Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av filosofie doktorsexamen i kemi med inriktning mot organisk kemi torsdagen den 15 oktober kl 10.00 i sal F3, KTH, Lindstedtsvägen 26, Stockholm. Avhandlingen försvaras på engelska. Opponent är Professor Benjamin L. Miller, University of Rochester, New York, USA.

Dynamic Covalent Resolution: Applications

for System Screening and Asymmetric Synthesis

Pornrapee Vongvilai

Doctoral Thesis

Stockholm 2009

พรระพี วงศว์ไิล

ISBN 978-91-7415-442-9 ISSN 1654-1081 TRITA-CHE-Report 2009:52 © Pornrapee Vongvilai, 2009 Universitetsservice US AB, Stockholm

Till Birgitta, Katrin and Namphung

Pornrapee Vongvilai, 2009: “Dynamic Covalent Resolution: Applications for System Screening and Asymmetric Synthesis” Organic Chemistry, KTH Chemical Science and Engineering, Royal Institute of Technology, SE-100 44 Stockholm, Sweden.

Abstract

Combined thermodynamic/kinetic events amount to a kinetically controlled Dynamic Combinatorial Resolution (DCR) process, where the lability of the molecules/aggregates are used to generate dynamics, and the species experiencing the lowest activation energy is selected via kinetic process. Both inter- and intramolecular processes can be performed using this concept, resulting in complete resolution and associated amplification of the selected species. When intermolecular processes are resolved using this method, an additional advantage is that only a catalytic amount of selector is required to control the system. In this thesis, the Henry and Strecker reactions were developed as efficient C–C bond-forming routes to single and multi-level dynamic covalent systems. These methods efficiently provided a vast variety of substrates from small numbers of starting compounds. These dynamic systems, generated under thermodynamic control at mild conditions, were coupled in one-pot processes with kinetically controlled lipase-mediated transacylation. The enzyme-mediated resolution of the dynamic nitroaldol system led to enantiomerically pure β-nitroacetates in high yield. Furthermore, combination of multi-level dynamic Strecker systems and lipase-mediated acylation resulted in the resolution of specific α-aminonitriles from the pool. In addition, the asymmetric synthesis of discrete β-nitroalkanol derivatives was simply achieved, resulting in high yields and high enantiomeric purities through the direct one-pot procedure. Moreover, racemase type activity of lipase enzyme through N-substituted α-aminonitrile structure has been discovered. By use of control experiments together with molecular modeling, the mechanism of the racemization process has been established. Asymmetric synthesis of N-methyl α-aminonitriles was also performed through the dual function of lipase, resulting in high yield and good enantioselectivity. Keywords: Dynamic covalent/kinetic/combinatorial resolution. Self-screening, Transesterification, Amidation, Enzyme catalysis, Nitroaldol reaction, Secondary alcohols, Strecker reaction, α-Aminonitriles, Racemase, Enzyme catalytic promiscuity.

Abbreviations

CALB Candida antarctica lipase B CCL Candida cylindracea Lipase CDC Constitutional Dynamic Chemistry CRL Candida rugosa Lipase DCC Dynamic Combinatorial Chemistry DCL Dynamic Combinatorial Library DCR Dynamic Combinatorial Resolution DIPEA N, N-diisopropylethylamine DMSO Dimethylsulfoxide E Enantiomeric ratio EC Enzyme Commission EDG Electron Donating Group ee Enantiomeric excess eq. Equivalent EWG Electron Withdrawing Group Glu Glutamine His Histidine KR Kinetic Resolution NMR Nuclear Magnetic Resonance PF Pseudomonas fluorescens Phe Phenylalanine PS-C Pseudomonas cepacia rac Racemic r.d.s. Rate determining step rt/r.t. Room temperature Ser Serine TEA Triethylamine TMSCN Trimethylsilyl cyanide Trp Tryptophan VA Vinyl acetate

“I have not failed. I've just found 10,000 ways that won't work.”

Thomas Alva Edison (1847 –1931)

List of Publications

This thesis is based on the following papers, referred to in the text by their Roman numerals I-VI:

I. Dynamic Combinatorial Resolution: Direct Asymmetric Lipase-Mediated Screening of a Dynamic Nitroaldol Library Pornrapee Vongvilai, Marcus Angelin, Rikard Larsson and Olof Ramström Angew. Chem. Int. Ed. 2007, 46, 948 – 950.

II. Tandem Driven Dynamic Combinatorial Resolution via Henry–Iminolactone Rearrangement Marcus Angelin, Pornrapee Vongvilai, Andreas Fischer and Olof Ramström Chem. Commun., 2008, 768 – 770.

III. Dynamic Asymmetric Multicomponent Resolution: Lipase-Mediated Amidation of a Double Dynamic System Pornrapee Vongvilai and Olof Ramström J. Am. Chem. Soc. in press

IV. In Situ Evaluation of Lipase Performances Through Dynamic Asymmetric Cyanohydrin Resolution Morakot Sakulsombat, Pornrapee Vongvilai and Olof Ramström Manuscript

V. Direct Asymmetric Dynamic Kinetic Resolution by Combined Lipase Catalysis and Nitroaldol (Henry) Reaction Pornrapee Vongvilai, Rikard Larsson and Olof Ramström Adv. Synth. Catal. 2008, 350, 448 – 452.

VI. Racemase Activity of B. cepacia Lipase Leads to Direct Asymmetric Dynamic Kinetic Resolution of α-Aminonitriles Pornrapee Vongvilai, Mats Linder, Morakot Sakulsombat, Tore Brinck and Olof Ramström Manuscript submitted

Paper not included in this thesis: VII. Phosphine-Catalyzed Disulfide Metathesis

Rémi Caraballo, Martin Rahm, Pornrapee Vongvilai, Tore Brinck and Olof Ramström Chem. Commun., 2008, 6603 – 6605.

Table of Contents

Abstract Abbreviations List of publications 1. Introduction ............................................................................... 1

1.1. Dynamic Combinatorial Chemistry (DCC) ........................................... 2 1.1.1. The reversible reaction .................................................................................. 3 1.1.2. Types of selecting formats in DCC ................................................................ 6

1.2. Dynamic Combinatorial Resolution (DCR) ........................................... 8 1.3. Dynamic Kinetic Resolution (DKR) ...................................................... 8 1.4. The Aim of This Thesis ....................................................................... 9

2. Explorations and Optimizations of Reversible C-C Bond Formation for Creation of Chiral Dynamic Systems .............. 10

2.1. Introduction ...................................................................................... 10 2.2. Nitroaldol (Henry) reaction ................................................................ 10 2.3. Strecker reaction and the generation of double dynamic covalent

systems ........................................................................................... 15 2.3.1. Strecker Reaction......................................................................................... 15 2.3.2. Double dynamic covalent systems .............................................................. 18 2.3.3. Thermodynamic studies following expansion of the double dynamic

covalent systems .......................................................................................... 19 3. Target Species: Lipases .......................................................... 25

3.1. Introduction ...................................................................................... 25 3.2. Lipases ............................................................................................ 25

4. Dynamic Combinatorial Resolutions: Lipase-Mediated Asymmetric Screening of Dynamic Nitroaldol, Cyanohydrin and Double Dynamic Aminonitrile Systems ........................... 28

4.1. Lipase-mediated resolution of dynamic nitroaldol systems ................. 28 4.1.1. Optimization of kinetic resolution of β-nitroalcohol substrates using

various enzymes and acyl donors ............................................................... 29 4.1.2. Direct Asymmetric Lipase-Mediated Screening of a dynamic nitroaldol

system .......................................................................................................... 32 4.2. Henry-iminolactone rearrangement: Tandem reaction-mediated

resolution of a dynamic nitroaldol system .......................................... 33 4.3. Lipase-mediated resolution of double dynamic aminonitrile systems .. 36

4.3.1. Optimization of kinetic resolution of N-substituted α-aminonitrile substrates using various enzymes and acyl donors ................................... 37

4.3.2. Direct asymmetric lipase-mediated screening of double dynamic aminonitrile systems..................................................................................... 38

4.4. Lipase-mediated resolution of dynamic cyanohydrin systems ............ 44 5. Chemoenzymatic Dynamic Kinetic Resolutions: Asymmetric

Synthesis of β-Nitroalcohols and N-substituted α- Aminonitriles ....................................................................... 49

5.1. Introduction ...................................................................................... 49 5.2. Dynamic kinetic resolution of nitroaldol adducts by combined lipase-

catalysis .......................................................................................... 49 5.3. Dynamic kinetic resolution of α-aminonitrile substrates by combined

lipase-catalysis ................................................................................ 52 5.3.1. Enzyme catalytic promiscuity....................................................................... 52 5.3.2. Racemase activity of Pseudomonas cepacia lipase leads to direct

asymmetric dynamic kinetic resolution of α-aminonitriles .......................... 53 5.3.3. Molecular dynamics (MD) and density functional theory (DFT)

calculations ................................................................................................... 57 6. Concluding Remarks ............................................................... 61 Acknowledgements Appendix References

1

1. Introduction

Nature is composed of various types of astonishingly complex systems, ranging from the subatomic to the cosmic scale. These systems relate not only to the phenomena of the physical world, but have an absolute impact also on life itself. The origin and mechanism of life is inherently intriguing, and understanding phenomena in Nature is of crucial interest for most scientists, conferring both fundamental knowledge and advancement for the human race. The increased understanding is generally incremental, but may witness important quantum leaps when information is put together in an intriguing way. For example, in 1859, Charles Darwin established the groundbreaking concept of "Natural selection", in part based on new studies on isolated animal populations. The concept has its roots in the idea of the survival of the fittest, in this case living organisms, where individuals best adapted to their environment are more likely to survive and reproduce.[1] A common example of natural selection in action is the development of antibiotic resistance in microorganisms (Figure 1). Mutations in the genetic material continuously create variants in the population with different abilities to survive, and one of these variants may be resistant to the antibiotic agent. The resistant bacterium is not the primary population from the start, but the selection pressure will favor the resistant strain's reproduction over several generations. As a result, the resistant strain becomes more dominant and will constitute the major population in the end.[2, 3] The Darwinian concept has experienced tremendous impact and underlies much of contemporary research in genetics. It also remains the primary, largely unchallenged, model for the evolution of species..

Figure 1. Resistance to antibiotic of bacteria, a) Before selection: a bunch of bacteria including resistant variety, b) After selection: most of normal bacteria die when bathed in antibiotics, c) The resistant bacteria creating population and become more common, d) Finally, entire infection evolves into resistant strain.

2

However, life is fundamentally built up from small chemical entities, which interact and communicate as a result of fundamental Natural laws and chemical principles. This complex interplay creates subsystems at the molecular level, also obeying the deterministic selection systems of Nature. The more limited molecular systems, being of lower overall complexity, serve as prototypes for the macroscopic network of entire living organisms. Over the last decade, based in part on the principles of supramolecular chemistry, the study of the characteristics of complex mixtures of interacting molecules with or without biological species has gained increasing interest.[4-6] This new scientific field (Systems Chemistry), leads to new understanding of complex chemical processes, and formulates principles for the study of larger biological systems.

1.1. Dynamic Combinatorial Chemistry (DCC) Dynamic Combinatorial Chemistry (DCC) is a concept for the generation of systems of components under thermodynamic control, relying on both molecular - dynamic covalent - and supramolecular interconnections, which allow for spontaneously adaptive behavior upon the input of an external stimulus.[12-15] This selective factor can either be the addition of a target molecule, or the alteration of the environment (pH, light, etc).[16-19] In principle, the adaptive nature of the DCC process, whereby the dynamic systems reconstitute themselves to adopt the best overall arrangement, make them especially interesting for coupled secondary processes. The overall system promotes the survival of the fittest and expels the weakest constituents from the system. The DCC process can be initialized by the selection of suitable building blocks. They must have functional groups or elements that are able to undergo reversible exchange. Next, these building blocks are all transformed by the use of non-covalent interactions or reversible covalent bonds to generate all the possible combinations of components. Then, the dynamic system is exposed to a receptor, which allows for the virtual selection process to find the component possessing the best binding affinity from the pool. Thus, if the receptor itself can act as a trap for a given component, the ensemble of candidates will be forced by kinetic or thermodynamic to rearrange in order to produce a majority of this species (Figure 2).

3

Figure 2. Schematic representation of the concepts behind dynamic combinatorial chemistry. A dynamic system of interchanging species is generated from reversibly connected fragments of building blocks. The best binder is selected, forcing the dynamic systems to rearrange so as to produce more of this member.

This possibility to spontaneously adapt to an external selection pressure is the major advantage of dynamic systems, and has led the applications of DCC to broaden. Examples include identification of ligands and inhibitors for receptors and enzymes,[9, 20-35] exploring molecular recognition,[36-40] the study hosts-guest systems and discovery of novel reactions[41, 42] as well as materials.[43-45]

1.1.1. The reversible reactions The key feature of DCC is the reversible reaction that mediates exchange of the building blocks between the different components. In an ideal case, every format of reaction can be used in DCC.[13, 15] However, in reality, several requirements are needed for the reversible reactions to be efficient in a dynamic system. The important features are: - Reactions need to be reversible on a reasonable time scale. - Reactions need to be compatible and controllable with the system used. - Reactions need to be stable, and not interrupted by other conditions used. - Reaction conditions need to be mild (temperature, pressure, concentration), due to the sensitivity of biological targets applied.

4

R1 NH2H R2

O

H R2

NR1

Imine formation

Disulf ide formation

H2O

R1S

SR2

R1S

SR2

R1 SH HS R2

Disulfide exchange HS R3 R3S

SR2 HS R1

Imine exchangeH R2

NR3

R1 NH2H R2

NR1

H2N R3

Transthioesterification R1 S

OR2 HS R3 HS R2R1 S

OR3

Diels-Alder reaction

Metathesis reaction

R2

R1R4

R3R4

R3

R3R4

R4

R3

R1 R2

R1R2 R3

R4 R1R4 R3

R2

Aldol type reaction R1 H

OEWG

R3

R2R1

OHEWG

R2 R3

Mm+ nL [MLn]m+Metal coordination

Electrostatic interaction R1 COO H3N R2

O

OR1 H3N R2

Hydrogen bonding OR2

R1O

HN

R4

R3

OR2

R1

OHN

R4

R3

Because of these criteria, the reversible reactions, which can be employed in DCC systems, are still limited. Imine formation/exchange and thiol–disulfide exchange, as well as metal coordination are most often chosen.[7] Hence, the development of efficient reversible reactions has become an important challenge. Some reversible reactions used for generating dynamic systems are shown in Figure 3.

Covalent bonds

Non-covalent bonds

Figure 3. Reversible reactions used for dynamic combinatorial chemistry to date.[10,

23, 28, 46-54]

5

R1NH2 R2

NH2 O

OH

N

OH

N

OH

R1 R2

N

O

R1N

O

R2

Zn

Zn2+

-H2O N

O

R1O

OZn

H

HN

O

R1 OOZn

H

DNAResin Selected

Components

Dynamic systems can in principle be composed by a sequence of several reversible reactions, either of the same or different type each which can carry its own dimension in structural space.[7] However, when two or more chemistries are combined, this results in additional complications when addressing the chemistries independently. The advantage of such “multilevel” systems is the extension of the diversity within the components since the building blocks are held together by a variety of linkages. Thus far, only a few examples of dynamic system involving more than one exchange process have been reported.[7] The first example of coupled reactions was demonstrated by Benjamin Miller.[55, 56] The structural diversity was accomplished via reversible imine formation and transition-metal (Zinc) complexation (Figure 4a).

Figure 4. a) Dynamic systems with multistage equilibrium process by Miller,[56]

b) Double-level “orthogonal” dynamic systems by Lehn.[57]

A couple of years later, Alexey Eliseev and Jean-Marie Lehn reported an example of double-level “orthogonal” dynamic systems. The processes were composed of terpyridine-based ligand coordination to the transition metal, Co(III), template and imine formation with ligands.[57] Lehn and co-worker

a)

b)

6

S SH

CO2

HS SH

CO2

S S

CO2

O OO2 x

pH = 6.9

H2O

X SH

CO2

X SH

CO2pH = 6.9

H2O, O2 X S

CO2

S X

CO2

X S

CO2

S X

CO2

pH = 6.9

H2O

Y S

CO2

S Y

CO2X S

CO2

S Y

CO2

x 2

have described the processes as Orthogonal systems, where the two exchange reactions can be addressed independently (Figure 4b). In contrast, the example where two exchange reactions can be performed simultaneously and communicate with each other, has been termed as Nonorthogonal system, and was firstly reported by Sijbren Otto and co-workers (Figure 5).[58] The building blocks carrying a thiol and a thioester functionality were exposed to atmospheric oxygen, causing disulfide oxidation and, subsequently, disulfide exchange. Thus, the two exchange processes can be activated consecutively but then proceed simultaneously.

