*2.9. Antipsychotic Drug: Enzymatic Reduction of 1-(4-Fluorophenyl)4-[4-(5-Fluoro-2-Pyrimidinyl)1-Piperazinyl]-1-Butanone*

The sigma receptor system in the brain and endocrine tissue has been target for development of new class of antipsychotic drugs [67,68]. Compound (*R*)-**31** (Figure 9) is a sigma ligand and has a high affinity for sigma binding site and antipsychotic efficacy. The enantioselective microbial reduction process was developed for the conversion of ketone **32** to both enantiomers of alcohol **31** [69]. Various microorganisms screened for the enatioselective reduction of 1-(4-fluorophenyl)4-[4-(5-fluoro-2-pyrimidinyl)1-piperazinyl]-1butanone **32**. From this screen, *Mortierella ramanniana* ATCC 38191 was identified to predominantly reduced compound **32** to (*R*)-**31**, while *Pullularia pullulans* ATCC 16623 was identified to predominantly reduced compound **32** to (*S*)-**31**. A single stage fermentation/biotransformation process was developed. Cells of *M. ramanniana* were grown in a 20-L fermentor and after 40 h growth period, the biotransformation process was initiated by addition of 40 g ketone **32** and 400 g glucose. The biotransformation process was completed in 24 h with a reaction yield of 100% and an e.e. of 98.9% for (*R*)-**31**. At the end of the biotransformation process, cells were removed by filtration and product was recovered from the filtrate in overall 80% yield [69].

**Figure 8.** Anticancer drug: Enzymatic preparation of C-13 paclitaxel side-chain synthon.

*2.10. Retinoic Acid Receptpor Agonist: Enzymatic Preparation of 2-(R)-Hydroxy-2-(1',2',3',4'-Tetrahydro-1',1',4',4'-Tetramethyl-6'-Naphthalenyl)Acetate* 

Retinoic acid and its natural and synthetic analogs (retinoids) exert a wide variety of biological effects by binding to or activating a specific receptor or sets of receptors [70]. They have been shown to effect cellular growth and differentiation and are promising drugs for the treatment of cancers [71]. A few retinoids are already in clinical use for the treatment of dermatological diseases such as acne and psoriasis. (*R*)-3-Fluoro-4-[[hydroxy-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl -2-naphthalenyl)-acetyl]amino]benzoic acid **33** (Figure 10) is a retinoic acid receptor gamma-specific agonist potentially useful as a dermatological and anticancer drug [72].

**Figure 9.** Antipsychotic drug: Enzymatic reduction of 1-(4-fluorophenyl)4 -[4-(5-fluoro-2-pyrimidinyl)1-piperazinyl]-1-butanone.

**Figure 10.** Retinoic acid receptor agonist: Enzymatic preparation of 2-(*R*)-hydroxy-2-(1ƍ,2ƍ,3ƍ,4ƍ-tetrahydro-1ƍ,1ƍ,4ƍ,4ƍ-tetramethyl-6ƍ-naphthalenyl)acetate.

Ethyl 2-(*R*)-hydroxy-2-(1ƍ,2ƍ,3ƍ,4ƍ-tetrahydro-1ƍ,1ƍ,4ƍ,4ƍ-tetramethyl-6ƍ-naphthalenyl)acetate **34** and the corresponding acid **35** were prepared as intermediates in the synthesis of the retinoic acid receptor gamma-specific agonist [73]. Enantioselective microbial reduction of ethyl 2-oxo-2-(1ƍ,2ƍ,3ƍ,4ƍ-tetrahydro-1ƍ,1ƍ,4ƍ,4ƍ-tetramethyl-6-naphthalenyl) acetate **36** to alcohol **34** was carried out using *Aureobasidium pullulans* SC 13849 in 98% yield and with an e.e. of 96%. At the end of the reaction, hydroxyester **34** was adsorbed onto XAD-16 resin and, after filtration, recovered in 94% yield from the resin with acetonitrile extraction. The recovered (*R*)-hydroxyester **34** was treated with Chirazyme L-2 or pig liver esterase to convert it to the corresponding (*R*)-hydroxyacid **35** in quantitative yield. The enantioselective microbial reduction of ketoamide **37** to the corresponding (*R*)-hydroxyamide **38** by *A. pullulans* SC 13849 has also been demonstrated [73].

