DNA and Aptamer Selection for Diagnostic Applications

vorgelegt von Diplom-Ingenieurin Janine Michel aus Berlin

Von der Fakultät III – Prozesswissenschaften der Technischen Universität Berlin zur Erlangung des akademischen Grades Doktorin der Ingenieurwissenschaften -Dr.-Ing.-

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr. Leif-A. Garbe Berichter: Prof. Dr. Jens Kurreck Berichter: PD Dr. Andreas Nitsche

Tag der wissenschaftlichen Aussprache: 27.09.2013

Berlin 2013 D83

To Olaf and my loving family, especially to grandpa Bernd. I miss you!

I

Acknowledgments

This work would have been impossible to complete without the help of many persons, including colleagues, family, and friends. Since there are so many of them I cannot acknowledge every single contribution by name, but I would like to thank everyone who helped me through this demanding and challenging but interesting time, regardless of the type of support. Above all, I would like to thank Andreas Nitsche for giving me the opportunity to do my PhD project and for the continuous support. Many thanks go to my dear colleagues Lilija Miller and Daniel Stern who helped me especially in the beginning of this project. Thank you for introducing me to the basics of and aptamers and for the fruitful and valuable scientific discussions and for frequent encouragement. I am grateful to Lilija Miller who helped me with phage display selections and subsequent peptide characterizations. Further, I would like to thank all members of the ZBS1 group for the friendly atmosphere and the company during lunch. A considerable contribution was made by students I supervised during my PhD project. Carolin Ulbricht, Daniel John and Alina Sobiech contributed to the “DNA aptamer selection and characterization project” during their bachelor’s thesis, internships, and master’s thesis, respectively. Guido Vogt contributed to the “peptide aptamer selection and characterization project” during his internship and his subsequent bachelor’s thesis. Wojtek Dabrowski contributed to the same project by programming the “Library Insert Finder” software, making the analysis of the numerous phage clones a lot easier. Furthermore, Wojtek Dabrowski performed the analysis of the NGS data and provided me with a workable amount of DNA sequences. In this context I would like to thank Aleksandar Radonić for introducing me to NGS library preparation and for the of these libraries. The staff of the sequencing lab, namely Julia Tesch, Julia Hinzmann, Silvia Muschter, Marlies Panzer, and Angelina Targosz, provided me with numerous DNA sequences. Thank you for this, I know it has been a lot of work. Many thanks go to Jörg Döllinger who performed the mass spectrometric analysis, to Janett Piesker who provided me with these wonderful EM pictures, and to Kazimierz Madela for the great help with the confocal laser scanning microscope pictures. Sincere thanks go to Jeffrey Drimmer and Ursula Erikli for copy-editing and continuously improving my English language skills. I am most grateful to my parents and my sister for always believing in me and supporting me in whatever I do. Thank you so much! Last but not least, I would like to thank my partner Olaf for the continuous support and always having an ear and the right words for me in times of despair.

II

Declaration of Authorship

I certify that the work presented here is, to the best of my knowledge and belief, original and the result of my own investigations, except as acknowledged. The present work has not been submitted, either in part or in its entirety, for a degree at this or any other University.

Berlin,

Janine Michel

III

Abstract

The availability of stable and reproducible detection methods is an important factor in the reliable detection and identification of common clinical infections, newly emerging pathogens, and bio-threat agents. Classical detection methods are based on the detection of nucleic acids or , using monoclonal and polyclonal as detection molecules. While -based detection methods depend on the detection of the genetic material of pathogens, antibodies have some considerable limitations, including tedious and expensive production, limited durability, and stability. To circumvent the many limitations of -based detection systems, there is an increasing interest in employing alternative recognition molecules for the detection and identification of pathogens. Such alternative molecules can be represented by aptamers. Aptamers are short synthetic molecules that can comprise of nucleic or amino acids. They are selected in vitro from huge libraries of molecules containing up to several billion randomly created sequences that can bind with high affinity and specificity to literally any target. Screenings of random synthetic DNA- and bacteriophage-based libraries were utilized in the present study to evaluate the applicability of synthetic nucleic acid and peptide aptamers for poxvirus detection. First, a suitable protocol for the selection of DNA aptamers using SELEX (systematic evolution of ligands by exponential enrichment) was established, and the amplification step was optimized using the Taguchi approach. Together with different random DNA libraries against native particles, SELEX resulted in the identification of 6 sequences, using the classical cloning and sequencing approach. The massively parallel sequencing of enriched DNA libraries resulted in identification of further 24 oligonucleotide sequences. These oligonucleotide sequences were characterized to determine affinities, specificities, and cross-reactivities. Of these 24 clones, 15 oligonucleotide sequences bound specifically to vaccinia virus particles and were characterized further. These 15 aptamers could be utilized for the detection of different poxvirus species in a sandwich assay in combination with a monoclonal anti-vaccinia antibody. Additionally, peptide aptamers for the specific detection of poxvirus particles were selected. For this, the combinatorial phage display methodology was used. Affinity selections of random peptide phage display libraries resulted in 17 recurring , indicating the enrichment of specific vaccinia virus-binding phage clones. After characterization of these 17 phage clones, five peptide sequences were synthesized and characterized. One phage- derived synthetic peptide (VV1) was able to bind specifically to vaccinia virus particles with the poxviral surface D8 as the interaction partner of VV1. The functionality of VV1 as capture molecule in combination with a polyclonal anti-vaccinia antibody could be shown. Furthermore, VV1 was successfully applied for the detection of different poxvirus species. At last, the successful combination of the selected and characterized SELEX-derived DNA- and phage display-derived peptide aptamers for the detection of vaccinia in an aptamer-based sandwich assay could be shown.

IV

Zusammenfassung

Ein wichtiger Faktor für die zuverlässige Detektion und Identifizierung verbreiteter klinischer Infektionen, neuartiger Pathogene und Bioterror-Agenzien ist die Verfügbarkeit stabiler und reproduzierbarer Detektionsmethoden. Klassische Detektionsmethoden basieren auf dem Nachweis von Nukleinsäuren oder Proteinen unter Verwendung monoklonaler und polyklonaler Antikörper. Während Nukleinsäure-basierte Detektionsmethoden darauf angewiesen sind, das genetische Material von Pathogenen nachzuweisen, weisen Antikörper erhebliche Nachteile auf. Diese Nachteile umfassen deren aufwändige und teure Produktion und eingeschränkte Stabilitäten. Auf Grund der vielfältigen Limitationen Antikörper-basierter Detektionsmethoden gibt es ein gesteigertes Interesse an der Verwendung alternativer Detektionsmoleküle. Aptamere können solch alternative Moleküle darstellen. Aptamere sind kurze, synthetische Moleküle, die aus Nuklein- oder Aminosäuren bestehen und mit hoher Affinität und Spezifität an buchstäblich jedes Zielmolekül binden können. Sie werden in vitro aus riesigen Bibliotheken selektiert, die mehrere Milliarden randomisierte Sequenzen enthalten können. Hier wurden randomisierte DNA- und Bakteriophagen-basierte Bibliotheken verwendet, um die Anwendbarkeit synthetischer Nukleinsäure- und Peptidaptamere für die Detektion von Pockenviren zu evaluieren. Zunächst wurde ein Protokoll für die Selektion von DNA- Aptameren unter Verwendung der SELEX (systematic evolution of ligands by exponential enrichment) Technologie etabliert und der Amplifikationsschritt mit Hilfe der Taguchi- Methode optimiert. Die Selektion gegen native Vaccinia-Virus-Partikel, unter Anwendung des klassischen Klonierungs- und Sequenzierungsansatzes, resultierte in der Identifizierung von 6 Oligonukleotiden. Die Hochdurchsatzsequenzierung angereicherter DNA-Bibliotheken führte zur Identifizierung 24 weiterer Oligonukleotide. Diese Oligonukleotide wurden hinsichtlich ihrer Affinitäten, Spezifitäten und Kreuzreaktivitäten charakterisiert. Von den 24 Oligonukleotiden banden 15 spezifisch an Vaccinia-Virus-Partikel, wurden weiter charakterisiert und konnten erfolgreich für die Detektion verschiedener Pockenviren in Kombination mit einem monoklonalen anti-Vaccinia-Antikörper eingesetzt werden. Zusätzlich zu den DNA-Aptameren über SELEX wurden Peptidaptamere als Pockenvirus- spezifische Detektionsmoleküle über Phagen-Display selektiert. Die Affinitätsselektion randomisierter Peptid-Phagen-Display-Bibliotheken resultierte in 17 wiederkehrenden Peptiden. Nach der Charakterisierung dieser 17 Phagenklone wurden fünf Peptidsequenzen synthetisiert und charakterisiert. Eines dieser synthetischen Peptide (VV1) band spezifisch an native Vaccinia-Virus-Partikel, wobei das pockenvirale Oberflächenprotein D8 als Interaktionspartner für VV1 identifiziert wurde. Die Funktionalität von VV1 als Fängermolekül in Kombination mit einem polyklonalen anti-Vaccinia-Antikörper und als Molekül für die Detektion weiterer Pockenvirus-Spezies konnte gezeigt werden. Als Letztes konnte die erfolgreiche Verwendung der selektieren DNA- und Peptid- Aptamere für die Detektion von Vaccinia-Viren in einem Aptamer-basierten Sandwich-Assay gezeigt werden.

V

Table of contents Acknowledgments ...... II Declaration of Authorship ...... III Abstract ...... IV Zusammenfassung ...... V 1. Introduction ...... - 1 - 1.1 Aptamers – alternative molecules for pathogen detection ...... - 1 - 1.1.1 Nucleic acid aptamers and their applications ...... - 2 - 1.1.2 Peptide aptamers and their applications...... - 4 - 1.2 Selection methods for aptamers ...... - 4 - 1.2.1 Principle of SELEX and possible variations for selection of nucleic acid aptamers ...... - 4 - 1.2.2 Principle of phage display and possible variations for the selection of peptide aptamers ...... - 7 - 1.3 Optimization of processes using the Taguchi approach ...... - 10 - 1.4 Application of high-throughput sequencing in the selection of DNA aptamers .. - 12 - 1.5 as potential targets for aptamer selection ...... - 13 - 1.5.1 Overview of different IMV and EEV envelope proteins ...... - 15 - 2. Aims of study ...... - 16 - 3. Materials and Methods ...... - 17 - 3.1 Materials ...... - 17 - 3.2 Cell culture and virus propagation ...... - 23 - 3.2.1 Cell preservation and recovery ...... - 23 - 3.2.2 Maintenance and subculture routine ...... - 24 - 3.2.3 Virus propagation ...... - 24 - 3.3 Nucleic acid aptamer selection via SELEX ...... - 24 - 3.3.1 Optimization of individual SELEX steps ...... - 25 - 3.3.2 Selection of aptamers against recombinant A27 ...... - 29 - 3.3.3 Selection of aptamers against native vaccinia virus particles ...... - 30 - 3.3.4 Quantification of ssDNA and binding assays ...... - 31 - 3.3.5 Cloning of enriched DNA library ...... - 32 - 3.3.6 Plasmid DNA isolation ...... - 32 - 3.3.7 Sanger Sequencing of DNA inserts ...... - 32 - 3.4 High-throughput sequencing for identification of enriched aptamer libraries from VACV selection ...... - 33 - 3.4.1 Library preparation...... - 33 - 3.4.2 Validation of DNA libraries for NGS ...... - 34 - 3.4.3 Spotting flow cell and NGS sequencing run ...... - 35 - 3.4.4 NGS data processing ...... - 35 - 3.5 Peptide aptamer selection by phage display ...... - 36 - 3.5.1 Random peptide phage display libraries ...... - 36 - 3.5.2 Bacterial strain maintenance and culture ...... - 36 - 3.5.3 Phage titering ...... - 37 - 3.6 Identification of peptide ligands to recombinant OPV proteins and infectious VACV particles ...... - 37 - 3.6.1 Biopanning against recombinant A27 ...... - 37 - 3.6.2 Biopanning against VACV particles ...... - 37 - 3.6.3 Amplification of phage clones ...... - 38 - 3.6.4 DNA sequencing of selected phage clones ...... - 38 - 3.6.5 Identification of peptide consensus sequences ...... - 38 -

VI

Table of contents

3.6.6 Phage ELISA ...... - 39 - 3.7 Selection of sequences for peptide synthesis ...... - 39 - 3.8 Characterization of selected aptamers ...... - 40 - 3.8.1 Pre-characterization of anti-A27 DNA aptamers ...... - 40 - 3.8.2 Indirect ELONA with anti-A27 DNA aptamers for affinity testing ...... - 40 - 3.8.3 Estimation of limit of detection (LOD) of anti-A27 DNA aptamers ...... - 41 - 3.8.4 Sandwich ELONA with anti-VACV DNA aptamers for affinity testing ...... - 41 - 3.8.5 Specificity testing with anti-VACV DNA aptamers ...... - 41 - 3.8.6 Sandwich ELONA with anti-VACV DNA aptamers as capture molecules...... - 41 - 3.8.7 Indirect ELISA with anti-A27 or anti-VACV peptides for affinity testing ..... - 42 - 3.8.8 Specificity testing of synthetic peptides ...... - 42 - 3.8.9 Sandwich ELISA with synthetic peptides as capture molecules ...... - 42 - 3.9 Identification of binding partner for peptide VV1 ...... - 43 - 3.9.1 Peptide ELISA with recombinant poxviral proteins ...... - 43 - 3.9.2 Surface plasmon resonance (SPR) assay ...... - 43 - 3.9.3 Pull-down assay, SDS PAGE and subsequent mass spectrometric analysis - 45 - 4. Results ...... - 47 - 4.1 Selection of DNA aptamers against A27 ...... - 47 - 4.2 Estimation of specificity, affinity, and limit of detection of selected anti-A27 DNA aptamers ...... - 48 - 4.3 Selection of DNA aptamers against native VACV particles ...... - 51 - 4.3.1 Optimization of individual SELEX steps ...... - 51 - 4.3.2 Cloning and sequencing of enriched DNA libraries ...... - 57 - 4.3.3 High-throughput sequencing of enriched DNA libraries ...... - 58 - 4.4 Characterization of anti-VACV DNA aptamers ...... - 66 - 4.4.1 Affinity testing of DNA aptamers ...... - 66 - 4.4.2 Specificity testing of DNA aptamers ...... - 72 - 4.4.3 Identification of binding partner for DNA aptamer clones ...... - 74 - 4.5 Selection of peptide aptamers via Phage Display ...... - 76 - 4.5.1 Identification of consensus peptide sequences ...... - 76 - 4.5.2 Evaluation of selectivity of enriched phage clones by phage ELISA ...... - 78 - 4.5.3 Selection of peptide sequences for synthesis ...... - 82 - 4.6 Characterization of selected peptide ligands ...... - 83 - 4.6.1 Affinity testing of synthetic peptide ligands ...... - 83 - 4.6.2 Specificity testing of anti-VACV peptide VV1 using different detection systems ...... - 86 - 4.6.3 Identification of binding partner for peptide VV1 ...... - 89 - 4.6.4 Surface plasmon resonance (SPR)/Biacore analysis of interaction of peptide VV1 with D8 ...... - 90 - 4.6.5 Visualization of VV1 OPV interaction by electron microscopy ...... - 92 - 4.7 Usability of peptide VV1 in IFA and Western Blot ...... - 93 - 4.8 Detection of OPV in an aptamer-based sandwich assay ...... - 95 - 5. Discussion ...... - 97 - 5.1 Selection and characterization of A27-specific DNA aptamers ...... - 97 - 5.2 Optimization of individual SELEX steps ...... - 98 - 5.2.1 Target presentation, elution and recovery of ssDNA ligands in the selection of VACV-specific DNA aptamers ...... - 98 - 5.2.2 Optimization of amplification of random DNA libraries using the Taguchi approach is fast and efficient ...... - 99 - 5.3 The decision of when to end the selection process is challenging ...... - 100 -

VII

Table of contents

5.4 HTS can be used to identify high affine DNA aptamers that might not be found with the classical cloning and sequencing approach...... - 102 - 5.4.1 DNA aptamers can be used as detection molecules for OPVs, but not as capture molecules ...... - 103 - 5.4.2 Single nucleotide changes can alter the binding characteristics of DNA aptamers significantly ...... - 104 - 5.5 Further applications of DNA aptamers ...... - 105 - 5.5.1 Usability of DNA aptamers to detect OPV-infected cells by fluorescence labeling ...... - 105 - 5.5.2 The selected DNA aptamers do not have any antiviral activity ...... - 105 - 5.5.3 Usability of DNA aptamers for the detection of OPV proteins in Western Blot ...... - 106 - 5.6 Identification of binding partner of the selected anti-VACV DNA aptamers ..... - 106 - 5.7 Selection of peptide aptamers using phage display ...... - 107 - 5.8 Binding properties of phage-free peptides and phage-attached peptides can differ significantly ...... - 107 - 5.9 Peptide VV1 can be used for the selective detection of different OPVs ...... - 108 - 5.10 Further applications of peptide aptamer VV1 ...... - 108 - 5.11 Approaches to optimize binding properties of synthetic DNA- and peptide ligands ...... - 110 - 5.12 DNA and peptide aptamers can be combined in a sandwich assay for OPV detection ...... - 111 - 5.12.1 Synthetic ligands as surrogate antibodies for pathogen detection? ...... - 112 - 5.12.2 Implementing synthetic DNA and peptide aptamers for the detection of pathogens ...... - 113 - 5.13 Comparison of SELEX and phage display methodology ...... - 115 - 5.14 Conclusions and perspective ...... - 116 - 6. Abbreviations ...... - 118 - 7. Figures ...... - 120 - 8. Equations ...... - 121 - 9. Tables ...... - 122 - 10. References ...... - 124 -

VIII

1. Introduction

An important factor for the reliable detection of pathogenic agents causing clinical infections, newly emerging pathogens, and bio-threat agents, is the availability of stable and reproducible detection methods. Numerous methods for the detection and identification of human pathogens have been established, and are therefore well comprehended [1]. While these methods are reliable, they generally cannot be used in the field, nor do they allow for a rapid detection of any pathogens. Nucleic acid-based detection methods, such as polymerase chain reaction (PCR) and microarrays, are fast and sensitive [2]. Nevertheless, they are limited to the detection of pathogens containing DNA (deoxyribonucleic acid) or RNA (ribonucleic acid). These methods are not able to detect protein threat agents, such as toxins and prions, nor do they give any information about the infectivity of the agent. Furthermore, only known pathogens can be detected. An alternative to PCR and microarrays is the use of antibody-based bioassays which are routinely used in clinical, environmental and food analysis [1-3]. These assays are based on polyclonal or monoclonal antibodies, which remain the most versatile and widely used protein-binding agents. Antibodies can bind to their target with high affinity and high specificity, being able to recognize only the target protein in a crude cellular extract containing many different proteins [4, 5]. However, antibodies have some considerable limitations. These include: tedious and expensive production, limited shelf-life, and the necessity of using animals. Furthermore, antibodies must maintain a relatively delicate three-dimensional structure in order to function, limiting the conditions under which they can be deployed [5, 6]. To circumvent the many limitations of nucleic acid- and antibody-based detection systems, there is an increasing interest in employing alternative recognition molecules for the detection and identification of pathogens [6, 7]. Ideally, these molecules display similar or superior sensitivities when compared to nucleic acid- and antibody-based systems; they exhibit enhanced stability in extreme environments (e.g. high temperature, pH, or acidic conditions); they are immune to interferences from components of complex sample matrices; and they are able to identify and quantify the target present [7]. Aptamers could have the potential to represent such alternative molecules.

1.1 Aptamers – alternative molecules for pathogen detection

Aptamers are small synthetic molecules, selected in vitro from huge libraries of molecules containing up to several billion randomly created sequences that can bind with high affinity and specificity to their target [7]. Aptamers can be divided into two types: 1) nucleic acid and 2) peptide aptamers [7, 8]. Both have several advantages over antibodies: they are small in size (5-25 kDa vs. 150 kDa), chemically stable, and easy to modify. Once selected, the sequence is always available. Their synthesis is fast, cheap and easy, and therefore a scale-up

- 1 -

Introduction is possible. No batch-to-batch variations, as known for antibodies, are observed. Aptamers have a low immunogenicity, and no animals or cell cultures are needed for their production. Selections can be performed against toxins and highly pathogenic agents, whereas the generation of antibodies can be challenging, because the animals die before their bodies produce antibodies. Aptamers can even be selected without prior knowledge of the target, leading to new molecular structures for diagnosis and treatment of diseases [3, 5, 7, 9-14]. Features, applications and selection methods of aptamers are described in the following.

1.1.1 Nucleic acid aptamers and their applications Nucleic acid aptamers are short single-stranded (ss) DNA or RNA molecules that adopt specific three-dimensional structures and can be selected from complex combinatorial libraries by a repetitive cyclic in vitro process called “Systematic Evolution of Ligands by EXponential enrichment” (SELEX) [3, 10]. They can bind with high affinity and specificity to their target molecule. Nucleic acid aptamers and SELEX were first described independently in 1990 by the Ellington [15] and Tuerk [16] laboratories. Since then, aptamers have been selected against a variety of targets. These include small molecules (such as organic dyes [15], amino acids [17], human [18]), ions [19, 20], nucleic acid binding proteins (such as DNA polymerases [16, 21]), peptides [22], proteins [23-25], toxins [26, 27], and more complex structures (such as parasites [28, 29], [30, 31], viruses [32, 33], and even entire cells [34, 35]). Numerous applications for nucleic acid aptamers have been developed. Aptamers have been used as bio-threat sensing ligands for the detection of Bacillus anthracis spores, influenza viruses, ricin toxin, and other pathogenic molecules in different aptasensing platforms and detection assays (reviewed by Fisher et al. [36]). Other reviews deal with the application of aptamers as therapeutics [10, 37-41] and their use for detection of cell- or disease-specific markers [34, 41-43]. In general, published applications of nucleic acid aptamers in bio-analytical assays include 1) aptamer-based affinity PCR [44] and electrochemical [45, 46] for protein detection, 2) exonuclease protection assays [47], 3) small molecule sensors for detection of adenosine and [48], and 4) affinity chromatography for separation of enantiomers of chiral drugs for example (e.g.) D- and L- peptides of vasopressin [49]. Further applications can be found in the analysis of environmental and food samples (with the recognition of carcinogenic aromatic amines [50] and for the recently described detection of streptomycin in honey [51]). A stable three-dimensional structure of the nucleic acid aptamers is essential for their functionality. This structure is dependent on the primary sequence, the length of the nucleic acid molecule, and the environmental conditions. The length of known nucleic acid aptamers varies from 25 to 90 bases. Typical structural motifs are stems, purine rich bulges, hairpin structures, internal loops, pseudo knots, kissing complexes, a G-quadruplex structure, and tetraloops [3]. Some examples of these structures are displayed in Figure 1.

- 2 -

Introduction

Typical equilibrium dissociation constants (KD) for nucleic acid aptamer binding are in the low pico- to micromolar range. Generally, it has been described that KD values for large size targets are in the order of pico- to nanomolar concentrations, while small molecule targets lead to aptamers with binding properties in the micro- to millimolar range [14, 52, 53]. In comparison, affinities for antibodies are described in the range of low pico- to micromolar [54, 55]. Here as well, affinities can depend on the size of the antigen. Furthermore, avidity and the time point of antibody isolation can have an impact on the antibody-binding properties known as affinity maturation [56]. With appropriate optimization protocols, even femtomolar KD values can be achieved [5].

C A B

Figure 1: Examples of different structural motifs of nucleic acid aptamers A) Hairpin or stem loop, B) -binding aptamer sequence folded as pseudoknot, C) G- quadruplex (modified from [57])

Nucleic acid aptamers have the ability to refold to their active conformation after denaturation. Apart from this and numerous further advantages, a drawback is their susceptibility towards which are commonly present in biological fluids [13]. This drawback can be overcome by the introduction of various modifications, for example to the oligonucleotide backbone [58]. These modifications can be introduced during or after the SELEX procedure. If introduced during the SELEX procedure, the modifications have to be compatible with the standard required for the SELEX process [59]. On the other hand, post-selection modifications of an aptamer can be challenging. It has been described that LNA-modified aptamers can switch the conformation of the binding region to a structure that is no longer recognized by the target molecule [60]. In contrast, the LNA modification of a tenascin-C-binding aptamer exhibited significant stem stabilization and improved plasma stability, while maintaining high affinity to the target molecule [60]. Overall, the number and position of the modifications seem to be critical parameters to improve the aptamer function [61]. It has been described that certain modifications at the 2’ position of make RNA resistant [62]. Furthermore, most of the endonucleases appear to be specific, so that substitutions at the 2’ positions of pyrimidine nucleotides are sufficient to protect RNA sequences from degradation [62]. Modifications, such as amino (NH2) and fluoro (F) functional groups at the 2’ position of the sugar, are substrates for enzymes used in the SELEX process. Therefore, successful selections with libraries containing with 2’-NH2 and 2’-F functional groups have been performed, resulting in aptamers with

- 3 -

Introduction enhanced survival times in biological fluids [63, 64]. Further modifications of natural nucleotides, the introduction of unnatural base pairs, and the importance of polymerases used for SELEX have been reviewed by Jayasena [13] and Kusser [59].

1.1.2 Peptide aptamers and their applications Two different types of peptide aptamers are found in the literature. Peptide aptamers were first described as combinatorial proteins, consisting of a variable peptide inserted into an inert, but constant scaffold protein [3, 8, 65]. The peptide ligands obtained from phage display are the amino acid equivalents to nucleic acid aptamers [7, 66]. Using phage display, peptide aptamers have been selected against titanium-based implants [67], Bacillus spores [68, 69], Newcastle disease virus [70], the cucumber mosaic virus coat protein [71], and against many other targets (such as small organic ligands, whole cells, or even organs), reviewed by Smith and Petrenko [72]. For peptide aptamers inserted into a scaffold protein, their use in microarrays [73] and the inhibition of intracellular processes [74] and biological pathways [75] has been described. While the application of nucleic acid aptamers in various bioanalytical assays has been described in a plethora of publications (compare 1.1.1), there are only few publications regarding the bioanalytical application of peptide aptamers derived from phage display. Many peptide ligands obtained from phage display act as agonists and antagonists of receptors [76]. They have been selected to define substrate sites for different enzymes [77, 78], or to identify sequence motifs necessary for binding [79]. Furthermore, the successful selection of peptides for the detection of biological threat agents has been described and reviewed by Petrenko and Vodyanoy [66].

1.2 Selection methods for aptamers

Many protocols for the selection of aptamers have been developed since their first description. While nucleic acid aptamers are always selected from one of the many variations of SELEX, the selection of peptide aptamers has been described to use different biomolecular methods (such as two-hybrid, mRNA display, , and phage display [3]). In the present work, SELEX and phage display were used to select aptamers. Therefore, the principle of both methods, including some of the many possible variations, will be explained in the following.

1.2.1 Principle of SELEX and possible variations for selection of nucleic acid aptamers Even though there are many variations of the SELEX procedure, all protocols use the common fundamental steps of binding, washing, elution, amplification, and generation of ss . The principle of the SELEX procedure is displayed in Figure 2. The initial and very important step of the selection process is the synthesis of an appropriate library (1). Typical SELEX libraries are composed of either RNA or single stranded DNA (ssDNA). These libraries contain a randomized region of 20 to 60 nucleotides (nts), flanked

- 4 -

Introduction by fixed primer sequences that are needed for PCR and/or RT-PCR amplification [52]. The complex libraries (containing 1013-1018 different nucleotide sequences) are then incubated with the target molecule (2). In doing so, the incubation of the target and the library is performed in solution, or the target is immobilized on a solid support (2a). Unbound nucleotide sequences are washed away (3), while bound sequences are eluted (4) and amplified (5) by PCR and/or RT-PCR. RNA aptamers have first to be reverse-transcribed (RT), thus an RNA polymerase promoter (usually T7) is introduced into the primer region during library construction. The final step is the generation of ss oligonucleotides (6), creating an enriched library (7) which is then introduced into a new SELEX round. This process is repeated for several rounds (typically 8-20), aiming for the reduction of the initial random oligonucleotide library in order to obtain few ligands with the highest affinity and specificity for the target [14]. The enriched library is usually cloned after the last selection round. This creates individual ligands which can then be sequenced and analyzed in binding assays. A recently introduced alternative to cloning is the high-throughput sequencing of enriched DNA ligands obtained after each selection round (compare 1.4). (1) RNA library Target DNA Enriched oligonucleotide (7) ssDNA library ssDNA library library + Primer Primer

(2) randomized region (6) ssOligonucleotides Binding 20-60nt

8-20 (2a) Target immobilization SELEX rounds (3) Washing (5) Amplification - PCR - RT-PCR (4) Elution

Last Round - Sequencing Cloning Individual aptamers - Binding assays

Figure 2: Schematic representation of the SELEX procedure SELEX is a cyclic multistep process which starts with the synthesis of a DNA oligonucleotide library (1), containing a randomized region of 20 to 60 nt, which are flanked by fixed primer regions that are needed for PCR amplification. To start the actual selection, the library (ssDNA or RNA) is incubated with the target (2). Here, the incubation in solution is possible. In many cases the target is immobilized (2a) on a solid support, prior to the incubation with the target. Unbound molecules are washed away (3), while bound molecules are eluted (4) and amplified (5) by PCR and/or RT- PCR (reverse transcriptase-PCR for RNA). The final step of each selection round is the generation of ss Oligonucleotides (6) which display the enriched library (7) that is entered into the next selection round. This process is repeated 8-20 times, until specific binders with high affinity have been enriched. In the last selection round the enriched pool is cloned, and monoclonal ligands are characterized by sequencing and binding assays (modified from [14]).

- 5 -

Introduction

For selection of high affine binders, the stringency is increased over the selection rounds. This can be done by reducing the amount of target used for the selection, reducing the incubation time of the library with the target, and by more stringent washing procedures. For this, the volume of washing buffer, the incubation time, and the number of washing steps can be increased. Furthermore, the introduction of negative selection steps, using for example the selection matrix, minimizes the enrichment of unspecific binding ligands [14]. To estimate the amount of bound and unbound oligonucleotides during the SELEX process, radioactive labels were commonly used in the past. This method is very sensitive and enables the detection of very low amounts of DNA. The disadvantage of this method is that the whole SELEX procedure has to be conducted in an isotope laboratory, making the method expensive and environmentally incompatible [11]. Another crucial step in the SELEX procedure is the separation of target-bound and unbound nucleotides. Conventional separation methods are affinity chromatography (for protein targets) and filtration through nitrocellulose membranes (normally for small molecules) [80]. Both techniques have low separation efficiencies, and large amounts of target and library are required in their initial incubation [52]. With capillary electrophoresis SELEX [81] and FluMag-SELEX [11], an efficient and easy separation of bound and free molecules is possible. In FluMag-SELEX the use of magnetic beads has been described. The use of magnetic beads requires less target, and aptamers with high binding affinity (KD values of 57-85 nM) can be obtained. Another feature of FluMag-SELEX is the direct fluorescent labeling of the selected DNA which can be used for strand separation and quantification. One of the major processes during SELEX is the amplification of the selected and eluted ligands by PCR. Musheev et al. showed that PCR amplification of random DNA libraries differs significantly from that of homogenous samples and can lead to a complete loss of specific products if not controlled properly [82]. Therefore, the PCR amplification of random DNA libraries needs to be optimized carefully. The Taguchi approach is a fast method which can be used to achieve a prudent optimization of PCR amplification of random DNA libraries. This will be addressed later on in more detail (compare 1.3). For the generation of ss oligonucleotides, various methods have been described. One possibility is the labeling of one primer with fluorescein which allows the separation of differently sized DNA strands by preparative PAGE (polyacryl gel electrophoresis) [11]. Further possibilities for the generation of ssDNA are asymmetric PCR [83], alkaline denaturation [84], and lambda exonuclease digestion [85]. Lambda exonuclease is a highly processive 5’  3’ exodeoxyribonuclease that selectively digests the 5’-phosphorylated strand of dsDNA. For this purpose, a 5’-phosphate is introduced into the undesired strand by PCR over the amplification primer. Asymmetric PCR and PAGE are laborious methods which yield low amounts of ssDNA [86]. In SELEX experiments, this is undesirable due to the massive loss of target-binding ligands. Paul et al. demonstrated the detachment of covalently-bonded streptavidin from the bead surface after an alkaline treatment [87], that resulted in the presence of the undesired biotinylated strand in the eluate. The undesired

- 6 -

Introduction strand could then re-anneal to the complementary strand which then loses its tertiary structure that is important for the target-binding ability. Furthermore, another potential target, streptavidin, is then present during the selection. Avci-Adali et al. demonstrated that lambda exonuclease digestion resulted in ssDNA with higher yields and purity, when compared to the alkaline treatment [86]. Some of the possible modified SELEX procedures are summarized in Table 1. In general, the appropriate selection procedure depends on the target (cell SELEX) or describes special features of the library (genomic SELEX) used for selection or of the aptamers (Mod-SELEX) obtained after the selection. With Tailored-SELEX [88, 89] the size of the aptamers is reduced with minimal primer or primer-free selection. Another interesting approach is Toggle-SELEX [90]. The aim of Toggle-SELEX is the generation of species cross-reactive aptamers, which has been performed successfully to select an aptamer binding to human and porcine thrombin.

Table 1: Overview of different SELEX processes (modified from [3]) SELEX method SELEX modifications Aim Ref. Automated Use of automated systems for the Reduction of time needed for [91] SELEX selection procedure the selection Improvement of the separation Capillary Use of capillary electrophoresis for process between sequences electrophoresis [81] sequence partitioning bound to the target and the SELEX other sequences Selection of aptamers for Cell SELEX Use of whole living cells as target [34] unknown cellular markers Use of very small amounts of Use of fluorescent labels for DNA target for the aptamer selection, FluMag-SELEX quantification and use of magnetic [11] rapid and efficient separation of beads for target immobilization bound and free molecules Library composed of fragmented Selection of natural sequences Genomic SELEX [92] genomic DNA binding bioactive proteins Use of mirror analogues of natural Mirror-image Selection of nuclease-resistant nucleotides ( -ribose or [89, 93] SELEX L aptamers/Spiegelmers L-deoxyribose) Production of stable (nucleases- resistant) aptamers and SELEX on a library of generation of aptamers with Mod-SELEX oligonucleotides with chemical [94] conformations and target- substitutions binding surfaces not accessible when using DNA or RNA Library with reduced or no fixed Minimization of the aptamer Tailored SELEX [88, 89] regions size Selection of aptamers which can Toggle-SELEX Use of different targets in selection [90] bind to several related proteins

1.2.2 Principle of phage display and possible variations for the selection of peptide aptamers Phage display has first been described in 1985 by George P. Smith [95]. It is a molecular methodology by which a library of proteins or peptides is expressed on the surface of phage particles, while the genetic information encoding these proteins or peptides resides inside [96]. Filamentous phage strains such as M13, fd, and f1 are thread-shaped bacterial viruses

- 7 -

Introduction that are commonly used as vectors in phage display work [66, 72]. The (E.coli)-specific phage M13 is about 1 µM long and 6 nm in diameter (Figure 3) [72]. Filamentous phages are composed mainly of the major coat protein VIII (pVIII) helically arranged over the length of the particle. Inside this tube lies the ssDNA, coding for another four minor coat proteins. There are five copies of each of the minor coat proteins; pIII and pVI cap the one end, while pVII and pIX are located at the opposite end of the particle [97].

Figure 3: Schematic representation of an M13 phage particle The bacteriophage M13 is widely used for phage display. Foreign proteins have been fused to three coat proteins (pIII, pVIII, and pVI). Here, the fusion with the coat proteins pIII and pVI is limited to a maximum number of 5, while numerous copies of the protein can be expressed on the phage surface, when pVIII is used as a fusion partner. The approximate number of each M13 coat protein is indicated in parentheses (with kind permission from Lilija Miller [98]).

Since its first description, numerous applications and modifications of phage display have been described [96, 99]. Next to the identification of protein-protein interactions, epitope discovery, and affinity maturation of antibodies, the construction and screening of combinatorial peptide libraries is an important application [100, 101]. These libraries are constructed by cloning (such as synthetic oligonucleotides with a fixed length but unspecified codons) as fusions with the M13 genes III or VIII, where they are expressed as peptide-capsid fusion proteins (Figure 4) [100].

Population of DNA variants Packaging

Population of peptide variants

Figure 4: Random peptide phage display libraries Using the phage display methodology, a library of variant nucleotide sequences can be converted into a library of variant peptides or proteins. The displayed peptides or proteins can be natural peptides, synthetic random peptides, protein domains or whole proteins, and even antibody fragments. The amount of displayed foreign peptides or proteins depends on the coat protein chosen as a fusion partner and the phage display system used [72]. In such a library each phage clone bears its own individual amino acid sequence. Because of this amino acid variability, random peptide libraries can be screened against numerous targets to isolate phages displaying peptides with the desired properties (modified from [101]).

- 8 -

Introduction

In the resulting peptide library, each phage clone displays a single peptide (typically ranging in length from 5 to 20 amino acids [101]). However, a library as a whole may represent a large variety of amino acid sequences (107-109) altogether [66]. Once a library is acquired, the task is to screen it in such a way that the original diversity of the library is reduced to a manageable number of clones. These can then be analyzed in detail, most often done by using a form of affinity selection that is also known as “biopanning” [96, 101]. Numerous targets have been applied in biopanning experiments performed in vitro and in vivo [72, 102-104]. In general, the selection of peptide aptamers with phage display is feasible for any conceivable target. Possible targets include but are not limited to parasites [104, 105], viral particles [70, 71], enzymes [106], and even whole cells [72, 102, 107, 108]. The biopanning procedure can be performed with antigen columns, coated antigen, biotinylated antigen, against cells, or in vivo [100]. Even though there are numerous biopanning protocols, all consist of multiple steps illustrated in Figure 5.

Figure 5: Principle of a biopanning procedure In the simplest form of a biopanning procedure, the target molecule of interest is immobilized on a solid support by passive adsorption and incubated with the random peptide library to allow phage binding. Unbound phages are washed away, while bound phages are eluted and titered. The eluted phages are then amplified, titered again, and applied to the next selection round. Depending on target complexity, this procedure is repeated 3-6 times. After each selection round, individual phage clones from the unamplified eluate are analyzed by DNA-sequencing (with kind permission from Lilija Miller [98]).

- 9 -

Introduction

In its simplest form, the biopanning procedure is carried out by the immobilization of the target molecule onto a surface (surface panning). Subsequently, the random peptide library is added to allow phage binding. Unbound phages are washed away, while bound phage clones are eluted. Elution can be carried out by incubation with a solution containing either free target or a competing . Due to the stability of filamentous phages, another possibility is the use of extremes of pH, denaturants, or ionic strength for unspecific elution of the phages. The eluted phages are titered on agar plates to estimate the yield after each selection round. Part of the eluted phage clones are randomly selected from these plates and analyzed by DNA sequencing. Following elution and titration, the phages are amplified by infection of bacterial host cells, and then the amplified eluate is applied to another round of affinity selection. Depending on target complexity, the biopanning procedure is repeated 3 to 6 times. Following the last selection round and the data analysis from DNA sequencing, individual phage clones are propagated and analyzed to identify a target-binding peptide sequence [96, 101, 104]. As in the SELEX procedure, in phage display many modifications of the individual steps (binding, washing, elution, and amplification) are possible. Depending on the desired functionality of the selected peptide, the immobilization method can be chosen. Using passive adsorption, it is possible that the immobilized molecule is forced out of its functional configuration [101], which is undesirable if protein ligands are sought to functional versions of a target. For the selection of detection molecules, it is favorable to design the selection process in a close relation to the desired application of the ligands being obtained. The stringency of the selection process can be enhanced by modified washing steps. Here, the amount of washing buffer, the washing time, and the amount of detergent present in the washing buffer can be increased. The purpose of washing is to remove non-binding phage from the selection process so that binding phages are selectively enriched [101]. The washing step deserves some consideration, because a balance between specificity and avidity of selected phage clones is required. If washing is too stringent, highly specific but weak binders may be lost. If washing is not stringent enough, populations of selected clones may be dominated by strong binders with low specificity. A successful biopanning experiment results in the identification of consensus peptide sequences or short amino acid motives. The obtained phage clones can be tested for their specificity to the target used during selection. However, for more detailed binding studies, the selected ligands can be synthesized as free, soluble peptides. In doing so, the peptides can be modified during synthesis and can be applied at defined concentrations for analysis using various methods. These phage-borne peptides can be used as probes for detection of pathogens, either alone or in combination with antibodies or even nucleic acid aptamers.

1.3 Optimization of processes using the Taguchi approach

The Taguchi approach is a statistics-based design-of-experiments method which was developed by Genechi Taguchi, with the purpose to develop products with high quality and

- 10 -

Introduction functionality at low cost and within a short time frame [109]. Since then, many multifactorial scientific processes such as ELISA (-linked immunosorbent assay) [110], 2-D- gelelectrophoreses [111], PCR [112], and microarrays [113] were optimized using the Taguchi approach. These multifactorial processes are usually optimized by variation of one factor at a time, while the other factors are kept at a constant level. This is slow, costly, and inefficient compared to rationally designed optimization methods [113]. In the Taguchi approach, especially designed orthogonal arrays are employed to optimize the amount of information obtained from a limited number of experiments [114]. This means that results for an optimization process can be achieved with less experimental effort within a shorter time frame. For example, using the one-factor-at-a-time method, the optimization of a process with four reaction components at three levels would require 34=81 individual experiments, whereas the optimization using one of Taguchi’s orthogonal arrays (L9) needs only 9 experiments. Many different orthogonal arrays are available from the referenced literature. A selection of different orthogonal arrays is displayed in Table 2. To evaluate the optimal experimental conditions, the so-called signal-to-noise (SN) ratios are calculated. Usually, three types of SN ratios are available: 1) lower is better, 2) nominal is best, and 3) higher is better. The SN ratios are calculated with Taguchi’s corresponding quadratic loss functions.

Table 2: Orthogonal arrays for 4 (L9), 5 (L16), or 7 (L18) factors at three (L9 and L18) or four (L16) levels The levels (1-4) of the factors A-H are combined in such a way that depending on the array used 9- 18 different experimental setups are created. From each individual experiment a yield can be estimated from which a corresponding signal-to-noise (SN) ratio is calculated (see text for more details). Three different SN ratios are applicable which are called: lower is better, nominal the best, and higher is better [115]. L9 (34) L16 (45) L18 (21 × 37) Exp. Exp. Exp. A B C D A B C D E Aa) B C D E F G H Nr. Nr. Nr. 1 1 1 1 1 1 1 1 1 1 1 1 / 1 1 1 1 1 1 1 2 1 2 2 2 2 1 2 2 2 2 2 / 1 2 2 2 2 2 2 3 1 3 3 3 3 1 3 3 3 3 3 / 1 3 3 3 3 3 3 4 2 1 2 3 4 1 4 4 4 4 4 / 2 1 1 2 2 3 3 5 2 2 3 1 5 2 1 2 3 4 5 / 2 2 2 3 3 1 1 6 2 3 1 2 6 2 2 1 4 3 6 / 2 3 3 1 1 2 2 7 3 1 3 2 7 2 3 4 1 2 7 / 3 1 2 1 3 2 3 8 3 2 1 3 8 2 4 3 2 1 8 / 3 2 3 2 1 3 1 9 3 3 2 1 9 3 1 3 4 2 9 / 3 3 1 3 2 1 2 10 3 2 4 3 1 10 / 1 1 3 3 2 2 1 11 3 3 1 2 4 11 / 1 2 1 1 3 3 2 12 3 4 2 1 3 12 / 1 3 2 2 1 1 3 13 4 1 4 2 3 13 / 2 1 2 3 1 3 2 14 4 2 3 1 4 14 / 2 2 3 1 2 1 3 15 4 3 2 4 1 15 / 2 3 1 2 3 2 1 16 4 4 1 3 2 16 / 3 1 3 2 3 1 2 17 / 3 2 1 3 1 2 3 18 / 3 3 2 1 2 3 1 a) Each orthogonal array can be modified according to the user’s needs. If not all factors are needed, a factor can be left out.