Figure 5. Simultaneous exchange of a) thioester and b) disulfide linkages.[58]

1.1.2. Types of selection formats in DCC The adaptive nature of DCC systems, whereby the dynamic systems reconstitute themselves to adopt the best overall arrangement, make them especially interesting for coupled secondary processes. Thus, in the selecting framework of DCC, three major processes may be considered. The selection depends on whether a receptor, a substrate or a component acts as template for the assembling of the other partners. - Casting: receptor-induced assembly of a substrate that fits the receptor (Figure 6a).[11] - Molding: substrate or template influences assembly of an overall geometry of receptor that optimally binds/fits the substrate (Figure 6b).[36, 39, 59]

a)

b)

7

- Self assembling: internal and/or intermolecular noncovalent interactions-induced self assembly of a substrate that forms the most stable structure (Figure 6c).[42, 45, 60, 61]

Figure 6. Different selection processes in DCC: a) selection of a guest by a separately introduced host or receptor (casting), b) selection of a host by a separately introduced guest or template (molding). c) selection of self-assembling molecules on the basis of noncovalent interactions between components (Self assembling).

a)

b)

c)

8

1.2. Dynamic Combinatorial Resolution (DCR) Dynamic combinatorial resolution (DCR) has been revealed as an application of DCC, wherein a kinetically controlled selection pressure is coupled with reversible thermodynamic systems. In this strategy, the selected component is resolved in an irreversible process and expelled from the binding site. Then, the site is free to host more of the components, and the reactions of the dynamic system are thus forced to drive resolution to completion. This is observed by the amplification of reaction products (Figure 7). An additional advantage with this approach is that in principle only catalytic amounts of the target species are needed. In the last few years, this application has been used as a new tool to prove thermodynamic property in dynamic systems. It was also demonstrated the potential for DCC application to the screening for enzyme substrates as well as reaction discovery.[22, 28, 34, 42, 51, 60]

Figure 7. Concept of Dynamic Combinatorial Resolution (DCR).

1.3. Dynamic Kinetic Resolution (DKR)

Due to the increasing demand for chiral building blocks in not only the pharmaceutical but also the agrochemical industry, the search for new and efficient methods for the synthesis of enantiomerically pure compounds has been a major active area of research in organic chemistry.[62] Nevertheless, resolution of racemic mixtures, especially by biocatalysts, is still the most common way to prepare enantiopure compounds on an industrial scale, because of economic efficiency as well as environmental impact.[63-65] Dynamic kinetic resolution (DKR) can be viewed as the individual version of dynamic combinatorial resolution. DKR emerged 1989, through work by Ryoji Noyori.[66, 67] The concept is used to synthesize a specific compound via a one pot reaction. The process is based on in situ racemization of a chiral reactant and then the faster reacting enantiomer will continuously be consumed by a kinetic resolution (KR) process (Figure 8). In principle, it is possible to obtain

9

100% yield and enantiomeric excess through the DKR process. This technique overcomes the drawbacks related to traditional Kinetic Resolution (KR), wherein a maximum 50% yield of desired product can be obtained and the remaining starting material must be separated from the product.

R

OH

R

OH

Catalyst

EnzymeAcyl donor

EnzymeAcyl donor

fast

slow

R

O

R

O

O

O

Major product

Minor product

Figure 8. General concept of dynamic kinetic resolution.

Dynamic Kinetic Resolution (DKR) has proven to be a powerful and efficient method for the preparation of enantiomerically pure compounds.[68-71] The combination of microbial kinetic resolution with the in situ enzyme, metal or base-catalysed racemisation has been employed in many cases of DKR process.[68, 71] However, the compatibility among the two catalysts together with substrate must be achieved and remains the major obstacle when designing these systems.[69]

1.4. The Aim of This Thesis

The aim of this work was to develop and expand the scope of dynamic combinatorial resolution (DCR). The dynamic chemistry involved is covalent C-C bond breaking/forming reaction, simultaneously generating chirality. The goal of the DCR techniques is to use an asymmetric screening of the components in a one-pot process. Thus, the exploration of suitable reversible reactions has initially been addressed. Due to the creation of chirality in many of the dynamic systems, lipase enzymes have been selected as target species in the DCR processes. Then, the compatibility between dynamic systems and secondary (selector) processes needed to be optimized. Furthermore, from the information of DCR processes, the development of a new synthetic method for the synthesis of individual components has been further developed.

10

2. Explorations and Optimizations of

Reversible C-C Bond Formation for Creation of Chiral Dynamic Systems.

(Papers I-III)

2.1. Introduction

Reversibility is a phenomenon that traditionally has troubled the synthetic chemist, forcing the use of excess reagents and limiting reaction yields. With the introduction of dynamic combinatorial chemistry (DCC) in the mid-1990s, reversibility in contrast became increasingly interesting. Nevertheless, one of the most important reactions in organic chemistry, C-C bond formation, is still quite rare in this matter and this is especially the case when generation of chirality is involved in the system.[7, 14]

2.2. Nitroaldol (Henry) reaction

The nitroaldol or Henry reaction consists of the addition of a nitroalkane compound to the carbonyl moiety of an aldehyde or ketone to yield β-nitroalcohol adducts. It was first reported in 1859 (Figure 9),[72] and has become a powerful C-C bond-forming process in organic chemistry, originating from the potential offered by nitro compounds for transformation into valuable functionalized structural motifs such as 1,2-amino alcohols and α-hydroxy carboxylic acids (Figure 10).[73-77]

R1 NO2 R2 H (or R3)

O

R2

OHNO2

R1

base

nitroalkane aldehyde (or ketone) β-nitroalcohol

Figure 9. The nitroaldol (Henry) reaction.

11

H

O

O2N

NO2

baseOH

O2N

NO2

1 2 10a

R2

OHNO2

R1

β-nitroalcohol

Reduction Oxidation

Denitration Dehydration

R2

OHNH2

R1

R2

ONO2

R1

R2NO2

R1

R2

OH

R1

DenitrationR2

O

R1

Diels-Aldersreaction

Michael addition

Figure 10. The β-Nitroalcohol adduct and its applications.

The reversibility of the Henry reaction has already been described in the literature and also presents a challenge when good enantioselectivity of the product is desired.[77] However, to create dynamic systems by using covalent C-C bond formation, reversibility was an important property of the thermodynamically controlled of nitroaldol reaction. Thus, this reversibility was an important advantage in the creation of β-nitroalcohol dynamic systems formed efficiently from the reversible nitroaldol formation. Generally, the Henry reaction can be accelerated by different basic reagents, for example organic- and inorganic bases, quaternary ammonium salts, or by use of ionic liquids.[73, 74, 76, 78, 79] Therefore, several bases were initially screened to find suitable conditions for dynamic system generation, in which the equilibration process should be rapid and stable, not interrupting the secondary enzymatic reaction used as the selection process.

Table 1. Thermodynamic study of Henry reaction with various bases.a

Entry Base Conversion (%)bTime (h)

1

2

3

4

TEA (0.2 eq.) 7 21

TEA (1 eq.) 3 45

DIPEA (1 eq.) 6 25

DABCO (1 eq.) 6 45

5 Morpholine (1 eq.) 6 <5

5 Piperazine (1 eq.) 6 <5

a Reactions were carried out with 0.025 mmol of compounds 1 and 2 and base in 0.6 mL of toluene at room temperature. b Determined by 1H NMR spectroscopy.

12

The reactions were conducted with one equivalent each of 2-nitropropane (1), 4-nitrobenzaldehyde (2), in the presence of base. 1H NMR was used to follow the reactions by comparing the signals of the aldehyde and nitroalcohol adduct (10a or 2-1, Table 1). Among the various bases, triethylamine was finally chosen as the best catalyst for the reactions, yielding the fastest equilibration times of around three hours at room temperature. It also proved to be compatible with the desired enzymatic reaction. First, a dynamic system (DS-I) of ten nitroaldol adducts (3-1 to 7-1) was generated from equimolar amounts of 4-trifluoromethylbenzaldehyde (3), 2-fluorobenzaldehyde (4), 3-nitrobenzaldehyde (5), 2-chlorobenzaldehyde (6) and 2,4-dichlorobenzaldehyde (7), 2-nitropropane (1) and the catalyst, triethylamine (Figure 11). To achieve a reasonable equilibration rate for the slightly different ratio between aldehydes:nitroalkane (5:1) the amounts of base was increased up to 10 equivalents. Furthermore, these benzaldehydes were chosen in view of the closely individual activity of nitroaldol reaction, and structural variety, and also showing close to isoenergetic behavior in dynamic system.

H

O

H

O

H

O

H

O

H

O

O2N

F3C

F

Cl

Cl

Cl

3

4

5

6

7

NO2

NEt3

OH

F3C

NO2

3-1(R) 3-1(S)

4-1(R) 4-1(S)

5-1(R) 5-1(S)

6-1(R) 6-1(S)

7-1(R) 7-1(S)

1

OH

F3C

NO2

Figure 11. Generation of dynamic nitroaldol system (DS-I) from five aldehydes ((3)-(7)) and nitropropane (1) in the presence of base.

The dynamic system was followed easily by 1H NMR analysis (Figure 12). At the starting point, in the absence of base, no generation of dynamic β-nitroalcohol system was observed (Figure 12a). After initiation of the dynamic system by addition of excess triethylamine at 40 oC, the process reached equilibrium in 18 hours (Figure 12b). From the 1H NMR spectrum, five reversible nitroaldol isomers (3-1 to 7-1) were clearly distinguished with different ratios depending upon the reactivity of starting aldehydes and stability of corresponding nitroaldol adducts. Furthermore, a higher conversion

(Ten reversible nitroaldol adducts)

13

of β-nitroalcohol compounds was noticed at ambient temperature. In this case, longer equilibration time was needed. The dynamic system (DS-I) is also a stable process, showing no side reactions or decomposition, even after several days.

Figure 12. 1H NMR of dynamic nitroaldol system (DS-I): a) Before system generation, b) DS-I at equilibrium (18 h).

The ability of nitroaldol (Henry) reaction to create a stereogenic center led us to further investigate the generation of a dynamic system which constructs more than one site of chirality. Thus, nitroethane (8) was selected as the nitro compound. In this case, two chiral centers are created in the β-nitroalcohol compounds, extending the applicability of the system. Thus, the dynamic nitroaldol system (DS-II) was investigated by addition of equimolar amounts of four selected aldehydes (4-nitrobenzaldehyde (2), 2-chlorobenzaldehyde (6), 2,4-dichlorobenzaldehyde (7) and 2-nitrobenzaldehyde (9)) and one equivalent of nitroethane (8) (Figure 13). Excess triethylamine was also proven to be a suitable catalyst with reasonable reaction rate for the dynamic process.

a)

b)

14

H

O

O2N2

H

O

Cl

Cl

7

H

O

6

H

ONO2

9

Cl

NO2

8

NEt3

OH

O2N

NO2

OH

O2N

NO2

2-8 (R, S) 2-8 (S, R)

OH

O2N

NO2

OH

O2N

NO2

2-8 (R, R) 2-8 (S, S)

6-8 (R, S) 6-8 (S, R)

6-8 (R, R) 6-8 (S, S)

7-8 (R, S) 7-8 (S, R)

7-8 (R, R) 7-8 (S, S)

9-8 (R, S) 9-8 (S, R)

9-8 (R, R) 9-8 (S, S) Figure 13. Generation of dynamic nitroaldol system (DS-II) from four aldehydes (2, 6, 7 and 9) and nitroethane (8) in the presence of base.

1H NMR spectroscopy was also employed to follow the reaction. After initiation of the dynamic system by addition of triethylamine, the reversible generation of eight diastereomers of the β-nitroalcohol compounds (sixteen compounds in total) was observed (Figure 14). At high concentration, the dynamic nitroaldol system (DS-II) has reached equilibrium within 3 hours.

Figure 14. 1H NMR of dynamic nitroaldol system (DS-II): a) DS-II at equilibrium (3 hours.

(Sixteen reversible nitroaldol adducts)

a)

15

2.3. Strecker reaction and the generation of double dynamic covalent systems.

As seen from the previous study, the nitroaldol (Henry) reaction was shown to be a potential reaction protocol for generating the desired dynamic covalent system. Moreover, even in a compact structure, generating dynamic chirality can play an important role in creation of substrate diversity and also asymmetry. In principle, both sides around the stereogenic center of nitroaldol adduct are exchangeable (R1 and R2, left structure, Figure 15). However, the limitation of the reaction as well as type of the structure has become an important issue in practice. Enhanced diversity of the restricted structural element can be expanded if more than one axis at the same stereogenic center is exchangeable. In this case, the two-dimensional (2D) diversity is in principle extended into three dimensions (3D), more efficiently covering the chemical space (right structure, Figure 15).[80] Therefore, novel types of reaction and structure have to be investigated.

OH

R2R1

R3

R2R1

Two dimentional (2D)diversity

Three dimentional (3D)diversity

Figure 15. General concept of multi-axis exchangeability around a chiral center.

2.3.1. Strecker reaction The Strecker reaction is one of the most important multicomponent reactions to synthesize α-amino acids via the transformation of α-aminonitriles.[81-83] This original reaction comprises a condensation of an aldehyde and ammonia, followed by addition of cyanide in a one pot process. Finally, the resulting α-aminonitrile is hydrolyzed (Figure 16).

R H

O

R CN

NH2

R CO2H

NH2KCN

NH4Cl

H

-Aminonitrileα -Amino acidα

Figure 16. General concept of the Strecker reaction.

16

When using amines instead of ammonia, the hydrocyanation of preformed imines, instead of the one-pot synthesis, represents a popular and widely used alternative route (Figure 17).[82, 84, 85] These bifunctional compounds, α-aminonitriles, have subsequently been shown to be versatile intermediates in a number of synthetic applications (Figure 17). In addition, the Strecker reaction exhibits one of the simplest and most economical methods for the preparation of α-amino acids in the lab as well on an industrial scale.[86-97]

R1 H

NR2 CN

R1 CN

HNR2

α−Aminonitrile

H3OR1 CO2H

HNR2

base

R1 CN

HNR2 R3X

R1 CN

HNR2

R3

LiAlH4

R1

HNR2

NH2

Imine

Figure 17. Generation of α-aminonitriles from imines and a cyanide source and their further transformations.

According to the original report of the Strecker reaction, water was used as a solvent, and the reaction is thermodynamically controlled. However, a calculation study of the reversible process has shown that retro-Strecker reaction, is much slower (107) than the forward rate, and slightly acidic conditions are required.[98] In organic solvent, some examples have shown the reversible property through crystallization or epimerization phenomena.[99, 100] Nevertheless, most of these studies are not general, requiring not only a particular chemical structure but also specific reaction conditions. Therefore, the search for precise and mild conditions to control the thermodynamic properties of the Strecker reaction was a necessary task. In the preliminary studies, we started with preparation of imine 6-A1 (or compound 11) by mixing 2-chlorobenzaldehyde (6), methylamine (A1) and anhydrous magnesium sulphate (Figure 18). Various cyanide reagents were tested with the imine 6-A1, subsequently forming the N-substituted α-aminonitrile. Trimethylsilyl cyanide (TMSCN (12)) proved to be a suitable reagent for preparing the desired Strecker product, without the formation of any side products. The common activator to release the cyanide source is methanol.[82, 83] However, in this study, it was revealed as inefficient, possibly due to inertness of the imine used, possessing neither electron withdrawing or

17

stabilizing groups attached to the imine nitrogen. Acetic acid proved to be an alternative activator. Thus, the Strecker reaction was generated by adding TMSCN (12) to imine 6-A1 in CDCl3 followed by acetic acid. After three hours, the reaction goes to almost completion (99% conversion) to form Strecker compound 6-A1-12 as monitored directly by 1H-NMR (Figure 18). Then, to study the thermodynamics of the Strecker reaction, another imine 6-A2 was added to the reaction mixture. In the absence of any catalyst, reversibility was not observed even for prolonged reaction times (>2 days). Attempts to add different organic acids or bases, for example trifluoroacetic acid, triethylamine, or DBU in order to reverse the reaction, failed, and none of these agents were satisfactory. Promisingly, raising the temperature up to 50 oC, the Strecker compound slowly reversed back to starting imine (6-A1). Unfortunately, the formation of the other Strecker product (6-A2-12) was not observed.

