Information transfer through base pairing in self-assembly driven self-replication

(1)

Information transfer through base pairing in self-assembly driven self-replication

MSc Thesis

Stratingh Institute for Chemistry University of Groningen

By

Priscilla F. Pieters

Supervisors

Prof. Dr. S. Otto, Dr. Charalampos Pappas

17/07/2018

(2)

1

One of the grand challenges in science, is understanding how biology transitions into chemistry.

From investigating the origin of life, to designing a fully synthetic chemical system to create life-like behavior, taking into the account the essential elements of life: replication, metabolism, selection, compartmentalization, and information transfer, is of great importance. However, combining these elements into one synthetic system has proven a challenging task. A promising way to achieve and study such a system is by dynamic combinatorial chemistry.

With this objective in mind, a system was designed to use self-replication and base-pairing to achieve information transfer and self-assembly in nucleobase amino acid chimeric building block containing systems. These building blocks can form macrocycles via their dithiol core and stack into fibers.

The fibers exhibit exponential growth via an elongation/breakage mechanism, creating a self-replicator, for which proof was obtained by seeding experiments. The systems under consideration were studied by UPLC, LC-MS, TEM and ITC.

The results obtained illustrate that in select libraries of these nucleobase amino acid chimeric building blocks, an influence of base-pairing can be observed when the library is templated with complementary DNA. ITC data shows that cytosine-containing fibers bind to complementary DNA of 10 nucleobases long. In addition, mixed libraries of cytosine- and thymine-containing building blocks showed a shift in product distribution when being templated by complementary DNA.

Mixing these promising building blocks with the well-established XGLKFK building block, showed that even in these mixtures, base-pairing could be observed. Besides, when combining two nucleobase building blocks with XGLKFK in excess, a state could be achieved where there is a co-existence of mixed XGLKFK with cytosine building block trimers, and mixed XGLKFK with thymine building block hexamers. This co-existence is made possible because the different species feed upon different food sources.

Although this research does not yet provide a straightforward synthetic system where self- replication is combined with information transfer via base-pairing, it represents an important first step forward to achieving this objective.

(3)

2

Chapter 1 Towards de-novo life 1.1 Introduction

Understanding the transition of chemistry into biology is one of the great challenges that we are facing nowadays. To understand how biology originated from abiotic molecules and how biology can be created from chemistry, we need to understand what ingredients are essential for life, and how these can be combined into one system. Looking into the origin of life, and understanding how these essential elements, replication, metabolism, selection, and compartmentalization, came together in one chemical system, can give inspiration towards the design of fully synthetic chemical systems to create life-like behavior.

1.2 Origin of life

In order to answer questions about the origin of life and to go towards mimicking life in synthetic systems, first of all we need to understand what life actually is. A commonly used definition of life was formulated by NASA and it states that: ‘Life is a self-sustained chemical system capable of undergoing Darwinian evolution’.¹ However, a definition like this leaves out life-forms that we would intuitively identify as living (are infertile mammals, like a mule, not alive?). In addition, none of the definitions used for life give any insight into how a living system can be created from inert matter and what actually makes it alive.

Therefore, instead of focusing on a definition of life, we can define certain characteristic that are necessary for life to exist. The main prerequisites for life as we know it, are for it to be compartmentalized, to have a metabolism, to exhibit selectivity, and to be able to self-reproduce.² In addition, it should be capable of Darwinian evolution, meaning that the system should be able to adapt by for example errors in the information transfer, but it should also retain enough information during its replication. In this project, the focus will be on the aspects of self-replication and information transfer.

1.2.1 The RNA world hypothesis

Life as we know it, utilizes proteins as catalysts, and DNA and RNA as information biomolecules. However, when we want to establish which of these originated first, we run into a problem similar to the chicken- and-egg paradox: ‘What came first?’. Proteins on the one hand, are synthesized based on DNA templates.

On the other hand, DNA is synthesized by proteins. Therefore, DNA and proteins seem to be codependent.

One hypothesis provided to solve this problem is the so called ‘RNA world hypothesis’, where RNA is considered an important precursor to life.³ This hypothesis is based on the finding that, besides the fact that RNA can act as information carrier, it has the unique potential to also act as an enzyme-like catalyst, for the discovery of which Altman and Cech were rewarded the Nobel Prize in 1989.^4,5 Regarding that in modern day life nucleic acids carry the information, as the base-pairing allows for information transfer as well as replication, biopolymers containing these nucleobases such as DNA and RNA are most probably the first reproducing molecules. Even though there are differing views on how exactly the RNA world existed, it is built on three basic assumptions: At some point in the evolution timeline, genetic continuity was achieved by the replication of RNA; Watson-Crick base-pairing was the mechanism of information transfer and mediator of this replication; There were no genetically encoded proteins involved as catalyst.⁶

(6)

5

However promising this hypothesis seems, there are still quite a few challenges to be solved to be able to prove the existence of such a ‘RNA world’. Even though having some catalytic activity, the activity of RNA is quite low especially considering its ligase abilities. One important problem is the emergence of RNA in prebiotic conditions, as the synthesis and polymerization of nucleotides under these conditions has proven to be quite difficult. Therefore, RNA is most probably already the product of evolution.

As nowadays the functions of catalysis and information transfer are found separately in peptides and DNA, an alternative hypothesis to the origin of life is the idea that nucleic acids and peptides evolved side-by- side.^7,8Such systems are only now starting to be explored, and offer a promising alternative for the origin of life from prebiotic chemistry. However, independent of the hypothesis on the formation of life, is the central role and importance of self-replication in these evolutionary scenarios.

1.2.2 Self-replication

An essential property of living organisms is their ability to make copies of their genetic information through replication. DNA replication involves the template-directed synthesis of a new complementary strand of DNA through templating of the parent strand. However, this template-directed synthesis is mediated by a large number of enzymes (which are produced by information encoded in the DNA itself). In prebiotic conditions, this complex biomolecular machinery would not have been present, and simpler pathways towards self-replication would have to be in place.

