Design, Synthesis, and Characterization of Functional Click Nucleic

DESIGN, SYNTHESIS, AND CHARACTERIZATION OF FUNCTIONAL CLICK NUCLEIC

ACID POLYMERS AND CONJUGATES FOR BIOLOGICAL APPLICATIONS

ALEX J. ANDERSON

B.S., Vanderbilt University, 2015

M.S., University of Colorado, 2017

A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Chemical and Biological Engineering

2020

Thesis committee: Dr. Christopher N. Bowman, Chair Dr. Stephanie J. Bryant Dr. Virginia L. Ferguson Dr. Jennifer N. Cha Dr. Frank Vernerey Abstract

Anderson, Alex J. (PhD. Chemical and Biological Engineering)

Design, Synthesis, and Characterization of Functional Click Nucleic Acid Polymers and

Conjugates for Biological Applications

Thesis directed by Christopher N. Bowman and Stephanie J. Bryant

Oligonucleotides are a powerful class of biopolymer which, through Watson-and-Crick hydrogen bonding, are capable of directing the self-assembly of materials as well as recognizing and binding to very specific targets. These features have made oligonucleotides an attractive tool for uses in materials science and biotechnology. However, there are several drawbacks to the implementation of oligonucleotides with a natural backbone (i.e. DNA or RNA) including susceptibility to degradation and a limited scale of production. Click Nucleic Acids (CNAs) represent a new class of oligonucleotide that was designed to alleviate these issues. The distinguishing feature of CNAs is that they are polymerized via radical mediated thiol-ene click chemistry, which vastly increases the speed and scale of oligonucleotide synthesis.

This thesis focuses on developing CNA polymers and conjugates and demonstrating the utility of CNAs as an alternative oligonucleotide for various biological applications. Initially, linear polymers and copolymers were synthesized and CNA’s cytocompatibility and ability to interact with cellular components was evaluated. Specifically, it was found that linear PEG-CNA conjugates were taken up by cells quickly through an apparently passive mechanism and did not exhibit significant cytotoxicity. In addition, the ability of CNA homopolymers to bind to mRNA was demonstrated by evaluating its use as an mRNA isolation technique. Taking advantage of the

ii inherent insolubility of CNA, functional mRNA was effectively isolated from total RNA extracts in yields comparable to commercially available products as determined by in vitro protein translation and RT-qPCR. The second part of this thesis details the synthesis of a branched CNA conjugate and its ability to form physically crosslinked networks upon the introduction of complementary ssDNA. These gels were found to be viscoelastic and completely thermoreversible, exhibiting a melting transition around 60°C. In all, microscale thermophoresis, circular dichroism spectroscopy, and rheology were all used to study the complex crosslinking interaction between CNA and DNA. Finally, higher order, CNA-microparticle conjugates were developed and their ability to load and delivery ssDNA was assessed. After confirming successful conjugation, CNA functionalized microparticles were found to load approximately 6 pmol ssDNA

/ mg microparticle and loading was sensitive to ssDNA length and sequence. Release of loaded ssDNA was determined to be temperature dependent, but stable to buffer pH. Further, phagocytosis of microparticles was observed via fluorescence microscopy and corroborated by biochemical analysis. In all, this thesis demonstrates the uses of CNA as a functional oligonucleotide for a variety of biological applications through the design of various conjugate architectures.

iii

To my Dad and Mom

Dad, for inspiring me, every single day

Mom, for the love and support you have always provided

Acknowledgements

This dissertation is the culmination of not only years of work, but also years of relationships with my mentors, colleagues, and friends. There is not a way to adequately thank all of the people who have helped me during this process, but I will do my best here.

First, I would like to thank my advisors, Christopher Bowman and Stephanie Bryant. I could not have asked for a more supportive pair of mentors. I often got a lot of questions about being a co-advised student and I always say I got the best of both worlds. Your varied expertise challenged me to think about problems from both a materials science and biological perspective which I think is one of the most important skills I developed in graduate school. I thank you both for your guidance and mentorship which have prepared me to take the next steps in my career.

I would like to thank my committee, Ginger Ferguson, Jennifer Cha, and Franck Vernerey for their valuable input throughout my years of research. This project got to where it is today through the help of you and your students. In addition, I want to thank all of my collaborators that have helped me on this project. In particular, I want to thank Heidi Culver, Ben Fairbanks, Jasmine

Sinha, and Mingtao Chen for lending scientific advice and allowing me to vent when things went awry.

This project was made possible through a National Science Foundation supported Materials

Research Science and Engineering Center (DMR 1420736). Personally, I was funded by a US

Department of Education Graduate Assistantship in Areas of National Need (GAANN) grant.

I also would like to thank my undergraduate advisors for encouraging my interest in research and for providing me with opportunities for me to learn new skills and explore different areas of research. Dr. Guelcher and Dr. Duvall, thank you for your willingness to take on an inexperienced undergraduate student and for helping me find my way.

The best piece of advice I can give to new graduate students is to make sure you are surrounded by people you love and who love you. At the end of the day, these are the people who will be your support when things become stressful due to deadlines, important presentations, or more often failures. In my case, I am lucky that I have had such an incredible support system at

CU-Boulder, and I would not trade my experience for the world.

My labmates in both the Bowman and Bryant research groups are all wonderful people and

I cannot thank them all enough for making lab a fun place to work. You all were an integral part to my experience here and made me feel like family from day one. I especially want to thank

Elizabeth and Stanley for being awesome lab mentors and more importantly, great friends. I know

I asked a lot of stupid questions, but I want to thank you for answering them earnestly and making me feel like I was a part of the lab. Margaret, thank you for being a great labmate, roommate, role model, and friend. When the goings got rough, you always provided a level-headed perspective and weren’t afraid to let me know when I was overreacting, which I sincerely appreciate. Parker also told me to thank Dorie for being his best friends. And Leila, thank you for being my best friend. Your dad was right, your grad school friends will be your friends for life, despite anything anyone says (including me). Oh, and Parker wanted me to thank Butler for being his other best friend.

Finally, to all my family and friends everywhere, thank you for your constant love and support!

Table of Contents

Chapter 1 - Introduction ...... 1 1.1 Introduction ...... 1

1.2 Uses of DNA-Polymer Conjugates ...... 3

1.2.1 Linear Architectures ...... 4

1.2.2 Branched Architectures ...... 5

1.2.3 Higher Order Architectures ...... 6

1.3 Development of XNA’s and their Polymer Conjugates ...... 7

1.3.1 Peptide Nucleic Acids ...... 8

1.3.2 Locked Nucleic Acids ...... 11

1.3.3 Morpholino Nucleic Acids ...... 12

1.3.4 Other types of XNAs ...... 14

1.4 Development of Click Nucleic Acids ...... 15

1.5 Thesis overview ...... 17

1.6 References ...... 18

Chapter 2 - Objectives and Scope ...... 28 2.1 Aim 1: Identify interactions between linear CNA polymers and copolymers with cells cellular components ...... 29

2.1.1 Sub-aim 1.1 ...... 29

2.1.2 Sub-aim 1.2 ...... 30

2.2 Aim 2: Utilize the interactions between a multi-armed CNA-polymer conjugate and DNA to create a stimuli-responsive gel network ...... 30

vii

2.3 Aim 3: Develop a higher order CNA-polymer conjugate capable of loading DNA for therapeutic delivery applications ...... 30

Chapter 3 - Cytocompatibility and Cellular Internalization of PEGylated “Clickable” Nucleic Acid Oligomers ...... 32 3.1 Abstract ...... 32

3.2 Introduction ...... 33

3.3 Experimental ...... 35

3.3.1 PEG-CNA-RHO and PEG-RHO conjugate Synthesis...... 35

3.3.2 Cell Culture...... 36

3.3.3 Cell viability...... 36

3.3.4 Cellular uptake of fluorescent conjugates...... 37

3.3.5 Time dependent cellular uptake and localization...... 37

3.3.6 Immunofluorescence...... 37

3.3.7 Fluorescence microscopy...... 38

3.3.8 Statistical Analysis...... 38

3.4 Results and Discussion ...... 39

3.5 Conclusions ...... 47

3.6 Acknowledgements ...... 48

3.7 References ...... 48

3.8 Supporting Information ...... 53

Chapter 4 - Messenger RNA Enrichment Using Synthetic Oligo(T) Click Nucleic Acids ...... 63 4.1 Abstract ...... 63

4.2 Introduction ...... 64

4.3 Results and Discussion ...... 66

viii

4.4 Materials and Methods ...... 72

4.4.1 CNA synthesis, polymerization, and characterization...... 72

4.4.2 Optimization of Cy-5 labelled RNA Pulldown...... 72

4.4.3 Optimization of Cy-5 labelled mRNA pulldown...... 73

4.4.4 In-Vitro Translation...... 73

4.4.5 Optimization of mRNA release...... 74

4.4.6 Pulldown specificity – rRNA vs. mRNA...... 74

4.4.7 Pulldown and release from Total RNA...... 75

4.4.8 RT-PCR...... 76

4.5 Acknowledgements ...... 76

4.6 References ...... 77

4.7 Supporting Information ...... 80

Chapter 5 - Viscoelastic and Thermoreversible Networks Crosslinked by Non-covalent Interactions Between “Clickable” Nucleic Acids Oligomers and DNA ...... 85 5.1 Abstract ...... 85

5.2 Introduction ...... 86

5.3 Materials and Methods ...... 88

5.3.1 Synthesis of 8PEG-T macromolecule ...... 88

5.3.2 Microscale Thermophoresis (MST) Titration ...... 89

5.3.3 Circular Dichroism ...... 90

5.3.4 Gel Formation ...... 90

5.3.5 Rheological Characterization ...... 90

5.3.6 Statistical Analysis ...... 91

5.4 Results and Discussion ...... 91

5.4.1 Synthesis of 8PEG-T macromolecule ...... 91

5.4.2 Binding of 8PEG-T to Polyadenine DNA ...... 93

5.4.3 Formation of CNA/DNA gels ...... 95

5.4.4 CNA/DNA gels are thermoreversible ...... 97

5.4.5 Understanding the CNA/DNA interaction within the gel network ...... 100

5.5 Conclusion ...... 103

5.6 Acknowledgments...... 103

5.7 References ...... 104

5.8 Supporting Information ...... 110

Chapter 6 - Synthesis and Characterization of Click Nucleic Acid Conjugated Polymeric Microparticles for DNA Delivery Applications ...... 116 6.1. Abstract ...... 116

6.2. Introduction ...... 117

6.3. Experimental ...... 118

6.3.1 Materials...... 118

6.3.2 Synthesis of monodisperse, step-growth microparticles...... 118

6.3.3 Synthesis of CNA decorated microparticles ...... 119

6.3.4 Characterization of CNA/microparticle copolymerization ...... 119

6.3.5 Preparation of nucleic acid loaded microparticles ...... 120

6.3.6 Evaluation of DNA release properties ...... 120

6.3.7 Cell Culture and Microparticle Uptake ...... 121

6.3.8 Metabolic Activity and Cell Viability Assays...... 121

6.3.9 Statistical Analysis ...... 122

6.4. Results and discussion ...... 122

6.5. Conclusion ...... 132

6.6. Supporting Information ...... 132

6.7. Acknowledgements ...... 133

6.8 References ...... 133

6.9 Supporting Information ...... 139

Chapter 7 - Conclusions and Recommendations ...... 142 7.1 Conclusions ...... 142

7.1.1 Linear polymers and copolymers ...... 142

7.1.2 Branched Conjugates...... 143

7.1.3 Higher order conjugates ...... 144

7.2 Recommendations ...... 145

7.3 References ...... 146

Bibliography ...... 148

List of Tables

Supporting Table 4.1 – Summary of optimized buffer compositions for CNA-facilitated RNA enrichment...... 84

Supporting Table 4.2 – List of qPCR Primer Sequences and Efficiencies ...... 84

xii

List of Figures

Figure 1.1 – Representations of linear, branched, and higher order nucleic acid-based conjugates...... 3

Figure 3.2 – Cellular internalization of PEG-CNA-RHO, PEG-RHO, and RHO. Cells were treated with 100 μg/mL of conjugate or an equimolar amount of rhodamine for 1 hr, washed, fixed and imaged by fluorescence microscopy. A significantly greater average fluorescence intensity per cell after washing was determined for cells treated with PEG-CNA-RHO (*=p < 0.001) than for cells treated with PEG-RHO or RHO controls, indicating CNA dependent cellular uptake. (n>1000, error bars represent standard deviations)...... 42

Figure 3.4 – Three representative images of immunofluorescence microscopy images highlight the cellular localization of PEG-CNA-rho conjugates. Cells were treated with 100 μg/mL of PEG-CNA-rho conjugates for 1 hr, then fixed, permeabilized, and stained. Blue represents the nucleus (Hoechst 33342), red represents the PEG-CNA-RHO conjugate, and green represents either endosomes, the endoplasmic reticulum, mitochondria, or lysosomes. Areas of yellow indicate colocalization between the PEG-CNA-rho conjugate and the indicated organelle.

xiii

Areas of magenta indicate colocalization between the PEG-CNA-rho conjugate and the nucleus...... 44

Figure 3.5 – Temperature dependent uptake of PEG-CNA-RHO. Cells were incubated with 100 μg/mL of PEG-CNA-RHO for 1 hour at either 4°C or 37°C. Cellular uptake of PEG-CNA- RHO was not significantly affected by incubation temperature, suggesting a passive uptake mechanism...... 46

Figure 4.1 – (a) Comparison of a natural DNA repeat unit to the CNA repeat unit. The thogether backbone removes backbone charge, but the 6-atom spacing allows for binding of complementary nucleic acids. (b) General process of mRNA isolation procedure. The oligo(T) CNA binds to and helps precipitate mRNA in solution and the mRNA can be released through a heating and reconstitution step...... 65

Figure 4.2 – Pulldown of A20 RNA as a function of oligo(T) CNA concentration. Oligo(T) CNA at sufficiently high concentrations achieved >90% pulldown of complementary RNA while effectively no pulldown was observed regardless of concentration for non-complementary sequences. (b) Release of RNA is achieved by heating samples to 75°C to dissociate the hybridization between CNA and RNA. Data is represented as the mean of at least 3 replicates and error bars represent standard deviations...... 67

Figure 4.3 – (a) Effective pulldown of fluorescent EGFP mRNA was achieved at CNA concentrations of 125 μM and higher. (b) Under optimized conditions, greater than 90% pulldown efficiency could be achieved at biologically relevant mRNA concentrations. (c) In vitro translation of enriched mRNA reveals that complementary CNA is needed for isolation to occur. (d) Enrichment of mRNA is only possible after the heating and reconstitution step. Data is represented as the mean of at least 3 replicates and error bars represent standard deviations...... 69

xiv

Figure 5.1 – (a) Synthesis of 8PEG-T was achieved through a copolymerization technique. The solid product was obtained after dialysis and deprotection. (b) Comparison between a DNA monomer and CNA monomer. The CNA repeat unit maintains the 6-atom spacing allowing for the hybridization to complementary nucleic acids...... 92

Figure 5.2 – (a) Microscale thermophoresis titration experiment showing binding between 8PEG-

T and a Cy-5 labelled A20 ssDNA. Solvent was 20% H2O in DMSO. A20 DNA was supplied at 50 nM. Data points are averages of technical replicates, n=3. (b) Circular dichroism of A20 ssDNA only and A20 bound to 8PEG-T. The profile of the 8PEG-T alone was subtracted out of the spectra showing binding (red curve). Spectra are averages of technical replicates, n=3...... 94

Figure 5.3 – (a) Gels only form when the ssDNA crosslinker is complementary and sufficiently long enough to bridge two adjacent 8PEG-T molecules (b) The CNA/DNA gel’s modulus was tracked as a function of time, rising to over 1 kPa in around 10 min. Rheology done at 0.1% strain and 1 Hz. (c) Hypothesized mechanism of gelation including cycles and the potential for multiple CNA oligomers to bind to one ssDNA crosslinker...... 95

Figure 5.4 – (a) Melting and reconstitution of the CNA/DNA gel. Gels were made at 5% w/v

8PEG-T, with an A40 ssDNA, at 1:1 A:T, in 20% H2O in DMSO. Rheology was performed at 5% strain at 1 Hz. (b) Thermal cycles of CNA/DNA gels showing complete reversibility. .. 97

Figure 5.5 – (a) Circular dichroism of the melting transition of 8PEG-T and A20. The signal at 290 nm was tracked as a function of temperature and fit to a sigmoidal decay curve yielding a melting point of 24.5°C. Data points are averages of technical replicates, n=3. (b) CNA/DNA gels were prepared at either 5% w/v or 3% w/v 8PEG-T and subjected to thermally induced melting. The gels were found to melt quicker and at a lower temperature at 3% w/v than at 5% w/v...... 99

Figure 5.6 – (a) Oscillation sweep of 5% w/v CNA/DNA gel. (b) Frequency sweep of 5% w/v

CNA/DNA gel. Gels were made with a 1:1 A:T ratio, in 20% H2O in DMSO. Plots are representative of 3 replicates...... 100

Figure 5.7 – Representative stress relaxation plots for (a) A40 crosslinker and (b) A20 crosslinker. Relaxation profiles overlap regardless of the temperature with A40 as the crosslinker, but xv

exhibit temperature dependent relaxation with the A20 crosslinker. Rheology was performed with a 10% immediate initial strain in both cases...... 102

Figure 6.1 – (a) UV-Vis absorbance spectra show the characteristic absorbance of nucleobases in the degradation products of microparticles functionalized with adenine and thymine. (b) The GPC trace of the degradation products indicated the presence of short oligomers. Comparison to an internal standard revealed the molecular weights to be 1300 and 1200 g/mol for adenine and thymine, respectively...... 124

Figure 6.4 – (a) The interaction between thymine CNA and adenine DNA is hypothesized to be mediated by hydrogen bonding between complementary base pairs. (b) Microparticles functionalized with thymine CNA oligomers showed significantly more loading of a complementary A10-Cy5 ssDNA sequence. Negligible loading was observed for adenine functionalized particles and unfunctionalized particles (n = 3, * p < 0.05). Loading was visually confirmed with fluorescence microscopy for (c) MP+oT, but not (d) MP+oA, or (e) MP only...... 127

Figure 6.5 – The rate of DNA release is dependent on the temperature, where higher temperatures lead to faster, burst release and lower temperatures lead to a slower more steady release. After 48 hours all samples were kept at 55°C, which initiated another burst release, demonstrating this use of this system for triggered DNA delivery...... 129

Figure 6.6 – (a) Macrophages appear to phagocytose A10-Cy5 loaded microparticles as evidenced by the clear association between the particles (green; the color was changed to provide better contrast) and cells. Nuclei are stained in blue. (b) 3D images of cells treated with A10-Cy5 xvi loaded microparticles showed that microparticles existed on the same plane as nuclei and occurred within the cell membrane, suggesting internalization. The cell membrane was stained in red. (c) The metabolic activity of macrophages treated with microparticles significantly increased up to 90% after 24 hours, supporting the conclusion that the microparticles are being phagocytosed. (n = 4, * p < 0.05) (d) CNA-functionalized microparticles do not cause significant cell death. (n = 4) ...... 131

xvii

List of Supporting Figures

Supporting Figure 3.1 – Gel Permeation Chromatography (Refractive Index) of copolymerization product. The shift of the poly(ethylene glycol) peak to higher molecular weights after copolymerizing with CNA monomer indicates the successful formation of the PEG-CNA block copolymer. The peak seen in the “Copolymerized PEG-CNA” trace at 18.5 minutes has been attributed to cyclized monomer...... 53

Supporting Figure 3.2 – Sample 1H NMR of PEG-SH. Integration of peaks were done with respect to peak B, the PEG methoxy protons...... 54

Supporting Figure 3.3 – Sample 1H NMR of the protected thymine, thiol-ene monomer...... 55

Supporting Figure 3.4 – Sample 1H NMR of the PEG-CNA conjugate, immediately after polymerization. Integration of peaks were done with respect to peak E, the PEG methoxy protons...... 56

Supporting Figure 3.5 – Sample 1H NMR of the water-soluble PEG-CNA fraction. Integration of peaks were done with respect to peak E, the PEG methoxy protons...... 57

of dye per mole of conjugate: 퐴560 푛푚/(휀560 푛푚 ∗ 푙) ∗ (푀푊푐표푛푗푢푔푎푡푒/(푚푔푐표푛푗푢푔푎푡푒/푚퐿)) =

(푚표푙푒푠 푑푦푒 / 푚표푙푒푠 푐표푛푗푢푔푎푡푒). Here, A560 nm is the absorbance of the solution at 560 nm, ε560 nm is the molar absorptivity of the dye at 560 nm, l is the path length, MWconjugate¬ is the molecular weight of the entire conjugate, and mg¬conjugate/mL is the concentration of the conjugate solution. The absorbance of PEG-CNA-RHO at 560 nm corresponds to an extent of conjugation of 0.48 moles dye/moles conjugate. However, this may be a slight overestimation

xviii

since there is evidence of some free fluorophore in the conjugate. In general, these values represent typical conjugation percentages for this reaction...... 59

Supporting Figure 3.8 – Gel permeation chromatography of the crude PEG-CNA conjugate compared to dye-functionalized PEG-CNA-RHO conjugate. The peak corresponding to the cyclized monomer disappears, but a slightly larger molecular weight peak appears just before 18 minutes. This peak has been attributed to unreacted dye that remains in the product. GPC analysis indicates that this peak accounts for only 4 mol% of the product. The presence of a small amount of unreacted dye was deemed tolerable since we have shown that the dye is not get taken up by cells on its own. In addition, the conjugate peak appears to be shifted to lower molecular weights. This behavior is hypothesized to be because conjugates with high degrees of polymerization (high molecular weights) were not solubilized by the water...... 60

Supporting Figure 3.9 – 1H NMR of the PEG-RHO conjugate. The presence of aromatic hydrogens that appear between 6.80 – 8.30 ppm indicates the presence of the rhodamine dye. Integration of the unique PEG peaks at 3.24 and 3.45-3.80 ppm show maintenance of the PEG structure after the reaction...... 61

Supporting Figure 3.10 – DLS measurements of PEG-CNA-RHO. DLS showed that PEG-CNA- RHO forms particles in solution at the concentration that is delivered to cells. Micelles had a volume average diameter of 28 ± 22 nm. DLS measurements were made with a Malvern Zetasizer at a detection angle of 173°. PEG-CNA-RHO was dissolved in PBS at a concentration of 0.1 mg/mL, and the size distribution was measured 3 times to verify repeatability...... 62

Supporting Figure 4.1 – Gel Permeation Chromatography of oligo(T) CNA. The mean retention time corresponded to a degree of polymerization of 16 ± 3 repeat units and a PDI of 1.5 ± 0.2...... 80

Supporting Figure 4.2 – Pulldown efficiency of A20 increased as the % DMSO decreased with the highest percentage occurring at 5%. This was attributed to the fact that a larger aqueous phase better facilitated CNA precipitation. Results are represented as averages of at least 3 replicates and error bars as standard deviations...... 80

xix

Supporting Figure 4.3 – The addition of LiCl salt to the binding buffer resulted in an increase in pulldown efficiency. This was attributed to the ability of Li+ ions to facilitate RNA precipitation...... 81

Supporting Figure 4.4 – Pulldown efficiency remained >90% for concentrations up to 64 μg/mL, demonstrating the robustness of this technique...... 81

Supporting Figure 4.5 – In order to correlate the EGFP fluorescence to input mRNA concentration, a standard curve was generated with known concentrations of EGFP mRNA that were not subjected to pulldown and release. From this curve, it was determined that the fluorescence of the translated protein was easily correlated to starting mRNA amounts in the range of 0-300 ng. For in vitro translation experiments, the mRNA yield was low (<100 ng)...... 83

Supporting Figure 4.6 – The % release of mRNA was found to be a function of mRNA concentrations using release buffer R1 which contains cationic salts that facilitates mRNA precipitation. It was observed that higher mRNA concentrations led to more efficient release...... 83

Supporting Figure 4.7 – Oligo(T) mediated precipitation of mRNA compared to rRNA confirmed specificity for mRNA...... 84

Supporting Figure 5.1 – Representative characterization of fully synthesized 8PEG-T macromers compared to the 8PEG-SH precursor...... 110

Supporting Figure 5.2 – Gel permeation chromatography (GPC) of the unfunctionalized 8-arm poly(ethylene glycol) precursor (red) and the 8PEG-T conjugate (black) following copolymerization. The shift to lower retention times in the primary peak indicates an increase in molecular weight. The shoulder that appears in the 8PEG-T trace is most likely due to the polymer polydispersity, which is a side effect of the quasi-step growth polymerization mechanism, or incomplete separation of smaller molecular weight products. Estimated molecular weights were determined via 1H NMR integration...... 111

Supporting Figure 5.3 – The mean final storage modulus of CNA/DNA gels after 3 heat/cool cycles shows a stiffening of the gel with each cycle number. Gels were made at 5% w/v 8PEG- T, 1:1 A:T ratio, and 20% H2O in DMSO. One cycle refers to a temperature ramp to 70°C xx

followed by a cooling to 22°C and equilibration for 20 minutes. Statistical significance is denoted by * (p<0.05)...... 112

Supporting Figure 5.4 – Temperature sweep data from 22°C to 60°C showing that at 60°C, the gel does not melt, indicating melting temperatures are higher than 60°C...... 113

Supporting Figure 5.5 – Arrhenius plot of characteristic relaxation times. A linear fit gives an activation energy of 110 ± 20 kJ/mol...... 115

Supporting Figure 6.1 – Loading of A10-Cy5 DNA is observed qualitatively by the blue hue of the pellet in the MP+oT samples. The MP+oA and MP only samples did not have the same color...... 139

Supporting Figure 6.2 – Calibration curves to quantify the loading of different types of Cy-5 labelled ssDNA: (a) A10-Cy5, (b) T10-Cy5, (c) (GA)10-Cy5, and (d) A20-Cy5...... 139

Supporting Figure 6.3 – MP+oT microparticles were loaded with A10-Cy5 DNA, A20-Cy5 DNA, T10-Cy5 DNA, and GA10-Cy5 DNA. Loading was significantly greater for the A10- Cy5 DNA, but significant loading was also observed for A20-Cy5 DNA. There was negligible loading for the non-complementary T10-Cy5 DNA and the periodically mismatched GA10- Cy5 DNA. These results support the specificity of the CNA/DNA mediated loading and demonstrates the microparticles ability to load DNA of different sizes...... 140

Supporting Figure 6.4 – Calibration curves to quantify the release of A10-Cy5 from microparticles under various conditions. (a) Calibration curve for long term, temperature dependent release. (b), (c), and (d) Calibration curves for pH dependent release. Calibration curves were made for each pH due to the possible dependence of Cy5 fluorescence on pH...... 141

Supporting Figure 6.5 – More representative images of microparticle uptake in RAW 264.7 cells. Nuclei are stained in blue and loaded microparticles are represented as magenta...... 141

xxi

List of Schemes

Scheme 3.1 - In vitro strategy for evaluating the cytocompatibility of PEGylated CNA copolymers ...... 40

Scheme 6.1 – Polymeric microparticles were synthesized with a hexafunctional thiol (Di-PETHP) and triacrylate (TMPTA) with a 20% excess thiol. This thiol served as a conjugation point for pendant CNA oligonucleotides...... 123

xxii

Chapter 1 - Introduction

1.1 Introduction

Natural nucleic acids, such as deoxyribonucleic acid (DNA) and ribonucleic acid (RNA), represent a powerful class of biopolymers that exhibit a complex and diverse array of functionalities that are primarily dictated by interactions on the monomer scale. Nucleic acid polymers are composed of four distinct nucleobases, thymine (T), adenine (A), cytosine (C), and guanine (G), all of which have specific hydrogen bonding capabilities. Particularly, T and A have specific affinities for each other which makes them a Watson and Crick hydrogen bonded pair.

The same is true for the C/G pair. These specific interactions have important consequences in

DNA’s and RNA’s role in a biological context. First, the primary sequence of nucleic acids is used to store information, the most important example being our genetic code. Second, the monomeric interactions have the ability to direct the assembly of higher order structures through hybridization, a great example being the complex structure of tRNA. Because of these unique attributes, nucleic acids have been used in a wide array of research settings to take advantage of these functionalities.

