Metals in Dynamic Chemistry:

Selection & Catalysis

Brian J.J. Timmer

Doctoral Thesis

Stockholm 2017

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av doktorsexamen i kemi med inriktning mot organisk kemi fredagen den 8 september kl 09.00 i sal F3, KTH, Lindstedtsvägen 26, Stockholm. Avhandlingen försvaras på engelska. Opponent är Dr. Stéphane Bellemin-Laponnaz, University of Strasbourg

II

Brian J.J. Timmer, 2017: ”Metals in Dynamic Chemistry: Selection & Catalysis”, KTH Chemical Science and Engineering, Royal Institute of Technology, SE-100 44 Stockholm, Sweden.

ISBN 978-91-7729-470-2 ISSN 1654-1081 TRITA-CHE-Report 2017:32 © Brian J.J. Timmer, 2017 Universitetsservice US-AB, Stockholm

III

Abstract

In the adaptation to the oxidative environment on earth, metals played a crucial role for the evolution of life. The presence of metals also allowed access to advanced functions due to their unique coordination sphere and reactivity. This thesis focused on exploiting these unique properties for further development of the field of dynamic chemistry – a field in which adaptation plays a central role as well.

The first part of the thesis aimed to create a better understanding of multivalent effects in carbohydrate-lectin interactions. By reversible ligand coordination to zinc ions one of the nanoplatforms, the Borromean rings, could be selectively obtained. After carbohydrate functionalization the binding events were monitored by quartz crystal microbalance technology and compared to glycosylated fullerenes and dodecaamide cages. Overall, this investigation indicated that statistical and polyelectrolyte effects play a considerable role in the observed multivalent effects.

The second part of the thesis aimed to design and synthesize a new catalyst for application in aqueous olefin metathesis. This afforded a ruthenium based catalyst that was applied in the self- and cross-metathesis of highly functionalized substrates, such as carbohydrates. In addition, it was shown that the presence of a small amount of acetic acid prevented undesired double bond isomerization.

The last part of the thesis aimed to explore new methods to discover transition metal catalysts. Dynamic exchange of directing groups generated a pool of potential substrates for C-H activation. Combining this pool of substrates with a pool of potential catalysts resulted in amplification of a reactive substrate/metal combination. By iterative deconvolution in combination with mass spectrometry this “intermediate” could be identified from the mixture, proving applicability of this alternative approach to catalyst discovery.

Keywords: constitutional dynamic chemistry, selection, multivalent

interactions, carbohydrates, quartz crystal microbalance, catalyst screening, aqueous olefin metathesis, C-H activation

IV

Sammanfattning på svenska

I anpassningen till den oxidativa miljön på jorden spelade metaller en avgörande roll för livets utveckling. Närvaron av metaller gav också tillgång till avancerade funktioner på grund av deras unika koordinationssfär och reaktivitet. Denna avhandling fokuserade på att utnyttja dessa unika egenskaper för vidareutveckling av dynamisk kemi – ett område där adaption också spelar en central roll.

Den första delen av avhandlingen syftade till att skapa en bättre förståelse av multivalenta effekter i kolhydrat-lektininteraktioner. Genom reversibla ligandkoordineringar till zinkjoner kan en av nanoplattformerna, de Borromeiska ringarna, erhållas selektivt. Efter kolhydratfunktionalisering studerades inbindningar med kvartskristall mikrobalans teknologi och jämfördes med glykosylerade fullerener och dodekaamidburar. Undersökningen indikerar att statistiska och polyelektrolytiska effekter spelar en betydande roll i de observerade multivalenta effekterna.

Den andra delen av avhandlingen syftade till att utveckla och syntetisera en ny katalysator för olefinmetates i vatten. Den modifierade ruteniumbaserade katalysatorn användes i själv- och korsmetatesen av mycket funktionaliserade substrat, såsom kolhydrater. Studien visade också att tillsats av en liten mängd av ättiksyra förhindrade oönskade dubbelbindningsisomerisering.

Den sista delen av avhandlingen syftade till att utforska nya metoder för att upptäcka övergångsmetallkatalysatorer. Dynamisk utbyte av styrande grupper genererade en samling av potentiella substrat för C-H aktivering. Kombinering av denna samling av substrat med en grupp av potentiella katalysatorer resulterade i amplifiering av en reaktiv substrat/metallkombination. Genom iterativ dekonvolution i kombination med massspektrometri kunde denna ”intermediär” identifieras från blandningen, vilket bevisar användbarheten av denna alternativa metod för att upptäcka katalysatorer.

Nyckelord: konstitutionell dynamisk kemi, selektion, multivalenta

interaktioner, kolhydrater, kvartskristall mikrobalans, katalysator screening, vattenhaltig olefinmetates, C-H-aktivering

V

Abbreviations

BSA bovine serum albumin CDC constitutional dynamic chemistry conA concanavalin A CuAAC CuI-catalysed azide-alkyne cycloaddition DBU 1,8-Diazabicyclo[5.4.0]undec-7-ene DG directing group DNA deoxyribonucleic acid EDC N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide·HCl ESI-MS electrospray ionization mass spectrometry equiv. equivalents GNA Galanthus nivalis agglutinin LA Lewis acid Mes mesityl MIC minimum inhibitory concentration MS molecular sieves NMR nuclear magnetic resonance RDS rate determining step rt room temperature sulfo-NHS N-hydroxysulfosuccinimide sodium salt TS transition state TSA transition state analogue QCM quartz crystal microbalance

VI

List of Publications

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

I. Spatially Well-Defined Carbohydrate Nanoplatforms:

Synthesis, Characterization and Lectin Interaction

Study Brian J.J. Timmer,‡ Marta Abellán-Flos,‡ Lars Mønster Jørgensen, Davide Proverbio, Samuel Altun, Olof Ramström, Teodor Aastrup and Stéphane P. Vincent Chem. Commun. 2016, 52, 12326-12329

II. Key Factors for Multivalency Effects in Carbohydrate

Nanoplatform-Lectin Interactions Marta Abellán-Flos,‡ Brian J.J. Timmer,‡ Samuel Altun, Teodor Aastrup, Stéphane P. Vincent and Olof Ramström Draft manuscript

III. Acid-Assisted Direct Olefin Metathesis of Unprotected

Carbohydrates in Water

Brian J.J. Timmer and Olof Ramström To be submitted

IV. Selective Cross-Metathesis of Highly Functionalized

Substrates in Aqueous Media Brian J.J. Timmer, Oleksandr Kravchenko and Olof Ramström Draft manuscript

V. Resolving a Reactive Organometallic Intermediate from

Dynamic Directing Group Systems by Selective C-H

Activation Fredrik Schaufelberger, Brian J. J. Timmer and Olof Ramström To be submitted

VII

Papers not included in this thesis:

I. Constitutional Dynamic Chemistry Lei Hu, Fredrik Schaufelberger, Brian J. J. Timmer, Marta Abellán-Flos and Olof Ramström In: Kirk-Othmer Encyclopedia of Chemical Technology, Wiley-VCH, Weinheim, 2014

II. Principles of Dynamic Covalent Chemistry Fredrik Schaufelberger, Brian J. J. Timmer and Olof Ramström In: Dynamic Covalent Chemistry: Principles, Reactions

and Applications (ed. W. Zhang), John Wiley & Sons, New York, 2017 In press

III. Combining Sustainable Chemistry Education with

Undergraduate Research – Student-Driven

“Greenification” of an Organic Chemistry Laboratory

Course Fredrik Schaufelberger, Sara Starrsjö, Olof Sjölin, Allan Starkholm, Brian J. J. Timmer and Olof Ramström To be submitted

IV. Simple and Effective Integration of Green Chemistry

and Sustainability Education into an Existing Organic

Chemistry Course Brian J. J. Timmer, Fredrik Schaufelberger, Daniel Hammarberg, Johan Franzén, Olof Ramström and Peter Dinér To be submitted

VIII

Table of Contents

Abstract ................................................................................ III Sammanfattning på svenska ................................................. IV

Abbreviations ......................................................................... V

List of Publications ................................................................ VI Table of Contents ................................................................ VIII 1. Introduction ....................................................................... 1

1.1 Dynamic Chemistry ............................................................................. 2

1.2 Metals in Dynamic Chemistry ............................................................. 4

1.2.1 Selection ................................................................................................ 5

1.2.2 Catalysis ................................................................................................ 8

1.3 Contents of This Thesis .................................................................... 12

2. Selection of Nanoplatforms for Application in Studying Multivalent Interactions ................................................... 13

2.1 Multivalent Interactions in Biological Systems .................................. 13

2.1.1 Carbohydrates in Biological Interactions .............................................. 13

2.2 Multifunctional Nanoplatforms .......................................................... 14

2.2.1 Nanoplatforms by CDC ........................................................................ 14

2.2.2 Selected Nanoplatforms in Current Study ............................................ 15

2.3 Nanoplatform Synthesis and Functionalization ................................. 17

2.4 Characterization of Multivalent Interactions ...................................... 21

2.4.1 Quartz Crystal Microbalance Technology ............................................. 21

2.4.2 Set-up of Binding Experiments ............................................................. 23

2.4.3 Binding Experiments with Concanavalin A ........................................... 24

2.4.4 Binding Experiments with Galanthus nivalis Agglutinin ......................... 28

2.4.5 Binding Experiments with Bacteria ....................................................... 31

2.5 Conclusions ...................................................................................... 31

3. Catalysis of Olefin Exchange Under Aqueous Conditions 33

3.1 Olefin Metathesis .............................................................................. 33

3.1.1 Catalysts for Olefin Metathesis............................................................. 33

3.1.2 Olefin Metathesis in CDC ..................................................................... 34

3.1.3 Aqueous Olefin Metathesis .................................................................. 35

3.2 Olefin Metathesis of Unprotected Carbohydrates ............................. 37

3.2.1 Catalyst Design and Synthesis............................................................. 37

3.2.2 Optimization of Metathesis Conditions ................................................. 38

3.2.3 Non-productive Chelation by Carbohydrates ........................................ 40

3.2.4 Application Range of Carbohydrates in Olefin Metathesis .................... 42

3.3 Olefin Metathesis of Other Chelating Substrates .............................. 44

3.3.1 Activity in Self-Metathesis .................................................................... 44

IX

3.3.2 Cross-Metathesis with Allyl Alcohol ...................................................... 45

3.4 Conclusions ...................................................................................... 47

4. Metal Catalysts as Selectors in Dynamic Systems .......... 49

4.1 Catalyst Discovery ............................................................................ 49

4.1.1 Catalyst Discovery in Dynamic Chemistry ............................................ 50

4.2 Hydroacylation Reaction ................................................................... 52

4.2.1 Trapping and Detection of Reactive Combination ................................ 53

4.2.2 Parameters in Substrate and Catalyst Pool .......................................... 54

4.2.3 Deconvolution of Substrate and Catalyst Pool ..................................... 55

4.2.4 Efficiency of the Selection .................................................................... 57

4.2.5 Catalytic Role of the Reactive Intermediate .......................................... 60

4.3 Conclusions ...................................................................................... 60

5. Concluding Remarks ...................................................... 61

Acknowledgements .............................................................. 62

Appendix .............................................................................. 64

References .......................................................................... 65

X

1

1. Introduction

Living matter consists for more than 99% of the six elements carbon, hydrogen, oxygen, nitrogen, phosphorus and sulphur. These six elements form the framework and inherent genetic information of every living organism. However, even though very few elements constitute the absolute majority of the building blocks of life, emergence of some of the advanced features of living systems required the presence of elements with more complex reactivity such as metals. Initially, the primary importance of metals was their reductive nature that lead to the formation of essential precursors of biomolecules in the oxidative environment on earth. Now biological systems have evolved to employ metals in other processes as well such as selection, transport and catalysis (Figure 1).1 Many such processes are unravelled every year, yet the mystery of how exactly life emerged is still under heavy investigation. One statement supported throughout all fields of science is that life emerged through the adaptive behaviour of molecular systems. The study and exploitation of such adaptive behaviour in molecular systems is part of a field called dynamic chemistry.2

Figure 1. Two examples of the appearance of metals in more advanced processes.

On the left adenosine triphosphate bound to a magnesium ion is shown, which activates it towards nucleophilic attack. On the right heme B is shown, in which the

central iron ion is required to bind oxygen to allow for its transportation.

2

1.1 Dynamic Chemistry

For centuries, it was desirable for chemists to work with transformations that provide single reaction products. However, the continuous development of sensitive analytical methods opened up new possibilities such as the ability to study mixtures instead of only characterizing single components. This ability was vital for formation of the field of dynamic chemistry, often also called adaptive chemistry or constitutional dynamic chemistry (CDC).3-6 As mentioned before, the field of dynamic chemistry studies and exploits the adaptive behaviour of molecular systems. It is therefore not only essential to be able to analyse such systems, but it is equally important that the mixture can exert adaptive behaviour. In other words, instead of requiring bonds to be kinetically stable, CDC uses bonds which are labile and thus reversibly formed. This is a property that is found in the majority of natural molecular systems. A classic example of a molecular system based on reversible bond formation is the DNA molecule (Figure 2). The DNA molecule secures biological information by masking the functionality present in the nucleotides of the separate strands in its double helix structure. This masking is done by using supramolecular interactions like H-bonding and π-π interactions. Access to this information is made possible by entities that are capable of breaking these H-bonding and π-π interactions (e.g. DNA Helicase).

Figure 2. Fragment of DNA with two nucleotide pairs enlarged to indicate the π-π and H-bonding interactions that are partly responsible for the unique shape and

properties of this class of molecules.

3

H-bonding, π-π and other non-covalent interactions are part of a division of CDC called dynamic non-covalent chemistry. Non-covalent interactions are well-known to be weak and to gain stability from additivity or cooperativity.7 For example, in the formation of self-assembled monolayers the stability of the assembly depends partly on the additivity of the van der Waals interactions. In smaller molecules only lower multiplicity of interactions can be achieved. Although this provides reversibility to the system, this also decreases the robustness of the assemblies. To support the demand for robustness, CDC also encompasses the field dynamic covalent chemistry.8-10 This field focuses on generating conditions to reversibly form and break covalent bonds. Up to date, many different reversible covalent linkages have been discovered amongst which imine, disulfide and boronic ester bonds have been most widely utilized (Figure 3).

Figure 3. The field of constitutional dynamic chemistry utilizes reversible behaviour in molecules over a wide range of bond strengths which can be divided in dynamic

non-covalent chemistry and dynamic covalent chemistry.

From the examples touched upon thus far, it is noticeable that the framework of CDC adopts an equal subset of elements as present in living systems. These elements form the functional groups that interact and adapt to the environment to form the most stable mixture of constituents. Similar to living systems, metals have also played a prominent role in the development of dynamic chemistry.

4

1.2 Metals in Dynamic Chemistry

Metals have played an important role in the development of dynamic chemistry due to their unique reactivity.11 This reactivity is primarily originating from the higher atomic number, which provides metals with more electrons and hereby higher variety of electron configurations. For the first three periods in the periodic table, the electrons mainly occupy s- and p-orbitals. Combination of these two orbital types gives access to sp, sp2 and sp3 hybridized orbitals which provide the molecules with a certain geometry. When continuing down the periodic table, where the majority of the metals are located, the electrons can also occupy d-orbitals (Figure 4A). Due to the unique shape and directionality of the d-

orbitals, more complex molecular geometries can be accessed when valence shell electrons occupy these orbitals. However, not only the geometry around the metal ions differs depending on their electron configuration, but their ability to donate and accept electrons alters as well. The reactivity of metal complexes allows, for example, for reactions such as oxidative addition and reductive elimination to occur, but also ligand association and dissociation (Figure 4).

Figure 4. (A)Shape and directionality of d-orbitals and (B) a few of the possible

geometries and reactions for metal complexes having d-orbitals in the valence shell.

Both these characteristics of metals play a very distinctive role in CDC. The shape of the ion can be used in combination with reversible metal coordination to select for different species in molecular systems. On the other hand, the reactivity of metals allows for reversibility of new bonds by acting as the catalyst in the reaction.

