Generation and regulation of enzyme-catalyzed DNA network growth using branched DNA motives

(1)

network growth using branched DNA motives

Dissertation

zur Erlangung des akademischen Grades des Doktors der Naturwissenschaften

(Dr. rer. nat.)

an der Universität Konstanz

Mathematisch-Naturwissenschaftliche Sektion Fachbereich Chemie

vorgelegt von

Sascha Keller

aus Mindersdorf

2013

(2)

Dissertation an der Universität Konstanz Prüfungsvorsitzender: Prof. Dr. Helmut Cölfen 1. Referent: Prof. Dr. Andreas Marx

2. Referent: Prof. Dr. Jörg Hartig

(3)

J. Am. Chem. Soc. 130, 13188-13189 (2008)

Keller, S., Wang, J., Chandra, M., Berger, R. & Marx, A.

DNA Polymerase-Catalyzed DNA Network Growth.

Chem. Soc. Rev. 40, 5690-5697 (2011) Keller, S. & Marx, A.

The use of enzymes for construction of DNA-based objects and assemblies.

(4)

(5)

1 Introduction

From nucleotides to DNA-based objects

DNA is the basic material for storage and transfer of genetic information in all living organisms. This biopolymer displays the fundament for stimulating and regulating thousands of biochemical processes. Thus, DNA became the target molecule in various research fields over the last 50 years.

A milestone for understanding the complex role of DNA for evolution, regulation and replication was the elucidation of its structure by Watson and Crick in 1953.^[1] Their proposal of a complementary double helical structure for nucleic acids out of the building blocks adenosine, cytidine, guanosine and thymidine still marks the current way of thinking in molecular genetics and other DNA-related scientific disciplines.

Since then, a huge amount of knowledge was generated by thousands of researchers. Many disciplines like phylogenetics, genetic engineering and forensics were founded during this development and gave us deep insights into the complex system of life.

The understanding of DNA as a linear, well-ordered biopolymer changed in the 1970’s. Insights into the replication process revealed that DNA displays a certain conformational variability, backbone flexibility and the possible formation of junctions.^[2-6] The elucidation of the recombination junction as the so-called Holliday junction opened the mind for more architectural possibilities in DNA´s structure. A pioneer in this field was Nadrian C.

Seeman. He predicted in 1982 a possibility to construct two- and three dimensional architectures by using DNA’s well known geometrical parameters and its complementary binding properties. The idea of constructing and predicting DNA junctions founded a new research field: DNA-based nanotechnology. As will be discussed in greater detail later, this nanotechnology is a bottom-up approach which attempts to hybridize short DNA-strands thereby creating architectures leading to functional materials ^[7-

10]. In the beginning, the access to short DNA oligomers was challenging.

(10)

Thereby, Caruthers developed the chemical synthesis of short DNA strands.^[11]

With the development of the fully automated solid phase supported DNA synthesis in 1985, research was boosted and created a plethora of new applications which are discussed in the following chapters. With this in hands, researcher’s creativity was not limited by sequence specific problems any more because DNA strands could be designed in every thinkable sequence or length up to 100 nucleotides.

About 30 years after the first idea of Seeman, the research field of DNA-based nanotechnology is still growing. Thereby, the construction methods of DNA- based nanomaterials ranges from exploiting the physiological properties of DNA to the use of living organisms. These methods are described in more detail in the following chapters.

(11)

1.1 Structural DNA nanotechnology

1.1.1 Unique Properties

There are many issues that make DNA an excellent nanoconstruction material. First, the favored Watson-Crick base pairing results in highly predictable hybridization of DNA strands.^[1] Second, the structure of the B- form resulting from hybridization of two complementary DNA is well- understood.^[12] Third, DNA possesses both structural stiffness as well as flexibility.^[13] The combination of relatively rigid dsDNA and flexible single- stranded DNA (ssDNA) opens up the possibility to modulate and design DNA motifs with desired geometry. Fourth, the progress in research fields as organic chemistry and molecular biology offers a rich toolbox for readily synthesizing, modifying, and replicating DNA.^[11,14-15]

In detail, DNA is a polynucleotide and subjected to specific molecular interactions known as A-T and G-C Watson–Crick hydrogen-bonding. This fact allows the convenient programming of artificial DNA receptor moieties through the simple four-letter alphabet.^[16] Important for architectural constructions is the feasible use of different geometries and topologies to generate a huge diversity of structure or to modify the properties of designed material.

Further, ssDNA is highly flexible to form sequence depended tight loops or even 180° turns.^[17] When two complementary ssDNA strands hybridize, they form a regular right-handed double helix which forms under physiological conditions a B-DNA with diameter of 2 nm and helical repeat of 3.4 nm-based on 10.5 bases.^[12,18] In case of alternating polypurine segments and low hydration conditions, a wider right-handed double helix called A-DNA occurs.

A left-handed double-helical form named Z-DNA is favored by increased ionic strength and sequences that alternate purines and pyrimidines.^[12,19]

However, the flexibility of dsDNA could also be observed in chromosomes concerning the packaging density of DNA on histones.^[20-22] The combination of ssDNA and dsDNA leads to tailored building blocks with tunable flexibility and rigidity.^[16] Short dsDNA do not show this flexibility. It was found that a length up to 150 base pairs can be considered as straight and rigid. This length is called the persistence length of DNA.^[23]

(12)

1.1.2 Self-assembled DNA architectures

In the early 1980’s, one idea of Seeman was to employ DNA as a construction material to crystallize biomolecules for x-ray analysis.^[5] Thereby, DNA should be used as a defined scaffold where biomolecules can be assigned at a specific position in the same orientation mimicking the order of a crystal.^[24] A key-step was Seeman’s idea which was inspired by naturally existing self-assembled DNA architectures known as Holliday junction.^[5] These junctions occur during the process of genetic recombination where two double helices partially interact with temporary branched structures. Seeman envisioned the approach to organize DNA with branched DNA (bDNA).^[25] These branched architectures became the key requirement for an effective scaffolding and construction material.^[26] Remarkable steps towards this goal were the first rigid DNA triangles consisting out of four-arm junctions^[27], two-dimensional DNA arrays,^[28-31] especially DNA-origami ^[32-36] and DNA crystals.^[30,37-39]

About three decades later, his initial concept was put into practice. In 2009, a well-ordered macromolecular 3D crystalline lattices of DNA was designed and self-assembled with precise control to a dimension of about 250 µm.^[40] In addition, it was possible to crystallize this DNA scaffolds with guest molecules to achieve Seeman’s postulated goal of DNA templated crystallization.^[41]

Using the abovementioned principle a myriad of approaches were reported which are illustrated in the following chapter.