Some of these applications include, gene therapy, drug delivery, biosensing, information storage, molecular electronics, and the engineering of nanostructures.1–8

Unfunctionalized nucleic acids were initially used in the aforementioned applications, and the oligonucleotide’s (ODNs) performance was tied to the properties of the natural nucleic acid strand, limiting its functionality.9–12 On the other hand, the field of synthetic polymers and other materials offers incredible control over polymer size, architecture, and material properties. A variety of different techniques can be used to obtain linear, branched, dendritic, and even higher order polymeric materials with a variety of monomer functionalities. In order to access the advantages of both nucleic acids and polymer chemistry, nucleic acids can be covalently 1 conjugated to synthetic polymers to enable widely applicable and more robust functionalities. In all, a variety of nucleic acid-polymer conjugate compositions and architectures have been explored, each having unique properties that are dictated by the nucleic acid type and sequence as well as the polymer type and architecture. For example, in the design of linear DNA diblock copolymers, polymer type and size significantly affect the self-assembly and hybridization behavior of the conjugate (see Section 1.2.1 for more details).13,14

However, for as powerful as nucleic acid-polymer conjugates are, there are drawbacks to using natural nucleic acids of DNA and RNA. The primary reason is the general lack of scale of

DNA and RNA synthesis. The predominant nucleic acid synthesis scheme relies on an iterative, solid-phase mechanism that severely limits scale polymer size and scale of production.15,16 The other reason is that the natural nucleic acid structure is not the most efficient for target identification and binding. Specifically, the natural phosphodiester linkage is susceptible to degradation and introduces charges that can lead to electrostatic repulsion. Thus, many groups have sought to synthesize nucleic acid mimics that retain the primary functionalities of DNA and

RNA but resolve one or both of these issues. Some examples such as phosphothioate nucleic acids only slightly alter the natural nucleic acid structure. Others such as locked nucleic acids (LNAs) and peptide nucleic acids (PNAs) replace the natural backbone with a different one while retaining the general monomer structure (this latter group is referred to as xeno-nucleic acids or XNAs). The last class of nucleic acid mimics utilize nucleobase functionalized monomers that can participate in common polymerization techniques like RAFT and ROMP. These mimics bare the least resemblance to natural nucleic acid structure (e.g. have different monomer spacings), but still exhibit typical nucleobase functionalities.

Figure 1.1 – Representations of linear, branched, and higher order nucleic acid-based conjugates. This review will attempt to summarize the progress made in nucleic acid conjugated

polymers and their biological applications with particular focus on macromolecular architecture

(i.e. linear, branched, higher order). First, a discussion on DNA-polymer conjugates will be

presented. Next will be an examination on how the field has included XNA technology. Finally,

other nucleobase functionalized synthetic polymers will be considered.

1.2 Uses of DNA-Polymer Conjugates

The union of DNA and synthetic polymers has allowed nucleic acid technology to be

applied in a variety of situations. For instance, in gene therapy, antisense ODNs have been

conjugated to polymers to increase their pharmacokinetic profiles by increasing nuclease

resistance.17,18 In bioelectronics, various types of electrodes have been decorated with specific

DNA sequences (called aptamers) designed to bind to targets of interest. The binding event

significantly alters the electrode’s conductivity which allows for analyte detection. As one might

imagine, careful design of the conjugate architecture is required to for each specific application

and functionality. This section will discuss three main architectures (depicted in Figure 1.1) and

the various applications that have been demonstrated with such structures.

1.2.1 Linear Architectures

Linear conjugates normally consist of an ON block and a synthetic polymer block and are most often achieved either by covalent conjugation in the solution phase or incorporation of the polymeric block during solid-phase synthesis.19 The polymer block can be designed to be either hydrophilic or hydrophobic and can be adjusted in length to achieve the desired conjugate properties. One of the most prominent classes of linear structures are amphiphilic ODN-polymer conjugates. These structures self-assemble in aqueous environments to create a variety of structures with a hydrophobic, polymer core and a hydrophilic ON corona.14 Oftentimes, the hydrophobic corona of the micellar formations contains a hydrophobic small molecule drug, such as doxorubicin or paclitaxel, for controlled drug delivery.20 In other cases, it was found that the micellar structure was improved ON cell uptake and lead to a greater antisense response upon polymer degradation.21,22 If the synthetic polymer is designed correctly, a vesicular structure is possible, which has been shown useful for drug delivery and optoelectronics applications.23,24

Alternatively, if water soluble polymers are used, such as poly(ethylene glycol) (PEG), no hierarchical structure forms, but these conjugates have been found to be useful in other ways. For instance, a DNA-PEG conjugate was used by Vreeland and coworkers to assist in polymer molecular weight analysis.25 The monodispersity and predictable charge density of the DNA strand led to a more sensitive separation of a polydisperse PEG sample using electrophoretic analysis.

Other reasons hydrophilic polymers are used in DNA conjugation are to augment the pharmacokinetic profiles of antisense ODNs.17 However, such conjugations can also detrimentally affect hybridization thermodynamics.13,26 Although, if the conjugate is designed correctly, one can achieve the improved pharmacokinetics while not affecting the DNA functionality.27

1.2.2 Branched Architectures

Branched DNA-polymer conjugates usually consist of multiple (i.e. > 2) ODNs attached to a polymeric backbone. These structures can be brush-like, where ODNs are grafted from a

“backbone”, or star-like, where the polymer structure itself is branched, and is functionalized with pendant ODNs. One of the primary advantages of branched conjugates is the presence of multiple binding sites per macromolecule which leads to more favorable binding kinetics.28–30 For instance, the larger number of DNA interactions have been shown to be useful for the electrochemical detection of specific nucleic acid targets by enabling sharp and robust hybridization kinetics.30 In addition, Fouz and coworkers used a bottlebrush DNA-polymer architecture as a nanotag for antibody targets. The conjugate was capable of sequestering thousands of fluorescent dyes without apparent self-quenching, leading to exceptionally bright signals.29

Another benefit of having multiple binding sites per macromolecule is the possibility for network formation. ON hybrids can act as crosslinkers if they bridge two branched macromolecules. Importantly, due to the properties of the non-covalent hybridization interaction, these crosslinks are stimuli-responsive and are susceptible to changes in temperature, pH, and competitive binders. In some cases, these crosslinks are included along-side covalent crosslinks, which leads to macroscopic changes in gel properties upon the introduction of a stimulus.31–33 Just as often, the networks are crosslinked completely by nucleic acid interactions and exhibit sol-gel transitions in response to the stimulus.34–39 In any case, these responsive gel systems have been used in applications such as drug delivery,40,41 biosensing and actuation,32 and even tissue engineering.42

1.2.3 Higher Order Architectures

Higher order DNA conjugates consist of high surface area materials (i.e. nanoparticles and nanosheets) grafted with ODNs at nonspecific locations. The higher density of nucleic acid coverage greatly enhances the number of interactions that are possible but can also significantly alter the properties of the conjugate materials. For instance, Mirkin and coworkers have found that the organization of single stranded nucleic acids on the surface of nanoparticles (the conjugate named Spherical Nucleic Acids, or SNAs) leads to increased binding constants, sharper melting transitions, efficient intracellular delivery, and increased nuclease resistance.43,44 These materials not only have been applied as therapeutic technologies, but also have great importance in the design of self-assembled atom equivalents.45 Through multivalent cation exchange, these structures are capable of dynamic structural control, with interesting implications for optical, magnetic, and mechanical material properties. Other groups have taken similar approaches to decorating metal-organic frameworks (MOFs) which are known for their porous structure and potential therapeutic applications.46

Alternatively, high surface area nanosheets made of inorganic metals or organic conducting materials also have unique properties that benefit from DNA conjugation. These architectures are usually applied as biosensors that detect analytes through DNA hybridization. In most cases, a single stranded DNA molecule designed to bind to a specific target is immobilized on a conducting surface. When the target molecule is introduced, the binding event results in a change in output signal (usually an increase in current impedance in the nanosheet but can also include optoelectronic signals like FRET responses). The type of conducting material is often tailored to the targeted application. For instance, graphene oxide (GO) is used because of its aqueous dispersibility, biocompatibility and surface area;4 gold has attractive optoelectronic properties and

6 favorable DNA immobilization;47 and synthetic conductive polymers are highly tunable.48 In addition, materials can be combined to exploit the advantages of each individual part.3 In all, these nanosheet-based conjugate biosensors have been shown to be highly specific for an array of analytes and obtain pico- to femto-molar detection limits.3,4,47

1.3 Development of XNA’s and their Polymer Conjugates

The field of xenonucleic acids (XNAs) was developed to address the challenges of using

DNA in the above applications.49 In a biological context, DNA is susceptible to degradation from cellular nucleases and exhibits poor bioavailability. DNA is also very sensitive to solution ionic strength, due to the overall negative charge that originates from the phosphate group. This charge is also the source of electrostatic repulsion that, when combined with backbone flexibility, makes hybridization less favorable. Some modifications are designed to specifically prevent ON degradation, eliminate backbone charge, or introduce rigidity to the backbone.

Another class of XNAs were developed to also combat another inefficiency of DNA technology, namely the arduous solid-phase synthetic scheme. The solid-phase strategy has dominated the field of nucleic acid synthesis, primarily because of the ability to precisely control monomer sequence is unmatched. However, for what this technique offers in specificity, it lacks in scalability. Each monomer addition is the result of many intermediate reactions and ultimately relatively low overall yields. This primary limitation results in length restrictions, relatively higher costs, and insufficient material yields that are necessary for many applications. Conversely, XNA synthesis methods in this class seek to use template directed and other solution-phase approaches for ON synthesis. However, most have not been able to rival the ability to design non-trivial sequences, involve high labor costs, or do not bind to natural nucleic acids due to structural mismatches. Regardless, these materials are still of interest because they retain many of the same 7

Figure 1.2 – Repeat unit structures of common XNAs (b-d) compared to the structure of natural DNA. Click-chemistry synthesized nucleic acids are shown in the bottom row (e-g). CNAs, the XNA discussed in this dissertation is shown in (f) and (g). nucleobase functionalities such as base-specific hydrogen bonding. The following sections will discuss prominent classes of XNAs and their polymer conjugates.

1.3.1 Peptide Nucleic Acids

Peptide Nucleic Acids (PNAs) is one of the most promising classes of XNAs in the field.

As opposed to a deoxyribose sugar and phosphate linked backbone, PNAs make use of a peptide linkage which eliminates backbone charge as well as nuclease susceptibility (Figure 1.2b).

Importantly, these ONs can be synthesized on a traditional peptide synthesizer which allows for the same level of sequence control that is offered through DNA solid-phase synthesis.50,51 PNAs are known to efficiently bind complementary DNA, adopting a similar double helix secondary

8 structure, and have been shown to even have more favorable binding thermodynamics due to the lack of electrostatic repulsion.52 Because of this, PNA/DNA duplexes are much more sensitive to single-pair mismatches than DNA/DNA duplexes.53 In addition, PNA has been found to infiltrate

DNA/DNA hybrids and supplant the complementary strand or even create triplexes through

Hoogsteen binding (hydrogen bonding in the major groove).52 For these reasons, PNA is most often used as an antisense therapy, where a specific PNA sequence is used to stop the translation of mutated genes that cause disease.54,55 However, one of the major drawbacks of PNAs is their decreased solubility because of the lack of backbone charge, which has been addressed through conjugation to solubilizing moieties or polymer conjugation.56,57

1.3.1.1 Linear PNA Conjugates

Like DNA, PNA functionality can be significantly improved and expanded by its conjugation to other polymers. A significant chunk of this class of architecture consist of PNA- peptide conjugates. Due to PNAs limited solubility, its ability to traverse the cellular membrane is lacking, and peptides known as cell penetrating peptides (CPPs) have been shown to help.58–60

CPPs are usually positively charged, amphiphilic molecules and have been used to shuttle PNA antisense sequences into diseased cells, achieving effective knockdown of gene expression.61,62 As an example, when a 16-mer antisense PNA ON targeting HeLa cell trans-activation was attached to a variety of well-known CPPs, trans-activation was effectively inhibited whereas no inhibition was detected without the CPPs.60

Other examples of linear PNA-based conjugates attempt to exploit the favorable binding thermodynamics of PNAs in self-assembly and biosensing applications. PNA based micelles, made of self-complementary PNA sequences attached to a hydrophobic alkyl block, formed architectures controlled by the copolymer composition.63 As discussed in Section 1.2.1, such

9 structures have interesting applications in biomedicine and drug delivery. Further, PNA conjugated to PEG was found to be useful as a probe in detecting single-nucleotide polymorphisms in hairpin

DNA structures through electrophoretic analysis.64 Briefly, PEG-PNA conjugates were able to disrupt the secondary structure of hairpinned ssDNA, bind to its target sequence, and due to the added size of the PEG block, retard the electrophoretic movement of the complex. Importantly, the PEG-PNA probe could effectively discriminate single base-pair mismatches leading to enhanced biodetection.

1.3.1.2 Branched PNA Conjugates

In literature, branched PNA-polymer architectures have primarily been applied in the fields of drug delivery and novel hydrogel development. Realizing the potential of PNA as an antisense therapeutic, Berthold and coworkers took a different approach to intracellular PNA delivery by conjugating an antisense strand to a branched poly(ethylenimine) (PEI), a cationic polymer often used for DNA transfection.65 It was found that the branched structure led to an enhanced antisense response compared to a comparable linear conjugate. In a materials science context, the excellent binding properties of PNAs were used to design hybrid hydrogels crosslinked by PNA/DNA duplexes and triplexes. The gels formed, while comparably soft, showed characteristic behavior of transient networks and could be applicable as a drug-free therapeutic.

1.3.1.3 Higher Order PNA Conjugates

The high specificity and enhanced stability of PNA duplexes makes them an ideal candidate for biosensing and self-assembly applications. To achieve the required response, higher- order architectures are often desired, usually in the form of high-surface area nanoparticles or nanosheets. The underlying nanomaterial substrates that have been utilized have ranged from polymeric materials (due to their enhanced ability to be manipulated and functionalized) to

10 inorganic metals (for their optoelectrical properties). For instance, Turner et al. showed that PNAs conjugated to polymeric shell crosslinked (SCK) nanoparticles were able to predictably direct the self-assembly of nanomaterials.66 Further, gold can serve as the underlying nanomaterial substrate if the PNA sequences contain thiols, as thiols have been shown to have a particular affinity for gold. These materials exhibit similar properties to DNA/Au conjugates, with the added benefit of increased binding affinity and biological stability.67–69

In a therapeutic sense, higher order conjugates can also lead to increased solubilization and intracellular deliver of PNAs, much like SNAs do. Galli and coworkers used this strategy to design a paramagnetic iron oxide nanoparticle functionalized with PNAs in an attempt to combine the high specificity interactions of PNAs, the magnetic properties of the particles, and the aforementioned combined benefits of conjugation.70 Similarly, Shrestha et al. found that a polymeric nanoparticle scaffold provided more customizability than an inorganic substrate.71 Shell crosslinked knedel-like nanoparticles functionalized with PNA were designed to recognize and inhibit the production of inducible nitric oxide synthase. Attesting to the tunability of polymeric scaffolds, the nanoparticle was also functionalized with a cell penetrating peptide to aid in the cellular uptake of the complex.

1.3.2 Locked Nucleic Acids

Another well-developed class of XNAs are Locked Nucleic Acids (LNAs). In contrast to

PNAs, LNAs retain much of the natural structure of DNA (i.e. the phosphodiester linkage and the ribose sugar), but the structure is conformationally “locked” with an oxymethylene linkage between the 2’ and 4’ ribose carbons (Figure 1.2c).72,73 This bridge enforces a favorable preorganization in the ON and ultimately reduces the entropic penalty of nucleic acid hybridization, resulting in better complement recognition and binding thermodynamics.72,74

Importantly, this new type of nucleic acid is compatible with standard solid-phase synthetic approach for DNA. For these reasons, scientists have been interested in incorporating LNAs into applications that utilize DNA functionalities, primarily as a dopant to enhance the ON’s nuclease resistance. Because it’s been found that simply doping in a few LNA nucleotides into a DNA sequence can enhance the ON’s properties, not many examples of complete LNA sequences exist.

However, some researchers have recognized LNAs superior binding capabilities and have designed higher order conjugate structures to develop novel biosensors.

1.3.2.1 Higher order LNA conjugates

LNA functionalized gold surfaces were employed as biosensors for the detection of specific DNA with high single base pair mismatch selectivity.75,76 While the LNA/DNA hybridization thermodynamics showed superior performance when compared to DNA/DNA hybridization, it should be noted that LNAs conserve the negatively backbone charge, making them sensitive to solution ionic strength and counter-ion type.76 In a similar vein, LNA functionalized gold surfaces have also been used as biosensors in “sandwich” mode, where the capture LNA probe (attached to the gold surface) hybridizes with a section of target DNA and the remainder is bound to a reporter ON. In this case the reporter ON contained a redox active enzyme which led to signal amplification and thus target detection.77 As with the previous examples, the biosensor was able to effectively distinguish between fully complementary and single base pair mismatches.

1.3.3 Morpholino Nucleic Acids

Like PNAs, morpholino nucleic acids completely change the natural phosphodiester-ribose sugar backbone of DNA and RNA to one with a morpholine ring and a phosphorodiamidate linkage (Figure 1.2d). Morpholinos were designed to originate from readily available and easily

12 transformable starting materials while requiring nuclease resistance. Even with the drastic change in backbone structure, morpholinos are able to bind to complementary DNA and RNA sequences with good specificity. They are also nuclease resistant and are much more water soluble than rival

PNAs. However, contrary to PNAs and LNAs, binding is no more efficient than natural DNA, therefore the primary advantage of such substitutions is a cheaper monomer unit and enhanced tolerance to solution ionic strength. Even so, morpholinos have been conjugated to a variety of substrate types and architectures and employed as an effective biosensor and therapeutic.

1.3.3.1 Branched Morpholino Conjugates

There are few examples of branched morpholino-polymer architectures, however they are an important class of conjugates because of the morpholino’s good water solubility despite a neutral backbone charge. These features made a morpholino functionalized, brush-like poly(N-(2- dyroxypropyl)methacrylamide) (HPMA) copolymer a key component of an apoptosis inducing therapeutic strategy.78 In brief, the multifunctional aspect of the HPMA-morpholino conjugate was imperative to the clustering of specific cell receptors that ultimately led to cell apoptosis. In addition, the strengths of morpholino nucleic acids were attractive for the development of a salt- free, hydrogel biosensor. Here, partially complementary morpholinos were included in an acrylamide hydrogel to achieve additional reversible crosslinks.79 Upon addition of a fully complementary strand these crosslinks would break, decreasing the crosslink density and increasing the swelling capacity of the gel.

1.3.3.2 Higher Order Morpholino Conjugates

Higher order, surface-attached morpholinos have been successfully employed as a capture probe for DNA-detection biosensors. In one example, morpholinos attached to nanochannels in a poly(ethylene terephthalate) (PET) membrane selectively hybridized with target ssDNA

13 introducing a net negative charge density that could be detected within the membrane.80 In another surface attached example, the negative charge of the bound target DNA led to the electrostatic binding of a cationic redox polymer which could be detected through amperometric measurements.81 Other researchers have taken different strategies at DNA detection using higher order morpholino architectures. Zu et al. synthesized two different morpholino-gold nanoparticle conjugates that could be connected by a target DNA sequence.82 This crosslinking interaction ultimately lead to a change in the optoelectronic properties of the gold particles and a detectable signal change. In all cases, the biosensors were efficiently able to distinguish between fully complementary and single base-pair mismatched sequences with low ionic strength requirements.

1.3.4 Other types of XNAs

An important distinction to point out regarding the previously described XNAs is that they all are synthesized through a type of solid-phase approach. As discussed in Section 2, the solid- phase approach is relatively inefficient compared to other polymerization strategies and very high conversions are needed to achieve an adequate yield. As an example, if an ON with 20 repeat units is desired, and the conversion of each monomer addition is 99%, the overall conversion becomes

82%. If the step conversion drops by 4% to 95%, the overall conversion will then be 36%. To counter this shortcoming, there have been attempts to develop ONs through higher yield, solution- phase mechanisms. Some examples of these include acrylate-mediated free radical polymerization techniques,83–85 RAFT polymerization,86,87 ring-opening metathesis polymerization (ROMP),88–90 and click-chemistry based polymerizations.16,91 The primary drawbacks to using such techniques include a lack of sequence control and the inability to bind natural nucleic acids. However, for applications where random or single-base sequences can be tolerated, these nucleic acid polymers still hold tremendous value. Many researchers have shown that base-specific interactions are still

14 attainable and useful for applications in nanostructure assembly, drug delivery, and hydrogel formation.

Further, the types of polymerizations mentioned here are attractive for oligonucleotide synthesis because copolymerization is trivial, as there are many techniques for creating copolymers with a variety of architectures. For instance, acrylate free radical polymerization and RAFT have been used for making di- and tri-block copolymers with pendant nucleobases.83–87 In solution, diblock conjugate materials were found to exhibit base dependent self-assembly facilitated by hydrogen bonding between complementary nucleobases. In one example, these self-assembled structures were used to deliver a small molecule cancer drug.83 In the solid phase, tri-block copolymers containing both adenine and thymine self-assembled into regular patterns with hard phases that exhibited characteristics of complementary hydrogen bonding.84 Di-block copolymers synthesized through ROMP exhibited similar self-assembly properties in solution and the resulting nanoaggregates were used to deliver the same cancer drug to cells.89,90 Interestingly, it’s been shown that complementary bases are necessarily needed to elicit self-assembly. Adenine functionalized di-block copolymers synthesized through ROMP were found to self-assemble without their thymine complement.88 Indeed, the self-hybridization of adenine has been observed in research on natural DNA as well.92

1.4 Development of Click Nucleic Acids

Each of the classes of nucleic acid conjugates detailed in the previous section have their advantages and drawbacks. PNA, LNA, and morpholino ONs have either superior binding characteristics, nuclease resistance, or both, but still hindered by the dependence on inefficient solid-phase synthesis. ONs synthesized by traditional polymerization techniques (i.e. RAFT,

ROMP, free radical) address the issue of yield and scale but are unable to interact with natural 15

Figure 1.3 – (a) Schematic of the thiol-ene polymerization mechanism. Reaction takes place between a thiol and alkene and proceeds in a step-growth manner. (b) This reaction scheme is utilized to synthesize linear oligomers of CNAs. nucleic acid polymers. Thus, there is a need for the development of an oligonucleotide that can be synthesized by high-yielding chemistry, bind to natural nucleic acid recognizing structures, and can be easily conjugated to other synthetic polymers. Step growth polymerization mechanisms, especially those that involve “click” reactions, are a promising alternative to the traditional polymerization strategies. These reactions are scalable, efficient, often orthogonal, and have a grander scope of monomers with more unique functionalities.

To date, there are at least two different approaches to using “click” chemistries to develop

XNAs. One, utilizing the azide-alkyne cycloaddition reaction (structure depicted in Figure 1.2e), has been shown to bind to DNA and even be recognized by polymerase enzymes, but this strategy is unable to achieve degrees of polymerization higher than 3 in solution without iterative deprotection steps.16 The other, deemed Click-Nucleic Acids (CNAs) (shown in Figures 1.2f and 16

1.2g), was developed by our lab in 2015 and involves a thiol-ene polymerization approach, the mechanism of which is detailed in Figure 1.3a.91 True to the paradigm of the “click” reaction, previous reports have shown that nucleic acid containing oligomers can be easily synthesized through this approach in good yields, quickly, and under atmospheric conditions (Figure 1.3b).91,93

Further, the thiol-ene polymerization mechanism is ideal for creating stable oligonucleotide conjugates quickly and efficiently as the thiol moiety is reactive to many different functional groups. So far, previous work has demonstrated CNA’s ability to interact with complementary structures both in solution and in confined conformations where binding is entropically favored.93,94

1.5 Thesis overview

CNAs represent a promising new type of XNA due to their facile, click-chemistry mediated synthesis and their ability to interact specifically with complementary oligonucleotides. This dissertation focuses on understanding the fundamental interactions of CNA and their conjugates and attempts to exploit these interactions to achieve various functions. Linear CNA oligomers and copolymers are discussed in Chapters 3 and 4. In Chapter 3, a linear CNA-PEG conjugate was synthesized and its cellular uptake and cytocompatibility were investigated. In addition, the localization of these conjugates was determined. Chapter 4 discusses the use of CNA as an mRNA isolation tool. This work takes advantage of oligo(T) CNA’s ability to recognize and bind to the poly(A) tail and CNA’s inherent insolubility in water to precipitate mRNA from total RNA.

Branched CNA conjugates are discussed in Chapter 5. This work uses a copolymerization strategy to synthesize multi-armed CNA-PEG copolymers that crosslink upon the introduction of complementary DNA. The resulting gel networks were characterized and their stress relaxation and thermoreversibility are reported. Finally, Chapter 6 focuses on a higher order CNA conjugate. 17

Monodisperse microparticles, synthesized through thiol-Michael step-growth polymerization, were doped into CNA polymerizations, yielding CNA-microparticle conjugates. The ability of these conjugates to load complementary DNA and their subsequent release properties were investigated. In all, this dissertation demonstrates how the fundamental interactions of CNA as a oligonucleotide can be exploited to achieve a variety of functions, and that these functions arise from the conjugate architecture.

1.6 References

1. Zhang, Q. et al. DNA Origami as an In Vivo Drug Delivery Vehicle for Cancer Therapy. ACS Nano 8, 6633–6643 (2014).

2. Han, D. et al. DNA Origami with Complex Curvatures in Three-Dimensional Space. Science 332, 342–346 (2011).

3. Huang, H., Bai, W., Dong, C., Guo, R. & Liu, Z. An ultrasensitive electrochemical DNA biosensor based on graphene/Au nanorod/polythionine for human papillomavirus DNA detection. Biosens. Bioelectron. 68, 442–446 (2015).

4. Singh, A. et al. Graphene oxide-chitosan nanocomposite based electrochemical DNA biosensor for detection of typhoid. Sens. Actuators B Chem. 185, 675–684 (2013).

5. Young, K. L. et al. Hollow Spherical Nucleic Acids for Intracellular Gene Regulation Based upon Biocompatible Silica Shells. Nano Lett. 12, 3867–3871 (2012).

6. Church, G. M., Gao, Y. & Kosuri, S. Next-Generation Digital Information Storage in DNA. Science 337, 1628–1628 (2012).

7. Nishikawa, M. et al. Biodegradable CpG DNA hydrogels for sustained delivery of doxorubicin and immunostimulatory signals in tumor-bearing mice. Biomaterials 32, 488–494 (2011).

8. Zhang, C. et al. Biodegradable DNA-Brush Block Copolymer Spherical Nucleic Acids Enable Transfection Agent-Free Intracellular Gene Regulation. Small 11, 5360–5368 (2015). 18

9. Moret, I. et al. Stability of PEI–DNA and DOTAP–DNA complexes: effect of alkaline pH, heparin and serum. J. Controlled Release 76, 169–181 (2001).

10. Liu, F., Shollenberger, L. M., Conwell, C. C., Yuan, X. & Huang, L. Mechanism of naked DNA clearance after intravenous injection. J. Gene Med. 9, 613–619 (2007).

11. Barry, M. E. et al. Role of Endogenous Endonucleases and Tissue Site in Transfection and CpG-Mediated Immune Activation after Naked DNA Injection. Hum. Gene Ther. 10, 2461– 2480 (1999).

12. Bureau, M. F. et al. Intramuscular plasmid DNA electrotransfer: Biodistribution and degradation. Biochim. Biophys. Acta BBA - Gene Struct. Expr. 1676, 138–148 (2004).

13. Jia, F. et al. Effect of PEG Architecture on the Hybridization Thermodynamics and Protein Accessibility of PEGylated Oligonucleotides. Angew. Chem. Int. Ed. 56, 1239–1243 (2017).

14. Peterson, A. M. & Heemstra, J. M. Controlling self-assembly of DNA-polymer conjugates for applications in imaging and drug delivery. WIREs Nanomedicine Nanobiotechnology 7, 282– 297 (2015).

15. Matteucci, M. D. & Caruthers, M. H. Synthesis of deoxyoligonucleotides on a polymer support. J. Am. Chem. Soc. 103, 3185–3191 (1981).

16. Isobe, H., Fujino, T., Yamazaki, N., Guillot-Nieckowski, M. & Nakamura, E. Triazole-Linked Analogue of Deoxyribonucleic Acid (TLDNA): Design, Synthesis, and Double-Strand Formation with Natural DNA. Org. Lett. 10, 3729–3732 (2008).

17. Harris, J. M., Martin, N. E. & Modi, M. Pegylation: A Novel Process for Modifying Pharmacokinetics. Clin. Pharmacokinet. 40, 539–551 (2001).

18. Harada, A., Togawa, H. & Kataoka, K. Physicochemical properties and nuclease resistance of antisense-oligodeoxynucleotides entrapped in the core of polyion complex micelles composed of poly (ethylene glycol)–poly (l-lysine) block copolymers. Eur. J. Pharm. Sci. 13, 35–42 (2001).

19. Li, Z., Zhang, Y., Fullhart, P. & Mirkin, C. A. Reversible and Chemically Programmable Micelle Assembly with DNA Block-Copolymer Amphiphiles. Nano Lett. 4, 1055–1058 (2004).

20. Alemdaroglu, F. E., Alemdaroglu, N. C., Langguth, P. & Herrmann, A. DNA Block Copolymer Micelles – A Combinatorial Tool for Cancer Nanotechnology. Adv. Mater. 20, 899–902 (2008).

21. Kim, B. S. et al. A 50-nm-Sized Micellar Assembly of Thermoresponsive Polymer-Antisense Oligonucleotide Conjugates for Enhanced Gene Knockdown in Lung Cancer by Intratracheal Administration. Adv. Ther. 3, 1900123 (2020).

22. Jeong, J. H. & Park, T. G. Novel Polymer−DNA Hybrid Polymeric Micelles Composed of Hydrophobic Poly(d,l-lactic-co-glycolic Acid) and Hydrophilic Oligonucleotides. Bioconjug. Chem. 12, 917–923 (2001).