When designing a fully synthetic system containing self-replicating molecules, the most typical and simple design is the templated ligation of two halves of the replicator, to produce a dimer of the autocatalyst in duplex form, which can subsequently dissociate. This simplest form of a self-replicating system is called a minimal self-replicating system.⁹ A schematic overview of this kind of replication is shown in Figure 1A . The template molecule T is self-complementary, and can be formed by the ligation of the two parts A and B. First, A and B reversibly bind to template T, to yield a termolecular complex. As the reactive ends of the molecules A and B are now in close proximity, the covalent bond formation between A and B to form T is facilitated. This results in a duplex of two T molecules, that can reversibly dissociate to give two template molecules T, each of which can on its own start a new replication cycle.

Figure 1. (A) A schematic overview of the mechanism of a minimal self-replicating system. T is the template molecule, and A and B are two complementary parts to T that can form T by ligation, reproduced from Ref [10], and (B) kinetic profiles of exponential and parabolic growth of self-replicating molecules, reproduced from Ref [11].

However, there are some obstacles that need to be overcome, and have been proven difficult in this type of autocatalytic system. A very important requirement for a replicator in order to allow for Darwinian

A B

(7)

6

evolution, and thus selection and survival of the fittest when there are competing replicators, is an exponential and not parabolic growth of the replicator.¹²

In the ideal case, where the autocatalytic reaction is exponential, the order of the reaction p = 1. However, in these template-based self-replication reactions, this exponential growth is often limited by the fact that the dimer duplex is thermodynamically more stable than the termolecular complex, most of the times because of entropic reasons, while the interactions of the termolecular and duplex complex stay the same.

This leads to slow duplex dissociation and thus product inhibition, as there is a reduced concentration of free catalyst, resulting in a parabolic growth where p = ½. This is called the square-root law of autocatalysis.¹³Proof of autocatalysis can be obtained by performing seeding experiments. By comparing the initial reaction rates to the initial concentration of the seeded template, the parameter p can be determined.¹¹

In terms of competing replicators, the difference between parabolic and exponential growth is of vital importance. In the case of parabolic growth, there will be an indefinite coexistence of different replicators feeding upon the same food. In this case, you will have survival of everybody, and as there is no extinction possible, it is limiting for evolution purposes. When exponential growth can be achieved in the competition of multiple replicators feeding upon the same food, there is an extinction of the ‘weakest’ replicator. This allows for survival of the fittest, and thus proper Darwinian evolution.¹⁴

1.2.3 Information transfer

A crucial part of the origin of life is the propagation of information in the self-replicating system. The transfer of information is often closely related to the autocatalytic cycle, as the formation of a new molecule of itself should contain the information that was contained in the parent molecule. This information transfer should take place with a high fidelity, but should also allow for errors and mutations, in order to allow for adaptation to changes in the environment.

This balance between information retention and information change by mutation can be described by the Eigen limit.¹⁵ The higher the rate of error occurrence is, the shorter the sequence of information is that can be retained during replication. With too many errors per replication, there is too much loss of information and after a few replication cycles, no information will be retained. In the case of RNA and DNA, in order to achieve a long sequence (> 100 bases), an error correction enzyme is needed. However, in order to encode for such an enzyme, a sequence larger than 100 bases is required.² The solution to this so called Eigen paradox is still being debated.

Even though information in a system can be contained in a variety of ways, such as macrocycle size and composition, side chain length, and supramolecular assembly, the most common way to properly transfer information is by base-pairing in biopolymers, such as in RNA and DNA. This kind of information transfer is of great importance in life and in the origin of life. Life has evolved nucleic acids arranged in a specific sequence in the biopolymer as the means to store and propagate information.

(8)

7

1.2.4 Peptide Nucleic Acids

Besides RNA and DNA, peptide nucleic acids (PNAs) are of great interest when considering information transfer via base-pairing. PNA is a synthetic nucleic acid, introduced in 1991 by Nielsen et al..¹⁶ PNAs exhibit stronger base-pairing interactions than DNA or RNA, and in addition a PNA/DNA duplex is more stable than a DNA/DNA duplex. Besides, it shows great specificity for base-pairing.¹⁷ The added stability can be explained by the fact that the backbone of PNA does not contain charged phosphate groups, as is the case for RNA and DNA, and thus the binding in not hindered by electrostatic repulsion. In addition, functional groups can be added to the backbone, allowing for tuning of the properties of the molecule.¹⁸ These advantages make PNA of great interest as a biopolymer containing nucleobases in order to create synthetic life-like systems. In addition, it is even hypothesized that PNA can be a precursor in the origin of life for DNA and RNA.¹⁹

1.3 Base-pairing in self-replicating supramolecular systems

The combination of synthetic supramolecular systems with biomolecular parts such as amino acids and nucleobases are a promising way to advance towards de-novo life. There have been a few examples of supramolecular systems that use base-pairing in their interactions for replication, without the use of enzymes, however these systems are still far from combining the essential elements of life into one system.

The first example of a non-enzymatic self-replicating system containing nucleobases was reported by von Kiedrowski, 1986.²⁰ He showed autocatalysis in a template-directed condensation of two trinucleotides in the presence of EDC, leading to a hexameric template with a palindromic sequence, shown in Figure 2. The reaction follows the minimal self-replicating system, exhibiting parabolic growth with p = ½, and cannot be used for long sequences. Moreover, the uncatalyzed pathway of condensation actually dominates over the autocatalytic pathway.

Figure 2. The self-replication of the palindromic hexamer 3 by EDC mediated ligation of 1 and 2 of which the reactive ends are brought into close proximity when bound to 3, reproduced from Ref [21].

(9)

8

In addition, Lincoln and Joyce were able to achieve replication with RNA oligonucleotides in a cross- catalytic way, using the ligation of two halves of the template to create the complementary strand to the template molecule.²² These examples from von Kiedrowski and Joyce show the ability of self-replication with oligonucleotides using base-pairing interactions. However, these systems require pre-synthesized oligonucleotides.

In 1987, an autocatalytic system consisting of nucleic acid-like oligomers bearing an artificial backbone structure was reported by Zielinksy and Orgel.²³ Two diribonucleotide analogues were ligated using a self- complementary tetraribonucleotide template in the presence of EDC. This system, based on base-pairing with PNAs, exhibits parabolic growth with p = ½.

