-isomer of the compound of formula (II) refer to the following configurations:

In a second aspect, the invention relates to a process for the production of brivaracetam, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (la)

wherein R is a carboxy protecting group or hydrogen, is contacted with an (R)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (la) and obtain a compound of the following formula (Ila)

(Ila) wherein R¹ is the same as in formula (la), wherein the process optionally further comprises a step of converting the compound of formula (Ila) obtained in the stereoselective reductive amination step into a compound of the following formula (Ilia)

(Ilia) and wherein the compound of formula (Ila) or (Ilia) is further converted into brivaracetam having the following formula (IVa):

(IVa).

In a third aspect, the invention relates to a process for the production of pregabalin, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (lb)

wherein R¹ is a carboxy protecting group or hydrogen, is contacted with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (lb) and obtain a compound of the following formula (lib)

(lib) wherein R¹ is the same as in formula (lb), wherein, if R¹ is different from hydrogen, the process further comprises a step of converting the compound of formula (Mb) into pregabalin having the following formula (Vb) l-fclM^

Brief description of the figures

Figure 1 : Schematic illustration of the combined racemization and stereoselective reductive amination step in the processes of the invention.

Detailed description of the invention

As described above, in a first aspect the present invention provides a process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (I)

wherein R¹ is a carboxy protecting group or hydrogen, and R² is alkyl or arylalkyi, is contacted either with an (R)-selective ω-transaminase or with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (I) and obtain a compound of the following formula (I I)

(I I) wherein R¹ and R² are the same as in formula (I).

R¹ is a carboxy protecting group (such as alkyl, alkenyl, aryl, or arylalkyi) or hydrogen. Preferably, R¹ is selected from C_1-8 alkyl, C₂.₈ alkenyl, C₆.₁ aryl, (C₆.₁ aryl)-C-|.₈ alkyl, and hydrogen. More preferably, R¹ is selected from C-._A alkyl (e.g. , methyl, ethyl, propyl, or butyl), C₂-4 alkenyl (e.g . , vinyl, propenyl (such as ally I) , or butenyl), phenyl, phenyl-(C₁. alkyl) (e.g ., benzyl or phenethyl) , and hydrogen. Even more preferably, R¹ is selected from methyl, ethyl, n-propyl, isopropyl, fe/f-butyl, vinyl, allyl, benzyl, and hydrogen. Yet even more preferably, R¹ is selected from ethyl, isopropyl, fe/f-butyl, and benzyl.

R² is alkyl or arylalkyi, preferably R² is selected from Ci_-8 alkyl and (C_6-14 aryl)-Ci.₈ alkyl. More preferably, R² is selected from d ₄ alkyl and phenyl-(Ci alkyl), particularly from d.₄ alkyl. Even more preferably, R² is selected from ethyl, propyl (e.g . , n-propyl or isopropyl) , butyl (e.g . , n-butyl or isobutyl), benzyl, and phenethyl. Still more preferably, R² is selected from ethyl, n-propyl, isopropyl, isobutyl, and benzyl, and yet even more preferably R² is n-propyl or isobutyl.

This process allows the stereoselective synthesis of a compound of formula (II) having either the (/^-configuration or the (S)-config uration at the asymmetric carbon atom carrying the aminomethyl group, which carbon atom is marked with an asterisk in formula (II) . As indicated by the term "stereoselective", this process allows the production of either the (R)-isomer or the (S)-isomer of a compound of formula (I I) in excess over the respective other isomer. In particular, if the compound of formula (I) is contacted with an (R)-selective ω-transaminase, a compound of formula (I I) having the ( Reconfiguration at the carbon atom carrying the aminomethyl group will be obtained in excess over the corresponding compound of formula (II) having the (S)-configuration at this carbon atom . Conversely, if the compound of formula (I) is contacted with an (S)-selective ω-transaminase, a compound of formula (II) having the (S)-configuration at the carbon atom carrying the aminomethyl group will be obtained in excess over the corresponding (R)-isomer of the compound of formula (II).

In the context of the present invention, it has surprisingly been found that the compounds of formula (I) having the (R)-configuration or the (S)-configuration at the carbon atom carrying the aldehyde (formyl) group undergo a racemization reaction when they are contacted with an ω-transaminase under reaction conditions, as also demonstrated in Example 5. The contacting of a compound of formula (I) with an (f?)-selective ω-transaminase or an (S)-selective ω-transaminase will hence result in the formation and maintenance of an equilibrium between the two stereoisomers of the compound of formula (I) having either the ( Reconfiguration or the (S)-configuration at the carbon atom carrying the aldehyde group, and will furthermore result in the conversion of one of these two stereoisomers of the compound of formula (I) into a specific stereoisomer of the compound of formula (II) having either the (/^-configuration or the (S)-configuration at the asymmetric carbon atom carrying the aminomethyl group, as illustrated in Figure 1 . Due to the ongoing racemization of the compound of formula (I), the specific stereoisomer of this compound which is reductively aminated into a compound of formula (II) will be replenished throughout the reaction. The process of the present invention thus allows a theoretically quantitative conversion of the aldehyde compound of formula (I) into a specific stereoisomer of the corresponding amine compound of formula (II), regardless of whether any particular stereoisomer of the compound of formula (I) having either the (/^-configuration or the (S)-configuration at the asymmetric carbon atom carrying the aldehyde group or any mixture of such stereoisomers is employed as the starting material. This is highly advantageous since a theoretical yield of 100% can be obtained even if a racemic mixture of the compound of formula (I) or any other mixture of stereoisomers of the compound of formula (I) is used.

While the above-described process of the invention allows a theoretically quantitative (100%) stereoselective conversion of the compound of formula (I) into the compound of formula (II), it will be understood that in practice a lower conversion rate will actually be obtained. The present invention particularly relates to the above process, wherein at least 60 wt-%, preferably at least 70 wt-%, more preferably at least 80 wt-% , even more preferably at least 85 wt-%, and yet even more preferably at least 90 wt-% of the compound of formula (I) which is employed in the combined racemization and stereoselective reductive amination step is converted into the corresponding compound of formula (II). As explained above, throughout the present specification and independently from the choice of the group R², the (f?)-isomer and the (S)-isomer of the compound of formula (II) refer to the following configurations:

The further formulae (W-R), (l l l-R) , (IV-R), (ll-S), (l ll-S) and (V-S) as provided in the present specification also refer to the specific configuration that is depicted in the corresponding formula, regardless of the choice of the group R².

As described above, any stereoisomer of the compound of formula (I) having either the ( Reconfig uration or the (S)-configuration at the carbon atom carrying the aldehyde group or any mixture of such stereoisomers can be employed in the process according to the first aspect of the invention. Preferably, the compound of formula (I) is employed as a racemic or a non-racemic mixture of the (R)-!somer and the (S)-isomer, and most preferably the compound of formula (I) is employed as a racemate (i.e. , as a mixture of eq ual molar amounts of the compound of formula (I) having the ( Reconfig uration at the carbon atom carrying the aldehyde group and of the compound of formula (I) having the (S)-configuration at the same carbon atom). The use of the racemate of the compound of formula (I) is advantageous as it can be prepared or obtained more easily and at lower costs than a specific stereoisomer of the compound of formula (I) but nevertheless allows to obtain a theoretical yield of 100% of the desired compound of formula (II) having a specific stereochemical configuration at the carbon atom carrying the aminomethyl group in the combined racemization and stereoselective reductive amination step.

Pyridoxal-5'-phosphate (PLP) or pyridoxamine-5'-phosphate (PMP) are cofaciors of the (R)-selective or (S)-selective ω-transaminase. In the presence of the ω-transaminase, PMP reacts with the aldehyde compound of formula (I) to form a Schiff base (imine). Upon tautomerization of this Schiff base and reaction with the ω-transaminase, an amine compound of formula (II) is released and an ω-transaminase-PLP Schiff base is formed. The ω-transaminase-PLP Schiff base can react with an amine (also referred to as an amino donor), such as alanine or isopropylamine, whereby an amine-PLP Schiff base is formed and the ω-transaminase is released. Upon tautomerization of the amine-PLP Schiff base to a carbonyl-PMP Schiff base and subsequent hydrolysis, PMP is regenerated and the carbonyl (formed from the amino donor) is released. PMP again reacts with an aldehyde compound of formula (I) and the ω-transaminase, as described above, to form an amine compound of formula (II) and an ω-transaminase-PLP Schiff base. Both PMP and PLP can thus be employed as a cofactor for the ω-transaminase reaction. Accordingly, the combined racemization and stereoselective reductive amination step can be conducted in the presence of PMP and/or PLP, e.g., in an aqueous medium containing PMP and/or PLP. In terms of cost-effectiveness, the use of PLP is preferable over the use of PMP. Thus, the combined racemization and stereoselective reductive amination step can advantageously be conducted in the presence of PLP, e.g., in an aqueous medium containing PLP. However, it is also possible to use the (f?)-seiective or (S)-selective ω-transaminase without the addition of PLP or PMP, in which case the transamination reaction will proceed via PLP or PMP that is already covalently bound to the ω-transaminase as a prosthetic group (forming a Schiff base with the ω-transaminase). This possibility of conducting the combined racemization and stereoselective reductive amination step using an (R)-selective or (S)-selective ω-transaminase without adding PLP or PMP is most preferable in terms of costs, particularly if the process according to the first aspect of the invention is conducted on an industrial scale. Accordingly, the combined racemization and stereoselective reductive amination step can also be conducted in the absence of free (unbound) PLP and free (unbound) PMP.

In the combined racemization and stereoselective reductive amination step, the compound of formula (I) is contacted with the ω-transaminase in the presence of an amine (also referred to as 'amino donor"), As described above, the ω-transaminase catalyzes the transfer of an amino group from its cofactor pyridoxamine-5'-phosphate (PMP) to the aldehyde compound of formula (I), whereby PMP is oxidized to pyridoxal-5'-phosphate (PLP). The amine is used by the ω-transaminase for the reductive amination of PLP, whereby PMP is regenerated and the amine is converted into the corresponding carbonyl compound.

The amine (i.e., the amino donor) is not particularly limited, provided that it comprises at least one amino group (-NH₂), and it is preferably selected from an amino acid (e.g., alanine, arginine, asparagine, aspartate, cysteine, glutamate, glutamine, glycine, histidine, isoleucine, leucine, lysine, methionine, phenylalanine, serine, threonine, tryptophan, tyrosine, valine, ornithine, β-alanine, or 3-aminobutyric acid, particularly the L-isomer or the D-isomer of any one of the aforementioned amino acids), an alkylamine (e.g., a Ci_₈ alkylamine). an alkenylamine (e.g., a C₂_₈ alkenylamine), an arylamine (e.g., a C₆.1 arylamine), an aralkylamine (e.g., a (C₆_₁₄ aryl)-(C- ₃ alkyl)amine), an arylalkenylamine (e.g., a (C₆-i₄ aryI)-(C₂-e alkenyl)amine), a heteroarylamtne (e.g., a C₅.u heteroarylamine), a heteroarylalkylamine (e.g., a (C₅_ ₄ heteroaryl)-(Ci.₈ alkyl)amine), a heteroarylalkenylamine (e.g., a (C5-14 heteroaryl)-(C₂-₈ alkenyl)amine), pyridoxamine-5'-phosphate (PMP), and any of the aforementioned compounds substituted with at least one further amino group (e.g., a diamine of any of the aforementioned compounds). More preferably, the amine is selected from alanine (e.g., L-alanine or D-alanine), glutamate (e.g., L-glutamate or D-glutamate), ethylamine, 1 ,2-diaminoethane, n-propylamine, 1 ,3-diaminopropane, isopropylamine (i.e., 2-propylamine), 1 ,2-diaminopropane, 2-butylamine, 1 -indolamine, 1-indanamine, 1 -aminotetralin, and 1-phenylethylamine. Even more preferably, the amine is selected from alanine, glutamate, ethylamine, n-propylamine, isopropylamine, 2-butylamine, 1-indolamine, 1-indanamine, and 1 -phenylethylamine. Most preferably, the amine is selected from alanine (e.g., L-alanine or D-alanine) and isopropylamine.

The amine can be employed in the combined racemization and stereoselective reductive amination step in molar excess (e.g., in about 2-fold to about 50-fold molar excess, particularly in about 10-fold to about 25-fold molar excess) in relation to the compound of formula (I) in order to shift the reaction equilibrium to the product side. To this end, the carbonyl that is formed from the deamination of the amine can also be degraded or removed from the reaction. For example, pyruvate that is formed from alanine (used as the amino donor) can be degraded or transformed enzymatically, e.g., using acetolactate synthase, pyruvate decarboxylase or lactate dehydrogenase. The compound of formula (II) can also be removed from the reaction media in order to shift the reaction equilibrium to the product side.

Thus, in order to transform the pyruvate that is formed from alanine (used as the amino donor) with lactate dehydrogenase as described above, the combined racemization and stereoselective reductive amination step can be conducted in a reaction medium comprising:

(i) lactate dehydrogenase (LDH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (in the form of NADH or NAD⁺), and a formate salt (such as, e.g., ammonium formate); or

(ii) lactate dehydrogenase (LDH), glucose dehydrogenase (GDH), nicotinamide adenine dinucleotide (in the form of NADH or NAD^*), and glucose; or

(iii) lactate dehydrogenase (LDH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (in the form of NADH or NAD⁺), and a phosphite salt (such as, e.g., ammonium phosphite).

The lactate dehydrogenase (LDH) will reduce the pyruvate that is formed from alanine (the amino donor) using NADH and will thus produce lactate, whereby product inhibition that may be caused by the accumulation of pyruvate formed from alanine can be avoided. The NADH required for the reduction of pyruvate to lactate is oxidized to NAD⁺ and is regenerated by formate dehydrogenase (FDH), glucose dehydrogenase (GDH) or phosphite dehydrogenase (PTDH) using formate, glucose or phosphite as the corresponding reducing agent.

If the amine is alanine, it is particularly preferred that the combined racemization and stereoselective reductive amination step is conducted in a reaction medium comprising:

(i) alanine dehydrogenase (Ala-DH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (in the form of NADH or NAD⁺), and ammonium formate; or

(ii) alanine dehydrogenase (Ala-DH), glucose dehydrogenase (GDH), nicotinamide adenine dinucleotide (in the form of NADH or NAD⁺), glucose, and an ammonium salt (such as, e.g., ammonium formate, ammonium phosphite, ammonium phosphate, or ammonium acetate); or

(iii) alanine dehydrogenase (Ala-DH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (in the form of NADH or NAD⁺), and ammonium phosphite. The alanine dehydrogenase (Ala-DH) will reduce the pyruvate that is formed from alanine (the amino donor) using ammonium and NADH and will thus regenerate L-alanine. This recycling of L-alanine is advantageous as it allows to use lower amounts of L-alanine and to avoid product inhibition that may be caused by the accumulation of pyruvate formed from alanine. The NADH required for the reduction of pyruvate to L-alanine is oxidized to NAD^" and is regenerated by formate dehydrogenase (FDH), glucose dehydrogenase (GDH) or phosphite dehydrogenase (PTDH) using formate, glucose or phosphite as the corresponding reducing agent. In the presence of an alanine racemase (e.g., alanine racemase from E. colt), L-alanine is in equilibrium with D-alanine, which is the prefered amino donor for some ω-transaminases (such as, e.g., ArR-ω-ΤΑ, ΗΝ-ω-ΤΑ, NF-ω-ΤΑ, ΑΤ-ω-ΤΑ, or GZ-ω-ΤΑ, which are further described herein below). The interconversion of L-alanine and D-alanine can be sped up by the addition or coexpression of an alanine racemase. Accordingly, if D-alanine is used as the amino donor (e.g., in the case that ArR-ω-ΤΑ, ΗΝ-ω-ΤΑ. NF-ω-ΤΑ, ΑΤ-ω-ΤΑ. or GZ-ω-ΤΑ is used), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in a reaction medium (i), (ii) or (iii) as defined above, which additionally contains alanine racemase (EC 5.1.1.1).

The above-mentioned enzymes alanine dehydrogenase (Ala-DH) and either one of formate dehydrogenase (FDH), glucose dehydrogenase (GDH) and phosphite dehydrogenase (PTDH) can also be co-expressed with the corresponding (R)-selective or (S)-selective ω-transaminase in a microorganism (e.g., in a bacterium), preferably in E. coli. This is particularly advantageous if the ω-transaminase is used in the form of permeabilized microbial cells, in the form of an extract or a lysate of microbial cells, or in the form of rehydrated lyophilized microbial cells (as also described herein below).

In one embodiment of the process according to the first aspect of the invention, the ω-transaminase is an ( ?)-selective ω-transaminase. The compound of formula (II) having the (/^-configuration at the carbon atom carrying the aminomethyl group will thus be obtained in excess over the corresponding (S)-isomer in the combined racemization and stereoselective reductive amination step.

Accordingly, in this embodiment the invention provides a process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (I)

(I) wherein R¹ is a carboxy protecting group or hydrogen, and R² is alkyl or arylalkyi, is contacted with an (R)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (I) and obtain a compound of the following formula (ll-R)

(li-R) wherein R¹ and R² are the same as in formula (I).

In this embodiment, R¹ is a carboxy protecting group (such as alkyl, alkenyl, aryl, or arylalkyi) or hydrogen. Preferably, R¹ is selected from Ci_₈ alkyl, C₂.₈ alkenyl, C₆_₁₄ aryl, (C₆_₁ arylJ-C^s alkyl, and hydrogen. More preferably, R¹ is selected from d. alkyl (e.g., methyl, ethyl, propyl, or butyl), C₂-₄ alkenyl (e.g., vinyl, propenyl (such as allyl), or butenyl), phenyl, phenyl-(Ci_₄ alkyl) (e.g., benzyl or phenethyl), and hydrogen. Even more preferably, R¹ is selected from methyl, ethyl, n-propyl, isopropyl, fe/f-butyl, vinyl, allyl, benzyl, and hydrogen. Yet even more preferably, R¹ is selected from ethyl, isopropyl, fert-butyl, and benzyl. In this embodiment, R² is alkyl or arylalkyl, preferably R² is selected from C _-8 alky! and (CV₁₄ aryl)-Ci.8 alkyl. More preferably, R² is selected from C^₄ alkyl and phenyl-(C_1-4 alkyl) (e.g., benzyl or phenethyl). Even more preferably, R² is alkyl. Yet even more preferably, R² is selected from ethyl, propyl (e.g., n-propyl or isopropyl), and butyl (e.g., n-butyl or isobutyl). Most preferably, R² is n-propyl.

It is particularly preferred that R² is alkyl (such as, e.g., ethyl, n-propyl, isopropyl, n-butyl, or isobutyl) and R¹ is selected from methyl, ethyl, n-propyl, isopropyl, terf-butyl, vinyl, allyl, benzyl, and hydrogen; even more preferably, R² is n-propyl and R¹ is selected from ethyl, isopropyl, ferf-butyl, and benzyl.

The (R)-selective ω-transaminase may be any ω-transaminase that is capable of catalyzing the stereoselective reductive amination of a compound of formula (I) into the corresponding compound of formula (II) so that a greater molar amount of the (R)-isomer of the compound of formula (II), i.e. the compound of formula (II) having the ( Reconfiguration at the carbon atom carrying the aminomethyl group, than of the (S)-isomer of the compound of formula (II), i.e. the compound of formula (II) having the (S)-configuration at the carbon atom carrying the aminomethyl group, is obtained. This capability can be tested, e.g., following any of the protocols described in the Examples in order to identify suitable (R)-selective ω-transaminases.

Preferably, the (R)-selective ω-transaminase is selected from:

ω-transaminase from Hyphomonas neptunium ("ΗΝ-ω-ΤΑ");

ω-transaminase from Arthrobacter citreus ("ArS-ω-ΤΑ");

ω-transaminase from Chromobacterium violaceum DSM 30191 ("CV-ω-ΤΑ");

ω-transaminase from Pseudomonas fluorescens ("PF-ω-ΤΑ");

ω-transaminase from Pseudomonas putida KT2440 gene PP2180 ("ΡΡ2-ω-ΤΑ");

ω-transaminase from Ochrobactrum anthropi ("ΟΑ-ω-ΤΑ");

ω-transaminase from Si!icibacter pomeroyi ("SP-ω-ΤΑ");

ω-transaminase from Vibrio fluvialis ("VF-ω-ΤΑ") ;

uj-transaminase from Neosartorya fischeri (" F-ω-ΤΑ");

ω-transaminase from Paracoccus denitrificans ("PD-ω-ΤΑ");

ω-transaminase from Pseudomonas putida KT2440 gene PP5182 ("ΡΡ1 -ω-ΤΑ");

ω-transaminase from Arthrobacter sp. KNK168 ("ArR-ω-ΤΑ");

ArR-G131 F-u)-TA;

ArRmut1 1-M1 17F-G279A-uj-TA; and a mutein of any one of the aforementioned ω-transaminases.

More preferably, the (R)-selective ω-transaminase is selected from ΗΝ-ω-ΤΑ, ArS-ω-ΤΑ, CV-ω-ΤΑ, PF-ω-ΤΑ, ΡΡ2-ω-ΤΑ, ΟΑ-ω-ΤΑ, SP-ω-ΤΑ, VF-ω-ΤΑ, NF-ω-ΤΑ, PD-ω-ΤΑ, and a mutein of any of these ω-TAs. Particularly preferred (R)-selective ω-transaminases are ΗΝ-ω- TA and muteins thereof, with ΗΝ-ω-ΤΑ being most preferred.

The stereoselectivity of an ω-transaminase generally depends on the structure of the substrate to be aminated. When a compound of formula (I) is used as a substrate, the extent of the stereoselectivity of the above-mentioned (R)-selective ω-transaminases will thus vary depending on the choice of the groups R¹ and R², and in particular on the choice of R since this group has been found to have a considerably greater influence on the stereoselectivity of the ω-transaminases than the group R², as also demonstrated in Examples 4 and 7. Depending on the choice of the group R¹ in the compound of formula (I), the use of certain (R)-selective ω-transaminases will thus allow a particularly advantageous stereoselective conversion.

Accordingly, if R¹ in the compound of formula (I) is ethyl and R² is d_4 alkyl (e.g., n-propyl), it is preferred to use an (^-selective ω-transaminase chosen from ΗΝ-ω-ΤΑ, ArS-ω-ΤΑ and muteins thereof, more preferably chosen from ΗΝ-ω-ΤΑ and ArS-ω-ΤΑ, and most preferably ΗΝ-ω-ΤΑ.

If R¹ in the compound of formula (I) is isopropyl and R² is C_1-4 alkyl (e.g., n-propyl), it is preferred to use an (R)-selective ω-transaminase chosen from ΗΝ-ω-ΤΑ and muteins thereof, most preferably ΗΝ-ω-ΤΑ.

If R¹ in the compound of formula (I) is /erf-butyl and R² is d . alkyl (e.g., n-propyl), it is preferred to use an (R)-selective ω-transaminase chosen from ΗΝ-ω-ΤΑ, PD-ω-ΤΑ, PF-ω-ΤΑ, VF-ω-ΤΑ, ΟΑ-ω-ΤΑ, GZ-ω-ΤΑ and muteins thereof, more preferably chosen from ΗΝ-ω-ΤΑ and VF-ω-ΤΑ, and most preferably ΗΝ-ω-ΤΑ.

If R^" in the compound of formula (I) is benzyl and R² is C, alkyl (e.g., n-propyl), it is preferred to use an (R)-selective ω-transaminase chosen from ΗΝ-ω-ΤΑ, ArR-G131 F-u>-TA and muteins thereof, and most preferably ΗΝ-ω-ΤΑ.

It is particularly preferred that the (R)-selective ω-transaminase is ΗΝ-ω-ΤΑ and that R¹ is isopropyl or terf-butyl. Even more preferably, the (R)-selective ω-transaminase is ΗΝ-ω-ΤΑ, R¹ is isopropyl or fe/f-butyl, and R² is C-^ alkyl (such as, e.g. , ethyl, n-propyl , isopropyl, n-butyl, or isobutyl). Most preferably, the ( )-seiective ω-transaminase is ΗΝ-ω-ΤΑ, R¹ is isopropyl or ferf-butyl, and R² is n-propyl. The combined racemization and stereoselective reductive amination step in this embodiment of the process according to the first aspect of the invention allows to obtain the (R)-isomer of the amine compound of formula (II), i.e. the compound of formula (l l-R), in advantageously high enantiomeric excess. Preferably, the compound of formula (ll-R) is obtained in an enantiomeric excess (ee) of at least about 50%, more preferably at least about 60%, even more preferably at least about 70% , even more preferably at least about 80% , and yet even more preferably at least about 90%.

The combined racemization and stereoselective reductive amination step can be conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% (e.g. , about 1 0 vol-% to about 30 vol-%) of at least one organic cosolvent. The organic cosolvent may, e.g., be selected from 1 , 2-dimethoxyethane (DME), dimethylformamide (DMF) , dimethyl sulfoxide (DMSO), tetrahydrofuran (THF), acetonitriie (MeCN), C ₆ alkanols (e.g. , methanol, ethanol, n-propanol or isopropanol) , and mixtures thereof, and it is preferably selected from DME, DMF and DMSO.

As demonstrated in Example 6, the stereoselectivity of an ω-transaminase is also influenced by the presence of such organic cosolvents in the aqueous reaction medium. In particular, the use of organic cosolvents such as DME, DMF and DMSO was found to improve the stereoselectivity of (f?)-selective ω-transaminases such as PF-ω-ΤΑ, VF-ω-ΤΑ, PD-ω-ΤΑ, ArS-ω-ΤΑ and ΟΑ-ω-ΤΑ in the reductive amination of compounds of formula (I). In the case of Η Ν-ω-ΤΑ, however, the use of these organic cosolvents was found to negatively affect the conversion of compounds of formula (I).

Thus, if ΗΝ-ω-ΤΑ is used as the (R)-selective ω-transaminase (and particularly if ΗΝ-ω-ΤΑ is used , R¹ in the compound of formula (I) is iert-butyl, and R² is

alkyl such as n-propyl), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising less than about 5 vol-% DME, less than about 5 vol-% DMF, and less than about 5 vol-% DMSO. More preferably, the combined racemization and stereoselective reductive amination step using ΗΝ-ω-ΤΑ is conducted in an aqueous medium comprising less than about 3 vol-% DME, less than about 3 vol-% DMF, and less than about 3 vol-% DMSO. Even more preferably, the combined racemization and stereoselective reductive amination step using ΗΝ-ω-ΤΑ is conducted in an aqueous medium comprising less than about 1 vol-% DME, less than about 1 vol-% DMF, and less than about 1 vol-% DMSO. Even more preferably, the combined racemization and stereoselective reductive amination step using ΗΝ-ω-ΤΑ is conducted in an aqueous medium that is free of DME, DMF and DMSO. It is particularly preferred that, if ΗΝ-ω-ΤΑ is used as the (/^-selective ω-transaminase, the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising less than about 5 vol-%), more preferably less than about 3 vol-%, and even more preferably less than about 1 vol-% of any organic cosolvents (in total). If PF-ω-ΤΑ is used as the (R)-selective ω-transaminase (and particularly if PF-ω-ΤΑ is used, R¹ in the compound of formula (I) is fe/f-butyl, and R² is C1.4 alkyl such as n-propyl), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent selected from DME, DMF and DMSO, preferably from DME and DMF. More preferably, the combined racemization and stereoselective reductive amination step using PF-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME (more preferably about 5 vol-% to about 25 vol-% DME, even more preferably about 15 vol-% to about 25 vol-% DME, and yet even more preferably about 20 vol-% DME) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF (more preferably about 15 vol-% to about 35 vol-% DMF, even more preferably about 25 vol-% to about 35 vol-% DMF, and yet even more preferably about 30 vol-% DMF). Even more preferably, the combined racemization and stereoselective reductive amination step using PF-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME, yet more preferably about 5 vol-% to about 25 vol-% DME, still more preferably about 15 vol-% to about 25 vol-% DME, and most preferably about 20 vol-% DME.

If VF-ω-ΤΑ is used as (R)-selective ω-transaminase (and particularly if VF-ω-ΤΑ is used, R¹ in the compound of formula (I) is terf- butyl, and R² is C-, alkyl such as n-propyl), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent selected from DME, DMF and DMSO, preferably DME. More preferably, the combined racemization and stereoselective reductive amination step using VF-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME (more preferably about 25 vol-% to about 35 vol-% DME, and even more preferably about 30 vol-% DME) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF (more preferably about 5 vol-% to about 25 vol-% DMF, even more preferably about 15 vol-% to about 25 vol-% DMF, and yet even more preferably about 20 vol-% DMF) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMSO (more preferably about 5 vol-% to about 25 vol-% DMSO, even more preferably about 5 vol-% to about 15 vol-% DMSO, and yet even more preferably about 10 vol-% DMSO). Even more preferably, the combined racemization and stereoselective reductive amination step using VF-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME, yet even more preferably about 25 vol-% to about 35 vol-% DME, and most preferably about 30 vol-% DME. The use of VF-ω-ΤΑ in an aqueous medium comprising about 30 vol-% DME is particularly advantageous with respect to the stereoselectivity of the reductive amination reaction. However, the use of VF-ω-ΤΑ in an aqueous medium comprising about 5 vol-% to about 25 vol-% DME (e.g. , about 10 vol-% or about 20 voi-% DME) is also advantageous as it provides a favorable conversion rate and, at the same time, a favorable albeit lower stereoselectivity than that obtained with 30 vol-% DME.

Accordingly, it is particularly preferred that the (R)-selective ω-transaminase is VF-ω-ΤΑ, that R¹ in the compound of formula (I) is feri-butyl, that R² is C_1-4 alkyl (particularly n-propyl), and that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME, more preferably about 25 vol-% to about 35 vol-% DME, and even more preferably about 30 vol-% DME. If PD-ω-ΤΑ is used as (R)-selective ω-transaminase (and particularly if PD-ω-ΤΑ is used, R¹ in the compound of formula (I) is ferf-butyl, and R² is C_1-4 alkyl such as n-propyl) , it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent selected from DME, DMF and DMSO, preferably from DME and DMF. More preferably, the combined racemization and stereoselective reductive amination step using PD-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME (more preferably about 5 vol-% to about 25 vol-% DME, even more preferably about 5 vol-% to about 15 vol-% DME, and yet even more preferably about 10 vol-% DME) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF (more preferably about 5 vol-% to about 25 vol-% DMF, even more preferably about 5 vo!-% to about 15 vol-% DMF, and yet even more preferably about 10 vol-% DMF) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMSO (more preferably about 5 vol-% to about 25 vol-% DMSO, even more preferably about 5 vol-% to about 15 vol-% DMSO, and yet even more preferably about 10 vol-% DMSO). Even more preferably, the combined racemization and stereoselective reductive amination step using PD-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME (more preferably about 5 vol-% to about 25 vol-% DME, even more preferably about 5 vol-% to about 15 vol-% DME, and yet even more preferably about 10 vol-% DME) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF (more preferably about 5 vol-% to about 25 vol-% DMF, even more preferably about 5 vol-% to about 15 vol-% DMF, and yet even more preferably about 10 vol-% DMF). Yet even more preferably, the combined racemization and stereoselective reductive amination step using PD-us-TA is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME, even more preferably about 5 vol-% to about 25 vol-% DME, still more preferably about 5 vol-% to about 15 vol-% DME, and most preferably about 10 vol-% DME.

If ArS-ω-ΤΑ is used as (/?)-selective ω-transaminase (and particularly if ArS-ω-ΤΑ is used, R¹ in the compound of formula (I) is ie/f-butyl, and R² is C-|.₄ aikyl such as n-propyl), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of DME, more preferably about 5 vol-%) to about 35 vol-% DME, even more preferably about 15 vol-% to about 35 vol-% DME, yet even more preferably about 25 vol-% to about 35 vol-% DME, and most preferably about 30 vol-% DME. The use of ArS-ω-ΤΑ in an aqueous medium comprising about 30 vol-% DME is particularly advantageous with respect to the stereoselectivity of the reductive amination reaction. However, the use of ArS-ω-ΤΑ in an aqueous medium comprising about 5 vol-% to about 25 vol-% DME (e.g. , about 10 vol-% DME or about 20 vol-% DME) is also advantageous as it provides a favorable conversion rate and, at the same time, a favorable albeit lower stereoselectivity than that obtained with 30 vol-% DME.