23. Kamps, A. C., Cativo, Ma. H. M., Chen, X.-J. & Park, S.-J. Self-Assembly of DNA-Coupled Semiconducting Block Copolymers. Macromolecules 47, 3720–3726 (2014).

24. Rodríguez‐Pulido, A. et al. Light-Triggered Sequence-Specific Cargo Release from DNA Block Copolymer–Lipid Vesicles. Angew. Chem. Int. Ed. 52, 1008–1012 (2013).

25. Vreeland, W. N. et al. Molar Mass Profiling of Synthetic Polymers by Free-Solution Capillary Electrophoresis of DNA−Polymer Conjugates. Anal. Chem. 73, 1795–1803 (2001).

26. Ghobadi, A. F. & Jayaraman, A. Effects of Polymer Conjugation on Hybridization Thermodynamics of Oligonucleic Acids. J. Phys. Chem. B 120, 9788–9799 (2016).

27. Shokrzadeh, N., Winkler, A.-M., Dirin, M. & Winkler, J. Oligonucleotides conjugated with short chemically defined polyethylene glycol chains are efficient antisense agents. Bioorg. Med. Chem. Lett. 24, 5758–5761 (2014).

28. Yang, L. et al. Self-Assembled Aptamer-Grafted Hyperbranched Polymer Nanocarrier for Targeted and Photoresponsive Drug Delivery. Angew. Chem. 130, 17294–17298 (2018).

29. Fouz, M. F. et al. Bright Fluorescent Nanotags from Bottlebrush Polymers with DNA-Tipped Bristles. ACS Cent. Sci. 1, 431–438 (2015).

30. Gibbs, J. M. et al. Polymer−DNA Hybrids as Electrochemical Probes for the Detection of DNA. J. Am. Chem. Soc. 127, 1170–1178 (2005).

31. Sicilia, G. et al. Programmable polymer-DNA hydrogels with dual input and multiscale responses. Biomater. Sci. 2, 203–211 (2014).

32. Murakami, Y. & Maeda, M. Hybrid hydrogels to which single-stranded (ss) DNA probe is incorporated can recognize specific ssDNA. Macromolecules 38, 1535–1537 (2005).

33. Peng, L. et al. Macroscopic Volume Change of Dynamic Hydrogels Induced by Reversible DNA Hybridization. J. Am. Chem. Soc. 134, 12302–12307 (2012).

34. Averick, S., Paredes, E., Li, W., Matyjaszewski, K. & Das, S. R. Direct DNA Conjugation to Star Polymers for Controlled Reversible Assemblies. Bioconjug. Chem. 22, 2030–2037 (2011).

35. Chen, P., Li, C., Liu, D. & Li, Z. DNA-Grafted Polypeptide Molecular Bottlebrush Prepared via Ring-Opening Polymerization and Click Chemistry. Macromolecules 45, 9579–9584 (2012).

36. Tanaka, S. et al. Bulk pH-Responsive DNA Quadruplex Hydrogels Prepared by Liquid-Phase, Large-Scale DNA Synthesis. ACS Macro Lett. 7, 295–299 (2018).

37. Guo, W. et al. Switchable Bifunctional Stimuli-Triggered Poly-N-Isopropylacrylamide/DNA Hydrogels. Angew. Chem. Int. Ed. 53, 10134–10138 (2014).

38. Yurke, B. Mechanical Properties of a Reversible, DNA-Crosslinked Polyacrylamide Hydrogel. J. Biomech. Eng. 126, 104 (2004).

39. Du, C. & Hill, R. J. Complementary-DNA-Strand Cross-Linked Polyacrylamide Hydrogels. Macromolecules 52, 6683–6697 (2019).

40. Liedl, T., Dietz, H., Yurke, B. & Simmel, F. Controlled Trapping and Release of Quantum Dots in a DNA-Switchable Hydrogel. Small 3, 1688–1693 (2007).

41. Wei, B., Cheng, I., Luo, K. Q. & Mi, Y. Capture and Release of Protein by a Reversible DNA- Induced Sol–Gel Transition System. Angew. Chem. Int. Ed. 47, 331–333 (2008).

42. Jiang, F. X., Yurke, B., Firestein, B. L. & Langrana, N. A. Neurite Outgrowth on a DNA Crosslinked Hydrogel with Tunable Stiffnesses. Ann. Biomed. Eng. 36, 1565–1579 (2008).

43. Cutler, J. I., Auyeung, E. & Mirkin, C. A. Spherical Nucleic Acids. J. Am. Chem. Soc. 134, 1376–1391 (2012).

44. Choi, C. H. J., Hao, L., Narayan, S. P., Auyeung, E. & Mirkin, C. A. Mechanism for the endocytosis of spherical nucleic acid nanoparticle conjugates. Proc. Natl. Acad. Sci. 110, 7625–7630 (2013).

45. Samanta, D. et al. Multivalent Cation-Induced Actuation of DNA-Mediated Colloidal Superlattices. J. Am. Chem. Soc. 141, 19973–19977 (2019).

46. Morris, W., Briley, W. E., Auyeung, E., Cabezas, M. D. & Mirkin, C. A. Nucleic Acid–Metal Organic Framework (MOF) Nanoparticle Conjugates. J. Am. Chem. Soc. 136, 7261–7264 (2014).

47. Karimizefreh, A., Mahyari, F. A., VaezJalali, M., Mohammadpour, R. & Sasanpour, P. Impedimetic biosensor for the DNA of the human papilloma virus based on the use of gold nanosheets. Microchim. Acta 184, 1729–1737 (2017).

48. Lee, K., Rouillard, J.-M., Pham, T., Gulari, E. & Kim, J. Signal-Amplifying Conjugated Polymer–DNA Hybrid Chips. Angew. Chem. Int. Ed. 46, 4667–4670 (2007).

49. Fairbanks, B. D., Culver, H. R., Mavila, S. & Bowman, C. N. Towards High-Efficiency Synthesis of Xenonucleic Acids. Trends Chem. 2, 43–56 (2020).

50. Egholm, M., Buchardt, O., Nielsen, P. E. & Berg, R. H. Peptide nucleic acids (PNA). Oligonucleotide analogs with an achiral peptide backbone. J. Am. Chem. Soc. 114, 1895–1897 (1992).

51. Porcheddu, A. & Giacomelli, G. Peptide nucleic acids (PNAs), a chemical overview. Curr. Med. Chem. 12, 2561–2599 (2005).

52. Egholm, M. et al. PNA hybridizes to complementary oligonucleotides obeying the Watson- Crick hydrogen bonding rules. Nature 365, 566–568 (1993).

53. Zhang, G.-J. et al. Highly sensitive measurements of PNA-DNA hybridization using oxide- etched silicon nanowire biosensors. Biosens. Bioelectron. 23, 1701–1707 (2008).

54. Shiraishi, T. & Nielsen, P. E. Down-regulation of MDM2 and activation of p53 in human cancer cells by antisense 9-aminoacridine–PNA (peptide nucleic acid) conjugates. Nucleic Acids Res. 32, 4893–4902 (2004).

55. Shiraishi, T., Hamzavi, R. & Nielsen, P. E. Subnanomolar antisense activity of phosphonate- peptide nucleic acid (PNA) conjugates delivered by cationic lipids to HeLa cells. Nucleic Acids Res. 36, 4424–4432 (2008).

56. Gildea, B. D. et al. PNA solubility enhancers. Tetrahedron Lett. 39, 7255–7258 (1998).

57. Nielsen, P. E., Haaima, G., Lohse, A. & Buchardt, O. Peptide Nucleic Acids (PNAs) Containing Thymine Monomers Derived from Chiral Amino Acids: Hybridization and Solubility Properties of D-Lysine PNA. Angew. Chem. Int. Ed. Engl. 35, 1939–1942 (1996).

58. Mier, W., Eritja, R., Mohammed, A., Haberkorn, U. & Eisenhut, M. Peptide–PNA Conjugates: Targeted Transport of Antisense Therapeutics into Tumors. Angew. Chem. Int. Ed. 42, 1968– 1971 (2003).

59. Lee, S. H., Moroz, E., Castagner, B. & Leroux, J.-C. Activatable Cell Penetrating Peptide– Peptide Nucleic Acid Conjugate via Reduction of Azobenzene PEG Chains. J. Am. Chem. Soc. 136, 12868–12871 (2014).

60. Turner, J. J. et al. Cell-penetrating peptide conjugates of peptide nucleic acids (PNA) as inhibitors of HIV-1 Tat-dependent trans -activation in cells. Nucleic Acids Res. 33, 6837–6849 (2005).

61. Pujals, S. & Giralt, E. Proline-rich, amphipathic cell-penetrating peptides. Adv. Drug Deliv. Rev. 60, 473–484 (2008).

62. Zorko, M. & Langel, Ü. Cell-penetrating peptides: mechanism and kinetics of cargo delivery. Adv. Drug Deliv. Rev. 57, 529–545 (2005).

63. Liu, L.-H. et al. Self-Assembly of Hybridized Peptide Nucleic Acid Amphiphiles. ACS Macro Lett. 3, 467–471 (2014).

64. Tsukada, H. et al. Quantitative single-nucleotide polymorphism analysis in secondary- structured DNA by affinity capillary electrophoresis using a polyethylene glycol–peptide nucleic acid block copolymer. Anal. Biochem. 433, 150–152 (2013).

65. Berthold, P. R., Shiraishi, T. & Nielsen, P. E. Cellular Delivery and Antisense Effects of Peptide Nucleic Acid Conjugated to Polyethyleneimine via Disulfide Linkers. Bioconjug. Chem. 21, 1933–1938 (2010).

66. L. Turner, J., L. Becker, M., Li, X., A. Taylor, J.-S. & L. Wooley, K. PNA-directed solution- and surface- assembly of shell crosslinked (SCK) nanoparticle conjugates. Soft Matter 1, 69– 78 (2005).

67. Khadsai, S. et al. Poly(acrylic acid)-grafted magnetite nanoparticle conjugated with pyrrolidinyl peptide nucleic acid for specific adsorption with real DNA. Colloids Surf. B Biointerfaces 165, 243–251 (2018).

68. Mateo-Martí, E., Briones, C., Pradier, C. M. & Martín-Gago, J. A. A DNA biosensor based on peptide nucleic acids on gold surfaces. Biosens. Bioelectron. 22, 1926–1932 (2007).

69. Hüsken, N., Gębala, M., Schuhmann, W. & Metzler-Nolte, N. A Single-Electrode, Dual- Potential Ferrocene–PNA Biosensor for the Detection of DNA. ChemBioChem 11, 1754–1761 (2010).

70. Galli, M. et al. Superparamagnetic iron oxide nanoparticles functionalized by peptide nucleic acids. RSC Adv. 7, 15500–15512 (2017).

71. Shrestha, R., Shen, Y., Pollack, K. A., Taylor, J.-S. A. & Wooley, K. L. Dual Peptide Nucleic Acid- and Peptide-Functionalized Shell Cross-Linked Nanoparticles Designed to Target mRNA toward the Diagnosis and Treatment of Acute Lung Injury. Bioconjug. Chem. 23, 574– 585 (2012).

72. Singh, S. K., Koshkin, A. A., Wengel, J. & Nielsen, P. LNA (locked nucleic acids): synthesis and high-affinity nucleic acid recognition. Chem. Commun. 0, 455–456 (1998).

73. Koshkin, A. A. et al. LNA (Locked Nucleic Acids): Synthesis of the adenine, cytosine, guanine, 5-methylcytosine, thymine and uracil bicyclonucleoside monomers, oligomerisation, and unprecedented nucleic acid recognition. Tetrahedron 54, 3607–3630 (1998).

74. Koshkin, A. A. et al. LNA (Locked Nucleic Acid): An RNA Mimic Forming Exceedingly Stable LNA:LNA Duplexes. J. Am. Chem. Soc. 120, 13252–13253 (1998).

75. Mishra, S., Ghosh, S. & Mukhopadhyay, R. Ordered Self-Assembled Locked Nucleic Acid (LNA) Structures on Gold(111) Surface with Enhanced Single Base Mismatch Recognition Capability. Langmuir 28, 4325–4333 (2012).

76. Mishra, S., Ghosh, S. & Mukhopadhyay, R. Maximizing Mismatch Discrimination by Surface- Tethered Locked Nucleic Acid Probes via Ionic Tuning. Anal. Chem. 85, 1615–1623 (2013).

77. Wang, K. et al. Design of a sandwich-mode amperometric biosensor for detection of PML/RARα fusion gene using locked nucleic acids on gold electrode. Biosens. Bioelectron. 26, 2870–2876 (2011).

78. Chu, T.-W., Yang, J., Zhang, R., Sima, M. & Kopeček, J. Cell Surface Self-Assembly of Hybrid Nanoconjugates via Oligonucleotide Hybridization Induces Apoptosis. ACS Nano 8, 719–730 (2014).

79. Langford, G. J., Raeburn, J., Ferrier, D. C., Hands, P. J. W. & Shaver, M. P. Morpholino Oligonucleotide Cross-Linked Hydrogels as Portable Optical Oligonucleotide Biosensors. ACS Sens. 4, 185–191 (2019).

80. Liao, T. et al. Ultrasensitive Detection of MicroRNAs with Morpholino-Functionalized Nanochannel Biosensor. Anal. Chem. 89, 5511–5518 (2017).

81. Gao, Z. & Ping Ting, B. A DNA biosensor based on a morpholino oligomer coated indium- tin oxide electrode and a cationic redox polymer. Analyst 134, 952–957 (2009).

82. Zu, Y., Ting, A. L., Yi, G. & Gao, Z. Sequence-Selective Recognition of Nucleic Acids under Extremely Low Salt Conditions Using Nanoparticle Probes. Anal. Chem. 83, 4090–4094 (2011).

83. Fan, J., Zeng, F., Wu, S. & Wang, X. Polymer Micelle with pH-Triggered Hydrophobic– Hydrophilic Transition and De-Cross-Linking Process in the Core and Its Application for Targeted Anticancer Drug Delivery. Biomacromolecules 13, 4126–4137 (2012).

84. Mather, B. D. et al. Supramolecular Triblock Copolymers Containing Complementary Nucleobase Molecular Recognition. Macromolecules 40, 6834–6845 (2007).

85. Spijker, H. J., Dirks, A. J. & Hest, J. C. M. van. Synthesis and assembly behavior of nucleobase-functionalized block copolymers. J. Polym. Sci. Part Polym. Chem. 44, 4242–4250 (2006).

86. Kang, Y. et al. Use of complementary nucleobase-containing synthetic polymers to prepare complex self-assembled morphologies in water. Polym. Chem. 7, 2836–2846 (2016).

87. Hua, Z. et al. Micellar nanoparticles with tuneable morphologies through interactions between nucleobase-containing synthetic polymers in aqueous solution. Polym. Chem. 7, 4254–4262 (2016).

88. Bazzi, H. S. & Sleiman, H. F. Adenine-Containing Block Copolymers via Ring-Opening Metathesis Polymerization: Synthesis and Self-Assembly into Rod Morphologies. Macromolecules 35, 9617–9620 (2002).

89. Kuang, H. et al. Biodegradable Amphiphilic Copolymer Containing Nucleobase: Synthesis, Self-Assembly in Aqueous Solutions, and Potential Use in Controlled Drug Delivery. Biomacromolecules 13, 3004–3012 (2012).

90. Kuang, H. et al. Core-crosslinked amphiphilic biodegradable copolymer based on the complementary multiple hydrogen bonds of nucleobases : synthesis, self-assembly and in vitro drug delivery. J. Mater. Chem. 22, 24832–24840 (2012).

91. Xi, W. et al. Clickable Nucleic Acids: Sequence-Controlled Periodic Copolymer/Oligomer Synthesis by Orthogonal Thiol-X Reactions. Angew. Chem. Int. Ed. 54, 14462–14467 (2015).

92. Saenger, W. Principles of Nucleic Acid Structure. (Springer Science & Business Media, 2013).

93. Han, X. et al. New Generation of Clickable Nucleic Acids: Synthesis and Active Hybridization with DNA. Biomacromolecules 19, 4139–4146 (2018).

94. Harguindey, A. et al. Click Nucleic Acid Mediated Loading of Prodrug Activating Enzymes in PEG–PLGA Nanoparticles for Combination Chemotherapy. Biomacromolecules 20, 1683– 1690 (2019).

Chapter 2 - Objectives and Scope

The field of nucleic acid research is fast growing. Chapter 1 discussed the various ways that oligonucleotides, natural or modified/non-natural, have been exploited for their ability to interact specifically with various types of molecules. However, the biggest setback for most strategies to date has been the ability to synthesize oligonucleotides quickly and at scale. Click

Nucleic Acids (CNAs), synthesized through a thiol-ene mediated polymerization mechanism, were developed as a promising alternative to existing oligonucleotides. The use of thiol-ene polymerization begets two important features of CNA synthesis. The first, is the quick reaction kinetics and tolerance to environmental conditions. Thiol-ene reactions normally reach full conversion in seconds to minutes and are tolerant to many temperatures, solvent conditions, and exposure to air (unlike other forms of radical polymerization). The second feature is that the resulting CNA oligomer maintains a free thiol, which can be easily functionalized with a variety of click-chemistry approaches (i.e. thiol/maleimide, thiol/halogen, or thiol/nitrile). Further, thiol- containing molecules and polymers can be doped into the polymerization as a simple, one-pot approach to create CNA conjugates.

The scope of this dissertation is to formulate a basic understanding of how the interactions between CNA containing polymers and conjugates can be utilized to develop functional materials with a variety of potential applications. Specifically, by using a copolymerization technique between CNA monomers and thiolated, poly(ethylene glycol) (PEG) precursors, this work aim to establish a synthetic strategy for creating a family of materials with CNA functionalities that is applicable to a wide variety of polymer architectures. The main objectives for this work are to demonstrate that CNA homopolymers and hybrid copolymers interact with and bind to nucleic

28 acid recognizing structures and to show that copolymer structure can be leveraged to enable new functionalities that arise from the CNAs’ fundamental interactions. These goals will be achieved through the following specific aims:

2.1 Aim 1: Identify interactions between linear CNA polymers and copolymers with cells

cellular components

This Aim’s goal is to identify the fundamental interactions of linear, CNA polymers and copolymers with in vitro biological systems to establish an understanding of their properties in the context of cellular structures. As an unfunctionalized oligomer, CNA is not water-soluble, primarily due to the lack of backbone charge. To investigate the ability of CNA to interact with structures in a primarily aqueous environment, it must be conjugated to a water-soluble polymer.

Chapter 3 reports a linear PEG-CNA block copolymer a investigates its cytocompatibility, cellular uptake, and cellular localization. However, Chapter 4 finds that despite its hydrophobicity, unfunctionalized CNA interacts with complementary nucleic acids in majority aqueous environments and facilitates coprecipitation. This work exploits this property as a simple and economic mRNA isolation strategy.

2.1.1 Sub-aim 1.1

As a nucleic acid mimic, CNA has obvious applications in biotechnology and one such application is as an antisense oligonucleotide. To realize this potential application, a fundamental understanding of the interactions between CNA and cells is necessary, which is the focus of this sub-aim. Chapter 3 covers work completed towards this goal. Once solubilized with a linear PEG and tagged with a fluorophore, the conjugate was incubated with cultured cells and its cytocompatibility, cellular uptake, and cellular localization was determined.

2.1.2 Sub-aim 1.2

This sub-aim is addressed in Chapter 4, where linear oligo(T) CNA is used as an mRNA isolation tool, capitalizing on its ability to bind to complementary nucleic acids. While conjugation is required to solubilize CNA in aqueous conditions, its inherent insolubility can be used to instigate precipitation of bound molecules. Here, optimization of the isolation procedure is performed, and the isolation yield is compared to the yield obtained through a commercially available product. Finally, the functionality of the mRNA is assessed by in vitro translation and

RT-qPCR.

2.2 Aim 2: Utilize the interactions between a multi-armed CNA-polymer conjugate and

DNA to create a stimuli-responsive gel network

Aim 2’s purpose is to investigate the ability of multifunctional PEG-CNA copolymers to form viscoelastic and stimuli-responsive 3D gel networks through nucleic acid hybridization with single stranded DNA (ssDNA) and characterize the material properties of the gels. Dynamic, stimuli-responsive networks usually make use of dynamic covalent bonds (i.e. Diels-Alder or thioester reactions) or physical interactions (i.e. electrostatics or hydrogen bonding). In Chapter

5, the hydrogen bonding capability of CNAs is exploited as a crosslinking interaction to give rise to dynamic networks with interesting properties. Specifically, these complex crosslinking interactions that lead to gel formation are studied, primarily by evaluating the gels viscoelasticity, thermoreversibility, and stress relaxation.

2.3 Aim 3: Develop a higher order CNA-polymer conjugate capable of loading DNA for

therapeutic delivery applications

Finally, the objective of Aim 3 is to demonstrate the polymerizability of CNA monomers in the presence of monodisperse, thiolated microspheres to create highly functionalized particles 30 and study their interactions with complementary nucleic acids. Microparticles have been used extensively for gene delivery applications, particularly to macrophages. Conjugating CNA to these particles, as shown in Chapter 6, introduces the ability to specifically interact with nucleic acids, which provides more control over the loading and release of the oligonucleotide. Specifically, this aim characterizes the CNA functionalized microparticles and demonstrates their ability to load and release oligonucleotides as a form of drug delivery.

Chapter 3 - Cytocompatibility and Cellular Internalization of PEGylated “Clickable” Nucleic Acid Oligomers

As published in Biomacromolecules, 2018

3.1 Abstract

The recently developed synthetic oligonucleotides referred to as “Click” Nucleic Acids

(CNAs) are promising due to their relatively simple synthesis based on thiol-X reactions with numerous potential applications in biotechnology, biodetection, gene silencing, and drug delivery.

Here, the cytocompatibility and cellular uptake of rhodamine tagged, PEGylated CNA copolymers

(PEG-CNA-RHO) were evaluated. NIH 3T3 fibroblast cells treated for 1 hour with 1, 10 or 100

μg/mL PEG-CNA-RHO maintained an average cell viability of 86%, which was not significantly different from the untreated control. Cellular uptake of PEG-CNA-RHO was detected within 30 seconds and the amount internalized increased over the course of 1 hour. Moreover, these copolymers were internalized within cells to a higher degree than controls consisting of either rhodamine tagged PEG or the rhodamine alone. Uptake was not affected by temperature (i.e., 4°C or 37°C), suggesting a passive uptake mechanism. Subcellular colocalization analysis failed to indicate significant correlations between the internalized PEG-CNA-RHO and the organelles examined (mitochondria, endoplasmic reticulum, endosomes and lysosomes). These results indicate that CNA copolymers are cytocompatible and are readily internalized by cells, supporting the idea that CNAs are a promising alternative to DNA in antisense therapy applications.

3.2 Introduction

Nucleic acids represent one of the most powerful, functional biological materials due to their ability to control and direct biological function at all levels and to transmit genetic information from one generation to the next. These functions are possible solely because of the specific nucleotide sequence of each polymer and the nucleobase’s ability to recognize and hybridize with its complementary base. These characteristics are what has made nucleic acids a popular topic in both material science and biotechnology. Specifically, the ability of DNA to recognize and bind to highly specific genetic sequences has led to its use as antisense agents in gene therapy. Significant reductions in protein expression, and thereby a modulation of disease symptoms, have been achieved by delivering oligomeric DNA that is complementary to targeted sequences of either DNA or mRNA1–3. While gene therapy has been a major area of research4,5, the field has largely been hindered by the inefficient cellular delivery of DNA, instability of foreign

DNA in biological environments, and the limited scale of DNA synthesis.

Several techniques have been developed to circumvent these shortcomings. One common strategy is to condense antisense oligonucleotides into complexed nanoparticle delivery systems6.

This strategy has been shown to increase transfection efficiency as well as protect them from extracellular degradation. However, these nanoparticles have a complicated mechanism of action, often require the use of cytotoxic polymers7, and once the oligonucleotides are released into the cytosol, they are vulnerable to intracellular degradation8,9. To prevent oligonucleotide degradation, the nucleotide structure can be chemically modified such as in phosphorothioate nucleic acids (PS-

DNA) or peptide nucleic acids (PNAs). By altering or eliminating the phosphodiester bond, PS-

DNA, PNAs, and other types of modified nucleic acids have proven to be nuclease resistant10–12.

In some cases, they have been shown to have higher transfection efficiencies13. However, like 33

DNA synthesis, the assembly of most modified nucleic acids oligomers relies on a complicated solid-phase strategy14,15, which is low yield and difficult to purify, time consuming, and can require the use of hazardous chemicals and solvents16.

In contrast, recently reported “Click” Nucleic Acids (CNAs) are nucleic acid oligomers synthesized by taking advantage of thiol-mediated click reactions17. CNA monomers bare a resemblance to DNA nucleotides in that they contain a pendent nucleobase; however, they differ in that the backbone contains a thioether structure that arises from the thiol-X reactions that are used to form the CNA. By employing conditions that favor the correct thiol-X reaction, carefully designed CNA monomers are polymerized into CNA polymer chains quickly and under ambient settings, making them an interesting and attractive alternative form of oligomeric nucleobases.

CNA structure eliminates the phosphodiester bond, rendering it invisible to both endo- and exonucleases18,19. Furthermore, while complete specific and arbitrary sequence control is still out of reach, sets of orthogonal thiol-X reactions can be used to build nucleic acid sequences that are polymerized in a single step to yield repetitive sequences of nucleotides. Thus, CNAs have the potential to be used for antisense gene therapy strategies, either as the antisense agent itself for treatment of diseases caused by repeat expansions20 or as a complexation agent for DNA delivery21. In addition, they may offer many advantages over current antisense agents, such as nuclease resistance and a more facile synthesis. However, before CNAs can be realized for such applications, it is necessary to understand whether they will be compatible with biological systems.

This study investigated cytocompatibility, cell uptake properties, and interactions with subcellular organelles of CNAs in vitro. Copolymers of CNA and poly(ethylene glycol) (PEG) were synthesized to create water soluble CNA conjugates. PEGylated CNAs functionalized with a fluorophore were incubated with cells and cytocompatibility was determined by measuring cell 34 metabolism. Cellular internalization was assessed by visualizing up-take with fluorescence microscopy. Finally, immunofluorescent techniques were employed to study subcellular colocalization of CNA copolymers upon internalization.

3.3 Experimental

3.3.1 PEG-CNA-RHO and PEG-RHO conjugate Synthesis.

The thymine CNA monomer was synthesized as previously described16. Photo-initiated copolymerization was carried out by dissolving thiolated poly(ethylene glycol) (PEG-SH) (MW

2000 g/mol, Sigma Aldrich) at 100 mM and thymine CNA monomer at 1 M in DMF containing

0.01 wt. % 2,2-Dimethoxy-2-phenylacetophenone (DMPA). After irradiation with 365 nm light for 15 minutes, the resulting material was precipitated into cold diethyl ether. The supernatant was decanted leaving the solid copolymer (PEG-CNA), which was dried and resuspended in DI H2O.

After agitating the polymer suspension overnight, the insoluble polymer fraction was spun down and the supernatant was collected and lyophilized, yielding water soluble PEG-CNA whose presence was confirmed by 1H NMR (Bruker AV-III 400 MHz) and GPC (Tosoh HLC®-

8320GPC). In all cases, GPC samples were run in DMSO using two detectors, Refractive Index

(RI) and Ultraviolet (UV). Average molecular weights were calculated using PMMA standards.

PEG-CNA was then dissolved at 500 μM in PBS in a scintillation vial and dithiols were reduced with tris(2-carboxyethyl)phosphine (TCEP) (Chem-Impex International) for 20 minutes while the solution was purged with inert gas. Then, under an inert atmosphere, 2.5 eq. of Rhodamine Red

(RHO) C2 maleimide dissolved in DMSO (7.4 mM) was added to the stirring PEG-CNA solution and allowed to react for 2-3 hours in the dark. The reaction solution was then dialyzed in the dark with a 1 kDa MWCO against DI H2O for 2 days (until the diasylate appeared colorless) yielding a pink fluorescent solution. The dialyzed PEG-CNA-RHO was then lyophilized and analyzed by 1H 35

NMR and GPC. The extent of labelling by the dye was determined by UV-VIS. The lyophilized

PEG-CNA-RHO powder was then reconstituted at a concentration of 1 mg/mL in PBS. This solution was sterile filtered with a 0.22 μm, PVDF, sterile syringe filter and frozen in aliquots.

PEG-RHO was synthesized in a similar manner by reacting PEG-SH (MW 2000 g/mol) with

Rhodamine Red C2 maleimide. Once the fluorescent molecule was introduced, care was taken to protect the compound from exposure to light.

3.3.2 Cell Culture.

NIH 3T3 fibroblast cells obtained from ATCC were cultured in DMEM growth media supplemented with 10% FBS and 1% antibiotics at 37°C and 5% CO2. Cells were passaged every

3-5 days or when 90% confluency was reached with 0.5% trypsin-EDTA and replating at a 1:10 ratio.

