Femtosecond Transient Absorption Study Of

Femtosecond Transient Absorption Study of Excited-State Dynamics in DNA Model Systems: Thymine-dimer Containing Trinucleotides, Alternate Nucleobases, and Modified Backbone Dinucleosides

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Jinquan Chen

Graduate Program in Chemistry

The Ohio State University

2012

Dissertation Committee:

Bern Kohler, Co-adviser

Terry L. Gustafson, Co-adviser

John M. Herbert

Dongping Zhong

Jinquan Chen

2012

ABSTRACT

DNA is the genomic information carrier for all life on Earth. The nucleobases that build up DNA absorb strongly in the UV region of the spectrum. A thorough understanding of the relaxation pathways for the energy gained from the UV radiation in

DNA is critical to our knowledge of life. The aim of this work is to answer three major questions: (1) can a excited purine base repair the adjacent thymine dimer by a photolyase-like electron transfer mechanism; (2) what are the excited state lifetimes of several xanthine derivatives; (3) what is the origin of the long-lived excited states exist in single- and double-stranded DNA.

Transient absorption study was carried on single strand model purine-containing trinucleotide. No evidence was found for that the adjacent purines adenine and guanine could repair thymine dimer by a photolyase-like electron transfer mechanism within experimental uncertainty. Excited-state dynamics of alternate nucleobases were studied by femtosecond transient absorption spectroscopy. Subpicosecond excited-state lifetimes were observed in hypoxanthine and several methylxanthine derivatives. All the compounds studied could be the candidates of possible precursor of today’s canonical nucleobases because of their photostablility property. At last, transient absorption study of adenine dinucleosides with different backbones reveals that the origin of the long-lived states in DNA oligomers is base stacking and base stacking is still presented even at high ii temperature. Meanwhile, experimental results on conjugated AT oligonucleotides suggest that the ultrafast monomer-like excited state nonradiative decay channel is presented in the well-stacked DNA systems.

iii

Dedication

This work is dedicated to my family.

ACKNOWLEDGMENTS

I would like first to thank my advisor, Bern Kohler, for his helping along this five years graduate school study.

I would also like to thank Terry L. Gustafson for willing to being my advisor after

Bern left Ohio State University and his encouragement for me during my research project.

I would like to thank the current and former members of Kohler group for their assistance and company through graduate school. Special thanks to Joseph Henrich who has taught me everything since I joined this group and his useful help from Coherent. I also want to thank Wolfgang Schreier for his guide in how to do good science in lab as well as all the travels and dinners we did together.

Finally, I would like to thank all my family members for their support during my study abroad time. Special thanks to my wife Mengyi Liu for all the encouragement and understanding during the last three years.

Vita

June 4, 1984 ...... Born - Chongqing, China

2007...... B.S. Chemistry, Nanjing University

2007 - present ...... Graduate Teaching and Research Associate,

The Ohio State University

Publications

1. Chen, J. and Kohler, B., “Ultrafast Nonradiative Decay by Hypoxanthine and Several

Methylxanthines in Aqueous and Acetonitrile Solution” Physical Chemistry Chemical

Physics, 14, 10677-10682 (2012)

2. Pan, Z.; Chen, J.; Schreier, W.J.; Kohler, B.; Lewis, F.D., “Thymine Dimer

Photoreversal in Purine-Containing Trinucleotides” Journal of Physical Chemistry B, 116,

698-704 (2012)

Fields of Study

Major Field: Chemistry

Table of Contents

Abstract ...... ii

Dedication ...... v

Acknowledgments...... v

Vita ...... vi

List of Tables ...... xii

List of Figures ...... xiii

List of Schemes ...... xviii

Chapter 1: Introduction ...... 1

References ...... 6

Chapter 2: Methods ...... 11

2.1 Transient absorption spectrometers ...... 11

2.1.1UV-pump/UV-Visible-probe measurements ...... 12

2.2. Steady-state measurements ...... 14

2.2.1. UV/Vis absorption measurements ...... 14

2.2.2. Circular dichroism measurements ...... 14

2.3. Sample handling...... 15

vii

2.3.1 Materials ...... 15

2.3.2 Spinning cell and temperature-controlled flow cell ...... 15

2.4 Data treatment ...... 16

References ...... 18

Chapter 3: Thymine dimer photoreversal in purine-containing trinucleotides ...... 26