5

1.2.1 Selection

A key feature of CDC is its adaptive behaviour and as the bonds employed are reversibly being formed and broken, the molecular mixtures will head towards thermodynamic equilibrium. Upon applying internal or external pressure, the molecular system will rearrange until a new thermodynamic equilibrium is achieved. The ‘pressure’ applied can thus select for different components in the mixture.12-15 To achieve external pressure a selector needs to be added, which depending on its nature can results in casting or moulding of the components of the mixture (Figure 5).16 In casting of a substrate the ´keys´ are assembled from different building blocks that exchange to form a mixture of all building block combinations. Addition of a ‘keyhole’ stabilizes and hereby amplifies the best fitting key. This approach is especially attractive for discovery of drugs as strong binders are amplified from the mixture.17 When the selector is the ‘key’ it is possible to select for a ‘lock’ or receptor, a feature useful in the amplification of potential catalysts and complex molecular architectures.

Figure 5. The concepts of casting of a substrate and moulding of a receptor in CDC,

illustrated using a ´lock and key´ analogy.

6

One of the initial examples of receptor moulding in CDC was made by Lehn, where selection of a single constituent was achieved in a dynamic mixture of metal ions and ligands.18-19 Reversible metal-coordination of a trimeric bipyridine ligand with FeX2 gave a mixture of circular helicates consisting of four to six ligands. Depending on the counterion used, however, one of the helicates could be selectively formed from the mixture (Figure 6).

Figure 6. Counterion selection of pentameric and hexameric circular helicates from

a molecular mixture of iron(II) ions and trimeric bipyridine ligands.

In the previous example, the selection of single or multiple constituents from the mixture was directed by templating of the counterion. However, metal ions can also act as templates and thus act as the ‘keys’ that define the shape of the amplified ‘locks’. A prominent example where the metal ion is used as template is made by the von Delius group. They developed Brønsted acid catalysed conditions to reversibly exchange alcohols on orthoesters.20 By introduction of diols, a dynamic mixture is obtained, which upon templating with sodium ions selectively forms cryptates encapsulating this ion.21 Further investigation showed the adaptive behaviour of these cryptates as different metal ions could be encapsulated when appropriate diols are introduced to increase the size of the cavity in the cryptate (Figure 7).22

7

Figure 7. Reversible orthoester exchange and selection of cryptates based on metal

ion templating.

From these examples it is already apparent that metal ions provide accessibility to a great variety of complex molecular architectures. The thermodynamic nature of the reversible metal-ligand coordination can result in selection of the most stable species in high yields. This also provides a certain degree of programmability to the system, where selection can be driven by the information supplied.23-25 In other words, by careful choice of metal/ligand combinations the outcome can be predicted and selected for. In a study by Lehn, it is shown that different complex molecular architectures can be achieved from the same ligand containing bi- and tridentate binding sites. By changing the information provided to the system, it can be programmed to give different outcomes. The information provided in their example is metal ions having different coordination spheres, ranging from tetra- to hexacoordinate (Figure 8).26

Figure 8. Programmable selection of different complex molecular architectures.

8

In summary, metal ions in dynamic chemistry provide a unique opportunity to select for single constituents from mixtures of compounds. As metal ions exhibit a pre-set coordination sphere, this information can be used in combination with rational design of the ligand. Altogether, this provides a facile route to access complex molecular architectures as compared to traditional synthesis.

1.2.2 Catalysis

The second role that metals play in CDC is as catalysts of reversible exchange. A property of metal ions that plays a major role is their inherent Lewis acidity. By coordination to a metal ion an electrophile can be activated, facilitating nucleophilic addition that takes place during the exchange process (Figure 9). As a result, exchange rates increase and equilibrium conditions are met in shorter time frames. In the case of exergonic reactions, this type of catalysis can even render reverse processes accessible.

Figure 9. Activation of electrophile for nucleophilic attack lowering the activation

barrier.

A prime example of an exchange reaction catalysed by Lewis acids (LAs) is the transimination reaction (Figure 10A). Several metal salts can be used for this process, and zinc(II) and scandium(III) salts have been shown to be especially efficient in facilitating this exchange.27-28 Compared to Brønsted acids, LAs usually give less deactivation of the nucleophile and the catalyst during transimination. This aspect is especially apparent in hydrazone exchange, where strong acid catalysis results in decomposition, but efficient exchange can be achieved using Lewis acid catalysis (Figure 10B).29 Further examples of LA catalysed reactions include transamidation,30-32 Strecker reactions33 and thioacetal formations (Figure 10C-E).34

9

Figure 10. Examples of Lewis acid (LA) promoted reversible exchange reactions

including (A) transimination, (B) hydrazone exchange, (C) transamidation, (D) Strecker reaction and (E) thioacetal formation.

In LA catalysis the metal is required to accept electrons and therefore high oxidation states of the metals are utilized. With lower oxidation states of transition metals, however, it is also possible for the metal to donate electrons to a substrate instead. In this process, the metal loses electrons and increases in oxidation state. In pioneering work of Amatore it is shown that formation of cationic π-allylpalladium(II) complexes by oxidative addition of allyl acetate is reversible (Scheme 1).35 This reversibility introduces thermodynamic control to the reaction as well, a property utilized by Sanders to synthesize porphyrin macrocycles.36

Scheme 1. Palladium-catalysed reversible allyl transesterification.

10

Another type of dynamic organometallic reactions are the metathesis type reactions. In this type of reactions double displacement of the substituents occurs, resulting in scrambling of the different groups. Two types of metal catalysed metathesis reactions have been used often in CDC – alkyne and olefin metathesis. In alkyne metathesis two alkyne fragments are reacted to form two new alkynes with scrambled substituents. Catalysts for alkyne metathesis are generally based on either tungsten or molybdenum. In dynamic applications the more novel Schrock-type molybdenum(VI) based catalysts are employed due to their significantly decreased moisture and air sensitivities (Scheme 2).37-38

Scheme 2. General reaction scheme for alkyne metathesis and a selection of

molybdenum(VI) catalysts used for dynamic exchange.

Unfortunately, the instability of the catalyst limits the use of alkyne metathesis in CDC, as many turnovers are commonly required to reach and maintain equilibrium. Instead the high rigidity and directionality of alkyne exchange has been employed in the programmed assembly of macrocycles,39-40 cages41 and catenanes.42

Despite all the developments in alkyne metathesis, it is still not as commonly employed in CDC as its more powerful olefin counterpart. In olefin metathesis two olefin fragments are reacted together to form two new olefins with scrambled substituents. Ever since the discovery of the reaction many catalysts have been developed, amongst which the most commonly used are the commercially available Grubbs-type catalysts (Scheme 3).43

11

Scheme 3. General reaction scheme for alkene metathesis and a selection of

Grubbs-type ruthenium(II)-based catalysts used for dynamic exchange.

The reaction progress for olefin metathesis is similar to that of alkyne metathesis and follows a [2+2]-cycloaddition pathway followed by reverse cycloaddition to form the new set of alkenes. This new set of alkenes generally contains both the E- and Z-isomers, although the latter is usually favoured (Scheme 4). All steps of this reaction are in principle reversible, however there is usually a large reactivity difference between terminal and internal olefins. This can limit the application in CDC as the reaction is prone to kinetic traps. The high catalyst activity allows for fast equilibration, but makes it hard to halt equilibration during analysis as even trace amounts of catalyst can promote further exchange.44

Scheme 4.General mechanism for the olefin metathesis reaction.

The advantages of olefin metathesis include the mild reaction conditions and wide functional group tolerance, which makes it applicable without interfering with other dynamic exchange reactions.45-

46 Overall, these properties make olefin metathesis an interesting option for use in CDC.

12

1.3 Contents of This Thesis

This thesis broadly concerns the unique properties and applications of metals in CDC. The second chapter focuses on the ability of metal ions to select for complex molecular architectures. This property is adopted to synthesize novel carbohydrate nanoplatforms, which are used to gain increased understanding of carbohydrate/receptor binding. The third chapter deals with the unique reactivity of metal catalysts in olefin metathesis. By adjusting the catalyst and reaction conditions an expansion of the scope of olefin metathesis is made. The last chapter focuses on catalyst discovery and combines the ability of metals in CDC to select and catalyse. Herein, dynamic exchange of directing groups in C-H activation results in selection of an active catalyst/substrate combination, which is successfully identified from the mixture.

13

2. Selection of Nanoplatforms for Application

in Studying Multivalent Interactions

(Paper I-II) The first scientific chapter will discuss the use of metals in selection

of complex molecular architectures using the toolbox provided by CDC. This ability to select can provide easy access to unique and well-defined nanoplatforms from mixtures of components. Herein, the selection, functionalization and application of two nanoplatforms in the study of multivalent interactions in biological systems is described.

2.1 Multivalent Interactions in Biological Systems

Many of the interactions in biological systems are of non-covalent nature. As most of these non-covalent interactions are relatively unstable, multiplicity is required for the formation of stable constructs. Constructs which are stabilized by multivalent interactions are for example DNA molecules, proteins and cell membranes. However, the role of multivalent interactions in biological systems is not only to stabilize constructs, they also play a role in the transfer of information. A key class of molecules responsible for this transfer of information is the carbohydrates.

2.1.1 Carbohydrates in Biological Interactions

Carbohydrates play an essential role in living systems. For example, by presentation of carbohydrates on the cell surface in form of the glycocalyx, cells can interact with other entities such as lectins, antibodies, viruses and enzymes (Figure 11). Due to these interactions carbohydrates have played a critical role in infection, modulation, regulation and adhesion processes.47 However, for these processes to work efficiently, a strong interaction between the carbohydrates and their related receptor is required. To strengthen the interaction between cells and other biological entities nature uses multiplicity. By presentation of multiple carbohydrates, the affinity can be increased resulting in enhancement of any coupled processes.48-49 However, it is often observed

14

that this increase in affinity is not linear, but an enhanced affinity on a per mole carbohydrate basis is observed instead. This enhanced affinity is often referred to as the ´cluster glycoside effect´.50-52

Figure 11. Representation of the cell surface and the glycocalyx. Presentation of multiple carbohydrates can result in multivalent binding to a variety of receptors.

2.2 Multifunctional Nanoplatforms

To be able to better understand and control the processes in which carbohydrates are involved, it is essential to study this so-called ´cluster glycoside effect´. To study this effect, much effort has been placed in the synthesis of multifunctional nanoplatforms and characterization of their binding to known receptors. This has resulted in the appearance of a wide variety of glycosylated architectures, e.g. glycodendrimers, glycopolymers, glyconanoparticles and glycoclusters.52-57 However, most of these platforms offer very little to no control over the spatial arrangement and local density of the carbohydrates they present. To overcome this problem, and to be able to better characterize multivalent effects, it is therefore crucial to design well-defined nanoplatforms. A field providing facile access to well-defined nanoplatforms is CDC.

2.2.1 Nanoplatforms by CDC

Due to the self-sorting nature of CDC, this field offers facile access to complex three-dimensional molecular architectures.15 By using rational design and strategically chosen building blocks, a wide variety of structures could be synthesized including cages58-64, molecular knots65-

66 and other mechanically interlocked structures.44, 67-71 Metal ions also play a prominent role in the selection of some of these complex architectures. An interesting example of this is provided by Nitschke,

15

who used reversible ligand coordination to iron(II)-ions to form tetrahedral cages (Figure 12).72 Further research showed that by using a variety of metals and linkers they were even able to select different molecular architectures.73

Figure 12. Formation of tetrahedral cages from diamine linkers and iron(II) ions.

2.2.2 Selected Nanoplatforms in Current Study

From all the three-dimensional structures available in the toolbox of CDC, a selection can easily be made to fit with the intended research goal. The starting point for this study was the multivalent effects observed for the well-studied glycofullerenes. In previous research on glycofullerenes, focus was placed on varying carbohydrates and investigating the influence of the linker type and linker length.74-77 To create an even deeper understanding of the nature of these multivalent effects, this study instead focuses on the spatial presentation of the carbohydrates and characteristics of the platform. To achieve this goal, comparable core structures to that of the fullerene are required. The fullerene itself consists of sixty carbons, arranged in a globular truncated icosahedral core structure. Functionalization of this core structure allows for six substituents to be added following an octrahedral Th-symmetrical addition pattern. Every substituent on this scaffold can easily be functionalized with two ligands, providing a dodecavalent core (Figure 13).78

16

Figure 13. Symmetry and size of functionalized fullerene.

An intriguing scaffold, comparable to the fullerene core structure is that of the molecular Borromean rings. A Borromean ring consists of three rings that are interlocked with each other, however, breaking any one of the rings results in disassembly of the construct. The first synthesis of a molecular Borromean ring was reported in 1997 by Seeman using DNA,79 succeeded by a wholly synthetic procedure in 2004 by the group of Stoddart.80 In the latter example, elegant use of building-block design, reversible transimination and ligand coordination to zinc(II)-ions, led to the formation of this intriguing S6-symmetrical scaffold with 78% isolated yield. Similar to the fullerenes, this core can be hexafunctionalized following a similar octahedral addition pattern of the substituents.81-82 Addition of two carbohydrate ligands on every position yields a dodecavalent core with a similar spatial presentation of the ligands as the fullerene (Figure 14). On the other hand, the Borromeates are slightly more than twice as large as fullerenes and the platform itself is positively charged surrounded by trifluoroacetate ions, instead of neutral like the fullerenes.

Figure 14. Symmetry and size of functionalized molecular Borromeate.

The third core structure of interest for this study was chosen to differentiate itself from the others by its spatial presentation of the

17

carbohydrate ligands. The selected core matching the requirements is the dodecaimine scaffold as published by the groups of Gawronski and Cooper.83-84 This Td-symmetrical core can be derivatized by reduction followed by amidation of the formed amines.85 Now, instead of the octahedral addition pattern of the ligands an icosahedral addition pattern of the twelve carbohydrates is obtained (Figure 15).

Figure 15. Symmetry and size of functionalized dodecaamine cage.

2.3 Nanoplatform Synthesis and Functionalization

The glycosylated Borromeate was prepared according to Scheme 5. Alkyne presenting dialdehyde 1 was condensed with diammonium salt 2 in the presence of the templating ZnII-ions. The obtained hexaalkyne BR·(R1)6 was glycosylated by a CuI-catalysed azide-alkyne cycloaddition (CuAAC) using divalent azido-functionalized mannoside 3. Overall, this route yielded the desired dodecavalent mannoside BR·(R2)6 in excellent yield.

The synthetic route towards the dodecaglycosylated cage and fullerene is depicted in Scheme 6A and B. The dodecaamine cage was prepared according to literature procedure by spontaneous self-organization of an imine mixture formed from trialdehyde 4 and ethylene diamine. In situ reduction of the formed dodecaimine gave the desired dodecaamine C·(R3)12.86 Acylation of this cage with acid chloride 5 affords dodecaalkyne C·(R4)12. This alkynylated scaffold can be functionalized by CuAAC with 6 to give the desired glycosylated cage C·(R5)12. The alkynylated fullerene F·(R6)12 could be obtained through the previously reported Bingel reaction between unsubstituted fullerene and malonic ester 7.87 Glycosylation of F·(R6)12 was performed through CuAAC with azide 6 to obtain the dodecavalent mannoside F·(R7)12.

18

Scheme 5. Synthetic route to obtain dodecaglycosylated Borromeate BR·(R2)6.

19

Scheme 6. Synthetic routes to obtain (A) dodecaglycosylated cage C·(R5)12 and (B)

dodecaglycosylated fullerene F·(R7)12.

Besides only comparing these three dodecamannosides with each other, further investigation was done to support the obtained results. For this three additional mannosides were synthesized. The first additional platform, the hexamannose Borromeate BR·(R8)6, was prepared by direct CuAAC of the hexaalkyne Borromeate BR·(R1)6 with monovalent mannoside 6 (Scheme 7A). This compound has the lowest local density of carbohydrates of all multivalent nanoplatforms in this study and served as control for enhanced affinity caused by the charged scaffold of

20

the Borromean rings. The second additional platform is the dodecamannoside cage C·(R10)12, which instead has the carbohydrates linked to the core with an aliphatic linker. This cage was prepared by acylation of unfunctionalized cage C·(R3)12 with acid chloride 7, followed by CuAAC with the same monovalent mannoside 6 (Scheme 7B). The aliphatic linker was chosen for its reduced rigidity providing a higher degree of conformational freedom to the linked carbohydrates. The last additional mannoside is the monovalent reference MR prepared by CuAAC of alkyne 8 and monovalent mannoside 6 (Scheme 7C). This monovalent reference is used to investigate if multivalent presentation of the carbohydrates results in enhancement of affinity on a per mole basis for the investigated glycosides, this so-called ”cluster glycoside effect”.