3.3.3 Cell viability.

Viability was assessed by an MTT assay. Briefly, cells were seeded in a 96 well plate at

800 cells/well and allowed to attach and grow for 48 hours. Cells were then treated with varying concentrations of PEG-CNA-RHO and PEG-RHO conjugate (0-100 μg/mL) in triplicate for 1 hour. A positive control was performed by adding a small punch of a powdered latex glove to the wells. After treatment, cells were washed with PBS and cultured in fresh media for another 48 hours. The assay was performed by replacing the media in each well with 100 μL of fresh media and adding 10 μL of 12 mM MTT to each well after which the plate was incubated at 37°C and

5% CO2 for 4-6 hours. The resulting crystal precipitates were dissolved by adding 100 μL of 0.1 g/mL sodium dodecyl sulfate in 0.1 M HCl after which the solution’s absorbance was measured at

570 nm.

3.3.4 Cellular uptake of fluorescent conjugates.

Cells were plated in a high quality, glass bottom 96 well plate (ibidi® μ-Plate) at a concentration of 4,000-7,000 cells/well and allowed to adhere and grow overnight at 37°C and 5%

CO2. The culture medium was replaced with 135 μL of fresh media to which 15 μL of the 1 mg/mL

PEG-CNA-RHO solution was added. The plate was swirled gently to allow mixing. After incubation for 1 hour, the cells were washed with PBS, and fresh media was added to the wells.

3.3.5 Time dependent cellular uptake and localization.

Cells were plated and treated with PEG-CNA-RHO as described previously, but with varying incubation times from 30 seconds to 6 hours. After incubation, the cells were washed with

PBS and immediately fixed with 4% paraformaldehyde for 10 minutes at room temperature. After washing with PBS, cells were incubated with fresh PBS overnight at 4°C, protected from light.

3.3.6 Immunofluorescence.

Cells were plated and treated with PEG-CNA-RHO the same way as before. After treatment, fresh media was added to the wells and the cells were immediately imaged to verify uptake. Cells were then fixed with 4% paraformaldehyde for 20 minutes at room temperature, after which they were washed twice with PBS. Cells were permeabilized with 0.5% Triton X-100 for 5 minutes at room temperature, washed twice with PBS, and then blocked with 3% BSA for 60 minutes at room temperature. After rinsing with PBS, cells were stained with primary antibodies

(Endoplasmic Reticulum: calreticuluin antibody, rabbit anti-mouse, Novus Biologicals®, used at

1:90 dilution in 3% BSA; Mitochondria: ATPB antibody, mouse monoclonal, Abcam®, used at

1:500 dilution in 3% BSA; Endosome: EEA1 antibody, mouse monoclonal, R&D Systems ®, used at 1:50 dilution in 3% BSA) for 1 hour at room temperature. After rinsing wells three times with

PBS, all traces of the primary antibody were removed by incubating with 0.01% Tween-20 for 1 37 hour at room temperature. The cells were washed with PBS twice and then stained with the fluorescent secondary antibody (AlexaFluor 488, donkey anti-mouse, Invitrogen™, used at 1:200 dilution in 3% BSA; AlexaFluor 647 donkey anti-rabbit, Invitrogen™, used at 1:200 dilution in

3% BSA) for 1 hour at room temperature in the dark. After washing with PBS three times, F-actin was stained with and phallotoxins (Phalloidin 488 and Phalloidin 647, Invitrogen™, used at 1:60 dilution in 3% BSA) for 20 minutes at room temperature. Cells were washed with PBS twice, and then stained with Hoechst 33342 (Invitrogen™, used at 1 μg/mL in 3% BSA) for 5 minutes at room temperature. Finally, cells were washed with PBS twice and stored in PBS at 4°C until imaged.

3.3.7 Fluorescence microscopy.

Cells with fluorescent components were imaged on a Nikon Spinning Disc Confocal microscope. To evaluate the extent of cellular uptake, a 40x air objective was used. For single cell images, a 100x oil immersion objective was used. For all cases, multiple images across multiple replicate wells were taken. Reported images are representative of the corresponding experimental conditions. All image analysis was completed with the ImageJ software.

3.3.8 Statistical Analysis.

Data presented are means of replicates (n, indicated in figure captions) with error bars representing standard deviations. Statistical analysis for the MTT assay and cell uptake assay was performed using a One-Way ANOVA with respect to cell viability and average fluorescence intensity per cell, respectively. Statistical significance was defined at the 95% confidence limit (p

< 0.05). In cases where significant differences were present, comparisons were made using

Tukey’s Post-Hoc analysis to identify differences between groups.

3.4 Results and Discussion

To begin evaluating CNAs’ cytocompatibility, linear, PEGylated CNAs were synthesized to increase the hydrophilicity of the CNA chain. Due to the lack of charge and hydrophilic groups on the CNA backbone, CNA homopolymers exhibit poor water solubility. It has also been shown that the addition of PEG to biomaterials can help increase the material’s bioavailability22, protect the material from degradation23, and stabilize nucleic acid hybridization interactions24.

Photoinitiated copolymerization of a thiolated PEG (molecular weight of 2000 g/mol) and thymine

CNA monomers yielded a linear block-copolymer, named PEG-CNA. This type of copolymerization was chosen as the preferred method of synthesis because it is a simple, one-step reaction and because it results in a thiol end group on the CNA block which can subsequently be functionalized with a fluorophore. Gel permeation chromatography of this copolymer showed an increase in molecular weight from the PEG-SH, suggesting successful copolymerization, along with a low-molecular weight peak that is attributed to the cyclized monomer (Supporting Figure

3.1). The water-soluble fraction of the block-copolymer was isolated, and GPC and 1H NMR analysis confirmed that this copolymer contained an average of 5 thymine repeat units per chain

(as obtained by peak integration with respect to the PEG methoxy peak) (See Supporting Figure

3.5). It should be noted that the degree of polymerization calculated here is less than our original molar excess of thymine monomer (10:1, thymine monomer:PEG-SH). This behavior is because the cyclized monomer that is observed in the GPC trace, unreacted CNA homopolymers, and copolymers with higher degrees of polymerization are less likely to be solubilized by the water due to their high CNA content.

Finally, a fluorophore (Rhodamine Red C2 maleimide) was conjugated to the sulfhydryl terminated end of the CNA chain, to yield the PEG-CNA-RHO conjugate. Effective conjugation 39

Scheme 3.1 - In vitro strategy for evaluating the cytocompatibility of PEGylated CNA copolymers

was confirmed by 1H NMR analysis and UV-VIS (Supporting Figures 3.6 and 3.7). The GPC trace of the resulting compound showed that the major peak is maintained and that many low molecular weight peaks from the initial copolymerization disappear (Supporting Figure 3.8).

Experiments were conducted using the strategy summarized in Scheme 3.1. Internalization experiments were performed by incubating cells with known concentrations of PEG-CNA-RHO, fluorescent PEG (PEG-RHO), or rhodamine dye alone (RHO), followed by subsequent washing.

In this study, NIH 3T3 fibroblast cells were employed, which have been used previously to investigate cytocompatibility and cell uptake25–27.

The cytocompatibility of PEG-CNA-RHO was determined by the MTT assay, which measures cellular metabolism. Since metabolic activity is directly proportional to cell number, the

MTT assay is an indirect measurement of cell viability, or how many cells survive after being treated with a molecule of interest. Concentrations of PEG-CNA-RHO ranging from 0 to 100

Figure 3.1 – An MTT assay confirms the low toxicity of the synthetic conjugates delivered in this study. Tukey’s post hoc comparison failed to identify significant differ-ences in viability at a 95% confidence limit. (n = 3, error bars represent standard deviations). μg/mL were incubated with NIH 3T3 cells for a period of one hour. Over the concentration range tested, cell metabolic activity was not statistically different from the un-treated control (Figure

3.1) indicating cytocompatibility. As a positive control, cells were incubated with a powdered latex glove for the same amount of time, which reduced cell metabolism to ~20%28. Cells were also treated with PEG-RHO in the same concentration range as above. Cell viability after treatment with PEG-CNA-RHO was not statistically different from the viability of cells treated with PEG-

RHO, which indicates that the CNA block does not add any significant toxicity.

To examine the effect of the PEG-CNA-RHO copolymer structure on cell uptake, cells were incubated with PEG-CNA-RHO, fluorescent PEG (PEG-RHO), and the fluorescent dye alone

(RHO). Figures 2 shows the extent of cellular uptake of representative samples after treatment

3.2 are cells that have internalized PEG-CNA-RHO. In contrast, the lack of fluorescence in the other panels (PEG-RHO and RHO alone, respectively) indicates that comparatively less uptake was observed. Semi-quantitative analysis confirms that cells incubated with PEG-CNA-RHO exhibited a statistically higher average fluorescence intensity than cells treated with PEG-RHO or

RHO alone. These results suggest that the presence of CNA in the copolymer facilitates uptake by cells.

Uptake kinetics were investigated by incubating cells with 100 μg/mL of PEG-CNA-RHO for various times ranging from 30 seconds to 1 hour. This concentration was chosen because it was shown to be cytocompatible as well as visible with microscopy. Representative confocal 42

Figure 3.3 – Incubating cells for increasingly longer times leads to an increased average cell- associated fluorescence. Representative images show cell fluorescence at 30 seconds, 10 minutes, 30 minutes, and 1 hour. Fluorescent intensity appears to begin to level off on the order of hours. (n>30, error bars represent standard deviations). microscopy images and semi-quantitative image analysis show an increase in cell-associated fluorescence over time, which are shown in Figure 3.3. All further experiments were conducted with a 1 hour incubation step because it appears that PEG-CNA-RHO uptake levels are easily visualized on this time scale.

Subcellular localization of the CNA copolymers was studied through immunofluorescence microscopy. In short, cells were treated with a known concentration of PEG-CNA-RHO (red fluorescence) for 1 hour, fixed, and then stained with organelle-specific fluorophore-antibody conjugates (green fluorescence). By measuring the presence and intensity of red fluorescence against green fluorescence for each pixel of an image, a semi-quantitative degree of colocalization is determined in the form of the Pearson coefficient or Mander’s coefficient29,30. This phenomenon is also visually identified in an overlay image by the presence of yellow pixels.

In all, four organelles were stained: endosomes, endoplasmic reticulum, mitochondria, and lysosomes. Additionally, nuclei and actin were stained to aid in cellular identification. Resulting

Figure 3.4 – Three representative images of immunofluorescence microscopy images highlight the cellular localization of PEG-CNA-rho conjugates. Cells were treated with 100 μg/mL of PEG- CNA-rho conjugates for 1 hr, then fixed, permeabilized, and stained. Blue represents the nucleus (Hoechst 33342), red represents the PEG-CNA-RHO conjugate, and green represents either endosomes, the endoplasmic reticulum, mitochondria, or lysosomes. Areas of yellow indicate colocalization between the PEG-CNA-rho conjugate and the indicated organelle. Areas of magenta indicate colocalization between the PEG-CNA-rho conjugate and the nucleus. overlay images of individual cells are shown in Figure 3.4. The localization of PEG-CNA-RHO

(red fluorescence) in each image shows a similar motif: the presence of small punctate areas of high fluorescence amid a mostly diffuse pattern within the cell. The small punctate areas within these images could suggest the sequestration of PEG-CNA-RHO by an organelle, or could simply be aggregation of the polymer in the highly complex intracellular environment. One phenomenon 44 to mention is that in some images it appears that the PEG-CNA-RHO and organelles look to be negatively correlated. This behavior is seen in the images staining for mitochondria, lysosomes, and the endoplasmic reticulum and are indicated by the white arrows. It should also be noted that there is evidence of PEG-CNA-RHO in the nucleus, which is seen by the magenta color in the images in Figure 3.4. Previous studies have shown that nuclear uptake of PS-DNA (and to some extent, natural phosphodiester linked DNA) occurs to an extent31,32. However, some reports have warned that this result could be an artifact of the fixation and permeabilization process33.

To calculate the Pearson and Mander’s coefficients, the JACoP ImageJ plugin was used to determine the overlap between the green and red channels of the overlay images. The Pearson coefficient generally measures how well the intensities of each channel are correlated for every pixel of an image. The Mander’s coefficient takes this one step further and can differentiate between overlap of the green channel with the red, or the red channel with the green. For both coefficients, only values close to 1 indicate significant colocalization, while all other values do not allow conclusions to be drawn34. In all images, colocalization coefficients were no higher than 0.4, suggesting that no conclusions can be made regarding colocalization between PEG-CNA-RHO and any of the organelles examined. However, this does not mean that there is no colocalization between PEG-CNA-RHO and subcellular organelles. It is possible that significant interactions are occurring, but the uncertainty associated with immunofluorescence techniques does not allow for its detection. For instance, given that the CNA block contains repeats of thymine, we might expect the copolymer to bind to structures containing adenine residues, like the poly-A tail of mRNA.

However, the immunofluorescence techniques here do not allow for the detection of such specific interactions.

Figure 3.5 – Temperature dependent uptake of PEG-CNA-RHO. Cells were incubated with 100 μg/mL of PEG-CNA-RHO for 1 hour at either 4°C or 37°C. Cellular uptake of PEG-CNA-RHO was not significantly affected by incubation temperature, suggesting a passive uptake mechanism.

To identify the mechanism by which uptake of CNA occurs, experiments were performed at 4°C or 37°C to probe passive and active transport mechanisms, respectively. At 4°C, active endocytotic uptake mechanisms are severely limited due to the lack of energy (i.e. ATP production)35, but passive transport pathways are not. Thus, the relative extent of uptake that is observed at 4°C provides information about whether uptake is passive or active. Cells were incubated at either 4°C or 37°C and then treated with PEG-CNA-RHO. As seen in Figure 3.5, the difference in average cellular fluorescence was determined to be insignificant between 4°C or

37°C, suggesting that uptake is primarily through a passive mechanism (i.e. free or facilitated diffusion).

It is likely that some of these results are explained by unique structural elements of the

CNAs. First, the PEG-CNA-RHO exhibits block amphiphilicity due to the hydrophilic PEG block and hydrophobic CNA block. It is possible that the hydrophobic CNA block is facilitating interaction with the phospholipids within the cell membrane, leading to increased translocation.

Previous reports of similarly structured amphiphilic molecules have indeed shown that increasing

46 hydrophobicity leads to greater cell uptake36. Further, studies on amphiphilic cell-penetrating peptides (CPPs) have concluded that insertion of the CPPs hydrophobic region into the cell membrane is crucial for the initial steps of membrane translocation37,38. It is also worth noting that like many other amphiphilic molecules, these polymers self-assemble into particles or micelles in solution (Supporting Figure 3.10). There have been other reports of self-assembled structures that are internalized by energy-independent pathways; however, the exact mechanisms are not fully understood.39,40. In addition, CNAs maintain a net neutral charge under biological conditions, which could also affect their interaction with the cell membrane. It has been shown that electrostatics affect the hydrophobic partitioning of amphiphiles into lipid-bilayer membranes38,41.

This behavior suggests that neutral CNAs have an increased association with the cell membrane and ultimately increased uptake compared to negatively charged DNA, which experiences electrostatic repulsion between the negatively charged backbone of DNA and the negatively charged cell membrane surface.42,43

3.5 Conclusions

CNAs have been developed as an alternative approach to modifying nucleic acids. Their ability to participate in thiol-X reactions allows for facile synthesis of nucleic acid oligomers.

Similar oligomers have been shown to have effective antisense characteristics in gene therapy applications. In the present study, the evaluation of CNA-based copolymers as potential antisense oligonucleotides was initiated by evaluating their cytocompatibility and cellular internalization characteristics. CNA copolymers were determined to be non-toxic to cells at concentrations up to

100 μg/mL. Additionally, PEG-CNA-RHO was taken up by cells within 1 hour and to a greater extent than PEG-RHO and RHO alone and that the extent of uptake was not affected by temperature, suggesting a passive, CNA dependent uptake mechanism. It is possible that this 47 uptake is because the lipophilic CNA block allows for better interaction with the lipid cell membrane. The distribution of PEG-CNA-RHO within cells exhibited both diffuse and punctate patterns with no apparent specific organelle colocalization. These results show that CNAs are cytocompatible, readily taken up by cells, and thus warrant future studies to assess their potential as a synthetically simple alternative to current antisense approaches.

3.6 Acknowledgements

This work was completed with support from an NSF MRSEC grant (DMR 1420736) and from a US Department of Education GAANN Fellowship to Alex Anderson. E.B.P. was supported by the NIH T32 National Institutional Research Service Award T32 HL07670. The imaging work was performed at the BioFrontiers Institute Advanced Light Microscopy Core. Spinning disc confocal microscopy was performed on Nikon Ti-E microscope supported by the BioFrontiers

Institute and the Howard Hughes Medical Institute.

3.7 References

1. Olie, R. A. et al. A novel antisense oligonucleotide targeting survivin expression induces apoptosis and sensitizes lung cancer cells to chemotherapy. Cancer Res. 60, 2805–2809 (2000).

2. Savage, D. B. et al. Reversal of diet-induced hepatic steatosis and hepatic insulin resistance by antisense oligonucleotide inhibitors of acetyl-CoA carboxylases 1 and 2. J. Clin. Invest. 116, 817–824 (2006).

3. Zinker, B. A. et al. PTP1B antisense oligonucleotide lowers PTP1B protein, normalizes blood glucose, and improves insulin sensitivity in diabetic mice. Proc. Natl. Acad. Sci. 99, 11357– 11362 (2002).

4. Paterson, B. M., Roberts, B. E. & Kuff, E. L. Structural gene identification and mapping by DNA-mRNA hybrid-arrested cell-free translation. Proc. Natl. Acad. Sci. 74, 4370–4374 (1977).

5. Zamecnik, P. C. & Stephenson, M. L. Inhibition of Rous sarcoma virus replication and cell transformation by a specific oligodeoxynucleotide. Proc. Natl. Acad. Sci. 75, 280–284 (1978).

6. Shi, B. et al. Challenges in DNA Delivery and Recent Advances in a Multifunctional Polymeric DNA Delivery Systems. Biomacromolecules (2017) doi:10.1021/acs.biomac.7b00803.

7. Kafil, V. & Omidi, Y. Cytotoxic Impacts of Linear and Branched Polyethylenimine Nanostructures in A431 Cells. BioImpacts BI 1, 23–30 (2011).

8. Breunig, M., Lungwitz, U., Liebl, R. & Goepferich, A. Breaking up the correlation between efficacy and toxicity for nonviral gene delivery. Proc. Natl. Acad. Sci. 104, 14454–14459 (2007).

9. Lechardeur, D. et al. Metabolic instability of plasmid DNA in the cytosol: a potential barrier to gene transfer. Gene Ther. 6, 482 (1999).

10. Eder, P. S., DeVINE, R. J., Dagle, J. M. & Walder, J. A. Substrate Specificity and Kinetics of Degradation of Antisense Oligonucleotides by a 3′ Exonuclease in Plasma. Antisense Res. Dev. 1, 141–151 (1991).

11. Wickstrom, E. Oligodeoxynucleotide stability in subcellular extracts and culture media. J. Biochem. Biophys. Methods 13, 97–102 (1986).

12. Grunweller, A. Comparison of different antisense strategies in mammalian cells using locked nucleic acids, 2’-O-methyl RNA, phosphorothioates and small interfering RNA. Nucleic Acids Res. 31, 3185–3193 (2003).

13. Sazani, P. et al. Nuclear antisense effects of neutral, anionic and cationic oligonucleotide analogs. Nucleic Acids Res. 29, 3965–3974 (2001).

14. Dueholm, K. L. et al. Synthesis of peptide nucleic acid monomers containing the four natural nucleobases: thymine, cytosine, adenine, and guanine and their oligomerization. J. Org. Chem. 59, 5767–5773 (1994).

15. Koshkin, A. A. et al. LNA (Locked Nucleic Acids): Synthesis of the adenine, cytosine, guanine, 5-methylcytosine, thymine and uracil bicyclonucleoside monomers, oligomerisation, and unprecedented nucleic acid recognition. Tetrahedron 54, 3607–3630 (1998).

16. Badi, N. & Lutz, J.-F. Sequence control in polymer synthesis. Chem. Soc. Rev. 38, 3383 (2009).

17. Xi, W. et al. Clickable Nucleic Acids: Sequence-Controlled Periodic Copolymer/Oligomer Synthesis by Orthogonal Thiol-X Reactions. Angew. Chem. Int. Ed. 54, 14462–14467 (2015).

18. Spitzer, S. & Eckstein, F. Inhibition of deoxyribonucleases by phosphorothioate groups in oligodeoxyribonucleotides. Nucleic Acids Res. 16, 11691–11704 (1988).

19. Crinelli, R., Bianchi, M., Gentilini, L. & Magnani, M. Design and characterization of decoy oligonucleotides containing locked nucleic acids. Nucleic Acids Res. 30, 2435–2443 (2002).

20. Aronin, N. & DiFiglia, M. Huntingtin-lowering strategies in Huntington’s disease: Antisense oligonucleotides, small RNAs, and gene editing: HUNTINGTIN-LOWERING STRATEGIES IN HD. Mov. Disord. 29, 1455–1461 (2014).

21. Harguindey, A. et al. Synthesis and Assembly of Click-Nucleic-Acid-Containing PEG–PLGA Nanoparticles for DNA Delivery. Adv. Mater. 29, n/a-n/a (2017).

22. Harris, J. M., Martin, N. E. & Modi, M. Pegylation: A Novel Process for Modifying Pharmacokinetics. Clin. Pharmacokinet. 40, 539–551 (2001).

23. Harada, A., Togawa, H. & Kataoka, K. Physicochemical properties and nuclease resistance of antisense-oligodeoxynucleotides entrapped in the core of polyion complex micelles composed of poly (ethylene glycol)–poly (l-lysine) block copolymers. Eur. J. Pharm. Sci. 13, 35–42 (2001).

24. Jia, F. et al. Effect of PEG Architecture on the Hybridization Thermodynamics and Protein Accessibility of PEGylated Oligonucleotides. Angew. Chem. Int. Ed. 56, 1239–1243 (2017).

25. Chan, W. S., Svensen, R., Phillips, D. & Hart, I. R. Cell uptake, distribution and response to aluminium chloro sulphonated phthalocyanine, a potential anti-tumour photosensitizer. Br. J. Cancer 53, 255–263 (1986).

26. Bryant, S. J., Nuttelman, C. R. & Anseth, K. S. Cytocompatibility of UV and visible light photoinitiating systems on cultured NIH/3T3 fibroblasts in vitro. J. Biomater. Sci. Polym. Ed. 11, 439–457 (2000).

27. Jin, H., Heller, D. A. & Strano, M. S. Single-Particle Tracking of Endocytosis and Exocytosis of Single-Walled Carbon Nanotubes in NIH-3T3 Cells. Nano Lett. 8, 1577–1585 (2008).

28. Lönnroth, E.-C. Toxicity of Medical Glove Materials: A Pilot Study. Int. J. Occup. Saf. Ergon. 11, 131–139 (2005).

29. Zinchuk, V. & Grossenbacher-Zinchuk, O. Quantitative Colocalization Analysis of Confocal Fluorescence Microscopy Images. in Current Protocols in Cell Biology (eds. Bonifacino, J. S., Dasso, M., Harford, J. B., Lippincott-Schwartz, J. & Yamada, K. M.) (John Wiley & Sons, Inc., 2011).

30. Manders, E. M. M., Verbeek, F. J. & Aten, J. A. Measurement of co-localization of objects in dual-colour confocal images. J. Microsc. 169, 375–382 (1993).

31. Lorenz, P., Baker, B. F., Bennett, C. F. & Spector, D. L. Phosphorothioate Antisense Oligonucleotides Induce the Formation of Nuclear Bodies. Mol. Biol. Cell 9, 1007–1023 (1998).

32. Leonetti, J. P., Mechti, N., Degols, G., Gagnor, C. & Lebleu, B. Intracellular distribution of microinjected antisense oligonucleotides. Proc. Natl. Acad. Sci. 88, 2702–2706 (1991).

33. Melan, M. A. & Sluder, G. Redistribution and differential extraction of soluble proteins in permeabilized cultured cells. Implications for immunofluorescence microscopy. J. Cell Sci. 101 ( Pt 4), 731–743 (1992).

34. Bolte, S. & Cordelieres, F. P. A guided tour into subcellular colocalization analysis in light microscopy. J. Microsc. 224, 213–232 (2006).

35. Schmid, S. L. & Carter, L. L. ATP is required for receptor-mediated endocytosis in intact cells. J. Cell Biol. 111, 2307–2318 (1990).

36. Luxenhofer, R. et al. Structure-property relationship in cytotoxicity and cell uptake of poly(2- oxazoline) amphiphiles. J. Controlled Release 153, 73–82 (2011).

37. Deshayes, S. et al. Insight into the Mechanism of Internalization of the Cell-Penetrating Carrier Peptide Pep-1 through Conformational Analysis. Biochemistry 43, 1449–1457 (2004).

38. Seelig, J. Thermodynamics of lipid–peptide interactions. Biochim. Biophys. Acta BBA - Biomembr. 1666, 40–50 (2004).

39. Pujals, S., Fernández-Carneado, J., López-Iglesias, C., Kogan, M. J. & Giralt, E. Mechanistic aspects of CPP-mediated intracellular drug delivery: Relevance of CPP self-assembly. Biochim. Biophys. Acta BBA - Biomembr. 1758, 264–279 (2006).

40. Pujals, S. & Giralt, E. Proline-rich, amphipathic cell-penetrating peptides. Adv. Drug Deliv. Rev. 60, 473–484 (2008).

41. Li, Y., Heitz, F., Le Grimellec, C. & Cole, R. B. Fusion Peptide−Phospholipid Noncovalent Interactions As Observed by Nanoelectrospray FTICR−MS. Anal. Chem. 77, 1556–1565 (2005).

42. Patil, S. D., Rhodes, D. G. & Burgess, D. J. DNA-based Therapeutics and DNA Delivery Systems: A Comprehensive Review. AAPS J. 7, E61–E77 (2005).

43. Meade, B. R. & Dowdy, S. F. Exogenous siRNA delivery using peptide transduction domains/cell penetrating peptides. Adv. Drug Deliv. Rev. 59, 134–140 (2007). 52

3.8 Supporting Information

D B

A C

Supporting Figure 3.2 – Sample 1H NMR of PEG-SH. Integration of peaks were done with respect to peak B, the PEG methoxy protons. 1H NMR (400 MHz, DMSO-d6) δ 3.79 – 3.39 (m, 182H), 3.24 (s, 3H), 2.62 (q, J = 7.1 Hz, 2H), 2.31 (t, J = 8.1 Hz, 1H).

Supporting Figure 3.3 – Sample 1H NMR of the protected thymine, thiol-ene monomer. 1H NMR (400 MHz, DMSO-d6) δ 11.29 (d, J = 4.8 Hz, 1H), 7.51 – 7.05 (m, 17H), 5.62 (dddt, J = 61.8, 17.1, 10.6, 5.2 Hz, 1H), 5.23 – 4.83 (m, 2H), 4.35 (d, J = 78.4 Hz, 2H), 3.65 (dd, J = 40.7, 5.3 Hz, 2H), 3.00 (t, J = 7.4 Hz, 2H), 2.63 – 2.19 (m, 4H), 1.74 (d, J = 1.2 Hz, 3H).

G A

B C F G E

D D F

Supporting Figure 3.4 – Sample 1H NMR of the PEG-CNA conjugate, immediately after polymerization. Integration of peaks were done with respect to peak E, the PEG methoxy protons. 1H NMR (400 MHz, DMSO-d6) δ 11.30 (s, 11H), 7.59-7.23 (m, 11 H), 4.73-4.41 (m, 22H), 3.51 (s, 226 H), 3.24 (s, 3H), 2.90-2.54 (m, 44H), 2.06-1.61 (m, 55H)

H A

F C G E g D D F

Supporting Figure 3.5 – Sample 1H NMR of the water-soluble PEG-CNA fraction. Integration of peaks were done with respect to peak E, the PEG methoxy protons. 1H NMR (400 MHz, DMSO-d6) δ 11.28 (s, 5H), 7.56 – 7.25 (m, 5H), 4.82 – 4.44 (m, 10H), 3.73 – 3.37 (m, 202H), 3.24 (s, 3H), 2.90 – 2.55 (m, 20H), 2.06 – 1.79 (m, 10H), 1.78 – 1.71 (m, 15H).

F A G G B H E C

D D

Supporting Figure 3.6 – Sample 1H NMR of the PEG-CNA-RHO conjugate. Integration of peaks were done with respect to peak E, the PEG methoxy protons. The starred peak at ~6.75 ppm is likely the maleimide peak of the unreacted rhodamine dye. When comparing the integration of this peak to others attributed to the dye, it is smaller than expected, indicating that most of the dye did react, or was dialyzed out. 1H NMR (400 MHz, DMSO-d6) δ 11.30 (s, 5H), 8.45 – 7.84 (m, 3H), 7.49 (d, J = 7.8 Hz, 3H), 7.39 (d, 5H), 7.12 – 6.85 (m, 3H), 4.57 (s, 10H), 3.72 – 3.44 (m, 181H), 3.24 (s, 3H), 1.74 (s, 15H), 1.21 (t, J = 13.2, 6.2 Hz, 12H).

