Biocatalysis in Asymmetric Synthesis

Biocatalysis in Asymmetric Synthesis, a new volume in the Foundations and Frontiers of Enzymology series, offers an applied discussion of synthesizing biological catalysts using asymmetric synthesis, for applications across research and industry. Here, global experts in the field analyze a wide variety of biocatalysts and their physical states, process conditions for their asymmetric synthesis, solvents required during synthesis, and even downstream procedures for the recovery of final products. The book adopts an interdisciplinary approach, merging fundamental biology and its synthetic applications across industries, with a wide range of practical examples, from directed evolution to biotransformation and production of novel enzymes and non-conventional catalysts. Throughout the book, the impact and application of biocatalysis in sustainable processing is considered in-depth. This book will also help non-experts in biocatalysis to apply this knowledge in their own research, providing a thorough overview of the ways asymmetric biocatalytic approaches may be adapted for different disciplines and downstream products.

• Explores biocatalysts as exquisite catalysts for fine chiral compound synthesis in different reaction media
• Features both foundational overviews and applied, practical examples across research and industry
• Includes chapter contributions from international leaders in the field

• Language: English
• Publisher: Academic Press
• Release date: May 4, 2024
• ISBN: 9780443190582

Book preview

Preface

Gonzalo de Gonzalo and Andrés R. Alcántara

Since time immemorial, humanity has harnessed biotransformations involving microorganisms to acquire valuable products such as bread, vinegar, wine, or beer. These practices marked the outset of utilizing biological catalysts. However, it was not until the 1980s that the genuine significance of biocatalysis began to emerge as a valuable and complementary alternative to other catalytic methodologies for crafting high-value compounds. Notably, advancements in biochemistry and molecular biology over the past century have deepened our understanding of enzyme structure and catalytic mechanisms. Furthermore, the revelation of enzymatic activity in nonaqueous environments broadened the spectrum of applications for enzymes. In this century, studies in genetic engineering and directed evolution have facilitated the customization and optimization of enzymes for specific purposes. Tailoring enzymes to exhibit desired properties has notably enhanced their utility, particularly in the realm of asymmetric synthesis.

Undoubtedly, asymmetric biocatalysis stands as a dynamic and swiftly evolving field at the nexus of biology and chemistry, where biological catalysts assume a pivotal role in fashioning chiral molecules. The enzymatic capacity to selectively produce single enantiomers holds profound implications and significant advancements in the pharmaceutical, agrochemical, and fine chemical industries, offering a sustainable approach to synthesizing complex molecules.

This book aims to explore various facets of preparing chiral valuable compounds using biocatalysts. All elements related to the synthesis of these desired compounds will be discussed, employing an interdisciplinary approach to gain a deeper understanding of how this form of catalysis can be applied to the synthesis of chiral molecules. We trust that this book will kindle curiosity, inspire creativity, and cultivate a profound appreciation for the intricate interplay between biology and chemistry. Whether you are a student, a researcher, or an industry professional, we encourage you to delve into the forthcoming pages and accompany us on a journey into the realm of asymmetric biocatalysis.

Chapter 1

Introduction to asymmetric synthesis employing biocatalysts

Andrés R. Alcántara¹ and Gonzalo de Gonzalo²,    ¹Department of Chemistry in Pharmaceutical Sciences, Faculty of Pharmacy, Complutense University of Madrid, Madrid, Spain,    ²Department of Organic Chemistry, University of Seville, Seville, Spain

Abstract

Biocatalysis is a powerful tool in asymmetric synthesis, which refers to the production of chiral compounds in a selective manner. Enzymes, which are natural biocatalysts, have evolved to perform stereospecific reactions with high efficiency and selectivity, making them highly valuable in asymmetric synthesis. Biocatalysis can be used for the asymmetric synthesis of a wide range of chiral compounds, including pharmaceuticals, agrochemicals, and flavors. Enzymes such as hydrolases and oxidoreductases, among many others, have been widely used in various applications of asymmetric synthesis. One of the main advantages of these catalysts is that their use often allows for milder reaction conditions compared to traditional chemical methods. Additionally, biocatalysis is habitually more environmentally friendly, since it typically requires less energy and produces less waste, contributing in this way to the development of greener procedures, one of the goals looked for by chemists in recent times. Overall, since its discovery as a valuable synthetic tool by the end of the last century, biocatalysis has emerged as an important tool in asymmetric synthesis, with the potential to provide a more sustainable and efficient approach to the production of pure chiral compounds. In this chapter, some general concepts and methods related to the use of biocatalysts for asymmetric syntheses will be presented.

Keywords

Asymmetric biocatalysis; chirality; enantioselectivity parameters; kinetic resolution; dynamic-kinetic resolution; deracemizations; biocatalysts preparation; novel-to-nature processes

1.1 Introduction

Although human beings have used nature-made biocatalysts since the dawn of time as microbial processes (fermentations for bread, beer, vinegar, or wine production), their application in organic synthesis (a short definition of the term biocatalysis) started in the first third of the 19th century, with Liebig and Wöhler describing in 1837 for the first time the application of an enzyme: the crude preparation of hydroxynitrile lyase (termed emulsin by the authors) from the almond Prunus amygdalus (nowadays PaHNL), which was able to catalyze the release of HCN from (R)-mandelonitrile, also isolated from those almonds [1]. This and some other pioneer studies are considered the beginning of the first wave of biocatalysis [2–5]. Since then, some subsequent achievements have been described, associated to some crucial developments in biocatalysis that boosted its progress.