3.1 Introduction ...... 26

3.2 Results ...... 28

3.2.1 Steady-state UV absorption measurements ...... 28

3.2.2 Transient absorption measurements of cis-syn thymine dimer ...... 30

3.2.3 Transient absorption measurements of purine-containing trinucleotides .... 31

3.3 Discussion ...... 31

3.3.1 Direct photoreversal of thymine dimer under UV irradiation ...... 31

3.3.2 Conditional quantum yield ...... 33

3.3.3 Transient absorption signals from aqueous cis-syn thymine dimer solution 37

3.3.4 Transient absorption aignals from purine-containing trinucleotides ...... 40

3.4 Conclusions ...... 42

Reference ...... 42

Chapter 4: UV hardiness of alternate nucleobasses ...... 65

4.1 Introduction ...... 65

viii

4.2 Results ...... 67

4.2.1 Steady-state UV absorption measurements ...... 67

4.2.2 Transient absorption measurements ...... 68

4.3 Discussion ...... 70

4.3.1. Ultrafast lifetimes observed in hypoxanthine and xanthine derivatives ...... 70

4.3.2. Comparison of excited-state decay mechanism between canonical and

alternate nucleobases ...... 72

4.4 Conclusions ...... 74

References ...... 74

Chapter 5: Long-lived excited states in adenine dinucleosides with different

backbones and conjugated AT system ...... 90

5.1 Introduction ...... 90

5.2 Results ...... 93

5.2.1 Steady-state UV-Vis absorption measurements ...... 93

5.2.2 Steady-state circular dichroism measurements ...... 94

5.2.3 Temperature dependent circular dichroism measurements ...... 95

5.2.4 Transient absorption measurements ...... 96

5.2.5 Temperature dependent transient absorption measurements ...... 98

5.3 Discussion ...... 99

5.3.1 Different base stacking conformation in five dinucleosides ...... 99

5.3.2 Dinucleoside structure in methanol-water solutions ...... 105

5.3.3 Thermodynamics analysis of temperature dependent results of

dinucleosides ...... 108

5.3.4 The fast decay channel is still present in well-stacked DNA system ...... 113

5.4 Conclusions ...... 116

References ...... 117

Bibliography ...... 154

List of Tables

Table 2.1 Description of components of Harrick Scientific TFC-M25-3 temperature-

controlled flow cell as labeled in Figure 2.3...... 23

Table 2.2 Description of items in temperature-controlled flow cellat working condition

as labeled in Figure 2.4...... 24

Table 2.3 Temperature differences between water bath setting and thermocouple

reading...... 25

Table 3.1 Molar absorption coefficients at wavelengths used in this study...... 61

Table 3.2 Cleavage quantum yields and conditional cleavage quantum yields...... 62

Table 3.3 Global Fitting parameter of TMP, T<>T and T<>T at the photostationary state.

...... 63

Table 3.4 Best-fit parameters for transient absorption signals in Figure 3.9 and 3.11. ... 64

Table 4.1 Lamda max values of xanthine derivatives in aqueous and acetonitrile solution

...... 86

Table 4.2 Lifetimes of xanthine derivatives in aqueous and acetonitrile solutions...... 87

Table 4.3 Lifetime and amplitude of xanthine derivatives in aqueous solutions at 290 nm

probe...... 88

Table 4.3 Lifetime and amplitude of xanthine derivatives in aqueous solutions at 280 nm

probe...... 89

Table 5.1 Global fitting parameters of five dinucleosides in buffer solutions ...... 147 xi

Table 5.2 Global fitting parameters of five dinucleosides in methanol-water solutions

...... 148

Table 5.3 Global fitting parameters of temperature dependent TA data of dinucleosides

...... 149

Table 5.4 Global fitting parameters of temperature dependent TA data of hairpin and

dumbbell conjugate A·T systems...... 150

Table 5.5 Thermodynamics parameters of five dinucleosides calculated in this study . 151

Table 5.6 Summary of thermodynamics parameters of dApdA and ApA in early studies

...... 152

xii

List of Figures

Figure 1.1 Structures of four DNA bases ...... 9

Figure 2.1 Coherent Libra femtosecond laser system ...... 19

Figure 2.2 UV-pump / UV-Visible-probe transient absorption setup on table ...... 20