NCl

NCl

TMSCN (12)Acetic acid

CDCl3

HNCl

CN

6-A1 6-A1-12> 99% conversion

6-A2

HNCl

CN

6-A2-12

Cat.

Cl

6

H

O NH2

(A1)

MgSO4

Figure 18. Model reactions to test the reversibility of the Strecker.

Instead, a variety of Lewis acids were examined. Interestingly, zinc halides, scandium triflate, gallium(III) triflate and also cadmium acetate were shown to be potentially reactive as catalysts for the retro-Strecker reaction, and formation of the α-aminonitrile 6-A2-12 could be observed. Moreover, most of these metal Lewis acids also catalyzed the forward process, which is also described in the literature.[82, 101-103] Zinc bromide proved optimal for use in the present system, resulting in the most rapid and stable equilibration, and, very importantly, also showed compatibility with the subsequent enzymatic reaction.

18

2.3.2. Double dynamic covalent systems From the previous study, the novel reversible C-C formation can be achieved utilizing the Lewis acid catalyzed retro-Strecker reaction. Next, to complete the system with multi-axial variation around a single stereogenic center, another efficient reversible reaction was considered. Initially, imine exchange was selected as a coupled process. However, it did not work when combined with the Strecker reaction. Much slower rate of the reversible Strecker reaction was observed, probably due to the fact that water, from imine exchange interfered with the catalyst. Instead, aldehyde-free transimination,[53] was used to established a reversible reaction, and chosen as a suitable process to couple with the Strecker reaction. Transimination is self-sufficient and the exchange also occurs without the disruption of the C=N unit into free amine and carbonyl. In addition, the reaction does not involve the mediation of a third constituent (i.e., water) which proved detrimental to the selection process. Again, the reaction conditions were screened since the literature reported that scandium triflate is an appropriate catalyst in transimination. The control experiment is similar to previously described Strecker reaction. In this case, amine A4 was added to the reaction mixture instead of imine 6-A2 (Figure 19).

NCl TMSCN (12)Acetic acid

CDCl3

HNCl

CN

6-A1 6-A1-12> 99% conversion

HNCl

CN

6-A4-12

Cat.

Cl

6

H

O NH2

(A1)

MgSO4

H2N

(A4)

Figure 19. Model reactions to test the reversibility of the double dynamic covalent reaction.

Focusing on Lewis acids, various catalysts were investigated. Zinc bromide proved to be a suitable catalyst in this matter (Table 2). At optimal conditions, (0.6 equivalent of ZnBr2 in DMSO-d6 0.7 M), the reaction reached the equilibrium within three hours. The combination between Lewis acids was also tested. However, the results showed no significant improvement and incompatibility of the system were observed by the dectection of the free aldehyde and the formation of a precipitate (Table 2). This suggests a strong chelation between the metal ion and the amine. In addition, it was found that ZnBr2 is as efficient in catalyzing the transimination process as Sc(OTf)3. From these results, we have successfully combined two dynamic covalent reactions around the same chiral center, generating dynamic diversity in not two, but, in principle, three dimensions via compact structure.

19

Table 2. Reversibility test of double dynamic covalent reactions by Lewis acids.a

Entry Lewis acid(0.3 eq.)b

Conversion (%)c

of 6-A4-12 (OCOS)d

1

2

3

4

5

6

ZnCl2

ZnBr2

ZnI2

Sc(OTf)3

Ga(OTf)3

Cd(OAc)2. 2H2O

41 (93)

30 (65)

7

22 (60)

20 (80)

37 (83)

Time (h)

ZnBr2 (0.6 eq.)8

ZnBr2 + Sc(OTf)39e

12

Zn(OTf)2

3 40 (85)

7

20 (60)24

>24

7

>24 20 (85)

n.d. n.d.

12

a Reaction were carried out with 0.1 mmol of imine 6-A1 mixed with 1 eq.TMSCN (12) and acetic acid. After three hours 1 eq. amine A4 was added tothe reaction mixture, (0.6 ml CDCl3, rt). b Catalyst was dissolved in DMSO-d6.c Determined by 1H NMR spectroscopy. d OCOS is overall conversion ofStrecker compound at equilibrium. e Reactions were not anylyzed due to theobservation of released aldehydes.

2.3.3. Thermodynamic studies following the expansion of the double dynamic covalent systems

We have demonstrated the new type of dynamic systems by combining two reversible covalent reactions, Strecker reaction (C-C bond) and transimination (C=N bond). In principle, a large variety in system components can be provided by small number of starting compounds since more than one axis is exchangeable. To prove the hypothesis, the thermodynamic studies of double dynamic covalent systems have been carried out through the expansion of libraries. Thus, the investigation started from generation of double dynamic system I (DDS-I) (Figure 20).

20

NCl

N

F

H2N

H2N H2N

Transimination

Dynamic System 1(6 imines)

6-A1

13-A2

13-A1 6-A2

6-A4 13-A4

CN

HNCl

TMSCN

ZnBr2

6-A1-12 (R)

13-A1-12 (R)

6-A2-12 (R)

13-A2-12 (R)

6-A4-12 (R)

13-A4-12 (R)

6-A2-12 (S)

13-A2-12 (S)

6-A4-12 (S)

13-A4-12 (S)

Dynamic System 2 (12 Strecker adducts)

A4

A1 A2

12

6-A1-12 (S)

13-A1-12 (S)

Figure 20. Double dynamic covalent system I (DDS-I).

Initially, two imines, 6-A1 and 13-A2, were mixed together in an NMR tube with benzene-d6 as a solvent (Figure 21a), a dynamic system was generated by adding one amine, A4, followed by ZnBr2 (60 mol%) (Figure 21b). This resulted in the formation of six different imines, immediately observed upon mixing. Subsequently, the second dynamic covalent reaction was achieved by the addition of two equivalents of TMSCN and acetic acid as an activator. Six isomeric Strecker products (12 compounds in total) were in this case clearly observed by 1H-NMR (Figure 21c). In addition, the Strecker reaction reached equilibrium in 18 hours (>99% conversion) in benzene-d6, slower than in CDCl3 and acetonitrile-d3. However, the better separation of the peaks was observed. With this method, reversible generation of twelve α-aminonitriles starting from two imines and one amine, followed by TMSCN, could be achieved. Next, one equivalent of amine A3 was added to the controlled reaction. The system underwent re-equilibration and, in 18 hours, two more isomeric Strecker products (4 compounds) were formed (Figure 21d).

21

6-A3-12 (R)(12 Strecker adducts

or 6 isomers)

DDS-IH2N

A36-A3-12 (S)

13-A3-12 (R) 13-A3-12 (S)

CN

HNCl

(4 New Strecker adducts formed)

Added

Figure 21. 1H-NMR analysis of the double dynamic system upon addition of amine A3 to DDS-I.

On the other hand, when instead one equivalent of imine 14-A3 was added to the controlled reaction (DDS-I), the double dynamic system re-equilibrated and a total of 24 α-aminonitriles (12 isomers) were formed after 24 hours (Figure 22), as confirmed by ESI-MS (see supporting information of paper III for more details).

a)

b)

c)

d)

6-A1 13-A2

2 imines started

6 imines formed

6 Strecker isomers formed

8 Strecker isomers formed

22

6-A3-12 (R)

(12 Strecker adductsor 6 isomers)

DDS-I 14-A36-A3-12 (S)

13-A3-12 (R) 13-A3-12 (S)

CN

HNCl

(12 New Strecker adducts formed)

NF

F

Added

14-A3-12 (R) 14-A3-12 (S)

14-A1-12 (R) 14-A1-12 (S)

14-A2-12 (R) 14-A2-12 (S)

14-A4-12 (R) 14-A4-12 (S)

Figure 22. 1H-NMR analysis of the double dynamic systems upon addition of imine 14-A3 to DDS-I.

We have shown that double dynamic covalent systems are occurring simultaneously. By adding amine A3 or imine 14-A3 to the controlled dynamic reactions, both dynamic covalent reactions undergo re-equilibration processes, communicating to each other, to achieve all the possible components. This process also demonstrated the efficiency of the process as “atom economic synthesis”. By adding just one more amine or imine into the double dynamic system, the number of Strecker compounds in the system increased by 33% or 100%, respectively.

a)

b)

6 Strecker isomers formed

12 Strecker isomers formed

23

Next, the generation of double dynamic covalent systems, DDS-II, was addressed. Both the stability and the optimal conditions for further study in a dynamic combinatorial resolution (DCR) process were investigated (Figure 23).

NCl

N

F

H2N

H2N H2N

Transimination

Dynamic System 1(12 imines)

6-A1

13-A2

6-A2 6-A3

CN

HNCl

TMSCN

ZnBr2

Dynamic System 2 (24 Strecker adducts)

A4

A1 A2

12

N

F14-A3

13-A1

13-A3 14-A1 14-A2

6-A4 13-A4 14-A4

H2NA3

3 eq.

F

6-A1-12 (R)6-A1-12 (S)

13-A1-12 (R) 13-A1-12 (S)

14-A1-12 (R) 14-A1-12 (S)

6-A2-12 (R) 6-A2-12 (S)

13-A2-12 (R) 13-A2-12 (S)

14-A2-12 (R) 14-A2-12 (S)

6-A3-12 (R) 6-A3-12 (S)

13-A3-12 (R) 13-A3-12 (S)

14-A3-12 (R) 14-A3-12 (S)

6-A4-12 (R) 6-A4-12 (S)

13-A4-12 (R) 13-A4-12 (S)

14-A4-12 (R) 14-A4-12 (S)

Figure 23. Double dynamic covalent system II (DDS-II).

By use of a powerful double dynamic system, a reversible set of 24 α-aminonitriles (12 enantiomeric pairs) was originally constructed from equimolar amounts of three imines 6-A1, 13-A2 and 14-A3, together with one equivalent of isopropylamine (A4). This was followed by addition of three equivalents of TMSCN (12) and acetic acid as an activator and also 60 Mol% ZnBr2 (in DMSO-d6 0.7 M,). These aldimines and the amine were chosen in view of their similar individual reactivity not only in the transimination process but also in the imine cyanation reaction, with close to isoenergetic behavior in the double dynamic system. In addition, a variety of aromatric and nitrogen substituted moieties were screened for this purpose. 1H NMR was used to follow each step of the process (Figure 24 a-c).

24

Figure 24. 1H NMR of double dynamic covalent system (DDS-II): a) Before system generation: started with 3 imines, b) Dynamic system 1 generation (transimination) in 2 hours: 12 imines formed, c) Dynamic system 2 generation (Strecker reaction) in 18 hours: 24 Strecker compounds formed.

According to the studies in this chapter, we have demonstrated that the nitroaldol (Henry) reaction is an efficient C-C bond-forming route to use for the generation of dynamic systems. The Henry reaction created a chiral center in the nitroaldol adduct, which can be used as an advantage to generate a variety of the components as shown in DS-II. Furthermore, we have developed the Strecker reaction as a new and efficient asymmetric C-C bond-forming, and also nitrogen containing, route to complex dynamic system generation, catalyzed by Lewis acids. We also showed the systems when two reversible covalent reactions were combined, transimination and the Strecker reaction. The process allowed us to access variation of multi-axis exchangeable around a stereogenic center, which was the advantage of generating a diversity of components from a low number of starting materials. In addition, mild reaction conditions were used and the high stability of the dynamic systems are appropriate to employ for the further study in dynamic combinatorial resolution (DCR) process.

a) b)

c)

25

3. Target Species: Lipases.

3.1. Introduction

One objective of DCC is to use it as a tool for the discovery of new ligands or substrates for biomolecules in general and for drug discovery in particular. Several examples have been demonstrated by combined dynamic systems with enzymes or receptor proteins [8, 10, 11, 34] or nucleic acids [104-106] to receive the amplification, based mostly on substrate preferential. An extra reagent, i.e. reducing agent, was often added, in order to trap the selected components so they do not reverse back to the starting compounds. By use of reversible Henry or Strecker reactions, the generations of dynamic systems (DS-I, DS-II, DDS-I and DDS-II) were obtained with diversity of not only chemical anatomy but also chirality. Thus, these results allow us to search for target species which can identify both substrate- and stereospecificity at the same time. After surveying different possibilities, we came across the lipase enzymes as target biomolecules for the selection process. Lipases have proven to be potential biocatalysts which can resolve chiral compounds, recognize a broad range of non-natural substrates.[107-110] Moreover, lipases have a high commercial availability, do not require expensive cofactors, and are easily recoverable. These properties make this family of enzymes potentially highly useful in a DCR system, and were thus probed in the present study.

3.2. Lipases Among the six categories of enzymes catalyzed reaction, hydrolases, particularly esterase families, have been widely studied and their properties are well documented. This might be due to their application ranges from ingredients in detergents, various uses in food, pulp and paper industries to use as catalysts in the chemical industry.[111-113] Lipases [EC 3.1.1.3] are the esterase belongs to the class of hydrolases. The natural reaction is the hydrolysis of triglycerides into fatty acids and glycerol at a water-oil interface (Figure 25a).[114] They are found in bacteria, fungi, plants and animals.[111] Compared with the esterases, lipases are more robust enzymes.[111] The reaction mechanism is used by all serine hydrolases,

26

involving a catalytic triad of serine, histidine and glutamate (or aspartate Figure 25b). The amino acids in the triad play an important role to activate the nucleophilicity of the hydroxyl group in serine which then can attack to the carbonyl carbon of the substrate. This leads, through a series of events, to the formation of an acyl-enzyme intermediate where the substrate is covalently bond to the enzyme. Next, water acts as a nucleophile attacking the acyl-enzyme intermediate, is releasing carboxylic acid product and regenerating the enzyme (Figure 25b).[111]

HCO

OO

CORCORCOR

Lipase

H2OHC

OH

OHOH

R H

Ox 3

GlycerolTriglycerides Fatty acids

O

O

HN N

His

Glu/AspO

Ser

HR

OOR1

Substrate

O

O

HN N

His

Glu/AspO

Ser

Catalytic triad Catalytic triad

H

O R

OR1

Tetrahedral intermediate(for example triglycerides)

O

O

HN N

His

Glu/AspO

Ser

Catalytic triad

- R1OH

RO

Acyl-EnzymeIntermediate

OHR1

O

O

HN N

His

Glu/AspO

Ser

Catalytic triad

RO

Acylation

OH

H

H2O

Deacylation by water molecule

O

O

HN N

His

Glu/AspO

Ser

Catalytic triad

H

O R

OH

Tetrahedral intermediate

O

O

HN N

His

Glu/AspHO

Ser

Catalytic triad

R OH

O

Hydrolysis product(for example fatty acids)

R O

OR1

- RCO2H

Figure 25. a) The biological function of lipases, b) Catalytic mechanism of lipases.

Lipases also recognize a broad range of unnatural substrates in both aqueous and non-aqueous media. When the enzyme is operating in an environment of low water activity, i.e. in organic solvent, any other nucleophile can compete with the water for the acyl-enzyme intermediate leading to a number of synthetically useful transformations, such as transacylation (Figure 26). In the lipase-catalyzed transacylation, the mechanisms has been characterized as Ping-Pong bi-bi mechanism which is a two-step reaction, similar to the one in

b)

a)

27

figure 25b.[115] But, instead of the water molecule attacking the acyl-enzyme intermediate, alternative nucleophiles, alcohols, amines or thiols, perform the deacylation (Figure 26). Moreover, when racemic compounds are used in transacylation processes, the recognition of enantiomeric molecules or enantiotopic groups on prostereogenic molecules with high enantioselectivity has been obtained. This is one feature that makes lipases very important for organic synthesis, and it has been widely applied in the case of kinetic resolution (KR) or dynamic kinetic resolution (DKR) until now.[70, 71]

Acyl-OR ROH(acyl donor) (released alcohol)

Enz Acyl-Enz

R1OHAcyl-OR1

R1NH2

R1SH

Acyl-NHR1

Acyl-SR1

(Various nucleophiles)(Transacylation products)

Figure 26. Lipase-Catalyzed Transacylation with various nucleophiles.