Thus the second wave, starting around the mid-1980s, was marked by the use of gene technology, facilitating the cloning and expression of the desired enzyme in appropriate microbial hosts and allowing rational and/or random mutagenesis of the biocatalysts. This technique enabled the production of new biocatalysts capable of being used in industrial processes, generally after an immobilization procedure, a technique already known but strongly developed during this second wave and still being useful [6–12]. Also in the mid-1980s, the pioneer studies from Klibanov and coworkers [13–18] showed how some enzymes (mainly lipases) were capable of working efficiently in pure water-immiscible organic solvents, therefore opening a new and extremely useful research area [19–24], which has even questioned the general assumption that water is the best solvent for biocatalysis [25]. The crest of this wave, and the beginning of the next one, was the Nobel Prize awarded to Kary Mullis in 1993 for the development of the polymerase chain reaction (PCR) technology [26].

In fact, the third wave started with the full and generalized implementation of the advanced protein engineering methods developed in the second wave, more specifically the DNA shuffling and error-prone polymerase chain reaction combined with high-throughput screening methods. This process, known as directed evolution, granted Frances Arnold the Nobel Prize in 2018 [27] and has facilitated the creation of à la carte biocatalysts, engineered to fit the process demands in terms of selectivity and robustness, even allowing biotransformations not previously known using enzymes [28–36].

Consequently, we are now facing what has been called the golden age of biocatalysis [37], as the rate of generation of new innovations is faster than ever. Actually, in recent publications, new trends in biocatalysis have been presented [5,34,38,39]; among them, artificial intelligence applied to the increasingly faster design of new biocatalysts [40,41], the implementation of biocatalysis in flow systems [42–44], and photobiocatalysis [45–47] are probably the techniques that would lead biocatalysis to the fourth wave.

Another aspect to be considered is the close relationship existing between biocatalysis and green chemistry. A brief and concise definition of this last one, as proposed by Roger Sheldon [48], is the "chemistry which efficiently utilizes (preferably renewable) raw materials, eliminates waste and avoids the use of toxic and/or hazardous reagents and solvents in the manufacture and application of chemical products. Accordingly, green chemistry intends primarily pollution prevention rather than waste remediation, following a benign by design" philosophy [49]. In this sense, most of the advantages associated to the use of biocatalyzed reactions are perfectly aligned with the principles of green chemistry, as defined by Anastas and Warner [50]. In fact,

• biocatalysis is obviously catalytic rather than stoichiometric, thus improving the ease of processing (principle #9, use of catalysts)

• biocatalysts are produced from readily available renewable resources (principle #7, use renewable feedstocks)

• biocatalysts are biodegradable (principle #10, design of biodegradable reagents and/or products)

• biocatalysts are basically nonhazardous and nontoxic (principle #3, less hazardous chemical synthesis; principle #4, safer chemicals and products)

• biocatalysts can substitute toxic metal-based catalysts, therefore avoiding costs associated with removing traces of these metals from end products (principle #3, less hazardous chemical synthesis; principle #4, safer chemicals and products)

• biotransformations are generally performed under mild conditions, at room temperature and atmospheric pressure (principle #6, increase energy efficiency)

• biotransformations are generally conducted in water or in sustainable biosolvents (principle #5, safer solvents and reaction conditions)

• last but not least, because of the highly enzymatic precision, biotransformations are carried out with excellent chemo-, regio-, and stereoselectivities, circumventing the need for functional-group activation, protection, and deprotection steps [principle #1, generate less waste; principle #6, increase energy efficiency; principle #8, avoid chemical derivatives (protecting groups)].

So, it is not surprising the high number of recent publications highlighting the green credentials of biocatalysis and its impact on the development of sustainable processes, both in the laboratory as well as in large-scale manufacturing of chemicals in industry [9,51–60]. Specifically, biocatalysis plays a preeminent role in the sustainable preparation of chiral building blocks for drug synthesis [61–72]. This is a clear consequence of the inherent enzymatic capability for recognizing asymmetry [67,73,74] derived from their unique selectivity, which makes them suitable for recognizing chirality (resolution of racemates), prochirality (either associated to sp² or sp³ prochiral compounds), and even axial chirality in atropoisomers, as shown in the next section.

1.2 Type of enzymatic processes for generating asymmetry

1.2.1 Chirality associated to sp³ carbon atoms

Enzyme’s capability for recognizing stereocenters associated to sp³ carbon atoms inside racemates has been traditionally explained by attending to the three-point attachment (TPA) model. This model, originally described independently by Easson and Stedman [75] and Ogston [76], postulates that the best recognized enantiomer of a chiral substrate is attached to a protein surface simultaneously at three sites, while the optical antipode, in any conformation, cannot be recognized at the same three sites, as shown in Fig. 1.1 (for compounds possessing only one stereocenter). Later, this model was renamed the three-point interaction (TPI) model, as the recognition between substrate and the enzyme is generally mediated by noncovalent interactions [77,78].

Figure 1.1 TPI schematic representation. TPI, Three-point interaction.

Some modifications of this TPI model have been proposed, such as the four-location (FL) model, assuming an extra interaction [78,79], but the TPI model is generally accepted. Whatever the model selected, there are different methodologies by which a biocatalyst can distinguish between enantiomers of a racemate, as we will comment in the following sections.

1.2.1.1 Kinetic resolutions (KRs)

This is a process based on the different rates at which both enantiomers of the racemic mixture react in the presence of the biocatalyst, as shown in Fig. 1.2. Under ideal conditions, the kinetic resolutions (KR) would lead to a mixture of two enantiopure compounds, namely, PR (the product resulting from the conversion of the fast-reacting enantiomer of the substrate (SR)) and SS (the remanent not-converted substrate); therefore because of the different chemical nature of PR and SS, they can be easily separated.