(13)

1.1.3 Branched DNA constructs

Many of already reported approaches to generate DNA-nanoarchitectures rely on branching elements which base on the natural occurring phenomena in a DNA-based organism. To mention triple branched replication forks occurring during semi-conservative replication^[1] as well as four-arm branched Holliday junction observed as an intermediate in genetic recombination.^[42] Thereby, more than two DNA-strands pair with each other forming a branching element. To connect these branching elements, the complementary fit of hybridized DNA strands called the sticky-ends are necessary. Doing so, these small single stranded overhangs at the end of a DNA strand generate an extension of the constructs. These constructs are called stably branched DNA since they were thermodynamically stable under physiological conditions and conserves a geometric motive.^[43] Based on this approach DNA-hydrogels^[44]

and DNA-arrays^[29,33,38] were formed.

Another class of branching elements are molecules, which connect three or more DNA-strands covalently. The formation of DNA networks does not rely on the non-covalent interactions of DNA. In contrast to the methods based on the hybridization of DNA strands, several covalently and non-covalently artificial branching molecules are reported in chapter 1.1.3.1 and 1.1.3.2.

(14)

1.1.3.1 Non-nucleoside building blocks for covalently branched DNA junctions For nanotechnology the main challenge is to establish systems bearing the ability to connect and assemble molecules. Therefore, chemically modified branching molecules were transferred to the known concepts of DNA-based self-assemblies to construct of DNA architectures, illustrated in chapter 1.1.2.

Figure 1: Covalent DNA-branching molecules: a ^[45], b ^[46], c ^[47], d ^[48], e ^[49], f ^[50], g ^[50]

Branching molecules were used to create junctions generating branched DNA (bDNA). These bDNA motifs could be synthesized from nucleosides or non- nucleoside building blocks. Here, the focus is on possible branching of DNA with non-nucleoside building blocks. As mentioned before, bDNA is generated by clever predesigned nanoarchitectures based on self-assembled oligonucleotides (ODNs). The geometry and shape of the branchpoint (BP) influence the assembly and geometry significantly.

(15)

Synthetic BPs allow a similar way of design. One can incorporate the desired branching unit via solid phase supported DNA synthesis to result in two- (Figure 1b, f), three-(Figure 1a, d, g) and four-way (Figure 1c, e) DNA junctions. In addition the geometry of branching molecules showed significant influence on the generated architectures.^[45-47,49]

The presented BPs can be characterized into three classes. The first class is created to introduce junctions without a fixed geometry (Figure 1a). The second class was developed to introduce fixed angles into a DNA strand (Figure 1b), often a benzene ring, methylene centre or geometry of a metal ion complex are used. This class can be used also for the introduction of branching points (Figure 1b, c). The third class of molecules cross-links DNA strands with reactive groups thereby creating DNA-networks (Figure 1d).

As a first example, the branching motif of a benzene ring was used in several approaches.^[45-47,49] Branch points bearing the benzene ring in a centre demonstrated the ability to organize DNA-templated assemble. The research group of Sleiman used chemically modified DNA junctions that allow the generation of extended structures up to millimeter scale like DNA nanotubes or nanofibers. They developed a strategy for the generation of long DNA nanotubes^[51] of tunable geometry and long-range assembly of DNA into nanofibers.^[52] Thereby the branching moiety depicted in Figure 1 b was used.^[46] Further, the two-dimensional aromatic molecules (Figure 1d, g) demonstrated the ability to form three-dimensional architectures arranged by the coordination sphere of metal ions.^[53] Interestingly, DNA polyhedra could be also determined using the persistence length of DNA in combination with varying number of branching molecules.^[54]

The synthesis of specific structures with a desired molecular function is still in its infancy and has to be further explored. To develop this approach, our lab synthesized a rigid three-way branched adamantane motif, which is capable of forming highly stable DNA networks. The generated moiety serve as a useful building block for DNA-based nanoconstructions based on a core retaining complex geometry.^[55] However, chemically branched DNA is a useful tool to generate DNA-based nanostructures but still bears in many cases the limitation of time-consuming synthesis of the branching molecules.

(16)

1.1.3.2 Nucleoside building blocks for covalent branched DNA junctions

Besides the possibility to use non-nucleoside building blocks to mimic DNA junctions, using nucleosides themselves as branching unit is an approach with high impact. Nucleosides bear several possible positions to introduce modifications which might not interfere with the Watson-Crick base-pairing and might show reduced disturbing influence on DNA’s conformation (Figure 2). If the nucleoside bears a modification ready to introduce a second DNA strand, a BP is created to branch ODNs. A powerful approach to introduce modified nucleosides into a DNA strand is the automated solid phase supported DNA synthesis. Therefore, the design of building block demands an appropriate protecting group strategy based on phosphoramidite chemistry.^[11]

Inspired by advances in synthesis of artificial ODNs, Horn and Urdea realized the first synthesis of branched ODNs.^[56] The synthesis strategy is based on branch point a in Figure 2. The molecule enabled synthesis of fork and comb- like ordered DNA oligomers that can be further used in assembly strategies. In fact, branch point Figure 2a showed interesting properties and is developed towards commercial pathogen tests for example for HIV (human immunodeficiency virus) and HCV (hepatitis C virus).^[57-59]

Figure 2: Nucleoside-based branching molecules for DNA

Marx and co-workers started to enforce the synthesis of nucleoside-based BPs used to synthesis branched ODNs. Branch points (f,g Figure 2) were synthesized. However, it turned out that the these BPs were not suitable for an efficient solid phase supported synthesis (unpublished work). Branch points (d,e Figure 2) showed that the synthesis of modified nucleoside derivatives as

(17)

self-assembled DNA constructs.^[60] Therefore, Marx and co-workers reported results for the synthesis, characterization, and assembling properties of asymmetrical bDNA molecules that are able to generate linear and circular bDNA constructs.^[61] Marx and co-workers employed a combination of chemical and enzymatic approaches to generate these objects, which represent a novel class of bDNA constructs. This concept appears to be useful and to be a main pillar for my research which will be discussed in the following chapters.

(18)

1.2 Enzymes for the construction of DNA-based object and assemblies

DNA has found wide applications in nanotechnology due to its simplicity and predictability of its secondary structure.^[26] Thereby, DNA nanoarchitectures are mainly based on intelligent sequence design leading to defined hybridized structures out of short chemically synthesized oligodeoxynucleotides (ODNs).

Based on this principle a variety of two- and three-dimensional nano-objects was reported.^[10,16] To enlarge the dimension of DNA-based structures periodical repeats of nano-objects were successfully employed to reach the microscopic scale.^[37,51] Chemical self replication in solution^[62] and on surfaces^[63] was used to amplify nanoobjects. In future, this concept might circumvent the scale-up processes that hitherto limit broad applications.

1.2.1 The tool-box of DNA modifying enzymes

One prospective advantage of DNA for nanoconstruction is its inherent potential for manipulation by DNA modifying enzymes. This topic is only at the start to be fully explored. The application of enzymes in nanotechnology with its myriad of tools available for manipulation of DNA in a sequence specific manner, holds great potential for using DNA as a construction material.