28

4. Dynamic Combinatorial Resolutions: Lipase-Mediated Asymmetric Screening of Dynamic Nitroaldol and Double Dynamic Aminonitrile

Systems. (Papers I-IV)

4.1. Lipase-mediated resolution of dynamic nitroaldol systems As discussed in chapter 2, dynamic nitroaldol or β-nitroalcohol systems (DS-I and DS-II), generated from a reversible nitroaldol (Henry) reaction, possess an exchangeable chirality through the secondary alcohols in the pool. Due to well demonstrated reactions between lipases and a diversity of chiral secondary alcohols,[116] the formed β-nitroalcohol substrates could possibly be further applied to this kind of target species. Ideally, lipase will select the best substrate and also show asymmetric discrimination in the transesterification process.[28] Thus, under the general concept of DCR, the dynamic nitroaldol systems, generated under thermodynamic control, will be coupled in a one pot process with kinetically controlled lipase-mediated transesterification (Figure 27a and b).

H

O

NO2R4

R3

R1, R2

baseOH

R1, R2

NO2

R3 R4

lipase

acyl donor

∗

OAc

R1, R2

NO2

R3 R4

Dynamic nitroaldol system Resolution

Figure 27. a) The concept of DCR. A dynamic library is formed from i components A and j components B, A specific constituent An-Bm is selectively recognized by a selector (e.g., an enzyme), enabling the formation of specific product Cnm under kinetic control. b) Dynamic Combinatorial Resolution (DCR) of nitroaldol systems.

b)

a)

29

4.1.1. Optimization of kinetic resolution of β-nitroalcohol substrates using various enzymes and acyl donors

Before combining the enzymatic reaction with a complex dynamic system, optimal conditions for kinetic resolution (KR) must be estrablished. A less complicated system was used initially, thus allowing us to understand the enzymatic reaction of β-nitroalcohol substrates more deeply. In the kinetic resolution (KR) process, achieved by lipase-catalyzed transesterification, most enzymes recognize broad ranges of substrates.[116] A size difference of the substituents at the secondary alcohol position of the substrates is, however, needed to obtain high enantiomeric excess.[117-119] In the present case, the size occupancies around the stereogenic position of the nitro alcohol adduct are similar to each other. Initially, a series of enzymes were screened using vinyl acetate (VA), commonly used for transesterification, as an acyl donor, and pure racemic 2-methyl-2-nitro-1-(4-nitrophenyl) propan-1-ol (10a) as an alcohol substrate. The reactions were followed for 24 h in dry toluene under argon atmosphere (Table 3).

Table 3. Kinetic Resolution of 10a using various enzymesa.

O2N

OHNO2

O2N

∗

ONO2

O

Lipase5 eq. VA

Toluene

Entry Enzyme Conversion [%][b] ee [%][c] E[d]

1

CALB

5

78

8

2 CRL 0 0 0 3 CCL 0 0 0 4 PS 10 0 1 5 PS-C I 11 99 >200 6 PS-C II 10 90 21 7 PF 7 93 30

8[e] PS-C I 29 99 >200 9[f] PS-C I 46 99 >200

[a] Reactions were carried out with 0.05 mmol (12 mg) of rac-10a with 10 mg ofenzyme and 5 equiv. of vinyl acetate in 0.3 mL of toluene at RT andArgonatmosphere for 24 h. [b] Determined by 1HMR spectra. [c] Determined by HPLC analysis using OD column. [d] Enantiomeric ratio [e] 30 mg of enzyme was used. [f] 30 mg enzyme, the reaction was run at 40 oC for 24 h.

Of the enzymes tested, lipases from Pseudomonas cepacia (PS, PS-CI and PS-CII) and Pseudomonas fluorescens (PF) showed transesterification activity with high enantioselectivity. In contrast, enzymes from different Candida

30

species; Candida antarctica lipase B (Novozyme 435), and Candida rugosa (CRL), showed low or no reactivity with this substrate (Table 3, entries 1, 2, 5, 6 and 7). Even for the better enzymes, however, the conversion was still quite low. This might be due to the difficulty for the sterically secondary alcohol substrate to access the enzyme pocket. By use of higher temperature as well as higher amounts of enzyme, acceleration of kinetic resolution process was observed, allowing for reasonable conversion, and enantiomeric ratio, on an acceptable time scale (Table 3, entries 8 and 9) The acyl donor is another important factor in the transesterification process by lipases, since in the first step it has to form an acyl-enzyme intermediate and displace the by-product, generally an alcohol. For example, if the acyl donor is too inert, the enzymatic process becomes slower. On the other hand, if the acyl donor is too reactive, the substrate might react and yield the acylation product without enzyme catalysis, leading to low or zero enantioselectivity of the product outcome. Moreover, by-products from the acyl donor should not react with any other reactant in the process. This could cause the system decomposition over time. Thus, various acyl donors have to be screened to find the best candidate for this system (Table 4). The results showed low conversions in the KR process when isopropyl acetate and isopropenyl acetate were used as acyl donors (Table 4, entries 1 and 2). Better conversion was observed when ethoxy vinyl acetate, p-chlorophenyl acetate and vinyl acetate were used (Table 4, entries 3-5). Unfortunately, only p-chlorophenyl acetate proved to be useful in the further DKR and DCR process (Table 4, entry 4). The other two acyl donors, ethoxy vinyl acetate and vinyl acetate, showed side reaction during the DKR process, with the nitroalkane attacking the acyl position of ethoxyvinyl acetate and attacking to the by-product released from vinyl acetate (Figure 28).

EtO O

O

NO2

O

NO2 EtOAcbaseToo reactive

acyl donor

O

NO2

OH

NO2baseH

Too reactiveby-product react further

with lipase

Figure 28. Side reactions from the acyl donors used.

31

Table 4. Kinetic Resolution of 10a using various acyl donorsa.

O2N

OHNO2

O2N

∗

ONO2

O

Lipase

Toluene

Acyl donor

Entry Acyl donor Conversion [%][b] ee [%][c] E[d]

1 OAc

10

98

110

2 OAc

7

95

42

3 EtO OAc

24

98

134

4

Cl

OAc 40

99

>200

5 OAc

46 99 >200

6[e] OAc

36 99 >200

7[f]

Cl

OAc

0

0

0

[a] Reactions were carried out with 0.05 mmol (12 mg) of rac-10a with 30 mg of enzyme and 5 equiv. of acyl donor in 0.3 mL of toluene at 40 oC and Argon atmosphere for 24 h. [b] Determined by 1HMR spectra. [c] Determined by HPLC analysis using OD column. [d] Enantiomeric ratio. [e] Vinyl acetate was used as a solvent. [f] p-Chlorophenyl acetate was used as a solvent

According to the literature,[119] increasing the concentration of enzyme-substrate complex by using the acyl donor as solvent can lead to improved results. Surprisingly, this approach gave no reaction for p-chlorophenyl acetate (Table 2, entry 6). Therefore, five equivalents of p-chlorophenyl acetate were used instead, even though this donor revealed slower conversion compared to vinyl acetate. However, no side reactions were observed in this case.

32

4.1.2. Direct asymmetric lipase-mediated screening of a dynamic nitroaldol system

After optimized the reaction conditions, lipase from Pseudomonas cepacia (PS-C I), and p-chlorophenyl acetate were selected as lipase and acyl donor, respectively. To achieve the aim of this study, the dynamic nitroaldol system I (DS-I) was generated as described in chapter 2. The PS-C I lipase together with five equivalents of p-chlorophenyl acetate, were then added to the dynamic nitroaldol system, run at 40 oC without stirring. This resulted in the transesterification of selected β-nitroalcohol structures, yielding the corresponding acetylated products, clearly visualized by 1H-NMR-spectra (Figure 29).

Figure 29. a) Dynamic system I in the presence of PS-C I and p-chlorophenyl acetate (t = 24h), b) Dynamic system I in the presence of PS-C I and p-chlorophenyl acetate (t = 14d).

b)

a)

33

As can be seen, two out of five nitroaldol isomers of DS-I were resolved by the DCR process. The preference of the selected products could be observed already in the early stages of the resolution (24h). This system required longer reaction time to reach completion (14 days). Interestingly, compound 5-1 ester, derived from β-nitroalcohol 5-1, is not the most prefers species in the DS-I. In DCR, on the other hand, lipase has selected this substrate (5-1) as the fittest, forcing the dynamic system to amplify and then subsequently yielding compound 5-1 ester as a major product. According to literature,[68, 71] para-substituted aromatic compounds showed the best reactivity to lipase, compared to other substituted positions. In this particular case, however, the meta-substituted aromatic β-nitroalcohol was proven to be a better substrate than some para-substituted compounds. In addition, molecules with ortho-substitution showed no reactivity. Moreover, higher substrate selectivity was observed at room temperature. Unfortunately, lower conversion was also noticed, due to slower enzymatic reaction. By use of HPLC analysis, the result clearly revealed the asymmetric discrimination of the resolution process, in which 99% and 98% ee were obtained for the 5-1 ester and 3-1 ester, respectively. Thus, this DCR process has efficiently selected two out of ten β-nitroalcohol substrates.

4.2. Henry-iminolactone rearrangement: tandem reaction-mediated resolution of adynamic nitroaldol system

Dynamic combinatorial resolution processes, where dynamic nitroaldol systems (DS-I) are kinetically resolved by enzyme-catalyzed reactions, has been successfully demonstrated in the previous section. The possibility of controlling dynamic systems by an internal kinetic selection pressure is a previously unexplored approach. During the investigation of the scope of the DCR technique, we discovered that the properties of one of the individual system members, created an internal selection pressure without an external biocatalyst. This represents a highly straightforward one-pot process to selective product formation from a pool of candidates, without the addition of intervening external factors. The phenomenon was coincidently found while nitroethane (8) and 2-cyanobenzaldehyde (16) were used in the dynamic nitroaldol system. The concept is demonstrated in Figure 30a and b. A dynamic system from sets of components A and B is allowed to form by reversible bond formation. A specific constituent can subsequently react further in a consecutive reaction to form a kinetically stable product. The thermodynamically controlled system

34

will simultaneously undergo continuous re-equilibration, and a clear amplification of the coupled tandem reaction product will be observed.

Figure 30. a) The concept of tandem driven, internal DCR. A dynamic system is formed from i components A and j components B. A specific combination An–Bm subsequently undergoes the selective formation of kinetically stable product Cnm. b) Internal dynamic combinatorial resolution (iDCR) of a nitroaldol system.

Subsequently, dynamic nitroaldol system II (DS-II), described in chapter 2, was mixed together with an equivalent of 2-cyanobenzaldehyde (16) in NMR tube, with acetonitrile-d3 as a solvent. In the absence of triethylamine, no reaction was catalyzed. Adding the same amount of base used as in the creation of DS-II, a total of 20 individual β-nitroalcohol adducts, including enantiomers and diastereoisomers, was initially generated (Figure 31a). The 1H-NMR analyses displayed a similar pattern to DS-II (Figure 14), with different nitroalcohols forming competitively. However, as time approached thirty minutes, an amplification of what was assumed to be the cyclic iminolactone (17) was observed (Figure 31a). This internal amplification process then gradually proceeded until all previously formed nitroalcohols, as well as all nitroethane, had been consumed (Figure 31b). An interesting observation was made upon isolation and characterization of the amplified product. NMR spectroscopic data together with supporting X-ray diffraction analysis clearly proved the compound not to be iminolactone (17), but rather lactam (18). Further mechanistic studies are yet to be made for the proposed rearrangement, but a few related systems are reported.[120-122]

b)

a)

35

Dynamic nitroaldol systems II(DS-II: 12 nitroaldoladducts generated)

H

OCN

16

OHN

NO2

17

NHO

NO2

18

Figure 31. Enlarged areas of 1H-NMR spectra for iDCR. (a) iDCR of nitroaldol system at t = 30 min. (b) iDCR of nitroaldol system after completed tandem reaction (t = 24 h).

Detailed time-dependent NMR-studies were also performed in order to evaluate the kinetic profile of the tandem reaction.[42] Experimental data was coherent with a reversibly-irreversibly coupled model, suggesting the proposed rearrangement step to be fast compared to the iminolactone cyclization (compound 17). The studies also revealed the forward nitroaldol formation to be of comparable rate to tandem cyclization, while the reverse nitroaldol reaction was slower and thereby rate determining for the overall DCR process. The 3-substituted isoindoline-1-one represents an interesting motif present in a variety of natural products and drug compounds.[123-125] Furthermore,

b)

a)

Rearrangement product

36

compound 18 belongs to a group of synthetic precursors to 1,2-diamines. Reduction of the nitro group and further hydrolysis of the lactam, would lead to these structures which are of broad utility, ranging from antitumor reagents to ligands in stereoselective organic synthesis.[126-128]

4.3. Lipase-mediated resolution of double dynamic aminonitrile systems

In this section, double dynamic systems have been subjected to a lipase-catalyzed resolution (transacylation) process. This was done to extend the scope of DCR and also show the efficiency of the screening method a with complex variety of compounds. Furthermore, in the last few years, lipase mediated asymmetric acylation of chiral primary amines is becoming increasingly common.[129, 130] Surprisingly, there are rarely examples of the resolution of secondary amines in the literature.[131-136] Thus, we decided to investigate a coupled double dynamic covalent system with lipase to carry out a double dynamic combinatorial resolution (DDCR) process. The concept is displayed below (Figure 32).

Figure 32. a) Double dynamic multicomponent resolution. A double dynamic system is formed from i components A and j components B followed by k components C. A specific constituent An-Bm-Cp is selectively recognized by a selector (e.g. an enzyme), enabling the formation of specific product Dnmp under kinetic control, b) Double dynamic combinatorial resolution of aminonitrile systems.

b)

a)

37

4.3.1. Optimization of kinetic resolution of N-substituted α-aminonitrile substrates using various enzymes and acyl donors

As far as we know, lipase-catalyzed transacylation of N-substituted α-aminonitrile substrates have never been reported. Therefore, the reaction conditions were optimized by an initial screening of a series of enzymes followed by a series of acyl donors. Regarding the DDS-II, 2-(4-fluorophenyl)-2-(methylamino)acetonitrile. Compound 13-A1-12 was selected as a model substrate. From our experiences, using 4 Å molecular sieves, it would keep the water molecule out of contact with the enzymatic reaction. Nevertheless, with this particular substrate, the molecular sieves showed inefficient reactivity which was observed by the decomposition of Strecker compound (13-A1-12) and also no observation of the desired product. Generally, kinetic resolution with various enzymes and acyl donors have been carried out by 0.05 mmol compound 13-A1-12 mixed with three equivalents of acyl donor in dry toluene. This was then subjected to a vial-contained lipase (50 mg), under argon atmosphere (Table 5 and 6).

Table 5. Kinetic Resolution using various enzymes of 13-A1-12

Kinetic resolution by lipase PS-C I gave the highest conversion to the acylation product (Table 5). In addition, p-chlorophenyl acetate and phenyl acetate were revealed as appropriate acyl donors for the chemoenzymatic reaction (entries 6 and 7, Table 6). In the end, phenyl acetate proved to be the best co-substrate since less side reactions were observed upon further study.

38

Table 6. Kinetic Resolution of 13-A1-12 using various acyl donor

On the basis of these results, the lipase PS-C I from Pseudomonas cepacia and phenyl acetate were selected as lipase and acyl donor, respectively, for the DDCR system.

4.3.2. Direct asymmetric lipase-mediated screening of double dynamic aminonitrile systems

Initially, the double dynamic aminonitrile system II (DDS-II) was contructed as described in chapter 2 (Figure 23 and 24). The method demonstrated reversible generation of 24 N-substituted α-aminonitriles (12 isomers) in two step reaction processes and also finished within hours.