Figure 1.2 Schematic representation of an enzymatic kinetic resolution (R isomer of substrate (SR) is considered the fast-reacting enantiomer).

To measure the enzymatic stereoselectivity (enantioselectivity in this case), the enantiomeric excess (e.e.) is commonly used. This term was defined by Morrison and Mosher [80]: "Assuming a linear relationship between rotation and composition and no experimental error, percent ‘optical purity’ [op] is equated with the percent of one enantiomer over the other, which we shall designate percent enantiomeric excess (% e.e.)." Thus a value of e.e. > 99.9% (leaving a 0.1% error margin in the detection of the other enantiomer) represents an enantiopure compound (the term homochiral as a synonym for enantiopure is strongly discouraged by IUPAC [81]). On the other hand, it has been proposed that the use of e.e. should be substituted by enantiomeric ratio (e.r.), as the increasing implementation of spectroscopic and chromatographic methods has replaced polarimetry for determining enantiomeric composition [82]. This enantiomeric ratio can be expressed either as a ratio normalized as a percent (e.g., e.r.=80:20, also known as enantiomeric composition, e.c. [83,84]) or as a number, in that case denoted as q, which is a ratio having a denominator of 1 (e.g., q.=4/1 for the same example above). The use of q is recommended for an easier comparison of relative compositions. Anyhow, the main reason why this parameter has not been generally accepted is because of its inherent definition [85]: in fact, if q is defined as R/S and if R ≥ S, q will vary from ∞ (enantiopure R) down to 1 (racemic mixture). Contrariwise, if S ≥ R, q will vary from 0 (enantiopure S) up to 1 (racemic mixture). The same arguments are valid for measuring diastereomeric excess (d.e.) and diastereomeric ratio (d.r.).

Specifically, for KRs of racemates involving two homocompetitive reactions, as it occurs when using biocatalysts, the ratio of rate constants for each enantiomer (kR/kS, once again being R the fast-converting enantiomer) is a good measure of the enzymatic enantioselectivity. In this case, this value can be calculated according to the classical equation initially defined for Sharpless epoxidation [86] (Eq. 1.1)

Equation (1.1)

where e.e. (enantiomeric excess of the reacting substrate) and C (conversion) are ≤1. This same equation was later applied to enzymatic resolution, using the E notation [87] to refer to the ratio between the specificity constants (kcat/KM, also called kinetic efficiency) for the transformation of both enantiomers of substrate (A and B, in this case, A being transformed faster than B), as shown in the following equation:

Equation (1.2)

where once again, e.e.S (enantiomeric excess of the reacting substrate) and C (conversion) are ≤1. If the enantiomeric excess of the product (e.e.P) is measured, the enantiomeric ratio is calculated according to the following equation:

Equation (1.3)

This so-defined enantiomeric ratio E is an adimensional value that can be calculated from reaction parameters such as the time, the conversion, and e.e.S or e.e.P [88]. These two ways of calculating the enantiomeric ratio have become very popular among researchers working in biocatalysis. Nevertheless, the use of this chemometric parameter has some important drawbacks:

1. Eqs. (1.2) and (1.3) are derived for simply enzymatic resolutions following classic Michelis–Menten kinetics, which implies only one substrate and one product, irreversible reactions, and neither enzyme- nor substrate-inhibition. This is not the usual scenario for enzyme catalysis, so all the abovementioned factors should be considered for the accurate calculation of E. This fact introduces a much higher degree of complexity in the mathematical equations (for a detailed explanation, please see the seminal paper from Straathof and Jongejan [88], Faber’s book [89], the book chapter from Diaz-Rodriguez and Lavandera [90], or the recent article by Mitchell and Krieger [91]). Anyhow, and independently of all these considerations, most authors do use the simplified formulae whatever the type of enzymatic resolution may be.

2. Apart from this major point, the interpretation of this parameter is not that straightforward; thus because logarithms must be calculated, the scale for quantifying the enantioselectivity is not that evident. As a rule of thumb, E values below 15 are inacceptable for practical purposes, being moderate to good in the range of 15–30 and excellent above this value. Still, in many papers, E values are expressed with several decimal digits, which are obviously senseless because of the experimental error in the calculation of C, e.e.S or e.e.P. A generalized way of expressing a perfect KR is reporting E >200.

KRs were the first class of enzymatic processes in use for the preparation of enantiopure compounds [92–94]. Therefore there are a huge number of enzymatic KRs reported in literature, as shown in Fig. 1.3. In any case, the main drawback of a KR is the limited maximum theoretical yield for each one of the desired enantiopure enantiomers (either PR or SS), only 50%.

Figure 1.3 Different types of enzyme-catalyzed KRs: (A.1) Using hydrolases (lipases) via enantioselective hydrolysis; (A.2) using hydrolases (lipases) via enantioselective acylation; (B) using epoxide hydrolases; (C) using alcohol dehydrogenases; (D) using transaminases; (E.1) using Baeyer–Villiger monooxygenases (normal KR); (E.2) using Baeyer–Villiger monooxygenases (parallel KR); (E.3) using styrene monooxygenases. KR, Kinetic resolution.

Undoubtedly, the most typical examples are those KRs catalyzed by hydrolases (lipases, mainly), where these enzymes mediate a transacylation between an acyl donor (an acid, anhydride, or ester) and a nucleophile (water, alcohol, and amine) [95,96]. These reversible reactions can occur in aqueous media (Fig. 1.3A.1 hydrolysis of esters, amides, thioesters, etc. [96]) or in low-water-containing environments (Fig. 1.3A.2, esterification, transesterification, interesterification, acidolysis, alcoholysis, or aminolysis) [97,98] and have been thoroughly revised in several papers [65,74,95,96,99–107]. Inside this book, Chapter 5 from Sandoval covers the use of hydrolases in asymmetric synthesis.