Figure 2.3 Harrick Scientific TFC-M25-3 temperature-controlled flow cell ...... 21

Figure 2.4 Temperature-control flow cell at working condition ...... 22

Figure 3.1 Structure and nomenclature of the trinucleotides investigated ...... 48

Figure 3.2 Normalized UV/vis spectra of the cis-syn thymine dimer (T<>T) solution and

TpT solution...... 49

Figure 3.3 (a) UV/vis spectra of the T<>T solution after the indicated irradiation times at

at 240 nm; (b) T<>T solution reached photoequilibrium in the spin cell ring

excited by the pump pulse...... 50

Figure 3.4 UV/vis spectra of the T<>TA after the indicated irradiation times at

254 nm...... 51

Figure 3.5 UV/vis spectra of AT<>T (a) and T<>TA (b) after the indicated irradiation

times during pump-probe experiments ...... 52

Figure 3.6 UV/vis spectra of GT<>T (a) and T<>TG (b) after the indicated irradiation

times during pump-probe experiments ...... 53

Figure 3.7 Average transient absorption signal recorded between 0 and 38 minutes of

240 nm excitation and the transient absorption signal recorded between 50 xiii

and 65 minutes irradiation time ...... 54

Figure 3.8 Transient absorption signal from a reference solution of TMP ...... 55

Figure 3.9 Transient absorption signals from aqueous solutions of AMP and AT<>T . 56

Figure 3.10 Transient absorption signals from aqueous solutions of GMP and GT<>T 57

Figure 3.11 Transient absorption signals from aqueous solutions of T<>TA ...... 58

Figure 3.12 Transient absorption signals from aqueous solutions of T<>TG ...... 59

Figure 3.13 Dimer level at equilibrium predicted for the indicated values of the quantum

yield ratio...... 60

Figure 4.1 Structures of adenine, guanine, hypoxanthine and xanthine ...... 78

Figure 4.2 Structures of four xanthine derivatives in this study ...... 79

Figure 4.3 UV/Vis absorption spectra of xanthine derivatives studied in (a) aqueous

solutions and (b) in acetonitrile solutions...... 80

Figure 4.4 Transient absorption signal recorded at 570 nm of four xanthine derivatives

in aqueous solutions...... 81

Figure 4.5 Transient absorption signal recorded at 570 nm of four xanthine derivatives

in acetonitrile solutions...... 82

Figure 4.6 Transient absorption signal recorded at 250 nm of xanthine derivatives

studied in (a) aqueous solutions and (b) acetonitrile solutions...... 83

Figure 4.7 Transient absorption signal recorded at 290 nm of four xanthine derivatives

in aqueous solutions...... 84

Figure 4.8 Transient absorption signal recorded at 280 nm of paraxanthine, caffeine and

theophylline in aqueous solutions...... 85

xiv

Figure 5.1 Structures of the dinucleosides investigated ...... 122

Figure 5.2 Structuresof DNA hairpin and dumbbell conjugate A·T systems ...... 125

Figure 5.3 UV-Vis absorption spectra of (a) dAMP, dApdA, dAp(abasic)pdA and

dApc3pdA in buffer solution; (b) ATP, Ap4A and Ap5A in buffer solution.

...... 126

Figure 5.4 UV-Vis absorption spectra of (a) dApdA, dAp(abasic)pdA and dApc3pdA in4

in methanol-water solution; (b)Ap4A and Ap5A in methanol-water solution.

...... 127

Figure 5.5 UV-Vis absorption spectra of DNA hairpin and dumbbell conjugate AT

systems ...... 128

Figure 5.6 Circular dichroism spectra of (a) dAMP, dApdA, dAp(abasic)pdA and

dApc3pdA in buffer solution; (b) ATP, Ap4A and Ap5A in buffer solution.