Enzymes are highly selective and processive nanofactories in biological processes. Over the last few decades, many DNA-modifying enzymes have been characterized and a huge variety of possible substrates is commercially available. Up to now, the development of enzyme catalyzed construction of DNA-based nanostructures has been focused on implementing well-known commercially available enzymes that are briefly introduced in this chapter. For nanotechnology the enzymatic synthesis of long DNA strands offers the chance to expand the scale dimensions of the bottom-up approaches and opens the door to extended assemblies up to millimeter scale. Thereby, the enzyme family of ligases engaged an important function. Ligases covalently connect the 5´-phosphate group of one strand and the 3´-hydroxyl group of another strand, thus forming a phosphodiester bond allowing to connect multiple DNA strands.^[64] As an antagonist, restriction enzymes catalyze sequence specific cleavage of phosphodiester bonds. Other useful enzymes are the terminal

(19)

independently a DNA strand from its 3´-hydroxy terminus, thus generating single stranded DNA (ssDNA).^[65] Furthermore, DNA polymerases represent an important class of enzymes.^[66-67] These are used in the polymerase chain reaction (PCR), presenting a tool for the exponential amplification of DNA in a short time scale. Considering scaling up processes, DNA polymerases are a class of enzymes that could play a potent role in further developments. Beyond that, several DNA polymerases can even proceed on circulized templates producing a periodical sequence in the rolling circle amplification (RCA)^[68-69]

and therefore are a candidate for scaling up processes or replicating in enzyme-based nanotechnology.^[70-71]

(20)

1.2.2 Enzymatic manipulation of DNA bound to surfaces or AuNP

This separate research field was strongly influenced by Chilkoti and co- workers who developed a proof of principle approach of an enzyme-catalyzed reaction on an ODN modified Au-surface (Figure 3.1).^[72] Their approach is based on self-assembled monolayers (SAM) presenting ODNs. The catalytic enzyme is adsorbed on an atomic force microscope (AFM) tip (Figure 3.1B).

Taking benefit of the well-evolved dip-pen nanolithography (DPN)^[73], it was possible to deposit the nuclease DNAseI nanometer precisely on the self- assembled monolayers. Digestion was initiated in the next step by coating the surface with a magnesium-ion containing buffer. This led to nanotrenches observed by AFM. In this case, DNAseI was able to modify a DNA assembly on the surface with nanometer precision.

Figure 3: Two examples for enzymatic manipulation on SAM: (1) ODN terminated SAM on Au get selectively digested (A), AFM-Tip inked with DNase I (B), adsorbed enzyme on SAM (C), local digest of DNA induced by Mg²⁺ ions in solution(D).^[72] (2) AFM images of gold arrays with 5`-SH-(CH2)6-T25 DNA-SAM: After treated with heat-inactivated TdT (A), active TdT (B), and active TdT followed by exonuclease I digestion (C). The line profiles of these figures are shown in D.^[74] (Figure 3.1 Adapted with permission from Jinho Hyun et al.;

Further investigations by the same group employ TdT for extension of DNA on surfaces (Figure 3.2).^[74] Instead of digesting an existing monolayer of DNA, they extend template independent a self-assembled monolayer consisting of ODNs. TdT is able to extend ODN strands that are nanopatterned on the surface. Thereby, the enzyme is freely diffusing in solution which also differs from the above mentioned DPN method. Further, the extended DNA could be digested afterwards by exonuclease I (Figure 3.2C). Next, the TdT properties on self-assembled monolayers were studied in more detail.^[75] Thereby, a dependence on the extension process by mainly two factors was observed.

First, the surface-initiated enzymatic polymerization process showed clear

(21)

monolayers. On the other hand, the choice of used nucleotides showed significant influence on the extension behavior as well. The authors speculate about extending the approach by introducing modified nucleotides acting as affinity tags. Doing so, functional molecular scaffolds might be generated to dock nanoparticles and biomolecules on a surface in order to fabricate hybrid structures at the meso-microscale and to provide new supramolecular architectures for biosensors and clinical diagnostics. Another approach for enzyme mediated manipulations of ODN immobilized on surfaces was reported by Scoles and co-workers.^[76] They demonstrated that sequence specific digestion catalyzed by DpnII was possible but dependent on the density of immobilized double-stranded DNA (dsDNA). It was observed that efficient digestion only appeared if the distance between the ODNs was twofold the diameter of the used restriction enzyme. In principle, the approach illustrated that restriction enzymes can be used for manipulations of DNA assemblies on self-assembled monolayers. Recently, Daube and co-workers reported the consecutive use of enzymes for manipulations of self-assembled monolayers.^[77]

They realized their concept of generating a patterned surface bearing different, spatially separated DNA sequences. These sequences were processed by using a restriction enzyme (BamHI) and a DNA polymerase I (Klenow fragment of E.

coli DNA polymerase I). Thereby, nanopatterned ODNs with different sequences are selectively digested first. Resulting DNA layers allowed sequence-specific extension by addition of modified nucleotides.

The mentioned examples illustrate that enzymes have a high impact on sequence specific manipulations by restriction, digestion and template dependent and independent DNA strand extensions. The use of enzymes and their precise action demonstrates their potential for further applications in surface structuring.

The fact that optical properties of Au-nanoparticles (AuNP) are governed by the size of AuNP aggregates induced a variety of approaches of DNA-controlled assembly strategies. The pioneering work by Mirkin and his co-workers on DNA-mediated assemblies of AuNP has opened the door to a variety of potential applications in biological sensing, medical diagnostics and drug delivery.^[78] Along these lines, researchers used enzymes as DNA specific and regulating tools for AuNP aggregation.

(22)

Brust and co-workers adapted enzyme catalyzed manipulations of DNA for the generation of nanoscale building blocks based on AuNPs.^[79-80] Thereby, they used the stabilizing effect of surface-bound ODN on AuNP in solution. DNA bound to AuNP can be described as a protecting moiety for spontaneous aggregation of nanoparticles (Figure 4).^[80] Brust and co-workers extended the existing approach of assembling single stranded DNA-capped AuNPs from Mirkin and co-workers ^[78] by sequence directed hybridization to a programmed assembly approach which could be regulated by restriction enzymes. Doing so, AuNPs were coated with a ligand shell of thio-modified single stranded DNA (ssDNA), allowing a complementary DNA strand to hybridize with the ssDNA on the surface and to form a stable duplex. This resulted in modified AuNPs that were not able to aggregate with another AuNP. Further, they designed the DNA strands in a way that a recognition site for EcoRI was included (Figure 4A). Doing so, the used restriction enzyme EcoRI can be used to cleave the double strands sequence specifically and release cohesive ends capable for

Figure 4: Schematic description of the method. (A) Gold nanoparticles derivatized with double- stranded DNA are treated with restriction enzyme EcoRI, which cleaves the DNA to yield cohesive ends. The red color represents the recognition site of the enzyme, and the arrows indicate the sites of cleavage on each strand. In reality, each 15 nm particle has around 100 DNA ligands; (B) Two cohesive ends hybridize leading to a weak association of particles; (C) The DNA backbones are covalently joined at the hybridized site by DNA ligase to yield a double-stranded link between particles.^[80] (Figure 4 Kanaras, A. et al., Angew. Chem., Int. Ed. (2003), Copyright Wiley-VCH Verlag GmbH & Co. KGaA.