#### *2.11. Anti-Alzheimer's Drugs: Enzymatic Reduction of 5-Oxohexanoate and 5-Oxohexanenitrile*

Ethyl-(*S*)-5-hydroxyhexanoate **39** and (*S*)-5-hydroxyhexanenitrile **40** (Figure 11) are key chiral intermediates in the synthesis of anti-Alzheimer's drugs [74]. Both chiral compounds have been prepared by enantioselective reduction of ethyl-5-oxohexanoate **41** and 5-oxohexanenitrile 42 by *Pichia methanolica* SC 16116 [75]. Reaction yields of 80%–90% and >95% e.e.s were obtained for each compound. In an alternate approach, the enzymatic resolution of racemic 5-hydroxyhexane nitrile **43** by enzymatic succinylation was demonstrated using immobilized lipase PS-30 to obtain (*S*)-5-hydroxyhexanenitrile **40** in 35% yield (maximum yield is 50%). (*S*)-5-Acetoxy-hexanenitrile **44** was prepared by enantioselective enzymatic hydrolysis of racemic 5-acetoxyhexanenitrile **45** by *Candida antarctica* lipase. A reaction yield of 42% and an e.e. of >99% were obtained [75].

**Figure 11.** Anti-Alzheimer's drugs: Enzymatic reduction of 5-oxohexanoate and 5-oxohexanenitrile.

*2.12. Enantioselective Microbial Reduction of Substituted Acetophenone* 

The chiral intermediates (*S*)-1-(2ƍ-bromo-4ƍ-fluorophenyl)ethanol 46 and (*S*)-methyl 4-(2ƍ-acetyl-5ƍ-fluorophenyl)-butanol **47** are potential intermediates for the synthesis of several potential anti-Alzheimer's drugs [76]. The chiral intermediate (*S*)-1-(2ƍ-bromo-4ƍ-fluoro phenyl)ethanol **46** (Figure 12A) was prepared by the enantioselective microbial reduction of 2-bromo-4-fluoro acetophenone **48** [77]. Organisms from genus *Candida*, *Hansenula*, *Pichia*, *Rhodotorula*, *Saccharomyces*, *Sphingomonas* and Baker's yeast reduced **48** to **46** in >90% yield and 99% enantiomeric excess (e.e.).

In an alternate approach, the enantioselective microbial reduction of methyl-4-(2ƍ-acetyl-5ƍ-fluorophenyl) butanoates **49** (Figure 12B) was demonstrated using strains of *Candida* and *Pichia*. Reaction yields of 40%–53% and e.e.s of 90%–99% were obtained for the corresponding (*S*)-hydroxy esters **47**. The reductase which catalyzed the enantioselective reduction of ketoesters was purified to homogeneity from cell extracts of *Pichia methanolica* SC 13825. It was cloned and expressed in *Escherichia coli* and recombinant cultures were used for the enantioselective reduction of the keto-methyl ester **49** to the corresponding (*S*)-hydroxy methyl ester **47**. On preparative scale, a reaction yield of 98% with an enantiomeric excess of 99% for **47** was obtained [77].

**Figure 12.** (**A**) Anti-Alzheimer's drugs: Enantioselective microbial reduction of substituted acetophenone; (**B**) Enantioselective microbial reduction of methyl-4-(2'-acetyl-5'-fluorophenyl) butanoates.

*2.13. Anticancer Drug: Enzymatic Preparation of (S)-2-Chloro-1-(3-Chlorophenyl)Ethanol* 

The synthesis of the leading candidate compound **50** [78] in an anticancer program (IGF-1 receptor inhibitors) [79,80] required (*S)*-2-chloro-1-(3-chlorophenyl)ethanol **51** (Figure 13) as an intermediate. Other possible candidate compounds used are analogs of (*S*)-alcohol **51**. From microbial screen of the reduction of ketone **52** to (*S*)-alcohol **51**, two cultures namely *Hansenula polymorpha* SC13824 (73.8% enantiomeric excess) and *Rhodococcus globerulus* SC SC16305 (71.8% enantiomeric excess) were identified that had the highest enantioselectivity. A ketoreductase from *Hansenula polymorpha*, after purification to homogeneity, gave (*S*)-alcohol **51** with 100% ee [81]. The ketoreductase was cloned and expressed in *E. coli* together with a glucose-6-phosphate dehydrogenase from *Saccharomyces cerevisiae* to allow regeneration of the NADPH required for the reduction process. An extract of *E. coli* containing the two recombinant enzymes was used to reduce 2-chloro-1-(3-chloro-4fluorophenyl)ethanone **52**. Intact *E. coli* cells provided with glucose were used to prepare (*S)*-2-chloro-1-(3-chloro-4-fluorophenyl)ethanol **51** in 89% yield with 100% e.e. [81].