Some recent examples of these processes illustrate the lipase-catalyzed KRs. Thus through a lipase-catalyzed hydrolytic process (Fig. 1.3, strategy A.1), the multikilogram synthesis of arbaclofen placarbil 3, employed for treating gastroesophageal reflux disease (Fig. 1.4), has been reported [108]. This prodrug is now facing a new life, as baclofen is under study as an addiction medicine to treat alcohol use disorders [109].

Figure 1.4 Kinetic resolution using a commercial preparation (IMMCALA-T2–150) of lipase, en route to the pro-drug arbaclofen placarbil.

For future development of the prodrug, a demand of 20,000–30,000 kg/year was considered, so that it was necessary to replace the original chemical synthesis of the prodrug, which required a chromatographic purification of the product. For this purpose, a chromatography-free process involving a lipase-based KR to furnish the required stereochemistry was carried out based on the enantioselective hydrolysis of racemic thiocarbonate (rac-1) [108]. Then, 60 commercially available enzymes were tested for the KR of (rac-1). Finally, authors selected a commercial preparation of Candida antarctica lipase A, immobilized on polymethacrylate (IMMCA-LA-T2-150 lipase) to get rid of the undesired (R)-2, which decomposes in situ, so that N2 sparging was required to remove the side-products. With a pH control using ammonium hydroxide, it was possible to generate 293.4 kg of enantiopure (S)-1, isolated out of three batches (60% w/w purity) with an e.e. of 99.6%. This intermediate was used to produce, through different chemical steps, 69 kg of enantiopure arbaclofen placarbil 3.

As commented before, lipases are the most used enzymes for catalyzing KRs because they can operate not only in aqueous but also in organic media (Fig. 1.3, strategy A.2). An interesting example of this type of procedure is illustrated in Fig. 1.5, showing the preparation of a rac-homopropargyl alcohol, chiral intermediate, for the synthesis of an essential building block (C14–C26) en route to Eribulin (Halaven™, Eisai), a truncated version of marine natural product halichondrin B, used for the treatment of metastatic breast cancer [110].

Figure 1.5 Kinetic resolution of a racemic alcohol in MTBE for the preparation of a chiral intermediate in the synthesis of C14–C26 building block of eribulin. MTBE, Methyl tert-butyl ether.

After obtaining an excellent KR in a stirred-tank reactor (STR) with Amano lipase PS as catalyst and methyl tert-butyl ether (MTBE) as solvent (see Fig. 1.5), the authors developed a continuous reaction in a packed-bed reactor (PBR) using the same solvent. Once the reaction finished (after 1 week of operating continuously), it produced 10.2 g of a mixture of (R)-4 and (S)-5 (overall yield 96%, productivity 1.12 mM/h/g). Both compounds were separated and converted to the desired propargylic alcohol (S)-4; in this way, the nonconverted (R)-4 alcohol is also transformed into the desired product, so that the 50% limitation of theoretical yield is circumvented. Finally, (S)-4 was used to produce the final compound (R)-6.

Apart from the external transformation of the nonconverted enantiomer (SS, according to Fig. 1.2) into the desired one (PR), there is another approach to solve the 50% yield limitation. In fact, the racemization of nonconverted SS and its subsequent re-addition to the reaction mixture is a strategy sometimes termed repeated kinetic resolution rKRs (Fig. 1.6). This sequential procedure is not generally applicable on a small scale, but it can be very useful at an industrial scale for continuous processes so that the racemized material can be re-fed into the reaction batch.

Figure 1.6 Schematic representation of a rKR. rKR, Repeated kinetic resolution.

Through these rKRs, the generation of residues is reduced, because Ss is not accumulated but rather recirculated. Therefore rKRs are preferable from an environmental point of view. Unfortunately, the urgent necessity for producing an enantiopure active pharmaceutical ingredient (API) in the pharmaceutical industry generally takes preference over a green design of the process concerning the generation and posterior removal of the waste (in KRs, the nonconverted enantiomer). In fact, it has been reported that pharma industry generates a huge amount of waste [53,65,72,111,112], with E(nvironmental)-factor (E, mass of waste/mass of product, usually expressed as kgs/kg [113,114]) values up to 200 in preclinical phases. In this book, Chapter 2 deals with chemometrics in biocatalyzed reactions.

A very classical example of a rKR is the resolution of racemic H-Phe-OiPr (isopropyl phenylalaninate, rac-8, Fig. 1.7), developed by Coca-Cola [115] for the preparation of the enantiopure (R)-isomer, subsequently used as a building block for the synthesis of aspartame, a noncaloric sweetener.

Figure 1.7 Preparation of enantiopure (R)-isopropyl phenylalaninate via rKR. rKR, Repeated kinetic resolution.

Thus by using a continuous process based on the use of a hollow fiber/liquid membrane (SLM) reactor, a feeding of rac-8 (hydrochloride) at 2 mM/h to 0.7 g/L subtilisin Carlsberg at pH = 7.5 and 25°C leaded to a steady-state reaction furnishing (S)-9 (incapable of permeating through the liquid membrane) with 95% e.e. at the theoretical. On the other hand, nonconverted (R)-8 (80% e.e.) can permeate down to the counter-flow aqueous phase at pH=3.5, becoming protonated and therefore unable to return back to the slightly basic phase. Notably, (R)-8 can be racemized quantitatively by heating with a salicylaldehyde catalyst in refluxing toluene, through the intermediate imine 10. With this system, an overall yield of 73% could be obtained.