Initially, the combination of double dynamic system (DS-II) with the enzymatic reaction was screened with N-substituted α-aminonitrile substrates at room temperature. This resulted in mainly N-methylacetylamide (19) which, might be due to the strong side reaction of primary amine A1, released from the transimination process, with the acyl donor via both catalyzed and/or uncatalyzed enzymatic reaction (Figure 33).

39

NH2

(A1)

NH

OPhenyl acetateO

O

OH

(19) Phenol

Figure 33. Side reaction between amine and acyl donor used in DDCR

To avoid this undesired reaction, several other conditions were tested. The best result was obtained when the reaction was carried out at 0 oC. In this case, we could avoid the by-products even when the reaction time was 14 days (less than 5% conversion). However, the desired acylation products were still hardly observed from the DDCR process. Thus, we hypothesized that the ZnBr2 solution, might bind to the polar amino acid residues on the protein surface. This coordination may block the active site or decrease the protein flexibility essential to catalytic activity.[137] Fortunately, we serendipitously came across a condition using solid state ZnBr2 as a heterogeneous catalysts for the double dynamic covalent systems and also run at 0 oC, and under argon atmosphere. This condition allowed the enzymatic process to work and we were able to screen N-substituted α-aminonitrile substrates from DDS-II in a one pot process (Figure 34). We proposed that solid state catalyst is not interfering with the chemoenzymatic reaction since they are mostly separated from each other. Moreover, rate of the side reaction has been dropped down due to strong chelation between amine and ZnBr2 in heterogeneous system.

40

LipasePS-C I

phenyl acetate

CN

N

F

13-A1-12 amide (R)

CN

N

F

14-A1-12 amide (R)

CN

N

6-A1-12 amide (R)

F

Cl

Double dynamic aminonitrile systems II(DDS-II: 24 N -subtsituted α-aminonitrile

compounds generated)

O

O

O

Figure 34. 1H-NMR spectrum of the reaction mixture of the double dynamic resolution process: signals of acylated Strecker compounds (N-alkyl protons) and remaining Strecker compounds (α-protons) of the dynamic system 2 in the presence of PS-C I and phenyl acetate (t = 40 d).

From 1H NMR spectrum (Figure 34), the results clearly indicated that compound 13-A1-12 amide, derived from 13-A1-12, is the fittest substrate of the DDCR process. However, for the other substrates, HPLC technique was used to analyze the data since the overlaps of significant peaks was observed in

41

1H NMR spectrum. Using a combination of achiral (Zorbax) and chiral (OD-H) column, all of the final products could be distinguished from each other (Figure 35a-c). This technique allowed us to determine not only conversion, but also enantiomeric excess, directly from the crude reaction mixture. According to both analyses (Figure 34 and 35), the major product was confirmed to be the 13-A1-12 amide, produced from aldimine 13-A1, obtained via transimination, and TMSCN (12). The relative concentration of the 13-A1-12 α-aminonitrile was however not among the highest compound to be generated in DDS-II. Nevertheless, 13-A1-12 was the main compound selected by the lipase for the amidation.

Figure 35. Conversion (%) of acylated N-substituted α-aminonitriles from the double dynamic resolution process at: a) 2 d; b) 18 d; and c) 40 d.

Interestingly, the other products were found to be 14-A1-12 amide and 6-A1-12 amide, showing the second and the third highest conversion, respectively. From previous experiences with β-nitroalcohols, the enzyme lipase does not prefer a substrates, possessing an ortho substituted phenyl moiety, not even a small entity such as fluoride.[138] In this study, ortho substituted aromatic α-aminonitrile substrates, not only fluoride but also chloride, showed the highest accessibility to the lipase active site to perform amidation. Moreover, at another axis around the stereogenic center, the results demonstrated that only a methyl group at the N-substituted moiety can fit the enzymatic reaction. After comparison of crude reaction mixtures and calculating of maximum concentration of standard final products by HPLC, conversions of each product

a) b)

c)

42

in the crude reaction mixture, throughout the time-scale, could be obtained. All final products were obtained in a combined overall yield of 5% after two days. Better yields were observed after longer reaction times and the products were obtained in over 39% and 65% yields after 18 and 40 days, respectively (Figure 35). The reaction has been left for 48 days where 1H NMR spectrum depicted only trace peaks of remaining target intermediates in DDS-II. The slightly increasing overall conversion (69%) was achieved. The slower rate at longer reaction times may be due to slow evaporization of methylamine, a major component in all three final products. Attempts to perform the reaction at -18 oC, resulted in stable dynamic systems, but no reactivity of the biocatalyst, even after 30 days. Equilibration of the double dynamic covalent library could be observed in 1H NMR spectra, not only by decreasing peaks of the substrate-contained right moiety, i.e. the para substituted fluoro phenyl specie, but also provides some of their peak of non-selected entities. A significant amplification effect was recorded by the HPLC chromatograms. After a two days (Figure 35a) reaction time, similar ratio of the final product, 13-A1-12 amide (2%) and 6-A1-12 amide (2%) was observed and also almost no conversion for 14-A1-12 amide (1%). For longer reaction times (18 days, Figure 35b), the ratio then changed to 22% (13-A1-12 amide), 8% (14-A1-12 amide) and 9% (6-A1-12 amide) which are clearly indicated that 13-A1-12 is the best binder in the system. At 40 days reaction time (Figure 35c), 14-A1-12 has proved to become a second best substrate since the ratio was changed to 37% (13-A1-12 amide), 15% (3-A1-6 amide) and 13% (6-A1-12 amide). The N -substituted α-aminonitrile–lipase DDCR process did not only amplify specific α-aminonitriles derivatives, but also caused asymmetric discrimination. HPLC analysis showed that the highest enantioselectivity of the process was around 5-10% ee, in cases of 13-A1-12 amide and 6-A1-12 amide. These results forced us to examine the stability of acidic α-proton in the final product with various conditions. The reaction was carried out in a NMR tube, using CDCl3 as solvent. The final compound, rac 13-A1-12 amide, was tested with a variety of organic acids or bases, together with a drop of D2O to notice the exchange of the α-proton. Unfortunately, fast racemization was observed in basic conditions, especially for the primary amine. Moreover, the exchange reaction could also be detected in diluted condition or when using the Strecker compound as a base (Table 7). Thus, these results clearly explained the low enantioselectivity outcome of the final products in the DDCR process since the primary amines and Strecker compounds are always present in the reaction mixture.

43

Table 7. Racemization Test of compound 13-A1-12 amide

F

CN

N

O

HF

CN

N

O

D

1.5 eq.D2Oadditive

Solvent

13-A1-12 amide deuterated-13-A1-12 amide

Entry Additive Solvent Conversion (%) Time (h) 1 TEA CDCl3 > 95 12 2 TBAOAc CDCl3 > 95 12 3 Isopropylamine CDCl3 80 24 4 13-A1-12 C6D6 3 24 5a Isopropylamine CDCl3 1 24 6b AcOH CDCl3 0 24 7b TFA CDCl3 0 24

8b, c PS-C I Toluene 0 24 9b, d PS-C I C6D6 0 24 10b No additive CDCl3 0 24

a Lower amount of 13-A1-12 amide (0.0018 mmol) used.b Reaction performed for up to 3 days with no visible exchange. c Reaction performed in seal-cap vial with 200 mg of enzyme. d Reaction performed in NMR tube with 20 mg of enzyme. However, replacement of the solvent to tert-butyl methyl ether (TBME), commonly used for lipases, resulted in prevention of the racemization process. These improved conditions were subsequently applied to the DDCR process. As expected, high enantioselectivity was observed for all three final products, 90% ee (13-A1-12 amide), 92% ee (6-A1-12 amide) and 73% ee (14-A1-12 amide) with approximately 14% conversion in 18 days (see supporting information of Paper III for more details). The rate of the resolution process was in this case lower than in toluene. These results strongly confirmed the efficiency of the dynamic resolution process by selecting three compounds from a pool of 24.

44

4.4. Lipase-mediated resolution of dynamic cyanohydrin systems

The cyanohydrins reaction, potential carbon-carbon bond formation, can be generated by reaction between an aldehyde or ketone and cyanide source. Asymmetric cyanohydrin building blocks are very versatile since they can be transformed into many chiral functional groups. Thus, many attempts of synthetic pathways have been reported to achieve the enantioselectivity.[139-142] In addition, thermodynamic property of cyanohydrins reaction has also been established, in which it could successfully be applied to the chemoenzymatic dynamic kinetic resolution process.[143-146] According to the literature, the reversibility of cyanohydrins reaction in organic solvent has proven to be efficient and could be catalyzed by simple organic bases.[145, 146] Lipases were used as a resolving agent in DKR process which resulted in broad range of cyanohydrins substrates were catalyzed via transesterification. Because of these reasons, dynamic combinatorial resolutions of cyanohydrins substrate were conducted to identify lipase preferential substrate from the selected systems. Moreover, an efficiency of thermodynamic property of cyanohydrins reaction could allow us to vary the condition of DCR process and investigate the enhancement of amplification in real-time. Although, the reversible cyanohydrin reactions can be generated by different cyanide sources,[140, 147, 148] one of the most convenient and less harmful methods is using acetone cyanohydrin (20) in the presence of base to mildly release cyanide ion and acetone as by-product. Then, five different aromatic aldehydes, 2-trifluoromethylbenzaldehyde (21), 2-naphthaldehyde (22), 2-methylbenzaldehyde (23), benzaldehyde (24) and 4-methoxybenzaldehyde (25) were chosen in order to demonstrate the application of the reversible cyanohydrin reaction and also substrate diversity (Figure 36).

Figure 36. Lipase screening of a dynamic cyanohydrin system.

45

Moreover, their corresponding intermediates are stable and their reactions did not generate any side product. The equimolar amount of these aldehydes and acetone cyanohydrin (20) and also triethylamine as a catalyst were mixed in chloroform-d to generate dynamic cyanohydrin system. NMR spectroscopy was used to monitor the ratios of intermediates in dynamic system and product formations. Because of their similar structure, their α–protons appear in the same regions, 5.40-5.95 and 6.30-6.70 ppm for cyanohydrin intermediates and ester products, respectively in NMR spectra (Figure 37). The dynamic system can reached equilibrium in three hours (Figure 37a).

Figure 37. NMR spectra (enlarged areas): a) Dynamic cyanohydrin at equilibrium; b) Dynamic lipase screening cyanohydrin system using lipase 3 mg at 0°C.

Isopropenyl acetate was chosen as acyl donor in lipase catalytic resolution. This acylating reagent is commonly used in lipase mediated transesterification process. Acetone, released as by product, would not interfere in either dynamic system or enzymatic reaction. Although, many different lipases are commercial available,[110, 149, 150] Pseudomonus cepacia lipase was found to have highest enzyme activities. Initially, dynamic cyanohydrin system was applied to 10 mg of lipase PS-C I and isopropenyl acetate at room temperature. To control acivity of water in enzymatic reaction, molecular sieve 4Å was also applied into the reaction. The resolution by the enzymatic process was finished in 7 days. The ratios of acylated cyanohydrin products, (21-25)-20 ester, demonstrate an effect of lipase substrate selectivity. The major intermediate 21-20 in dynamic system

46

was turned out not to be a best substrate for lipase. Instead, compound 24-20 was catalyzed and amplified to achieve the fittest, compound 24-20 ester of the system (Figure 38A).

Figure 38. Substrate preferential factor of dynamic cyanohydrin system using A. 10 mg lipase at RT; B. 6 mg lipase at RT; C. 3 mg lipase at RT and D. 3 mg lipase at 0°C.

To enhance the selectivity of DCR process, reaction conditions were varied. Many studies reported that lipase activities could be improved by the exchange of acylating reagents,[151] solvents,[152] enzymes[153] and/or reaction temperature.[154] First, various acyl donors (2-naphthalenyl acetate, phenyl acetate and its derivatives, see paper IV) were applied to the resolution process. The results indicated that these acyl donors can enhance only the rate of enzymatic reactions but not for the outcome in substrate selectivity. Then, a variety of solvents were screened in the enzymatic reaction. The reaction using toluene as a solvent showed similar rate of DCR process and also ratio of substrate specificity as chloroform-d. When polar solvents, for example, tert-butyl methyl ether, dimethyl ether, acetonitrile, dimethyl sulfoxide and

47

dimethyl formamide, were used as solvent, their lipase catalyzed reactions were much slower. Next, the amount of lipase PS-C I was decreased down to 6 and 3 mg which led DCR processes finished in 10 and 14 days, respectively. The ratios of final ester products (21-25)-20 ester were monitored. By use of the calculation of preferential factor, the amplifications of substrate selectivity have been visualized (Figure 38B and 38C). The ratio of compound 24-20 ester was clearly enhanced as amount of lipase was decreased. In contrast, the reverse trend for cyanohydrin intermediates 21-20 and 23-20 was observed. Temperature dependence in enzymatic reaction has been studied in many biocatalysts, especially increasing enantioselectivity of enzyme activity by decreasing temperature.[154-156] To decrease lipase resolution process and enhance lipase selectivity, low temperature method was also tried in dynamic cyanohydrin screening system. The same dynamic cyanohydrin system was applied with 3 mg of lipase enzyme and the enzymatic reaction was performed at 0°C (Figure 37b and 38D). Because of lower reaction temperature, the enzymatic resolution was retarded and finished in 30 days. Following the DCR process at various time-scales by proton NMR spectra revealed that the ratio of cyanohydrin intermediates (21-25)-20 was similar for the whole resolution process. Hence, these results indicated that the thermodynamic property of cyanohydrin reaction was still efficient until lipase resolution was finished even at lower temperature. Again, in case of lower temperature, substrate selectivity has been enhanced in DCR process. Compounds 24-20 ester and 22-20 ester have obviously been more amplified by lipase as respect to the first and second preferential substrates. To explore enantioselectivity of final cyanohydrins ester products, chiral HPLC was used to analyze. The HPLC chromatograms showed enantiomeric ratio of individual ester products. All ester products, except the unfavor product 21-20 ester, were asymmetric resolved by lipase transesterification. Interestingly, the highest enantiomeric ratio among these ester products is from cyanohydrin ester 24-20 ester, 83% ee. This indicated that lipase catalyzed reaction was powerful screening tool to resolve optically active cyanohydrin ester product. In conclusion, we have demonstrated that the generated dynamic system (DS-I and DS-II) or double dynamic systems (DDS-II) were successfully coupled with kinetically controlled secondary processes, either biological catalyst or irreversible reaction. By use of lipase enzymes, the results showed the efficient substrate-screening method with mild condition as well as performed in a one pot process. Moreover, the internal selection pressure from a consecutive intramolecular tandem reaction to the prototype dynamic nitroaldol system

48

(DS-II) also influenced the thermodynamically controlled process and caused a quantitative amplification effect. The investigation showed that the tandem reaction was followed by an unexplored type of intramolecular rearrangement. Moreover, the efficiency of reversible cyanohydrins reaction has been applied to DCR process by lipase-mediated transesterification. The results demonstrated that the enhancement of substrate preferential can possibly monitor by vary the reactions condition, in which DCR process was used to visualize the effect in real time. The information from the DCR and DDCR methods (substrate selectivity and stereoselectivity) led us to develop a practical chemoenzymatic route for the preparation of optically active β-nitroalcohol and N-methyl aromatic α-aminonitrile derivatives through lipase-mediated amidation (see chapter 5).