If ΟΑ-ω-ΤΑ is used as (R)-selective ω-transaminase (and particularly if ΟΑ-ω-ΤΑ is used, R¹ in the compound of formula (I) is fe/f-butyl, and R² is C,.₄ alkyl such as n-propyl), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent selected from DME, DMF and DMSO, preferably from DME and DMF, and particularly DMF. More preferably, the combined racemization and stereoselective reductive amination step using ΟΑ-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME (more preferably about 5 vol-% to about 25 vol-% DME, even more preferably about 15 vol-% to about 25 vol-% DME, and yet even more preferably about 20 vol-% DME) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF (more preferably about 5 vol-% to about 25 vol-% DMF, even more preferably about 15 vol-% to about 25 vol-% DMF, and yet even more preferably about 20 vol-% DMF) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMSO (more preferably about 5 vol-% to about 25 vol-% DMSO, even more preferably about 15 vol-% to about 25 vol-% DMSO, and yet even more preferably about 20 vol-% DMSO). Even more preferably, the combined racemization and stereoselective reductive amination step using ΟΑ-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DME (more preferably about 5 vol-% to about 25 vol-% DME, even more preferably about 15 vol-% to about 25 vol-% DME, and yet even more preferably about 20 vol-% DME) or in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF (more preferably about 5 vol-% to about 25 vol-% DMF, even more preferably about 15 vol-% to about 25 vol-% DMF, and yet even more preferably about 20 vol-% DMF). Yet even more preferably, the combined racemization and stereoselective reductive amination step using ΟΑ-ω-ΤΑ is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% DMF, more preferably about 5 vol-% to about 25 vol-% DMF, even more preferably about 15 vol-% to about 25 vol-% DMF, and most preferably about 20 vol-% DMF.

The (R)-selective ω-transaminase to be used in the combined racemization and stereoselective reductive amination step in the process according to the first aspect of the invention, including any of the specific ( )-selective ω-transaminases described herein, can be prepared by recombinant expression in a microorganism, e.g., in a bacterium such as Escherichia coli, or can be prepared by peptide synthesis. It is preferred to use an (R)-selective ω-transaminase (including any of the specific (R)-selective ω-transaminases described herein) that has been expressed in a bacterium, and most preferably in E. coli, in the combined racemization and stereoselective reductive amination step. The (/^-selective ω-transaminase can be used, e.g., in purified form, in partly purified form, in the form of permeabilized microbial cells, in the form of extracts or lysates of microbial cells, or in the form of rehydrated lyophilized microbial cells (i.e., in the form of a rehydrated microbial cell lyophilisate). Moreover, the ( )-selective ω-transaminase to be used in the combined racemization and stereoselective reductive amination step, including any of the specific (R)-selective ω-transaminases described herein, can be employed in immobilized form. For example, the (R)-selective ω-transaminase can be chemically, physically or by means of genetic engineering methods adsorbed at and/or covalently bound at/in a chemically organic or inorganic support material, wherein the catalytic activity of the enzyme is retained and the immobilized (R)-selective ω-transaminase can be used repeatedly and continuously. Accordingly, the (R)-selective ω-transaminase to be used in the combined racemization and stereoselective reductive amination step can be immobilized on a solid support material. Suitable solid support materials are known in the art and include, e.g., polymer matrices such as calcium alginate, polyacrylamide. polymethacrylate, polystyrene, polystyrene/polymethyacrylate, Sepabeads^®, ReliZyme™ Diaion^® and Eupergit^® as well as inorganic matrices such as Celite^®. Further specific solid support materials that can be used in the processes of the invention as well as immobilization techniques are described in: Sheldon, 2007; Homaei et al. , 2013; Sumitra et al. , 2013; Knezevic-Jugovic et al. , 201 1 ; Bornscheuer, 2003; and Bickerstaff Jr GF, 1997. The use of the ω-transaminase in immobilized form is advantageous because the enzyme and the product can be easily separated and the immobilized enzyme may be recycled and reused, thus rending the process more economic. It is also possible to use cells (e.g., permeabilized microbial cells, such as permeabilized E. coli cells) containing the (R)-selective ω-transaminase, which cells are immobilized on a solid support material (e.g., any of the specific solid support materials described above). The combined racemization and stereoselective reductive amination step can be conducted in an aqueous medium at any suitable pH, for example, at a pH in the range of about 6 to about 1 1. The aqueous medium may comprise an aqueous buffer system such as, e.g. , a phosphate buffer (e.g., an admixture of K₂HP0₄ and KH₂P0 ), a Tris buffer (e.g. , an admixture of tris(hydroxymethyl)aminomethane and a strong inorganic acid, such as hydrochloric acid) , a PIPES buffer (e.g. , an admixture of piperazine-N,N'-bis(2-ethanesulfonic acid) and an alkali hydroxide, such as NaOH or KOH), or a HEPES buffer (e.g., an admixture of 4-(2- hydroxyethyl)-1 -piperazineethanesulfonic acid and an alkali hydroxide, such as NaOH or KOH). It is also possible that the aqueous medium does not comprise any buffer, i.e., does not comprise any buffering agent.

Moreover, the combined racemization and stereoselective reductive amination step can be carried out at any suitable temperature, e.g., at a temperature of about 20°C to about 50°C, preferably about 25°C to about 45°C. If ArR-ω-ΤΑ is used as (R)-selective ω-transaminase, it is preferred that the combined racemization and stereoselective reductive amination step is conducted at a temperature of about 35°C to about 50°C, more preferably at about 45°C.

The process may further comprise a step of converting the compound of formula (ll- ) into a compound of the following formula (\\\-R)

(ll l-R) wherein R² is the same as in formula (ll-R). This cyclization reaction of the compound of formula {W-R) into the compound of formula (ll l-R) can occur spontaneously in the reaction medium by addition of a base, e.g . , a strong base like NaOH but can also be avioded, e.g. , by using a weak base like NaHC0₃. The compound of formula (W-R) or (l ll-R) can be used for the preparation of compounds of the racetam family, such as brivaracetam. Accordingly, the compound of formula (W-R) or (l l l-R) can be further converted into a compound of the following formula (IV-R):

(IV-R) wherein R² is the same as in formula (l l-R) or (l ll-R) .

Such a conversion can be carried out, e.g . , in accordance with or in analogy to the corresponding synthetic route described in Kenda et al. , 2004. Thus, the compound of formula (ll l-R) can be subjected to an alkylation reaction with a (S)-2-halobutyric ester (e.g . , a (C_1-6 alkyl) (S)-2-halobutyrate, such as methyl (S)-2-bromobutyrate or ethyl (S)-2-bromobutyrate) and an aminolysis reaction (e.g ., by reacting the product obtained in the alkylation step with ammonia) to obtain the corresponding compound of formula (W-R). For example, ethyl (S)- 2-bromobutyrate can be refluxed in acetonitrile with the compound of formula (lll-R) under addition of sodium hydrid, and after aminolysis and column chromatography the corresponding compound of formula (IV-R) is obtained. The conversion of a compound of formula (l l l-R) into a compound of formula (iV-R) can also be carried out, e.g . , in accordance with or in analogy to any of the synthetic routes described in WO 2007/031263.

Brivaracetam corresponds to the compound of formula (IV-R) wherein R² is n-propyl:

Brivaracetam In a further embodiment of the process according to the first aspect of the invention , the ω-transaminase is an (S)-selective ω-transaminase. The compound of formula (I I) having the (S)-configuration at the carbon atom carrying the aminomethyl group will thus be obtained in excess over the corresponding (f?)-isomer in the combined racemization and stereoselective reductive amination step.

wherein is a carboxy protecting group or hydrogen, and R² is alkyl or arylalkyl, is contacted with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (I) and obtain a compound of the following formula (II- S)

(ll-S) wherein R¹ and R² are the same as in formula (I).

In this embodiment, R is a carboxy protecting group (such as alkyl, alkenyl, aryl, or arylalkyl) or hydrogen. Preferably, R¹ is selected from C_1-8 alkyl, C₂.₈ alkenyl, C₆. ₄ aryl, (C₆_₁₄ aryl)-C_1-8 alkyl, and hydrogen. Even more preferably, R¹ is selected from C- alkyl (e.g. , methyl, ethyl, propyl, or butyl) , C₂- alkenyl (e.g . , vinyl, propenyl (such as allyl) , or butenyl), phenyl, phenyl- (Ci_₄ alkyl) (e.g . , benzyl or phenethyl) , and hydrogen. Even more preferably, R¹ is selected from methyl, ethyl, n-propyl, isopropyl, ferf-butyl, vinyl, allyl, benzyl, and hydrogen. Yet even more preferably, R¹ is selected from ethyl, isopropyl, ferf-butyl, and benzyl.

In this embodiment, R² is alkyl or arylalkyl, preferably R² is selected from Ci_₈ alkyl and (C₆.₁ aryl)-C_1-8 alkyl. More preferably, R² is selected from d_₄ alkyl and phenyl-(C, .₄ alkyl) . Even more preferably, R² is selected from ethyl, propyl (e.g., n-propyl or isopropyl), butyl (e.g., n-butyl or isobutyl), benzyl, and phenethyl. Yet even more preferably, R² is selected from isobutyl, n-propyl, isopropyl, ethyl, and benzyl. Most preferably, R² is isobutyl. It is particularly preferred that R² is d.₄ alkyl (such as, e.g., ethyl, n-propyl, isopropyl, n-butyl, or isobutyl) or phenyl-(C_{1 -4} alkyl) (such as, e.g., benzyl or phenethyl) and R¹ is selected from methyl, ethyl, n-propyl, isopropyl, rerf-butyl, vinyl, allyl, benzyl, and hydrogen; even more preferably, R² is isobutyl and R¹ is selected from ethyl, isopropyl, ferf-butyl, and benzyl. The (S)-selective ω-transaminase may be any ω-transaminase that is capable of catalyzing the stereoselective reductive amination of a compound of formula (I) into the corresponding compound of formula (II) so that a greater molar amount of the (S)-isomer of the compound of formula (II), i.e. the compound of formula (I I) having the (S)-configuration at the carbon atom carrying the aminomethyl group, than of the (R)-isomer of the compound of formula (II), i.e. the compound of formula (II) having the (/^-configuration at the carbon atom carrying the aminomethyl group, is obtained. This capability can be tested, e.g., following any of the protocols described in the Examples in order to identify suitable (S)-selective ω-transaminases. Preferably, the (S)-seiective ω-transaminase is selected from:

ω-transaminase from Gibberella zeae ("GZ-ω-ΤΑ");

ω-transaminase from Aspergillus terreus ("ΑΤ-ω-ΤΑ");

ω-transaminase from Arthrobacter sp. KNK168 ("ArR-ω-ΤΑ");

ω-transaminase from Ralstonia eutropha ("RE-ω-ΤΑ");

ω-transaminase from Bacillus megaterium ("ΒΜ-ω-ΤΑ");

ArRmutl 1-ω-ΤΑ;

ArR-S218P-G131 F-o TA;

ArRmutl 1 -M1 17F-W-TA;

ArRmutl 1 -G279A-to-TA;

ArRmutl 1-M 117F-A60V-u TA;

ArRmutl 1-M 1 17F-A60V-G279A-iu-TA;

ArRmutl 1-M 1 17F-A60V-G279V-UJ-TA;

ArRmutl 1-M 1 17F-A60V-G279L-co-TA:

ArRmutl 1 -M1 17F-A6QV-G279l-oo-TA;

ArRmutl 1 -M 1 17F-A60V-G279F-GJ-TA;

ArRmutl 1-M117F-A60V-G279V-I 152V-U)-TA;

ArRmutl 1 -M 1 17F-A60V-G279V-S277T-U-TA: GZ-L56V- J-TA;

GZ-νδΟΤ-ω-ΤΑ;

ΘΖ-Ε1 15Τ-ω-ΤΑ;

GZ-T273S-OJ-TA;

GZ-EH SD-ω-ΤΑ;

GZ-V148A-oo-TA;

GZ-L56V-T273S-u>-TA;

GZ-T273S-A275V-UJ-TA; and

a mutein of any one of the aforementioned ω-transaminases.

More preferably, the (S)-selective ω-transaminase is selected from GZ-ω-ΤΑ, ΑΤ-ω-ΤΑ, ArR- S218P-G 131 F-ω-ΤΑ, ArRmutl 1 -G279A-0J-TA, ArRmutl 1 -M1 17F-A60V-G279A-co-TA, ArRmutl 1 -M 1 1 7F-A60V-G279V-u>-TA, ArRmutl 1 -M1 17F-A60V-G279L-u)-TA, ArRmutl 1 - M 1 1 7F-A60V-G279I-UJ-TA, ArRmutl 1 -M 1 17F-A60V-G279F-aa-TA, ArRmutl 1 -M1 17F-A60V- G279V-I 152V-00-TA. ArRmutl 1 -M 1 1 7F-A60V-G279V-S277T-(jo-TA, GZ-L56V-u>-TA, GZ-T273S-UJ-TA, GZ-E1 5D-u)-TA, GZ-L56V-T273S-u>-TA, and a mutein of any of these ω- TAs. Even more preferably, the (S)-selective ω-transaminase is selected from GZ-ω-ΤΑ, ArRmutl 1 -M 1 1 7F-A60V-G279V-oo-TA, ArRmutl 1 -M1 17F-A60V-G279L-U)-TA, ArRmutl 1 - M1 1 7F-A60V-G279I-U)-TA, ArRmutl 1 -M 1 17F-A60V-G279F-00-TA, ArRmutl 1 -M1 17F-A60V- G279A- -TA, ArRmutl 1 -M 1 7F-A60V-G279V-I 152V-U)-TA, GZ-L56V-u)-TA, GZ-T273S-U)-TA, GZ-E1 15D-oj-TA, and GZ-L56V-T273S-oo-TA. Particularly preferred (S)-selective ω-transaminases are GZ-ω-ΤΑ, ArRmutl 1 -M 1 17F-A60V-G279V-co-TA, ArRmutl 1 -M 1 1 7F- A60V-G279l- -TA, GZ-L56V- -TA, and GZ-T273S^-TA. The stereoselectivity of an ω-transaminase generally depends on the structure of the substrate to be aminated . When a compound of formula (I) is used as a substrate, the extent of the stereoselectivity of the above-mentioned (S)-selective ω-transaminases will thus vary depending on the choice of the groups R¹ and R², and in particular on the choice of R¹ since this group has been found to have a considerably greater influence on the stereoselectivity of the ω-transaminases than the group R², as also demonstrated in Examples 4 and 7. Depending on the choice of the group R¹ in the compound of formula (I), the use of certain (S)-selective ω-transaminases will thus allow a particularly advantageous stereoselective conversion . Accordingly, if R¹ in the compound of formula (I) is ethyl and R² is Ci„₄ alkyl (e.g. , isobutyl or n-propyl) , it is preferred to use an (S)-selective ω-transaminase chosen from ArRmutl 1 - M1 17F-A60V-G279A-u>-TA, ArRmutl 1 -M 17F-A60V-G279V-u TA, and ArRmutl 1 -M1 17F- A60V-G279F-U)-TA, most preferably ArRmutl 1-M1 17F-A60V-G279V-u)-TA.

If R¹ in the compound of formula (I) is isopropyl and R² is C- alkyl (e.g. , isobutyl or n-propyl), it is preferred to use ΑΤ-ω-ΤΑ as (S)-selective ω-transaminase.

If R¹ in the compound of formula (I) is ferf-butyl and R² is Ci.₄ alkyl (e.g., isobutyl or n-propyl), it is preferred to use an (S)-selective ω-transaminase chosen from GZ-ω-ΤΑ, ArRmutl 1 - 1 17F- A60V-G279V-oo-TA, ArRmutl 1 -M1 17F-A60V-G279L-U0-TA, ArRmutl 1 -M1 17F-A60V-G279I-U⁾- TA, ArRmutl 1- 1 17F-A60V-G279V-l 152V-u>-TA, ArRmutl 1 -M1 17F-A60V-G279V-S277T-OJ- TA, GZ-L56V-u)-TA, GZ-T273S-00-TA, GZ-E1 15D-u)-TA, and GZ-L56V-T273S-oo-TA, more preferably chosen from GZ-ω-ΤΑ, ArRmutl 1 -M1 17F-A60V-G279V-oo-TA, ArRmutl 1 -M1 17F- A60V-G279L-u)-TA, ArRmutl 1 -M1 17F-A60V-G279l-to-TA. ArRmutl 1 -M 1 17F-A60V-G279V- Ι 152ν-ω-ΤΑ, GZ-L56V-00-TA, GZ-T273S-QJ-TA, and GZ-L56V-T273S-u>-TA, even more preferably chosen from GZ-ω-ΤΑ, ArRmutl 1 -M1 17F~A60V-G279V-oo-TA, ArRmutl 1- 1 17F- A60V-G279I-UJ-TA, GZ-L56V-u)-TA, and GZ-T273S-oo-TA, yet even more preferably chosen from ArRmutl 1 - 1 17F-A60V-G279V- )-TA, ArRmutl 1 -M1 17F-A60V-G279l-u)-TA, and GZ- L56V-OJ-TA, and most preferably chosen from ArRmutl 1 -M1 17F-A60V-G279V-oo-TA and GZ- L56V-uo-TA.

If R¹ in the compound of formula (I) is benzyl and R² is C_:..«. alkyl (e.g., isobutyl or n-propyl), it is preferred to use an (S)-selective ω-transaminase chosen from ArRmutl 1 -M1 17F-A60V- G279V-00-TA, ArRmutl 1-M1 7F-A60V-G279L-W-TA, ArRmutl 1 -M1 17F-A60V-G279l- -TA, and ArRmutl 1 -M 1 17F-A60V-G279F^-TA. more preferably chosen from ArRmutl 1 -M1 17F- A6QV-G279V-u>-TA and ArRmutl 1 -M1 17F-A60V-G279F- -TA.

It is particularly preferred that the (S)-selective ω-transaminase is chosen from ArRmutl 1 - M1 17F-A60V-G279V-W-TA, ArRmutl 1 -M 1 1 7F-A60V-G279l- -TA and GZ-L56V^-TA and that R¹ is ie/f-butyl. Even more preferably, the (S)-selective ω-transaminase is chosen from ArRmutl 1-M 1 17F-A60V-G279V-u TA, ArRmutl 1 - 1 17F-A60V-G279I-LO-TA and GZ-L56V- - TA, R¹ is iert-butyl, and R² is Ci_₄ alkyl (e.g. , isobutyl or n-propyl, particularly isobutyl). Yet even more preferably, the (S)-selective ω-transaminase is ArRmutl 1 -M 1 17F-A60V-G279V-u- TA or GZ-L56V- -TA, R¹ is ferf-butyl, and R² is C _-4 alkyl (particularly isobutyl). The combined racemizaiion and stereoselective reductive amination step in this embodiment of the process according to the first aspect of the invention allows obtaining the (S)-isomer of the amine compound of formula (II), i.e. the compound of formula (l l-S), in advantageously high enantiomeric excess. Preferably, the compound of formula (W-S) is obtained in an enantiomeric excess (ee) of at least about 30% , more preferably at least about 40% , even more preferably at least about 50%, even more preferably at least about 60% , and yet even more preferably at least about 70%.

The combined racemization and stereoselective reductive amination step can be conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% (e.g. , about 10 vol-% to about 30 vol-%) of at least one organic cosolvent. The organic cosolvent may, e.g. , be selected from 1 ,2-dimethoxyethane (DME), dimethylformamide (DMF), dimethyl sulfoxide (DMSO), tetrahydrofuran (THF), acetonitriie (MeCN), d.₆ alkanols (e.g . , methanol, ethanol, n-propanol or isopropanol), and mixtures thereof, and it is preferably selected from DME, DMF and DMSO.

As demonstrated in Example 6, the stereoselectivity of an ω-transaminase is also influenced by the presence of such organic cosolvents in the aqueous reaction medium. In particular, the use of organic cosolvents such as DME, DMF and DMSO was found to improve the stereoselectivity of (S)-selective ω-transaminases such as ArRmutI 1 -M 1 17F-A60V-G279V-oo- TA and ArR-ω-ΤΑ in the reductive amination of compounds of formula (I) . If a mutein/variant of ArRmutI 1 -ω-ΤΑ, such as ArRmutI 1 -M1 17F-A60V-G279V-OJ-TA, is used as (S)-selective ω-transaminase (and particularly if ArRmutI 1 -M1 17F-A60V-G279V-io-TA is used , R¹ in the compound of formula (I) is fe/f-butyl, and R² is C_{1 -4} alkyl such as isobutyl or n-propyl) , it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent which is preferably selected from DME, DMF and DMSO , and is more preferably DMF. More preferably, the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 1 0 vol-% to about 25 vol-% DME (more preferably about 20 vol-% DME) or in an aqueous medium comprising about 10 vol-% to about 25 vol-% DMF (more preferably about 20 vol-% DMF) or in an aqueous medium comprising about 10 vol-% to about 25 vol-% DMSO (more preferably about 20 vol-% DMSO) . Even more preferably, the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 1 0 vol-% to about 25 vol-% DMF (more preferably about 20 vol-% DMF) or in an aqueous medium comprising about 10 vol-% to about 25 vol-% DMSO (more preferably about 20 vol-% DMSO). Yet even more preferably, the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 10 vol-% to about 25 vol-% DMF, particularly about 20 vol-% DMF. The use of a mutein/variant of ArRmutI 1 -ω-ΤΑ, such as ArRmutl 1-M 117F-A60V-G279V-UJ-TA, in an aqueous medium comprising about 20 vol-% DMF is particularly advantageous with respect to the stereoselectivity of the reductive amination reaction. However, the use of a mutein/variant of ArRmutl 1 -ω-ΤΑ, such as ArRmutl 1 -M1 17F-A60V-G279V-OJ-TA, in an aqueous medium comprising about 1 0 vol-% to about 25 vol-% DMSO (e.g., about 20 vol-% DMSO) is also advantageous as it provides a favorable conversion rate and, at the same time, a favorable albeit lower stereoselectivity than that obtained with 20 vol-% DMF.

It is thus particularly preferred that the (S)-selective ω-transaminase is ArRmutl 1 -M1 17F- A6QV-G279V-u)-TA, R¹ is fert-butyl, R² is C alkyl (most preferably isobutyl), and that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 10 vol-% to about 25 vol-% DMF (e.g., about 20 vol-% DMF) or about 10 vol-% to about 25 vol-% DMSO (e.g. , about 20 vol-% DMSO). Even more preferably, the (S)-selective ω-transaminase is ArRmutl 1-M1 17F-A60V-G279V-DJ-TA, R¹ is terf-butyl, R² is C alkyl (most preferably isobutyl), and the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 10 vol-% to about 25 vol-% DMF, particularly about 20 vol-% DMF.

If ArR-ω-ΤΑ is used as (S)-selective ω-transaminase (and particularly if ArR-ω-ΤΑ is used, R¹ in the compound of formula (I) is fert-butyl, and R² is C_1-4 alkyl such as isobutyl or n-propyl), it is preferred that the combined racemization and stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent which is preferably selected from DME, DMF and DMSO, more preferably DME. More preferably, the combined racemization and stereoselective reductive amination step using ArR-ω-ΤΑ is conducted in an aqueous medium comprising about 10 vol-% of DME, DMF or DMSO, and even more preferably about 10 vol-% of DME.

The (S)-selective ω-transaminase to be used in the combined racemization and stereoselective reductive amination step in the process according to the first aspect of the invention, including any of the specific (S)-selective ω-transaminases described herein, can be prepared by recombinant expression in a microorganism, e.g., in a bacterium such as Escherichia coli, or can be prepared by peptide synthesis. It is preferred to use an (S)-selective ω-transaminase (including any of the specific (S)-selective ω-transaminases described herein) that has been expressed in a bacterium, and most preferably in E. coli, in the combined racemization and stereoselective reductive amination step. The (S)-selective ω-transaminase can be used, e.g., in purified form, in partly purified form, in the form of permeabilized microbial cells, in the form of extracts or lysates of microbial cells, or in the form of rehydrated lyophilized microbial cells (i.e., in the form of a rehydrated microbial cell lyophilisate).

Moreover, the (S)-selective ω-transaminase to be used in the combined racemization and stereoselective reductive amination step, including any of the specific (S)-selective ω-transaminases described herein, can be employed in immobilized form. For example, the (S)-selective ω-transaminase can be chemically, physically or by means of genetic engineering methods adsorbed at and/or covalently bound at/in a chemically organic or inorganic support material, wherein the catalytic activity of the enzyme is retained and the immobilized (S)-selective ω-transaminase can be used repeatedly and continuously. Accordingly, the (S)-selective ω-transaminase to be used in the combined racemization and stereoselective reductive amination step can be immobilized on a solid support material. Suitable solid support materials are known in the art and include, e.g. , polymer matrices such as calcium alginate, polyacrylamide, poiymethacrylate, polystyrene, polystyrene/polymethyacrylate, Sepabeads^®, ReliZyme™ Diaion^® and Eupergit^® as well as inorganic matrices such as Celite^®. Further specific solid support materials that can be used in the processes of the invention as well as immobilization techniques are described in: Sheldon, 2007; Homaei et al., 2013; Sumitra et al., 2013; Knezevic-Jugovic et al., 201 1 ; Bornscheuer, 2003; and Bickerstaff Jr GF, 1997. The use of the ω-transaminase in immobilized form is advantageous because the enzyme and the product can be easily separated and the immobilized enzyme may be recycled and reused, thus rending the process more economic. It is also possible to use cells (e.g. , permeabilized microbial cells, such as permeabilized E. coli cells) containing the (f?)-selective ω-transaminase, which cells are immobilized on a solid support material (e.g. , any of the specific solid support materials described above).

The combined racemization and stereoselective reductive amination step can be conducted in an aqueous medium at any suitable pH, for example, at a pH in the range of about 6 to about 1 1. The aqueous medium may comprise an aqueous buffer system such as, e.g., a phosphate buffer (e.g., an admixture of K₂HP0₄ and KH₂P0₄) , a Tris buffer (e.g., an admixture of tris(hydroxymethyl)aminomethane and a strong inorganic acid, such as hydrochloric acid), a PIPES buffer (e.g . , an admixture of piperazine-N,N'-bis(2-ethanesulfonic acid) and an alkali hydroxide, such as NaOH or KOH), or a HEPES buffer (e.g., an admixture of 4-(2- hydroxyethyl)-1 -piperazineethanesulfonic acid and an alkali hydroxide, such as NaOH or KOH). It is also possible that the aqueous medium does not comprise any buffer, i.e. , does not comprise any buffering agent. Moreover, the combined racemization and stereoselective reductive amination step can be carried out at any suitable temperature, e.g., at a temperature of about 20°C to about 50°C, preferably about 25°C to about 45°C. If ArR-ω-ΤΑ is used as (S)-selective ω-transaminase, it is preferred that the combined racemization and stereoselective reductive amination step is conducted at a temperature of about 20°C to about 30^oC, more preferably at about 25°C.

The process may further comprise a step of converting the compound of formula (ll-S) into a compound of the following formula (lll-S)

(Ill-S) wherein R² is the same as in formula (Il-S). This cyclization reaction of the compound of formula (Il-S) into the compound of formula (Ill-S) can occur under basic conditions in the reaction medium, e.g., by addition of a strong base like

NaOH,

Alternatively, the process may also comprise a step of converting the compound of formula (Il-S), wherein R¹ is different from hydrogen, into a compound of the following formula (V-S)

(V-S) wherein R² is the same as in formula (Il-S).

This reaction can be conducted using any method of carboxylic ester cleavage known in the art. In particular, this reaction can be carried out in accordance with or in analogy to the approach described in Example 8 or 9. Thus, the ester group comprised in the compound of formula (Il-S), wherein R is different from hydrogen, can be hydrolyzed under acidic conditions using, e.g., an aqueous solution of hydrochloric acid, to provide the compound of formula

(V-S) The above-described process leading to a compound of formula (V-S) is particularly useful for the preparation of pregabalin which corresponds to the compound of formula (V-S) wherein R² isobutyl:

Pregabalin

In a second aspect, the present invention provides a process for the production of brivaracetam, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (la)

wherein R¹ is a carboxy protecting group or hydrogen, is contacted with an (R)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (la) and obtain a compound of the following formula (Ma)

wherein R is the same as in formula (la), wherein the process optionally further comprises a step of converting the compound of formula (Ha) obtained in the stereoselective reductive amination step into a compound of the following formula (Ilia)

(Ilia) and wherein the compound of formula (1 la) or (Ilia) is further converted into brivaracetam having the following formula (IVa):

The process according to the second aspect of the invention is preferably conducted as described above with respect to the first aspect of the invention. Accordingly, the features and embodiments described in connection with the process of the first aspect of the invention, including all preferred features and embodiments, also apply to the process of the second aspect of the invention, provided that the ω-transaminase is an (R)-selective ω-transaminase and that R² is n-propyl.

In a third aspect, the present invention provides a process for the production of pregabalin, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (lb)

(lb) wherein R¹ is a carboxy protecting group or hydrogen, is contacted with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (lb) and obtain a compound of the following formula (lib)

(lib) wherein R¹ is the same as in formula (lb), wherein, if R¹ is different from hydrogen, the process further comprises a step of converting the compound of formula (lib) into pregabalin having the following formula (Vb)

The process according to the third aspect of the invention is preferably conducted as described above with respect to the first aspect of the invention. Accordingly, the features and embodiments described in connection with the process of the first aspect of the invention, including all preferred features and embodiments, also apply to the process of the third aspect of the invention, provided that the ω-transaminase is an (S)-selective ω-transaminase and that R² is isobutyl. The compounds of formula (I) to be used in the processes according to the various aspects of the present invention, including the compounds of formula (la) or (lb), can be prepared by methods known in the field of synthetic chemistry. For example, the compounds of formula (I) can be prepared in accordance with or in analogy to the synthetic routes described in the appended examples, or in WO 2005/027856, or in Hodgson et al., 2009.

In particular, the compounds of formula (I) can be prepared in accordance with the following scheme, in which the groups R¹ and R² have the same meanings as the corresponding groups comprised in the compound of formula (I):

As shown in the above scheme, the compounds of formula (I) can be prepared from compounds of formula (VII) and diisobutylamine and refluxed in toluene. Reaction water can be removed by the addition of molecular sieves. R¹ bromo acetate is added. The reaction is refluxed until the solution turns brown. Acetate buffer is added, the suspension is stirred for 1 hour at room temperature (RT) and the organic phase is washed with Na₂C0₃ saturated solution. The organic phase is dried over Na₂S0₄ and the solvent is evaporated. The resulting compound of formula (I) can be purified by column chromatography. In the case of the compound of formula (lb), the corresponding compound of formula (VII) is not commercially available. The aldehyde 8 (i.e., the corresponding compound of formula (VII) in which R² is isobutyl) can be synthesized from the corresponding alcohol by chemical or enzymatical oxidation, e.g., using pyridium chlorochromate (PCC; e.g., as described in Schmidt et al., 201 1 ) or using an alcohol dehydrogenase (ADH-ht).

PCC

Acetaldehyde ,

Phosphate

buffer pH 8

The following definitions apply throughout the present specification, unless specifically indicated otherwise.