- 11 -

Introduction

One of the most important steps in SELEX is the amplification of the selected random DNA libraries by PCR. If not controlled properly, the amplification of random DNA libraries can lead to a complete loss of the specific product [82], and thus to the loss of high affine and specific ligands, and finally even to the failure of selection [116]. The reaction solution of a

PCR contains several different components (buffer, MgCl2, dNTP’s, primer, and Taq polymerase) that can greatly influence the purity and the yield of the product [117, 118]. Furthermore, the initial amount of template, the primer annealing temperature, and the number of performed cycles can have a great impact on product specificity. The situation is further complicated by the fact that some of the variables are quite interdependent [119]. With Taguchi’s orthogonal arrays, these interdependent parameters are combined during one experiment. In SELEX PCRs, a high yield of specific product is favorable, therefore the quadratic loss function “better is best” is used for the calculation of SN ratios.

1.4 Application of high-throughput sequencing in the selection of DNA aptamers

The initial DNA libraries used in SELEX usually have a complexity of up to 1018 different nucleotide sequences [14]. Through the amplification of random DNA libraries with PCR, other nucleotide sequences not present in the initial library can evolve. The actual aim of SELEX is the reduction of random nucleotide sequences to a manageable amount. However, with classical cloning and Sanger sequencing, only a small fraction of these selected nucleotide sequences can be analyzed. An alternative to this approach is the application of next generation sequencing (NGS) or high-throughput sequencing (HTS) for the analysis of the selected oligonucleotides [120].

Table 3: Overview of NGS platforms available at the Robert Koch Institute Platform NGS chemistry Read length Run Reads per (bases) time run in mio Roche/ 454’s GS FLX Pyrosequencing 800 20 h 1 Reversible Illumina HiSeq 1500 100 (×2) 4 (8) d 1,500 terminators Similar to Ion Torrent PGM 400 6h 5 pyrosequencinga) a) The signal after incorporation of nucleobases is measured by differences in pH values.

Many different NGS platforms exist [121]. Table 3 gives an overview of NGS platforms that are available at the Robert Koch Institute. These platforms differ in library preparation, sequencing chemistry, read length (bases), running time, giga bases produced per run, and the acquisition costs for the sequencing machine and chemicals. One of these platforms was commercialized by Illumina. Using this platform, up to 1.5 billion sequences can be analyzed in parallel. The principle of the Illumina platform is based on sequencing-by-synthesis, using reversible terminator nucleotides for the 4 bases, each labeled with a different fluorescent dye. Before sequencing, the dsDNA fragments are ligated at both ends to adapters which are used as primers for amplification and later on for the identification of valid sequences. After

- 12 -

Introduction denaturation, the ssDNA fragments are hybridized to adapter-complementary oligonucleotides which are attached to the surface of a glass flow cell containing 8 separate lanes. To obtain sufficient light signal intensity for the reliable detection of the added bases, the hybridized fragments are clonally amplified in so called “clusters”. The flow cell is then placed into the sequencer, and each cluster is supplied with polymerase and the four fluorescence labeled terminator nucleotides. The 3’-OH of these nucleotides are chemically inactivated to assure the incorporation of only a single base per cycle. After the incorporation into the DNA strand, the nucleotide and its position is detected and identified via its fluorescent dye by imaging. The fluorescent dye is then removed and the 3’ end of the nucleotide deblocked for the next base incorporation cycle [122, 123]. Sequences with 100 base pairs (bp) in length are generated, which is sufficient for sequencing of DNA aptamers.

1.5 Orthopoxviruses as potential targets for aptamer selection

For the selection of peptide and DNA aptamers, the vaccinia virus (VACV) strain New York City Board of Health (NY) was chosen. The VACV belongs to the family of . The Poxviridae comprise a family of large, complex DNA viruses that infect , birds, and insects [124]. Poxviruses are classified into two subfamilies: the Entomopoxvirinae and the . The first one can cause infections and diseases in arthropods, while the latter can infect vertebrates. Nine genera of the Chordopoxvirinae are known, one of them being (OPV) [125]. The most prominent member of the OPV is the variola virus (VARV), the causative agent of . Since VARV only infected humans, it could be eradicated by a worldwide vaccination campaign by the WHO (World Health Organization) [126]. The OPV are further divided into different virus species, including camelpox virus (CMLV), virus (CPXV), virus (MPXV), ectromelia virus (ECTV), and VACV (some of which are important zoonotic pathogens [124, 125]). Different OPVs are immunologically cross-reactive and cross-protective [127], so that after infection with any member of this genus a protection against an infection with any other member of this genus is obtained. In the last two decades, the number of CPXV infections in cats, exotic animals, and humans has been increasing [127-129]. Furthermore, emerging human MPXV outbreaks in Africa and North America were reported [130-132], and cases of VACV infections in humans in South America and India were described [133-135]. Since its eradication, the only legitimately held VARV stocks are stored at the Centers for Disease Control and Prevention (CDC) in the United States of America (USA), and at the State Research Center for Virology and (VECTOR) in Russia [136]. However, it cannot be ruled out that unknown stocks of VARV exist in spite of these regulations. Infections of humans with OPVs might become more numerous owing to an absent or inadequate immune status of the population, because of the abrogated smallpox vaccination in the 1980s. With the low immune status of the population, and the increased fear of the

- 13 -

Introduction use of VARV as biowarfare agent [137-139], it is still important to research OPVs and to assure an accurate diagnostics of OPV and OPV infections. The linear dsDNA genome of OPV has a size of about 200 to 230 kbp and codes for approximately 150 to 200 genes which are divided into early, intermediate, and late genes [140]. OPV replicate in the cytoplasm, in so-called “virus factories” of infected host cells, and develop enveloped brick-shaped virions with a size of approximately 140-170 nm × 300- 350 nm × 250 nm [127]. During OPV replication, four different types of virions are produced that differ from one another by their outer membrane [141]. There are two major infectious forms of VACV: the intracellular mature virus (IMVs) and the extracellular enveloped virus (EEV) (Figure 6).

Figure 6: Schematic virion structure of an OPV particle The virus core contains the S-shaped linear dsDNA genome of OPV which is wrapped by core- associated enzymatic and non-enzymatic proteins. Lateral bodies (grey) flank the core. The core is surrounded by one (IMV) or two (EEV) lipid membranes, into which several viral proteins are embedded. The function of the different poxvirus proteins is explained in the legend (kindly provided by Daniel Stern, Robert Koch Institute, Berlin).

The IMVs represent the majority of infectious progeny; they are surrounded by a single membrane and remain inside the cell until cell lysis. EEVs are surrounded by a second membrane and are exported from the cell before cell death [142]. The EEV outer membrane is an extremely fragile structure which is damaged by virus purification. Nevertheless, the particle retains full infectivity as an IMV. IMVs are thought to be important for long-term stability and transmission of the virus between hosts in the environment [143]. In contrast, EEVs are responsible for fast cell-to-cell spread and long-range dissemination [142, 144].

- 14 -

Introduction

However, different viral proteins are found within either lipid membrane, which makes IMV and EEV structurally, antigenically, and functionally different. The differences between IMVs and EEVs affect virus attachment and entry into cells, egress from cells, and virus dissemination [145].

1.5.1 Overview of different IMV and EEV envelope proteins Several non-glycosylated viral proteins are embedded in the lipid membrane of IMVs. Some examples are A27, D8, F9, L1, and H3 (Table 4) [146]. A27 (MW 12.6 kDa) forms a triple coiled-coil structure and interacts with A17 (MW 21 kDa) through a C-terminal α-helix [147]. Furthermore, A27 is known to interact with cell surface glycosaminoglycans, mainly heparan sulfate, and is therefore associated with virus-cell fusion and entry into the cell [148]. D8 (MW 35 kDa) binds to cell surface chondroitin sulfate and mediates the adsorption of IMV to cells. Soluble D8 can block the binding of IMV to cells and therefore can inhibit VACV infection at the adsorption stage [149, 150]. F9 is associated with virus entry and cell- cell fusion. L1 (MW 27 kDa) is a myristoylated protein with 20 % sequence homology to F9 and is known to be involved in cell entry, fusion of IMV with the cytoplasm membrane, and virus assembly [151-153]. The H3 protein is able to bind to heparin sulfate, and is involved in virion assembly and virus infection [154].

Table 4: Overview of different VACV membrane proteins Protein Localization Function Mol %/ Weight %a) A17 IMV Assembly/fusion 0.29/0.26 Attachment/fusion A27 IMV 8.41/4.09 microtubule transport D8 IMV Attachment/entry 1.74/2.39 Attachment/IMV H3 IMV 2.13/3.08 maturation L1 IMV Entry/fusion 0.14/0.15 A33 EEV Spreading -/- B5 EEV Spreading -/- a)Data according to [146]

The proteins A33 and B5 are located in the membrane of EEVs. The B5 protein (MW 42 kDa) is essential for efficient EEV formation and is associated with spreading of the virus [155, 156]. The protein A33 is involved in binding of actin tails and colocalizes with B5 and A36 [157]. Furthermore, all proteins mentioned are known to induce neutralizing antibodies [158, 159].

- 15 -

2. Aims of study

Classical methods for the detection of common clinical infections, newly emerging infections, and bio-threat agents are based on PCR and immunological detection assays using antibodies. While PCR-based methods are not able to detect particles, antibodies exhibit considerable limitations. Novel detection molecules, such as peptide and DNA aptamers could overcome these limitations. Methods for the selection of such aptamers are SELEX and phage display. The aim of this thesis was to develop and evaluate DNA and peptide aptamers as detection molecules for OPVs. Thus, the two main objectives were:

I. The selection of DNA aptamers, using SELEX, and

II. The selection of peptide aptamers, using phage display in a BSL2 laboratory.

For the selection of DNA aptamers, a suitable protocol needed to be established. For this, the following working steps were required:

1. Establishment of a suitable SELEX protocol, including optimization of different SELEX steps (immobilization, elution, and amplification),

2. Screening of different random DNA libraries against OPV proteins and native VACV particles,

3. Assessment of the applicability of NGS for the identification of high affine DNA aptamers against native VACV particles, and comparison to the classical cloning and sequencing approach,

4. Selection of DNA sequences for synthesis, and

5. Characterization of potential target-binding DNA ligands.

Additionally, the selection of peptide aptamers necessitated the following working steps:

1. Screening of random phage display libraries against OPV proteins and native VACV particles,

2. Characterization of potential target binding phage clones,

3. Selection of sequences for peptide synthesis, and

4. Characterization of synthetic peptides.

At last, a DNA and peptide aptamer-based sandwich assay for the detection of VACV particles was to be developed.

- 16 -

3. Materials and Methods

The following chapter summarizes materials and technical equipment and will provide a description of experimental approaches and methods used in this study. All materials not listed in 3.1 will be provided later in this thesis along with method descriptions.

3.1 Materials

Additives and chemicals

3,3´,5,5´-Tetramethylbenzidine (TMB) tablets Sigma, St. Louis, USA 4',6-Diamidino-2-phenylindol (DAPI) Invitrogen, Darmstadt, Germany 5-Bromo-4-chloro-3-indolyl-β-D-galactoside Fermentas, St. Leon-Rot, Germany (X-Gal) Acetic acid (CH3COOH) Merck, Darmstadt, Germany Ammonium bicarbonate Sigma, St. Louis, USA Bovine serum albumin Carl Roth, Karlsruhe, Germany Calcium chloride dihydrate (CaCl2*2H2O) Merck, Darmstadt, Germany Casein sodium salt Sigma-Aldrich, Deisenhofen, Germany Dimethyl formamide (DMF) Merck, Darmstadt, Germany Dimethyl sulfoxide (DMSO) Serva, Heidelberg, Germany Disodium phosphate (Na2HPO4) Merck, Darmstadt, Germany Ethanol Carl Roth, Karlsruhe, Germany Ethylenediaminetetraacetic acid (EDTA) Gibco BRL®, Eggenstein, Germany UltraPure™ 0,5 M pH 8 Ethylenediaminetetraacetic acid disodium salt Sigma, St. Louis, USA dihydrate (Na2EDTA) Glycerol (Ultrapure™) Invitrogen, Darmstadt, Germany GE Healthcare UK Limited, Buckinghamshire, Glycin HCl pH 1.7 England HEPES Carl Roth, Karlsruhe, Germany Hydrochloric acid (HCl) Carl Roth, Karlsruhe, Germany Hydrogen peroxide (H2O2) Sigma, St. Louis, USA Imidazole ≥ 99 % (Glyoxalin, Carl Roth, Karlsruhe, Germany 1,3-Diazza-2,4-cyclopentadien) Isopropyl-β-D-thiogalactoside (IPTG) Fermentas, St. Leon-Rot, Germany Magnesium chloride Hexahydrate Carl Roth, Karlsruhe, Germany (MgCl2*6H2O) Polyethylene glycol (PEG) SigmaUltra Sigma, St. Louis, USA (MW:8,000) Ribonucleic acid, transfer from baker's yeast Sigma-Aldrich, Deisenhofen, Germany (S. cerevisiae) t-RNA Sodium chloride (NaCl) Merck, Darmstadt, Germany Sodium hydroxide (NaOH) Carl Roth, Karlsruhe, Germany Sodium iodide dihydrate (NaI*2H2O) Sigma, St. Louis, USA Sucrose Merck, Darmstadt, Germany Tetracycline hydrochloride Sigma, St. Louis, USA Triton X-100 Sigma, St. Louis, USA Trizma® hydrochloride (Tris-HCl) Sigma, St. Louis, USA Trizma®Base (Tris-Base) Sigma. St. Louis, USA Tryptone Gibco BRL®, Eggenstein, Germany Tween 20 Sigma-Aldrich, Deisenhofen, Germany UltraPure™ Agarose Invitrogen, Darmstadt, Germany

- 17 -

Materials and Methods

Yeast extract Gibco BRL®, Eggenstein, Germany

Agarose gel electrophoresis

6 × Loading dye Fermentas, St. Leon-Rot, Germany 6 × Orange DNA Loading Dye Fermentas, St. Leon-Rot, Germany NuSieve Agarose FMC BioProducts, Rockland, USA Low Range Ultra Agarose (for NGS) Bio-Rad Laboratories, Munich, Germany Ethidium bromide (EtBr) Carl. Roth, Karlsruhe, Germany GeneRuler™ 100 bp DNA-Ladder Fermentas, St. Leon-Rot, Germany O’GeneRuler™ Ultra Low Range DNA Ladder, Fermentas, St. Leon-Rot, Germany ready-to-use SYBR® Gold nucleic acid gel stain Invitrogen, Darmstadt, Germany TAE for NGS Illumina, Eindhoven, Netherlands

Antibodies

Polyclonal Antibody to Vaccinia Virus (Lister Acris Antibodies, Herford, Germany strain) – Purified (rabαVACV) Polyclonal Antibody to Vaccinia Virus (Lister Acris Antibodies, Herford, Germany strain) – Biotin (rabαVACV-Biotin) Vaccinia Virus antibody (20-VR69) Fitzgerald, Acton, USA Vaccinia Virus (Polyclonal Antibody) MyBiosource.com Monoclonal anti-A27 A3/710-20 Robert Koch Institute (RKI), Berlin, Germany Monoclonal anti-A27 A3/710-20 -Biotin RKI, Berlin, Germany Monoclonal anti-A27 A1/40-1 RKI, Berlin, Germany Monoclonal anti-A27 A1/40-1-Biotin RKI, Berlin, Germany Monoclonal anti-A27 A1/6-15 RKI, Berlin, Germany

Bacterial Strains

E.coli ER2738 New England Biolabs, Ipswich, USA One Shot® TOP10 Electrocompetent E.coli Invitrogen, Darmstadt, Germany

Buffers, solutions, and media

Phosphate buffered saline (PBS, without Mg2+ 8.0 g NaCl, 0.2 g KCl, 1.44 g Na2HPO4 , ad 1 L and Ca2+) ddH2O 80 g NaCl, 2 g KCl, 14.4 g Na HPO , ad 1 L 10 × PBS 2 4 ddH2O Agarose gel electrophoresis 242 g Tris Base, 57.1 g glacial acetic acid, Tris-acetate-EDTA (TAE)-buffer (50 x, pH 8) 100 mL 0.5 M EDTA, ad 1 L ddH2O 20 mL 50 × TAE-buffer, 100 µL EtBr TAE-EtBr buffer (1 ×) (10 mg/mL), ad 1 L ddH2O Phage Display 10 g NaCl, 10 g tryptone, 5 g yeast extract, ad Luria-Bertani (LB) broth 1 L ddH2O, adjust to pH 7.0 with 5 N NaOH, autoclave LB agar 1 L LB broth, 20 g agar, pour into petri dishes LB top agar 1 L LB broth, 0.7 % (w/v) agarose, autoclave 1 L LB broth, 1.25 mL of 10 mg/mL-filter- LB-Tetracycline broth sterilized tetracycline; store in dark, cool place 1 L LB agar cooled to 55°C, 1.25 mL of LB-Tetracycline agar 10 mg/mL-filter-sterilized tetracycline, pour into petri dishes; store in dark, cool place

- 18 -

Materials and Methods

1.25 g IPTG, 1 g X-Gal, 25 mL DMF, store at IPTG / X-Gal stock -20°C 1 L LB agar cooled to 55°C, 1 mL IPTG /X-Gal LB / IPTG / X-Gal plates stock, pour into petri dishes; store in dark, cool place 20 mg/mL in 1:1 Ethanol: ddH O, store at Tetracycline stock 2 -20°C 20 % (w/v) PEG-8,000, 2.5 M NaCl, autoclave, PEG / NaCl store at room temperature (RT) 10 mM Tris-HCl pH 8.0, 1 mM EDTA, 4 M Iodide buffer sodium iodide (NaI), autoclave, store at RT in the dark, discard if color is evident Coating buffer 0.1 M NaHCO3 pH 8.6 Blocking buffer 1 PBS, 0.05 % (w/v) Casein Blocking buffer 2 PBS, 2 % (w/v) BSA PBS, 0.05 % (w/v) Casein, 0.5 % (v/v) Dilution buffer 1 Tween 20 Dilution buffer 2 PBS, 2 % (w/v) BSA, 0.5 % (v/v) Tween 20 Washing buffer PBS, 0.5 % (v/v) Tween 20 Elution buffer (nonspecific disruption of 0.2 M Glycine-HCl pH 2.2 binding) Neutralization buffer (elution step) 1 M Tris-HCl pH 9.1 Pull-down assay and SDS PAGE Ammonium bicarbonate buffer 0.1 M (NH4)HCO3 12.11 g Tris, 87.66 g NaCl, ad 1 L ddHO , 10 × TBS running buffer 2 pH 7.6 SELEX Magnesium chloride hexahydrate (for 1 M) 20.33 g ad to 100 ml ddH2O Calcium chloride dihydrate (for 1 M) 14.7 g ad to 100 ml ddH2O 100 mL 10 × PBS, 1 mL MgCl2 (1 M), 1 mL Selection buffer (SB) CaCl2 (1 M), 10 mL Tween 20 (0.5 %), ad to 1 L, pH 7.6 Blocking buffer SB, 3 % (w/v) BSA, 0.1 mg/mL t-RNA Washing buffer SB, 0.005 % (w/v) Tween 20 10 × (v/v) PCR buffer, 500 mM imidazole, His elution buffer 2 mM MgCl2, ad to 1 mL ddH2O Elution buffer (VACV selection) RLT buffer, Qiagen, Hilden, Germany Biacore 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, Running buffer HBS-EP 0.05 % Tween 20 Regeneration buffer 10 mM Glycine HCl pH 1.7

Cell culture Cell culture flasks Nunclon ™ ∆ Surface Nunc™, Wiesbaden, Germany (25-175 cm2) CryoTubes™ (1 mL and 1.8 mL) Nunc™, Wiesbaden, Germany Serological pipettes (1-50 mL) Nunc™, Wiesbaden, Germany Nunc Multidishes Nunclon™ ∆ (24 and 96 Nunc™, Wiesbaden, Germany wells) Falcon tubes (15 mL and 50 mL) TPP,Trasadingen, Switzerland D-MEM culture medium Gibco BRL®, Eggenstein, Germany E-MEM culture medium Gibco BRL®, Eggenstein, Germany Fetal calf serum (FCS) Gibco BRL®, Eggenstein, Germany L-glutamine PAA, Pasching, Germany

- 19 -

Materials and Methods

1:2 mixture of Trypsin / EDTA PAA, Pasching, Germany Neubauer cell counting chamber Blau, Wertheim, Germany HEp2 - Human cervix carcinoma, HeLa European Collection of Cell Cultures (ECACC): derivate 86030501 Vero76 – Kidney tissues of an African green ECACC: 85020205 monkey (Ceropithecus aethiops) Composition of cell culture media For Vero D-MEM, 10 % FCS, 2 mM L-glutamine For HEp2 D-MEM, 5 % FCS, 2 mM L-glutamine

Enzyme conjugates Goat anti-mouse IgG (Fc Fragment) conjugated Dianova, Hamburg, Germany to horseradish peroxidase (HRP) HRP / Anti-M13 monoclonal conjugate GE Healthcare Europe, Freiburg, Germany Peroxidase-conjugated AffiniPure Goat Anti- Dianova, Hamburg, Germany Rabbit IgG (HR+L) Peroxidase-conjugated Streptavidin Dianova, Hamburg, Germany

Kits 660 nm Protein Assay Pierce, Rockford, USA Dynabeads® Co-Immunoprecipitation Kit Invitrogen, Darmstadt, Germany NucleoSpin® Extract II Macherey-Nagel, Düren, Germany NucleoSpin® Plasmid QuickPure Macherey-Nagel, Düren, Germany ssDNA / RNA Clean & Concentrator™ Zymo Research, Irvine, USA Agilent High Sensitivity DNA Kit Agilent Technologies, Waldbronn, Germany TrueSeq™ SR ClusterKit v3 Illumina, Eindhoven, Netherlands TrueSeq™ SBS Kit v3 (200 cycles) Illumina, Eindhoven, Netherlands Quant-iT™ OliGreen® ssDNA Reagent and Kit Invitrogen™, Darmstadt, Germany Invitrogen™, Molecular Probes®, Darmstadt, Qubit™ ssDNA Assay Kit Germany TOPO TA Cloning® (pCR®2.1-TOPO®) Invitrogen, Darmstadt, Germany BigDye® Terminator v.3.1 Cycle Sequencing Applied Biosystems, Foster City, USA Kit Invitrogen™, Molecular Probes®, Darmstadt, Qubit® ds BR Assay Kit (2-1,000 ng) Germany MinElute® Gel Extraction Kit Qiagen, Hilden, Germany KAPA SYBR® FAST Universal qPCR Kit KAPABiosystems, Woburn, USA

Oligonucleotides

M13 F (5’-GTAAAACgACggCCAg) Metabion, Martinsried, Germany M13 R (5’-CAggAAACAgCTATgAC) Metabion, Martinsried, Germany Invitrogen, Darmstadt, IBA, Göttingen, For DS (5’-ATgCCAgCTTATTCAACC)a) Germany Invitrogen, Darmstadt, IBA, Göttingen, Rev DS (5’-AgACTgCACCgAATAACC)b) Germany DS pool (90 mer) ATgCCAgCTTATTCAACC- IBA, Göttingen, Germany (N54)-ggTTATTCggTgCAgTCT Invitrogen, Darmstadt, Metabion, Martinsried, FOR JM (5’-CTgATgCTgACTgCgACC)a) Germany Invitrogen Darmstadt, Metabion, Martinsried, Rev JM (5’-gTCAgCgTCgAgCAgTACC)b) Germany JM pool (67 mer) 5’-CTgATgCTgACTgCgACC- TiB MOLBIOL, Berlin, Germany (N30)-ggTACTgCTCgACgCTgAC - 20 -

Materials and Methods

JM pool (77 mer) 5’-CTgATgCTgACTgCgACC- TiB MOLBIOL, Berlin, Germany (N40)- ggTACTgCTCgACgCTgAC 96 gIII primer (5’-CCCTCATAgTTAgCgTAACg) Invitrogen, Darmstadt, Germany a) The oligonucleotides For DS and For JM have also been ordered with 5’-biotin and 5’-FAM. b) The oligonucleotides Rev DS and Rev JM have also been ordered with 5’-phosphate.

Other consumables

Microcentrifuge tubes, safe-lock (0.5-2.0 mL) Sarstedt, Numbrecht, Germany Nunc-Immuno™ Plates Polystyrene, MaxiSorp Nunc, Wiesbaden, Germany F96 Pierce® NeutrAvidin® Coated 96-Well Plates Pierce, Rockford, USA Qubit™ assay tubes Invitrogen™, Karlsruhe, Germany SA Chip GE Healthcare, Munich, Germany

Phage Display

Ph.D.-12 linear dodecapeptide library New England Biolabs, Ipswich, USA

Polymerase Chain Reaction (PCR)

Platinum® Taq DNA-Polymerase Invitrogen™, Karlsruhe, Germany 10 x PCR buffer Invitrogen™, Karlsruhe, Germany MgCl2 Invitrogen™, Karlsruhe, Germany dNTP (Deoxyribonucleotide triphosphate) Amersham, Freiburg, Germany PCR water (DNase free, Fluka) Sigma-Aldrich, Deisenhofen, Germany TaqMan PCR-Plates (96 well) Applied Biosystems, Weiterstadt, Germany Optical Adhesive Covers Applied Biosystems, Weiterstadt, Germany PCR tubes (0.2 mL) Thermo Fisher Scientific, Rockford, USA PCR stripes Thermo Fisher Scientific, Rockford, USA PCR stripe cover Thermo Fisher Scientific, Rockford, USA

Proteins A27 (NR-2626) BEI Resources, Manassas, USA A27 E.coli RKI, Berlin, Germany A33 (NR-545) BEI Resources, Manassas, USA B5 (NR-546) BEI Resources, Manassas, USA B5 E.coli GenExpress, Berlin, Germany D8 E.coli GenExpress, Berlin, Germany F13 E.coli GenExpress, Berlin, Germany F9 (NR-2626) BEI Resources, Manassas, USA H3 E.coli GenExpress, Berlin, Germany L1 (NR-2625) BEI Resources, Manassas, USA

Pull-down assay and SDS PAGE

Dynabeads® M-280 Streptavidin Invitrogen Dynal AS, Oslo, Norway Flamingo fluorescent gel stain Bio-Rad Laboratories, Munich, Germany Halt Protease Inhibitor Cocktail (PI) 100 x Thermo Fisher Scientific, Rockford, USA 6 × Laemmli sample buffer Bio-Rad Laboratories, Munich, Germany Precise™ Protein Gels (8-16 % & 4-20 %, Pierce, Rockford, USA 12 wells) Precision Plus Protein Standard All Blue Bio-Rad Laboratories, Munich, Germany RIPA Buffer Pierce, Rockford, USA

- 21 -

Materials and Methods

SELEX

Dynabeads® His-Tag Isolation and Pulldown Invitrogen Dynal AS, Oslo, Norway DynaMag™-2 Magnet Invitrogen Dynal AS, Oslo, Norway Lambda Exonuclease Fermentas, St. Leon-Rot, Germany RLT buffer (elution VACV selection) Qiagen, Hilden, Germany Random nucleotide libraries (67 mer, 77 mer) TIB MOLBIOL, Berlin, Germany Random nucleotide library (90 mer) IBA, Göttingen, Germany

Software

7500 Software v2.0.5 Applied Biosystems™, Foster City, USA Adobe Photoshop CS5 v12.0.3 Adobe Systems Incorporated, San Jose, USA Agilent Bioanalyzer Software Agilent Technologies, Waldbronn, Germany Biacore X-100 Evaluation Software 1.0 GE Healthcare, Munich, Germany BioEdit v7.0.9.0 Tom Hall, Ibis Biosciences, Carlsbad, USA EndNote X5.0.1 Thomson Reuters, New York, USA GraphPad Prism 5 v5.01 GraphPad Software, San Diego, USA E.A.S.Y Win32 Herolab, Wiesloch, Germany Image Lab v4.0.1 Bio-Rad Laboratories, München, Germany ImageJ v1.42q National Institutes of Health, USA Library Insert Finder (Author: Wojtek RKI, Berlin, Germany Dabrowski) Magellan-Data analysis software v6&7 Tecan Group, Männedorf, Switzerland MxPro - Mx3005P v4.10 Stratagene, USA Biocomputing Lab, School of Computer PseudoViewer3 Web Application Science and Engineering Inha University, Inchon, Korea Sequence Scanner v1.0 Applied Biosystems™, Foster City, USA ValFold 1.0.0 Joey Akitomi,[160]

Technical equipment

7500 Applied Biosystems™, Foster City, USA Agilent Bioanalyzer 1000 Agilent Technologies, Waldbronn, Germany Biacore X-100 GE Healthcare, Munich, Germany cBot Illumina, Eindhoven, Netherlands ChemiDoc Bio-Rad Laboratories, Munich, Germany Consort EV231 Consort bvba, Turnhout, Belgium Dark Reader® DR89X transilluminator Clare chemical research, Dolores, USA Easy Nano LC and LTQ Orbitrap Discovery™ Thermo Fisher Scientific, Rockford, USA mass spectrometer Fast Semi-Dry Blotter Pierce, Rockford, USA Gel electrophoresis chamber Horizon 58 Gibco BRL®, Eggenstein, Germany Gel electrophoresis power supply ST304 Gibco BRL®, Eggenstein, Germany Heraeus Megafuge 16R Centrifuge Thermo Fisher Scientific, Waltham, USA HydroFlex™ microplate washer Tecan Group, Männedorf, Switzerland Illumina Sequencer Illumina, Infinite® 200 PRO microplate reader Tecan Group, Männedorf, Switzerland Mastercycler® ep gradient Eppendorf, Hamburg, Germany Microwave oven Privileg 8520 Privileg, Furth, Germany Mini-Protean® 3 Cell Bio-Rad Laboratories, Munich, Germany Mx3000P QPCR System Agilent Technologies, Santa Clara, USA NanoDrop™ ND-1000 Spectrophotometer PeQ Lab , Erlangen, Germany NanoDrop™ ND-3300 Spectrophotometer PeQ Lab Biotechnologies, Erlangen, Germany Perfect Blue™ Horizontal MiniGel Systems PeQ Lab Biotechnologies, Erlangen, Germany

- 22 -

Materials and Methods

Qubit® 2.0 Fluorometer Invitrogen™, Molecular Probes®, Thermomixer comfort Eppendorf, Hamburg, Germany Video documentation system Herolab, Wiesloch, Germany

Viruses

Isolated from a bat [161], propagated on Adenovirus-2 PPV1 (Bat AdV-2 PPV1) Vero76 cells Camelpox virus, strain CP-19, kindly provided CMLV CP-19 from Sandra Essbauer IMB, Munich, propagated on HEp2 cells Cowpox virus, strain Brighton Red, ATCC VR- CPXV BR 302™, propagated on REF A1 cells Cowpox virus, strain GuWi, isolated from an CPXV GuWi infected elephant [128], propagated on HEp2 cells Feline calicivirus (FCV) ATCC VR-2057™, propagated on CRFK cells Herpesvirus isolate 10-43-001-01, RKI, Berlin, HSV-1 Germany M13K07 Helper Phage Invitrogen™, Karlsruhe, Germany Ectromelia virus (Mousepox), Sandra MP-Nü Essbauer, IMB, Munich, propagated on HEp2 cells , Sandra Essbauer, IMB, MPXV München Munich, Germany MPXV X 12-06-011-01 Monkeypox virus, propagated on HEp2 cells Parapox virus (PPXV) Strain D1701, Tübingen, Hanns-Joachim Rziha Porcine Parvovirus (PPV) ATCC VR-742™, propagated on PK13 cells Vaccinia virus, strain IHD-W, ATCC VR-1354™, VACV IHD-W propagated on HEK 293T cells Vaccinia Virus, strain New York City Board of VACV NYCBOH Health, ATCC VR-1536®, propagated on HEp2 cells Vaccinia virus, strain Western reserve, ATCC VACV WR VR-1441™, propagated on HEK 293T cells

3.2 Cell culture and virus propagation

HEp2 is an adherent continuous cell line originating from a human cervix carcinoma. HEp2 cells were used for generation of IFA slides, which were used to evaluate the ability of the selected aptamers to fluorescence label OPV-infected cells, for antiviral activity testing of selected aptamers, and for propagation of some of the viruses utilized in this thesis. Vero76 is an adherent continuous cell line, originating from kidney tissue of an African green monkey. Vero76 cells were used for propagation of the adenovirus and for titration of virus suspensions.

3.2.1 Cell preservation and recovery

All cell lines used were preserved by freezing (-80°C) for short-term storage or cryopreserved (gas phase of liquid nitrogen) for long-term storage. For this, sub-confluent (70-80 %) cell cultures were trypsinated and resuspended in culture medium. Cell

- 23 -

Materials and Methods concentration was determined using a “Neubauer cell counting chamber”. The cell suspensions were centrifuged for 10 min at 1,000×g. The supernatant was then discarded, and the cells resuspended in cold Freeze-mix (10 % DMSO, 90 % FCS) to a final cell number of 2-4×106 cells/mL. Afterwards, the cells were dispensed into cryovials, cooled down on ice, and frozen at -20°C. Finally, the cells were stored at -80°C or in the gas phase of a liquid nitrogen freezer. To recover frozen cells, vials were thawed at 37°C and resuspended in fresh culture medium. DMSO is toxic for cell lines and therefore needs to be removed. Cell suspensions were centrifuged for 10 min at 1,000×g, and the cells were then washed with culture medium before they were seeded into culture flasks.

3.2.2 Maintenance and subculture routine

All the cell lines were maintained according to the recommendations of the European Collection of Cell Culture (ECACC) and the American Type Culture Collection (ATCC). All the cell lines were cultured at 37°C, 5 % CO2, and 90 % humidity and passaged 1:3 to 1:10, depending on cell density and demands, twice a week. For this, medium was aspirated and the cells were washed once with PBS. After washing, the cells were trypsinated by adding a trypsin/EDTA mixture and incubated at 37°C for 5-10 minutes. For splitting to 1:5, fresh

1 medium was added and /5 was transferred to a new cell culture flask. The flask was then filled with 4/5 of new culture medium. When a defined cell number was needed, cells were counted using a “Neubauer cell counting chamber”. The required cell dilution was seeded into appropriate cell culture plates or flasks.

3.2.3 Virus propagation

All viruses used in this work are listed above, and were provided from colleagues. VACV NYCBOH, used for the selection of DNA and peptide aptamers, was propagated and purified as described before [162]. All the other viruses listed were used for specificity testing of the selected aptamers.

3.3 Nucleic acid aptamer selection via SELEX

SELEX is a cyclic repetitive process that enriches highly specific aptamers (ssDNA/RNA) binding to a target of choice out of a complex nucleotide library of up to 1015 different molecules. Three different DNA libraries with randomized regions of 30, 40, or 54 nucleotides (nt) respectively, were chosen for the selection. All three libraries are flanked by fixed primer regions with a length of 18 to 19 nt which were used for the amplification of respective libraries. This resulted in libraries with a total length of 67 nt, 77 nt, and 90 nt. Primers for the 90 mer were based on published sequences [39] and slightly modified. The primers for the other libraries (67 mer and 77 mer) were developed based on information stated in Design, Synthesis, and Amplification of DNA Pools for In Vitro Selection [163]. Single SELEX steps include: 1. target immobilization, 2. binding of nucleotides to the target, 3.

- 24 -

Materials and Methods washing away unbound nucleotides, 4. the elution of bound nucleotides, 5. the amplification of eluted nucleotides, and 6. finally the generation of ss nucleotides to start a new SELEX round.

3.3.1 Optimization of individual SELEX steps

Factors influencing successful aptamer selection include target presentation, PCR amplification, efficient elution, and library complexity. To generate aptamers capable of detecting OPVs in a sandwich ELONA (enzyme-linked oligonucleotide assay), selection was carried out against native VACV particles captured by an anti-OPV (mAb) for up to 12 consecutive rounds. To regain bound nucleotides, the elution step had to be carefully optimized. Robust intermittent DNA amplification was assured by thorough PCR optimization employing the Taguchi methodology. The influence of library complexity was assessed by parallel selection using three ssDNA libraries (67 mer, 77 mer, and 90 mer).

Target immobilization and elution of bound nucleotides Aptamers were selected against recombinant vaccinia virus proteins and against native vaccinia virus particles. Recombinant proteins were expressed with C- or N-terminal His-tag and could be coupled to magnetic beads. Bound protein and therefore bound nucleotides were eluted using imidazole and heat. To evaluate the optimal elution conditions, a bead ELISA was performed by comparing 3 elution times (5, 7, and 10 min) and 3 different imidazole concentrations (0.2, 0.5, and 1 M) for His-tagged proteins. For the optimization, 1 µg of recombinant A27 (rA27) was mixed with 5 µL Dynabeads® and incubated for 10 min at room temperature (RT) in a total volume of 100 µL. After that, the beads were washed 4 times (2 min with 500 µL SB at 800 rpm on a Thermomixer Comfort). Finally, the protein was eluted for 5, 7, or 10 min with 0.2 M, 0.5 M, or 1 M imidazole. An additional incubation step was performed for each sample for 5 min at 95°C. The supernatant was discarded; the beads were incubated with anti-A27 monoclonal antibody (mAb) A1/6-15 (1:10,000, 30 min, RT) then washed 4 times with 1 mL SB each. The anti-A27 mAb was incubated with goat anti-mouse HRP (horse radish peroxidase)- labeled secondary antibody (1:2,500, 30 min, RT) and washed 4 times for 2 min with 500 µL SB each. A TMB (3,3’5,5’-tetramethylbenzidine) solution was used as a substrate for the enzyme HRP. The reaction was stopped with 2 M H2SO4. The solution was then transferred into wells of an ELISA plate and assayed at 450 nm with a reference measurement at 620 nm using the Infinite®200 PRO microplate reader. Additionally, the same experiment was performed without the elution step. The sample with the highest difference in OD signal (non-eluted –eluted sample) represented the best elution condition. The elution step for the selection against native VACV particles was optimized comparing different lysis buffers (Cytos All, RIPA, AL Qiagen, and RLT Qiagen). To estimate which lysis buffer was most suitable for the elution, an OPV real-time PCR assay was used (Table 5) to quantify genomic pox-viral DNA. The real-time PCR was performed in a 25 µL reaction

- 25 -

Materials and Methods containing the reaction components listed in Table 6. For quantification, a plasmid standard with the OPV target sequence was measured in each run. The real-time PCR was performed on the Mx3000P QPCR System under cycling conditions listed in Table 6.

Table 5: Oligonucleotides used for real-time PCR Name Sequence OPV F TAATACTTCgATTgCTCATCCAgg OPV R ACTTCTCACAAATggATTTgAAAATC OPV TMGB FAM‐TCCTTTACgTgATAAATCAT NFQ MGB The base “G” is given in lower letters to avoid confusion with “C”. F - Forward, R - Reverse, FAM – FAM label, MGB – Minor Groove Binder, NFQ – Non-fluorescent quencher

Following the elution, the ssDNA libraries had to be purified to remove genomic pox-viral DNA, which would interfere with ssDNA quantification, and the lysis buffer, which could inhibit library amplification. Therefore, two purification kits (NucleoSpin® Extract II [Macherey-Nagel] and the ssDNA / RNA Clean & Concentrator™ [Zymo Research]) were compared for their suitability to eliminate genomic pox-viral DNA and to preserve the ssDNA libraries. The manufacturer’s protocol was followed. To eliminate genomic poxviral DNA, two columns for dsDNA were used, or the columns for dsDNA were used twice for the same sample. To quantify the reduction of genomic pox-viral DNA, again a real-time PCR assay was used. To check if the library could still be detected, a conventional PCR with library-specific primers was performed.

Table 6: Components and cycling conditions for real-time PCR 25 µL reaction volume containing Cycling conditions PCR water 16.6 µL Temperature Duration Cycle number PCR buffer (10 ×) 2.5 µL 95°C 10 min 1 × MgCl (50 mM) 2.0 µL 95°C 15 sec 2 45 × dNTPs (25 mM) 1.0 µL 60°C 34 sec Primer F (10 µM) 0.75 µL Primer R (10 µM) 0.75 µL Probe (10 µM) 0.25 µL Taq (5 U/µL) 0.15 µL Sample Volume (DNA) 1.0 µL

PCR optimization Musheev and Krylov have shown that amplification of heterogeneous DNA libraries differs significantly from amplification of homogenous samples, and that, if not controlled properly, PCR amplification of random libraries could lead to a complete loss of specific products by rapid accumulation of by-products [82]. During the SELEX process, 3 different PCR amplification steps were conducted: 1. pre-amplification of eluted ssDNA ligands to generate dsDNA ligands, 2. amplification of dsDNA ligands, produced in PCR 1 to be used in the next SELEX round, and 3. amplification of single clones to be used in binding assays. For these 3 PCRs, 3 primer sets were used (Table 7).

- 26 -

Materials and Methods

Table 7: Used primer sets for amplification of aptamers Name Sequence 5’ to 3’ Used for PCR FOR DS ATgCCAgCTTATTCAACC 1 and 2 FOR DS Bio Biotin-ATgCCAgCTTATTCAACC 3 Rev DS AgACTgCACCgAATAACC 1 p-Rev DS Phosphate-AgACTgCACCgAATAACC 2 and 3 FOR JM CTgATgCTgACTgCgACC 1 and 2 FOR JM Bio Biotin-CTgATgCTgACTgCgACC 3 Rev JM gTCAgCgTCgAgCAgTACC 1 p-Rev JM Phosphate-TCAgCgTCgAgCAgTACC 2 and 3

In PCR 1, the ssDNA ligands were amplified using unmodified primers. This allows the repetition of a SELEX round without needing to start all over again, to clone the respective round for sequencing of single ligands, and to preserve material to prepare a library for high- throughput sequencing with the Illumina sequencer (see 3.4). PCR 2 was performed to obtain a sufficient amount of material to be used in the next SELEX round. The reverse primer was 5’-phosphorylated to generate ssDNA through λ-Exonuclease digestion.

Table 8: Cycling conditions for PCR optimization with Taguchi arrays Temperature Duration Cycle number 94°C 3 min 1 × 94°C 15 sec 58-64°Ca) 15 sec 12-24 ×b) 72°C 15 sec 72°C 5 min 1 × 4°C hold a) Temperature depending on array used, for L9 and L16 array 64°C were used for 90 mer, and 62°C for 67 mer and 77 mer b) Cycle number depending on array used, the L9 array was performed for 12 & 16 cycles (90 mer) or 16 & 20 cycles (67 mer and 77 mer)

PCR 3 was performed to biotinylate single clones, to be used in an ELONA-binding assay (see 3.3.2). Since all 3 PCRs used different primer combinations, all 3 PCRs had to be optimized individually. The Taguchi approach was used to optimize all PCRs. To estimate which orthogonal array was most suitable, the arrays L9, L16, and L18 were used to optimize PCR 1 for all three ssDNA libraries. Factors to be optimized and their concentration levels are listed in Table 9. All PCRs were performed in triplicate in 25 µL reactions, with cycling conditions listed in Table 8.