Supporting Figure 3.7 – Labelling efficiency of the PEG-CNA-RHO. UV-VIS was used to assess the extent of functionalization of conjugates. Briefly, absorbance spectra were gathered for each sample with a NanoDrop 1000, and the following equation was used to obtain the moles of dye 퐴560 푛푚 푀푊푐표푛푗푢푔푎푡푒 푚표푙푒푠 푑푦푒 per mole of conjugate: ∗ = . Here, A560 nm is the 휀 ∗푙 푚푔푐표푛푗푢푔푎푡푒 푚표푙푒푠 푐표푛푗푢푔푎푡푒 560 푛푚 ൗ푚퐿 absorbance of the solution at 560 nm, ε560 nm is the molar absorptivity of the dye at 560 nm, l is the path length, MWconjugate¬ is the molecular weight of the entire conjugate, and mg¬conjugate/mL is the concentration of the conjugate solution. The absorbance of PEG-CNA- RHO at 560 nm corresponds to an extent of conjugation of 0.48 moles dye/moles conjugate. However, this may be a slight overestimation since there is evidence of some free fluorophore in the conjugate. In general, these values represent typical conjugation percentages for this reaction.

C D

Chapter 4 - Messenger RNA Enrichment Using Synthetic Oligo(T) Click Nucleic Acids

As published in Chemical Communications, 2020

4.1 Abstract

Enrichment of mRNA is a key step in a number of molecular biology techniques, particularly in the rapidly growing field of transcriptomics. Currently, mRNA is isolated using oligo(thymine) DNA (oligo(dT)) immobilized on solid supports, which binds to the poly(A) tail of mRNA to pull the mRNA out of solution through the use of magnets or centrifugal filters. Here, a simple method to isolate mRNA by complexing it with synthetic click nucleic acids (CNAs) is described. Oligo(T) CNA bound efficiently to mRNA, and because of the insolubility of CNA in water, >90% of mRNA was readily removed from solution using this method. Simple washing, buffer exchange, and heating steps enabled mRNA’s enrichment from total RNA, with a yield of

3.1 ± 1.5% of the input total RNA by mass, comparable to the yield from commercially available mRNA enrichment beads. Further, the integrity and activity of mRNA after CNA-facilitated pulldown and release was evaluated through two assays. In vitro translation of EGFP mRNA confirmed the translatability of mRNA into functional protein and RT-qPCR was used to amplify enriched mRNA from total RNA extracts and compare gene expression to results obtained using commercially available products.

4.2 Introduction

Several molecular biology techniques benefit from or require enrichment of messenger

RNA (mRNA), including quantitative polymerase chain reaction (qPCR), in vitro translation experiments, and next-generation sequencing of RNA (RNA-seq). In such assays, mRNA provides information about gene expression and abnormal gene fusions to better understand different cellular processes and disease states.1–7 In particular, mRNA provides information about the protein-coding regions of the transcriptome.8–10 Amplification of mRNAs in qPCR is also a useful diagnostic technique, especially for identifying overexpressed proteins in breast, prostate, and other types of cancers.6,11–13 Enrichment of mRNA is achieved using commercial kits containing oligo(dT) immobilized on solid supports (e.g., magnetic beads or cellulose) to concentrate poly(A) mRNA selectively, which makes up just 1-5% of total RNA in the cell.14-18 While this method is effective, it was anticipated that click nucleic acid (CNA) oligonucleotides, a newly developed type of xenonucleic acid (XNA), could be exploited in an alternative, simple strategy for mRNA isolation.

The CNAs used in this work (Figure 4.1a) have a six atom spacing per repeat unit similar to DNA/RNA, facilitating binding to native DNA/RNA.19-21 While there are reports of using other types of modified nucleic acids to selectively isolate mRNA,22-23 two key attributes give CNA oligonucleotides a distinct advantage over these materials for such applications. Primarily, the thiol-ene “click” reaction used to polymerize CNAs is ideal for producing mononucleotide repeat sequences, particularly oligo(thymine) (oligo(T)), which is synthesized and purified at the hundreds-of-milligram to gram scale in a few hours. Conversely, solid phase synthesis of DNA,

RNA, or other xeno nucleic acid (XNA) oligonucleotides suffers from low yields, increased

In addition, CNA oligonucleotides are attractive for mRNA enrichment due to their nonpolar, thioether backbone. Unlike native nucleic acids, which have good solubility in water because of their negatively charged phosphodiester backbone, non-functionalized CNAs are water insoluble. Despite this insolubility, stable CNA-DNA or CNA-RNA hybrids form if CNA oligonucleotides are dissolved in DMSO first and then added to an aqueous solution containing native nucleic acids.20 Another benefit of the non-ionic CNA backbone is improved hybrid stability, even at low ionic strength,20 relative to native DNA-DNA or DNA-RNA duplexes, which must overcome electrostatic repulsion between strands to hybridize. Therefore, it was hypothesized that mRNA enrichment could be achieved by a simple process exploiting the 65 insolubility of CNA along with its ability to bind complementary RNA (Figure 4.1b). Herein, conditions under which mRNA is pulled out of solution and subsequently released from oligo(T)

CNA are identified and the functionality of mRNA enriched via this method is demonstrated in two molecular assays.

4.3 Results and Discussion

Oligo(T) CNA was synthesized as previously described.19,20 The number average degree of polymerization for the CNAs used in this work was 16 ± 3 with a polydispersity of 1.5 ± 0.2

(Supporting Figure 4.1). Because CNA solubility depends on length, it was anticipated that CNA precipitation and, consequently, mRNA pulldown efficiency would be affected by buffer composition. To test this hypothesis, the ratio of DMSO to water was adjusted to balance CNA solubilization and RNA binding. Pulldown efficiency for an A20 RNA oligonucleotide was found to be highest (95  1%) at the lowest DMSO concentrations tested (<20 vol%) (Supporting Figure

4.2). This finding is favorable for downstream applications of the isolated mRNA, such as RT-

PCR, because DMSO is well-tolerated, even favorable, when included at low concentrations (<10 vol%).24,25 As such, all subsequent studies were performed using 5 vol% DMSO.

Next, the effects of salt concentration were examined. Methods relying on immobilized oligo(dT), which are densely packed on solid supports, require optimized ionic strength to maximize pulldown efficiency.26 Sufficiently high cation concentrations are needed to enable hybridization between the oligo(dT) and poly(A).26–28 Further, salt content has been shown to influence the structure and stability of target RNA in solution.29 Compared to other monovalent cations, Li+ ions are less effective at precipitating DNA,30 allowing for selective precipitation of

RNA over DNA, and have better solubility in organic cosolvents, including DMSO. In a buffer

DIRECT™ kit, 10 mM Tris pH 7.5, 1M LiCl, 1 mM EDTA), 93% of mRNA was pulled out of solution. Unlike methods relying on oligo(dT), it is noteworthy that mRNA pulldown with CNA was also possible in the absence of salts, albeit with lower efficiency (70%) (Supporting Figure

4.3). All subsequent pulldown studies were carried out using this LiCl containing buffer (Buffer

B, Table S4.1).

To begin optimizing the CNA concentration required to isolate RNA, the concentration of

CNA was varied while keeping an A20 RNA concentration fixed at 125 nM. Greater than 75% pulldown efficiency was achieved using oligo(T) CNA concentrations as low as 7.5 μM (30

μg/mL) (Figure 4.2a). When a non-complementary RNA oligonucleotide, U20, was used in place of the complementary A20 oligonucleotide, no pulldown was observed (Figure 4.2a), demonstrating that CNA-mediated pulldown of RNA oligonucleotides was a result of sequence-

67 specific hybridization rather than non-specific, hydrophobic interactions. As with traditional mRNA isolation procedures, RNA was released by heating (Figure 4.2b).

Higher CNA concentrations were required to precipitate EGFP mRNA compared to short

RNA oligonucleotides used in the prior study, which is attributed to its larger size (996 nucleotides) and increased hydrophilicity. Still, it was found that greater than 75% of mRNA was removed using a CNA concentration of at least 125 μM (500 μg/mL) (Figure 4.3a). For even higher CNA concentrations (150-250 μM), between 90-92% pulldown efficiency was consistently achieved for

EGFP mRNA concentrations over the range of 30-90 ng/mL (Figure 4.3b). Furthermore, even when relatively high concentrations of mRNA were tested (2 – 64 μg/mL), the pulldown efficiency remained above 90% (Supporting Figure 4.4).

Having demonstrated efficient pulldown, it was necessary to demonstrate that the released mRNA remained functional, ultimately confirming the utility of CNA as an mRNA enrichment tool. To do so, an in vitro translation (IVT) kit was used to translate EGFP mRNA recovered after pulldown and release from oligo(T) CNA, non-complementary oligo(A) CNA, or no CNA. After washing and release, the concentration of functional EGFP mRNA, or mRNA that could be translated into fluorescent protein, was measured via IVT, where higher EGFP fluorescence indicated a higher concentration of functional mRNA (Supporting Figure 4.5). The ratio of the fluorescence after pulldown and release to the fluorescence where no pulldown occurred (“relative

EGFP fluorescence”) was taken as a measure of the enrichment of specifically functional mRNA

(see SI for more details). A relative EGFP fluorescence of 1 would indicate no pulldown or release.

For mRNA mixed with complementary oligo(T) CNA and released at 75°C, relative EGFP fluorescence was 3.2 ± 0.4 (Figure 4.3c), while for both negative controls (i.e., oligo(A) CNA or no CNA), relative EGFP fluorescence was close to 1, (1.2 ± 0.1 and 1.0 ± 0.4, respectively). In the

absence of heating, EGFP fluorescence for the oligo(T) CNA samples was negligible, comparable

to levels observed for controls with no mRNA (Figure 4.3d). These results support the hypothesis

that mRNA enriched through the oligo(T) CNA isolation method is readily translated into protein.

In prior studies, it was observed that while RNA pulldown was consistently above 90%,

the release efficiency was more variable and sometimes quite poor (e.g. ~34% recovery, Figure

4.2b). After additional optimization of release conditions, the monovalent salt:mRNA

concentration ratio in the release buffer was found to be the largest driver of variability due to salt-

induced precipitation of mRNA (Supporting Figure 4.6). The washing and release procedure

For applications in molecular biology, it was useful to compare the CNA isolation method with commercially available protocols for enriching mRNA. First, oligo(T)’s ability to selectively precipitate mRNA was confirmed by comparing the pulldown efficiency to that with rRNA, which accounts for ~80% of total RNA (Supporting Figure 4.7). Next, using the optimized procedure detailed above, side-by-side isolations of mRNA from cell-isolated total RNA were conducted using oligo(T) CNA and the Dynabeads®, which are magnetic beads functionalized with oligo

(dT)25. After two washes and heat mediated release, the total mass of enriched mRNA was 100 ±

50 ng and 90 ± 20 ng for CNA and Dynabeads®, respectively (Figure 4.4a). These concentrations accounted for 3.1 ± 1.5% and 2.9 ± 0.8% of the input total RNA for the CNA and Dynabeads® methods, respectively, which fall within the expected range of mRNA abundance in total RNA (1-

5%). Overall, the yield of mRNA after isolation with CNA is comparable to the yield using the commercially available Dynabeads®.

RT-qPCR was then performed on isolated mRNA to confirm that it would remain useful for downstream bioanalysis. Results were assessed for accuracy by comparing relative gene expressions to a sample of freshly isolated total RNA and mRNA isolated with Dynabeads®.

Complementary DNA was synthesized using an oligo(dT12-18) primer rather than a collection of random primers to ensure that only mRNA was reverse transcribed. Two genes of interest were chosen, IL1B (coding for interleukin-1 beta) and MMP2 (coding for matrix metalloproteinase 2)

(Table S4.2), due to their relevance to the cell line used (MCF10a cells, non-tumorigenic mammary gland cells). GAPDH was used as the housekeeping gene. Figure 4.4b shows the

Figure 4.4 –(a) Using optimized buffer conditions, the performance of oligo(T) CNA compared to Dynabeads showed no statistical difference in mass yield. (b) There was also no statistical difference in relative expression levels measured using mRNA input from different isolation procedures. Data is represented as the mean of at least 3 replicates and error bars represent standard deviations. relative expression (RE) levels of IL1B and MMP2 compared to the housekeeping gene. For both genes examined, RE levels of mRNA isolated with CNA were not statistically different than either mRNA isolated with Dynabeads® or total RNA samples. These results support the hypothesis that mRNA isolated using CNA maintains its integrity. Further, the isolation process does not bias isolation of specific mRNAs by length, at least for the two genes investigated herein. If the CNA- mediated isolation method did experience length dependent isolation efficiencies, significant decreases in RE would be expected for larger mRNAs. In addition to the fact that this was not observed in this experiment, it is worth noting that the mRNA for MMP2 is nearly twice the size of the mRNA for IL1B, depending on the transcript.

This study reports an alternative strategy of mRNA isolation using hydrophobic oligo(T)

CNAs, which have the ability to pull mRNA out of solution via binding to the poly(A) tail.

Through the optimization of buffer conditions and material concentrations, mRNA yields as high as 94% were observed. Given the relative cost and scalability of CNA compared to traditional

71 mRNA enrichment methods, CNAs represent a favorable method for situations when a large amount of reagent is necessary. In comparison to the Dynabeads® mRNA DIRECT™ kit, the

CNA-mediated isolation method described herein resulted in similar overall yield and quality of mRNA as determined by in vitro translation and RT-qPCR analysis. These results indicate that mRNA isolation using hydrophobic CNA offers a competitive alternative to traditional, DNA dependent strategies.

4.4 Materials and Methods

4.4.1 CNA synthesis, polymerization, and characterization.

Thymine CNA monomer was synthesized as previously described with a slight modification.17 Specifically, the thymine imide was protected by an acid-labile tert-butoxide

(BOC) group prior to backbone addition. Oligo(T) was synthesized as previously described via a light-initiated thiol-ene reaction.27 The BOC protecting groups on the resulting oligomers were then deprotected by concentrated (37.2% w/w) hydrochloric acid. The product was then washed extensively with DI H2O and Acetone to ensure complete removal of any salts. Finally, the average molecular weight and polydispersity index (PDI) was evaluated by gel permeation chromatography (TOSOH – HLC8320GPC) using an internal standard of short CNA oligomers.

4.4.2 Optimization of Cy-5 labelled RNA Pulldown.

Oligo(T) was first dissolved in DMSO at prescribed concentrations while the RNA (A20) was dissolved in the Binding buffer (B, see Table S1 for buffer compositions). The effect of DMSO concentration in the binding mixture was evaluated with four different concentrations, 5%, 10%,

20% and 50%. After mixing CNA solution with the RNA solution (final CNA concentration of

500 μM CNA, RNA concentration of 125 nM), the solutions were centrifuged at 6,000 g for 2 min to pellet the precipitate. To evaluate pulldown efficiency, Cy5 fluorescence of the supernatant was 72 measured and compared to a negative control without oligo(T) CNA. To evaluate the effect of salt concentration on pulldown efficiency, pulldown was performed in solutions containing 5% DMSO and either no salt or 1 M LiCl. Pulldown efficiency was evaluated as described previously.

Using the optimized buffer and the procedure dictated above, the amount of CNA needed to effectively precipitate small RNA strands such as A20 was determined by a CNA titration.

Pulldown efficiency was determined for a range of oligo(T) concentrations from 0.015 – 500 μM with a fixed RNA concentration of 125 nM. Samples were performed in triplicate. As a negative control, a non-complementary sequence (U20) was used to confirm base-specific interactions.

Release of the A20 RNA strand was achieved by reconstituting the pellet in release buffer R1 and heating to 75°C for 5 minutes. The solution was then spun at 15,000 g for 2 min to quickly pellet the unbound CNA, and the RNA concentration in the supernatant was measured by Cy5 fluorescence. As a negative control, the resulting RNA concentrations of samples that received no heat were also measured.

4.4.3 Optimization of Cy-5 labelled mRNA pulldown.

Using the optimized pulldown procedure described above, Cy5 labelled EGFP mRNA was used to evaluate oligo(T)’s ability to precipitate larger RNA strands. To evaluate the amount of

CNA required for efficient pulldown, same titration experiment as described above was performed with the same fixed RNA concentration and the same concentration range of oligo(T). To demonstrate the range of mRNA concentrations that can be precipitated by this method, 250 μM of oligo(T) CNA was used to pull down mRNA in a concentration range of 2 – 64 μg/mL.

4.4.4 In-Vitro Translation.

In-vitro translation was performed using Retic Lysate IVT™ Kit from Thermo Fisher

Scientific following manufacturer’s instructions. Samples of EGFP mRNA were subjected to the

73 basic pulldown procedure as detailed above. To ensure the amount of mRNA used for IVT input remained in a workable range, an initial concentration of 62.5 ng/µL of EGFP mRNA was used for mRNA enrichment experiments. As negative controls, mRNA was incubated by itself or with oligo(A) CNA rather than oligo(T) CNA. After pulldown, samples were washed once with 0.67X

SSC, 5% DMSO and then either kept at room temperature or heated at 75°C for 15 minutes in the same buffer before immediately centrifuging to remove any precipitated CNA and mRNA.

Following this step, 5 μL of the supernatant was used as input for the IVT reactions. IVT was performed according to the manufacturer’s protocols and a the resulting EGFP fluorescence was measured. A relative fluorescence factor was then calculated to determine how effective each condition was at concentrating EGFP mRNA. A factor of 1 corresponded to the fluorescence signal obtained from mRNA that was not pulled down. See Supporting Information for additional details.

4.4.5 Optimization of mRNA release.

After observing poor mRNA release at low mRNA concentrations, washing and release steps were altered to gradually reduce the buffer’s LiCl salt content. Specifically, CNA-mRNA pellets were washed once with buffer W1 (10 mM Tris pH 7.5, 150 mM LiCl, 1 mM EDTA) and once with buffer W2 (10 mM Tris pH 7.5, 5% DMSO). Release was accomplished by adding buffer R2 (10 mM Tris pH 7.5, 5% DMSO), mixing, and heating at 75°C for 5 minutes.

4.4.6 Pulldown specificity – rRNA vs. mRNA.

To make sure oligo(T) CNA could selectively precipitate mRNA instead of other types of

RNA, pulldown efficiency was compared between inputs of mRNA and rRNA. For the rRNA, the

100 ng/uL rRNA supplied with the Qubit™ RNA BR Assay was used and for mRNA, CleanCap®

EGFP mRNA from TriLink Biotechnologies was used. For pulldown experiments, oligo(T) and

RNA were supplied at 500 μM and 50 ng/μL respectively, and the procedure utilized the optimized

74 binding buffer, B, detailed above. Pulldown efficiency was evaluated by measuring the RNA remaining in the supernatant using the Qubit™ RNA BR Assay according to manufacturer’s protocols. The resulting fluorescence of each sample was used as an indicator for the amount of

RNA not pulled out of solution by the CNA.

4.4.7 Pulldown and release from Total RNA.

Total RNA extracts were obtained from MCF10a cells with TRIzol™ Reagent using the manufacturer’s guidelines. The total RNA was then diluted to an initial concentration of 750

µg/mL. 10 µL of this concentration was added to individual tubes that were then subject to different pulldown conditions. Pulldown with CNA was accomplished by first mixing the total

RNA with a 2x concentration of buffer B to a volume of 19 μL. After efficient mixing, 1 uL of 3 mM oligo(T) CNA was added to the total RNA solution. As a positive control, pulldown was performed with Dynabeads™ from an mRNA DIRECT™ Purification Kit. In this case, the total

RNA was mixed with beads suspended in 10 µL of the supplied binding buffer. The samples were then incubated at room temperature for 5 minutes after which the CNA samples were centrifuged at 10,000 g for 2 minutes while the Dynabeads™ samples were placed on a magnet for 2 minutes.

In both cases 15 uL of supernatant was taken and saved for RNA quantification. Both samples were then washed once with the wash buffer W1 and once with buffer W2. For mRNA release, pellets were reconstituted in buffer R2 and each sample was placed in a thermocycler and held at

75C for 5 minutes. As quickly as possible, the CNA samples were spun at 6000 g for 2 min while the Dynabeads™ sample was placed on a magnet. The supernatant was then collected for quantification and downstream assays.

4.4.8 RT-PCR.

Stocks of mRNA was obtained by the previously described pulldown conditions. As a negative control, RNA was degraded with Nucleoside Digestion Mix (New England Biolabs™).

Transcription to cDNA was performed with a high capacity reverse transcription kit (Applied

Biosystems) using an oligo(dT)12-18 primer (Invitrogen™). The oligo(dT) primer was used to ensure that only mRNA was transcribed to cDNA, which was necessary to control for the fact that the total RNA aliquots also contained rRNA and tRNA. Quantitative PCR (qPCR) was conducted with Fast SYBR Green Master Mix (Applied Biosystems) on a 7500 Fast Real-time PCR Machine.

Relative expression (RE) is defined as the ratio of expression of a gene of interest (GOI) to a reference gene, GAPDH, through the following formula:

퐶푡푟푒푓 (퐸푟푒푓) 푅퐸 = 퐶푡 (퐸퐺푂퐼) 퐺푂퐼

where E refers to the true efficiency of each primer pair and Ct is the number of cycles needed to reach a prescribed signal threshold.

4.5 Acknowledgements

This work was completed with support from an NSF MRSEC grant (DMR 1420736), the

Colorado Office of Economic Development and International Trade (CTGG12016-2273), and from a US Department of Education GAANN Fellowship to Alex Anderson. MST experiments were performed in the Shared Instruments Pool of the Department of Biochemistry at the

University of Colorado Boulder. The MST instrument was funded by NIH Shared Instrumentation

Grant S10OD21603.

4.6 References

1. Stark, R., Grzelak, M. & Hadfield, J. RNA sequencing: the teenage years. Nature Reviews Genetics 20, 631–656 (2019).

2. Chepelev, I., Wei, G., Tang, Q. & Zhao, K. Detection of single nucleotide variations in expressed exons of the human genome using RNA-Seq. Nucleic Acids Res. 37, e106 (2009).

3. Cirulli, E. T. et al. Screening the human exome: a comparison of whole genome and whole transcriptome sequencing. Genome Biol. 11, R57 (2010).

4. Todd, E. V., Black, M. A. & Gemmell, N. J. The power and promise of RNA-seq in ecology and evolution. Mol Ecol 25, 1224–1241 (2016).

5. Bustin, S. A. Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. Journal of Molecular Endocrinology 25, 169–193 (2000).

6. Devonshire, A. S. et al. Application of next generation qPCR and sequencing platforms to mRNA biomarker analysis. Methods 59, 89–100 (2013).

7. Wong, M. L. & Medrano, J. F. Real-time PCR for mRNA quantitation. BioTechniques 39, 75– 85 (2005).

8. Tang, F. et al. mRNA-Seq whole-transcriptome analysis of a single cell. Nature Methods 6, 377–382 (2009).

9. Cloonan, N. et al. Stem cell transcriptome profiling via massive-scale mRNA sequencing. Nature Methods 5, 613–619 (2008).

10. Mortazavi, A., Williams, B. A., McCue, K., Schaeffer, L. & Wold, B. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nature Methods 5, 621–628 (2008).

11. Wang, Z. et al. Circulating MACC1 as a novel diagnostic and prognostic biomarker for nonsmall cell lung cancer. J Cancer Res Clin Oncol 141, 1353–1361 (2015).

12. Wu, N. C. et al. Comparison of central laboratory assessments of ER, PR, HER2, and Ki67 by IHC/FISH and the corresponding mRNAs (ESR1, PGR, ERBB2, and MKi67) by RT-qPCR on an automated, broadly deployed diagnostic platform. Breast Cancer Res Treat 172, 327– 338 (2018).

13. de Souza, M. F. et al. Circulating mRNAs and miRNAs as candidate markers for the diagnosis and prognosis of prostate cancer. PLoS One 12, (2017).

14. Hornes, E. & Korsnes, L. Magnetic DNA hybridization properties of oligonucleotide probes attached to superparamagnetic beads and their use in the isolation of poly(A) mRNA from eukaryotic cells. Gene Analysis Techniques 7, 145–150 (1990).

15. Albretsen, C., Kalland, K.-H., Haukanes, B.-I., Håvarstein, L.-S. & Kleppe, K. Applications of magnetic beads with covalently attached oligonucleotides in hybridization: Isolation and detection of specific measles virus mRNA from a crude cell lysate. Analytical Biochemistry 189, 40–50 (1990).

16. Pemberton, R. E., Liberti, P. & Baglioni, C. Isolation of messenger RNA from polysomes by chromatography on oligo(dT)-cellulose. Analytical Biochemistry 66, 18–28 (1975).

17. Rowenow, C., Saxena, R.M., Durst, M. & Gingeras, T.R. Prokaryotic RNA preparation methods useful for high density array analysis: comparison of two approaches. Nucleic Acid Res. 29, e112 (2001).

18. Wu, J. et al. Ribogenomics: the Science and Knowledge of RNA. Genomics, Proteomics & Bioinformatics 12, 57-63 (2014).

19. Han, X. et al. New Generation of Clickable Nucleic Acids: Synthesis and Active Hybridization with DNA. Biomacromolecules 19, 4139–4146 (2018).

20. Culver, H. R. et al. Click Nucleic Acid-DNA Binding Behavior: Dependence on Length, Sequence, Ionic Strength, and Other Factors. Biomacromolecules 21, 4205-4211 (2020).

21. Anderson, A. J., Culver, H. R., Bryant, S. J. & Bowman, C. N. Viscoelastic and thermoreversible networks crosslinked by non-covalent interactions between “clickable” nucleic acid oligomers and DNA. Polym Chem 11, 2959-2968 (2020).

22. Jacobsen, N. et al. Direct isolation of poly(A)+ RNA from 4 M guanidine thiocyanate‐lysed cell extracts using locked nucleic acid‐oligo(T) capture. Nucleic Acid Res. 32 e64-e64 (2004).

23. Phelan, D., Hondorp, K., Choob, M., Efimov, V. & Fernandez, J. Messenger Rna Isolation Using Novel Pna Analogues. Nucleosides, Nucleotides & Nucleic Acids Res 20, 1107-1111 (2001).

24. Jensen, M. A., Fukushima, M. & Davis, R. W. DMSO and Betaine Greatly Improve Amplification of GC-Rich Constructs in De Novo Synthesis. PLoS One 5, (2010).

25. Strien, J., Sanft, J. & Mall, G. Enhancement of PCR Amplification of Moderate GC-Containing and Highly GC-Rich DNA Sequences. Mol Biotechnol 54, 1048–1054 (2013).

26. Gong, P. & Levicky, R. DNA surface hybridization regimes. PNAS 105, 5301–5306 (2008).

27. Wong, I. Y. & Melosh, N. A. An Electrostatic Model for DNA Surface Hybridization. Biophysical Journal 98, 2954–2963 (2010).

28. Irving, D., Gong, P. & Levicky, R. DNA Surface Hybridization: Comparison of Theory and Experiment. J. Phys. Chem. B 114, 7631–7640 (2010).

29. Draper, D. E. A guide to ions and RNA structure. RNA 10, 335–343 (2004).

30. Zinchenko, A. A. & Yoshikawa, K. Na+ Shows a Markedly Higher Potential than K+ in DNA Compaction in a Crowded Environment. Biophysical Journal 88, 4118–4123 (2005).

4.7 Supporting Information

600

6 2

400

c n

a 200

b A 0 10 15 20 Time (min) Supporting Figure 4.1 – Gel Permeation Chromatography of oligo(T) CNA. The mean retention time corresponded to a degree of polymerization of 16 ± 3 repeat units and a PDI of 1.5 ± 0.2.

Supporting Figure 4.4 – Pulldown efficiency remained >90% for concentrations up to 64 μg/mL, demonstrating the robustness of this technique.

Design of In Vitro Translation Experiments

To ensure the amount of mRNA used for IVT input stayed below 300 ng, an initial concentration of 62.5 ng/µL of EGFP mRNA was used for mRNA enrichment experiments, as this concentration would correspond to 297 ng of mRNA in 5 uL of the supernatant after release if

100% of the mRNA was recovered. At the opposite end, if there was no specific or non-specific pulldown, the concentration of mRNA remaining in 5 uL of the supernatant after release would be just 6.25% of the original. This value was determined based on the total dilution (1/16) of the initial concentration over the course of the procedure (1/4 dilution for washing step and 1/4 dilution for release step). The fluorescence that resulted from the in vitro translation assay was taken as a signal of the concentration of functional mRNA. A relative fluorescence factor was then calculated to determine how effective each condition was at concentrating specifically functional EGFP mRNA, where:

퐹푙. 푠푎푚푝푙푒 푅푒푙푎푡푖푣푒 퐹푙푢표푟푒푠푐푒푛푐푒 퐹푎푐푡표푟 = 퐹푙. 표푓 푠푎푚푝푙푒 푎푠푠푢푚푖푛푔 푛표 푒푛푟푖푐ℎ푚푒푛푡

A factor greater than 1 indicated the concentration of functional mRNA and a factor of 1 corresponded to a situation where no concentration (i.e. pulldown) occurred.

Supporting Figure 4.7 – Oligo(T) mediated precipitation of mRNA compared to rRNA confirmed specificity for mRNA.