Reproduced with permission).

new hybridization events resulting in AuNP aggregates (Figure 4B). In principle, this concept can be transferred to a myriad of other restriction enzymes thereby controlling assembly processes. The characterization of AuNP assemblies was monitored by a visible color change induced by aggregation. As a next step, the hybridized AuNP assemblies were further stabilized with T4 DNA ligase (Figure 4C).^[80] Thus, T4 DNA ligase catalyzed the formation of

(23)

assemblies. Further studies indicated that this process is reversible and the ligated DNA-AuNP aggregates can be subsequently cleaved by treatment with a restriction enzyme.^[79] The system was put forward to a dynamic system in which hybridization of ODN-modified AuNP was followed by a ligation step and a restriction digest illustrating that the dynamic control of AuNP assembly processes can be controlled by enzymatic reactions.^[81] Up to date, a plethora of restriction enzymes has been discovered and hundreds are commercially available. This fact allows a large number of selective preparations of nanostructure building blocks in future.

The number of enzymes adaptable to DNA-modified AuNP was increased by the Klenow fragment of E. coli DNA polymerase I. This enzyme was identified to extend an ODN bound to AuNP with the help of a hybridized template strand (Figure 5A).^[82] It was found, that the enzyme elongation efficiency was strongly dependent on the density of ODN bound to the AuNP and the linker length that connect the ODN with the Au surface. These sterical problems also influenced the efficient primer-template annealing which resulted in decreased extension efficiency.

One example for amplifying AuNP species connected with ssDNA rather than dsDNA, was reported by Alivisatos and co-workers.^[83] The researchers illustrated an approach called ligase chain reaction (LCR). In LCR, two short DNA strands were first hybridized to a DNA template forming a nicked double strand DNA region. The reaction mixture was afterwards incubated to allow Thermus aquaticus DNA ligase to form a covalent bond at the nick site.

Afterwards, the reaction mixture was heated to melt the template from the ligated dimers. Thereby, AuNP dimers bridged by ssDNA linker were formed and released cycle by cycle. Li and co-workers adapted another idea to DNA modified AuNP.^[84] They disclosed the performance of rolling circle amplification (RCA) on AuNP (Figure 5B). RCA is known to be a straightforward biochemical method that can generate long ssDNA with repeating sequence units.

(24)

Figure 5: (A) In step 1, the primers in the DNA-AuNP conjugates are annealed to the template strand followed by extension in step 2 accomplished by the addition of Klenow (the large fragment of DNA polymerase I).^[82] (B) Schematic illustration of RCA on AuNPs. DNA- AuNP conjugates produced by RCA as a scaffold templating the binding of the formation of 3D nanostructures (blue dots).^[84] (Figure 5A Adapted with permission from Sheila R.

et al., Angew. Chem., Int. Ed. (2006), Copyright Wiley-VCH Verlag GmbH & Co. KGaA.

First, thio-modified DNA primers were attached to AuNP and a circular template was annealed to the primer. Thereby, DNA polymerase Φ29, known to be proficient in RCA, was used to generate long ssDNA strands which were visualized by AFM. The resulted ssDNA scaffolds were further modified by nanospecies like AuNP bearing complementary ODN. The ODNs were designed to hybridize in an ordered manner on the repeating template sequence forming an ordered space-filling assembly.

Besides AuNP based progress, another class of nanoparticles consisting of DNA block-copolymers (DBC) appeared to be a target for enzymatic manipulations. Amphiphilic organic nanoparticles have great potential for future applications.^[85] These DBCs were built out of a non-polar block- copolymer and polar DNA parts exhibiting a shell of DNA and a core with a hydrophobic polymer in aqueous environment.^[86] Such DBC micelles have been investigated for potential applications in antisense oligonucleotide or drug delivery^[87] and combinatorial tools in nanotechnology-based therapeutic approaches.^[88] Recently, based on DBCs a virus-like particle with the potential to act as a bioinspired container was reported.^[89] In this context, enzymes showed to be a powerful tool for the regulation of particle and micelle growth.

(25)

independent at the 3´-hydroxy terminus leading to a post-synthetic extension of block copolymer aggregates.^[90] Thereby, the methods offered a possible control over the growth of nanoparticles by employing TdT under mild and isothermal conditions. Further, it was possible to observe micelle growth by scanning force microscopy (SFM) in a time dependent fashion.

1.2.3 DNA-based materials: Hydrogels

Besides the use of DNA for assembly and modifying DNA hybrid materials, DNA itself may be considered as a material. Natural DNA strands are either linear or circular depending on sequence and length. However, in DNA recombination events branched DNA topologies, so called Holliday junctions, are formed. They have inspired researchers in the field of DNA nanotechnology in the design of rigidified building blocks with predictable topology.^[25] The branched DNA constructs, so called tiles, have been employed numerous times e.g. for the bottom-up construction of structured DNA-based lattices.^[91]

Li and co-workers combined this concept to construct Y-shaped structure (Fig.

4A) and studied their self-assembly towards dendrimer-like DNA (DL-DNA).^[92]

The Y-structure consists of three partial complementary ssDNAs that hybridize by crossover in a programmable fashion to form a distinct structure. For a controlled assembly to a DL-DNA, sticky ends were implemented in the design for specific binding to a complementary ssDNA. For example, a Y-shaped structure (compare Figure 6A) was formed by hybridization of three ssDNA strands into a three-way-junction bearing three sticky ends. As a consequence, the junction presented sticky ends for hybridization. Thus, it was possible to assemble sequence- specific another three Y-shaped DNA motifs resulting in a dendrimer-like structure. The authors called this step the first generation of assembly which was stabilized by enzymatic ligation.^[92]

Repeating these steps, it was possible to generate stable DL-DNA up to the fifth generation in a highly controlled fashion. This design and assembly strategy was transferred to construct other DNA building blocks gaining access to even more complex nanostructures. In this context, the same

(26)

Figure 6: (A) Used Holliday-junctions, (B) Crosslinking of Holliday-junctions via enzymatic ligation, ^[44] (C) A schematic diagram of the gelation process through enzymatic crosslinking and cell-free expression with P-gel pads with a commercial transcription assay.^[93] (Figure 6A,B Adapted by permission from Macmillan Publishers Ltd: [Nature Materials] Um, S. H. et al. 5, 797-801, copyright (2006). Figure 6C Adapted by permission from Macmillan Publishers Ltd: [Nature Materials] Park, N. et al. 8, 432-437, copyright (2009)).

research group extended the variety of branched DNA motifs with T- and X- DNA structures (Figure 6A).^[44] These structures bear self-complementary sticky ends that resulted in the formation of complex networks (Figure 6B).