49

5. Chemoenzymatic Dynamic Kinetic

Resolutions: Asymmetric Synthesis of β-Nitroalcohols and N-substituted α-

Aminonitriles Compounds. (Papers V-VI)

5.1. Introduction Chirality is one of the most intriguing features of natural substances which is of importance for specific biological activity. Moreover, due to the increasing demand for such compounds as chiral building blocks in not only the pharmaceutical but also agrochemical industry, the search for new and efficient methods for the synthesis of enantiomerically pure compounds has been a major active area of research in organic chemistry.[62] However, a resolution of racemic mixtures, especially by biocatalyst, is still the most common way to prepare enantiopure compounds on an industrial scale, because of economic efficiency as well as environmental impact.[63-65] Among all the biocatalysts, lipases have the advantages of commercial availability, easy use and recycling. They have also proven to be very effective catalysts for the synthesis and resolution of a broad range of non-natural substrates.[110, 112, 129]

5.2. Dynamic kinetic resolution of nitroaldol addcucts by combined lipase-catalysis

Dynamic Kinetic Resolution (DKR) has proven to be a powerful and efficient method for the preparation of enantiomerically pure compounds.[70] It conquers the drawbacks of traditional Kinetic Resolution (KR), maximum 50% yield of desired product and also separated remaining starting material required. The combination of enzyme, chemocatalyst or microbial kinetic resolution with the in situ enzyme, metal or base-catalyzed racemisation of secondary alcohol has been employed in many cases of DKR process.[71] However, the compatibility among the two catalysts together with substrate must be achieved which most often still be the major obstacle.[69] In the previous chapter, we developed a method for self-screening of dynamic nitroaldol system through kinetically controlled lipase-mediated

50

transesterification.[28] To explore the scope of the reaction, we have further investigated a novel asymmetric synthesis of β-nitroalkanol derivatives, utilizing a direct one-pot reaction combining a nitroaldol (Henry) reaction and kinetic resolution (KR) through lipase-catalyzed transesterification (Figure 39).

R''NO2

R'

R H

O R

OHNO2

R' R''

R

OHNO2

R' R''

base

Kinetic resolution (KR)Henry reaction

R

OAcNO2

R' R''

R

OAcNO2

R' R''

fast

slow

Dynamic kinetic resolution (DKR)

Figure 39. General concept of the one-pot dynamic kinetic resolution process.

From experiences in chapter 2 and 4, the optimal condition was achieved. By using an excess of 2-nitropropane (1) together with p-nitrobenzaldehyde (2a), in the presence of PS-C I and p-chlorophenyl acetate (toluene, 40 oC and 2 days), compound 2a could be isolated in 90% yield and 99% ee (entry 1, Table 8). The biocatalyst, Pseudomonas cepacia (PS-CI), was easily recovered by filtration, washed with organic solvent and dried under vacuum without loss of activity. This experiment demonstrated that the method is an alternative method for the synthesis of nitroalkanol derivatives via a one-pot reaction in high yield and with high enantioselectivity. This new one-pot synthesis was applied to a range of different aldehyde substrates (Table 8). p-Cyanobenzaldehyde (2b) was used under the same conditions as for p-nitrobenzaldehyde but the results showed slightly lower enantiomeric excess (Table 8, entry 2). In the case of benzaldehydes 2c–f, five equivalents of 2-nitropropane were used. These results also showed high yields as well as enantiomeric excess (entries 3–5, Table 8) with the exception being 3,5-bis(trifluoromethyl) benzaldehyde (2f) where the enantiospecificity decreased to 80% ee (entry 6, Table 8). This effect may be due to the presence of two groups of similar size next to the stereogenic center.

51

Table 8. One-pot combined nitroaldol (Henry) reaction and lipase-catalyzed transesterification.[a]

NO2 R

O

O

NO2

10a-m

NEt3

R H

O

2a-m 1

PS-C Ip-Cl-C6H4OAc

Toluene

Entry R Product Time [d] yield [%][b] ee [%][c]

1

4-O2N-C6H4

10a

2

90

99

2 4-NC-C6H4 10b 2 89 91 3[d] 4-F3C-C6H4 10c 3 89 97 4[d] 3-O2N-C6H4 10d 3 90 91 5[d] 3-NC-C6H4 10e 3 92 97 6[d] 3.5-bis(F3C)-C6H4 10f 3 80 80 7[e] 4-F-C6H4 10g 4 85 98 8[e] 4-Cl-C6H4 10h 4 83 97 9[e] 4-Br-C6H4 10i 4 81 96 10[f] C6H4 10j 4 80 91 11[f] 4-CH3-C6H4 10k 4 35 93 12[f] 4-CH3O-C6H4 10l 4 28 99 13[f] Thiophene-2 10m 4 68 46 [a] Condition: 0.125 mmol benzaldehyde, 0.5 mmol 1, 0.25 mmol of TEA, 106.25 mg

p-chlorophenyl acetate, 90 mg of PS-C I, 0.25 mL of toluene, 40 oC, 2 days. [b] Isolated yield. [c] Determined by HPLC analysis using OD and OD-H column.

[d, e] 5 and 10 equivalents of 2-nitropropane and 2.5 and 5 equivalents of TEA were

used, respectively. [f] Condition: 0.125 mmol benzaldehyde, 1.25 mmol 1, 0.625 mmol of TEA, 106.25

mg, p-chlorophenyl acetate, 90 mg of PS-C I, 0.1 mL of toluene, 40 oC.

To achieve a reasonable equilibration rate for less electron withdrawing, unsubstituted, and electron-donating substituents and also heteroaromatic benzaldehydes, 2g–m, the amount of nitropropane was increased to ten equivalents. After a reaction time of 4 days, compounds 10g–j were obtained in good yields and with high enantiomeric excess (entries 7–10, Table 8). In contrast, benzaldehydes 2k and 2l resulted in rather low yields, as expected from the lower nitroaldol reaction rates, but retained high enantioselectivity (entries 11 and 12, Table 8). The heteroaromatic 2-thio phenecarboxaldehyde

52

(2m) resulted in a higher yield but considerably lower enantiospecificity of the β-nitroalkanol product within the same reaction time (entry 13, Table 8). In addition, several aliphatic aldehydes (butyraldehyde, isovalerylaldehyde, and octylaldehyde) were examined. However, all of them displayed unsatisfactory enantioselectivity. Thus, diminishing the aldehyde ring size, or the use of aliphatic aldehydes , lead to decreasing enantiospecificities. On the other hand, this one-pot procedure results in high stereospecificities for all benzaldehydes tested with the reaction yields depending on the different substituents.

5.3. Dynamic kinetic resolution of α-aminonitrile substrates by combined lipase-catalysis

Optically active secondary amines are important chiral building blocks in industries as well as applications in organic asymmetric synthesis as chiral auxiliaries, catalysts or resolving agents.[129, 157-160] As described in chapter 4, we have developed a method for self-screening of dynamic N-substituted α-aminonitrile systems through kinetically controlled lipase-mediated amidation.[34] Nevertheless, there is still a need to develop efficient enantioselective synthesis and technically feasible methods for this type of substrates. Hence, the techniques developed, and the information gained (substrate selectivity and stereoselectivity), from the complex self-screening systems led us to further investigate a practical process of chemoenzymatic dynamic kinetic resolution (DKR), for the preparation of optically active N-methyl aromatic α-aminonitrile derivatives.

5.3.1. Enzyme catalytic promiscuity Despite their great proficiency in action and robustness, enzymes exhibit a remarkable evolutionary adaptability (evolvability). Another significant ability of an enzyme, recently becoming an important subject for both enzymologist and organic chemists, is when a single active site catalyzes more than one type of chemical transformation. So called “enzyme catalytic promiscuity”.[161, 162] The transformations are different if the types of bonds cleaved and/or made are different, and if the transition states of the two reactions are different.[161] From a synthetic point of view, enzyme promiscuity is especially interesting when the type of reaction being catalyzed does not occur in nature. In addition, the discovery of promiscuous behavior has the potential of expanding the diversity of synthetic methodologies.[163]

53

Research over the last few years has uncovered many examples of this phenomenon. It has also become clear that catalytic promiscuity has implications in evolutionary relationships.[161-163] Examples include alkaline phosphatase catalyzed hydrolysis of p-nitrophenylsulfate,[164, 165] aminopeptidase-catalyzed hydrolysis of phosphoesters,[166] lipase-catalyzed Michael additions of N-,[167] O-,[168] S-,[169] and C-nucleophiles,[170] aldol additions,[171] oligomerization of siloxanes,[172] racemase catalyzed PLP-dependent aldol additions,[173] and arylmalonate decarboxylase catalyzed aldol additions.[174] Many of these studies involve protein engineering which was challenged by turning low promiscuous activity of an enzyme into the main activity or to use part of the existing mechanism and divert the reaction path to new products.

5.3.2. Racemase activity of Pseudomonas cepacia lipase leads to direct asymmetric dynamic kinetic resolution of α-aminonitriles

During optimization of dynamic kinetic resolution processes, we have discovered novel promiscuous racemase type activity of lipase enzymes through α-aminonitrile structures. We also found that the racemase ability communicated with the amidation which resulted in asymmetric synthesis of α-aminonitrile derivatives in good yield and high enantiomeric excess via a direct and mild one-pot DKR process (Figure 40).

Figure 40. The dual functions in lipase mediated dynamic asymmetric resolution of N-methyl α-aminonitriles.

To the best of our knowledge, no such case of enzyme promiscuity has been reported even with related mutant biocatalysts. Thus, we have demonstrated the first example of this phenomenon together with the investigation of the

54

plausible enzymatic mechanisms of the racemization process which was supported not only by experimental but also computational studies. Initially, the unexpected phenomenon was observed when compound 26a was used to optimize the KR procedure. The reaction was performed by mixing 26a with phenyl acetate as acyl donor and tert-butyl methyl ether (TBME), using as solvent, lipase PS-C I. 1H NMR was used to follow the reaction. After 1 day at 40 oC, the final product 27a was observed in 34% conversion. Surprisingly, leaving the reaction for another day, the conversion of final product 27a was determined to be 60%, exceeding the theoretical limit of KR, together with high enantioselectivity, 88% ee. The remaining starting material 26a, however, was observed with almost 0% ee. After 8 days, the process had reached completion, 94% conversion and 85% ee of final product 27a (Table 9, entry 1). Table 9. Dynamic kinetic resolutions of N-methyl α-aminonitriles 26a-c via the lipase-catalyzed amidation/racemization one-pot process.[a]

R

CN

HN

R

CN

NPS-C I

Phenyl acetate

TBME

26a-c 27a-c

40 oC

Ac

This result clearly indicated that the transformation of compound 26a to 27a was carried out through DKR process instead of traditional KR. Other N-methyl α-aminonitrile derivatives, 26b and 26c, were also subjected to these conditions. These compounds were chosen in view of their similar core

55

structure but differences in stereoelectronic effects, possibly providing more information for mechanistic study. As expected, the results were similar as compound 26a (Table 9, entry 2 and 3). Similar rate for the DKR process was observed in case of compound 26a and 26b respect to non-substitued and electron rich aromatic. At 2 days of reaction time, both compounds showed over 50 % conversion and completion was achieved within 7-8 days. But t½ of compound 26b (1.134) is a little bit shorter than for 26a (1.342). In contrast, compound 26c, electron poor aromatic, showed the slowest rate, taking almost 3 days for over 50% conversion (t½ = 2.357) and 13 days for completion (Table 8, see also Supporting information of Paper VI for details about the kinetic study). In principle, the racemization method of starting compound, i.e. 26a-c, is needed as a key step to complete the DKR procedure. According to the kinetic study, the plausible mechanisms were hypothesized either proceeding via base abstracted α-proton of starting N-methyl α-aminonitrile or bond cleavage via retro-Strecker reaction. Moreover, the racemization could either be an enzymatic or non-enzymatic process. To answer these questions, various control experiments and molecular modeling were performed. Initially, the stability of the α-proton in the starting compound was investigated. Previous studies demonstrated that the α-proton of the final product (for example 27c) is sensitive to basic conditions, and thereby leading to racemization.[34] Compound 26c was thus tested with a variety of organic acids and bases together with D2O, in order to record the exchange of the α-proton. However, no racemization was observed even when using of strong acid or base and also high temperature (see also Supporting information of Paper VI). Next, deuterated in the α-position (d-26c) of racemic compound 26c was prepared. Compound d-26c was also subjected to DKR conditions as in Table 8 and the reaction was run for 5 days at 40 oC. The outcome showed no racemization of the compound d-26c and 27c deuterated in the α-position (d-27c) was obtained in 65% yield and also 86% ee (see also Supporting information of Paper VI). These results strongly confirmed that the racemization mechanism does not pass through the abstraction of α-proton of starting Strecker compounds (26a-c). Thereafter, non-racemic compound 26c (10% ee) was prepared in situ. Then, the reaction mixture has divided in three different fractions which was subjected to different conditions. The first one contained no lipases. By HPLC analysis, no racemization of compound 26c (10% ee) was observed, even when heating up to 40 oC for two days (Figure 41).

56

CN

HN

F

26c (10% ee)

No lipasescontained

lipasescontained

Pre-incubatedlipases with

MHPNPcontained

40 oC,2 days

RT 2 days then40 oC 1 day

RT 2 days then40 oC 2 days No racemization

was observed

Racemization wasobserved from10% to 2 % ee

No racemizationwas observed

1

2

3

Figure 41. Racemization tests of non-racemic 26c using various conditions.

In contrast, the second fraction was subjected to lipase. As expected, the racemization of compound 26c (from 10% to 2% ee) was observed, showing a faster rate at 40 oC compare to room temperature (Figure 41). These results strongly confirmed that the recemization process was performed by the lipase. The last fraction was subjected to pre-incubated lipase PS-C I with methyl 4-nitrophenyl hexylphosphonate (MNPHP), a known inhibitor of CALB.[175] At 40 oC for 2 days, the result also shows no racemization of compound 26c which can be explained in term of competitive inhibitor, methyl hexylphosphonate from MNPHP has already blocked the active site of lipase (Figure 41). Hence, this study has proven that the racemization and transacylation processes were carried out at the same binding pocket in the lipase. In addition, compound 26c was again mixed with TBME and lipase PS-C I, and then left at 50 oC for 3 days. By 1H NMR analysis, the aldehyde peak was noticeably observed in much higher conversion than in the other controlled experiments, without enzyme. This result supported the hypothetical mechanisms where racemization occurs through a retro-Strecker reaction, carried out by the lipase. According to all of the control experiments, the results allowed us to propose the mechanisms of racemization process through C-C bond-breaking/forming, retro-Strecker/Strecker reaction in natural active site (Figure 42). To support and explain more mechanistic details, molecular modeling of enzyme lipase and the Strecker substrate were analyzed.

57

OO

HN

N

His286

Asp264

HOSer87

NC

H

N

NO

Gln88

HN

H

O

Leu17

OO

HN

N

His286

Asp264

HOSer87

NC

H

N

NO

Gln88

HN

H

O

Leu17

OO

HN

N

His286

Asp264

HOSer87

N

N

NO

Gln88

HN

H

O

Leu17

H

C

S-enantiomer

R-enantiomer

retr o-Streckerreaction

Rotation ofimine

OO

HN

N

His286

Asp264

HOSer87

N

N

NO

Gln88

HN

H

O

Leu17

H

C

Rotation

Streckerreaction

Figure 42. Proposed racemization mechanisms

5.3.3. Molecular dynamics (MD) and density functional theory (DFT) calculations

Additional support for the racemase-type activity was acquired from molecular dynamics (MD) and density functional theory (DFT) calculations. MD simulations were performed for both aminonitrile enantiomers complexed to the Pseudomonas cepacia lipase (see Supplementary information of Paper VI for details), and the results show that both enantiomers are tightly bound to the oxyanion hole via hydrogen bonds to the cyano group (Figure 43). Strong hydrogen bonding is also observed between the Nε of His286 and the substrate amino group (S3), locking the substrates in a rather rigid position in the active

58

site. Moreover, there is a fluctuating H-bond between Ser87 Oγ and the α-nitrogen (S4), which may play a role in the transition state. The hydrogen bond network of each enantiomer is consistent with a general base-catalyzed elimination mechanism, in which His286 abstracts the proton of the amino group, and the TS and the intermediate cyanide ion are stabilized by the oxyanion hole. Simulations of the intermediates indicate that the cyanide ion is held very tightly not only by the oxyanion hole donors, but also by Ser87. The imine showed higher flexibility but stayed in the vicinity of the cyanide ion, further supporting the suggested racemization mechanism.

Figure 43. Snapshot of MD simulations of the S- (left) and R-aminonitriles (right), with selected active site residues. Hydrogen bonds are depicted in green.