Supporting Table 4.1 – Summary of optimized buffer compositions for CNA-facilitated RNA enrichment Buffer Name Composition Binding (B) 10 mM Tris pH 7.5, 1 M LiCl, 1 mM EDTA Wash 1 (W1) 10 mM Tris pH 7.5, 150 mM LiCl, 1 mM EDTA Wash 2 (W2) 10 mM Tris pH 7.5, 5% DMSO Release (R1) 0.67X SSC, 5% DMSO Release (R2) 10 mM Tris pH 7.5, 5% DMSO

Supporting Table 4.2 – List of qPCR Primer Sequences and Efficiencies Gene Primer Sequence Efficiency

F: 5’ – GCAAGAGCACAAGAGGAAGAG – 3’ GAPDH 107% R: 5’ – AAGGGGTCTACATGGCAACT – 3’ F: 5’ – TACCTGTCCTGCGTGTTGAA – 3’ IL1B 92% R: 5’ – TCTTTGGGTAATTTTTGGGATCT – 3’ F: 5’ – AGAAGGCTGTGTTCTTTGCAG – 3’ MMP2 100% R: 5’ – AGGCTGGTCAGTGGCTTG – 3’

Chapter 5 - Viscoelastic and Thermoreversible Networks Crosslinked by Non-covalent Interactions Between “Clickable” Nucleic Acids Oligomers and DNA

As published in Polymer Chemistry, 2020

5.1 Abstract

An approach to efficient and scalable production of oligonucleotide-based gel networks is presented. Specifically, a new class of xenonucleic acid (XNA) synthesized through a scalable and efficient thiol-ene polymerization mechanism, “Clickable” Nucleic Acids (CNAs), were conjugated to a multifunctional poly(ethylene glycol), PEG. In the presence of complementary single stranded DNA (ssDNA), the macromolecular conjugate assembled into a crosslinked 3D gel capable of achieving storage moduli on the order of 1 kPa. Binding studies between the PEG-

CNA macromolecule and complementary ssDNA indicate that crosslinking is due to the

CNA/DNA interaction. Gel formation was specific to the base sequence and length of the ssDNA crosslinker. The gels were fully thermoreversible, completely melting at temperatures above 60°C and re-forming upon cooling over multiple cycles and with no apparent hysteresis. Shear stress relaxation experiments revealed that relaxation dynamics are dependent on crosslinker length, which is hypothesized to be an effect of the polydisperse CNA chains. Arrhenius analysis of characteristic relaxation times was only possible for shorter crosslinker lengths, and the activation energy for these gels was determined to be 110 ± 20 kJ/mol. Overall, the present work demonstrates that CNA is capable of participating in stimuli-responsive interactions that would be

85 expected from XNAs, and that these interactions support 3D gels that have potential uses in biological and materials science applications.

5.2 Introduction

Nucleic acids are powerful building blocks capable of directing the assembly of specific microstructures through Watson-Crick base pairing. Since the discovery of 2D and 3D DNA- directed nano-assembly of complex structures,1,2 researchers have sought to design nucleic acid sequences and nucleic acid polymer conjugates for the development of functional materials. One primary advantage of structured DNA materials is their stimuli-responsiveness, or their ability to switch between two different states in response to an environmental change, which makes them ideal candidates for many biomedical applications such as biosensing and drug delivery.3–6 For instance, Chaithongyot et al. applied DNA origami principles to design a cargo-loaded DNA nanosphere that opens in the presence of a cancer-specific protein marker.5 Kahn and colleagues conjugated oligonucleotides to drug-loaded metal-organic frameworks and demonstrated pH and potassium ion triggered release.6

Recently, there has been a push to design macroscopic, 3D, DNA-based gels. These types of networks are crosslinked and solvent swollen materials that are also responsive to stimulus such as heat, pH, stress, or competitive binders. There are numerous examples in the literature of gels made from extracted DNA, which form hydrogels through physical entanglements,7–9 or are chemically crosslinked by a bis-epoxide crosslinker.8–11 With the development of automated DNA synthesis, examples of precisely structured DNA gels formed with rationally designed DNA oligonucleotides are expanding. These gels are typically designed to self-assemble into multifunctional building blocks with complementary “sticky-ends”, which hybridize to other building blocks to create an interconnected 3D network.12–16 Yet another approach to creating 86 stimuli-responsive DNA gels has been to conjugate oligonucleotides onto synthetic polymers and use them as crosslinkers.17–27 These oligonucleotide-polymer hybrids offer many advantages such as diverse, multifunctional architectures, reduced cost due to shorter oligonucleotide requirements, and improved control over functionality and molecular weight.

While the potential applications of these materials have been demonstrated at a proof-of- concept level, the field has largely been held back by inherent drawbacks of oligonucleotide synthesis. Currently, the iterative, solid-phase approach is the prevailing synthetic scheme because

28,29 of its ability to produce monodisperse DNA strands with precise sequence definition. While the automation of this process has reduced the cost of oligonucleotide synthesis, the strategy still has characteristic disadvantages such as length restrictions and low yields, which severely limit large-scale implementation of DNA based technologies. Thus, there is a need for alternative synthetic approaches that are inexpensive, simple to execute, and have high yields while still affording the aforementioned advantages associated with DNA-based gels.

Click chemistries are ideally suited to address this issue, particularly when employed in a step-growth polymerization mechanism. To qualify as a click reaction, the transformation must reach quantitative conversions quickly and efficiently under ambient conditions with minimal byproducts. Some researchers have, for example, used a copper catalyzed, azide-alkyne cycloaddition to synthesize nucleic acid oligomers that bind to native DNA and are recognized by

DNA polymerase.30–32 However, this approach still requires a solid-phase support and iterative addition to reach lengths greater than 3 repeat units. Recently, a new generation of oligonucleotides, “Click” Nucleic Acids (CNAs), have been developed to circumvent these synthesis inefficiencies while maintaining the capacity to hybridize with complementary nucleic acid sequences.33,34 CNAs fall into a class of molecules known as xenonucleic acids (XNAs), or 87 nucleic acids with non-natural backbones.35,36 Specifically, CNAs are synthesized through a radical mediated, linear thiol-ene “click” polymerization mechanism. Previous work has shown that CNAs are readily conjugated to synthetic polymers through thiol-ene or thiol-Michael mechanisms, are capable of binding to complementary ssDNA strands, and are cytocompatible,37,38 making them promising candidates for nucleic acid-based biomaterials.

Herein, the development of a 3D gel reversibly crosslinked by sequence specific interactions between CNA and a complementary DNA strand is reported. These CNA/DNA gels are shown to exhibit complex viscoelastic responses to mechanical stimuli, as well as rapid and complete thermoreversibility. This work establishes CNA/DNA gels as a material capable of stimuli-responsive behavior with potential use in a variety of biological and materials science applications.

5.3 Materials and Methods

5.3.1 Synthesis of 8PEG-T macromolecule

Thymine CNA monomer was synthesized as described previously with a slight adjustment.34 Here, the thymine nucleobase was protected with a tert-butoxycarbonyl (BOC) group at the imide prior to backbone addition. The thiol functionality on the monomer was deprotected by dissolving the monomer in methanol (0.5 M) and at least two equivalents of sodium hydroxide

(NaOH) for 5-10 minutes. The reaction was neutralized with excess phosphate buffered saline

(PBS) and an equimolar amount of hydrochloric acid (HCl). The monomer was extracted into dichloromethane (DCM) and dried over sodium sulfate. The solvent was removed by vacuum yielding dry, deprotected CNA monomer. This monomer was combined with 8-arm poly(ethylene glycol) thiol (20,000 g/mol, Jenkem USA) at a 15:1 monomer to arm ratio, and dissolved at 17% weight (with respect to monomer) in dimethyl sulfoxide (DMSO) containing 0.1% weight DMPA 88 as a photoinitiator and 2.5% weight tris(2-carboxyethyl)phosphine HCl (TCEP HCl). The TCEP

HCl was included to protect against the formation of disulfides during the polymerization. The mixture was polymerized with 365 nm UV light at 12 mW/cm2 for 15 minutes. The crude product mixture was directly dialyzed against DMSO (20,000 MWCO) to remove unbound and/or unreacted CNA. The dry, solid product was obtained after lyophilization of the DMSO solvent.

Gel permeation chromatography (GPC) was used to qualitatively confirm the molecular weight increase of the copolymer and 1H NMR was used to quantify the number average degree of polymerization per PEG arm. The number average degree of polymerization was used to approximate the molecular weight of the 8PEG-T macromer. Prior to gel formation, the solid product was dissolved in a 1:2 ratio of DCM to trifluoroacetic acid (TFA) to remove the BOC protecting group and subsequently precipitated in ether. The product was washed extensively with ether to remove all TFA.

5.3.2 Microscale Thermophoresis (MST) Titration

8PEG-T was dissolved at 31.2 mM in DMSO to achieve a 250 uM CNA oligomer end group concentration. This stock was then serially diluted 1:2 to create a ligand titration curve. A

DNA “pre-mix” solution was prepared by mixing 150 μL DMSO, 50 μL MilliQ H2O, 25 μL 0.66X saline sodium citrate (SSC) buffer, and 25 μL of 1 μM A20-Cy5 DNA in 0.66X SSC buffer. Equal volumes of each 8PEG-T stock and the “pre-mix” solution were mixed and vortexed, for a final

DMSO concentration of 80 vol%, CNA concentration of 125 μM, and DNA concentration of 50 nM. Each sample was centrifuged to remove insoluble fractions and 13 μL were taken to load into

MST capillaries. MST traces were measured on a Nanotemper Monolith NT.115 at a temperature of 23°C using the red LED operating at 14% power and the 1475 nm infrared (IR) laser operating at 80% power. The ratio of the average fluorescence before (-1 – 0s, Fcold) and after (0.5 – 1.5 s, 89

Fhot) turning on the IR laser (at 0 s) was taken as the normalized fluorescence (i.e., Fnorm =

Fcold/Fhot). Fnorm was plotted against the log10 of CNA concentration and fit to the Hill Model to obtain an apparent dissociation constant (Kd’).

5.3.3 Circular Dichroism

For circular dichroism experiments, the copolymer was dissolved at 50 μM in 20% H2O in

DMSO containing 25 μM A20 ssDNA. These concentrations were chosen to ensure no gel formed in the cuvette. CD spectra were obtained on an Applied Photophysics Chirascan Plus CD spectrometer in a quartz cuvette with a 0.5 mm pathlength, using a 1 nm step size and a 1 nm bandwidth, with a time per point of 0.5 seconds. To determine CNA/DNA melting, the temperature was varied from 12.5°C to 50°C at a rate of 2°C/min using a Peltier controlled sample holder, and the temperature inside the cuvette was monitored using in-cell temperature sensors.

5.3.4 Gel Formation

All gels were made at a 5% w/v 8PEG-T concentration in 80% DMSO/20% DI H2O. The amount of ssDNA added was calculated by the required CNA:DNA base ratio. Gels were formed by separately dissolving 8PEG-T and ssDNA in solvent to confirm that the individual components alone did not form a gel upon macroscopic visualization. After mixing the two components, the solution was set at room temperature. Gel formation was verified by assessing its stability upon inversion and mechanical agitation.

5.3.5 Rheological Characterization

Oscillatory shear rheology was carried out on gels with an Ares G2 Rheometer (TA

Instruments). In a typical experiment, the gel was heated above its reverse gelation temperature and an aliquot was placed between a parallel plate geometry. The bottom plate consisted of the

Advanced Peltier Systems (TA Instruments) geometry, and the thermally insulating top plate was

8 mm in diameter. The evolution of modulus was monitored at 0.5% strain and 1 Hz. The thermal sensitivity of CNA gels was characterized at 5% strain and 1 Hz while cycling the temperature between 70°C and 22°C at 5°C/min. Strain sweeps were conducted at 1.6 Hz and frequency sweeps at 5% strain. Stress relaxation was performed with a 10% initial strain applied immediately. To prevent solvent evaporation during experiments, a thin layer of light mineral oil was added around the sample.

5.3.6 Statistical Analysis

Unless otherwise stated, data are presented are representative of multiple replicates.

Statistical significance was defined at the 95% confidence level (p<0.05) using appropriate

ANOVA analysis. In cases where significance was detected, Tukey’s Post Hoc analysis was used for pair-wise comparisons.

5.4 Results and Discussion

5.4.1 Synthesis of 8PEG-T macromolecule

The CNA polymerization follows a thiol-ene click mechanism that results in linear oligomers of nucleic acids (Figure 5.1a). To create a multifunctional macromolecule, a thiolated

8-armed poly(ethylene glycol) (8PEG-SH) star polymer was doped into the CNA polymerization at a 15:1 thymine monomer to PEG arm ratio. The mixture was combined with a photoinitiator

(0.1% DMPA) and TCEP HCl (a reducing agent to prevent disulfide formation) and irradiated with UV light. The terminal thiols of the 8PEG-SH served as conjugation points for the CNA, resulting in a star block-copolymer with pendant CNA oligomers. Linear thiol-ene polymerization has been shown to yield polymers and oligomers quickly and efficiently,39,40 but it is important to note that the thiol-ene polymerization mechanism for this particular reaction is quasi step-growth 91 and results in a polydisperse product. Therefore, there is a distribution of CNA monomers per arm, ranging, for example, from monomers or dimers to 20+-mers. For situations in which polydispersity can be tolerated, this synthetic scheme provides a clear advantage in simplicity and efficiency over other polymerizations strategies. After polymerization, the majority of unreacted

CNA monomer and shorter oligomers were removed by dialysis, and the CNA was deprotected with acid. After extensive washing, the resulting product was analyzed with GPC and 1H NMR

(Supporting Figures 5.1 and 5.2). Relative integration of protons indicated the achieved degree of polymerization was between 10 and 12 repeat units, depending on the batch, which effectively doubles the molecular weight of each arm (i.e. 2500 g/mol PEG and 2700 g/mol CNA). This scheme yields a co-polymer product with average molecular weights around 42,000 g/mol. The

92 molecular weight increase in GPC traces qualitatively supported this finding. 1H NMR also showed complete deprotection of the BOC group.

CNA monomers were specifically designed to bind to DNA.33,34 A comparison of a thymine CNA repeat unit and a thymine DNA repeat unit binding to a complementary nucleotide is shown in Figure 5.1b. The differences between the two monomers are primarily in the backbone, where the CNA monomer avoids the deoxyribose sugar and phosphate group and is instead linked by a thioether bond. However, the conserved 6-atom spacing between repeat units allows for efficient hydrogen bonding between complementary nucleotides.

5.4.2 Binding of 8PEG-T to Polyadenine DNA

The binding of thymine CNA oligomers and complementary adenine ssDNA has been investigated previously, and it has been shown that binding is achieved in solution at relatively low concentrations.34 MST and CD spectroscopy were employed to confirm that binding was not affected by the conjugation of the thymine oligomers to a PEG macromolecule. A typical MST titration experiment is used to evaluate the binding of a ligand to a fluorescent analyte by tracking the change in the rate of diffusion of the unbound analyte and the bound analyte in a microscale temperature gradient.41 By titrating in the ligand, the multiarmed 8PEG-T, while keeping the fluorescent analyte concentration constant, A20-Cy5 DNA at 50 nM, a binding curve was created and the apparent dissociation constant (Kd’) extrapolated.

Figure 5.2a shows an MST titration for the 8PEG-T macromolecule and a fluorescent A20 analyte in 80% DMSO/20% H2O (v/v). This solvent mixture was chosen for two reasons.

Primarily, the 8PEG-T macromolecule is not water-soluble and requires the presence of a significant amount of DMSO to dissolve; at higher DMSO concentrations 90% and above,

Figure 5.2 – (a) Microscale thermophoresis titration experiment showing binding between 8PEG- T and a Cy-5 labelled A20 ssDNA. Solvent was 20% H2O in DMSO. A20 DNA was supplied at 50 nM. Data points are averages of technical replicates, n=3. (b) Circular dichroism of A20 ssDNA only and A20 bound to 8PEG-T. The profile of the 8PEG-T alone was subtracted out of the spectra showing binding (red curve). Spectra are averages of technical replicates, n=3. however, 8PEG has limited solubility. Additionally, on-going studies have shown that CNA-DNA exhibits weak hybridization beyond 80% DMSO. The sigmoidal increase in normalized fluorescence indicates a significant binding event that corresponds to an apparent dissociation constant (Kd’) of 19.5 μM. The Kd’ of unconjugated, homothymine CNA polymers was previously found to be 70 μM in 50/50 aqueous/DMSO mixture.34 The difference between the experimental conditions is likely the primary reason for the decrease in Kd’ as the 80/20 system has been found to be better for CNA/DNA binding than the 50/50 system (unpublished observations). However, the result confirms that PEGylation of the CNA oligomer does not interfere with ssDNA binding.

Circular dichroism was then used to study the binding event directly and the effect it has on ssDNA conformation. Figure 5.2b shows the CD spectra of A20 DNA before and after the addition of the 8PEG-T macromolecule. In pure aqueous environments, the CD spectrum of single stranded poly(A) DNA exhibits a positive Cotton effect with zero crossing at 260 nm, which is

42 attributed to base stacking. However, in a 20% H2O/80% DMSO, A20 ssDNA exhibits a shallow,

94 negative peak around 250-260 nm, suggesting limited base stacking in this solvent system. Upon introduction of the complementary 8PEG-T macromolecule, the structure clearly changes, as evidenced by the inversion of the peak at 260 nm. This spectrum is similar to duplexes in solutions with a high ethanol fraction. This behavior suggests that the duplex adopts a dehydrated A-form, as evidenced by the maximum around 260 nm with a shoulder that persists to around 290 nm.42

However, additional studies must be performed to describe fully the molecular structure of the

CNA/DNA interaction in this solvent system.

5.4.3 Formation of CNA/DNA gels

Gels were created using the 8PEG-T macromolecule by mixing with different types and lengths of ssDNA. 8PEG-T and ssDNA crosslinker were separately dissolved in 20% H2O in

DMSO at the same concentrations to confirm that each component alone did not result in a gel. It is important to note here that the synthetic scheme leaves each 8PEG-T arm with a thiol end group that could potentially form disulfide bonds. However, the observation that 8PEG-T does not form gels on its own suggests that disulfides do not significantly contribute to crosslinking. Because each arm of 8PEG-T has on average 10-12 thymine repeat units, it was hypothesized that a 40-mer homopolymer of adenine ssDNA (A40) would be sufficient to connect adjacent 8PEG-T macromers. The two components were then simply mixed, vortexed, and allowed to set at room temperature. The final gel formulation consisted of 5% w/v 8PEG-T and a 1:1 adenine:thymine ratio (e.g. a 4:1 CNA oligomer to A40 ratio). At this ratio, each CNA base has the potential to interact with a DNA base to ensure the amount of DNA added is sufficient to crosslink the network.

Attempts to de-gel the sample by strand exchange with T40 ssDNA were unsuccessful due to complete destabilization of DNA/DNA duplexes in solutions with greater than 60% DMSO content.43 Gel formation was tracked by oscillatory shear rheology and began as soon as the experiment was initiated (Figure 5.3b). The storage modulus of this gel formulation evolved quickly, reaching a plateau in the range of 500-1000 Pa in about 15 minutes.

To test the sensitivity of the gel to the type of ssDNA crosslinker, a 40-mer homopolymer of thymine (T40) and a 5-mer homopolymer of adenine (A5) were used. The T40 ssDNA does not interact specifically with the thymine CNA and was used to assess CNA/DNA binding specificity.34 The A5 ssDNA is too short to bridge two adjacent CNA arms and was used to assess the role of crosslinker length. In each case 8PEG-T was held at 5% w/v and a 1:1 CNA base to

DNA base ratio was used. Neither crosslinker led to gelation, which was confirmed by the flow observed upon vial inversion. These findings, when coupled with data from binding studies, indicate that gelation is due to sequence specific base-pairing between the thymine CNA chains 96 and the complementary DNA crosslinker. A general structure of the gel and how it is crosslinked with DNA is proposed in Figure 5.3c, which shows possible effective crosslinks, the potential for dangling chains and cycles, and the possibility of more than two CNA oligomers to interact with one A40 strand.

5.4.4 CNA/DNA gels are thermoreversible

Because CNA/DNA gels are crosslinked by temperature sensitive hydrogen bonding, the gel mechanics should be sensitive to stimuli that disrupt these interactions. For instance, high temperatures lead to the dissociation of double stranded nucleic acids, and it would therefore be expected that gels made based on these interactions would undergo a gel-sol transition at elevated temperatures. This phenomenon is well documented for DNA/DNA gels and similar materials.10,13,44,45 Shear rheology was used to track the modulus as a function of temperature to study the thermosensitivity of CNA/DNA gels. Upon increasing the temperature from 22°C to

70°C, the gels change from a predominantly elastic solid to an unstructured liquid (Figure 5.4a).

Cooling the liquid back to room temperature caused the gel to regain its structure and evolve to a

Figure 5.4 – (a) Melting and reconstitution of the CNA/DNA gel. Gels were made at 5% w/v 8PEG-T, with an A40 ssDNA, at 1:1 A:T, in 20% H2O in DMSO. Rheology was performed at 5% strain at 1 Hz. (b) Thermal cycles of CNA/DNA gels showing complete reversibility.

97 modulus of the same order of magnitude as the initial gel. Importantly, the gel fully “melts” and reforms with repeated heat/cool cycles, indicating complete thermoreversibility through at least the three cycles assessed here (Figure 5.4b). One interesting observation of this thermoreversibility is that the gel stiffness appears to increase with each successive cycle

(Supporting Figure 5.3). This observation could be explained by artifactual effects such as unavoidable solvent loss during the heating process, or by material specific phenomena such as increased crosslinking with successive cycles. The latter effect has been observed previously in

DNA gels and has been hypothesized to be a result of the reorganization of the nucleic acid strands upon dissociation, leading to more crosslinks and entanglements upon reassociation.8,10

To understand further the gel-to-sol transition, the CD spectrum was measured as a function of temperature to investigate the CNA/DNA melting process (Figure 5.5a). The change in CD at 290 nm indicates that the melting temperature is 24.5°C in 80% DMSO/20% H2O. This temperature is comparable to the melting temperature of an A12/T12 DNA duplex (aqueous, 0.05

M NaCl) at a similar concentration.46,47 The melting curve obtained from CD measurements can be interpreted as a representation of the percent of intact CNA/DNA interactions as a function of temperature. Therefore, based on this curve, all duplexes would be expected to be dissociated at temperatures of 40°C and higher. In the context of the CNA/DNA gel, this would lead to complete reverse gelation. Experimentally, however, the gels tolerated higher temperatures, ultimately

“melting” above 60°C during rheological temperature sweeps (e.g. Figure 5.4a). To confirm this observation was not due to slow kinetics of de-gelation, a second temperature sweep was performed with 60°C as the maximum instead of 70°C (Supporting Figure 5.4). The gel maintained its elasticity at 60°C, confirming that a higher temperature is needed for reverse gelation to occur. 98

One possible explanation for this phenomenon is the concentration dependence of nucleic acid melting temperatures.7,48 In this study, the dilute concentration at which the melting temperature was measured (0.3% w/v) by CD and the concentration used to form a gel (5% w/v) differed by an order of magnitude, potentially explaining the large discrepancy. To investigate whether this dependence is present in the CNA/DNA system, gels were made at 3% w/v 8PEG-T and the same 1:1 A:T ratio and the shear modulus was tracked as a function of temperature. The temperature at which reverse gelation was achieved for the 3% w/v gel was 52 ± 5°C, which was significantly lower than 61 ± 3°C, the temperature at which the 5% w/v gel reached reverse gelation. To ensure this behavior was not due to differences in gel stiffness, the normalized modulus was plotted as a function of temperature to compare the “melting” rate of each gel (Figure

5.5b). The less concentrated gel “melted” faster and at lower temperatures than the more concentrated gel. These results confirm that concentration plays a significant role in the temperature dependence of CNA/DNA gels. However, other interactions may be contributing to

Figure 5.6 – (a) Oscillation sweep of 5% w/v CNA/DNA gel. (b) Frequency sweep of 5% w/v CNA/DNA gel. Gels were made with a 1:1 A:T ratio, in 20% H2O in DMSO. Plots are representative of 3 replicates. this discrepancy, such as phase separation between the hydrophilic PEG block and the relatively hydrophobic CNA block. These and other physical interactions could lead to additional non- covalent interactions with different temperature dependencies and result in the increased reverse gelation temperature.

5.4.5 Understanding the CNA/DNA interaction within the gel network

To understand further the CNA/DNA interactions and how they contribute to the mechanical properties of the gel, the viscoelastic response to various mechanical stimuli was investigated using shear rheology. First, strain amplitude sweeps were conducted on A40 crosslinked gels (5% w/v, 1:1 A:T) to identify the linear viscoelastic region (LVR). The LVR persisted to about 45% strain and the elasticity was maintained to about 200% strain, which is the point at which all crosslinks were disrupted, and the gel began to flow (Figure 5.6a). With knowledge of the LVR, frequency sweeps were performed to identify the viscoelastic response of the gel to an oscillating strain. Based on Semenov and Rubenstein’s theory of reversible networks, the material’s frequency response can be used to extract information about the lifetimes of the

100 reversible interactions. Specifically, the crossover between storage modulus (G’) and loss modulus

(G”) provides a good estimation of the lifetime of the crosslinking interaction. In the range examined, CNA/DNA gels exhibited clear frequency dependent behavior, primarily evident by the

U-shaped loss modulus curve. Furthermore, G’ and G” crossover was not captured in the experiment, but appears to occur at frequencies below 0.01 rad/s, indicating a long lifetime of the reversible crosslink (Figure 5.6b), consistent with studies performed with DNA crosslinked polyacrylamide hydrogels.25

Stress relaxation experiments were performed on A40 crosslinked gels (5% w/v, 1:1 A:T) and at different temperatures. These experiments were used to determine the kinetic dissociation rate constant, which is correlated to the lifetime of the crosslinking interaction.49,50 Further, by plotting either the relaxation rates or characteristic relaxation times as a function of temperature in an Arrhenius plot, the activation energy required for the crosslink breakage was also determined.

Stress relaxation plots were fit with a form of the Kohlrausch-Williams-Watts stretched exponential function, which has been shown to adequately describe viscoelastic systems.51–53

Surprisingly, the relaxation profiles did not change as a function of temperature from 15 C to as high as 50 C, nearing the reverse gelation temperature (Figure 5.7a). The calculated relaxation time constants also showed no difference with temperature.

One hypothesis for this behavior is related to the 8PEG-T structure. The CNA polymerization is a step-growth mechanism resulting in polydisperse CNA chains. This polydispersity will yield a spectrum of crosslinking interactions with different numbers of base pairs. This crosslinking landscape would result in a spectrum of binding energies and melting temperatures. Thus, as temperature is increased, low energy, short CNA/DNA interactions with

101

Figure 5.7 – Representative stress relaxation plots for (a) A40 crosslinker and (b) A20 crosslinker. Relaxation profiles overlap regardless of the temperature with A40 as the crosslinker, but exhibit temperature dependent relaxation with the A20 crosslinker. Rheology was performed with a 10% immediate initial strain in both cases.

relatively few base pairs will break, while long CNA/DNA interactions, which have higher energy from additional base pairing, will persist. As a result, the average number of base pairs contributing to a crosslink (as well as the thermodynamic stability of the crosslink) will increase with increasing temperature. Thus, the relaxation profiles in Figure 5.7a may be dominated by the long CNA/DNA chain interactions. It stands to reason if the longer crosslinking interactions with more base pairs are eliminated, the temperature independence due to differences in the crosslink lengths could be mitigated and an Arrhenius dependence would be observed. Indeed, when the DNA crosslinker was shortened to A20 while maintaining a 5% w/v 8PEG-T concentration and a 1:1 A:T ratio, temperature dependent stress relaxation was observed (Figure 5.7b). It should be noted that for

102 both the A40 and A20 crosslinkers, the stress relaxation profiles at the lower temperatures appear similar, relaxing about 50% of their original stress within 30 minutes. These profiles were fit to the KWW model and the characteristic relaxation time constants were calculated. The Arrhenius activation energy for crosslink relaxation in this system was determined to be 110 ± 20 kJ/mol

(Supporting Figure 5.5). As a comparison, the activation energy for dsDNA of a comparable length (in aqueous conditions, ~1M NaCl) has been found to be approximately 220 kJ/mol.54,55

5.5 Conclusion

In this work the ability of CNAs to participate in reversible interactions with complementary ssDNA was exploited to develop stimuli-responsive, 3D gels. Single molecule binding experiments confirmed the interaction between the CNA functionalized multi-arm macromolecule, 8PEG-T, and a complementary adenine ssDNA. When 8PEG-T and a complementary ssDNA crosslinker were mixed in a more concentrated solution, gels formed and reached a modulus on the order of 1 kPa. Furthermore, these gels were viscoelastic, thermoreversible, and stress relaxing, which is consistent for physically crosslinked materials.

Several features of the responsive behavior of the gel highlight unique attributes of the CNA/DNA system, which include a strong concentration dependent “melting” and complex hybridization dynamics due to the polydispersity of the CNA oligomers. This work demonstrates the ability for

CNA to facilitate stimuli responsive crosslinking and create reversible 3D gel materials for future biological and materials science applications.

