Responsive Molecular Systems Through Dynamic Covalent Chemistry

Home , Inherent chirality, Meso compound

Responsive Molecular Systems through

Dynamic Covalent Chemistry

Antanas Karalius

Doctoral Thesis

LIC OR DOC Thesis Stockholm 2020 Stockholm year

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 21 februari kl 14.30 i sal F3, KTH, Lindstedtsvägen 26, Stockholm. Avhandlingen försvaras på engelska. Opponent är Professor Stefan Matile, Université de Genève (UNIGE), Schweiz.

ISBN 978-91-7873-435-1

ISSN 1654-1081

TRITA-CBH-FOU-2020:9

Universitetsservice US AB, Stockholm

Tėčiui

III Antanas Karalius, 2020: “Responsive Molecular Systems through Dynamic Covalent Chemistry”, School of Engineering Sciences in Chemistry, Biotechnology and Health, KTH – Royal Institute of Technology, SE-100 44 Stockholm, Sweden. Abstract

Nature tends to inspire research in chemistry. Systems that emerge from molecules interacting via reaction networks is something that life has mastered over the course of evolution in order to produce complexity. Dynamic reactions are key in systems chemistry, where reaction networks give rise to complex, emergent behavior. This thesis aims to harness a special feature of selected dynamic reaction systems – responsiveness.

The first chapter of this thesis introduces dynamic covalent chemistry and a general approach to create simple reaction networks by connecting dynamic covalent reactions. Concepts in systems chemistry are introduced in terms of network topology, responsiveness and non-equilibrium processes, while drawing parallels to natural systems.

The second chapter explores the potential of the nitroaldol reaction for dynamic systems. Nitroaldol reactions are demonstrated for dynamic polymerization as well as formaldehyde-responsive breakdown of dynamic polymers. The simultaneous formation and breakdown of polymers create emergent non- equilibrium behavior. Furthermore, nitroaldol produced-diols are used in boronate ester formation. This reactivity produces interdependence over two reactions. Combining nitroaldol and boronate building blocks enabled boronate dynamers of different topology.

The third chapter explores metal coordination effects in dynamic reaction networks. Novel base-free nitroaldol reactivity is exploited in reaction networks with hemiacetals. A systemic response to metals is demonstrated by hemiacetal- metal coordination. In the second half of the chapter, a biomimetic dynamic imine complex is shown to produce emergent, π-π-interactions resembling a “draw-bridge”. Variation of metal charge, effective electrostatic character of substituent and ligands gives control over the system and its emergent π-π- interactions.

Keywords: stimuli-responsiveness, dynamic covalent reactions, systems chemistry, constitutional dynamics, non-equilibrium systems, nitroaldol reaction, dynamers, reversible polymerization, boronate ester, imine ligand, metal coordination, self–assembly, emergent properties, network topology.

IV Sammanfattning på svenska

Naturen tenderar att inspirera forskning inom kemi. System uppkommer från molekyler som interagerar i reaktionsnätverk, något som levande system genom hundratusentals år av evolution har utnyttjat för att producera ständigt växande komplexitet. Dynamiska reaktioner är en nyckelkomponent i systemkemi, där reaktionsnätverk ger upphov till komplexa, emergenta beteenden. Denna avhandling syftar till att utnyttja en speciell egenskap hos dynamiska reaktionssystem - responsivitet.

Det första kapitlet i denna avhandling introducerar dynamisk kovalent kemi och en allmän metod för att skapa enkla reaktionsnätverk genom att sammankoppla olika dynamiska kovalenta reaktioner. Begrepp inom systemkemi introduceras i termer av nätverkstopologi, responsivitet och reaktionssystem långt ifrån jämvikt, samtidigt som paralleller dras till naturliga system.

Det andra kapitlet undersöker potentialen för nitroaldolreaktioner för dynamiska system. Nitroaldol-reaktioner används här för dynamisk polymerisation såväl som formaldehyd-responsiv nedbrytning av dynamiska polymerer. Simultan bildning och nedbrytning av polymer skapar ett emergent icke- jämviktsbeteende. Vidare används nitroaldol-producerade dioler i boronatesterbildning. Denna reaktivitet skapar en sammankoppling av två reaktioner som därmed influerar varandra. Genom att kombinera nitroaldol- och boronat-system öppnas möjligheterna att skapa boronatdynamerer av olika topologi.

Det tredje kapitlet utforskar metallkoordinationseffekter i dynamiska reaktionsnätverk. Den nya basfria nitroaldolreaktionen beskriven i kapitel två utnyttjas här i ett reaktionsnätverk med hemiacetaler. Ett systemiskt svar på addition av metaller uppkommer således genom hemiacetal-metallkoordination. Under andra hälften av kapitlet demonstreras ett naturinspirerat dynamiskt iminkomplex som ger emergenta π-π-interaktioner som liknar en så kallad ”klaff-bro”. Variation av metallernas laddning samt den effektiva elektrostatiska karaktären hos substitutuenterna och liganderna ger kontroll över systemet och dess emergenta π-π-interaktioner.

Nyckelord: stimuli-responsivitet, dynamiska kovalenta reaktioner, systemkemi, konstitutionell dynamik, icke-jämviktssystem, nitroaldolreaktion, dynamerer, reversibel polymerisation, boronester, iminligand, metallkoordination, själv- organisation, emergenta egenskaper, nätverkstopologi.

V Abbreviations d Day d Deuterated Da Dalton DCC N,N′-Dicyclohexylcarbodiimide DFT Density Functional Theory DMSO Dimethyl Sulfoxide DNA Deoxyribonucleic Acid DOSY Diffusion Ordered Spectroscopy EDG Electron-Donating Group equiv. Equivalent Et Ethyl EWG Electron-Withdrawing Group EXSY Exchange Spectroscopy Gb Giga-Base Pairs h Hours kb Kilo-Base Pairs MALDI-TOF Matrix-Assisted Laser Desorption/Ionization Time-Of-Flight Mass Spectrometry MeOD Deuterated Methanol min Minute(s) NMR Nuclear Magnetic Resonance NOE Nuclear Overhauser Effect NOESY Nuclear Overhauser Effect Spectroscopy ppm Parts Per Million RNA Ribonucleic Acid TEA Triethylamine THF Tetrahydrofuran

VI List of Publications

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

I. Formation and Out-of-Equilibrium, High/Low State Switching of a Nitroaldol Dynamer in Neutral Aqueous Media Karalius, A., Zhang, Y., Kravchenko, O., Elofsson, U., Szabó, Z., Yan, M. and Ramström, O. Angew. Chem. Int. Ed., 2020, DOI:10.1002/anie.201911706

II. Rapidly Exchanging, Double-Dynamic, Catalyst-Free Nitroaldol-Hemiacetal Systems for Metal-Responsive Reversible Polymerization Karalius, A., Kravchenko, O., Elofsson, U., Szabó, Z., Yan, M. and Ramström, O. Manuscript

III. Control Over Emergent π-π-Interactions in Double-Dynamic Coordination Complexes Through a Nature-Inspired Coordination-Triggered System Karalius A., Grape, E. S., Inge, K., Kravchenko, O., Szabó, Z., Yan, M. and Ramström, O. Manuscript

IV. Interdependent, Dynamic Nitroaldol and Stereoselective Boronic Ester Reactions for Complex Dynamers of Different Topologies Karalius A., Szabó, Z. and Ramström, O. Manuscript

VII Table of Contents

Abstract...... IV Sammanfattning på svenska ...... V Abbreviations ...... VI List of Publications ...... VII Table of Contents ...... VIII 1. Introduction ...... 1 1.1 Dynamic chemistry ...... 2 1.2 Systems chemistry ...... 7 1.3 Dynamic-kinetic coupled systems ...... 8 1.4 Multi-dynamic covalent systems ...... 9 1.5 Dynamic covalent polymers ...... 9 1.6 Stimuli-responsiveness of dynamic covalent systems ...... 17 1.7 Non-equilibrium systems ...... 22 1.8 The aim of this thesis ...... 23 2. Nitroaldol-coupled dynamic systems ...... 25 2.1 Introduction ...... 25 2.2 Aqueous nitroaldol systems ...... 27 2.3 Boronate-nitroaldol systems ...... 36 3. Metal coordination-coupled multi-dynamic systems ...... 43 3.1 Introduction ...... 43 3.2 Metal coordination in nitroaldol-hemiacetal double dynamic system 44 3.3 Metal coordination-triggered emergence of π-π interactions ...... 55 4. Concluding remarks ...... 67 Acknowledgements ...... 69 Appendix ...... 71 References ...... 72

VIII 1. Introduction

“Life appears to us to be a dynamical process. There is no evolution of individuals. Evolution is a property of populations.”

Manfred Eigen

Living nature is often a source of inspiration for research in chemistry. The evolutionary process creates distinct solutions to unique problems for synthetic chemistry to draw from. Discovery of bioactive molecules, catalysts and materials is quite often rooted in their analogues found in living systems, such as natural products, enzymes and biomaterials exhibiting specialized and explicit properties and functions. Living organisms are capable to adapt to their environment through diversification and selection which are the two key steps in obtaining such new features (Figure 1).1

Figure 1. A. Evolution is often attributed to living systems, however it can originate at a chemical level. It can be understood as a repetitive two-step process involving diversification and selection. B. Protease2 – a peptide sequence that catalyzes the cleavage of peptide bonds C. Hammerhead ribozyme3 – a ribonucleic acid sequence catalyzing the reversible cleavage of phosphodiester bonds of ribonucleic acids.4

On a chemical level, the strategy for accessing complexity has been shown with a simple calculation already by Emil Fischer that 20 different amino acids may form 2.3·1018 different sequences in a 20-mer peptide.5 Normally catalysts such as enzymes require significant physical material to be assembled and, even though can be incredibly selective, are quite bulky by artificial synthetic

1 catalysis standards. However such practice can infact be chemically economical. The requirement to construct an enzyme is, in principal, the formation of one type of bond, in this case an amide bond. Remarkably, a peptide sequence made by a controlled arduous repetition of this reaction gives a unique sequence that acts, for instance, as a protease (Figure 1 B) which in fact catalyzes the reverse process – the hydrolysis of those same peptide bonds it is made of. Surprisingly it is not the only possible sequence to have the capacity do so. Organisms following different paths of evolution can arrive at the same solution via different means, in this case with completely different peptide sequences having identical active sites.6 Peptides are not the only case and in fact discoveries of ribozymes – ribonucleic acid (RNA) sequences that can also act as catalysts for RNA formation and cleavage, had a profound impact, since RNA is known to also act as storage of genetic information in some forms of life. There is speculation in how life on earth came to be, but this discovery gave rise to a popular theory of the RNA world7 which suggests that molecules capable of catalyzing their own formation and undergoing a prebiotic evolution leading to the first living cell could have been strands of RNA.

In the field of chemistry, chemical systems has recently become an active area of research.8-9 It is now titled systems chemistry, the name for the discipline being inspired by systems biology.10 Due to the ability to generate function as opposed to properties, systems are able to give rise to responsiveness, adaptation and emergence through their chemical connectivity network. One of the most important foundations in generation of both natural and artificial systems are dynamic, reversible processes.8

1.1 Dynamic chemistry Historically, the development of sophisticated analytical tools allowed for observing interactions of molecules and led to understanding of supramolecular chemistry – chemistry beyond the molecule. Subsequently, dynamic covalent chemistry, using dynamic covalent bonds, and constitutional dynamic networks (or dynamic systems) emerged which can be seen as subsets of supramolecular chemistry. All reversible interactions of matter serve as the basis for self- assembly, self-organization, responsiveness, feedback and adaptation on a chemical level.11 The idea that supramolecular chemistry is in essence dynamic chemistry, let to the exploration of new types of dynamics and gave rise to dynamic constitutional chemistry, where dynamic covalent interactions are used to generate libraries of compounds, also called dynamic covalent systems.12 These were then used to produce the process of selection via a number of ways. In essence this mimics the single step of evolution, where first diversity is created and the fittest molecule is selected out.13

Figure 2. Dynamic covalent chemistry deals with reversible reactions. It can be used as a cyclic process model. Furthermore a dynamic reaction is the simplest cyclic reaction from a network connectivity standpoint.14

Reversible reactions are quite unique since they are not static. It is a continuous two-way process of making and breaking which can reach thermodynamic equilibrium when these two opposing processes occur at equal rates. Re- cyclability15 not only appears to play an important role but is infact quite central in nature. In terms of network topology, a reversible chemical reaction is the simplest topologically-cyclic reaction. For this reason, dynamic covalent chemistry could be envisioned as a tool for studying cyclic processes (Figure 2).16 Conveniently, spectroscopic methods developed for studying individual molecules can be directly applied to study molecules in mixtures where they are interacting reversibly.16 The interactions can thus be quantified by parameters such as rates and equilibrium distributions.

Reversibility Reversibility can be achieved by different means. In a reversible chemical reaction, typically the dissociation and formation of the bond follows the same minimum energy path over the activation barrier.17 The forward and reverse reactions are then determined by the relative energy barriers in the opposing directions. Subsequently the overall equilibrium distribution is determined by the relative thermodynamic stability of starting material and product. However several irreversible reactions in a network can form a subset cyclic reaction network (Figure 3).

Figure 3. Microscopic (left) and (right) weak reversibility.18 Weak reversibility can occur over multiple steps, which can be reversible. Such larger cyclic processes are often an isolated subcomponent, while standalone macroscopic reversibility is a topic in non-equilibrium thermodynamics.19

Microscopic reversibility can be seen as the simplest case of a cyclic reaction. Cyclic reactions do not have a fundamental limit in how many chemical species

3 are in a cycle, but large cycles have some disfavored constraints in nature. On the other hand, many transition states appear crucial and can be identified in certain reactions. A cycle in such case can be somewhat less ambiguous than an individual reaction arrow or intermediate molecular species in terms of numeric countability. In living systems, reversibility is often in control and cyclic reactions are created under non-equilibrium conditions, consuming energy to drive reactions while enabling the reuse of building blocks through processes such as selective catalytic cleavage.

Self-organization Reversibility is a prerequisite for self-organization such as self-assembly or self- sorting.20 This means that any type of assembly has to be reversible to avoid kinetically trapped states.21 In living systems, lipid-membranes or formation of double helices through multiple combined interactions are prime examples that require microscopic reversibility to reach a global organized state.22

Figure 4. Dynamic process is prerequisite of self-assembly, self-sorting and self- organization which allows for emergence of structures from relatively simple building blocks.20-22

In artificial systems, reversibility has been used to generate dynamic libraries of small molecules, form dynamic polymers and template the formation of macrocycles (Figure 4). Furthermore, highly ordered shape persistent structures such as molecular cages and frameworks can be created, using dynamic covalent bonds. Here, metal coordination is often used to direct the geometry. This strategy is based on coordination numbers and coordination sphere geometry with designed synthetic or self-assembling ligands.23

Dynamic non-covalent interactions Non-covalent interactions are important in nature. The commonly encountered types can be loosely grouped into electrostatic, π-effects, van der Waals forces and solvophobic forces. In organisms this is used for instance in assembly and stabilization of higher order structure of proteins, exploiting different amino acid side chains to produce complementary ion pairing. On the other hand repulsive

4 effects are also important and often a net negative charge is common to prevent aggregation of large macromolecules. The hydrophobic effect is primary driving force between fatty acid chains, enabling the self-assembly of vesicles and membranes.

Figure 5. Important examples of non-covalent interactions. a. Ion pairing, observed in peptides helps to stabilize tertiary structure. b. Stabilizing Interactions between π- systems. c. Hydrogen bonding in a guanine and cytosine base pair.

Hydrogen bonding is crucial in DNA and RNA (Figure 5c), where it acts both for complementary base pair recognition as well as stabilization of the double helix structure in DNA – typically DNA having more guanine and cytosi. Furthermore, aromatic interactions (Figure 5b) between the base pairs act to further minimize the energy of the macroscopic structure in DNA, RNA and in some cases protein. The multivalence effect – multiple repeating interactions extended over space, serves to substantially stabilize the structures while maintaining the reversibility for any dismount or disassembly required for biological processes.17, 24

Dynamic covalent reactions Dynamic covalent reactions, as the name implies, reversibly form bonds that are covalent.25 The selection of most common dynamic bonds can be classified in several ways, but in general, they can be loosely classified into reversible formation and constituent exchange (Figure 6).

Figure 6. General types of dynamic covalent reactions. Reversible formation and constitutional exchange often act together in multi-constituent systems.

Any bond that can form reversibly can, in principal, undergo constituent exchange under the same reaction conditions, if additional reactants are present in the system. Further classification of formation can be according to functional

5 groups into homo/cross-coupling reactions and exchange can be further separated from metathesis. Additionally, isomerizations and transsymmetric reactions are an important class of reactions occurring on the same molecule.26 Classification by dynamic bond into carbon-carbon, carbon-heteroatom and heteroatom-heteroatom is commonplace. Interestingly, the characterization by reaction type is less common.27 For instance addition and condensation reactions differ since in case of addition, reactions are self-contained, while condensation reactions, that form an additional small molecule product, such as water, require it for the reverse reaction (Figure 7).

Figure 7. Commonly encountered representative examples of dynamic covalent reactions.28-29

Thus condensation reactions, such as ester or imine formations, could in-fact be seen as a special case of exchange since two molecules are formed from two initial molecules. Due to different mechanism of formation, the choice between exchange, condensation and addition can be an important factor for material design.27

Other processes There are a variety of physical phenomena that can occur together with dynamic reactions. For instance crystallization,30 evaporation,31 bidirectional interface crossing,32 phase change,33 as well as energy exchange are physical processes that sometimes are inseparable from, and often are crucial in chemical changes of molecules and materials.34

6 1.2 Systems chemistry Traditionally and especially in organic chemistry, challenges revolve around making synthetic molecules or materials. However chemical systems have become a topic of active research over the past few decades.35 The interplay between chemical species in systems gives rise to chemical reaction networks in which molecules can be made, broken up, replicated, repaired and recycled enabling a new level of complexity. Systemic functions such as responsiveness, use of chemical fuel and transient formation are gradually becoming not just interesting properties, but a part of the actual toolbox for developing novel smart materials, in biomedical applications,36 energy and origin of life studies.8 Emergence, by definition, is a property of systems and is not attributable to individual building blocks. The difference to the traditional chemical design approach is that systems are studied and engineered for function rather than properties, where structure can arise not just from atomic architecture but rather chemical reaction connectivity itself. Dynamic networks can be created by stacking robust dynamics exhibiting processes or components into a single multidynamic system.28 Not unlike nature, artificial systems can be created to operate out-of-equilibrium, consuming energy or fuel to transiently create or even maintain the system in certain non-equilibrium states, form dissipative structures or give rise to phenomena such as chemical oscillations.

Topology of reaction networks Chemical systems are not simply explained by the chemistry of the constituents. Networks understood in terms of their topology which can be represented by a network graph.37 In graph theory, a directed graph (or digraph) is a graph that is made up of a set of vertices or nodes connected by edges. When edges have a direction associated with them, they can also be called arrows (Figure 8).

Figure 8. Examples of network connectivity represented by directed graphs: a. Directed graph containing two cycles and an exit node (sink); b. Complete graph K5, where all nodes are directly connected to each other. C. a tree graph with a branching factor of 2.38 Well connected networks exhibit robustness, while the presence of a cycle is a requisite for a chemical system to be self-maintaining.14

Chemical reactions, even in traditional synthetic schemes are already typically represented by molecules which correspond to nodes in a digraph and

7 transformations of them are represented by arrows. Connectivity still arises guided by chemical principles, but the resulting network can be interpreted by the nature or degree of connectivity of the components. There are several levels of complexity in analyzing the network connectivity: relational, stoichiometric and kinetic.39 Different limitations are present for each level. For design implications in stimuli-responsive systems the relational approach can often be sufficient, while for instance modeling kinetics of complex biochemical reactions is computationally demanding. Representative network constituents, such as the arrows and the nodes, the degree of node connectivity, hierarchy, the number of cyclic processes14 of different degrees,15 can all be counted and assigned to the system in order to classify it for general comparison to other systems. The mathematical representation can not only be used for comparison but it also maintains the potential for distinguishing new, emergent processes in complex mixtures.

1.3 Dynamic-kinetic coupled systems When an irreversible process such as a kinetic reaction is coupled to a dynamic process, it gives rise to new phenomena with broad classification scope. Some common examples include switch-on dynamics,40 kinetic selection (resolution) and a case of catalysis (Figure 9).41

Figure 9. Dynamic reaction coupled to a unidirectional reaction gives rise to several phenomenon. Left - switch on-dynamics; middle - selection, also known as dynamic kinetic resolution (DKR); right – catalysis.42 Usually catalytic cycles have varying number of nodes that represent different intermediate species.

When dynamic kinetic resolution is employed in stereoselective synthesis,43 a stereoselective reaction in tandem with racemization allows for chiral resolution yielding up to 100% theoretical conversion. In contrast, resolving a racemic mixture by other means is limited, in theory, to yields of no more than 50%. Unidirectional processes can be of physical nature and systems where reaction- diffusion is coupled can yield emergent systemic properties.44 In graph theory and especially flow networks, the nodes that have associated arrows pointing only toward them, are sometimes called exit nodes or sinks. A system possessing at least a single such node is draining, meaning that it will be resolved over time and will not persist without addition of new starting or intermediate material as represented in Figure 9 middle and right illustrations, making it non-equilibrium system.