- 27 -

Materials and Methods

Table 9: Orthogonal arrays with optimized factors (A-H) and their concentration levels (1-4) The L9 array was performed with 10 fmol and 20 fmol ssDNA for 12 and 16 cycles (90 mer) and for 16 and 20 cycles (67 mer and 77 mer). The L16 array was performed with 20 fmol and 100 fmol ssDNA. L9 1 2 3 L18 1 2 3

A MgCl2 [mM] 2 4 8 A / / / /

B Taq [U] 0.5 1 1.5 B MgCl2 [mM] 0.5 2 6 C Primer [µM] 0.2 0.6 1.2 C Taq [U] 0.5 1 1.5 D dNTPs [µM] 50 200 350 D Primer [µM] 0.2 0.4 0.6 E dNTPs [µM] 50 200 400 L16 1 2 3 4 F cycles [x] 12 16 20

A MgCl2 [mM] 2 4 6 10 G annealing temp. [°C] 58 62 64 B Taq [U] 0.5 1 1.5 2 H DNA [fmol] 40 20 10 C Primer [µM] 0.2 0.4 0.6 0.8 D dNTPs [µM] 10 50 200 400 E cycles [x] 12 16 20 24

Visualization and analysis of PCR products For analysis, 5 µL of the PCR products were loaded with 1 µL 6 × Orange DNA Loading Dye onto 3.5-5 % agarose gels (containing 0.4 µg/mL EtBr) and then run with 1 × TAE-EtBr buffer in a horizontal gel chamber (Horizon) for 30-40 min at 90 V constant voltage. Additionally, as reference for PCR product length and comparison of signal intensity across gels, 1.5 µL of O’GeneRuler™ Ultra Low Range DNA Ladder was applied. The amplification products were visualized under ultra violet light (305 nm), photographed with E.A.S.Y Win32 software, and analyzed with ImageJ as described by Luke Miller [164] to calculate product yields. In doing so, the signal intensity of the gel band of the PCR product was related to the intensity of the 300 bp gel band of the DNA ladder (this band represents the signal intensity of 9 ng of a 300 bp band). If no PCR product was detectable, the signal intensity was set to 1 and then related to the signal intensity of the first band of the DNA ladder, because the yield must be >0 to fulfill Taguchi’s quadratic loss function. The product yields were then used to calculate the SN ratios for each factor at each factor level. For each factor, the optimal conditions are represented by the highest SN ratio. Alternatively, optimum reaction conditions can be estimated by maximizing the SN ratios using polynomial regression analysis. Both analysis methods were performed and the results compared. The SN ratios were calculated using Taguchi’s quadratic loss function higher is better:

Equation 1: Taguchi’s quadratic loss function higher is better n is the number of trails with given concentrations, yi is the product yield in corresponding trails

- 28 -

Materials and Methods

3.3.2 Selection of aptamers against recombinant A27

The selection against the OPV envelope protein A27 was performed with the 90 nt library. For the first selection round, the library was mixed with the components listed in Table 10 in a final volume of 1,660 µL, incubated for 15 min at 95°C with rocking at 800 rpm, and then actively cooled down to 23°C in a Thermomixer comfort. After denaturation and renaturation, Tween 20 was added to a final concentration of 0.005 %, and 15 µL were sampled for quantification of ssDNA.

Table 10: Components for denaturation and renaturation reaction for selection against A27 (selection round 1) Component Volume [µL] Final concentration PBS (10 ×) 166 1 × MgCl2 (50 mM) 33.2 1 mM CaCl2 (50 mM) 33.2 1 mM H2O 1,371.8 Library (100 µM) 16.6 1 µM (1.66 nmol)

For the actual selection step, the library was incubated with 1.66 nmol recombinant A27 (BEI resources) for 30 min at 23°C, using the Thermomixer comfort with interval shaking at 800 rpm (10 sec every 20 sec). Additionally, t-RNA was added to a final concentration of 0.1 mg/mL to prevent unspecific binding of ssDNA to the protein. To separate bound from unbound ssDNA, 28 µL of washed ‘Dynabeads® His-Tag Isolation & Pull-down’ were added to the protein/DNA mixture and incubated for 30 min at 23°C with interval shaking at 800 rpm. After incubation, beads were washed three times with 500 µL SB for 2 min at 1200 rpm to remove unbound DNA. The protein and the bound DNA were eluted by incubation with 250 µL His elution buffer for 5 min at 23°C (1200 rpm) and 10 min at 95°C. For quantification, 10 µL were sampled, the remaining sample was directly amplified with unmodified primers in a total volume of 500 µL, split to 20 × 25 µL, using reaction components and cycling conditions listed in Table 11. For amplification with marked primers, 16 µL of PCR1 were amplified in a total volume of 400 µL (16 × 25 µL) for 18 cycles, with reaction component concentrations as in Table 11.

Table 11: Components and cycling conditions for PCR1 (selection against A27) 500 µL reaction volume containing Cycling conditions PCR water 142.00 µL Temperature Duration Cycle number PCR buffer (10 ×) 26.0 µL 94°C 3 min 1 × MgCl2 (50 mM) 20.0 µL 94°C 15 sec dNTPs (2.5 mM) 40.0 µL 64°C 15 sec 20 × Primer F (10 µM) 15.00 µL 72°C 15 sec Primer R (10 µM) 15.00 µL 72°C 5 min 1 × Taq (5 U/µL) 2.0 µL 4°C hold Sample Volume (DNA) 240.0 µL

PCR products were analyzed by gel electrophoresis, as described under Visualization and analysis of PCR products. The PCR product of PCR2 was purified using the “NucleoSpin® ExtractII” Kit following the manufacturer’s instructions. PCR products were eluted with 50 µL per column after incubation for 5 min at RT. Concentrations of the samples were - 29 -

Materials and Methods determined by OD-measurement with the NanoDrop™1000. To generate ssDNA to be entered into the next SELEX round, samples/PCR products were diluted to 40 ng/µL and digested with λ-Exonuclease following the manufacturer’s instructions. The reaction was stopped by incubation at 80°C for 10 min. The sample was purified again with the “NucleoSpin® ExtractII” Kit, using buffer NTC for purification of ssDNA and eluted with 50 µL NE buffer after incubation for 5 min at RT. The amount of ssDNA was quantified by OD measurement with the NanoDrop™ 1000 spectrophotometer. Corresponding concentrations in pmol were calculated to assure addition of the same amount (molar mass) of protein in the next selection round to keep the ratio of DNA to protein 1:1. Starting in the second selection round, the library was diluted in a total volume of 70 µL using all components listed in Table 10 and the entire enriched amplified library from the prior round. To increase the stringency of subsequent selection rounds, the library was denatured for 5 min, and the incubation with the protein and the beads was reduced to 20 min and 10 min, respectively. Additional washing steps were added. The amount of added beads was dependent on the amount of added protein and exceeded the maximum binding capacity of the beads by at least 2-fold.

3.3.3 Selection of aptamers against native vaccinia virus particles

The final SELEX protocol was based on all optimization processes described above. The monoclonal anti-A27 capture antibody A3/710-20 was diluted to 2 µg/mL in coating buffer, and 100 µL were immobilized on MaxiSorp ELISA plates overnight at 4°C. The following day, wells were emptied, filled completely with blocking buffer (SB + 3 % BSA (w/v) + 0.1 mg/mL of t-RNA), and incubated for 1 hr at RT. In the meantime, the libraries were mixed with components listed in Table 12 in a final volume of 100 µL, incubated at 95°C for 10 min, and put on ice for an additional 10 min.

Table 12: Components for denaturation and renaturation reaction for selection against native VACV particles (selection round 1) Component Volume [µL] Final concentration PBS (10 ×) 10 1 × MgCl2 (50 mM) 2 1 mM CaCl2 (50 mM) 2 1 mM H2O 66-76a) Library (50-100 µM) 10-20b) 10 µM (1 nmol)c) a, b) Volume of H2O and library was dependent on library: the 90mer had a stock concentration of 100 µM and the other libraries had a stock concentration of 50 µM. From round two on, 50 µL of the enriched library were used c) Final library concentration in rounds 2-9 was dependent on obtained amounts in prior selection rounds

For quantification of ssDNA, 10 µL were sampled. To the remaining volume/sample, Tween 20 and t-RNA were added to a final concentration of 0.005 % (v/v) or 1 mg/mL, respectively. Blocked wells were washed 3 times with 300 µL of SB, 5 × 105 pfu of VACV 1536 were added, and the plate was incubated at 37°C for 1 hr. Wells were washed 5 times with

- 30 -

Materials and Methods

300 µL of SB, the denatured libraries were added, and the plate was incubated with gentle rocking (150 rpm) for 1 hr at RT. Starting in round 2, a counter selection step was performed by incubation of the library with the capture Ab for 1 hr prior the incubation with the VACV particles. After incubation with the libraries, wells were washed 10 times with 300 µL of SB in the first round. Washing steps were increased by 1 in consecutive rounds, not exceeding 15 washing steps. Bound nucleotides were eluted by incubation with 150 µL of RLT buffer for 15 min at 37°C. The eluate was transferred into sample tubes and incubated for 1 hr at 60°C to inactivate any remaining native VACV. All samples were purified with ssDNA/RNA Clean and Concentrator™ Kit (Zymo Research) following the manufacturer’s instructions, but using the column for removal of dsDNA twice. Samples were eluted with 50 µL of H2O after incubation for 5 min. 10 µL of the eluted libraries were used for quantification of ssDNA. The remaining sample was amplified with optimized conditions for PCR 1 as described before (Optimization of PCR using the Taguchi approach). To estimate the optimal cycle number, a test PCR for 16 × and 20 × cycles was performed. The PCR product of PCR 1 was purified using the NucleoSpin® ExtractII Kit following the manufacturer’s instructions. The PCR products were eluted with 50 µL per column after incubation for 5 min at RT. A part of PCR 1 was diluted and amplified with marked primers and optimized conditions for PCR2 for 16 to 20 cycles. PCR products were analyzed by gel electrophoresis as described under Visualization and analysis of PCR products. PCR products were again purified as described above. Generation of ssDNA was performed as described under 3.3.2. The selection process was repeated for a total of 12 selection rounds.

3.3.4 Quantification of ssDNA and binding assays

There are different ways to estimate/determine after which selection cycle to end the SELEX process. One possibility is to quantify ssDNA before and after each SELEX round and to calculate the yield of the respective round. The amount of ssDNA was quantified using the OliGreen® ssDNA Reagent and either the NanoDrop™ ND-3300 Spectrophotometer, the Infinite® 200 PRO microplate reader, or the Qubit® 2.0 Fluorometer following the manufacturer’s instructions for the OliGreen® ssDNA Reagent. Binding assays were performed after 6 and 12 SELEX rounds for the libraries used for VACV selection (77 mer and 90 mer), which is a second possibility to determine after which selection cycle to end the selection process. For this, the unselected library and enriched libraries from each round were amplified with biotinylated and phosphorylated primers and optimized PCR component concentrations for 12 to 16 cycles. PCR products were digested with λ-Exonuclease as described before. For the binding assay, MaxiSorp ELISA plates were coated over night with 200 ng/well of mAb A3/710-20 in 100 µL of coating buffer. On the following day, wells were aspirated and then completely filled with blocking buffer and incubated for 1 hr at 37°C. Amplified libraries from each selection round were diluted to 100 nM and treated as described under 3.3.3. Blocked wells were washed 3 times with 300 µL of SB. To one half of the wells 5×105 pfu/well of VACV were added and incubated for

- 31 -

Materials and Methods

1 hr at RT. The other half was filled with SB to estimate background binding of the libraries. Subsequently, all wells were washed 5 times with 300 µL of SB. A volume of 100 µL of each amplified library was added and incubated for 1 hr at RT. Wells were washed 8 times and 100 µL of streptavidin peroxidase (SA-POD) diluted 1:2,500 in SB + 0.25 % (w/v) BSA were added and incubated for 30 min at RT. Finally, the wells were washed four times, a TMB substrate was added, incubated for 15 min, and stopped with H2SO4. The color reaction was assayed at 450 nm with reference measurement at 620 nm using the Infinite® 200 PRO microplate reader.

3.3.5 Cloning of enriched DNA library

Specific ligands of the aptamer library were enriched after several selection rounds. To obtain individual ligand sequences from the 90 mer and the 77 mer selected against VACV, amplified libraries of selection rounds 2, 4, 5, 6, 8, 9, and 12 were cloned into vector pCR®-II- TOPO® using the “TOPO-TA Cloning®” Kit from Invitrogen and transformed into One Shot® TOP10 Electrocompetent E.coli following manufacturer’s instructions. To separate ligand sequences of the 90 mer selected against rA27, the enriched library of round nine was cloned and transformed. Positive clones were identified via blue-white screening and colony PCR as suggested by the manufacturer.

3.3.6 Plasmid DNA isolation

Positive clones were transferred to 3 mL of LB medium and amplified overnight at 37°C. Overnight cultures were used for isolation of plasmid DNA using the “NucleoSpin® Plasmid QuickPure” Kit according to manufacturer’s instructions. The plasmid DNA was eluted with 50 µL of buffer AE after incubation for 5 min at RT. The DNA concentration was determined using a NanoDrop™ ND-1000 Spectrophotometer and used directly for sequencing of insert DNA.

3.3.7 Sanger Sequencing of DNA inserts

The purified plasmid DNA was used to determine the aptamer sequence cloned into the vector. To analyze DNA sequences, the standard Sanger sequencing method was used. The sequencing reaction was performed using the “BigDye® Terminator v3.1 Cycle Sequencing” Kit. For the sequencing reaction, the M13 F plasmid primer was used. The cycling conditions and components applied in a sequencing reaction are listed in Table 13. The separation of the fragments was done using the 3130xl and 3500xl Dx Genetic Analyzer capillary electrophoresis instruments.

- 32 -

Materials and Methods

Table 13: Components, cycling conditions, and DNA amounts for a sequencing reaction Components Amount Cycling conditions 5 × ABI buffer 1.5 µL Temperature Duration Number of cycles Primer 0.5 µL 96°C 2 min 1 × BigDye 3.1 1 µL 96°C 10 sec DNA 1-5 µla) 55-60°Cb) 5 sec 25 × PCR-grade water ad 10 µL 60°C 4 min DNA amounts Plasmids 150-300 ng ssDNA 50-300 ng a) Depending on DNA concentration b) Depending on primer used

Data processing All obtained sequences were analyzed with the BioEdit software and aligned to the forward and reverse primer for each library in order to check the insert length and the orientation of the insert. When the reverse primer was found, the sequence was displayed reverse complement. Sequences were listed and compared.

Selection of DNA sequences for synthesis Single DNA sequences were chosen for synthesis to analyze their binding characteristics. Criteria for synthesis were the results from the binding assay from individual clones (A27 selection) or quantity of obtained DNA sequences (VACV selection).

3.4 High-throughput sequencing for identification of enriched aptamer libraries from VACV selection

To obtain a broader view of aptamer sequences obtained, the next generation sequencing (NGS) or high-throughput sequencing (HTS) technology was used. If each selection round is sequenced, several million sequences from each selection round can be compared. By this, it is possible to see if specific sequences have been enriched from round to round. Cho et al. have shown that enriched sequences can display better sequences in terms of affinity and specificity than those that only have the highest quantity [120]. Therefore, the enriched libraries from 8 out of 12 selection rounds of the 90 mer and the 77 mer libraries applied in the selection against native VACV particles were sequenced.

3.4.1 Library preparation

Illumina’s TruSeq DNA™ Sample Preparation Kit 48 Set A v2 was used to prepare the enriched aptamer libraries for sequencing on the Genome Analyzer II (Illumina). 1 µg of amplified libraries selected against VACV (see 3.3.3) was subjected to preparation for sequencing, which entails end repair, addition of adenosine to 3’ ends, adapter ligation, fragment size selection by gel extraction, and PCR following the manufacturer’s instructions. If less than 1 µg of amplified PCR product was available, the maximum amount available was used (see Table 14). All samples were quantified with the “Qubit® ds BR Assay” Kit (2-

- 33 -

Materials and Methods

1,000 ng) and the Qubit® 2.0 or 3.0 Fluorometer prior to library preparation. On an Illumina flow cell 8 lanes are available. To sequence all 16 selection rounds simultaneously, the adapter indices provided and also suggested by the manufacturer (Table 14) were used. Following adapter ligation, all samples were separated on a 2 % agarose gel, containing SyBr Gold, using a 100 bp Ladder as size marker to select the specific size range of molecules needed for proper cluster formation on the cluster station. The gel bands with a size of about 300 bp were excised and extracted, using the “MinElute gel extraction” Kit by Qiagen following Illumina’s instructions. DNA fragments were enriched and purified as suggested by Illumina, and the libraries were validated as described under 3.4.2.

Table 14: Amount used of amplified ligands obtained in PCR 1 and adapter indices used for NGS Selection round Used amount [ng] Index no. Used amount [µL]a) Lane 77-1 ?b) 4 0.25 1 & 2 77-2 134 5 0.25 1 & 2 77-3 1000 6 2.5 1 & 2 77-4 533 12 1.25 1 & 2 77-5 1000 4 2.5 3 & 4 77-6 1000 5 2.5 3 & 4 77-7 358 6 0.93 3 & 4 77-8 655 12 1.65 3 & 4 90-1 1000 4 2.5 5 & 6 90-2 417 5 1.0 5 & 6 90-3 890 6 2.08 5 & 6 90-4 1000 /c) / 5, 6 & 7 90-5 1000 4 2.5 7 & 8 90-6 372 5 0.93 7 & 8 90-7 196 6 0.5 7 & 8 90-8 753 12 1.85 7 & 8 a) Amount of index used was dependent on available amount of amplified library. The standard 1 index amount was 2.5 µL/1 µg of aptamer library. If only /10 of the needed library amount was 1 available, only /10 of the corresponding index was used. b) The amount of round 1 of the 77mer could not be determined, therefore all of the remaining volume was used for library preparation. The index amount was adapted to the lowest amount used. c) Round 4 of the 90mer was used to check if the protocol for the library preparation was suitable and to estimate best conditions for sequencing. Therefore, the library for round 4 was not prepared and sequenced again on this flow cell.

3.4.2 Validation of DNA libraries for NGS

Before the NGS libraries are spotted on the flow cell, the quality and the quantity of the prepared libraries should be controlled.

Quality control of DNA libraries The libraries used have a length of 90 nt or 77 nt. With the adapters added, a total length of about 200 bp will be achieved. To verify if the prepared enriched libraries have the correct length, all samples were diluted 1:50 and analyzed with the Agilent Bioanalyzer, using the “High Sensitivity DNA Assay” Kit following manufacturer’s instructions.

- 34 -

Materials and Methods

Quantification of DNA libraries To assure that the same amount of the enriched libraries of every selection round was hybridized on the flow cell, the DNA libraries had to be quantified. The “KAPA Library Quantification” Kit for Illumina Genome Analyzer platform for quantification was used following manufacturer’s instructions, with minor modifications. All samples and standards were assayed in duplicate. Samples were diluted 1:10,000 and 1:100,000 and ran on a real- time cycler 7500.

3.4.3 Spotting flow cell and NGS sequencing run

After validation, the prepared libraries were ready to load into the Illumina cluster station. The cluster station was used to hybridize the prepared aptamer libraries to a flow cell which would then be placed into the Genome Analyzer to obtain sequence data. The cluster station bridge amplified each single molecule 35 times to generate cloned clusters, which were large enough for individual nucleotides to be visualized by the optics of the Sequencer [120]. Quantified NGS libraries were adjusted to 2 nM. Subsequently, 10 µL of the 2 nM solution of selection rounds 1-4 and 5-8 for each library were pooled. Finally, 5 µL of 2 nM PhiX control was added to 5 µL of the pooled libraries, and the libraries were denatured as suggested by the manufacturer. The PhiX control is a reliable, adapter-ligated library used as a control for Illumina sequencing runs. The library is derived from the small, well- characterized PhiX (a bacteriophage) genome and can be used as a high-level spike-in for unbalanced samples. The libraries were loaded into the flow cell cluster station at a concentration of 10 pM with 50 % PhiX, and spotted onto the flow cell as indicated in Table 14. For clustering, the “TruSeq™ SR Cluster” Kit v3 was used following the Reagent Preparation Guide for c-Bot- HiSeq/HiScanSQ. The chemistry for the sequencing run was prepared according to the instructions for the “TruSeq™ SBS” Kit v3 (200 cycles). The flow cell was loaded into the Sequencer which ran for 101 cycles.

3.4.4 NGS data processing

After sequencing, the data were processed using Illumina Pipeline Software which performs base calling and quality filtering. Downstream analysis was performed using software developed internally. At first, sequences were scanned for indices to sort sequences according to their selection round. Then, sequences were scanned for forward and reverse primers and the insert length. Correct sequence reads would include 90 nt or 77 nt of sequence, depending on the library sequenced, including 18 nt of 5’ primer + 54 nt or 40 nt aptamer + 18 nt 3’ primer for the 90 mer or 19 nt for the 77 mer. Hereby, the last 5’ primer and the first 3’ primer were used to eliminate concatemeric sequences that only contained primer sequences. Any sequences not matching these criteria were filtered out. Primer sequences were trimmed out after filtering, leaving 54 nt or 40 nt for downstream analysis.

- 35 -

Materials and Methods

Reverse complement tags were combined with the sequenced forward tags and tags were count.

3.5 Peptide aptamer selection by phage display

The following sections describe the methodological approach for the selection of peptide ligands for a specific target using a random peptide phage display library. The protocols were adapted from the Ph.D.™ Phage Display Libraries - Instruction Manual (New England BioLabs®Inc., Ipswich, USA) and established at the Centre for Biological Threats and Special Pathogens at the Robert Koch Institute in Berlin by Lilija Miller [162]. All alterations to the protocols are stated below.

3.5.1 Random peptide phage display libraries

The phage display methodology is a biomolecular tool with versatile application options. Within the scope of this thesis one of these options, the phage display of random peptide libraries, was performed and compared to SELEX. With established materials and protocols for biopanning experiments, a commercially available random peptide library was tested for binding to recombinant VACV proteins and infectious/native VACV particles. All biopanning experiments were performed with a premade Ph.D.™ Phage Display library. The Ph.D.™ system is based on the M13KE cloning vector. The random peptides are fused to the minor coat protein pIII, and the random library consists of a linear dodecapeptide (Ph.D.-12) (Figure 7).

Figure 7: Premade Ph.D. random peptide phage display library The random peptide sequences (X12) are fused to the N-terminus of the minor coat protein pIII of the filamentous phage M13 (green oval). The library contains a short linker sequence (GGGS) between the displayed peptide and pIII. The library has a complexity on the order of 109 independent clones (modified from Ph.D.™ Phage Display Libraries - Instruction Manual, New England Biolabs® Inc.).

3.5.2 Bacterial strain maintenance and culture

The recommended E.coli host strain ER2738 was used to amplify and titer filamentous phages from the Ph.D.™ library. For maintenance, ER2738 was streaked onto LB-tetracycline agar plates from glycerol cultures, incubated at 37°C overnight, and stored at 4°C for a maximum of up to one month. For phage titering, 10 mL of LB medium were inoculated with a single bacterial colony and grown at 37°C and 200 rpm for 5-6 hr (OD600 ~ 0.5). 200 µL of this titration culture were used per titering plate. For amplification of phage eluates and individual phage clones, 20 mL of LB-tetracycline medium were inoculated with a single bacterial colony and grown overnight (~12-14 hr) at 37°C and 200 rpm. Overnight cultures were diluted 1:100 in LB medium and used for phage amplification.

- 36 -

Materials and Methods

3.5.3 Phage titering

The phage titering was performed according to the manufacturer’s instructions. The titer was calculated as described in Equation 2.

Equation 2: Calculation of phage titer pfu/ mL - plaque forming unit per mL

3.6 Identification of peptide ligands to recombinant OPV proteins and infectious VACV particles

The random peptide library was tested for binding to recombinant A27, a poxviral surface protein, and to native VACV particles.

3.6.1 Biopanning against recombinant A27

For identification of peptide ligands to recombinant A27, MaxiSorp ELISA plate wells were coated with 10 µg/mL of protein in 150 µL of coating buffer. One additional well per biopanning round was filled with blocking buffer 1 (PBS/0.05 % [w/v] Casein) for negative selection. The Ph.D.™-12 library was diluted to 2×1011 pfu/well in dilution buffer 1 (PBS/0.05 % [w/v] Casein / 0.5 % [v/v] Tween20). In the first selection round, a negative selection was performed against uncoated washed wells, by incubation for 1 hr at RT with gentle rocking (Thermomixer comfort). The supernatant from the negative selection well was transferred to target coated blocked wells, and incubated again for 1 hr at RT with gentle rocking. After selection, the supernatant was aspirated, and the wells were washed 20 times with 300 µL/well of washing buffer. Bound clones were eluted with 100 µL of elution buffer for nonspecific disruption of binding by incubation for 15 min at RT. The eluate was transferred to a microcentrifuge tube and neutralized with 15 µL of 1 M Tris-HCl pH 9.1. The eluted phages were titered as described above (3.5.3). Amplified titered phages were diluted to 2×1011 pfu/well and used for subsequent biopanning rounds as described above. A total of three selection rounds were performed. Prior to every positive selection round, a negative selection against uncoated (only 1st round) or Casein-coated wells was performed. To increase stringency in the subsequent selection rounds, additional washing steps were added.

3.6.2 Biopanning against VACV particles

The biopanning against infectious VACV particles was performed together with Lilija Miller as described above with minor modifications for a total of 4 selection rounds. For selection against VACV particles, 5×105 pfu/well in 150 μL of coating buffer were coated. The blocking and dilution buffer contained 2 % BSA instead of Casein. Negative selection was performed against blocked wells in the first 3 rounds and against HEp2 cell lysate after the

- 37 -

Materials and Methods fourth selection round to reduce the amount of unspecifically binding phage clones. Incubation times for negative and positive selection were 2 hr. The eluate of each round was incubated in a water bath for 1 hr at 60°C to inactivate infectious VACV particles that might be potentially present

3.6.3 Amplification of phage clones

To identify potential consensus sequences, selected binding phage clones were characterized by ELISA and sequencing. For this purpose, individual plaques were picked from the titration plates of unamplified eluates with a cut-off pipette tip. For characterization of binding phage clones, 30 (10) clones were randomly selected from the first round eluate, 50 (20) clones were selected from the second round eluate and finally 61 (20) clones were selected from the third round eluate for A27 selected clones (VACV selected clones). Four selection rounds were performed against VACV particles. In the last selection round, 68 clones were randomly picked. The selected clones were amplified following manufacturer’s instructions. The amplified phages were stored at 4°C until needed.

3.6.4 DNA sequencing of selected phage clones

Due to the fact that the library of peptide variants is expressed on the outside of a phage virion, while the genetic material encoding each variant resides on the inside, peptides could be identified through sequencing of single-stranded phage DNA. The ssDNA was purified from amplified phage clones, as described by the manufacturer. The DNA concentration was determined by using the NanoDrop™ ND-1000 Spectrophotometer, and prepared ssDNA was sequenced as described in 3.3.7 using the -96 gIII sequencing primer.

3.6.5 Identification of peptide consensus sequences

The Sequence Scanner v1.0 software was used to evaluate the quality of the DNA sequences. The sequences were then exported and saved in FASTA format. For the identification of randomized peptide sequences, these FASTA files were uploaded to “Library Insert Finder” software (this software was programmed at the Robert Koch Institute in Berlin for rapid and efficient identification of the randomized DNA region inserted into the phage clone genome). With this software, a contemporary translation of uploaded random DNA sequences identified into amino acid sequences was possible. The software first transformed the obtained DNA sequence into a complementary DNA strand. The position of the random DNA sequence was then identified through a definition of the constant flanking sequences (Figure 8). The random peptide sequences that had been obtained were then compared. If a random peptide sequence was identified more than once among the randomly selected binding clones, it was noted as a potential consensus sequence, and the clone was then characterized further.

- 38 -

Materials and Methods

Figure 8: Sequence of random peptide library-gIII fusion The library is expressed with an N-terminal leader sequence that was removed upon secretion at the position indicated by the arrow, resulting in the creation of the randomized peptide positioned directly at the N-terminus of the mature protein. The positions of the constant sequences flanking the randomized insert sequence are highlighted in blue. The position of the sequencing primer -96 gIII is highlighted in red (modified from Ph.D.™ Phage Display Libraries – Instruction Manual, New England Biolabs® Inc.).

3.6.6 Phage ELISA

Phage clones with potential consensus sequences and all the other clones from the last selection round (A27 selection) were applied to phage ELISA to confirm specific binding to the target. The ELISA was performed as described by the manufacturer with some minor changes. For phage clones selected against native VACV particles, MaxiSorp ELISA plates were coated with 105 pfu/well of VACV and with porcine parvovirus (PPV) as negative control. For phage clones selected against recombinant A27, MaxiSorp ELISA plates were coated with 200 ng/well of rA27, rL1, or 105 pfu/well of VACV, or parapox virus (PPXV) and AdV as negative controls. The plates were blocked with PBS/BSA (for clones from VACV selection) or PBS/Casein (for clones from A27 selection). To estimate background binding of selected phage clones to the corresponding blocking reagent, two wells per clone were coated with blocking buffer. Phage clones (1010 pfu/well) were pipetted into coated, blocked wells and incubated for 1-2 hr at RT. A TMB solution was used as substrate for the enzyme HRP. Plates were assayed at 450 nm with reference measurement at 620 nm using the Infinite®200 PRO microplate reader.

3.7 Selection of sequences for peptide synthesis

For further analysis, individual clones were chosen for synthesis. Selected clones were synthesized as free soluble synthetic peptides, with a biotin molecule at the C-terminus of the peptide. The selection was based on results from phage ELISA and DNA sequencing.

- 39 -

Materials and Methods

Peptide sequences of clones with the following characteristics were selected for synthesis:  for anti-VACV clones, those with stronger ELISA signal to VACV compared to BSA or PPV-coated wells,  for anti-A27 clones, those with stronger ELISA signal to rA27 and VACV compared to rL1, PPXV, AdV, and Casein,  and those with the highest quantity.

Peptides were ordered from PANATecs GmbH (Tuebingen, Germany) or Biomatik (Cambridge, Canada).

3.8 Characterization of selected aptamers

The synthetic aptamers ordered were delivered as lyophilized powder and had to be dissolved prior to further experimentation. The aptamers were dissolved in sterile, distilled water and stored in aliquots at -20°C. The nucleic acid aptamers were dissolved to a final concentration of 100 µM, and the peptide aptamers were dissolved to a final peptide concentration of 2 mg/mL. To characterize aptamers, binding kinetics were assessed and specificity testing was performed. Using native viral particles, all ELISA and ELONA steps were performed in a class-II biological safety cabinet with personal protective equipment. Experiments with monkeypox virus (MPXV) particles were performed in a BSL3 laboratory. The plates were washed using a multichannel pipette by repeatedly pipetting up and down. ELISA plates not containing any infectious viral particles were washed with an automated microplate washer (HydroFlex™ microplate washer, Tecan).

3.8.1 Pre-characterization of anti-A27 DNA aptamers

Eight DNA aptamer clones were identified and chosen for synthesis after selection against rA27. To pre-characterize these clones and minimize the amount of clones for further characterization, wells of a MaxiSorp ELISA plate were coated overnight with 200 ng/well of rA27 E.coli, rA27 BEI, rD8 E.coli, rL1 E.coli, rL1 BEI, rH3 E.coli, and lysate of VACV-infected cells (VACV RIPA). The following day, wells were aspirated and completely filled with blocking buffer. After washing, biotinylated DNA aptamers were added at 10 nM and detected with streptavidin conjugated to HRP (1:5,000). The color reaction was assayed as described before (3.6.6).

3.8.2 Indirect ELONA with anti-A27 DNA aptamers for affinity testing

Equilibrium dissociation constants (KD) of DNA aptamers to rA27 were determined by indirect ELONA. MaxiSorp ELISA plates were coated overnight with 200 ng/well of rA27. The following day, the supernatant was aspirated, and the wells were completely filled with blocking buffer and incubated for 1 hr at RT. Synthetic DNA aptamers were twofold serially

- 40 -

Materials and Methods diluted, starting at 1 µM, treated as described above (3.3.3), and added to the wells. Binding of DNA aptamers was detected with streptavidin conjugated to HRP and the color reaction was assayed as described before (3.6.6).

3.8.3 Estimation of limit of detection (LOD) of anti-A27 DNA aptamers

For the estimation of the LOD of the anti-A27 DNA aptamers, wells of a MaxiSorp ELISA plate were coated overnight with two-fold serially diluted rA27, starting at 10 µg/mL. The following incubation steps were performed as described under 3.8.1 with 100 nM of each of the aptamer clones 8, 11, 22, and 27. The LOD was calculated with non-linear regression analysis of log-transformed dilution steps of A27 using GraphPad Prism5. The cut off was set to 0.1, because mean background signals were below this value.

3.8.4 Sandwich ELONA with anti-VACV DNA aptamers for affinity testing

KDs of DNA aptamers to VACV particles were determined by sandwich ELONA. MaxiSorp ELISA plates were coated overnight with 200 ng/well of A3/710-20 capture antibody. Wells were aspirated, completely filled with blocking buffer, and incubated for 1 hr at 37°C. Synthetic DNA aptamers were two-fold serially diluted, starting at 1 µM, and treated as described above (3.3.3). Wells were washed 3 times, and then 5×105 pfu/well of VACV particles were applied to blocked, washed wells and incubated for 1 hr at RT. To estimate background binding of DNA aptamers to the capture antibody, the same number of wells filled with VACV particles were filled with SB. After incubation, wells were washed 3 to 5 times, and a volume of 100 µL of each DNA aptamer dilution was added to the wells and incubated for 1 hr at RT. Wells were washed 4 to 8 times, and SA-POD diluted 1:5,000 in blocking buffer was added and incubated for 30 min at RT. Wells were washed again 4 to 8 times, and TMB was added as substrate. The color reaction was detected as described before (3.6.6).

3.8.5 Specificity testing with anti-VACV DNA aptamers

For further specificity analysis of DNA aptamers, MaxiSorp ELISA plates were coated overnight with 5×105 pfu/well of purified VACV 1536, MP-Nü, CPXV BR, CPXV GuWi, CMLV CP-19, or VACV 1536, VACV IHD-W, VACV WR, PPXV, AdV, PPV, or FCV from cell culture supernatant. Wells were washed twice, filled with blocking buffer, and incubated for 1 hr at RT with gentle rocking. Wells were again washed twice, and 100 µL of 100 nM of the tested aptamer clones were added and incubated for 2 hr at RT. Wells were washed 6 times and incubated with SA-POD (1:5,000) for 1 hr at RT. Wells were again washed 6 times, and a TMB substrate solution was added. The color reaction was detected as described before (3.6.6).

3.8.6 Sandwich ELONA with anti-VACV DNA aptamers as capture molecules

To analyze if the synthetic DNA aptamers could be used as capture molecules in a sandwich ELONA, NeutrAvidin®-coated plates (Pierce) were used. Manufacturer’s

- 41 -

Materials and Methods instructions were followed with minor modifications. The DNA aptamers were diluted to 400 nM and treated as described above (3.3.3) before they were applied to NeutrAvidin®- coated plates (Pierce). 5×105 pfu/well of the viruses listed under 3.8.5 were added to the wells and incubated for 1 hr at RT.

3.8.7 Indirect ELISA with anti-A27 or anti-VACV peptides for affinity testing

Affinities of anti-A27 and anti-VACV peptides to A27 or VACV were determined by indirect ELISA. MaxiSorp ELISA plates were coated overnight with 200 ng/well of recombinant proteins (rA27, rL1, rB5) or 5×105 pfu/well of VACV. For further specificity testing of anti VACV peptides, wells were coated with the same amount of porcine parvovirus (PPV) and an adenovirus isolated from bats (AdV-2 PPV1), respectively. The following day, all wells were washed twice with 300 µL/well of washing buffer and were then completely filled with blocking buffer and incubated for 1 to 2 hr at RT. The peptides were twofold serially diluted in blocking buffer, and 100 µL/well of every dilution were added and incubated for 2 hr at RT. Subsequently, wells were washed 6 times, and 100 µL of SA-POD diluted 1:5,000 in blocking buffer was added and incubated for 1 hr. Wells were again washed 6 times, then a TMB substrate solution was added and the color reaction was detected (see 3.6.6).

3.8.8 Specificity testing of synthetic peptides

To further analyze cross reactivity and specificity of synthetic peptides, MaxiSorp ELISA plates were coated overnight with 5×105 pfu/well of purified VACV 1536, MP-Nü, CPXV BR, CPXV GuWi, CMLV CP-19, or VACV 1536, VACV IHD-W, VACV WR, MPXV X12, MPXV München, PPXV, AdV, PPV, or FCV from cell culture supernatant, heat- and UV-inactivated VACV 1536, 200 ng/well of HEp2 cell lysate, and 3 % BSA. The following incubation and washing steps were performed as described above (see 3.8.7). The peptide concentration used was 0.4 µM for the anti-VACV peptide and 3 µM for the anti-A27 peptides.

3.8.9 Sandwich ELISA with synthetic peptides as capture molecules

To analyze if the synthetic peptides could be used as capture molecules in sandwich ELISA, NeutrAvidin®-coated plates (Pierce) were used. Manufacturer’s instructions were followed with minor modifications. At first, to estimate a proper dilution, the peptide was two-fold serially diluted, starting at 50 µg/ml, and applied to the washed wells. 5×105 pfu/well of VACV 1536 were added in duplicate to each dilution and incubated for 1 hr at RT with gentle rocking. The VACV particles were detected by incubation with polyclonal rabbit anti-VACV antibody (1:2,500, 30 min, RT) and subsequent incubation with HRP-labeled goat anti-rabbit antibody (1:2,500, 30 min, RT). Again, TMB was used as substrate. Plates were assayed at 450 nm with 620 nm reference measurement. For specificity testing of synthetic peptides as capture molecules, 100 µL of the 0.4 µM peptide dilution were applied to the NeutrAvidin® plates. 5×105 pfu/well of the viruses

- 42 -

Materials and Methods listed under 3.8.8 were added to the wells in duplicate, and incubated for 1 hr at RT. All subsequent steps were performed as described above.

3.9 Identification of binding partner for peptide VV1

The binding partner for peptide VV1 was unknown. Since the selection was performed against whole VACV particles, it is likely that a poxviral surface protein is the binding partner. Therefore, an ELISA with various recombinant poxviral surface proteins was performed. Additionally, surface plasmon resonance assays with the Biacore® X-100 were conducted. And finally, the peptide was applied in a pull-down assay.

3.9.1 Peptide ELISA with recombinant poxviral proteins

For the identification of the binding partner for VV1, MaxiSorp ELISA plates were coated overnight with 400 ng/well of various poxviral surface proteins (rF9, rA27, rB5, rL1, rD8, rF13 and rA33), and a VACV RIPA lysate. All experimental steps were performed as described in 3.8.7. Peptides were two-fold serially diluted in blocking buffer, starting at 10 µg/mL.

3.9.2 Surface plasmon resonance (SPR) assay

Information of kinetics and affinity of biospecific interactions can be obtained by analysis of adsorption of analytes to immobilized ligands on a -based analytical system [165]. For this, the Biacore system can be used. The main features of a Biacore system are the sensor chip, a surface plasmon resonance (SPR) detector, and a special flow channel. The chip consists of a glass body, a thin gold film, and a surface matrix. The structure of the chip and the principle of the assay are shown in Figure 9.

Figure 9: Structure of sensor chip (left) and principle of surface plasmon resonance (SPR) detection (right) The sensor chip consists of a glass body, a thin gold film, and the surface matrix. The optical SPR- detector is located at the glass side of the chip and detects the change in reflected light angle according to the mass bound to the matrix (modified from [166]).

The analyte is applied in a continuous flow of sample, and if binding to the ligand occurs, a resonance signal resulting from SPR can be detected as a sensogram. Association and - 43 -

Materials and Methods dissociation kinetics of this interaction can be followed in real time (Figure 10). The chip can be regenerated and used for further analysis. Peptide VV1 seemed to bind specifically to VACV particles. To identify the binding partner of VV1, a streptavidin (SA)-coated sensor chip from GE Healthcare (Munich, Germany) was used and run in a Biacore X-100 SPR unit (GE Healthcare). To prepare the SA chip for binding, 1 M NaCl in 50 mM NaOH was injected 3 times. Biotinylated peptides, diluted in running buffer (HBS-EP+, 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.05 % Tween 20), were immobilized at surface densities of 269 RU VV5 (non-binding peptide) to flow cell 1 and 67 RU VV1 to flow cell 2.

Figure 10: Sensogram showing the response from the SPR detector The boxes show the condition on the chip surface, and the diagram shows the display on the screen. The sensogram starts at the baseline while only buffer is flowing over the chip surface. During injection of the analyte, the change in response signal (RU) results from two processes: association to and dissociation from the surface. After sample pulse, the buffer continues to flow over the surface, and the changes in RU result only from dissociation. Finally, the chip is regenerated, resulting in disruption of the binding of ligand and analyte, plus changes in RU back to the baseline (modified from [166]).

To test binding specificity of VV1 50 µg/ml of recombinant proteins expressed in E.coli (D8, B5, F13, A33, and H3) or a Baculovirus system (F9, L1, A33, B5, and A27) were injected over the sensor chip for 120 sec at a flow rate of 5 µl/min. In between, the flow cells were regenerated by injection with 10 mM glycine HCl pH 1.7 (GE Healthcare) for 30 sec at a flowrate of 10 µL/min. For affinity testing of VV1 and its potential binding partner, the protein D8 (MW 30 kDa) was two-fold serially diluted (3.25 nM – 1.67 µM; duplicate injection of 1.67 µM) and applied to both flow cells of the SA chip with a flow rate of 30 µL/min. Association measurement was performed for 120 sec, and dissociation was assessed for 60 sec. For analysis, the double referenced response signals were fitted with BIAevaluation Software 1.0 to either a 1:1

- 44 -

Materials and Methods

Langmuir binding model to determine binding kinetics or a Steady-State-Affinity model to determine binding affinity. All measurements were performed at 25°C. Finally, the binding mechanism of the VV1 D8 interaction was determined. To achieve this, D8 was diluted to 50 µg/mL and injected with a flow rate of 10 µL/min for 30, 60, 120, 240, and 480 sec. Response curves were referenced to the highest response curve (480 sec), and dissociation curves were compared.

3.9.3 Pull-down assay, SDS PAGE and subsequent mass spectrometric analysis

Results of the peptide ELISA against various recombinant proteins led to the potential binding partner of VV1. To confirm this finding, a pull-down assay with the biotinylated peptide and magnetic streptavidin beads (Dynabeads® M-280 Streptavidin, Invitrogen) was performed (principle is shown in Figure 11).