The next step, achieving self-replication in fully synthetic systems was demonstrated by von Kiedrowski in 1992, where they condense3-amino-benzamidines and 2-formylphenoxyacetic acids using a amidinium carboxylate salt bridge resembling base-pairing interactions as the non-covalent templating interaction.²⁴ This first step toward synthetic supramolecular replicating systems can elaborate the structural, functional, and dynamic principles that are important in the origin of life and de-novo life. However, also this system exhibits parabolic growth, limiting usage for evolution purposes.

1.3.1 Overcoming product inhibition

Even though autocatalytic systems using base-paring have been achieved, the major challenge is to overcome the product inhibition. One way to approach this, is by destabilizing the template duplex, in order to free the product and allow for proper exponential growth. An example of systems using the difference in spatial arrangement of intermediate and product in order to destabilize the duplex has been reported by Chmielewski in 2003.²⁵ The system presented exhibits p = 0.91, even though still being sub- exponential growth, this is a big step forward to p = 1. A helical peptide is used that can produce a coiled- coil structure of two strands, in which a proline is incorporated. Proline is known to produce a pronounced kink in an α-helix, bending the helical structure. When they incorporated this kink at the centre of the template, the product of the self-replication reaction has a reduced affinity for the template, resulting in significant less product inhibition.

The first exponential non-enzymatic amplification of oligonucleotides was realized by Luther et al. in 1998, using a replication procedure called SPREAD (Surface Promoted Replication and Exponential Amplification of DNA analogues).²⁶ A DNA strand immobilized on a support is hybridized with complimentary oligonucleotides, which are ligated by EDC. Next the strands are liberated, and subsequently immobilized separately again. Even though this procedure allows for exponential growth, this is a non-autonomous system.

Although these systems show some promise towards synthetic self-replicating systems, specifically containing nucleobases, there has still not been an example exhibiting autonomous, exponential growth.

In addition, it only uses the ability of base-pairing to a limited extent, as selectivity is not an issue regarding that there is no diversity of the options presented.

(10)

9

1.4 Dynamic combinatorial chemistry

One way to obtain self-replicators from simple starting materials, and allow for facile investigation of information transfer and templating effects is by using dynamic combinatorial chemistry. Dynamic combinatorial chemistry makes use of reversible bonds, resulting in the fact that in a dynamic combinatorial library (DCL), all constituents are in equilibrium as it is under thermodynamic control.²⁷ DCLs are a great tool for determining the thermodynamically most stable library member under certain conditions. This can help identify either intramolecular stabilizing forces such as in foldamers, or it can be used to determine the interactions of the library members when considering intermolecular interactions, such as biasing towards the library members that can form stable supramolecular structures. Moreover, the response of a DCL towards external influences can be investigated, such as temperature, pH, or even molecular recognition.^27,28

It is important to note that, when studying templating effects, even though in most cases the compound that binds the strongest is amplified, a DCL strives towards to lowest overall Gibbs energy. As all the library members are connected through equilibrium reactions, the templating effect will be represented in the complete product distribution, as all the members are interconnected. Therefore the final equilibrium distribution is the thermodynamically most stable product mixture.

1.4.1 Reversible exchange reactions

There are a variety of covalent exchange reactions that can be utilized in order to create a dynamic combinatorial library. One of these is the C=N exchange, resulting in e.g. an imine or an acyl hydrazone, with imine exchange being the most commonly used (Figure 3). The imine condensation between an aldehyde and amine is acid catalyzed, therefore imine-based DCLs are often generated at mildly acidic pH (aqueous buffers of around 5.0 pH).²⁷

Figure 3. The imine exchange reaction.

Secondly, CO-X bond exchange with X = S, O, NH, gives a variety of useful reactions that can be applied in dynamic combinatorial chemistry. These are examples of asymmetric exchange reactions, covalently binding two different functional groups. However, the next example couples two of the same functional groups, two thiols, in order to form a disulfide. Subsequent disulfide exchange, as shown in Figure 4, requires a deprotonated thiol, and is therefore very pH dependent and has to be performed under neutral to basic conditions (aqueous buffer of around 7-10 pH).²⁹

Figure 4. The disulfide exchange reaction.

1.4.2 Self-replicating systems using DCC

Using dynamic combinatorial chemistry, exponential growth in a synthetic system was achieved by Otto et al. in 2010.³⁰ A peptide building block was designed, attached to an aromatic dithiol core. Due to the

(11)

10

presence of two thiol groups, through oxidative disulfide formation and exchange macrocycles of a variety of sizes can be formed as illustrated in Figure 5.

Figure 5. Schematic representation of macrocycle formation and interconversion via disulfide exchange starting from a dithiol containing building block, reproduced from Ref [30].

They observed the emergence of a self-replicator from a small dynamic combinatorial library made from a dithiol-functionalized building peptide building block. The peptide sequence features alternating hydrophobic (leucine) and charged (lysine) α-amino acids, that are known to have a high propensity to self-assemble into β-sheet structures through noncovalent interactions.³¹ These β-sheet interactions strengthen the intermolecular interactions, causing stacking of certain macrocycle sizes. The structure of the building block is shown in Figure 6.

Figure 6. Schematic representation of the self-replication mechanism.

In the DCL, the different macrocycles are exchanging building block. Initially trimers and tetramers are the dominant species. However, at some point larger macrocycles such as hexamers nucleate by the

(12)

11

intermolecular peptide-peptide interactions that result in nanostructure formation and the formation of fibers, which are held together by β-sheet interactions. Next, the fibers elongate, resulting in linear growth.

However, when a mechanical force is applied, such as stirring of the DCL, the fibers will fragment, freeing more growing fiber ends and enabling exponential growth with p = 1.^10,30 This elongation/breakage mechanism is illustrated in Figure 6. The macrocycle size of the species that will emerge as replicator depends on the strength of the mechanical force applied, showing the ability of such a self-replicating dynamic system to respond to external stimuli. As this system does not face the challenge of product inhibition, this is the first synthetic system in which autonomous exponential growth is achieved. Proof of the autocatalytic nature of the growth process was achieved by seeding experiments.