5.6 Acknowledgments.

The authors would like to acknowledge Mingtao Chen, Jasmine Singha, and Benjamin

Fairbanks for helpful discussion and input in the preparation of this manuscript. This work was

103 completed with support from an NSF MRSEC grant (DMR 1420736) and from a US Department of Education GAANN Fellowship to Alex Anderson.

5.7 References

1. Seeman, N. C. Nucleic acid junctions and lattices. J. Theor. Biol. 99, 237–247 (1982).

2. Chen, J. & Seeman, N. C. Synthesis from DNA of a molecule with the connectivity of a cube. Nature 350, 631–633 (1991).

3. Wang, C., Ren, J. & Qu, X. A stimuli responsive DNA walking device. Chem. Commun. 47, 1428–1430 (2011).

4. Goda, T. & Miyahara, Y. A hairpin DNA aptamer coupled with groove binders as a smart switch for a field-effect transistor biosensor. Biosens. Bioelectron. 32, 244–249 (2012).

5. Chaithongyot, S., Chomanee, N., Charngkaew, K., Udomprasert, A. & Kangsamaksin, T. Aptamer-functionalized DNA nanosphere as a stimuli-responsive nanocarrier. Mater. Lett. 214, 72–75 (2018).

6. Kahn, J. S., Freage, L., Enkin, N., Garcia, M. A. A. & Willner, I. Stimuli-Responsive DNA- Functionalized Metal–Organic Frameworks (MOFs). Adv. Mater. 29, 1602782 (2017).

7. Arfin, N., Aswal, V. K., Kohlbrecher, J. & Bohidar, H. B. Relaxation dynamics and structural changes in DNA soft gels. Polymer 65, 175–182 (2015).

8. Okay, O. DNA hydrogels: New functional soft materials. J. Polym. Sci. Part B Polym. Phys. 49, 551–556 (2011).

9. Karacan, P., Cakmak, H. & Okay, O. Swelling behavior of physical and chemical DNA hydrogels. J. Appl. Polym. Sci. 128, 3330–3337 (2013).

10. Topuz, F. & Okay, O. Rheological Behavior of Responsive DNA Hydrogels. Macromolecules 41, 8847–8854 (2008).

104

11. Topuz, F., Singh, S., Albrecht, K., Möller, M. & Groll, J. DNA Nanogels To Snare Carcinogens: A Bioinspired Generic Approach with High Efficiency. Angew. Chem. Int. Ed. 55, 12210–12213 (2016).

12. Cao, R., Gu, Z., Hsu, L., Patterson, G. D. & Armitage, B. A. Synthesis and Characterization of Thermoreversible Biopolymer Microgels Based on Hydrogen Bonded Nucleobase Pairing. J. Am. Chem. Soc. 125, 10250–10256 (2003).

13. Xing, Y. et al. Self-Assembled DNA Hydrogels with Designable Thermal and Enzymatic Responsiveness. Adv. Mater. 23, 1117–1121 (2011).

14. Fernandez-Castanon, J., Bianchi, S., Saglimbeni, F., Leonardo, R. D. & Sciortino, F. Microrheology of DNA hydrogel gelling and melting on cooling. Soft Matter 14, 6431–6438 (2018).

15. Um, S. H. et al. Enzyme-catalysed assembly of DNA hydrogel. Nat. Mater. 5, 797–801 (2006).

16. Nishikawa, M. et al. Biodegradable CpG DNA hydrogels for sustained delivery of doxorubicin and immunostimulatory signals in tumor-bearing mice. Biomaterials 32, 488–494 (2011).

17. Liedl, T., Dietz, H., Yurke, B. & Simmel, F. Controlled Trapping and Release of Quantum Dots in a DNA-Switchable Hydrogel. Small 3, 1688–1693 (2007).

18. Wei, B., Cheng, I., Luo, K. Q. & Mi, Y. Capture and Release of Protein by a Reversible DNA- Induced Sol–Gel Transition System. Angew. Chem. Int. Ed. 47, 331–333 (2008).

19. Hu, Y., Guo, W., Kahn, J. S., Aleman‐Garcia, M. A. & Willner, I. A Shape-Memory DNA- Based Hydrogel Exhibiting Two Internal Memories. Angew. Chem. Int. Ed. 55, 4210–4214 (2016).

20. Murakami, Y. & Maeda, M. Hybrid hydrogels to which single-stranded (ss) DNA probe is incorporated can recognize specific ssDNA. Macromolecules 38, 1535–1537 (2005).

21. Peng, L. et al. Macroscopic Volume Change of Dynamic Hydrogels Induced by Reversible DNA Hybridization. J. Am. Chem. Soc. 134, 12302–12307 (2012). 105

22. Yurke, B. Mechanical Properties of a Reversible, DNA-Crosslinked Polyacrylamide Hydrogel. J. Biomech. Eng. 126, 104 (2004).

23. Jiang, F. X., Yurke, B., Firestein, B. L. & Langrana, N. A. Neurite Outgrowth on a DNA Crosslinked Hydrogel with Tunable Stiffnesses. Ann. Biomed. Eng. 36, 1565–1579 (2008).

24. Jiang, F. X., Yurke, B., Schloss, R. S., Firestein, B. L. & Langrana, N. A. The relationship between fibroblast growth and the dynamic stiffnesses of a DNA crosslinked hydrogel. Biomaterials 31, 1199–1212 (2010).

25. Du, C. & Hill, R. J. Complementary-DNA-Strand Cross-Linked Polyacrylamide Hydrogels. Macromolecules 52, 6683–6697 (2019).

26. Tanaka, S. et al. Bulk pH-Responsive DNA Quadruplex Hydrogels Prepared by Liquid-Phase, Large-Scale DNA Synthesis. ACS Macro Lett. 7, 295–299 (2018).

27. Tanaka, S. et al. Intelligent, Biodegradable, and Self-Healing Hydrogels Utilizing DNA Quadruplexes. Chem. – Asian J. 12, 2388–2392 (2017).

28. Matteucci, M. D. & Caruthers, M. H. Synthesis of deoxyoligonucleotides on a polymer support. J. Am. Chem. Soc. 103, 3185–3191 (1981).

29. Beaucage, S. L. & Caruthers, M. H. Deoxynucleoside phosphoramidites—A new class of key intermediates for deoxypolynucleotide synthesis. Tetrahedron Lett. 22, 1859–1862 (1981).

30. Isobe, H., Fujino, T., Yamazaki, N., Guillot-Nieckowski, M. & Nakamura, E. Triazole-Linked Analogue of Deoxyribonucleic Acid (TLDNA): Design, Synthesis, and Double-Strand Formation with Natural DNA. Org. Lett. 10, 3729–3732 (2008).

31. H. El-Sagheer, A. & Brown, T. Efficient RNA synthesis by in vitro transcription of a triazole -modified DNA template. Chem. Commun. 47, 12057–12058 (2011).

32. El-Sagheer, A. H., Sanzone, A. P., Gao, R., Tavassoli, A. & Brown, T. Biocompatible artificial DNA linker that is read through by DNA polymerases and is functional in Escherichia coli. Proc. Natl. Acad. Sci. 108, 11338–11343 (2011). 106

33. Xi, W. et al. Clickable Nucleic Acids: Sequence-Controlled Periodic Copolymer/Oligomer Synthesis by Orthogonal Thiol-X Reactions. Angew. Chem. Int. Ed. 54, 14462–14467 (2015).

34. Han, X. et al. New Generation of Clickable Nucleic Acids: Synthesis and Active Hybridization with DNA. Biomacromolecules 19, 4139–4146 (2018).

35. Fairbanks, B. D., Culver, H. R., Mavila, S. & Bowman, C. N. Towards High-Efficiency Synthesis of Xenonucleic Acids. Trends Chem. (2019) doi:10.1016/j.trechm.2019.06.004.

36. Morihiro, K., Kasahara, Y. & Obika, S. Biological applications of xeno nucleic acids. Mol. Biosyst. 13, 235–245 (2017).

37. Anderson, A. J. et al. Cytocompatibility and Cellular Internalization of PEGylated “Clickable” Nucleic Acid Oligomers. Biomacromolecules 19, 2535–2541 (2018).

38. Harguindey Albert et al. Synthesis and Assembly of Click‐Nucleic‐Acid‐Containing PEG– PLGA Nanoparticles for DNA Delivery. Adv. Mater. 29, 1700743 (2017).

39. Khire, V. S., Lee, T. Y. & Bowman, C. N. Synthesis, Characterization and Cleavage of Surface-Bound Linear Polymers Formed Using Thiol−Ene Photopolymerizations. Macromolecules 41, 7440–7447 (2008).

40. Sarapas, J. M. & Tew, G. N. Poly(ether–thioethers) by Thiol–Ene Click and Their Oxidized Analogues as Lithium Polymer Electrolytes. Macromolecules 49, 1154–1162 (2016).

41. Jerabek-Willemsen, M. et al. MicroScale Thermophoresis: Interaction analysis and beyond. J. Mol. Struct. 1077, 101–113 (2014).

42. Gray, D. M., Ratliff, R. L. & Vaughan, M. R. [19] Circular dichroism spectroscopy of DNA. in Methods in Enzymology vol. 211 389–406 (Academic Press, 1992).

43. Wang, X., Lim, H. J. & Son, A. Characterization of denaturation and renaturation of DNA for DNA hybridization. Environ. Health Toxicol. 29, (2014).

107

44. Nagahara, S. & Matsuda, T. Hydrogel formation via hybridization of oligonucleotides derivatized in water-soluble vinyl polymers. Polym. Gels Netw. 4, 111–127 (1996).

45. Kang, H. et al. Near-Infrared Light-Responsive Core–Shell Nanogels for Targeted Drug Delivery. ACS Nano 5, 5094–5099 (2011).

46. SantaLucia, J. & Hicks, D. The Thermodynamics of DNA Structural Motifs. Annu. Rev. Biophys. Biomol. Struct. 33, 415–440 (2004).

47. Freier, S. M. et al. Improved free-energy parameters for predictions of RNA duplex stability. Proc. Natl. Acad. Sci. U. S. A. 83, 9373–9377 (1986).

48. Mason, T. G., Dhople, A. & Wirtz, D. Linear viscoelastic moduli of concentrated DNA solutions. Macromolecules 31, 3600–3603 (1998).

49. Parada, G. A. & Zhao, X. Ideal reversible polymer networks. Soft Matter (2018) doi:10.1039/C8SM00646F.

50. Rubinstein, M. & Semenov, A. N. Thermoreversible gelation in solutions of associating polymers. 2. Linear dynamics. Macromolecules 31, 1386–1397 (1998).

51. Sasaki, N., Nakayama, Y., Yoshikawa, M. & Enyo, A. Stress relaxation function of bone and bone collagen. J. Biomech. 26, 1369–1376 (1993).

52. Hotta, A., Clarke, S. M. & Terentjev, E. M. Stress Relaxation in Transient Networks of Symmetric Triblock Styrene−Isoprene−Styrene Copolymer. Macromolecules 35, 271–277 (2002).

53. Meng, F., Pritchard, R. H. & Terentjev, E. M. Stress Relaxation, Dynamics, and Plasticity of Transient Polymer Networks. Macromolecules 49, 2843–2852 (2016).

54. Ikuta, S., Takagi, K., Wallace, R. B. & Itakura, K. Dissociation kinetics of 19 base paired oligonucleotide-DNA duplexes containing different single mismatched base pairs. Nucleic Acids Res. 15, 797–811 (1987).

108

55. Morrison, L. E. & Stols, L. M. Sensitive fluorescence-based thermodynamic and kinetic measurements of DNA hybridization in solution. Biochemistry 32, 3095–3104 (1993).

109

5.8 Supporting Information

Supporting Figure 5.1 – Representative characterization of fully synthesized 8PEG-T macromers compared to the 8PEG-SH precursor. 110

111

* * 1000

800

600

400

200 MeanFinal Modulus (Pa)

0 1 2 3 4 Cycle Number Supporting Figure 5.3 – The mean final storage modulus of CNA/DNA gels after 3 heat/cool cycles shows a stiffening of the gel with each cycle number. Gels were made at 5% w/v 8PEG-T, 1:1 A:T ratio, and 20% H2O in DMSO. One cycle refers to a temperature ramp to 70°C followed by a cooling to 22°C and equilibration for 20 minutes. Statistical significance is denoted by * (p<0.05).

112

Supporting Figure 5.4 – Temperature sweep data from 22°C to 60°C showing that at 60°C, the gel does not melt, indicating melting temperatures are higher than 60°C.

113

KWW Function

The Kohrausch-Williams-Watts (KWW) function was used to fit stress relaxation data.

This model is primarily a phenomenological fitting model, but describes the relaxation of real materials quite well. It modifies the traditional exponential decay function with a stretching factor,

β, which gives an indication of the distribution of relaxation times that exist in the material.

푡 훽 퐺(푡) = 퐺 exp [( ) ] 0 휏

By plotting the normalized modulus, only two parameters need to be fit: τ, which is the characteristic relaxation constant, and β, which as described above is the stretching factor.

Reversible gel theory predicts that the characteristic time constant, τ, has an Arrhenius dependence on temperature. Thus, by obtaining τ as a variety of temperatures, one can extract out an estimate of the activation energy of the relaxation process and creating an Arrhenius plot from the data.

퐸 퐸 휏 = 휏 exp ( 푎 ) → ln(휏) = 푎 + ln (퐴) 0 푅푇 푅푇

The below graph shows the Arrhenius dependence of the relaxation time constant on temperature for gels crosslinked with A20 ssDNA. The activation energy for the relaxation in these gels, i.e. slope, was determined to be 110 ± 20 kJ/mol.

114

Supporting Figure 5.5 – Arrhenius plot of characteristic relaxation times. A linear fit gives an activation energy of 110 ± 20 kJ/mol.

115

Chapter 6 - Synthesis and Characterization of Click Nucleic Acid Conjugated Polymeric Microparticles for DNA Delivery Applications

6.1. Abstract

Microparticle-mediated nucleic acid delivery is a popular strategy to achieve therapeutic outcomes via antisense gene therapy. However, current methods used to fabricate polymeric microparticles suffer from suboptimal properties such as particle polydispersity and low encapsulation efficiency. Here, a new particulate delivery system based on step-growth thiol-

Michael dispersion polymerization, is reported in which a low polydispersity microparticle is functionalized with a synthetic nucleic acid mimic, namely click nucleic acids (CNA). CNA oligomers, exhibiting an average length of approximately four nucleic acid repeat units per chain for both adenine and thymine bases, were successfully conjugated to excess thiols present in the microparticles. Effective DNA loading was obtained by simple mixing; and, up to 6 ± 2 pmol of complementary DNA/mg of particle was achieved, depending on the length of DNA used. In addition, DNA loading was orders of magnitude less for noncomplementary sequences and sequences containing an alternating base mismatch. The DNA release properties were evaluated, and it was found that release could be triggered by sudden changes in temperature but was unaffected over a range of pH. Finally, phagocytosis of loaded microparticles was observed by confocal microscopy and corroborated by an increase in cellular metabolic activity up to 90%.

116

Overall, this work suggests that CNA functionalized microparticles could be a promising platform for controlled DNA delivery.

6.2. Introduction

Gene therapy is a promising option for the treatment of numerous diseases. This strategy uses antisense oligonucleotides (ASOs) to inhibit the translation of mutated or overexpressed mRNA.1–5 However, the field has been hindered by the inefficient delivery of these nucleic acid molecules due to poor cellular uptake and instability.6,7 To combat these issues, particulate polymeric delivery systems have been identified as promising carriers for nucleic acids, as they can provide protection for the ASO and be engineered to improve cellular uptake, among many other benefits.8–12 Particularly, micro-sized particles (MPs) are regarded as an important class of particulate delivery systems. MPs are easily phagocytosed by macrophages and dendritic cells, providing a simple route to target these particular cell populations.13 For instance, microparticles are routinely used for the delivery of nucleic acids to pulmonary macrophages to reduce lung inflammation.14,15

Polymeric microparticle nucleic acid delivery systems are often made of neutral or slightly negatively charged polymers such as poly(lactic-co-glycolic acid) (PLGA), poly(ε-caprolactone)

(PCL), and poly(ethylene glycol) (PEG).16–19 While these polymeric systems have shown promise in DNA delivery, they also have significant drawbacks that affect their performance. For instance, the particle size distribution has a high polydispersity which leads to an inefficient cellular uptake and reduced therapeutic effect.20–22 In addition, loading usually occurs via basic entrapment, which is difficult to achieve, non-specific, and leads to lower delivery efficiencies.23 To avoid these issues, cationic polymers are often introduced to interact electrostatically with the nucleic acid.24–

26 These polymers, however, are frequently cytotoxic at high levels.27–29 117

Herein, a monodisperse microparticle was functionalized with a novel xeno-nucleic acid, namely click nucleic acid (CNA), that is made via a thiol-ene polymerization. The microparticles are synthesized via a thiol-Michael step-growth polymerization which allows for a homogenous network structure and the ability to incorporate excess functional groups for functionalization (e.g. thiols, as in this work). CNAs have been recently demonstrated to be cytocompatible, bind to complementary DNA and RNA, and have been previously used to complex single stranded DNA

(ssDNA) for oligonucleotide delivery on the nanoscale.30–32 Furthermore, their facile, click- chemistry mediated synthesis makes them an attractive alternative to DNA for applications in which a large amount of material is required. This study examines the CNA-polymer conjugate system as a potential microscale nucleic acid delivery platform. These conjugates were characterized for oligonucleotide attachment and evaluated for their ability to load complementary ssDNA. Finally, the release of loaded ssDNA was evaluated as a function of temperature, suggesting a facile method for controlled release, and their uptake in a macrophage cell line was determined.

6.3. Experimental

6.3.1 Materials.

Dipentaerythritol hexakis(3-mercaptopropionate) (Di-PETHP) was obtained from Bruno

Bock Thiochemicals. Trimethylolpropane triacrylate (TMPTA), polyvinylpyrrolidone (PVP), 4- methoxyphenol (MEHQ), and 2,2-dimethyoxy-2-phenylacetophenone (DMPA) were obtained from Sigma Aldrich.

6.3.2 Synthesis of monodisperse, step-growth microparticles.

Microparticles were synthesized following a previously established protocol.33,34 Briefly, a hexafunctional thiol (Di-PETHP) and a triacrylate (TMPTA) were mixed in methanol at a ratio 118 that resulted in a 20% excess of thiols. To this solution was added the surfactant, PVP, and the radical inhibitor, MEHQ. The solution was mixed thoroughly and the base catalyst (TEA, 5% by weight with respect to monomers) was added. The mixture was stirred at room temperature for at least 5 hours. The resulting particles were washed extensively with mixtures of acetone and methanol to remove excess monomer, surfactant, inhibitor, and catalyst. The particles were stored at 4°C in 50:50 acetone/methanol until needed.

6.3.3 Synthesis of CNA decorated microparticles

To obtain CNA functionalized microparticles, click nucleic acid monomers were polymerized as described previously,32,35 and the thiolated microparticles were included as a dopant in the polymerization. Briefly, CNA monomer was deprotected with 2 equivalents of

NaOH, neutralized, and then extracted into dichloromethane (DCM) and dried under vacuum.

Separately, thiolated microparticles were suspended in DMF (containing 0.1% by weight of the photoinitiator (DMPA)) at a concentration of 40 mg particle / mL. This solution was added to the dried CNA monomer to achieve a monomer concentration of 20% (wt./vol.). The mixture was then sonicated to enhance diffusion of the monomer into the particle and polymerized with 365 nm UV light at 12-15 mW/cm2 for 15 minutes. The resulting particles were washed at least 3 times with either acetone or DCM to remove the unreacted monomer and oligomer as well as excess photoinitiator.

6.3.4 Characterization of CNA/microparticle copolymerization

To characterize the morphology of the CNA decorated microparticles, scanning electron microscope (SEM) images were taken. Microparticles suspended in methanol were applied dropwise to a clean silica wafer. Following solvent evaporation, samples were sputter coated with

3 nm of gold and imaged with a Hitachi SU3500 (SEM). To verify CNA/microparticle

119 conjugation, aliquots of particles were degraded under basic conditions, neutralized, and dried by lyophilization. The degradation products were analyzed by UV-Vis spectroscopy and 1H NMR in

DMSO. In addition, the molecular weight and degree of polymerization of the CNA oligomers were calculated using gel permeation chromatography (GPC) (TOSOH – HLC8320GPC) with an internal standard of short CNA oligomers.

6.3.5 Preparation of nucleic acid loaded microparticles

DNA loaded microparticles were prepared by first obtaining a 2 mg aliquot of CNA functionalized microparticles. 200 μL of phosphate buffered saline (PBS) (10 mM, pH 7.5) was added to each sample followed by 1 μL of 100 mM Cy5 labelled A10 DNA (A10-Cy5). The particles were fully suspended with sonication and mixed overnight on a vortex shaker. Samples were then washed 3x with DI H2O to remove excess DNA and salt. Microparticles were imaged on a Nikon Spinning Disc Confocal with a 640 nm excitation laser. To determine the amount of

DNA that remained in the microparticles, the particles were degraded in 5M NaOH, neutralized with HCl, and lyophilized. 300 μL of DI H2O was added and each sample was heated to 50°C for

15 minutes. After heating, the samples were quickly centrifuged to remove any insoluble components and 200 uL were taken for fluorescence measurements using a Denovix

Spectrophotometer/Fluorometer. Molar amounts of A10-Cy5 DNA were obtained using a calibration curve which confirmed a linear relationship between A10-Cy5 DNA and fluorescence in the concentration range examined.

6.3.6 Evaluation of DNA release properties

DNA loaded microparticles were suspended in PBS and incubated at various temperatures and in various pH buffers. At prescribed timepoints, 200 μl aliquots were taken for fluorescence analysis. The samples were replenished with fresh PBS and the microparticles were resuspended. 120

6.3.7 Cell Culture and Microparticle Uptake

A murine macrophage cell line (RAW 264.7, ATCC) was cultured in Dulbecco’s Modified

Eagle’s Medium (DMEM) media containing 10% fetal bovine serum (FBS, Atlanta Biologics) and

1X penicillin/streptomycin (Corning) at 37°C and 5% CO2, per the recommended protocols. Near confluent cells were collected for sub-culture and plating using a cell scraper. For uptake experiments, cells were plated at a concentration of 12,500 cells/well in Ibidi μPlates with coverslip quality glass bottoms. After 2 days of growth, cells were treated with A10-Cy5 loaded microparticles at a concentration of 100 μg/mL for 4 hours at room temperature to allow for phagocytosis. Cells were then washed with PBS and fixed with 4% PFA at room temperature for

5-10 minutes. Cell nuclei were counterstained with Hoechst 33342 (Thermo Fisher) and the cell membrane with CellTracker Orange (Invitrogen). Imaging was performed on a Nikon Spinning

Disc Confocal microscope.

6.3.8 Metabolic Activity and Cell Viability Assays

The metabolic activity of cells treated with CNA functionalized microparticles was assessed with alamarBlue™ (Invitrogen) and cytocompatibility was determined by staining cells with calcein AM (Invitrogen). For each assay, RAW 264.7 cells were plated at a density of 20,000 cells/well and allowed to adhere and grow overnight. Cells were then treated with varying concentrations of oligo(thymine) (oligo(T)) functionalized microparticles (MP+oT). Samples of wells were left untreated or treated with 1% Triton-X, which served as controls for the responses of normally proliferative cells and dead cells, respectively. Each assay was then performed after

24 hours. For the alamarBlue™ assay, the alamarBlue™ reagent was added to each well to achieve a 1:10 dilution. Absorbance of each well was read at 570 and 600 nm after 6 hours and the relative metabolic activity with respect to the negative control was calculated using the following formula: 121

푠푎푚푝푙푒 푠푎푚푝푙푒 휀600푛푚퐴570푛푚 − 휀570푛푚퐴600푛푚 % 푀푒푡푎푏표푙푖푐 퐴푐푡푖푣푖푡푦 = 푛푒푔 푐표푛푡푟표푙 푛푒푔 푐표푛푡푟표푙 휀600푛푚퐴570푛푚 − 휀570푛푚퐴600푛푚

For the cytocompatibility assay, cells were stained with 2 μM calcein AM in PBS for 30 minutes. A sample of untreated wells was left unstained to control for background fluorescence.

After staining, the fluorescence of each well was read with a 485/535 excitation/emission wavelength pair. The percentage of live cells relative to the control was calculated by the following formula:

푠푎푚푝푙푒 푏푎푐푘푔푟표푢푛푑 퐹푙485 − 퐹푙485 % 퐿푖푣푒 퐶푒푙푙푠 = 푛푒푔 푐표푛푡푟표푙 푏푎푐푘푔푟표푢푛푑 퐹퐿485 − 퐹푙485

6.3.9 Statistical Analysis

To determine statistical significance, the appropriate test was applied to the data. For histogram analysis, the Kolmogorov-Smirnov test was used (n = 200) at the 0.05 confidence level.

Differences in means, all of which were determined to be homogenous and follow Gaussian distributions, were calculated by ANOVA with Tukey’s Post-Hoc test modification at the 0.05 confidence level.

6.4. Results and discussion

The microparticles used in this work have several advantages over traditional microparticle formulations used in drug delivery (e.g., PLGA). First, the particles are synthesized through a step- growth thiol-Michael dispersion polymerization, in which the particle size is primarily determined by the gel-point conversion, resulting in a narrow polydispersity.33,34 The second advantage is the ability of the step growth mechanism to utilize off-stoichiometric systems to leave residual functionality, thereby enabling facile subsequent functionalization through any type of thiol or acrylate mediated chemistry.33,36 The particles used here were polymerized using a hexafunctional

122

thiol (Di-PETHP) and a triacrylate (TMPTA). A 20% stoichiometric excess of thiols was introduced, enabling post-polymerization functionalization with CNA via copolymerization. It has been previously shown that thiolated polymers participate in the thiol-ene CNA polymerization reaction, leading to an end-functionalized, thioether linked copolymer.32,37 Both monomers were suspended in methanol along with a surfactant (PVP) and a radical inhibitor (MEHQ). Upon addition of the base catalyst (TEA) the solution turned a milky white, indicative of particle formation. After extensive washing to remove excess monomer, surfactant, and inhibitor, the particle size distribution and morphology were determined by SEM. The particles were spherical and had an average diameter of 3.0 ± 0.4 μm.

To synthesize CNA-functionalized microparticles, the thymine and adenine monomers were polymerized in the presence of the thiolated microparticles, as indicated in Scheme 6.1, yielding oligo(thymine) (oligo(T)) functionalized microparticles (MP+oT) and oligo(adenine)

(oligo(A)) functionalized microparticles (MP+oA). Specifically, CNA monomer was dissolved in a solution of photoinitiator (0.1 wt% DMPA in DMF) containing dispersed microparticles (30 mg/mL) so that the mass of microparticles was 20% that of the monomer. The mixture was

123

extensively with various organic solvents to remove unreacted monomers and unattached

oligomers.

To assess the success of the conjugation directly, the particles were degraded under basic

conditions and the degradation products were analyzed by UV-Vis, gel permeation

chromatography (GPC), 1H NMR, and SEM. The absorption spectra of the degraded CNA-

microparticle conjugates showed a characteristic peak around 280 nm (Figure 6.1a), indicating

the presence of nucleobases, whereas the blank control particles had no such peak. In addition, the

GPC traces of the degradation products showed the presence of oligomeric CNA (Figure 6.1b).

An internal standard of short CNA oligomers was used to determine that the degree of

polymerization for thymine and adenine functionalized particles was approximately 4 for both

bases. In addition, 1H NMR was used to confirm that these spectral data were indeed due to CNA

nucleobases (Figure 6.2). SEM was used to determine if the resulting CNA-microparticle

124

Figure 6.2 - Comparison of 1H NMR spectra of the degradation products of blank microparticles to the degradation products of microparticles functionalized with oT and oA. The highlighted regions are labelled to the corresponding, color-coded protons on the CNA monomers. The presence of these peaks supports the conclusion that CNA oligomers were successfully grafted to the free thiols on the microparticles. conjugates underwent a shape or size change after copolymerization (Figure 6.3a). Analysis of particle diameters (shown in Figure 6.3b) using the Kolmogorov-Smirnov test indicated that there was a shift in the histogram of particle diameters with means from 3.0 ± 0.4 μm to 3.1 ± 0.4 μm; addition of CNA oligomers led to a slight but statistically significant increase in the microparticle size.