As used herein, the term "alky!" refers to a monovalent saturated aliphatic (i.e., non-aromatic) acyclic hydrocarbon group (i.e., a group consisting of carbon atoms and hydrogen atoms) which may be linear or branched. Preferred alkyl groups are C _-8 alkyl groups. The term "alkenyl" refers to a monovalent unsaturated aliphatic acyclic hydrocarbon group which may be linear or branched and comprises at least one carbon-to-carbon double bond while it does not comprise any carbon-to-carbon triple bond. Preferred alkenyl groups are C alkenyl groups. The term "aryl" refers to a monovalent aromatic hydrocarbon group, including monocyclic as well as bridged ring and/or fused ring systems, containing at least one aromatic ring. The "aryl" may, e.g., have 6 to 14 ring atoms, particularly 6 to 10 ring atoms. "Aryl" may, e.g., refer to phenyl, naphthyl. anthracenyl, indanyi, or tetraliny!. The term "heteroaryl" refers to a monovalent aromatic ring group, including monocyclic as well as bridged ring and/or fused ring systems, containing at least one aromatic ring which comprises one or more (such as, e.g., one, two, or three) ring heteroatoms independently selected from O, S, and N. The "heteroaryl" may, e.g., have 5 to 14 ring atoms, particularly 5 to 10 ring atoms, and preferably 5 or 6 ring atoms. "Heteroaryl" may, e.g., refer to thienyl, furanyl, pyrrolyl, imidazolyl, pyrazolyl, pyridinyl, pyrazinyl, pyrimidinyl, pyridazinyl, oxazolyl, isoxazolyl, furazanyl, benzo[b]thienyl, naphtho[2,3-b]thienyl, thianthrenyl, isobenzofuranyl, chromenyl, xanthenyl, phenoxathiinyl, indolizinyl, isoindolyl, indolyl, indazolyl, purinyl, isoquinolyl, quinolyl, phthalazinyl, naphthyridinyl, quinoxalinyl, cinnolinyl, pteridinyl, carbazolyl, β-carbolinyl, phenanthridinyl, acridinyl, perimidinyl, phenanthrolinyl, phenazinyl, isothiazolyl, phenothiazinyl, phenoxazinyl, pyrazoio[1 ,5-a]pyrimidinyl, 1 ,2-benzoisoxazolyl, or benzimidazolyl.

The term "halogen" or "halo" refers to fluoro, chloro, bromo, or iodo. The term "(R)-isomer", when used in connection with a compound of formula (II), refers to the stereoisomer having the following specific configuration at the carbon atom carrying the aminomethyl group:

Likewise, the term "(S)-isomer", when used in connection with a compound of formula (II), refers to the stereoisomer having the following specific configuration at the carbon atom carrying the aminomethyl group:

The term "enantiomeric excess" (or "percent enantiomeric excess" or "ee") refers to the difference between the mole fraction of one specific enantiomer (i.e., the specified enantiomer) and the mole fraction of the other enantiomer in relation to the sum of the mole fractions of both enantiomers, expressed as a percent value, and thus describes the extent of the excess of one specific enantiomer in relation to the other enantiomer. If, for example, a specific enantiomer is provided in the absence of the other enantiomer, the enantiomeric excess will be 100%, while a racemate comprising equal molar amounts of the two enantiomers will have an enantiomeric excess of 0%. Accordingly, the "enantiomeric excess" is defined by the following formula:

ee [%] = (^rno'^e fr^acti°ⁿ °f ^tne specified enantiorner) - (mole fraction of the other enantiomer ) ^ ^ (mole fraction of the specified enantiorner) + (mole fraction of the other enantiorner) The enantiomeric excess can be determined, e.g., by chiral HPLC.

As used herein, the term "ω-transaminase from Arthrobacter sp. NK168" or "ArR-ω-ΤΑ" refers to an ω-transaminase that is found in Arthrobacter sp. KNK 68 and, in particular, to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEQ ID NO: 1 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 2;

(c) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid hybridizing under stringent conditions to the complementary strand of a nucleic acid as defined in (b); and

(d) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c).

The term "ω-transaminase from Arthrobacter sp. KNK168" or "ArR-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 1 .

As used herein, the term "ArRmutl 1 -ω-ΤΑ" or "ArRmutH " refers to an ω-transaminase as described in Savile et al. , 2010 and its supporting material as "Arthrobacter Round 1 1 Transaminase" or "Rd 1 1 TA" and, in particular, to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEQ ID NO: 3;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 4;

The term "ArRmutl 1 -ω-ΤΑ" or "ArRmutl 1 " preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 3. As used herein, the term "ω-transaminase from Bacillus megaterium" or "ΒΜ-ω-ΤΑ" refers to an ω-transaminase that is found in Bacillus megaterium and, in particular, to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEQ ID NO: 5; (b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in

SEQ ID NO: 6;

The term "ω-transaminase from Bacillus megaterium" or "ΒΜ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 5.

As used herein, the term "ω-transaminase from Alcaligenes denitrificans" or "AD-ω-ΤΑ" refers to an ω-transaminase that is found in Alcaligenes denitrificans and, in particular, to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEQ I D NO: 7; (b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in

SEQ ID NO: 8;

The term "ω-transaminase from Alcaligenes denitrificans" or "AD-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 7.

As used herein, the term "ω-transaminase from Chrome-bacterium violaceum DSM 30191 " or "CV-ω-ΤΑ" refers to an ω-transaminase that is found in Chrome-bacterium violaceum DSM 30191 (also referred to as Chromobacterium violaceum ATCC 12472) and, in particular, to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEQ ID NO: 9;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 10; (c) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid hybridizing under stringent conditions to the complementary strand of a nucleic acid as defined in (b); and

The term "ω-transaminase from Chromobacterium violaceum DSM 30191 " or "CV-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 9. As used herein, the term "ω-transaminase from Paracoccus denitrificans" or "PD-ω-ΤΑ" refers to an ω-transaminase that is found in Paracoccus denitrificans and, in particular, to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEQ iD NO:

1 1 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in

SEQ ID NO: 12;

The term "ω-transaminase from Paracoccus denitrificans" or "PD-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 1 1 .

As used herein, the term "ω-transaminase from Pseudomonas putida KT2440 gene PP5182" or "ΡΡ1 -ω-ΤΑ" refers to an ω-transaminase that is encoded by gene PP5182 of Pseudomonas putida KT2440 and, in particular, to a protein selected from:

13;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ I D NO: 14;

(c) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid hybridizing under stringent conditions to the complementary strand of a nucleic acid as defined in (b); and (d) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c).

The term "ω-transaminase from Pseudomonas putida KT2440 gene PP5182" or "ΡΡ1 -ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 13.

As used herein, the term "ω-transaminase from Pseudomonas putida KT2440 gene PP2180" or "ΡΡ2-ω-ΤΑ" refers to an ω-transaminase that is encoded by gene PP2180 of Pseudomonas putida KT2440 and, in particular, to a protein selected from:

15;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 16;

The term "ω-transaminase from Pseudomonas putida KT2440 gene PP2180" or "ΡΡ2-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 15.

As used herein, the term "ω-transaminase from Vibrio fluviaiis" or "VF-ω-ΤΑ" refers to an ω-transaminase that is found in Vibrio fluviaiis and, in particular, to a protein selected from:

17;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 18;

(d) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c),

The term "ω-transaminase from Vibrio fluviaiis" or "VF-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 17. As used herein, the term "ω-transaminase from Arthrobacter citreus" or "ArS-ω-ΤΑ" refers to an ω-transaminase that is found in Arthrobacter citreus and, in particular, to a protein selected from:

19;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ I D NO: 20;

The term "ω-transaminase from Arthrobacter citreus" or "ArS-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 9.

As used herein, the term "ω-transaminase from Ralstonia eutropha" or "RE-ω-ΤΑ" refers to an ω-transaminase that is found in Ralstonia eutropha and, in particular, to a protein selected from :

21 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 22;

The term "ω-transaminase from Ralstonia eutropha" or "RE-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 21 .

As used herein, the term "ω-transaminase from Variovorax paradoxus" refers to an ω-transaminase that is found in Variovorax paradoxus and, in particular, to a protein selected from:

23; (b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 24;

The term "ω-transaminase from Variovorax paradoxus" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 23.

As used herein, the term "ω-transaminase from Aspergillus terreus" or "ΑΤ-ω-ΤΑ" refers to an ω-transaminase that is found in Aspergillus terreus and, in particular, to a protein selected from:

25;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 26;

The term "ω-transaminase from Aspergillus terreus" or "ΑΤ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 25.

As used herein, the term "ω-transaminase from Hyphomonas neptunium" or "ΗΝ-ω-ΤΑ" refers to an ω-transaminase that is found in Hyphomonas neptunium and, in particular, to a protein selected from:

27;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 28;

The term "ω-transaminase from Hyphomonas neptunium" or "ΗΝ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 27.

As used herein, the term "ω-transaminase from Pseudomonas fluorescens" or "PF-ω-ΤΑ" refers to an ω-transaminase that is found in Pseudomonas fluorescens and, in particular, to a protein selected from:

29;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 30;

The term "ω-transaminase from Pseudomonas fluorescens" or "PF-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 29.

As used herein, the term "ω-transaminase from Silicibacter pomeroyr or "SP-ω-ΤΑ" refers to an ω-transaminase that is found in Silicibacter pomeroyi (see, e.g., Steffen-Munsberg et al., 2013(a); and Steffen-Munsberg et al., 2013(b)) and, in particular, to a protein selected from:

31 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 32;

(d) a protein comprising or consisting of an amino acid, sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c).

The term "ω-transaminase from Silicibacter pomeroyi' or "SP-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 31 . As used herein, the term "ω-transaminase from Ochrobactrum anthropr or "ΟΑ-ω-ΤΑ" refers to an ω-transaminase that is found in Ochrobactrum anthropi (see, e.g., Park et al., 2012) and, in particular, to a protein selected from:

33;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 34;

The term "ω-transaminase from Ochrobactrum anthropi"' or "ΟΑ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 33.

As used herein, the term "ω-transaminase from Neosartorya fischerf' or "NF-ω-ΤΑ" refers to an ω-transaminase that is found in Neosartorya fische (see, e.g., Hohne et al. , 2010) and, in particular, to a protein selected from:

35;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 36;

The term "ω-transaminase from Neosartorya fischerr or "NF-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 35.

As used herein, the term "ω-transaminase from Gibberella zeae" or "GZ-ω-ΤΑ" refers to an ω-transaminase that is found in Gibberella zeae (see, e.g., Hohne et al. , 2010) and, in particular, to a protein selected from: (a) a protein comprising or consisting of the amino acid sequence shown in SEQ ID NO:

37;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 38;

The term "ω-transaminase from Gibberella zeae" or "GZ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 37.

As used herein, the term "ArRmutI 1-G279A-u)-TA" refers to a protein selected from:

39;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 40;

The term "ArRmutI 1 -G279A-oo-T A" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 39.

As used herein, the term "ArRmutI 1-M1 17F-u>-TA" refers to a protein selected from:

41 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 42;

The term "ArRmutI 1-M1 17F-u>-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 41.

As used herein, the term "ArRmutI 1-M1 17F-G279A-oo-TA" refers to a protein selected from:

43;

SEQ ID NO: 44;

The term "ArRmutI 1-M1 17F-G279A-oo-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 43.

As used herein, the term "ArRmutI 1-M1 17F-A60V-u)-TA" refers to a protein selected from:

45;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 46;

The term "ArRmutI 1-M117F-A60V-u)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 45.

As used herein, the term "ArRmut1 1 -A60V-M 1 17F-G279A-to-TA" or "ArRmutI 1-M1 17F-A60V- G279A-(jj-TA" refers to a protein selected from: (a) a protein comprising or consisting of the amino acid sequence shown in SEQ ID NO: 47;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 48;

The term "ArRmutl 1 -A60V-M117F-G279A-to-TA" or "ArRmutl 1 -M1 17F-A60V-G279A-U)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 47.

It is to be understood that the order in which the mutated amino acid residues are indicated in the name of the ω-transaminase referred to in the above definition (and also in the names of any further ω-transaminases referred to in the present specification) is interchangeable. Thus, for example, the terms 'ArRmutl 1 -A60V-M1 17F-G279A-U)-TA", "ArRmutl 1 -A60V- G279A-M1 17F-<i TA", "ArRmutl 1 -M117F-A60V-G279A-U)-TA" "ArRmutl 1-M1 17F-G279A- Α60ν-ω-ΤΑ", "ArRmutl 1-G279A-A60V-M117F-W-T A" and "ArRmutl 1-G279A- 117F-A60V-to- TA" are used herein interchangeably.

As used herein, the term "ArRmutl 1 -M1 17F-A60V-G279V-u)-TA" or "ArRmutl 1-A60V-M1 17F- G279V-u)-TA" refers to a protein selected from:

49;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 50;

The term "ArRmutl 1-M1 17F-A60V-G279V-u>-TA" or "ArRmutl 1 -A60V-M1 17F-G279V-U)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 49. As used herein, the term "ArRmut1 1 -M1 17F-A60V-G279L- -TA" or "ArRmutl 1-A60V-M1 17F- G279L-u>-TA" refers to a protein selected from:

51 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 52;

The term "ArRmutl 1 -M1 17F-A60V-G279L-co-TA" or "ArRmutl 1 -A60V-M 1 17F-G279L-u)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 51.

As used herein, the term "ArRmutl 1-M1 17F-A60V-G279I-U)-TA" or "ArRmutl 1-A60V- 1 17F- G279l-u)-TA" refers to a protein selected from:

53;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 54;

The term "ArRmutl 1-M117F-A60V-G279l-u.-TA" or "ArRmutl 1-A60V-M 17F-G279l-co-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 53.

As used herein, the term "ArRmutl 1 -M1 17F-A60V-G279F- -TA" or "ArRmutl 1-A60V-M1 17F- G279F-u)-TA" refers to a protein selected from:

55;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 56; (c) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid hybridizing under stringent conditions to the complementary strand of a nucleic acid as defined in (b); and

The term "ArRmutl 1 -M 1 17F-A60V-G279F- -TA" or "ArRmutl 1 -A60V-M1 17F-G279F-u>-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 55. As used herein, the term "ArRmutl 1 -A60V-M1 17F-G279V-T64V- -TA" refers to a protein selected from:

57;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 58;

The term 'ArRmutl 1 -A60V-M1 17F-G279V-T64V-u>-T A" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 57. As used herein, the term "ArRmutl 1 -A60V-M1 7F-G279V-T1 19Ε-ω-ΤΑ" refers to a protein selected from:

59;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 60;

(d) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c). The term "ArRmutl 1-A60V-M1 17F-G279V-T1 19Ε-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 59.

As used herein, the term "ArRmut11-A60V-M1 17F-G279V-l152V-w-TA" refers to a protein selected from:

61 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 62;

The term "ArRmutl 1-A60V-M117F-G279V-1152V-oo-TA^M preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 61.

As used herein, the term "ArRmutl 1-A60V- 1 17F-G279V-S277T-oo-TA" refers to a protein selected from:

63;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 64;

The term "ArRmutl 1 -A60V-M1 17F-G279V-S277T-oo-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 63.

As used herein, the term "ArRmutl 1-A60V-M117F-G279V-S277N-ui-TA" refers to a protein selected from:

65; (b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 66;

The term "ArRmutl 1-A60V-M117F-G279V-S277N-u)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 65.

As used herein, the term "ArR-S218P-u>-TA" refers to a protein selected from:

(a) a protein comprising or consisting of the amino acid sequence shown in SEO iD NO:

67:

SEQ ID NO: 68;

The term "ArR-S218P-co-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 67.

As used herein, the term "ArR-G131 F-S218P-u)-TA" refers to a protein selected from:

69;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 70;

(d) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c). The term "ArR-G 131 F-S218Ρ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 69.

As used herein, the term "ArR-G 131 F-ω-ΤΑ" refers to a protein selected from:

71 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 72;

The term "ArR-G 131 F-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 71.

As used herein, the term "GZ-L56V-u>-TA" refers to a protein selected from:

73;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 74;

The term "GZ~L56V-u>-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 73.

As used herein, the term "GZ-V60T-u>-TA" refers to a protein selected from:

75;

SEQ ID NO: 76; (c) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid hybridizing under stringent conditions to the complementary strand of a nucleic acid as defined in (b); and

The term "GZ-V60T-u)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 75. As used herein, the term "GZ-E1 15Τ-ω-ΤΑ" refers to a protein selected from:

77;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 78;

The term "GZ-E1 15Τ-ω-ΤΑ" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 77.

As used herein, the term "GZ-V148l-u>-TA" refers to a protein selected from:

79;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 80;

The term "GZ-V148l-u)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 79. As used herein, the term "GZ-T273S-iu-TA" refers to a protein selected from:

81 ;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 82;

The term "GZ-T273S-u>-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 81.

As used herein, the term "GZ-T273N-u>-TA" refers to a protein selected from:

83;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 84;

The term "GZ-T273N-oo-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 83.

As used herein, the term "GZ-A275V-u)-TA" refers to a protein selected from:

85;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 86;

The term "GZ-A275V-cu-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 85.

As used herein, the term "GZ-Y58L-co-TA" refers to a protein selected from:

87;

SEQ ID NO: 88;

The term "GZ-Y58L-co-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 87.

As used herein, the term "GZ-E115D-u>-TA" refers to a protein selected from:

89;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 90;

The term "GZ-E1 15D-u)-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 89. As used herein, the term "GZ-V148A-u)-TA" refers to a protein selected from:

91 ; (b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 92;

The term "GZ-V148A-GJ-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 91.

As used herein, the term "GZ-L56V-A275V-u)-TA" refers to a protein selected from:

93;

SEQ ID NO: 94;

The term "GZ-L56V-A275V-u>-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 93.

As used herein, the term "GZ-L56V-T273S-u>-TA" refers to a protein selected from:

95;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 96;

(d) a protein comprising or consisting of an amino acid sequence encoded by a nucleic acid being degenerate as a result of the genetic code to the nucleotide sequence of a nucleic acid as defined in (b) or (c). The term "GZ-L56V-T273S-oo-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 95.

As used herein, the term "GZ-T273S-A275V-u>-TA" refers to a protein selected from:

97;

(b) a protein encoded by a nucleic acid consisting of the nucleic acid sequence depicted in SEQ ID NO: 98;

The term "GZ-T273S-A275V-u>-TA" preferably refers to a protein consisting of the amino acid sequence shown in SEQ ID NO: 97.

The amino acid sequences referenced above in connection with specific ω-transaminases are included in a separate sequence listing and are also shown in the following. In the case of conflict between any of the sequences specified in the sequence listing and the corresponding sequence shown further below, the present invention relates to both sequences. ω-ΤΑ from Arthrobacter sp. KNK168 - Amino acid sequence fSEQ ID NO: 1 )

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGYLHSDVTYTV FHVWNGNAFRLDDHIERLFSNAESMRII PPLTQDEVKEIALELVAKTELREAFVSVSITRGYSST PGERDITKHRPQVYMYAVPYQWIVPFDRIRDGVHAMVAQSVRRTPRSSIDPQVKNFQWGDLI RAVQETHDRGFEAPLLLDGDGLLAEGSGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLG HEAILADITLAELLDADEVLGCTTAGGVWPFVSVDGNPISDGVPGPITQSI IRRYWELNVESSSL LTPVQY

ArRmutl 1-ω-ΤΑ - Amino acid sequence (SEQ ID NO: 3)

MAFSADTPEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSD ATYTTFHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAMVTVTITR GYSSTPFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQ WGDLIRAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIA ESLGHEAILADITPAELYDA.DEVLGCSTGGGVWPFVSVDGNSISDGVPGPVTQSIIRRYWELN VEPSSLLTPVQYALEHHHHHH ω-TA from Bacillus meaaterium - Amino acid sequence (SEQ ID NO: 5)

MSLTVQKINWEQV EWDRKYLMRTFSTQNEYQPVPIESTEGDYLIMPDGTRLLDFFNQLYCV NLGQ KNQKVNAAI KEALD R YG FVWDTYAT D Y KAKAAKI 11 ED I LG DE DWPG KVRFVSTGS EAV ETALNIARLYTNRPLWTREHDYHGWTGGAATVTRLRSYRSGLVGENSESFSAQIPGSSYNS AVLMAPSPNMFQDSDGNLLKDENGELLSVKYTRRMIENYGPEQVAAVITEVSQGAGSAMPPY EYIPQIRKMTKELGVLWINDEVLTGFGRTGKWFGYQHYGVQPDIITMGKGLSSSSLPAGAVLV SKEIAAFMDKHRWESVSTYAGHPVAMAAVCANLEVMMEENFVEQAKDSGEYIRSKLELLQEK HKSIGNFDGYGLLWIVDIVNAKTKTPYVKLDRNFTHGMNPNQIPTQilMKKALEKGVLIGGVMP NT RIGASLNVSRGDIDKA DALDYALDYLESGEWQ ω-ΤΑ from Alcaligenes denitrificans - Amino acid sequence (SEQ ID NO: 7)

MSAAKLPDLSHLWMPFTANRQFKANPRLLASAKGMYYTSFDGRQILDGTAGLWCVNAGHCR EEIVSAIASQAGVMDYAPGFQLGHPLAFEAATAVAGLMPQGLDRVFFTNSGSESVDTALKIAL AYHRARGEAQRTRLIGRERGYHGVGFGGISVGGISPNRKTFSGALLPAVDHLPHTHSLEHNA FTRGQPEWGAHLADELERIIALHDASTIAAVIVEPMAGSTGVLVPPKGYLEKLREITARHGILLIF DEVITAYGRLGEATAAAYFGVTPDLITMAKGVSNAAVPAGAVAVRREVHDAIVNGPQGGIEFF HGYTYSAHPLAAAAVLATLDIYRREDLFARARKLSAAFEEAAHSLKGAPHVIDVRNIGLVAGIEL SPREGAPGARAAEAFQKCFDTGLMVRYTGDILAVSPPLiVDENQIGQ!FEGIGKVLKEVA ω-TA from Chromobacterium violaceum - Amino acid sequence (SEQ ID NO: 9)

MQKQRTTSQWRELDAAHHLHPFTDTASLNQAGARVMTRGEGVYLWDSEGNKIIDGMAGLW CVNVGYGRKDFAEAARRQMEELPFYNTFFKTTHPAWELSSLLAEVTPAGFDRVFYTNSGSE SVDTMIRMVRRYWDVQGKPEKKTLIGRWNGYHGSTIGGASLGGMKYMHEQGDLPIPGMAHI EQPWWYKHGKDMTPDEFGVVAARWLEEKILEIGADKVAAFVGEPIQGAGGVIVPPATYWPEI ERICRKYDVLLVADEVICGFGRTGEWFGHQHFGFQPDLFTAAKGLSSGYLPIGAVFVGKRVAE GLIAGGDFNHGFTYSGHPVCAAVAHANVAALRDEGIVQRVKDDIGPYMQKRWRETFSRFEHV DDVRGVGMVQAFTLVKNKAKRELFPDFGEIGTLCRDIFFRNNLI RACGDHiVSAPPLV TRA EVDEMLAVAERCLEEFEQTLKARGLA ω-ΤΑ from Paracoccus denitrificans - Amino acid sequence (SEQ ID NO: 1 1 )

NQPQSWEARAETYSLYGFTDMPSVHQRGTVWTHGEGPYIVDVHGRRYLDANSGLWNMVA GFDHKGLIEAAKAQYDRFPGYHAFFGRMSDQTVMLSEKLVEVSPFDNGRVFYTNSGSEAND TMVKMLWFLHAAEGKPQKRKILTRWNAYHGVTAVSASMTG PYNSVFGLPLPGFIHLTCPHY WRYGEEGETEAQFVARLARELEDTITREGADTIAGFFAEPVMGAGGVIPPAKGYFQAILPILRK YDIPMISDEVICGFGRTGNTWGCLTYDFMPDAIISSKNLTAGFFPMGAVILGPDLAKRVEAAVE AIEEFPHGFTASGHPVGCAIALKAIDWMNEGLAENVRRLAPRFEAGLKRIADRPNIGEYRGIG FMWALEAV DKPTKTPFDANLSVSERIANTCTDLGLICRPLGQSIVLCPPFILTEAQMDEMFEK LEKALDKVFAEVA ω-ΤΑ from Pseudomonas putida KT2440 gene PP5182 (ΡΡ1-ω-ΤΑ) - Amino acid sequence (SEQ ID NO: 13)

MSVNNPQTREWQTLSGEHHLAPFSDYKQLKEKGPRIITKAGGVHLWDSEGHKILDGMAGLW CVAVGYGREELVQAAEKQMRELPYYNLFFQTAHPPALELAKAITDVAPEGMTHVFFTGSGSE GNDTVLRMVRHY ALKGKPHKQTIIGRINGYHGSTFAGACLGGMSGMHEQGGLPIPGIVHIP QPYWFGEGGDMTPDAFGIWAAEQLEKKILEVGEDNVAAFIAEPIQGAGGVIIPPETY PKVKEI LAKYDI LFVADEVICGFGRTGEWFGSDYYDLKPDLMTiAKGLTSGYIP GGVIVRDKVAKVISE GGDFNHGFTYSGHPVAAAVGLENLRILRDEQIVEKARTEAAPYLQKRLRELQDHPLVGEVRG LGMLGAIELVKDKATRSRYEGKGVGMICRTFCFENGLIMRAVGDT IIAPPLVISHAEI DELVEK ARKCLDLTLEA!R ω-ΤΑ from Pseudomonas putida KT2440 gene PP2180 (ΡΡ2-ω-ΤΑ) - Amino acid sequence (SEQ ID NO: 15)

MSEQNSQTLAWQSMSRDHHLAPFSDVKQLAEKGPRIITSAKGVYLWDSEGNKILDGMAGLW CVAVGYGRDELAEVASQQMKQLPYYNLFFQTAHPPALELAKAIADVAPQGMNHVFFTGSGSE GNDTVLRMVRHYWALKGKKNKNVIIGRINGYHGSTVAGAALGGMSGMHQQGGVIPDIVH I PQ PYWFGEGGDMTEADFGWVAAEQLEKKILEVGVDNVAAFIAEPIQGAGGVI IPPQTYWPKVKEI LARYDILFVADEVICGFGRTGEWFGTDYYDLKPDLMTIAKGLTSGYIPMGGVIVRDEVAKVISE GGDFNHGFTYSGHPVAAAVGLENLRILRDEQIIQQVHDKTAPYLQQRLRELADHPLVGEVRGL G LGAiELVKDKATRARYEGKGVGMICRQHCFDNGLIMRAVGDTMI IAPPLVISI EEIDELVEKA RKCLDLTYEAVR ω-ΤΑ from Vibrio fluvia!is - Amino acid seguence (SEQ ID NO: 17)

MASRGSHHHHHHGANKPQSWEARAETYSLYGFTDMPSLHQRGTVWTHGEGPYIVDVNGR RYLDANSGLWNMVAGFDHKGLiDAAKAQYERFPGYHAFFGRMSDQTVMLSEKLVEVSPFDS GRVFYTNSGSEANDTMVKMLWFLHAAEGKPQKRKILTRWNAYHGVTAVSASMTGKPYNSVF GLPLPGFVHLTCPHYWRYGEEGETEEQFVARLARELEETIQREGADTIAGFFAEPV GAGGVI PPAKGYFQAI LPILRKYDIPVISDEVICGFGRTGNT GCVTYDFTPDAIISSKNLTAGFFPMGAVI LGPELSKRLETAIEAIEEFPHGFTASGHPVGCAIALKAIDWMNEGLAENVRRLAPRFEERLKHI AERPNIGEYRGIGFMWALEAVKDKASKTPFDGNLSVSERIANTCTDLGLICRPLGQSWLCPP FILTEAQMDEMFD LEKALD VFAEVA ω-ΤΑ from Arthrobacter citreus - Amino acid sequence (SEQ ID NO: 19)

MGLTVQKINWEQVKE DRKYLMRTFSTQNEYQPVPIESTEGDYLITPGGTRLLDFFNQLCCV NLGQKNQKVNAAIKEALDRYGFVWDTYATDYKAKAAKIIIEDILGDEDWPGKVRFVSTGSEAV ETALNIARLYTNRPLWTREHDYHGWTGGAATVTRLRSFRSGLVGENSESFSAQIPGSSCSS AVLMAPSSNTFQDSNGNYLKDENGELLSVKYTRRMIENYGPEQVAAVITEVSQGVGSTMPPY EYVPQIRKMTKELGVLWISDEVLTGFGRTGKWFGYQHYGVQPDIITMGKGLSSSSLPAGAW VSKEIAAFMDKHRWESVSTYAGHPVAMAAVCANLEVMMEENLVEQAKNSGEYIRSKLELLQE KHKSIGNFDGYGLLWIVDIVNAKTKTPYVKLDRNFRHGMNPNQIPTQI IMEKALEKGVLIGGAM PNTMRIGASLNVSRGDIDKAMDALDYALDYLESGEWQQS ω-ΤΑ from Ralstonia eutropha - Amino acid sequence (SEQ ID NO: 21 )

MDAAKTVIPDLDALWMPFTANRQYKAAPRLLASASGMYYTTHDGRQILDGCAGLWCVAAGH CRKEIAEAVARQAATLDYAPPFQMGHPLSFEAATKVAAIMPQGLDRIFFTNSGSESVDTALKIA LAYHRARGEGQRTRFIGRERGYHGVGFGG AVGGIGPNRKAFSANLMPGTDHLPATLNIAEA AFSKGQPTWGAHLADELERIVALHDPSTIAAVIVEPLAGSAGVLVPPVGYLDKLREITTKHGILLI FDEVITAFGRLGTATAAERFKVTPDLITMAKAINNAAVPMGAVAVRREVHDTWNSAAPGAIEL AHGYTYSGHPLAAAAAIATLDLYQRENLFGRAAELSPVFEAAVHSVRSAPHVKDIRNLGMVAG

lELEPRPGQPGARAYEAFLKCLERGVLVRYTGDILAFSPPLIISEAQIAELFDTVKQALQEVQAL EHHHHHH ω-ΤΑ from Variovorax paradoxus - Amino acid sequence (SEQ ID NO: 23)

MSDSAIDQSLAQAFQRFTEANPASRRQFEAQVRYMPGANSRSVLFYAPFPLTIAKGEGAALW DADGHRYADFIAEYTAGVYGHSAPEIRDAWEAMQGGINLTGHNLLEGRLAKTICERFPQIEQL RFTNSGTEANLMALTAALHFTGRRKIWFSGGYHGGVLGFGAKPLPTTVPFDFLVLPYNDVQA ASEQIEKHGSEIAAILVEPMQGASGCIPGRLDFLQALREGATKVGALLVFDEVMTSRLAPNGLA NKLGIRSDLTTLGKYIGGGMSFGAFGGRSDVMAQFDPRTGSLSHSGTFNNNVMT AAGYAG LTKLFTPEAAGRLAERGEAMRARLNALCAKEDVAMQFTGVGSLMNAHFVRGEVRRVDDLAA VDGRLRQLLFFHLLGEGIYASPRGFIVLSLPLTDADIDRFAAAIGRFIGEYRALLPSAK ω-ΤΑ from Aspergillus terreus - Amino acid sequence (SEQ ID NO: 25)

MASMDKVFAGYAARQAILESTETTNPFAKGIAWVEGELVPLAEARIPLLDQGFMHSDLTYDVP SVWDGRFFRLDDHITRLEASCTKLRLRLPLPRDQVKQILVEMVAKSGIRDAFVELIVTRGL GV RGTRPEDIVNNLYMFVQPYWWMEPDMQRVGGSAWARTVRRVPPGAIDPTVKNLQWGDL VRGMFEAADRGATYPFLTDGDAHLTEGSGFNIVLVKDGVLYTPDRGVLQGVTRKSVINAAEA FGIEVRVEFVPVELAYRCDEIFMCTTAGGIMPITTLDGMPVNGGQIGPITKKIWDGYWAMHYD AAYSFEIDYNERNLEHHHHHH ω-ΤΑ from Hyphomonas neptunium - Amino acid sequence (SEQ ID NO: 27) MLTFQKVLTGFQTRADARAERTDAFADGIAWIENEFVPIGKARIPILDQGFLHSDLTYDVPAV NGRIFRLDDHLDRLEVSCAKMRLPLPIARPELRRLVMELVSRSGLRDAYVEIIVTRGLKFLRGA QAEDIIPNLYLMAVPYVWILPLEYQNHGAPAWTRTVRRTPPGALDPTIKNLQWGDLVRGLME AGDRDSFFPILPDGDGNATEGAGYN!VLVRNGELHTPRRGVLEGITRRTVLEIAAARGLKTHVT EIPIQALYECDELF CSTAGGIMPLVLLDGNIVGDGTVGPVTRMIWEAYWDLHDDPQLSEPVT YAPLEHHHHHH ω-ΤΑ from Pseudomonas fluorescens - Amino acid sequence (SEQ ID NO: 29)