#### *2.14. Thrombin Inhibitor: Enzymatic Preparation of (R)-2-Hydroxy-3,3-Dimethylbutanoic Acid*

Thrombin is a trypsin-like protease enzyme that plays a critical role in intrinsic and extrinsic blood coagulation. As a result of the enzymatic activation of numerous coagulation factors, thrombin is activated to cleave fibrinogen, producing fibrin, which is directly responsible for blood clotting. An imbalance between these factors and their endogenous activators and inhibitors can give rise to a number of disease states such as myocardial infarction, unstable angina, stroke, ischemia, restenosis following angioplasty, pulmonary embolism, deep vein thrombosis, and arterial thrombosis [82,83]. Consequently, the aggressive search for a potent, selective, and bioavailable thrombin inhibitor is widespread [84]. An intensive effort by Merck has led to the identification of thrombin inhibitor **53** [85]. The synthesis of **53** required a key chiral intermediate (*R*)-hydroxy ester **54**. An enzymatic process was developed for the asymmetric reduction of ketoester **55** to (*R*)-**54** using commercially available ketoreductase KRED1001 (Figure 14). The cofactor NADPH required for this reaction was regenerated using glucose dehydrogenase. The hydroxy ester (*R*)-**54** was isolated as an oil and then saponified to the corresponding enantiomerically pure hydroxy acid (*R*)-**56** without epimerization [86]. The enantiomerically pure (*R*)-**56** was obtained in 82% isolated yield (>99.5% e.e.).

**Figure 13.** Anticancer drug: Enzymatic preparation of (*S*)-2-chloro-1-(3-chlorophenyl)ethanol.

**Figure 14.** Thrombin inhibitor: Enzymatic preparation of (*R*)-2-Hydroxy-3,3-dimethylbutanoic acid.

*2.15. Endothelin Receptor Antagonist: Enantioselective Microbial Reduction of Keto Ester and Chloroketone* 

Endothelin is present in elevated levels in the blood of patients with hypertension, acute myocardial infarction and pulmonary hypertension. Two endothelin receptor sub-types have been identified which bind endothelin, thus causing vasoconstriction [87,88]. Endothelin receptor antagonists such as compound **57** (Figure 15) have potential therapeutic value. Synthesis of compound **57** required two key chiral intermediates (*S*)-alcohols **58** and **59**. Enantioselective microbial reduction of a ketoester **60** and a chlorinated ketone **61** to their corresponding (*S*)-alcohols **58** and **59** was demonstrated using *Pichia delftensis* MY 1569 and *Rhodotorula piliminae* ATCC 32762 to afford desired products in >98% e.e. and >99% e.e, respectively [89]. Reductions were scaled up to 23 L to produce the desired (*S*)-alcohols in 88% and 97% yields, respectively.

**Figure 15.** Endothelin receptor antagonist: Enantioselective microbial reduction of keto ester and chloroketone.

*2.16. Calcium Channel Blocker: Preparation of [(3R-cis)-1,3,4,5-Tetrahydro-3-Hydroxy-4-(4-Methoxyphenyl)-6-(Trifluromethyl) -2H-1-Benzazepin-2-One]* 