The network building blocks were covalently connected by enzyme-mediated ligation resulting in stable DNA-based hydrogels. Thereby, the intrinsic geometrical character of the DNA branching motifs significantly influenced the morphology of the DNA hydrogels. Further, interesting properties like biocompatibility, biodegradability and easy handling in fabrication processes were observed. DNA scaffolds up to millimeter scale were feasible. Several applications were demonstrated based on DNA-hydrogels like the encapsulation of mammalian cells. These cells can be recovered alive by a

(27)

summary, the authors reported a new class of DNA-based material that might be exploited in a variety of applications including drug delivery, tissue engineering, three-dimensional cell culture that might be useful in cell transplant therapy.^[44]

Beside these studies, the same research group reported a further application of DNA hydrogels.^[93] They successfully designed a cell-free protein-producing gel (P-gel). The concept is based on the generation of a hydrogel with X-shaped DNA (X-DNA) and long DNA strands coding for a protein as building block (Figure 6C). In this approach ligases were employed for the covalent connection of X-DNA structures and the DNA bearing the respective genes forming the DNA hydrogel. In three studied examples the genes were successfully translated into proteins using a commercially available cell-free transcription and translation system. ^[93] Interestingly, the amount of generated protein was higher than the amount generated by the corresponding translation of the plasmid in solution. Thus, cross-linked plasmids embedded in DNA-hydrogels show increased protein expression efficiency than plasmids in solution. This finding might have an impact on high-throughput protein engineering and protein production.

(28)

1.2.4 Enzymatic replication and amplification of DNA nanojunctions

Many DNA-based nanostructures, especially geometric DNA objects, are generated out of ssDNA hybridized in a well-defined manner into highly ordered structures. As mentioned above (Figure 6), main progress in this field is based on the successful use of branched DNA junctions as building blocks for nanoconstruction.^[94] The access to large nanostructures in this field is limited by the amount of DNA oligomers and the scale

Figure 7: Design of the DNA octahedron. A, Three-dimensional structure involving twelve ODNs (octahedron edges) connected by six helper strands (octahedron vertices). Five of the ODNs are DX motifs (cyan) and seven are PX motifs (rainbow colors). The joints are four- way junctions that connect the core-layer double helices of each ODN. B, Secondary structure of the branched-tree folding intermediate. The structure consists of a single heavy chain (black) and five unique light chains (cyan). Like colors indicate half-PX loops whose sequence-specific cross-association generates a strut that serves as an edge of the DNA octahedron. Colored stripes coincide with strand crossover positions. Folding to the structure in the upper left is complete when all seven PX motifs have formed.^[95]

(Figure 7 Reprinted by permission from Macmillan Publishers Ltd: [Nature] Shih, W.M. et al. 427, 618-621, copyright (2004)).

of generated nanostructures. A straightforward approach to solve this problem would be the enzymatic replication and amplification of the DNA nanojunctions. However, the replication of nanostructures is rather difficult since DNA polymerases have to unfold secondary structures. One approach reported by Joyce and co-workers showed a possible way for the design and synthesis of geometric DNA nanostructures with a limited number of small

(29)

ODNs.^[95] They used small ODNs acting as helper strands for folding a 1.7 kb DNA strand into an octahedron (Figure 7A). The challenge was the synthesis and amplification of the scaffolding long DNA strand (Figure 7B). For a thermo-stable folding of a long DNA strand with short ODN, the scaffolding long DNA strand has to be single stranded. Therefore, they prepared a 1.7 kb template using small ODNs in a PCR-based assembly followed by ligation resulted in a dsDNA construct bearing necessary sequences for correct folding of the stable secondary structures paranemic-crossover^[96] (PX) and double- crossover^[97] (DX) (Figure 7B)

RCA was found to be a potent method for the amplification of nanojunctions for mainly two reasons (Figure 8).^[98] First, RCA is isothermal thus the nanojunction folding is not influenced. Second, heavily applied Φ29 DNA polymerase in RCA is known to overcome certain topological constraints such as circular padlock DNA probes hybridized with a linear DNA target or connected to a circular DNA strand.^[71] In the depicted scheme a four-arm nanojunction was chosen as a target motif to be replicated (Figure 8). This

Figure 8: Schematic diagram of the DNA nanojunction amplification cycle by RCA.^[98] (Figure 8 Lin, C. et al., Angew. Chem., Int. Ed. (2006), Copyright Wiley-VCH Verlag GmbH & Co. KGaA.

(30)

nanojunction was amplified successfully in multiple repeats. The structure was subsequently liberated by a restriction enzyme PstI. Thereby, many copies of the nanojunction were released. Since the complementary sequence (sense (-) strand) of the nanojunction was generated, a second round of RCA was needed to finally get the original sequence context resulting in a sense (+) strand.

These preliminary findings showed that it is possible to amplify a simple nanojunction with Φ29 DNA polymerase. In a following report by Lin et al. the amplification of a more complex structure was shown following the concept depicted in Figure 8.^[70] As a substrate for the amplification the known PX- motif, a DNA structure built out of several DNA strands that are crossed over for rigidifying the object, was chosen since it is frequently used in structural DNA nanotechnology.

Compared to the structure depicted in Figure 8, the structural properties of the PX-motif are more challenging for replication by Φ29 DNA polymerase. The hybridization density of the PX-motif results in a compact and circular template for RCA. An efficient replication was strongly dependent on the used primer length. With 60mer primer for example the PX-motif was partially unfolded and showed an excellent replication with sequence fidelity. Thus, the approach to use RCA to amplify nanostructures demonstrated its potential.

Problems due to complex topologies of the nanostructures are still restricting the replication efficiency and limit the application of this approach.

(31)

Figure 9: Schematic drawing showing the in-vivo replication of a DNA nanostructure. A single- stranded DNA nanostructure (four-arm nanojunction or PX-motif) is inserted into phagemid, transformed into XL1-Blue cells, and amplified in vivo in the presence of helper phages. High copy numbers of cloned nanostructures can be obtained readily by using standard molecular biology techniques.^[99] (Reprinted from: Lin, C. et al. Proc. Natl.

Acad. Sci. U.S.A. 105, 17626-17631. Copyright (2008) National Academy of Sciences, U.S.A.).