8 1.4 Multi-dynamic covalent systems Comparing multi-dynamic systems against one another is not straightforward. However, the number of dynamic processes, that a multicomponent system contains, is a good indication of the nature of its connectivity.14 There are substantial differences in terms of network topology of such systems arising from the use of orthogonal or non-orthogonal processes. In a chemical sense, reactions are orthogonal when they can occur without interfering, while non- orthogonal reactions do interfere. There is some variation in definitions in case of dynamic covalent systems, one being that orthogonality is achieved when a single reaction is occurring at a given time,45 while another approach describes simultaneous reactions taking place independently on different functional groups.46 In general, both approaches have inherent utility, for instance allowing for superior level of control over assembly of functional materials in the first case and accessing more simultaneous diversity in the second. On the other hand non-orthogonality can be utilized as well, especially in designing responsive molecular systems, since interdependence of reactions allows to form network connectivity, in turn enabling to transfer chemical reactivity, such as in a dynamic signaling cascade.47-48 Overall, multiple reactions can occur between both the different building blocks as well as the different functional groups on the same building block. This allows coupling different dynamic reactions and processes into a single dynamic system for creating simple molecular networks. An important case of such networks is dynamic covalent polymerization. The uniqueness of dynamic polymerization systems is the absence of a predetermined reaction network size.

1.5 Dynamic covalent polymers Dynamic covalent polymers are usually categorized by main chain dynamics (Figure 10 a, b) and side chain dynamics (Figure 10 c).49-51 Linear dynamic covalent polymers can be formed from a single self-complementary bifunctional building block.

Figure 10. Common types of dynamic polymers:11, 49-51 a. From a self-complementary monomer; b. From two complementary bifunctional monomers; c. Side chain dynamic polymer.

9 A directional dynamer forms when two functional groups are cross-coupled in self-complementary building blocks. This is also the most common case of polymerization in nature – for example peptides have the amino or carboxy terminals on the linear polymers. If a homo-coupling bifunctional molecule is used, there is no directionality in the resulting dynamer. Similarly, two homo- bifunctional building blocks cross couple to give a non-directional or alternating directional dynamer, analogous to a synthetic co-polymer. Corresponding examples of dynamer systems are presented in Figure 11.

Figure 11. a. Cyclic and open chain polydisulfide;52 b. Hydrazone dynamic co- polymer;53 c. Polylactic acid (PLA) is a reversible ester formation product and a linear directional polymer;5 d. Polyvinyl alcohol dynamic side chain esterification with borate, producing crosslinking after the second addition.54

Furthermore, macrocylization55-58 is a common process, especially if the dynamer is formed via ring opening that occurs through an exchange mechanism such as disulfide exchange between the same-type cyclic building blocks (Figure 11 a). Macrocyclization can occur with linear and directional molecules (Figure 11 c) and can hinder the determination of molecular weight from conversions.5 This holds true to directional and co-polymers.5, 11, 36, 49-51, 59-64 Macrocyclic products can sometimes become kinetically trapped and therefore undesired.65 Having control over the topology of macromolecules can thus be an advantage in dynamic systems. Crosslinking66-69 (Figure 11 d) and branching,70-71 occur when there is a possibility for more than bifunctional reactivity. By extension - frameworks72 are essentially dynamic polymerization

10 in three dimensions with a requisite of building block functionality being greater than two. Also, fixed dimension nano-objects, such as micelles,73-74 can form by dynamic polyfunctional reactivity - in this case the morphology and geometry assist and dictate the outcome into reproducible sized nano-objects. Side chain dynamics occurs when the side groups of a polymer with a static backbone is reacting to form dynamic bonds (Figure 10 c, d). In essence this is a case for polyfunctional reactivity rather than dynamic polymerization however it is an important case of dynamers due to the unique features this difference allows. In addition, the side chain dynamic feature can be used to express or create multivalency to self-assemble materials that have a fixed and predetermined number of dynamic bonds.59 There are important differences in dynamic reaction networks where dynamic polymerization can occur. Typically the reaction network topology for dynamer formation is not described in detail, however qualitative analysis of the reaction network connectivity, as well as the knowledge of the mathematical expressions for growth of the particular dynamer reaction networks can give insights into operating and controlling dynamic systems where the different types of dynamic polymerizations occur.

Terminal node-dynamic oligomerization A linearly propagating dynamic system network is created when dynamic oligomerizations occur from a single bifunctional building block, which adds and cleaves only at the terminus. It could be seen as a version of coupling or stacking of non-orthogonal dynamic reactions rather than network multiplication (Figure 12).

Figure 12. Terminal node-dynamic oligomerization scheme. Any given node exhibits only stepwise reversible formation. A close example of this system is a directional thioester75

A case of n-dynamic, or terminal node-dynamic oligomerization network can be attributed to the formation of biomolecules. For instance template-directed nucleic acid synthesis has the stepwise growth reaction occurring at the terminus, while the reverse stepwise process is achieved by an endonuclease, which cleaves the terminal building blocks one at a time. A synthetic biomimetic

11 biodynamer was recently reported,75 where growth and cleavage were shown to occur through the terminus. Scrambling was also shown to occur if nucleophilic thiol groups are presented in larger numbers, and thus dynamics can be mixed and condition dependent rather than terminal node-type. In terms of network connectivity, a complete dynamic system with nmax of four units would have four species and the number of first degree dynamic processes for the entire system (or the first degree dynamics order) would be equal to three. In a general case, the first degree dynamic order of the system would be N-1. Furthermore, all nodes except for the neighboring ones are weakly connected and the degree of separation between any two nodes is proportional to the distance between those two nodes.

Hyper-dynamic oligomerization A hyper-dynamic oligomerization system can form from single self- complementary bifunctional monomer. The forming chain can be broken at any point to form a smaller building block of any lesser size, and in opposite direction - any building blocks of two sizes can fuse to form a dynamer equal to the combined size (Figure 13). This is a more commonly encountered artificial dynamer than n-dynamic type. In terms of network topology, the dynamic system has every point connected to every other point in both directions (Figure 13), making each node microscopically reversible with every other node in the system.

Figure 13. Hyper-dynamic oligomerization. Each node is accessible from any other one by one single step (blue and gray arrows). In topological terms this system forms 76 n-complete graph set (Kn). Polyester formation from lactic acid is an example of such a system.5

This dynamic polymerization system exhibits much greater connectivity than n- dynamic. In fact, it is completely connected in terms of network topology forming a Kn graph set. Size nmax = 4 creates a system exhibiting six first-order dynamic processes and in a general case system of n sized molecules would have N(N-1)/2 first order dynamic processes. It is worthwhile noting that “mixed” dynamics are also possible in systems. For example, the templated synthesis of

12 nucleic acids occurs stepwise and when coupled with a rapid endonuclease scission of the main chain, would give a stepwise growth process together with a “hyper-” reverse process. On the opposite note, if a peptide chain stepwise growth is coupled with an exo-peptidase, which cleaves dipeptide fragments from the terminus, it would give a hypo-dynamic system: mixed system with dynamic order lower than n-dynamic.

Dynamic co-polymer analogues Copolymers are much more prevalent in artificial chemistry. The terminal- dynamic co-polymer system forms in a reversible cross-coupling reaction of two complementary homo-bifunctional components that react in a stepwise fashion, similarly to earlier terminal node-dynamic case (Figure 14 blue arrows). In the general case first-order dynamics of the system is 2·(N-1), which is twice the size as for a single building block terminal-dynamic system. Similarly to the hyper-dynamic oligomerization, a dynamic co-polymer formation also exhibits a type of hyper-connectivity, but with the addition of the emergence of macroscopic reversibility (Figure 14 blue and gray arrows). In contrast to the hyper-dynamic oligomerization case, this system is not completely connected, and the symmetric building blocks of type BAB and ABA are not directly connected by a reaction.

Figure 14. Hyper-dynamic co-oligomerization scheme of two complementary bifunctional building blocks. The complementary nodes ABA and BAB, representing symmetric molecules, are weakly connected to each other. An example of this system is hydrazone dynamer formation by reversible condensation of dhihydrazide and dialdehyde building blocks.53

The hyper-dynamic oligomerization system has first order dynamics of N(N-1). It also exhibits weak reversibility through formation of nodes ABA and BAB which are not directly connected. For these nodes to interconvert it takes at least two reactions. Example of such a system is a hydrazone formation (Figure 14 right).

13 Side chain-dynamic polymers Side chain dynamics is a case often attributable to dynamers. In reality, reversible reaction of a polymer with monofunctional building blocks, creates a dynamic system of fixed size, predetermined by the static length of the main chain (Figure 15). Thus such systems generate networks attributable to polyfunctional building blocks and the dynamics of the system is not a function as in earlier cases, but rather a network of fixed size. This is solely dictated by the length, or more precisely, by the number of functional side groups present on the backbone. Side chain dynamic 3-functional building block with a directional backbone, has a fixed dynamic network consisting of eight nodes and an order of 12 according to first-degree dynamics (Figure 15).

Figure 15. Reaction network of a side chain-dynamic tri-functional building block with a complementary mono-functional molecule. This type of network is of fixed size, depending on the number of functional groups on the starting polymer. An example system is the side group dynamics of chitosan.77

Such type of dynamic networks are created, for instance, in an artificial template-directed synthesis, where building blocks are dynamically assembled on a chain. The subsequent step is often a covalent derivatization to form a static complementary strand in artificial systems.78-80 On the other hand, multifunctional compound dynamic systems give better defined network topology81 because of having defined valency in comparison to a potential size distribution in case of static oligomers and polymers. As mentioned, the system does not exhibit growth and has a finite dynamic order predetermined by the main chain length. The backbone symmetry is also important, for example when the backbone is non-directional, a tri-functional building block with all functionalities being equivalent results in three stepwise consecutive reactions

14 in comparison to twelve reversible reactions with eight nodes, as is the case in Figure 15.

Biomacromolecules and biodynamers As mentioned earlier, nature exploits a strategy that allows for great diversity and complexity to arise economically from a chemical synthesis sense: 5- nucleobases to a phosphodiester backbone, and 20 amino acids connected through an amide bond is most of what it takes to make large macromolecules with a variety of functions. The varying nature of eukaryotic genome82 coupled with transposable elements83 and mutation rates which are progressively increasing while going downward in size84 could indeed serve the conceptual idea that although molecules such as proteins,85 prions,86 DNA and RNA are not “dynamers” in current understanding, they could indeed be considered as dynamic (Figure 16). In the more general sense, the ability to reverse the same reactions on demand allows to recycle the initial building blocks after the macromolecule has served its purpose.

Figure 16. Reversible cleavage of RNA87 catalyzed by ribozymes4 and acid-catalyzed depurination.24

Mutations – a type of natural constitutional dynamics, allows for diversification of a genome.88 The diversity of a genome that can be generated is extremely large,6 but not necessarily infinite as the hypothetical network topologies might imply. Extremes are encountered to natural molecules. For instance one of the largest known peptides is Titin, reaching up to 30k amino acids in length,24 while arguably, the largest known genome of 670 giga-base pairs (Gb) belongs to Polychaos dubium.82 The smallest genome for a living cell is a target of active field of research named the minimum genome project. Currently, the record holder is a synthetic organism named syn3.0 with a completely synthetic genome of 531 kb.89 Viruses can have as little as 250–400 nucleotides for plant Viroids, which are essentially plain RNA strands exploiting insects as vectors

15 for spreading as well as all the replication machinery from their hosts. “Spiegelman’s monster” was a result of experimental replication of a phage under error prone conditions.90 It is a strand of RNA that gradually evolved and exhibited fast replication - a strand of RNA with 218 base pairs. This is an interesting Drakes law illustration, where larger genomes are observed to have lower mutation rates.84 From a network topology standpoint, there is little difference between a network populated by dynamic covalent polymerization and a stationary population of quasispecies obtained by error prone replication.6 Extrapolating Drakes law downward to molecules such as oligomers, high mutation rate91 perhaps would appear like dynamic constitutional chemistry.92 In such case, going the opposite way and progressively stabilizing88 a dynamic oligomerization process could serve as a key to adaptive chemistry and subsequent bottom-up directed chemical evolution93 where the most persistent constitutional preference information of the system is gradually inscribed in a sequence by diminishing dynamics.94 Such an approach has potential for investigating the information entropy threshold by artificial means. In a recent by Otto group, a self-replicator molecule emerged from a dynamic covalent system. At its core, it had dynamic covalent disulfide bonds and was able to undergo diverisifcation.95 Artificial macromolecules, such as peptide nucleic acids, chemically modified DNA sequences as well as synthetic amino acid analogues bearing peptides, have been a target of applied and fundamental research.96 A novel strategy in systems chemistry is biodynamers – acting as artificial macromolecules capable of self-assembly and mimicking nucleic acids, proteins and polysaccharides (Figure 17).

Figure 17. Examples of biodynamers. a. Linear directional thioester;75 b. peptoid;34 Side group dynamics of chitosan.77

The goals of studying such dynamic systems and networks include development of templated synthesis and assembly,78-80, 97 creating artificial self-replication as well as understanding cooperative behavior and interactions between macromolecules,93, 98 with one of the ultimate goals being the elucidation of the origins of life. A variety of strategies are employed,99 varying from thermodynamically and kinetically controlled systems to exploiting non- equilibrium conditions.100

1.6 Stimuli-responsiveness of dynamic covalent systems Stimuli-responsiveness plays a vital role in life. The ability to respond to the environment and its changes is essential for survival of living organisms. For instance, vernalization is the ability of a plant to respond to prolonged cold. This allows to trigger flowering after winter hibernation.101 Phototaxis is light- triggered change of directionality of an organism observed in plants and bacteria.102 In artificial chemical systems, dynamic covalent reactions have inherent stimuli-responsiveness associated to them (Figure 18).

Figure 18. Possible strategies of harnessing inherent stimuli-responsiveness of dynamic covalent systems. Each dynamic covalent reaction has unique chemistry that can give rise to responsiveness to unique chemical stimuli.36, 49, 53, 62, 74, 103-108

Artificial chemical systems allow for novel materials that can change their physical or chemical nature with an externally introduced trigger. This research is gaining traction in biomedical applications such as controlled drug release, as well as manufacturing novel “smart” materials capable of changing shape or consistency.103 Furthermore, chemical signaling cascades allow for signal mediation in sensing applications. For example signal auto-amplification can give further increase in responsiveness, providing greater design possibilities.109- 111

Physical The control over a dynamic system by physical means is sought after, since it makes it easy to control such systems by external manipulation (Figure 19).

Figure 19. Selected illustrative examples of redox,36 light112 and thermo- responsiveness113 of dynamic covalent bonds as well as the type of systemic perturtbation

For example the ability to electrochemically reduce and oxidize back to a disulfide bond allowed for creation of a system, in which upon reduction, a dynamic association of thiolate with CO2 to form a thiocarbonate occurs. This is reversed upon oxidation of the thiolate back to the disulfide allowing to 40 control CO2 capture and release by the flip of a switch. Another example is diselenide exchange triggered upon introduction of a light source.112 Incorporation of such bonds into polymer structures allows for molecular weight change through generation of block copolymers upon dynamic exchange of different size building blocks. Size control can also be achieved via temperature controlled dynamic oligomerization of hydrazones.113 Growth of the dynamer is triggered by heat while chain braking occurs upon cooling the system.

Chemical Chemical control of the direction of existing dynamic systems is usually based on inherent chemical properties of functional groups present in the building blocks, or functional groups appearing in the product after formation of dynamic bonds (Figure 20). In the most general case, introducing a new direction for the system always changes the previous equilibria of the initial system.

Figure 20. Chemical responsiveness of selected dynamic bonds: a. change in pH,114 b. Metal coordination115 c. Competition with monofunctional building block116 d. Kinetic selection of a constituent,117 cat. = chiral catalyst.

The native properties of the dynamic reaction can be useful in designing systems with desired responses. For instance imine formation is inherently responsive to pH (Figure 20 a), and allows for fine-tuning either building blocks or the actual pH of the system in order to control the overall equilibrium. In some sense, the protonated amine can be considered as a new equilibrium product in a dynamic imine equilibrium, and the control of this new competing equilibrium can determine the ratio of free (protonated) amine and imine product.114 Several different responsive materials have been designed based on the inherent pH responsiveness of imines.104, 118-119 Alternatively, a new equilibrium process can be introduced to the system, resulting in amplification of a component in the dynamic mixture. For example metal coordination can induce amplification of hemiacetal molecule (Figure 20 b). Additional equilibria can change material properties such as in the case of imine coordination hydrogel. In a particular example,116 a competing monofunctional building block served to dissociate crosslinking of a diamine in a hydrogel which changed the overall mechanical properties of the gel from shape persistent to flowing. Nitroaldol reactions have proved as useful synthetic reactions, and their prochiral nature can be further utilized in case of dynamic kinetic resolution. In a common example, one of the reversibly forming beta-nitroalcohol diastereomers is captured by stereoselective acylation by an enzyme (Figure 20 d). Interestingly, nitroaldol reactions appear to be somewhat underrepresented in dynamic systems and materials.

Physicochemical In some instances, a physical signal can be transformed into a chemical and vice versa (Figure 21). Materials capable of such function can be useful in chemical

20 systems since, they allow for mediating between physical and chemical processes and can provide the systems a much broader applicability scope.

Figure 21. Reversible light triggered isomerization and cyclisation of a photoacid by light releases a proton (top).120 Switch on fluorescence upon reversible formation of boronate ester with a diol is pursued in sugar sensing applications (bottom).121

Pairing a chemical mediator to a dynamic system can allow to change a physical effect into a chemical one. This enables chemical-responsive systems to be triggered by physical stimuli, as was shown by combining self-assembly of a clatrotetrachelate cage with a photoacidic mediator in a system in which the generated acid triggered the disassembly of the cage. Since the acidity of the photoacid disappears when light is discontinued, the cage is able to re-assemble in the dark.120 On the opposite end, fluorescence is significantly milder effect and is useful for sensing applications. An ortho-aminomethylphenylboronic acid with an N-substituted fluorescent label showed an increase in fluorescence upon formation of a boronic ester. The mechanism of fluorescence increase appears to have been elusive. It was recently clarified to be a “loose bolt effect” where fluorescence is quenched by O–H bond vibrations, capable of accepting the excitation energy. Quenching is diminished by boronate ester formation resulting in increased fluorescence.121

Stimuli-responsiveness effects in multi-dynamic systems Similarly to a simple dynamic equilibrium, in a multi-dynamic system, all the components are connected, thus perturbation of one component or concentration perturbs the entire system. An important implication of this is that it allows for transfer of chemical properties through the chemical reaction network connectivity. To a similar extent, changes in one equilibrium affects the equilibrium positions in the entire system; however, it is important to consider

21 the connectivity in designing such systems. For instance, kinetic processes have a strong effect on the entire system since a sink node is introduced, creating a unidirectional flow for the system.122 On the other hand, well-connected networks exhibit systemic property of robustness (Figure 22).123

Figure 22. Illustrations of multi-dynamic system perturbation with introduction of stimuli as a new reaction pathways. a. Introduction of a new equilibrium to a well- connected (robust) system; b. Introduction of a unidirectional process to a lesser connected system.

This means that in order to get a sizeable response effect, the system requires a stronger signal due to ability compensate. This resistance capacity arises from possible alternative reaction pathways. Furthermore, kinetic effects start playing a role and thus slow reaction pathways can essentially be neglected in certain systems and have to be considered early on in design steps.44 Nevertheless kinetic effects can also be exploited through controlled catalyst activity giving temporal control over reaction rates.124 Kinetic signal amplification can be achieved by modular enzymatic cascade networks, which are controlled through strong enzyme catalyst inhibitors, where one enzyme cleaves the inhibitor of another.109

1.7 Non-equilibrium systems When a unidirectional process such as a kinetic reaction is introduced to a dynamic reaction network of a system, the equilibrium capacity of the system disappears and the system switches into a non-equilibrium state. It will eventually resolve to thermodynamically stable products. In order to maintain the system, additional energy125 or matter in the form of fuel has to be supplied, together with the removal of the kinetic product to prevent buildup that would eventually interfere with the system. If supply and removal of the kinetic product in the system occurs at equal rates, the system can achieve a non-equilibrium steady state,126 which unlike a thermodynamic equilibrium has to be actively maintained (Figure 23) and can be regulated by changes in rates of the kinetic reactions.127

Figure 23. A multi-dynamic non-equilibrium system is created when linear processes (kinetic asymmetry), such as consumption of fuel, are introduced into a dynamic system.127 The dotted line indicates boundary of the system and orange arrows indicate addition and removal of reagents and products (waste) to and from the system. A non-equilibrium system is created regardless of which product is added or removed from the system (gray arrows). A dynamic system in the flow of energy promotes self-organization and moves in the opposing direction of entropy increase inducing dissipative-adaptation to stimuli.128

Physical addition and removal of material (e.g. fuel) as well as energy can also create a system that operates outside of the equilibrium.129 If addition and removal rates match, the system is said to be operating at a steady state.126 This appears to be in close resemblance to living systems, where a steady state of an organism is called homeostasis. In other cases when the system is an out-of- equilibrium state, the intermediate product formation is temporal and is said to be transient. An overall linear kinetic process in which fuel or energy is consumed, creates energy dissipation through the system.130 An interesting implication is that in a multi-dynamic system (Figure 23) there is no inherent difference which intermediate is supplied and which is removed - the system would maintain the non-equilibrium condition regardless as long as the kinetic asymmetry condition is maintained. Non-equilibrium systems allow for dissipative self-assembly127, 131-133 and polymerization,126, 129 chemical- oscilators,134-135 materials130, 136-137 and systems with tunable lifetime,138 machineries that can operate as well as drive the system out-of-equilibrium.125, 139-141 Furthermore, natural closeness of non-equilibrium systems to living systems which are operating far-from-equilibrium, allows to elucidate and eventually borrow function such as adaptation, self-replication and autonomy from nature.142

1.8 The aim of this thesis The aim of this thesis is to expand the scope of coupling non-orthogonal processes by investigating dynamic reactions that are underrepresented in chemical systems. Broadly, the strategic approach explored is based on pairing non-orthogonal combinations of dynamic covalent reactions to create simple

23 chemical networks, with the aim of exploiting, transferring, mediating and altering the different stimuli-responsive properties of select dynamic covalent bonds. The first chapter introduced dynamic covalent reactions as a tool to generate dynamic systems. Drawing parallels to systems found in nature, strategies for applications in artificial systems and materials were described.