1.2.1.2 Parallel kinetic resolutions

Another type of process to generate chirality is the parallel KRs (PKRs). This is the case when both enantiomers or the racemic substrate are converted by the biocatalyst, furnishing two different products (PR and QS), which are not enantiomers, as shown in Fig. 1.8.

Figure 1.8 PKR. PKR, Parallel kinetic resolution.

Ideally, both processes leading to PR and QS should proceed at similar reaction rates, without interferences, and, of course, PR and QS must be easily isolated and separable from each other. Anyhow, the limitation of a maximal theoretical yield of 50% for both products (PR and QS) is still present. These PKRs are typical when using redox enzymes, as recently reviewed [116]. There are three general types of PKRs, namely, chemo-, regio-, and stereodivergent, depending on the enzymatic behavior [90]. Thus in chemodivergent PKRs, final products PR and QS are neither regio- nor stereoisomeric compounds but rather pseudoenantiomers (two molecules with opposite configurations at all of the stereocenters and differing at a position remote from them) or can be completely different, generally because of the to the transformation of one of the compounds through a series of reactions that cannot be avoided. An example of a chemodivergent PKR is the 2-oxoglutarate-dependent dioxygenase (2-ODD-PH)-catalyzed ring-closure of racemic hydroxy-derivative of yatein (11 and ent-11) to the corresponding podophyllotoxin 12 [117], a precursor of etoposide 14 and teniposide 15, two anticancer drugs employed as chemotherapeutic agents. As can be seen in Fig. 1.9, only one of the enantiomers (11, (3R,4R,1′S)) suffers the desired cyclization to furnish 12, while ent-11 becomes hydroxylated at the benzyilic position, leading to a mixture of diastereoisomers 13.

Figure 1.9 Chemodivergent PKR of enantiomers of the yatein derivative 11. PKR, Parallel kinetic resolution.

Two main types of regiodivergent PKRs are generally reported. In the first type, starting from a single functional group, two different regioisomeric products can be produced. The archetypical example is the monooxygenase-catalyzed regiodivergent Baeyer–Villiger oxidation of racemic ketones possessing a stereocenter in the vicinity of the keto group (Fig. 1.10, path B). In the excellent review by Harwood and coworkers, different examples of these two possible oxidative pathways using BVMO can be found [116].

Figure 1.10 Classical KR (Path A) versus regiodivergent PKR (Path B) in the Baeyer–Villiger oxidation of racemic ketones. PKR, Parallel kinetic resolution.

The second type of regiodivergent PKRs are those in which a racemic substrate possesses several reacting groups at different positions on the molecule, and those reacting groups become distinctly modified for each enantiomer. The regioselective acylation of racemic nucleosides catalyzed by a lipase (depicted in Fig. 1.11), reported by Martínez-Montero et al., is a good example [118].

Figure 1.11 Regiodivergent PKR in the lipase-catalyzed acylation of racemic nucleosides. PKR, Parallel kinetic resolution.

In fact, the treatment of N-benzoyl-β-D/L-2′-deoxy-cytidine (16a/16b) with acetonoxime levulinate in the presence of PSL-C at 30°C in THF afforded a mixture of two acylated products: β-D-3′-O-levulinyl ester 18a (93% yield, e.e. >99%) and β-L-5′-O-levulinyl derivative 19a (82% yield, e.e. >99%), formed because of the opposite acylation preference exhibited by PSL-C in β-D- and β-L-2′-deoxynucleosides, regioselectivity acylating at the 3′-hydroxyl group of β-D-nucleosides, while for the β-L-isomer, the acylation is exclusively performed in the 5′-hydroxyl group. Similar results were obtained for N⁶-benzoyl-2′-deoxy-β-D/L-adenosine (17a/17b), yielding β-D-3′-O-levulinyl ester 18b (80% yield, e.e. >99%) and β-L-5′-O-levulinyl derivative 19b (87% yield, e.e. >99%), although in this case reaction temperature had to be increased from 30°C to 45°C.

Finally, a stereodivergent PKR implies the creation of a new chirality center (of the same configuration) for both enantiomers of the substrate, so that different diastereoisomers are formed. Some typical examples are the bioreduction of racemic ketones catalyzed by alcohol dehydrogenases (ADHs, Fig. 1.12A) or the addition of hydrogen cyanide to racemic aldehydes mediated by hydroxynitrile lyases (HNLs, Fig. 1.12B).

Figure 1.12 Stereodivergent PKRs: (A) Reduction of racemic ketones. (B) Addition of HCN to racemic aldehydes. HCN, Hydrogen cyanide; PKR, Parallel kinetic resolution.

For both cases, by selecting the biocatalyst with the appropriate stereoselectivity (Prelog or anti-Prelog) for ADHs, delivering the hydride equivalent through the Re or Si faces of the ketone [119], or (R)- or (S)-selective HNLs [120], it is possible to obtain a fixed configuration of the newly formed stereocenter. For more information on selective bioreductions, see Chapter 6.

1.2.1.3 Deracemizations

This methodology, as depicted in Fig. 1.13, would imply the conversion of a racemic mixture into an enantioenriched mixture or, more ideally, into a single enantiomer. Deracemizations, frequently used for preparing optically pure alcohols and amines [121], are very atom-efficient processes [122–124], as 50% of the desired compound (SR in this case) is in fact present in the starting racemic mixture, so that only the other enantiomer (SS) has to be transformed into SR via a stereoinversion.

Figure 1.13 Deracemizations: (A) Two possibilities: one-pot one-step direct stereoinversion and one-pot two-step stereoinversion via intermediate. (B) Detailed description of the last strategy.