The success in the enzymatic replication of DNA nanostructures brought up the question if replication was also possible in cells or viruses. A concept was presented by Seeman and co-workers illustrating the possibility of in vivo amplification of DNA nanostructures.^[99] Related to a former work of Lin mentioned above (Figure 8), a four-armed nanojunction and a PX-motif were chosen and inserted in phagemids, transformed into XL1-Blue cells and amplified in vivo in the presence of helper phages (Figure 9). This approach uses standard molecular biological techniques leading to a high copy number of the cloned nanostructures. In detail, an single-stranded PX-DNA was inserted into a double-stranded phagemid vector bearing a single- stranded M13 DNA replication origin. The vector was transferred in E. coli cells (XL1-Blue). A single colony of the E. coli cells is then cultured before the phage-containing E. coli gets infected by a helper phage. The infected cells were then amplified and the phages were precipitated (Figure 9). Phage DNA was extracted and restricted. The exponential growth of the phage resulted in a high number of copies of the DNA nanostructure. Comparing with the in-

(32)

vitro amplification method mentioned before, in-vivo replication provides a higher efficiency by taking advantage of the naturally existing cellular machinery. A further advantage of in vivo replication is the little amount needed to start the replication process, and once correctly inserted in a vector, it can be easily stored. However, in vivo replication was only performed for rather simple secondary structures. Complex or even knotted topologies seem to be a great challenge for the replication machinery in cells. Nevertheless, the success of this method and the straightforward molecular biological methods bear the potential to be commonly used for preparing moderate amounts of long ssDNA strands with designed sequences.

(33)

1.3 Analytical Methods in Nanotechnology

At the beginning of nanotechnology research the progress in DNA-based nanotechnology was limited due to restricted analytical methods resulting in rather difficult conditions for detailed examinations. The main question is often how one can study the system at nanometer resolution. The best possibility to understand and manipulate a nanoarchitecture is to visualize it.

Visualization helps to understand the shape of complex assembled architectures and generate innovative concepts for construction and manipulation of these.

Concerning the research field of DNA based nanotechnology, it took almost a decade from 1982 of N. Seeman’s initial concept of using DNA self-assembly for the construction of two-dimensional arrays to its first experimental realization.^[5,26] In 1991, a DNA molecule was identified via radiolabeled native polyacrylamide electrophoresis to indicate the connectivity of a cube.^[43] In the early 1990s, the development of analytics like NMR, MS or IR-spectroscopy were well established but were not helpful to study such an ensemble of molecules resulting in nanostructured assemblies.[31,97,100]

One milestone in characterizing nanostructures was the development of the atomic force microscopy (AFM).^[101] Supported by unexpectedly rapid development, visualization of objects from molecular to atomic resolution was achieved.^[102] The field of DNA-based nanotechnology was boosted with the first image of circular single stranded DNA by Bustamante et al. in 1992.^[103] After visualization of further native DNA strands by Hansma^[102] and Lybuchenko^[104], AFM was established as a main analytical method to proof architectures at the nanometer scale. Again Seeman’s research group was the first who published a two-dimensional Holliday junction analogues via AFM.^[29]

Hansma^[102] and Lyubchenko^[104] first studied the imaging of DNA on mica in liquids and air and found crucial facts that support the design of sample preparation methods for AFM. Now, it is common practice to image DNA on bare mica, after adsorbing the DNA from a Ni²⁺containing solution.^[104] The Ni²⁺ shows stronger binding of DNA than Mg²⁺ due to the bigger ion radius.

Thereby, Ni²⁺ facilitates the absorption of DNA to the mica surface. For the sample preparation it pointed out that binding of DNA is also dependent on

(34)

the salt concentration and the pH of the buffer that should be between 6.5 and 7.0. One can vary the imaging by performing the studies in aqueous buffer solution or in air. The method in such an aqueous buffer solution was used by our collaboration partner from the MPI in Mainz (detail in chapter 5.2.4). The method generate pictures of high resolution but is only working for long DNA strand (> 200 base pairs from buffer solution containing divalent cations, e.g.

Mg²⁺ and Ni²⁺).^[105] The related mechanism of the binding process is a proposed model that describes a counterion correlation, which pulls the DNA onto the mica by the attraction force of the divalent cations that were bound to the surface.^[106] Another possibility to gain a stronger binding of DNA to mica is a sample preparation called AP Mica.^[104] Hereby, the mica is modified by silanization with 3-aminopropyltriethoxy silane (APTES) thus creating covalently bound positive charges presenting themselves protonated primary amines for the binding of the negatively charged DNA.

In 2004, the group of Joyce used cryo transmission electron microscopy (cryo- TEM)^[107] allowing the visualization a three dimensional architecture of a octahedron.^[95] This rather complex analytical method was extensively used for three dimensional architectures and give realistic insights in small objects geometry.^[108]

Nowadays, analytical methods try to gain more insights of the dynamic properties of nanoarchitectures. Since already used analytical method are limited to give information about one static picture of a certain state, observations of dynamic processes of nanostructures´ shape in motion are of high interest.^[109-110]

(35)

1.4 Preliminary work

As mentioned in chapter 1.1.3.2 the concept to gain branched DNA (bDNA) is based on an alkyl-chain modified nucleoside that introduce covalent branching building blocks into DNA. The following chapter intends to the concept in greater detail.

Based on the work of Horn and Urdea a reminiscent synthesis strategy of branched ODNs was developed in our laboratory.^[56,61] The approach provides access to symmetrical and asymmetric bDNA constructs (Figure 10).

Figure 10: General synthetic strategy for asymmetrical bDNA. BP=branching point, A- E=synthesized DNA strands.

For asymmetric synthesis, one requires a protecting-group strategy at the branching point (BP, Figure 10) that enables selective manipulations to be carried out to elongate branch B (Figure 10) without affecting the extension of linear strand A (Figure 10). Phosphoramidite building block 1 (Figure 11) is equipped with two orthogonal protecting groups and was used in the first studies. Incidentally, the already commercially available N4-(6-hydroxyhexyl)- 5-methyl-2’-deoxycytidine building block b (Figure 2) exhibits an inverted protecting-group strategy with respect to N4-(6-hydroxyhexyl)- and 5’-OH- protection. Our strategy allows the synthesis of multiple branched ODN where every introduced branch can bear a desired DNA sequence. Therefore, we

(36)

developed a straightforward synthetic strategy for desired phosphoramidite 1 (Figure 11). The synthesis starts with 3’-O-TBS-protected thymidine 2 ^[111] that was treated with 2-chloro-1-methylpyridinium iodide in the presence of 1,4- diazabicyclo [2.2.2]octane (DABCO) and levulinic acid to yield compound 3.

Thymidine analogue 3 was subsequently treated with 1,2,4-triazole in the presence of POCl3 to give triazole derivative 4, which was then converted into 2’-deoxycytosine derivative 5 after treatment with 6- aminohexanol.