MNSNNKAWLKEHNTVHMMHPMQDPKALHEQRPLIIQSGKGVHITDVDGRRFIDCQGGLWCV NAGYGRREI IDAVTRQMEELAYYSLFPGSTNAPAIALSQKLTEVAAEEGMVKASFGLGGSDAV ETALKIARQYWKLEGQPDKVKFVSLYNGYHGLNFGGMSACGGNAWKSSYEPLMPGFFQVES PHLYRNPFTNDPEELAEICAQILERQiEMQAPGTVAALIAEPIQGAGGVIVPPASYWPRLRQICD KYDILLIADEVITGLGRSGSLFGSRGWGVKPDIMCLAKGISSGYVPLSATLVNSRVARAWERD AGFTSVYMHGYTYSGHPVSCAAALAAIDIVLQENLAENARVVGDYFLEKLLILKDKHRAIGDVR GKGLMLAVELVKERATKEPFGPADAYPLAISEACVNNGVMIRTIVNKLIISPPLTFTTEHVDEVIE VLDRAFVANPW ω-ΤΑ from Silicibacter pomeroyi - Amino acid sequence (SEQ ID NO: 31 )

MSLATITNHMPTAELQALDAAHHLHPFSANNALGEEGTRVITRARGVWLNDSEGEEILDAMAG LWCVNIGYGRDELAEVAARQMRELPYYNTFFKTTHVPAIALAQKLAELAPGDLNHVFFAGGG SEANDTNIRMVRTYWQNKGQPEKTVI !SRKNAYHGSTVASSALGG AGMHAQSGLI PDVHHI NQPNWWAEGGDMDPEEFGLARARELEEAILELGENRVAAFIAEPVQGAGGVIVAPDSYWPEI QRICDKYDILLIADEVICGFGRTGNWFGTQTMGIRPHIMTIAKGLSSGYAPIGGSIVCDEVAHVI GKDEFNHGYTYSGHPVAAAVALENLRILEEENILDHVRNVAAPYLKEKWEALTDHPLVGEAKI VGMMASIALTPNKASRAKFASEPGTIGYICRERCFANNLIMRHVGDRMIISPPLVITPAEIDEMF VRIRKSLDEAQAEIEKQGLMKSEG ω-ΤΑ from Ochrobactrum anthropi - Amino acid sequence (SEQ ID NO: 33)

MTAQPNSLEARDIRYHLHSYTDAVRLEAEGPLVIERGDGIYVEDVSGKRYIEAMSGLWSVGVG FSEPRLAEAAARQMKKLPFYHTFSYRSHGPVIDLAEKLVSMAPVPMS AYFTNSGSEANDTV VKLIWYRSNALGEPERKKI ISRKRGYHGVTIASASLTGLPNNHRSFDLPIDRILHTGCPHFYREG QAGESEEQFATRLADELEQLIIAEGPHTIAAFIGEPVMGAGGVWPPKTYWEKVQAVLKRYDIL LIADEVICGFGRTGNLFGSQTFDMKPDILVMSKQLSSSYLPISAFLINERVYAPIAEESHKIGTLG TGFTASGHPVAAAVALENLAIIEERDLVANARDRGTY QKRLRELQDHPLVGEVRGVG LIAGV ELVTDKQAKTGLEPTGALGAKANAVLQERGVISRAMGDTLAFCPPLIINDQQVDTMVSALEAT LNDVQASLTR ω-ΤΑ from Neosartorya fischeri - Amino acid sequence (SEQ I D NO: 35)

MASMDKVFSGYHARQKLLERSDNPFSKGIAYVEGKLVLPSDARIPLLDEGFMHGDLTYDVTTV WDGRFFRLDDHMQRILESCDKMRLKFPLAPSTVKNILAEMVAKSGIRDAFVEVIVTRGLTGVR GSKPEDLYNNNIYLLVLPYVWVMAPENQLLGGSAIITRTVRRTPPGAFDPTIKNLQWGDLTKG LFEAMDRGATYPFLTDGDTNLTEGSGFNIVLVKNGIIYTPDRGVLRGITRKSVIDVARANNIDIR LEVVPVEQVYHSDEIFMCTTAGGIMPITLLDGQPVNDGQVGPIT KIWDGYWEMHYNPAYSFP VDYGSG ω-ΤΑ from Gibberella zeae - Amino acid sequence (SEQ ID NO: 37)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EA PGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRI PPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVE AYEADEIFMCTTAGGI PITTMDGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

ArRmutl 1 -G279A-co-TA - amino acid sequence (SEQ ID NO: 39)

HMTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDATYT TFHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAMVTVTITRGYSS TPFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDL IRAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTAGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY

ArRmut1 1 -M1 17F-u)-TA - amino acid sequence (SEQ ID NO: 41)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDATYTT FHVWNGNAFRLGDHIERLFSNAESIRLI PPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSiDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAiLADITPAELYDADEVLGCSTGGGVWPFVSVDGNSISDGVPGPITQSI IRRYWELNVEPSSL LTPVQY

ArRmut1 1 - 1 17F-G279A-o TA - amino acid sequence (SEQ I D NO: 43)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAA IEGAFVPPSEARISIFDQGFYTSDATYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHL VAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTAGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY ArRmutl 1-A60V-M1 17F-0--TA - amino acid sequence (SEQ ID NO: 45)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHiERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNWVIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILAD!TPAELYDADEVLGCSTGGGVWPFVSVDGNSISDGVPGPiTQSMRRYWELNVEPSSL LTPVQY

ArRmutl 1 -A60V-M117F-G279A-u)-TA - amino acid sequence (SEQ ID NO: 47)

MTSE!VYTHDTGLDYiTYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQWMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVW!KDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTAGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY

ArRmutl 1-A60V-M117F-G279V-U)-TA - Amino acid sequence (SEQ ID NO: 49)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTVGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY

ArRmutl 1-A60V-M1 17F-G279L-OJ-TA - Amino acid sequence (SEQ ID NO: 51 )

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQ GDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGVVRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTLGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSLL TPVQY ArRmutl 1 -A60V-M1 17F-G279i-g)-TA - Amino acid sequence (SEQ ID NO: 53) MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEAR!SIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQ IVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNNAA/IKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTIGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSLL TPVQY

ArRmutl 1-A60V-M117F-G279F-oo-TA - Amino acid sequence (SEQ ID NO: 55)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHV NGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNNAA IKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTFGGWVPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSLL TPVQY

ArRmutl 1 -A60V-M117F-G279V-T64V-u)-TA - Amino acid sequence (SEQ ID NO: 57)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTV FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTVGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY ArRmutl 1 -A60V-M117F-G279V-TH 9E-u)-TA - Amino acid sequence (SEQ ID NO: 59)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVEVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTVGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY

ArRmutl 1 -A60V-M117F-G279V-l152V-u TA - Amino acid sequence (SEQ ID NO: 61 )

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWWPFDRIRDGVHL VAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCSTVGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY

ArRmutl 1-A60V-M117F-G279V-S277T-u)-TA - Amino acid sequence (SEQ ID NO: 63)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHVWNGNAFRLGDHIERLFSNAESIRLIPPLTQDEVKEIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCTTVGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSLL TPVQY

ArRmutl 1-A60V-M117F-G279V-S277N-u)-TA - Amino acid sequence (SEQ ID NO: 65)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGFYTSDVTYTT FHWVNGNAFRLGDHIERLFSNAESIRLIPPLTQDEV EIALELVAKTELREAFVTVTITRGYSST PFERDITKHRPQVYMSACPYQWIVPFDRIRDGVHLMVAQSVRRTPRSSIDPQVKNFQWGDLI RAIQETHDRGFELPLLLDCDNLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLGH EAILADITPAELYDADEVLGCNTVGGVWPFVSVDGNSISDGVPGPITQSIIRRYWELNVEPSSL LTPVQY

ArR-S21 SP-uj-TA - Amino acid sequence (SEQ ID NO: 67)

TSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGYLHSDVTYTV FHVWNGNAFRLDDHIERLFSNAESMRIIPPLTQDEVKEIALELVAKTELREAFVSVSITRGYSST PGERDITKHRPQVYMYAVPYQWIVPFDRIRDGVHAMVAQSVRRTPRSSIDPQVKNFQWGDLI RAVQETHDRGFEAPLLLDGDGLLAEGPGFNVWIKDGWRSPGRAALPGITRKTVLEIAESLG HEAILADITLAELLDADEVLGCTTAGGVWPFVSVDGNPISDGVPGPITQSIIRRYWELNVESSSL LTPVQY

ArR-G 131 F-S218P-UJ-TA - Amino acid sequence (SEQ ID NO: 69)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGYLHSDVTYTV FHVWNGNAFRLDDHIERLFSNAESMRIIPPLTQDEVKEIALELVAKTELREAFVSVSITRGYSST PFERDITKHRPQVYMYAVPYQWIVPFDRIRDGVHAMVAQSVRRTPRSSIDPQVKNFQWGDLI RAVQETHDRGFEAPLLLDGDGLLAEGPGFNWVIKDGWRSPGRAALPGITRKTVLEIAESLG HEAILADITLAELLDADEVLGCTTAGGVWPFVSVDGNPISDGVPGPITQSIIRRYWELNVESSSL LTPVQY ArR-G131 F-oo-TA - Amino acid sequence (SEQ ID NO: 71)

MTSEIVYTHDTGLDYITYSDYELDPANPLAGGAAWIEGAFVPPSEARISIFDQGYLHSDVTYTV FHVWNGNAFRLDDHIERLFSNAESMRIIPPLTQDEVKEIALELVAKTELREAFVSVSITRGYSST PFERDITKHRPQVYMYAVPYQWIVPFDRIRDGVHAMVAQSVRRTPRSSIDPQVKNFQWGDLI RAVQETHDRGFEAPLLLDGDGLLAEGSGFNWVIKDGWRSPGRAALPGITRKTVLEIAESLG HEAILADITLAELLDADEVLGCTTAGGVWPFVSVDGNPISDGVPGPITQSIIRRYWELNVESSSL LTPVQY

GZ-L56V-o)-TA - Amino acid sequence (SEQ ID NO: 73)

MSTMDKIFAGHAQRQATLVASDNIFANG!AWIQGELVPLNEARIPLMDQGFMHGDVTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPY WVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADE I FMCTTAGG I MPITTM DG KPVKDG KVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-V60T- -TA - Amino acid sequence (SEQ ID NO: 75)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDTPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTAGGI MPITTM DGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL GZ-E1 15T-u)-TA - Amino acid sequence (SEQ ID NO: 77)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVTLIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNVVFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTAGGIMPITTM DGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-V148l-a TA - Amino acid sequence (SEQ ID NO: 79)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWIMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTRG MFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCGHE VRVEYVPVEMAYEADEIFMCTTAGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMHWEDE FSFKIDYQKLKL

GZ-T273S-UJ-TA - Amino acid sequence (SEQ ID NO: 81)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEAR!PLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCSTAGGIMPITT DGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-T273N-co-TA - Amino acid sequence (SEQ ID NO: 83)

STMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCNTAGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL GZ-A275V-oo-TA - Amino acid sequence (SEQ ID NO: 85)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKK RMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNViYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTVGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-Y58L-u)-TA - Amino acid sequence (SEQ ID NO: 87)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTLDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTAGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL GZ-E115D-o)-TA - Amino acid sequence (SEQ ID NO: 89)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASV KMRMQFPIPRDEIRMTLLDMLAKSGiKDAFVDLIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTAGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-V148A-u TA - Amino acid sequence (SEQ ID NO: 91)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGI DAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWAMSPEAQYVGGNAVIARTVRRiPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTAGGIMPITT DGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-L56V-A275V-oo-TA - Amino acid sequence (SEQ ID NO: 93)

MSTMDKIFAGHAQRQATLVASDN!FANGIAWIQGELVPLNEARIPLMDQGFMHGDVTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLD LAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYWWMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCTTVGGIMPITT DGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL GZ-L56V-T273S-U)-TA - Amino acid sequence (SEQ ID NO: 95)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDVTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYWWMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCSTAGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMHWE DEFSFKIDYQKLKL

GZ-T273S-A275V-a)-TA - Amino acid sequence (SEQ ID NO: 97)

MSTMDKIFAGHAQRQATLVASDNIFANGIAWIQGELVPLNEARIPLMDQGFMHGDLTYDVPAV WDGRFFRLDDHLDRLEASVKKMRMQFPIPRDEIRMTLLDMLAKSGIKDAFVELIVTRGLKPVR EAKPGEVLNNHLYLIVQPYVWVMSPEAQYVGGNAVIARTVRRIPPGSMDPTIKNLQWSDFTR GMFEAYDRGAQYPFLTDGDTNITEGSGFNWFVKNNVIYTPNRGVLQGITRKSVIDAAKWCG HEVRVEYVPVEMAYEADEIFMCSTVGGIMPITTMDGKPVKDGKVGPVTKAIWDRYWAMH E DEFSFKIDYQKLKL

The term "hybridization" or "hybridizing", as used herein in connection with nucleic acids, relates to hybridizations under conditions of any degree of stringency. In general, hybridizations of nucleic acids, such as Southern or Northern hybridizations, can be performed under experimental conditions of various degrees of stringency. Usually a nucleic acid immobilized on a solid support such as a membrane is contacted with a liquid containing another, similar nucleic acid (called probe) under suitable buffer and temperature conditions in order to selectively allow the interaction of the probe with the immobilized nucleic acid, wherein the probe has a certain degree of sequence identity to the immobilized nucleic acid to be tested. Commonly the buffer used for the hybridization, in particular the washing steps after hybridization, is standard sodium citrate buffer (SSC; also referred to as saline sodium citrate buffer). A 20-fold concentrated SSC buffer contains 3 M NaCI and 0.3 M sodium citrate, adjusted to pH 7.0 using HCI, and is commercially available, e.g., from Sigma Aldrich. The Na⁺ concentration of a corresponding 20-fold SSC buffer is 3.3 M (i.e., 3.3 mol/L), and it is 1.65 M for a 10-fold SSC buffer, 0.825 M for a 5-fold SSC buffer, 0.33 M for a 2-fold SSC buffer, 0.165 M for a 1-fold SSC buffer, and 0.0165 M for a 0.1 -fold SSC buffer. Formamide or sodium dodecyl sulfate (SDS) can be added to the SSC buffer to reduce unspecific binding of the probe. The stringency of the hybridization depends on the percentage of the nucleotides G and C present in the sequence of the probe (%G+C) and the hybridization conditions, particularly the temperature, the concentration of Na^" and the concentration of formamide or SDS (if present). In general, the higher the hybridization temperature and the lower the sodium (Na^*) concentration, the higher will be the stringency. The stringency of the hybridization can thus be controlled by appropriately choosing the temperature for the hybridization, the concentration of the SSC buffer (and thereby the sodium concentration) and optionally the concentration of formamide (or SDS) added to the SSC buffer. If different concentrations of SSC buffer are used in different steps of the hybridization procedure, the concentrations of sodium and formamide in the most concentrated SSC buffer (which is typically the buffer used for the final washing step) are decisive.

Accordingly, at a given GC-content of the probe and at specific concentrations of sodium and formamide (if present) in the most concentrated buffer used for the washing after hybridization (typically the buffer used for the final washing step), the stringency of the hybridization can be controlled by adjusting the hybridization temperature to a specific number of degrees Celsius (e.g., 25°C or less) below the effective melting temperature (Tm) which can be calculated using the following formula (for DNA-DNA-hybridizations): Tm = 81.5 + 16.6(log M [Na⁺]) + 0.41 (%G+C) - 0.72(% formamide)

In the above formula, "Tm" is the temperature under which the sequence of the immobilized nucleic acid to be tested needs to match 100% of the sequence of the probe in order for both sequences to hybridize to each other; "log M [Na⁺]" is the logarithm to base 10 (log-io) of the concentration of sodium (Na⁺) in mol/L in the buffer; "%G+C" is the percentage of the nucleotides G and C in the sequence of the probe (GC-content); and "% formamide" is the concentration of formamide in %(volume/volume) in the buffer.

The farther the hybridization temperature is below the Tm, the lower will be the stringency of the hybridization. In particular, for each 1.4°C the hybridization temperature is below the calculated Tm, the hybridization will still occur in the presence of 1% sequence mismatch, i.e., a mismatch of x% of the sequences of the probe and the immobilized nucleic acid to be tested will still lead to hybridization if the hybridization temperature is at least x · 1 .4°C below the calculated Tm. For example, if the Tm is calculated to be 90°C and the hybridization experiment is conducted at 65°C (i.e., 25°C below the Tm), nucleic acid sequences matching at least 82.1 % of the sequence of the probe will hybridize to the probe (i.e., 25°C/1.4°C = 17.9, meaning that 100% - 17.9% = 82.1 %).

As used herein, hybridization under "stringent conditions" preferably means that the hybridization temperature is about 25°C or less below the Tm (calculated using the formula explained above), which corresponds to a minimum sequence identity of about 82.1 % required for hybridization to occur (i.e., 100% - (25°C/1.4°C)%). More preferably, hybridization under stringent conditions means that the hybridization temperature is about 20°C or less below the Tm (corresponding to a minimum sequence identity of about 85.7% required for hybridization), even more preferably about 15"C or less below the Tm (corresponding to a minimum sequence identity of about 89.3% required for hybridization), even more preferably about 10°C or less below the Tm (corresponding to a minimum sequence identity of about 92.9% required for hybridization), even more preferably about 7°C or less below the Tm (corresponding to a minimum sequence identity of about 95.0% required for hybridization), yet even more preferably about 5°C or less below the Tm (corresponding to a minimum sequence identity of about 96.4% required for hybridization), and still more preferably about 3°C or less below the Tm (corresponding to a minimum sequence identity of about 97.9% required for hybridization). Conversely, hybridization under "non-stringent conditions" means that the hybridization temperature is below the above-defined temperature required for stringent hybridization. As used herein, the term "mutein" refers to a protein or polypeptide differing by one or more amino acids from a given reference protein or polypeptide, wherein such difference is caused by the addition, substitution or deletion of at least one amino acid or any combination of such addition(s), substitution(s) and/or deletion(s). In particular, a "mutein" of an ω-transaminase refers to a protein/polypeptide having at least 70% (preferably at least 80% , more preferably at least 85%, even more preferably at least 90%, and yet even more preferably at least 95%) amino acid sequence identity with the corresponding reference ω-transaminase and having the enzymatic activity of an ω-transaminase. Accordingly, a "mutein" of an (^-selective ω-transaminase refers to a protein/polypeptide having at least 70% (preferably at least 80%, more preferably at least 85%, even more preferably at least 90%, and yet even more preferably at least 95%) amino acid sequence identity with the corresponding (R)-selective ω-transaminase and having the enzymatic activity of an (ft)-selective ω-transaminase with regard to catalyzing the stereoselective reductive amination of a compound of formula (I). A "mutein" of an (S)-selective ω-transaminase refers to a protein/polypeptide having at least 70% (preferably at least 80%, more preferably at least 85%, even more preferably at least 90%, and yet even more preferably at least 95%) amino acid sequence identity with the corresponding (S)-selective ω-transaminase and having the enzymatic activity of an (S)-selective ω-transaminase with regard to catalyzing the stereoselective reductive amination of a compound of formula (I).

The term "about" preferably refers to ±10% of the indicated numerical value, more preferably to ±5% of the indicated numerical value, and in particular to the exact numerical value indicated.

It is to be understood that the present invention specifically relates to each and every combination of features and embodiments described herein, including any combination of general and/or preferred features/embodiments. In particular, the invention specifically relates to all combinations of preferred features (including all degrees of preference) of the processes provided herein. In this specification, a number of documents including patent applications and scientific literature are cited. The disclosure of these documents, while not considered relevant for the patentability of this invention, is herewith incorporated by reference in its entirety. More specifically, all referenced documents are incorporated by reference to the same extent as if each individual document was specifically and individually indicated to be incorporated by reference. The present invention particularly relates to the following items:

1. A process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (I)

(I) wherein R¹ is a carboxy protecting group or hydrogen, and R² is alkyl or arylalkyl, is contacted either with an (R)-selective ω-transaminase or with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (I) and obtain a compound of the following formula (II)

(II) wherein R¹ and R² are the same as in formula (I), and further wherein the (R)-isomer and the (S)-isomer of the compound of formula refer to the following configurations:

2. The process of item 1 , wherein at least 70 wt-% of the compound of formula (I) employed in the combined racemization and stereoselective reductive amination step is converted into the compound of formula (II). The process of item 1 or 2, wherein at least 90 wt-% of the compound of formula (I) employed in the combined racemization and stereoselective reductive amination step is converted into the compound of formula (I I).

The process of any one of items 1 to 3, wherein R¹ is selected from alkyl, alkenyl, aryl, arylalkyl, and hydrogen.

The process of any one of items 1 to 4, wherein R¹ is selected from methyl, ethyl, n-propyl, isopropyl, ferf-butyl, vinyl, allyl, benzyl, and hydrogen.

The process of any one of items 1 to 5, wherein R² is selected from ethyl, n-propyl, isopropyl, isobutyl, and benzyl.

The process of any one of items 1 to 6, wherein the compound of formula (I) is a racemate.

The process of any one of items 1 to 7, wherein the ω-transaminase is an (R)-selective ω-transaminase, and wherein the compound of formula (II) is a compound of formula (W-R)

wherein R¹ and R² are the same as in formula The process of item 8. wherein R² is n-propyl. The process of item 8 or 9, wherein the (R)-selective ω-transaminase is selected from ω-transaminase from Hyphomonas neptunium, ω-transaminase from Arthrobacter citreus, ω-transaminase from Chromobacterium violaceum DSM 30191 , ω-transaminase from Pseudomonas fluorescens, ω-transaminase from Pseudomonas putida KT2440 gene PP2180, ω-transaminase from Ochrobactrum anthropi, ω-transaminase from Silicibacter pomeroyi, ω-transaminase from Vibrio fluvialis, ω-transaminase from Neosartorya fischeri, ω-transaminase from Paracoccus denitrificans, ω-transaminase from Pseudomonas putida KT2440 gene PP5182, ω-transaminase from Arthrobacter sp. KNK168, ArR-G131 F-u)-TA, ArRmutl 1 -M1 17F-G279A-u)-TA, and a mutein of any one of the aforementioned ω-transaminases.

1 1. The process of any one of items 8 to 10, wherein the (^-selective ω-transaminase is ω-transaminase from Hyphomonas neptunium or a mutein thereof.

12. The process of item 8, wherein the (f?)-selective ω-transaminase is ω-transaminase from Hyphomonas neptunium, wherein R¹ is isopropyl or ferf-butyl, and wherein R² is n-propyl. 13. The process of any one of items 8 to 12, wherein the compound of formula (ll-f?) is obtained in an enantiomeric excess of at least about 70%. 14. The process of any one of items 8 to 13, wherein the compound of formula (ll-f?) is obtained in an enantiomeric excess of at least about 90%.

15. The process of any one of items 8 to 14, further comprising a step of converting the compound of formula (ll-f?) into a compound of the following formula (lll-f?)

wherein R² is the same as in formula (ll-f?).

16. The process of any one of items 8 to 14, wherein the process optionally further comprises a step of converting the compound of formula (ll-f?) into a compound of the following formula (lll-f?)

(lll-f?) wherein R² is the same as in formula (ll-f?), and wherein the compound of formula (W-R) or (lll-R) is further converted into a compound having the following formula (IV-f?):

(IV-R) wherein R² is the same as in formula (ll-R) or (lll-R).

The process of any one of items 1 to 7, wherein the ω-transaminase is an (S)-selective ω-transaminase, and wherein the compound of formula (II) is a compound of formula (II-

wherein R¹ and R² are the same as in formula (I).

The process of item 17, wherein R² is selected from isobutyl, n-propyl, isopropyl, ethyl, and benzyl.

The process of item 17 or 18, wherein R² is isobutyl.

The process of any one of items 17 to 19, wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibbereila zeae, ω-transaminase from Aspergillus terreus, ω-transaminase from Arthrobacter sp. KNK168, ω-transaminase from Ralstonia eutropha, ω-transaminase from Bacillus megaterium, ArRmutl 1 -oo-TA, ArR-S218P- G131 F-W-TA, ArRmutl 1 -M1 17F-oj-TA. ArRmutl 1 -G279A-0J-TA, ArRmutl 1 -M 1 17F- A60V-OJ-TA, ArRmutl 1 -M1 17F-A60V-G279A-IU-TA, ArRmutl 1 -M 117F-A60V-G279V-o TA, ArRmutl 1 -M 1 17F-A60V-G279L-oo-TA, ArRmutl 1 -M 17F-A6QV-G279l-uj-TA. ArRmutl 1 -M 1 1 7F-A60V-G279F-OJ-TA. ArRmutl 1 -M 1 17F-A60V-G279V-I 1 52V-W-TA, ArRmutl 1 -M1 17F-A60V-G279V-S277T^-TA, GZ-L56V^-TA, GZ-V60T- -TA. GZ- Ε1 15Τ-ω-ΤΑ, GZ-T273S-OJ-TA, GZ-EH SD-ω-ΤΑ, GZ-V148A-oo-TA, GZ-L56V-T273S-io- ΤΑ, GZ-T273S-A275V-UJ-TA, and a mutein of any one of the aforementioned ω-transaminases.

The process of any one of items 17 to 20, wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibberella zeae, ArRmutl 1-M1 17F-A60V-G279V-oo- TA, ArRmutl 1-M1 17F-A60V-G279L-co-TA, ArRmutl 1-M117F-A60V-G279I-0J-TA, ArRmutl 1 -M1 17F-A60V-G279F-0J-TA, ArRmutl 1 -M1 17F-A60V-G279A-oo-TA,

ArRmutl 1 -M1 17F-A60V-G279V-I152V-UJ-TA, GZ-L56V-io-TA, GZ-T273S-OJ-TA, GZ- E1 15D-UJ-TA, and GZ-L56V-T273S-CO-TA.

22. The process of any one of items 17 to 21 , wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibberella zeae, ArRmutl 1 -M117F-A60V-G279V-u>- TA, ArRmutl 1 -M1 17F-A60V-G279I-U)-TA, GZ-L56V-G0-TA, and GZ-T273S-00-TA.

23. The process of item 17, wherein the (S)-selective ω-transaminase is selected from ArRmutl 1 -M1 17F-A60V-G279V-ou-TA, ArRmutl 1 -M1 17F-A60V-G279I-OJ-TA and GZ-L56V-DJ-TA, R¹ is terf-butyl, and R² is alkyl. 24. The process of item 17, wherein the (S)-selective ω-transaminase is ArRmut11 -M1 17F- Α60ν-σ279ν-ω-ΤΑ or GZ-L56V-io-TA, R¹ is tert-butyl, and R² is isobutyl. 25. The process of any one of items 17 to 24, wherein the compound of formula (ll-S) is obtained in an enantiomeric excess of at least about 50%.

26. The process of any one of items 17 to 25, wherein the compound of formula (ll-S) is obtained in an enantiomeric excess of at least about 70%. 27. The process of any one of items 17 to 26, further comprising a step of converting the compound of formula (ll-S), wherein R¹ is different from hydrogen, into a compound of the following formula (V-S)

(V-S) wherein R² is the same as in formula (ll-S). . The process of any one of items 17 to 26, further comprising a step of converting the compound of formula (ll-S) into a compound of the following formula (ll-l-S)

wherein R² is the same as in formula (ll-S). . The process of any one of items 1 to 28, wherein the stereoselective reductive amination step is conducted in the presence of pyridoxal-5'-phosphate (PLP) and/or pyridoxamine- 5'-phosphate (PMP). . The process of any one of items 1 to 29, wherein the stereoselective reductive amination step is conducted in the presence of pyridoxal-5'-phosphate (PLP). . The process of any one of items 1 to 28, wherein the stereoselective reductive amination step is conducted without adding pyridoxal-5'-phosphate (PLP) or pyridoxamine- 5 -phosphate (PMP). , The process of any one of items 1 to 31 , wherein the amine, in the presence of which the compound of formula (I) is contacted with the ω-transaminase, is selected from an amino acid, an alkylamine, an alkenylamine, an arylamine, an aralkylamine, an arylalkenylamine, a heteroarylamine, a heteroarylalkylamine, a heteroarylalkenylamine, pyridoxamine-5^'-phosphate (PMP), and any of the aforementioned compounds substituted with at least one further amino group. . The process of item 32, wherein the amine is selected from alanine, glutamate, ethylamine, 1 ,2-diaminoethane, n-propylamine, 1 ,3-diaminopropane, isopropylamine, 1 ,2-diaminopropane, 2-butylamine, 1-indolamine, 1-indanamine, 1 -aminotetralin, and 1-phenylethylamine. . The process of item 32 or 33, wherein the amine is alanine. 35. The process of item 34, wherein the stereoselective reductive amination step is conducted in a reaction medium comprising:

(i) alanine dehydrogenase (Ala-DH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and ammonium formate; or (ii) alanine dehydrogenase (Ala-DH), glucose dehydrogenase (GDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), glucose, and an ammonium salt; or

(iii) alanine dehydrogenase (Ala-DH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and ammonium phosphite. 36. The process of item 34, wherein the stereoselective reductive amination step is conducted in a reaction medium comprising:

(i) lactate dehydrogenase (LDH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (NADH or NAD'), and a formate salt; or

(ii) lactate dehydrogenase (LDH), glucose dehydrogenase (GDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and glucose; or

(iii) lactate dehydrogenase (LDH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and a phosphite salt.

37. The process of item 32 or 33, wherein the amine is isopropylamine.

38. The process of any one of items 1 to 37, wherein the stereoselective reductive amination step is conducted in an aqueous medium at a pH in the range of about 6 to about 1 1.

39. The process of item 38, wherein the aqueous medium comprises a phosphate buffer, a Tris buffer, a PIPES buffer, or a HEPES buffer.

40. The process of item 38, wherein the aqueous medium does not comprise any buffering agent. 41. The process of any one of items 1 to 40, wherein the stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent.

42. The process of item 41 , wherein the organic cosolvent is selected from 1 ,2-dimethoxyethane, dimethylformamide, dimethyl sulfoxide, tetrahydrofuran, acetonitrile, a C_{1 6} alkanol, and mixtures thereof. 43. The process of item 41 or 42, wherein the organic cosolvent is selected from 1 ,2-dimethoxyethane, dimethylformamide, and dimethyl sulfoxide.

44. The process of any one of items 8 to 10, 13 to 16 or 29 to 40, wherein the (f?)-selective ω-transaminase is ω-transaminase from Vibrio fluvialis, R¹ is ierf-butyl, R² is n-propyl, and the stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% of 1 ,2-dimethoxyethane.

45. The process of any one of items 17 to 40, wherein the (S)-selective ω-transaminase is ArRmutI 1 -M1 17F-G279V-A60V-u)-TA and the stereoselective reductive amination step is conducted in an aqueous medium comprising about 10 vol-% to about 25 vol-% of dimethylformamide or about 10 vol-% to about 25 vol-% of dimethyl sulfoxide.

46. The process of item 45, wherein the stereoselective reductive amination step is conducted in an aqueous medium comprising about 20 vol-% dimethylformamide.

47. The process of any one of items 1 to 46, wherein the stereoselective reductive amination step is conducted at a temperature of about 20°C to about 50°C. 48. The process of any one of items 1 to 47, wherein the ω-transaminase has been expressed in Escherichia coli.

49. The process of any one of items 1 to 48, wherein the ω-transaminase is immobilized on a solid support material, or wherein cells containing the ω-transaminase are immobilized on a solid support material.

50. A process for the production of brivaracetam, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (la)

(la) wherein R¹ is a carboxy protecting group or hydrogen, is contacted with an (R)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (la) and obtain a compound of the following formula (lla)

(lla) wherein R is the same as in formula (la), wherein the process optionally further comprises a step of converting the compound of formula (lla) obtained in the stereoselective reductive amination step into a compound of the following formula (Ilia)

(Ilia) and wherein the compound of formula (lla) or (Ilia) is further converted into brivaracetam having the following formula (IVa):

The process of item 50, wherein at least 70 wt-% of the compound of formula (la) employed in the combined racemization and stereoselective reductive amination step is converted into the compound of formula (lla).