Scheme 7. Synthetic route to control samples (A) hexamannose Borromeate BR·(R8)6, (B) dodecamannoside cage with aliphatic linker C·(R10)12 and (C)

monovalent reference MR.

21

2.4 Characterization of Multivalent Interactions

Different methods have been described to evaluate the binding between proteins and carbohydrates. These methods include the hapten inhibition assays, enzyme-linked lectin assays, isothermal titration microcalorimetry and even force spectroscopy.50, 88 However, these techniques only provide meagre information about the binding including IC50-values or binding constants. Technologies providing more detailed information about specific binding events are surface plasmon resonance89 and quartz crystal microbalance (QCM) technology.90-91 Both technologies provide a real-time analysis of the binding event based on either optics (for surface plasmon resonance) or acoustics (for QCM). The latter technology was used to characterize the affinity of the glycosylated nanoplatforms in this study.

2.4.1 Quartz Crystal Microbalance Technology

Measurement of protein-carbohydrate interactions by QCM technology is based on a phenomenon called the reverse piezoelectric effect. Application of an alternating current to opposite faces of a quartz crystal induces a mechanical vibration at the crystal’s fundamental frequency (f0). In liquid systems, the change of mass on the surface (Δm) is related to the observed frequency difference (Δf) based on the equation shown below.90

∆� =����

�∆�

+ ��

���

��(∆��∆��)

��

Other factors contributing to this equation are the QCM/mass

sensitivity (Cf, also called the Sauerbrey constant), the area of the electrode (A) and the change in absolute viscosity (ΔηL) and density (ΔρL). The latter two terms indicate that modifications of the buffer system also give observable changes in oscillation frequency, which is caused by the difference in pressure applied on the surface. The working principle of QCM technology is outlined in Figure 16.

22

Figure 16. Change of the oscillation frequency in response to a change in the mass

on the surface, arrows indicate shear motion of the crystal wafer.

When coupled to a frequency counter, the change in mass on the surface can be observed in real-time. To illustrate this principle, the observed response during the immobilization of the first studied lectin concanavalin A (conA) is shown in Figure 17. For these experiments, the starting gold surface of the QCM-chip has a self-assembled monolayer containing carboxyl end-groups. Two QCM-chips are used in parallel to be able to correct for non-specific binding (channel 1 and control channel 2). Functionalization of the chips is performed by activation using EDC and sulfo-NHS to provide a reactive ester surface (step I). As observable in the sensorgrams of both channel 1 and 2, the pressure applied by the reagent solution is different than that of the running buffer resulting in a large response. After injection is stopped, the signal drops back and the sensorgram shows a small shift in frequency. This indicates that only little mass has been added to the surface as expected. Injection of a solution of conA to only channel 1 (step II) provides 60 Hz functionalization on that chip, while channel 2 presents no conA on its surface. The reactive groups on both chips are eventually removed by injection of ethanolamine providing the starting QCM-chips for injection of the glycosylated nanoplatforms (step III).

23

Figure 17. (A) Schematic representation of the three steps required for

immobilization of conA on a QCM-chip. (B) The sensorgrams obtained for the conA positive channel 1 and unfunctionalized control channel 2, to illustrate the response

obtained after increasing the mass on the surface of the chip.

2.4.2 Set-up of Binding Experiments

The binding event was monitored by injection of eight different concentrations of the respective samples for 84 seconds (35 µL at a 25 µL/min rate). The concentrations of the samples ranged from 1.43 to 500 µg/mL. Every separate concentration of sample was injected in duplicate and simultaneously over both the lectin functionalized channel and the unfunctionalized control channel. After injection of the samples, the surfaces were left for at least 300 seconds to monitor the dissociation of any bound sample. The obtained sensorgrams were corrected for non-specific binding by subtraction of the obtained response in the negative channel. After analysis of each sample, the binding pockets of the proteins were regenerated by injection of a 10 mM solution of glycine at pH 1.5-2.5 (as exemplified in Figure 18). This regeneration proved most efficient for Borromeates BR·(R2)6 and BR·(R8)6 due to instability of these assemblies at this pH. For cages C·(R5)12 and C·(R10)12 almost fully regenerated after the assigned dissociation time. On the other hand, surfaces treated with fullerene F·(R7)12 only partly regenerated after treatment with glycine, probably due to hydrophobic interactions of the core with the self-assembled monolayer on the sensor surface. As the sensor surfaces remain unsaturated throughout the binding experiments, it was still possible to retrieve accurate binding data for the samples that did not allow for full regeneration.

24

Figure 18. Surface regeneration using 10 mM glycine at pH 2.5 after injection of

BR·(R2)6, C·(R5)12 and F·(R7)12 over a conA functionalized QCM-chip. Sensorgrams shown are without subtraction of the control channel.

The obtained sensorgrams were examined for spikes and other factors that could interfere with proper fitting of the kinetic model. The selected sensorgrams were subjected to kinetic fitting using a 1 : 2 interaction model to accommodate for both a monovalent and multivalent style of binding. The association rates (ka), dissociation rates (kd) and corresponding dissociation constants (KD) were extracted from the fitted models.

2.4.3 Binding Experiments with Concanavalin A

For the first part of the binding study the lectin concanavalin A (Figure 19) was immobilized on the QCM-chip as shown before in Figure 17. This 102 kDa homotetrameric lectin has a total of four mannose binding sites, which are separated from each other by approximately 65Å. As a result of this distance, it should be impossible for the nanoplatforms in this study to bridge between two binding pockets. The degree of surface functionalization of the QCM-chip should not allow for binding of the nanoplatforms in between two separate conA molecules. The extracted kinetic and thermodynamic data for binding of the tested mannosides to this lectin can be found in Table 1.

Figure 19. Structural data for concanavalin A.

25

Table 1. Association rates, dissociation rates and dissociation constants obtained for binding of the different samples to a conA functionalized QCM-chip.

Glycosylated scaffold

ka1

(s-1)

kd1

(s-1)

KD1

(M)

ka2

(s-1)

kd2

(s-1)

KD2

(M)

BR·(R2)6 4.72∙104

±1.34 ∙103

5.59∙10-2

±1.19∙10-5

1.19∙10-6

±3.38∙10-8

1.31∙103

±3.92∙102

1.93∙10-4

±3.98∙10-5

1.47∙10-7

±8.17∙10-8

C·(R5)12 9.52∙103

±6.82 ∙102

1.01∙10-1

±6.34∙10-5

1.06∙10-5

±8.22∙10-8

1.67∙103

±3.19 ∙103

2.02∙10-3

±5.23∙10-5

1.21∙10-6

±8.86∙10-7

F·(R7)12 2.73 ∙104

±1.35 ∙103

8.70∙10-2

±7.75∙10-6

3.18∙10-6

±1.57∙10-7

6.50∙102

±3.26∙102

6.63∙10-4

±8.28∙10-6

1.02∙10-6

±7.00∙10-7

BR·(R8)6 6.87 ∙103

±7.35 ∙102

1.04∙10-3

±3.18∙10-5

1.52∙10-7

±2.11∙10-8

1.15∙104

±4.35∙101

5.49∙10-2

±2.50∙10-4

4.76∙10-6

±3.96∙10-8

C·(R10)12 6.86 ∙103

±4.12 ∙101

7.83∙10-2

±3.54∙10-4

1.14∙10-5

±1.20∙10-7

6.93∙102

±8.47∙101

2.55∙10-3

±1.12∙10-4

3.68∙10-6

±6.21∙10-7

MR* - - - - - 1.09∙10-4

±0.19∙10-4

* Determined from saturation binding due to low response observed in the sensorgrams.

The initial study on conA functionalized surfaces included dodecavalent mannosides BR·(R2)6, C·(R5)12 and F·(R7)12. The sensorgrams obtained for these nanoplatforms are shown in Figure 20. From these sensorgrams it can be clearly seen that binding of the nanoplatforms does not follow a simple one to one binding profile. The extracted kinetic data also shows that there are two distinctive events, a fast association and dissociation event giving a weaker affinity and a slower association and dissociation event indicative for a more stable multivalent binding. When directly comparing the sensorgrams, it is noticeable that the dissociation rate of the icosahedrally glycosylated cage C·(R5)12 is significantly higher than that of the two octahedrally substituted scaffolds. Overall, the estimated dissociation constants for these three nanoplatforms are 147 nM, 1.21 µM and 1.02 µM for BR·(R2)6, C·(R5)12 and F·(R7)12, respectively. Interestingly, even though the sensorgram clearly indicates slower dissociation for the dodecaglycosylated fullerene, the combination with a slower association rate gives this nanoplatform a similar affinity to that observed for the dodecaglycosylated cage.

26

Figure 20. Selected curves (dotted lines) and fitted data (gray lines) for

dodecavalent mannosides BR·(R2)6, C·(R5)12 and F·(R7)12. Different dotted lines indicate injections at different concentrations or duplicates.

Assessment of the extent of the multivalent effect was done by performing the same experiment with monovalent reference MR (Figure 21). With a molecular weight of merely 524.6 g/mol the limitations of QCM technology were tested as the expected frequency shift would be much lower. As anticipated, only a low change in frequency was observed during the binding event and, as a result, a kinetic model could not be accurately fitted to this data. However, as clear saturation was observed, saturation binding could be determined for monovalent reference MR providing a KD-value of 109 µM. From these results it can be deduced that there is approximately an 66-, 8- and 10-fold increase in affinity per carbohydrate for BR·(R2)6, C·(R5)12 and F·(R7)12, respectively.

Figure 21. Sensorgram obtained for binding of the monovalent reference MR. The dissociation constant was determined by saturation kinetics. Different lines in the

sensorgram indicate injections at different concentrations or duplicates.

As regular chelation with conA is inaccessible for BR·(R2)6, C·(R5)12 and F·(R7)12, different causes for the observed cooperativity must be operating. For icosahedrally substituted C·(R5)12, the lower local concentration of carbohydrates might result in weaker secondary interactions (i.e. interactions with sites other than the mannose binding pocket), resulting in the observed saturation. Both octahedrally substituted BR·(R2)6 and F·(R7)12 have a less distributed presentation of the carbohydrates on their periphery, but they still show a clear difference in binding affinity to conA. This difference could be caused by the

27

difference in size and charge of the nanoplatforms. Additionally, the cyclopropane motif on the fullerene core forces a higher spatial distribution of the carbohydrates on F·(R7)12.

Overall, these results indicate that secondary chelation in combination with statistical and polyelectrolyte effects are the major contributors to the observed increase in affinity. To further support these observations, two modified nanoplatforms were tested in the same binding experiments. The first of the modified nanoplatforms is the hexavalent Borromeate BR·(R8)6, having the same octahedral presentation of the carbohydrates but only a single carbohydrate per position. The second modified nanoplatform is the dodecavalent cage C·(R10)12. In this platform the linker directly connected to the scaffold was changed from an aromatic to an aliphatic linker. This change is partially expected to change the entropic contribution to the affinity, due to the increase in conformational freedom of the linked carbohydrates. The observed sensorgrams and the fitted curves for binding of BR·(R8)6

and C·(R10)12 to conA are provided in Figure 22. The extracted kinetic and thermodynamic data from the fitted curves can be found in Table 1.

Figure 22. Selected curves (dotted lines) and fitted data (gray lines) for mannose functionalized hexavalent Borromeate BR·(R8)6 and aliphatic linker cage C·(R10)12. Different dotted lines indicate injections at different concentrations or duplicates.

When comparing the sensorgrams obtained for BR·(R8)6 and C·(R10)12 to those obtained for BR·(R2)6 and C·(R5)12, it is immediately apparent from the frequency shift that less material is binding to the surface at similar concentrations. The increase in conformational freedom of C·(R10)12 shows a reduction in association rate corresponding to the multivalent binding in combination with an increased dissociation rate. As a result, the overall dissociation constant increases from 1.02 µM for C·(R5)12 to 3.68 µM for C·(R10)12. For the Borromeates the situation is a little more complicated as similar overall dissociation constants of 147 and 152 nM are observed for BR·(R2)6 and BR·(R8)6, respectively.

28

However, from the obtained sensorgrams it can be seen that saturation binding of the surface is achieved with BR·(R8)6, which is a result of the higher dissociation rate for the multivalent binding event. The similar dissociation constant is originating from the increase in association rate of the hexavalent Borromeate, which might be a result of the shorter linker between the charged platform and the carbohydrate. In total, these results are in agreement with the importance of statistical and polyelectrolyte effects as contributing factors to the observed increase in affinity of multivalent glyconanoplatforms.

2.4.4 Binding Experiments with Galanthus nivalis Agglutinin

To further investigate the factors contributing to multivalent binding effects, another lectin was investigated. The lectin chosen for this part of the study is Galanthus nivalis agglutinin (GNA, Figure 23). This 52 kDa homotetrameric lectin was immobilized on the QCM-chip in a similar fashion as performed for conA. GNA has a total of twelve mannose binding sites, which are separated from each other by approximately 19Å. This shorter distance between the binding sites should allow bridging of two binding pockets by the investigated glyconanoplatforms.

Figure 23. Structural data for Galanthus nivalis agglutinin.

In comparison to conA, GNA shows much weaker affinity to monovalent mannosides.92 This weaker affinity could be counteracted by the greater number of mannose binding sites, potentially resulting in a sharp increase of the relative potency of multivalent ligands. The extracted kinetic and thermodynamic data for binding of the tested mannosides to GNA can be found in Table 2. As expected, the weaker binding event to the GNA functionalized surface shows high dissociation rates, as a result of the weak monovalent binding.

29

Table 2. Association rates, dissociation rates and dissociation constants obtained for binding of the different samples to a GNA functionalized QCM-chip.

Glycosylated scaffold

ka1

(s-1)

kd1

(s-1)

KD1

(M)

ka2

(s-1)

kd2

(s-1)

KD2

(M)

BR·(R2)6 1.06∙104

±1.52∙103 8.14∙10-4

±3.20∙10-6 7.66∙10-8

±1.15∙10-8

2.69∙104 ±2.16∙103

1.13∙10-1 ±7.74∙10-6

4.18∙10-6 ±3.39∙10-7

C·(R5)12 5.96∙102

±2.40∙102 6.68∙10-4

±2.53∙10-6 1.12∙10-6

±5.44∙10-7

1.93∙103 ±4.11∙102

1.20∙10-1 ±1.86∙10-6

6.12∙10-5 ±1.39∙10-5

F·(R7)12 4.70∙102

±7.24∙101 7.00∙10-4

±1.81∙10-6 1.49∙10-6

±2.39∙10-7

2.14∙104 ±6.37∙103

1.52∙10-1 ±3.71∙10-5

7.08∙10-5 ±2.31∙10-6

BR·(R8)6 4.02∙103

±9.40∙102 8.30∙10-4

±1.91∙10-6 2.06∙10-7

±5.16∙10-8

1.38∙104 ±5.19∙103

8.62∙10-2 ±1.81∙10-5

6.23∙10-6 ±2.72∙10-6

C·(R10)12* - - - - - -

MR* - - - - - -

* Kinetic data could not be determined using QCM technology due to too weak binding to the sensor surface.