Next, the ability of CNA-functionalized microparticles to load complementary single stranded DNA was investigated. Previous research on CNAs has demonstrated their ability to interact with and bind to complementary oligonucleotides in entropically confined systems and in 125

Figure 6.3 – Analysis of microparticle diameters obtained through SEM shows a significant increase after CNA copolymerization. The diameter of the microparticles prior to copolymerization was 3.0 ± 0.4, which increased to 3.1 ± 0.4 after copolymerization. solution.30,31,35 It stands to reason that CNA-functionalized microparticles would show an increased loading of complementary DNA mediated through nucleobase-specific hydrogen bonding. Prior to loading, the microparticles were treated with 1M HCl to remove the Boc protecting group on the nucleobases, neutralized, and washed extensively with water and methanol. Unfunctionalized microparticles were also similarly treated to control for any change to the microparticle structure itself. Equal aliquots of unfunctionalized microparticles, MP+oT, and

MP+oA were incubated with a Cy5 labelled 10-mer of adenine DNA (A10-Cy5) in PBS with steady mixing for 24 hours. After loading, the particles were centrifuged and washed at least three times with DI H2O to remove unbound DNA. A10-Cy5 loading was qualitatively observed in the

MP+oT samples by a light blue color to the microparticle pellet which was not observed in the other samples (MPs only, and MP+oA) (Supporting Figure 6.1). 126

To directly quantify the A10-Cy5 loading, microparticles were degraded under basic

conditions, neutralized, and lyophilized. The A10-Cy5 strands were dissolved in water and aliquots

were taken for fluorescence analysis. The resulting fluorescence was compared to a calibration

curve of A10-Cy5 that underwent the same process. The MP+oT samples incorporated

significantly more A10-Cy5 DNA than either the MP+oA or MPs by multiple orders of magnitude

(Figure 6.4b). The amount of A10-Cy5 DNA loaded in the microparticle was 6 ± 2 pmol per mg

particle of MP+oT samples, which was ten-fold higher than the MP+oA samples. While the

127 amount of loading achieved here is smaller (~ a tenth) than what has been previously reported in

PLGA microparticle with CNA,30 further studies can take advantage of this highly tunable system.

For example, the thiol-Michael and thiol-ene polymerizations used herein make it possible to tune particle size, porosity, and excess thiol functionality, all of which can affect the amount of particle- bound DNA to maximize DNA loading.

To demonstrate the specificity of the CNA for the nucleic acid load, MP+oT microparticles were also incubated with other types of ssDNA. Loading was performed with a 20-mer of adenine

(A20) (to evaluate the effect of DNA length), a 10-mer of thymine (T10) (to demonstrate specificity), and a 20-mer of a guanine/adenine alternating sequence (GA10) (to probe the tolerance to mismatches) (Supporting Figure 6.2). While A10-Cy5 ssDNA had the best loading, there was also significant loading for the A20-Cy5 ssDNA, demonstrating the ability of these particles to bind to longer sequences. This binding becomes important for gene therapy applications because many antisense oligonucleotides are between 10 and 20 bases in length. There was negligible loading for both the T10 and GA10 sequences, supporting the idea that loading is due to a sequence specific interaction between CNA and a complementary oligonucleotide.

The conditions under which DNA release could be achieved were then evaluated. Because loading is primarily due to hydrogen bonding between complementary oligonucleotides, environmental triggers such as heat would have a significant effect on the release kinetics. To test this hypothesis, equal aliquots of MP+oT were loaded with A10-Cy5 and suspended in PBS at three different temperatures: 37°C, 22°C, and 4°C. The supernatant was collected at prescribed time points and analyzed for Cy5 fluorescence. Figure 6.5a details the effect temperature has on

DNA release. The samples maintained at 37°C showed an immediate, burst release which is attributed to the dissociation of most of the CNA/DNA interactions (i.e. hybridization “melting”).

128

Interestingly, for samples maintained at lower temperatures, an initial burst release is followed by

a slow and steady release over time. The initial burst release (i.e. release in the first 30 minutes) is

attributed to the release of weakly bound ssDNA. After the initial burst, the release rate is governed

by the kinetics of dissociation, which is dependent on temperature. Higher temperatures lead to

faster dissociation kinetics and therefore faster release. Such behavior has been observed before in

CNA-based materials.37 To further test this hypothesis, after 48 hours at their initial temperatures,

all samples were incubated at 55°C. Samples originally at 22°C and 4°C underwent a rapid burst

release of ssDNA, while the samples at 37°C underwent a small increase in DNA release. In

addition, A10 release as a function of pH was evaluated (Figure 6.5b). Over the range of pHs

tested (5.8 – 8.5) DNA release remained consistent, indicating good pH stability. The pH range

here represents potential subcellular environments from endosomes (pH ~5.5) to mitochondria (pH

~8.0).38

129

These results suggest that this system could be a promising controlled DNA delivery system using heat as a trigger. While Figure 6.5a indicates that nearly 90% of the total DNA is released at 37°C (i.e. body temperature), it is possible to tune this response to prevent burst release at this temperature and make triggered release more physiologically relevant. The large differences in release profiles in Figure 6.5a suggest there might be a sharp transition from slow to burst release. If the microparticle formulation and CNA polymerization conditions are optimized to control the average degree of polymerization per particle, then this transition can be adjusted to prevent burst release at 37°C but trigger it at just a few degrees higher. Local heating could then be used to trigger release and efficient control over DNA delivery can be achieved.

To be used as a DNA delivery strategy, loaded MP+oT must first be internalized by phagocytes. Previous work has shown that macrophages phagocytose micro-sized particles, but that the extent of internalization is affected by the surface characteristics of the particle.13,21,39,40

To determine whether DNA loaded, CNA-functionalized microparticles are phagocytosed by macrophages, a combination of confocal microscopy and biochemical assays was used. For these experiments, RAW 264.7 cells (murine macrophages) were used for their phagocytotic behavior.

Qualitatively, Figure 6.6a shows that cells treated with A10-Cy5 loaded MP+oT appear to internalize the microparticles as evidenced by the clear association between the particles and the cells. To confirm that the particles were indeed internalized and not adhered to the surface of the cells, three dimensional images were taken (Figure 6.6b, Supplemental Movie). These images showed that the majority of the particles resided on the same plane as the nucleus and could be

130

Figure 6.6 – (a) Macrophages appear to phagocytose A10-Cy5 loaded microparticles as evidenced by the clear association between the particles (green; the color was changed to provide better contrast) and cells. Nuclei are stained in blue. (b) 3D images of cells treated with A10-Cy5 loaded microparticles showed that microparticles existed on the same plane as nuclei and occurred within the cell membrane, suggesting internalization. The cell membrane was stained in red. (c) The metabolic activity of macrophages treated with microparticles significantly increased up to 90% after 24 hours, supporting the conclusion that the microparticles are being phagocytosed. (n = 4, * p < 0.05) (d) CNA-functionalized microparticles do not cause significant cell death. (n = 4) found within the confines of the cell membrane. To support further this conclusion, the metabolic

activity of cells treated with microparticles was evaluated with an alamarBlue™ assay. The

alamarBlue™ assay measures the concentration of high energy molecules like NADPH, whose

increase in production has been shown to be correlated with the process of phagocytosis.41,42 It

was found that cells treated with microparticles had significantly higher relative metabolic

131 activities than cells that remained untreated (Figure 6.6c). The increase in energetic molecule concentration supports the conclusion that the macrophages are taking up the CNA functionalized microparticles. Finally, the cytocompatibility of unloaded MP+oT was evaluated by a

LIVE/DEAD assay. For microparticle concentrations up to 100 μg/mL, cells experienced no significant toxicity (Figure 6.6d). Taken together, these results show that CNA-functionalized microparticles are internalized by phagocytotic macrophages and exhibit no inherent toxicity, further supporting their use as a potential DNA delivery vector.

6.5. Conclusion

This work reports a CNA-microparticle conjugate capable of interacting with and loading complementary ssDNA. The successful incorporation of CNA after copolymerization was demonstrated by 1H NMR, GPC, and UV-Vis. SEM analysis revealed a shift in the particle diameter distribution with mean diameters increasing from 3.0 ± 0.4 μm to 3.1 ± 0.4 μm upon

CNA functionalization. CNA functionalized microparticles were able to achieve a ssDNA load of

6 ± 2 pmol/mg particle. Further, the ssDNA loading was specific to sequence and could carry oligonucleotides of at least 20 bases. As expected, the ssDNA release was dependent on temperature where a simple increase in temperature led to rapid release. Finally, confocal microscopy showed phagocytosis of loaded CNA-functionalized microparticles, which was corroborated by an increase in metabolic activity up to 90%. These results suggest that the reported

CNA-microparticle conjugate system is a promising ssDNA delivery strategy, capable of rapid, temperature-dependent ssDNA release.

6.6. Supporting Information

Supporting Information is available and contains: Supporting Figure1, Qualitative images of A10-Cy5 DNA loading; Supporting Figure 2, Calibration curves to quantify DNA loading; 132

Supporting Figure 3, Loading of other DNA types; Supporting Figure 4, Calibration curves to quantify DNA release; Supporting Figure 5, More images of microparticle uptake; and a

Supporting Movie (Movie S1) showing 3D microparticle uptake.

6.7. Acknowledgements

This work was completed with support from an NSF MRSEC grant (DMR 1420736), an

NIH grant from NHLBI (5R01HL148335), a US Department of Education GAANN Fellowship to Alex Anderson, an NSF GRF award to Emerson Grey (DGE 1650115), and a U.S. National

Institutes of Health (1 F31 DE027861-01A1) Fellowship award to Nicholas Bongiardina. SEM images were taken in the Colorado Shared Instrumentation in Nanofabrication and

Characterization facility, College of Engineering and Applied Science, University of Colorado

Boulder. The authors would like to acknowledge the support of the staff and facility that have aided in this work. The imaging work was performed at the BioFrontiers Institute Advanced Light

Microscopy Core. Spinning disc confocal microscopy was performed on Nikon Ti-E microscope supported by the BioFrontiers Institute and the Howard Hughes Medical Institute.

6.8 References

1. Evers, M. M., Toonen, L. J. A. & van Roon-Mom, W. M. C. Antisense oligonucleotides in therapy for neurodegenerative disorders. Advanced Drug Delivery Reviews 87, 90–103 (2015).

2. Bennett, C. F. & Swayze, E. E. RNA Targeting Therapeutics: Molecular Mechanisms of Antisense Oligonucleotides as a Therapeutic Platform. Annual Review of Pharmacology and Toxicology 50, 259–293 (2010).

3. Passini, M. A. et al. Antisense Oligonucleotides Delivered to the Mouse CNS Ameliorate Symptoms of Severe Spinal Muscular Atrophy. Science Translational Medicine 3, 72ra18- 72ra18 (2011).

133

4. Carroll, J. B. et al. Potent and Selective Antisense Oligonucleotides Targeting Single- Nucleotide Polymorphisms in the Huntington Disease Gene / Allele-Specific Silencing of Mutant Huntingtin. Molecular Therapy 19, 2178–2185 (2011).

5. Lentz, J. J. et al. Rescue of hearing and vestibular function by antisense oligonucleotides in a mouse model of human deafness. Nat Med 19, 345–350 (2013).

6. Moschos, S. A. et al. Uptake, Efficacy, and Systemic Distribution of Naked, Inhaled Short Interfering RNA (siRNA) and Locked Nucleic Acid (LNA) Antisense. Molecular Therapy 19, 2163–2168 (2011).

7. Dirin, M. & Winkler, J. Influence of diverse chemical modifications on the ADME characteristics and toxicology of antisense oligonucleotides. Expert Opinion on Biological Therapy 13, 875–888 (2013).

8. Ferreiro, M. G., Tillman, L. G., Hardee, G. & Bodmeier, R. Alginate/Poly-L-Lysine Microparticles for the Intestinal Delivery of Antisense Oligonucleotides. Pharm Res 19, 755– 764 (2002).

9. Ahmed, A. R. & Bodmeier, R. Preparation of preformed porous PLGA microparticles and antisense oligonucleotides loading. European Journal of Pharmaceutics and Biopharmaceutics 71, 264–270 (2009).

10. De Rosa, G., Quaglia, F., La Rotonda, M. I., Besnard, M. & Fattal, E. Biodegradable microparticles for the controlled delivery of oligonucleotides. International Journal of Pharmaceutics 242, 225–228 (2002).

11. Wang, Y. et al. Systemic Delivery of Modified mRNA Encoding Herpes Simplex Virus 1 Thymidine Kinase for Targeted Cancer Gene Therapy. Molecular Therapy 21, 358–367 (2013).

12. Jiang, S., Eltoukhy, A. A., Love, K. T., Langer, R. & Anderson, D. G. Lipidoid-Coated Iron Oxide Nanoparticles for Efficient DNA and siRNA delivery. Nano Lett. 13, 1059–1064 (2013).

134

13. Tabata, Y. & Ikada, Y. Effect of the size and surface charge of polymer microspheres on their phagocytosis by macrophage. Biomaterials 9, 356–362 (1988).

14. Gayakwad, S. G. et al. Formulation and in vitro characterization of spray-dried antisense oligonucleotide to NF-κB encapsulated albumin microspheres. Journal of Microencapsulation 26, 692–700 (2009).

15. McKiernan, P. J., Lynch, P., Ramsey, J. M., Cryan, S. A. & Greene, C. M. Knockdown of Gene Expression in Macrophages by microRNA Mimic-Containing Poly (Lactic-co-glycolic Acid) Microparticles. Medicines 5, 133 (2018).

16. Hedley, M. L., Curley, J. & Urban, R. Microspheres containing plasmid-encoded antigens elicit cytotoxic T-cell responses. Nat Med 4, 365–368 (1998).

17. Jones, D. H., Corris, S., McDonald, S., Clegg, J. C. S. & Farrar, G. H. Poly(dl-lactide-co- glycolide)-encapsulated plasmid DNA elicits systemic and mucosal antibody responses to encoded protein after oral administration. Vaccine 15, 814–817 (1997).

18. Deveza, L. et al. Microfluidic Synthesis of Biodegradable Polyethylene-Glycol Microspheres for Controlled Delivery of Proteins and DNA Nanoparticles. ACS Biomater. Sci. Eng. 1, 157– 165 (2015).

19. Vroman, B., Ferreira, I., Jérôme, C., Jérôme, R. & Préat, V. PEGylated quaternized copolymer/DNA complexes for gene delivery. International Journal of Pharmaceutics 344, 88–95 (2007).

20. Brandhonneur, N. et al. Specific and non-specific phagocytosis of ligand-grafted PLGA microspheres by macrophages. European Journal of Pharmaceutical Sciences 36, 474–485 (2009).

21. Ayhan, H., Tuncel, A., Bor, N. & Pişkin, E. Phagocytosis of monosize polystyrene-based microspheres having different size and surface properties. Journal of Biomaterials Science, Polymer Edition 7, 329–342 (1996).

135

22. Katare, Y. K., Muthukumaran, T. & Panda, A. K. Influence of particle size, antigen load, dose and additional adjuvant on the immune response from antigen loaded PLA microparticles. International Journal of Pharmaceutics 301, 149–160 (2005).

23. Hsu, Y.-Y., Hao, T. & Hedley, M. L. Comparison of Process Parameters for Microencapsulation of Plasmid DNA in Poly(D,L-Lactic-co-Glycolic) Acid Microspheres. Journal of Drug Targeting 7, 313–323 (1999).

24. Capan, Y., Woo, B. H., Gebrekidan, S., Ahmed, S. & DeLuca, P. P. Preparation and Characterization of Poly (D,L-Lactide-Co-Glycolide) Microspheres for Controlled Release of Poly(L-Lysine) Complexed Plasmid DNA. Pharm Res 16, 509–513 (1999).

25. Kasturi, S. P., Sachaphibulkij, K. & Roy, K. Covalent conjugation of polyethyleneimine on biodegradable microparticles for delivery of plasmid DNA vaccines. Biomaterials 26, 6375– 6385 (2005).

26. Zhang, X.-Q., Intra, J. & Salem, A. K. Comparative study of poly (lactic-co-glycolic acid)- poly ethyleneimine-plasmid DNA microparticles prepared using double emulsion methods. Journal of Microencapsulation 25, 1–12 (2008).

27. Brunot, C. et al. Cytotoxicity of polyethyleneimine (PEI), precursor base layer of polyelectrolyte multilayer films. Biomaterials 28, 632–640 (2007).

28. Moghimi, S. M. et al. A two-stage poly(ethylenimine)-mediated cytotoxicity: implications for gene transfer/therapy. Molecular Therapy 11, 990–995 (2005).

29. Hill, I. R. C., Garnett, M. C., Bignotti, F. & Davis, S. S. In vitro cytotoxicity of poly(amidoamine)s: relevance to DNA delivery. Biochimica et Biophysica Acta (BBA) - General Subjects 1427, 161–174 (1999).

30. Harguindey Albert et al. Synthesis and Assembly of Click‐Nucleic‐Acid‐Containing PEG– PLGA Nanoparticles for DNA Delivery. Advanced Materials 29, 1700743 (2017).

136

31. Culver, H. R. et al. Click Nucleic Acid-DNA Binding Behavior: Dependence on Length, Sequence, Ionic Strength, and Other Factors. Biomacromolecules.

32. Anderson, A. J. et al. Cytocompatibility and Cellular Internalization of PEGylated “Clickable” Nucleic Acid Oligomers. Biomacromolecules 19, 2535–2541 (2018).

33. Wang, C. et al. Monodispersity/Narrow Polydispersity Cross-Linked Microparticles Prepared by Step-Growth Thiol–Michael Addition Dispersion Polymerizations. Macromolecules 48, 8461–8470 (2015).

34. Wang, C., Podgórski, M. & N. Bowman, C. Monodisperse functional microspheres from step- growth “click” polymerizations: preparation, functionalization and implementation. Materials Horizons 1, 535–539 (2014).

35. Han, X. et al. New Generation of Clickable Nucleic Acids: Synthesis and Active Hybridization with DNA. Biomacromolecules 19, 4139–4146 (2018).

36. Wang, C. et al. Photoinduced Tetrazole-Based Functionalization of Off-Stoichiometric Clickable Microparticles. Advanced Functional Materials 27, 1605317 (2017).

37. Anderson, A. J., Culver, H. R., Bryant, S. J. & Bowman, C. N. Viscoelastic and thermoreversible networks crosslinked by non-covalent interactions between “clickable” nucleic acid oligomers and DNA. Polym. Chem. (2020) doi:10.1039/D0PY00165A.

38. Casey, J. R., Grinstein, S. & Orlowski, J. Sensors and regulators of intracellular pH. Nature Reviews Molecular Cell Biology 11, 50–61 (2010).

39. Thiele, L., Merkle, H. P. & Walter, E. Phagocytosis and Phagosomal Fate of Surface-Modified Microparticles in Dendritic Cells and Macrophages. Pharm Res 20, 221–228 (2003).

40. Pacheco, P., White, D. & Sulchek, T. Effects of Microparticle Size and Fc Density on Macrophage Phagocytosis. PLOS ONE 8, e60989 (2013).

137

41. Romeo, D., Zabucchi, G., Soranzo, M. R. & Rossi, F. Macrophage metabolism: Activation of NADPH oxidation by phagocytosis. Biochemical and Biophysical Research Communications 45, 1056–1062 (1971).

42. Rossi, F. The O2−-forming NADPH oxidase of the phagocytes: nature, mechanisms of activation and function. Biochimica et Biophysica Acta (BBA) - Reviews on Bioenergetics 853, 65–89 (1986).

138

6.9 Supporting Information

Supporting Figure 6.1 – Loading of A10-Cy5 DNA is observed qualitatively by the blue hue of the pellet in the MP+oT samples. The MP+oA and MP only samples did not have the same color.

Supporting Figure 6.2 – Calibration curves to quantify the loading of different types of Cy-5 labelled ssDNA: (a) A10-Cy5, (b) T10-Cy5, (c) (GA)10-Cy5, and (d) A20-Cy5. 139

Supporting Figure 6.3 – MP+oT microparticles were loaded with A10-Cy5 DNA, A20-Cy5 DNA, T10-Cy5 DNA, and GA10-Cy5 DNA. Loading was significantly greater for the A10-Cy5 DNA, but significant loading was also observed for A20-Cy5 DNA. There was negligible loading for the non-complementary T10-Cy5 DNA and the periodically mismatched GA10-Cy5 DNA. These results support the specificity of the CNA/DNA mediated loading and demonstrates the microparticles ability to load DNA of different sizes.

140

Supporting Figure 6.5 – More representative images of microparticle uptake in RAW 264.7 cells. Nuclei are stained in blue and loaded microparticles are represented as magenta.

141

Chapter 7 - Conclusions and Recommendations

7.1 Conclusions

Oligonucleotides are powerful natural polymers that have been targeted for uses in materials science and biotechnology because of their abilities to direct self-assembly and bind to specific targets. The implementation of these biopolymers can be difficult, however with potential issues relating to stability in biological conditions and synthetic scalability. To address these issues, the field has developed a variety of modifications to the natural oligonucleotide structure, from changing functional groups of the phosphate or sugar moieties to completely substituting the natural backbone.1,2 The next step in oligonucleotide modifications has been the development of

Click Nucleic Acids (CNAs). The work presented in this dissertation aims to demonstrate CNA’s functionality as an oligonucleotide and its use as a viable alternative to natural or other modified nucleic acids. Central to the research has been the strategy of designing CNA-polymer conjugates to access new or enhanced functionalities for different biological applications. Conjugating CNA to polymeric materials achieves two important goals: 1) it increases the solubility of the polymer; and 2) allows for control over copolymer architecture. The findings of this work support the idea that CNA is exceptionally suited to be conjugated to a variety of polymeric architectures and that

CNA’s unique characteristics can be used to design materials for a variety of applications.

7.1.1 Linear polymers and copolymers

Because of its potential as a bioactive polymer, initial studies focused on CNA’s cytocompatibility and cellular interactions. However, as a homopolymer, CNA is hydrophobic due to its non-ionic, thioether backbone, and therefore its application in an aqueous, biological environment is severely limited. Initial work (Chapter 3) sought to develop a water-soluble CNA-

142 polymer conjugate and investigate its interactions with cells in vitro. Linear PEG-CNA conjugates were synthesized through a copolymerization scheme, which successfully led to a stable thioether linkage. PEG-CNA conjugates were found to have little cytotoxicity up to 100 μg/mL and to be taken up by cells within minutes through a seemingly passive uptake mechanism. Finally, colocalization analysis concluded that intercellular PEG-CNA was mostly diffuse in the cytoplasm and was not trafficked to any of the organelles studied. These studies confirmed the cytocompatibility of CNA-based copolymers and their use in biological contexts.

Work performed in Chapter 4 took the concept of linear CNA as a bioactive polymer a step further and studied its interaction with mRNA. Previous studies have confirmed CNA’s ability to specifically bind to complementary oligonucleotides and this research sought to apply these concepts to biologically relevant nucleic acids. However, in this work, the insolubility of CNA homopolymers was exploited to induce coprecipitation of the mRNA, leading to effective mRNA isolation. After careful optimization of buffer conditions, nearly 100% of mRNA could be precipitated from solution upon addition of an oligo(thymine) CNA homopolymer. This procedure was also demonstrated to work for isolation from total RNA extracts as well, giving comparable yields to commercially available products. The precipitated mRNA could then be released and analyzed via bioanalytical methods such as in vitro translation and RT-qPCR with no detectible change in mRNA functionality.

7.1.2 Branched Conjugates

In Chapter 5, the same copolymerization scheme used in Chapter 3 was employed to synthesize an 8-armed PEG-CNA conjugate. The increase in CNA chains per molecule from 1 to

8 enabled the formation of a crosslinked network capable of achieving a 3D macroscopic structure, which is a good example of how conjugate architecture can be leveraged to enable new

143 functionalities. The synthesized 8-arm PEG-CNA copolymer was found to crosslink upon the addition of complementary single stranded DNA to create a viscoelastic and thermoresponsive network. Shear rheology was used to characterize the mechanical properties of the gel while microscale thermophoresis (MST) and circular dichroism (CD) was used to directly evaluate the

CNA/DNA interaction that led to the crosslinking. Gels were capable of achieving moduli of up to 1 kPa, readily dissociated at temperatures as high as 60°C, and exhibited frequency responses characteristic of a viscoelastic material. Further, MST and CD analysis of the CNA/DNA interaction indicated that nucleic acid hybridization occurs under the gelation conditions used.

These findings support the idea that the CNA/DNA binding interaction can be exploited to design macroscopic materials by controlling the conjugate architecture.

7.1.3 Higher order conjugates

The last section of this dissertation developed a higher-order CNA conjugate as a DNA delivery system. In this context, loading of a model therapeutic DNA strand is made possible because of the vastly increased number of CNA chains per particle. This enables interactions with more oligonucleotides than either the linear or branched architectures which is important to achieve efficacious delivery. In this work, monodisperse, thiolated microparticles were included in CNA polymerizations in the same manner as in Chapters 3 and 5, resulting in CNA decorated microparticles. Effective conjugation was confirmed by analyzing the degradation products of the microparticles with UV-Vis, GPC, and 1H NMR, and CNA chains were found to be between 4-6 repeat units on average. The ability of these CNA decorated particles to load complementary ssDNA was evaluated and it was found that oligo(thymine) functionalized particles were able to carry up to nearly 6 pmol of DNA / mg particle. In addition, it was found that the length and sequence of DNA also affected the loading capacity. The release of this ssDNA was characterized

144 as a function of temperature and pH and it was found that while pH did not seem to influence release kinetics, temperature’s effect was pronounced. Finally, phagocytosis of loaded microparticles was observed via fluorescence microscopy and corroborated by analysis of relative cellular metabolic activity. In all, these results suggest that higher order CNA-microparticle conjugates are useful as a triggerable DNA delivery system.

7.2 Recommendations

Put together, this work establishes a strategy to synthesize CNA-polymer conjugates and demonstrates that these conjugates can be designed to have functionalities that are useful in a variety of biological applications. CNAs have the potential to become a widely used xeno-nucleic acid. Their development addressed two main issues that normally the implementation of other oligonucleotides: 1) the thioether linked backbone is not susceptible to enzymatic or hydrolytic degradation, increasing the stability of the polymer in biological conditions; and 2) the thiol-ene polymerization mechanism significantly improves scale of production and decreases time of synthesis. However, in order to fully realize the potential of CNAs, the ongoing study of their fundamental interactions is key.

As CNA chemistry evolves to make use of new nucleobases and polymerization techniques, we must further understand what effects these changes may have in the context of their applications. For example, all of the work presented in this dissertation makes use of homo- sequences of CNA (i.e. oligo(thymine) and oligo(adenine)). Ongoing work is focused on introducing sequence control into CNA design via orthogonal, thiol-mediated “click” reactions.3

As the capabilities of synthesizing non-trivial CNA sequences expands, the pool of potential applications does as well. For instance, these orthogonal reactions allow for the synthesis of unique

CNA triplets which can be polymerized via thiol-ene to achieve triplet repeats.3 Triplet repeats 145 occur naturally in the human genome and oftentimes are locations of mutations that cause neurodegenerative diseases.4,5 Antisense gene therapy is a promising treatment for such diseases and usually involves the delivery of oligonucleotides containing triplet repeats. Thus, one of the long-term goals of the CNA project is to design polymers to recognize and bind to specific, consequential sequences. To achieve this, much more can be done to evaluate CNA’s behavior in vitro and eventually in vivo as well. While Chapter 3 initiated this work and provided a framework of understanding, specific mechanisms of uptake, cellular trafficking, and bioactivity need to be elucidated.

Another thrust in CNA development will be in the self-assembly of more complex constructs. In this area, the development of some sequence control is also important for enabling the design of more intricate structures than what was achievable in this dissertation with single nucleobase sequences. Further, while CNA structure, sequence, and binding environment can have a large impact on the self-assembly behavior, there are still unknowns regarding exactly how these factors affect binding thermodynamics. As an illustration of this concept, the shortening of the

CNA repeat unit from 7 atoms to 6 atoms significantly increased the binding of thymine homopolymers,6 but this has not been evaluated for the other nucleobases. In addition, some recent studies have been done that look into how some of these factors affect CNA/DNA binding,7 but more research is needed on the topic. In order to develop design rules for the direction of self- assembly using CNAs, these fundamental thermodynamic questions should be answered.

7.3 References

1. Verma, S. & Eckstein, F. Modified oligonucleotides: synthesis and strategy for users. Annual review of biochemistry 67, 99–134 (1998).

146

2. Fairbanks, B. D., Culver, H. R., Mavila, S. & Bowman, C. N. Towards High-Efficiency Synthesis of Xenonucleic Acids. Trends in Chemistry (2019) doi:10.1016/j.trechm.2019.06.004.

3. Han, X., Fairbanks, B. D., Sinha, J. & Bowman, C. N. Sequence-Controlled Synthesis of Advanced Clickable Synthetic Oligonucleotides. Macromolecular Rapid Communications n/a, 2000327.

4. Evers, M. M., Toonen, L. J. A. & van Roon-Mom, W. M. C. Antisense oligonucleotides in therapy for neurodegenerative disorders. Advanced Drug Delivery Reviews 87, 90–103 (2015).

5. Smith, R. A. et al. Antisense oligonucleotide therapy for neurodegenerative disease. J Clin Invest 116, 2290–2296 (2006).

6. Han, X. et al. New Generation of Clickable Nucleic Acids: Synthesis and Active Hybridization with DNA. Biomacromolecules 19, 4139–4146 (2018).

7. Culver, H. R. et al. Click Nucleic Acid-DNA Binding Behavior: Dependence on Length, Sequence, Ionic Strength, and Other Factors. Biomacromolecules.

147

Bibliography

Chapter 1