In the second chapter, nitroaldol reactions under neutral aqueous conditions are studied. A dynamic copolymer based on nitroaldol reaction is made using nitromethane as a bifunctional building block. This dynamer is further investigated for its responsive properties, as well as transient out-of-equilibrium formation modes. Furthermore, the dynamic nitroaldol systems are extended to include an interdependent boronate formation. This shows to have an inhibitory effect to the rates of nitroaldol reaction in both directions, stabilizing the system against degradation by action of base. By coupling nitroaldol building blocks with boronate formation, a number of dynamers with different topologies can be produced, including side chain and main chain-dynamic type as well as crosslinking.

In the third chapter – metal coordination coupling to dynamic systems is investigated. Firstly, nitroaldol building blocks, exhibiting catalysis-free reactivity are studied for systems using metal coordination. It is demonstrated that in a nitroaldol-hemiacetal coupled system, metal coordination to hemiacetal species downregulates the nitroaldol reaction, through network connectivity. This new-found reactivity is also extended to nitroaldol dynamers, which form spontaneously upon dissolution of building blocks and their size being initial concentration dependent. Furthermore, downregulation of the dynamer size is demonstrated by metal coordination. Lastly, in a nature-inspired coordination system, emergent π-π-interactions are demonstrated in bis-meridional complexes with zinc and gallium. The π-π-interaction is shown to be induced by metal coordination, while number of π-bridges was controlled by changing opposing ligand as well as metal charge in the coordination complex, allowing for future extensions to redox-responsive systems.

24 2. Nitroaldol-coupled dynamic systems

(Papers I, IV)

“An investigator starts research in a new field with faith, a foggy idea, and a few wild experiments. Eventually the interplay of negative and positive results guides the work. By the time the research is completed, he or she knows how it should have been started and conducted“

Donald J. Cram

2.1 Introduction Dynamic covalent reactions enable a large scope of chemical systems34, 50, 79, 100, 124, 143-145 and materials.55, 60, 74, 107, 146-147 Currently, this covers self- assembling,108 self-healing,51, 148 stimuli-responsive56, 75, 106-107, 149 materials for a variety of applications. Furthermore chemical reversibility in multi-dynamic systems allows to explore complex behavior.109, 150-151 Systems have been design to create self-replicators,93, 152 chemical oscillators153 and out-of-equilibrium operation,139, 154 which can introduce functions such as tunable lifetime into materials.138 Often a single dynamic covalent bond can be the basis for emergence of function in a system, with each type of bond enabling a whole domain of unique complexity and function attributable to its chemistry.45, 139, 147, 155-162 At a bare minimum, it allows for responsiveness to molecules, bearing functional groups that are homologues to ones making the dynamic covalent bond, i.e. an imine bond can be broken by exchange with both aldehydes and amines. The unique chemistries of dynamic covalent reactions enable transfer, inhibition or emergence of new systemic properties through reaction pathways connected into networks.44, 163 Thus, it makes it an advantageous strategy to integrate different dynamic covalent bonds into systems and explore new systemic effects.

Due to its stereogenic nature, simultaneous nitro- and hydroxyl-functionality introduction and a reversible C-C bond, the nitroaldol (Henry) reaction is quite unique amongst a limited number of available dynamic covalent reactions that are commonly used.12, 28-29, 36, 49-50, 55, 62, 103, 105, 164-166 Nevertheless nitroaldol reaction appears to be underrepresented in systems chemistry. It is essentially a self-contained addition reaction and does not require the product to react with additional molecules to reverse the bond formation, contrasting to condensation

25 reactions such as imine formation. Therefore, dynamic covalent systems enable research into stimuli responsiveness and adaptation with the intent of addressing important challenges in chemistry, biology and materials sciences.166-170 Water is an especially attractive media for tackling these challenges. Furthermore, it can serve as the prerequisite for dynamic exchange required for self-assembly,21 and in the most direct case - dynamic oligomerization. Dynamic covalent polymers bear inherent responsive nature which makes them well-suited for a variety of applications including drug delivery as well as creating smart, cyclable materials capable of self-assembly or self-healing.51, 145, 147, 165, 171 Recently, the complex behavior between oligomers having complementary functionalities that allow multivalency effect-driven assembly of unnatural peptide strands have been demonstrated.60, 79, 145 Dynamic supramolecular and covalent polymers are gaining a lot of attention, especially in terms of bio- inspired functions.98, 130 For example the ability to introduce multiple exchanging functionalities to a macromolecule with as little as one type of dynamic bond50, 75 is an approach that closely resembles the strategy living of systems. While reversibility of certain chemical bonds, such as amides,132 typically requires continuous chemical energy input, biodynamers are a good example of a workaround approach for trying to replicate the complexity of natural systems using bonds that are dynamic at ambient conditions.172 Even though the toolbox of dynamic covalent chemistry is gradually expanding,45, 161, 173 a handful of bonds maintain popularity for specific types of applications. For instance, boronic acids are excellent in terms of chemo-selectivity towards diols and thus have continuously been investigated for recognition,174-176 as well as formation of various structures, e.g. macrocycles177 and cages.146 Due to their utility in synthesis, a variety of boronic acid building blocks are commercially available, making boronate chemistry attractive for creating and evaluating dynamic systems.106, 178 Evolution, which is suggested to emerge at a chemical level,179 can be seen as a two-step process, requiring diversification and selection. Polymeric macromolecules are excellent in this regard, since only several initial building blocks enable exponential number of different configurations. Similarly to natural oligomers,75, 180 hypothetical binary dynamer systems, where a single unit/bond introduces two configurations, provides access to over 215 possible configurations of unique sequence dynamic molecules, with as little as 17 repeating units. If stereoconfiguration is ever to be used for diversification, the nitroaldol reaction is one of the few C–C bond forming dynamic covalent reactions, combined with the introduction of β- nitroalcohol functionalities which can be achieved in potentially excellent stereochemical purity and in high yields.181-186

26 2.2 Aqueous nitroaldol systems The nitroaldol reaction has been studied for synthetic applications in water-rich solutions to some extent. Reports include the use of different catalysts,187-190 while stereoselectivity can be greatly enhanced in some examples by coupling with lipase-mediated resolution processes.117, 191-193 The reaction has not been explored for dynamic chemistry in aqueous media however, other than in our preliminary studies.194 To our knowledge the formation of dynamic covalent polymers (dynamers) or frameworks have not been shown earlier as well.

Non-equilibrium systems chemistry has seen a lot of progress in recent years.8, 93, 126, 131-132, 137-139, 153-154, 195-199 It enables new function in chemical systems, as for example transient self-assembly. Non-equilibrium systems are classified by the nature of formation and dissipation. Sorrenti et al129 suggested that when a state of a system requires fuel or mass exchange with environment to maintain that state, it is a non-equilibrium state. Furthermore - a transiently forming structure in such system is dissipative. Kinetic asymmetry was proposed to be prerequisite for a system to be non-equilibrium,127 meaning that fuel is used to drive the system into a less thermodynamically favored state.200 A large interest remains in non-equilibrium oligomerization due to the direct relevance to biological systems,7, 93, 132, 136, 198, 201-203 which tend to naturally exploit oligomeric species. A non-equilibrium state that exists by means of consumption of energy/fuel is transient (short-lived) if the energy/fuel is limited. Artificially maintaining a steady, out-of-equilibrium state for prolonged time is not trivial and can require elaborate experimental set up.126

pH influence on reaction rates

Figure 24. Water soluble building blocks pyridine-2-carboxaldehyde 1 and nitroethanol 2 were employed to make a model dynamic nitroaldol system in aqueous buffer.

With the aim of achieving dynamic nitroaldol oligomerization in aqueous media, we employed two model systems for the elucidation of essential reaction parameters. First, reaction rate and pD dependence was established in the initial model system study of 2-nitroethanol 1 and pyridine-2-carboxaldehyde 2

27 (Figure 24). Several different pD values in 0.4 M phosphate buffered aqueous solutions were used.

Figure 25. Model nitroaldol reaction of 1 and 2 at different pD values.

Formation of nitroaldol product 3 (Figure 24) was rapid in all cases with a lower comparative rate at pD 6.0. Equilibrium constants were estimated to be >500 M- 1, indicating greater thermodynamic stability of the product vs. the starting materials. The hydrated aldehyde species appeared at a relatively constant level throughout the reaction. At pD = 7.4, the forward rate was 2.9 M-1 min-1 and χ50 (50% reaction advancement degree/equilibration degree)204-205 was reached in approximately 6 min. This rate is greater than some other reversible covalent reactions under similar conditions, e.g. hydrazone/oxime formation,206 as well as more rapid than reported in earlier nitroaldol reaction studies in organic solvents.117, 192

di-Henry dynamic systems For the subsequent studies with nitromethane, pD 7.4 was selected as the optimal pD, having excellent equilibration rate while displaying good systemic stability over prolonged period of time. In comparison to the reactivity of 2-nitroethanol, nitromethane 4 can react twice, acting as a bifunctional molecule. This leads to multiple simultaneous equilibria (Figure 26). For this reason, equilibration was studied from both the substrate- as well as the intermediate/product directions, initiating the dynamic system from compounds 1 and 4, intermediate 5, or product 6 (Figure 26).

Figure 26. Nitromethane 4 reacting as a bifunctional molecule with pyridine-2- carboxaldehye 1.

The buffer was proposed to be the catalyst acting as a general base. The concentration of the buffer was decreased to increase accuracy by decreasing the rate and minimizing formation of the trace amounts of the deprotonated product species that were observed in the system. The deprotonated species decreased to negligible amounts upon decreasing the buffer concentration from 0.4 M to 0.1 M. Observation of both of the species simultaneously can likely be attributable to the nitroalkane anomaly.207-208

Figure 27. Different possible diastereomers of 6, and their corresponding 1H NMR signals as observed in the dynamic system after equilibration and H-D exchange.

A hydrogen-bonded six-membered transition state with a water molecule has been proposed to induce syn-stereoselectivity in aqueous media.187 This preference would suggest that the syn,syn-producing reaction could have the faster formation rate, but the subsequent isomerization between the meso- isomers is likely to be rapid as well, thus the basis for the faster formation is not sufficient to establish the absolute configuration in the present case even though one of the meso-diastereomers appears dominant at equilibrium (Figure 27, Figure 28).

29 100

Conversion, Conversion, % 20

0 1/4 5 1/4 6 (1:1) (2:1)

Figure 28. Conversions at equilibrium in different starting scenarios: 1 and 4 (1:1 and 2:1 ratio), 5 and 6 . Gray bar indicates fraction of aldehyde 1, blue bar indicates fraction of mono-Henry product 5, orange bar indicates di-Henry product 6. Conversion is calculated according to aldehyde functional groups. Maximum conversion to mono-Henry product in case of starting with a 2:1 ratio of 1 and 4 as well as 6 is 50%.

The different systems (Figure 28) were monitored by 1H NMR. Distributions appeared to be similar between the complementary pairs of experiments. (Figure 26, Figure 28). The equimolar 1 and 4 dynamic system resulted in a higher -1 amount of nitroaldol product (K1: 4500 M ) compared to the result with 2- nitroethanol (K: 1020 M-1). This effect is creditable to the nitromethane having lower steric hindrance in comparison to the secondary nitrocompound 2. It is interesting to note that one meso-isomer of product 6 dominated in aqueous media, while the corresponding non-meso compound formed exclusively in organic solvent. The non-meso diastereomers are in fact a racemic pair (Figure 27).

Nitroaldol dynamer The reaction between building blocks 7 and 4 was then addressed to establish dynamic polymerization (Figure 29). Reaction was followed by 1H NMR under neutral aqueous conditions, resulting in significantly broadened peaks. This hindered the determination of the extent of the reaction. However at and above 40 mM of initial concentrations, precipitation of the polymeric adduct gradually occurred and the formed dynamer 8 could be isolated by filtration.

Figure 29. Formation of nitroaldol dynamer under neutral aqueous conditions based on the bifuntional reactivity of nitromethane 4 and pyridine-2,6-dialdehyde 7.

Since precipitation occurs due to a threshold in molecular weight formation, dynamer formation could be easily upscaled without affecting the average size distribution. Significant quantities of dynamer 8 could therefore be obtained for further characterization and analysis. MALDI-TOF mass spectrometry showed peaks up to 3 kDa in weight, with the major peaks being 196 Da apart, corresponding to a repeating single unit of dialdehyde/nitromethane adduct. Some minor peaks appearing at the same frequency are attributable to the different possible combinations of terminal groups, such as aldehyde/hemiacetal or nitroalcohol. The highest observed signals corresponded to approximately 15 repeating units at around 3 kDa. Additional lower-weight signals could be attributed to multiple ionizations and fragmentation in the ion trap. Diffusion NMR (DOSY) experiments in THF-d8 were used to additionally estimate the molecular weight by correlating molecular weight to diffusion constant. When used with a reference of known mass, the correlation is expressed as Dsol/Dref = 1/3 (Mref/Msol) where Dsol is the diffusion constant for the dynamer and Dref in this case is the diffusion constant recorded for dimer 6 as an external reference under the same conditions (Figure 30).209 The diffusion constant of the oligomer corresponded to approximately 3.3 kDa representing an oligomerization degree of ca. 17 repeating units, as obtained from the signal decay fitting. The diffusion constant is calculated for spherical species according to the equation. The variation of the diffusion constant can therefore occur on solubilization and self- folding for the dynamer. If larger sizes were to be investigated, these parameters would have to be additionally checked using different solvents.

Figure 30. Pseudo 2D-DOSY NMR plot for the dynamer 8 (left) and di-Henry product 6 (right) in THF-d8.

The stereopreference of the model system allows to anticipate similarirties in the multiplied configurations over the length oligomer chain. Solvent effects observed in model systems enable the possibility to change between the racemic-like and the meso-like configurations, albeit both size control and extent of stereopreference would have to be re-established in cases for the dynamer.

Stimuli-response studies When the initial building block concentrations were kept below 20 mM, no precipitation occurred during the equilibration. After approximately 6 h, the aldehyde signals decreased to 3.5% of the initial values and no further change was observed over several days. The mixture appeared stable and this allowed studying the stimuli-responsiveness of the dynamer. The inherent responsiveness of the nitroaldol systems to both aldehydes and nitro compounds was tested for perturbation or reversal of the equilibrated dynamer system (Figure 31).

Figure 31. Stimuli responsiveness of the dynamer 8 to monofunctional building blocks: Nitroethanol (left) and formaldehyde (right).

Formaldehyde was chosen as the stimulus for evaluating the nature of stimuli- responsiveness of the dynamer. Remarkably, no formaldehyde-responsive materials appear to have been studied, even though the molecule is of biological significance.210-211

Figure 32. NMR spectra of formation and formaldehyde triggered breakdown of dynamer 8. Bottom – starting mixture of aldedhyde 7 and nitromethane 4, middle – dynamic polymer at equilibrium, top – dialdehyde 7 after formaldehyde-driven disassembly of dynamic polymer.

The control reaction between nitromethane and formaldehyde appeared to be slow. However, high conversion (approximately 95%) to nitroethanol and its subsequent addition products (Figure 31) was reached within 8 days. Building

33 on this result, 10 equiv. of formaldehyde were introduced to an equilibrated system of the dynamer 8. Perturbation of the equilibrium over 4 days resulted in re-emergence of the dialdehyde 7 at 97% of the initial mixture (Figure 32). The experimental results demonstrate that recycling of dialdehyde can be achieved using formaldehyde. When 10 equiv. of 2-nitroethanol was introduced, a new equilibrium reaching 89% conversion to the dinitroaldol product was achieved within 4 days, and remained stable over the next 4 days.

Transient polymerization in non-equilibrium systems Differences in nitroaldol product rates and stabilities in the competing reactions gave an indication that an out-of-equilibrium oligomer formation could be achieved in a reaction network (Figure 33).

Figure 33. A non-equilibrium system created when formation of oligomer is occurring together with formaldehyde induced breakdown.

NMR was used to follow the progress of the system with signal broadening being a qualitative indicator of change of the system between high (polymeric) and low (monomeric) states (Figure 34). Three separate experiments were set up. In the first case, 1 equiv. of nitromethane, 1 equiv. of dialdehyde 7 and 10 equiv. of formaldehyde were introduced simultaneously. This resulted in a harsh environment where no substantial oligomerization occurred. The nitromethane was substantially digested by formaldehyde before any larger oligomers could form. Upon subsequent repeated addition of nitromethane and formaldehyde the same result was observed. In the second case, 2 equiv. of nitromethane were used, leading to significant oligomerization. According to the model system, this should not have been the case in the absence of formaldehyde due to saturation of the dialdehyde (Figure 28). This situation is indicative of transient oligomer formation, where the system operates out-of-equilibrium when the produced waste products are inert and the building block ratio is maintained in the process. The behavior of the system can be seen as emergent since a transient balance is achieved through interplay of two processes. In the third experiment the system was pre-equilibrated to a higher oligomeric state with 1 equiv. of nitromethane, but without formaldehyde. After apparent equilibrium, an additional 1 equiv. of nitromethane and 10 equiv. of formaldehyde were added simultaneously. This

34 noticeably perturbed the system. However a significant delay occurred before the system moved toward the lower molecular weight extreme. In this case, the excessive nitromethane acted to protect and maintain the system at a high state.

Figure 34. Qualitative investigation of non-equilibrium dynamer 8 formation under three different scenarios. Yellow line indicates transient “High” molecular weight state. A) When 1 equiv. of nitromethane 4 is introduced to a solution of 7 in presence of excessive formaldehyde. B) When 2 equiv. of nitromethane 4 is introduced under same initial conditions as A; C) When 1 equiv. of nitromethane 4 together with excessive formaldehyde is introduced to a pre-formed dynamer 8 system; D) Snippets of 1H NMR spectra of the system at “High” and “Low” molecular weight states.

According to classification mentioned in the introduction, the system forms the oligomers transiently because of the discontinuous experimental set-up, with nitromethane acting as fuel in the formaldehyde-rich system. Interestingly, in these conditions both aldehydes react fairly rapidly with free nitromethane. However, if the dynamer is already pre-formed, it seems to suppress the presence of any free nitromethane and therefore slow down the digestion process. In the model system (Figure 28), the formation of the competing dimer and monomer has kinetic “overshoots”, which can be observed for intermediate products. This is suggestive of a kinetic and thermodynamic disparity between stereoisomers. If this effect is also present in dynamer formation, this implies that the dissipative systemic state, which is continuously recycling the transient larger dynamer, could have an effect on the distribution of preferential

35 stereoconfigurations on the main chain. It could suppress the thermodynamically preferred stereoconfiguration, which is likely to lead to some meta-stable chain configurations under normal thermodynamic control. Furthermore, the upregulation of the kinetic effect might, in principal, allow access to more of the stereodiversity. This could occur by minimization of both kinetic and thermodynamic traps through the continuous mass reflux, enabling higher expression of alternative configurations on the chain.

The difference observed by changing the starting point for the experiments suggests that creating a steady state for this system would require the content of the nitromethane to be in some excess. Reaching the steady state would likely require continuous addition of nitromethane, and the addition rate would be a function of oligomer size as well as free nitromethane/formaldehyde concentration ratio. Overall, the higher size-state requires lower concentrations of free nitromethane, which would in turn slow down the digestion of the excess, resulting in the protecting effect. Furthermore, depending on the desired molecular weight-state, the system might require prolonged reaction time to reach a steady state if the starting point of the system is the low initial state and could likely be shortened by starting from the high molecular weight state.

2.3 Boronate-nitroaldol systems Building on the results on novel nitroaldol dynamers described earlier in this chapter, we wanted to exploit the 1,3-diol functional groups forming in the nitroaldol reaction with nitromethane and 2-nitroethanol in dynamic boronate formation. Boronate formation was first established between model substrates 1 and a in different solvents by 1H NMR (Figure 35). As described earlier, 1 is a double-addition product of 2-pyridinecarboxaldehyde to nitromethane.

Figure 35. Boronate ester 1-a formation from di-Henry product 1 and phenyl boronic acid a in different solvents.

Boronate ester 1-a appears favored in all tested aprotic solvents, while methanol disfavors its formation. 10% v/v of D2O in acetonitrile reverses the equilibrium to starting materials. Consequently, acetonitrile was chosen as the most universal solvent, exhibiting good solubility of both starting materials, as well

36 as displaying reasonable stability for 1 in solution over prolonged time, being only slightly worse than THF-d8.

Coupled nitroaldol reaction and boronate ester formation As the aqueous conditions were unsuitable for boronate formation, we could not use the aqueous buffer employed in the initial study (see section 2.2.2). Therefore basic conditions for nitroaldol reaction were investigated using TEA in acetonitrile in three separate experiments (Figure 36).