1.4.3 DCC for nucleobase-containing supramolecular systems

Recent advances in creating synthetic informational polymers have shown the great importance of using dynamic combinatorial chemistry in order to create nucleobase-containing supramolecular systems. An important step towards creating an informational polymer was achieved by Ghadiri et al. in 2009, when they created a self-assembling sequence-adaptive peptide nucleic acid (PNA).³² Using reversible covalent thioester bonds, they managed to anchor by self-assembly, nucleobases onto a simple oligo-dipeptide backbone, which undergo dynamic sequence modifications in response to changing templates in solution under enzyme-free conditions. The oligomers self-pair with complementary thioester-PNA, and cross-pairs with RNA and DNA through base-pairing. This system shows the advantages of using PNA to create synthetic biopolymers for information transfer purposes.

Figure 7. (A) Mechanism for reversible covalent assembly of the oligomers via anchoring of thioester-derived recognition units onto an oligopeptide backbone, and (B) illustration of oligomer assembly and binding to a complementary oligonucleotide template, reproduced from Ref [32].

More recently Brunsveld et al. showed the advantages of using a combination of oligonucleotides and synthetic supramolecular systems in order to create responsive nucleobase-containing supramolecular wires.³³ Using the stacking of bis-pyridine amphiphilic discotic molecules, that are known to spontaneously self-assemble into supramolecular wires, and combining this with complementary oligonucleotides, the positioning of the discotic molecules in the supramolecular wire can be regulated. It is responsive to stimuli like salt concentration, and changing the number of base pairs. In addition, changing the content of non- functionalized discotic monomers regulates the extent of DNA-duplex formation.

In addition, Gazit showed in 2015 that assemblies of PNAs have the ability to give rise to interesting physical properties such as the emittance of light.³⁴ They achieved ordered self-assembled structures of guanine-containing di-PNA, that coordinates by stacking interactions as well as Watson-Crick base pairing.

The assemblies were found to exhibit optical properties such as voltage-dependent electroluminescence

A B

(13)

12

and wide-range excitation-dependent fluorescence in the visible region. This system thus shows great promise of PNAs in technological applications in bionanotechnology and material science.

(14)

13

1.5 Aim of the project and outline of the thesis

This thesis aims to use self-replication and base-pairing to achieve information transfer and self-assembly in nucleobase amino acid chimeric building block containing systems. The design of the building blocks is a modification of the system described previously by Otto, using a dithiol core to achieve macrocycle formation, fiber formation/elongation/breakage in a dynamic combinatorial library.³⁰ Additionally, a nucleobase is incorporated into the small peptide chain attached to the dithiol core, allowing for base- pairing which can drive and enhance self-replication and information transfer. The systems were studied by UPLC, LC-MS, TEM, and ITC as the main analysis techniques.

In chapter 2, dynamic combinatorial libraries of the nucleobase containing building blocks are analyzed in non-mixed and mixed composition. Additionally, templating effect are investigated by adding short DNA strands to the libraries. The libraries that showed signs of base-pairing were further investigated by NMR and ITC.

In chapter 3, the nucleobase containing building blocks with histidine as amino acid were mixed with the well-established XGLKFK system. The possible effects and enhancement of base-pairing in these systems were investigated by DNA templating experiments. In addition, selected libraries consisting of three different building blocks were investigated.

(15)

14

1.6 References

[1] S. A. Benner, Defining life, Astrobiology, 2010, 10, 1021.

[2] B. M. Rode, D. Fitz, T. Jakschitz, The First Steps of Chemical Evolution towards the Origin of Life, Chem. Biodivers., 2007, 4, 2674.

[3] W. Gilbert, Origin of Life: The RNA World, Nature, 1986, 319, 618.

[4] T. R. Cech, A. J. Zaug, P. J. Grabowski, In Vitro Splicing of the Ribosomal RNA Precursor of

Tetrahymena: Involvement of a Guanosine Nucleotide in the Excision of the Intervening Sequence, Cell, 1981, 27, 487.

[5] B. C. Stark, R. Kole, E. J. Bowman, S. Altman, Ribonuclease P: an Enzyme with an Essential RNA Component, Proc. Natl. Acad. Sci USA., 1978, 75, 3717.

[6] M. P. Robertson, G. F. Joyce, The Origins of the RNA World, Cold Spring Harb Perspect Biol., 2012, 4, a003608.

[7] B. H. Patel, C. Percivalle, D. J. Ritson, C. D. Duffy, J. D. Sutherland, Common Origins of RNA, Protein and Lipid Precursors in a Cyanosulfidic Protometabolism, Nature Chem., 2015, 7, 301.

[8] C. Gibard, S. Bhowmik, M. Karki, E.-K. Kim, R. Krishnamurthy, Phosphorylation, Oligomerization and Self-Assembly in Water under Potential Prebiotic Conditions, Nature Chem., 2018, 10, 212.

[9] H. Duim, S. Otto, Towards Open-Ended Evolution in Self-Replicating Molecular Systems, Beilstein J.

Org. Chem., 2017, 13, 1189.

[10] M. Colomb-Delsuc, E. Mattia, J. W. Sadownik, S. Otto, Exponential Self-Replication Enabled Through a Fibre Elongation/Breakage Mechanism, Nat Commun., 2015, 6, 7427.

[11] Z. Dadon, N. Wagner, G. Ashkenasy, The Road to Non-Enzymatic Molecular Networks, Angew. Chem.

Int. Ed., 2008, 47, 6128.

[12] E. Szathmáry, I. Gladkih, Sub-Exponential Growth and Coexistence of Non-Enzymatically Replicating Templates, J. Theor. Biol., 1989, 138, 55.

[13] G. Von Kiedrowski, Minimal Replicator Theory I: Parabolic Versus Exponential Growth, In: H. Dugas, F. P. Schmidtchen (eds), Bioorganic Chemistry Frontiers, 1993, vol 3. Springer, Berlin, Heidelberg.

[14] I. Scheuring, E. Szathmáry, Survival of Replicators with Parabolic Growth Tendency and Exponential Decay, J. Theor. Biol., 2001, 212, 99.