The direct stereoinversion is the simplest but also the more difficult, so the one-pot two-step stereoinversion via intermediate (Fig. 1.13, path B) is the most usual. In this case, it is mandatory to use two different biocatalysts with opposite stereoselectivity (or a multienzymatic system expressed in a whole cell) to obtain the final desired SR; this fact clearly implies a higher overall cost. Archetypical examples of this type of process are the deracemization of secondary alcohols by using a pair of complementary ADHs (Prelog and anti-Prelog) or the deracemization of secondary amines employing (R) and (S)-ω-transaminases, presented in Fig. 1.14.

Figure 1.14 Deracemization of: (A) secondary alcohols, (B) secondary amines.

In the previous strategies, both semireactions must be enantioselective to ensure the optical purity of the final compounds. However, if one of the processes is not selective, it is mandatory to perform several reaction cycles to reach high optical purity in the final product. This is known as cyclic deracemization. A typical example is the synthesis of crispine A (Figs. 1.15, 20), an alkaloid isolated from Carduus crispus. Deracemization of rac-20 is performed employing a monoamine oxidase variant (MAO-N5) combined with amine–borane complex, leading to the chiral compound in 48% yield and >97% e.e. after 40 hours [125].

Figure 1.15 Cyclic deracemization of crispine A.

Thus MAO catalyzes the selective oxidation of the (S)-20 amine to the corresponding iminium 21, which is nonselectively reduced by the ammonia–borane complex. After several redox cycles, it is possible to obtain the enantioenriched starting material. A rational design of the MAO led to a mutant enzyme MAO-N-C9, which showed a 990-fold increase in the specific activity compared to the initial enzyme MAO, leading to optically active (R)-Crispine A (e.e. >97%) in 2 hours.

1.2.1.4 Enantioconvergent processes

In this type of process, a stereocomplementary reaction of both enantiomers of the substrate occurs, with one of them suffering a retention of the configuration and the other one through stereoinversion, leading to one single chiral product. There are different biocatalytic processes that can be described as enantioconvergent processes (ECPs). Through the simplest one, shown in Fig. 1.16, no further modifications are required after the enzymatic action on both enantiomers.

Figure 1.16 Direct Enantioconvergent Processes catalyzed by EHs (A) or sulfatases (B). EHs, Epoxide hydrolases.

A classic example is the enantioconvergent opening of epoxides catalyzed by epoxide hydrolases (EHs), which can promote both the Markovnikov (inversion) or anti-Markovnikov attack (retention) on racemic styrene oxide, leading exclusively to the (R)-diol (Fig. 1.16, path A). This ECP was originally reported using a pair of stereocomplementary EHs from Aspergillus niger and Beauveria sulfurescens [126]. The same research group later described the same ECP, but mediated by the EH from Solanum tuberosum [127]. Another case of direct ECP is the use of a pair of enantiocomplementary sulfatases [from Pseudomonas aeruginosa (retaining) and Rhodopirellula baltica (inverting)] for the preparation of optically pure secondary alcohols [128], as depicted in Fig. 1.16B.

There are some other types of ECPs in which additional nonenzymatic steps (inverting or retaining configuration) are required after the initial KR. For instance, racemic esters shown in Fig. 1.17 were resolved by a lipase-catalyzed hydrolysis, and subsequently, an inverting Mitsunobu of the alcohol obtained in the KR allowed the obtention of the enantiomeric pure (S)-acetoxytosilates [129]. Interestingly, the use of an insoluble polymer-attached PPh3 avoided the issues associated to the removal of the nonreacted PPh3.

Figure 1.17 Enantioconvergent processes combining a lipase-catalyzed KR and a Mitsunobu inversion. KR, Kinetic resolution.

In another example, depicted in Fig. 1.18, using only an inverting sulfatase (as shown in Fig. 1.16), Faber and coworkers described the ECP as combining the KR of racemic secondary alcohols using the sulfatase from P. aeruginosa with a (retaining) acid hydrolysis to afford pure alcohols [130]. Although more complex, another example of ECP was reported by Szymański et al. [131]. This process, shown in Fig. 1.19, combined a KR of α-bromoamides catalyzed by a haloalkane dehalogenase with a (retaining) mesylation of the resulting alcohol and a subsequent inverting SN2 of both α-mesyl- and α-bromo-amides to furnish α-substituted amides with high yield and optical purity.

Figure 1.18 Enantioconvergent processes combining a sulfatase-catalyzed KR and a retaining acid hydrolysis. KR, Kinetic resolution.

Figure 1.19 Enantioconvergent processes combining a haloalkane dehydrogenase, a mesylation and an inverting SN2.

1.2.1.5 Dynamic kinetic resolutions

To solve the inherent problem of KRs (maximum theoretical conversion of 50%), a common strategy is to transform KRs into dynamic kinetic resolutions (DKRs). In these systems, both substrate enantiomers become dynamically interconverted via an in situ racemization (difference with examples shown in Section 1.2.1.3, where the racemization is performed after the KR). Therefore in DKRs, the enzymatic conversion of the fast-reacting enantiomer SR can reach completion because of the continuous racemization of the slow-reacting antipode SS (Fig. 1.20).

Figure 1.20 General scheme of a DKR. DKR, Dynamic kinetic resolution.

There are some requirements for an ideal DKR. Firstly, the biocatalyst must be at least moderately selective for SR versus SS, reaching a minimum of E>20 for a KR of both enantiomers; in fact, this theoretical KR should not be as perfect as required for classical KRs, as racemization step guarantees a constant concentration for SR and SS. Secondly, the racemization step should proceed at a rate at least ten times faster than the enzyme-catalyzed conversion of SR into PR. Finally, it is mandatory that final product PR does not racemize and must be stable under the reaction conditions.