This compound was treated with 4,4’-dimethoxytriphenylmethylchloride (DMTr-Cl) in the presence of (N,N-dimethylamino)pyridine (DMAP) to yield the protected intermediate 6. The 3’-O-TBS group was then removed by using a

Figure 11: Synthesis of phosphoramidite 1: a) 2-chloro-1-methylpyridinium iodide, DABCO, levulinic acid, CH3CN/dioxane, 90%; b) 1,2,4-triazole, NEt3, POCl3, CH3CN, 100%; c) 6- aminohexanol, CH3CN, 88%; d) DMTr-Cl, DMAP, pyridine, 88%; e) AcOH/TBAF, THF, 90%;

f) DIPEA, O-(2-cyanoethyl)-N,N’-diisopropylchlorophoramidite, CH2Cl2, 97%. (TBS=tert- butyldimethylsilyl).

(37)

mixture of AcOH and tetrabutyl ammonium fluoride (TBAF) to give compound 7, which was converted into target phosphoramidite 1 by reacting 7 with O-(2- cyanoethyl)- N,N’-diisopropylchlorophoramidite in presence of N,N- diisopropylethyl amine (DIPEA). The general synthesis of bDNA outlined in Figure 10 shows that after incorporation of branched monomer 1, the DMTr group attached to the nucleobase moiety is cleaved and the synthesis of segment B is continued. The DMTr group of the last incorporated monomer was deprotected by using standard automated DNA synthetic procedures, and was subsequently acetylated by using the capping protocol of the synthesizer.

The levulinyl group on the 5’-OH position of the branch point was cleaved by treatment of the controlled pore glass (CPG) loaded column with hydrazine (0.5 M) in a mixture of pyridine/acetic acid (1:1). Synthesis was continued to extend the branch and form strand C. The yield of every coupling step was found to be >98%. Remarkably, the depicted synthetic route towards bDNA is compatible with conventional and inverse phosphoramidites (5’-CE phosphoramidites, CE=cyanoethyl).

Figure 12: Denaturing PAGE analysis of bDNA ligation with T4 DNA ligase.

A) Depiction of proposed structures. B) Denaturating PAGE analysis of bDNA objects: a) lane 1=bDNA I; lanes 2–6=ligation with bDNA II for 5 min, 1 h, 3 h, 5 h, and 7 h, respectively, by using the ab splint; b) lane 1=bDNA I; lanes 2–6=ligation with bDNA II for 5 min, 1 h, 3 h, 5 h, and 7 h, respectively, by using the cd splint; c): lane 1=bDNA I; lanes 2–6=ligation with bDNA II for 5 min, 1 h, 3 h, 5 h, and 7 h, respectively, by using the ab and cd splint.

With this strategy one could synthesize bDNAs bearing at each side arm a different nucleotide sequence. Marx and co-workers could show that bDNA constructs could be uniquely addressed by complementary DNA strands.^[61]

(38)

DNA strands which are complementary to two different sequences thereby connecting two separated bDNAs are called splints. Hybridization and ligation experiment using this splints could demonstrate the ability to form non- covalently and covalently bound bDNA constructs. The results could be monitored by means of analytical polyacrylamide gel electrophoresis (PAGE) with ³²P-labeled bDNA (see Figure 12).

After the proof of principle that the artificial bDNA are accepted by T4 DNA ligase^[61], we envisaged expanding this approach by further DNA modifying enzymes. My first work in this context was done within my diploma thesis.^[112]

In this work it was possible to reproduce the synthesis strategy and to illustrate the possible ligation of small sequences of about 7 nucleotides (nt) next to a BP assisted by a splint. Additionally, the following digest of the ligated construct using EcoRI was feasible showing that a reversible ligation- digest manipulation of bDNA motives is achievable. It was possible to realize the synthesis of ODN bearing three BPs with a determined mass of about 23 kDa.

(39)

1.5 Aim of my work

The aim of my work was to generate branched DNA (bDNA) constructs using a chemically synthesized, covalently linked, three-way branch point and to exploit its effect on enzymatic reactions like DNA ligation, DNA digestion and DNA amplification to gain biocompatible DNA networks using a bottom-up approach. In my diploma thesis^[112], it was illustrated that chemically synthesized three-way branching points can be used indeed for the synthesis of such DNA constructs. Their characterization showed assembly properties that are able to generate linear and circular bDNA constructs. We employed a combination of chemical and enzymatic approaches to generate these objects which represented a novel class of constructs at this time.

Chemical modified precursor with possible pre-organizing properties on the generation of bDNA constructs should be the starting point of my doctoral thesis. The chemical modified precursor could be based on the combination of covalently connected bDNA Y-motifs that act as primer strands in PCR leading to DNA networks. For this purpose a robust synthesis route should be developed to gain fast access to branching points, which allow large scale synthesis of modified ODNs. These ODNs should be designed to be accepted by DNA polymerases as primers for enzymatic DNA amplification. The main goal should be the access of three-dimensional bDNA constructs. The generated bDNA constructs should be characterized using state-of-the-art analytics and imaging methodologies.

In the second part of my work, the potential of incorporating modified deoxynucleotides into the bDNA networks should be exploited to create highly functionalized bDNA by enzymatic synthesis. Further, biophysical properties of the networks should be tuned by variation of the template length used within PCR experiments. As a main step, the determination and adaption of analytical methods for visualization and monitoring of generated bDNA structures should be achieved.

Finally, the influence on the geometry of bDNAs by varying the branching core molecule should be studied. Therefore, the synthesis of a new DNA branching building block bearing an aromatic center in the core should be examined.

(40)

This flat and rigid center should arrange three DNA oligomers which are covalently attached in meta position to each other at a benzene ring. The conserved angle might conserve the two-dimensional extension for the generation of DNA nanostructures. In addition, effects on the two-dimensional extension using long DNA templates stiffened by locked nucleic acid (LNA) moieties should be studied because of a potential increase the rigidity of DNA networks. Finally, generated DNA networks should be characterized by multiple analytical methods to gain further insights into the conformational properties of the respective DNA, their geometric extension and macroscopic appearance.

(41)

2 Results and Discussion

2.1 DNA-based networks using flexible branched primer constructs

Many of the already described DNA nanoarchitectures rely on branching elements, which base on the naturally occurring Holliday junction. Thereby;

more than two DNA-strands pair with each other, thereby forming the branching element. Another class of branching elements are molecules, which connect three or more DNA-strands covalently. The formation of DNA networks does not rely on the non covalent interactions of DNA. These branchpoints can be introduced into DNA via automated solid phase supported DNA synthesis. We decided to use this as a starting point for our strategy. The before mentioned synthesis strategy (Figure 11) illustrated that increased efforts for synthesis of ODNs are necessary. Therefore, the first issue

was to

Figure 13: Synthetically approaches to accessing symmetrically branched DNA building blocks Figure was modified from Figure S1 (Adapted from Aldaye, F. A. et al.; Copyright 2007 American Chemical Society ^[53]

(42)

reduce efforts in synthesis of ODNs. It is common knowledge that the solid phase synthesis is limited to its length of connected nucleotides. Our idea was to synthesize two strands in parallel to gain a branched ODN bearing two identical sequences. A comparable access to branched symmetrical ODNs was published at the same time in the laboratories of H.F. Sleiman.^[53] Their approach (see Figure 13) used branching molecules bearing the same protection group strategy for the solid phase supported synthesis. They described three pathways to access symmetrically branched DNA branches:

the most suitable approach for our concept is depicted in Figure 13c. Thereby step i shows the standard DNA synthesis, step ii the incorporation of the branching building block, step iii the further extension on the solid phase support with further amidites and step iv the work-up and purification steps.