[15] M. Eigen, Selforganizatino of Matter and the Evolution of Biological Macromolecules, Naturwissenschaften, 1971, 58, 465.

[16] P. E. Nielsen, M. Egholm, R. H. Berg, O. Buchardt, Sequence-Selective Recognition of DNA by Strand Displacement with a Thymine-Substituted Polyamide, Science, 1991, 254, 1497.

(16)

15

[17] M. Egholm, O. Bulchardt, L. Christensen, C. Behrens, S. M. Freier, D. A. Driver, R. H. Berg, S. K. Kim, B.

Norden, P. E. Nielsen, PNA Hybridizes to Complementary Oligonucleotides Obeying the WatsonCrick Hydrogen-Bonding Rules, Nature, 1993, 365, 566.

[18] R. E. Kleiner, Y. Brudno, M. E. Birnbaum, D. R. Liu, DNA-Templated Polymerization of Side-Chain- Functionalized Peptide Nucleic Acid Aldehydes, J. Am. Chem. Soc., 2008, 130, 4646.

[19] S. A. Banack, J. S. Metcalf, L. Jian, D. Craighead, L. L. Ilag, P. A. Cox, Cyanobacteria Produce N-(2- Aminoethyl)Glycine, a Backbone for Peptide Nucleic Acids Which May Have Been the First Genetic Molecules for Life on Earth, PLoS ONE, 2012, 7, e49043.

[20] G. von Kiedrowski, A Self-Replicating Hexadeoxynucleotide, Angew. Chem. Int. Ed., 1986, 25, 932.

[21] V. Patzke, G. von Kiedrowski, Self Replicating Systems, ARKIVOC, 2007, 293.

[22] T. A. Lincoln, G. F. Joyce, Self-Sustained Replication of an RNA Enzyme, Science, 2009, 323, 1229.

[23] W. S. Zielinksi, L. E. Orgel, Autocatalytic Synthesis of a Tetranucleotide Analogue, Nature, 1987, 327, 346.

[24] A. Terfort, G. von Kiedrowski, Self-Replication by Condensation of 3-Aminobenzamidines and 2- Formylphenoxyacetic Acids, Angew. Chem. Int. Ed., 1992, 5, 654.

[25] X. Li, J. Chmielewski, Peptide Self-Replication Enhanced by a Proline Kink, J. Am. Chem. Soc., 2003, 125, 11820.

[26] A. Luther, R. Brandsch, G. von Kiedrowski, Surface-Promoted Replication and Exponential Amplification of DNA Analogues, Nature, 1998, 396, 245.

[27] P. T. Corbett, J. Leclaire, L. Vial, K. R. West, J.-L. Wietor, J. K. M. Sanders, S. Otto, Dynamic Combinatorial Chemistry, Chem. Rev., 2006, 106, 3652.

[28] J. D. Cheeseman, A. D. Corbett, J. L. Gleason, R. J. Kazlauskas, Receptor-Assisted Combinatorial Chemistry: Thermodynamics and Kinetics in Drug Discovery, Chem. Eur. J., 2005, 11, 1708.

[29] S. Otto, R. L. E. Furlan, J. K. M. Sanders, Dynamic Combinatorial Libraries of Macrocyclic Disulfides in Water, J. Am. Chem. Soc., 2000, 122, 12063.

[30] J. M. A. Carnall, C. A. Waudby, A. M. Belenguer, M. C. A. Stuart, J. J.-P. Peyralans, S. Otto, Mechanosensitive Self-Replication Driven by Self-Organization, Science, 2010, 327, 1502.

[31] W. F. DeGrado, J. D. Lear, Induction of Peptide Conformation at Apolar/Water Interfaces. 1. A Study with Model Peptides of Defined Hydrophobic Periodicity, J. Am. Chem. Soc., 1985, 107, 7684.

[32] Y. Ura, J. M. Beierle, L. J. Leman, L. E. Orgel, M. R. Ghadiri, Self-Assembling Sequence-Adaptive Peptide Nucleic Acids, Science, 2009, 325, 73.

(17)

16

[33] M. Á. A. García, E. M. Estirado. L.-G. Milroy, L. Brunsveld, Dual-Input Regulation and Positional Control in Hybrid Oligonucleotide/Discotic Supramolecular Wires, Angew. Chem. Int. Ed., 2018, 57, 4976.

[34] O. Berger, L. Alder-Abramovich, M. Levy-Sakin, A. Grunwald, Y. Liebel-Peer, M. Bacher, L. Buzhansky, E. Mossou, V. T. Forsyth, T. Schwartz, Y. Ebenstein, F. Frolow, L. J. W. Shimon, F. Patolsky, E. Gazit, Light- Emitting Self-Assembled Peptide Nucleic Acids Exhibit Both Stacking Interactions and Watson-Crick Base Pairing, Nature Nanotechnology, 2015, 10, 353.

(18)

17

Chapter 2. The influence of DNA templating on nucleobase containing self-replicators

2.1 Introduction

This thesis aims to use the macrocycle self-replication and base-pairing interaction to drive and enhance information transfer and self-assembly. With this approach, the system is extended from solely peptide- based building blocks to nucleobase-containing building blocks. Using and incorporating these in dynamic combinatorial libraries allows for the investigation of their supramolecular organization and the possibility of information transfer via base-pairing of nucleobase-containing building blocks.

A series of nucleobase functionalized building blocks, containing a dithiol core (X) appended via an amino ethyl glycine linker to nucleobase and terminated by an amino acid, was synthesized (Figure 8). The dithiol core enables dynamic covalent bond formation via oxidation to disulfides, creating macrocycles. The amino acid at the end of the chain gives the opportunity to introduce stabilizing interactions, enhancing supramolecular structure formation. The nucleobase is attached via an amino ethyl glycine linker, making this a peptide nucleic acid (PNA). PNAs have an advantage over DNA or RNA when binding to a DNA strand, as the backbone of a PNA does not contain charged phosphate groups, and thus there is no electrostatic repulsion. This is useful for the purpose of this project, as base-paring between the PNA and a DNA strand with templating experiments is desired.