Hybrid-DFT calculations were subsequently performed to validate the reaction mechanism (see Supplementary information of Paper VI for details). The model system, taken from the R-aminonitrile simulation, is displayed in Figure 43 and shows the optimized geometry of the transition state at the B3LYP/6-31+G(d) level. The calculations indicate a concerted mechanism in which proton abstraction of the amino-group precedes the leaving of the cyanide ion. The whole catalytic machinery of the lipase is active, but evidently, as predicted by MD, Ser87 acts more like an auxiliary hydrogen bond donor than a nucleophile (Figure 44). The activation energy was calculated to 21 kcal/mol, which is in reasonable agreement with the experimentally observed rate.

59

Figure 44. Model transition state (R-aminonitrile). The annotated distances are (Å) TS1=2.02, TS2=1.79, TS3=2.89, TS4=2.05, TS5=, TS6=1.91.

In summary, the techniques developed, and the information gained from the complex system screening, were further developed to practical synthetic methods of the individual chiral building blocks. By use of biological catalysts, lipases, asymmetric syntheses of β-nitroalcohol and α-aminonitrile derivatives were achieved in good yields and good enantioselectivities via dynamic kinetic resolution processes.

60

61

6.

Concluding Remarks The development of the new concept Dynamic Combinatorial Resolution (DCR) has been a major task in this thesis. In addition, chemoenzymatic asymmetric Dynamic Kinetic Resolution (DKR) has also been developed. Initially, the nitroaldol (Henry) reaction was investigated and proved to be an efficient method of a C-C bond forming route for the dynamic system formations (DS-I and DS-II). By using the thermodynamic property of the Henry reaction, it allowed us to simply generate a variety of β-nitroalcohol substrates, possessing chiral center and continuously interconverted in a one-pot process. To further explore dynamic covalent chemistry, the Strecker reaction, C-C bond-forming and nitrogen containing route, has also been developed for this purpose. Furthermore, we have demonstrated that the combination of two dynamic covalent systems, aldehyde-free transimination and imine cyanation, allowed for multiple exchanges in three dimensions around a single stereogenic center in a restricted structure. This multiple exchange process also demonstrated the vast expansion of the system (DDS-I) by addition of just one single component, an amine or imine. In addition, the dynamic and double dynamic covalent systems (DS-I and DDS-II) under thermodynamic control could also be successfully coupled to a secondary synthetic process mediated by a lipase under kinetic control. These DCR processes have been used to generate a collection of potential enzyme substrates, and to identify the best substrates for a lipase from Pseudomonas cepacia (PS-C I). By combining the dynamic nitroaldol system (DS-I) with lipase-mediated transesterification, complete, asymmetric resolution of the system could be achieved in a one-pot process, producing enantiomerically pure β-nitroacetates in high yield. Moreover, in case of double dynamic covalent systems (DDS-II), the use of solid state catalyst (ZnBr2) allowed the compatibility between complex thermodynamic and kinetic processes. The resolution process is a significant screening method since the diversity of Strecker compounds can be evaluated at the same time in one pot. Furthermore, the reduced need for purification as well as the low amount used of the toxic cyanide compound is advantageous and environmentally friendly. Moreover, the dynamic cyanohydrin system has also been applied in a DCR process. Due to the efficient thermodynamic property of cyanohydrin formation, it allowed us to visualize the enhancement of amplification through various conditions in real time.

62

We have also formulated and demonstrated the concept of intramolecular dynamic combinatorial resolution (iDCR). In a prototype dynamic nitroaldol system (DS-II), internal selection pressure from a consecutive intramolecular tandem reaction influenced the thermodynamically controlled process and caused a quantitative amplification effect. Further investigation also showed that the tandem reaction was followed by an unexplored type of intramolecular rearrangement, affording an interesting nitrosubstituted isoindolinone structure. This concept has intriguing potential applications: For example, it provides a possible route to systematization of reaction discovery and that tandem reactions could be a tool for controlling dynamic combinatorial systems and it provides a tool to demonstrate dynamics in biased equilibria. Finally, we have demonstrated for the first time that the asymmetric synthesis of β-nitroalkanol derivatives can be simply and efficiently achieved, in a one-pot procedure, by a combined lipase-catalyzed transesterification and nitroaldol (Henry) reaction. Simple tuning of the reaction conditions allows for the isolation of the desired products in good to excellent yield, together with high stereoselecitve, in reasonable reaction times. Moreover, we have discovered that wild-type lipase enzyme has racemase type activity specifically through the N-substituted α-aminonitrile (Strecker) core structure. By use of various control experiments and computational chemistry, the racemization mechanism was proposed to go through the C-C bond-breaking/forming, retro-Strecker/Strecker process, taking place in the same site as the natural binding pocket. We have also demonstrated that the asymmetric synthesis of N-methyl α-aminonitrile derivatives can be simple, green and efficiently achieved by a combined lipase promiscuous racemase activity and lipase-catalyzed transesterification in a direct one-pot reaction. The spontaneous communication between these two processes via a single catalyst afforded the desired products in good yield and with high enantioselectivity. The biocatalyst made the isolation of the desired products easy obtained without loss of selectivity. The catalyst also proved recoverable by ordinary methods, making these systems suitable for upscaling.

63

Acknowledgements

This thesis arose in part out of years of research that has been done since I came to Ramström’s group. By that time, I have worked with a great number of people whose contribution in assorted ways to the research and the making of the thesis deserved success. It is a pleasure to convey my sincere gratitude to them all in my acknowledgment. First and foremost, I would like to thank my supervisor Professor Olof Ramström, for making it possible for me to begin my Ph.D. in Organic synthesis. Thank you for all the ideas and inspiring guidance and for sharing your vast knowledge in chemistry, scientific writing and philosophy in general. I have learned a lot from you over the years which I am extremely grateful. Thank you very much for believing in me and being patience for my slowness, sometime. It was a great experience to work with you. Professor Tobjörn Norin for constructive discussion and valuable comments on this thesis. Brinton Seashore-Ludlow and members in Ramtröm group for taking time and efforts of proof-reading this thesis.

My co-authors: Professor Tore Brinck, Dr. Rikard Larsson, Macus angelin, Mats Linders and Morakot Sakulsombat for making all the work done.

I would also like to thank all present and former members of Ramström group: Zhichao and Hai for making me know more about the “impact factor” and “professor type” in China and for warm welcome during my early time at KTH. Rikard for helping in chemistry during my first year and also never-ending problem with “personbevis”. Marcus for your interesting and excellent suggestion in chemistry and for the “funny forehead”. Remi for being such a very nice friend and letting me know more about politeness of French people. Oscar for your perfective and organization. Morakot, Lingquan “Vince”, René, Catarina, Yan, Adrian and Chelsea for being good friends, for pleasant times in the lab and for spreading a nice atmosphere throughout the group. And my exchange students: Ye and Carole for skilful assistance.

Dr. Ulla Jacobsson and Dr. Zoltan Szabo for supporting with the NMR issues.

Henry Challis, Lena Skowron and Ann Enqvist for all kinds of help with administrative things.

64

All former and present colleagues at the chemistry department for a friendly and pleasant working atmosphere.

The Aulin-Erdtman foundation, Knut och Alice Wallenbergs Stiftelse, and Ragnar of Astrid Signeuls for conferences and traveling grant.

The Swedish Research Council, the Swedish Foundation for International Cooperation in Research and Higher Education and the Carl Trygger foundation are gratefully acknowledged for financial support.

____________________________________________________

I would also like to thank my former supervisor and co-supervisor, Professor Yodhathai Thebtaranonth and Dr. Masahiko Isaka, for inspiration and teaching me a fantastic Organic chemistry, not only in theory but also in practice.

Hesse family, especially Birgitta, Jörgen, Ylva and Tina for supporting and taking care of us since the first day in Sweden. Thank you very much.

My daughter Katrin (Mai) for a wonderful time and for teaching me some Svenska.

My wife, Namphung, for your endless love and support and also believe in me. Your patience, love and encouragement have upheld me. Thank you very much for coming to Sweden with me. I know it was not easy to decide since you got an offer from the great group in USA. Love you so much.

ขอบคณุพอ่, แม,่ พ่ีณัฐ และน้องพมุท่ีคอยดแูลและเป็นกําลังใจให้จอมมาตลอด และยงัเช่ือมัน่ในดวัจอมและปล่อยให้ทําตามอยา่งท่ีจอมต้องการครับ

Thank you very much

65

Appendix

The following is a description of my contribution to Publications I to VI, as requested by KTH. Paper I: I contributed to the formulation of the research problems and I performed majority of the lab work and I wrote the article. Paper II: I contributed to the formulation of the research problems and I performed majority of the lab work and I wrote the article. Paper III: I contributed to the formulation of the research problems and I performed all of the lab work and I wrote the article. Paper IV: I contributed to the formulation of the research problems and I performed some of the lab work Paper V: I contributed to the formulation of the research problems and I performed majority of the lab work and I wrote part of the article. Paper VI: I contributed to the formulation of the research problems and I performed all of the lab work, excluding the calculation study, and I wrote part of the article.

66

References

[1] C. Darwin, John Murray, London 1859. [2] B. E. Murray, R. C. J. Möllering, Medical Clinics of North America

1978, 62, 899. [3] N. Datta, V. M. Hughes, Nature 1983, 306, 616. [4] G. M. Whitesides, R. F. Ismagilov, Science 1999, 284, 89. [5] R. F. Ludlow, S. Otto, Chemical Society Reviews 2008, 37, 101. [6] J. W. Szostak, Nature 2009, 459, 171. [7] P. T. Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor, J. K. M.

Sanders, S. Otto, Chemical Reviews 2006, 106, 3652. [8] S. Ladame, Organic & Biomolecular Chemistry 2008, 6, 219. [9] P. J. Edwards, Drug Discovery Today 2009, 14, 108. [10] A. Herrmann, Organic & Biomolecular Chemistry 2009, 7, 3195. [11] M. F. Schmidt, J. Rademann, Trends in Biotechnology 2009, 27, 512. [12] H. L. L. Châtelier, C. R. Hebd. Seances Acad. Sci. 1884, 99, 786 [13] J. M. Lehn, Chemistry-A European Journal 1999, 5, 2455. [14] J. M. Lehn, Chemical Society Reviews 2007, 36, 151. [15] O. Ramström, J. M. Lehn, Nature Reviews Drug Discovery 2002, 1,

26. [16] N. Giuseppone, G. Fuks, J. M. Lehn, Chemistry-A European Journal

2006, 12, 1723. [17] N. Giuseppone, J. M. Lehn, Angewandte Chemie, International

Edition 2006, 45, 4619. [18] N. Giuseppone, J. M. Lehn, Chemistry-A European Journal 2006, 12,

1715. [19] L. A. Ingerman, M. L. Waters, Journal of Organic Chemistry 2009,

74, 111. [20] T. Bunyapaiboonsri, O. Ramström, S. Lohmann, J.-M. Lehn, L. Peng,

M. Goeldner, ChemBioChem 2001, 2, 438. [21] T. Bunyapaiboonsri, H. Ramström, O. Ramström, J. Haiech, J. M.

Lehn, Journal of Medicinal Chemistry 2003, 46, 5803. [22] R. Larsson, O. Ramström, European Journal of Organic Chemistry

2005, 285. [23] S. A. Poulsen, F. Bornaghi Laurent, Bioorganic & medicinal

chemistry 2006, 14, 3275. [24] B. Shi, R. Stevenson, D. J. Campopiano, M. F. Greaney, Journal of

the American Chemical Society 2006, 128, 8459. [25] A. Valade, D. Urban, J. M. Beau, ChemBioChem 2006, 7, 1023. [26] B. M. R. Lienard, N. Selevsek, N. J. Oldham, C. J. Schofield,

ChemMedChem 2007, 2, 175. [27] A. Valade, D. Urban, J. M. Beau, Journal of Combinatorial

Chemistry 2007, 9, 1. [28] P. Vongvilai, M. Angelin, R. Larsson, O. Ramström, Angewandte

Chemie, International Edition 2007, 46, 948. [29] C. Clavaud, J. Le Gal, R. Thai, M. Moutiez, C. Dugave,

Chembiochem 2008, 9, 1823.

67

[30] T. C. Mark, M. H. Molly, N. Viswanathan, L. S. Robert, M. Taylor, D. F. Amy, K. Cao, A. E. Daniel, Bioorganic & medicinal chemistry letters 2008, 18, 3978.

[31] B. M. R. Lienard, R. Hueting, P. Lassaux, M. Galleni, J. M. Frere, C. J. Schofield, Journal of Medicinal Chemistry 2008, 51, 684.

[32] G. Nasr, E. Petit, D. Vullo, J. Y. Winum, C. T. Supuran, M. Barboiu, Journal of Medicinal Chemistry 2009, 52, 4853.

[33] R. J. Williams, A. M. Smith, R. Collins, N. Hodson, A. K. Das, R. V. Ulijn, Nature Nanotechnology 2009, 4, 19.

[34] P. Vongvilai, O. Ramström, Journal of the American Chemical Society 2009, In press.

[35] M. F. Schmidt, A. El Dahshan, S. Keller, J. Rademann, Angewandte Chemie, International Edition 2009, 48, 6346.

[36] S. Otto, R. L. E. Furlan, J. K. M. Sanders, Science 2002, 297, 590. [37] A. Gonzalez-Alvarez, I. Alfonso, F. Lopez-Ortiz, A. Aguirre, S.

Garcia-Granda, V. Gotor, European Journal of Organic Chemistry 2004, 1117.

[38] R. F. Ludlow, J. Liu, H. Li, S. L. Roberts, J. K. M. Sanders, O. Sijbren, Angewandte Chemie, International Edition 2007, 46, 5762.

[39] K. R. West, R. F. Ludlow, P. T. Corbett, P. Besenius, F. M. Mansfeld, P. A. G. Cormack, D. C. Sherrington, J. M. Goodman, M. C. A. Stuart, S. Otto, Journal of the American Chemical Society 2008, 130, 10834.

[40] P. T. Corbett, J. K. M. Sanders, O. Sijbren, Chemistry-A European Journal 2008, 14, 2153.

[41] M. W. Kanan, M. M. Rozenman, K. Sakurai, T. M. Snyder, D. R. Liu, Nature 2004, 431, 545.

[42] M. Angelin, P. Vongvilai, A. Fischer, O. Ramstroem, Chemical Communications 2008, 768.

[43] C. F. Chow, S. Fujii, J. M. Lehn, Angewandte Chemie, International Edition 2007, 46, 5007.

[44] S. Ulrich, J. M. Lehn, Angewandte Chemie, International Edition 2008, 47, 4462.

[45] A. Ciesielski, G. Schaeffer, A. Petitjean, J.-M. Lehn, P. Samori, Angewandte Chemie, International Edition 2009, 48, 2039.

[46] A. V. Eliseev, M. I. Nelen, Journal of the American Chemical Society 1997, 119, 1147.

[47] S. Otto, R. L. E. Furlan, J. K. M. Sanders, Journal of the American Chemical Society 2000, 122, 12063.

[48] O. Ramström, J. M. Lehn, ChemBioChem 2000, 1, 41. [49] M. S. Goodman, V. Jubian, B. Linton, A. D. Hamilton, Journal of the

American Chemical Society 2002, 117, 11610. [50] P. J. Boul, P. Reutenauer, J. M. Lehn, Organic Letters 2004, 7, 15. [51] R. Larsson, Z. Pei, O. Ramström, Angewandte Chemie, International

Edition 2004, 43, 3716. [52] E. Kolomiets, J. M. Lehn, Chemical Communications 2005, 1519. [53] N. Giuseppone, J.-L. Schmitt, E. Schwartz, J.-M. Lehn, Journal of the

American Chemical Society 2005, 127, 5528. [54] B. R. McNaughton, B. L. Miller, Organic Letters 2006, 8, 1803.

68

[55] B. Klekota, M. H. Hammond, B. L. Miller, Tetrahedron Letters 1997, 38, 8639.

[56] B. Klekota, B. L. Miller, Tetrahedron 1999, 55, 11687. [57] V. Goral, M. I. Nelen, A. V. Eliseev, J. M. Lehn, Proceedings of the

National Academy of Sciences of the United States of America 2001, 98, 1347.

[58] J. Leclaire, L. Vial, S. Otto, J. K. M. Sanders, Chemical Communications 2005, 1959.

[59] R. T. S. Lam, A. Belenguer, S. L. Roberts, C. Naumann, T. Jarrosson, S. Otto, J. K. M. Sanders, Science 2005, 309, 879.