The process of item 50 or 51 , wherein at least 90 wt-% of the compound of formula (la) employed in the combined racemization and stereoselective reductive amination step is converted into the compound of formula (lla). 53. The process of any one of items 50 to 52, wherein R¹ is selected from alkyl, alkenyl, aryl, arylalkyl, and hydrogen. 54. The process of any one of items 50 to 53, wherein R is selected from methyl, ethyl, n-propyl, isopropyl, terf-butyl, vinyl, allyl, benzyl, and hydrogen.

55. The process of any one of items 50 to 54, wherein the compound of formula (la) is a racemate.

56. The process of any one of items 50 to 55, wherein the (R)-selective ω-transaminase is selected from ω-transaminase from Hyphomonas neptunium, ω-transaminase from Arthrobacter citreus, ω-transaminase from Chromobacterium violaceum DSM 30191 , ω-transaminase from Pseudomonas fluorescens, ω-transaminase from Pseudomonas putida KT2440 gene PP2180, ω-transaminase from Ochrobactrum anthropi. ω-transaminase from Silicibacter pomeroyi, ω-transaminase from Vibrio fluvialis, ω-transaminase from Neosartorya fischeri, ω-transaminase from Paracoccus denitrificans, ω-transaminase from Pseudomonas putida KT2440 gene PP5182, ω-transaminase from Arthrobacter sp. KNK168, ArR-G 131 F-w-TA, ArRmutl 1 -M117F- Θ279Α-ω-ΤΑ, and a mutein of any one of the aforementioned ω-transaminases.

57. The process of any one of items 50 to 56, wherein the (R)-selective ω-transaminase is ω-transaminase from Hyphomonas neptunium or a mutein thereof. 58. The process of item 50, wherein the (^-selective ω-transaminase is ω-transaminase from Hyphomonas neptunium, and wherein R¹ is isopropyl or terf-butyl.

59. The process of any one of items 50 to 58, wherein the stereoselective reductive amination step is conducted in the presence of pyridoxal-5^!-phosphate (PLP) and/or pyridoxamine-5'-phosphate (PMP).

60. The process of any one of items 50 to 59, wherein the stereoselective reductive amination step is conducted in the presence of pyridoxal-5'-phosphate (PLP). 61. The process of any one of items 50 to 58, wherein the stereoselective reductive amination step is conducted without adding pyridoxal-5'-phosphate (PLP) or pyridoxamine-5^:-phosphate (PMP). The process of any one of items 50 to 61 , wherein the amine, in the presence of which the compound of formula (la) is contacted with the ω-transaminase, is selected from an amino acid, an alkylamine, an alkenylamine, an arylamine, an aralkylamine, an arylalkenylamine, a heteroarylamine, a heteroarylalkylamine, a heteroarylalkenylamine, pyridoxamine-5'-phosphate (PMP), and any of the aforementioned compounds substituted with at least one further amino group.

The process of item 62, wherein the amine is selected from alanine, glutamate, ethy!amine, 1 ,2-diaminoethane, n-propylamine, 1 ,3-diaminopropane, isopropylamine, 1 ,2-diaminopropane, 2-butylamine, 1-indolamine, 1-indanamine, 1-aminotetralin, and 1-phenylethylamine.

The process of item 62 or 63, wherein the amine is alanine.

The process of item 64, wherein the stereoselective reductive amination step is conducted in a reaction medium comprising:

(i) alanine dehydrogenase (Ala-DH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and ammonium formate; or

(ii) alanine dehydrogenase (Ala-DH), glucose dehydrogenase (GDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), glucose, and an ammonium salt; or

(iii) alanine dehydrogenase (Ala-DH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (NADH or NAD ), and ammonium phosphite.

(i) lactate dehydrogenase (LDH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (NADH or NAD^'), and a formate salt; or

The process of item 62 or 63, wherein the amine is isopropylamine. 68. The process of any one of items 50 to 67, wherein the stereoselective reductive amination step is conducted in an aqueous medium at a pH in the range of about 6 to about 1 1 . 69. The process of item 68, wherein the aqueous medium comprises a phosphate buffer, a Tris buffer, a PIPES buffer, or a HEPES buffer.

70. The process of item 68, wherein the aqueous medium does not comprise any buffering agent.

71 . The process of any one of items 50 to 70, wherein the stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent. 72. The process of Item 71 , wherein the organic cosolvent is selected from 1 ,2-dimethoxyethane, dimethylformamide, dimethyl sulfoxide, tetrahydrofuran, acetonitrile, a C _e alkanol, and mixtures thereof.

73. The process of item 71 or 72, wherein the organic cosolvent is selected from 1 ,2-dimethoxyethane, dimethylformamide, and dimethyl sulfoxide.

74. The process of any one of items 50 to 56 or 59 to 70, wherein the (f?)-selective ω-transaminase is ω-transaminase from Vibrio fluvialis, R¹ is ieri-butyl, and the stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 35 vol-% of 1 ,2-dimethoxyethane.

75. The process of any one of items 50 to 74, wherein the stereoselective reductive amination step is conducted at a temperature of about 20^'C to about 50°C. 76. The process of any one of items 50 to 75. wherein the (R)-selective ω-transaminase has been expressed in Escherichia coli.

77. The process of any one of items 50 to 76, wherein the (R)-selective ω-transaminase is immobilized on a solid support material, or wherein cells containing the (R)-selective ω-transaminase are immobilized on a solid support material. A process for the production of pregabalin, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (lb)

wherein R¹ is the same as in formula (lb), wherein, if R¹ is different from hydrogen, the process further comprises a step of converting the compound of formula (lib) into pregabalin having the following formula (Vb)

The process of item 78, wherein at least 50 wt-% of the compound of formula (lb) employed in the combined racemization and stereoselective reductive amination step is converted into the compound of formula (lib). 80. The process of item 78 or 79, wherein at least 70 wt-% of the compound of formula (lb) employed in the combined racemization and stereoselective reductive amination step is converted into the compound of formula (lib). 81 . The process of any one of items 78 to 80, wherein R¹ is selected from alkyl, alkenyl, aryl, aryialkyl, and hydrogen.

82. The process of any one of items 78 to 81 , wherein R¹ is selected from methyl, ethyl, n-propyl, isopropyl, te/f-butyl, vinyl, allyl, benzyl, and hydrogen.

83. The process of any one of items 78 to 82, wherein the compound of formula (lb) Is a racemate.

84. The process of any one of items 78 to 83, wherein the (S)-selective ω-transaminase is selected from w-transaminase from Gibbere!la zeae, ω-transaminase from Aspergillus terreus, ω-transaminase from Arthrobacter sp. KNK168, ω-transaminase from Ralstonia eutropha, ω-transaminase from Bacillus megaterium, ArRmutl 1 -ω-ΤΑ, ArR-S218P- G 131 F-G0-TA, ArRmutl 1 -M 1 17 F-ω-ΤΑ, ArRmutl 1 -G279A-u>-TA, ArRmutl 1 -M1 17F- Α60ν-ω-ΤΑ, ArRmutl 1 -M l 17F-A60V-G279A-oo-TA, ArRmutl 1-M1 17F-A60V-G279V-U)- TA, ArRmutl 1-M1 17F-A60V-G279L-oo-TA, ArRmutl 1 -M1 17F-A60V-G279l-oo-TA,

ArRmutl 1 -M1 17F-A60V-G279F-to-TA, ArRmutl 1 -M1 17F-A60V-G279V-l 152V-oo-TA, ArRmutl 1 -M1 17F-A60V-G279V-S277T-W-TA, GZ-L56V-to-TA, GZ-V60T-oo-TA, GZ- Ε 1 15Τ-ω-ΤΑ, GZ-T273S-00-TA, GZ-E1 15D-G0-TA, GZ-V148A-UJ-TA, GZ-L56V-T273S-u>- TA, GZ-T273S-A275V-u)-TA, and a mutein of any one of the aforementioned ω-transaminases.

85. The process of any one of items 78 to 84, wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibberella zeae, ArRmutl 1 -M 1 17F-A60V-G279V- ω-ΤΑ, ArRmutl 1-M1 17F-A60V-G279L-u)-TA, ArRmutl 1-M1 17F-A60V-G279l-uj-TA, ArRmutl 1-M1 17F-A60V-G279F-OJ-TA, ArRmutl 1 -M1 17F-A60V-G279A-u>-TA,

ArRmutl 1-M1 17F-A60V-G279V-l 152V-u TA, GZ-L56V-u>-TA, GZ-T273S-u)-TA, GZ-E1 5D- -TA, and GZ-L56V-T273S-to-TA.

86. The process of any one of items 78 to 85, wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibberella zeae, ArRmutl 1 - 1 17F-A60V-G279V- ω-ΤΑ, ArRmutl 1 -M 1 17F-A60V-G279I^-TA, GZ-L55V^-TA, and GZ-T273S-W-TA. The process of item 78, wherein the (S)-selective ω-transaminase is selected from ArRmutl 1 -M1 17F-A60V-G279V-io-TA, ArRmutl 1 -M1 17F-A60V-G279l- -TA and GZ-L56V-u)-TA, and R¹ is ferf-butyl.

The process of item 78, wherein the (S)-selective ω-transaminase is ArRmutl 1 -M 1 1 7F- A60V-G279V-CJO-TA or GZ-L56V-u)-TA, and R¹ is ferf-butyl.

89. The process of any one of items 78 to 88, wherein the stereoselective reductive amination step is conducted in the presence of pyridoxal-5'-phosphate (PLP) and/or pyridoxamine-5'-phosphate (PMP) .

90. The process of any one of items 78 to 89, wherein the stereoselective reductive amination step is conducted in the presence of pyridoxal-5'-phosphate (PLP). 91 . The process of any one of items 78 to 88, wherein the stereoselective reductive amination step is conducted without adding pyridoxal-5'-phosphate (PLP) or pyridoxamine-5'-phosphate (PMP) .

92. The process of any one of items 78 to 91 , wherein the amine, in the presence of which the compound of formula (lb) is contacted with the ω-transaminase, is selected from an amino acid, an alkylamine, an alkenylamine, an arylamine, an aralkylamine, an arylalkenylamine, a heteroarylamine. a heteroarylalkylamine, a heteroarylalkenylamine, pyridoxamine-5^'-phosphate (PMP) , and any of the aforementioned compounds substituted with at least one further amino group.

93. The process of item 92, wherein the amine is selected from alanine, glutamate, ethylamine, 1 ,2-diaminoethane, n-propylamine, 1 ,3-diaminopropane, isopropylamine, 1 ,2-diaminopropane, 2-butylamine, 1-indolamine, 1-indanamine, 1-aminotetralin. and 1 -phenylethylamine.

94. The process of item 92 or 93, wherein the amine is alanine.

The process of item 94, wherein the stereoselective reductive amination step is conducted in a reaction medium comprising:

(i) alanine dehydrogenase (Ala-DH), formate dehydrogenase (FDH). nicotinamide adenine dinucleotide (NADH or NAD⁺), and ammonium formate; or (ii) alanine dehydrogenase (Ala-DH), glucose dehydrogenase (GDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), glucose, and an ammonium salt; or

(iii) alanine dehydrogenase (Ala-DH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and ammonium phosphite.

96. The process of item 94, wherein the stereoselective reductive amination step is conducted in a reaction medium comprising:

(i) lactate dehydrogenase (LDH), formate dehydrogenase (FDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and a formate salt; or

(iii) lactate dehydrogenase (LDH), phosphite dehydrogenase (PTDH), nicotinamide adenine dinucleotide (NADH or NAD⁺), and a phosphite salt. 97. The process of item 92 or 93, wherein the amine is isopropylamine.

98. The process of any one of items 78 to 97, wherein the stereoselective reductive amination step is conducted in an aqueous medium at a pH in the range of about 6 to about 11.

99. The process of item 98, wherein the aqueous medium comprises a phosphate buffer, a Tris buffer, a PIPES buffer, or a HEPES buffer.

100. The process of item 98, wherein the aqueous medium does not comprise any buffering agent.

101 . The process of any one of items 78 to 100, wherein the stereoselective reductive amination step is conducted in an aqueous medium comprising about 5 vol-% to about 45 vol-% of at least one organic cosolvent.

102. The process of item 101 , wherein the organic cosolvent is selected from 1.2-dimethoxyethane, dimethylformamide, dimethyl sulfoxide, tetrahydrofuran, acetonitrile, a C-|.₆ alkanol, and mixtures thereof. 103. The process of item 101 or 102, wherein the organic cosolvent is selected from 1 ,2-dimethoxyethane, dimethylformamide, and dimethyl sulfoxide. 104. The process of any one of items 78 to 100, wherein the (S)-selective ω-transaminase is ArRmutl 1 -M1 17F-G279V-A60V-o TA and the stereoselective reductive amination step is conducted in an aqueous medium comprising about 10 vol-% to about 25 vol-% of dimethylformamide or about 10 vol-% to about 25 vol-% of dimethyl sulfoxide.

105. The process of item 104, wherein the stereoselective reductive amination step is conducted in an aqueous medium comprising about 20 vol-% dimethylformamide.

106. The process of any one of items 78 to 105, wherein the stereoselective reductive amination step is conducted at a temperature of about 20°C to about 50°C.

107. The process of any one of items 78 to 106, wherein the (S)-selective ω-transaminase has been expressed in Escherichia coli. 108. The process of any one of items 78 to 107, wherein the (S)-selective co-transaminase is immobilized on a solid support material, or wherein cells containing the (S)-selective ω-transaminase are immobilized on a solid support material.

The invention will now be described by reference to the following examples which are merely illustrative and are not to be construed as a limitation of the scope of the present invention.

EXAMPLES

Analytical Methods

Determination of conversion GC-MS -Measurements

All gas chromatography-mass spectrometry (GC-MS) measurements were carried out with an Agilent 7890A GC system, equipped with an Agilent 5975C mass-selective detector (electron impact, 70 eV) and using a (5%-phenyl)-methylpolysiloxane phase column (Agilent HP-5ms, 30 m column length, 250 pm inner diameter of column, 0.25 μιτι film thickness (of the (5%- phenyl)-methylpolysiloxane phase)). Helium was used as carrier gas. GC program parameters: injector 250°C; constant flow 2 mL/min; temperature program 100°C/hold 0 min; 200°C/rate 10°C per min/hold 0 min; 300°C/rate 20°C per min/hold 5 min. Nuclear magnetic resonance (NMR)-Measurements:

¹H- and ³C-NMR spectra were recorded at 20°C on a Avance III ultrashielded 300 MHz Bruker NMR unit; chemical shifts are given in ppm relative to Me₄Si or the corresponding solvent signal ( H: CDCI₃ = 7.26 ppm; ¹³C CDCI₃ = 77.0 ppm).

GC Method 1

Conversions were measured by gas chromatography (GC) using an Agilent 7890 A GC system, equipped with an FID detector (Flame Ionization Detector) and using an Agilent J&W DB-1701 column (30 m, 250 pm, 0.25 pm). Helium was used as carrier gas. GC program parameters: injector 250°C; constant pressure 1 bar; temperature program: 807hold 6.5 min; 160°C/rate 10°C per min/hold 5 min; 200°C/rate 20°C per min/hold 2 min.

Retention times: compound 1a 11.9 min, compound 2a 12.3 min, compound 1 b 12.4 min, compound 2b 12.8 min, compound 1 c 12.8 min, compound 2c 13.2 min, compound 3 15.6 min, imine 4c 13,2 min, imine 5c 18.7 min, compound 10c 13.5 min, compound 12c 13.7 min, compound 13 16.8 min, imine 18c (10c+13) 21.2 min, imine 17c (10c+isopropylamine) 14.1 min, acetophenone 7.7 min, 1-methylbenzylamine (MBA) 5.8 min.

GC Method 2

Conversions were measured by GC using an Agilent 7890 A GC system, equipped with an FID detector and using an Agilent J&W DB-1701 column (30 m, 250 pm, 0.25 pm). Helium was used as carrier gas. GC program parameters: injector 250°C: constant pressure 1 bar; temperature program: 807hold 6.5 min; 160°C/rate 5°C per min/hold 5 min; 200°C/rate 20°C per min/hold 2 min.

Retention times: compound 1 d 30.3 min, compound 3 20.2 min. Determination of optical purity Method A

The enantiomeric excess (ee) was measured by GC using an Agilent 7890 A GC system, equipped with an FID detector and using a Macherey-Nagel Hydrodex^® β-TBDAc column (50 m, 250 pm). Hydrogen was used as carrier gas. GC program parameters: injector 250°C; constant flow 1.05 mL/min. Temperature program: 100°C/hold 0 min; 130°C/rate 10°C per min/hold 5 min; 160³ C/rate 5°C per min/hold 5 min; 220°C/rate 10°C per min/hold 10 min. Retention times: compound (S)-3 25.6 min, compound (R)-3 25.8 min. Method B

The ee was measured by GC using an Agilent 7890 A GC system, equipped with an FID detector and using a Restek Rt^®-b-DEXse column (30 m, 320 Mm, 0.25 pm). Hydrogen was used as carrier gas. GC program parameters: injector 250°C; constant flow 1 .3 mL/min.

Temperature program 100°C/hold 0 min; 130°C/rate 5°C per min/hold 0 min; 160°C/rate 1 °C per min/hold 0 min; 180°C/rate 10°C per min/hold 0 min.

Retention times: compound (R)-13 28.7 min, (S)-13 29.0 min

Materials

General: All starting materials were obtained from commercial suppliers and used as received with exception of racemic substrates 1a-1d, 10c and reference compounds. The synthesis of those compounds is described below. The reactions were carried out with standard Schlenk techniques under dry N₂ atmosphere in oven-dried (120°C) glassware.

Experimental procedure

Preparation of Enzymes and Muteins

The following ω-transaminases as well as different muteins/variants of ArR-ω-ΤΑ and of GZ-ω-ΤΑ (as described further below) were employed:

ΡΡ1 -ω-ΤΑ; co-transaminase from Pseudomonas putida KT2440 gene PP5182;

ΡΡ2-ω-ΤΑ: ca- -transaminase from Pseudomonas putida KT2440 gene PP2180;

PD-ω-ΤΑ: co-transaminase from Paracoccus denitrificans;

ArS-ω-ΤΑ: co- -transaminase from Arthrobacter citreus;

CV-ω-ΤΑ: ω· -transaminase from Chromobacterium violaceum DSM 30191 ;

PF-ω-ΤΑ: co- -transaminase from Pseudomonas fluoresceins;

VF-ω-ΤΑ: co-transaminase from Vibrio fluvialis;

ΑΤ-ω-ΤΑ: co- -transaminase from Aspergillus terreus;

ArR-ω-ΤΑ: co-transaminase from Arthrobacter sp. KNK168;

ΗΝ-ω-ΤΑ: co- -transaminase from Hyphomonas neptunium;

RE-ω-ΤΑ: ω· -transaminase from Ralstonia eutropha;

ΒΜ-ω-ΤΑ: co- -transaminase from Bacillus megaterium;

AD-ω-ΤΑ: co-transaminase from Aicaligenes denitrificans;

Mez-ω-ΤΑ: co- -transaminase from Mesorhizobium;

SP-ω-ΤΑ: ω- -transaminase from Silicibacter pomeroyi; ΟΑ-ω-ΤΑ: ω-transaminase from Ochrobactrum anthropi;

GZ-ω-ΤΑ: ω-transaminase from Gibberella zeae; and

NF-ω-ΤΑ: ω-transaminase from Neosartorya fischeri. Biocatalvtic reductive amination

All experiments were carried out on a one milliliter scale using 2 mL Eppendorf tubes as reaction vessels. All experiments were done in duplicate. Biocatalvtic reductive amination using alanine as amine donor

In the case of ArR-ω-ΤΑ, ΑΤ-ω-ΤΑ, ΗΝ-ω-ΤΑ, GZ-ω-ΤΑ and NF-ω-ΤΑ, D-alanine was used as amine donor. When using the other ω-TAs, L-alanine was applied as amine donor. Lyophilized cells of E. coli containing overexpressed ω-transaminase (20 mg) were rehydrated in sodium phosphate buffer (1 mL, pH 7, 100 mM) containing pyridoxal-5'-phosphate (PLP) (1 mM) and NADH free acid (1 mM) at 30°C and 120 revolutions per minute (rpm) for 30 min. Alanine dehydrogenase (Ala-DH) ( 15 pL, 12 U total activity), formate dehydrogenase (FDH) (5 mg, 11 U total activity), ammonium formate (9.5 mg, 150 mM) and alanine (22.3 mg, 250 mM) as well as the substrate (9 pL, 50 mM) were added. Reductive amination was carried out at 30°C in a thermo shaker (750 rpm) for 24 h. In case of cosolvent studies the appropriate amount of buffer was replaced by the organic solvent. Thereby, the cosolvent was added after rehydration. The reaction was quenched by addition of aqueous NaHC0₃ saturated solution (300 pL) or aqueous NaOH 10 N solution (200 pL). In the case that the reaction was quenched by NaOH, formation of the lactam 3 occurred immediately. After extraction with ethyl acetate (EtOAc) (2 x 500 pL) the combined organic phases were dried over Na₂S0₄ and analyzed via gas chromatography.

Biocatalvtic reductive amination using 2-propylamine as amine donor

Lyophilized cells of E. coli overexpressing ω-transaminase (20 mg cells) were rehydrated in phosphate buffer (800 pL, pH 7, 100 mM) containing PLP (1 mM) at 30°C and 750 rpm for 30 min. Dimethyl sulfoxide (DMSO) or the appropriate organic solvent (e.g. , DMF) (200 pL) and 2-propylamine (50 pL, 500 mM) as well as the substrate (9 pL, 50 mM) were added. Reductive amination was carried out at 45°C in a thermo shaker (750 rpm) for 24 h. The reaction was quenched by addition of NaHC0₃ saturated solution (300 pL) or NaOH 10 N solution (200 pL). When the reaction was quenched by NaOH, formation of the lactam 3 occurred immediately. When NaHC0₃ was applied, the lactam 3 was formed in the organic phase within a few days. After extraction with EtOAc (2 x 500 pL) the combined organic phases were dried over Na₂S0₄ and analyzed via gas chromatography. Biocatalytic reductive amination using a-methylbenzylamine (MBA) as amine donor

Lyophilized cells of E. coli containing overexpressed ω-transaminase (VFmut) (20 mg cells) were rehydrated in phosphate buffer (850 μΙ_, pH 7, 100 mM) containing PLP (0.5 mM) at 30°C and 120 rpm for 30 min. Racemic a-Methylbenzylamine (rac-MBA) (26 pL, 200 mM) and dimethyl sulfoxide (DMSO) ( 50 μΙ_) as well as the substrate (9 pL, 50 mM) were added (Park et al., 2013). Reductive amination was carried out at 37°C in a thermo shaker (750 rpm) for 24 h. The reaction was quenched by addition of aqueous NaOH 10 N solution (200 pL). After extraction with EtOAc (2 x 500 pl_) the combined organic phases were dried over Na₂S0₄ and analysed via gas chromatography.

Example 1 : Synthesis of brivaracetam precursors (compounds 1a - 1d) and formation of optically enriched 1c

Synthesis of compound 1a

Valeraldehyde (3.25 mL, 30 mM) and diisobutylamine (5.22 mL, 30 mM) were dissolved in toluene (120 mL) and refluxed for 8 hours using a Dean Stark apparatus. Molecular sieve was added and the mixture was stirred at room temperature overnight. Molecular sieves were removed by filtration over ceiite and bromo ethyl acetate (5.00 mL, 45 mM) was added. While the reaction was refluxed for 8 h the solution turned brown. Acetate buffer (70 mL; 35 g acetic acid, 35 g sodium acetate, 70 mL water) was added, the suspension was stirred for 1 hour at room temperature and the organic phase was washed with Na₂C0₃ saturated solution. The organic phase was dried over Na₂S0₄ and the solvent was evaporated. The product was purified by column chromatography (silica, eluent: hexane/EtOAc 97:3) to give a brown oil (1.4 g, 8 mmol, 26 % yield overall).

1a

¹H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.93 (t, ³J_y2 = 7.1 Hz, 3 H, 1-H), 1.21 (d, ³J_Me,r = 6.2 Hz, 3 h, /Pr-H), 1.22 (d, ³J_Me,r = 6.2 Hz, 3 H, iPr-H), 1 ,29-1.53 (m, 3 H, 2-H, 3-H_a), 1 ,70 (m_c, 1 H, 3-H_b), 2.36 (dd, ³J_5a,4 = 5.3 Hz, ²J_5a<5b = 16.4 Hz , 1 H, 5_a-H), 2.65 (dd, ³J_5b,4 = 8.1 Hz, ²J_5b|5a = 16.4 Hz , 1 H, 5_b-H), 2.81 (m_c, 1 H, 4-H), 5.00 (m_c, 1 H, 1'-H), 9.70 (s, 1 H, HCO). ¹³C-NMR (75MHz, CDCI₃): δ [ppm] = 14.21 (C-1 ), 20.15 (C-2), 21 .96 (2 C, /Pr), 30.80 (C-3), 33.66 (C-5), 47.74 (C-4), 68.32 (C-1 '), 171.60 (OCO), 203.21 (HCO). Synthesis of compound 1 b

Valeraldehyde (3.25 mL, 30 mM) and diisobutylamine (5.22 mL, 30 mM) were dissolved in toluene (120 mL) and refluxed for 8 hours using a Dean Stark apparatus. Molecular sieve was added and the mixture was stirred at room temperature overnight. Molecular sieve was removed by filtration over celite and isopropyl bromoacetate (5.83 ml, 45mM) was added. While the reaction was refluxed for 8 h the solution turned brown. Acetate buffer (70 mL; 35 g acetic acid, 35 g sodium acetate, 70 ml water) was added, the suspension was stirred for 1 hour at room temperature and the organic phase was washed with Na₂C0₃ saturated solution. The organic phase was dried over Na₂S0₄ and the solvent was evaporated. The product was purified by column chromatography (silica, eluent: hexane/EtOAc 97:3) to give a brown oil (1 .2 g, 6 mmol, 20 % yield overall).

1 b ¹H-NMR (300 MHz, CDCk): 5_H [ppm] = 0.93 (t, ³J_{l i2} = 7.1 Hz, 3 H, 1-H), 1 .21 (d, ³J_{Me r} = 6.2 Hz, 3 h, /Pr-H), 1.22 (d, ³J_Me,r = 6.2 Hz, 3 H, iPr-H), 1 ,29-1 .53 (m, 3 H, 2-H, 3-H_a), 1 ,70 (m_c, 1 H, 3-H_b), 2.36 (dd, ³J₅a_;4 = 5.3 Hz, ²J_5ai5b = 16.4 Hz , 1 H, 5_a-H), 2.65 (dd, ³J_5bA = 8.1 Hz, ²J₅b.5a = 16.4 Hz , 1 H, 5_b-H), 2.81 (m_Cl 1 H, 4-H), 5.00 (m_c, 1 H, t-H). 9.70 (s, 1 H, HCO). ¹³C-NMR (75MHz, CDCI₃): δ [ppm] = 14.21 (C-1 ), 20.15 (C-2), 21.96 (2 C, /Pr), 30.80 (C-3), 33.66 (C-5), 47.74 (C-4), 68.32 (C-1 '), 171.60 (OCO), 203.21 (HCO).

Synthesis of compound 1 c Valeraldehyde (3.25 mL, 30 mM) and diisobutylamine (5.22 mL, 30 mM) were dissolved in toluene (120 mL) and refluxed for 8 hours using a Dean Stark apparatus. Molecular sieve was added and the mixture was stirred at room temperature overnight. The molecular sieve was removed by filtration over celite and i-butyl bromoacetate (6.65 ml, 45 mM) was added. While the reaction was refluxed for 8 h the solution turned brown. Acetate buffer (70 mL; 35 g acetic acid, 35 g sodium acetate, 70 mL water) was added, the suspension was stirred for 1 hour at room temperature and the organic phase was washed with Na₂C0₃ saturated solution. The organic phase was dried over Na₂S0₄ and the solvent was evaporated. The product was purified by column chromatography (silica, eluent: hexane/EtOAc 97:3) to give a brown oil (1 .1 g, 6 mmol, 20 % yield overall).

1c

¹H-NMR (300 MHz, CDCI₃): δ_Η [ppmj = 0.92 (t, ³J_{1 2} = 7.15, 3 H, 1 -H), 1 .27-1 .51 (m, 12 H, 2-H, 3_a-H, iBu-H), 1.68 (m, 1 H, 3_b-H), 2.34 (dd, ³J_5a,4 = 5.3 Hz, ²J_5a,5b = 16.2 Hz , 1 H, 5₈-H), 2.60 (dd, ³J₅b,4 = 8.1 Hz, ²J_5b,5a = 16.2 Hz, 1 H, 5_b-H), 2.75 (m_c, 1 H, 4-H), 9.69 (d, ³J_Hco,4 = 10 Hz, 1 H, HCO).

¹³C-NMR (75MHz, CDCI₃): δ [ppm] = 14.37 (C-1 ), 20.34 (C-2), 28.32 (C-fBu), 30.95 (C-3), 34.85 (C-5), 48.04 (C-4), 81 .29 (C-1 '), 171.49 (OCO), 203.56 (HCO).

In order to prepare one specific enantiomer of the aldehyde compound 1 c, a racemic mixture of this compound was stereoselectively reduced by alcohol dehydrogenase (ADH-A; i.e., alcohol dehydrogenase from Rhodococcus ruber, as described, e.g., in Edegger et al.. 2006) to the corresponding alcohol 7c. Since this reduction proceeds via kinetic resolution, the remaining compound 1 c (which is not reduced by ADH-A) is obtained in enantioenriched form, as illustrated in Scheme 1 below, and can hence be used for racemization studies.

Specifically, ADH-A (20 μΙ_, 0.3 U) and compound 1 c (9 μΙ_, 50 mM) were added to sodium phosphate buffer (1 mL, 100 mM, pH 7) containing NADH (1 mM) and isopropanol (30 μί). The reduction was performed at 30°C and 700 rpm for 15 to 20 min. After extraction with EtOAc (2 x 500 μί) the combined organic phases were dried over Na₂S0₄ and analysed via

GC.

ADH-A

Scheme 1 : Preparation of compound 1 c in enantioenriched form via kinetic resolution by ADH-A reduction.

Synthesis of compound 1d

Valeraldehyde (3.25 mL, 30 mM) and diisobutylamine (5.22 mL, 30 mM) were dissolved in toluene (120 mL) and refluxed for 8 hours using a Dean Stark apparatus. Molecular sieve was added and the mixture was stirred at room temperature (RT) overnight. Molecular sieve was removed by filtration over celite and bromo benzyl acetate (7.15 mL, 10.308 g, 45 mM) was added. While the reaction was refluxed for 8 h the solution turned brown. Acetate buffer (70 mL; 35 g acetic acid , 35 g sodium acetate, 70 mL water) was added, the suspension was stirred for 1 hour at room temperature and the organic phase was washed with Na₂C0₃ saturated solution. The organic phase was dried over Na₂S0₄ and the solvent was evaporated. The product was purified by column chromatography (silica, eluent: hexane/EtOAc 97:3) to give a brown oil (1 .4 g, 6 mmol, 20 % yield overall) .

1 d Ή- MR (300 MHz, CDC ) : 5_H [ppm] = 0.90 (t, ³J_1>2 = 7.2 Hz, 3 H, 1 -H), 1 .55 - 1 .24 (m, 3 H, 2-H, 3-H_a) , 1 .70 (m_c, 1 H, 3-H_b), 2.38-2.51 (dd, ³J_5a.4 = 4.9 Hz, ²J_5a>5b = 16.3 Hz, 1 H, 5-H_a), 2.68-2.81 (dd, ³J₅ = 8.1 Hz, ²J_5bfia - 16.3 Hz, 1 H, 5-H_b) , 2.85 (m_c, 1 H, 4-H), 5.1 (s, 2 H, 1 '-H), 7.35 (m_c, 5 H, Ph-H) , 9.71 (d , ³J_6i4 = 0.9 Hz, 1 H, 6-H). ¹³C-NMR (75 M Hz. CDCI₃) : δ [ppm] = 14 03 ( 1 -C), 20 01 (2-C). 30 66 (3-C), 33 02 (5-C) , 47.54 (4-C), 66.61 (1 '-C) , 128.28 (Ph-C), 128.32 (Ph-C), 128.59 (Ph-C) , 135.70 (Ph-C) , 1 71 .88 (0-C=0), 202.91 (6-C).