Diltiazem **62** (Figure 16) a benzothiazepinone calcium channel blocking agent that inhibits influx of extracellular calcium through L-type voltage-operated calcium channels, has been widely used clinically in the treatment of hypertension and angina [90]. Since diltiazem has a relatively short duration of action [91], an 8-chloro derivative recently has been introduced into the clinic as a more potent analogue [92]. Lack of extended duration of action and little information on structure-activity relationships in this class of compounds led Floyd *et al*. [93] to prepare isosteric 1-benzazepin-2-ones; this led to identification of (*cis*)-3-(acetoxy)-1-[2-(dimethylamino)ethyl] -1,3,4,5-tetrahydro-4-(4-methoxyphenyl)-6-trifluoromethyl)-2H-1-benzazepin-2-one **63** as a longer lasting and more potent antihypertensive agent. A key intermediate in the synthesis of this compound was (3*R*-*cis*)-1,3,4,5-tetrahydro-3-hydroxy-4-(4-methoxyphenyl)-6-(trifluromethyl) -2*H*-1-benzazepin-2-one **64**. An enantioselective process was developed for the reduction of 4,5-dihydro-4-(4-methoxyphenyl)-6-(trifluoromethyl)-1*H*-1-benzazepin-2,3-dione **65** to **64** using *Nocardia salmonicolor* SC 6310, in 96% reaction yield with 99.8% e.e. [94].

**Figure 16.** Calcium channel blocker: Preparation of [(3*R*-*cis*)-1,3,4, 5-tetrahydro-3-hydroxy-4-(4-methoxyphenyl)-6-(trifluromethyl)-2*H*-1-benzazepin-2-one].

#### *2.17.* E*3-Receptor Agonist: Reduction of 4-Benzyloxy-3-Methanesulfonylamino-2'-Bromo-Acetophenone*

E3-Adrenergic receptors are found on the cell surfaces of both white and brown adipocytes and are responsible for lipolysis, thermogenesis, and relaxation of intestinal smooth muscle [95]. Consequently, several research groups are engaged in developing selective E3 agonists for the treatment of gastrointestinal disorders, type II diabetes, and obesity [96,97]. Biocatalytic syntheses of chiral intermediates required for the total synthesis of E3 receptor agonists **66** (Figure 17) has been demonstrated [98].

**Figure 17.** E3-Receptor agonist: Reduction of 4-benzyloxy-3-methanesulfonylamino-2-bromo-acetophenone.

The microbial reduction of 4-benzyloxy-3-methanesulfonylamino-2ƍ-bromo-acetophenone **67** to the corresponding (*R*)-alcohol **68** has been demonstrated [98] using *Sphingomonas. paucimobilis*  SC 16113. The growth of *S. paucimobilis* SC 16113 was carried out in a 750-L fermentor and harvested cells (60 kg) were used to conduct the biotransformation in 10-L and 200-L preparative batches using 20% (wt/vol, wet cells). In some batches, the fermentation broth was concentrated 3-fold by microfilteration and subsequently washed with buffer by diafilteration and used directly in the bioreduction process. In all the batches, reaction yields of >85% and e.e.s. of >98% were obtained.

The isolation of alcohol **68** from the 200-L batch gave 320 g (80% yield) of product with an e.e. of 99.5%.

In an alternate process, frozen cells of *S. paucimobilis* SC 16113 were used with XAD-16 hydrophobic resin (50 g/L) adsorbed substrate at 10 g/L concentration. In this process, an average reaction yield of 85% and an e.e. of >99% were obtained for alcohol **68**. At the end of the biotransformation, the reaction mixture was filtered on a 100 mesh (150 ȝm) stainless steel screen, and the resin retained by the screen was washed with water. The product was then desorbed from the resin with acetonitrile and crystallized in 75% overall yield with a 99.8% e.e.[98].

## *2.18. Penem and Carbapenem: Enzymatic Preparation of (R)-1,3-Butanediol and (R)-4-Chloro-3-Hydroxybutonoate*

(*R*)-1,3-Butanediol **69** (Figure 18) is a key starting material of azetidinone derivatives **70**, which are key chiral intermediates for the synthesis of penem **71** and carbapenem antibiotics [99]. From a microbial screen*,* the *Candida parapsilosis* strain IFO 1396 was identified which produced (*R*)-1,3-butanediol from the racemate. The (*S*)-1,3-butanediol oxidizing enzyme (CpSADH) which produced (*R*)-1,3-butanediol from the racemate was cloned in *Escherichia coli*. The recombinant culture catalyzed the enantioselective oxidation of secondary alcohols and also catalyzed the asymmetric reduction of aromatic and aliphatic ketones to their corresponding (*S*)-secondary alcohols. Using the recombinant enzyme, (*R*)-1,3-butanediol was produced in 97% yield and 95% e.e. using 150 g/L input of the racemate. Recombinant enzyme (CpSADH) was also used for reduction of ethyl 4-chloroacetoacetate **72** to produce ethyl-(*R*)-4-chloro-3-hydroxybutonoate **73** in 95% yield and 99% e.e. using 36 g/L substrate input. Isopropanol was used to regenerate the NADH required for this reduction. Ethyl-(*R*)-4-chloro-3-hydroxybutonoate is useful for the synthesis of L-carnitine **74** and (*R*)-4-hydroxyl pyrrolidone **75** [100,101]).