Figure 36. Investigation of non-orthogonal nitroaldol and boronate formation reactions in systems using TEA as base. A) Equilibrium position for nitroaldol reaction, 1 + TEA; B) 1-a boronate with TEA; C) Nitroaldol reaction (1 + TEA, 12 h) from mixture A after 12 h incubation (starting point) treated with phenylboronic acid a.

In the first case, the equilibrium direction for 1 by addition of TEA (Figure 36, A) was examined. It showed that the di-Henry product broke into the mono- Henry product and an equivalent amount of aldehyde almost immediately. It is clear that the second step of the nitroaldol reaction is unfavored under these conditions. In the second case (Figure 36, B), in situ pre-formed boronate was treated with TEA. Interestingly, almost no retro-nitroaldol reaction was observed. In the third case, addition of phenyl boronic acid a into a pre- equilibrated mixture of 1 and TEA was checked for any influence on di-Henry product formation and overall equilibrium (Figure 36, C). As anticipated, some emergence of 1-a was gradually noticeable, albeit seemed to be slow. Overall, trace signals of 1-a were observed in case C as well as aldehyde signals in case B, gradually appearing over the course of 12h. This shows that the equilibration rate of nitroaldol reaction is inhibited in both directions. Regardless of the slow rates, a selection of di-product is in fact observable. This means that the new equilibrium introduced into the dynamic system serves to increase the expression of di-Henry product 1 by an observable margin.

37 Stereopreference The di-Henry product of nitromethane, which is reacting as bifunctional molecule with pyridine-2-caboxaldehyde has three stereocenters, and in turn enable two meso diastereomers and a pair of diastereomers that are enantiomeric in regard to each other (Figure 37). During the nitroaldol reaction, as discussed in the earlier part of this chapter, stereochemistry is dictated by solvent effects. This is the determining factor for diastereomer distribution at equilibrium. In an aqueous system the one of the meso-diastereomers formed as a trace compound, while the other two diastereomers are present in near-equimolar amounts at equilibrium, with a slight preference for the major meso-diastereomer. Synthesis of the di-Henry product 1 in organic aprotic solvent exclusively yields the racemic diastereomer. The boronic ester of the three isolated compounds have different couplings on the 2-nitropropan-1,3-diol- bridge: 1-meso* is the minor meso-diastereomer formed at equilibrium in aqueous media. Only a single peak that corresponds to overlap of 3 H signals appears after the boronate 1meso*a formation. The disappearance of spin-coupling strongly suggests that the two equivalent protons next to the hydroxyl groups are close to 90° angle with the proton on the nitro-substituted carbon when the boronate cycle formation locks the configuration. Furthermore, the other 1mesoa compound shows coupling between the protons on the same position.

Figure 37. Stereopreferential boronate ester formation between 1 equiv. of 1 mixture with equal ratios of diastereomers, reacting with 0.33 equiv. of phenylboronic acid a.

The stereopreference of different boronate formations was investigated under equimolar conditions for the three diastereomers of 1. (Figure 37). The dynamic boronate formation showed a preference for racemic boronate, followed by the minor meso* diastereomer, while the major meso diastereomer appears to be the least-favored.

38 Additional studies were done with a selection of boronic acids (Figure 38). The system was simplified for this stereopreference study by omitting the minor meso* diastereomer and only the two major diastereomers were investigated. For this experiment, 2D COSY NMR was employed since the boronate signals of the two diastereomers form either 2 or 3-spin systems. It means that formation of peaks belonging to symmetric 2 spin or asymmetric 3 spin systems are attributable to meso and racemic diastereomers accordingly, except for two noted cases, where in the mixed system the corresponding ratio was derived from free diols (Figure 38).

Figure 38. Boronate esters forming from different boronic acids (a-g) and equimolar a b mixture of 1-rac and 1-meso compounds. Ratio of 1mesoa/1raca. Ratio derived from unreacted starting diols. cNo reaction observed, due to limited solubility of boronic acid d.

It was established, that a stereopreference of approximately a 2:1 ratio was common for the boronates regardless of the boronic acid used, with some deviations. No inversion of stereopreference was observed. Generally, 1raca forms preferentially, with an approximate 2:1 ratio.

Boronate dynamers of different topology In the formation of boronate ester-bearing dynamers several unique dynamer topologies can be envisioned (Figure 39). Combinations of mono-, bi-, and polyfunctional building blocks were investigated in terms of the forming dynamer size. DOSY NMR experiments yielded molecular weights (Figure 39 iv) which were obtained by correlating the diffusion constants with external reference standards of similar compositions and known molecular weights (as described in chapter 2.2.3).

39 i.

2-m ii. iv. Repeating unit* 2-p Dynamer right curve 3-m

3-p iii. label 3-a

3 Mw, Da

0 2,500 5,000 7,500 cross-link fold Figure 39. i. Coupling diol-bearing substrates with diboronic acids enable a boronate ester based-dynamers of different topology, molecular weight and properties. The Bis- Henry product 2 produces two sets of diols for boronate formation while oligomer 3 contains multiple reactive sites proportional to chain length ii. Schematic illustration of different possible topologies between bis-Henry product 2 and di-boronic acid building blocks p and m; iii. Schematic illustration of labeled, crosslinked or folded oligomer 3 arrangements by coupling to boronic acids a-m; iv. Molecular weights obtained from DOSY NMR experiments (see section 2.2.3). *Repeating unit mass calculated by taking single smallest repeating unit assuming complete diol and boronic acid conversion to boronate esters.

Interestingly, comparing of the molecular weight of phenyl boronic derivative 3-a to its corresponding nitroaldol dynamer 3 revealed that single-unit weight- increase for the boronate is smaller than the molecular weight obtained from the diffusion constant correlation. This discrepancy in weight increase suggests two possible effects: 1) additional aggregation or 2) unfolding of the dynamer by extension lengthwise due to increased rigidity of the backbone upon boronate derivatization. The increased rigidity would give a lower value, since the diffusion constant correlation is presumptive of spherical species. It does not account for diffusion change going from more packed to more unpacked states for the oligomer. Additionally, two boronate dynamers 2-p and 2-m are formed between the corresponding bifunctional building blocks. The geometry of the backbone (Figure 39, ii) is dictated by the diboronic acid groups on the phenyl ring. Despite this difference, the resulting dynamer molecular weight is similar in this case. The dynamer 3 mixtures with m and p diboronic acid building blocks lead to noticeable aggregations and a substantial increase in molecular weight in both cases, evident by significant NMR signal broadening. This

40 broadening results in higher error margins. Nonetheless, the increase of the molecular weight of 3-p appears to be smaller than for 3-m. This could be due to reactivity differences of di-boronic acids (i.e. p-diboronic is known to have a higher tendency to self-condense), different modes of crosslinking or even due to the absence of crosslinking due to mono-saturation of more reactive dynamer sites leading to some extent to presence of “label” topology (Figure 39 iii).

To sum up, we established the scope for the dynamic covalent boronate ester formation for diastereotopic 1,3-diols, which are formed in nitroaldol reactions. Boronate ester formation appears to be favored in most aprotic solvents. The presence of boronate ester on the diol inhibits its native favorable retro- nitroaldol reaction, substantially stabilizing the diol in the presence of base. A possible application for this effect is “AND” logic gate for a stimuli-responsive system. In this gate nitroaldol dynamics would be activated if two conditions are met: 1) boronate cleavage (e.g. by competing diol); 2) presence of a base catalyst. The stereopreference of boronate formation was elucidated for diastereomeric nitroaldol diols together with strong evidence supporting the absolute configuration of meso-diastereomers used in this study. The systems were extended to topologically-distinct boronate functional group-bearing dynamers. Furthermore, the formation of boronate esters on the nitroaldol dynamers appears to be influenced by conformation, which in accord to initial boronate stereopreference study with di-Henry products. Moreover, boronate ester formation changes the diffusion constant slightly unproportioned to the molecular weight change. Diboronic acid building blocks, with both meta- and para-substitution, enabled main chain formation of nitroaldol-boronate dynamer. Similar molecular weight dynamers were formed in both cases, while permitting variation differences in geometries. Additionally, nitroaldol dynamer interaction with diboronic acids leads to crosslinking of dynamer strands and substantial increase in the molecular weight. Fold-like geometries by intramolecular bind are also possible in this case.

Several future research directions can be envisioned based on the results of this study: 1) Investigation of sequence-selective cleavage for the nitroaldol dynamer by action of a base and a boronic acid in tandem. 2) Functionalization by complementary-bonding substituents; 3) Extensions to stimuli-responsive materials for biomedical applications.

42 3. Metal coordination-coupled multi-dynamic systems

(Papers II, III)

“Now OK, it works also, in less than 170 °C glycol, but if you want to finish your PhD in a reasonable timespan - you better heat it!”

J. M. Lehn

3.1 Introduction Novel reactivity that enables seamless, catalyst-free and spontaneous generation of a dynamic system is highly sought after in self-assembly,155, 212 dynamic covalent chemistry as well as adaptive31, 213 and responsive molecular systems.64 One of the most straightforward approaches for generation of such systems is coupling dynamic processes together. This enables extension of the responsive property of a dynamic cluster to the entire system, e.g. through a dynamic signaling cascade,47 or in more simple cases – through reaction network connectivity. Combining dynamic covalent bonds with other interactions such as π-stacking, metal coordination and diffusion allows for discovery of new complexity and function of materials 32-33, 45, 155-159, 173. Dynamic covalent systems using metal coordination receive a lot of attention32, 139, 146, 156-157, 159-162, 173 for a good reason: metal-ligand interactions can be used to control structural assembly with precision as well as induce response to mechanical systems. This is due to the predictable nature of coordination bonds and their spatial arrangements guided by metal coordination number, ligand denticity and rigidity.

In nature metal coordination plays key roles in photosynthesis, cellular respiration and numerous enzymatic functions.214 Sometimes metal complexes appear in unexpected places – for example, about 10% dry weight of a dipicolinic acid complex which is reported to be present in certain bacterial endospores was proposed to be implicated in preferential intercalation to DNA, binding to RNA as well as drying-stabilization of the species. The stabilization of RNA by various dynamic modes in the medical and prebiotic setting is openly being discussed7, 215-216 however, interactions with possible natural coordination complexes is not often considered in this regard. On the other hand

43 oligonucleotide conjugates with metal complexes are studied for a variety of interactions.217-228 In biomedical setting, metal complex intercalation is typically studied either for therapeutic219, 222 or labeling functions.222, 229-230 Conjugate complexes are used for DNA supramolecular and nano-technology224, 231 as well as sequence specific DNA cleavage96 which is suggested occurr through some level of metal complex intercalation. Both the mechanism and extent of coordination-driven dynamic assembly through electrostatic π-π-interactions and intercalation in natural nucleic acid systems appears unrepresented in literature. Although electrostatic properties232 and modes of dynamic (covalent) stabilization of RNA in the rough conditions of the hypothetical RNA world have been proposed,7, 233 there seems to be no overlap with the primitive chemistry exploited by organisms that are extremely resilient and long-lived in nature.234 Dipicolinate-related natural structures are known to help with shuttling of different metals in bacteria,235 and substituted pyridines such as pyridoxal,236-237 seem to be a prime suspect in pre-metabolic processes possibly involving metal coordination. It is therefore important to investigate natural substituted pyridine intercalation mechanisms, especially in cases that involve metal coordination.

3.2 Metal coordination in nitroaldol-hemiacetal double dynamic system In this study, the possibility of combining nitroaldol reactions with metal coordination in a single system was addressed. A common limitation of nitroaldol reactions is the requirement of base catalysis. In a continuation of work involving out-of-equilibrium systems exploiting nitroaldol reactions under neutral aqueous conditions,238 a nitroalkyl compound that exhibits tautomerism was noted (Figure 40). Interestingly, the proposed nitronate form is known to be the reactive form in the Henry reaction with aldehydes.

Figure 40. 2-(Nitromethyl)pyridine 1a and its tautomeric nitronate form.239

The pursuit of this system was carried out in methanol, because of its similarities to water, superior solubility of building blocks and mitigation of acidic metal aqua-complex formation, in consequent steps. Realizing a system for the nitroaldol reaction with 2-pyridinecarboxaldedehyde 2 was initially

44 investigated. The aldehyde forms hemiacetal 4 upon interaction with methanol as a second dynamic process in the system (Figure 41).

Figure 41. Nitroaldol reaction between 2-nitromethyl pyridine 1 and pyridine-2- carboxaldehyde 2 to form nitroaldol product 3 as well as aldehyde 2 equilibrium with its hemiacetal 4.

A rather unusual kinetic behavior was observed for this system (Figure 42 top). An overall “overshoot” is observable, which then equilibrates to the thermodynamic equilibrium. This “overshoot” typically occurs approximately 10–15 times earlier than the perceived systemic equilibrium is reached. This feature is further noted as “peak conversions” (Figure 42).

Figure 42. Aldehyde 2, hemiacetal 4, Henry product 3 (A, B, - are two observavable diastereomeres, 3A being the higher-shifted) distributions during the model nitroaldol reaction of 1 and 2 in d4-MeOD (top) and control reaction (bottom) where aldehyde- hemiacetal is pre-equilibrated by incubating in methanol before the start of the Henry Reaction.

To elucidate the cause, the multiple-equilibria effect was investigated first. For this, a control experiment was performed where aldehyde 2 is pre-equilibrated in methanol to form 4 before the addition of nitroaldol reaction (Figure 42 bottom). The reaction displays a similar overall kinetic profile both before and after pre-equilibration. Also, supramolecular interactions cannot be easily dismissed due to the high substrate/product display of hydrogen bond donating and accepting groups. This observable perturbed reactivity cannot be fitted using simple kinetic models; however, the characterization of these multiple equilibria systems using the observed peak conversion was used as a convenient time point

46 in recording reaction progress. Noting the similarities to the nitromethylene compound 1, ethyl nitroacetate 1a was also investigated for its capacity to undergo spontaneous Henry reactions. (Table 1).

Table 1. Comparison of active nitromethylene-bearing building blocks in double dynamic systems.

1 1Duration 2 3 3 4 4 A:B ratio 9.8 h** 58% 36% 0.58 1 80 h 44% 47% 0.47 47 min** 56% 29% 0.76 1a 12 h* 45% 47% 0.53 1Duration noted for different criteria: * when apparent equilibrium reached; ** at peak Henry conversions; after 80h, when equilibriation rate diminished. 2Overall conversion to Henry product (all diastereomers and nitronate species if present). 3Percent conversion to hemiacetal 4. 4Ratio between two observable sets of diastereomers at noted time: A – higher ppm shift diastereomer, B – lower ppm shift diastereomer, absolute configurations not established.

The nitroacetic acid showed good reactivity with comparable equilibrium distributions, albeit also displaying a kinetic overshoot. Similar results between the two building blocks indicate that this could be a feature of the system under these conditions. Good reactivity permitted to expand the scope of the system to dialdehydes which can be employed for nitroaldol reactions as well. Furthermore, work by Drahoňovský et al115 showed that hemiacetal formation is promoted by introduction of a coordinating metal, such as zinc (Figure 43). In that work, amongst various alcohols, methanol was also shown to form hemiacetals that were amplified by zinc coordination.

Figure 43. di-Hemiacetal formation and zinc coordination.

Both hemiacetal and nitroaldol products have a similarly furnished coordination sphere. This led us to question how a double-dynamic nitroaldol–hemiacetal system would behave upon introduction of metal ions. Unfortunately, the nitroaldol reaction kinetics were more difficult to plot and follow by NMR

47 spectroscopy due to multiple species displaying overlapping signals when dialdehyde 5 is used in contrast to monoaldehyde 2. Instead it was possible to estimate the overall conversion of the initial aldehyde to nitroaldol, as well as the hemiacetal products by % functional group conversion (Figure 44).

Figure 44. Dynamic nitroaldol system between 2-nitromethylpyridine 1 and dialdehyde 2. Reaction followed by NMR. 3* (Blue line) – conversion to nitroaldol, product by functional group conversion, 6 (red line) – hemiacetal.

Interestingly, even though there are substantially more intermediate equilibria that occur when a dialdehyde 5 is used (Figure 45), the conversions seem to peak in a similar time frames to the monoaldehyde 2 case (cf. Figure 42).

Figure 45. Dynamic nitroaldol system between 1 and dialdehyde 5. i. Multiple equilibria resulting from nitroaldol reaction and hemiacetal formation. ii. Extended dyanmic system - upon addition of zinc, dihemiacetal complex is amplified by complexation.

48 After establishing the dynamics for the system, zinc was introduced for coordinating ligands. In the presence of zinc, the amplification of dihemiacetal is quite clear from the 1H NMR (Figure 46). DOSY NMR experiments, the separate species could be further confirmed.

Figure 46. Systemic response to 3 equiv. Zn2+ addition. Bottom: equilibrated mixture; top – 4h after addition of 3 equiv. Zn2+. Blue color indicates free 1 signals, red - free 6 and coordinating 6-Zn. Pre-equilibrated mixture at a 2:1 ratio of 1 and 5 (left).

From these results, it appears that di-hemiacetal component is amplified more than five-fold from the initial mixture in under 4 h under consumption of nitroaldol products (Figure 47).

Figure 47. Systemic response to 3 eq Zn2+ addition. Pre-equilibrated system of 5 and 1 response to 3 equiv. of Zinc chloride. Red line 6 indicates content of dihemiacetal

49 by functional group and blue line 3 indicates nitroaldol overall conversion by functional group.

The system was further investigated using 1, 3 and 5 equivalents of zinc chloride to examine if the degree of the response would change (Table 2). The dihemiacetal component 6 gets amplified up to fives times in most cases, to maximum of 57% conversion. The overall conversion for the nitroaldol reaction drop dramatically in correlation to the excess used, from 71% at initial equilibrium to 22% in cases of using 5 equivalents. Time to peak conversion also decreases substantially with larger excess, suggesting a kinetic effect to the selection. These results indicate that zinc coordination makes the hemiacetal thermodynamically more favourable due to complexation, giving rise to change in product distribution in the double-dynamic system.

Table 2. Peak conversions and their change upon introduction of Zn2+ at different equivalents for dialdehyde 5 systems with 1.

Entry Conversion1 6 Time Equilibrated sample i 71% 9% 14 h 1 equiv. ZnCl2 46% 52% 7 h 3 equiv. ZnCl2 ii 38% 52% 3.7 h 5 equiv. ZnCl2 22% 57% 1.8 h 1Noted conversions are at maximum conversion change (cf. Figure 47). Conversions calculated as fraction by functional group. 2Overall nitroaldol conversion is calculated by functional group conversion which is corresponds to mixture of mono and di-Henry products.

To force the system to regenerate after Zn2+ addition, tetramethylammonium oxalate was used to scavenge the metal. Oxalate successfully precipitates zinc from the methanol solution, although the system does not regenerate fully and remains perturbed. This perturbation could in part be due to temporal pH fluctuation as well as the presence of additional species that are introduced i.e. the counterions changing the overall chemical environment. After complete removal of the Zn2+ ions, the system does not appear the same due to changes in some shifts of the NMR signals, but nonetheless the overall nitroaldol conversion seems to gradually return to the conversion values at the initial equilibrium distribution before metal addition, approx. 63% in comparison to 65% at the start of the experiment.

Figure 48. Cycling of the system into different states using metal coordination to suppres nitroaldol reaction and upregulate diacetal and oxalate was used to precipitate the zinc out of methanolic solution, restoring the initial state that can subsequently re-equilibrate.

To further diversify the system we were interested in investigating the combination of dialdehydes and dinitromethylene compounds. Although aldehydes, dialdehydes, and nitroacetic acid mono-esters are commercially available building blocks, bifunctional activated nitromethylene molecules require additional synthetic efforts. Direct nitration, as was employed in the synthesis of 1 from 2-picoline, on more than one methyl group has been reported to be difficult on a single pyridine scaffold.239 The reactivity of nitroacetic acid ethyl ester was suggestive of evaluating a nitroacetic acid diester (Figure 49), which can be prepared by reacting nitroacetic acid with dry ethylene glycol under mild conditions and employing a mild dehydrating agent such as N,N'- dicyclohexylcarbodiimide (DCC).240 Free nitroacetic acid is sensitive and tends to decarboxylate upon action of nucleophiles, especially in the presence of activating agents.240 However the mild reaction conditions however allow for the synthesis of polyfunctional nitroaldol building block 7 (Figure 49).

Figure 49. Ethylene di-nitroacetate bifunctional nitroaldol building block 7 and the nitroacetic acid model 1a.

After esterification, the diester appears to be bench-stable. Diester 7 was tested qualitatively with mono-aldehyde 2 in methanol and appears to have similar

51 reactivity to ethyl nitroacetate 1a. Overall, the peak conversion of diester 7 to nitroaldol products reached 53% in comparison to 56% of 1a (cf. Table 1).

Control of size in dynamic oligomerization Extending the system with dinitromethylene 7 to include dialdehyde 5, should provide extended chains of both starting materials after consequent Henry reactions. Indeed, oligomerization occurred spontaneously when complementary bifunctional building blocks were used, forming dynamer 8 (Figure 50). The system behaves similarly to the previous cases with monofunctional and bifunctional reactants, in terms of apparent rate and product distributions after equilibration.

Figure 50. Spontaneous formation of dynamer 8 from bifunctional building blocks 5 and 7.