[60] M. Angelin, A. Fischer, O. Ramström, The Journal of Organic Chemistry 2008, 73, 3593.

[61] R. Nguyen, L. Allouche, E. Buhler, N. Giuseppone, Angewandte Chemie, International Edition 2009, 48, 1093.

[62] R. Noyori, Nature Chemistry 2009, 1, 5. [63] B. Michael, D. Klaus, H. Tilo, H. Bernhard, K. E. Maria, S. Rainer,

Z. Thomas, Angewandte Chemie, International Edition 2004, 43, 788. [64] J. Sukumaran, U. Hanefeld, Chemical Society Reviews 2005, 34, 530. [65] E. Fogassy, M. Nogradi, D. Kozma, G. Egri, E. Palovics, V. Kiss,

Organic & Biomolecular Chemistry 2006, 4, 3011. [66] R. Noyori, T. Ikeda, T. Ohkuma, M. Widhalm, M. Kitamura, H.

Takaya, S. Akutagawa, N. Sayo, T. Saito, et al., Journal of the American Chemical Society 1989, 111, 9134.

[67] M. Kitamura, T. Ohkuma, M. Tokunaga, R. Noyori, Tetrahedron: Asymmetry 1990, 1, 1.

[68] O. Pamies, J. E. Bäckvall, Chemical Reviews 2003, 103, 3247. [69] B. Martin-Matute, J. E. Bäckvall, Current Opinion in Chemical

Biology 2007, 11, 226. [70] H. Pellissier, Tetrahedron 2008, 64, 1563. [71] A. H. Kamaruddin, M. H. Uzir, H. Y. Aboul-Enein, H. N. Abdul

Halim, Chirality 2009, 21, 449. [72] L. Henry, C. R. Hebd. Séances Acad. Sci. 1985, 120, 1265. [73] G. Rosini, Comprehensive Organic Synthesis, (Eds.: B.M. Trost, I.

Fleming, C. H. Heathcock), Pergamon, Oxford 1991, 2, 321. [74] F. A. Luzzio, Tetrahedron 2001, 57, 915. [75] N. Ono, The Nitro Group in Organic Synthesis, Wiley-VCH,

Weinheim 2001. [76] J. Boruwa, N. Gogoi, P. P. Saikia, N. C. Barua, Tetrahedron:

Asymmetry 2006, 17, 3315. [77] C. Palomo, M. Oiarbide, A. Laso, European Journal of Organic

Chemistry 2007, 2007, 2561. [78] T. Jiang, H. Gao, B. Han, G. Zhao, Y. Chang, W. Wu, L. Gao, G.

Yang, Tetrahedron Letters 2004, 45, 2699. [79] F. A. Khan, J. Dash, R. Satapathy, S. K. Upadhyay, Tetrahedron

Letters 2004, 45, 3055. [80] C. M. Dobson, Nature 2004, 432, 824. [81] A. Strecker, Ann. Chem. Pharm 1850, 75, 27. [82] H. Groger, Chemical Reviews 2003, 103, 2795.

69

[83] P. Merino, E. Marques-Lopez, T. Tejero, R. P. Herrera, Tetrahedron 2009, 65, 1219.

[84] S. Claude, Angewandte Chemie, International Edition 2004, 43, 1764. [85] G. K. Friestad, A. K. Mathies, Tetrahedron 2007, 63, 2541. [86] V. Banphavichit, W. Mansawat, W. Bhanthumnavin, T. Vilaivan,

Tetrahedron 2004, 60, 10559. [87] A. Berkessel, S. Mukherjee, J. Lex, Synlett 2006, 41. [88] E. J. Corey, M. J. Grogan, Organic Letters 1999, 1, 157. [89] H. Ishitani, S. Komiyama, S. Kobayashi, Angewandte Chemie,

International Edition 1998, 37, 3186. [90] C. A. Krueger, K. W. Kuntz, C. D. Dzierba, W. G. Wirschun, J. D.

Gleason, M. L. Snapper, A. H. Hoveyda, Journal of the American Chemical Society 1999, 121, 4284.

[91] S. Nakamura, N. Sato, M. Sugimoto, T. Toru, Tetrahedron: Asymmetry 2004, 15, 1513.

[92] M. S. Sigman, E. N. Jacobsen, Journal of the American Chemical Society 1998, 120, 5315.

[93] M. S. Sigman, E. N. Jacobsen, Journal of the American Chemical Society 1998, 120, 4901.

[94] M. Takamura, Y. Hamashima, H. Usuda, M. Kanai, M. Shibasaki, Angewandte Chemie, International Edition 2000, 39, 1650.

[95] P. Vachal, E. N. Jacobsen, Organic Letters 2000, 2, 867. [96] P. Vachal, E. N. Jacobsen, Journal of the American Chemical Society

2002, 124, 10012. [97] A. G. Wenzel, M. P. Lalonde, E. N. Jacobsen, Synlett 2003, 1919. [98] J. H. Atherton, J. Blacker, M. R. Crampton, C. Grosjean, Organic &

Biomolecular Chemistry 2004, 2, 2567. [99] W. H. J. Boesten, J.-P. G. Seerden, B. de Lange, H. J. A. Dielemans,

H. L. M. Elsenberg, B. Kaptein, H. M. Moody, R. M. Kellogg, Q. B. Broxterman, Organic Letters 2001, 3, 1121.

[100] R. Sakurai, S. Suzuki, J. Hashimoto, M. Baba, O. Itoh, A. Uchida, T. Hattori, S. Miyano, M. Yamaura, Organic Letters 2004, 6, 2241.

[101] B. A. Horenstein, K. Nakanishi, Journal of the American Chemical Society 1989, 111, 6242.

[102] J. Mulzer, A. Meier, J. Buschmann, P. Luger, Synthesis-Stuttgart 1996, 123.

[103] G. K. S. Prakash, T. Mathew, C. Panja, S. Alconcel, H. Vaghoo, C. Do, G. A. Olah, Proceedings of the National Academy of Sciences of the United States of America 2007, 104, 3703.

[104] A. M. Whitney, S. Ladame, S. Balasubramanian, Angewandte Chemie, International Edition 2004, 43, 1143.

[105] S. Ladame, A. M. Whitney, S. Balasubramanian, Angewandte Chemie, International Edition 2005, 44, 5736.

[106] P. C. Gareiss, K. Sobczak, B. R. McNaughton, P. B. Palde, C. A. Thornton, B. L. Miller, Journal of the American Chemical Society 2008, 130, 16254.

[107] A. Schmid, J. S. Dordick, B. Hauer, A. Kiener, M. Wubbolts, B. Witholt, Nature 2001, 409, 258.

70

[108] H. E. Schoemaker, D. Mink, M. G. Wubbolts, Science 2003, 299, 1694.

[109] V. Gotor-Fernandez, R. Brieva, V. Gotor, Journal of Molecular Catalysis B:Enzymatic 2006, 40, 111.

[110] A. Ghanem, Tetrahedron 2007, 63, 1721. [111] R. D. Schmid, R. Verger, Angewandte Chemie, International Edition

1998, 37, 1608. [112] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Kesseler, R.

Sturmer, T. Zelinski, Angewandte Chemie, International Edition 2004, 43, 788.

[113] U. T. Bornscheuer, R. J. Kazlauskas, Hydrolases in Organic Synthesis: Regio- and Stereoselective Biotransformations, 2nd edition Wiley-VCH:Weinheim 2005.

[114] Protein Data Bank, http://www.rcsb.org/pdb/. [115] A. Zaks, A. M. Klibanov, Proceedings of the National Academy of

Sciences of the United States of America 1985, 82, 3192. [116] Q. Jing, R. J. Kazlauskas, Chirality 2008, 20, 724. [117] T. Ema, J. Kobayashi, S. Maeno, T. Sakai, M. Utaka, Bulletin of the

Chemical Society of Japan 1998, 71, 443. [118] M. Cygler, P. Grochulski, R. J. Kazlauskas, J. D. Schrag, F.

Bouthillier, B. Rubin, A. N. Serreqi, A. K. Gupta, Journal of the American Chemical Society 2002, 116, 3180.

[119] R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport, L. A. Cuccia, Journal of Organic Chemistry 1991, 56, 2656.

[120] R. Sato, T. Senzaki, T. Goto, M. Saito, Chemistry Letters 1984, 13, 1599.

[121] R. Sato, M. Ohmori, F. Kaitani, A. Kurosawa, T. Senzaki, T. Goto, M. Saito, Bulletin of the Chemical Society of Japan 1988, 7, 2481.

[122] S. Young Seok, L. Chang Hoon, L. Kee-Jung, Journal of Heterocyclic Chemistry 2003, 40, 939.

[123] D. L. Comins, S. Schilling, Y. Zhang, Organic Letters 2005, 7, 95. [124] I. R. Hardcastle, S. U. Ahmed, H. Atkins, G. Farnie, B. T. Golding,

R. J. Griffin, S. Guyenne, C. Hutton, P. Kallblad, S. J. Kemp, M. S. Kitching, D. R. Newell, S. Norbedo, J. S. Northen, R. J. Reid, K. Saravanan, H. M. G. Willems, J. Lunec, Journal of Medicinal Chemistry 2006, 49, 6209.

[125] M. Lamblin, A. Couture, E. Deniau, P. Grandclaudon, Organic & Biomolecular Chemistry 2007, 5, 1466.

[126] C. Kison, N. Meyer, T. Opatz, Angewandte Chemie, International Edition 2005, 44, 5662.

[127] D. Lucet, T. L. Gall, C. Mioskowski, Angewandte Chemie, International Edition 1998, 37, 2580.

[128] L. Denis, S. Stephane, K. Olivier, G. Thierry Le, M. Charles, European Journal of Organic Chemistry 1999, 1999, 2583.

[129] F. van Rantwijk, R. A. Sheldon, Tetrahedron 2004, 60, 501. [130] V. Gotor-Fernandez, E. Busto, V. Gotor, Advanced Synthesis &

Catalysis 2006, 348, 797. [131] G. Asensio, C. Andreu, J. A. Marco, Tetrahedron Letters 1991, 32,

4197.

71

[132] A. J. Blacker, M. J. Stirling, M. I. Page, Organic Process Research & Development 2007, 11, 642.

[133] G. F. Breen, Tetrahedron: Asymmetry 2004, 15, 1427. [134] T. W. Chiou, C. C. Chang, C. T. Lai, D. F. Tai, Bioorganic &

Medicinal Chemistry Letters 1997, 7, 433. [135] V. Gotor-Fernandez, P. Fernandez-Torres, V. Gotor, Tetrahedron:

Asymmetry 2006, 17, 2558. [136] A. Liljeblad, H. M. Kavenius, P. Taehtinen, L. T. Kanerva,

Tetrahedron: Asymmetry 2007, 18, 181. [137] T. Ema, M. Jittani, K. Furuie, M. Utaka, T. Sakai, Journal of Organic

Chemistry 2002, 67, 2144. [138] P. Vongvilai, R. Larsson, O. Ramström, Advanced Synthesis &

Catalysis 2008, 350, 448. [139] Y. N. Belokon, A. J. Blacker, L. A. Clutterbuck, D. Hogg, M. North,

C. Reeve, European Journal of Organic Chemistry 2006, 4609. [140] J. Holt, U. Hanefeld, Current Organic Synthesis 2009, 6, 15. [141] D. V. Johnson, A. A. Zabelinskaja-Mackova, H. Griengl, Current

Opinion in Chemical Biology 2000, 4, 103. [142] M. North, D. L. Usanov, C. Young, Chemical Reviews 2008, 108,

5146. [143] M. Inagaki, J. Hiratake, T. Nishioka, J. Oda, Bulletin of the Institute

for Chemical Research 1989, 67, 132. [144] M. Inagaki, J. Hiratake, T. Nishioka, J. Oda, Journal of the American

Chemical Society 1991, 113, 9360. [145] M. Inagaki, A. Hatanaka, M. Mimura, J. Hiratake, T. Nishioka, J.

Oda, Bulletin of the Chemical Society of Japan 1992, 65, 111. [146] M. Inagaki, J. Hiratake, T. Nishioka, J. Oda, Journal of Organic

Chemistry 1992, 57, 5643. [147] H. Griengl, H. Schwab, M. Fechter, Trends in Biotechnology 2000,

18, 252. [148] L. Veum, L. T. Kanerva, P. J. Halling, T. Maschmeyer, U. Hanefeld,

Advanced Synthesis & Catalysis 2005, 347, 1015. [149] E. M. Anderson, K. M. Larsson, O. Kirk, Biocatalysis and

Biotransformation 1998, 16, 181. [150] E. Fukusaki, S. Satoda, Journal of Molecular Catalysis B: Enzymatic

1997, 2, 257. [151] T. Ema, S. Maeno, Y. Takaya, T. Sakai, M. Utaka, Tetrahedron:

Asymmetry 1996, 7, 625. [152] M. Persson, D. Costes, E. Wehtje, P. Adlercreutz, Enzyme and

Microbial Technology 2002, 30, 916. [153] T. Ema, Tetrahedron: Asymmetry 2004, 15, 2765. [154] T. Sakai, Tetrahedron: Asymmetry 2004, 15, 2749. [155] T. Sakai, K. Hayashi, F. Yano, M. Takami, M. Ino, T. Korenaga, T.

Ema, Bulletin of the Chemical Society of Japan 2003, 76, 1441. [156] T. Sakai, T. Kishimoto, Y. Tanaka, T. Ema, M. Utaka, Tetrahedron

Letters 1998, 39, 7881. [157] M. Berger, B. Albrecht, A. Berces, P. Ettmayer, W. Neruda, M.

Woisetschlager, Journal of Medicinal Chemistry 2001, 44, 3031.

72

[158] J. W. Nieuwenhuijzen, R. F. P. Grimbergen, C. Koopman, R. M. Kellogg, T. R. Vries, K. Pouwer, E. van Echten, B. Kaptein, L. A. Hulshof, Q. B. Broxterman, Angewandte Chemie, International Edition 2002, 41, 4281.

[159] M. F. A. Adamo, V. K. Aggarwal, M. A. Sage, Journal of the American Chemical Society 2000, 122, 8317.

[160] K. W. Henderson, W. J. Kerr, J. H. Moir, Chemical Communications 2000, 479.

[161] U. T. Bornscheuer, R. J. Kazlauskas, Angewandte Chemie, International Edition 2004, 43, 6032.

[162] A. Aharoni, L. Gaidukov, O. Khersonsky, S. M. Gould, C. Roodveldt, D. S. Tawfik, Nature Genetics 2005, 37, 73.

[163] K. Hult, P. Berglund, Trends in Biotechnology 2007, 25, 231. [164] P. J. O'Brien, D. Herschlag, Journal of the American Chemical

Society 1998, 120, 12369. [165] I. Catrina, P. J. O'Brien, J. Purcell, I. Nikolic-Hughes, J. G. Zalatan,

A. C. Hengge, D. Herschlag, Journal of the American Chemical Society 2007, 129, 5760.

[166] A. Ercan, H. I. Park, L.-J. Ming, Biochemistry 2006, 45, 13779. [167] T. Kitazume, T. Ikeya, K. Murata, Journal of the Chemical Society,

Chemical Communications 1986, 1331 [168] O. Torre, I. Alfonso, V. Gotor, Chemical Communications 2004,

1724. [169] P. Carlqvist, M. Svedendahl, C. Branneby, K. Hult, T. Brinck, P.

Berglund, ChemBioChem 2005, 6, 331. [170] M. Svedendahl, K. Hult, P. Berglund, Journal of the American

Chemical Society 2005, 127, 17988. [171] C. Branneby, P. Carlqvist, A. Magnusson, K. Hult, T. Brinck, P.

Berglund, Journal of the American Chemical Society 2003, 125, 874. [172] Q.-M. Wang, T. C. W. Mak, Chemical Communications 2002, 2682. [173] F. P. Seebeck, D. Hilvert, Journal of the American Chemical Society

2003, 125, 10158. [174] Y. Terao, K. Miyamoto, H. Ohta, Chemistry Letters 2007, 36, 420. [175] P. Carlqvist, M. Svedendahl, C. Branneby, K. Hult, T. Brinck, P.

Berglund, ChemBioChem 2005, 6, 331.