**Figure 18.** Penem and carbapenem: Enzymatic preparation of (*R*)-1,3-butanediol and (*R*)-4-chloro-3-hydroxybutonoate.

#### *2.19. Integrin Receptor Agonist: Enzymatic Preparation of (R)-Allylic Alcohol*

(*R*)-allylic alcohol **76** (Figure 19) was required as an intermediate for the synthesis of a desired monanoic derivate useful as an integrin receptor antagonist for the inhibition of bone desorption and treatment of osteoporosis [102]. A pilot scale whole cell process was developed for the enantioselective 1,2-reduction of prochiral alpha,beta-unsaturated ketone **77** to (*R*) allylic alcohol, (*R*)-**76** using *Candida chilensis* [103]. Initial development showed high enantiomeric excess (>95%) but low product yield (10%). Further process development, using a combination of statistically designed screening and optimization experiments, improved the desired alcohol yield to 90%. The fermentation growth stage, particularly medium composition and growth pH, had a significant impact on the bioconversion while process characterization identified diverse challenges including the presence of multiple enzymes, substrate/product toxicity, and biphasic cellular morphology. Manipulating the fermentation media allowed control of the whole cell morphology to a predominantly unicellular broth, away from the viscous pseudohyphae, which were detrimental to the bioconversion. The activity of a competing enzyme, which produced the undesired saturated ketone **78** and (*R*)-saturated alcohol 79, was minimized to < or =5% by controlling the reaction pH, temperature, substrate concentration, and biomass level. Despite the toxicity effects limiting the volumetric productivity, a reproducible and saleable process was demonstrated at pilot scale with high enantioselectivity (e.e. > 95%) and overall yield greater than 80% [104]. The whole cell approach proved to be a valuable alternative to chemical reduction routes.

**Figure 19.** Integrin receptor agonist: Enzymatic preparation of (*R*)-allylic alcohol.

*2.20. NK1 Receptor Antagonists: Enzymatic Synthesis of (S)-3,5-Bistrifluoromethylphenyl Ethanol* 

The synthesis of (*S*)-3,5-bistrifluoromethylphenyl ethanol, (*S*)-**80**, (Figure 20), an intermediate for the synthesis of NK-1 receptor antagonists **81** [104] was demonstrated from a ketone **82** via asymmetric enzymatic reduction process [105]. The isolated enzyme alcohol dehydrogenase from *Rhodococcus erythropolis* reduced the poorly water soluble substrate with an excellent enantiomeric excess (>99.9%) and good conversion (>98%). The optimized process was demonstrated up to pilot scale using concentration (390 mM) using a easy isolation process achieving overall isolation yields

(>90%). Process improvements at preparative scale, demonstrated increase in the substrate input to 580 mM achieving a space time yield of 260 g/L/day [105].

**Figure 20.** NK1 receptor antagonists: Enzymatic synthesis of (*S*)-3,5-bistrifluoromethylphenyl ethanol.

#### **3. Enzymatic Preparation of Chiral Amino Acids**

The reductive amination of D-keto acids using amino acid dehydrogenases to be one of the most useful methods because the enzymes have good stability, broad substrate specificity and very high enantioselectivity and can be used at high substrate concentrations as keto acids are soluble in aqueous system. The reductive aminations process coupled to an enzymatic cofactor regeneration system are most prominent method for preparation of chiral amino acids. For most enzymes, the required cofactor is NADH but NADPH is required in some cases. Yeast formate dehydrogenase is commonly used for NADH regeneration and glucose dehydrogenase usually from *Bacillus* species may be used for either NADH or NADPH regeneration. There are excellent reviews on the amino acid dehydrogenases and examples of their synthetic utilities [106–109].