For the size of the forming dynamer 8 the determining factor is the conversion. Average molecular size and conversion interdependency5 can be expressed 1 according to Carothers equation: 푋̅ = where p is conversion. Despite the 1−푝 similarity between the model monofunctional building blocks and oligomers the conversion parameter could not be pinpointed for oligomer systems due to significant signal overlap and broadening of the NMR spectra. Instead, oligomer size can be probed using DOSY NMR by establishing the diffusion coefficient (D) for molecular average of dynamer 8, and compare the obtained D value with a reference compound with similar chemical properties, and known molecular weight. The diffusion constant can be correlated to molecular weight according

퐷 3 푀 to equation 푠표푙 = √ 푟푒푓 209. The diffusion coefficient is solvent parameter- 퐷푟푒푓 푀푠표푙 dependent. Changes in viscosity, especially in the higher molecular weight range are not accounted for and thus the values are indicative without additional experimental confirmation.

Figure 51. Average size of dynamer at equilibrium on initial concentration of bifunctional building blocks 5a and 7a. The molecular size estimated using DOSY experiments using the diHenry product formin in 5a and 1a system under the same conditions as external reference. The unit size of the dynamer corresponds to a repeating conjugate unit comprised of 5a+7a.

Dynamic oligomerization experiments indicate a correlation between molecular weight increase and building block concentration (Figure 51). The slope of the y axis provided the size in base 2 logarithmic scale. This gave an apparent linear trend for the weight increase of systems with different initial concentrations measured after equilibration. The data points range from 20 mM to 320 mM. It is worth noting that at high concentrations, a change in viscosity would have to be considered for larger dynamic oligomerization systems.

This result confirms that dynamer size can be regulated by changing the concentrations of building blocks. The change in conversion is more difficult to see than in Table 4 due to signal drift and overlap in pre-equilibrated dynamer system (Figure 52). The next target was to investigate if the dynamer, which is a Henry product, is also downregulated by metal coordination. Qualitative response is similar to cases with dialdehyde 5a and 2-(nitromethyl)pyridine 1, where the conversion of nitroaldol product appears to decrease, while amplifying the diacetal 6 (Figure 52). The overall conversion cannot be exactly determined but the disappearance of broad signals at ca 7.5 ppm as well as 4.2 ppm in the NMR spectrum indicates consumption of 8. Indeed, the formation of the hemiacetal product is clearly observed by enhancement of the NMR signals at ca 8.2, 7.9 and 5.8 ppm.

On the opposite note, dihemiacetal formation reaches 25% upon Zn2+ addition.

Figure 52. Systemic response to 3 equiv. Zn2+ addition of a pre-equilibrated mixtures of equimolar 7a and 5a at 40 mM initial concentration. Bottom – preequilibrated dynamer system, top – system after addition of zinc. Blue signals indicate free 7a, while free and zinc coordinating di-hemiacetal component 6a signals are makred red. 6a gets amplified to 25% (top) from the trace in the equilibrated mixture (bottom).

The interaction of the dialdehyde 5a – diester 7a system with zinc showed that the response of diester is less than that of the 2-nitromethyl pyridine 1 system with 5a. This result indicates that even though the systemic metal coordination response originates from hemiacetal coordination, the differences in competing molecules 1 and 7a creates systems with differing properties. As a result, this leads to variation in response magnitude, regardless of similar equilibration in model cases.

In summary, the unprecedented reactivity of electron-withdrawing group- bearing nitromethylene building blocks was demonstrated. These building blocks spontaneously create dynamic systems at ambient conditions and without external catalysis. This dynamic nitroaldol system allowed for assessing the assessing systemic responses upon metal coordination. It was demonstrated, that a double dynamic nitroaldol-hemiacetal system formed from 2,6-pyridine dicarboxaldehyde and 2-nitromethyl pyridine as well as nitroacetic acid ester building blocks in methanol is responsive to metal coordination. The dynamic behavior is incurred through coordination and upregulation of dihemiacetal

54 species, perturbing the equilibrium of the entire system and significantly reversing the nitroaldol reaction. This system can be recovered by subsequent removal of coordinating metal ions, through oxalate coordination. These unique features could be utilized in dynamic oligomerization, where size control can be achieved by simple regulation of the constituent concentrations. In addition, the dynamer size could be downregulated through metal coordination. This newly discovered reactivity of reported nitroaldol substrates can find utility for building up and breaking down complexity in dynamic covalent systems exploiting metal coordination.

3.3 Metal coordination-triggered emergence of π-π interactions For effective DNA intercalation, π-π-interactions play a major role. In this work models for ligand-metal coordination effect on π-π-interaction are investigated. The study is inspired by a natural dipicolinate complex, which is suspected to stabilize the DNA in certain organisms by intercalation. The primary question addressed in this study is whether π-π-interaction, being the prime interaction in intercalation, could be triggered by metal coordination. The naturally occurring dipicolinate and dipicolinate-based metal chelators were investigated (Figure 53). Methanol was chosen for this study since it is a polar protic solvent, similar to, but more practical than aqueous solutions, where the more hydrophobic molecules as well as imines are investigated.

For the model study, Zn2+ and Ga3+ were chosen, since both metals tend to form well-defined octahedral complexes (Figure 53), are redox stable and diamagnetic, enabling rapid characterization by NMR spectroscopy. The similarity between these two metals and iron at +2 and +3 oxidation states, would also enable direct extrapolation of the results to a redox responsive system in the future.

Figure 53. Similarity between the building blocks A and B. Both building blocks can act as tridentate ligands taking meridional (mer) configuration in octahedral coordination complexes (right).

Initially the effects of electron donating (EDG) and withdrawing substituents (EWG) were compared, changing the effective electrostatic character of the flexible and unconjugated aryl group (Figure 54).

Figure 54. In situ formation of imine ligands Am and Af.

Initial observations upon introduction of 1 equiv. zinc triflate to 2 equiv. of in situ formed Am ligand displayed a notable downfield shift of the pyridine unit (Figure 55 red color) and up-field shift of the entire trimethoxybenzyl fragment signals (Figure 55 blue color). The up-field shift in NMR spectra is typical for functional groups in proximity to an electron-rich system, such as an aryl fragment, due to electron shielding. Since the benzylic group is freely flexible over the methylene bridge, homo and hetero interactions are possible, i.e. the substituted benzyl groups can be interacting with one another or with the pyridine π-system.

Figure 55. NMR of in situ generated ligand Am (top) and the observed shielding effect from a 2:1 interaction with zinc2+ ions.

To further verify the presence of π-π-interactions NOESY NMR experiments were performed. The results revealed cross-relaxation between the pyridine ring protons and methoxy substituents on the benzylic group (Figure 56 a). The methyl protons at the 3’ and 5’ positions showed a NOESY cross-peak with all protons in the pyridine ring while the 4-MeO- showed most intense cross-peaks

56 with the proton next to the carboxyl group of the pyridine ring. On the other hand, the cross-peak with the para-proton of the pyridine is negligible.

The shielding effect observed by the 1H NMR spectroscopy together with the NOESY experiment suggests that the electron-rich aryl group stacks with the more π-deficient pyridine in a parallel offset fashion. Both of the benzyl protons show cross-peaks, which indicates that the two planes of the system could be moving over the conjugated π-system of the pyridine, giving a single average peak (Figure 56 b).

Figure 56. a) 2D NMR spectrum of the AmAm-Zn complex indicating the NOE cross- peaks of the “lowered-bridge” and EXSY exchange cross-peak of the aryl group in the “raised-bridge”-position, labeled bridge flip; b) structure illustration indicating the spatial proximity of the fragments (circled) in lowered-bridge arrangement giving rise to NOE cross-peaks c) Crystal structure of AmAm-Zn complex.

A feature of the NOESY spectra is that the EXSY cross-peaks appear from dynamic processes, such as the chemical exchange between functional groups. In this case, the raised and lowered bridge-state cross-peaks of the aryl group (Figure 56 a). Exchange cross-peaks can result from a few other dynamic processes such as imine or free ligand exchange. Unfortunately, the raised- bridge complex signals are poorly resolved from the free ligand due to similar chemical shifts and are thus hard to quantify. However, the imine part of the raised-bridge system seems observable from an unshielded trace imine signal observed downfield, which likely belongs to this complex. Over prolonged time the complex tends to crystallize from d4-methanol. A high-enough purity crystal was withdrawn from an NMR tube for X-ray diffraction experiments. The crystal structure of the corresponding complex also showed a preferential stack

57 of the two ligands (Figure 56 c), in accordance with DFT-predicted geometry (Figure 57 a) and the results from NMR spectroscopy. The two aryl groups are locked above opposite parts of the picolinate, one appearing centered over the carboxy group and the other over the imine. The non-equivalence in π-π- stacking modes of the two identical ligands is a unique feature of the present complex. In order to understand why the aryl groups are aligned differently, π- π-interaction energies between trimethoxyphenyl moiety and the different sides of picolinate were evaluated computationally. Despite the fact that interaction with the carboxylate side of picolinate was found to be favored in isolated ligands (12.33 kcal/mol vs 9.25 kcal/mol or imine side stacking), analysis of conformations of zinc complex with dangling aryl groups revealed that the energy gain upon stacking with the imine part of picolinate is larger (7.4 kcal/mol vs 5.0 kcal/mol for carboxylate side stacking, Figure 57 b). It appears that when one trimethoxyphenyl moiety is aligned over the imine, it sterically blocks this stacking mode for the second trimethoxyphenyl moiety. This steric hindrances forces the trimethoxyphenyl moiety to align on the carboxylate side of the picolinate (Figure 57 c). Since the opposing zinc-pyridine bond angle is below 180°, the sides become inequivalent. It is possible that this is a steric, electrostatic or crystal lattice-preferential configuration or a combination of these effects. Nonetheless, this can be also explained by strong π-π-interaction effects which are comparable with the Zn-imine bond energy241 and thus can overcome the ideal octahedral configuration. An asymmetry in π-π-effects as discussed above leads to asymmetry in the ligand coordination, representing an uncommon example of direct influence of second coordination sphere effects on the first coordination sphere.

Figure 57. a) Geometry-optimized structure of AmAm-Zn complex, H atoms omitted for clarity. b) Overlaid optimized structures of AmAm-Zn and its stable conformers with rotated aryl groups, H atoms omitted for clarity. c) Induction of disfavored stacking mode via sterical hindrance by preferred stacking mode.

To see the influence of electronics on the π-π-interactions, the more electron poor trifluorophenyl derivative was used. This trifluorophenyl derivative is expected to present a less favorable stacking with the electron-deficient pyridine ring. The initial imine ligand formation from the starting fluorinated benzylamine Af (Figure 54) appeared slower and less favored than in the case of the trimethoxy substituent, indicating that there is an inherent limitation to the use of withdrawing groups for in-situ formed imines in this system. Upon introduction of 1 equiv. zinc triflate to 2 equiv. in situ formed Af ligand also produced a similar shielding effect, however, in this case the mixture gradually produced a precipitate. Dilution was required to inhibit precipitation. The dynamic self-organization process is indeed slow in this case and, while in the previous case equilibration was near immediate, the dilute system took overnight to equilibrate. The NOESY spectrum produced multiple cross-peaks between shielded and deshielded benzyl group, which could indicate intermolecular interactions. In AmAm-Zn complex configuration, which is similar to AfAf-Zn, this was not observed and the protons of the aryl groups would be fairly distant in the closed arrangement. Overall, the interaction is favored but association has a prolonged kinetic effect both due to hindered imine formation as well as diminished electrostatic effects.

59 Next, the factors affecting the π-π-interactions in these systems were further investigated, by creating a mixed system with dipicolinate ligand B (Figure 58). In particular, the change in net charge of the stack-accepting ligand in the bridge- stack was investigated for any possible changes in stacking preference.

Figure 58. Structures for mixed mer-ligand system exisiting in equilibrium with the homo complexes and their charge and number of π-π-interactions in mixed ligand coordination systems with zinc2+.

The NMR spectra of the mixed system was further processed to enhance the signals of the mixed AB systems (Figure 59). The separate spectra of homoleptic complexes were subtracted from the heteroleptic-complex mixture to make the mixed ligand complex NMR signals more discernable from the NMR spectrum the spectrum.

Figure 59. The spectra of homoleptic complexes with zinc as well as new heteroleptic complex species, emergent in the mixed ligand system. The imine peak at ca 8.8 ppm from AmB-Zn complex is evident due to lack of shielding from a benzyl, resulting in a higher shift.

The mixed ligand complex AmB-Zn showed larger downfield shifts for the imine proton. This result demonstrated that in case of the AmAm-Zn complex, the shielding effect is also present for the pyridine ligand part from the benzyl group. However, the electrophilic interaction with the metal produces stronger deshielding, leading to a net downfield shift, albeit lower than either of the cases separate. The net complex charge can also be controlled by coordinating metal charge. For comparison Ga3+ was introduced, allowing to additionally check for a possible net charge effect on the distribution of the different species in the mixed ligand system (Figure 60). In this way the net charge of the mixed complex is 0 while the electron-withdrawing capacity can be enhanced through stronger electrophilic activation. However, it seemed that one of the imine ligands is typically broken and only the hemiacetal-imine mixed coordination is favored. This effect could be attributed to a combination of the electrophilic activation and the asymmetry between the ligands observed in the zinc crystal structure data – the opposing pyridine ligands bond zinc in a less-than-straight angle in the distorted octahedral coordination mode, which could be even further distorted between two ligands in a gallium imine complex, due to further change in length of the coordination bonds as was observed for structurally related

61 amine-based hexadentate chelators coordinating to Ga3+.242 The strain is likely reduced when the hemiacetal is coordinating instead. The interchange between acetal, hemiacetal and aminal is still possible, but was not investigated further in this study.

Figure 60. Structures for mixed mer-ligand system and their charge and number of π- π-interactions in mixed ligand coordination systems with gallium.

The strong electrophilic activation and oxophilicity of gallium leads to the loss of the benzylamine unit as hemiacetal coordination becomes more prefered, and for the mixed system AB nearly no cross π-stacking was observed from the shifts in 1H NMR spectra (Figure 61). This is an interesting effect due to possible utility in redox-triggered capture-release of amines, with applications in redox- responsive release of primary amine-containing bioactive moieties, extrapolating this system for redox active metals such as iron.

Table 3. Net observed number of π-π interactions observed by NMR in homo and hetero ligand systems and AB/AA distributions of mixed systems after equilibration.

Coordinating metal Zn2+ Ga3+

Benzylamine R groups -OMe -F -OMe -F

1Homo complex AA-Mn+ 2x π-π 2x π-π 1x π-π 1x π-π

2Hetero complex AB-Mn+ 1x π-π 1x π-π 0x π-π 0x π-π

3Approximate AB-Mn+ / AA- 3/1 4/1 4/1 3/1 Mn+distribution 1Homo complex in case of zinc is the diimine complex, while for gallium the main species observed is the mixed hemiacetal/imine complex. 2Hetero complex for zinc is the imine and dipicolinate complex whilst for gallium the main species appears to be

62 hemiacetal-dipicolinate complex which appears the same in both cases. 3distributions in case of zinc are evaluated from imine peak integral ratios appearing at 8.8–8.6 ppm while in gallium case the hemiacetal signals were used and appear at 6.2–6.1 ppm. Minor species appear as trace signals in case of gallium complexes and were not accounted for.

The 1H NMR spectra for the mixed systems are displayed in Figure 61 using signal suppression of the homo species by NMR subtraction. Note that in the case where there is some shift-drift the signal suppression is limited, since it is achieved by subtraction of entire spectra.

Figure 61. Comparison of the mixed ligand complexes with zinc and gallium: Signals of homoligand complexes existing in equilibrium were suppressed through spectral subtraction. For the zinc complexes, both trimethoxy- and trifluorobenzylic group bearing complexes produce an up-field shift of the the corresponding protons indicating an interaction with the dipicolinate ligand, while in case of gallium the hybrid complex appears to be the hemiacetal with both amines free in the system

For both of the mixed gallium complexes, the shifts suggest that the complexes are identical. For the mixed complexes, generally good solubility was observed, and no complex crystallized out of the solutions during the experiments.

Since the π-interaction is observed in both EWG and EDG containing aryl groups, additional substituents were evaluated (Figure 62). The p-methoxy and p-CF3 groups were chosen to compare for any steric interaction against tested EWG and EDG containing benzylamines. In addition, an amine containing the fluorophore pyrene was tested. Fluorophores are relevant for sensing applications, where in this case the proximity of two pyrenyl groups next to each other is expected to promote excimer fluorescence243 while stacking interaction of pyrene with an electron-deficient π-system is known to produce quenching.244

Figure 62. Mono-substituted benzylamines b and e selected to further investigate steric influences in the ligand as well as 1-aminomethylpyrene based ligand p to check for a competing homoleptic π-π-interaction, exploiting large surface, proximity and potential for excimer formation.

The following part of the study addressed the π-stacking effects. As π-stacking effects were absent in AB-Ga complexes, the mixed ligand systems with dipicolinic acid were omitted and focus was placed on the homoleptic complexes.

Figure 63. The scope study of AA coordination mode for zinc and gallium complexes. The π-π-bridge effect appeared in all case in the same fashion as for the tri- substituted benzyl amines showing that the system is applicable for a variety of binders in further development.

The observable NMR effects were similar as in cases of the tri-substituted aryl groups for ligands Ab and Ae. The more π-deficient trifluorobenzylamine based complex AeAe-Zn tends to crystallize out overnight into fine, cotton hair-like crystals, unsuitable for X-ray crystallography. As for the pyrenyl based-imine ligand, in case of zinc it produced a highly orderly structure according to NMR spectroscopy. The most interesting feature is that in this case the pyridine protons are shifted downfield. Again, no NOESY cross-peak could be observed between the pyridine and aromatic pyrene protons, as in previous cases. However, the strongly pronounced upfield shift of the pyridine protons indicate that the pyridine appears to be well overlapping with the central part of the pyrene ring. Also, the splitting of the benzyl group into two distinct proton signals appears downfield.

Figure 64. a) The ApAp-Zn complex structure with pyrenyl and pyridine group intramolecular π-π-interaction affected the marked protons b) NOESY spectrum of the complex with the spatially isolated three proton crosspeak system corresponding to pyridine protons and the spatially locked benzylic protons.

The NOESY cross-peaks do not appear between the fragments, which, together with the strong shielding effect suggests that the electron-deficient pyridine protons end up above the epicenter of the pyrene π-system. Also, interesting to note is that each pyrene, due to asymmetric shape, can also interact by flipping to one of the two inequivalent sides in the complex. In this case it is difficult to suggest which of the rotational isomers is favored. In case of the gallium complexes, the mixture displayed poor solubility resulting in the less resolved NMR spectra which appear similar to the corresponding zinc complex. Similar to the previously studied gallium complexes, it is evident that the mixed hemiacetal and imine ligands are obtained.

In summary, a π-π-interaction induced by metal coordination was investigated. The π-π-interaction is established between two meridional ligands on octahedral zinc and gallium complexes, acting as a drawbridge. For homoleptic complexes, two π-π-interactions are present. In mixed system, with the higher negatively charged dipicolinate ligand, the π-interaction is fairly strong but more of the open-bridge peaks appear when the approaching benzyl group is π-rich. Altogether, the results support the notion that the interaction is electrostatic in nature and is partially charge dependent.

In case of complexes with gallium3+, one benzylamine fragment is cleaved per coordination complex, leading to only one π-interaction for a single ligand system and no π-π-interactions are observed for mixed ligand systems,

65 regardless of the substituents present on the benzylamine. This feature suggests the potential for redox-responsive stack and release ability.

In perspective, the extension of this system to iron2+ and iron3+ complexes would enable capture and release of a range of amino-group bearing substrates, as well as capture or release of bioactive amino-functionalized entities. One of the challenges for switching to iron2+ and iron3+ complexes is the necessity to directly monitor the system by other means than NMR spectroscopy. Other spectroscopic signatures for these systems have to be considered to be able to directly compare whether π-stacking effects are present. This system can also be developed into sensor systems, and expanded to show color effect through e.g. extension of the π-surface.

These complexes can further be used for competitive release of fluorophores, such as pyrene, upon intercalation, in a manner which would produce fluorescence.

Furthermore, extension of the bridge ligand would immediately yield the sensing capability of bioactive compounds in the phenethylamine-class, such as dopamine. Moreover, owing to inherent chirality of meridional complexes, imine formation from chiral primary amines could enable chiral enrichment sensing of said amines by new dynamic covalent means.245

We propose that a nucleic acid dynamic association with metal chelate auxiliary could allow for catalytic function in a biotic or prebiotic setting through exploitation of a variety of metal-catalyzed reactions. It was demonstrated that both zinc and gallium can induce the association between π-systems of opposing charge, which would suggest that different metal complexes could have the capacity to insert in-between the stacked nucleic π-system. For now, one can only speculate what kind of catalytic or protective function these nucleic acid dynamic conjugates could bear with different metals. Further studies in this direction could find utility for both developing selective intercalators as well as a new approach on dynamic nucleic acid enzymes, which could be incorporating an auxiliary via π-stacked coordination complex for catalytic activity.

66 4. Concluding remarks

The first chapter discussed the design strategies based on dynamic covalent reaction networks for creating responsive molecular systems, as well as implications of multi-dynamic processes to systemic stability. The extensions of systems involving dynamic covalent polymerization to biomimetic molecules, templated synthesis and non-equilibrium systems were elaborated, and a gradual systemic stabilization strategy was proposed for adaptive behavior, as a stimuli- responsiveness extension.