The obtained sensorgrams for the first three dodecavalent mannosides BR·(R2)6, C·(R5)12 and F·(R7)12 are given in Figure 24. Interestingly, a similar dissociation rate is observed for all three nanoplatforms for the multivalent binding event, both visible in the extracted kinetic data and the sensorgrams. The main contribution to the multivalent effect in this case seems to originate from the ability to bridge different binding sites. As the distance between the separate mannose binding sites on GNA is only 19Å, it is expected that spatial presentation of the carbohydrates BR·(R2)6, C·(R5)12 and F·(R7)12 is of less importance as all of the studied nanoplatforms present carbohydrates with this respective distance to each other on different faces of the platforms. The dissociation constants observed are 77 nM for BR·(R2)6 versus 1.12 µM and 1.49 µM for C·(R5)12 and F·(R7)12, respectively. The significantly lower dissociation constant detected for BR·(R2)6 again indicates the importance of polyelectrolyte effects in multivalent binding, providing this positively charged platform with a faster association rate to the negatively charged GNA functionalized surface (GNA, pI < 4.6).

30

Figure 24. Selected curves (dotted lines) and fitted data (gray lines) for

dodecavalent mannosides BR·(R2)6, C·(R5)12 and F·(R7)12. Different dotted lines indicate injections at different concentrations or duplicates.

Similar to the conA based experiments, binding of the hexavalent Borromeate BR·(R8)6, dodecavalent cage with aliphatic linkers C·(R10)12

and monovalent reference MR was tested against the GNA functionalized QCM-chip as well (Figure 25). Due to the weak binding of monovalent substrates to GNA, the monovalent mannoside MR shows no measurable affinity to the surface. Interestingly, also the cage C·(R10)12 shows no quantifiable affinity to the surfaces, confirming that an increase in rotational freedom again leads to a decrease in affinity. For the hexavalent Borromeate BR·(R8)6 an increase in dissociation constant to 206 nM was observed, as compared to the 77 nM for the dodecavalent Borromeate BR·(R2)6. Interestingly, the dissociation rate observed for the multivalent binding event is similar for both platforms and the difference in dissociation constant is mainly due to the decrease in the association rate for BR·(R8)6. This difference in association rate is most likely due to the lower availability of correctly oriented carbohydrates for the multivalent binding.

Figure 25. Selected curves (dotted lines) and fitted data (gray lines) for hexavalent Borromeate BR·(R8)6, no binding was observed for cage C·(R10)12 and monovalent

reference MR. Different dotted lines indicate injections at different concentrations or duplicates.

31

2.4.5 Binding Experiments with Bacteria

An attempt was made to evaluate the binding of the glyconanoplatforms to bacteria immobilized on the QCM-chip. For this two Escherichia coli strains ORN178 (type 1 fimbriae positive, binds mannosides) and ORN208 (type 1 fimbriae negative, does not bind mannosides) were immobilized on the two different channels. Unfortunately, the immobilized quantity of bacteria was either too low, or the binding of the studied nanoplatforms to the bacteria is too weak to produce a quantifiable response. Interestingly, the Borromeate BR·(R2)6

shows high non-specific affinity to both surfaces due to charge based interactions of the positively charged platform with the negatively charged bacteria (Figure 26).

Figure 26. Observed frequency shift upon injection of BR·(R2)6, C·(R5)12 and F·(R7)12 on channel 1 containing surface immobilized ORN178 (gray line) and

channel 2 containing surface immobilized ORN208 (dotted line).

2.5 Conclusions

This chapter has shown the synthesis of two novel highly ordered carbohydrate nanoplatforms using the toolbox provided by CDC, the glycoborromeates (BR·(R2)6 and BR·(R8)6) and glycocages (C·(R5)12

and C·(R10)12). The affinities of these nanoplatforms towards conA, GNA and bacteria functionalized surfaces were assessed using QCM technology and compared with a glycofullerene (F·(R7)12) and monovalent reference (MR). Overall, the binding studies demonstrate different contributors to the observed multivalent effects. In absence of the possibility to bridge different binding sites for conA, more optimal local densitities of the carbohydrates in the octahedrally substituted platforms can stabilize potential secondary interactions. In presence of the ability to bridge different binding sites for GNA, the spatial distribution of carbohydrates has limited influence and similar

32

dissociation rates were observed for all platforms. In conclusion, for both lectin functionalized surfaces statistical (entropic) and polyelectrolyte effects provide a large contribution to the multivalent binding effects observed between proteins and carbohydrates in this study. It is also shown that subtle changes can have large effects on the observed binding profile.

33

3. Catalysis of Olefin Exchange Under

Aqueous Conditions

(Paper III-IV) The previous chapter discussed the use of metals in the selection of

the Borromeates that were used to investigate the contributing effects in carbohydrate-lectin interactions. In the following chapter, the metal will function as the reactive centre in the catalyst for olefin metathesis. Herein, aqueous conditions are developed to avoid the problem of undesired isomerization reactions. Under these conditions it was possible to perform olefin metathesis on unprotected carbohydrates among others.

3.1 Olefin Metathesis

Olefin metathesis is one of the most useful organic transformations to directly break and form carbon-carbon double bonds. For example, this transformation has been widely employed in the synthesis of pharmaceuticals and advanced polymers.93 Ever since its discovery in the 1950s, much effort has been placed in understanding the reaction mechanism and development of more efficient and stable catalysts. This work eventually led to the Nobel Prize in 2005, awarded to Chauvin, Schrock and Grubbs. Even though olefin metathesis has been widely investigated, the limits and boundaries of this transformation are still constantly being rediscovered.

3.1.1 Catalysts for Olefin Metathesis

A wide variety of specialized metal catalysts has been developed for olefin metathesis including Z-selective,94-96 ethenolysis,97-98 enantioselective,99 recyclable,100-101 and many other potent catalysts for reactions in organic solvents.102-103 Most of these catalysts are derived from either the Grubbs104 or Schrock105 alkylidenes as shown in Figure 27. The Schrock alkylidenes are among the most reactive available for the olefin metathesis reaction, however their high reactivity also renders this type of catalysts unstable in the presence of air and moisture. Albeit less reactive, the Grubbs alkylidenes offer a more stable alternative.

34

Figure 27. Structures of early Grubbs and Schrock alkylidene complexes.

Further development on olefin metathesis catalysts has resulted in the commercial availability of several variants of the Schrock and Grubbs alkylidenes,43, 106-114 with more being added to retailers catalogues continuously.

3.1.2 Olefin Metathesis in CDC

Several examples of olefin metathesis have appeared in CDC. Because the reaction itself only scrambles the two substituents on two different alkenes, no by-products are formed. This attribute makes dynamic olefin metathesis an excellent reaction for application in self-healing polymers as shown by Guan.115-116 By incorporation of a low quantity of Grubbs 2nd generation catalyst into a polybutadiene cross-linked polymer network, effective self-healing of the material could be achieved (Figure 28). This type of error-correcting behaviour by olefin metathesis has also been applied in the synthesis of macrocycles and complex molecular architectures.65, 117-119

Figure 28. Self-healing of polybutadiene cross-linked polymer by incorporated

Grubbs 2nd generation catalyst.

35

Due to the stability of the carbon-carbon double bond, dynamic olefin metathesis also has a potential future in chemical biology.120 Early experiments by Ward show that self-metathesis of methyl oleate forms a new mixture of alkenes at equilibrium.121 This scrambling of alkenes has later shown to be effective in discovery of inhibitors for carbonic anhydrase II by Poulsen.122 The thermodynamic nature of the equilibrating alkene mixture can also be used for tandem catalysis as shown by Hartwig.123 In their work, dynamic exchange of the alkenes leads to continuous reformation of the substrate for a metalloenzyme catalysed epoxidation (Figure 29). Overall, this approach has resulted in a 1.5 fold increase in yield compared to the stepwise process.

Figure 29. Dynamic exchange of olefins continuously reforming the substrate

selected by a kinetic transformation by cytochrome p450 BM3.

Even though several applications of dynamic olefin metathesis have been published, the reaction still has clear practical limitations. Among these limitations is the low functional group tolerance during dynamic exchange.124-125 Furthermore, the potential of dynamic olefin exchange in chemical biology would increase if the reaction is applicable with substrates that are insoluble in apolar solvents, like a great part of biologically relevant substrates is. Therefore, better procedures for running the exchange directly under aqueous conditions are required.

3.1.3 Aqueous Olefin Metathesis

Several approaches have been reported for running olefin metathesis under aqueous conditions.126-129 Included in these approaches are acoustic emulsification,130 micellar catalysis131-135 and supramolecular additives.136-137 Although these examples have offered great potential for moving reactions from hazardous organic solvents to aqueous solutions, their applicability to water-soluble substrates is limited. Another interesting, and perhaps the most biocompatible approach reported up to

36

date, is the formation of artificial metallo-enzymes containing Grubbs’ catalyst. This approach has been demonstrated by covalently attaching the catalyst to the protein.138-139 Alternatively, this is also possible by docking the catalyst into (strept)avidin using biotin interactions (Figure 30).140-141 These metallo-enzymes have shown to be active in ring-closing metathesis and cross-metathesis, but as the catalyst is generally surrounded by the protein, accessibility of certain substrates to the active site is limited.

Figure 30. Artificial metallo-enzyme using the biotin-streptavidin interaction as

reported by Ward.

A more direct approach to aqueous olefin metathesis is by designing molecular catalysts that can be directly used in aqueous mixtures. By addition of modifications to the catalysts core structure, its solubility in aqueous mixtures can be increased. These modifications included the introduction of quaternary ammonium groups,142-147 protonated amines143, 148 and ethylene glycol chains149-151 (examples provided in Figure 31). This approach to aqueous olefin metathesis would be most suitable for application in CDC as there is a no risk of interaction of the substrates in the mixture with any additives that were required for the previously mentioned approaches. Although some of the existing water-soluble catalysts show excellent activity in olefin metathesis, its application in CDC is still limited as the reported substrate scope consists almost exclusively of ammonium tagged simple alkenes.

37

Figure 31. Examples of reported catalysts for direct application in aqueous

mixtures.

3.2 Olefin Metathesis of Unprotected Carbohydrates

As discussed in chapter 2, carbohydrates play an important role in a wide variety of biological processes. Therefore, this class of molecules would provide a valuable expansion of the substrate scope of olefin metathesis in aqueous media. Pioneering studies by Davis152-154 and Harding124 show direct application of unprotected carbohydrates as substrates in cross-metathesis. As their studies used Hoveyda-Grubbs 2nd generation catalyst, mixed solvent systems were required. Besides requiring mixed solvent systems, the reactions were also limited to high catalyst loadings and privileged alkenes such as allylic sulphides or selenides. In an earlier report by Blechert, the first example of self-metathesis of a C-glycoside with lower catalyst loading was shown in absence of any privileged linkers. Although their solid-supported catalyst could initiate self-metathesis in methanol, a high degree of isomerization was observed and no reaction occurred in water.155

3.2.1 Catalyst Design and Synthesis

To further investigate the possibility of olefin metathesis of unprotected carbohydrates in aqueous media, a simple catalyst construct was desired. The catalyst design contained two quaternary ammonium groups on the N-heterocyclic carbene to ensure solubility throughout the whole reaction, while avoiding the need of highly acidic conditions. Symmetrical placement of the amines in combination with late-stage methylation would allow for facile synthesis and analysis. Synthesis of a carbene precursor matching these design features had already been reported by Plenio.156 Adapting their route, the carbene precursor imidazolinium chloride 12 could be synthesized in four steps from 2,6-

38

dimethylaniline. Deprotonation of 12 afforded the carbene which was coupled to Hoveyda-Grubbs 1st generation catalyst. The non-methylated catalyst Cnm was treated with methyl triflate to afford the desired catalyst C, in a total of six steps from commercially available starting materials (Scheme 8).

Scheme 8. Synthesis of catalyst C in six steps starting from commercially available

2,6-dimethylaniline.

3.2.2 Optimization of Metathesis Conditions

To test the self-metathesis of carbohydrates, first a 0.1 M solution of allyl mannoside (AllMan) in D2O at 40°C was reacted in presence of 2.5 mol% of catalyst C. Unfortunately, only isomerization could be observed under these reaction conditions. As previous investigations suggest non-productive chelation as limiting factor in olefin metathesis of functionalized molecules,148, 157 the linker between the carbohydrate and the alkene was increased. Similarly, 3-butenyl mannoside (ButMan)

39

only showed isomerization as well. However, when 4-pentenyl mannoside (PenMan) was subjected to the same conditions, 5% self-metathesis could be observed (Table 3, entry 1). Increasing the reaction temperature did not result in a higher degree of self-metathesis, but an increase in isomerization was observed instead (Table 3, entry 2-3). Self-metathesis of PenMan could be improved to 13% in the presence of 5 mol% C, however an increase of the undesired double-bond isomerization was also observed (Table 3, entry 4).

Due to the similar undesired isomerization observed for C in comparison to previously reported water-soluble catalysts, it was necessary to find conditions that prevent isomerization. Fogg has shown that catalyst decomposition in the presence of Brønsted bases potentially causes formation of ruthenium hydride species.158 Based on this and previous observations by Grubbs,149, 159 self-metathesis of PenMan was tried under mildly acidic conditions by addition of a small portion of acetic acid. Interestingly, addition of 2.5 v/v% effectively decreased the degree of isomerization while increasing the self-metathesis yield to 26% (Table 3, entry 5). This indicates that under acidic conditions the rate of catalyst decomposition decreases. Increasing the quantity of acetic acid decreased the amount of self-metathesis (Table 3, entry 6), with a total loss of reactivity in pure acetic acid.

As under prolonged reaction times isomerization of PenMan was still observed, further modifications of the reaction conditions were investigated. When decreasing the reaction temperature to 30°C no isomerization was observed, however a drop in self-metathesis rate was observed resulting in only 18% self-metathesis after prolonged reaction time (Table 3, entry 7). The optimal temperature for this reaction was shown to be 35°C, where self-metathesis achieves similar yields while only traces of isomerization were observed (Table 3, entry 8).

40

Table 3. Optimization of self-metathesis reaction of PenMan using catalyst C.a

Entry C

(mol%) T

(°C) Solvent

Self-metathesis (%)

Isomerization (%)

1 2.5 40 D2O 5 Traces

2 2.5 50 D2O 6 44

3 2.5 60 D2O 5 87

4 5 40 D2O 13 28

5 5 40 D2O + 2.5% CD3COOD 26/26b 0/6b

6c 5 40 H2O + 5% CH3COOH 19/22b 0/13b

7 5 30 D2O + 2.5% CD3COOD 16/18b 0/0b

8 5 35 D2O + 2.5% CD3COOD 18/27b 0/tracesb

a The reactions were performed at an initial substrate concentration of 0.1 M in the described solvent. Conversions were determined by 1H-NMR spectroscopy after 45-130 minutes reaction time, and are given as percentage for the combined cis- and trans-isomers.b Conversion determined at t = 18 h. c Test run in non-deuterated solvents to determine if deuterium atoms were incorporated in the alkene of the isomerized products.

3.2.3 Non-productive Chelation by Carbohydrates

To further probe the diminishing effect of non-productive chelation on the self-metathesis of carbohydrates, the optimized reaction conditions were applied to a series of alkene-functionalized α- (mannose/Man) and β-linked (galactose/Gal) carbohydrates containing different linker lengths (allyl/All, 3-butenyl/But, 4-pentenyl/Pen and 5-hexenyl/Hex). As the optimal reaction conditions used 2.5 v/v% acetic acid, the necessity of the hazardous methylation step was questioned and the substrates were also tested by direct catalysis with non-methylated catalyst Cnm.

Initially, the self-metathesis of allyl alcohol was studied as non-productive chelation has less influence on its reactivity due to formation of a disfavoured four-membered chelated state. This substrate showed fast initiation and up to 93% conversion was observed after 300 minutes of reaction time, with similar results obtained for both catalysts C and Cnm (Table 4, entry 1). As expected from previous results, both AllMan

41

and AllGal did not show any self-metathesis products (Table 4, entry 2 and 3), but up to 13% conversion to the dimers could now be observed for ButMan and ButGal (Table 4, entry 4 and 5). Increasing the linker length to that in PenMan/PenGal further improved the yield (Table 4, entry 6 and 7), with the trend continuing to HexMan/HexGal providing up to 50% self-metathesis yields after prolonged reaction time (Table 4, entry 8 and 9).