Magnetic SA-bead DNA with biotinylated aptamers

Peptide + Separation Elution and washing of beads VACV lysate

SDS-PAGE In-gel digestion Mass spectrometric analysis

Figure 11: Principle of pull-down for the identification of interaction partners of aptamers The biotinylated peptide or DNA aptamers are immobilized onto magnetic streptavidin beads and are then incubated with the lysate of purified washed viral particles. The magnetic beads are collected with a magnet and unbound lysate components are washed away. The bound lysate components are eluted and separated on an SDS-PAGE. Obtained proteins can then be identified with mass spectrometric analysis.

For the pull-down assay, sugar cushion-purified VACV particles were lysed. The particles were washed with 100 mM ammoniumbicarbonate buffer and spun down (1 hr, 26,000×g, 4°C) to remove any remaining free proteins. The viral particles were resuspended in lysis buffer (IP buffer from Dynabeads + 100 mM NaCl + protease inhibitor) and incubated on ice for 15 min. The samples were again spun down (5 min, 26,000×g, 4°C). The supernatant containing viral proteins was transferred into a fresh microcentrifuge tube and used for the pull-down assay. Before use, the beads were washed three times with 500 µL of PBS. The

- 45 -

Materials and Methods biotinylated peptides VV1 and VV5 (non-binding control peptide) were diluted to 600 nM and coupled to the magnetic beads (30 min, 22°C, 1,400 rpm). The beads were then washed five times with washing buffer (PBS/ 0.01 % [w/v] BSA/0.05 % [v/v] Tween 20). A volume of 100 µL of the VACV lysate was then added and incubated with the beads for 1 hr at RT. The beads were washed 3 times with washing buffer and 3 times with PBS. All washing fractions were collected and reduced with a speedvac to a remaining volume of 50 µL. To elute bound proteins, the beads were incubated with 25 µL of RIPA buffer for 15 min on ice and for an additional 5 min at 95°C. The lysate, the wash fractions 1, 4, 5, and 6, and the eluate were each mixed with 6× Laemmli buffer (Bio-Rad, Germany) and incubated for 5 min at 95°C. A volume of 30 µL of each sample was applied to wells of an 8-16 % premade SDS gel (Pierce, USA). The gel ran using the Mini-Protean® 3 Cell for 10 min at 80 V and for 60 min at 120 V until the sample buffer reached the end of the gel. Afterwards, the gel was incubated for 1 hr with 40 % ethanol/10 % acetic acid solution to fix the proteins in the gel. Subsequently, the gel was washed twice with 30 % ethanol for 20 min and once with H2O. The gel was stained overnight using flamingo fluorescent gel stain (Bio-Rad, Germany) and then analyzed with Image Lab software v4.0.1 at the ChemiDoc (Bio-Rad, Germany) with an excitation wavelength of 512 nm. In the lane of the eluate, one distinct gel band was detected and afterwards cut from the gel for trypsination and mass spectrometric analysis. All experimental steps for mass spectrometry, including the preparation of samples, were performed by Jörg Döllinger, RKI, as described before [167]. The peptides were separated with an Easy-nanoLC (Thermo Fisher Scientific, Rockford, USA) on a 10 cm, ID 75 µm, 3 µm, C18 column using a linear 30 min gradient of Acetonitrile in 0.1 % Formic Acid at 300 nL/min. The liquid chromatograph (LC) was coupled online to an LTQ Orbitrap Discovery™ mass spectrometer (Thermo Fisher Scientific, Rockford, USA) which was operated in a data-dependent manner in the m/Z range of 300 – 1,700 by selecting the Top7 2+ and 3+ charged ions for CID type fragmentation. The proteins were identified using sequest algorithm via Proteome Discoverer 1.3 (Thermo Fisher Scientific, Rockford, USA). Therefore, a custom-made database was generated from all VACV entries of the UniProt Knowledge database; proteins with a false discovery rate smaller than 1 % were regarded as identified.

- 46 -

4. Results

The objective of this thesis was to develop synthetic ligands, which can be used as detection molecules for highly pathogenic agents. The model organism used, was the Vaccinia virus, a member of the family of Poxviridae. All OPV species are cross-reactive and cross-protective. Therefore, it is most likely that synthetic ligands selected against VACV particles can also detect other members of the genus OPV. To select synthetic ligands (DNA and peptide aptamers), SELEX and phage display were the methods of choice. The following sections first summarize the results of DNA aptamer selection and characterization, including the optimization of individual SELEX steps. This is followed by presenting the results of peptide aptamer selection and characterization. Aptamers can be selected against various targets, including proteins or more complex structures such as whole viral particles. To compare the selection processes and the obtained aptamers, SELEX and phage display were performed against a dominant surface protein of the OPV (A27) and against native VACV particles.

4.1 Selection of DNA aptamers against A27

DNA aptamer selection against recombinant A27 (rA27) was performed in solution, as described under 3.3.2 with a 90 mer library. Before this, the optimal elution conditions were assessed as described under 3.3.1. Bead ELISA results revealed that elution for 5 min with His elution buffer containing 500 mM imidazole was most suitable for the His-tagged proteins (data not shown). In total, 9 selection rounds were performed. Samples for quantification of ssDNA were taken before incubation of the ligands with the target and after elution of bound ligands. All samples and standards were diluted with TE buffer, mixed 1:1 with Oligreen, and fluorescence intensities were measured using the Infinite® 200 PRO microplate reader and the NanoDrop ND-3300. The calculated yields of the respective rounds are displayed in Figure 12. From round 1 to round 9, a steady increase of eluted ssDNA could be observed with both measuring methods. Due to the yield of over 100 % in round 9 (measured with the microplate reader), the enriched library of this round was amplified, cloned, and transformed into E.coli. The following day, 29 white colonies and 1 blue colony were randomly picked, and applied to colony PCR. 20 colony PCR-positive clones were amplified in LB medium overnight and were then subjected to plasmid preparation and sequencing. Sequencing revealed 18 individual oligonucleotide sequences that were amplified and biotinylated at the 5’-end and applied to binding assays (data not shown).

- 47 -

Results

120 Nano Drop ELISA 100 80 60 40 20 0 Yieldof eluted aptamers [%] 1 2 3 4 5 6 7 8 9 Selection round

Figure 12: Yield of eluted aptamers after each selection round (A27 selection) Samples for quantification of ssDNA were taken before incubation with the target and after elution. These samples were diluted and mixed 1:1 with OliGreen, and resulting fluorescence intensities were measured. From this, the yields of eluted aptamers were calculated. Yields were calculated from fluorescence intensity measurements obtained with a microplate reader (green bars) and the NanoDrop ND-3300 (blue bars).

Those DNA aptamer sequences with higher ELONA signal for rA27 or VACV RIPA (lysate of VACV-infected cells) in comparison to rD8 as control were chosen for synthesis (Table 15).

Table 15: Synthetic DNA aptamer sequences derived from SELEX against rA27 Clone # DNA sequence a) A27 8 CCAAAgATAgAAgCggTGCggggggggggAgCgTggCggAgAAACAAgTgAgTAg A27 9 CCACACGAGCAACATGGCTACGGAAGAGCGGGTGTATACTAgTACgCAAAgTgg A27 11 CGGAgCgTgAgATCAAATggTACgACgCCggggggTTACAAAgAgggTACgTgg A27 16 CCACACgCgACACAgAAAACAgAgTgAggTCAgggAggggCggCAAAgCTgAgg A27 20 CCACTTgAgTCTggAACgTggTCgCAAggTCCgggAgTggggggTgCCggCTg A27 22 CCACAATAgCgTgCTAggACAgggggCgTAggCgggCgAAgggCgTATgATgAg A27 25 CCACgCAgCACATACTCAgTTgACATATACgCTgTAggggTTCCCAggCTgTCg A27 27 CCAgTgACATgCAAggCAggAggAggTACTAAggAACgACggTTgAgATggggg a) DNA aptamer sequences are displayed from 5’ to 3’ without flanking primer sequences. All aptamer sequences were synthesized with a 5’ biotin from Invitrogen (Darmstadt, Germany).

4.2 Estimation of specificity, affinity, and limit of detection of selected anti-A27 DNA aptamers

The synthesized aptamers selected against rA27 were applied to an indirect ELONA against various recombinant poxviral proteins at a defined concentration, for pre- characterization (Figure 13). All aptamer clones showed only low or no binding to the control proteins, with OD values below 0.1. Incubation with aptamer clones 8, 11, 20, and 27 resulted in highest OD values (0.6 to 0.9) to rA27, and these were subjected to further characterization. Affinities to rA27 were determined (Figure 14). All four aptamer clones

- 48 -

Results showed characteristic binding curves, with KD values ranging from 25 nM (clone 8) to 58 nm (clone 16).

rA27 E.coli rA27 Baculo rD8 E.coli rL1 E.coli rL1 Baculo rH3 E.coli VACV RIPA 1.0

0.8

0.6

0.4

OD OD (450-620nm) 0.2

0.0 8 9 11 16 20 22 25 27 Clone number

Figure 13: Binding assay with synthetic DNA aptamers (A27 selection) To test binding of synthesized DNA aptamers and their target specificity, wells were coated overnight with 200 ng/well of rA27 E.coli, rA27 Baculo, rD8 E.coli, rL1 E.coli, rL1 Baculo, rH3 E.coli, and lysate of VACV-infected cells (VACV RIPA). Synthesized biotinylated DNA aptamers (8, 9, 11, 16, 20, 22, 25, and 27) were added at 10 nM and detected with streptavidin conjugated to HRP. Aptamer sequences 8, 11, 20, and 27 showed specific binding to rA27 BEI, and were used for further characterization. Baculo: Proteins were recombinantly expressed in a Baculovirus system

A B 2.5 2.5

2.0 2.0

8 1.5 1.5 16 Bmax/2 20 1.0 1.0 27

8 16 20 27

OD (450-620nm) OD OD (450-620nm) OD 0.5 0.5 Kd 24.65 58.25 35.34 37.97

0.0 0.0 0 Kd 50 100 150 200 250 0 50 100 150 200 250 Aptamer concentration [nM] Aptamer concentration [nM]

Figure 14: Binding responses and curve fit to calculate the steady-state affinities for DNA aptamer - A27 interactions (A) Measured binding responses were fit using non-linear regression analysis “one site - total binding” equation (GraphPad Prism 5 v5.01). The equilibrium dissociation constant (KD) was calculated as DNA aptamer concentration with ½ Bmax binding response. (B) Wells of a MaxiSorp ELISA plate were coated with 200 ng/well of rA27 and incubated with a two-fold serial dilution series of synthetic biotinylated DNA aptamers derived from aptamer clones numbers 8, 16, 20, and 27. After washing, the wells were incubated with streptavidin conjugated to HRP. The DNA aptamer binding was detected using TMB color reaction. All samples were measured in duplicate. Error bars represent standard deviation. The table shows the calculated KD values in nM.

- 49 -

Results

For detection molecules, not only affinity and specificity are important factors. Another important characteristic is the detection limit. To estimate the detection limit for the 4 aptamer clones, wells of a MaxiSorp ELISA plate were coated with two-fold serially diluted rA27 and incubated with 100 nM of the biotinylated aptamer clones. The lower limit of detection (LOD) was calculated using non-linear regression analysis of log-transformed dilution steps, using software GraphPad Prism 5 (Figure 15). The titration curves for the calculation of the LOD looked very similar for all clones tested. The LODs varied between 78 ng/mL rA27 for clone 8 and 84 ng/mL rA27 for clone 27.

3.0 8 2.5 11 2.0 20 1.5 27

1.0 OD (450-620nm) OD 0.5

0.0 0 1 2 3 4 (log 10) A27 [ng/ml]

Figure 15: Titration curves for calculation of lower limit of detection (LOD) of rA27, determined with anti-A27 aptamers using indirect ELONA Wells of a MaxiSorp ELISA plate were coated with a two-fold serial dilution series of rA27, starting at 10 µg/mL, and incubated with 100 nM of anti-A27 aptamer clones 8, 11, 20 and 27. The cut off was set to 0.1 because the mean background signal was lower (0.02). For all tested anti-A27 aptamers, the LOD was between 78 ng/mL and 84 ng/mL.

Results of sandwich ELONA with rA27 showed that the tested aptamer clones 8, 11, 20, and 27 are also functional as detection molecules in combination with monoclonal anti-A27 antibodies (data not shown). SELEX against other poxviral surface proteins (H3, D8, and L1) was also performed. But the obtained DNA ligands from these selections did not bind to their protein target (data not shown).

- 50 -

Results

4.3 Selection of DNA aptamers against native VACV particles

Selection of DNA aptamers against native VACV particles was performed as described in section 3.3.3 with two different libraries for a total of 12 selection rounds. Before the selection process could be started, individual selection steps had to be optimized carefully.

4.3.1 Optimization of individual SELEX steps

Elution of ssDNA libraries In contrast to the elution during selection against rA27 with magnetic beads performed in plastic tubes, the nucleotides during selection against native VACV particles performed in ELISA plates could not be eluted simply by heat. The objective was to lyse the viral particles to disrupt the binding of the viral particles to the capture antibody. There were several different lysis buffers available (Cytos All, RIPA, AL Qiagen, and RLT Qiagen). To test which lysis buffer would be most efficient, a SELEX round without library was simulated, elution with the different lysis buffers was performed, and poxviral DNA was quantified by quantitative real-time PCR as described in section 3.3.1. Additionally, the samples were heat inactivated, and then part of the samples was purified to reduce the amount of genomic pox- viral DNA which would interfere with the quantification of ssDNA with OliGreen. Furthermore, the lysis buffer which could inhibit library amplification should be eliminated. For this, two different purification kits were used (NucleoSpin® Extract II [Macherey-Nagel], and the ssDNA / RNA Clean & Concentrator™ [Zymo Research]). To evaluate if the reduction of genomic pox-viral DNA could be increased, the column for dsDNA were used twice and two columns for reduction of dsDNA were also used.

A) B) C)

Figure 16: Analysis of lysed VACV samples spiked with ssDNA library after purification with kit from Macherey-Nagel (A, C) and from Zymo Research (B) VACV particles were lysed with RLT buffer, spiked with an ssDNA library, and purified using the NucleoSpin® Extract II Kit (A, C) or the ssDNA / RNA Clean & Concentrator™ Kit (B). Purified samples were analyzed using a High Sensitivity dsDNA chip and the Agilent bioanalyzer. The sample purified with the kit from Macherey-Nagel showed many peaks (C) or bands (A) that were not present using the kit from Zymo research (B). Here, only the standards (green and violet bands) and the ssDNA library could be detected. - 51 -

Results

Quantitative real-time PCR assays revealed that the lysis with the RLT buffer was most efficient and that purification with the ssDNA / RNA Clean & Concentrator™ Kit from Zymo Research led to the best reduction of genomic pox-viral DNA (data not shown). Next, it was assessed if the lysis buffer inhibited the purification of the ssDNA library. To test this, viral particles were mixed with lysis buffer, and ssDNA library was spiked into the samples. The samples were purified with the 2 purification kits, and a conventional PCR with library- specific primers was performed. Analysis of conventional PCR showed specific product bands of the correct size after purification with the ssDNA / RNA Clean & Concentrator™ Kit. Additionally, the samples were analyzed with the Agilent Bioanalyzer to check the purity of the samples (Figure 16). The samples purified using the NucleoSpin® Extract II Kit showed up next to the peaks for the internal standards and the ssDNA library multiple bands (Figure 16 A & C), that were not present using the ssDNA / RNA Clean & Concentrator™ Kit (Figure 16 B).

Optimization of PCR using the Taguchi approach An important step during the SELEX process was the amplification of the enriched library via PCR. Excessive amplification and unfavorable reaction conditions could lead to the accumulation of unwanted by-products and, thus the loss of specific products. Therefore, the amplification step had to be optimized individually for all 3 libraries used. This was done by using the Taguchi approach, comparing three different orthogonal arrays (L 9, L 16, and L 18). It was estimated which array was most suitable for optimization of heterogeneous DNA libraries, and which analyzing method (non-linear regression or analysis via highest SN value) led to the best results. The optimization was conducted as described in PCR optimization. The product yields were calculated from the pixel densities of the obtained PCR products as described before. An example of the calculated product yields for the L9 array is shown in Table 16. From these product yields, the corresponding SN values were calculated (Table

17). An example calculation is shown for level 1 of the component MgCl2 (Equation 3).

Table 16: Mean yields for each reaction component at each level tested (example for L9 array) Factor 1 2 3 [A] MgCl2 0.31…0.91…0.51 1.41…2.23…1.89 3.29…2.03…3.86 [B] Taq 0.31…1.41…3.29 0.91…2.23…2.03 0.51…1.89…3.86 [C] Primer 0.31…1.89…2.03 091…1.41…3.86 0.51…2.23…3.29 [D] dNTPs 0.31…2.23…3.86 0.91…1.89…3.29 0.51…1.41…2.03

Equation 3: Example for calculation of the SN value for the lowest MgCl2 level

- 52 -

Results

Table 17: Signal-to-noise (SN) ratios for reaction components calculated for the reaction components of the L9 array using the mean of the yields 1 2 3 [A] MgCl -7.22 4.85 8.72 [B] Taq -5.76 2.60 -1.53 [C] Primer -5.73 2.28 -1.47 [D] dNTPs -5,63 2,79 -1,92

The calculated optimal PCR conditions had to be verified by a confirmation PCR, and were compared to the original PCR conditions. The results of the optimization of PCR 1 with the different arrays for the 90 mer library are shown in Figure 17. Not only product intensity but also product specificity was important when amplifying heterogeneous DNA libraries. Results after optimization with the L 9 array showed increased signal intensities (Figure 17 A lanes 2-9 compared to lanes 10-13), while results after optimization with the L 16 or L 18 array showed not only increased signal intensities, but also unspecific by-products (Figure 17 B, lanes 2-7 compared to lanes 8-12). The reaction conditions leading to the PCR product shown in lane 5 of Figure 17 A (indicated by red ellipse) were the optimal conditions for the amplification of the 90 mer, because these conditions generated the most specific product. This product was generated with the L 9 array using analysis with the highest SN values.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 A

B

Figure 17: Gel analysis of PCR products of the confirmation PCR for the 90 mer library A. 1: O’GeneRuler™ DNA ladder, 2-9: products of optima for PCR 1 of the 90 mer obtained from array L 9. 10-13 products of original PCR protocol for 90 mer library at different template concentrations. B. 1: O’GeneRuler™ DNA ladder, 2-3: optima for PCR 1 of the 90 mer obtained from array L 18, 4-7: optima for PCR 1 of the 90 mer obtained from array L 16, 8-12: products of original PCR protocol for 90 mer library at different template concentrations, 13-14: no template control. Several optima for the L 9 array were obtained, because the array was performed for 10 fmol and 20 fmol DNA with 12 × and 16 × cycles. Furthermore, the data were analyzed with and without non-linear regression analysis. Multiple optima for the L 16 array were due to the analysis methods chosen. For the L 18 array, the analysis methods led to the same results, but two maxima could be observed. The optimal reaction conditions led to the PCR product highlighted with the red ellipse.

Results of the optimization of PCR 1 with the different arrays of the 67 mer and 77 mer libraries are shown in Figure 18. An increase of signal intensity compared to the original PCR protocol could be achieved with all 3 arrays used for both libraries. For the 67 mer library, the most intense and specific products were generated with the L 9 array using non-

- 53 -

Results linear regression analysis (Figure 18 A, lane 3). For the 77 mer library, the most intense and specific products were generated with the L 9 array using the analysis with highest SN values (Figure 18 B, lane 2). The optimal amplification conditions for PCR 1 for all three libraries are listed in Table 18.

1 2 3 4 5 6 7 8 9 10 11 12 13 A

B

Figure 18: Gel analysis of PCR products of the confirmation PCR 1 for the 67 mer (A) and 77 mer (B) libraries 1 & 13: O’GeneRuler™ DNA ladder, 2-3: products of optima for PCR 1 obtained from L 9 array, 4-7: products of optima for PCR 1 obtained from L 16 array, 8-9: products of optima for PCR 1 obtained from L 18 array, 10-12: products of original PCR protocol at different template concentrations at 20 × cycles (see text for more details). The optimal reaction conditions led to the PCR product highlighted with the red ellipse.

For all 3 libraries, a significant increase in signal intensity, but varying specificity could be achieved with all arrays used. The optimization process led to drastic increases of MgCl2 and primer concentrations. All arrays confirmed that excessive cycling and too much initial template results in generation of unwanted by-products. It could be shown that the choice of parameter concentration levels could influence significantly the outcome of the optimization process (low MgCl2 in L18 array resulted in no PCR product). The most specific PCR products were generated with the L 9 array. Therefore, this array was used for the optimization of PCR 2.

Table 18: Optimal and original reaction conditions for PCR 1 for all three libraries as analyzed by Taguchi optimization Factor 67 mer 77 mer 90 mer original

MgCl2 [mM] 5.98 8 8 2 dNTPs [µM] 142 200 350 200 Primers [nM] 760 1200 1200 280 Taq [U] 0.7 1 1.5 0.5 Cycles 20 20 12 12-20

PCR 2 was optimized comparing 12 × cycles and 16 × cycles (90 mer) and 16 × cycles and 20 × cycles (67 mer and 77 mer) with 2 different DNA concentrations (10 fmol and 20 fmol for 67 mer and 77 mer, and 3 fmol and 7.6 fmol for 90 mer). Again, the analysis methods of non-linear regression and highest SN value were applied. Optimization of PCR 2 led to a significant increase in product yield and product specificity for all 3 libraries with all compared factors (Figure 19 for 67 mer and 77 mer, Figure 20 for 90 mer). The optimal

- 54 -

Results amplification conditions for PCR 2 for all three libraries are listed in Table 19, and these were obtained by analysis with highest SN values.

1 2 3 4 5 6 7 8 9 10 A

B

Figure 19: Gel analysis of PCR products of the confirmation PCR 2 for the 67 mer (A) and 77 mer (B) libraries 1 & 10: O’GeneRuler™ DNA ladder, 2-5: products of optima for PCR 2 obtained from L 9 array, 6-7: PCR products of optima for PCR 2 obtained from L 18 array, 8-9: products of original PCR protocol at 10 fmol (8) and 20 fmol (9) dsDNA. A significant increase of signal intensity after optimization could be observed for both libraries. The most intense and specific bands were obtained with L 9 array, using 10 fmol DNA (20 × cycles for the 67 mer [A], 16 × cycles for the 77 mer [B]), and analysis after highest SN value (lane 2, highlighted by red ellipse).

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Figure 20: Gel analysis of PCR products of the confirmation PCR 2 for the 90 mer library 1 & 14: O’GeneRuler™ DNA ladder, 2: sample 6 from L 9 array, 3: sample 7 from L 9 array, 4-7: products of optima for PCR 2 obtained from L 9 array, analysis with non-linear regression, 8-11: products of optima for PCR 2 obtained from L 9 array, analysis after highest SN value, 12-13: products of original PCR protocol after 12 (lane 12) and 16 (lane 13) cycles with 7.6 fmol DNA. A significant increase of signal intensity and specificity could be observed after optimization with the L 9 array (compare lanes 12 and 13 with lanes 4-11). The most intense and specific band could be detected in lane 9.

Table 19: Optimal and original reaction conditions for PCR 2 for all three libraries as analyzed by Taguchi optimization Factor 67 mer 77 mer 90 mer original

MgCl2 [mM] 8 4 8 2 dNTPs [µM] 350 350 200 200 Primers [nM] 1200 1200 600 280 Taq [U] 1 1 1 0.5 Cycles 20 16 12 12-20

- 55 -

Results

Quantification of ssDNA and binding assays To calculate the yield for each selection round, a reliable quantification of ssDNA is important. The amount of eluted ssDNA after each round can be very low; therefore, a sensitive assay for quantification is needed. The highly sensitive fluorescent dye OliGreen® that is able to detect as little as 200 pg of ssDNA/sample was used for quantification of ssDNA. Before and after each selection round, samples for quantification of ssDNA were taken and assayed as described in section 3.3.4. Before the samples from the selection rounds were assayed, a standard dilution series for all 3 libraries was measured with a microplate reader, the NanoDrop 3300, and the Qubit 2.0. Results revealed that the dynamic range for the 90 mer library was between 10 ng/mL and 1000 ng/mL, and for the shorter libraries, between 250 ng/mL and 1000 ng/mL (independent of the device used). A reliable quantification, especially for the low ssDNA amounts, was impossible for the 67 mer and 77 mer libraries. Furthermore, the quantification of the 90 mer library did not allow a clear decision after which selection round the selection process should be ended because of greatly fluctuating yields (data not shown). To still decide when to end the selection process, two binding assays for the 90 mer and the 77 mer libraries were performed, one after 6 and one after 12 selection rounds. The results for the binding assays after 12 selection rounds are displayed in Figure 21. For the 77 mer library, an increase of binding from round 1 to round 5 could be observed, and thereafter a steady decrease in binding. The results were similar for the 90 mer library, with a negative peak in round four, and another increase starting in round nine.

VACV 90mer mAb 90mer VACV 77mer mAb 77mer 2.0

1.5

1.0

OD OD (450-620nm) 0.5

0.0

Bib Rd1 Rd2 Rd3 Rd4 Rd6 Rd7 Rd8 Rd9 Rd5 Rd10 Rd11 Rd12

Figure 21: Binding assay of 77 mer and 90 mer libraries after 12 selection rounds Binding assays of the enriched libraries were performed after 12 selection rounds. For this, 200 ng of the anti-A27 capture antibody mAb A3/710-20 were coated into wells of a MaxiSorp ELISA plate overnight. The next day, 5×106 pfu/mL of VACV were added. The virus particles were incubated with 10 nM of the unselected libraries (Bib), and the enriched amplified biotinylated libraries from rounds 1-12 (purple line for the 90 mer library, and green line for the 77 mer library). The libraries were detected with SA-POD. To check background binding to the capture Ab, the enriched libraries were incubated with the mAb without VACV particles (orange and red lines for 90 mer and 77 mer library, respectively). Data points show mean values of duplicates ±SD.

- 56 -

Results

Both libraries show constantly only low background binding to the capture antibody (Figure 21). The data again, did not show clear results to decide after which selection round to end the selection process. This is why the enriched libraries of multiple selection rounds were cloned.

4.3.2 Cloning and sequencing of enriched DNA libraries For the cloning of enriched libraries, rounds 2, 4, 6, and 8 were chosen. Later on, also enriched libraries of round 5 for both libraries were cloned, and those of rounds 9 and 12 for the 90 mer library. With this, it was assessed if sequences that were found in rounds 6 and 8 could also be found in rounds 5 and 9. Round 12 of the 90 mer library was cloned to estimate if the sequences of this selection round could give a clue for the increase of the OD signals observed in the binding assays. From each cloned round, 10-15 clones were picked and sequenced and obtained sequences were compared. For the 90 mer library, 93 of the 105 sequenced clones were evaluable and compared. The remaining 12 sequences were not evaluable because either the plasmid did not contain any insert sequence, more than one clone was present in the sample, or because the sequencing reaction did work properly. For the enriched 90 mer library, 77 individual sequences were identified. Starting in round 9, sequences that were longer than expected (up to 168 nucleotides) were found. These sequences contained additional copies of either one or both primers and/or some nucleotides of unknown origin. The most dominant sequence in the 90 mer library was clone 90.6.1. This clone was found in rounds 6, 8, 9, and 12 - altogether 13 times. Additionally, clones that differed only by one nucleotide from this sequence (90.6.7, 90.6.8, and 90.6.9) were found. Clone 90.8.10 was first found in round 8. More copies of this clone were found in round 9. But again, this sequence contained multiple copies of the primers. For the 77 mer library, 47 of the 75 sequenced clones were evaluable and compared. Hereby, 45 individual sequences were identified. Here as well, sequences that were longer than expected were found, but to a lesser amount compared to the 90 mer library. Furthermore, sequences that were 10 nucleotides shorter than expected were also found. This indicated a contamination with the 67 mer library. The remaining sequences were not evaluable because of reasons described above. For both libraries, sequences that were found multiple times and those that were similar to these sequences (Table 20) were chosen for synthesis to be characterized further.

- 57 -

Results

Table 20: Synthetic DNA aptamer sequences derived from SELEX against VACV particles (cloning) Clonea) DNA sequenceb) Frequencyc) 90.6.1 TTggACCgCggTgggTAgTCAggTATACTCCAAAATgCTTTATTTAgCACAAgg 13 90.6.7 TTggACCgCggTgggTAgTCgggTATACTCCAAAATgCTTTATTTAgCACAAgg 1 90.6.8 TTggACCgCggTgggTAgTCAggTATACTCCAAAATgCTTTATTTAgCACAAAg 2 90.6.9 TTggACCGCggTgggTAgCCAggTATACTCCAAAATgCTTTATTTAgCACAAgg 1 90.8.10 TCCgTATAATAgTgCTgTACTAAgCAAATTTATAgTTCTCTAgAAAgTgCCCgC 3 77.4.8 CTAgCggTCAAATCATTgCACACTTCCggATATgCTCggg 2 77.8.8 TAgACggTATAggTTTTATggAAgCTggTTAAgCAgAgAg 2 a) The clone name is composed of the length of originating library (90 mer or 77 mer), the round it first appeared in, and the actual number of the clone in this round. b) DNA aptamer sequences are displayed from 5’ to 3’ without corresponding primer sequences. All DNA aptamer sequences were synthesized with a 5’ biotin by Metabion (Martinsried, Germany). Differences to clone 90.6.1 are highlighted in red. c) Frequency of how often this sequence was found in all rounds together.

4.3.3 High-throughput sequencing of enriched DNA libraries Sequencing of cloned enriched libraries was time consuming and gave only a limited view of the distribution of sequences present in the enriched libraries. As described before, analysis of data from high-throughput sequencing (HTS) of enriched DNA libraries could lead to the identification of ligands which bind with higher affinity and specificity to their target as compared to sequences obtained from cloning. In this context, an enrichment of sequences from round to round can be more important in terms of affinity and specificity than the frequency of this sequence. The rounds 1 to 8 of the 90 mer and the 77 mer library were sequenced with the Illumina Genome Analyzer II, as described in section 3.4. The total counts of sequences obtained from sequencing with the Illumina Genome Analyzer II, and the percentages of sequences that could be assigned to a certain round, are shown in Table 21. Overall, more than 1.2 billion sequences were obtained from HTS. Around 600 million sequences could be assigned to the eight selection rounds by correct index filtering, 482 million sequences of which were assigned to the 90 mer library, and 130 million sequences of which were assigned to the 77 mer library. Here, only 26 % of the 90 mer sequences, and only 18 % of the 77 mer, could be assigned explicitly to a certain selection round. In general, 50 % of all obtained sequences are those of the PhiX control. Furthermore, a multiplex approach to sequence the clones of 4 enriched libraries in parallel was used. The index sequences for exact identification are only 6 nucleotides in length. If only 1 nucleotide cannot be identified correctly, the corresponding sequence cannot be assigned to a specific selection round.

- 58 -

Results

Table 21: Number of reads obtained from sequencing with the Illumina Genome Analyzer II and distribution of obtained sequences among 90 mer and 77 mer libraries

90 mer1) 77 mer2) Lane Reads Round Readsa) % assignedb) Readsa) % assignedb) 1 145,849,910 1 52,089,964 14.12 6,748,079 2.37 2 138,450,427 2 44,186,351 11.98 3,430,391 1.21 3 145,622,714 3 42,227,032 11.45 11,593,442 4.08 4 147,189,906 4 252,276,222 /c) 12,708,707 4.47 5 185,073,884 5 29,724,357 10.06 27,796,256 9.49 6 183,861,791 6 29,750,931 10.07 21,814,992 7.45 7 148,252,213 7 21,228,968 7.18 22,362,846 7.64 8 147,250,964 8 10,701,221 3.62 24,244,811 8.28 Sum 1,241,551,809 482,185,046 130,699,524 1) Reads of the 90 mer library that were found in lanes 5 & 6 (rounds 1-3) and 7 & 8 (rounds 5-8) of the flow cell 2) Reads of the 77 mer library that were found in lanes 1&2 (rounds 1-4) and 3&4 (rounds 5-8) of the flow cell a) Reads that could be assigned to the corresponding rounds by correct index filtering b) % of sequences that could be assigned to corresponding rounds by correct index and primer filtering c) Round 4 of the 90 mer library was sequenced first to estimate optimal sequencing conditions, therefore no index filtering was performed

Table 22 and Table 23 give an overview of the HTS data analysis, containing total sequence counts of the 90 mer and 77 mer libraries that could be assigned to the corresponding round, the number of unique sequences, the average of how often each sequence was found in the respective selection round, and the percentage of sequences that were found at least twice in this selection round. For the 90 mer library, an increase of duplicates from 0.51 % in round 1 to 58.64 % in round 8 was observed, indicating that ligands specifically binding to the target had been enriched and amplified (Table 22).

Table 22: Overview of sequence counts of 90 mer library after HTS data analysis Round Totala) Uniqueb) Averagec) % of duplicatesd) 1 17,863,278 17,772,884 1.01 0.51 2 15,060,634 12,602,963 1.2 16.32 3 14,347,950 11,823,027 1.21 17.60 4 45,971,170 29,573,471 1.55 35.67 5 12,316,503 9,906,904 1.24 19.56 6 11,852,052 5,140,704 2.31 56.63 7 5,966,801 2,430,728 2.45 59.26 8 2,279,435 942,849 2.42 58.64 a) Total number of sequences for each round obtained after correct index and primer filtering b) Number of sequences that were found only once in this round c) Average of how often each sequence was found in this round d) % count of sequences that were found at least twice in this round

- 59 -

Results

The data from the analysis of the 77 mer library showed a different picture. Here, in all these selection rounds, around ¾ of all sequences were duplicates, and only a little increase of duplicates from round 1 to round 4 could be observed. From round 5 onwards, the number of duplicates decreased (Table 23).

Table 23: Overview of sequence counts of 77 mer library after HTS data analysis

Round Totala) Uniqueb) Averagec) % of duplicatesd) 1 614,380 148,240 4.14 75.87 2 537,842 144,672 3.72 73.10 3 2,509,455 690,990 3.63 72.46 4 3,012,659 585,247 5.15 80.57 5 5,560,898 1,271,729 4.37 77.13 6 4,331,177 1,050,973 4.12 75.73 7 3,960,638 1,143,833 3.46 71.12 8 3,900,594 1,195,953 3.26 69.34 a) Total number of sequences for each round obtained after correct index and primer filtering b) Number of sequences that were found only once in this round c) Average of how often each sequence was found in this round d) % count of sequences that were found at least twice in this round

The total counts and the percentage of the top ten sequences of the 90 mer and the 77 mer are listed in Table 24 and Table 25. The percentages of the top ten sequences give an overview of the sequences that were enriched during the eight selection rounds. When looking at the top ten sequences of the 90 mer library (Table 24), it is conspicuous that the top ten sequences from round 1 could only be found in this round, and that all the other top ten sequences appeared in round 2 or even later. The sequences that were identified with cloning and sequencing (90.6.1, 90.8.10, and 90.6.8) were first found in round 2 with HTS. All sequences that are similar to these sequences appeared in round 3 or round 4. It is possible that these sequences evolved during PCR amplification.

- 60 -

Results

Table 24: Counts and percentage of top 10 sequences obtained after analysis of HTS data of 90 mer library Counts % Namea) R1 R2 R3 R4 R5 R6 R7 R8 R1 R2 R3 R4 R5 R6 R7 R8 1 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 2 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 3 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 4 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 5 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 6 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 7 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 8 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 9 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 10 3 0 0 0 0 0 0 0 2E-5 0 0 0 0 0 0 0 DS 2.1 0 110 125 307 70 121 42 13 0 7E-4 9E-4 7E-4 6E-4 1E-3 7E-4 6E-4 12 0 79 104 253 110 111 26 8 0 5E-4 7E-4 6E-4 9E-4 9E-4 4E-4 4E-4 DS 2.3 0 49 68 140 33 85 27 4 0 3E-4 5E-4 3E-4 3E-4 7E-4 5E-4 2E-4 14 0 40 37 75 35 19 13 8 0 3E-4 3E-4 2E-4 3E-4 2E-4 2E-4 4E-4 DS 2.5 0 34 43 263 118 126 86 33 0 2E-4 3E-4 6E-4 1E-3 11E-3 14E-3 0.0014 90.6.8 0 20 1.450 4.820 41.834 439.197 110.813 20.324 0 1E-4 0.01 1E-2 0.34 3.71 1.86 0.89 17 0 19 23 81 31 14 5 0 0 1E-4 2E-4 2E-4 3E-4 1E-4 1E-4 0 18 0 18 27 77 16 23 5 2 0 1E-4 2E-4 2E-4 1E-4 2E-4 1E-4 1E-4 19 0 18 24 64 12 29 14 1 0 1E-4 2E-4 1E-4 1E-4 2E-4 2E-4 0.0000 20 0 13 19 69 51 57 58 24 0 1E-4 1E-4 2E-4 4E-4 5E-4 1E-3 11E-3 90.8.10 0 5 475 1.788 51.830 1.198.439 830.712 336.434 0 3E-5 33E-3 4E-3 0.42 10.11 13.92 14.76 90.6.1 0 2 174 1.252 38.557 1.552.270 866.748 356.482 0 1E-5 1E-3 3E-3 0.31 13.10 14.53 15.64 DS 4.4 0 1 66 335 3.565 49.053 14.359 3.067 0 7E-6 5E-4 7E-4 0.03 0.41 0.24 0.13 DS 4.8 0 0 34 147 4.098 101.844 74.394 32.909 0 0 2E-4 3E-4 0.03 0.86 1.25 1.44 DS 8.6 0 0 28 113 2.985 67.089 48.969 20.464 0 0 2E-4 2E-4 0.02 0.57 0.82 0.90 DS 5.7 0 0 24 81 2.002 47.258 31.943 13.234 0 0 2E-4 2E-4 0.02 0.40 0.54 0.58 DS 5.8 0 0 3 23 1.423 62.646 90.338 69.480 0 0 2E-5 1E-4 0.01 0.53 1.51 3.05 DS 5.9 0 0 12 61 1.357 78.456 55.557 35.730 0 0 1E-4 1E-4 0.01 0.66 0.93 1.57

- 61 -

Results

DS 5.10 0 0 29 109 991 16.186 4.948 1.315 0 0 2E-4 2E-4 0.01 0.14 0.08 0.06 DS 6.10 0 0 2 12 306 21.479 15.928 10.041 0 0 1E-4 3E-5 2E-3 0.18 0.27 0.44 DS 7.9 0 0 0 9 324 15.667 19.863 15.973 0 0 0 2E-5 3E-3 0.13 0.33 0.70 DS 8.8 0 0 0 2 58 5.656 10.914 17.605 0 0 0 4E-6 5E-4 0.05 0.18 0.77 DS 8.10 0 0 1 5 236 9.480 15.167 13.776 0 0 7E-6 1E-5 2E-3 0.08 0.25 0.60 a) All clones were numbered consecutively. Clones that were selected for synthesis were renamed according to their occurrence in the top ten list or after the name obtained from cloning. Sequences that were similar to sequence 90.6.1 were highlighted in dark red and sequences that were similar to sequences 90.8.10 were highlighted in blue.

Table 25: Counts and percentage of top 10 sequences obtained after analysis of HTS data of 77 mer library Counts % Namea) R1 R2 R3 R4 R5 R6 R7 R8 R1 R2 R3 R4 R5 R6 R7 R8 77.8.8 4185 2394 15402 18717 56902 45244 39109 36957 0.681 0.445 0.614 0.621 1.023 1.045 0.987 0.947 JM 1.2 243 225 746 715 354 141 64 51 0.04 0.042 0.03 0.024 0.006 0.003 0.002 0.001 JM 1.3 96 137 393 604 1825 1445 1461 1327 0.016 0.025 0.016 0.02 0.033 0.033 0.037 0.034 4 92 116 506 787 1982 1859 1793 1748 0.015 0.022 0.02 0.026 0.036 0.043 0.045 0.045 5 88 60 573 848 1918 1381 1210 1101 0.014 0.011 0.023 0.028 0.034 0.032 0.031 0.028 6 84 23 251 302 621 416 274 244 0.014 0.004 0.01 0.01 0.011 0.01 0.007 0.006 7 84 69 423 887 2471 2471 2622 2640 0.014 0.013 0.017 0.029 0.044 0.057 0.066 0.068 8 80 139 210 183 284 172 69 62 0.013 0.026 0.008 0.006 0.005 0.004 0.002 0.002 9 77 72 539 1066 3527 3539 3645 3345 0.013 0.013 0.021 0.035 0.063 0.082 0.092 0.086 10 77 36 101 40 26 6 6 2 0.013 0.007 0.004 0.001 0 0 0 0 11 19 277 120 150 151 64 23 18 0.003 0.052 0.005 0.005 0.003 0.001 0.001 0 12 34 252 140 151 159 65 33 30 0.006 0.047 0.006 0.005 0.003 0.002 0.001 0.001 13 46 249 159 180 196 105 60 37 0.007 0.046 0.006 0.006 0.004 0.002 0.002 0.001 14 28 239 204 224 209 71 45 23 0.005 0.044 0.008 0.007 0.004 0.002 0.001 0.001 15 19 236 58 11 10 1 1 0 0.003 0.044 0.002 0 0 0 0 0 16 12 226 70 4 4 0 0 0 0.002 0.042 0.003 0 0 0 0 0 17 8 225 45 30 19 0 1 0 0.001 0.042 0.002 0.001 0 0 0 0 JM 2.10 1 221 29 25 10 12 5 3 0.0002 0.041 0.0012 0.001 0.0002 0.0003 0.0001 0.0001 19 70 128 509 592 1422 1070 889 808 0.011 0.024 0.02 0.02 0.026 0.025 0.022 0.021

- 62 -

Results

JM 3.7 52 57 488 629 3027 3084 3282 3635 0.008 0.011 0.019 0.021 0.054 0.071 0.083 0.093 JM 3.8 63 29 433 878 3964 4651 5625 6446 0.01 0.005 0.017 0.029 0.071 0.107 0.142 0.165 JM 3.9 44 51 433 646 2538 2502 2771 3007 0.007 0.009 0.017 0.021 0.046 0.058 0.07 0.077 JM 4.8 53 27 323 694 4143 5818 8845 11441 0.009 0.005 0.013 0.023 0.075 0.134 0.223 0.293 24 52 43 383 637 1544 1186 1027 1032 0.008 0.008 0.015 0.021 0.028 0.027 0.026 0.026 JM 5.6 66 86 361 594 3020 4405 6235 7745 0.011 0.016 0.014 0.02 0.054 0.102 0.157 0.199 JM 5.7 37 10 243 479 2895 4355 7738 9309 0.006 0.002 0.01 0.016 0.052 0.101 0.195 0.239 27 50 44 252 546 2796 4011 6246 6925 0.008 0.008 0.01 0.018 0.05 0.093 0.158 0.178 JM 5.9 54 21 211 517 2788 4458 7083 9192 0.009 0.004 0.008 0.017 0.05 0.103 0.179 0.236 29 27 83 191 442 1820 3658 6060 7864 0.004 0.015 0.008 0.015 0.033 0.084 0.153 0.202 30 40 44 129 288 1887 3602 6353 8502 0.007 0.008 0.005 0.01 0.034 0.083 0.16 0.218 31 44 22 237 385 2429 3314 5147 6084 0.007 0.004 0.009 0.013 0.044 0.077 0.13 0.156 32 11 7 75 165 1332 2448 4524 6610 0.002 0.001 0.003 0.005 0.024 0.057 0.114 0.169 a) All clones were numbered consecutively. Clones that were selected for synthesis were renamed according to their occurrence in the top ten list or after the name obtained from cloning. Sequences similar to 77.8.8 were highlighted in orange.