The second chapter investigated dynamic nitroaldol systems. First, dynamic oligomerization in aqueous media was demonstrated. The inherent responsiveness to formaldehyde was addressed and exploited to create a dynamic covalent polymerization system, in which a nitroaldol dynamer formed transiently. The persistence of the dynamer was shown to be dependent on the initial conditions of the system. Furthermore, the nitroaldol reaction with nitroethanol and nitromethane enables functional groups (1,3-diols) that are able to form boronates. The interaction of nitroaldol and boronate systems was shown to be inhibitory – the nitroaldol reaction is strongly inhibited in both directions by the presence of boronate ester formation. The stereoselectivity of boronic acid was observed for different diastereomers of nitroaldol products. Several types of boronate- nitroaldol dynamers were prepared, including side-chain, main chain and a crosslinking dynamer. Overall results demonstrate the potential of using nitroaldol reactions in non-equilibrium, multi-dynamic as well as responsive molecular systems.

In the third chapter, metal coordination role was addressed in terms of control over supramolecular processes. A dynamic nitroaldol – hemiacetal system, where the nitroaldol reaction was shown to occur without external catalysis, was demonstrated to be responsive to metal coordination by emergence of a new dynamic reaction – metal - hemiacetal coordination, effectively showing that the control over one dynamic process gained influence over the other, through the connectivity of the network. This was extended to dynamic polymerization, in which the nitroaldol dynamer spontaneously formed upon dissolution of reagents. The size of the dynamer is effectively controlled either by increasing the concentration, making the dynamer larger, or by metal coordination, thereby disrupting the formation. The extent of the effects is effectively predetermined

67 by the reaction network and the interplay of systemic robustness as well as choice and specificity of stimulus response. Next, metal coordination-triggered π-π-assembly was demonstrated based on nature inspired systems. A tridentate meridional ligand was shown to have emergent π-stacking effects with an opposing ligand. The extent of these effects was investigated in a multi-dynamic system in which complexes having 2, 1 or 0 π-π-interactions form. In case of increased metal charge, one amine molecule was demonstrated to cleave to form a hemiacetal instead. Overall, the system showed metal-triggered assembly and emergent π-π-interactions. Metal charge was demonstrated to have control over one of the dynamic covalent bonds of the ligand. This system shows promise for developing sensors based on π-π- interactions, such as intercalation and redox-controlled capture-release of amino-functionalized molecules in the future.

The approach of combining dynamic covalent reactions into reaction networks has been demonstrated to be a viable strategy for discovery of new systemic stimuli-responsive behaviors, emerging from connectivity in the dynamic systems, operating in equilibrium or non-equilibrium conditions.

68 Acknowledgements

“One day I will find the right words, and they will be simple.”

Jack Kerouac

First and foremost I have to admit that I am a very lucky person on many levels. At the top of everything that makes me feel so very lucky are the amazing people I had the chance to meet and spend time with during this part of my journey. I have to express my sincerest gratitude to the following persons who made it possible:

My supervisors Olof Ramström and Zoltan Szabo. Olof, I am sincerely grateful to you for giving me the opportunity to become a PhD student under your supervision. I learned a lot from you regardless of circumstances. I want to thank for the trust and the freedom to pursue ideas as well as sharing wisdom on what is worth pursuing in the end. Zoltán, I wish to thank you for taking up the task and becoming a great supervisor. I picked up quite a few things from you during the time you supervised me and for that I am grateful.

Co-supervisors Mingdi Yan, Ulla Elofsson, Per Berglund and Licheng Sun for sharing the responsibility and for the valuable discussions and feedback over the years.

Sincere gratitude to my mentor Raimondas Mozūraitis. Ačiū.

A very big thanks to Markus Kärkas for thesis preview as well as Kristina Druceikaitė, Giampiero Proietti, Oleksandr Kravchenko, Yansong Ren Fredrik Schaufelberger and Brian Timmer for all the rounds of review. Truly thanks!

Special thanks to Alex for never opting out from helping out and being a mate!

Co-authors that made the research possible: Olof Ramström, Mingdi Yan, Ulla Elofsson, Zoltán Szabó, Yang Zhang, Oleksandr Kravchenko, Erik S. Grape, A. Ken Inge.

All the former group members: Alex, Alexey, Allan, Brian, Fredrik, Giampiero, Joakim, Juan, Karolina, Kequan, Laura, Lei, Linnea, Linqin, Lizhou, Maurice, Na, Qijun, Ruzal, Sheng, Sonja, Yang, Yanmiao, Yansong, Iwona and Vercha for being the best possible random combination of people to work with! I had great fun with all of you and thanks for making the dull weather outside unnoticeable!

69 Big thanks to all the colleagues in the division and beyond who made this place lively by making and owning traditions of Friday lunch, beer, games, sports and routine humor over the years: Quentin, Fredrik, Yansong, Alex, Giampiero, Kequan, Björn, Andrey, Daniele, Valentina, Cristiana, Tamara, Linqin, Xia, Tianqi, Piret, Bo, Qijun, Fuguo, Biaobiao, Leila, Lizhou – I thank you all!

Office mates: Alex, Bo, Hao and Qijun. It is great fun to share the workspace with you guys! Someone will have to water the PhD trees when I am not there!

Peter, Zoltán, Lena, Ulla, Ann, Alex, Andrey, Giampiero, Daniele, Biaobiao, and everyone who kept the division and school going! Peter, Istvan, Inger and Mikael for problem solving, doing the right thing and being proper adults. The brave climbers of the chemistry department: Björn, Giampiero, Maria, Tamara, Maurice, Daniele – who do not let go easy! Special thanks to Ulla Elofsson for hosting me at RISE as well as Lucas Boge for all the help with ellipsometry. A kind thank you to prof. Aldrik Velders for hosting me at WUR. I would like to thank the excellent group of wonderful people I met there and to express additional gratitude to Camilla, Bânû, Pieter and Vittorio. I want to thank everyone who was involved in the ResMoSys network. Albano, Camilla, Suzanne, John, Ligia, Marion, Guille, Vanessa, Qiang, Giuseppe, Lianjin - without you folks the experience would not have been anywhere near to what it was! What a crazy bunch of people from all over! I am very happy to have met you all. Lastly, the most important people: my family. Kiprai, Aukse ir Simona ačiū jums už palaikymą ir vizitus. Aš jumis labai džiaugiuosi ir didžiuojuosi. Finally, my Kristina. Be tavęs, tavo meilės, rūpesčio ir paramos man nebūtų pavykę. Ačiū kad atsibeldei iš paskos, myliu ir bučiuoju ir dėkoju tau labai. One last thing, if you are reading this, you most likely are person that I forgot to thank for something. And for that something I thank you.

70 Appendix

The following is a description of my contributions to Papers I–IV:

Paper I: I contributed to the formulation of the research problem, performed the majority of the experimental work and wrote the manuscript.

Paper II: I formulated the research problem and designed the experiments, performed the experimental work and wrote the manuscript.

Paper III: I formulated the research problem and designed the experiments, performed the experimental work and wrote the manuscript.

Paper IV: I formulated the research problem and designed the experiments, performed the experimental work and wrote the manuscript.

71 References

1. Pross, A., What is life?: How chemistry becomes biology. Oxford University Press: 2016. 2. Image of 1LVB (Phan, J., Zdanov, A., Evdokimov, A.G., Tropea, J.E., Peters III, H.K., Kapust, R.B., Li, M., Wlodawer, A., Waugh, D.S. Structural basis for the substrate specificity of tobacco etch virus protease (2002) J. Biol. Chem. 277: 50564-50572) created with NGL (A.S. Rose, A.R. Bradley, Y. Valasatava, J.D. Duarte, A. Prlić, P.W. Rose (2018) NGL viewer: web-based molecular graphics for large complexes. Bioinformatics 34: 3755–3758). 3. Image of 3ZD5 (Martick, M., Scott, W.G. Tertiary contacts distant from the active site prime a ribozyme for catalysis (2006) Cell 126: 309- 320) created with NGL (A.S. Rose, A.R. Bradley, Y. Valasatava, J.D. Duarte, A. Prlić, P.W. Rose (2018) NGL viewer: web-based molecular graphics for large complexes. Bioinformatics 34: 3755–3758). 4. Fedor, M. J.; Williamson, J. R., The catalytic diversity of RNAs. Nat. Rev. Mol. Cell Biol. 2005, 6, 399-412. 5. Carothers, W. H., Polymers and polyfunctionality. Trans. Faraday Soc. 1936, 32, 39-49. 6. Eigen, M., Macromolecular Evolution: Dynamical Ordering in Sequence Space. Ber. Bunsen-Ges. Phys. Chem 1985, 89, 658-667. 7. Benner, S. A.; Bell, E. A.; Biondi, E.; Brasser, R.; Carell, T.; Kim, H. J.; Mojzsis, S. J.; Omran, A.; Pasek, M. A.; Trail, D., When Did Life Likely Emerge on Earth in an RNA‐First Process? ChemSystemsChem 2019, 10.1002/syst.201900035. 8. Ashkenasy, G.; Hermans, T. M.; Otto, S.; Taylor, A. F., Systems chemistry. Chem. Soc. Rev. 2017, 46, 2543-2554. 9. Ludlow, R. F.; Otto, S., Systems chemistry. Chem. Soc. Rev. 2008, 37, 101-8. 10. von Kiedrowski, G.; Otto, S.; Herdewijn, P., Welcome Home, Systems Chemists! J. Syst. Chem. 2010, 1, 1. 11. Lehn, J. M., From supramolecular chemistry towards constitutional dynamic chemistry and adaptive chemistry. Chem. Soc. Rev. 2007, 36, 151-60. 12. Hu, L.; Schaufelberger, F.; Timmer, B. J. J.; Flos, M. A.; Ramström, O., Constitutional Dynamic Chemistry. In Kirk-Othmer Encycl. Chem. Technol., John Wiley & Sons, Inc.: 2014, 10.1002/0471238961.consrams.a01, p 1-25. 13. Ramstrom, O.; Bunyapaiboonsri, T.; Lohmann, S.; Lehn, J. M., Chemical biology of dynamic combinatorial libraries. Biochim. Biophys. Acta 2002, 1572, 178-86. 14. Kreyssig, P.; Escuela, G.; Reynaert, B.; Veloz, T.; Ibrahim, B.; Dittrich, P., Cycles and the qualitative evolution of chemical systems. PLoS One 2012, 7, e45772.

72 15. Gleiss, P. M.; Stadler, P. F.; Wagner, A.; Fell, D. A., Relevant Cycles in Chemical Reaction Networks. Adv. Complex. Syst. 2011, 04, 207-226. 16. Kritz, M. V.; Trindade Dos Santos, M.; Urrutia, S.; Schwartz, J. M., Organising metabolic networks: Cycles in flux distributions. J. Theor. Biol. 2010, 265, 250-60. 17. Anslyn, E. V.; Dougherty, D. A., Modern physical organic chemistry. University science books: 2006. 18. Feinberg, M.; Horn, F. J. M., Dynamics of open chemical systems and the algebraic structure of the underlying reaction network. Chem. Eng. Sci. 1974, 29, 775-787. 19. Meixner, J., Macroscopic and microscopic reversibility. Rep. Math. Phys. 1975, 7, 37-57. 20. Lehn, J. M., Toward self-organization and complex matter. Science 2002, 295, 2400-3. 21. Whitesides, G. M.; Grzybowski, B., Self-assembly at all scales. Science 2002, 295, 2418-21. 22. Philp, D.; Stoddart, J. F., Self-assembly in natural and unnatural systems. Angew. Chem. Int. Ed. 1996, 35, 1154-1196. 23. Chakrabarty, R.; Mukherjee, P. S.; Stang, P. J., Supramolecular coordination: self-assembly of finite two- and three-dimensional ensembles. Chem. Rev. 2011, 111, 6810-918. 24. Nelson, D. L.; Lehninger, A. L.; Cox, M. M., Lehninger principles of biochemistry. Macmillan: 2008. 25. Rowan, S. J.; Cantrill, S. J.; Cousins, G. R.; Sanders, J. K.; Stoddart, J. F., Dynamic covalent chemistry. Angew. Chem. Int. Ed. 2002, 41, 898-952. 26. Schaufelberger, F., Timmer, B.J. and Ramström, O., Principles of Dynamic Covalent Chemistry. In Dynamic Covalent Chemistry, Jin, W. Z. a. Y., Ed. 2017, 10.1002/9781119075738.ch1, p 1-30. 27. Kloxin, C. J.; Bowman, C. N., Covalent adaptable networks: smart, reconfigurable and responsive network systems. Chem. Soc. Rev. 2013, 42, 7161-73. 28. Wilson, A.; Gasparini, G.; Matile, S., Functional systems with orthogonal dynamic covalent bonds. Chem. Soc. Rev. 2014, 43, 1948-62. 29. Herrmann, A., Dynamic combinatorial/covalent chemistry: a tool to read, generate and modulate the bioactivity of compounds and compound mixtures. Chem. Soc. Rev. 2014, 43, 1899-933. 30. Angelin, M.; Vongvilai, P.; Fischer, A.; Ramström, O., Crystallization-Driven Asymmetric Synthesis of Pyridine-β-nitroalcohols via Discovery-Oriented Self-Resolution of a Dynamic System. Eur. J. Org. Chem. 2010, 2010, 6315-6318. 31. Armao, J. J. t.; Lehn, J. M., Adaptive Chemical Networks under Non-Equilibrium Conditions: The Evaporating Droplet. Angew. Chem. Int. Ed. 2016, 55, 13450-13454.

73 32. Saggiomo, V.; Luning, U., Transport of calcium ions through a bulk membrane by use of a dynamic combinatorial library. Chem. Commun. 2009, 10.1039/b902847a, 3711-3. 33. Alonci, G.; Fiorini, F.; Riva, P.; Monroy, F.; López-Montero, I.; Perretta, S.; De Cola, L., Injectable Hybrid Hydrogels, with Cell-Responsive Degradation, for Tumor Resection. ACS Appl. Bio Mater. 2018, 1, 1301-1310. 34. Chen, C.; Tan, J.; Hsieh, M. C.; Pan, T.; Goodwin, J. T.; Mehta, A. K.; Grover, M. A.; Lynn, D. G., Design of multi-phase dynamic chemical networks. Nat. Chem. 2017, 9, 799-804. 35. Grzybowski, B.; Otto, S.; Philp, D., Systems chemistry: a web themed issue. Chem. Commun. 2014, 50, 14924-5. 36. Zhang, Y.; Qi, Y.; Ulrich, S.; Barboiu, M.; Ramström, O., Dynamic covalent polymers for biomedical applications. Mater. Chem. Front. 2020, 10.1039/c9qm00598f. 37. Temkin, O. N.; Bonchev, D. G., Application of graph theory to chemical kinetics: Part 1. Kinetics of complex reactions. J. Chem. Educ. 1992, 69, 544. 38. Eigen, M., Selforganization of matter and the evolution of biological macromolecules. Naturwissenschaften 1971, 58, 465-523. 39. Veloz, T.; Razeto-Barry, P., Reaction Networks as a Language for Systemic Modeling: Fundamentals and Examples. Systems 2017, 5, 11. 40. Singh, P.; Rheinhardt, J. H.; Olson, J. Z.; Tarakeshwar, P.; Mujica, V.; Buttry, D. A., Electrochemical Capture and Release of Carbon Dioxide Using a Disulfide-Thiocarbonate Redox Cycle. J. Am. Chem. Soc. 2017, 139, 1033-1036. 41. Benkö, G.; Flamm, C.; Stadler, P. F. In Generic Properties of Chemical Networks: Artificial Chemistry Based on Graph Rewriting, Advances in Artificial Life, Berlin, Heidelberg, 2003//; Banzhaf, W.; Ziegler, J.; Christaller, T.; Dittrich, P.; Kim, J. T., Eds. Springer Berlin Heidelberg: Berlin, Heidelberg, 2003; pp 10-19. 42. Kozuch, S., Steady State Kinetics of Any Catalytic Network: Graph Theory, the Energy Span Model, the Analogy between Catalysis and Electrical Circuits, and the Meaning of “Mechanism”. ACS Catal. 2015, 5, 5242-5255. 43. Zhang, Y.; Vongvilai, P.; Sakulsombat, M.; Fischer, A.; Ramstrom, O., Asymmetric Synthesis of Substituted Thiolanes through Domino Thia-Michael-Henry Dynamic Covalent Systemic Resolution using Lipase Catalysis. Adv. Synth. Catal. 2014, 356, 987-992. 44. Grzybowski, B. A.; Fitzner, K.; Paczesny, J.; Granick, S., From dynamic self-assembly to networked chemical systems. Chem. Soc. Rev. 2017, 46, 5647-5678. 45. Lascano, S.; Zhang, K. D.; Wehlauch, R.; Gademann, K.; Sakai, N.; Matile, S., The third orthogonal dynamic covalent bond. Chem. Sci. 2016, 7, 4720-4724.

74 46. Seifert, H. M.; Ramirez Trejo, K.; Anslyn, E. V., Four Simultaneously Dynamic Covalent Reactions. Experimental Proof of Orthogonality. J. Am. Chem. Soc. 2016, 138, 10916-24. 47. Ren, Y.; You, L., Dynamic signaling cascades: reversible covalent reaction-coupled molecular switches. J. Am. Chem. Soc. 2015, 137, 14220-8. 48. Sun, X.; Anslyn, E. V., An Auto-Inductive Cascade for the Optical Sensing of Thiols in Aqueous Media: Application in the Detection of a VX Nerve Agent Mimic. Angew. Chem. Int. Ed. 2017, 56, 9522-9526. 49. Chakma, P.; Konkolewicz, D., Dynamic Covalent Bonds in Polymeric Materials. Angew. Chem. Int. Ed. 2019, 58, 9682-9695. 50. Liu, Y.; Lehn, J. M.; Hirsch, A. K., Molecular Biodynamers: Dynamic Covalent Analogues of Biopolymers. Acc. Chem. Res. 2017, 50, 376- 386. 51. Roy, N.; Bruchmann, B.; Lehn, J. M., DYNAMERS: dynamic polymers as self-healing materials. Chem. Soc. Rev. 2015, 44, 3786-807. 52. Bang, E. K.; Lista, M.; Sforazzini, G.; Sakai, N.; Matile, S., Poly(disulfide)s. Chem. Sci. 2012, 3, 1752-1763. 53. Khan, I.; Wang, J.; Zou, H.; Ye, H.; Zha, D.; Zhang, Y.; You, L., Noncovalent and Dynamic Covalent Chemistry Strategies for Driving Thermoresponsive Phase Transition with Multistimuli and Controlled Encapsulation/Release. ACS Appl. Mater. Interfaces 2020, 12, 2962-2973. 54. Casassa, E. Z.; Sarquis, A. M.; Vandyke, C. H., The Gelation of Polyvinyl-Alcohol with Borax - a Novel Class Participation Experiment Involving the Preparation and Properties of a Slime. J. Chem. Educ. 1986, 63, 57-60. 55. Jin, Y.; Wang, Q.; Taynton, P.; Zhang, W., Dynamic covalent chemistry approaches toward macrocycles, molecular cages, and polymers. Acc. Chem. Res. 2014, 47, 1575-86. 56. Liu, Y.; Jia, Y.; Wu, Q.; Moore, J. S., Architecture-Controlled Ring-Opening Polymerization for Dynamic Covalent Poly(disulfide)s. J. Am. Chem. Soc. 2019, 141, 17075-17080. 57. Gross, D. E.; Moore, J. S., Arylene–Ethynylene Macrocycles via Depolymerization–Macrocyclization. Macromolecules 2011, 44, 3685- 3687. 58. Liu, B.; Pappas, C. G.; Zangrando, E.; Demitri, N.; Chmielewski, P. J.; Otto, S., Complex Molecules That Fold Like Proteins Can Emerge Spontaneously. J. Am. Chem. Soc. 2019, 141, 1685-1689. 59. Mahon, E.; Barboiu, M., Synthetic multivalency for biological applications. Org. Biomol. Chem. 2015, 13, 10590-9. 60. Webber, M. J.; Appel, E. A.; Meijer, E. W.; Langer, R., Supramolecular biomaterials. Nat. Mater. 2016, 15, 13-26.