Table 4. Self-metathesis reactions catalysed by C or (Cnm).a

a Self-metathesis reactions were performed using 5 mol % of C or (Cnm) at 35°C for the stated duration. The initial substrate concentration was 0.1 M in D2O containing 2.5 v/v% CD3COOD. Conversions were determined by 1H-NMR spectroscopy and are given as percentage for the combined cis- and trans-isomers.b Conversion determined at t = 24 h.

From this data it was apparent that both the mannosides and galactosides showed similar reactivity, with only slightly reduced yields obtained for the galactosides. Surprisingly, non-methylated catalyst Cnm performed slightly better in the self-metathesis of carbohydrates.

Entry Substrate Conversion (%)

t = 15 min t = 60 min t = 300 min

1 55 (58) 78 (74) 93 (82)

2

0 (0) 0 (0) 0 (0)

3

0 (0) 0 (0) 0 (0)

4

2 (4) 7 (11) 11 (13)

5

1 (2) 2 (4) 6 (5)

6

6 (13) 14 (18) 16 (24)

7

6 (13) 14 (21) 16 (22)

8

24 (28) 38 (33) 41 (41) 45 (50)b

9

18 (25) 25 (31) 29 (33) 34 (47)b

42

This observation could be explained by the slightly increased solubility of the protonated catalyst in comparison to C, which contains the partly hydrophobic triflate anions introduced by the methylation reaction. Overall, these results support the major role non-productive chelation plays in metathesis reactions performed on highly functionalized substrates, as illustrated in Figure 32.

Figure 32. Illustration of potential non-productive chelation of AllMan to Cnm.

According to observations by Davis154 and Hirota160, addition of metal chloride salts would prevent deactivation of the catalyst by reducing chelation. Even though addition of 100 mM of magnesium chloride led to increased self-metathesis rates for Cnm, overall lower conversions were observed to the self-metathesis products. Interestingly, although conversions were lower, traces of product could now be observed for the allyl substituted glycosides AllMan and AllGal.

3.2.4 Application Range of Carbohydrates in Olefin Metathesis

To further investigate the potential of olefin metathesis of unprotected carbohydrates in biological systems, several additional experiments were performed with the best reacting substrates HexMan and HexGal in presence of catalyst Cnm (Table 5). Further increasing the catalyst loading to 10 mol% increased the self-metathesis yields to up to 81%. Second, as application in drug discovery would require the presence of an external template that could act as an external chelator as well, self-metathesis was attempted in presence of bovine serum albumin (BSA). When 0.01 mM BSA was added to the mixture, the self-metathesis remained unretarded and equal conversions were observed for both glycosides. Only at 0.06 mM loading of BSA (100 wt% compared to the catalyst Cnm), a slight decrease of conversion was observed to 36% and 28% for HexMan and HexGal, respectively.

43

Table 5. Self-metathesis of HexMan and HexGal at increased catalyst loading or in the presence of BSA.a

Substrate Conditions Conversion (%)

t = 15 min

t = 60 min

t = 300 min

t = 24 h

5 mol% Cnm 28 33 41 50

10 mol% Cnm 48 65 76 81

0.01 mM BSA 5 mol% Cnm

29 36 41 50

0.06 mM BSAb

5 mol% Cnm 23 35 35 36

5 mol% Cnm 25 31 33 47

10 mol% Cnm 38 43 49 63

0.01 mM BSA 5 mol% Cnm

27 32 31 41%

0.06 mM BSAb

5 mol% Cnm 17 26 26 28

a Self-metathesis reactions were performed under the given conditions at 35°C for the stated duration. The initial substrate concentration was 0.1M in D2O containing 2.5 v/v% CD3COOD. Conversions were determined by 1H-NMR spectroscopy and are given as single percentage for the combined cis- and trans-isomers. b Equals 100 wt% of BSA compared to Cnm.

Having established conditions for the self-metathesis of carbohydrates, next the more synthetically relevant cross-metathesis was investigated. To achieve cross-metathesis similar conditions were employed whilst adding an equivalent concentration of allyl alcohol to the mixture as cross-coupling partner. Appearance of a new set of peaks in the 1H-NMR spectrum (Figure 33) and mass spectrometry confirmed the formation of the cross-metathesis product. After only 15 minutes of reaction time 35 and 36% of cross-metathesis was observed for HexMan and HexGal, respectively. Surprisingly, after prolonged reaction times significant amounts of self-metathesis of the glycosides could be observed as well, while the cross-metathesis yields only increased by a few percent to 40 and 39%. This result could point towards the thermodynamic nature of the olefin exchange reaction.

44

Figure 33. Alkene region of 1H-NMR spectrum showing (A) the self-metathesis of

allyl alcohol after 300 minutes, (B) the self-metathesis of HexMan after 300 minutes and (C) cross-metathesis between equimolar amounts of allyl alcohol and HexMan

after 300 minutes and 24 hours of reaction time.

3.3 Olefin Metathesis of Other Chelating Substrates

To further study the limitations of olefin metathesis due to non-productive chelation by functional groups, a new series of substrates was investigated. This series of substrates included allyl to 5-hexenyl (All to Hex) substituted pentaethylene glycol (PEG) and the 5-hexenyl ester of diglycine (HexGlyGly).

3.3.1 Activity in Self-Metathesis

Initially, the self-metathesis of the new substrates was tested under similar conditions as for the carbohydrates. The two shorter alkene linker substituted PEGs, AllPEG and ButPEG, showed a similar low reactivity in self-metathesis as observed for the glycosides. For these two substrates only around 5% conversion to the self-metathesis products was observed (Table 6, entry 1 and 2). Increasing the distance between the Brønsted basic groups and the reacting alkene increased the reactivity with yields of 20% and 28% for PenPEG and HexPEG, respectively (Table 6, entry 3 and 4). These results support the hypothesis of non-productive

45

chelation being determinant for the reaction outcome, with slightly lower yields observed in comparison to the glycosides. This slight decrease in reactivity could be explained by the increased availability of the more basic ether groups in the PEG functionalized alkenes. Self-metathesis of HexGlyGly containing an amide bond and a terminal ammonium group only provided a low yield of 9% (Table 6, entry 5). This last example would indicate deactivation of the catalyst by introduction of the primary ammonium end-group, or stronger chelation by the functionality present in this molecule.

Table 6. Self-metathesis reactions of alkene substituted PEGs and HexGlyGly catalysed by Cnm.a

a Self-metathesis reactions were performed using 5 mol % of Cnm at 35°C for the stated duration. The initial substrate concentration was 0.1 M in D2O containing 2.5 v/v% CD3COOD. Conversions were determined by 1H-NMR spectroscopy and are given as percentage for the combined cis- and trans-isomers.

3.3.2 Cross-Metathesis with Allyl Alcohol

After establishing the reactivity of these substrates in self-metathesis, they were subjected to cross-metathesis with an equimolar quantity of allyl alcohol. Cross-metathesis of AllPEG with allyl alcohol could be observed, however the conversion could not be determined due to overlapping signals in the 1H-NMR spectrum (Table 7, entry 1). Interestingly, even though only 5% self-metathesis was observed for ButPEG in the absence of allyl alcohol, cross-metathesis with allyl alcohol gave a yield of 21% (Table 7, entry 2). Increasing the linker

Entry Substrate Conversion (%)

t = 15 min t = 60 min t = 300 min

1

1 5 6

2

3 5 5

3

15 19 20

4

22 26 28

5

4 9 9

46

length, as in PenPEG and HexPEG, increased the cross-metathesis yields up to 43% (Table 7, entry 3 and 4). For HexGlyGly cross-metathesis also proved to be higher than two-fold more efficient than self-metathesis, with a conversion of 28% (Table 7, entry 5). As in all these cross-metathesis reactions less than 5% of self-metathesis was observed, this indicates that chelating alkenes behave as type II olefins under these conditions even though they are terminal olefins.161

Table 7. Cross-metathesis reactions of alkene substituted PEGs and HexGlyGly with equimolar amounts of allyl alcohol, catalysed by Cnm.a

a Cross-metathesis reactions were performed using 5 mol % of Cnm at 35°C for the stated duration. The initial substrate concentration was 0.1M in D2O containing 2.5 v/v% CD3COOD and an equimolar amount of allyl alcohol. Conversions were determined by 1H-NMR spectroscopy and are given as percentage for the combined cis- and trans-isomers. b Conversion to cross-metathesis product could not be determined due to overlapping signals in the 1H-NMR spectrum.

To further increase selectivity in the cross-metathesis reaction, the two 5-hexenyl substituted substrates HexPEG and HexGlyGly were reacted with allyl alcohol at lower temperature. Lowering the temperature was expected to decrease the rate of the slow self-metathesis even further. Indeed, a small decrease in rate was observed under these conditions. However, the cross-metathesis yields did not follow the same trend, and only a slightly higher yield was observed (Table 8, entry 1 and 2). To further increase the cross-metathesis yield the ratio between the alkenes was adjusted. Maintaining the same catalyst to alkene ratio whilst decreasing the quantity of HexPEG and HexGlyGly provided higher cross-metathesis yields of 62% and 41%, respectively (Table 8, entry 3 and 4).

Entry Substrates Conversion (%)

t = 15 min t = 60 min t = 300 min

1 AllPEG NDb NDb NDb

2 ButPEG 13 20 21

3 PenPEG 31 42 43

4 HexPEG 40 43 43

5 HexGlyGly 19 27 28

47

Table 8. Cross-metathesis reactions of HexPEG and HexGlyGly with allyl alcohol at lower temperature and with varied substrate loadings, catalysed by Cnm.a

a Cross-metathesis reactions were performed using 2.5 mol % of Cnm (based on total alkene content) at the given temperature for the stated duration. The total initial alkene concentration for both substrates was 0.2M in D2O containing 2.5 v/v% CD3COOD. Conversions were determined by 1H-NMR spectroscopy and are given as percentage for the combined cis- and trans-isomers.

3.4 Conclusions

This chapter has shown that self-metathesis and cross-metathesis of highly functionalized molecules is feasible under catalytic conditions. Addition of a small quantity of acetic acid deters isomerization and allows for direct use of non-methylated catalyst Cnm. The latter aspect effectively removes the hazardous methylation step in the catalyst synthesis, while maintaining an environmentally benign solvent mixture. Further investigation confirmed that non-productive chelation reduces the degree of self-metathesis, which allows for higher selectivity in cross-metathesis reactions. Even though the stability of the current catalyst is insufficient for application in CDC, its activity is retained in the presence of an equivalent weight of protein further increasing its potential for biological applications. Overall, this work greatly expands the substrate scope for aqueous olefin metathesis.

Entry Temperature Substrates Conversion (%)

t = 15 min t = 60 min t = 300 min

1 rt HexPEG (0.1 M) + allyl alcohol (0.1 M)

41% 49% 50%

2 rt HexGlyGly (0.1 M) + allyl alcohol (0.1 M)

18% 27% 33%

3 35°C HexPEG (0.05 M) + allyl alcohol (0.15 M)

39% 58% 62%

4 35°C HexGlyGly (0.05 M) + allyl alcohol (0.15 M)

31% 38% 41%

48

49

4. Metal Catalysts as Selectors in Dynamic

Systems

(Paper V) The previous two chapters have touched upon the use of metals in

selection of nanoplatforms and in catalysis of olefin exchange. The last chapter will combine the two areas of selection and catalysis. Herein, the selection of a reactive substrate/metal catalyst combination, from an exchanging mixture of substrates, will be described. This reactive substrate is generated in situ by co-catalysed dynamic exchange of directing groups (DGs).

4.1 Catalyst Discovery

The discovery of new catalysts has been of great importance for improvement of existing chemical processes, but also for the discovery of new ways to functionalize molecules by rendering previously inert bonds reactive. New catalysts are continuously being developed and tested for applicability in given reactions. The results from these experiments are rationalized, the catalyst is redesigned and another trial is conducted. This process is continuously repeated until the catalyst obtained meets desired criteria (Figure 34). This type of trial-and-error discovery of new catalysts has shown to be functional, but often costly and time-consuming.

Figure 34. Traditional approach for catalyst synthesis, following iterative

redesigning of the catalyst based on rationalization of observed conversions.

50

To speed up catalyst discovery, much effort has been put into development of new methodologies, such as combinatorial catalysis.162 In combinatorial catalysis, multiple catalysts are synthesized in one-pot and the catalyst mixture is tested for its activity in a set transformation. An active mixture is further evaluated by a process called deconvolution. During deconvolution, the catalyst mixture is subdivided into smaller groups, which are again tested in the same transformation. This process is repeated until a single ‘best’ catalyst is selected from the mixture. Following this approach, it is unnecessary to synthesize, isolate and characterize every potential catalyst (Figure 35).163

Figure 35. Catalyst discovery by combinatorial catalysis and iterative deconvolution.

However, combinatorial catalysis also presents some drawbacks. To successfully discover the best catalyst with this approach, every compound needs to be strictly non-interacting with other components in the mixture in order to avoid false negatives and positives. The best scenario for combinatorial catalysis is if catalysts either work or do not catalyse the reaction at all. If this is not the case, the pool with several poor catalysts might outcompete the pool containing the most active one, resulting in false positives instead. Therefore, combinatorial catalysis is best for quantitative rather than qualitative analysis.

4.1.1 Catalyst Discovery in Dynamic Chemistry

A principle aspect of catalyst discovery in dynamic chemistry includes the direct analysis of mixtures of catalysts for finding species showing catalytic potential.164 The main difference with combinatorial catalysis is that the mixture of catalysts is now formed under thermodynamic instead of kinetic control. It is impossible to directly observe active species in the mixture using this approach as the transition

51

state (TS), representing the catalyst bound to the substrate, is a saddle-point on the potential energy surface and not the required thermodynamic minimum. To be able to directly amplify potential catalysts from the mixture, a so-called transition analogue (TSA) can be used instead. The TSA resembles the structure of the transition state, but is kinetically inert under the reaction conditions allowing it to be used as external selector. Any member of the interchanging mixture of components that favours binding to the TSA would be expected to also stabilize the true TS. This methodology has been successfully applied in the discovery of catalysts for the Diels-Alder,165 acetal hydrolysis166 and ester hydrolysis reactions.167-168 Although this is an elegant approach to discover new variations of catalysts, it loses efficiency in reactions with a late TS due to product inhibition. Moreover, for every reaction of interest it is necessary to design and synthesize a new TSA.

A slightly different approach to direct catalyst discovery using CDC is by focusing at the reaction outcome instead of the TSA. Our group has shown that through reversible imine exchange, a bifunctional catalyst mixture could be generated for catalysis of the Morita-Baylis-Hillman reaction. By deconvolution the best catalyst could be selected from the mixture. However, similar to combinatorial catalysis, this approach worked best with on-off type behaviour of the potential catalysts.169 Interestingly, when this reversible modification of a catalytic unit is performed in situ, the activity of the catalyst can be regulated through feedback (Figure 36).170

Figure 36. Catalyst up- and down-regulation by reversible exchange of modifying

group.

52

Research in catalyst discovery using CDC has mainly focused on reactions without stable intermediates along the reaction pathway (i.e. reactions with a single TS). When looking at transition metal catalysis, there are multiple intermediates on the catalytic cycle as well as resting states and off-cycle intermediates. It has been shown that the concentration of such intermediates can be correlated to the catalytic activity.171-172 However, the intermediates on the catalytic cycle are generally short lived and often only the intermediates with a longer lifetime, directly in front of the rate determining step, can be detected. For this reason, information on catalytic cycles is often retrieved by trapping transient intermediates. An investigation was made to probe if it is also possible to trap an intermediate active substrate/metal catalyst combination from a mixture of exchanging substrates and metal catalysts. Identification of such an intermediate would provide information on the catalytically active species and its preferred substrate. The model reaction that was chosen to test this novel approach is the metal-catalysed hydroacylation reaction.

4.2 Hydroacylation Reaction

The ubiquity of C-H bonds in organic molecules makes C-H functionalizations of great interest for the scientific community.173 The two challenges with C-H activations are reactivity and selectivity. To overcome these challenges, directing groups are being used to bring the reactive metal centre in proximity to the functionalizable C-H bond. This preorganization effectively lowers the activation barrier for the initial oxidative addition and provides regioselectivity to the reaction.174-175 The C-H functionalization that was chosen as the model for this study is the hydroacylation reaction (Scheme 9). This reaction has been well-studied and the intermediate substrate/metal complexes have shown to be reasonably stable.176-178 Moreover, catalysis of the hydroacylation reaction has been reported with a wide variety of metals, including ruthenium, platinum, rhodium, iridium and cobalt.179-184

Scheme 9. Metal-catalysed hydroacylation of aldehydes to yield ketones.