- 63 -

Results

Looking at the sequences obtained from the 77 mer library, one sequence stands out (77.8.8). This sequence could be found in every selection round with the highest quantity, but no continuous enrichment of this sequence could be observed. With HTS, only one sequence (JM 1.3) was found that was similar to the clone 77.8.8. This sequence did not show any significant enrichment either. The other clone obtained from cloning (77.4.8) was found only 3 times with HTS, and no sequences were found that were similar to this sequence. Nevertheless, individual sequences were enriched from selection round to selection round, and were chosen for synthesis.

Selection of DNA sequences for synthesis Over 600 million sequences were obtained from HTS. These sequences were to be tested for affinity and specificity to VACV particles. However, due to their high number, not all of these sequences could be tested. Therefore, in order to minimize the sequences that should be synthesized for each library, criteria were defined to sort out sequences. For both DNA libraries, the top ten sequences of each selection round were listed and compared, leaving 33 sequences for the 90 mer library (Table 24) and 32 sequences for the 77 mer library (Table 25). The percentage for each top ten sequence in each round, and the enrichment from selection round to selection round based on the percentage, were calculated. In Figure 22 and Figure 23 the percentage distribution of selected top ten sequences among the eight selection rounds is shown. For the 90 mer library, two sequences (90.6.1 and 90.8.10) (Figure 22) were dominant. These two sequences were also found to be the dominant sequences during cloning of enriched libraries.

18 90.6.8

16

14 90.8.10 12 90.6.1

10 DS 4.4 8 DS 4.8

6 DS 8.6 of clones clones of 4 DS 5.8 2

0 percentage of occurrence ofoccurrence percentage 1 2 3 4 5 6 7 8 Round Figure 22: Percentage of selected top ten sequences from the 90 mer library among the eight selection rounds HTS data analysis revealed 2 dominant sequences (90.6.1, blue rectangle, and 90.8.10, dark red triangle) that were enriched during the selection process and represented each up to 15 % of all sequences in round 8. These 2 sequences were also found during cloning experiments. Sequence 90.6.8 (dark blue plus sign), that was also found during cloning experiments, showed enrichment from round 1 to round 6 from 0.0001 % to 3.71 % and a decrease thereafter. This sequence differs only in 1 nucleotide from that of clone 90.6.1.

- 64 -

Results

Selection of sequences for synthesis was dependent on the library. For the 90 mer, 14 sequences were synthesized which fulfilled one of the following criteria, (Table 26):  Sequence has similarity to sequence obtained from cloning, to compare binding characteristics (11),  Sequence was enriched during the 8 selection rounds (1)  Sequence was decreased during the 8 selection rounds (2)

Table 26: Synthetic DNA aptamer sequences derived from SELEX against VACV particles with 90 mer library (HTS) Clone DNA sequencea) Remark DS 2.1 CCCAgCgATCAAATTCTTCCAgCATAAATACggTCTAgCTAATCCCTAgTggCG decreased DS 2.3 CCACTTgAgTCTggAACgTggTCgCAAggTCCgggAgTggggggTgCCggCTg decreased DS 2.5 CCAgTgACATgCAAggCAggAggAggTACTAAggAACgACggTTgAgATggggg enriched DS 4.4 TTggACCgCggTgggTAgACAggTATACTCCAAAATgCTTTATTTAgCACAAAg similar to 90.6.1 DS 4.8 TCCgTATAATAgTgCTgTACTAAgCAAATTTgTAgTTCTCTAgAAAgTgCCCgC similar to 90.8.10 DS 5.7 TCCgTATAATAgTgCTgTACTAAgCAAATTTATAgTACTCTAgAAAgTgCCCgC similar to 90.8.10 DS 5.8 TCCgTATAATAgTgCTgTACTAAgCAAgTTTATAgTTCTCTAgAAAgTgCCCgC similar to 90.8.10 DS 5.9 TTggACCgCggTgggTAgTCAggTATACTCCAAAATgCTTTATTTAgCACAAg similar to 90.6.1 DS 5.10 TTggACCgCggTggTAgTCAggTATACTCCAAAATgCTTTATTTAgCACAAAg similar to 90.6.1 DS 6.10 TTggACCgCggTggTAgTCAggTATACTCCAAAATgCTTTATTTAgCACAAgg similar to 90.6.1 DS 7.9 ACCgTATAATAgTgCTgTACTAAgCAAATTTATAgTTCTCTAgAAAgTgCCCgC similar to 90.8.10 DS 8.6 TCCgTATAATAgTgCTgTACTAAgCAAATTTATAgTCCTCTAgAAAgTgCCCgC similar to 90.8.10 DS 8.8 TTggTCCgCggTgggTAgTCAggTATACTCCAAAATgCTTTATTTAgCACAAgg similar to 90.6.1 DS 8.10 TCCgTATAATAgTgCTgTACTAAgCAACTTTATAgTTCTCTAgAAAgTgCCCgC similar to 90.8.10 a) DNA aptamer sequences are displayed from 5’ to 3’ without corresponding primer sequences. All DNA aptamer sequences were synthesized with a 5’ biotin by Metabion (Martinsried, Germany).

1,2 77.8.8 1,0 JM 1.2 0,8 JM 1.3 JM 2.10 0,6 JM 3.8

0,4 JM 4.8 of clones of 0,2 JM 5.6 JM 5.7

percentage of occurrence ofoccurrence percentage 0,0 R1 R2 R3 R4 R5 R6 R7 R8

Round Figure 23: Percentage of selected top ten sequences from the 77 mer library among the eight selection rounds HTS data analysis showed 1 dominant sequence (77.8.8, dark blue diamond) which was also found with cloning. This sequence was found in every selection round, but did not show any significant enrichment over the 8 selection rounds. Starting in round 4, several clones occurred more often.

- 65 -

Results

For the 77 mer, 10 sequences were synthesized which fulfilled one or more of the following criteria, (Table 27):  Sequence was present in each selection round (all),  Sequence was listed among the top ten sequences in at least one selection round (all),  Similar to sequence obtained from cloning (1),  Sequence was enriched during the 8 selection rounds (6)  Sequence was inconsistent during the 8 selection rounds (1),  Sequence was shorter than expected (2). All synthesized DNA sequences were analyzed by ELONA for binding to check if the occurrence of the sequence could be correlated with affinity and/or specificity.

Table 27: Synthetic DNA aptamer sequences derived from SELEX against VACV particles with 77 mer library (HTS) Clone DNA sequencea) Remark JM 1.2 TATCCTTCAggTgCTTAAACTTgCTTgTCT shorter & inconsistent JM 1.3 TAgACggTATAggCTTTATggAAgCTggTTAAgCAgAgAg similar to 77.8.8 JM 2.10 AgCTgAgTCTATAAgTCTATCggTTggTCA shorter JM 3.7 CCCCCTATggACCAACCTACCACCAAACCCCAATACCTCT top ten up & enriched JM 3.8 CgACCgCCATCCCCCTTAgAACACACgTAACAATgCCCAC top ten up & enriched JM 3.9 ACCACCCAggTATTCCAgCCTTgTCCACggTCCTCCATgC enriched JM 4.8 ACCACACACgCCAgATCCCCAgACCgCgTACTCCCAAACCACg enriched JM 5.6 ACgAgCAAACCCACAgCTgCAATCggTTAggAACCgCgAT enriched JM 5.7 CgAgACAgAAATAggATggCggACCAgggTAggTAAAgCC top ten up JM 5.9 CCATCgCATACgAgCCCggCACAATACgATCCAgCgCAAg enriched a) DNA aptamer sequences are displayed from 5’ to 3’ without corresponding primer sequences. All DNA aptamer sequences were synthesized with a 5’ biotin by Metabion (Martinsried, Germany).

4.4 Characterization of anti-VACV DNA aptamers

Selection against native VACV particles was performed in sandwich ELONA format with a monoclonal anti-A27 capture antibody. To evaluate binding kinetics, all synthesized DNA aptamers obtained from cloning and HTS were applied as detection molecules in sandwich ELONAs against captured VACV particles and against the capture antibody. If clones showed affinity to VACV particles but not to the capture antibody, they were further analyzed for specificity and cross-reactivity.

4.4.1 Affinity testing of DNA aptamers Titration curves of DNA aptamers obtained from cloning are shown in Figure 24. The 90 mer clones 90.6.1 (Figure 24 A) and 90.8.10 (Figure 24 B) bound specifically to VACV with affinities of 27.7 ± 3.8 nM and 73.6 ± 8.8 nM, respectively, and showed only low background binding to the capture antibody (cAb). Clone 77.8.8 (Figure 24 C) did not show a characteristic binding curve, whereas clone 77.4.8 bound specifically to VACV (26.2 ± 2.8 nM) but not to the cAb (Figure 24 D).

- 66 -

Results

A B 1.2 1.2 1.0 1.0 0.8 0.8 0.6 0.6

0.4 0.4 OD (450-620nm) OD 0.2 (450-620nm) OD 0.2 0.0 0.0 0 50 100 150 200 250 0 50 100 150 200 250 Aptamer concentration in [nM] Aptamer concentration in [nM] C D 1.2 1.2 VACV captured mAb 1.0 1.0 0.8 0.8 0.6 0.6

0.4 0.4 OD (450-620nm) OD 0.2 (450-620nm) OD 0.2 0.0 0.0 0 50 100 150 200 250 0 50 100 150 200 250 Aptamer concentration in [nM] Aptamer concentration in [nM]

Figure 24: Titration of DNA aptamers by sandwich ELONA (clones obtained from the classical cloning and sequencing approach) Wells of a MaxiSorp ELISA plate were coated overnight with mAb 710-20. The following day, VACV particles were added and subsequently incubated with twofold serially diluted biotinylated DNA aptamer clones 90.6.1 (A), 90.8.10 (B), 77.8.8 (C), and 77.4.8 (D) (blue). Background binding of aptamers to the capture antibody was assessed by incubation without VACV particles (red). All samples were measured in duplicate. Error bars represent the SD of means. Steady-state affinity (KD) was calculated with GraphPad Prism, using non-linear regression equation “total and non- specific binding”.

Cloning and HTS data analysis led to 8 clones that were similar to clone 90.6.1. The characterization of these clones in terms of affinity and similarity to clone 90.6.1 is summarized in Table 28. The aptamer clones DS 5.9 and DS 5.10 bound more strongly to VACV particles than clone 90.6.1 and were characterized further. All other clones similar to 90.6.1:  did not bind to VACV particles,  bound to the cAb,  showed no or only shallow binding curves, and were discarded. Furthermore, 6 clones similar to clone 90.8.10 were identified. The characterization of these clones is summarized in Table 29. Three clones (DS 5.7, DS 5.8, and DS 7.9) bound more strongly to VACV than clone 90.8.10 and were characterized further. All other clones similar to 90.8.10 did not bind to VACV and were discarded. The characterization of the 3 remaining 90 mer sequences obtained from HTS data analysis is summarized in Table 30. All 3 sequences bound to the cAb and were discarded.

- 67 -

Results

Table 28: Characterization of sequences similar to 90.6.1

Sequence Difference to 90.6.1 KD in nMa) Remark 90.6.1 --- 27.67 Most frequent sequence from cloning 90.6.7 1 nt exchanged 49.84 90.6.8 1 nt exchanged 85.28 No characteristic binding curve 90.6.9 1 nt exchanged 62.26 Shallow binding curve DS 4.4 2 nt exchanged 96.41 No characteristic binding curve DS 5.9 89 nt 7.81 Characterized further DS 5.10 89 nt 6.00 Characterized further DS 6.10 89 nt 0.75 Binding to capture mAb DS 8.8 1 nt exchanged / No binding a) Equilibrium dissociation constant KD was calculated with GraphPad Prism, using non-linear regression equation “total and non-specific binding”.

Table 29: Characterization of sequences similar to 90.8.10

Sequence Difference to 90.8.10 KD in nMa) Remark 90.8.10 --- 73.63 From cloning DS 4.8 1 nt exchanged 0.004 No characteristic binding curve DS 5.7 1 nt exchanged 3.34 Characterized further DS 5.8 1 nt exchanged 22.32 Characterized further DS 7.9 1 nt exchanged 17.87 Characterized further DS 8.6 1 nt exchanged / No binding DS 8.10 1 nt exchanged / No binding a) Equilibrium dissociation constant KD was calculated with GraphPad Prism, using non-linear regression equation “total and non-specific binding”.

Table 30: Characterization of remaining 90 mer sequences

Sequence KD in nMa) Remark DS 2.1 7.45 Binding to capture antibody DS 2.3 9.03 Binding to capture antibody DS 2.5 1.35 Binding to capture antibody a) Equilibrium dissociation constant KD was calculated with GraphPad Prism, using non-linear regression equation “total and non-specific binding”.

For all synthesized DNA aptamers, potential secondary structures were predicted with ValFold, and visualized with Pseudoviewer 3.0 (Figure 25, Figure 26, Figure 27 for 90 mer sequences, and Figure 28 for 77 mer sequences). Figure 25 and Figure 26 show clearly that single nucleotide exchanges can lead to completely different secondary structures. In addition, Table 28 and Table 29 demonstrate that these minor sequence modifications can have a great impact on the binding characteristics of the respective clone. For example, in clone DS 5.7 (similar to 90.8.10) the T on position 55 is substituted by an A, resulting in a completely different secondary structure (Figure 26) and a KD value that is 22 times better. In contrast, clone DS 8.8 has a similar secondary structure to clone 90.6.1 (Figure 25) but a complete loss of binding can be observed. Here, the A on position 23 is substituted by a T.

- 68 -

Results

90.6.1 DS 8.8 DS 5.10 90.6.8

90.6.7 90.6.9 DS 4.4 DS 5.9 DS 6.10

Figure 25: Potential secondary structures of ssDNA aptamers with similarity to clone 90.6.1, predicted with ValFold and visualized with Pseudoviewer 3.0 The DNA sequences shown above differ by only one nucleotide. This can lead to completely different secondary structures (compare clone 90.6.1 and 90.6.8).

90.8.10

DS 5.7 DS 4.8

DS 5.8 DS 8.6 DS 8.10 DS 7.9

Figure 26: Potential secondary structures of ssDNA aptamers with similarity to clone 90.8.10, predicted with ValFold and visualized with Pseudoviewer 3.0

- 69 -

Results

DS 2.1 DS 2.5 DS 2.3

Figure 27: Potential secondary structures of remaining 90mer aptamers, predicted with ValFold and visualized with Pseudoviewer 3.0 These 90 mer sequences did not have any sequence or structure similarity to other clones identified with HTS but were enriched or decreased during the eight selection rounds. All these clones bound to the anti-A27 capture antibody.

77.8.8 JM 1.3 77.4.8

JM 1.2 JM 3.7

JM 4.8

JM 2.10

JM 3.8

JM 5.7 JM 5.6

JM 5.9

Figure 28: Potential secondary structures of ssDNA aptamers (77 mer), predicted with ValFold and visualized with Pseudoviewer 3.0 The clones 77.8.8 and JM 1.3 differ by only one nucleotide and form completely different secondary structures. While clone 77.8.8 did not bind to native VACV particles, clone JM 1.3 did.

- 70 -

Results

In general, various DNA ligands from HTS data analysis were obtained that showed diverse binding characteristics (Figure 29). For both libraries, clones were found that bound better to VACV particles than clones obtained from cloning (Figure 29 A & B). Additionally, non-binding clones (Figure 29 C) and clones that bound to the capture antibody (Figure 29 D) were found. In total, 29 different DNA aptamer clones were analyzed, 14 of which bound to VACV particles but not to the cAb and were then characterized further.

A B 1.0 1.2 0.8 1.0 0.8 0.6 0.6 0.4 0.4

0.2 OD (450-620nm) OD OD (450-620nm) OD 0.2 0.0 0.0 0 50 100 150 200 250 0 50 100 150 200 250 Aptamer concentration in [nM] Aptamer concentration in [nM] C D 2.0 1.2 VACV 1.0 1.5 mAb 0.8 1.0 0.6 0.4

0.5 OD (450-620nm) OD OD (450-620nm) OD 0.2 0.0 0.0 0 50 100 150 200 250 0 50 100 150 200 250 Aptamer concentration in [nM] Aptamer concentration in [nM]

Figure 29: Titration of DNA aptamers by sandwich ELONA (clones obtained from HTS data analysis) Wells of a MaxiSorp ELISA plate were coated overnight with mAb 710-20. The following day, VACV particles were added and subsequently incubated with two-fold serially diluted biotinylated DNA aptamer clones DS 5.10 (A), JM 1.3 (B), DS 2.5 (C), and DS 8.6 (D). Background binding of aptamers to the capture antibody was assessed by incubation without VACV particles. Clones were all obtained after HTS data analysis. Titration curves show examples of clones that show better binding than similar sequences (A & B), unspecific binding (C) and no binding (D) (see text for more details). All samples were measured in duplicate. Error bars represent SD of means. Steady-state affinities (KD) were calculated with GraphPad Prism, using non-linear regression equation “total and non-specific binding”.

The characterization of the 77 mer clones is summarized in Table 31. Clone JM 1.3 which differs from clone 77.8.8 by only one nt (T  C at P 32) bound specifically to VACV particles with an affinity of 11.9 ± 2.3 nM, whereas clone 77.8.8 did not show any characteristic binding to VACV particles. Three 77 mer clones displayed very shallow titration curves, and were discarded. All the other 77 mer clones had affinities to VACV particles in the range of 1.7 to 8 nM (Table 31), showed only low background binding to the cAb, and were characterized further. None of the 77 mer sequences showed binding to the capture antibody such as the 90 mer sequences, leading to the assumption, that short aptamer sequences can bind more specifically.

- 71 -

Results

Table 31: Characterization of 77 mer sequences

Sequence KD in nMa) Remark JM 1.2 7.45 Characterized further JM 1.3 9.03 Similar to 77.8.8, characterized further JM 2.10 1.35 Shallow binding curve JM 3.7 17.87 Shallow binding curve JM 3.8 7.99 Characterized further JM 3.9 1.69 Characterized further JM 4.8 2.09 Characterized further JM 5.6 7.60 Characterized further JM 5.7 1.81 Characterized further JM 5.9 33.20 Shallow binding curve a) Equilibrium dissociation constant KD was calculated with GraphPad Prism using non-linear regression equation total, and non-specific binding.

4.4.2 Specificity testing of DNA aptamers With specificity ELONA, it was evaluated if the aptamer clones could detect other OPV strains, and if they could distinguish OPV from other related or unrelated viruses. To do so, an indirect ELONA was performed as described in section 3.8.5. The aptamer clones were tested for interaction with different VACV strains, the camelpox virus strain CP-19, the cowpox virus strain GuWi, the OPV-related parapoxvirus, and 3 unrelated DNA viruses (AdV, PPV, and FCV). Results for clones JM 3.9 and DS 7.9 are shown exemplarily (Figure 30).

2.5 JM 3.9 2.0 DS 7.9 1.5

1.0

OD (450-620nm) OD 0.5

0.0

AdV PPV FCV PPXV VACV WR MP-Nü SC VACV NY SCVACV NYVACV CC IHD-W CMLV CP-19CPXV SC GuWi SC

Figure 30: Specificity testing of VACV selected DNA aptamers as detection molecules Wells of a MaxiSorp ELISA plate were coated overnight with different OPV strains, PPXV, AdV, PPV and FCV, and were incubated with 10 nM of biotinylated DNA aptamer clones JM 3.9 (striped bars) and DS 7.9 (squared bars) to test for specificity and cross-reactivity. All the OPV strains were detected. Non-OPV samples showed low or no ELONA signals. SC = virus particles purified over a sugar cushion, CC = virus particles obtained from supernatant of infected cell cultures, VACV = vaccinia virus, VACV NY = VACV 1536 strain New York City Board of Health, VACV IHD-W = vaccinia virus strain IHD-W, VACV WR = vaccinia virus strain Western Reserve, CMLV CP-19 = camelpox virus strain CP-19, CPXV GuWi = cowpox virus strain GuWi, MP-Nü = mousepox virus isolate, PPXV = parapoxvirus, AdV = adenovirus, PPV = porcine parvovirus, FCV = feline calicivirus.

- 72 -

Results

The two displayed clones bound to all OPV strains tested and showed no or only background binding (max. OD = 0.3) to PPXV, AdV, PPV, and FCV. Most of the other tested DNA aptamer clones led to comparable results (not shown). However, some tested aptamers, especially the ones obtained from cloning, showed increased background binding to AdV and PPV with OD values up to 0.8. This could not be seen with aptamer clones closely related to these sequences, indicating an increase in specificity because of point mutations in the sequence (not shown). In further experiments, it was evaluated if the DNA aptamer clones capable of detecting VACV particles could be used as capture molecules in a sandwich ELONA in combination with a polyclonal anti-VACV antibody (Figure 31). To estimate the background signal, the direct binding of the VACV particles to the NeutrAvidin plate was assessed. The VACV particles bound slightly to the NeutrAvidin plate, but an increase in signal intensity was achieved when the aptamer clones were used as capture molecules. In contrast, no increase in signal intensity could be observed when PPXV were tested (blue bars). Additionally, aptamer clone DS 8.6 which did not show a typical binding curve, was used as non-binding control (red bars). All tested clones showed no or only slightly increased signal intensities when compared to clone DS 8.6, leading to the assumption that the DNA aptamers are more suitable as detection molecules than as capture molecules.

0.5 77.4.8 JM 1.2 90.6.1 JM 1.3 0.4 90.8.10 JM 3.8 0.3 DS 5.7 JM 3.9 DS 5.8 JM 4.8 0.2 DS 5.9 JM 5.6 DS 5.10 JM 5.7 OD (450-620nm) OD 0.1 DS 7.9 DS 8.6 Virus to plate 0.0

VACV PPXV

Figure 31: Specificity testing of DNA aptamers as capture molecules on NeutrAvidin plates Wells of a NeutrAvidin-coated plate were incubated with 400 nM of biotinylated DNA aptamer clones to capture VACV NY (vaccinia virus strain New York City Board of Health) and parapoxvirus (PPXV). Furthermore, the direct binding of the viruses to the NeutrAvidin plate was assessed (blue bar). For all aptamer clones, an increase in signal intensity was observed when DNA aptamers were used as capture molecules for VACV particles. Clone DS 8.6 was used as non-binding control (red bars). Only some aptamer clones showed increased signal intensities compared to clone DS 8.6. Bars represent means of duplicates ± SD.

- 73 -

Results

4.4.3 Identification of binding partner for DNA aptamer clones The specificity and affinity of several DNA aptamer clones was assessed in former experiments. All clones interacted with various OPV strains, but the interaction partners of theses clones remain unknown. To evaluate the binding partner of the selected DNA aptamer clones, different approaches were used. Since the selection was performed against native VACV particles, one approach was to perform an ELONA with different poxviral surface proteins (Figure 32). The other approach was to perform a pull-down assay with subsequent SDS-PAGE (Figure 33) and MALDI/MS analysis of obtained protein bands. According to protein ELONA results, all tested aptamer clones bound to lysate of VACV particles. DNA aptamer clones 90.8.10 (Figure 32 A), 90.6.1, DS 5.9 (Figure 32 B), JM 1.2, JM 3.8 and JM 3.9 (Figure 32 C & D) seemed to bind to the major poxviral surface protein A27 (OD varied from 0.7 to 1.6). For all other DNA aptamer clones, an explicit interaction partner could not be identified by ELONA.

A B

1.4 1.8 90.8.10 90.6.1 1.2 1.6 DS 5.9 DS 5.7 1.4 1.0 DS 5.10 DS 5.8 1.2 SA-POD 0.8 DS 7.9 1.0 0.6 0.8 0.4 0.6

0.4 OD (450-620nm) OD 0.2 (450-620nm) OD 0.2 0.0 0.0 H3 L1 D8 F9 B5 H3 L1 D8 F9 B5 A27 A33 F13 BSA A27 A33 F13 BSA Lysate Lysate C D 1.6 1.6 JM 1.2 1.4 1.4 JM 3.9 JM 1.3 1.2 1.2 JM 4.8 JM 3.8 JM 5.6 1.0 1.0 JM 5.7 0.8 0.8 0.6 0.6

0.4 0.4 OD (450-620nm) OD 0.2 (450-620nm) OD 0.2 0.0 0.0 H3 L1 D8 F9 B5 H3 L1 D8 F9 B5 A27 A33 F13 BSA A27 A33 F13 BSA Lysate Lysate

Figure 32: Identification of binding partner for DNA aptamers with protein ELONA Wells of a MaxiSorp ELISA plate were coated with 200 ng of various recombinant poxviral proteins (A27, H3, L1, D8, F9, A33, B5, and F13) or lysate of purified VACV and were incubated with 10 nM of biotinylated DNA aptamer clones similar to clone 90.8.10 (A), similar to clone 90.6.1 (B) or with different 77 mer clones (C+D). Bars represent means of duplicates ± SD.

To confirm the interaction of clones above listed with A27, and to identify the interaction partner of all other clones, a pull-down assay was performed. For the pull-down assay, the biotinylated DNA ligands were coupled to magnetic beads and were then incubated with lysate of purified washed VACV particles (see 3.9.3 for more details). The beads were washed several times, and potentially bound lysate components were eluted with RIPA buffer. The

- 74 -

Results eluates were applied to SDS-PAGE, and the gels were stained with ‘Blue Silver staining’ solution to visualize proteins. The analysis of the SDS-PAGE (Figure 33) showed the following banding pattern:  two dominant protein bands (one between 100 kDa and 130 kDa, and one between 10 kDa and 15 kDa), also present in the initial lysate (lanes 7 & 14),  in lanes representing 90 mer clones (Figure 33lanes 4, 5, 12, and 13), an additional band between 25 kDa and 35 kDa,  in lanes representing 77 mer clones (lanes 2, 3, 6, 9, 10, and 11), an additional band between 15 kDa and 25 kDa, not visible in the initial lysate,  additional 1-2 faint bands between 15 kDa and 25 kDa in some lanes. For identification, the protein bands were cut out and prepared for MALDI/MS analysis. Due to technical problems with the MALDI in the institute, the identification of the protein bands has not been possible so far.

Figure 33: SDS-PAGE of eluted lysate components captured with DNA aptamers 1 & 8: Prestained protein marker, rows 2-13 contain the eluate of beads coupled with DNA aptamer clones, 2: 77.4.8, 3: 77.8.8, 4: 90.6.1, 5: 90.8.10, 6: JM 1.3, 7 & 14 lysate, 9: JM 4.8, 10: JM 5.6, 11: JM 5.7, 12: DS 5.9, and 13: DS 5.10. VACV particles were lysed with IP buffer and added to magnetic beads containing the DNA aptamer clones listed above. After several washing steps, bound lysate components were eluted with RIPA buffer. Eluates were applied to SDS-PAGE. Gels were stained with ‘Blue Silver Staining’ solution. For all aptamer clones, two dominant protein bands are visible that are also present in the initial lysate: one between 100 kDa and 130 kDa, and one between 10 kDa and 15 kDa. Depending on clone length, another dominant band can be detected: one between 15 kDa and 25 kDa (for 77 mer clones) and one between 25 kDa and 35 kDa (for 90 mer clones), which cannot be seen in the lysate. For some clones, additional protein bands between 15 kDa and 25 kDa were detectable. These protein bands were present in the initial lysate as well. All protein bands were cut out of the gel and prepared for MALDI/MS analysis.

- 75 -

Results

4.5 Selection of peptide aptamers via Phage Display1

Selection of peptide aptamers was performed as described under 3.6.1 and 3.6.2 against the recombinant poxviral surface protein A27 and against infectious VACV particles, for 3 and 4 selection rounds, respectively. After each selection round, the titer of the eluate was determined (see 3.5.3), and individual phage clones were picked and sequenced (see 3.6.4). An increase in titers was observed after each selection round (Figure 34), indicating that individual target-binding phage clones have been enriched.

10 8

10 7 A27

Titer [pfu/mL] VACV 10 6 1 2 3 4 Selection round

Figure 34: Titers of phage eluates determined after each selection round The eluates containing binding phages were titered after each selection round. The selections were performed against immobilized native VACV particles and against rA27. An increase in titers from the first to the last selection round could be observed for both targets. This indicated an enrichment of target-binding phage clones. (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.)

4.5.1 Identification of consensus peptide sequences After each selection round, individual phage clones were picked from agar plates used for titration to identify encoded peptide sequences. For this purpose, altogether 141 (A27) and 103 (VACV) plaques were picked. The ssDNA was extracted and sequenced. The obtained DNA sequences were analyzed with the “Library Insert Finder” software developed in-house for rapid identification of encoded peptide sequences. For the selection, a premade dodecapeptide (Ph.D.-12) library was used. Therefore, all identified peptide sequences had to contain 12 amino acids. For each selection (A27 and VACV), an enrichment of individual phage clones (identified by their DNA) was observed from the first to the last selection round. The frequency of identified peptide sequences ranged from 1 to 6 (for A27) or 8 (for VACV) in individual eluates, respectively. Peptide sequences that were found more than once and all other peptide sequences from the last selection round (for anti-A27 phages) were

1 The phage display selection against native VACV particles and the pre-characterization of the obtained peptides was performed together with Lilija Miller, distributed in equal parts.

- 76 -

Results declared to be consensus sequences. Consensus peptide sequences for selection against A27 and VACV particles are listed in Table 32 and Table 33. The occurrence of these consensus peptide sequences after multiple selection rounds indicated an enrichment of specifically binding clones. Nevertheless, it was useful to prove by phage ELISA if theses phage clones really bound to the target.

Table 32: Consensus peptide sequences of A27-selected phage clones Clone # Peptide sequence Frequencya) 1st eluate 2nd eluate 3rd eluate 1 RTENMMMSPLSL 1 2 RLDTQNSWTLIH 1 3 TLNPGDVTVSPR 5 4 4 TIGIASNMDRRP 1 6 5 AANNNALSFKTL 1 1 6 MHEPPGPAQNSS 2 7 TGTYTFTTHSLR 2 8 KVYDLHKLHVAQ 2 2 9 LPLSATAIPRET 1 4 10 KIWVHQTPVSRA 5 11 SLKVWDLNKNRV 3 12 FHPGKAWVIAMK 1 13 VESRHTDPSQWA 1 14 WSVDYGRKLWLI 1 15 ATAPAQRDLVTK 1 16 SLMVTNANTNAT 1 17 KVWTFSPNLPIR 1 18 FTTIATTDPLWH 3 19 TNSSLKSDWPGT 1 20 TNENMTTSRTGS 3 21 DGNLLFWDMQVG 1 22 KIWNIAPEPAMP 1 23 APALPMATNTRY 1 24 SGQTTPSTYHTT 2 25 VYPFRLEHNLPQ 1 26 DHIAKNQHQYTT 1 a) Frequency of how often the peptide sequence was found in the respective phage eluate

- 77 -

Results

Table 33: Consensus peptide sequences of VACV-selected phage clones (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.) Clone # Peptide sequence Frequencya) 2nd eluate 3rd eluate 4th eluate 1 KVWLPPRHEHQY 2 1 3 2 KVFYPAAANPNQ 1 1 3 QALLEGNAKGGN 1 2 4 NRPDSAQFWLHH 3 5 TADKLLYGLFKS 1 6 6 DEWDALLMRIRT 1 3 7 NPTPYPMLPLRG 8 8 KPTYSWDPAQLK 2 9 GPTFSWDHLRGQ 2 10 KIFQLPQISPPM 4 11 GDLASWIITSFK 8 12 NMELHPHSLPRP 2 13 ANTTKHSVLAAI 2 14 EALNDWVNDSEY 2 15 APTAYNKNDWAL 2 16 DPWWRGNEARAA 2 17 TPVWSWEPPLQE 4 a) Frequency of how often the peptide sequence was found in the respective phage eluate

4.5.2 Evaluation of selectivity of enriched phage clones by phage ELISA Phage clones bearing consensus peptide sequences, were applied to phage ELISA (see 3.6.6) to determine if these clones bound to the respective target (rA27 or VACV particles). For this, the target was coated to wells of an ELISA plate. This was followed by incubation with individual, amplified phage clones, which were detected using an anti-phage Ab. Results of ELISA with potentially A27-binding phages are shown in Figure 35 and those with potentially VACV-binding phages in Figure 36. In addition to the targets, multiple controls were utilized. If available, the eluates of all selection rounds were measured to evaluate an enrichment of phages binding to the target during consecutive selection rounds. Furthermore, the unselected Ph.D.-12 library and the wildtype phage M13KO7 were used as negative controls to assess background binding. To test anti-A27 consensus phage clones for specificity, wells were coated with:  A27, expressed in E.coli and in a Baculovirus system,  native VACV particles,  L1, another poxviral surface protein,  casein, the blocking reagent,  parapoxvirus (PPXV), as a related DNA virus,  and adenovirus (AdV), an unrelated DNA virus. The addition of phage clones to wells containing the controls (rL1, Casein, PPXV, and AdV) resulted in low signals (OD = 0.01-0.5) for all samples tested, except for two (clones # 18 and 19). In general, all individual phage clones bound better to rA27 and VACV than to the controls, with ELISA signals ranging from 0.9 to 3.0 OD. For rA27, an increase in ELISA signal

- 78 -

Results intensity for the eluates was observed from round 1 to round 3. None of the samples tested showed binding to A27 expressed in E.coli (Figure 35). To test anti-VACV consensus phage clones for specificity, wells were coated with:  native VACV particles,  porcine parvovirus (PPV), an unrelated DNA virus,  and BSA, the blocking reagent. The addition of phage clones to wells containing BSA resulted in low signals (OD = 0.03- 0.3) for all samples tested. In general, all individual phage clones tested showed better binding to VACV than to PPV, with ELISA signals ranging from 0.4 to 2.0 OD (Figure 36). All control phages, including the eluates, the Ph.D.-12-library, and the wild type phage, bound more strongly to VACV than to PPV. For the eluates, an increase in ELISA signal intensity for VACV could be observed from round 2 to round 3, indicating an enrichment of high affine binders (Figure 36).

- 79 -

Results

A27 VACV Casein L1 A27 E.coli PPXV AdV 3.5  3.0    2.5    2.0  1.5

1.0 OD (450-620nm) OD 0.5 0.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Phage clone number M13KO7Ph.D.-12 1st2nd eluate3rd eluate eluate

Figure 35: Anti-A27 phage clones tested for specificity in phage ELISA Wells were coated with 200 ng of recombinant protein (rA27 and rL1) or 105 pfu/well VACV, PPXV or AdV particles overnight, to test target binding of individual phage clones. Wells coated with PBS containing 0.05% Casein were used to test for phages binding to the blocking reagent used during the selection. Consensus phage clones (1-26) were added at 1010 pfu/well. Additionally, the 1st, 2nd, and 3rd eluate, as well as the unselected Ph.D.-12- library and the wildtype phage M13KO7 were tested. Binding of phages was detected by incubation with monoclonal anti-M13 antibody. Bars represent means ± SD of duplicates. Red arrows point to phage clones which were selected for synthesis of peptides. VACV = vaccinia virus strain New York City Board of Health, PPXV = parapoxvirus, AdV = Adenovirus

- 80 -

Results

2.5 VACV PPV BSA  2.0  1.5    1.0

OD (450-620nm) OD 0.5

0.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Phage clone number Ph.D.-12M13KO7 2nd eluate3rd eluate4th eluate

Figure 36: Anti-VACV phage clones tested for specificity in phage ELISA Wells were coated with 105 pfu/well of VACV or PPV particles overnight, to test target binding of individual phage clones. Wells coated with PBS containing 2 % BSA were used to test for phage clones binding to the blocking reagent. Consensus phage clones (1-17) were added at 1010 pfu/well. Additionally, the second, third and fourth eluate as well as the unselected Ph.D.-12 library and the wildtype phage M13KO7 were tested. Binding of phages was detected by incubation with monoclonal anti-M13 antibody. Bars represent means ± SD of duplicates. Red arrows point to phage clones which were selected for synthesis of peptides. VACV = vaccinia virus strain New York City Board of Health, PPV = porcine parvovirus, BSA = bovine serum albumin. (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.)

- 81 -

Results

4.5.3 Selection of peptide sequences for synthesis A phage ELISA can give an impression of how specific a peptide displayed by a phage clone can bind to its target. However, the selected random peptide is still attached to the phage particle. Therefore, it is possible that interactions of other phage surface proteins can occur during phage ELISA. To study the binding of a peptide to its target, and to rule out unspecific interactions, selected peptide sequences were synthesized as free, soluble peptides. The selection of peptide sequences for synthesis was based on the frequency of occurrence of the respective phage clone (Table 32 & Table 33) and the corresponding ELISA signal (Figure 35 & Figure 36). The objective of the peptides selected against rA27 is the application as detection molecules, not only for the recombinant protein rA27 but also for OPV particles. Therefore, peptides displayed by phage clones with higher OD signal for rA27 and for VACV than for the controls (PPV, BSA) were chosen for synthesis (indicated by red arrows and listed in Table 34). Additionally, clone number 18 was chosen for synthesis to test if not only the phage clone but the peptide also showed high background binding. As for the phage clones selected against VACV particles, clones with high signal for VACV and concomitant low signals for PPV and BSA were chosen for synthesis (Table 35).

Table 34: Synthetic anti-A27 peptides derived from phage display Clone # Peptide sequencea) MWb) c) 2 Ac-RLDTQNSWTLIHGGGSK-NH2 (Biotin) 2,137 3 Ac-TLNPGDVTVSPRGGGSK-NH2 (Biotin) 1,909 4 Ac-TIGIASNMDRRPGGGSK-NH2 (Biotin) 1,984 8 Ac-KVYDLHKLHVAQGGGSK-NH2 (Biotin) 2,104 9 Ac-LPLSATAIPRETGGGSK-NH2 (Biotin) 1,923 10 Ac-KIWVHQTPVSRAGGGSK-NH2 (Biotin) 2,076 17 Ac-KVWTFSPNLPIRGGGSK-NH2 (Biotin) 2,113 18 Ac-FTTIATTDPLWHGGGSK-NH2 (Biotin) 2,057 a) Sequences are provided using the one letter amino acid code from N-terminus to C-terminus. All peptides were ordered from Biomatik (Cambridge, Canada) b) Molecular weight (MW) in Dalton c) Peptides were synthesized with C-terminal Biotin, and abbreviations Ac and NH2 refer to an N- terminal acetyl group and a C-terminal amidation

Table 35: Synthetic anti-VACV peptides derived from phage display (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.) Clone # Peptide sequencea) MWb) c) 1 Ac‐KVWLPPRHEHQYGGGSK‐NH2 (Biotin) 2,243 5 Ac‐TADKLLYGLFKSGGGSK‐NH2 (Biotin) 2,009 8 Ac‐KPTYSWDPAQLKGGGSK‐NH2 (Biotin) 2,087 13 Ac‐ANTTKHSVLAAIGGGSK‐NH2 (Biotin) 1,856 14 Ac‐EALNDWVNDSEYGGGSK‐NH2 (Biotin) 2,085 a) Sequences are provided using the one letter amino acid code from N-terminus to C-terminus. All peptides were ordered from PANATecs GmbH (Tuebingen, Germany) b) Molecular weight (MW) in Dalton c) Peptides were synthesized with C-terminal Biotin, and abbreviations Ac and NH2 refer to an N- terminal acetyl group and a C-terminal amidation

- 82 -

Results

Synthetic peptides were tested as potential detection molecules for rA27 and VACV particles. From the anti-A27 selection, 8 out of 26 consensus peptide sequences were synthesized (Table 34), and from the anti-VACV selection, 5 out of 17 (Table 35) consensus peptide sequences. The Ph.D.-12library contains a short linker sequence (Gly-Gly-Gly-Ser) between the displayed peptide and the phage protein pIII. To obtain similar reaction conditions as the ones during the selection process, this sequence was added at the C- terminus of the peptides. For characterization in an ELISA, a C-terminal biotin was attached over a lysine residue. Altogether, synthetic peptides were ordered with the following criteria/characteristics:  Free N-terminus (as present during the selection),  Acetylated N-terminus (to block positive charge),  Linker sequence and biotin added to C-terminus,  Amidated C-terminal carboxylate (to block negative charge),  Peptide purity of over 85 %,  Peptide amount of 1-4 mg.

4.6 Characterization of selected peptide ligands

The phage-free soluble peptides should be evaluated according to their potential as detection molecules for OPV. For this purpose, the affinity and specificity of the synthesized peptides was assessed with an indirect ELISA. If peptides showed any significant affinity, the specificity and cross-reactivity of the corresponding peptide were determined.

4.6.1 Affinity testing of synthetic peptide ligands For affinity testing of synthetic peptides (see 3.8.7), recombinant proteins (for anti-A27 peptides) or viral particles (for anti-VACV peptides) were directly coated into wells of an ELISA plate and incubated with a two-fold serial dilution series of the biotinylated peptide. To determine specificity of anti-A27 or anti-VACV peptides, additional wells were coated with rL1 and rB5 (Figure 37) or two VACV-unrelated DNA viruses (PPV and AdV) (Figure 38). None of the anti-A27 peptides showed specific binding to rA27 (Figure 37). If binding occurred, the peptides bound to all proteins tested (Figure 37 C). The same could be observed for peptides obtained from phage display selection against two other poxviral proteins, rL1 and rB5 (Figure 39). Both peptides showed high affinity, but no specificity. In further experiments, the peptides showed the same effect when tested against various viral particles (VACV, PPXV, PPV, FCV, and AdV) (data not shown), leading to the assumption that theses peptides had a general affinity to proteins.

- 83 -

Results

A B 2.5 2.5 2.0 2.0 1.5 1.5 1.0 1.0

0.5 0.5

OD (450-620nm) OD OD (450-620nm) OD 0.0 0.0 0 5 10 15 20 25 0 5 10 15 20 25 Peptide concentration [µM] Peptide concentration [µM] C D 3.0 2.5 2.5 2.0 2.0 1.5 1.5 1.0 1.0

0.5 0.5

OD (450-620nm) OD (450-620nm) OD 0.0 0.0 0 5 10 15 20 25 0 5 10 15 20 25 Peptide concentration [µM] Peptide concentration [µM] E F 2.5 2.5 2.0 2.0 1.5 1.5 1.0 1.0

0.5 0.5

OD (450-620nm) OD OD (450-620nm) OD 0.0 0.0 0 5 10 15 20 25 0 5 10 15 20 25 Peptide concentration [µM] Peptide concentration [µM] G H 2.5 2.5 2.0 2.0 L1 A27 1.5 1.5 B5 1.0 1.0 Casein

0.5 0.5

OD (450-620nm) OD (450-620nm) OD 0.0 0.0 0 5 10 15 20 25 0 5 10 15 20 25 Peptide concentration [µM] Peptide concentration [µM]

Figure 37: Affinity testing of anti-A27 peptides by peptide ELISA Wells of a MaxiSorp ELISA plate were coated with rL1, rA27, rB5, or the blocking reagent casein (to estimate background binding). Then they were incubated with twofold serially diluted biotinylated peptide derived from (A) phage clone number 2, (B) clone number 3, (C) clone number 4, (D) clone number 8, (E) clone number 9, (F) clone number 10, (G) clone number 17, and (H) clone number 18. The peptide was detected by incubation with SA-POD through TMB color reaction. All samples were measured in duplicate. Error bars represent standard deviation. None of the tested peptides showed specific binding to rA27.