Synthesis of optically enriched lactam (R)-Z as reference compound

I n order to identify (R)- and (S)-enantiomers of the lactam 3, an optically enriched reference compound was synthesized via a 2-step chemoenzymatic method employing an ene-reductase (EBP 1 ) (Winkler et al. , 201 3; Schwarz et al. , 2005), as illustrated in Scheme 2 below: Tris-HCI buffer CN Methanol, CoCI₂ ^* 6 _ZO ^ β ₅pont. f ₀

^■ ₀" EBP1. NADH ^ ^"^ ^Q^ _NaBH₄ ^ ^ ^v O

(R)-2 (R)-3

(Il ia)

Scheme 2: Synthesis of an optically enriched reference compound, i.e. the (R)-enantiomer of lactam 3.

Step 1

An aliquot of ene reductase (EBP1 , 75 μΙ) was added into a 1 ,5 ml Eppendof tube containing a Tris-HCI buffer solution (0.8 ml, 50 mM, pH 7.5), the substrate (2 μΙ, 16.3 mM) and the cofactor NADH (15 mM). The mixture was shaken at 30°C and 450 rpm. After 24 hours, the product was extracted with EtOAc (2x 600 μ!_). The organic phase was dried over Na₂S0₄ and the solvent was evaporated (Winkler et al., 2013). This step was repeated (48 times) and the products were combined until an amount of 88 mg of crude product was obtained.

Step 2

(Ilia)

Cobalt chloride hexahydrate (333 mg, 1.4 mmol) was added to a solution of the crude product of step 1 (88 mg, 0.6 mmol) in methanol (10 ml) to give a deep purple colored solution. Sodium borohydride (52.8 mg, 1.4 mmol) was added portionwise over 10 minutes with caution to control the evolution of hydrogen and the exothermic reaction that ensued to give a black solution. The reaction mixture was stirred for 30 minutes under nitrogen atmosphere and then quenched carefully by addition of 10 N HCI. Concentrated NH₄OH was added until pH ~9 was reached. The mixture was extracted with ethyl acetate (4 x 15 ml). The combined organic phases were dried over Na₂S0₄ and concentrated to afford 20 mg of crude reference compound {R)-Z (Schwarz et al. , 2005). Example 2: Synthesis of a pregabalin precursor (compound 10c)

The aldehyde compound 10c can be used as a starting material in the process for the production of pregabalin according to the present invention. Compound 10c was prepared from the aldehyde 8 as shown in the following Scheme 3:

Scheme 3: Synthesis of compound 10c.

Since the aldehyde 8 is not commercially available, an ADH-hT (Cannio et al., 1994) catalyzed synthesis of compound 8 from 4-methyl-1 -pentanol (11 ) was established on a one milliliter scale, as illustrated in Scheme 4:

PCC

Aeeialdehvde .

Phosphate

buffer pH 8

Scheme 4: Synthesis of compound 8 by either ADH-hT catalyzed oxidation or PCC oxidation of 4-methyl-1 -pentanol

On one milliliter scale 60% conversion was obtained. This process can be used to prepare the starting material for the synthesis of pregabalin and pregabalin precursors.

4-Methyl-1 -pentanol (7 μΐ_, 50 mM), ADH-hT (25 pL strep-tag purified, see: Angew. Chem. Int. Ed. 2012, 51, 9156-9159) and acetaldehyde (25 μ|_, 500 mM) were added to sodium phosphate buffer (1 ml_, 100 mM, pH 8) containing NAD free acid (1 mM). Biotransformation was performed at 50°C and 700 rpm in a thermo mixer for 24 h. Samples were extracted in EtOAc (2 x 500 ml) , organic phases were dried over Na₂S0₄ and analysed via GC-MS. Synthesis of pregabalin percursor 10c via Nef-reaction and transesterification

In order to synthesize pregabalin precursor 10c as a substrate for ω-transaminase-catalyzed bioamination, the first intent was to prepare aldehyde 10e starting from nitro compound 14 via Nef-reaction, and then to transesterify methyl ester 10e to i-butyl ester 10c (see Scheme 5 below). The i-butyl ester 10c was desired since in case of a similar substrate (brivaracetam precursor), the highest stereoselectivity of ω-TAs in the transamination was observed with the -butyl ester c (see Example 4).

Scheme 5: Synthesis of pregabalin precursor 10c via Nef-reaction and transesterification.

Nef-reaction (see Scheme 5 above)

Sodium methoxide (NaOMe) (600 mg, 1 1.1 mmol) was dissolved in dry methanol (25 mL) on ice, ethyl 5-methyl-3-(nitromethyl)hexanoate 14 (1000 mg, 4.6 mmol) was added dropwise, ice bath was removed and the mixture was stirred for 2 h. The resultant suspension was added to a solution of concentrated H₂S0₄ (1 .1 mL) in dry methanol (25 mL) at -10°C. The reaction was continued for 20 min and then the milky suspension was poored onto CH₂CI₂ / water 50:50, the aqueous phase was basified with NaHC0₃ saturated solution and extracted 4 times with CH₂CI₂. The combined organic phases were dried over Na₂S0₄ and the solvent was evaporated. The resulting acetal 5 was identified via GC-MS analysis.

Acetal 15 was added to water (25 mL), refluxed for 5 h and the desired product 10e was isolated by basic extraction in CH₂CI₂. The combined organic phases were dried over Na₂SCX[ and the solvent was evaporated to give a yellow oil (Simoneau et al., 1988).

In case of ethyl ester 10a formed as by product (ca. 50 %), the two products were isolated by column chromatography (silica, hexanes/EtOAc 95:5). In case of remaining nitro compound 16 (ca. 50%), it was partly removed from the desired product by kugelrohr distillation. Otherwise the crude product was applied for transesterification.

Despite intermediate formation of the side products 16 and 10a the Nef reaction worked smoothly. Several attempts to transform the methyl ester 10e into f-butyl ester 10c were made (see Scheme 6 below). However, aldehyde 10e decomposed when reacted with sulfuric acid or potassium f-butoxide and remained unreacted when treated with ion exchanger Amberlite IR120 or Lipase B from Candida antarctica (CalB). f-BuOH

H₂S0₄

f-BuOH

f-BuOH

Cal-B

Scheme 6; Attempts of transesterification of the methyl ester 10e into the f-butyl ester 10c. Since the transesterification of methyl ester 10e was not successful, the pregabalin precursor 10c was synthesized in a similar fashion as the brivaracetam precursor 1 c (see Example 1 ). For this purpose alcohol 11 was oxidized to aldehyde 8 which was not commercially available, employing pyridinium chlorochromate. Due to the high volatility of aldehyde 8, it was isolated as enamine 9 which was further transformed to pregabalin precursor 10c in the same way as the brivaracetam precursor 1 c (see Example 1 ), as also shown in Scheme 7 below. This approach worked smoothly and aldehyde 10c was obtained in 20% overall yield.

Scheme 7: Synthesis of pregabalin precursor 10c (MS 3A = molecular sieve 3A).

NMR

Methyl 3-formyl-5-methylhexanoate (10e)

H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.93 (d, ³ _5,6 = 6.4 Hz, 3 H, -CH₃), 0.96 (d, ³J_5,6. = 6.6 Hz, 3 H, -CH₃), 1.27 (m_c, 1 H, -CH(CH₃)₂), 1.55-1.76 (m, 2 H, 4-H), 2.42 (dd,

5.0 Hz,

16.6 Hz, 1 H, 2-H ), 2.88 (m_c, 1 H, 3-H), 3.69 (s, 3 H, -OCH₃), 9.71 (d, ³J_CHO,₃ = 1.1 Hz, 1 H, -CHO) .

¹³C-NMR (75MHz, CDCI₃): 6_C [ppm] = 22.41 (-CH₃), 22.86 (-CH₃), 25.75 (-CH(CH₃)₂), 33.43(2- C), 37.84 (4-C), 45.99 (3-C), 52.00 (-OCH₃), 172.54 (OC=0), 203.25 (C=0).

Ethyl 3-formyl-5-methylhexanoate (10a)

¹H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.93 (d, = 6.7 Hz, 3 H, -CH₃), 0.96 (d, ³J_5,6. = 6.9 Hz, 3 H, -CH₃), 1.26 (s, 3 H, -0CH₂CH₃), 1.21-1.35 (m, 1 H, -CH(CH₃)₂), 1.55-1.77 (m, 2 H, 4- H), 2.40 (dd, ³J_2a,3= 5.1 Hz, ²J_2a,2b= 16.6 Hz, 1 H, 2-H_a), 2.68 (dd, ³J_2b,3= 8.3 Hz, ²J_2b,2a= 16.5 Hz, 1 H, 2-H_b), 2.88 (m_c, 1 H, 3-H), 4.14 (q, ³J _,2.= 7.1 Hz 2 H, -0CH₂CH₃), 9.71 (d, ³J_CHO.₃ = 1 -2 Hz, 1 H, -CHO).

¹³C-NMR (75MHz, CDCI₃): 5_C [ppm] = 14.32(2'-C), 22.55(-CH₃), 23.49(-CH₃), 25.75 (-CH(CH₃)₂), 33.79 (2-C), 37.85 (4-C), 46.02 (3-C), 60.91 (1 '-C), 172.06 (OC=0), 203.33 (C=0).

Methyl 5-meihyl-3-(nitromethyl)hexanoate (16)

H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.93 (d. ³J_{5 5} = 6.2 Hz, 3 H, -CH₃), 0.96 (d, ³ ₅.6^> = 7.5 Hz, 3 H, -CH₃), 1 .27 (m_c, 1 H, -CH(CH₃)₂), 1 .55- 1 .77 (m, 2 H, 4-H), 2.45 (d, ³J_2,3= 6.4 Hz, 2 H, 2-H), 2.60-2.75 (m, 1 H, 3-H), 3.68 (s, 3 H, -OCH₃), 4.44 (dd, ³J_NCH₃,3 = 5.7 Hz, J_NCHa.Hb = 12.3 Hz, 1 H, -CH_aN0₂), 4.51 (dd, ²J_{N C}Hb.z = 6.6 Hz, ²J_NCHb,Ha = 12.3 Hz, 1 H, -CH_bN0₂).

¹³C-NMR (75MHz, CDCI₃): S_c [ppm] = 22.39(-CH₃), 23.63(-CH₃), 25.18 (3-C), 32.12 (-CH(CH₃)₂), 35.89 (2-C), 40.65 (4-C), 52.00 (-OCH₃), 78.82 (-CH₂N0₂), 172.13 (OC=0). 4.) Transesterification Method 1

Methyl ester 10e (50 mg, 0.26 mmol) was dissolved in f-butanol (2 ml_), cone, sulfuric acid (5 μΙ_) was added and the mixture was stirred at 40 °C overnight. Aldehyde 10e was degraded and no i-butyl ester 10c was found.

Method 2

Methyl ester l Oe (57 mg, 0.33 mmol) was dissolved In i-butanol (10 ml_) containing ion exchanger Amberlite IR120 (35 mg). The mixture was stirred at 40°C. After 10 days still no conversion of aldehyde 10e was observed.

Method 3

Potassium i-butoxide was dissolved In f-butyl alcohol (5 mL) and methyl ester 10e (52 mg, 0.30 mmol) was added. The reaction was run for 100 min, CH₂CI₂ and NH₄CI sat. solution was added. The aquous phase was extracted 4 x with in CH₂CI₂, the combined organic phases were dried over Na₂S0₄ and the solvent was evaporated to give a residue containing of oil and solid (Ueno et al., 991). Aldehyde 10e was degraded and no i-butyl ester 10c was found.

Method 4

Methyl ester 10e (57 mg, 0.33 mmol) was dissolved in i-butanol (10 mL) and CalB (200 mg) was added. The mixture was stirred at 40°C. After 7 days still no conversion was observed. PCC-oxidation of 4-methyl-1 -pentanol

Pyridinium chloro chromate (9.36 g, 43.4 mmol), NaOAc (0.36 g, 4.4 mmol) and neutral alumina as adsorbent (47 g) were suspended in dry CH₂CI₂ (90 mL), cooled to 0°C on ice and 4-methyl-1 -pentanol 11 (2.2 g, 21.7 mmol) was added. The ice bath was removed and the reaction was run for 4 h. The mixture was filtered through a 1 cm layer of silica and the filter cake was washed with CH₂CI₂ (50 mL) (Schmidt et al. , 201 1 ). Due to the high volatility of aldehyde 8 it was isolated as enamine 9. After addition of diisobutylamine (2.71 g, 21 .0 mmol) and molecular sieves (3 A, 2 g) the mixture was stirred for 27 h, the molecular sieves were removed by filtration through Celite, filter cake was washed with CH₂CI₂, and the solvent was evaporated to give a slightly yellowish oil (4.123 g). The crude product 9 was employed for the synthesis of pregabalin precursor 10c. Synthesis of pregabalin precursor 10c

The crude product 9 (4 g) was dissolved in toluene (100 mL) and i-butyl bromoacetate (4.80 mL, 6.34 g, 32 mM) was added. The reaction was refluxed for 5 h and the solution turned brown. The mixture was cooled to room temperature, acetate buffer (70 mL; 35 g acetic acid, 35 g sodium acetate, 70 mL water) was added, the suspension was stirred for 1 hour at ambient temperature and the organic phase was washed with Na₂C0₃ saturated solution. The organic phase was dried over Na₂S0₄ and the solvent was evaporated to give a dark brown oil containing also a salt (10 g) (WO 2005/027856; Hodgson et al. , 2009). The product was purified by column chromatography (silica, eluent: hexane/EtOAc 96:4) to give a yellow oil (850 mg , 4 mmol, 20 % yield overall).

¹H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.94 (t, ³J_5,6 = 6.5 Hz, 6 H, -CH₃), 1.25 (m_c, 1 H, -CW(CH₃)₂), 1.44 (s, 9 H , -OC(CH₃)₃, 1.54-1.76 (m, 2 H, 4-H), 2.53 (dd, ³J_2a,3= 5.1 Hz, J_2a,2b= 16.3 Hz, 1 H, 2-H_a), 2.59 (dd, ³J_2b,3= 8.1 Hz, ²J_2b,2a= 16.3 Hz, 1 H, 2~H_b), 2.80 (m_c, 1 H, 3-H), 9.70 (dd, ³JCHO,3 = 1.4 Hz, 1 H, -CHO). ¹³C-NMR (75MHz, CDCI₃): 5_C [ppm] = 22.52 (-CH₃), 22.86 (-CH₃), 25.73 (-CH(CH₃)₂), 28.15 (-OC(CH₃)₃), 35.24 (2-C), 37.80 (4-C), 46.25 (3-C), 81.17 (-OC(CH₃)₃) , 171 .24(OC=0), 203.53 (C=0).

Synthesis of reference compounds rac 13 and (S)-13 rac 4-lsobutylpyrrolidin-2-one (rac 13)

rac Pregabalin (50 mg, 0.3 mmol) was disolved in MeOH (5 mL) and trimethylsilyl- diazomethane (TMS-diazomethane) (400 mL, 2 M solution in hexane) was added dropwise until the solution stayed yellow. Then the solution was stirred for 30 min and the solvent was removed under vacuum to give a brown oil (39 mg, 0.3 mmol).

(S)-4-lsobutylpyrrolidin-2-one ((S)-13)

(S)-Pregabalin (50 mg, 0.3 mmol) was disolved in MeOH (5 mL) and TMS-diazomethane (900 mL, 2 M solution in hexane) was added dropwise until the solution stayed yellow. Then the solution was stirred for 30 min and the solvent was removed under vacuum to give a brown oil (50 mg). The products (not pure) were identified by GC-MS analysis.

Example 3: Molecular biological and microbiological methods

Preparation of variants of ArRmutl 1- and GZ-ω-ΤΑ Primers

The applied primers were designed using the Agilent QuickChange primer design tool:

Variant Primer Mutation Method for quick change mutagenesis

ArRmutl 1 - ctgggttgtagcacacttggtggtgtttggccgtttg V279L A

M1 17F-A60V- and

G279L-to-TA caaacggccaaacaccaccaagtgtgctacaacccag

ArRmutl 1 - ctgggttgtagcacaattggtggtgtttggccgtttg V279I A

M1 17F-A60V- and

G279I-U)-TA caaacggccaaacaccaccaattgtgctacaacccag

ArRmutl 1 - ctgggttgiagcacatttggtggigtttggccgtttg V279F A

M1 17F-A60V- and

G279F-UJ-TA caaacggccaaacaccaccaaatgtgctacaacccag

ArRmutl 1 - cagcgacgttacctataccgtctttcatgtgtggaatggt T64V A

M1 17F-A60V- and

G279V-T64V- accattccacacatgaaagacggtataggtaacgtcgctg

ω-ΤΑ

ArRmutl 1 - cgaactgcgtgaagcatttgttgaggttaccattacccgtggttata T1 19E A

M1 17F-A60V- and

G279V-T1 19E- tataaccacgggtaatggtaacctcaacaaatgcttcacgcagttcg

ω-ΤΑ

ArRmutl 1 - catgtccgtatcagtgggttgttccgtttgatcgt 1152V A

M1 17F-A6QV- and

G 279 V- 1152 V- acgatcaaacggaacaacccactgatacggacatg

ω-ΤΑ ArRmutH- atgaagttctgggttgtaccacagttggtggtgt S277T A M1 17F-A60V- and

G279V-S277T- acaccaccaactgtggtacaacccagaacttcat

ω-ΤΑ

ArRmutH- gatgaagttctgggttgtaacacagttggtggtgtt S277N A M1 17F-A60V- and

G279V- aacaccaccaactgtgttacaacccagaacttcatc

S277N-0J-TA

GZ-L56V-U)-TA tttatgcatg g eg atgtg acctatg atgttccggca L56V A and

tgccggaacatcataggtcacatcgccatgcataaa

GZ-Y58L-OJ-TA caaactgccggaacatctaaggtcagatcgccatgc Y58L B and

g catggcg atctg accttag atgttccggcagtttg

GZ-V60T- g g eg atctg acctatg ataccccg g cag tttg g g atg gt V60T A ω-ΤΑ and

accatcccaaactgccggggtatcataggtcagatc

GZ-E115T- ggtattaaagatgcctttgttaccctgattgttacccgtggtctg E1 15T A ω-ΤΑ and

cagaccacgggtaacaatcagggtaacaaaggcatctttaatacc

GZ-E115D- gaccacgggtaacaatcagatcaacaaaggcatctttaa E1 15D B ω-ΤΑ and

tt a aag atg cctttgttg atctg attg ttacccg tg g tc

GZ-V148I- gtgcagccgtatgtttggattatgagtccggaagcacag V148! A ω-ΤΑ and

ctgtgcttccggactcataacccaaacatacggctgcac

GZ-V148A- tgcttccggactcatagcccaaacatacggctg V148A A ω-ΤΑ and

cagccgtatgtttgggctatgagtccggaagca

GZ-T273S- ggccgatgaaatctttatgtgtagcaccgcaggcgg T273S B ω-ΤΑ and

ccgcctgcggtgctacacataaagatttcatcggc

GZ-T273N- gccgatgaaatctttatgtgtaacaccgcaggcgg T273N B ω-ΤΑ and

ccgcctgcggtgttacacataaagatttcatcggc

GZ-A275V- ctttatgtgtaccaccgtaggcggtattatgccga A275V B ω-ΤΑ and

tcggcataataccgcctacggtggtacacataaag

GZ-L56V- ctttatgtgtaccaccgtaggcggtattatgccga A275V B

Α275ν-ω-ΤΑ and

tcggcataataccgcctacggtggtacacataaag

GZ-L56V- ggccgatgaaatctttatgtgtagcaccgcaggcgg T273S B

T273S-U)-TA and

ccgcctgcggtgctacacataaagatttcatcggc

GZ-T273S- ctttatgtgtagcaccgtaggcggtattatgccga A275V B

Α275ν-ω-ΤΑ and

tcggcataataccgcctacggtgctacacataaag

Table 1 : Primers used for construction of variants.

Quick change mutagenesis

Method A - Agilent QuickChange kit

Mutagenesis and following amplification of the mutated plasrnid in E. coli XL 10-gold ultracompetent cells was performed according to the manual of the Agilent QuickChange Lightning Side directed Mutagenesis kit. Successful mutation was confirmed by plasrnid sequencing (LGC Genomics, Berlin, Germany) and transaminase variants were overexpressed in E. coli BL21 (DE3).

Method B - without QuickChange kit

Mutant Strand Synthesis Reactions

PCR mixtures were prepared as given in Table 2. For each mutation two PCR tubes were prepared, one with the forward primer and one with the reverse primer. Moreover for each dsDNA template a Dpnl control sample was prepared without any primer to confirm mutagenesis. In case of the mutation A275V 6 vol-% DMSO was added to avoid hairpin formation of the primers. The two PCR steps were conducted as given in Table 3 and Table 4.

Compound Amount

polymerase buffer 5 pL

dsDNA template 0.5 pL

dNTP mix 2.5 pL Phusion DNA polymerase 0.25 pL

primer 1 pL

H₂0³ 14.75 μί.

Total amount 25 pL

Table 2: Preparation of PCR samples.

a in case of mutation A275V 6 vol-% DMSO was added. Step 1 : linear PCR

Table 3: Temperature program for linear PCR.

3 S0°C was applied in case of primers for A275V.

After Step 1 the content of tube 1 and tube 2 was combined and step 2 was performed.

Step 2: Exponential PCR

Table 4: Temperature program for exponential PCR

3 60°C was applied in case of primers for A275V. Dpn\ digestion of the amplification products

Samples for the Dpnl digest were prepared as given in Table 5, the digestion was conducted at 37°C for 7 h and the product was then heated to 80°C to denature the Dpnl enzyme.

Successful mutation was confirmed by plasmid sequencing (LGC Genomics, Berlin, Germany) and transaminase variants were overexpressed in £. coli BL21 (DE3).

Table 5: Preparation of Dpnl digest

Successful mutation was confirmed by plasmid sequencing (LGC Genomics, Berlin, Germany) and transaminase variants were overexpressed in E. coli BL21 (DE3).

Transformation of E. coli competent cells

To a shot of E. coli BL21 (DE3) competent cells the piasmide (10-100 ng) was added and the cells were incubated on ice for 30 min. Then the cells were heat-shocked at 42°C for 30 sec, preheated SOC medium (250 pL) was added and incubated at 37°C and 350 rpm for 1 h. The cells were plated and incubated at 37°C overnight. Overexpression of transaminases

For the preculture one colony was inoculated in 10 mL of LB medium containing ampicillin (100 mg/mL). The overnight culture (ONC) was incubated at 30°C with shaking at 120 rpm overnight. 2-5 mL of the ONC were inoculated in 330 mL LB medium containing the antibiotic ampicillin. After growing at 30°C and 120 rpm until OD₆₀₀ = 0.6 - 0.8 and induction (respective inducers are outlined in Table 6) the cultures were shaken at 120 rpm and 20°C overnight.

Cells were harvested by centrifugation (20 min, 8000 rpm), washed with phosphate buffer (100 mM, pH 7, 0.5 mM PLP), resuspended in phosphate buffer (100 mM, pH 7, 0.5 mM PLP) and lyophilized. ωΤΑ Origin Original vector Inducer

ArR-ω-ΤΑ Arthrobacter sp. KNK168 pET2 a IPTG

ArRmutH -ω-ΤΑ Variant of ArR-ω-ΤΑ pET21 a IPTG

Variants of ArRmutH- Variants of ArRmutl 1 -ω-ΤΑ pET21 a IPTG ω-ΤΑ

ΒΜ-ω-ΤΑ Bacillus megaterium pET21 a IPTG

AD-ω-ΤΑ Alcaligenes denitrificans pASK-IBA35+ ATC

CV-ω-ΤΑ Chromobacterium violaceum pET21 a IPTG

PD-ω-ΤΑ Paracoccus denitrificans pASK-!BA35+ ATC

ΡΡ1 -ω-ΤΑ Pseudomonas putida KT2440 pASK-IBA35+ ATC

ΡΡ2-ω-ΤΑ Pseudomonas putida KT2440 PASK-IBA35+ ATC

VF-ω-ΤΑ Vibrio fluvialis PASK-IBA35+ ATC

ArS-ω-ΤΑ Arthrobacter citreus pASK-IBA35+ ATC

RE-ω-ΤΑ Ralstonia eutropha pET21 a IPTG

VP-ω-ΤΑ Variovorax paradoxus pET21 a IPTG

ΑΤ-ω-ΤΑ Aspergillus terreus pET21 a IPTG

ΗΝ-ω-ΤΑ Hyphomonas neptunium pET21 a IPTG

PF-ω-ΤΑ Pseudomonas fluorescens pET21 a IPTG

SP-ω-ΤΑ Silicibacter pomeroyi pET21 a IPTG

ΟΑ-ω-ΤΑ Ochrobactrum anthropi pET21 a IPTG

NF-ω-ΤΑ Neosartorya fischeri pET21 a IPTG

GZ-ω-ΤΑ Gibberella zeae pET21 a IPTG

Variants of GZ-ω-ΤΑ Variants of GZ-ω-ΤΑ pET21 a IPTG

Table 6: Vectors used for subcloning of transaminases and inducers employed for overexpression. IPTG = Isopropyl-^-D-thiogalactopyranosid; ATC = anhydrotetracycline.

Example 4: Stereoselective enzymatic synthesis of a brivaracetam precursor using different ω-transaminases The aldehyde compounds 1 a to 1 d were biocatalytically transformed into the corresponding amines 2a to 2d using different ω-transaminases via dynamic kinetic resolution. These substrates were converted by all tested enzymes with high conversions and high stereoselectivity (up to 92% ee by ω-transaminase from Hyphomonas neptunium). The amines 2a to 2d furthermore formed the corresponding lactam 3, as illustrated in Scheme 8 below.

1a: R = ethyi 2a: R

1b: R = isopropyl 2b: R isopropyl

1c: = iett-butyi 2c: R tert-butyl

1 d: R = enzyl 2d: R benzyl

Scheme 8: Biocatalytic reductive amination of compounds 1a to 1d to compounds 2a to 2d, respectively, and conversion into the brivaracetam precursor 3.

The compounds 1a, 1c and 1 d were also reductiveiy aminated using different muteins/variants of ArR-ω-ΤΑ, as shown in Scheme 9 below. In these experiments, 20 vol-% of DMSO was added to the reaction medium as a cosolvent, 2-propylamine was used as amine donor, and the transformations were carried out at 45°C.

1a R = ethyi 2a R = ethyi

1c R = tert-butyl 2c R = tert-butyl

1 d R = benzy! 2d R = benzyl

Scheme 9: Biocatalytic reductive amination of compounds 1a, 1c and 1d to compounds 2a, 2c and 2d, respectively, and conversion into the brivaracetam precursor 3

The reductive amination of compound 1 c (and of the reference substrate acetophenone) was furthermore tested using the ω-transaminases SP-ω-ΤΑ, ΟΑ-ω-ΤΑ, GZ-ω-ΤΑ and NF-ω-ΤΑ which had been overexpressed in £ coli BL21 (DE 3). Results

The compounds 1a to 1d were aminated using different ω-transaminases originating from various microorganisms. In most cases the substrates 1a to 1d were transformed with high to full conversion, as shown in Tables 6 and 7 below. Subsequently, the amines 2a to 2d formed the lactam 3, when NaOH was used for the workup. When NaHC0₃ was applied, the ring closure was much slower taking a few days. An influence of the ester groups on the stereoselectivity of the ω-TAs can be observed. For example, the ee measured for the lactam 3 was highest when the ferf-butyl ester compound 1c or the benzyl ester compound 1 d were applied as substrate. The best enantiopreference was gained with ΗΝ-ω-ΤΑ (i.e., 92% ee). Interestingly, the stereopreference of ArR-ω-ΤΑ was inverted for substrates having different ester groups: This enzyme formed the other enantiomer of compound 2a than of compounds 2b to 2d which carry more sterically demanding ester moieties. Furthermore, dynamic kinetic resolution has taken place in the experiments since the products show significantly high enantiomeric excess even at full conversion.

Table 7: Reductive amination of compounds 1a to 1 d to the corresponding amines 2a to 2d, respectively. The conversion (conv.) and the optical purity of the resulting amines (enantiomeric excess (ee) of either the (R)-isomer or the (S)-isomer) are shown, rac = racemic; n.d. = not determined. As described above, the compounds 1 a, 1 c and 1 d were also reductively aminated using different muteins/variants of ArR-ω-ΤΑ (see Scheme 9). Under the conditions in these experiments (DMSO as cosolvent, 2-propylamine as amine donor, 45°C), unreacted aldehyde 1c was found to form the corresponding imine 4c with 2-propylamine.

The compounds 1 a, 1 c and 1 d were transformed into the corresponding amines 2a, 2c and 2d, respectively, by most of the tested ArR-ω-ΤΑ muteins/variants, as indicated in Table 8 below. It can be seen that the enantiopreference of most of these enzymes is contrary to the one of the majority of the ω-TAs shown in Table 7. ArR-ω-ΤΑ (see Table 7) provided the same enantiomer of the lactam 3 as most of its variants (see Table 8). Nevertheless, the ArR-ω-ΤΑ variants ArR-G13 F-oo-TA and ArRmutl 1 -M1 17F-G279A- TA provided the opposite enantiomer of the lactam 3 than all other ArR-ω-ΤΑ variants tested.

Table 8: Reductive amination of compounds 1 a, 1 c and 1 d to the corresponding amines 2a, 2c and 2d, respectively, using different muteins/variants of ArR-ω-ΤΑ. The conversion (conv.) and the optical purity of the resulting amines (enantiomeric excess (ee) of either the (R)-isomer or the (S)-isomer) are shown, rac = racemic; n.d. = not determined. It has thus been found that the amino acid residue in position 279 of the ω-transaminase ArRmutl 1 -ω-ΤΑ has a high impact on the enantioselectivity of this enzyme for the compounds of formula (I). Hence, increasing the size of the amino acid residue in position 279 from glycine to valine and thereby reducing the size of a binding pocket in this enzyme dramatically increased the enantiomeric excess in which the (S)-isomer of the corresponding amine was obtained. A further enlargement of the amino acid residue in position 279 to leucine, isoleucine or phenylalanine, however, reduced the conversion rate of the corresponding substrates and did not provide a significant increase in stereoselectivity. Furthermore, the ω-TAs from Si!icibacter pomeroyi, Ochrobactrum anthropi, Gibberel!a zeae and Neosartorya fischeri were found to transform the reference substrate acetophenone to 1 -phenyiethy!amine with perfect stereoselectivity. SP-ω-ΤΑ and ΟΑ-ω-ΤΑ produced (S)- phenylethylamine whereas GZ-ω-ΤΑ and NF-ω-ΤΑ were (R)-selective. as indicated in the literature. Compound 1 c was also converted stereoselectively to the corresponding amine 2c by these enzymes and further formed the lactam 3. The (R)-enantiomer of the product was formed in 52% ee using ΟΑ-ω-ΤΑ, and the (S)-enantiomer was formed in 55% ee using GZ-ω- TA, as also shown in the following Table 9:

Table 9: Reductive amination of compound 1c and the reference substrate acetophenone to the corresponding amine 2c and 1 -methylbenzylamine (MBA), respectively, using ω-TAs from Silicibacter pomeroyi, Ochrobactrum anthropi, Gibberella zeae and Neosartorya fischeri. The conversion rate (conv.) and the optical purity of the resulting amines (enantiomeric excess (ee) of either the ( )-isomer or the (S)-isomer) are shown.

It has thus been demonstrated that both enantiomers of the lactam 3 can be produced efficiently and with high enantioselectivity from the aldehyde compounds of formula (I) using either (R)-selective or (S)-selective ω-transaminases. In particular, the (R)-enantiomer of the lactam 3 could be produced from compound 1 c with 92% ee using ΗΝ-ω-ΤΑ, and the (S)-enantiomer of the lactam 3 could be produced from compound 1c with 64% ee using ArRmutl 1- M1 17F-G279V-A60V-0J-TA.