Figure 13b shows a variation of protection groups on the branching building block. The building block depicted in Figure 13a is very important for our approach because it is possible to generate three-way branched ODNs with three 3´-hydroxy termini. This 3´-ends allow the usage of the ODNs in the polymerase chain reaction (PCR) as primers.

For our approach, the leading structure of branching point 1 (Figure 11) was used where the levulinic acid moiety is exchanged by a DMTr protection group.

This literature known molecule (see Figure 14) is possible to synthesize within one month in 5 gram scale.^[56] With this branching point (OBP) it was possible to synthesize ODNs with reduced synthetical effort in high yield.

Figure 14: Phosphoramidite of symmetrical branching point (OBP)

Next, it was necessary to optimize the ODNs being modified or elongated by

(43)

polymerases are a class of enzymes that play a potent role in further developments because of their ability to amplify exponentially DNA in PCR.

To use DNA polymerase to enlarge branched ODNs it is necessary to synthesize 3´-hydroxy termini which can be extended. To adapt the solid phase supported synthesis which intrinsically generates 5´-hydroxy termini with our method to gain 3´-hydroxy termini could be realized using inverted amidites. The difference to the standard method is the use of inverted phosphoramidites (see Figure 15). By using inverse amidites one can generate

ODNs bearing a

Figure 15: Commercial available phosphor amidites; inverse amidite and standard amidite in comparison. B = A^Bz, G^DMF, C^Bz, T

3´-hydroxy terminus at each synthesized strand that is accessible for DNA polymerases to start their extensions with nucleotides. The established solid phase supported synthesis was evaluated within my diploma thesis together with M. Chandra.

(44)

Figure 16: Synthesized ODNs with three identical sequences and three 3´-hydroxy termini.

A: Primer bearing three 20 nt Faber primer sequences; B: Primer bearing three 20 nt Faber reverse primer sequences.

The first synthesized ODNs according to our strategy using OBP (Figure 14) as branching molecule are depicted in Figure 16. We studied the primer constructs in PCR where the amplification rate of linear in comparison to branched primers was determined as well as the influence of Mg²⁺

concentration on the extension of DNA strands covalently connected to one center. At last, the generated DNA constructs were tested whether they were still digestible by commercial available restriction enzymes.

(45)

2.1.1 Characterization of branched ODNs by thermal denaturation studies and CD spectroscopy.

Formation of stable duplexes with complementary DNA strands is a prerequisite for the employment of the Y-shaped bDNA primer and reverse primer in PCR experiments (Figure 16). Thus, we investigated duplex formation properties. In order to investigate whether the thymidine residues that are directly linked to OBP are amenable to participate in duplex formation, complementary oligonucleotides with varied lengths were investigated and compared to linear, non-branched reference duplexes. The thermal denaturation studies indicate that dependent on length, Y-shaped bDNA OBP I (Table 1) behave comparably to the linear non-branched counterpart 8 (Table 1). Increases in Tm were observed with increasing duplex length. Noteworthy, the obtained results show that the modified cytosine, used as branching molecule, is amenable to contribute to duplex stability. This was evidenced by an increase in Tm, when OBP I was hybridized to 11 in comparison to the one nucleotide shorter 10. If the hybridizing strand is one nucleotide longer and generates an overhang with ODN 12 one can observe a destabilization determined as a decrease in Tm of ≈2 °C.

Table 1: Thermal denaturation studies comparing linear and branched oligonucleotides hybridization. Incorporated phosphoramidite is depicted as OBP for branched ODN I.

Melting study: temperatures were determined with ± 0.5 °C accuracy.

Linear DNA bDNA

duplex Tm

[°C] duplex Tm

[°C]

5´-TGGTGATGGTGATGGT 3´-ACCACTACCACTACCACT

9

8 61.5 5´-TGGTGATGGTGATGGT

(3´-ACCACTACCACTACCACT)3 OBP 9

OBP I 59.9 5´-TGGTGATGGTGATGGTG

3´-ACCACTACCACTACCACT 10

8 63.0 5´-TGGTGATGGTGATGGTG

(3´-ACCACTACCACTACCACT)3 OBP 10

OBP I 63.0 5´-TGGTGATGGTGATGGTGA

3´-ACCACTACCACTACCACT

11

8 65.2 5´-TGGTGATGGTGATGGTGA (3´-ACCACTACCACTACCACT)3 OBP

11

OBP I 64.0 5´-TGGTGATGGTGATGGTGAC

3´-ACCACTACCACTACCACT

12

8 63.6 5´-TGGTGATGGTGATGGTGAC (3´-ACCACTACCACTACCACT)3 OBP

12

OBP I 63.0

To determine the influence of the modifications on the DNA conformation CD spectra were measured (Figure 17). The oligomer duplexes are labeled as followed: A. 9 + OBP I; B. 10 + OBP I; C. 11 + OBP I, D. 12 + OBP I. In the case of branched DNA motives linear DNA oligomers were used 6 µM and branched

(46)

DNA motives 2 µM to saturate all three arms of branched DNA motif. The CD spectra observed, for the branched duplex DNA showed slightly blue-shifted positive and conserved negative bands compared to the CD spectra of unmodified DNA. The crossover at 257 nm is typical for the B- conformation^[113]. The CD spectra of all modified ODNs show data of an overall B-form conformation similar to the unmodified DNA. These spectra, lead to the assumption that the introduced modification in OBP I allow to undergo an B- DNA conformation.

Figure 17: CD-spectra of DNA duplexes A. 9 + OBP I; B. 10 + OBP I; C. 11 + OBP I; D. 12 + OBP I

Generation and regulation of enzyme-catalyzed DNA network growth using branched DNA motives

network growth using branched DNA motives

Dissertation

zur Erlangung des akademischen Grades des Doktors der Naturwissenschaften

(Dr. rer. nat.)

an der Universität Konstanz

Mathematisch-Naturwissenschaftliche Sektion Fachbereich Chemie

vorgelegt von

Sascha Keller

aus Mindersdorf

2013

Table of Contents

1 Introduction

1.1 Structural DNA nanotechnology

1.2 Enzymes for the construction of DNA-based object and assemblies

1.3 Analytical Methods in Nanotechnology

1.4 Preliminary work

1.5 Aim of my work

2 Results and Discussion

2.1 DNA-based networks using flexible branched primer constructs