- 84 -

Results

As for the anti-VACV peptides, only the peptide resulting from clone 5 (Figure 38 B) showed binding to VACV particles. This peptide showed no significant binding to PPV or AdV particles. The KD value for clone 5 to VACV particles came to 78 ± 5.7 nM and was calculated using nonlinear regression analysis choosing the binding saturation equation for “monovalent total and nonspecific binding” (GraphPad Prism Software Version 5.01). In the following chapters, this peptide is referred to as VV1.

A B 2.0 2.0

1.5 1.5

1.0 1.0

0.5 0.5

OD(450-620nm) OD(450-620nm)

0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Peptide concentration [µM] Peptide concentration [µM] C D 2.0 2.0

1.5 1.5

1.0 1.0

0.5 0.5

OD(450-620nm) OD(450-620nm)

0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Peptide concentration [µM] Peptide concentration [µM]

E 2.0 VACV 1.5 PPV Adenovirus 1.0

0.5 OD(450-620nm)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Peptide concentration [µM]

Figure 38: Titration of anti-VACV peptides by peptide ELISA Wells of a MaxiSorp ELISA plate were coated with VACV, PPV, or AdV and incubated with the two- fold serially diluted biotinylated peptide derived from (A) phage clone number 1, (B) clone number 5, (C) clone number 8, (D) clone number 13, and (E) clone number 14. The peptide was detected by incubation with SA-POD through TMB color reaction. All samples were measured in duplicate in two independent experiments. Error bars represent the standard deviation. Only the peptide derived from clone number 5 (B) showed specific binding to VACV but not to PPV and AdV, with a computed KD value of 78 nM. (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.)

- 85 -

Results

A B 3.0 3.5 2.5 3.0 2.5 2.0 A27 2.0 1.5 B5 1.5 1.0 L1 1.0 BSA

0.5 OD (450-620nm) OD

OD (450-620nm) OD 0.5 0.0 0.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Peptide concentration [µM] Peptide concentration [µM]

Figure 39: Titration of anti-L1 (A) and anti-B5 (B) peptides by peptide ELISA Wells of a MaxiSorp ELISA plate were coated with rL1, rA27, rB5, or the blocking reagent BSA (to estimate background binding). Then they were incubated with a twofold serially diluted biotinylated peptide derived from (A) a phage clone selected against L1 and (B) a phage clone selected against B5. Peptides were detected by incubation with SA-POD through TMB color reaction. Both peptides showed that if binding occurred, the peptide would bind unspecifically to all the proteins tested. All samples were measured in duplicate. Error bars represent standard deviation.

4.6.2 Specificity testing of anti-VACV peptide VV1 using different detection systems From several peptides selected against rA27 and VACV, only one (VV1) showed specific affinity to OPV. To characterize VV1 binding in more detail, the suitability of the peptide as detection and as capture molecule for different OPV particles was assessed. The principle of these assays is shown in Figure 40. A B

secondary antibody labeled with HRP SA-POD anti-virus antibody

biotinylated peptide virus

virus biotinylated peptide

Figure 40: Principle of the two detection systems used, with the peptide as detection molecule (A) and as capture molecule (B) A) In this system the peptide is applied as detection molecule. Here, the viral particles are immobilized into wells of a multi-well plate and the biotinylated peptide is added and subsequently detected by incubation with SA-POD (streptavidin peroxidase). B) In this system, the peptide is applied as capture molecule. Here, the biotinylated peptide is immobilized into NeutrAvidin coated wells of a multi-well plate. The viral particles are added and detected by incubation with a specific anti-virus antibody, followed by the incubation with a secondary antibody labeled with HRP (horse radish peroxidase). In both assays, a TMB color reaction is detected with a microplate reader.

- 86 -

Results

To evaluate the peptides’ specificity as detection molecule, particles of several OPV strains and other related or unrelated virus particles were coated into wells of a MaxiSorp ELISA plate and incubated with 400 nM of peptide VV1 (Figure 41). Next to the sugar cushion (SC)-purified VACV NY (OD = 0.5), the camelpox virus (CMLV) strain CP-19, the cowpox virus (CPXV) strain GuWi, and a mousepox virus isolate (MP-NÜ) could be detected with OD signals ranging from 1.3 to 1.9. Samples from inactivated virus particles (UV and heat) as well as virus particles obtained from cell culture supernatants (CC) showed very low OD signals (0.07 to 0.1) undistinguishable from the negative controls (e.g. PPXV or HEp- 2 with OD values of 0.25). The differences in OD signals from native and inactivated VACV NY could be a hint that VV1 binds to a conformational epitope rather than a linear epitope.

2.0

1.5

1.0

OD OD (450-620nm) 0.5

0.0

BSA AdV CCPPV CCFCV CC HEp-2 PPXV CC MP-Nü SC VACV NYVACV SC NY CC CPXV BR SC VACV WR CC VACV NY RIPA CMLV CP-19CPXV SC GuWi SC VACV VACVNY SC NYUV SC heat VACV IHD-W CC

Figure 41: Specificity testing of VACV-selected peptide VV1 as detection molecule Wells of a MaxiSorp ELISA plate were coated with different OPV strains, PPXV, Adv, PPV, FCV, HEp- 2, and BSA, and were then incubated with 400 nM of biotinylated peptide VV1 to test for specificity and cross-reactivity. Only OPV could be detected, and samples with purified virus particles led to higher signals than virus particles from cell culture supernatants. All non-OPV samples showed low or no ELISA signals. SC = virus particles purified over a sugar cushion, CC = virus particles obtained from supernatant of infected cell cultures, VACV = vaccinia virus, VACV NY = VACV 1536 strain New York City Board of Health, VACV UV = UV-inactivated VACV 1536, VACV heat = heat-inactivated VACV 1536, VACV RIPA =lysate of VACV 1536-infected HEp-2 cells, CPXV BR = cowpox virus Brighton Red, VACV IHD-W = vaccinia virus strain IHD-W, VACV WR = vaccinia virus strain Western Reserve, CMLV CP-19 = camelpox virus strain CP-19, CPXV GuWi = cowpox virus strain GuWi, MP- Nü = mousepox virus isolate, PPXV = parapoxvirus, AdV = adenovirus, PPV = porcine parvovirus, FCV = feline calicivirus, HEp-2 = lysate of HEp-2 cells, BSA = bovine serum albumin

For further characterization of VV1-binding properties, it was analyzed if the peptide could be used as a capture molecule in a sandwich ELISA in combination, with a polyclonal anti-VACV antibody (Figure 42). At first, to determine the optimal peptide concentration, a twofold serial dilution series of peptide VV1 starting at 25 µM was added to NeutrAvidin- coated wells. VACV particles were added and detected by incubation with p-anti-VACV Ab. The resulting ELISA signals ranged from 1.3 OD (for 0.24 µM) to 1.5 OD(for 24.9 µM) with little variation in between. The addition of buffer instead of peptide resulted in an ELISA

- 87 -

Results signal of 0.1 OD, indicating that no VACV particles bound to the NeutrAvidin plate (data not shown). The binding capacity of the NeutrAvidin-coated wells was given as ~15 pmol D- biotin/well. To assure saturation of all binding sites with the biotinylated peptide, 400 nM (equaling 45 pmol in 100 µl) were chosen for further experiments. Sandwich ELISA results (Figure 42) showed that all tested OPV strains could be detected with ELISA signals ranging from 1.2 OD (MP-Nü) to 2.3 OD (VACV NY CC). Again, the inactivated VACV particles could not be detected. Furthermore, the control viruses (PPXV, PPV, and FCV), as well as HEp-2 and BSA, showed only background binding (OD =0.12). The increased ELISA signal intensity for AdV (OD = 0.6) was due to unspecific binding of the anti- VACV antibody to the AdV particles (data not shown) which bound to the NeutrAvidin plate. Experiments with a monoclonal anti-VACV Ab confirmed that the AdV suspension was not contaminated with OPV (data not shown).

2.5

2.0

1.5

1.0 OD OD (450-620nm) 0.5

0.0

BSA AdV CCPPV CCFCV CC HEp-2 PPXV CC MP-Nü SC VACV NYVACV SC NY CC CPXV BR SC VACV WR CC MPXV X12MPXV CC Mü CC CMLV CP-19CPXV SCGuWi SC VACV VACVNY SC NY UV SC heat VACV IHD-W CC

Figure 42: Specificity testing of peptide VV1 as capture molecule on NeutrAvidin plates Wells of a NeutrAvidin-coated plate were incubated with 400 nm of biotinylated peptide VV1 to capture various different VACV strains (VACV NY = vaccinia virus strain New York City Board of Health, VACV IHD-W = vaccinia virus strain IHD-W, VACV WR = vaccinia virus strain Western Reserve), a mousepox virus isolate (MP-NÜ), the camelpox virus strain CP-19 (CMLV CP-19), the cowpox virus strains Brighton Red and GuWi (CPXV BR and CPXV GuWi), and the monkeypox virus strains (MPXV X12 and MPXV Mü) to test for cross-reactivity, parapoxvirus (PPXV), bat adenovirus (AdV), porcine parvovirus (PPV), feline calicivirus (FCV), lysate of HEp-2 cells (HEp-2) and bovine serum albumin (BSA). SC = virus particles purified over a sugar cushion, CC = virus particles obtained from supernatant of infected cell cultures. All samples were measured in duplicate. Error bars represent the standard deviation.

Results of a third approach revealed that VV1 was not functional as a detection molecule for OPV particles in a sandwich ELISA with anti-VACV antibodies as capture molecules (data not shown). It has been described that aptamers do not show any differences between synthesis charges. To verify this, peptide VV1 was reordered from two different suppliers, and the affinity (to VACV and D8) and specificity ELISAs were repeated (data not shown). It could be

- 88 -

Results confirmed that peptides from different charges and different suppliers show no or only little variation for affinity and specificity.

4.6.3 Identification of binding partner for peptide VV1 In previous experiments, the affinity and specificity of peptide VV1 to native VACV particles could be shown. The peptide was obtained from a selection against whole virus particles, therefore the exact binding partner was unknown. To identify the binding partner, an indirect ELISA with various poxviral surface proteins was conducted (Figure 43). Peptide VV1 only bound to the poxviral surface protein D8, indicating that D8 was the interaction partner of VV1. The KD value came to 157 ± 15 nM, and was calculated using nonlinear regression analysis - choosing the binding saturation equation for “monovalent total binding” (GraphPad Prism Software Version 5.01).

3.0 D8 2.5 F9 A27 2.0 B5 L1 1.5 VACV RIPA F13 1.0

A33 OD OD (450-620nm) 0.5

0.0 0 1 2 3 4 5 Peptide concentration [µM]

Figure 43: Identification of binding partner for VV1 via indirect ELISA Wells of a MaxiSorp ELISA plate were coated with the poxviral proteins F9, A27, B5, L1, D8, F13, A33, and a lysate of VACV-infected cells (VACV RIPA), and were then incubated with a twofold serial dilution series of biotinylated peptide VV1. After washing, the wells were incubated with streptavidin conjugated to HRP. The peptide aptamer binding was detected using a TMB color reaction. VV1 showed only binding to D8. For all other proteins tested, only background binding could be observed. This indicated that D8 was the interaction partner for peptide VV1. All samples were measured in duplicate. Error bars represent standard deviation.

Another possibility to identify a binding partner was the use of a pull-down assay with subsequent nLC-MS/MS analysis. Such a pull-down assay was performed with peptide VV1, coupled to magnetic biotin-binding beads (compare 3.9.3). Lysate of purified VACV particles was prepared. The lysate was incubated with the magnetic beads and then washed several times before bond lysate components were eluted. The initial lysate, the wash fractions 1 and 5-6, as well as the eluate were applied to SDS-PAGE. The gel was stained with a fluorescent dye. As a control, the same experiment was performed with peptide VV5, a non- binding peptide. In Figure 44, the stained SDS gels are shown. In lane 1 (containing the lysate of VACV particles) of both gels, various protein bands are visible. Lane 2 shows a

- 89 -

Results smear of BSA protein bands, which was present in the first washing buffer. No proteins can be detected in lanes 3 and 4 (wash fractions 5 & 6). In lane 5 of Figure 44 A, one distinct band between 25 kDa and 37 kDa appeared which was identified as the poxviral surface protein D8 by MALDI/MS analysis, whereas in the corresponding lane using the control peptide (Figure 44 B) no protein band could be detected.

1 2 3 4 5 6 1 2 3 4 5 6 A B

Figure 44: Identification of binding partner for VV1 via pull-down assay, SDS-PAGE and subsequent nLC-MS/MS analysis Peptide VV1 (A) and peptide VV5 (B) were immobilized on magnetic beads and incubated with lysate of purified VACV particles. After several washing steps, bond lysate components were eluted. The initial lysate (lane 1), the wash fractions 1 (lane 2), 5 (lane 3), and 6 (lane 4), and the eluate (lane 5), as well as a protein marker (lane 6) were applied to an SDS-PAGE. The gel was stained with the fluorescent dye flamingo and photographed with Image Lab software (v4.0.1). In lane 5 (A), one distinct protein band between 25 kDa and 37 kDa could be detected (red ellipse). This protein band was cut out and applied to MALDI/MS analysis. The protein was identified as the poxviral surface protein D8. In comparison, the eluate of VV5-immobilized beads (B, lane 5) did not show any protein bands.

4.6.4 Surface plasmon resonance (SPR)/Biacore analysis of interaction of peptide VV1 with D8 The binding partner of peptide VV1 had been identified as the poxviral surface protein D8. To further analyze the interaction of VV1 and D8 and to verify specific binding of VV1 to D8, SPR/Biacore assays were performed. For these assays, the biotinylated peptide VV1 was spotted onto flow cell 2 (Fc 2), and the non-binding biotinylated peptide VV5 was spotted onto Fc 1 of a streptavidin chip. Experiments were performed as described in section 3.9.2. D8 was added at a twofold serial dilution series to both flow cells, and the difference (Fc 2-Fc 1) of the response signals (RU) was plotted and fitted (Figure 45). Response units (ΔRU) increased with higher D8 concentrations. When binding kinetics were determined by fitting to a 1:1 Langmuir binding model to determine binding kinetics, an association rate 4 -1 -1 -2 -1 constant ka of 8.4 × 10 M s , a dissociation rate constant kd of 8 × 10 s , and an equilibrium dissociation constant of 0.95 ×10-6 M were measured, while analysis of steady- state binding, to determine binding affinity, gave an equilibrium dissociation constant of 1.1 × 10-6 M. Hence, affinities determined by both methods were in close agreement; the interaction was characterized by rapid unbinding, leading to lowered affinity.

- 90 -

Results

200

100

RU (Fc2-Fc1)  0 0 50 100 150 time [s]

Figure 45: Binding response and curve fit for D8 binding to peptide VV1 Binding of a two-fold serial dilution series of D8 (3.25 nM – 1.67 µM; duplicate injection of 1.67 µM in HBS/EP+ buffer) to peptide VV1 (67 RU immobilized on Fc 2, negative control: peptide VV5, 269 RU immobilized on Fc 1). Association was assessed for 120 seconds and dissociation for 60 seconds. Measured binding responses (red lines) were fit using a two-state reaction model (black lines, BIAevaluation Software 1.0).

In further experiments, injection of D8 resulted in positive binding signals, whereas injection of various other recombinant poxviral proteins resulted in no or negative binding responses, which proved specific interaction between VV1 and D8 (Figure 46).

300 200 100 0

-100 RU (Fc2-Fc1)

 -200 -300

F9 L1 B5 A33 A27 Buffer D8 E. coli B5 E. coli H3 E.D8 coli E. coli F13 A33E. coli E. coli Recombinant proteins

Figure 46: Binding analysis to assess the specificity of VV1 peptide binding to D8 via SPR/Biacore Recombinant orthopoxviral surface proteins were diluted to 50 µg/mL in HBS-EP+ buffer and injected for 120 seconds at a flowrate of 5 µl/min over immobilized VV1 (Fc2) and VV5 (Fc1). Only D8 injection generated positive binding signals. Negative binding responses were caused by differential refractive changes on Fc1 and Fc2, most probably caused by buffer components used to stabilize recombinant proteins. The D8 protein was injected twice, once at the beginning and once at the end of the experiment.

A two-state binding mechanism for VV1-D8 interaction was postulated. This means that a conformational change takes place after the peptide has bound the protein. To prove this hypothesis, different D8 concentrations were injected for the same time period (Figure 47 A), followed by injection of a constant D8 concentration (50 µg/ml) (Figure 47 B) for different time periods, and then dissociation curves were compared. Injection of different D8

- 91 -

Results concentrations for the same time period led to identical dissociation curves, whereas longer injection of D8 led to slower dissociation, as expected for a two-state binding mechanism. A

100 50 µg/ml 25 µg/ml 12.5 µg/ml

6.25 µg/ml (Fc2-Fc1)

50 3.125 µg/ml RU

 50 µg/ml

0 Binding Binding response referenced 0 20 40 60 80 100 time [s] B 480 s 100 240 s 120 s

60 s (Fc2-Fc1)

50 30 s

RU 

0 Binding response Binding referenced 0 20 40 60 80 100 time [s]

Figure 47: Proof of two-state binding mechanism for D8-VV1 interaction A: Dissociation curves for different D8 concentrations binding to immobilized VV1 were referenced to the highest binding response (response for 50 µg/mL of D8 was set to 100). As all D8 concentrations were injected for the same time period (120 s), dissociation kinetics were identical. B: To prove the postulated two-state binding model, 50 µg/mL of D8 were injected for different time periods (ranging from 30 s to 480 s) at 10 µl/min. Binding curves were compared. As expected for a two-state reaction binding, longer injection times led to slower dissociation.

4.6.5 Visualization of VV1 OPV interaction by electron microscopy To visualize the binding of VV1 to OPV, immuno-negative staining was performed, and the samples were analyzed with an electron microscope (Figure 48). The experimental setup was performed by Janett Piesker (RKI). First, VACV particles were fixed with 2% paraformaldehyde in 0.05 M HEPES buffer and adsorbed onto a sample support (mesh copper grid coated with plastic film). Then, the samples were incubated with the biotinylated peptide VV1 and VV5 as negative control. This was followed by the incubation with an anti-biotin antibody coupled to 10 nm colloidal gold particles. The samples were then negative stained as previously described. With VACV, only limited binding of VV1 to the viral particles could be observed (data not shown). In specificity ELISA (Figure 41), it could be demonstrated that VV1 bound better to

- 92 -

Results mousepox virus particles (MP-Nü) than to VACV particles, therefore the experiment was repeated using MP-Nü particles. Here, specific binding of peptide VV1 to the viral particles was observed (black dots), while peptide VV5 did not show such binding (Figure 48).

A B

Figure 48: Electron negative contrast image of mousepox virus particles incubated with VV1 (A) and VV5 (B) Mousepox particles were fixed with paraformaldehyde and adsorbed onto a sample support (400 nm copper mesh grid coated with plastic film). The samples were incubated with biotinylated peptide VV1 (A) or VV5 (B) and subsequently with an anti-Biotin Ab coupled to 10 nm colloidal gold particles (black dots) and negatively stained with phosphotungstic acid. Specific interaction of VV1 with the viral particles can be observed (A, black dots), while no interaction of VV5 with the viral particles can be seen (B). Size bar equals 100 nm.

4.7 Usability of peptide VV1 in IFA and Western Blot

Next to PCR and ELISA, immunofluorescence assays (IFA) and Western Blot are, common diagnostic assays. Therefore, a protocol for the application of peptide VV1 in Western Blot and IFA was established during this project. For Western Blot, D8 and different recombinant OPV proteins (A27, A33, B5, F9, F13, H3, and L1) and VACV RIPA lysate were separated on a 4-20% SDS gel and transferred to a nitrocellulose membrane (0.45µM). Additionally, the D8 protein and the VACV RIPA lysate were heat inactivated and applied. The membrane was incubated with the biotinylated peptide and subsequently incubated with SA-POD. After substrate reaction, different protein bands were detectable. In the lanes containing D8 and heat-inactivated D8 (lanes 5 & 6), one distinct band between 25 and 35 kDa could be detected, two bands in the lane containing VACV RIPA lysate (lane 7), and no protein band in all other lanes containing different OPV proteins (Figure 49). The band detected in the lane with heat-inactivated D8 is less intense than the band detected in the lane with native D8. In the lane containing heat-inactivated VACV RIPA lysate, no band was visible. An ELISA with heat inactivated D8 and a two-fold dilution series of the biotinylated peptide resulted in an increase of the signal with rising peptide concentrations, but no characteristic binding curve could be observed. This indicated that VV1 binds to a heat-labile region of D8. The functionality of peptide VV1 could further be shown in a fluorescence assay. CPXV- infected HEp-2 cells were incubated with VV1 pre-incubated with streptavidin fluorescin isothiocyanat (SA-FITC), resulting in a specific fluorescence signal (green) (Figure 51), while non-infected cells showed only a low auto-fluorescence signal (Figure 50).

- 93 -

Results

kDa M ~170--- ~130--- ~100--- ~70---

~55--- ~40--- ~35--- ~25---

~15---

Figure 49: Application of peptide VV1 in western blot for the detection of OPV protein D8 A protein marker (lane 1), different poxviral proteins and VACV RIPA lysate (lanes 2-12: A27, A33, B5, D8, D8 heat-inactivated, VACV RIPA, VACV RIPA heat-inactivated, H3, F9, F13, and L1) were separated on a 4-20% premade SDS gel and subsequently transferred onto a nitrocellulose membrane. The membrane was then incubated with 500 ng/mL of biotinylated peptide VV1. The proteins were detected after several washing steps by incubation with SA-POD (1:50,000) and “SuperSignal West Dura Extended Duration” substrate using the “ChemiDoc MP Imaging” system and Image Lab software (v4.0.1).

Figure 50: Fluorescence image of non-infected HEp-2 cells incubated with peptide VV1 A: Staining of cellular DNA with DAPI (blue), B: staining of non-infected HEp-cells with SA-FITC pre- incubated with peptide VV1 (green), C: transmission light picture of uninfected HEp-2 cells, D: staining with polyclonal anti-D8 DyLight 649-coupled antibody, merge: overlay of A-D. The size bar equals 20 µm.

- 94 -

Results

Figure 51: Fluorescence image of CPXV-infected HEp-2 cells incubated with peptide VV1 A: Cell nucleus stained with DAPI (blue), B: CPXV-infected HEp-2 cells stained with SA-FITC pre- incubated with peptide VV1 (green), C: transmission light picture of infected HEp-2 cells, D: staining with anti-D8 DyLight 649-coupled antibody (red), merge: overlay of A-D. The size bar equals 20 µm.

4.8 Detection of OPV in an aptamer-based sandwich assay

The aim of this thesis was to develop alternative detection molecules to antibodies. In the above paragraphs, the selection of DNA and peptide aptamers against VACV particles has been described. One peptide aptamer capable to detect and capture OPV particles and several DNA aptamers which were able to detect OPV particles in an antibody-based sandwich assay could be successfully identified. These aptamers were now combined in an aptamer-based sandwich assay (Figure 52). Peptide VV1 was immobilized on amine-binding plates, and VACV NY particles were added and then incubated with different biotinylated DNA aptamers. A polyclonal anti-VACV antibody served as positive control. The total signal (light blue bars, A), the background signal (red bars, B) caused by binding of the VACV particles to the plate when no capture peptide was present, and possible interactions of peptide VV1 with the different DNA aptamers (purple bars, C) were assessed. For the DNA aptamers, the total signal varied from 0.2 (JM 1.3) to 0.9 (JM 4.8). However, the total signal was significantly higher when compared to the no-peptide and no-virus control signal for most of the tested DNA aptamers. To estimate a specific detection signal, the difference signals (total – no-peptide – no-virus, dark blue bars, D) were calculated. The difference signals clearly demonstrate that the detection of VACV in an aptamer-based sandwich assay is possible. Furthermore, the combination of VV1 with DNA aptamer JM 4.8 is comparable to the detection of VACV in a sandwich assay with a polyclonal antibody (pAb).

- 95 -

Results

3 2 1.0 A B C D total 0.8 no peptide no virus 0.6 difference

0.4 OD OD (450-620nm)

0.2

0.0 Ab Ab Ab Ab 90.6.1 77.4.8DS 5.7DS 5.8DS 5.9 DS 7.9JM 1.2JM 1.3JM 3.8JM 3.9JM 4.8JM 5.6JM 5.7 90.6.1 77.4.8DS 5.7DS 5.8DS 5.9 DS 7.9JM 1.2JM 1.3JM 3.8JM 3.9JM 4.8JM 5.6JM 5.7 90.6.1 77.4.8DS 5.7DS 5.8DS 5.9 DS 7.9JM 1.2JM 1.3JM 3.8JM 3.9JM 4.8JM 5.6JM 5.7 90.6.1 77.4.8DS 5.7DS 5.8DS 5.9 DS 7.9JM 1.2JM 1.3JM 3.8JM 3.9JM 4.8JM 5.6JM 5.7 90.8.10 DS 5.10 90.8.10 DS 5.10 90.8.10 DS 5.10 90.8.10 DS 5.10

Figure 52: Aptamer-based sandwich ELISA for the detection of OPV Wells of an amine-binding plate were immobilized with peptide VV1 via a free NH2 group. VACV particles were added and incubated with 100 nM of different characterized DNA aptamers. Next to the total signal (light blue bars, A), the background binding of VACV NY (red bars, B) to the plate when no peptide was present, as well as potential interactions of peptide and DNA aptamers (purple bars, C) were assessed. The dark blue bars display the difference signals (total – no peptide- no virus, D). As positive control served a polyclonal anti-VACV antibody (Ab). Bars represent means of duplicates ± SD.

- 96 -

5. Discussion

The detection and identification of clinical infections, emerging pathogens, and bio-threat agents depends on the availability of reliable detection methods. Common detection methods are based on the detection of nucleic acids or on immunological assays using monoclonal and polyclonal antibodies [1]. Nucleic acid-based detection methods, like PCR and microarrays, are usually unable to detect whole particles. With antibodies, the detection of whole particles is possible, but antibodies have some considerable limitations. These limitations could be overcome by aptamers [2, 6, 104]. Therefore, the aim of this thesis was to select DNA and peptide aptamers and to assess their suitability as detection molecules for a model pathogen, in this case OPVs. In more detail, the aptamers were selected against poxviral proteins as well as against native VACV particles to compare different selection strategies. Obtained ligands were analyzed to assess their potential to be used in different diagnostic assays (ELISA, Western Blot, and IFA). In the following sections, the selected aptamers and the two selection methods, SELEX and phage display, are compared. Finally, the potential of aptamers to rival antibodies in diagnostics are discussed.

5.1 Selection and characterization of A27-specific DNA aptamers

Several different SELEX protocols for the selection of nucleic aptamers against various targets have been published in the past. One of these approaches includes the micromagnetic selection. With this approach, DNA aptamers were described that were selected against the light chain of recombinant Botulinum neurotoxin A [168], platelet- derived BB (PDGF-BB) [120], and streptavidin-coated beads [11, 169] with affinities in the nM range. In the present project, magnetic beads were used for the selection of DNA aptamers against rA27, a poxviral surface protein. The complex of rA27 and a random DNA library was captured by Ni-coated magnetic beads via the His-tag of rA27 (compare 3.3.2). Using this approach, nucleotide sequences that do not interact with the target could be washed away easily, while bound nucleotide sequences could be eluted and quantified. The quantity of ssDNA was assessed twice: before incubation with the target and after elution. The corresponding yields were then calculated. In round 9, a yield of over 100 % was calculated (Figure 12). Hence, the enriched DNA library of round 9 was cloned, resulting in the identification of 4 aptamer clones which bound selectively to rA27 with apparent KD values of 24.6 nM to 58.3 nM (assessed with ELONA). This is comparable to previously described DNA aptamers selected against various viral proteins [23, 24, 170]. The selected anti-A27 aptamers could be used successfully as detection molecules for rA27 in an indirect (Figure 13) and in a sandwich ELONA (not shown). The detection limit of the selected anti-A27 aptamers was determined by indirect ELONA and was calculated to be in the range of 7.8-8.4 ng of A27 (MW = 14 kDa) (Figure 15). In comparison, Cho et al. were able to detect 0.92 µg of SARS corona virus nucleocapsid protein with an ssDNA

- 97 -

Discussion aptamer in a Western Blot analysis [24]. Drolet et al. calculated the theoretical minimum detection limit of an anti-VEGF aptamer to be 25 pg/mL, using ELONA [171].

5.2 Optimization of individual SELEX steps

For the selection of DNA aptamers against native VACV particles, a suitable SELEX protocol was established during this project. Factors that influence successful aptamer selection include:  target presentation,  elution and recovery of ssDNA ligands,  PCR amplification,  library complexity,  and the overall experimental protocol. Variations of these factors can result in differences in aptamer affinity and specificity as well as in the selection efficiency [52]. Therefore, individual SELEX steps were optimized thoroughly, and the selection against native VACV particles was performed with two different DNA libraries.

5.2.1 Target presentation, elution and recovery of ssDNA ligands in the selection of VACV-specific DNA aptamers In general, there are two variations for the target molecule presentation. In the first variation, the target molecule and the nucleotide library are incubated in solution; and in the second variation, the target molecule is immobilized on a solid support and is then incubated with the nucleotide library [52]. Fixing a target molecule to a solid support could prevent the conjugation side of the target molecule from interacting with the nucleic acid library [52]. Furthermore, proteins may change conformation when being immobilized on a surface [172], which could result in the loss of the binding epitope. To prevent this, DNA aptamers against VACV were selected in close relation to the sandwich detection assay they should be used in later. For this purpose, a monoclonal anti-A27 Ab was immobilized to capture the native viral particles. This capture antibody displayed another target during selection, which could lead to an accumulation of nonspecific binders. Therefore, a counter- selection step against the capture antibody was performed before every selection round. Furthermore, the obtained DNA ligands should always be tested for binding to the sample matrix used for selection to assess specific binding. Different procedures for the elution and recovery of nucleic acid ligands, dependent on the selection target, are described [24, 173]. The easiest method to elute nucleic acids from their targets, is an incubation step with an elution buffer at high temperatures (70-95 °C), which is a commonly used approach for the selection against proteins [11, 51, 174]. When using the in-solution selection with magnetic beads, the nucleic acids can be simply eluted by heat. In contrast, this is not possible when immobilizing the target to a solid support. Therefore, an appropriate elution method needed to be identified. Different lysis

- 98 -

Discussion buffers were tested with the objective to disrupt the binding of the viral particles to the mAb (compare 3.3.1). Subsequently, the solution was transferred to a microcentrifuge tube and incubated for 1 h at 60°C to inactivate the remaining native VACV particles and to elute the DNA ligands from the viral particles. To recover and concentrate the ssDNA ligands, as well as to reduce the amount of genomic poxviral DNA, two different purification kits were tested. The eluate was analyzed for purity using an Agilent Bioanalyzer and by qRT-PCR to evaluate the amount of genomic poxviral DNA. The RLT buffer proved to be most suitable for the lysis of VACV particles (assessed by qRT-PCR) and the “ssDNA/RNA Clean and Concentrator™” kit to reduce the amount of genomic poxviral DNA significantly, while at the same time maintaining the ssDNA ligands (Figure 16). Therefore, this purification kit and the RLT buffer were used during the selection. Nevertheless, it is possible that not all viral particles were lysed by the RLT buffer. This would lead to the loss of high affine VACV-binding ssDNA ligands that therefore could not be identified by sequencing. With the immobilization of a capture antibody and the elution with a lysis buffer, a novel SELEX protocol was established.

5.2.2 Optimization of amplification of random DNA libraries using the Taguchi approach is fast and efficient The amplification of random DNA libraries is an important step in SELEX. If this step is not controlled properly, the amplification of random DNA libraries can lead to a complete loss of the desired ssDNA ligands [82], and therefore the failure of the selection process is possible [116]. In this project, the Taguchi approach, which had been successfully applied before in the optimization of multifactorial scientific processes (compare 1.3), was used for the optimization of the amplification of different random DNA libraries (see 4.3.1). Three orthogonal arrays (L 9, L 16, and L 18) were compared, coming to the conclusion that the L 9 array in combination with the analysis of highest SN values was most suitable for the optimization of random DNA libraries. The optimization with all arrays tested resulted in significantly increased signal intensities. But the optimization with the L 9 array resulted in the generation of most specific products (Figure 17 & Figure 18). Various protocols for the optimization of PCR amplification can be found in the literature

[175-177]. Most of them deal with the optimization of the compounds MgCl2, dNTPs, primer, Taq polymerase, annealing temperature, and the number of cycles. The situation is even more complicated by the fact that some of the variables are quite interdependent. For example, dNTPs directly chelate with Mg2+ ions, therefore an increase of dNTP concentration reduces the concentration of free Mg2+ ions available to influence polymerase function [175]. Using the Taguchi approach, the interaction of these compounds can be analyzed.

The optimization of the SELEX PCRs in this project resulted in a drastic increase of MgCl2 (by up to 8 mM) and primer concentrations (by up to 1200 nM), while the dNTP and Taq polymerase concentrations did not increase as much. The optimal annealing temperature was set to 64°C.

- 99 -

Discussion

It has been described that for most primer pairs relatively low Mg2+ concentrations (1.5- 2.5 mM with 200 µM dNTPs) and a relatively high annealing temperature (60 °C) provide the most stringent amplification conditions and minimize mispriming [117]. With up to 8 mM, the required amount of MgCl2 is significantly higher than concentrations described in the literature, but, as stated before, the optimal concentration can greatly depend on the primers used. The number of amplification cycles can have a great impact on product specificity, since excessive cycling can lead to the conversion of PCR products into random-length higher molecular weight fragments [178]. For heterogeneous DNA templates, it has been shown that by-products appear as early as in the fifth cycle of PCR [82]. Furthermore, when product formation reached a maximum level, five additional cycles completely converted the product to by-products. For this reason, the optimal number of amplification cycles for the enriched libraries was estimated after each selection round. From the results and the literature discussed above, it can be concluded that with the Taguchi approach the amplification of heterogeneous DNA libraries can be optimized rapidly and with high efficiency. Nevertheless, a good knowledge of the process to be optimized is required.

5.3 The decision of when to end the selection process is challenging

A common procedure to determine the end of the selection process is to calculate the yield of eluted nucleic acid ligands. Radioactive labels are usually used to quantify the bound and unbound fractions of nucleic acid ligands during the SELEX process [11]. This is a sensitive method that enables the detection of low concentrations of nucleic acids. But this method is rather expensive and environmentally incompatible. Fluorescent labels are an alternative to radioactive labels [11]. With this method, the fluorescent label (e.g. fluorescein or Cy5) is introduced over a 5’-modified amplification primer which is then incorporated into the DNA ligands by PCR amplification. Depending on the target, the fluorescence emission of the nucleic acid aptamers is assessed by flow cytometric analysis [35] or by the use of fluorescence microplate readers [11]. The fluorescence intensities can then be used to calculate the amount of bound and eluted aptamers. Another possibility to quantify ssDNA is the use of ultra-sensitive fluorescent nucleic acid stains such as OliGreen® [179]. This reagent is specific for ssDNA and enables the quantification of as little as 100 pg/mL of nucleic acid with a standard spectrofluorometer that uses fluorescein excitation and emission wavelength [180]. The successful application of OliGreen® in the selection of aptamers [181] and for the detection of adenosine with adenosine-binding aptamers has been described [182]. In this project, OliGreen® was used to quantify the ssDNA ligands before incubation with the target and after elution, and the yield of the eluted ssDNA ligands was calculated. For the measurement of the fluorescence intensities different devices were used and compared. The linear detection range of 100 pg/mL to 1 µg/mL (using an 18 nt M13 primer) described by

- 100 -

Discussion the manufacturer (Invitrogen.com, Molecular Probes Handbook, http://de- de.invitrogen.com/site/de/de/home/References/Molecular-Probes-The- Handbook/Nucleic-Acid-Detection-and-Genomics-Technology/Nucleic-Acid-Detection-and- Quantitation-in-Solution.html#head4) could not be confirmed with the two libraries used in this project. A possible explanation for this could be the base selectivity of OliGreen®. The manufacturer describes a fluorescence enhancement for poly(dT) and poly(dG), and low signals for poly(dA) and poly(dC). In random DNA libraries, it is not unlikely that homopolymer sequences are present. Furthermore, the fluorescence signal can be decreased in the presence of sodium acetate, sodium chloride, and magnesium chloride. The combination of the presence of homopolymers in the random DNA libraries, and the use of sodium chloride and magnesium chloride containing SELEX selection buffer, could explain the decreased linear detection range. The calculation of the yield of eluted ssDNA ligands after each selection round with the two libraries resulted in a strong scattering of measuring points (data not shown). This might be explained by the fact that the amount of ssDNA ligands given in the selection rounds varied and was dependent on the amount obtained after the generation of ssDNA ligands. Furthermore, t-RNA was added to the samples before a selection round to minimize unspecific binding of the ssDNA ligands. Moreover, the selection was performed against whole viral particles. Even though the samples were purified after elution, it is possible that traces of t-RNA and genomic poxviral DNA were still present in the samples. The manufacturer described that the OliGreen reagent exhibits fluorescence enhancement when bound to dsDNA and RNA. Rhinn et al. described the successful quantification of cDNA in the presence of RNA with OliGreen using a real-time PCR thermocycler [183]. They found out that at 80 °C the RNA/OliGreen-generated fluorescence signal is undistinguishable from the background level observed in wells containing no RNA. At the same time, a significant level of fluorescence is still emitted at 95 °C when DNA is present. Finally, they found that the influence of RNA on OliGreen-emitted fluorescence is bound to less than 10 % in mixtures of ssDNA and RNA when measured at 80°C. Therefore, the quantification of ssDNA ligands in SELEX may be improved by fluorescence measurements at 80 °C using OliGreen and a real-time PCR thermocycler. An alternative to the quantification of ssDNA ligands is represented by monitoring the enrichment of binding DNA ligands over the course of a selection by binding assays. In this project, such binding assays were performed with the biotinylated enriched 77 mer and 90 mer DNA libraries (Figure 21). No significant binding to the capture antibody over the course of selection could be seen, whereas an increase of binding of the enriched 77 mer and 90 mer libraries to the target between the first and the fifth selection round could be observed. Thereafter, the binding signal for the target decreased for both libraries. For the 90 mer library, another increase of the binding signal could be seen after round nine, accompanied by an increase of the binding signal against the capture antibody. Cloning and sequencing of the enriched libraries of selection rounds 8 to 12 revealed that some DNA

- 101 -

Discussion ligands of the 90 mer library were longer than expected. These sequences contained one to multiple extra copies of either the forward or reverse primer or of both. For cloning, the PCR product was used without prior purification. Therefore, excess primers could be cloned into the bacterial vector along with the desired ssDNA ligands. Furthermore, some sequences contained additional nucleotides of unknown origin. It is possible that these nucleotides were incorporated during the PCR amplification step. Different random ssDNA ligands might have hybridized, which would then result in a template for the DNA polymerase. However, longer ssDNA ligands can form a larger variety of secondary structures than shorter ssDNA ligands. It is therefore possible that these ligands form structural motifs that are able to bind to the capture antibody that was present during the selections. The highest binding signals were detected with the libraries of selection round 5, indicating that these libraries contained more high affine binders than the libraries from the other selection rounds. But in order to be able to draw a conclusion about the affinities of the enriched libraries to the native VACV particles, a serial dilution series of the biotinylated ssDNA ligands should be analyzed. As an alternative to the biotin label an incubation with OliGreen to monitor the enrichment of binding DNA molecules could be performed [181]. This would allow for the analysis of unmodified, unlabeled DNA ligands. However, the quantification of the eluted ssDNA ligands and the binding assays did not allow a decision of when to end the selection process and which enriched libraries of which selection round should be cloned. Therefore, the enriched libraries of several selection rounds were cloned. By this, the clones that occurred in several selection rounds and those that were enriched from round to round could be identified. An alternative to the classical cloning and sequencing approach is HTS. With HTS, the identification of high affine binders is possible as early as after three selection rounds [120].

5.4 HTS can be used to identify high affine DNA aptamers that might not be found with the classical cloning and sequencing approach

The classical cloning and sequencing approach led to the identification of five 90 mer DNA sequences and two 77 mer DNA sequences (Table 20) that were synthesized and characterized. With HTS, additional 90 mer and 77 mer clones were identified (Table 26 & Table 27) and chosen for synthesis and then characterized. Some of these ligands had 3-22- fold lower KD values, and therefore higher binding affinities than the ligands obtained from cloning and sequencing. This is similar to or even better than findings of Cho et al. who identified ssDNA aptamers against PDGF-BB with HTS that had a 3-8-fold higher affinity and a 2-4-fold higher specificity, compared to clones discovered through the conventional cloning and sequencing approach [120]. For the 77 mer library, one clone (JM 1.3) identified by HTS was highly similar to one clone (77.8.8) that was found with cloning and sequencing. The clone JM 1.3 bound specifically to VACV particles, while the clone 77.8.8 did not. These two clones differed only

- 102 -

Discussion by one nucleotide. The effect of single nucleotide changes on binding characteristics of aptamers is discussed in more detail in 5.4.2. Taking together the results of the analysis of the binding characteristics of the DNA aptamers obtained from the conventional cloning and sequencing and the HTS approaches, one can conclude that HTS is a useful approach that can lead to the identification of high affine and selectively binding DNA sequences that might otherwise not be found by using the conventional approach.

5.4.1 DNA aptamers can be used as detection molecules for OPVs, but not as capture molecules In general, three different binding kinetics of the selected DNA aptamers could be observed (Figure 29):  characteristic binding to VACV particles,  binding to the capture Ab,  and no binding. Here, none of the 77 mer sequences bound unspecifically to the capture antibody, while three of the 90 mer sequences did. One possible explanation could be the larger structural variety of longer aptamers.

The estimated KD values of the characterized DNA aptamers for VACV particles were in the range of 1.7-73.6 nM. This is comparable to the binding affinities described for complex target molecules [184]. All OPV species are immunologically cross-reactive and cross-protective [127], therefore it was assessed if the DNA aptamers selected during this project were able to detect not only VACV particles but also other OPV species. All DNA aptamers that bound specifically to VACV particles were also able to detect different OPV species, but not the related PPXV and the non-related viruses (AdV, PPV, and FCV) (Figure 30). The DNA aptamers identified by HTS generated lower background signals than the clones identified with the conventional cloning and sequencing approach, which implies that these clones are more specific. In further experiments, the suitability of the selected DNA aptamers as capture molecules was assessed. For this, the biotinylated DNA aptamers were immobilized on NeutrAvidin- coated plates. VACV or PPXV particles were then added and subsequently incubated with a polyclonal anti-VACV Ab. To assess background binding, the viral particles were applied directly to the wells, without prior immobilization of any DNA aptamers. With all the tested DNA aptamers, an increase of the ELONA signal compared to the background signal could be observed. But when compared to a control DNA aptamer that did not show characteristic binding to VACV (clone DS 8.6), the increase was only marginal (Figure 31). A possible explanation could be that the DNA aptamers were immobilized directly over their 5’ biotin to the NeutrAvidin coated plates. Due to their tetrameric nature, avidins can bind four biotin molecules [185]. Therefore, four biotinylated DNA aptamers would come into close proximity. This might prevent the direct interaction of the DNA aptamer with its target molecule due to steric hindrance that influences nucleic acid folding

- 103 -

Discussion and structure. To still use the tetrameric construct, but to render the DNA aptamers more accessible to their target molecules, dodecyl (12-carbon) or oligodeoxythymidine (oligo(dT)) spacers could be used [186-189]. Lao et al. could show a 100-1000-fold enhancement of microarray sensitivity using an oligo(dT)-modified thrombin-binding aptamer. Furthermore, they could show that aptamer avidity was another effective strategy to enhance the microarray detection. They spotted two distinct oligo(dT)-modified, thrombin-binding aptamers on the same microarray spot, and obtained higher signals when compared to the monovalent binding of either aptamer. Taking together the results from the indirect and the sandwich ELONA of the selected anti-VACV DNA aptamers, it can be concluded that the selected DNA aptamers can be used as detection molecules for OPVs, but are not functional as capture molecules for virus particles. Techniques that might improve the binding characteristics of aptamers are discussed below in more detail (5.9).