75 61. Maeda, T.; Otsuka, H.; Takahara, A., Dynamic covalent polymers: Reorganizable polymers with dynamic covalent bonds. Prog. Polym. Sci. 2009, 34, 581-604. 62. Garcia, F.; Smulders, M. M., Dynamic covalent polymers. J. Polym. Sci., Part A: Polym. Chem. 2016, 54, 3551-3577. 63. Griebel, J. J.; Nguyen, N. A.; Astashkin, A. V.; Glass, R. S.; Mackay, M. E.; Char, K.; Pyun, J., Preparation of Dynamic Covalent Polymers via Inverse Vulcanization of Elemental Sulfur. ACS Macro Lett. 2014, 3, 1258- 1261. 64. Wojtecki, R. J.; Meador, M. A.; Rowan, S. J., Using the dynamic bond to access macroscopically responsive structurally dynamic polymers. Nat. Mater. 2011, 10, 14-27. 65. Beeren, S. R.; Pittelkow, M.; Sanders, J. K., From static to dynamic: escaping kinetic traps in hydrazone-based dynamic combinatorial libraries. Chem. Commun. 2011, 47, 7359-61. 66. Wei, Z.; Yang, J. H.; Zhou, J.; Xu, F.; Zrinyi, M.; Dussault, P. H.; Osada, Y.; Chen, Y. M., Self-healing gels based on constitutional dynamic chemistry and their potential applications. Chem. Soc. Rev. 2014, 43, 8114-31. 67. Luisier, N.; Schenk, K.; Severin, K., A four-component organogel based on orthogonal chemical interactions. Chem. Commun. 2014, 50, 10233-6. 68. Bapat, A. P.; Ray, J. G.; Savin, D. A.; Hoff, E. A.; Patton, D. L.; Sumerlin, B. S., Dynamic-covalent nanostructures prepared by Diels-Alder reactions of styrene-maleic anhydride-derived copolymers obtained by one-step cascade block copolymerization. Polym. Chem. 2012, 3, 3112-3120. 69. Cromwell, O. R.; Chung, J.; Guan, Z., Malleable and Self- Healing Covalent Polymer Networks through Tunable Dynamic Boronic Ester Bonds. J. Am. Chem. Soc. 2015, 137, 6492-5. 70. Sun, H.; Kabb, C. P.; Sumerlin, B. S., Thermally-labile segmented hyperbranched copolymers: using reversible-covalent chemistry to investigate the mechanism of self-condensing vinyl copolymerization. Chem. Sci. 2014, 5, 4646-4655. 71. Jones, C. D.; Simmons, H. T. D.; Horner, K. E.; Liu, K.; Thompson, R. L.; Steed, J. W., Braiding, branching and chiral amplification of nanofibres in supramolecular gels. Nat. Chem. 2019, 11, 375-381. 72. Zhang, Y.; Li, Y.; Su, C.; Barboiu, M., Dynameric Frameworks with Aggregation-Induced Emission for Selective Detection of Adenosine Triphosphate. ChemPlusChem 2018, 83, 506-513. 73. Wang, J. Y.; Stuart, M. A. C.; Marcelis, A. T. M.; Colomb- Delsuc, M.; Otto, S.; van der Gucht, J., Stable Polymer Micelles Formed by Metal Coordination. Macromolecules 2012, 45, 7179-7185. 74. Deng, R.; Derry, M. J.; Mable, C. J.; Ning, Y.; Armes, S. P., Using Dynamic Covalent Chemistry To Drive Morphological Transitions:

76 Controlled Release of Encapsulated Nanoparticles from Block Copolymer Vesicles. J. Am. Chem. Soc. 2017, 139, 7616-7623. 75. Mavila, S.; Worrell, B. T.; Culver, H. R.; Goldman, T. M.; Wang, C.; Lim, C. H.; Domaille, D. W.; Pattanayak, S.; McBride, M. K.; Musgrave, C. B.; Bowman, C. N., Dynamic and Responsive DNA-like Polymers. J. Am. Chem. Soc. 2018, 140, 13594-13598. 76. Garcia-Domenech, R.; Galvez, J.; de Julian-Ortiz, J. V.; Pogliani, L., Some new trends in chemical graph theory. Chem. Rev. 2008, 108, 1127-69. 77. Marin, L.; Simionescu, B.; Barboiu, M., Imino-chitosan biodynamers. Chem. Commun. 2012, 48, 8778-80. 78. Nunez-Villanueva, D.; Ciaccia, M.; Hunter, C. A., Cap control: cyclic versus linear oligomerisation in covalent template-directed synthesis. RSC Adv. 2019, 9, 29566-29569. 79. Hebel, M.; Riegger, A.; Zegota, M. M.; Kizilsavas, G.; Gacanin, J.; Pieszka, M.; Luckerath, T.; Coelho, J. A. S.; Wagner, M.; Gois, P. M. P.; Ng, D. Y. W.; Weil, T., Sequence Programming with Dynamic Boronic Acid/Catechol Binary Codes. J. Am. Chem. Soc. 2019, 141, 14026-14031. 80. Gabrielli, L.; Nunez-Villanueva, D.; Hunter, C. A., Two- component assembly of recognition-encoded oligomers that form stable H- bonded duplexes. Chem. Sci. 2020, 11, 561-566. 81. Mirzahosseini, A.; Noszal, B., Species-Specific Standard Redox Potential of Thiol-Disulfide Systems: A Key Parameter to Develop Agents against Oxidative Stress. Sci. Rep. 2016, 6, 37596. 82. Parfrey, L. W.; Lahr, D. J.; Katz, L. A., The dynamic nature of eukaryotic genomes. Mol. Biol. Evol. 2008, 25, 787-94. 83. Wicker, T.; Sabot, F.; Hua-Van, A.; Bennetzen, J. L.; Capy, P.; Chalhoub, B.; Flavell, A.; Leroy, P.; Morgante, M.; Panaud, O.; Paux, E.; SanMiguel, P.; Schulman, A. H., A unified classification system for eukaryotic transposable elements. Nat. Rev. Genet. 2007, 8, 973-82. 84. Holmes, E. C., On being the right size. Nat. Genet. 2005, 37, 923-4. 85. Ruff, Y.; Garavini, V.; Giuseppone, N., Reversible native chemical ligation: a facile access to dynamic covalent peptides. J. Am. Chem. Soc. 2014, 136, 6333-9. 86. Prusiner, S. B., Prions. Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 13363-83. 87. Symons, R. H., Self-cleavage of RNA in the replication of small pathogens of plants and animals. Trends Biochem. Sci 1989, 14, 445-50. 88. van Nimwegen, E.; Crutchfield, J. P.; Huynen, M., Neutral evolution of mutational robustness. Proc. Natl. Acad. Sci. U. S. A. 1999, 96, 9716-20. 89. Hutchison, C. A., 3rd; Chuang, R. Y.; Noskov, V. N.; Assad- Garcia, N.; Deerinck, T. J.; Ellisman, M. H.; Gill, J.; Kannan, K.; Karas, B. J.;

77 Ma, L.; Pelletier, J. F.; Qi, Z. Q.; Richter, R. A.; Strychalski, E. A.; Sun, L.; Suzuki, Y.; Tsvetanova, B.; Wise, K. S.; Smith, H. O.; Glass, J. I.; Merryman, C.; Gibson, D. G.; Venter, J. C., Design and synthesis of a minimal bacterial genome. Science 2016, 351, aad6253. 90. Kacian, D. L.; Mills, D. R.; Kramer, F. R.; Spiegelman, S., A replicating RNA molecule suitable for a detailed analysis of extracellular evolution and replication. Proc. Natl. Acad. Sci. U. S. A. 1972, 69, 3038-42. 91. Sniegowski, P.; Raynes, Y., Mutation rates: how low can you go? Curr Biol 2013, 23, R147-9. 92. Mutschler, H.; Wochner, A.; Holliger, P., Freeze-thaw cycles as drivers of complex ribozyme assembly. Nat. Chem. 2015, 7, 502-8. 93. Edeleva, E.; Salditt, A.; Stamp, J.; Schwintek, P.; Boekhoven, J.; Braun, D., Continuous nonenzymatic cross-replication of DNA strands with in situ activated DNA oligonucleotides. Chem. Sci. 2019, 10, 5807-5814. 94. Wagner, A., Robustness and evolvability: a paradox resolved. Proc. R. Soc. B 2008, 275, 91-100. 95. Sadownik, J. W.; Mattia, E.; Nowak, P.; Otto, S., Diversification of self-replicating molecules. Nat. Chem. 2016, 8, 264-9. 96. Ghidini, A.; Murtola, M.; Strömberg, R., Oligonucleotide Based Artificial Ribonucleases (OBANs). In DNA in Supramolecular Chemistry and Nanotechnology, 2017, 10.1002/9781118696880.ch3.2, p 158-171. 97. Nowak, P.; Saggiomo, V.; Salehian, F.; Colomb-Delsuc, M.; Han, Y.; Otto, S., Localized template-driven functionalization of nanoparticles by dynamic combinatorial chemistry. Angew. Chem. Int. Ed. 2015, 54, 4192-7. 98. Gimenez Molina, A.; Barvik, I.; Muller, S.; Vasseur, J. J.; Smietana, M., RNA-based boronate internucleosidic linkages: an entry into reversible templated ligation and loop formation. Org. Biomol. Chem. 2018, 16, 8824-8830. 99. Pasparakis, G.; Krasnogor, N.; Cronin, L.; Davis, B. G.; Alexander, C., Controlled polymer synthesis--from biomimicry towards synthetic biology. Chem. Soc. Rev. 2010, 39, 286-300. 100. Bai, Y.; Chotera, A.; Taran, O.; Liang, C.; Ashkenasy, G.; Lynn, D. G., Achieving biopolymer synergy in systems chemistry. Chem. Soc. Rev. 2018, 47, 5444-5456. 101. Sung, S.; Amasino, R. M., Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 2004, 427, 159- 64. 102. Madigan, M. T.; Bender, K. S.; Buckley, D. H.; Stahl, D. A.; Sattley, W. M., Brock Biology of Microorganisms, Books a la Carte Edition. Pearson Education: 2017. 103. Mukherjee, S.; Cash, J. J.; Sumerlin, B. S., Responsive Dynamic Covalent Polymers. In Dynamic Covalent Chemistry, 2017, 10.1002/9781119075738.ch8, p 321-358.

78 104. Hatai, J.; Hirschhauser, C.; Niemeyer, J.; Schmuck, C., Multi- Stimuli-Responsive Supramolecular Polymers Based on Noncovalent and Dynamic Covalent Bonds. ACS Appl. Mater. Interfaces 2020, 12, 2107-2115. 105. Ulrich, S., Growing Prospects of Dynamic Covalent Chemistry in Delivery Applications. Acc. Chem. Res. 2019, 52, 510-519. 106. Mellerup, S. K.; Wang, S., Boron-based stimuli responsive materials. Chem. Soc. Rev. 2019, 48, 3537-3549. 107. Houck, H. A.; Blasco, E.; Du Prez, F. E.; Barner-Kowollik, C., Light-Stabilized Dynamic Materials. J. Am. Chem. Soc. 2019, 141, 12329- 12337. 108. Grzelczak, M.; Liz-Marzan, L. M.; Klajn, R., Stimuli- responsive self-assembly of nanoparticles. Chem. Soc. Rev. 2019, 48, 1342- 1361. 109. Pogodaev, A. A.; Fernandez Regueiro, C. L.; Jakstaite, M.; Hollander, M. J.; Huck, W. T. S., Modular Design of Small Enzymatic Reaction Networks Based on Reversible and Cleavable Inhibitors. Angew. Chem. Int. Ed. 2019, 58, 14539-14543. 110. Roszak, R.; Bajczyk, M. D.; Gajewska, E. P.; Hołyst, R.; Grzybowski, B. A., Propagation of Oscillating Chemical Signals through Reaction Networks. Angew. Chem. 2018, 131, 4568-4573. 111. Zhang, J.; Luijten, E.; Grzybowski, B. A.; Granick, S., Active colloids with collective mobility status and research opportunities. Chem. Soc. Rev. 2017, 46, 5551-5569. 112. Ji, S.; Cao, W.; Yu, Y.; Xu, H., Dynamic diselenide bonds: exchange reaction induced by visible light without catalysis. Angew. Chem. Int. Ed. 2014, 53, 6781-5. 113. Folmer-Andersen, J. F.; Lehn, J. M., Thermoresponsive dynamers: thermally induced, reversible chain elongation of amphiphilic poly(acylhydrazones). J. Am. Chem. Soc. 2011, 133, 10966-73. 114. Godoy-Alcantar, C.; Yatsimirsky, A. K.; Lehn, J. M., Structure- stability correlations for imine formation in aqueous solution. J. Phys. Org. Chem. 2005, 18, 979-985. 115. Drahonovsky, D.; Lehn, J. M., Hemiacetals in dynamic covalent chemistry: formation, exchange, selection, and modulation processes. J. Org. Chem. 2009, 74, 8428-32. 116. Garcia, F.; Pelss, J.; Zuilhof, H.; Smulders, M. M., Multi- responsive coordination polymers utilising metal-stabilised, dynamic covalent imine bonds. Chem. Commun. 2016, 52, 9059-62. 117. Vongvilai, P.; Angelin, M.; Larsson, R.; Ramstrom, O., Dynamic combinatorial resolution: direct asymmetric lipase-mediated screening of a dynamic nitroaldol library. Angew. Chem. Int. Ed. 2007, 46, 948-50. 118. Nguyen, R.; Jouault, N.; Zanirati, S.; Rawiso, M.; Allouche, L.; Fuks, G.; Buhler, E.; Giuseppone, N., Core-shell inversion by pH modulation in dynamic covalent micelles. Soft Matter 2014, 10, 3926-37.

79 119. Wu, X.; Busschaert, N.; Wells, N. J.; Jiang, Y. B.; Gale, P. A., Dynamic covalent transport of amino acids across lipid bilayers. J. Am. Chem. Soc. 2015, 137, 1476-84. 120. Jansze, S. M.; Cecot, G.; Severin, K., Reversible disassembly of metallasupramolecular structures mediated by a metastable-state photoacid. Chem. Sci. 2018, 9, 4253-4257. 121. Sun, X.; Chapin, B. M.; Metola, P.; Collins, B.; Wang, B.; James, T. D.; Anslyn, E. V., The mechanisms of boronate ester formation and fluorescent turn-on in ortho-aminomethylphenylboronic acids. Nat. Chem. 2019, 11, 768-778. 122. Binti Hasan, N. A. S.; Balasubramanian, P., Exact Solution for the Kinetic Equations of First Order Reversible Reaction Systems through Flow Graph Theory Approach. Ind. Eng. Chem. Res. 2013, 52, 10594-10600. 123. Cafferty, B. J.; Wong, A. S. Y.; Semenov, S. N.; Belding, L.; Gmur, S.; Huck, W. T. S.; Whitesides, G. M., Robustness, Entrainment, and Hybridization in Dissipative Molecular Networks, and the Origin of Life. J. Am. Chem. Soc. 2019, 141, 8289-8295. 124. Schaufelberger, F.; Ramstrom, O., Kinetic Self-Sorting of Dynamic Covalent Catalysts with Systemic Feedback Regulation. J. Am. Chem. Soc. 2016, 138, 7836-9. 125. Pezzato, C.; Cheng, C.; Stoddart, J. F.; Astumian, R. D., Mastering the non-equilibrium assembly and operation of molecular machines. Chem. Soc. Rev. 2017, 46, 5491-5507. 126. Sorrenti, A.; Leira-Iglesias, J.; Sato, A.; Hermans, T. M., Non- equilibrium steady states in supramolecular polymerization. Nat. Commun. 2017, 8, 15899. 127. Ragazzon, G.; Prins, L. J., Energy consumption in chemical fuel-driven self-assembly. Nat. Nanotechnol. 2018, 13, 882-889. 128. England, J. L., Dissipative adaptation in driven self-assembly. Nat. Nanotechnol. 2015, 10, 919-23. 129. Sorrenti, A.; Leira-Iglesias, J.; Markvoort, A. J.; de Greef, T. F. A.; Hermans, T. M., Non-equilibrium supramolecular polymerization. Chem. Soc. Rev. 2017, 46, 5476-5490. 130. Merindol, R.; Walther, A., Materials learning from life: concepts for active, adaptive and autonomous molecular systems. Chem. Soc. Rev. 2017, 46, 5588-5619. 131. Wanzke, C.; Jussupow, A.; Kohler, F.; Dietz, H.; Kaila, V. R. I.; Boekhoven, J., Dynamic Vesicles Formed By Dissipative Self‐Assembly. ChemSystemsChem 2019, 2. 132. Tena‐Solsona, M.; Boekhoven, J., Dissipative Self‐Assembly of Peptides. Isr. J. Chem. 2019, 59, 898-905. 133. van Rossum, S. A. P.; Tena-Solsona, M.; van Esch, J. H.; Eelkema, R.; Boekhoven, J., Dissipative out-of-equilibrium assembly of man- made supramolecular materials. Chem. Soc. Rev. 2017, 46, 5519-5535.

80 134. Semenov, S. N.; Wong, A. S.; van der Made, R. M.; Postma, S. G.; Groen, J.; van Roekel, H. W.; de Greef, T. F.; Huck, W. T., Rational design of functional and tunable oscillating enzymatic networks. Nat. Chem. 2015, 7, 160-5. 135. Hermans, T. M.; Stewart, P. S.; Grzybowski, B. A., pH Oscillator Stretched in Space but Frozen in Time. J. Phys. Chem. Lett. 2015, 6, 760-6. 136. Kumar, M.; Ing, N. L.; Narang, V.; Wijerathne, N. K.; Hochbaum, A. I.; Ulijn, R. V., Amino-acid-encoded biocatalytic self-assembly enables the formation of transient conducting nanostructures. Nat. Chem. 2018, 10, 696-703. 137. Boekhoven, J.; Hendriksen, W. E.; Koper, G. J.; Eelkema, R.; van Esch, J. H., Transient assembly of active materials fueled by a chemical reaction. Science 2015, 349, 1075-9. 138. Rieß, B.; Boekhoven, J., Applications of Dissipative Supramolecular Materials with a Tunable Lifetime. ChemNanoMat 2018, 4, 710-719. 139. Biagini, C.; Fielden, S. D. P.; Leigh, D. A.; Schaufelberger, F.; Di Stefano, S.; Thomas, D., Dissipative Catalysis with a Molecular Machine. Angew. Chem. Int. Ed. 2019, 58, 9876-9880. 140. Kathan, M.; Hecht, S., Photoswitchable molecules as key ingredients to drive systems away from the global thermodynamic minimum. Chem. Soc. Rev. 2017, 46, 5536-5550. 141. Sen, A., Self-propelled motors: Light-seeking synthetic trees. Nat. Nanotechnol. 2016, 11, 1003-1004. 142. Hess, H.; Ross, J. L., Non-equilibrium assembly of microtubules: from molecules to autonomous chemical robots. Chem. Soc. Rev. 2017, 46, 5570-5587. 143. Orrillo, A. G.; Escalante, A. M.; Martinez-Amezaga, M.; Cabezudo, I.; Furlan, R. L. E., Molecular Networks in Dynamic Multilevel Systems. Chem. Eur. J. 2019, 25, 1118-1127. 144. Osypenko, A.; Dhers, S.; Lehn, J. M., Pattern Generation and Information Transfer through a Liquid/Liquid Interface in 3D Constitutional Dynamic Networks of Imine Ligands in Response to Metal Cation Effectors. J. Am. Chem. Soc. 2019, 141, 12724-12737. 145. Pappas, C. G.; Shafi, R.; Sasselli, I. R.; Siccardi, H.; Wang, T.; Narang, V.; Abzalimov, R.; Wijerathne, N.; Ulijn, R. V., Dynamic peptide libraries for the discovery of supramolecular nanomaterials. Nat. Nanotechnol. 2016, 11, 960-967. 146. Jansze, S. M.; Severin, K., Clathrochelate Metalloligands in Supramolecular Chemistry and Materials Science. Acc. Chem. Res. 2018, 51, 2139-2147. 147. Omosun, T. O.; Hsieh, M. C.; Childers, W. S.; Das, D.; Mehta, A. K.; Anthony, N. R.; Pan, T.; Grover, M. A.; Berland, K. M.; Lynn, D. G.,

81 Catalytic diversity in self-propagating peptide assemblies. Nat. Chem. 2017, 9, 805-809. 148. Garcia, J. M.; Jones, G. O.; Virwani, K.; McCloskey, B. D.; Boday, D. J.; ter Huurne, G. M.; Horn, H. W.; Coady, D. J.; Bintaleb, A. M.; Alabdulrahman, A. M.; Alsewailem, F.; Almegren, H. A.; Hedrick, J. L., Recyclable, strong thermosets and organogels via paraformaldehyde condensation with diamines. Science 2014, 344, 732-5. 149. Li, L. Y.; Yuan, C. H.; Dai, L. Z.; Thayumanavan, S., Thermoresponsive Polymeric Nanoparticles: Nucleation from Cooperative Polymerization Driven by Dative Bonds. Macromolecules 2014, 47, 5869-5876. 150. Xing, P.; Chen, H.; Bai, L.; Hao, A.; Zhao, Y., Superstructure Formation and Topological Evolution Achieved by Self-Organization of a Highly Adaptive Dynamer. ACS Nano 2016, 10, 2716-27. 151. Billiet, S.; De Bruycker, K.; Driessen, F.; Goossens, H.; Van Speybroeck, V.; Winne, J. M.; Du Prez, F. E., Triazolinediones enable ultrafast and reversible click chemistry for the design of dynamic polymer systems. Nat. Chem. 2014, 6, 815-21. 152. Saghatelian, A.; Yokobayashi, Y.; Soltani, K.; Ghadiri, M. R., A chiroselective peptide replicator. Nature 2001, 409, 797-801. 153. Leira-Iglesias, J.; Tassoni, A.; Adachi, T.; Stich, M.; Hermans, T. M., Oscillations, travelling fronts and patterns in a supramolecular system. Nat. Nanotechnol. 2018, 13, 1021-1027. 154. Tena-Solsona, M.; Wanzke, C.; Riess, B.; Bausch, A. R.; Boekhoven, J., Self-selection of dissipative assemblies driven by primitive chemical reaction networks. Nat. Commun. 2018, 9, 2044. 155. Zhang, K. D.; Matile, S., Complex Functional Systems with Three Different Types of Dynamic Covalent Bonds. Angew. Chem. Int. Ed. 2015, 54, 8980-3. 156. Galan, A.; Escudero-Adan, E. C.; Ballester, P., Template- directed self-assembly of dynamic covalent capsules with polar interiors. Chem. Sci. 2017, 8, 7746-7750. 157. Facciotti, C.; Saggiomo, V.; Bunschoten, A.; Fokkink, R.; Hove, J. B. T.; Wang, J.; Velders, A. H., Cyclodextrin-based complex coacervate core micelles with tuneable supramolecular host-guest, metal-to- ligand and charge interactions. Soft Matter 2018, 14, 9542-9549. 158. Cvrtila, I.; Fanlo-Virgos, H.; Schaeffer, G.; Monreal Santiago, G.; Otto, S., Redox Control over Acyl Hydrazone Photoswitches. J. Am. Chem. Soc. 2017, 139, 12459-12465. 159. Boterashvili, M.; Shirman, T.; Popovitz-Biro, R.; Wen, Q.; Lahav, M.; van der Boom, M. E., Nanocrystallinity and direct cross-linkage as key-factors for the assembly of gold nanoparticle-superlattices. Chem. Commun. 2016, 52, 8079-82.