In DKRs, the enantiomeric factor can be calculated attending to e.e.P as shown in the following equation [90]:

Equation (1.4)

Theoretically, E value does not depend on the reaction rate, but e.e.P does change as the reaction progresses if racemization rate is not fast enough. In fact, there are large amounts of SR at the initial stages, but a significant drop of e.e.P would happen at high conversion because of the increasingly lower concentration of SR. Therefore it is necessary to manipulate the overall process by controlling the amount of enzyme and/or racemization catalyst, and by making a rational choice of the reaction conditions.

DKRs can be classified according to the type of catalyst used for the in situ racemization of substrates. An archetypical case is the combined use of lipases and transition metal catalysts for the DKR of secondary alcohols or amines, proceeding through a prochiral intermediate (ketone or imine), as shown in Fig. 1.21.

Figure 1.21 DKR of secondary alcohols or amines via combination of lipases and metal catalysts. DKR, Dynamic kinetic resolution.

Many examples of this type of DKR can be found in the literature, by combining lipase-catalyzed acylation with ruthenium, rhodium, or iridium catalysts [102,105,132–135]. Only some recent examples will be commented on here. For instance, the use of oxovanadium catalysts allows the DKR of tertiary alcohols, as the racemization mechanism proceeds through an addition/elimination mechanism involving the formation of carbocation intermediates [135]. In this sense, the DKR of 1-methyl-1-tetralol [21] has been recently reported (Fig. 1.22A) [136]. Highest racemization rates were obtained with V-MPS4, a vanadium catalyst immobilized onto mesoporous silica of 4.0 nm, operating at temperatures as low as 15°C to avoid the formation of side products. Reaction medium was diisopropyl ether (DIPE), in which the KR of racemic 1-methyl-1-tetralol catalyzed by C. antarctica lipase A (CAL-A) performed smoothly, furnishing (R)-1-methyl-1-tetralol acetate [22] with excellent optical purity and 32% conversion after 48 hours. In the DKR with V-MPS4, enantiopure (R)-23 was obtained with only 29% conversion because of a partial deactivation of the lipase by the metal catalyst. As previously commented, to overcome this issue, it was necessary to act on the reaction system, in this case implementing a continuous addition of fresh active CAL-A, leading to an enantiopure final product with 77% conversion after 312 hours.

Figure 1.22 Some examples of DKRs via combination of lipases and metal catalysts employing as substrates: (A) tertiary alcohols and (B) primary amine. DKR, Dynamic kinetic resolution.

An attractive DKR of racemic amines (24) has been reported using a hybrid biocatalyst made of palladium nanoparticles attached onto cross-linking enzyme aggregates of C. antarctica B lipase (CAL-B-CLEAs), as shown in Scheme 1.22 B [137]. By using this coimmobilized catalyst, it was possible to reach high process efficiency, since it was possible to reduce the presence of high percentages of inert material, one of the major disadvantages associated to immobilized catalysts. So, CAL-B-CLEAs were able to catalyze the selective acylation of the racemic amines using ethyl methoxyacetate as an acyl donor, while Pd(0) nanoparticles mediated the efficient racemization of the unreacted enantiomer at high temperature in hydrogen atmosphere, requiring a base (Na2CO3). The best solvent proved 1,4-dioxane at 90°C, affording chiral (R)-amides (25) with high conversions and optical purities after 18 hours, but with around 10% of ethylbenzene as a subproduct. Conversely, the use of THF at 90°C led to lower conversions at the same time, but no secondary products were detected.

Another possibility for DKRs is performing the racemization of the starting material using chemocatalysts [138], generally acids or bases. If an acid is used, a reversible equilibrium between a chiral and an achiral species mediated by the acid (e.g., keto-enol tautomerism) is the key step. Otherwise, the presence of a protonable moiety at the chirality center acting as a good leaving group and promoting the formation of a carbocation is another alternative, as reported for the DKR of several secondary by combining a lipase-catalyzed acylation with a racemization using zeolites [139].

Base-catalyzed racemization is maybe one of the most employed methods to undertake enzymatic DKRs. In this strategy, it is required that the substrate have an acidic hydrogen atom at the stereocenter, which is feasible to remove under basic conditions. For instance, alcalase (subtilisin A) adsorbed on ethyl-graftmidazolesous silica gel was used to catalyze the DKR of N-Boc-Phe thioethyl ester rac-(26) with benzylamine (Fig. 1.23) using 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as base [140].

Figure 1.23 DKR of thioester combining a protease and a basic-catalyzed racemization. DKR, Dynamic kinetic resolution.

The reaction was performed using a continuous-flow mode was carried out in a system with two multicolumn (packed-bed) reactors, employing six of them for the enzymatic KR (41% conversion after 24 hours (E >200) at 50°C) and the other five columns for the racemization at 150°C. Thus the overall DKR took place with 79% conversion and 8.17 g/L h of volumetric productivity after 3 hours, affording 79% of (S)-amide 27 with 98% enantiomeric excess (Fig. 1.24).

Figure 1.24 DKR of racemic aldehydes via ω-transaminase-catalyzed bioamination combined with base-catalyzed racemization. DKR, Dynamic kinetic resolution.

In another example, different racemic aldehydes possessing aliphatic residues at a chiral center in α position to the aldehyde moiety have been used as starting materials for the amination reactions (Fig. 1.25) [141] catalyzed by ω-transaminases (ωTAs).

Figure 1.25 Bienzymatic DKR of racemic amino acid amides combining racemases and hydrolases to furnish enantiopure D- or L-amino acids. DKR, Dynamic kinetic resolution.