#### *3.1. Tigemonam: Enzymatic Synthesis of (S)-ȕ-Hydroxyvaline*

(*S*)-ȕ-hydroxyvaline **83** (Figure 21), is a key chiral intermediate required for the total synthesis of orally active monobactam [110], Tigemonam **84**. Chiral amino acids have been made from corresponding keto acids by reductive amination process [111]. The synthesis of (*S*)-ȕ-hydroxyvaline **83** from D-keto-ȕ-hydroxyisovalerate **85** by reductive amination using leucine dehydrogenase from *Bacillus sphaericus* ATCC 4525 has been demonstrated [112]. The NADH required for this reaction was regenerated by either formate dehydrogenase from *Candida boidinii* or glucose dehydrogenase from *Bacillus megaterium*. The required substrate **85** was generated either from D-keto-ȕ-bromoisovalerate or its ethyl esters by hydrolysis with sodium hydroxide *in situ.* In this process, an overall reaction yield of 98% and an enantiomeric excess of 99.8% were obtained for the L-ȕ-hydroxyvaline **83**.

**Figure 21.** Tigemonam: Enzymatic synthesis of (*S*)-ȕ-hydroxyvaline.

#### *3.2. Atazanavir: Enzymatic Synthesis of (S)-Tertiary-Leucine*

Atazanavir **86** is an acyclic aza-peptidomimetic, a potent HIV protease inhibitor [53,54]. Synthesis of atazanavir required (*S*)-tertiary leucine **87** (Figure 22). An enzymatic reductive amination of ketoacid **88** to amino acid **87** by recombinant *Escherichia coli* expressing leucine dehydrogenase from *Thermoactinimyces intermedius* has been demonstrated. The reaction required ammonia and NADH as a cofactor. NAD produced during the reaction was converted back to NADH using recombinant *Escherichia coli* expressing formate dehydrogenase from *Pichia pastoris*. A reaction yield of >95% with an e.e. of >99.5% was obtained for **87** at 100 g/L substrate [113]. Leucine dehydrogenase from *Bacillus* strain has also been cloned and expressed and used in reductive amination process [114,115].

#### *3.3. Vanlev: Enzymatic Synthesis of (S)-6-Hydroxynorleucine*

Vanlev **89** (Figure 23) is an antihypertensive drug which acts by inhibiting angiotensin-converting enzyme (ACE) and neutral endopeptidase (NEP) [116]. (*S*)-6-Hydroxynorleucine **90** is a key intermediate in the synthesis of Vanlev. The synthesis and complete conversion of 2-keto-6-hydroxyhexanoic acid **91** to (*S*)-6-hydroxynorleucine **90** was demonstrated by reductive amination using beef liver glutamate dehydrogenase [117]. As depicted, compound **91**, in equilibrium with 2-hydroxytetrahydropyran-2-carboxylic acid sodium salt **92**, was converted to **90**. The reaction requires ammonia and NADH. NAD produced during the reaction was recycled to NADH by the oxidation of glucose to gluconic acid using glucose dehydrogenase from *Bacillus megaterium*. The reaction was complete in about 3 h at 100 g/L substrate input with a reaction yields of 92% and e.e. of 99.8% for (*S*)-6-hydroxynorleucine. The synthesis and isolation of keto acid **91** required several steps. In a second, more convenient process the ketoacid was prepared by treatment of racemic 6-hydroxy norleucine **90** [produced by hydrolysis of 5-(4-hydroxybutyl) hydantoin **93**] with (*R*)-amino acid oxidase (Figure 24) After the e.e. of the unreacted (*S*)-6-hydroxynorleucine had risen to 99.8%, the reductive amination procedure was used to convert the mixture containing the 2-keto-6-hydroxyhexanoic acid entirely to (*S*)-6-hydroxynorleucine in 97% yield with 99.8% e.e. from racemic 6-hydroxynorleucine at 100 g/L substrate input [117]. The (*S*)-6-hydroxynorleucine prepared by the enzymatic process was converted chemically to Valev **89** [118].