5.4.2 Single nucleotide changes can alter the binding characteristics of DNA aptamers significantly The most prominent 90 mer clone (90.6.1) was found 13 times altogether in the enriched libraries of round 6, 8, 9, and 12. The apparent KD value for this clone was calculated to be 27.7 ± 3.8 nM. Several clones which differed only by one nucleotide to this clone were identified. Some of these clones bound to the VACV particles with higher affinities, some bound to the capture antibody, and some did not show any characteristic binding curve, when compared to clone 90.6.1 (Table 28). This leads to the conclusion that single nucleotide changes can significantly alter the binding characteristics of DNA aptamers. Nicol et al. observed that RNA aptamers against the human papilloma virus 16 E7 oncoprotein, with single nucleotide exchanges and sequence similarities of over 97 %, had

KD values of 87, 107, and 251 nM, respectively [23]. They postulated that the aptamers with lower affinities may have arisen late in the selection process by PCR-induced mutation of the original sequence, and that it was possible that these aptamers would have been competed out of the selected pool when the selection had continued beyond their 10 selection rounds. They further postulated that different secondary structures could be the reason for the altered binding affinities, because it is known that small differences in the sequence can have large effects on RNA folding. Potential secondary structures of the ssDNA aptamers obtained in this project were predicted using ValFold [160]. These structures demonstrate that single nucleotide changes in ssDNA ligands can have a great impact on ssDNA folding (Figure 25). The same could be observed for clones that were highly similar to clones 90.8.10 and 77.8.8, respectively (Figure 26 & Figure 28). Since the clones 90.6.7 and 90.6.9 were not found by HTS, it is also possible that these sequences had been generated by sequencing errors. However, it is likely that the secondary structure predictions do not accurately reflect the in-solution structure of the selected and analyzed aptamers. Furthermore, the 5’ biotin label was not taken into account when

- 104 -

Discussion predicting the secondary structures. Further work would be needed in order to address this issue and characterize the aptamers structurally.

5.5 Further applications of DNA aptamers

The successful applicability of the selected anti-VACV DNA aptamers in ELONA could be shown. In further experiments their suitability in detection assays, such as Western Blotting and immunofluorescence assay (IFA), and their potential to inhibit the virus infection of HEp-2 cells was analyzed. The general functionality of nucleic acid aptamers in bioimaging methods [30, 34, 190, 191], their use in blotting assays (Western Blotting, dot blot, and quantum dots) [31, 192-194], and their potential to inhibit the activity of different pathogenic agents [26, 170, 195] has been described in the past.

5.5.1 Usability of DNA aptamers to detect OPV-infected cells by fluorescence labeling The usability of the selected DNA aptamers to fluorescently label CPXV-infected HEp-2 cells was analyzed during this project. For this, the slides with fixed CPXV-infected and uninfected HEp-2 cells were incubated with the biotinylated DNA aptamers and SA-FITC. The slides were analyzed using a confocal laser scanning microscope. None of the tested DNA aptamers was able to generate a specific, distinct fluorescence signal that could be distinguished from the background signal observed for uninfected HEp- 2 cells (data not shown). Avci-Adali et al. demonstrated that dead cells lead to unspecific uptake/binding of ssDNA [196]. It is possible that the fixing and permeabilization of the HEp-2 cells has led to free nucleic acid binding motifs that caused the high background signal. However, the selected DNA aptamers were not functional in classical fluorescence imaging assays.

5.5.2 The selected DNA aptamers do not have any antiviral activity Different protocols to test the antiviral activity of a compound have been described. In this project, a protocol variant where the ssDNA aptamers were preincubated with VACV particles was used, modeled after the publication of Witkowski et al. [197]. In this investigation, the xCELLigence™ RT-CES system was used. This system offers the possibility to non-invasively quantify the cell status of infected and non-infected cells in 96-well E- plates in real time. To evaluate the potential antiviral activity of the selected DNA aptamers, a two-fold dilution series was tested. The VACV particles were preincubated with the DNA aptamers, and the mixture was added to wells containing HEp-2 cells. Additionally, only DNA aptamers and no viral particles were added to wells containing HEp-2 cells to assess the cytotoxicity of these DNA aptamers. No cytopathic effect of the HEp-2 cells using either DNA aptamer could be observed. Nevertheless, the tested DNA aptamers did not show any antiviral activity either (data not shown).

- 105 -

Discussion

There are many possible explanations of the lacking antiviral activity of the selected DNA aptamers. Even though the stability of an unmodified DNA aptamer (selected against inactivated VACV particles) in medium has been shown previously [198], one possible explanation could be the susceptibility of unmodified DNA aptamers to nucleases, and leading to their rapid degradation. Another explanation could be that the DNA aptamers cannot fold correctly in the cell culture medium used, due to too low cation concentrations that are required for proper folding [199]. The indirect ELONA with different OPV proteins led to the assumption that the interaction partner of some of the selected aptamers could be A27. Even though the capability of A27 and other OPV proteins to induce neutralizing antibodies has been described before [158, 159], it is possible that the selected aptamers bind to epitopes that are not involved in virus infection.

5.5.3 Usability of DNA aptamers for the detection of OPV proteins in Western Blot Within this project, a suitable protocol for the detection of proteins with the selected DNA aptamers in Western Blot was established (data not shown). The incubation of a nitrocellulose membrane with 100 nM ssDNA biotinylated aptamer overnight, a subsequent incubation with SA-POD for 1 hr, and the use of the “SuperSignal® West Dura Extended Duration Substrate” proved to be most suitable. This is comparable to findings of Ramos et al. who could successfully detect Leishmania infantum H3 proteins using up to 100 nM of digoxigenin-labeled aptamer populations in immunoblot assays [28]. In this project, the lysate of sugar cushion-purified VACV particles was used for Western Blot analysis. After incubation of nitrocellulose strips with the biotinylated DNA aptamers, up to four bands of different molecular weights (16, 19, 60, and 130 kDa) could be detected. The clear identification of the detected bands only by their molecular weight was not possible. Before the separation on SDS gels, the samples were not denatured nor did the sample buffer contain any reducing agents. Therefore, it might be possible that the bands were multimers or complexes of different proteins. For example, it is known that, dependent on the concentration, A27 can form mono-, di-, tri-, and hexamers. Furthermore, the VACV A27 protein forms stable complexes with A26, tethered to mature virions by association with the A17 transmembrane protein [147, 200]. However, for the clear identification of the detected bands, and therefore an identification of DNA aptamer interaction partners, mass spectrometric analysis could be performed.

5.6 Identification of binding partner of the selected anti-VACV DNA aptamers

For the identification of possible interaction partners of the selected DNA aptamers, an ELONA with different OPV proteins was performed (Figure 32). The results showed that some of the tested DNA aptamers seemed to bind to the IMV protein A27. With 8.09 Mol%, A27 is the most frequent IMV membrane protein [146].

- 106 -

Discussion

For all of the other DNA aptamers, a signal above background level was observed for the virus lysate, but for none of the tested OPV proteins. The selection was performed against native VACV particles captured by a mAb. Therefore, it is possible that the selected DNA aptamers bind not just to one OPV protein, but to a complex of several proteins. For A27, the complex formation with the OPV proteins A25, A26, and A17 has been described [200]. It is possible that the DNA aptamers can only form their correct secondary structure when all interaction partners are present. In protein ELONA, using recombinant OPV proteins, the display of this complexity was not possible. To identify the interaction partner of the DNA aptamers binding to the VACV lysate, but to none of the tested proteins, as well as to confirm the finding of A27 as an interaction partner, a pull-down assay (Figure 33) was performed with subsequent mass spectrometric analysis. The results of the mass spectrometric analysis are still pending.

5.7 Selection of peptide aptamers using phage display

An alternative to DNA aptamers to be used as detection molecules is represented by peptide aptamers. These can be selected by the screening of combinatorial phage display peptide libraries [66, 102]. In this project, libraries of filamentous phage clones carrying short random peptide sequences were screened against different OPV proteins and against native immobilized virus particles. The screening resulted in the identification of recurring OPV-binding phage clones carrying different peptide sequences. Individual phage clones obtained after each biopanning round were assayed by phage ELISA, not only against the respective targets (A27 and VACV) and the blocking agents (BSA and Casein) which are commonly tested as non-binding control [25, 201], but also against another OPV protein (L1), against a related (PPXV), and against two non-related (PPV, AdV) DNA viruses (Figure 35 & Figure 36). This early stringent characterization already demonstrated that some of the enriched phage clones, despite their seemingly specific binding to the target and no binding to BSA and casein, bound to the control protein (phage clone 18 in Figure 35) or the control virus (phage clone 6 in Figure 36). Therefore, mostly peptides of these phage clones for which the phage ELISA resulted in a high signal for the target and a concomitant low signal for the controls (PPV for anti-VACV clones and L1, PPXV, and AdV for anti-A27 clones) were selected for synthesis. These peptides were then synthesized as free soluble peptides that can be evaluated for their use in diagnostic assays. For this purpose, the selected peptides needed to be characterized thoroughly.

5.8 Binding properties of phage-free peptides and phage-attached peptides can differ significantly

While the synthesized peptides selected against different OPV proteins did not bind at all or unspecifically to any protein tested, one out of five synthesized anti-VACV peptides bound specifically to its target. This underlines the fact that phage ELISA is only an estimation of the binding properties of the peptide still attached to the phage. The filamentous phages

- 107 -

Discussion have five different surface proteins [72]. All of them can theoretically interact with other proteins present during phage ELISA, resulting in unwanted signals. The selection of the right phage clone during phage ELISA is even more complicated by the fact that the titers of the amplified individual phage clones cannot be determined so precisely as to add all phage clones at the same number. Moreover, not every identified phage clone expresses exactly five copies of the pIII-fused peptide. An alternative explanation could be that the peptide modifications (acetylation and amidation) chosen for synthesis were not optimal for all peptide clones. Furthermore, it is possible that the phage-free peptides form secondary structures that are different from those of the phage-attached peptides. This and the modifications could lead to a wrong conformation and a loss of binding It would be interesting to analyze what impact the peptide modifications can have on the binding properties of phage display-derived peptides.

5.9 Peptide VV1 can be used for the selective detection of different OPVs

After the successful demonstration of the specific binding of peptide VV1 to native VACV particles (Figure 38 B), its ability to also detect related OPVs was analyzed. VV1 could be utilized not only for the detection of VACV, but also of other OPVs, as demonstrated by electron microscopy (Figure 48) and different ELISA formats (Figure 41 & Figure 42). For this, the peptide worked better when used as a capture molecule than as a detection molecule. This might be explained by the immobilization of the biotinylated peptide to the NeutrAvidin-coated plates. Avidins can bind four biotin molecules, resulting in a construct with four peptide binding sites (compare 5.11) [185] and increasing the functional binding avidity of the peptide. When VV1 was used as a detection molecule, it was only monovalent. To improve the binding affinity of the peptide, it could be used as multimer (this aspect is discussed later (5.11)). It was further interesting to find out which surface protein is targeted by VV1. Using three different approaches (ELISA, SPR/Biacore, and mass spectrometry), it could be demonstrated that the peptide is directed against the VACV envelope protein D8. The D8 protein is highly conserved among OPV. The VACV D8 is to 98 % identical to the CPXV protein, 96 % to CMLV and VARV, 94 % to MPXV, and 93 % to ECTV, as determined by a protein database search using a protein query with the algorithm blastp. It could be shown that VV1 is also suitable for the detection of the clinically relevant CPXV and MPXV. Since even ECTV could be successfully detected by VV1, it is most likely that the peptide could also detect the biothreat-relevant VARV due to the high sequence homology.

5.10 Further applications of peptide aptamer VV1

Analogous to the DNA aptamers selected during this project, the functionality of peptide VV1 in IFA and Western Blot was assessed. Furthermore, the peptides’ potential to inhibit virus infection was analyzed.

- 108 -

Discussion

The VACV envelope protein D8, previously identified as the interaction partner of VV1, binds to chondroitin sulfate on cells and plays a role in virus attachment to the cell surface [149]. Soluble D8 interferes with the binding of IMV to cells and therefore can inhibit VACV infection at the adsorption stage [149]. However, peptide VV1 did not exhibit virus- neutralizing properties tested by a plaque reduction assay (data not shown), indicating a binding site outside the chondroitin sulfate crevice [150]. Furthermore, it was shown by Rodriguez et al. that VACV infectivity in BALB/c mice was reduced in D8 deletion strains [202], while the D8-negative virus replicates efficiently in cultured cells [203]. This suggests that VACV utilizes alternate routes of host cell adhesion and infection, which greatly reduces the chance of aptamer-mediated neutralization by targeting a single VACV envelope protein. Further experiments led to the assumption that VV1 binds to a heat-labile region of D8. To investigate this, wells of a MaxiSorp ELISA plate were coated with heat-denatured D8 and incubated with a two-fold dilution series of the biotinylated D8. With an increased peptide concentration, the ELISA signal rose, but no typical binding curve was observed (data not shown). This is similar to the finding that the peptide bound better to native VACV particles than to heat- and UV-inactivated viral particles (Figure 41 & Figure 42). The results from Western Blot analysis further support the idea of VV1 binding to a heat- labile region of D8. Despite the application of the same amount of recombinant protein (500 ng), the protein band detected in the lane containing heat-denatured D8 (Figure 49, lane 6) was less intense than the protein band detected in lane 5 (native D8). No protein band was detectable in the lane containing the heat-denatured VACV RIPA lysate (applied at 3 µg of total protein), while two intense protein bands were detected in the lane containing native VACV RIPA lysate. The virus lysate contains a number of different proteins of which D8 displays only 1.74 Mol% [146] (equaling around 2,000-4,000 D8 molecules on the virus surface). It could be possible that all D8 molecules in the heat-treated lysate were denatured, while a large fraction of the recombinant D8 was still in its native conformation, enabling VV1 to interact with the protein. The ~140 kDa band detected in the lane with the native virus lysate could be explained by the fact that the full-length D8 ectodomain (AA 1-264) forms tetramers [150]. These tetramers are formed by the noncovalent association of two disulfide-linked dimers, which could explain the ~70 kDa protein band in Western Blot analysis. The recombinant D8 used for ELISA, Western Blot, and Biacore measurements (AA 2-260) was monovalent due to the lack of an unpaired cysteine at position 261 which is critical for multimerization [150]. This would explain why only one ~35 kDa protein band could be detected in the lanes containing the recombinant protein (Figure 49). In further bioimaging experiments, VV1 was successfully applied to fluorescently label CPXV-infected HEp-2 cells (Figure 51). This and the results discussed above indicate that synthetic peptides selected by phage display of random peptide libraries can be used as alternative to antibodies in OPV detection.

- 109 -

Discussion

5.11 Approaches to optimize binding properties of synthetic DNA- and peptide ligands

One possibility to improve the binding properties of synthetic DNA and peptide ligands is their multimerization which results in an increase of their valency [102]. The valency of the selected aptamers could be increased by taking advantage of the biotin modification. This modification was introduced during peptide synthesis at the C-terminus of the peptide and during DNA synthesis at the 5’-end of the oligonucleotide. The C-terminus of the peptide is fused to the phage protein pIII during affinity selection and is therefore not involved in target binding. There are two options to increase the valency of the synthetic biotinylated aptamers. First, multiwell ELISA plates pre-coated with any avidin [185] could be used to immobilize the aptamers. In this case, the aptamers would be used as capture molecules for virus detection. Through the binding of the biotin attached to the aptamers, multiple aptamers would come into close proximity, which could potentially result in an increase of the avidity. This was tested for the DNA and peptide aptamers identified during this project. While the peptide aptamer proved to be a suitable capture molecule (Figure 42) for OPVs, the DNA aptamers were not as potent (Figure 31). The introduction of dodecyl or oligo(dT) spacers into the DNA sequences could improve their functionality [204]. An approach for the successful improvement of DNA aptamer affinity had been shown by Hasegawa et al. [205]. They combined two thrombin-binding aptamers (using flexible poly(dT) linkers) that recognize a different part of the thrombin molecule and achieved a 10-fold increased KD value compared to that of the monomers. Nonaka et al. improved the affinity of a VEGF- binding DNA aptamer by truncation [206]. With this, they aimed to minimize the structure of the aptamer to improve the accessibility of its VEGF-binding domain. Moreover, avidin and the biotinylated aptamers could be pre-incubated together in solution. This would result in a construct with four binding sites due to the tetrameric nature of avidins [185]. This construct could then be used to detect the immobilized virus particles. Experiments with the peptide in fluorescence imaging assays showed its increased binding capability when multimerized. While the incubation of CPXV-infected HEp-2 cells with the monovalent peptide resulted in no specific fluorescence signal (data not shown), the incubation with VV1 pre-incubated with SA-FITC did (Figure 51). This supports the thesis that VV1 binding to D8 could be greatly enhanced by peptide multimerization. Additionally, improvements of the selected DNA and peptide aptamers could be achieved through site-directed mutagenesis, as had been shown by Murase et al. for peptides of 18-20 amino acids [207]. The principle of the in vitro site-directed mutagenesis was derived from observations about antibody-antigen interactions. It could be shown that even small changes of the amino acid residues that form the binding site can alter the strength of an antibody- antigen interaction significantly [208]. To improve the binding properties of the identified DNA and peptide aptamers, random mutations could be introduced. This could either be

- 110 -

Discussion done by error-prone PCR or by synthesis, resulting in the generation of new phage display peptide or DNA libraries. These new libraries could then be subjected to another affinity selection under stringent conditions to identify even more affine phage clones or ssDNA ligands [209].

5.12 DNA and peptide aptamers can be combined in a sandwich assay for OPV detection

Within this project, one peptide aptamer functional as an OPV-specific detection- and capture molecule, and several DNA aptamers functional as OPV-specific detection molecules were identified. In one approach, the suitability of these aptamers in a sandwich assay was assessed (4.7). Here, the peptide aptamer with a free NH2 group at the C-terminus was used as capture molecule, while the biotinylated DNA aptamers were used as detection molecules (Figure 52). The principle of this assay is shown in Figure 53.

A

B C D Peptide with free

NH2 NH2 group Viral particle

Biotinylated DNA aptamer

Streptavidin NH NH peroxidase

Maleic anhydride activated plate

Figure 53: Principle of aptamer-based sandwich detection assay A: Principle of peptide binding to maleic anhydride activated plate (modified from manufacturer manual, Pierce). B, C, and D: Principle of aptamer-based sandwich assay with the peptide as capture molecule and the biotinylated DNA aptamers as detection molecules. The peptide is immobilized onto maleic anhydride activated plates via its free NH2 group. After blocking, the viral particles are added and detected with the biotinylated DNA aptamers (B). To assess background binding of the viral particles to the maleic anhydride activated plates, no peptide was immobilized (C). To further assess unspecific interaction of the DNA aptamers with the peptide, no viral particles were added (D).

To evaluate the functionality of this sandwich assay, the viral particles were added to the wells without prior immobilization of the peptide (Figure 53 C). The resulting signal and the signal obtained from potential interactions of the peptide- and DNA aptamers (Figure 53 D) were then subtracted from the total signal. The functionality of the peptide as capture

- 111 -

Discussion molecule in combination with a pAb has already been shown before (Figure 42). Therefore, this pAb was used as a positive control. When looking at the difference signals, the combination of peptide VV1 with the DNA aptamer clone JM 4.8 led to results comparable to those with the pAb (Figure 52). This sandwich assay showed the proof of principle for aptamer-based OPV detection. To further improve OPV detection with aptamers, this assay can be optimized. The capture ability of peptide VV1 could be enhanced by immobilization of the peptide on NeutrAvidin-coated plates, which would result in a tetrameric structure in comparison to the monomeric structure when using the amine-binding plates. But, to still use the DNA aptamers as detection molecules, another labeling is required. An alternative to biotin could be the direct tagging with HRP or a fluorescent marker, such as FITC. To assure the functionality of the synthetic HRP- or FITC-labeled DNA aptamer as detection molecule, a repetition of the ELONA assays is then required.

5.12.1 Synthetic ligands as surrogate antibodies for pathogen detection? Synthetic peptide and nucleic acid aptamers have several advantages over antibodies. These important advantages have led to the suggestion of using nucleic acid and peptide aptamers as capture and detection elements in [1]. These aptamers have been proposed as alternatives to antibodies, and their potential to replace antibodies in some biotechnical approaches one day is described [5, 6]. Nevertheless, existing limitations of synthetic molecules should be noted.

Only a minority of the described synthetic aptamers bind their targets with a KD of 10 nM or lower, and in the majority of cases the specificity of binding remains to be determined [5]. In comparison, a good antibody has an affinity to its target in the low picomolar range [54].

With appropriate optimization protocols even femtomolar KD values can be achieved [5]. For the specific detection of pathogens, meaning the avoidance of false-positive and false- negative results, specificity and selectivity of the detection molecule used is essential. For example, the ability of synthetic molecules to detect their target in the presence of a large excess of diverse competitors should be addressed [5]. Unfortunately, most of the previous studies have not addressed this issue. The results presented in this project demonstrate that phage-derived synthetic peptides and ssDNA ligands can recognize their targets specifically, even in the presence of non-target compounds. Here, different viruses (PPV, PPXV, AdV, and FCV) were used as a control to test for binding specificity. These viruses were propagated on cell lines, and were subsequently semi-purified by disrupting the cells and by centrifuging down the cell debris. This resulted in virus suspensions that still contained a considerable amount of cell debris. Therefore, it can be concluded that peptide VV1 and the ssDNA aptamers bound specifically VACV particles, but not PPV, PPXV, AdV, and FCV, or the cell debris. Until today, various biopanning experiments have been performed against different pathogens, including viral proteins and viruses, parasites, and bacteria. Specific peptides have been successfully selected against viral envelope proteins of human immunodeficiency

- 112 -

Discussion virus type 1 [210, 211], cucumber mosaic virus [71], and West Nile virus [212]. In contrast, virus particles have been used less frequently in biopanning experiments. Nevertheless, specific peptides have been successfully selected against Newcastle Disease virus [70] and the avian influenza virus H9N2 [213]. All of the above-mentioned selections against viral proteins or viruses, except for the selection against the cucumber mosaic virus, were performed to identify peptides with antiviral activity. The phage display methodology is predominantly used for the identification of molecules targeting different biological structures. Thus, there are relatively few applications of phage display in the detection area [66]. Nevertheless, peptides for the detection of cucumber mosaic virus [71], Bacillus spores [68, 69, 214], and of staphylococcal enterotoxin B [215] have been successfully selected. While numerous biopanning experiments selecting nucleic acid aptamers against viral proteins have been performed (reviewed in Ref. [216]), there are only few publications in which the target is the whole virus particle. These include an RNA aptamer selected against Rous sarcoma virus [32], a DNA aptamer selected against vaccinia virus [198], and DNA aptamers selected against avian influenza virus H5N1 [217]. Furthermore, Labib et al. recently selected a DNA aptamer against viable VACV particles which was implemented in an aptamer-based viability impedimetric sensor for viruses which is able to distinguish between viable and non-viable VACV [33]. With this and the results discussed above, it can be concluded that aptamers could indeed rival antibodies in OPV detection in individual cases.

5.12.2 Implementing synthetic DNA and peptide aptamers for the detection of pathogens The essential element of a real-time, mobile detector is a probe that binds spore, bacterium, toxin, or virus, and as a part of an analytical platform generates a measureable signal [218]. The choice of the detection probes is dictated by their specificity, selectivity, performance, storage, and operational and environmental stability. In most platforms, these probes are monoclonal or polyclonal antibodies. Antibodies are well-established detection molecules with excellent binding properties. However, polyclonal Abs are not selective, because they recognize all antigens to which the immunized animal has been exposed in the past. Monoclonal Abs are more selective, but their application is hindered by the high production costs and the inherent sensitivity to unfavorable environmental conditions [218]. Both types of antibodies are relatively unstable and require refrigeration [219]. Synthetic molecules such as DNA and peptide aptamers, on the other hand, are chemically very stable. While the suitability of the implementation of synthetic peptides into various diagnostic methods and platforms remains to be demonstrated, various nucleic acid-based assays for the detection of protein have been described. These include assays:  for the detection of cytokines and growth factors by fluorescence anisotropy [220],  to detect protein molecular variants by fluorescence quenching [221],

- 113 -

Discussion

 for electrochemical detection of picomolar levels of platelet-derived growth factor in serum [222],  for the quantification of thrombin using a fluorescence sensor [223],  to monitor clinical samples for illegal- or prescription drugs [224, 225] using a “dipstick” format. Furthermore, aptamers for the detection of biological threat agents (such as bacterial cells, spores, viruses, and toxins) and foodborne pathogens and their application in different detection assays have also been described (reviewed by Tombelli et al. [226]). Kirby et al. described the adaptation of an anti-ricin RNA aptamer in a chip-based microsphere array [227]. The system was composed of a flow cell connected to a fast performance liquid chromatography pump, and a fluorescence microscope for observation. The flow cells contained silicon chips with multiple wells in which beads modified with the sensor elements were deposited. They then used commercially available SA-beads and modified them with biotinylated anti-ricin RNA aptamers. These were used to demonstrate the possibility of quantifying the fluorescence labeled protein. One promising platform that allows the rapid detection of the analyte and can be used in the field is based on an immunoaffinity column assay. This assay has been patented and commercialized under the name Abicap® by Senova GmbH (Weimar, Germany) and has already been successfully applied for the rapid detection of botulinum neurotoxin type A [228] and the Ebola virus [229]. This immunoaffinity assay is based on columns prefilled with filters on which a pathogen-specific antibody is immobilized. The sample containing the pathogen is passed over the column by gravitation, where the analyte is trapped and detected by incubation with a specific biotinylated antibody, followed by the substrate reaction [228]. The signal produced in the column is directly quantified with a small dedicated photometer. In this system, the capture and/or the detection antibody could possibly be substituted by peptide and/or DNA aptamers, allowing for simpler storage conditions of the assay. For nucleic acid aptamers, surface plasmon resonance, quartz crystal microbalances, and surface acoustic wave technique could also be applied as aptamer-based biosensing devices [230]. With these systems, high specificities and low limits of detection can be achieved. However, they might not be applicable for the simultaneous detection of multiple analytes. Another promising application of the synthetic peptide and DNA aptamers is the implementation in peptide and DNA microarrays. Here, different solid supports for microarray production could be used, including glass or plastic microscope slides, nitrocellulose membranes, or gold surfaces [53, 231]. The advantage of such microarrays is the fact that several aptamers with specificities against different pathogens can be spotted onto the solid support. This allows for a rapid multiplex detection of different pathogens. To address this issue, Taitt et al. described microarrays with naturally occurring antimicrobial peptides for the detection of inactivated biothreat agents [232]. Here, the peptides were utilized to capture the pathogen, while antibodies were used as detection molecules.

- 114 -

Discussion

5.13 Comparison of SELEX and phage display methodology

In this project, SELEX and phage display were used for the selection of DNA and peptide aptamers as specific detection molecules for OPVs. Both methods are cyclic repetitive processes with the aim to reduce the number of combinatorial molecules in a randomly created library to a manageable amount by screening against a desired target [3, 6]. For both methods, the initial library is composed of DNA. While this DNA library can be directly applied in SELEX, a random peptide phage display library needs to be generated for phage display. Today, different phage display libraries are commercially available. Some important parameters of SELEX and phage display are compared in Table 36. With up to 1015 different ssDNA sequences [14], the initial library in SELEX has a much higher complexity than the phage display library which contains up to 1010 phage clones [102], displaying different peptide sequences. This increases the chance to select a specific ligand for the desired target with SELEX. While a constant region for amplification is required in SELEX, in addition to the randomized region, the peptide sequences are amplified by infection of bacterial host cells with the corresponding phages. In both methods, high affine binders can be lost if the amplification conditions are unfavorable for this clone. Peptide sequences are selected while still attached to the phage surface and need to be synthesized to assess the binding potential of the free soluble peptide. In contrast, the ssDNA ligands obtained after selection can be directly characterized.

Table 36: Comparison of important parameters of SELEX and phage display SELEX Phage Display Starting material Library of up to 1015 different Library with up to 109 ssDNA sequences different phage clones Length of random library 30-54 nt (67-90 nt) 7-12 aa (11-16 aa) (total) Costs of library Low (approx. 100 €) Medium (approx. 350 €) Effort Low-medium Medium-high Amplification PCR Infection of host cell Duration of one selection 1-1.5 days 3 days round Rounds needed 8-20 3-6 Expertise Necessary Necessary Time needed for synthesis Up to 2 weeks 3 weeks to 3 months Costs per synthetic ligand 50-100 € 270-600 € Time needed for complete Approx. 6-8 weeks 2.5-4.5 months selection (12 rounds) (4 rounds) Total costs Low-medium Medium-high nt: nucleotides, aa: amino acids

One round of SELEX can be accomplished within 1-1.5 working days with low to medium effort, while one phage display round needs 3 days to complete. This is due to the amplification and titration steps for which bacterial host cells are infected [72]. The number of rounds needed to select specific ligands depends on the complexity of the target, diversity

- 115 -

Discussion of the library, affinity towards the target molecule, and stringency of selection [91]. Typically, the number of rounds needed varies from 8-20 selection rounds for SELEX, while only 3-6 selection rounds are needed for phage display. Therefore, the SELEX process can be finished in 8-30 working days, while a phage display selection can be completed in 9-18 working days. Considerable differences can be observed in the synthesis of the selected ligands. While the synthesis and delivery of ssDNA sequences can vary from 2 days to 2 weeks, synthesis of peptides takes at least 3 weeks, or even up to 3 months, depending on the supplier. Furthermore, the costs for the synthesis of the synthetic ligands can differ greatly. While ssDNA sequences cost 50-100 €/sequence, one peptide can cost 270-600 €. Due to the long time needed for synthesis, the total time from the first selection round to the synthesized peptide ligand can be as long as 4.5 months, while only 6-8 weeks are needed to obtain the ssDNA ligands. The selectivity and affinity of the synthesized ligands still needs to be determined, but when using identical screening assays, the same time frame will be required for both types of aptamers. In general, the SELEX procedure is faster, cheaper, and requires less effort than phage display. But for both methods, an experienced scientist is necessary, and there is no guarantee that ligands are obtained at the end of the experiment that bind specifically and with high affinity to the target.

5.14 Conclusions and perspective

Although still in its infancy, the identification and optimization of synthetic peptide and DNA aptamers might provide a source of detection molecules that are easily accessible and more ethical than the hybridoma technology. The most important benefits of these synthetic molecules lie in their chemical stability, the option of synthesis under controlled conditions, and no need for animal experiments. Phage display and SELEX represent two suitable methods for the selection of aptamers. Nevertheless, it should be noted that appropriate protocols need to be established, and that single selection steps need to be optimized thoroughly. For the phage display methodology, it can be concluded that the selection against viral particles is more fruitful than the selection against viral proteins, since peptides selected against viral proteins, that showed affinity to their protein target, bound also unspecifically to the protein controls. For the SELEX methodology, DNA libraries with different length and two different protocols were compared. Therefore, it can be concluded that:  shorter DNA libraries can lead to ligands with higher specificity when compared to longer libraries,  in-solution selection as well as surface panning can lead to specific ligands,  and the decision of when to end the selection process can be very challenging.

- 116 -

Discussion

In this project, it could be shown that this problem can be addressed by the cloning and sequencing of multiple enriched selection rounds as well as the application of high throughput sequencing (HTS). With HTS, a higher number of clones could be analyzed in parallel, and sequences that were not found with the classical cloning and sequencing approach could be identified. Furthermore, the enrichment of individual oligonucleotide sequences could be observed. At last, the functionality of the selected DNA and peptide aptamers for the detection of OPVs in a sandwich assay could be shown. The optimization of this assay can result in the development of a method for OPV detection, utilizing only synthetic DNA and peptide aptamers, which is highly desirable.

- 117 -

6. Abbreviations

µg Microgram µl Microliter µM Micromolar Abs Antibodies AdV Adenovirus ATCC American Type Culture Collection bp(s) Base pair(s) BSA Bovine serum albumin CC Cell culture supernatant CMC Carboxymethyl cellulose CMLV Camelpox virus CMLV CP-19 Camelpox virus strain CP-19 CPXV Cowpox virus CPXV BR Cowpox virus strain Brighton Red CPXV GuWi Cowpox virus strain GuWi CRFK Continuous cell line Da Dalton DNA Deoxyribonucleic acid dNTP Deoxyribonucleotide triphosphate ds double-stranded E. coli Escherichia coli ECACC European Collection of Cell Cultures EDTA Ethylenediaminetetraacetic acid EEV(s) Extracellular enveloped virus(es) ELISA Enzyme-linked immunosorbent assay ELONA Enzyme-linked oligonucleotide assay EtBr Ethidium bromide EV(s) Extracellular virions FCS Fetal calf serum FCV Feline calicivirus fmol Femtomol g Earth’s gravitational acceleration HeLa Continuous cell line HEp-2 Continuous cell line hr Hour(s) HRP Horse radish peroxidase HTS High-throughput sequencing IFA Immunofluorescence assay IgG Immunoglobulin G IMV(s) Intracellular mature virus(es) IPTG Isopropyl-β-D-thiogalactopyranoside kbp Kilo Base pairs KD Equilibrium dissociation constant kDa Kilo Dalton L Liter LB Luria Bertani LC Liquid chromatograph LOD Limit of detection M Molar mAb Monoclonal antibody min Minute(s) mL Milliliter

- 118 -

Abbreviations mM Millimolar MOI Multiplicity of infection MP-Nü Ectromelia virus (Mousepox virus) isolate MV(s) Mature virus(es) MVA Modified Vaccinia virus Ankara ng Nanogram nm Nanometer nt Nucleotide(s) OD450 Optical density at 450nm OPV(s) Orthopox virus(es) pAb Polyclonal antibody PBS Phosphate buffered saline PBS-T Phosphate buffered saline containing Tween 20 PCR Polymerase chain reaction pfu Plaque forming units PK13 Continuous cell line pmol Picomol PPV Porcine parvovirus PPXV Parapoxvirus Recombinant poxviral protein A27 expressed in rA27 Baculovirus REF A1 (rat embryonic fibroblasts) Continuous cell line RKI Robert Koch Institute RNA Ribonucleic acid rpm Revolutions per minute RT Room temperature SA-chip Streptavidin-coated sensor chip SA-FITC Streptavidin conjugated to fluorescein isothiocyanate SA-POD Streptavidin conjugated to horse radish peroxidase SC Sugar cushion sec Second(s) Systematic Evolution of Ligands by EXponential SELEX enrichment SPR Surface plasmon resonance ss single-stranded TBS Tris buffered saline TBS-T Tris buffered saline containing Tween 20 TMB 3,3’5,5’-Tetramethylbenzidine U Unit (of enzyme activity) v/v Volume per volume VACV Vaccinia virus VACV 1536 Vaccinia virus strain New York City Board of Health VACV IHD-W Vaccinia virus strain IHD-W VACV WR Vaccinia virus strain Western Reserve Vero76 Continuous cell line w/v Weight per volume X-Gal 5-Bromo-4-chloro-3-indolyl-β-D-galcatopyranoside

- 119 -

7. Figures

Figure 1: Examples of different structural motifs of nucleic acid aptamers ...... - 3 - Figure 2: Schematic representation of the SELEX procedure ...... - 5 - Figure 3: Schematic representation of an M13 phage particle ...... - 8 - Figure 4: Random peptide phage display libraries ...... - 8 - Figure 5: Principle of a biopanning procedure ...... - 9 - Figure 6: Schematic virion structure of an OPV particle ...... - 14 - Figure 7: Premade Ph.D. random peptide phage display library ...... - 36 - Figure 8: Sequence of random peptide library-gIII fusion ...... - 39 - Figure 9: Structure of sensor chip (left) and principle of surface plasmon resonance (SPR) detection (right) ...... - 43 - Figure 10: Sensogram showing the response from the SPR detector ...... - 44 - Figure 11: Principle of pull-down for the identification of interaction partners of aptamers ...... - 45 - Figure 12: Yield of eluted aptamers after each selection round (A27 selection) ...... - 48 - Figure 13: Binding assay with synthetic DNA aptamers (A27 selection) ...... - 49 - Figure 14: Binding responses and curve fit to calculate the steady-state affinities for DNA aptamer - A27 interactions ...... - 49 - Figure 15: Titration curves for calculation of lower limit of detection (LOD) of rA27, determined with anti- A27 aptamers using indirect ELONA ...... - 50 - Figure 16: Analysis of lysed VACV samples spiked with ssDNA library after purification with kit from Macherey-Nagel (A, C) and from Zymo Research (B) ...... - 51 - Figure 17: Gel analysis of PCR products of the confirmation PCR for the 90 mer library ...... - 53 - Figure 18: Gel analysis of PCR products of the confirmation PCR 1 for the 67 mer (A) and 77 mer (B) libraries ...... - 54 - Figure 19: Gel analysis of PCR products of the confirmation PCR 2 for the 67 mer (A) and 77 mer (B) libraries ...... - 55 - Figure 20: Gel analysis of PCR products of the confirmation PCR 2 for the 90 mer library ...... - 55 - Figure 21: Binding assay of 77 mer and 90 mer libraries after 12 selection rounds ...... - 56 - Figure 22: Percentage of selected top ten sequences from the 90 mer library among the eight selection rounds ...... - 64 - Figure 23: Percentage of selected top ten sequences from the 77 mer library among the eight selection rounds ...... - 65 - Figure 24: Titration of DNA aptamers by sandwich ELONA (clones obtained from the classical cloning and sequencing approach) ...... - 67 - Figure 25: Potential secondary structures of ssDNA aptamers with similarity to clone 90.6.1, predicted with ValFold and visualized with Pseudoviewer 3.0 ...... - 69 - Figure 26: Potential secondary structures of ssDNA aptamers with similarity to clone 90.8.10, predicted with ValFold and visualized with Pseudoviewer 3.0 ...... - 69 - Figure 27: Potential secondary structures of remaining 90mer aptamers, predicted with ValFold and visualized with Pseudoviewer 3.0 ...... - 70 - Figure 28: Potential secondary structures of ssDNA aptamers (77 mer), predicted with ValFold and visualized with Pseudoviewer 3.0 ...... - 70 - Figure 29: Titration of DNA aptamers by sandwich ELONA (clones obtained from HTS data analysis) . - 71 - Figure 30: Specificity testing of VACV selected DNA aptamers as detection molecules ...... - 72 - Figure 31: Specificity testing of DNA aptamers as capture molecules on NeutrAvidin plates ...... - 73 - Figure 32: Identification of binding partner for DNA aptamers with protein ELONA ...... - 74 -

- 120 -

Equations

Figure 33: SDS-PAGE of eluted lysate components captured with DNA aptamers ...... - 75 - Figure 34: Titers of phage eluates determined after each selection round ...... - 76 - Figure 35: Anti-A27 phage clones tested for specificity in phage ELISA ...... - 80 - Figure 36: Anti-VACV phage clones tested for specificity in phage ELISA ...... - 81 - Figure 37: Affinity testing of anti-A27 peptides by peptide ELISA ...... - 84 - Figure 38: Titration of anti-VACV peptides by peptide ELISA ...... - 85 - Figure 39: Titration of anti-L1 (A) and anti-B5 (B) peptides by peptide ELISA ...... - 86 - Figure 40: Principle of the two detection systems used, with the peptide as detection molecule (A) and as capture molecule (B) ...... - 86 - Figure 41: Specificity testing of VACV-selected peptide VV1 as detection molecule...... - 87 - Figure 42: Specificity testing of peptide VV1 as capture molecule on NeutrAvidin plates ...... - 88 - Figure 43: Identification of binding partner for VV1 via indirect ELISA...... - 89 - Figure 44: Identification of binding partner for VV1 via pull-down assay, SDS-PAGE and subsequent nLC- MS/MS analysis ...... - 90 - Figure 45: Binding response and curve fit for D8 binding to peptide VV1 ...... - 91 - Figure 46: Binding analysis to assess the specificity of VV1 peptide binding to D8 via SPR/Biacore ...... - 91 - Figure 47: Proof of two-state binding mechanism for D8-VV1 interaction ...... - 92 - Figure 48: Electron negative contrast image of mousepox virus particles incubated with VV1 (A) and VV5 (B) ...... - 93 - Figure 49: Application of peptide VV1 in western blot for the detection of OPV protein D8 ...... - 94 - Figure 50: Fluorescence image of non-infected HEp-2 cells incubated with peptide VV1 ...... - 94 - Figure 51: Fluorescence image of CPXV-infected HEp-2 cells incubated with peptide VV1 ...... - 95 - Figure 52: Aptamer-based sandwich ELISA for the detection of OPV...... - 96 - Figure 53: Principle of aptamer-based sandwich detection assay ...... - 111 -

8. Equations

Equation 1: Taguchi’s quadratic loss function higher is better ...... - 28 - Equation 2: Calculation of phage titer ...... - 37 -

Equation 3: Example for calculation of the SN value for the lowest MgCl2 level ...... - 52 -