Figure 8. The design of the nucleobase-containing building blocks. The dithiol core is shown in yellow and denoted X. The functional group R1 is one of the four nucleobases (C, G, T, or A), and R2 is either the amino acid phenylalanine or histidine.

Using dynamic combinatorial chemistry, libraries were prepared of these nucleobase-containing molecules. The libraries were analyzed by UPLC and LC-MS to determine the product distribution over time. In addition, on selected libraries, TEM imaging was performed. First of all, the behavior of the libraries made from a single building block. Secondly, their behavior in mixed building block libraries will be discussed. Additionally, DNA templating experiments were performed to investigate the influence and possibility of base-pairing in these systems. If there are changes observed in the system that are caused by base-pairing, this may enable information transfer in these systems of self-replicating molecules. In the libraries that showed signs of possible base-pairing interactions, follow-up experiments were performed with NMR (STD, DOSY) and ITC.

X

(19)

18

2.2 Results and discussion

The building blocks containing the cytosine nucleobase will be denoted with C, with guanine G, with thymine T, and with adenine A. To non-specifically indicate any of these nucleobases, B will be used. If the amino acid at the end of the building block is phenylalanine, F will be added as subscript (e.g. CF for a cytosine an phenylalanine containing building block). When the amino acid is histidine, H will be added (e.g. GH). When analyzing the product distribution of a library over time, the composition of the different macrocycles will be denoted by the nucleobases the building blocks contain (e.g. a trimer of two GH

molecules and one TH is denoted G2T).

2.2.1 Phenylalanine containing building blocks

First of all, libraries were prepared of the four nucleobase- containing building blocks with phenylalanine as amino acid at 1.0 mM in 50 mM borate buffer (pH = 8.2), and they were monitored over time by UPLC and LC-MS. The concentration of 1.0 mM was chosen in order to prevent lateral association of fibers that might happen at higher concentrations, as that could hinder base-pairing effects in later templating experiments.

As can be seen from Figure 9A, the cytosine-containing building block CF does not show a significant preference for either tetramer or trimer formation, with just a slight excess of tetramer when the library was equilibrated. Interestingly, with this building block also a small amount of pentamer is formed, not observed in the other cases. By TEM imaging (Figure 9B), fibrillar structures were observed.

Figure 9. (A) The change in the product distribution with time of a DCL made from of CF (1.0 mM) and (B) TEM image of the fibrillar structures of CF at day 37.

Secondly, a DCL made from the guanine-containing building block GF shows a strong preference for tetramer formation as is shown in Figure 10A. In addition, besides fiber formation, there are tape-like structures present in the TEM images (Figure 10B). As will be seen from the TEM image for the GF libraries (section 2.2.2), this tape-like structure is characteristic for guanine-containing building blocks.

0 20 40 60 80 100

0 10 20 30 40

Fraction of total peak area (%)

Time (days)

1mer 3mer 4mer 5mer

200 nm

A B

(20)

19

Figure 10. (A) The change in the product distribution with time of a DCL made from GF (1.0 mM) and (B) TEM image of the fibrillar and tape-like structures of GF at day 37.

In case of a DCL made from the thymine-containing building block TF, the formation of tetramer is preferred, although there is a significant amount of trimer still present when the library has equilibrated (Figure 11A). The TEM image shown in Figure 11B shows the formation of fibrillar structures.

Figure 11. (A) The change in the product distribution with time of a DCL made from TF (1.0 mM) and (B) TEM image of the fibrillar structures of TF at day 37.

As can be seen from Figure 12A, in the case of a DCL made from the adenine-containing building block AF, the formation of tetramer is observed, with only a small amount of trimer present. From the TEM image in Figure 12B fiber formation can be observed, although not very pronounced.

0 20 40 60 80 100

0 10 20 30 40

Time (days)

1mer 2mer 3mer 4mer

0 20 40 60 80 100

0 10 20 30 40

Time (days)

1mer 2mer 3mer 4mer

500 nm

200 nm

A B

(21)

20

Figure 12. (A) The change in the product distribution with time of a DCL made from AF (1.0 mM) and (B) TEM image of the fibrillar structures of AF at day 37.

All the different XBF libraries form supramolecular structures, CF and TF with only a small preference for tetramer formation over trimer formation, and GF and AF with a strong preference for tetramer formation.

The corresponding mass data can be found in Figure S1.

2.2.1.1 Mixed BF libraries

Subsequently, the different building blocks were mixed in a 1:1 ratio with a total concentration of 2.0 mM in borate buffer (50 mM, pH = 8.2), in order to investigate the products a mixed library would yield. When combining different nucleobase-containing building blocks in one system, there might be selective interactions with DNA in subsequent templating experiments, as it will only bind to the structures with complementary nucleobases. However, separation of the different compounds by UPLC and LC-MS was a challenging task, hindering proper analysis. In Figure S2, the UPLC chromatograms of the different mixed libraries at day 35 are shown, together with their peak assignment. In all cases, mixed tetramers and often also mixed trimers are found. These mixtures also show supramolecular structures, of which an example is shown in Figure S3.

2.2.1.2 DNA templating with BF

The next step is investigating if we can detect an influence of base-pairing on the different systems. This was achieved by adding 0.1 mM of a DNA strand of 10 of the same nucleobases long (B10, with B = C, G, T, A) when creating the library. The libraries were monitored over time by UPLC and LC-MS. First of all, the effect of DNA templating on single building block libraries were investigated. For all the four nucleobase- containing building blocks, a library of 1.0 mM was prepared in borate buffer (50 mM, 8.2 pH), with 0.1 mM DNA, with the DNA either being the complementary base to the building block, or the same nucleobase that is incorporated in the building block. Unfortunately, in all these cases, no significant effect of the DNA templating was observed (see Figure 13 and Figure S4).

0 20 40 60 80 100

0 10 20 30 40

Time (days)

1mer 3mer 4mer

200 nm

A B

(22)

21

Figure 13. The change in the product distribution with time of a DCL made from AF (1.0 mM) either with 0.1 mM A10 or 0.1 mM T10 added.