53

Similar to other C-H functionalizations, hydroacylations benefit from the presence of a directing group. To avoid having to install and remove directing groups in separate steps, a so-called transient directing group can be used.185-188 These transient directing groups can be attached using existing reversible covalent bonds and can therefore be formed and broken repeatedly during the catalytic cycle. As a result, only catalytic amounts of the directing group are required. In hydroacylation reactions, installation of the transient directing group has been performed both through imine bond formation and transimination reactions (Scheme 10).180, 183, 189-190

N

R H

R' O

R R'

N

R H

DG

H2N DG

N

R

DG

R'

N

RH

DG

O

R H

H2O

imineformation

transiminationreaction

H2N DG

Scheme 10.Transient directing group catalysed hydroacylation.

4.2.1 Trapping and Detection of Reactive Combination

As this approach to catalyst discovery depends on detection of a reactive substrate/metal catalyst combination of the hydroacylation reaction, two considerations need to be made. First of all, efficient exchange of directing groups needs to ensure that different potential reactive substrates are represented in the mixture. The second prerequisite of this approach is the presence of a detectable intermediate that is straightforward to track during the reaction. To be able to rationally design this selection experiment, it is crucial to investigate the catalytic cycle (Scheme 11A).

The hydroacylation reaction proceeds through oxidative addition of the aldimine C-H bond to the metal center. Coordination and insertion of the alkene, followed by reductive elimination affords the ketimine product. As has been shown for the hydroacylation reaction, transimination can be used to exchange the directing group in order to form more of the reactive substrate.183 Additionally, it has been demonstrated that the rate-determining step (RDS) of the reaction is the reductive elimination.191

54

Scheme 11. (A) General catalytic cycle of the transient directing group catalysed hydroacylation reaction. (B) The set-up of this study, where removal of the alkene substrate results in an incomplete catalytic cycle and accumulation of a detectable

intermediate.

The strategy to screen for reactive substrate/metal catalyst combinations for this reaction is shown in Scheme 11B. By removal of the alkene substrate the reaction is trapped at the intermediate metal hydride stage, a species that has previously shown to be relatively stable.191 Another advantage of this species is the ease of observation of the hydride signal by 1H-NMR spectroscopy, as hydride signals are highly shielded and appear between -2 and -25 ppm. A similar transimination strategy is used to continuously generate more of the reactive substrate that is consumed and subsequently removed from the reaction mixture. To ensure generation of higher concentrations of the reactive intermediate, higher catalyst loadings can be used.

4.2.2 Parameters in Substrate and Catalyst Pool

For generation of the substrate pool, six aromatic amines A2-A7 containing different potential directing groups were selected. The building blocks were chosen for their different electronics and geometry for the donation of the lone pair to the metal ion. By mixing with unfunctionalized imine I1, a new equilibrium was achieved, in which all amines A1-A7 and all imines I1-I7 were expressed. As Jun has shown that a potential limiting factor of the hydroacylation reaction could be release of the directing group and subsequent reformation of the reactive substrate.190 An additional reaction parameter thus important in this system is the Brønsted acid transimination catalyst. Therefore, eight

55

different Brønsted acids H1-H8 were selected based on pKa-value and steric differences. The last parameter included in the screening is the metal catalysts. Eight different metal catalysts M1-M8 were selected, four based on preceding activity in C-H activation reactions (IrI, RhI, Pt0, RuII) and four control compounds without any expected activity (MnII, ZnII, NiII, CuI).

Scheme 12. Directing groups, acid co-catalysts and metal catalysts used in this

study.

4.2.3 Deconvolution of Substrate and Catalyst Pool

To select for the optimal reactive substrate/metal catalyst combination in the presence of the most efficient co-catalyst, a multi-parameter iterative deconvolution approach was used.163 In this approach, the catalysts are separated in different batches and screened against the directing group pool. Iterating this deconvolution with the best performing batch leads to selection of the best performing catalyst combination. The results of this iterative deconvolution are shown in Figure 38.

For the first round of screening, I1 was mixed with A2-A7 in the presence of a subset of the acid catalysts H1-H4 or H5-H8. Initial NMR studies confirmed that the system equilibrated under mild heating, with five new imine peaks being observable (I7 was undetected in the NMR

56

spectrum). These two pools of potential substrates with different acid mixtures were split and added to two different pools of metals, M1-M4 and M5-M8. The four generated reaction pools were analysed after 30 minutes by 1H-NMR spectroscopy for formation of a new quartet hydride signal around -11 ppm (Figure 37).

Figure 37. Observed hydride signal in the 1H-NMR spectrum in toluene-d8 of a

confirmed working catalysts/substrate combination.

The reaction pool with the highest concentration of hydride species could be determined via internal standard. For the second round, the reaction pool containing the highest concentration of hydride species was divided into smaller pools of its acids (H1-H2 and H3-H4) and metals (M5-M6 and M7-M8). Following the same procedure as described above, the reaction pool with the highest concentration of hydride species was selected. Repeating the procedure for a third round resulted in detection of H2 (benzoic acid) and M7

(tris(triphenylphosphine)rhodium(I) chloride) as the ideal combination for generation of a reactive intermediate from the pool of different directing groups.

During the deconvolution procedure, two of the reaction parameters were taken into account – the acid and the metal. For determination of the third parameter – the active directing group –a similar deconvolution protocol could be conducted. However, given that the reactive intermediate that is formed during this screening is formed by oxidative addition of the reactive substrate to the metal it should have the highest molecular weight of any species in the pool. Therefore, ESI-MS was chosen as straightforward way to identify the hydride intermediate in

situ. Subjecting the selected pool to ESI-MS revealed I2M7 as the obtained structure (Figure 38). This means that I7 is the reactive substrate, which contains A2 (2-amino-3-picoline) as directing group. To confirm that the hydride I2M7 was the species observed by ESI-MS and

57

to differentiate from the complex I2•M7, the experiment was performed in toluene-d8. This experiment showed a correlation between the hydride formation and consumption of imine I2. Interestingly, I2M7 was obtained with a yield of 25-30%, which is higher than the equilibrium concentration of I2. This indicates the resolving nature of the system, as more reactive substrate is formed while it siphoned out of the reaction by formation of the hydride species.

Figure 38. Screening of directing group system using a multi-parameter iterative

deconvolution. The values in the matrices indicate the concentration (mM, average from duplicate experiment) of the metal hydride species around -11 ppm.

Conditions: I1 (0.03 mmol), A2-A7 (0.015 mmol, 25 mM each), acids (3.0 μmol combined, 10 mol%), toluene (0.60 mL), 80°C, argon, 1 h, then metal catalysts

(0.015 mmol each), 4Å MS (20 mg), 80°C, 30 min.

4.2.4 Efficiency of the Selection

As the highest efficiency in selection of the reactive substrate/metal catalyst combination was only 25-30%, the reaction was further investigated. First the stability of the formed complex I2M7 was studied under the previously described reaction conditions with H2 and M7 in the presence of all amines. The hydride peak was monitored over time by 1H-NMR spectroscopy. As observed from the concentration profile over time (Figure 39), I2M7 rapidly formed but the rate of decomposition begins to outcompete the rate of formation after circa 70 minutes of reaction time.

58

Figure 39. Concentration variation of I2M7 over time during selection experiment.

Conditions: I1 (0.03 mmol), A2-A7 (0.015 mmol, 25 mM each), H2 (3.0 μmol), toluene (0.60 mL), 80°C, argon, 1 h, then M7 (0.015 mmol), 4Å MS (20 mg), 80°C

for the indicated time.

From these results, it is apparent that the formed I2M7 is not persistent under the current reaction conditions, and the reason might be the presence of high concentrations of free amine in the mixture (6:1 ratio, amine to M7). Ligand exchange of the triphenylphosphine ligands on the metal center with free amines could potentially result in the formation of unstable products. To confirm this hypothesis, the effects of increasing and reducing the pool of free amines was tested. By addition or removal of new potential directing groups, the free amine concentration could be altered as shown in Table 9. From these results, a correlation could be observed where lower free amine concentrations result in formation of a higher concentration of I2M7.

59

Table 9. Effect of altering the free amine concentration by introduction or removal of additional directing group containing amines.

N

Toluene, 80°C, Ar

H

I1

In-Im

An-Am

Rh(PPh3)3Cl4 Å MS

60 min

OH

O

N N

Rh

HClPh3P

PPh3

I2M7

An-Am+

NMR yield

(I2M7) A2 A3 A4 A5 A6 A7 A8 A9

15

32

27

57

73 *Checkmark indicates presence of specified amine in the reaction. Conditions: I1 (0.03 mmol), amines A2-A9 (0.015 mmol each, if included), benzoic acid (3.0 μmol), toluene (0.60 mL), 80°C, argon, 60 min, then Rh(PPh3)3Cl (0.015 mmol), 4 Å MS (ground, 20 mg), 80°C, 60 min. NMR yield of I2M7 determined by 1H-NMR spectroscopy with PhSiMe3 as internal standard.

With the knowledge that higher free amine concentrations reduce the possibility of producing a hit, an alternative protocol was devised decreasing the presence of free amines in the system. To achieve this, all amines A1-A7 were condensed with a slightly substoichiometric quantity of benzaldehyde in the presence of molecular sieves (MS) and benzoic acid H2. After one day of reaction time, 97% consumption of benzaldehyde was observed and the formed imine mixture was subjected to Wilkinson’s catalyst M7. Formation of hydride I2M7 was observed, now with a more than three times improved NMR yield of 81% (Scheme 13). Even though this approach utilizes more aldehyde, the highly increased efficiency would decrease the chance of false negatives.

O

Toluene, 4 Å MS 80°C, Ar, 24 h

HI1-I7

A1-A7 60 min

OH

O

N N

Rh

HClPh3P

PPh3

I2M7

A1-A7+Rh(PPh3)3Cl

6 equiv. 1 equiv. each

81%

Scheme 13. Reduction of free amine content by direct condensation and selective

C-H activation from the formed imine pool.

60

4.2.5 Catalytic Role of the Reactive Intermediate

As the hydride I2M7 was captured from the catalytic cycle by removal of the alkene substrate, it is crucial to confirm that this species is actually a reactive intermediate. To confirm that this intermediate can form the hydroacylation product and is not an off-cycle structure, I2M7 was independently synthesized and isolated. Pure I2M7 was then subjected to hydroacylation conditions as shown in Scheme 14. When I2M7 was reacted with 1-octene followed by hydrolysis of the obtained ketimine product, the expected hydroacylation product was obtained with an isolated yield of 80%.

Scheme 14. Synthesis of I2M7 and conversion to the hydroacylation product by

reaction with 1-octene and hydrolysis of the ketimine.

4.3 Conclusions

This chapter has shown a novel approach to select and identify a reactive substrate/metal catalyst combination for the hydroacylation reaction. Dynamic exchange of directing groups resulted in reformation of the reactive substrate in the mixture, which is siphoned out by the active catalyst. This results in amplification of the reactive intermediate, increasing effectiveness and applicability of this approach in catalyst discovery. Overall, this study demonstrates the investigation of 448 possible directing group, acid and metal catalyst combinations in just 12 separate experiments. Careful consideration is however still required for more subtle differences in reactivity like those observed with the acids. Care also needs to be taken to avoid interaction of different members of the pool as this could give inhibitory and synergistic effects resulting in observation of false positives or false negatives. It is worthwhile noting that I2M7 is the most reactive structure to be resolved from a dynamic mixture, showing that transient or metastable species can also provide a kinetic sink for such systems.

61

5. Concluding Remarks

The title – Metals in Dynamic Chemistry: Selection & Catalysis – suggests that the focus of this thesis was placed on applying metals in the field of dynamic chemistry. In prospect this holds true, however this occurred serendipitously and the recurrence of metals in this thesis is solely a consequence of their ability to realize processes that are otherwise inaccessible.

In chapter two, two novel carbohydrate nanoplatforms were synthesized – the dodecaglycosylated cages and Borromeates – from which the latter was selected by templating with zinc ions. Using quartz crystal microbalance technology the real-time binding behaviour to concanavalin A and Galanthus nivalis agglutinin was observed. Comparison of the binding behaviour of the two novel nanoplatforms with dodecaglycosylated fullerenes indicated the importance of statistical and polyelectrolyte effects for observed enhancement of affinity.

In chapter three, a ruthenium-based olefin metathesis catalyst was prepared for application in aqueous media. During the study of self-metathesis of highly functionalized substrates, such as carbohydrates, it was discovered that addition of acetic acid prevents undesired isomerization. Additionally, it was shown that non-productive chelation limits the reactivity of highly functionalized substrates in self-metathesis, a property that could be used for increasing selectivity in cross-metathesis.

In the last chapter of this thesis, a proof of concept was provided for an alternative method to discover catalytic activity from mixtures. By combining iterative deconvolution and mass spectrometry, a reactive substrate/metal catalyst combination could be amplified and detected from a mixture of acid catalysts, transition metal catalysts and transient directing groups.

In summary, this thesis thus demonstrated how metals can be used in

dynamic chemistry to select nanoplatforms and catalyse olefin exchange. Furthermore, it has been shown that the catalyst can also be used to select the directing group that provides a reactive combination of the two.

62

Acknowledgements

This thesis would not have come to this point without the influence of others. Whether it was by aiding in administration, by conveying knowledge, by providing good and/or bad examples, by providing opportunity or by just being you (whatever that may mean) – you deserve the Dutch way of expressing gratitude – Bedankt! As everyone in the following list has influenced me through their attitude in approaching problems, I have chosen to present this list in alphabetical order.

Administrative and other personnel – Without you, who would have time for other things? Fortunately, there is many of you including Ulla, Lena, Ann, Ilona, Vera, Daniel, Nina, Susanne, Johanna, Inger, Henry, Pia, Mats, Andra and all of those that work behind the scenes whose existence is rarely acknowledged.

Collaborators – Science is not trying to improve the world on your own, but by standing strong together. Thank you Fredrik, Lei, Oleksandr, Marta, Sara, Olof2, Allan, Daniel, Lars, Davide, Samuel, Teodor, Stéphane, Johan and Peter for standing by my side or allowing me to stand by yours.

Forgotten in translation – �� = ∆� ∙ √�

���� ← This equation shows the

correlation between the persons that are forgotten here (Fp), with the time difference in which we last had contact (Δt), the square root of my stupidity (S) divided by my ability to remember all the people that needed to receive a thank you here (ARty).

Financial Support – The things you can do if you are provided with financial support are amazing. Therefore my gratitude to the European Commission for financing DYNANO and Aulin-Erdtman foundation for financing my attendance to conferences all over the world.

Group – Over the years many people have been in the group, but many people have left as well including Fredrik, Yan, Yang, Lei, Juan, Sheng, Boo, Niranjan, Qi, Laura, Ruzal, Ryo, Sesha, Maurice, Julia, Björn, Javier, Veronika, Allan, Linnea, Sonja, Karolina, Sara and Olof. Now it is my time to leave Oleksandr, Yanmiao, Antanas, Yansong, Giampiero, Kequan and Joakim behind. Thank you for the atmosphere you have provided in the offices and on the laboratory over all these years.

63

Networkers – The interdisciplinary environment provided by Silviya, Marta, Susanne, Calin, Ioanna, Eva, Valentina, Xabier, Erol, Muhammet, Luca, Alina, Romina, Allessandro, Lars, Madalina, Yan, Junjun, Marion, Hélène, Laurence, Sadika, Mihail, Nino, Antonio, Claudiu, Dolores, Edit, Jesús, Nicolas, Michaela and Stéphane often left me not knowing what everyone is talking about, but provided a highly stimulating environment to learn more.

Students – Many students have passed over the years, those I have guided personally (Björn, Pamina and Allan) and those I have guided as a group (Organisk Kemi 1 and Organisk Kemi 2). Hope you have at least learned as much from me as I have learned from teaching you.