Thus the basic amine donor (L- or D-alanine) is responsible for the racemization of the aldehyde chiral center, promoting the DKR process to furnish the final amines with theoretical yields up to 100%. Several racemic oxoesters rac-28 were used, leading to both enantiomers of amine 29 (precursors of brivaracetam and pregabalin, valuable drugs for the treatment of neurological diseases), depending on the TAs employed, at preparative scale (200 mg, 20 mL). For the precursor of brivaracetam, Escherichia coli cells expressing Paracoccus denitrificans TA were employed in phosphate buffer pH 7.0 containing 1,2-dimethoxyethane as cosolvent, achieving the (R)-amine in 65% isolated yield and 90% ee after 24 hours at 30°C. In the case of the pregabalin intermediate, E. coli cells expressing Vibrio fluvialis-TA led to a 55% yield of the (R)-amine at the same reaction conditions, but in this case with 60% ee.

The use of racemases to mediate the interconversion of substrate enantiomers is another possibility. However, their narrow substrate specificity and low stability in organic solvents have precluded their general use, being limited to processes involving amino acids, hydroxy acids, and their derivatives [142–144]. As an example, Asano and Yamaguchi [145] reported a bienzymatic DKR method for converting L-alaninamide ((S)-31, R=Me) to the corresponding D-alanine ((R)-32, R=Me), as depicted in Fig. 1.25.

These authors employed α-amino-ε-caprolactam (ACL) racemase and D-aminopeptidase from Achromobacter obao; by using the bienzymatic system, 45 mM L-alanine amide ((S)-31, R = Me) was completely converted to D-alanine ((R)-32, R=Me) in 7 hours (final yield more than 99.7%, e.e. >99%). In a subsequent study, Yasukawa et al. [146] expanded this bienzymatic strategy to the deracemization of α-aminonitriles (rac-30) via DKR. To this purpose, a trienzymatic system composed of nitrile hydratase (NHase) from Rhodococcus opacus 71D (catalyzing the nonstereoselective hydrolysis of rac-30), D-aminopeptidase from Ochrobactrum anthropic (selective hydrolysis of (R)-31) and ACL racemase from Achromobacter obae (racemization of the slow reacting (S)-30) was used, furnishing D-amino acids (R)-32 in excellent yield and enantiopurity. Additionally, the use of NHase, ACL racemase, and L-amino acid amidase from Brevundimonas diminuta TPU 5720 (instead of D-aminopeptidase) allowed the synthesis of optically pure L-amino acids (S)-32 also in excellent yields.

1.2.2 Pro-chirality associated to sp² carbon atoms

Biotransformations have been widely employed in the preparation of enantiomerically enriched compounds, starting from prochiral molecules associated to sp² carbon atoms. The chiral products can be obtained with a maximum theoretical yield of 100% upon an enzymatic attack through the prochiral Re- or Si-faces. In these cases, stereoselectivity is determined by the relative constants of the conversion of the starting material in each of the enantiomers of the final product. A graphic representation of the enzymatic conversion of planar prochiral substrates is shown in Fig. 1.26 (assuming C>B>A priority according to Cahn/Ingold/Prelog rules).

Figure 1.26 Prochiral-discriminating enzymatic conversion of substrates through Si- or Re-face attack.

The TPI model can also be applied to rationalize the ability of enzymes for this enantioface differentiation [89], as shown in Fig. 1.27. In fact, the optimal match between the functional groups of prochiral substrate (left) and the active site promotes the Si-attack of the chemical operator (D) on the sp² carbon atom from the top-side. The alternative fit of substrate (right) would not be recognized in the active site because of a mismatch in binding; therefore the attack by the chemical operator D from the opposite side (Re-face) will not be possible.

Figure 1.27 TPI model differentiating Si-face (allowed, left) and Re-face (prohibited, right) attack. TPI, Three-point interaction.

The archetypical example of this type of enzymatic stereo-recognition is the bioreduction of prochiral ketones using ADHs, already mentioned in this chapter (see Fig. 1.12), which will be thoroughly commented on in Chapter 6.

1.3 Biocatalysts preparations

Enzymes can be applied for chemical transformations in different types of preparations, either as (semi)purified free enzymes, as extracts free or cells or as whole cell systems. Depending on the type of biocatalytic process that will be performed, one preparation or another will be more suitable.

Biocatalytic reactions carried out by whole cells represent the simpler approach to develop a biotransformation [147–152]. In fact, no catalyst purification is required, something that can be troublesome for certain enzymes. Cells also contain all the necessary ingredients to perform the reactions; thus there is no need to add expensive cofactors (nicotinamides, flavins, PLP, etc.) to the reaction medium. Anyhow, the use of these preparations requires a certain expertise in growing cells and in fermentation procedures, which usually requires more sophisticated equipment compared to those used for isolated enzymes. Whole cell systems also suffer mass transport limitations, especially when the desired biocatalyst is located inside the cells. Low substrate and product concentrations are required since organic molecules can disrupt the cell membranes. In general, biotransformations carried out directly by whole cell systems show lower volumetric productivity. In addition, once the reactions have arrived at completion, the final product must be separated from the cell’s biomass, which requires an additional purification step.

After the corresponding fermentation, the whole cells containing the desired biocatalyst can be partially or totally purified to achieve the corresponding enzymes. Initial steps of this protein purification usually consist of breaking the cells by different methods (sonication, enzymatic methods, or high pressure) to obtain a crude mixture of the desired enzyme in solution with solid cell fragments. After removing the solid part by centrifugation, the liquid phase achieved is the so-called cell-free extract, which can be straightforwardly employed as a biocatalyst in different biotransformations. These crude preparations consist of a small fraction of the desired enzyme mixed with proteins from the host microorganism and other cellular components. This can lead to problems when carrying out the desired biocatalyzed reactions, as emulsions can be formed and side-reactions can occur because of the presence of other nondesired