Example 5: Examination of racemization of compounds of formula (I) with ω-ΤΑ- transfected E. coli cell lysate and with the ω-transaminase reaction system

Brivaracetam- and pregabalin precursors such as compounds of formula (I) were aminated with quantitative conversion and high ee values. This strongly indicates that the reaction proceeds via dynamic kinetic resolution, caused by racemisation of the substrate aldehydes. Notably, the racemisation of these compounds is not obvious as in case of 2-phenylpropionaldehyde derivatives (Fuchs et al. , 2014) since the chiral centre of aldehydes employed here is not in benzyiic position. Hence the racemisation of these substrates in the presence of various components of the transaminase reaction system was investigated based on brivaracetam precursor 1c.

Preparation of optically enriched a -chiral aldehyde

Enantioenriched substrate 1 c was prepared by the use of alcohol dehydrogenase ADH-A from Rhodococcus ruber (overexpressed in E. coli and partially purified by heat pulse) (Karabec et al., 2010).

This enzyme reduces the aldehyde 1 c to the corresponding alcohol 7c by kinetic resolution meaning that one enantiomer is transformed and the other one remains optically enriched. Nevertheless ADH-A showed low enantioselectivity and after transforming one enantiomer also the other one was converted. Therefore the reaction had to be stopped in the optimal case at 50% conversion. Hence depending on the actual reaction time, the starting ee for each experiment was determined.

ADH-A

Scheme 10: ADH mediated preparation of enantioenriched brivaracetam precursor 1 c. Racemisation study

Enantioenriched aldehyde 1 c was incubated in phosphate buffer containing various compounds of the transaminase reaction system and samples were analysed after certain time periods to detect possible changes of the ee value.

1 c racAc

Scheme 11 : Racemisation of enantioenriched aldehyde 1 c.

First aldehyde 1 c was incubated in pure sodium phosphate buffer pH 7, which is the medium for transamination reactions. As expected, no spontaneous racemisation of the substrate was observed (Table 10). On the other hand, when lyophilised cells of E. coli BL21 (DE3) (host) were added, the ee dropped dramatically (72 % ee to 27% ee in 2 h, Table 1 1 ). entry time (min) initial ee 1c (%) final ee 1c (%) Aee 1 c (%)

1 1 72 64 8

2 10 72 64 8

3 30 64 60 4

4 60 72 65 7

5 120 68 60 8

Table 10 : Racemisation of aldehyde 1 c in phosphate buffer pH 7. entry time (min) initial ee 1c (%) final ee 1c (%) Aee c (%)

1 1 66 60 6

2 10 66 58 8

3 30 66 50 16

4 60 73 40 33

5 120 72 27 45

Table 11 : Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing lyophilized cells of the E. coli host.

When enantioenriched substrate 1c was incubated in phosphate buffer containing the cofactor PMP no significant influence on the ee was detected (Table 12). Also when PLP, the non aminated form of the cofactor, was present, racemisation of aldehyde 1c did not occur (see Table 16). entry time (min) initial ee 1 c (%) final ee 1 c (%) Aee c (%)

1 1 70 64 6

2 10 74 66 8

3 30 74 66 8

4 60 70 62 8

5 120 72 62 10

Table 12: Racemisation of aldehyde 1c in phosphate buffer pH 7 containing PMP (0.5 mM).

To test whether racemisation is induced by amino acid catalysis, the substrate was incubated in phosphate buffer containing L-proline or L-lysine, respectively, at 5 mM concentration in combination with pyridoxamine-5'-phosphate. However no racemisation of aldehyde 1c was observed (Table 13).

L-proline L-lysine

time initial final initial final

entry Lee 1 c (%) Aee 1 c (%) (min) ee c (%} ee 1c (%) ee 1c (%) ee 1 c (%)

1 1 54 47 7 56 50 6

2 10 54 50 4 60 54 6

3 30 58 50 8 66 58 Q

4 60 58 50 8 54 48 6

5 120 58 50 8 46 40 6

Table 13: Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing PMP (0.5 mM) and L-proline (5 mM) or L-lysine (5 mM), respectively.

When enantioenriched aldehyde 1c was incubated with alanine in the same concentration (250 mM) as applied for biotransformations, a slight racemisation was observed within four hours (see Table 14). However racemisation under these conditions was far not as strong as observed with the E. coli background. entry time (min) initial ee 1 (%) final ee 1 (%) Aee 1 (%)

1 1 70 70 0

2 10 70 70 0

3 30 70 68 2 4 60 72 70 2

5 120 72 66 6

4 240 72 64 8

Tab!e 14: Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing rac-aianine (250 mM).

Based on these results (Tables 13 and 14), racemisation of substrate 1 c seems not to be induced by amino acids catalysis.

Moreover the effect of an ω-transaminase on the racemization of brivaracetam precursor 1c was investigated by addition of heat-shock purified transaminase ArRmutl 1 -ω-ΤΑ (Table 15) in the absence or presence of PLP, PMP (Table 16), or 2-propylamine (Table 18) or combinations thereof (Table 17, Table 18). entry time (min) initial ee 1 (%) final ee 1 (%) Aee 1 (%)

1 1 47 43 4

2 10 44 40 4

3 30 50 47 3

4 60 49 44 5

5 120 46 39 7

Table 15: Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing ArRmutl 1 -ω-ΤΑ. ωΤΑ, PLP ωΤΑ, PMP

time initial final initial final

entry Aee 1 c (%) Aee 1 c (%) (min) ee 1 c (%) ee 1 c (%) ee 1 c (%) ee 1c (%)

1 1 70 70 0 70 70 0

2 10 68 68 0 70 70 0

3 30 66 66 0 70 70 0

4 60 66 66 0 78 70 8

5 120 70 70 0 80 70 10

Table 16: Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing ArRmutl 1-ω-ΤΑ and PLP or PMP (1 mM), respectively.

The addition of transaminase did not significantly affect the ee value, neither in pure form nor under addition of the cofactor PLP. Nevertheless, when aldehyde 1 c was incubated with a combination of ArRmutl 1 -ω-ΤΑ and PMP, the ee decreased by 10% within two hours. Now also a combination of ArRmutl 1-ω-ΤΑ, PLP and 2-propylamine, a system in which PMP is generated in situ, was investigated and also in this case clear racemisation of aldehyde 1 c was observed (Table 17). In this system the concentration of substrate 1 c obviously decreases over time due to ω-ΤΑ mediated amination of the aldehyde moiety. However, from the results given here it is not possible to specify whether the racemisation is caused by an enzymatic process or simply by selective consumption of the enantiomer in excess in the amination. entry time (min) initial ee 1 c (%) final ee 1 c (%) Aee 1c (%)

1 1 54 46 8

2 10 50 44 6

3 30 60 50 10

4 60 54 44 10

5 120 68 44 24

6 240 70 10 60

Table 17: Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing ArRmutl 1 -ω-ΤΑ, IPA (50 mM) and PLP (1 mM).

When investigating changes in the ee of substrate 1 c under addition of 2-propylamine or a combination of ω-ΤΑ and 2-propylamine it turned out that this amine in 50 mM concentration induces racemisation of aldehyde 1 c. ωΤΑ, IPA IPA

time initial final initial final

entry Aee 1c (%) Aee 1 c (%

(min) ee 1 c (%) ee 1c (%) ee 1 c (%) ee 1 (%)

1 1 64 63 1 54 48 6

2 10 68 68 0 64 58 6

3 30 66 60 6 62 56 6

4 60 60 54 6 63 54 9

5 120 65 52 13 68 54 14

6 240 60 26 34 68 34 34

Table 18: Racemisation of aldehyde 1c in phosphate buffer pH 7 containing 2-propylamine (IPA. 50 mM) in the presence or absence of ArRmutl 1 -ωΤΑ.

Finally the racemisation of substrate 1c was investigated in the presence of ω-ΤΑ, PMP and D-alanine. In this experiment the ee decreased from 70% to 48% over four hours (Table 19), whereby no amination of the aldehyde was observed with alanine as amine donor. entry time (min) initial ee 1 c (%) final ee 1 c (%) Aee 1 c (%)

1 1 66 64 2

2 60 60 56 4

3 240 70 48 22

Table 19: Racemisation of aldehyde 1 c in phosphate buffer pH 7 containing ArRmutI 1 -ω-ΤΑ, D-Ala (250 mM) and PMP (1 mM).

These results indicate that substrate 1 c can be racemised by low amounts of PMP (ca. 0.04 eq) in combination with an ω-transaminase or with an amine (e.g. , 2-propylamine) in equimolar amounts. Moreover these results approve the assumption that racemisation of aldehyde 1 c is performed via the formation of a Schiff base via the proposed mechanism outlined in Scheme 12. Aldehyde 1c is assumed to form the imine from IPA, which undergoes reversible rearrangement to the corresponding enamine and further to the opposite enantiomer.

However, despite racemisation being observable in the presence of selected compounds of the reaction system, the by far strongest racemisation is observed by E. coli background most likely due to the various PLP-dependent enzymes present.

Scheme 12: Mechanism proposed for the racemisation of substrate 1 c via imine/enamine formation.

Nevertheless, the combination of ArRmutI 1 -ω-ΤΑ, D-Ala and PMP shows deary a racemization (Table 19). A combination of ArRmutI 1-ω-ΤΑ and PMP showed also a slight decrease of ee (Table 16). This racemization is not observed when using the components on their own (see Table 14 for rac-Ala, Table 15 for ArRmutI 1 -ω-ΤΑ and Table 12 for PMP) providing a reaction set-up for high yield and high ee values. Experimental Procedures

Racemization of brivaracetam precursor 1 c Deracemisation of brivaracetam precursor raolc

Purified (heat pulse) ADH-A (20 μΐ_, 0,3 U) and brivaracetam precursor rac-1 c (9 pL, 50 mM) were added to sodium phosphate buffer (1 mL, 100 mM, pH 7.0, 1 mM NAD free acid) and iPrOH (30 μΙ_). The reduction was performed at 30°C and 750 rpm for 15-20 min. After extraction with ethyl acetate (2 x 500 pL) the combined organic phases were dried over sodium sulphate and analysed via gas chromatography.

Racemisation

The organic solvent ethyl acetate was removed under air flow and enantioenriched brivaracetam precursor 1c was incubated in one of the following mixtures:

· Sodium phosphate buffer (1 mL, 100 mM, pH 7)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing lyophilized cells of the E. coli BL21 (DE3) (20 mg, host for ω-ΤΑ overexpression)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing PMP (0.5 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing PMP (0.5 mM) and L-proline (0.3 mg, 5 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing PMP (0.5 mM) and L-leucine (0.4 mg, 5 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing rac-alanine (22.3 mg, 250 mM)

· Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing ArRmutl 1 -ω-ΤΑ (250 pL, heat pulse purified preparation)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing ArRmutl 1 -ω-ΤΑ (250 pL, heat pulse purified preparation) and PLP (1 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing ArRmutl 1-ωΤΑ (250 pL, heat pulse purified preparation) and PMP (1 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing IPA (50 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing IPA (50 mM) and ArRmutl 1 -ω-ΤΑ (250 pL, heat pulse purified preparation)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing iPA (50 mM), ArRmutl 1- ω-ΤΑ (250 pL, heat pulse purified preparation) and PLP (1 mM)

• Sodium phosphate buffer (1 mL, 100 mM, pH 7) containing D-alanine (22.3 mg, 250 mM), ArRmutl 1 -ω-ΤΑ (250 pL, heat pulse purified preparation) and PMP (1 mM) Racemisation tests were performed at 30°C and 750 rpm. Samples were taken after 1 min, 10 min, 30 min, 1 h, 2 h or 4 h. After extraction with EtOAc (2 x 500 μΙ_) the combined organic phases were dried over Na₂S0₄ and analysed via chiral gas chromatography

Example 6: Influences of organic cosolvents and different reaction temperatures in the stereoselective enzymatic synthesis of a brivaracetam precursor using ω-transaminases Cosolvent study

The effect of the organic cosolvents 1 ,2-dimethoxyethane (DME), dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) on the stereoselectivity of ω-transaminases in the reductive amination of the aldehyde compound 1c was tested using concentrations of 10 to 30 vol-% of the different organic cosolvents, as illustrated in Scheme 13:

Scheme 13: Transamination of compound 1 c in the presence of organic cosolvents.

Despite a general decrease of conversion (after 24 h), the use of cosolvents resulted in many cases in an enhancement of ee values.

The results obtained with the different cosolvents are shown in the following Tables 20 to 23: ω-transaminase DME (vol-%) conv. (%) ee (%) of lactam 3

ΗΝ-ω-ΤΑ 10 <1 -

ArR-ω-ΤΑ 10 60 33 (S)

PF-ω-ΤΑ 10 89 80 (R)

20 93 85 (R)

30 28 80 (R)

VF-ω-ΤΑ 10 99 90 (R) 20 99 92 (R)

30 78 93 (R)

PD-ω-ΤΑ 10 99 86 (R)

20 99 80 (R)

ArS-ω-ΤΑ 10 99 51 (R)

20 99 56 (R)

30 94 65 (R)

ΟΑ-ω-ΤΑ 10 99 77 (R)

20 99 86 (R)

Table 20: Reductive amination using 1 ,2-dimethoxyethane (DME) as cosolvent.

The ω-transaminases PF-ω-ΤΑ, VF-ω-ΤΑ and ArS-ω-ΤΑ showed the highest tolerance towards DME, as shown in Table 20. VF-ω-ΤΑ gave up to 93% ee for the (R)-isomer of the lactam 3 (in 30 % DME).

Table 21 : Reductive amination using dimethylformamide (DMF) as cosolvent. n.d. = not determined (due to low conversion).

When DMF was used as cosolvent (see Table 21 ), the ee value of lactam 3 was increased to 86% ee for the (R)-isomer when VF-ω-ΤΑ and 20 vol-% DMF were used . ω-transaminase DMSO (vol-%) conv. (%) ee (%) of lactam 3

ΗΝ-ω-ΤΑ 10 <1 n.d.

ArR-ω-ΤΑ 10 85 31 (S)

PF-ω-ΤΑ 10 91 59 (R)

VF-ω-ΤΑ 10 99 84 (R)

20 99 84 (R)

PD-ω-ΤΑ 10 99 65 (R)

20 99 64 (R)

ArS-ω-ΤΑ 10 99 14 (R)

ΟΑ-ω-ΤΑ 10 99 64 (R)

20 99 73 (R)

Table 22: Reductive amination using dimethyl sulfoxide (DMSO) as cosoivent.

The use of DMSO as a cosoivent showed only little effect on the stereoselectivity of the tested ω-TAs (see Table 22). The obtained conversions were also less affected by this cosoivent than by DME or DMF.

The ω-transaminases PF-ω-ΤΑ, VF-ω-ΤΑ. PD-ω-ΤΑ and ΟΑ-ω-ΤΑ showed the highest tolerance towards the organic cosolvents DME, DMF and DMSO whereas ΗΝ-ω-ΤΑ did not convert substrate 1c at all in presence of the tested cosolvents. The highest ee values (93% ee (R), VF-ω-ΤΑ, 30 vol-% DME) were obtained by application of DME which, however, had the most reducing effect on the conversion rates.

Since mainly the (R)-isomer of lactam 3 was produced by the enzymes investigated above, also the (S)-selective ArR-ω-ΤΑ variant ArRmutl 1-M1 17F-G279V-A60V-u TA was tested employing the organic cosolvents DME, DMF and DMSO in concentrations of 20 to 30 vol-% (see Scheme 14 below).

5c

Scheme 14: Transamination of compound 1 c using ArRmutl 1 -M1 17F-G279V-A60V-OJ-TA in the presence of organic cosolvents.

In these experiments, the imine 4c (formed from unreacted substrate 1 c and 2-propylamine) as well as the imine 5c (formed from unreacted aldehyde 1 c and product amine 2c) were found in all samples. These two Schiff bases were identified via GC-MS measurements and imine 4c was also formed in blank reactions when only substrate 1 c and 2-propylamine were combined in aqueous buffer.

Table 23 Reductive amination using ArRmutl 1 -M 1 1 7F-G279V-A60V-DJ-TA and cosolvents. As shown in Table 23 above, it was found that this variant of ArRmutl 1 -ω-ΤΑ, i.e. ArRmutl 1 - M 1 1 7F-G279V-A60V-00-TA. has a high tolerance towards organic cosolvents. No sig nificant difference in the ee of the (S)-isomer of lactam 3 was observed when employing various cosolvents in different amounts. The highest stereoselectivity of ArRmutl 1 -M 1 17F-G279V- A60V- TA was observed when DMF was applied (80% ee) whereas the use of DMSO resulted in the highest conversions to the product 3 (61 % conv. , 20% DMSO).

The influences of organic cosolvents and temperature on the biocatalytic reductive amination of a compound of formula (I), i.e. the aldehyde 1 c, to the corresponding amine 2c and its conversion to lactam 3 were investigated. The ee value of the product 3 could be enhanced to 93% (R) using VF-ω-ΤΑ with 30 vol-% DME as cosolvent and to 80% (S) using ArRmutH - M 17F-G279V-A60V -TA with 30 vol-% DMF as cosolvent. Temperature study

The temperature dependence of five ω-TAs was investigated (see Scheme 15 below).

wTA, PLP

Scheme 15: Transamination of compound 1 c using ω-transaminases at different temperatures.

As shown in Table 24, all five tested ω-transaminases transformed aldehyde 1 c to the desired product in a broad temperature range (25-45°C). ω-transaminase temperature conv. ee (%)

(°C) (%) of

lactam 3

ArR-ω-ΤΑ 25 99 17 (S)

30 95 1 1 (S)

35 93 13 (R)

40 92 20 {R)

45 61 23 (R)

ArS-ω-ΤΑ 25 >99 25 (R)

30 >99 25 (R)

35 >99 29 (R)

40 >99 17 (R) 45 >99 12 (R)

PD-ω-ΤΑ 25 >99 31 (R)

30 >99 31 (R)

35 >99 42 (R)

40 >99 40 (R)

45 >99 45 (R)

PF-ω-ΤΑ 25 >99 40 (R)

30 >99 41 (R)

35 >99 49 (R)

40 >99 42 (R)

45 >99 44 (R)

VF-ω-ΤΑ 25 >99 40 (R)

30 >99 38 (R)

35 >99 54 (R)

40 >99 42 (R)

45 >99 54 (R)

Table 24: Reductive amination using different ω-transaminases at varied temperatures.

The ee values of the formed product 3, however, were hardly affected by varied temperatures. A notable exception was again ArR-ω-ΤΑ which inverted its enantioselectivity between 30°C and 35°C. Thus, the (S)-isomer of lactam 3 was formed in excess at 30°C and below whereas the (R)-isomer of lactam 3 was preferentially produced at 35°C and above.

In this regard, a-methylbenzylamine (MBA) was chosen to be tested as an alternative amine donor (see Scheme 1 6 below) since it is reported to be a potent amine donor in the amination of various ketones using ArR-ω-ΤΑ or ΟΑ-ω-ΤΑ (Park et al. , 2013) .

Scheme 16: ω-Transaminase-catalyzed reductive amination of compound 1 c using a-methylbenzyiamine (MBA) as amine donor.

When substrate 1 c was employed, however, mainly a Schiff base formed from the aldehyde (1 c) and MBA was produced. In nearly all experiments substrate 1 c was fully converted to this imine. Only when ArS-ω-ΤΑ was applied, a small amount of the desired product 3 was found with 50% ee (R). These results show that the use of a-methylbenzylamine as an amine donor in the ω-ΤΑ-catalyzed reductive amination of aldehyde compounds of formula (I) is disadvantageous, particularly when compared to the use of amine donors such as 2-propylamine or alanine.

Example 7: Stereoselective enzymatic synthesis of pregabaiin precursors using ω-transaminases

The pregabaiin precursor 10c was synthesized in a four-step process with 20% overall yield. Then it was transformed by ω-transaminase-mediated reductive amination to the corresponding amine 12c which cyclizes under basic conditions to form the lactam 13, as illustrated in Scheme 1 7 below. The pharmaceutically interesting (S)-enantiomer was synthesized in this way with up to 76% ee. Moreover, to further improve the enantioselectivity, site-directed mutagenesis of two (S)-selective ω-TAs was performed.

10c 12c 13

Scheme 17: ω-ΤΑ-mediated reductive amination of pregabaiin precursor 10c

Bioamination of pregabaiin precursor 10c

Scheme 18: ω-ΤΑ-mediated reductive amination of pregabaiin precursor 0c. As illustrated in Scheme 1 8, the pregabalin precursor 10c was transformed with the different ω-transaminases under the conditions which were found to give the best results for the corresponding brivaracetam precursor 1 c (see Examples 4 and 6) . The additional methyl group of aldehyde 10c had no significant effect on the enantioselectivity of the ω-transaminases but on conversions. In general, the conversions of aldehyde 10c were lower than the ones of compound 1 c and some ω-TAs (i.e. , ArR-ω-ΤΑ and ΑΤ-ω-ΤΑ) did not convert substrate 10c. The desired (S)-enantiomer of lactam 13 was obtained in up to 76% ee (using ArRmutl 1 -M1 17F-A60V-G279V-u>-TA). The results obtained with the different ω-transaminases for both the brivaracetam precursor 1 c and the pregabalin precursor 0c are summarized in the following Table 25:

Table 25: Reductive amination of brivaracetam precursor 1 c and pregabalin precursor 10c. The ee of lactam 13 was determined. DMSO - dimethyl sulfoxide; DMF = dimethylformamide; DME = 1 ,2-dimethoxyethane; n.d. = not determined (¹ - product was only detected as imine 18c formed from compound 10c and 12c; ² - not determined due to low conversion). The reductive amination of methyl ester 10e employing ArRmutl 1 -M1 17F-A60V-G279V-oo-TA and GZ-ω-ΤΑ was also investigated. However, when ArRmutl 1 -M 1 17F-A60V-G279V-to-TA was employed with 2-propylamine as amine donor, neither substrate nor product could be extracted from the reaction medium. GZ-ω-ΤΑ (using D-alanine as amine donor) transformed aldehyde 10e to the corresponding amine or lactam with 90% conversion and 10% ee (S).

Variants of ArRmutl 1 -M 17F-A60V-G279V-uj-TA and GZ-ω-ΤΑ Homology models of ArRmutl 1 -M 1 17F-A60V-G279V-co-TA and GZ-ω-ΤΑ (the structure of ArR-ω-ΤΑ was used as template), the two ω-transaminases showing the highest stereoselectivity for brivaracetam precursor 1 c, were compared and five amino acids of ArRmutl 1 -M1 17F-A60V-G279V-W-TA and seven amino acids of GZ-ω-ΤΑ were selected to be mutated by site-directed mutagenesis (see experimental procedure described above and Example 3).

The mutations were performed using an Agilent QuickChange Lightning Site Directed Mutagenesis kit and plasmids carrying genes of eleven of the desired twelve genes were successfully prepared. The corresponding variants of ArRmutl 1 -ω-ΤΑ and GZ-ω-ΤΑ were overexpressed in E. coli BL21 (DE3), and the lyophilized cells were tested for the amination of pregabalin precursor 10c.

The bioamination of the pregabalin precursor 10c was performed using these variants of ArRmutl 1 -ω-ΤΑ and GZ-ω-ΤΑ at conditions found to be best for ArRmutl 1-M 17F-A60V- G279V-u)-TA and GZ-ω-ΤΑ. Due to the basic workup, product 12c reacted further to lactam 13 which was analysed via GC measurements. The bioamination of the brivaracetam precursor 1 c was also tested using these variants of ArRmutl 1 -ω-ΤΑ and GZ-ω-ΤΑ, and the corresponding lactam 3 was analysed via GC measurements. The results of these experiments are summarized in Table 26 below. In particular, the variants GZ-L56V-u)-TA and GZ-T273S-UJ-TA were found to give the desired (S)-enantiomer of lactam 13 at >60 % conversion and an enantioselectivity of 80% ee and 72% ee, respectively, which renders these enzymes highly advantageous as (S)-selective ω-TAs, e.g., in the preparation of pregabalin or pregabalin precursors. Analogy of

ω-transaminase mutation number XX

10c 1c Kk for ArRmut*

conv. (%) ee conv. (%) ee

ArRmut* 27 76 (S) 54 80 (S)

ArRmut*-T64V <1 n.d.² <1 n.d.²

ArRmut*-T1 19E <1 n.d.² <1 n.d.²

ArRmut*-l 152V <1 76 (S) >99 77 (S)

ArRmut*-S277T 15 58 (S) >99 56 (S)

ArRmut*-S277N 10 n.d.³ <1 n.d.²

GZ-ω-ΤΑ 40 n.d.³ 70 55 (S)

GZ-L56V-u)-TA L60V 63 80 (S) 62 48 (S)

GZ-V60T- -TA V64T 3 42 (S) 48 rac

GZ-E1 15T-co-TA E1 19T 60 15 (S) 86 6 (S)

GZ-V148I-W-TA V152I <1 n.d.² <1 n.d.²

GZ-T273S-oo-TA T277S 65 72 (S) 94 34 (S)

GZ-T273N-u>-TA T277N 25 n.d.³ <1 n.d.²

Table 26: Reductive amination of pregabalin precursor 10c and brivaracetam precursor 1 c using ArRmutl 1 -M1 17F-A60V-G279V-U-TA (abbreviated in this table as ArRmut*), GZ-ω-ΤΑ, or different variants thereof.

Reaction conditions: ArRmut* (and variants thereof): substrate 10c or 1 c (50 mM), ω-ΤΑ (lyophiiised cells, 20 mg mL^"1) in phosphate buffer (100 mM, pH 7, 1 mM PLP) containing 20 ol% DMF and 2-propylamine (500 mM) at 30°C, 120 rpm, 24 h; GZ-ω-ΤΑ (and variants thereof): substrate 10c or 1 c (50 mM), ω-ΤΑ (lyophiiised cells, 20 mg mL^"1), D-alanine (250 mM), Ala-DH (12 U), FDH (1 1 U), NAD (1 mM) and ammonium formate (1 50 mM in phosphate buffer (100 mM, pH 7, 1 mM PLP) at 30°C, 120 rpm, 24 h).

¹ ee determined of the corresponding lactam 13 or 3

² not determined due to low conversion

³ not determined, product was only detected as imine of substrate and product In view of the results obtained with the ω-TAs described above as well as the investigation of homology models, seven further variants of GZ-ω-ΤΑ were constructed (see Table 27 below). Since the primers for all variants containing the mutation A275V are likely to form hairpins with high melting temperature, mutagenesis was performed following a protocol alternative to the Agilent QuickChange Lightning site directed mutagenesis kit, which was applied for the construction of variants previously. In this concept, PCR is divided into two steps, first the two DNA strands were amplified separately (linear step), then the two PCR mixtures are combined to amplify the whole plasmid (exponential step). Mutants obtained in this way were overexpressed in E. coli BL21 (DE3), and lyophilised cells were employed as catalysts for the reductive amination.

The brivaracetam precursor 1c was aminated by four of the prepared GZ-ω-ΤΑ variants, mostly in low yields and mediocre ee values. The pregabalin precursor 10c was transformed only by GZ-L56V-T273S-u TA, which is combining the mutations of GZ-L56V-u)-TA and GZ-T273S-UU-TA, the two variants showing the highest stereoselectivity in previous experiments (see Table 26 above). GZ-L56V-T273S- -TA was found to convert 92% of substrate 10c in 24 hours and lactam 13 was formed in 60% ee (S). The results obtained for the seven further variants of GZ-ω-ΤΑ are summarized in the following Table 27.

Analogy of

mutation J VL \ 1

ω-transaminase 10c 1c

number for

ArRmut^*

conv. (%) ee (%)¹ conv. (%) ee (%)¹

GZ-Y58L-0J-TA Y62L <1 n.d.² <1 n.d.²

GZ-E1 15D-u>-TA E1 19D <1 n.d.² 35 72 (S)

GZ-V148A-0J-TA V152A <1 n.d.² 18 40 (S)

GZ-A275V-UJ-TA A279V <1 n.d.² <1 n.d.²

GZ-L56V-T273S-W-TA L60V-T277S 92 60 (S) 97 25 (S)

GZ-L56V-A275V-10-TA L60V-A279V <1 n.d.² <1 n.d.²

GZ-T273S-A275V-u TA T277S-A279V <1 n.d.² 4 55 (S)

Table 27: Reductive amination of pregabalin precursor 10c and brivaracetam precursor 1c using different GZ-ω-ΤΑ variants (ArRmut* = ArRmutl 1 -M 117F-A60V-G279V-oo-TA) .

Reaction conditions: substrate 10c or 1c (50 mM), ω-ΤΑ (lyophilised cells, 20 mg ml_^"1) in phosphate buffer (100 mM, pH 7, 1 mM PLP) at 30°C, 120 rpm, 24 h; D-alanine (250 mM), Ala-DH (12 U), FDH (11 U), NAD⁺ (1 mM) and ammonium formate (150 mM).

¹ ee determined of the corresponding lactam 13 or 3

² not determined due to low conversion

³ not determined, product was only detected as imine of substrate and product The amination of pregabalin precursor 10c and brivaracetam precursor 1 c was furthermore tested using the variants of ArRmutI 1 -ω-ΤΑ with D-alanine as amino donor and using the GZ-ω-ΤΑ variants in combination with 2-propylamine as amino donor. No conversion to the corresponding amine was observed in these experiments.

If the formation of the lactam 13 (see Schemes 17 and 18 above) is undesired, it was found that the lactam formation can be avoided and the corresponding open-chain amino ester can be isolated, e.g., by replacing the reaction medium after the reductive amination step with a saturated NaHC0₃ solution.

This was investigated subsequent to the reductive amination of the aldehyde compound 1 c to the amine 2c, as shown in Scheme 19 below. The open-chain amino ester 2c was isolated from the reaction medium by replacing 10 N NaOH solution by saturated NaHC0₃ solution in the basic extraction of amino compounds. The reaction was first tested on analytical scale (>99% conv.), and was then performed on a 10 mL-scale and the product 2c was identified by NMR (36 % yield).

Scheme 19: Reductive amination compound 1 c to the amine 2c

Synthesis of amine 2c: Lyophilized cells of E. coli containing overexpressed PD-ω-ΤΑ (200 mg) and FDH (50 mg, 1 10 U) were rehydrated in phosphate buffer (9 mL, pH 7, 100 mM) containing PLP (1 mM) and NADH free acid (1 mM) at 30°C and 120 rpm for 30 min. Ala-DH (150 μΐ, 120 U), ammonium formate (95 mg, 150 mM), L-aianine (223 mg, 250 mM) and cosolvent ,2-dimsthoxyethane (100 mL) as well as the substrate (100 mg. 0,5 mmol) were added. Reductive amination was carried out at 30°C in an orbital shaker (120 rpm) for 24 h. Cell material was removed by centrifugation and the supernatant was acidified with 2 M HCI solution (2 mL) and extracted with r-butyl methyl ether (2 * 5 mL). The aqueous layer was basified by addition of saturated NaHC0₃ solution (3.5 mL) and was extracted again with f-butyl methyl ether (3 x 10 mL). The combined organic phases of the basic extraction were dried over Na₂S0₄ and the solvent was evaporated to give a yellow oil (36 mg, 36 % yield).

¹ H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.92 (t, ³J_{6 5} = 6.9 Hz, 3 H, CH₃) , 1 .20-1 .50 (m, 6 H, H-4, H-5, H₂), 1.45 (s, 9 H, C(CH₃)₃), 1 .86 (m_c, 1 H, CH), 2.14-2.22 (dd, ³J_23i3 = 6.3 Hz, ²J_2a,_2b = 14.6 Hz, 1 H, 2-H_a), 2.20-2.30 (dd, ³J_2b,3 = 7.1 Hz, ²J_2b,2a = 14.6 Hz, 1 H, 2-H_b), 2.56-2.65 (dd, 3^NCHa,3 = 6.3 Hz, ²J _{N CH}a. NCHb = 12.9 Hz, 1 H, NCH₂), 2.65-2.75 (dd, ³J_3iNc_Hi) = 5.6 Hz, ²J_NCHb, NCHa ⁼ 12.9 Hz, 1 H, HCH₂) 3C-NMR (75 MHz, CDCI₃): 5_C [ppm] = 14.43(-CH₃), 19.99 (5-C), 28.23 (-CH(CH₃)₃), 34.22 (3-C, 4-C), 38.44 (2-C), 45.52 (NCH₂), 80.31 (CO), 172.95 (OCO).