Sidekicks – Some superheroes need sidekicks to teach the students crucial practical laboratory skills. If sidekicks would work together like Fredrik, Björn, Rebec(c/k)a, Robin, Olof, Viktor, Guillermo, Nils, Max, Joakim, Alex, Lele, Tove and me did over all the years the superhero movies would have been a lot more boring. You learn from teaching, but you learn a lot more from teaching together. Every one of you had their own style providing me with great examples of how I could improve my own teaching abilities.

Superheroes – The real heroes in science are those that convey their knowledge to those that are in need of some extra. Even though many of the courses repeated a lot of information that I had already received from previous superheroes there was still a lot of things that Christina, Mingdi, Zoltan, Torbjörn, Pino, Kalman, Fahmi and Jan-Erling managed to teach me.

Supervisor – Freedom is not always a given, but in the group of Olof it definitely is. Thank you for channelling some of the funding to me to support me as a PhD-student in your group and for guiding me in what to do and what not to do both, in the past and for the future.

The OC – The division would not have been the same without colleagues, whether lazy or active. Special thanks to those that spend considerable time to make the division more professional looking including Fredrik, Quentin, Anna, Robin, Oleksandr, Joakim, Yansong, Christina and Peter.

Zpecial person – There is always a person that you are willing to cheat the system for and for me that is Miranda. Hope that says enough!

64

Appendix

The following is a description of my contributions to Publications I to V, as requested by KTH: Paper I: I contributed to the formulation of the research problem, shared the experimental work and wrote the draft manuscript. Paper II: I contributed to the formulation of the research problem, shared the experimental work and wrote the draft manuscript. Paper III: I formulated the research problem, performed the experimental work and wrote the draft manuscript. Paper IV: I formulated the research problem, performed majority of the experimental work and wrote the draft manuscript. Paper V: I contributed to discussion of the research problems, performed part of the experimental work and proofread the draft manuscript.

65

References 1. Monosson, E., Metal Planet in Evolution in a Toxic World: How Life Responds

to Chemical Threats, Island Press: Washington, 2012; 48-64. 2. Reek, J. N. H.; Otto, S., Dynamic Combinatorial Chemistry. Wiley-VCH:

Weinheim, 2010. 3. Lehn, J.-M., Chem. Soc. Rev. 2007, 36, 151-160. 4. Lehn, J.-M., Angew. Chem. Int. Ed. 2015, 54, 3276-3289. 5. Lehn, J.-M., Angew. Chem. Int. Ed. 2013, 52, 2836-2850. 6. Barboiu, M., Constitutional Dynamic Chemistry. Springer: Heidelberg, 2012. 7. Tebben, L.; Mück-Lichtenfeld, C.; Fernández, G.; Grimme, S.; Studer, A.,

Chem. Eur. J. 2017, 23, 5864-5873. 8. Rowan, S. J.; Cantrill, S. J.; Cousins, G. R. L.; Sanders, J. K. M.; Stoddart, J.

F., Angew. Chem. Int. Ed. 2002, 41, 898-952. 9. Zhang, W., Dynamic Covalent Chemistry: Principles, Reactions, and

Applications. John Wiley & Sons, Limited: 2017. 10. Li, J.; Nowak, P.; Otto, S., J. Am. Chem. Soc. 2013, 135, 9222-9239. 11. van Leeuwen, P. W. N. M., Homogeneous Catalysis. Kluwer Academic

Publishers: Dordrecht, 2004. 12. Sakulsombat, M.; Zhang, Y.; Ramström, O., Dynamic Systemic Resolution in

Constitutional Dynamic Chemistry, Barboiu, M., Ed. Springer Berlin Heidelberg: Berlin, Heidelberg, 2012; 55-86.

13. Ji, Q.; Lirag, R. C.; Miljanić, O. Š., Chem. Soc. Rev. 2014, 43, 1873-1884. 14. Osowska, K.; Miljanić, O. Š., Synlett 2011, 1643-1648. 15. He, Z.; Jiang, W.; Schalley, C. A., Chem. Soc. Rev. 2015, 44, 779-789. 16. Huc, I.; Lehn, J.-M., Proc. Natl. Acad. Sci. USA 1997, 94, 2106-2110. 17. Ramström, O.; Lehn, J.-M., Nat. Rev. Drug Discov. 2002, 1, 26-36. 18. Hasenknopf, B.; Lehn, J.-M.; Kneisel, B. O.; Baum, G.; Fenske, D., Angew.

Chem. Int. Ed. 1996, 35, 1838-1840. 19. Hasenknopf, B.; Lehn, J.-M.; Boumediene, N.; Dupont-Gervais, A.; Van

Dorsselaer, A.; Kneisel, B.; Fenske, D., J. Am. Chem. Soc. 1997, 119, 10956-10962.

20. Brachvogel, R.-C.; von Delius, M., Chem. Sci. 2015, 6, 1399-1403. 21. Brachvogel, R.-C.; Hampel, F.; von Delius, M., Nat. Commun. 2015, 6, 7129. 22. Shyshov, O.; Brachvogel, R.-C.; Bachmann, T.; Srikantharajah, R.; Segets, D.;

Hampel, F.; Puchta, R.; von Delius, M., Angew. Chem. Int. Ed. 2017, 56, 776-781.

23. Ruben, M.; Rojo, J.; Romero-Salguero, F. J.; Uppadine, L. H.; Lehn, J.-M., Angew. Chem. Int. Ed. 2004, 43, 3644-3662.

24. Kramer, R.; Lehn, J. M.; Marquis-Rigault, A., Proc. Natl. Acad. Sci. USA 1993, 90, 5394-5398.

25. Kocsis, I.; Dumitrescu, D.; Legrand, Y.-M.; van der Lee, A.; Grosu, I.; Barboiu, M., Chem. Commun. 2014, 50, 2621-2623.

26. Funeriu, D. P.; Lehn, J.-M.; Fromm, K. M.; Fenske, D., Chem. Eur. J. 2000, 6, 2103-2111.

27. Leussing, D. L.; Leach, B. E., J. Am. Chem. Soc. 1971, 93, 3377-3384. 28. Giuseppone, N.; Schmitt, J.-L.; Schwartz, E.; Lehn, J.-M., J. Am. Chem. Soc.

2005, 127, 5528-5539.

66

29. Giuseppone, N.; Schmitt, J.-L.; Lehn, J.-M., Angew. Chem. Int. Ed. 2004, 43, 4902-4906.

30. Eldred, S. E.; Stone, D. A.; Gellman, S. H.; Stahl, S. S., J. Am. Chem. Soc.

2003, 125, 3422-3423. 31. Stephenson, N. A.; Zhu, J.; Gellman, S. H.; Stahl, S. S., J. Am. Chem. Soc.

2009, 131, 10003-10008. 32. Becerra-Figueroa, L.; Ojeda-Porras, A.; Gamba-Sánchez, D., J. Org. Chem.

2014, 79, 4544-4552. 33. Vongvilai, P.; Ramström, O., J. Am. Chem. Soc. 2009, 131, 14419-14425. 34. Sutton, L. R.; Donaubauer, W. A.; Hampel, F.; Hirsch, A., Chem. Commun.

2004, 1758-1759. 35. Amatore, C.; Jutand, A.; Meyer, G.; Mottier, L., Chem. Eur. J. 1999, 5, 466-

473. 36. Kaiser, G.; Sanders, J. K. M., Chem. Commun. 2000, 1763-1764. 37. Heppekausen, J.; Stade, R.; Goddard, R.; Fürstner, A., J. Am. Chem. Soc. 2010,

132, 11045-11057. 38. Jyothish, K.; Zhang, W., Angew. Chem. Int. Ed. 2011, 50, 8478-8480. 39. Gross, D. E.; Moore, J. S., Macromolecules 2011, 44, 3685-3687. 40. Sisco, S. W.; Moore, J. S., J. Am. Chem. Soc. 2012, 134, 9114-9117. 41. Wang, Q.; Yu, C.; Zhang, C.; Long, H.; Azarnoush, S.; Jin, Y.; Zhang, W.,

Chem. Sci. 2016, 7, 3370-3376. 42. Wang, Q.; Yu, C.; Long, H.; Du, Y.; Jin, Y.; Zhang, W., Angew. Chem. Int. Ed.

2015, 54, 7550-7554. 43. Garber, S. B.; Kingsbury, J. S.; Gray, B. L.; Hoveyda, A. H., J. Am. Chem. Soc.

2000, 122, 8168-8179. 44. Leigh, D. A.; Pritchard, R. G.; Stephens, A. J., Nat. Chem. 2014, 6, 978-982. 45. Okochi, K. D.; Jin, Y.; Zhang, W., Chem. Commun. 2013, 49, 4418-4420. 46. Wilson, A.; Gasparini, G.; Matile, S., Chem. Soc. Rev. 2014, 43, 1948-1962. 47. Varki, A.; Cummings, R. D.; Esko, J. D.; Freeze, H. H.; Stanley, P.; Bertozzi,

C. R.; Hart, G. W.; Etzler, M. E., Essentials of Glycobiology, 2nd edition. Cold Spring Harbor Laboratory Press: New York, 2009.

48. Mammen, M.; Choi, S.-K.; Whitesides, G. M., Angew. Chem. Int. Ed. 1998, 37, 2754-2794.

49. Hunter, C. A.; Anderson, H. L., Angew. Chem. Int. Ed. 2009, 48, 7488-7499. 50. Lundquist, J. J.; Toone, E. J., Chem. Rev. 2002, 102, 555-578. 51. Cecioni, S.; Imberty, A.; Vidal, S., Chem. Rev. 2015, 115, 525-561. 52. Chabre, Y. M.; Roy, R., Adv. Carbohydr. Chem. Biochem. 2010, 63, 165-393. 53. Chen, X.; Ramström, O.; Yan, M., Nano Res. 2014, 7, 1381-1403. 54. Wang, X.; Ramström, O.; Yan, M., Adv. Mater. 2010, 22, 1946-1953. 55. Bernardi, A.; Jiménez-Barbero, J.; Casnati, A.; De Castro, C.; Darbre, T.;

Fieschi, F.; Finne, J.; Funken, H.; Jaeger, K.-E.; Lahmann, M.; Lindhorst, T. K.; Marradi, M.; Messner, P.; Molinaro, A.; Murphy, P. V.; Nativi, C.; Oscarson, S.; Penades, S.; Peri, F.; Pieters, R. J.; Renaudet, O.; Reymond, J.-L.; Richichi, B.; Rojo, J.; Sansone, F.; Schaffer, C.; Turnbull, W. B.; Velasco-Torrijos, T.; Vidal, S.; Vincent, S.; Wennekes, T.; Zuilhof, H.; Imberty, A., Chem. Soc. Rev. 2013, 42, 4709-4727.

56. Wittmann, V.; Pieters, R. J., Chem. Soc. Rev. 2013, 42, 4492-4503. 57. Ramström, O.; Yan, M., Chem. Eur. J. 2015, 21, 16310-16317. 58. Xu, D.; Warmuth, R., J. Am. Chem. Soc. 2008, 130, 7520-7521.

67

59. Skowronek, P.; Warzajtis, B.; Rychlewska, U.; Gawronski, J., Chem. Commun.

2013, 49, 2524-2526. 60. Sun, J.; Warmuth, R., Chem. Commun. 2011, 47, 9351-9353. 61. Mastalerz, M., Chem. Commun. 2008, 4756-4758. 62. Wang, Q.; Zhang, C.; Noll, B. C.; Long, H.; Jin, Y.; Zhang, W., Angew. Chem.

Int. Ed. 2014, 53, 10663-10667. 63. Klotzbach, S.; Scherpf, T.; Beuerle, F., Chem. Commun. 2014, 50, 12454-

12457. 64. Tönnemann, J.; Scopelliti, R.; Zhurov, K. O.; Menin, L.; Dehnen, S.; Severin,

K., Chem. Eur. J. 2012, 18, 9939-9945. 65. Ayme, J.-F.; Beves, J. E.; Leigh, D. A.; McBurney, R. T.; Rissanen, K.;

Schultz, D., Nat. Chem. 2012, 4, 15-20. 66. Ponnuswamy, N.; Cougnon, F. B. L.; Clough, J. M.; Pantoş, G. D.; Sanders, J.

K. M., Science 2012, 338, 783-785. 67. Prakasam, T.; Lusi, M.; Elhabiri, M.; Platas-Iglesias, C.; Olsen, J.-C.; Asfari,

Z.; Cianférani-Sanglier, S.; Debaene, F.; Charbonnière, L. J.; Trabolsi, A., Angew. Chem. Int. Ed. 2013, 52, 9956-9960.

68. Patel, K.; Miljanić, O. Š.; Stoddart, J. F., Chem. Commun. 2008, 1853-1855. 69. Furusho, Y.; Oku, T.; Rajkumar, G. A.; Takata, T., Chem. Lett. 2004, 33, 52-

53. 70. Li, H.; Zhang, H.; Lammer, A. D.; Wang, M.; Li, X.; Lynch, V. M.; Sessler, J.

L., Nat. Chem. 2015, 7, 1003-1008. 71. Satoshi, S.; Takayuki, M.; Hiroyuki, M.; Hiroshi, T., Supramol. Chem. 2011,

23, 244-248. 72. Mal, P.; Schultz, D.; Beyeh, K.; Rissanen, K.; Nitschke, J. R., Angew. Chem.

Int. Ed. 2008, 47, 8297-8301. 73. Jansze, S. M.; Cecot, G.; Wise, M. D.; Zhurov, K. O.; Ronson, T. K.; Castilla,

A. M.; Finelli, A.; Pattison, P.; Solari, E.; Scopelliti, R.; Zelinskii, G. E.; Vologzhanina, A. V.; Voloshin, Y. Z.; Nitschke, J. R.; Severin, K., J. Am.

Chem. Soc. 2016, 138, 2046-2054. 74. Cecioni, S.; Oerthel, V.; Iehl, J.; Holler, M.; Goyard, D.; Praly, J.-P.; Imberty,

A.; Nierengarten, J.-F.; Vidal, S., Chem. Eur. J. 2011, 17, 3252-3261. 75. Durka, M.; Buffet, K.; Iehl, J.; Holler, M.; Nierengarten, J.-F.; Vincent, S. P.,

Chem. Eur. J. 2012, 18, 641-651. 76. Nierengarten, J.-F.; Iehl, J.; Oerthel, V.; Holler, M.; Illescas, B. M.; Munoz,

A.; Martin, N.; Rojo, J.; Sanchez-Navarro, M.; Cecioni, S.; Vidal, S.; Buffet, K.; Durka, M.; Vincent, S. P., Chem. Commun. 2010, 46, 3860-3862.

77. Sánchez-Navarro, M.; Muñoz, A.; Illescas, B. M.; Rojo, J.; Martín, N., Chem.

Eur. J. 2011, 17, 766-769. 78. Hirsch, A.; Vostrowsky, O., Eur. J. Org. Chem. 2001, 829-848. 79. Mao, C.; Sun, W.; Seeman, N. C., Nature 1997, 386, 137-138. 80. Chichak, K. S.; Cantrill, S. J.; Pease, A. R.; Chiu, S.-H.; Cave, G. W. V.;

Atwood, J. L.; Stoddart, J. F., Science 2004, 304, 1308-1312. 81. Chichak, K. S.; Peters, A. J.; Cantrill, S. J.; Stoddart, J. F., J. Org. Chem. 2005,

70, 7956-7962. 82. Yates, C. R.; Benítez, D.; Khan, S. I.; Stoddart, J. F., Org. Lett. 2007, 9, 2433-

2436. 83. Skowronek, P.; Gawronski, J., Org. Lett. 2008, 10, 4755-4758.

68

84. Tozawa, T.; Jones, J. T. A.; Swamy, S. I.; Jiang, S.; Adams, D. J.; Shakespeare, S.; Clowes, R.; Bradshaw, D.; Hasell, T.; Chong, S. Y.; Tang, C.; Thompson, S.; Parker, J.; Trewin, A.; Bacsa, J.; Slawin, A. M. Z.; Steiner, A.; Cooper, A. I., Nat. Mater. 2009, 8, 973-978.