82 160. Carpenter, J. P.; McTernan, C. T.; Ronson, T. K.; Nitschke, J. R., Anion Pairs Template a Trigonal Prism with Disilver Vertices. J. Am. Chem. Soc. 2019, 141, 11409-11413. 161. Ren, Y.; Xie, S.; Svensson Grape, E.; Inge, A. K.; Ramstrom, O., Multistimuli-Responsive Enaminitrile Molecular Switches Displaying H(+)- Induced Aggregate Emission, Metal Ion-Induced Turn-On Fluorescence, and Organogelation Properties. J. Am. Chem. Soc. 2018, 140, 13640-13643. 162. Kieffer, M.; Pilgrim, B. S.; Ronson, T. K.; Roberts, D. A.; Aleksanyan, M.; Nitschke, J. R., Perfluorinated Ligands Induce Meridional Metal Stereochemistry to Generate M8L12, M10L15, and M12L18 Prisms. J. Am. Chem. Soc. 2016, 138, 6813-21. 163. Lehn, J. M., Perspectives in chemistry--steps towards complex matter. Angew. Chem. Int. Ed. 2013, 52, 2836-50. 164. Frei, P.; Hevey, R.; Ernst, B., Dynamic Combinatorial Chemistry: A New Methodology Comes of Age. Chem. Eur. J. 2019, 25, 60-73. 165. Zhang, Y.; Barboiu, M., Constitutional Dynamic Materials-- Toward Natural Selection of Function. Chem. Rev. 2016, 116, 809-34. 166. Lehn, J. M., Perspectives in chemistry--aspects of adaptive chemistry and materials. Angew. Chem. Int. Ed. 2015, 54, 3276-89. 167. Miljanic, O. S., Small-Molecule Systems Chemistry. Chem 2017, 2, 502-524. 168. Jin, Y.; Yu, C.; Denman, R. J.; Zhang, W., Recent advances in dynamic covalent chemistry. Chem. Soc. Rev. 2013, 42, 6634-54. 169. Barboiu, M., Constitutional Dynamic Chemistry. Springer Verlag: Berlin Heidelberg, 2012. 170. Miller, B. L., Dynamic Combinatorial Chemistry: In Drug Discovery, Bioorganic Chemistry and Materials Science. John Wiley & Soncs, Inc. : Hoboken, NJ, 2010. 171. Zou, W.; Dong, J.; Luo, Y.; Zhao, Q.; Xie, T., Dynamic Covalent Polymer Networks: from Old Chemistry to Modern Day Innovations. Adv. Mater. 2017, 29. 172. Hirsch, A. K.; Buhler, E.; Lehn, J. M., Biodynamers: self- organization-driven formation of doubly dynamic proteoids. J. Am. Chem. Soc. 2012, 134, 4177-83. 173. Timmer, B. J. J.; Ramstrom, O., Acid-Assisted Direct Olefin Metathesis of Unprotected Carbohydrates in Water. Chem. Eur. J. 2019, 25, 14408-14413. 174. Furikado, Y.; Nagahata, T.; Okamoto, T.; Sugaya, T.; Iwatsuki, S.; Inamo, M.; Takagi, H. D.; Odani, A.; Ishihara, K., Universal reaction mechanism of boronic acids with diols in aqueous solution: kinetics and the basic concept of a conditional formation constant. Chem. Eur. J. 2014, 20, 13194-202. 175. Bull, S. D.; Davidson, M. G.; van den Elsen, J. M.; Fossey, J. S.; Jenkins, A. T.; Jiang, Y. B.; Kubo, Y.; Marken, F.; Sakurai, K.; Zhao, J.;

83 James, T. D., Exploiting the reversible covalent bonding of boronic acids: recognition, sensing, and assembly. Acc. Chem. Res. 2013, 46, 312-26. 176. Teichert, J. F.; Mazunin, D.; Bode, J. W., Chemical sensing of polyols with shapeshifting boronic acids as a self-contained sensor array. J. Am. Chem. Soc. 2013, 135, 11314-21. 177. Martinez-Aguirre, M. A.; del Campo, J. M.; Escalante-Tovar, S.; Yatsimirsky, A. K., Self-assembly and recognition properties of a tetraanionic macrocyclic boronate ester in aqueous medium. RSC Adv. 2015, 5, 30075-30083. 178. Zhao, F. L.; Dong, A. J.; Deng, L. D.; Guo, R. W.; Zhang, J. H., Morphology control and property design of boronate dynamic nanostructures. Polym. Chem. 2019, 10, 2436-2446. 179. Pascal, R.; Pross, A.; Sutherland, J. D., Towards an evolutionary theory of the origin of life based on kinetics and thermodynamics. Open Biol. 2013, 3, 130156. 180. Dill, K. A.; Ozkan, S. B.; Shell, M. S.; Weikl, T. R., The protein folding problem. Annu. Rev. Biophys. 2008, 37, 289-316. 181. Yadav, G. D.; Deepa; Singh, S., Prolinamide‐Catalysed Asymmetric Organic Transformations. ChemistrySelect 2019, 4, 5591-5618. 182. Karasawa, T.; Oriez, R.; Kumagai, N.; Shibasaki, M., anti- Selective Catalytic Asymmetric Nitroaldol Reaction of alpha-Keto Esters: Intriguing Solvent Effect, Flow Reaction, and Synthesis of Active Pharmaceutical Ingredients. J. Am. Chem. Soc. 2018, 140, 12290-12295. 183. Yamada, T.; Kuwata, M.; Takakura, R.; Monguchi, Y.; Sajiki, H.; Sawama, Y., Organocatalytic Nitroaldol Reaction Associated with Deuterium-Labeling. Adv. Synth. Catal. 2018, 360, 637-641. 184. Alegre-Requena, J. V.; Marques-Lopez, E.; Herreraa, R. P., Organocatalyzed Enantioselective Aldol and Henry Reactions Starting from Benzylic Alcohols. Adv. Synth. Catal. 2018, 360, 124-129. 185. Wang, S.; Guo, Z.; Chen, S.; Jiang, Y.; Zhang, F.; Liu, X.; Chen, W.; Sheng, C., Organocatalytic Asymmetric Synthesis of Spiro- Tetrahydrothiophene Oxindoles Bearing Four Contiguous Stereocenters by One-Pot Michael-Henry-Cascade-Rearrangement Reactions. Chem. Eur. J. 2018, 24, 62-66. 186. Rossi, S.; Porta, R.; Brenna, D.; Puglisi, A.; Benaglia, M., Stereoselective Catalytic Synthesis of Active Pharmaceutical Ingredients in Homemade 3D-Printed Mesoreactors. Angew. Chem. Int. Ed. 2017, 56, 4290- 4294. 187. Bora, P. P.; Bez, G., Henry Reaction in Aqueous Media at Neutral pH. Eur. J. Org. Chem. 2013, 2013, 2922-2929. 188. Lai, G.; Guo, F.; Zheng, Y.; Fang, Y.; Song, H.; Xu, K.; Wang, S.; Zha, Z.; Wang, Z., Highly enantioselective Henry reactions in water catalyzed by a copper tertiary amine complex and applied in the synthesis of (S)-N-trans-feruloyl octopamine. Chem. Eur. J. 2011, 17, 1114-7.

84 189. Fan, J.; Sun, G.; Wan, C.; Wang, Z.; Li, Y., Investigation of DNA as a catalyst for Henry reaction in water. Chem. Commun. 2008, 10.1039/b805767b, 3792-4. 190. Ballini, R.; Bosica, G., Nitroaldol Reaction in Aqueous Media: An Important Improvement of the Henry Reaction. J. Org. Chem. 1997, 62, 425- 427. 191. Zhang, Y.; Ramstrom, O., Thiazolidinones derived from dynamic systemic resolution of complex reversible-reaction networks. Chem. Eur. J. 2014, 20, 3288-91. 192. Vongvilai, P.; Larsson, R.; Ramstrom, O., Direct asymmetric dynamic kinetic resolution by combined lipase catalysis and nitroaldol (Henry) reaction. Adv. Synth. Catal. 2008, 350, 448-452. 193. Angelin, M.; Fischer, A.; Ramstrom, O., Crystallization- induced secondary selection from a tandem driven dynamic combinatorial resolution process. J. Org. Chem. 2008, 73, 3593-5. 194. Zhang, Y. Dynamic Systems: Enzymatic Synthesis, Exchange Reactions and Applications in Materials Science. Doctoral, Royal Institute of Technology, Stockholm, Sweden, 2015. 195. Della Sala, F.; Neri, S.; Maiti, S.; Chen, J. L.; Prins, L. J., Transient self-assembly of molecular nanostructures driven by chemical fuels. Curr. Opin. Biotechnol. 2017, 46, 27-33. 196. Chen, J. L.; Maiti, S.; Fortunati, I.; Ferrante, C.; Prins, L. J., Temporal Control over Transient Chemical Systems using Structurally Diverse Chemical Fuels. Chem. Eur. J. 2017, 23, 11549-11559. 197. Della Sala, F.; Maiti, S.; Bonanni, A.; Scrimin, P.; Prins, L. J., Fuel-Selective Transient Activation of Nanosystems for Signal Generation. Angew. Chem. Int. Ed. 2018, 57, 1611-1615. 198. Boekhoven, J.; Brizard, A. M.; Kowlgi, K. N.; Koper, G. J.; Eelkema, R.; van Esch, J. H., Dissipative self-assembly of a molecular gelator by using a chemical fuel. Angew. Chem. Int. Ed. 2010, 49, 4825-8. 199. Boekhoven, J., Complexity from small molecules. Nat. Nanotechnol. 2018, 13, 979-980. 200. Nguyen, M.; Vaikuntanathan, S., Design principles for nonequilibrium self-assembly. Proc. Natl. Acad. Sci. U. S. A. 2016, 113, 14231- 14236. 201. Pappas, C. G.; Sasselli, I. R.; Ulijn, R. V., Biocatalytic Pathway Selection in Transient Tripeptide Nanostructures. Angew. Chem. Int. Ed. 2015, 54, 8119-23. 202. Grzybowski, B. A.; Huck, W. T., The nanotechnology of life- inspired systems. Nat. Nanotechnol. 2016, 11, 585-92. 203. Del Grosso, E.; Ragazzon, G.; Prins, L. J.; Ricci, F., Fuel- Responsive Allosteric DNA-Based Aptamers for the Transient Release of ATP and Cocaine. Angew. Chem. Int. Ed. 2019, 58, 5582-5586.

85 204. Ciaccia, M.; Cacciapaglia, R.; Mencarelli, P.; Mandolini, L.; Di Stefano, S., Fast transimination in organic solvents in the absence of proton and metal catalysts. A key to imine metathesis catalyzed by primary amines under mild conditions. Chem. Sci. 2013, 4. 205. Nowotny, J.; Bak, T.; Burg, T., Electrical properties of polycrystalline TiO2: Equilibration kinetics. Ionics 2007, 13, 71-78. 206. Crisalli, P.; Kool, E. T., Water-soluble organocatalysts for hydrazone and oxime formation. J. Org. Chem. 2013, 78, 1184-9. 207. Ando, K.; Shimazu, Y.; Seki, N.; Yamataka, H., Kinetic study of proton-transfer reactions of phenylnitromethanes. Implication for the origin of nitroalkane anomaly. J. Org. Chem. 2011, 76, 3937-45. 208. Kresge, A. J., The Nitroalkane Anomaly. Can. J. Chem. 1974, 52, 1897-1903. 209. Claridge, T. D. W., Diffusion NMR Spectroscopy. In High- Resolution NMR Techniques in Organic Chemistry, Claridge, T. D. W., Ed. Elsevier: Boston, 2016, 10.1016/b978-0-08-099986-9.00010-5, p 381-419. 210. Kalapos, M. P., A possible evolutionary role of formaldehyde. Exp. Mol. Med. 1999, 31, 1-4. 211. Burgos-Barragan, G.; Wit, N.; Meiser, J.; Dingler, F. A.; Pietzke, M.; Mulderrig, L.; Pontel, L. B.; Rosado, I. V.; Brewer, T. F.; Cordell, R. L.; Monks, P. S.; Chang, C. J.; Vazquez, A.; Patel, K. J., Mammals divert endogenous genotoxic formaldehyde into one-carbon metabolism. Nature 2017, 548, 549. 212. Brachvogel, R.-C.; Hampel, F.; von Delius, M., Self-assembly of dynamic orthoester cryptates. Nat Commun 2015, 6. 213. Lehn, J.-M., Perspectives in Chemistry—Aspects of Adaptive Chemistry and Materials. Angewandte Chemie International Edition 2015, 10.1002/anie.201409399, n/a-n/a. 214. Monosson, E., Metal Planet. In Evolution in a Toxic World, Island Press/Center for Resource Economics: Washington, DC, 2012, 10.5822/978-1-61091-221-1_4, p 48-64. 215. Antonio, J. P. M.; Russo, R.; Carvalho, C. P.; Cal, P.; Gois, P. M. P., Boronic acids as building blocks for the construction of therapeutically useful bioconjugates. Chem. Soc. Rev. 2019, 48, 3513-3536. 216. Ruiz-Mirazo, K.; Briones, C.; de la Escosura, A., Prebiotic systems chemistry: new perspectives for the origins of life. Chem. Rev. 2014, 114, 285-366. 217. Abraham, A. T.; Zhou, X.; Hecht, S. M., Metallobleomycin- mediated cleavage of DNA not involving a threading-intercalation mechanism. J. Am. Chem. Soc. 2001, 123, 5167-75. 218. Drozdz, W.; Walczak, A.; Bessin, Y.; Gervais, V.; Cao, X. Y.; Lehn, J. M.; Ulrich, S.; Stefankiewicz, A. R., Multivalent Metallosupramolecular Assemblies as Effective DNA Binding Agents. Chem. Eur. J. 2018, 24, 10802-10811.

86 219. Galindo-Murillo, R.; Garcia-Ramos, J. C.; Ruiz-Azuara, L.; Cheatham, T. E., 3rd; Cortes-Guzman, F., Intercalation processes of copper complexes in DNA. Nucleic Acids Res. 2015, 43, 5364-76. 220. Karmakar, S.; Harcourt, E. M.; Hewings, D. S.; Scherer, F.; Lovejoy, A. F.; Kurtz, D. M.; Ehrenschwender, T.; Barandun, L. J.; Roost, C.; Alizadeh, A. A.; Kool, E. T., Organocatalytic removal of formaldehyde adducts from RNA and DNA bases. Nat. Chem. 2015, 7, 752-758. 221. Lipscomb, L. A.; Zhou, F. X.; Presnell, S. R.; Woo, R. J.; Peek, M. E.; Plaskon, R. R.; Williams, L. D., Structure of DNA-porphyrin complex. Biochemistry 1996, 35, 2818-23. 222. Pages, B. J.; Ang, D. L.; Wright, E. P.; Aldrich-Wright, J. R., Metal complex interactions with DNA. Dalton Trans. 2015, 44, 3505-26. 223. Rokita, S. E.; Burrows, C. J., Salen–Metal Complexes. In Small Molecule DNA and RNA Binders, 2002, 10.1002/3527601783.ch6, p 126-145. 224. Takezawa, Y.; Duprey, J.-L.; Shionoya, M., Metal-Aided Construction of Unusual DNA Structural Motifs. In DNA in Supramolecular Chemistry and Nanotechnology, 2017, 10.1002/9781118696880.ch1.5, p 65-77. 225. Tan, J. D.; Hudson, S. E.; Brown, S. J.; Olmstead, M. M.; Mascharak, P. K., Syntheses, structures, and reactivities of synthetic analogs of the three forms of cobalt(III)-bleomycin: proposed mode of light-induced DNA damage by the cobalt(III) chelate of the drug. J. Am. Chem. Soc. 1992, 114, 3841-3853. 226. Uluçcam, G.; Beynek, N.; Seller, Z.; Akalın, G.; Turan, G.; Benkli, K., Synthesis, Characterization of Some Transition-Metal Complexes of a New Heptadentate N5S2 Schiff-Base Ligand and the Effects of These Metal Complexes on U2OS Cells Cytotoxicity and DNA Cleavage Activity. Phosphorus, Sulfur, and Silicon and the Related Elements 2008, 183, 2237- 2247. 227. Zeglis, B. M.; Pierre, V. C.; Barton, J. K., Metallo-intercalators and metallo-insertors. Chem. Commun. 2007, 10.1039/b710949k, 4565-79. 228. Zhu, J.; Haynes, C. J. E.; Kieffer, M.; Greenfield, J. L.; Greenhalgh, R. D.; Nitschke, J. R.; Keyser, U. F., Fe(II)4L4 Tetrahedron Binds to Nonpaired DNA Bases. J. Am. Chem. Soc. 2019, 141, 11358-11362. 229. Zeglis, B. M.; Barton, J. K., A mismatch-selective bifunctional rhodium-Oregon Green conjugate: a fluorescent probe for mismatched DNA. J. Am. Chem. Soc. 2006, 128, 5654-5. 230. Duprey, J. L.; Carr-Smith, J.; Horswell, S. L.; Kowalski, J.; Tucker, J. H., Macrocyclic Metal Complex-DNA Conjugates for Electrochemical Sensing of Single Nucleobase Changes in DNA. J. Am. Chem. Soc. 2016, 138, 746-9. 231. Lippert, B.; Sanz Miguel, P. J., The Renaissance of Metal- Pyrimidine Nucleobase Coordination Chemistry. Acc. Chem. Res. 2016, 49, 1537-45.

87 232. Xu, D.; Landon, T.; Greenbaum, N. L.; Fenley, M. O., The electrostatic characteristics of G.U wobble base pairs. Nucleic Acids Res. 2007, 35, 3836-47. 233. Grew, E. S.; Bada, J. L.; Hazen, R. M., Borate minerals and origin of the RNA world. Orig. Life Evol. Biosph. 2011, 41, 307-16. 234. Kong, L.; Setlow, P.; Li, Y. Q., Analysis of the Raman spectra of Ca(2+)-dipicolinic acid alone and in the bacterial spore core in both aqueous and dehydrated environments. Analyst 2012, 137, 3683-9. 235. Cortese, M. S.; Paszczynski, A.; Lewis, T. A.; Sebat, J. L.; Borek, V.; Crawford, R. L., Metal chelating properties of pyridine-2,6- bis(thiocarboxylic acid) produced by Pseudomonas spp. and the biological activities of the formed complexes. Biometals 2002, 15, 103-20. 236. Aylward, N.; Bofinger, N., A plausible prebiotic synthesis of pyridoxal phosphate: vitamin B6 - a computational study. Biophys. Chem. 2006, 123, 113-21. 237. Austin, S. M.; Waddell, T. G., Prebiotic synthesis of vitamin B6-type compounds. Orig. Life Evol. Biosph. 1999, 29, 287-96. 238. Karalius, A.; Zhang, Y.; Kravchenko, O.; Elofsson, U.; Szabo, Z.; Yan, M.; Ramstrom, O., Formation and Out-of-Equilibrium, High/Low State Switching of a Nitroaldol Dynamer in Neutral Aqueous Media. Angew. Chem. Int. Ed. 2019, n/a. 239. Feuer, H.; Lawrence, J. P., Alkyl nitrate nitration of active methylene compounds. IX. Nitration of alkyl substituted heterocyclic compounds. J. Org. Chem. 1972, 37, 3662-3670. 240. Sylvain, C.; Wagner, A.; Mioskowski, C., An efficient procedure for the esterification of nitroacetic acid: Application to the preparation of Merrifield resin-bound nitroacetate. Tetrahedron Lett. 1999, 40, 875-878. 241. The Stereochemistry of the Mechanical Bond. In The Nature of the Mechanical Bond, 2016, 10.1002/9781119044123.ch5, p 471-554. 242. Boros, E.; Ferreira, C. L.; Cawthray, J. F.; Price, E. W.; Patrick, B. O.; Wester, D. W.; Adam, M. J.; Orvig, C., Acyclic chelate with ideal properties for (68)Ga PET imaging agent elaboration. J. Am. Chem. Soc. 2010, 132, 15726-33. 243. Bains, G. K.; Kim, S. H.; Sorin, E. J.; Narayanaswami, V., The extent of pyrene excimer fluorescence emission is a reflector of distance and flexibility: analysis of the segment linking the LDL receptor-binding and tetramerization domains of apolipoprotein E3. Biochemistry 2012, 51, 6207-19. 244. Perry, M.; Carra, C.; Chretien, M. N.; Scaiano, J. C., Effect of hexafluorobenzene on the photophysics of pyrene. J. Phys. Chem. A 2007, 111, 4884-9. 245. Tickell, D. A.; Lampard, E. V.; Lowe, J. P.; James, T. D.; Bull, S. D., A Protocol for NMR Analysis of the Enantiomeric Excess of Chiral Diols Using an Achiral Diboronic Acid Template. J. Org. Chem. 2016, 81, 6795-9.