**Figure 22.** Atazanavir (anti-viral agent): Enzymatic synthesis of (*S*)-tertiary-leucine.

**Figure 24.** Vanlev: Enzymatic conversion of racemic 6-hydroxy norleucine to (*S*)-6-hydroxymorleucine.

#### *3.4. Vanlev: Enzymatic Synthesis of Allysine Ethylene Acetal*

(*S*)-2-Amino-5-(1,3-dioxolan-2-yl)-pentanoic acid [(*S*)-allysine ethylene acetal] **94** (Figure 25) is one of three building blocks used in an alternative synthesis of Vanlev **89**. Synthesis of **94** was demonstrated by reductive amination of ketoacid acetal **95** using phenylalanine dehydrogenase [PDH] from *Thermoactinomyces intermedius* [119]. The reaction required ammonia and NADH; NAD produced during the reaction was recyled to NADH by the oxidation of formate to CO2 using formate dehydrogenase [FDH].*T. intermedius* PDH was cloned and expressed in *Escherichia coli* and recombinant culture was used as a source of PDH. Expression of *T. intermedius* PDH in *P. pastoris*, inducible by methanol, allowed generation of both enzymes in a single fermentation as methanol grown cells of *P. pastoris* also contained formate dehydrogease. A total of 197 kg of **94** was produced in three 1600-L batches using a 5% concentration of substrate **95** with an average yield of 91 M % and e.e. >98% [119]. (*S*)-allysine ethylene acetal was converted to Vanlev **89** [118].

## *3.5. Saxagliptin: Enzymatic Reductive Amination of 2-(3-Hydroxy-1-Adamantyl)-2-Oxoethanoic Acid*

Dipeptidyl peptidase 4 (DPP-4) is a ubiquitous proline-specific serine protease responsible for the rapid inactivation of incretins, including glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic peptide. To alleviate the inactivation of GLP-1, inhibitors of DPP-IV are being evaluated for their ability to provide improved control of blood glucose for diabetics [120–123]. Januvia developed by Merck is a marketed DPP4 Inhibitor [122].

**Figure 25.** Vanlev: Enzymatic synthesis of allysine ethylene acetal.

Saxagliptin **96** [121,122] (Figure 26), a DPP-IV inhibitor developed by Bristol-Myers Squibb and now approved for type 2 diabetic treatment by Food and Drug administration, requires (*S*)-*N*-boc-3-hydroxyadamantylglycine **97** as an intermediate. A process for conversion of the keto acids **98** to the corresponding amino acid **99** using (*S*)-amino acid dehydrogenases was developed. A modified form of a recombinant phenylalanine dehydrogenase cloned from *Thermoactinomyces intermedius* and expressed in *Pichia pastoris* or *Escherichia coli* was used for this process. NAD produced during the reaction was recycled to NADH using formate dehydrogenase. The modified phenylalanine dehydrogenase contains two amino acid changes at the *C*-terminus and a 12 amino acid extension of the *C*-terminus [124].

**Figure 26.** Saxagliptin: Enzymatic reductive amination of 2-(3-hydroxy-1-adamantyl)-2-oxoethanoic acid.

Production of multi-kg batches was originally carried out with extracts of *Pichia pastoris*  expressing the modified phenylalanine dehydrogenase from *Thermoactinomyces intermedius* and endogenous formate dehydrogenase. The reductive amination process was further scaled up using a preparation of the two enzymes expressed in single recombinant *E. coli*. The amino acid **99** was directly protected as its boc derivative without isolation to afford intermediate. Yields before isolation were close to 98% with 100% e.e. [124].

Reductive amination was also conducted using cell extracts from *E. coli* strain SC16496 expressing PDHmod and cloned FDH from *Pichia pastoris*. Cell extracts after polyethyleneamine treatment, clarification and concentration were used to complete the reaction in 30 h with >96% yield and >99.9% e.e. of product **99**. This process has now been used to prepare several hundred kg of boc-protected amino acid **97** to support the development of Saxagliptin [124].

#### *3.6. Enzymatic Synthesis of (S)-Neopentylglycine*

The enantioselective synthesis of (*S*)-neopentylglycine **100** (Figure 27) has been developed by Groeger *et al.* [125]. Recombinant whole cell containing leucine dehydrogenase and formate dehydrogenase was used in the reductive amination of the corresponding D-keto acid **101**. The desired (*S*)-neopentylglycine was obtained with >95% conversion and a high enantioselectivity of >99% e.e. at substrate concentrations of up to 88 g/L. Spiroheterocyclic compounds [morpholine-4-carboxylic acid amides of heterocyclic cyclohexylalanine and neopentylglycine derivatives and their analogs] are useful as reversible inhibitors of cysteine proteases such as cathepsin S useful in the treatment of variety of autoimmune diseases [126].