85. Culshaw, J. L.; Cheng, G.; Schmidtmann, M.; Hasell, T.; Liu, M.; Adams, D. J.; Cooper, A. I., J. Am. Chem. Soc. 2013, 135, 10007-10010.

86. Swamy, S. I.; Bacsa, J.; Jones, J. T. A.; Stylianou, K. C.; Steiner, A.; Ritchie, L. K.; Hasell, T.; Gould, J. A.; Laybourn, A.; Khimyak, Y. Z.; Adams, D. J.; Rosseinsky, M. J.; Cooper, A. I., J. Am. Chem. Soc. 2010, 132, 12773-12775.

87. Iehl, J.; Nierengarten, J.-F., Chem. Eur. J. 2009, 15, 7306-7309. 88. Hu, Q.; Yang, H.; Wang, Y.; Xu, S., Chem. Commun. 2016, 52, 3705-3708. 89. Duverger, E.; Frison, N.; Roche, A.-C.; Monsigny, M., Biochimie 2003, 85,

167-179. 90. Cheng, C. I.; Chang, Y.-P.; Chu, Y.-H., Chem. Soc. Rev. 2012, 41, 1947-1971. 91. Pei, Z.; Anderson, H.; Aastrup, T.; Ramström, O., Biosens. Bioelectron. 2005,

21, 60-66. 92. Diederich, F.; Künzer, H., Recent Trends in Molecular Recognition. 1st ed.;

Springer Berlin Heidelberg: New York, 1998. 93. Grubbs, R. H.; Wenzel, A. G.; O’Leary, D. J.; Khosravi, E., Handbook of

Metathesis. Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2015. 94. Keitz, B. K.; Endo, K.; Herbert, M. B.; Grubbs, R. H., J. Am. Chem. Soc. 2011,

133, 9686-9688. 95. Koh, M. J.; Khan, R. K. M.; Torker, S.; Yu, M.; Mikus, M. S.; Hoveyda, A. H.,

Nature 2015, 517, 181-186. 96. Smit, W.; Koudriavtsev, V.; Occhipinti, G.; Törnroos, K. W.; Jensen, V. R.,

Organometallics 2016, 35, 1825-1837. 97. Marx, V. M.; Sullivan, A. H.; Melaimi, M.; Virgil, S. C.; Keitz, B. K.;

Weinberger, D. S.; Bertrand, G.; Grubbs, R. H., Angew. Chem. Int. Ed. 2015, 54, 1919-1923.

98. Bidange, J.; Fischmeister, C.; Bruneau, C., Chem. Eur. J. 2016, 22, 12226-12244.

99. Malcolmson, S. J.; Meek, S. J.; Sattely, E. S.; Schrock, R. R.; Hoveyda, A. H., Nature 2008, 456, 933-937.

100. Schoeps, D.; Buhr, K.; Dijkstra, M.; Ebert, K.; Plenio, H., Chem. Eur. J. 2009, 15, 2960-2965.

101. Rix, D.; Caijo, F.; Laurent, I.; Gułajski, Ł.; Grela, K.; Mauduit, M., Chem.

Commun. 2007, 3771-3773. 102. Hoveyda, A. H.; Zhugralin, A. R., Nature 2007, 450, 243-251. 103. Grubbs, R. H., Tetrahedron 2004, 60, 7117-7140. 104. Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W., J. Am. Chem. Soc.

1992, 114, 3974-3975. 105. Murdzek, J. S.; Schrock, R. R., Organometallics 1987, 6, 1373-1374. 106. Schrock, R. R.; Murdzek, J. S.; Bazan, G. C.; Robbins, J.; DiMare, M.;

O'Regan, M., J. Am. Chem. Soc. 1990, 112, 3875-3886. 107. Monsaert, S.; Drozdzak, R.; Dragutan, V.; Dragutan, I.; Verpoort, F., Eur. J.

Inorg. Chem. 2008, 432-440. 108. Trnka, T. M.; Morgan, J. P.; Sanford, M. S.; Wilhelm, T. E.; Scholl, M.; Choi,

T.-L.; Ding, S.; Day, M. W.; Grubbs, R. H., J. Am. Chem. Soc. 2003, 125, 2546-2558.

69

109. Torborg, C.; Szczepaniak, G.; Zielinski, A.; Malinska, M.; Wozniak, K.; Grela, K., Chem. Commun. 2013, 49, 3188-3190.

110. Love, J. A.; Morgan, J. P.; Trnka, T. M.; Grubbs, R. H., Angew. Chem. Int. Ed.

2002, 41, 4035-4037. 111. Kingsbury, J. S.; Harrity, J. P. A.; Bonitatebus, P. J.; Hoveyda, A. H., J. Am.

Chem. Soc. 1999, 121, 791-799. 112. Gessler, S.; Randl, S.; Blechert, S., Tetrahedron Lett. 2000, 41, 9973-9976. 113. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H., Org. Lett. 1999, 1, 953-956. 114. Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H., Angew. Chem. Int. Ed.

1995, 34, 2039-2041. 115. Lu, Y.-X.; Guan, Z., J. Am. Chem. Soc. 2012, 134, 14226-14231. 116. Neal, J. A.; Mozhdehi, D.; Guan, Z., J. Am. Chem. Soc. 2015, 137, 4846-4850. 117. Fürstner, A.; Thiel, O. R.; Ackermann, L., Org. Lett. 2001, 3, 449-451. 118. van Gerven, P. C. M.; Elemans, J. A. A. W.; Gerritsen, J. W.; Speller, S.; Nolte,

R. J. M.; Rowan, A. E., Chem. Commun. 2005, 3535-3537. 119. Jin, Y.; Zhang, A.; Huang, Y.; Zhang, W., Chem. Commun. 2010, 46, 8258-

8260. 120. Binder, J. B.; Raines, R. T., Curr. Opin. Chem. Biol. 2008, 12, 767-773. 121. Brändli, C.; Ward, T. R., Helv. Chim. Acta 1998, 81, 1616-1621. 122. Poulsen, S.-A.; Bornaghi, L. F., Bioorg. Med. Chem. 2006, 14, 3275-3284. 123. Denard, C. A.; Huang, H.; Bartlett, M. J.; Lu, L.; Tan, Y.; Zhao, H.; Hartwig,

J. F., Angew. Chem. Int. Ed. 2014, 53, 465-469. 124. Hunter, L.; Condie, G. C.; Harding, M. M., Tetrahedron Lett. 2010, 51, 5064-

5067. 125. McNaughton, B. R.; Bucholtz, K. M.; Camaaño-Moure, A.; Miller, B. L., Org.

Lett. 2005, 7, 733-736. 126. Tomasek, J.; Schatz, J., Green Chem. 2013, 15, 2317-2338. 127. Burtscher, D.; Grela, K., Angew. Chem. Int. Ed. 2009, 48, 442-454. 128. Grela, K.; Gułajski, Ł.; Skowerski, K., Alkene Metathesis in Water in Metal-

Catalyzed Reactions in Water, Wiley-VCH Verlag GmbH & Co. KGaA: 2013; 291-336.

129. Levin, E.; Ivry, E.; Diesendruck, C. E.; Lemcoff, N. G., Chem. Rev. 2015, 115, 4607-4692.

130. Gułajski, Ł.; Sledz, P.; Lupa, A.; Grela, K., Green Chem. 2008, 10, 271-274. 131. Lynn, D. M.; Kanaoka, S.; Grubbs, R. H., J. Am. Chem. Soc. 1996, 118, 784-

790. 132. Lipshutz, B. H.; Ghorai, S.; Abela, A. R.; Moser, R.; Nishikata, T.; Duplais,

C.; Krasovskiy, A.; Gaston, R. D.; Gadwood, R. C., J. Org. Chem. 2011, 76, 4379-4391.

133. Davis, K. J.; Sinou, D., J. Mol. Catal. A: Chem. 2002, 177, 173-178. 134. Lipshutz, B. H.; Aguinaldo, G. T.; Ghorai, S.; Voigtritter, K., Org. Lett. 2008,

10, 1325-1328. 135. Lipshutz, B. H.; Ghorai, S., Org. Lett. 2009, 11, 705-708. 136. Diallo, A. K.; Boisselier, E.; Liang, L.; Ruiz, J.; Astruc, D., Chem. Eur. J. 2010,

16, 11832-11835. 137. Brendgen, T.; Fahlbusch, T.; Frank, M.; Schühle, D. T.; Seßler, M.; Schatz, J.,

Adv. Synth. Catal. 2009, 351, 303-307. 138. Mayer, C.; Gillingham, D. G.; Ward, T. R.; Hilvert, D., Chem. Commun. 2011,

47, 12068-12070.

70

139. Basauri-Molina, M.; Verhoeven, D. G. A.; van Schaik, A. J.; Kleijn, H.; Klein Gebbink, R. J. M., Chem. Eur. J. 2015, 21, 15676-15685.

140. Lo, C.; Ringenberg, M. R.; Gnandt, D.; Wilson, Y.; Ward, T. R., Chem.

Commun. 2011, 47, 12065-12067. 141. Jeschek, M.; Reuter, R.; Heinisch, T.; Trindler, C.; Klehr, J.; Panke, S.; Ward,

T. R., Nature 2016, 537, 661-665. 142. Kirkland, T. A.; Lynn, D. M.; Grubbs, R. H., J. Org. Chem. 1998, 63, 9904-

9909. 143. Gułajski, Ł.; Michrowska, A.; Narożnik, J.; Kaczmarska, Z.; Rupnicki, L.;

Grela, K., ChemSusChem 2008, 1, 103-109. 144. Michrowska, A.; Gułajski, Ł.; Kaczmarska, Z.; Mennecke, K.; Kirschning, A.;

Grela, K., Green Chem. 2006, 8, 685-688. 145. Jordan, J. P.; Grubbs, R. H., Angew. Chem. Int. Ed. 2007, 46, 5152-5155. 146. Skowerski, K.; Szczepaniak, G.; Wierzbicka, C.; Gułajski, Ł.; Bieniek, M.;

Grela, K., Catal. Sci. Tech. 2012, 2, 2424-2427. 147. Mohr, B.; Lynn, D. M.; Grubbs, R. H., Organometallics 1996, 15, 4317-4325. 148. Wang, Z. J.; Jackson, W. R.; Robinson, A. J., Green Chem. 2015, 17, 3407-

3414. 149. Gallivan, J. P.; Jordan, J. P.; Grubbs, R. H., Tetrahedron Lett. 2005, 46, 2577-

2580. 150. Hong, S. H.; Grubbs, R. H., J. Am. Chem. Soc. 2006, 128, 3508-3509. 151. Gleeson, E. C.; Wang, Z. J.; Jackson, W. R.; Robinson, A. J., J. Org. Chem.

2015, 80, 7205-7211. 152. Lin, Y. A.; Chalker, J. M.; Floyd, N.; Bernardes, G. J. L.; Davis, B. G., J. Am.

Chem. Soc. 2008, 130, 9642-9643. 153. Lin, Y. A.; Chalker, J. M.; Davis, B. G., ChemBioChem 2009, 10, 959-969. 154. Lin, Y. A.; Chalker, J. M.; Davis, B. G., J. Am. Chem. Soc. 2010, 132, 16805-

16811. 155. Connon, S. J.; Blechert, S., Bioorg. Med. Chem. Lett. 2002, 12, 1873-1876. 156. Sashuk, V.; Schoeps, D.; Plenio, H., Chem. Commun. 2009, 770-772. 157. Miller, B. L., Dynamic Combinatorial Chemistry: In Drug Discovery,

Bioorganic Chemistry, and Materials Science. Wiley: Hoboken, 2010. 158. Ireland, B. J.; Dobigny, B. T.; Fogg, D. E., ACS Catal. 2015, 5, 4690-4698. 159. Hong, S. H.; Sanders, D. P.; Lee, C. W.; Grubbs, R. H., J. Am. Chem. Soc.

2005, 127, 17160-17161. 160. Matsuo, T.; Yoshida, T.; Fujii, A.; Kawahara, K.; Hirota, S., Organometallics

2013, 32, 5313-5319. 161. Chatterjee, A. K.; Choi, T.-L.; Sanders, D. P.; Grubbs, R. H., J. Am. Chem.

Soc. 2003, 125, 11360-11370. 162. Dahmen, S.; Bräse, S., Synthesis 2001, 1431-1449. 163. Wolf, E.; Richmond, E.; Moran, J., Chem. Sci. 2015, 6, 2501-2505. 164. Gasparini, G.; Dal Molin, M.; Prins, L. J., Eur. J. Org. Chem. 2010, 2429-2440. 165. Brisig, B.; Sanders, J. K. M.; Otto, S., Angew. Chem. Int. Ed. 2003, 42, 1270-

1273. 166. Vial, L.; Sanders, J. K. M.; Otto, S., New J. Chem. 2005, 29, 1001-1003. 167. Kannappan, R.; Nicholas, K. M., ACS. Comb. Sci. 2013, 15, 90-100. 168. Gasparini, G.; Prins, L. J.; Scrimin, P., Angew. Chem. 2008, 120, 2509-2513. 169. Schaufelberger, F.; Ramström, O., Chem. Eur. J. 2015, 21, 12735-12740. 170. Schaufelberger, F.; Ramström, O., J. Am. Chem. Soc. 2016, 138, 7836-7839.

71

171. Markert, C.; Pfaltz, A., Angew. Chem. Int. Ed. 2004, 43, 2498-2500. 172. Muller, C. A.; Markert, C.; Teichert, A. M.; Pfaltz, A., Chem. Commun. 2009,

1607-1618. 173. Shilov, A. E.; Shul'pin, G. B., Chem. Rev. 1997, 97, 2879-2932. 174. Zhang, M.; Zhang, Y.; Jie, X.; Zhao, H.; Li, G.; Su, W., Org. Chem. Front.

2014, 1, 843-895. 175. Davies, H. M. L.; Morton, D., J. Org. Chem. 2016, 81, 343-350. 176. Jun, C.-H.; Jo, E.-A.; Park, J.-W., Eur. J. Org. Chem. 2007, 1869-1881. 177. Ghosh, A.; Johnson, K. F.; Vickerman, K. L.; Walker, J. A.; Stanley, L. M.,

Org. Chem. Front. 2016, 3, 639-644. 178. Albinati, A.; Arz, C.; Pregosin, P. S., J. Organomet. Chem. 1987, 335, 379-

394. 179. Nagamoto, M.; Nishimura, T., Chem. Commun. 2015, 51, 13791-13794. 180. Yang, J.; Seto, Y. W.; Yoshikai, N., ACS Catal. 2015, 5, 3054-3057. 181. Kondo, T.; Tsuji, Y.; Watanabe, Y., Tetrahedron Lett. 1987, 28, 6229-6230. 182. Ohtaka, A.; Kato, N.; Kurosawa, H., Organometallics 2002, 21, 5464-5466. 183. Jun, C.-H.; Hong, J.-B., Org. Lett. 1999, 1, 887-889. 184. Jo, E.-A.; Jun, C.-H., Tetrahedron Lett. 2009, 50, 3338-3340. 185. Bedford, R. B.; Coles, S. J.; Hursthouse, M. B.; Limmert, M. E., Angew. Chem.

Int. Ed. 2003, 42, 112-114. 186. Mo, F.; Dong, G., Science 2014, 345, 68-72. 187. Zhang, F.-L.; Hong, K.; Li, T.-J.; Park, H.; Yu, J.-Q., Science 2016, 351, 252-

256. 188. Sun, H.; Guimond, N.; Huang, Y., Org. Biomol. Chem. 2016, 14, 8389-8397. 189. Jun, C.-H.; Lee, H.; Hong, J.-B., J. Org. Chem. 1997, 62, 1200-1201. 190. Jun, C.-H.; Lee, D.-Y.; Lee, H.; Hong, J.-B., Angew. Chem. Int. Ed. 2000, 39,

3070-3072. 191. Marcé, P.; Godard, C.; Feliz, M.; Yáñez, X.; Bo, C.; Castillón, S.,

Organometallics 2009, 28, 2976-2985.