- 121 -

9. Tables

Table 1: Overview of different SELEX processes (modified from [3]) ...... - 7 - Table 2: Orthogonal arrays for 4 (L9), 5 (L16), or 7 (L18) factors at three (L9 and L18) or four (L16) levels ...... - 11 - Table 3: Overview of NGS platforms available at the Robert Koch Institute...... - 12 - Table 4: Overview of different VACV membrane proteins ...... - 15 - Table 5: Oligonucleotides used for real-time PCR ...... - 26 - Table 6: Components and cycling conditions for real-time PCR ...... - 26 - Table 7: Used primer sets for amplification of aptamers ...... - 27 - Table 8: Cycling conditions for PCR optimization with Taguchi arrays ...... - 27 - Table 9: Orthogonal arrays with optimized factors (A-H) and their concentration levels (1-4) ...... - 28 - Table 10: Components for denaturation and renaturation reaction for selection against A27 (selection round 1) ...... - 29 - Table 11: Components and cycling conditions for PCR1 (selection against A27) ...... - 29 - Table 12: Components for denaturation and renaturation reaction for selection against native VACV particles (selection round 1) ...... - 30 - Table 13: Components, cycling conditions, and DNA amounts for a sequencing reaction ...... - 33 - Table 14: Amount used of amplified ligands obtained in PCR 1 and adapter indices used for NGS ...... - 34 - Table 15: Synthetic DNA aptamer sequences derived from SELEX against rA27 ...... - 48 - Table 16: Mean yields for each reaction component at each level tested (example for L9 array) ...... - 52 - Table 17: Signal-to-noise (SN) ratios for reaction components calculated for the reaction components of the L9 array using the mean of the yields ...... - 53 - Table 18: Optimal and original reaction conditions for PCR 1 for all three libraries as analyzed by Taguchi optimization ...... - 54 - Table 19: Optimal and original reaction conditions for PCR 2 for all three libraries as analyzed by Taguchi optimization ...... - 55 - Table 20: Synthetic DNA aptamer sequences derived from SELEX against VACV particles (cloning) ...... - 58 - Table 21: Number of reads obtained from sequencing with the Illumina Genome Analyzer II and distribution of obtained sequences among 90 mer and 77 mer libraries ...... - 59 - Table 22: Overview of sequence counts of 90 mer library after HTS data analysis ...... - 59 - Table 23: Overview of sequence counts of 77 mer library after HTS data analysis ...... - 60 - Table 24: Counts and percentage of top 10 sequences obtained after analysis of HTS data of 90 mer library ...... - 61 - Table 25: Counts and percentage of top 10 sequences obtained after analysis of HTS data of 77 mer library ...... - 62 - Table 26: Synthetic DNA aptamer sequences derived from SELEX against VACV particles with 90 mer library (HTS) ...... - 65 - Table 27: Synthetic DNA aptamer sequences derived from SELEX against VACV particles with 77 mer library (HTS) ...... - 66 - Table 28: Characterization of sequences similar to 90.6.1 ...... - 68 - Table 29: Characterization of sequences similar to 90.8.10 ...... - 68 - Table 30: Characterization of remaining 90 mer sequences ...... - 68 - Table 31: Characterization of 77 mer sequences ...... - 72 - Table 32: Consensus peptide sequences of A27-selected phage clones ...... - 77 -

- 122 -

Tables

Table 33: Consensus peptide sequences of VACV-selected phage clones (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.) ...... - 78 - Table 34: Synthetic anti-A27 peptides derived from phage display ...... - 82 - Table 35: Synthetic anti-VACV peptides derived from phage display (These data were generated in cooperation with Lilija Miller and can also be found in her doctoral thesis [98]. The data are shown here as well to be able to follow the line of argument.) ...... - 82 - Table 36: Comparison of important parameters of SELEX and phage display...... - 115 -

- 123 -

10. References

1. Lim, D.V., et al., Current and developing technologies for monitoring agents of bioterrorism and biowarfare. Clin Microbiol Rev, 2005. 18(4): p. 583-607. 2. Iqbal, S.S., et al., A review of molecular recognition technologies for detection of biological threat agents. Biosens Bioelectron, 2000. 15(11-12): p. 549-78. 3. Mascini, M., I. Palchetti, and S. Tombelli, Nucleic acid and peptide aptamers: fundamentals and bioanalytical aspects. Angew Chem Int Ed Engl, 2012. 51(6): p. 1316-32. 4. Lipman, N.S., et al., Monoclonal versus polyclonal antibodies: distinguishing characteristics, applications, and information resources. ILAR J, 2005. 46(3): p. 258-68. 5. Kodadek, T., et al., Synthetic molecules as antibody replacements. Acc Chem Res, 2004. 37(9): p. 711-8. 6. Ruigrok, V.J., et al., Alternative affinity tools: more attractive than antibodies? Biochem J, 2011. 436(1): p. 1-13. 7. Ngundi, M.M., et al., Nonantibody-based recognition: alternative molecules for detection of pathogens. Expert Rev Proteomics, 2006. 3(5): p. 511-24. 8. Conidi, A., V. van den Berghe, and D. Huylebroeck, Aptamers and Their Potential to Selectively Target Aspects of EGF, Wnt/beta-Catenin and TGFbeta-Smad Family Signaling. Int J Mol Sci, 2013. 14(4): p. 6690-719. 9. Lopez-Colon, D., et al., Aptamers: turning the spotlight on cells. Wiley Interdiscip Rev Nanomed Nanobiotechnol, 2011. 3(3): p. 328-40. 10. Khati, M., The future of aptamers in medicine. J Clin Pathol, 2010. 63(6): p. 480-7. 11. Stoltenburg, R., C. Reinemann, and B. Strehlitz, FluMag-SELEX as an advantageous method for DNA aptamer selection. Anal Bioanal Chem, 2005. 383(1): p. 83-91. 12. Zhang, Z., M. Blank, and H.J. Schluesener, Nucleic acid aptamers in human . Arch Immunol Ther Exp (Warsz), 2004. 52(5): p. 307-15. 13. Jayasena, S.D., Aptamers: An Emerging Class of Molecules That Rival Antibodies in Diagnostics. Clin Chem, 1999. 45(9): p. 1628-1650. 14. Stoltenburg, R., C. Reinemann, and B. Strehlitz, SELEX--a (r)evolutionary method to generate high-affinity nucleic acid ligands. Biomol Eng, 2007. 24(4): p. 381-403. 15. Ellington, A.D. and J.W. Szostak, In vitro selection of RNA molecules that bind specific ligands. Nature, 1990. 346(6287): p. 818-822. 16. Tuerk, C. and L. Gold, Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase. Science, 1990. 249(4968): p. 505-10. 17. Connell, G.J., M. Illangesekare, and M. Yarus, Three small ribooligonucleotides with specific arginine sites. Biochemistry, 1993. 32(21): p. 5497-502. 18. Bock, L.C., et al., Selection of single-stranded DNA molecules that bind and inhibit human thrombin. Nature, 1992. 355(6360): p. 564-566. 19. Ciesiolka, J., J. Gorski, and M. Yarus, Selection of an RNA domain that binds Zn2+. RNA, 1995. 1(5): p. 538-50. 20. Hofmann, H.P., et al., Ni2+-binding RNA motifs with an asymmetric purine-rich internal loop and a G-A base pair. RNA, 1997. 3(11): p. 1289-300. 21. Tuerk, C. and S. MacDougal-Waugh, In vitro evolution of functional nucleic acids: high-affinity RNA ligands of HIV-1 proteins. Gene, 1993. 137(1): p. 33-9. 22. Xu, W. and A.D. Ellington, Anti-peptide aptamers recognize amino acid sequence and bind a protein epitope. Proc Natl Acad Sci U S A, 1996. 93(15): p. 7475-80. 23. Nicol, C., et al., Effects of single nucleotide changes on the binding and activity of RNA aptamers to human papillomavirus 16 E7 oncoprotein. Biochem Biophys Res Commun, 2011. 405(3): p. 417-21. 24. Cho, S.J., et al., Novel system for detecting SARS coronavirus nucleocapsid protein using an ssDNA aptamer. J Biosci Bioeng, 2011. 25. Choi, J.S., et al., Screening and characterization of high-affinity ssDNA aptamers against anthrax protective antigen. J Biomol Screen, 2011. 16(2): p. 266-71.

- 124 -

References

26. Chang, T.W., et al., In vitro selection of RNA aptamers that inhibit the activity of type A botulinum neurotoxin. Biochem Biophys Res Commun, 2010. 396(4): p. 854-60. 27. Hesselberth, J.R., et al., In vitro selection of RNA molecules that inhibit the activity of ricin A- chain. J Biol Chem, 2000. 275(7): p. 4937-42. 28. Ramos, E., et al., In vitro selection of Leishmania infantum H3-binding ssDNA aptamers. Oligonucleotides, 2010. 20(4): p. 207-13. 29. Göringer, H.U., M. Homann, and M. Lorger, In vitro selection of high-affinity nucleic acid ligands to parasite target molecules. International Journal for Parasitology, 2003. 33(12): p. 1309-1317. 30. Cao, X., et al., Combining use of a panel of ssDNA aptamers in the detection of Staphylococcus aureus. Nucleic Acids Res, 2009. 37(14): p. 4621-8. 31. Joshi, R., et al., Selection, characterization, and application of DNA aptamers for the capture and detection of Salmonella enterica serovars. Mol Cell Probes, 2009. 23(1): p. 20-8. 32. Pan, W., et al., Isolation of virus-neutralizing from a large pool of random sequences. Proc Natl Acad Sci U S A, 1995. 92(25): p. 11509-13. 33. Labib, M., et al., Aptamer-based viability impedimetric sensor for viruses. Anal Chem, 2012. 84(4): p. 1813-6. 34. Shangguan, D., et al., Identification of liver cancer-specific aptamers using whole live cells. Anal Chem, 2008. 80(3): p. 721-8. 35. Tang, Z., et al., Generating aptamers for recognition of virus-infected cells. Clin Chem, 2009. 55(4): p. 813-22. 36. Fischer, N.O., T.M. Tarasow, and J.B. Tok, Aptasensors for biosecurity applications. Curr Opin Chem Biol, 2007. 11(3): p. 316-28. 37. Bunka, D.H., O. Platonova, and P.G. Stockley, Development of aptamer therapeutics. Curr Opin Pharmacol, 2010. 10(5): p. 557-62. 38. Keefe, A.D., S. Pai, and A. Ellington, Aptamers as therapeutics. Nat Rev Drug Discov, 2010. 9(7): p. 537-50. 39. Shangguan, D., et al., Aptamers evolved from live cells as effective molecular probes for cancer study. Proc Natl Acad Sci U S A, 2006. 103(32): p. 11838-43. 40. Nimjee, S.M., C.P. Rusconi, and B.A. Sullenger, Aptamers: an emerging class of therapeutics. Annu Rev Med, 2005. 56: p. 555-83. 41. Ferreira, C.S., et al., DNA aptamers against the MUC1 tumour marker: design of aptamer- antibody sandwich ELISA for the early diagnosis of epithelial tumours. Anal Bioanal Chem, 2008. 390(4): p. 1039-50. 42. Hu, M. and K. Zhang, The application of aptamers in cancer research: an up-to-date review. Future Oncol, 2013. 9(3): p. 369-76. 43. Dua, P., S. Kim, and D.K. Lee, Nucleic acid aptamers targeting cell-surface proteins. Methods, 2011. 54(2): p. 215-25. 44. Di Giusto, D.A., et al., Proximity extension of circular DNA aptamers with real-time protein detection. Nucleic Acids Res, 2005. 33(6): p. e64. 45. Labib, M., et al., Electrochemical differentiation of epitope-specific aptamers. Anal Chem, 2012. 84(5): p. 2548-56. 46. Ikebukuro, K., C. Kiyohara, and K. Sode, Novel electrochemical sensor system for protein using the aptamers in sandwich manner. Biosensors and Bioelectronics, 2005. 20(10): p. 2168- 2172. 47. Wang, X.L., et al., Ultrasensitive detection of protein using an aptamer-based exonuclease protection assay. Anal Chem, 2004. 76(19): p. 5605-10. 48. Yan, X., et al., DNA aptamer folding on magnetic beads for sequential detection of adenosine and cocaine by substrate-resolved chemiluminescence technology. Analyst, 2010. 135(9): p. 2400-7. 49. Michaud, M., et al., A DNA Aptamer as a New Target-Specific Chiral Selector for HPLC. Journal of the American Chemical Society, 2003. 125(28): p. 8672-8679.

- 125 -

References

50. Brockstedt, U., et al., In vitro evolution of RNA aptamers recognizing carcinogenic aromatic amines. Biochem Biophys Res Commun, 2004. 313(4): p. 1004-8. 51. Zhou, N., et al., Selection and identification of streptomycin-specific single-stranded DNA aptamers and the application in the detection of streptomycin in honey. Talanta, 2013. 108: p. 109-16. 52. Hamula, C., et al., Selection and analytical applications of aptamers. TrAC Trends in Analytical Chemistry, 2006. 25(7): p. 681-691. 53. Smuc, T., I.Y. Ahn, and H. Ulrich, Nucleic acid aptamers as high affinity ligands in biotechnology and biosensorics. J Pharm Biomed Anal, 2013. 81-82: p. 210-7. 54. Foote, J. and H.N. Eisen, Kinetic and affinity limits on antibodies produced during immune responses. Proceedings of the National Academy of Sciences, 1995. 92(5): p. 1254-1256. 55. Poulsen, T.R., et al., Kinetic, Affinity, and Diversity Limits of Human Polyclonal Antibody Responses against Tetanus Toxoid. The Journal of Immunology, 2007. 179(6): p. 3841-3850. 56. Eisen, H.N. and G.W. Siskind, Variations in Affinities of Antibodies during the Immune Response. Biochemistry, 1964. 3: p. 996-1008. 57. Radom, F., et al., Aptamers: Molecules of great potential. Biotechnol Adv, 2013. 58. Agrawal, S., Antisense oligonucleotides: towards clinical trials. Trends Biotechnol, 1996. 14(10): p. 376-87. 59. Kusser, W., Chemically modified nucleic acid aptamers for in vitro selections: evolving evolution. J Biotechnol, 2000. 74(1): p. 27-38. 60. Barciszewski, J., et al., Locked nucleic acid aptamers. Methods Mol Biol, 2009. 535: p. 165-86. 61. Lee, K.Y., et al., Bioimaging of nucleolin aptamer-containing 5-(N-benzylcarboxyamide)-2'- deoxyuridine more capable of specific binding to targets in cancer cells. J Biomed Biotechnol, 2010. 2010: p. 168306. 62. Pieken, W.A., et al., Kinetic characterization of ribonuclease-resistant 2'-modified hammerhead . Science, 1991. 253(5017): p. 314-7. 63. Lin, Y., et al., Modified RNA sequence pools for in vitro selection. Nucleic Acids Res, 1994. 22(24): p. 5229-34. 64. Pagratis, N.C., et al., Potent 2'-amino-, and 2'-fluoro-2'-deoxyribonucleotide RNA inhibitors of keratinocyte growth factor. Nat Biotechnol, 1997. 15(1): p. 68-73. 65. Colas, P., et al., Genetic selection of peptide aptamers that recognize and inhibit cyclin- dependent kinase 2. Nature, 1996. 380(6574): p. 548-550. 66. Petrenko, V., Phage display for detection of biological threat agents. Journal of Microbiological Methods, 2003. 53(2): p. 253-262. 67. Liu, Y., et al., Peptide aptamers against titanium-based implants identified through phage display. J Mater Sci Mater Med, 2010. 21(4): p. 1103-7. 68. Turnbough, C.L., Jr., Discovery of phage display peptide ligands for species-specific detection of Bacillus spores. J Microbiol Methods, 2003. 53(2): p. 263-71. 69. Knurr, J., et al., Peptide ligands that bind selectively to spores of Bacillus subtilis and closely related species. Appl Environ Microbiol, 2003. 69(11): p. 6841-7. 70. Ozawa, M., K. Ohashi, and M. Onuma, Identification and characterization of peptides binding to newcastle disease virus by phage display. J Vet Med Sci, 2005. 67(12): p. 1237-41. 71. Gough, K.C., W. Cockburn, and G.C. Whitelam, Selection of phage-display peptides that bind to cucumber mosaic virus coat protein. J Virol Methods, 1999. 79(2): p. 169-80. 72. Smith, G.P. and V.A. Petrenko, Phage Display. Chem Rev, 1997. 97(2): p. 391-410. 73. Evans, D., et al., Electrical protein detection in cell lysates using high-density peptide-aptamer microarrays. J Biol, 2008. 7(1): p. 3. 74. Blum, J.H., et al., Isolation of peptide aptamers that inhibit intracellular processes. Proc Natl Acad Sci U S A, 2000. 97(5): p. 2241-6. 75. Norman, T.C., et al., Genetic selection of peptide inhibitors of biological pathways. Science, 1999. 285(5427): p. 591-5. 76. Doorbar, J. and G. Winter, Isolation of a peptide antagonist to the thrombin receptor using phage display. J Mol Biol, 1994. 244(4): p. 361-9.

- 126 -

References

77. Matthews, D.J. and J.A. Wells, Substrate phage: selection of protease substrates by monovalent phage display. Science, 1993. 260(5111): p. 1113-7. 78. Ohkubo, S., et al., Substrate phage as a tool to identify novel substrate sequences of proteases. Comb Chem High Throughput Screen, 2001. 4(7): p. 573-83. 79. Stephen, C.W. and D.P. Lane, Mutant conformation of p53. Precise epitope mapping using a filamentous phage epitope library. J Mol Biol, 1992. 225(3): p. 577-83. 80. Tombelli, S., M. Minunni, and M. Mascini, Analytical applications of aptamers. Biosens Bioelectron, 2005. 20(12): p. 2424-34. 81. Mosing, R.K. and M.T. Bowser, Microfluidic selection and applications of aptamers. Journal of Separation Science, 2007. 30(10): p. 1420-1426. 82. Musheev, M.U. and S.N. Krylov, Selection of aptamers by systematic evolution of ligands by exponential enrichment: addressing the polymerase chain reaction issue. Anal Chim Acta, 2006. 564(1): p. 91-6. 83. Gyllensten, U.B. and H.A. Erlich, Generation of single-stranded DNA by the polymerase chain reaction and its application to direct sequencing of the HLA-DQA locus. Proc Natl Acad Sci U S A, 1988. 85(20): p. 7652-6. 84. Espelund, M., R.A. Stacy, and K.S. Jakobsen, A simple method for generating single-stranded DNA probes labeled to high activities. Nucleic Acids Res, 1990. 18(20): p. 6157-8. 85. Higuchi, R.G. and H. Ochman, Production of single-stranded DNA templates by exonuclease digestion following the polymerase chain reaction. Nucleic Acids Res, 1989. 17(14): p. 5865. 86. Avci-Adali, M., et al., Upgrading SELEX technology by using lambda exonuclease digestion for single-stranded DNA generation. Molecules, 2010. 15(1): p. 1-11. 87. Paul, A., et al., Streptavidin-coated magnetic beads for DNA strand separation implicate a multitude of problems during cell-SELEX. Oligonucleotides, 2009. 19(3): p. 243-54. 88. Pan, W., P. Xin, and G.A. Clawson, Minimal primer and primer-free SELEX protocols for selection of aptamers from random DNA libraries. Biotechniques, 2008. 44(3): p. 351-60. 89. Vater, A., et al., Short bioactive Spiegelmers to migraine‐associated calcitonin gene‐related peptide rapidly identified by a novel approach: Tailored‐SELEX. Nucleic Acids Res, 2003. 31(21): p. e130. 90. White, R., et al., Generation of species cross-reactive aptamers using "toggle" SELEX. Mol Ther, 2001. 4(6): p. 567-73. 91. Glokler, J., T. Schutze, and Z. Konthur, Automation in the high-throughput selection of random combinatorial libraries--different approaches for select applications. Molecules, 2010. 15(4): p. 2478-90. 92. Wen, J.D. and D.M. Gray, Selection of genomic sequences that bind tightly to Ff gene 5 protein: primer-free genomic SELEX. Nucleic Acids Res, 2004. 32(22): p. e182. 93. Eulberg, D., et al., Development of an automated in vitro selection protocol to obtain RNA- based aptamers: identification of a biostable substance P antagonist. Nucleic Acids Res, 2005. 33(4): p. e45. 94. Keefe, A.D. and S.T. Cload, SELEX with modified nucleotides. Curr Opin Chem Biol, 2008. 12(4): p. 448-56. 95. Smith, G.P., Filamentous fusion phage: novel expression vectors that display cloned antigens on the virion surface. Science, 1985. 228(4705): p. 1315-7. 96. Pande, J., M.M. Szewczyk, and A.K. Grover, Phage display: concept, innovations, applications and future. Biotechnol Adv, 2010. 28(6): p. 849-58. 97. Sidhu, S.S., Engineering M13 for phage display. Biomol Eng, 2001. 18(2): p. 57-63. 98. Miller, L., Expression Libraries as Tools for the Development of Subunit Vaccines and Novel Detection Molecules for Orthopoxviruses, in Fakultät III, Prozesswissenschaften 2011, Technische Universität Berlin: D83. p. 113. 99. Dunn, I.S., Phage display of proteins. Curr Opin Biotechnol, 1996. 7(5): p. 547-53. 100. Azzazy, H.M. and W.E. Highsmith, Jr., Phage display technology: clinical applications and recent innovations. Clin Biochem, 2002. 35(6): p. 425-45.

- 127 -

References

101. Willats, W.G.T., Phage display: practicalities and prospects. Plant Molecular Biology, 2002. 50: p. 837-854. 102. Szardenings, M., Phage display of random peptide libraries: applications, limits, and potential. J Recept Signal Transduct Res, 2003. 23(4): p. 307-49. 103. Babickova, J., et al., In vivo phage display - A discovery tool in molecular biomedicine. Biotechnol Adv, 2013. 104. Tonelli, R.R., W. Colli, and M.J. Alves, Selection of binding targets in parasites using phage- display and aptamer libraries in vivo and in vitro. Front Immunol, 2012. 3: p. 419. 105. Huang, J.X., S.L. Bishop-Hurley, and M.A. Cooper, Development of anti-infectives using phage display: biological agents against bacteria, viruses, and parasites. Antimicrob Agents Chemother, 2012. 56(9): p. 4569-82. 106. Huang, L., et al., Novel peptide inhibitors of angiotensin-converting enzyme 2. J Biol Chem, 2003. 278(18): p. 15532-40. 107. Szardenings, M., et al., Phage display selection on whole cells yields a peptide specific for melanocortin receptor 1. J Biol Chem, 1997. 272(44): p. 27943-8. 108. Askoxylakis, V., et al., Binding of the Phage Display Derived Peptide CaIX-P1 on Human Colorectal Carcinoma Cells Correlates with the Expression of Carbonic Anhydrase IX. Int J Mol Sci, 2012. 13(10): p. 13030-48. 109. Taguchi, G. and A.P. Organization, Introduction to quality engineering: designing quality into products and processes. 1986: The Organization. 110. Jeney, C., et al., Taguchi optimisation of ELISA procedures. J Immunol Methods, 1999. 223(2): p. 137-46. 111. Khoudoli, G.A., et al., Optimisation of the two-dimensional gel electrophoresis protocol using the Taguchi approach. Proteome Sci, 2004. 2(1): p. 6. 112. Cobb, B.D. and J.M. CIarkson, A simple procedure for optimising the polymerase chain reaction (PCR) using modified Taguchi methods. Nucleic Acids Res, 1994. 22(18): p. 3801- 3805. 113. Luo, W., M. Pla-Roca, and D. Juncker, Taguchi design-based optimization of sandwich immunoassay microarrays for detecting breast cancer biomarkers. Anal Chem, 2011. 83(14): p. 5767-74. 114. Al-Refaie A. , L.M.H. Alpha Risk of Taguchi Method with L18 Array for NTB Type QCH by Simulation. in World Congress on Engineering. 2008. London, UK. 115. Zolfaghari, G., et al., Taguchi optimization approach for Pb(II) and Hg(II) removal from aqueous solutions using modified mesoporous carbon. J Hazard Mater, 2011. 192(3): p. 1046- 55. 116. Shao, K., et al., Emulsion PCR: A High Efficient Way of PCR Amplification of Random DNA Libraries in Aptamer Selection. PLoS ONE, 2011. 6(9): p. e24910. 117. Erlich, H.A., Polymerase chain reaction. J Clin Immunol, 1989. 9(6): p. 437-47. 118. Saiki, R.K., et al., Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science, 1988. 239(4839): p. 487-91. 119. Roux, K.H., Optimization and troubleshooting in PCR. Cold Spring Harb Protoc, 2009. 2009(4): p. pdb ip66. 120. Cho, M., et al., Quantitative selection of DNA aptamers through microfluidic selection and high-throughput sequencing. Proceedings of the National Academy of Sciences, 2010. 121. Metzker, M.L., Sequencing technologies - the next generation. Nat Rev Genet, 2010. 11(1): p. 31-46. 122. Mardis, E.R., The impact of next-generation sequencing technology on genetics. Trends Genet, 2008. 24(3): p. 133-41. 123. Ansorge, W.J., Next-generation DNA sequencing techniques. N Biotechnol, 2009. 25(4): p. 195-203. 124. Modrow S, F.D., Truyen U, Schätzl H, Pockenviren, in Molekulare Virologie. 2010, Spektrum Akademischer Verlag: Heidelberg. p. 610-630.

- 128 -

References

125. Moss, B., Poxviridae: The Viruses and their Replication, in DM Knipe and PM Howley (eds.), Fields Virology. 2007, PA. Lippincott Williams & Wilkins: Philadelphia. p. 2905-2945. 126. WHO, Declaration of global eradication of smallpox. Wkly Epidemiol Rec, 1980. 55: p. 2127- 2137. 127. Essbauer, S., M. Pfeffer, and H. Meyer, Zoonotic poxviruses. Vet Microbiol, 2010. 140(3-4): p. 229-36. 128. Kurth, A., et al., Rat-to-elephant-to-human transmission of cowpox virus. Emerg Infect Dis, 2008. 14(4): p. 670-1. 129. Coras, B., et al., Cowpox and a cat. Lancet, 2005. 365(9457): p. 446. 130. Damon, I.K., C.E. Roth, and V. Chowdhary, Discovery of monkeypox in Sudan. N Engl J Med, 2006. 355(9): p. 962-3. 131. Sejvar, J.J., et al., Human monkeypox infection: a family cluster in the midwestern United States. J Infect Dis, 2004. 190(10): p. 1833-40. 132. Kile, J.C., et al., Transmission of monkeypox among persons exposed to infected prairie dogs in Indiana in 2003. Arch Pediatr Adolesc Med, 2005. 159(11): p. 1022-5. 133. Silva, D.C., et al., Clinical signs, diagnosis, and case reports of Vaccinia virus infections. Braz J Infect Dis, 2010. 14(2): p. 129-34. 134. Trindade, G.S., et al., Brazilian vaccinia viruses and their origins. Emerg Infect Dis, 2007. 13(7): p. 965-72. 135. Singh, R.K., et al., : an emerging and re-emerging . Anim Health Res Rev, 2007. 8(1): p. 105-14. 136. Moore, Z.S., J.F. Seward, and J.M. Lane, Smallpox. Lancet, 2006. 367(9508): p. 425-35. 137. Alibek, K., Smallpox: a disease and a weapon. Int J Infect Dis, 2004. 8 Suppl 2: p. S3-8. 138. Mahy, B.W., An overview on the use of a viral pathogen as a bioterrorism agent: why smallpox? Antiviral Res, 2003. 57(1-2): p. 1-5. 139. Halloran, M.E., et al., Containing bioterrorist smallpox. Science, 2002. 298(5597): p. 1428-32. 140. Resch, W., et al., Protein composition of the vaccinia virus mature virion. Virology, 2007. 358(1): p. 233-47. 141. Moss, B., Poxvirus entry and membrane fusion. Virology, 2006. 344(1): p. 48-54. 142. Vanderplasschen, A., M. Hollinshead, and G.L. Smith, Intracellular and extracellular vaccinia virions enter cells by different mechanisms. J Gen Virol, 1998. 79 ( Pt 4): p. 877-87. 143. Condit, R.C., N. Moussatche, and P. Traktman, In a nutshell: structure and assembly of the vaccinia virion. Adv Virus Res, 2006. 66: p. 31-124. 144. Doceul, V., et al., Repulsion of superinfecting virions: a mechanism for rapid virus spread. Science, 2010. 327(5967): p. 873-6. 145. Roberts, K.L. and G.L. Smith, Vaccinia virus morphogenesis and dissemination. Trends Microbiol, 2008. 16(10): p. 472-9. 146. Chung, C.S., et al., Vaccinia virus proteome: identification of proteins in vaccinia virus intracellular mature virion particles. J Virol, 2006. 80(5): p. 2127-40. 147. Vazquez, M.I., et al., The vaccinia virus 14-kilodalton (A27L) fusion protein forms a triple coiled-coil structure and interacts with the 21-kilodalton (A17L) virus membrane protein through a C-terminal alpha-helix. J Virol, 1998. 72(12): p. 10126-37. 148. Chung, C.S., et al., A27L protein mediates vaccinia virus interaction with cell surface heparan sulfate. J Virol, 1998. 72(2): p. 1577-85. 149. Hsiao, J.C., C.S. Chung, and W. Chang, Vaccinia virus envelope D8L protein binds to cell surface chondroitin sulfate and mediates the adsorption of intracellular mature virions to cells. J Virol, 1999. 73(10): p. 8750-61. 150. Matho, M.H., et al., Structural and biochemical characterization of the vaccinia virus envelope protein D8 and its recognition by the antibody LA5. J Virol, 2012. 86(15): p. 8050-8. 151. Bisht, H., A.S. Weisberg, and B. Moss, Vaccinia virus L1 protein is required for cell entry and membrane fusion. J Virol, 2008. 82(17): p. 8687-94. 152. Foo, C.H., et al., The myristate moiety and amino-terminus of the vaccinia virus L1 constitute a bipartite functional region needed for entry. J Virol, 2012.

- 129 -

References

153. Brown, E., T.G. Senkevich, and B. Moss, Vaccinia virus F9 virion membrane protein is required for entry but not virus assembly, in contrast to the related L1 protein. J Virol, 2006. 80(19): p. 9455-64. 154. Lin, C.-L., et al., Vaccinia Virus Envelope H3L Protein Binds to Cell Surface Heparan Sulfate and Is Important for Intracellular Mature Virion Morphogenesis and Virus Infection In Vitro and In Vivo. J. Virol., 2000. 74(7): p. 3353-3365. 155. Roberts, K.L., et al., Acidic residues in the membrane-proximal stalk region of vaccinia virus protein B5 are required for glycosaminoglycan-mediated disruption of the extracellular enveloped virus outer membrane. J Gen Virol, 2009. 90(Pt 7): p. 1582-91. 156. Aldaz-Carroll, L., et al., Epitope-mapping studies define two major neutralization sites on the vaccinia virus extracellular enveloped virus glycoprotein B5R. J Virol, 2005. 79(10): p. 6260- 71. 157. Perdiguero, B. and R. Blasco, Interaction between vaccinia virus extracellular virus envelope A33 and B5 glycoproteins. J Virol, 2006. 80(17): p. 8763-77. 158. Moss, B., Smallpox vaccines: targets of protective immunity. Immunol Rev, 2011. 239(1): p. 8- 26. 159. Berhanu, A., et al., Vaccination of BALB/c mice with Escherichia coli-expressed vaccinia virus proteins A27L, B5R, and D8L protects mice from lethal vaccinia virus challenge. J Virol, 2008. 82(7): p. 3517-29. 160. Akitomi, J., et al., ValFold: Program for the aptamer truncation process. Bioinformation, 2011. 7(1): p. 38-40. 161. Sonntag, M., et al., New adenovirus in bats, Germany. Emerg Infect Dis, 2009. 15(12): p. 2052-5. 162. Miller, L., Expression Libraries as Tools for the Development of Subunit Vaccines and Novel Detection Molecules for Orthopoxviruses, 2011, Technische Universität Berlin. 163. Hall, B., et al., Design, synthesis, and amplification of DNA pools for in vitro selection. Curr Protoc Mol Biol, 2009. Chapter 24: p. Unit 24 2. 164. Miller, L. Analyzing gels and western blots with ImageJ. 2010 [cited 2010; Available from: http://lukemiller.org/index.php/2010/11/analyzing-gels-and-western-blots-with-image-j/. 165. Malmqvist, M., Surface plasmon resonance for detection and measurement of antibody- antigen affinity and kinetics. Current Opinion in Immunology, 1993. 5(2): p. 282-286. 166. Gillaspy, D.A. Biomolecular Interactions with the Biacore T100. [cited 2012 27.11.2012]; Available from: http://microgen.ouhsc.edu/biacore.htm. 167. Shevchenko, A., et al., In-gel digestion for mass spectrometric characterization of proteins and proteomes. Nat. Protocols, 2007. 1(6): p. 2856-2860. 168. Lou, X., et al., Micromagnetic selection of aptamers in microfluidic channels. Proc Natl Acad Sci U S A, 2009. 106(9): p. 2989-94. 169. Qian, J., et al., Generation of highly specific aptamers via micromagnetic selection. Anal Chem, 2009. 81(13): p. 5490-5. 170. Gopinath, S.C., et al., An efficient RNA aptamer against human influenza hemagglutinin. J Biochem, 2006. 139(5): p. 837-46. 171. Drolet, D.W., L. Moon-McDermott, and T.S. Romig, An enzyme-linked oligonucleotide assay. Nat Biotech, 1996. 14(8): p. 1021-1025. 172. Mierendorf, R.C., Jr. and R.L. Dimond, Functional heterogeneity of monoclonal antibodies obtained using different screening assays. Anal Biochem, 1983. 135(1): p. 221-9. 173. Daniels, D.A., et al., A tenascin-C aptamer identified by tumor cell SELEX: systematic evolution of ligands by exponential enrichment. Proc Natl Acad Sci U S A, 2003. 100(26): p. 15416-21. 174. Kumar, P.K., et al., Isolation of RNA aptamers specific to the NS3 protein of hepatitis C virus from a pool of completely random RNA. Virology, 1997. 237(2): p. 270-82. 175. Roux, K.H., Optimization and troubleshooting in PCR. PCR Methods Appl, 1995. 4(5): p. S185- 94. 176. Kramer, M.F. and D.M. Coen, Enzymatic amplification of DNA by PCR: standard procedures and optimization. Curr Protoc Immunol, 2001. Chapter 10: p. Unit 10 20.

- 130 -

References

177. Li, Y., et al., Development and validation of a new PCR optimization method by combining experimental design and artificial neural network. Appl Biochem Biotechnol, 2010. 160(1): p. 269-79. 178. Bell, D.A. and D.M. DeMarini, Excessive cycling converts PCR products to random-length higher molecular weight fragments. Nucleic Acids Res, 1991. 19(18): p. 5079. 179. Reyderman, L. and S. Stavchansky, Determination of single-stranded oligodeoxynucleotides by capillary gel electrophoresis with laser induced fluorescence and on column derivatization. J Chromatogr A, 1996. 755(2): p. 271-80. 180. Gray, G.D. and E. Wickstrom, Rapid measurement of modified oligonucleotide levels in plasma samples with a fluorophore specific for single-stranded DNA. Antisense Nucleic Acid Drug Dev, 1997. 7(3): p. 133-40. 181. Wochner, A. and J. Glokler, Nonradioactive fluorescence microtiter plate assay monitoring aptamer selections. Biotechniques, 2007. 42(5): p. 578, 580, 582. 182. Chen, S.J., C.C. Huang, and H.T. Chang, Enrichment and fluorescence enhancement of adenosine using aptamer-gold , PDGF aptamer, and Oligreen. Talanta, 2010. 81(1-2): p. 493-8. 183. Rhinn, H., D. Scherman, and V. Escriou, One-step quantification of single-stranded DNA in the presence of RNA using Oligreen in a real-time polymerase chain reaction thermocycler. Anal Biochem, 2008. 372(1): p. 116-8. 184. Liu, J., Z. Cao, and Y. Lu, Functional nucleic acid sensors. Chem Rev, 2009. 109(5): p. 1948-98. 185. Laitinen, O.H., et al., Brave new (strept)avidins in biotechnology. Trends Biotechnol, 2007. 25(6): p. 269-77. 186. Cho, E.J., et al., Optimization of aptamer microarray technology for multiple protein targets. Anal Chim Acta, 2006. 564(1): p. 82-90. 187. Katilius, E., C. Flores, and N.W. Woodbury, Exploring the sequence space of a DNA aptamer using microarrays. Nucleic Acids Res, 2007. 35(22): p. 7626-35. 188. Heyduk, E. and T. Heyduk, Nucleic Acid-Based Fluorescence Sensors for Detecting Proteins. Anal Chem, 2005. 77(4): p. 1147-1156. 189. Xiao, Y., et al., Label-free electronic detection of thrombin in blood serum by using an aptamer-based sensor. Angew Chem Int Ed Engl, 2005. 44(34): p. 5456-9. 190. Xiong, X., et al., DNA aptamer-mediated cell targeting. Angew Chem Int Ed Engl, 2013. 52(5): p. 1472-6. 191. Ulrich, H., A.H. Martins, and J.B. Pesquero, RNA and DNA aptamers in cytomics analysis. Cytometry A, 2004. 59(2): p. 220-31. 192. Murphy, M.B., et al., An improved method for the in vitro evolution of aptamers and applications in protein detection and purification. Nucleic Acids Res, 2003. 31(18): p. e110. 193. Conrad, R. and A.D. Ellington, Detecting Immobilized Protein Kinase C Isozymes with RNA Aptamers. Analytical Biochemistry, 1996. 242(2): p. 261-265. 194. Shin, S., et al., An alternative to Western blot analysis using RNA aptamer-functionalized quantum dots. Bioorg Med Chem Lett, 2010. 20(11): p. 3322-5. 195. Cheng, C., et al., Potent inhibition of human influenza H5N1 virus by oligonucleotides derived by SELEX. Biochem Biophys Res Commun, 2008. 366(3): p. 670-4. 196. Avci-Adali, M., et al., Pitfalls of cell-systematic evolution of ligands by exponential enrichment (SELEX): existing dead cells during in vitro selection anticipate the enrichment of specific aptamers. Oligonucleotides, 2010. 20(6): p. 317-23. 197. Witkowski, P.T., et al., Cellular impedance measurement as a new tool for poxvirus titration, antibody neutralization testing and evaluation of antiviral substances. Biochem Biophys Res Commun, 2010. 401(1): p. 37-41. 198. Nitsche, A., et al., One-step selection of Vaccinia virus-binding DNA aptamers by MonoLEX. BMC Biotechnol, 2007. 7: p. 48. 199. Girardot, M., P. Gareil, and A. Varenne, Interaction study of a -binding aptamer with mono- and divalent cations by ACE. Electrophoresis, 2010. 31(3): p. 546-55.

- 131 -

References

200. Howard, A.R., T.G. Senkevich, and B. Moss, Vaccinia virus A26 and A27 proteins form a stable complex tethered to mature virions by association with the A17 transmembrane protein. J Virol, 2008. 82(24): p. 12384-91. 201. Rogers, J.D., et al., Identification and characterization of a Peptide affinity reagent for detection of noroviruses in clinical samples. J Clin Microbiol, 2013. 51(6): p. 1803-8. 202. Rodriguez, J.R., D. Rodriguez, and M. Esteban, Insertional inactivation of the vaccinia virus 32- kilodalton gene is associated with attenuation in mice and reduction of viral gene expression in polarized epithelial cells. J Virol, 1992. 66(1): p. 183-9. 203. Niles, E.G. and J. Seto, Vaccinia virus gene D8 encodes a virion transmembrane protein. J Virol, 1988. 62(10): p. 3772-8. 204. Lao, Y.-H., K. Peck, and L.-C. Chen, Enhancement of Aptamer Microarray Sensitivity through Spacer Optimization and Avidity Effect. Anal Chem, 2009. 81(5): p. 1747-1754. 205. Hasegawa, H., et al., Improvement of Aptamer Affinity by Dimerization. Sensors (Basel), 2008. 8(2): p. 1090-1098. 206. Nonaka, Y., K. Sode, and K. Ikebukuro, Screening and improvement of an anti-VEGF DNA aptamer. Molecules, 2010. 15(1): p. 215-25. 207. Murase, K., et al., EF-Tu binding peptides identified, dissected, and affinity optimized by phage display. Chem Biol, 2003. 10(2): p. 161-8. 208. Harlow E, L.D., Antibody-Antigen Interactions. Antibodies- A Laboratory Manual. 1st ed. 1988, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. 209. Thie, H., et al., Affinity maturation by phage display. Methods Mol Biol, 2009. 525: p. 309-22, xv. 210. Ferrer, M. and S.C. Harrison, Peptide ligands to human immunodeficiency virus type 1 gp120 identified from phage display libraries. J Virol, 1999. 73(7): p. 5795-802. 211. Biorn, A.C., et al., Mode of action for linear peptide inhibitors of HIV-1 gp120 interactions. Biochemistry, 2004. 43(7): p. 1928-38. 212. Bai, F., et al., Antiviral peptides targeting the west nile virus envelope protein. J Virol, 2007. 81(4): p. 2047-55. 213. Rajik, M., et al., Identification and characterisation of a novel anti-viral peptide against avian influenza virus H9N2. Virol J, 2009. 6(1): p. 74. 214. Williams, D.D., O. Benedek, and C.L. Turnbough, Jr., Species-specific peptide ligands for the detection of Bacillus anthracis spores. Appl Environ Microbiol, 2003. 69(10): p. 6288-93. 215. Goldman, E.R., et al., Phage-displayed peptides as reagents. J Mol Recognit, 2000. 13(6): p. 382-7. 216. Gopinath, S.C., Antiviral aptamers. Arch Virol, 2007. 152(12): p. 2137-57. 217. Wang, R., et al., Selection and characterization of DNA aptamers for use in detection of avian influenza virus H5N1. J Virol Methods, 2013. 189(2): p. 362-369. 218. Petrenko, V.A. and I.B. Sorokulova, Detection of biological threats. A challenge for directed molecular evolution. J Microbiol Methods, 2004. 58(2): p. 147-68. 219. Dover, J.E., et al., Recent advances in peptide probe-based biosensors for detection of infectious agents. J Microbiol Methods, 2009. 78(1): p. 10-9. 220. Fang, X., et al., Molecular aptamer for real-time oncoprotein platelet-derived growth factor monitoring by fluorescence anisotropy. Anal Chem, 2001. 73(23): p. 5752-7. 221. Fang, X., et al., Synthetic DNA aptamers to detect protein molecular variants in a high- throughput fluorescence quenching assay. Chembiochem, 2003. 4(9): p. 829-34. 222. Lai, R.Y., K.W. Plaxco, and A.J. Heeger, Aptamer-based electrochemical detection of picomolar platelet-derived growth factor directly in blood serum. Anal Chem, 2007. 79(1): p. 229-33. 223. Heyduk, E. and T. Heyduk, Nucleic acid-based fluorescence sensors for detecting proteins. Anal Chem, 2005. 77(4): p. 1147-56. 224. Baker, B.R., et al., An electronic, aptamer-based small-molecule sensor for the rapid, label- free detection of cocaine in adulterated samples and biological fluids. J Am Chem Soc, 2006. 128(10): p. 3138-9.

- 132 -

References

225. Win, M.N., J.S. Klein, and C.D. Smolke, Codeine-binding RNA aptamers and rapid determination of their binding constants using a direct coupling surface plasmon resonance assay. Nucleic Acids Res, 2006. 34(19): p. 5670-82. 226. Tombelli, S., M. Minunni, and M. Mascini, Aptamers-based assays for diagnostics, environmental and food analysis. Biomol Eng, 2007. 24(2): p. 191-200. 227. Kirby, R., et al., Aptamer-based sensor arrays for the detection and quantitation of proteins. Anal Chem, 2004. 76(14): p. 4066-75. 228. Attree, O., et al., Development and comparison of two immunoassay formats for rapid detection of botulinum neurotoxin type A. J Immunol Methods, 2007. 325(1-2): p. 78-87. 229. Lucht, A., et al., Development of an immunofiltration-based antigen-detection assay for rapid diagnosis of Ebola virus infection. J Infect Dis, 2007. 196 Suppl 2: p. S184-92. 230. Zhou, J., M.R. Battig, and Y. Wang, Aptamer-based molecular recognition for biosensor development. Anal Bioanal Chem, 2010. 398(6): p. 2471-80. 231. Marik, J. and K.S. Lam, Peptide and small-molecule microarrays. Methods Mol Biol, 2005. 310: p. 217-26. 232. Taitt, C.R., S.H. North, and N.V. Kulagina, Antimicrobial peptide arrays for detection of inactivated biothreat agents. Methods Mol Biol, 2009. 570: p. 233-55.

- 133 -