DNA templating experiments were also performed on the mixed libraries of XBF. The libraries were prepared by adding DNA composed of the complementary base for either one of the nucleobase- containing building blocks. Unfortunately, also in these mixed building block libraries, no significant DNA templating effect was observed (an example shown in Figure S5). A possible explanation for the absence of an effect of base-pairing is the fact that there might be electrostatic repulsion between the negatively charged fibers and the negatively charged DNA. From literature, it is known that magnesium ions increase the base-pairing strength in DNA. This stabilization is explained by the fact that the electrostatic repulsion between phosphate groups is decreased upon binding of positively charged Mg²⁺ ions.² Therefore, to a selection of libraries containing either a single building block or a mixture of two building blocks with DNA, 50 mM of MgCl2 was added. However, even in these cases, no base-pairing effect was observed (Figure S6 and S7).

Because of the lack of any evidence of base-pairing interaction in the building blocks containing the amino acid phenylalanine, no further investigation into these libraries was pursued. Instead, the focus was shifted to nucleobase-containing building block with histidine as amino acid (XBH). Histidine is positively charged, which can result in an electrostatic interaction between the template DNA and the fibers, facilitating possible base-pairing.

2.2.2 Histidine containing building blocks

Libraries were prepared of the four nucleobase-containing building blocks with histidine as amino acid at 1 mM in 50 mM borate buffer (pH = 8.2), and they were monitored by UPLC and LC-MS over time.

First of all, in the case of a DCL made from the cytosine-containing building block CH, a concentration of 1.0 mM was too low to achieve fiber formation, so a concentration of 2.0 mM was used. In addition, the library was pre-oxidized with sodium perborate (NaBO3). From Figure 14A, it can be concluded that CH

preferentially forms trimers. The TEM image in Figure 14B shows the formation of fibrillar structures.

0 20 40 60 80 100

0 5 10 15 20 25

Time (days) 4mer + A10 3mer + A10 4mer + T10 3mer + T10

(23)

22

Figure 14. (A) The change in the product distribution with time of a DCL made from CH (2.0 mM) and (B) TEM image of the fibrillar structures of CH at day 8.

Secondly, a library at a concentration of 1.0 mM of the guanine-containing building block GH was prepared.

Separation of the different macrocycle sizes for this building block on the UPLC is difficult to achieve as GH

macrocycles give broad peaks, but LC-MS analysis showed that first trimer and tetramers are formed, and over time only tetramer remains. This can also be recognized in the sharpness of the peak as shown in Figure 15A. These macrocycles form tape-like structure, shown in Figure 15B. This tape-like structure was previously also observed for GF, indicating that the guanine-containing building blocks prefer to organize in these tape-like structures instead of simple fibers.

Figure 15. (A) UPLC chromatography analyses (monitored at 254 nm) of the product mixtures obtained after 2 days (top) and 56 days (bottom) of a 2.0 mM solution of GH, and (B) TEM image of the tape-like structures of GH at day 8.

As can be seen from Figure 16A, in the case a DCL made from the thymine-containing building block TH, in the beginning the trimer and tetramer macrocycles are formed with similar rates. However, at a certain point (around day 30) tetramer formation takes over and the amount of trimer decreases. From the TEM image in Figure 16B, it can be seen that these building blocks form supramolecular fibers.

0 10 20 30 40 50 60 70 80 90 100

0 5 10 15 20 25

Time (days)

1mer 3mer 4mer

500 nm 500 nm A

500 nm

D8

B

4mer + 3mer

4mer A

B D2

D56

500 nm

(24)

23

Figure 16. (A) The change in the product distribution with time of a DCL made from TH (1.0 mM) and (B) TEM image of the fibrillar structures of GH at day 8.

Lastly, a library at a concentration of 3.8 mM of the adenine-containing building block AH was prepared.

Tetramer macrocycles were formed preferentially (Figure 17A), which form fibers that laterally associate as shown in Figure 17B.

Figure 17. (A) The change in the product distribution with time of a DCL made from AH (3.8 mM) and (B) TEM image of the fibrillar structures of AH.

All the different XBH libraries form supramolecular structures, CH and TH with a preference for trimer formation over tetramer formation, and GH and AH with a strong preference for tetramer formation. The corresponding mass data can be found in Figure S8.

To investigate if these species are self-replicators, seeding experiments were performed. To a pre-oxidized library, 10% of the fiber was added. The formation of the potential self-replicator was followed over time in a non-seeded and seeded library. When comparing the rate of formation of the trimer of CH in a seeded and non-seeded library, it shows that the addition of seed significantly accelerates the formation of the trimer, shown in Figure 18A. From this it can be concluded that the trimer of CH is a self-replicator.

0 10 20 30 40 50 60 70 80 90 100

0 20 40 60 80

Time (days)

1mer 2mer 3mer 4mer

0 10 20 30 40 50 60 70 80 90 100

0 2 4 6

Time (days)

1mer 3mer 4mer 5mer

200 nm

A B

(25)

24

Figure 18. Growth curve of (A) the trimer of a 2.0 mM solution of CH in borate buffer and (B) the tetramer of a 3.8 mM solution of AH in borate buffer, with and without 10% seed added at t = 0.

Determining if the tetramer of GH is a replicator is less straightforward, as the peaks of the trimer and tetramer overlap in the UPLC data. However, from LC-MS and observing the shape of the peak, it is clear that in a seeded library, there is an immediate formation of tetramer with almost no trimer present. An indication of this is the sharpness of the peak in the UPLC chromatogram, which is not observed in the non-seeded library (Figure S9) On the other hand, in the non-seeded library, the trimer species stays a significant part of the product over the first few days. These results are a strong indication that the tetramer of GH is a self-replicator. To determine the initial rate of the growth of tetramer of GH, deconvolution of the UPLC peak could be performed to determine the amounts of trimer and tetramer separately.

When performing seeding experiment on TH no significant seeding effect can be found, as the library has a significant amount of trimer as well as tetramer. However, in the case of AH, seeding experiments clearly showed that the tetramer is a self-replicator, as shown in Figure 18B. From these results, it can be concluded that the trimer of CH, the tetramer of GH and the tetramer of AH are replicators.