Example 8; Ester hydrolysis of brivaracetam precusor

The ester hydrolysis of a brivaracetam ierf-butyl ester derivative was successfully conducted on analytical scale using acidic conditions.

In order to test the cleavage of the ferf-butyl ester group of amino ester 2c, first the reductive amination of aldehyde 1 c was performed on a preparative scale (20 mL) employing PD-ω-ΤΑ, which gave high conversion on the analytical scale, as illustrated in Scheme 20:

1 c 2c 6

Scheme 20: ω-ΤΑ-mediated reductive amination and subsequent ester cleavage.

The first attempt to form compound 6 was to test enzymes (lipase, esterase) for the ester cleavage. The activity of Candida antarctica lipase A (CalA) (Henke et al., 2002; Kourist et al. , 2008; Dominguez de Maria et al., 2005; Schmidt et al. , 2005; Krishna et al. , 2002), BS2 from Bacillus subtilis (Kourist et al., 2008) and PLE (Kourist et al., 2008; Schmidt et al. , 2005) towards esters of tertiary alcohols is known from literature. These hydrolases bear the same GGG(A)X motif in the active site. Moreover, also Candida antarctica lipase B (CalB) and Candida rugosa esterase were tested for the cleavage of compound 2c due to their broad substrate scope and wide applicability. Ester cleavage experiments were conducted in phosphate buffer pH 7 or pH 8 at 40°C and ^• 700 rpm. The progress of the reaction was followed by Liquid-Chromatography-Mass- Spectrometry (LC-MS) over 10 days. However, only partial ester hydrolysis took place in all samples and was also observed in blank reactions (substrate 2c was incubated in buffer, without enzyme) in the same range.

Cleavage of esters similar to compound 2c is reported to work under acidic conditions in the presence of trifluoro acetic acid or formic acid (Schelkun et al. , 2006; US 6,710, 190). Taking into account the results observed in the blank reaction mentioned above, ester cleavage was now tested by heating a suspension of compound 2c in water to 95°C and by incubating compound 2c in an aqueous solution containing 0.3 M HCI at 60°C. Since this second method led to faster conversion, it was used to cleave ester 2c on a preparative scale (0.2 mM), where 20 mg of crude product 6 was isolated.

Experimental procedure

Synthesis of brivaracetam precursor 2c

Lyophilized cells of £ coli containing overexpressed PD-ω-ΤΑ (400 mg) and FDH (100 mg, 220 U) were rehydrated in phosphate buffer (18 mL, pH 7, 100 mM) containing PLP (1 mM) and NADH free acid (1 mM) at 30°C and 120 rpm for 30 min. Ala-DH (300 μΐ_, 240 U), ammonium formate (180 mg, 150 mM), alanine (446 mg, 250 mM) and cosolvent 1 ,2-dimethoxyethane (200 mL) as well as the substrate (201 mg, 1 mmol) was added. Reductive amination was carried out at 30°C in an orbital shaker (120 rpm) for 24 h. Cell material was removed by centrifugation and the supernatant was basified by addition of saturated NaHC0₃ solution (3.5 mL) and was subsequently extracted with ferf-butyl methyl ether (7 ^ 10 mL). The combined organic phases were dried over Na₂S0₄ and the solvent was evaporated to give the product as a yellow oil (130 mg, 65% yield).

Ή-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.92 (t, ³J_{5 5} = 6.9 Hz, 3 H, CH₃), 1.20-1 .50 (m, 6 H, H-4, H-5, H₂), 1 .45 (s, 9 H, C(CH₃)₃), 1.86 (m_c, 1 H, 14.6 Hz, 1 H , 2-H_a). 2.20-2.30 (dd, ³J_{2b 3}= 7.1 Hz,

(dd, ³J_{NCHa 3}= 6.3 Hz, ²J _NCHa! CHb= 12.9 Hz, 1 H, NCH₂), 2.65-2.75 (dd, J _NCHb = 5.6 Hz, ² J _m _NCHa= 12.9 Hz, 1 H, NCH₂)

¹³C-NMR (75MHz, CDCI₃): _c [ppm] = 14.43(-CH₃), 19.99 (5-C), 28.23 (-CH(CH₃)₃), 34.22 (3-C, 4-C), 38.44 (2-C), 45,52 (NCH₂), 80.31 (CO), 172.95 (OCO). Ester hydrolysis

Four 2 mL Eppendorf tubes containing water (1 mL), HCI (200 ί of 2 M solution) and substrate 2c were incubated at 60°C, 700 rpm for 5 h. Content of the four reaction vessels was combined and the solvent was evaporated to give a brown oil. To remove hydrophobic impurities the crude product was dissolved in water and extracted with EtOAc (2 x 50 ml_). The aqueous phase was separated and the solvent was evaporated to give a brown residue (20 mg product, not pure). H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.95 (m_0l 3H, CH₃), 1.29-1.51 (m, 4H, CH₂) , 2.17 (m_c, 1 H, CH), 2.46 (d,

6.5 Hz, 2 H, NCH₂)

¹³C-NMR (75MHz, CDCI₃): 6_C [ppm] = 14.32 (-CH₃), 20.5 (5-C), 34.61 ( -C), 34.74 (3-C), 36.71 (2-C), 44.04 (NCH₂), 174.33 (OCOH)

Example S: isolation of amino ester 12c and cleavage of the 'Bu-group for the preparation of pregabalin

In order to isolate pregabalin. first precursor 10c was transformed to the corresponding amine 12c and the tert-butyl ester was cleaved under acidic conditions to isolate the HCI salt of pregabalin 20.

Scheme 21 : ω-ΤΑ mediated reductive amination and following ester cleavage.

Simultaneously to the amination and ester cleavage of the corresponding brivaracetam precursor (as described in Example 8) also pregabalin was synthesised. First the reductive amination of aldehyde 10c was performed on a preparative scale (20 ml_) employing VF-ω-ΤΑ, which gave high conversion on the analytical scale. Cleavage of the ester moiety of 12c was then performed in a 0.3 M aqueous solution of HCI at 60°C to isolate the desired product in the form of the HCI salt 20 in 63% yield. Experimental procedure

Biotransformation on preparative scale

Lyophilized cells of E. coli containing overexpressed VF-ω-ΤΑ (400 mg) and FDH (100 mg, 220 U) were rehydrated in phosphate buffer (20 mL, pH 7, 100 mM) containing PLP (1 mM) and NADH free acid (1 mM) at 30°C and 120 rpm for 30 min. Ala-DH (300 μΙ_, 240 U), ammonium formate (180 mg, 150 mM), alanine (446 mg, 250 mM) and the substrate 10c (220 mg, 1 mmol) was added. Reductive amination was carried out at 30°C in an orbital shaker (120 rpm) for 24 h. Cell material was removed by centrifugation and the supernatant was basified by addition of saturated NaHC0₃ solution (3.5 mL) and was extracted with f- butyl methyl ether (7 x 10 mL). The combined organic phases were dried over Na₂S0₄ and the solvent was evaporated. The product was purified by column chromatography (silica, CH₂CI₂ : MeOH 100/0 to 94/6) and isolated as a orange oil (108 mg, 0.4 mM, 50 % yield). ¹ H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.89 (d, ³J_e>5 = 6.5 Hz, 3 H, CH₂). 0.90 (d, ³J_6,5 = 6.5 Hz, 3 H, CH₃). 1 .14 (m_c, 2H, 4-H), 1 44 (s, 9H, C(CH₃)₃), 1.60 (bs, 2H, H₂), 1.64 (m_c, 1 H, 5-H), 1.93 (m_c, 1 H, 3-H), 2.13-2.21 (dd, ³ _{2a 3} = 6.1 Hz, ² _{2a b}= 14.6 Hz, 1 H, 2-H_a), 2.21 -2.30 (dd, V_2B,3 = 7.3 Hz, ²J_{2b a}= 14.6 Hz, 1 H, 2-H_b), 2.55-2.65 (dd, ³J_NCHa.₃= 6.5 Hz, ²J _NCHa, CH₀= 12.9 Hz, 1 H, NCH₂), 2.66-2 76 (dd.

12 9 Hz. 1 H. NCH₂).

^{, 3}C-NMR (75MHz, CDCI₃): δ₀ [ppm] = 22.94(-CH₃), 25.29 (5-C), 28.24 (-CH(CH₃)₃), 36.30 (3- C), 38.60 (2-C). 41 .52 (4-C), 45.71 (NCH₂), 80.36 (CO), 172.88 (OCO).

NMR spectra are not reported in literature.

Ester hydrolysis

f-Bu-ester 12c (140 mg, 65 mM) was diluted with water (5 mL) and HCI (1 mL of 2 M solution) was added. The mixture was heated to 65°C for 12 h. After cooling to room temperature the solvent was evaporated. The remaining brown oil was dissolved in a small amount of water extracted with EtOAc (3 x 1 mL). The aqueous phase was removed under vacuum to give the product as a yellow oil (104 mg, 0.41 mM, 63 % yield).

¹ H-NMR (300 MHz. CD₃OD): 0.93 (d, ³J_{6 5} = 6.7 Hz, 3 H, CH₃), 0.95 (d, ³J_{6 5} = 6.7 Hz, 3 H, CH₃), 1.27 (t, ³J = 7.2 Hz, 2 H, 4-H), 1.67 (m_c, 1 H, 5-H), 2.22 (m_C! 1 H, 3-H), 2.45 (d, ³J_{2 3} = 6.2 Hz. 2 H, 2-H), 2.97 (d, ³J_{2 3} = 6.3 Hz. 2 H, NCH,). ¹³C-NMR (75MHz, CDCI₃): 5_C [ppm] = 22.52 (-CH₃), 23.18 (-CH₃), 26.14 (5-C), 32.60 (3-C), 37.1 1 (2-C), 42.03 (4-C), 44.50 (NCH₂), 175.79 (OCO).

NMR spectra are in agreement with literature,

Example 10: Alternative synthetic routes to pregabalin precursor 10c

The synthesis of the pregabalin precursor 10c was established in Example 2 starting from the aldehyde 8, as illustrated in the following Scheme 22:

Scheme 22: Synthesis of compound 10c (see Example 2).

As described in Example 2, pyridinium chiorochromate (PCC) can be used for the oxidation of 4-methyl-1-pentanol to the corresponding aldehyde 8 which is not commercially available. Although this method generally worked well, it is desirable to avoid using PCC due to its high toxicity. Therefore, two alternative approaches were investigated, on the one hand the ADH mediated oxidation of alcohol 1 and on the other hand the hydrogenation of unsaturated aldehyde 19.

ADH mediated oxidation Unlike the PCC method, the application of alcohol dehydrogenases (see Scheme 23 below) provides a green and ecologically benign method.

Scheme 23: Oxidation of alcohol 11 employing ADH. A panel of alcohol dehydrogenases was investigated and besides two examples producing the corresponding carboxylic acid and two examples not transforming the substrate, ADH-hT from Bacillus stearothermophilus (Cannio et al., 1994; Ceccarelli et al. , 2004) and lyophilised whole cells of Janibacter terrae (Orbegozo et al., 2010) gave the most promising results to be optimised, as also shown in the following Table 28: biocatalyst conv. (%)¹

J. terrae whole cells² 42

ADH-hT purified³ 45

Codexis primary ADH 101³

Codexis primary ADH 103³

ADH-CP³ 8⁴

ADH-T³ 30⁴

Table 28: Potential biocatalysts for the oxidation of 4-methylpentanol 11.

1 determined by GC-MS analysis

2 Reaction conditions: substrate 1 (80 mM), J, terrae (lyophilised whole cells, 35 mg ml_^"1) in phosphate buffer (100 mM, pH 7) containing acetaldehyde; at 30 °C, 750 rpm, 24 h.

3 Reaction conditions: substrate 1 1 (50 mM), ADH (30 μί_) in phosphate buffer (1 00 mM. pH 7.0, 1 mM NAD(P)H) containing acetaldehyde/acetone (10 eq); determined by GC-MS analysis.

4 Signifcant amount of acid formation was detected.

ADH mediated oxidation - standard conditions: 30 pL ADH added to 1 ml_ phosphate buffer (1 mM, pH 7.0, 100 mM). 4-Methylpentanol 11 (6.2 μΐ_, 50 mM) was added. Acetaldehyde/acetone (28 μΙ_, 10 eq) was added and the reactions were vertically shaken for 24 h at 30°C and 650 rpm. After extraction with EtOAc (2 x 500 pL) the combined organic phases were dried over Na₂S0₄ and analysed via GC-MS.

Janibacter terrae mediated oxidation - standard conditions: 35 mg lyophilised whole cells Janibacter terrae were hydrated in phosphate buffer (1 ml_. pH 7, 100 mM) for 30 min at 30°C and 120 rpm. Acetaldehyde (53 pL, 10 eq) and 4-methylpentanol 11 (10 pL, 80 mM) were added and vertically shaken for 24 h at 30°C and 650 rpm. After extraction with EtOAc (2 x 500 μΙ_) the combined organic phases were dried over Na₂S0₄ and analysed via GC-MS.

Optimisation of the process employing Janibacter terrae whole cells was performed in terms of hydrogen acceptor to substrate ratio, reaction temperature, pH value and substrate concentration. The most efficient parameter to enhance the conversion in this process was the increase of pH up to a value of 8.5 to 9 (see Table 29). pH value conv.

pH 6.5 64

pH 7.0 67

pH 7.5 81

pH 8.0 82

pH 8.5 84

pH 9.0 84 Table 29: Influence of pH on the oxidation of 4-methylpentanol by J. terrae whole cells

¹ Reaction conditions: substrate 11 (80 mM), J. terrae (lyophilised whole cells, 35 mg ml¹) in phosphate buffer (100 mM) containing acetaldehyde (15 eq); at 30°C, 750 rpm, 24 h; determined by GC-MS analysis. When substrate concentration was increased, the total conversion dropped. However, the molar amount of substrate converted was still increased (see Table 30 below). substrate [mM] conv. (%)¹ conv. (mmol)

80 84 67

160 62 99

240 54 130

Table 30: Study of employed substrate concentration

Reaction conditions: substrate 11 , J terrae (lyophilised whole cells, 35 mg mL^"1) in phosphate buffer (100 mM, pH 8) containing acetaldehyde (15 eq); at 30°C, 750 rpm, 24 h; determined by GC-MS analysis.

Interestingly, it turned out that the addition of either cofactor NADPH or NADH was not favorable for the formation of the desired aldehyde 8. Hence, when NADH was added, the oxidation of alcohol was inhibited, whereas when adding NADPH the formed aldehyde was further oxidised to the corresponding carboxylic acid. Therefore, NADPH is assumed to be the required cofactor of the enzymes involved in the oxidation process; nevertheless, no benefit for this purpose was obtained by its addition (Wuensch et a!., 2013). The aldehyde 8 can further be isolated from the aqueous phase and separated from the starting material and from the hydrogen acceptor acetaldehyde, before it is employed in the synthesis of the pregabalin precursor 10c. Moreover, the above-described enzymatic oxidation can be further optimized on the preparative scale.

Hvdroqenation of α,β-unsaturated aldehyde

19 rt, 24 h

Scheme 24: Hydrogenation of unsaturated aldehyde 19 employing 10% Pd/C.

The hydrogenation of unsaturated aldehyde 19 was first tested in ethanol and methanol as these are generally suitable solvents for hydration reactions. Formation of the desired product was quantitative and no hydrogenation of the aldehyde moiety was observed. Nevertheless, the isolation of the product 8 from methanol and ethanol turned out to be challenging, in the form of aldehyde as well as in the form of enamine 9 which was isolated in the PCC method (see Example 2). Thus, dichloromethane and toluene were investigated as solvents for the hydrogenation, and quantitative conversion to the aldehyde 8 was obtained in both cases.

The hydrogenation of 4-methyl-2-pentenal 19 in toluene with 10% palladium on activated charcoal was performed on a preparative scale (30 mmol) (see Scheme 24 above), and the formed aldehyde was applied for the synthesis of pregabalin precursor 10c, which was obtained in 15% isolated yield, as described in the following.

4-Methyl-2-pentenal 19 (3.5 mL, 30 mmol, 1.0 eq) was dissolved in toluene (120 mL) and 10% Pd/C (250 mg) was added and the mixture was stirred under H₂ atmosphere for 24 h. Pd/C was removed over celite and diisobutylamine (5.22 mL, 30 mmol, 1.0 eq) was added to the mixture and refluxed for 8 h under Dean Stark apparatus. 3 A molecular sieve was added and the mixture was stirred overnight. Molecular sieve was removed by filtration over celite and bromo fer-butyl acetate (6.68 mL, 45 mmol, 1.5 eq) was added. The reaction was refluxed for 8 h, then acetate buffer (70 mL, 35 g acetic acid, 58 g sodium acetate trihydrate, 70 mL H₂0) was added, the suspension was stirred for 1 h at rt and the organic phase was washed with saturated sodium carbonate solution. The organic phase was dried over sodium sulphate and the solvent was evaporated. The product was purified by column chromatography (silica, petrolether/EtOAc 97:3) to give an orange oil 10c (957 mg, 4.4 mmol, 15 % yield overall). NMR: fer-butyl 3-formyl-5-methylhexanoate

¹H-NMR (300 MHz, CDCI₃): δ_Η [ppm] = 0.94 (t, ³J_5,6 = 6.6 Hz, 6 H, -CH₃), 1.25 (m_c, 1 H, -CH(CH₃)₂) , 1.44 (s, 9 H, -OC(CH₃)₃, 1 .56-1.72 (m, 2 H, 4-H), 2.55 (dd, ³J_2ai3= 5.1 Hz, ²J_2a,2b= 16.2 Hz, 1 H, 2-H_a), 2.60 (dd, ³J_2b>3= 8.1 Hz, ²J_2b2s= 15.0 Hz, 1 H, 2-H_b), 2.80 (m_c, 1 H, 3-H), 9.70 (dd, ³JCHO,3 = 1.5 Hz, 1 H, -CHO).

¹³C-NMR (75MHz, CDCI₃): 5_C [ppm] = 22.40 (-CH₃), 22.73 (-CH₃), 25.60 (-CH(CH₃)₂), 28.02 (-OC(CH₃)₃), 35.12 (2-C), 37.69 (4-C), 46.12 (3-C), 81.04 (-OC(CH₃)₃), 171.10 (OOO), 203.38 (C=0).

Moreover, the hydrogenation of 4-methyl-2-pentenal 19 was also investigated using Pt0₂ as catalyst. However, due to its tendency to reduce also the aldehyde moiety to the alcohol, Pd/C was chosen to be the more suitable catalyst.

Example 11 : Attempt to reproduce the transaminase-based synthesis of pregabalin taught in WO 2008/127646 (Comparative Example)

WO 2008/127646 describes the synthesis of pregabalin via the chemical or enzymatic amination of an aldehyde precursor. While the use of transaminases is mentioned in this document, WO 2008/127646 fails to provide any corresponding experimental examples. It was attempted to reproduce the teaching of this document using five of the specific transaminases listed therein for the amination of the oxo-acid precursor of pregabalin. However, as described in the following , three of these five transaminases (i.e. , ω-transaminase from Alcaligenes denithficans (AD-ω-ΤΑ) and two ω-transaminases from Pseudomonas putida (ΡΡ1 -ω-ΤΑ and ΡΡ2-ω-ΤΑ)) did not convert the substrate at all, the fourth transaminase (i.e., ω-transaminase from Chromobacterium violaceum (CV-ω-ΤΑ)) only gave the racemic product, and the fifth transaminase (i.e. , ω-transaminase from Vibrio fluvialis (VF-ω-ΤΑ)) gave pregabalin only at a low ee (24% of the (S)-isomer).

10e I0f 12f

Scheme 25: Demethyiation and subsequent reductive amination to form 12f (salt form of pregabalin). As shown in Scheme 25 above, the methyl group of the ester 10e was cleaved employing Lipase B from Candida antarctica (CalB) to form the free acid 10f. This ester cleavage was performed in phosphate buffer pH 7 and the progress of the reaction was followed via Thin Layer Chromatography (TLC) and LC-MS. After 2 h the demethylation was completed and lyophilized cells of E. coli containing overexpressed transaminases as well as alanine and all components of the Ala-recycling system were added to the reaction vessel.

After 24 h the formation of pregabalin-salt 12f was verified via LC-MS and the amino acid was then methylated again in order to perform chiral analysis of the product. As shown in Scheme 26, trimethylsilyI-(TMS)diazomethane was employed for the methylation, and spontaneous cyclisation of the product 12e gave lactam 13 which was analysed via GC on a chiral stationary phase.

Scheme 26: Methylation and subsequent spontaneous lactam formation. Experimental procedure All experiments were performed on a 1 mL scale using 2 mL Eppendorf tubes and were done in duplicate.

Substrate 10e (10 mg, 50 mM) and CalB (15 mg, 75 U) were incubated in phosphate buffer (1 mL, pH 7, 100 mM) at 30°C and 700 rpm for 3 h. The reaction was followed via TLC (hexanes/EtOAc 80:20) and LC-MS. Then lyophiiised cells of E. coli containing overexpressed ω-ΤΑ (20 mg), PLP (1 mM), L-alanine (22 mg, 250 mM), Ala-DH (12 U total activity), FDH (1 1 U total activity), NAD free acid (1 mM), and ammonium formate (9.5 mg, 150 mM) were added. The transamination was performed at 30°C and 700 rpm for 24h. The formation of pregabalin- salt 12f was checked via LC-MS.

In order to determine the ee of the product, the amino acid 12f was esterified: cell material was removed by centrifugation (10 min, 13000 rpm), the supernatant was separated and solvent was removed in a centrifugal evaporator. The residue was dissolved in methanol (700 μί), TMS-diazomethane (200 μΙ_, 2M solution in hexane) was added, and after 15 min the solvent was removed under vacuum. The residue was dissolved in NaHC0₃ sat. aqueous solution (300 μΙ_) and extracted with EtOAc (2 x 500 μ!_). The combined organic phases were dried over Na₂S0₄ and analysed via GC on a chiral stationary phase.

Results

The results obtained in this experiment are summarized in the following Table 31 :

Table 31 : Formation of pregabalin. Reaction conditions: substrate 10f (50 mM), ω-ΤΑ (lyophilised cells, 20 mg mL^~1) in phosphate buffer (100 mM, pH 7, 1 mM PLP) at 30°C, 120 rpm, 24 h. L-alanine (250 mM), Ala-DH (12 U total activity), FDH (1 1 U total activity), NAD^' (1 mM) and ammonium formate (150 mM). n.d. = not determined due to low conversion.

The amination of the free oxo acid 10f was tested employing five transaminases in accordance with WO 2008/127646. However, only CV-ω-ΤΑ and VF-ω-ΤΑ were found to convert the substrate 10f. Moreover, the product obtained by CV-ω-ΤΑ was racemic, and VF-ω-ΤΑ formed the (S)-enantiomer of amino acid 12f (pregabalin) in only 24 % ee.

These results demonstrate that the teaching of WO 2008/127646 with respect to the use of transaminases for the synthesis of pregabalin is largely defective and lacks clear instructions that would directly and necessarily lead to the stereoselective production of pregabalin having the (S)-configuration.

REFERENCES

Andruszkiewicz R et al. Synthesis. 1989; 12:953-5.

Bickerstaff Jr GF. Immobilization of enzymes and cells. Humana Press, 1997.

Bornscheuer UT. Angew Chem Int Ed Engl. 2003;42(29):3336-7. Cannio R et al. Eur J Biochem. 1994;222(2):345-52.

Ceccarelli C et al. Biochemistry. 2004;43:5266-77.

Dominguez de Maria P et al. J Mot Cat B: Enzym. 2005;37:36-46.

Edegger K et al. Chem Commun (Camb). 2006;22:2402-4.

Fuchs CS et al. Application of ω-transaminases for the asymmetric synthesis of β-chiral amines. Poster Presentation PO-1 15 at 15. Osterreichische Chemietage - Joint Meeting of the Hungarian & Austrian Chemical Societies that took place from September 23 to 26, 2013 in Graz, Austria.

Fuchs CS et al. Adv Synth Catal. 2014;356:2257-65.

Henke E et ai. Angew Chem int Ed. 2002;41 :321 1-3.

Hodgson DM et al. J Org Chem. 2009:74(3): 1019-28.

Hoekstra MS et al. Org Process Res Dev. 1997; 1 (1 ):26-38.

Hdhne M et al. Nat Chem Biol. 2010;6(1 1):807-13.

Homaei AA et al. J Chem Biol. 2013;6(4): 185-205.

Karabec M et al. Chem Commun. 2010;46:6314-6.

Kenda BM et al. J Med Chem. 2004;47(3):530-49.

Knezevic-Jugovic ZD et al. Methods Mol Biol. 201 ;679:99-1 1.

Koszelewski D et al. Journal of Molecular Catalysis B: Enzymatic. 2009;60(3-4): 191 -4.

Koszelewski D et al. Trends Biotechno!. 2010;28(6):324-32.

Kourist R et al. ChemBioChem. 2008;9:491 -8.

Krishna SH et al. Tetrahedron. Asymmetry 2002;13:2693-6.

Malik MS et al. Appl Microbiol Biotechnol. 2012;94(5): 163-71.

Ooi T et al. J Am Chem Soc. 2005; 127: 1 19-25.

Orbegozo T et al. Eur J Org Chem. 2010; 18:3445-8.

Park ES et al. Appl Microbiol Biotechnol. 2012;93(6):2425-35.

Park ES et al. ChemCatChem. 2013;5(7): 1734-8.

Savile CK et al. Science. 2010;329(5989):305-9.

Schelkun RM et al. Bioorg Med Chem Lett. 2006; 16:2329-32.

Schmidt M et al. J Org Chem. 2005;70:3737-40.

Schmidt Y et al. Chemistry. 2011 : 17(42): 1 780-8.

Schwarz JB et al. J Med Chem. 2005:48(8):3026-35.

Sheldon RA. Adv Synth Catal. 2007;349: 289-1307.

Shin JS et al. Biotechnol Bioeng. 1999;65(2):206-1 1 .

Shin JS et al. Biosci Biotechnol Biochem. 2001 ;65(8): 1782-8.

Simoneau B et al. Tetrahedron. 1988;44(4): 1015-22.

Steffen-Munsberg F et al. ChemCatChem. 2013;5(1): 150-3. (a)

Steffen-Munsberg F et al. ChemCatChem. 2013; 5(1 ): 154-7. (b) Sumitra D et al. 3 Biotech. 2013;3(1): 1-9.

Ueno H et al. J Med Chem. 1991 ;34(8):2468-73. Winkler CK et al. J Org Chem. 2013;78(4): 1525-33 Wuensch C et al. ChemCatChem. 2013;5:1744-8.

Claims

(I) wherein R¹ is a carboxy protecting group or hydrogen, and R² is alkyl or arylalkyl, is contacted either with an (ft)-seiective ω-transaminase or with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (I) and obtain a compound of the following formula (II)

(II) wherein R and R² are the same as in formula (I), and further wherein the (R)-isomer and the (S)-isomer of the compound of formula (II) refer to the following configurations:

2. The process of claim 1 , wherein R is selected from alkyl, alkenyl, aryl, arylalkyl, and hydrogen. The process of claim 1 or 2, wherein the compound of formula (I) is a racemate.

The process of any one of claims 1 to 3, wherein the ω-transaminase is an (R)-selective ω-transaminase, and wherein the compound of formula (II) is a compound of formula (ll-R)

(W-R) wherein R¹ and R² are the same as in formula (I).

5. The process of claim 4, wherein R² is n-propyl.

6. The process of claim 4 or 5, wherein the (R)-selective ω-transaminase is selected from ω-transaminase from Hyphomonas neptunium, ω-transaminase from Arthrobacter citreus, ω-transaminase from Chromobacterium violaceum DSM 30191 , ω-transaminase from Pseudomonas fluorescens, ω-transaminase from Pseudomonas putida KT2440 gene PP2180, ω-transaminase from Ochrobactrum anthropi, ω-transaminase from Silicibacter pomeroyi, ω-transaminase from Vibrio fluvialis, ω-transaminase from Neosartorya fischeri, ω-transaminase from Paracoccus denitrificans, ω-transaminase from Pseudomonas putida KT2440 gene PP5182, ω-transaminase from Arthrobacter sp. KNK168, ArR-G131 F-oo-TA, ArRmutl 1 -M1 17F-G279A-U)-TA, and a mutein of any one of the aforementioned ω-transaminases.

7. The process of any one of claims 4 to 6, wherein the ( )-selective ω-transaminase is ω-transaminase from Hyphomonas neptunium or a mutein thereof.

8. The process of any one of claims 1 to 3, wherein the ω-transaminase is an (S)-selective ω-transaminase, and wherein the compound of formula (II) is a compound of formula (ll-S)

(US) wherein R¹ and R² are the same as in formula (I).

9. The process of claim 8, wherein R² is selected from isobutyl, n-propyl, isopropyl, ethyl, and benzyl.

10. The process of claim 8 or 9, wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibberella zeae, ω-transaminase from Aspergillus terreus, ω-transaminase from Arthrobacter sp. KNK168, ω-transaminase from Ralstonia eutropha, ω-transaminase from Bacillus megaterium, ArRmutl 1 -ω-ΤΑ, ArR-S218P- G131 F-to-TA, ArRmutl 1 - 1 17F-00-TA, ArRmutl 1 -G279A-U)-TA, ArRmutl 1 -M1 17F- Α60ν-ω-ΤΑ, ArRmutl 1 -M 1 17F-A60V-G279A-u)-TA, ArRmutl 1-M1 17F-A60V-G279V-U)- TA, ArRmutl 1-M1 17F-A60V-G279L-00-TA, ArRmutl 1 -M1 17F-A60V-G279I-U)-TA, ArRmutl 1 -M 1 17F-A60V-G279F-u>-TA, ArRmutl 1 -M 1 17F-A60V-G279V-1152V-00-TA, ArRmutl 1-M1 17F-A60V-G279V-S277T- -TA, GZ-L56V-U)-TA. GZ-V6OT-00-TA. GZ-E1 15T-u>-TA, GZ-T273S-oo-TA, GZ-E1 15D-C0-TA, GZ-V148A-to-TA, GZ-L56V- T273S-U-TA, GZ-T273S-A275V-u>-TA, and a mutein of any one of the aforementioned uj-transaminases.

1 1. The process of any one of claims 8 to 10, wherein the (S)-selective ω-transaminase is selected from ω-transaminase from Gibberella zeae, ArRmutl 1 -M 1 17F-A60V- G279V-00-TA, ArRmutl 1 -M1 17F-A60V-G279I-U)-TA, GZ-L56V- -TA, and GZ-T273S-0J-TA.

12. The process of any one of claims 1 to 1 1 , wherein the amine, in the presence of which the compound of formula (I) is contacted with the ω-transaminase, is selected from an amino acid, an alkylamine, an alkenylamine, an arylamine, an aralkylamine, an arylalkenylamine, a heteroarylamine, a heteroarylalkylamine, a heteroarylalkenylamine, pyridoxamine-5'-phosphate (PMP), and any of the aforementioned compounds substituted with at least one further amino group.

13. The process of claim 12, wherein the amine is selected from alanine, glutamate, ethylamine, 1 ,2-diaminoethane, n-propylamine, 1 ,3-diaminopropane, isopropylamine, 1 ,2-diaminopropane, 2-butylamine, 1 -indolamine, 1 -indanamine, 1 -aminotetralin, and 1 -phenylethylamine.

14. A process for the production of brivaracetam, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (la)

wherein R is the same as in formula (la), wherein the process optionally further comprises a step of converting the compound of formula (lla) obtained in the stereoselective reductive amination step into a compound of the following formula (Ilia)

(IVa).

A process for the production of pregabalin, the process comprising a combined racemization and stereoselective reductive amination step in which a compound of the following formula (Ib)

(Ib) wherein R¹ is a carboxy protecting group or hydrogen, is contacted with an (S)-selective ω-transaminase in the presence of an amine to racemize the compound of formula (Ib) and obtain a compound of the following formula (lib)

(lib) wherein R¹ is the same as in formula (lb), wherein, if R is different from hydrogen, the process further comprises a step of converting the compound of formula (lib) into pregabalin having the following formula (Vb)