**Figure 27.** Enzymatic synthesis of (*S*)-neopentylglycine.

*3.7. Glucogen like Peptide: Enzymatic Deracemization Racemic Amino Acid to (S)-Amino Acid* 

The (*S*)-amino-3-[3-{6-(2-methylphenyl)}pyridyl]-propionic acid 102 (Figure 28) is a key intermediate required for synthesis of GLP-1 mimics or GLP-1 receptor modulators. Such receptor modulators are potentially useful for the treatment of type II diabetes treatment [127,128].

(*S*)-Amino-3-[3-{6-(2-methylphenyl)}pyridyl]-propionic acid was prepared by enzymatic deracemization process [129] in 72% isolated yield with >99.4% e.e. from racemic amino acid **103** using combination of two enzymes (*R*)-amino acid oxidase from *Trigonopsis variabilis* expressed in *Escherichia coli* and (*S*)-aminotransferase from *Sporosarcina ureae* cloned and expressed in *Escherichia coli*. (*S*)-aspartate was used as amino donor. A (*S*)-aminotransferase was also purified from a soil organism identified as *Burkholderia* sp. and cloned and expressed in *Escherichia coli* and used in this process [131]. In enzymatic process racemic amino acid was first treated with (*R*)-amino acid oxidase for 4 h to convert racemic amino acid to mixture of (*S*)-amino acid and keto acid **104**. Subsequently in the same reaction mixture (*S*)-aminotransferase was charged to convert keto acid **104** to (*S*)-amino acid **102** to get 85% yield at the end of the biotransformation process. This process was scaled up to 100 L scale at a substrate input of 1.5 kg.

In an alternate process, the enzymatic dynamic resolution of racemic amino acid **103** was also demonstrated. (*R*)-selective oxidation with celite-immobilized (*R*)-amino acid oxidase from *Trigonopsis variabilis* expressed in *Escherichia coli* in combination with chemical imine reduction with borane-ammonia gave a 75% in process yield and 100 e.e. of (*S*)-amino acid **102** [129].

#### *3.8. Preparation of (R)-Amino Acid*

(*R*)-Amino acids are increasingly becoming important building blocks in the production of pharmaceuticals and fine chemicals, and as chiral directing auxiliaries and chiral synthons in organic synthesis [130,131]. Using both rational and random mutagenesis, Rozzell and Novick [132] have created the broad substrate range, nicotinamide cofactor dependent, and highly stereoselective (*R*)-amino acid dehydrogenase. This new enzyme is capable of producing (*R*)-amino acids via the reductive amination of the corresponding 2-keto acid with ammonia. This biocatalyst was the result of three rounds of mutagenesis and screening performed on the enzyme *meso*-diaminopimelate (*R*)-dehydrogenase from *Corynebacterium glutamicum*. The first round targeted the active site of the wild-type enzyme and produced mutants that were no longer strictly dependent on the native substrate. The second and third rounds produced mutants that had an increased substrate range including straight- and branched-aliphatic amino acids and aromatic amino acids. The very high selectivity toward the (*R*)-enantiomer (95% to >99% e.e.) was shown to be preserved three rounds of mutagenesis and screening [132]. This new enzyme was active against variety of amino acids could complement and improve upon current methods for (*R*)-amino acid synthesis. The synthesis of (*R*)-cyclohexylalanine 105 (Figure 29) was developed by reductive amination of cyclohexylpyruvate **106** to yield (*R*)-**105** in 98% yield and >99% e.e. (*R*)-105 is a potential chiral intermediate for the synthesis of thrombin inhibitor Inogatran **107** [133].

The deracemisation of DL-amino acids using L-amino acid oxidase from *Proteus myxofaciens*  and amine-boranes as chemical reducing agents has been investigated. Amine-boranes were found to be of particular interest in terms of reactivity and chemoselectivity compared to sodium borohydride and cyanoborohydride. Starting from the racemic amino acids, a range of D-amino acids were prepared in yields of up to 90% and e.e. >99% [134].