2.2.2.1 Mixed BH libraries

Subsequently, the different building blocks were mixed in a 1:1 ratio with a total concentration of 2.0 mM in borate buffer (50 mM, 8.2 pH), in order to investigate the products a combined library would yield.

However, GH containing libraries have broad peaks in the UPLC chromatograms, hindering proper integration and therefore analysis of the library product distribution.

Figure 19 shows the product distribution of the mixed library containing GH and TH in a 1:1 ratio. It shows the emergence of both mixed trimers and tetramers, with the latter being the predominant product.

0 20 40 60 80 100

0 50 100 150

Time (hours)

non-seeded 10% seed

0 10 20 30 40 50 60 70 80 90 100

0 2 4 6

Time (days)

non-seeded 10% seed

A B

(26)

25

Figure 19. The change in the product distribution with time of a mixed DCL made from GH (1.0 mM) and TH (1.0 mM).

When altering the ratio of building buildings blocks to either GH:TH 1:2 or 2:1, the chromatograms shown in Figure S10 are obtained. An increase in relative amount of GH results in tetramers that predominantly contain GH (so G4 and TG3), while increasing the relative amount of TH results in more TH rich trimer formation (T3 and T1G2), with a decrease in the amount of tetramer. This shows that TH triggers a preference for trimer, and GH has a preference for tetramer formation, which coincides with the product these building blocks form in a non-mixed library.

In the case of a mixed DCL made from CH:GH with a ratio of 1:1 at a concentration of 2.0 mM in borate buffer, the separation of the different species became even harder. LC-MS analysis shows the formation of all the mixed tetramers and trimers (Figure S11). However, with UPLC analysis the different tetramers and trimers could not be separated. When increasing the relative amount of CH in this mixed library (e.g.

CH:GH with ratio 3:1), besides the trimers and tetramers, also pentamers are formed. This shows that in certain conditions, CH can also stimulate the formation of pentamers.

Finally, the mixed library with CH:TH with a ratio of 1:1 at 2.0 mM concentration in borate buffer was investigated. When following the evolution of this mixture with UPLC over time, first as well the mixed tetramers as mixed trimers are formed. However, the relative amount of tetramer goes down after approximately 8 days, and the mixed trimers C2T1 and C1T2 are the dominant species in the mixture, shown in Figure 20A. In addition, these mixed trimers form fibrillar structures as shown in Figure 20B.

0 10 20 30 40 50 60 70 80 90 100

0 10 20 30

Time (days)

G1 T1

G2T1 G1T2

T3 4mers

(27)

26

Figure 20. (A) The change in the product distribution with time of a mixed DCL made from CH (1.0 mM) and TH (1.0 mM), and (B) TEM image of the fibrillar structures of the mixed library of CH and TH at day 42.

When investigating different ratio’s (CH:TH = 1:3, 1:2, 2:1, 3:1), all the libraries give mixed trimers as the predominant product (Figure S12). These mixed trimers are promising for further research, as they are most probably self-replicators consisting of mixed trimers. The libraries are built of non-complementary nucleobases facilitating templating experiments, and they give well defined peaks in the UPLC chromatograms. In addition, they form fibrillar structures.

Seeding experiments were performed on the mixed library with a ratio of CH:TH of 1:1. The experiments were performed with two different kinds of seed, firstly with 10% mixed trimer, and secondly with 10%

trimer of CH, to see if this species can crosscatalyze the formation of the mixed trimer. The results are shown in Figure S13. For as well the C2T1 as the C1T2 species, significant influence on the rate of formation can be observed with both seeds. When seeded, the initial rate of formation is faster than in the absence of seed. These results show that these CH with TH mixed trimers are self-replicators and can be crosscatalyzed by the trimer of CH. The corresponding mass data can be found in Figure S14.

2.2.2.2 DNA templating with BH

The next step is investigating if we can detect an influence of base-pairing on the different systems. These experiments were performed in a similar way as previously discussed for the XBF libraries in section 2.2.1.2. When examining the influence of DNA templating on single building block containing libraries, a small influence of templating was observed on the libraries with CH.However, the influence was too small to follow up on these results. For GH and TH, no significant influence was observed (Figure S15).

Next, the DNA templating effect on mixed libraries was investigated. For the mixed DCL made from CH and GH, no templating experiment were performed as it was not possible to analyze the product distribution.

For the mixed GH and TH libraries, no DNA templating effect could be observed by analyzing these libraries with UPLC (Figure S16). However, in the case of the mixed CH with TH libraries, a shift in product distribution can be seen when performing templating experiments. When 0.1 mM A10 DNA is added to a library with a total concentration of 2.0 mM of mixed CH:TH 1:1, pentameric species become a significant part of the final products, shown in Figure 21.

0 10 20 30 40 50 60 70 80 90 100

0 10 20 30 40 50

Time (days)

C1 T1

C2T1 C1T2

T3 C3T1

C2T2 C1T3

A B

200 nm

(28)

27

Figure 21. UPLC chromatography analyses (monitored at 254 nm) of the product mixtures obtained after 58 days of the mixed library CH and TH with ratio 1:1, (A) with no DNA added, and (B) with 0.1 mM A10 DNA added.

Moreover, this change is not observed for the other DNA species, as shown in Figure 22. Therefore, this interaction seems to be base-pair specific. This selective change in preferred macrocycle upon templating with complementary DNA, shows great promise for achieving base-pairing in this system.

Figure 22. UPLC chromatography analyses (monitored at 254 nm) of the product mixtures obtained after 58 days of the mixed library CH and TH with ratio 1:1, (A) with 0.1 mM G10 DNA added, (B) with 0.1 mM T10 DNA added, and (C) with 0.1 mM C10 DNA added.

T3 C1T2

C2T1 C2T3 C3T2 With 0.1 mM A10

No DNA

With 0.1 mM G10

With 0.1 mM T10

With 0.1 mM C10

4mers

C1T2 C2T1

A

B

A

B

C

T3 C1T2 C2T1

C2T3 C3T2

Information transfer through base pairing in self-assembly driven self-replication