...... 129

Figure 5.7 Circular dichroism spectra of (a) dAMP, dApdA, dAp(abasic)pdA and

dApc3pdA in methanol-water solution; (b)Ap4A and Ap5A in methanol-

water solutions ...... 130

Figure 5.8 Temperature dependent circular dichroism spectra of dAMP ...... 131

Figure 5.9 Temperature dependent circular dichroism spectra of dApdA ...... 132

Figure 5.10 Temperature dependent circular dichroism spectra of dAp(abasic)pdA .... 133

Figure 5.11 Temperature dependent circular dichroism spectra of dApc3pdA ...... 134

Figure 5.12 Temperature dependent circular dichroism spectra of Ap4A ...... 135

Figure 5.13 Temperature dependent circular dichroism spectra of Ap5A ...... 136

Figure 5.14 (a) Temperature dependent circular dichroism spectra and (b) UV melting

curve of dumbbell AKT-12...... 137

Figure 5.15 (a) Temperature dependent circular dichroism spectra and (b) UV melting

curve of hairpin AKT-13...... 138

Figure 5.16 Transient absorption signals at 250 nm following excitation at 266nm of

(a)AMP and dApdA in buffer solution; (b) dApdA dAp(abasic)pdA and

dApc3pdA in buffer solution; (c) dApdA, dAp(abasic)pdA and dApc3pdA

in methanol-water solution ...... 139

Figure 5.17 Transient absorption signals from (a) dAMP and ATP; (b) ATP, Ap4A and

Ap5A in buffer solution; (c) Ap4Aand Ap5A in methanol-water solution.

...... 140

Figure 5.18 Transient absorption signals from DNA hairpin and dumbbell conjugate A·T

systems in buffer solution...... 141

Figure 5.19 Temperature dependent transient absorption signals from (a)dApdA and (b)

dAp(abasic)pdA in aqueous solution at different temperature indicated. .. 142

Figure 5.20 Temperature dependent transient absorption signals from (a) AKT-12 and

(b) AKT-13 in aqueous solution at different temperature indicated ...... 143

Figure 5.21 Different base stacking conformation suggested by previous studies. (a) α-

face of adenine base; (b) parallel α-β stacking; (c) anti-parallel α-α or β-β

stacking...... 144

Figure 5.22 (a) Van’t Hoff plot of five dinucleosides calculated from temperature

xvi

dependent CD data; (b) Van’t Hoff plot of dApdA with assumed melting

temperatures...... 145

Figure 5.23 (a) long-lived state amplitude change vs temperature; (b) Van’t Hoff plot of

dApdA and dAp(abasic)pdA calculated from temperature dependent

TA data ...... 146

xvii

List of Schemes

Scheme 3.1 Thymine dimer (T<>T) formation and cleavage photoreactions ...... 47

xviii

CHAPTER1

INTRODUCTION

Sunlight is essential for life on earth due to the large amount of energy it provides.

However, it also contains significant amounts of harmful UV (ultraviolet, λ< 400nm) radiation. These UV photons can potentially damage DNA and lead to skin cancer.1

Adenine, thymine, guanine and cytosine (Figure 1.1), the nucleobases that build up DNA, absorb strongly in the UV region of the solar spectrum. A thorough understanding of how

DNA dissipates extra energy gained from the UV radiation is critical to our knowledge of life.

Experimental and theoretical studies publised since 2000 have focused on revealing the photophysics of nucleobases and DNA oligomers.2,3 Time-resolved femtosecond transient absorption experiments have demonstrated that all natural DNA bases can follow an ultrafast nonradiative decay channel to return to the ground state within one picosecond after UV excitation to the 1ππ* state.2-9 Theoretical studies have showed that conical intersections (CIs) provide a barrierless internal conversion decay pathway from the excited state to the ground state.10-12 However, this is not the whole story. Recent studies discovered that relatively long-lived states could also be populated during UV excitation. There is a dark 1nπ* state that exists in single pyrimidine bases in solution.

The lifetime of this dark state is 30 ps to 4 ns depending on the solvent.13-14 Pyrimidine

1 bases can also undergo intersystem crossing to the triplet states, which have μs lifetimes.13-14 Even more complex photophysics exists in single- and double-stranded

DNA oligomers. Due to base stacking and base pairing interactions, higher-ordered structures are formed in single- and double-stranded DNA oligomers. Long-lived excited states with lifetimes of tens of picoseconds and longer were observed in these oligomers.15-23 The nature of these intermediate states is still a hot research topic.

Although the photophysics of DNA has been studied for more than ten years, there are still a lot of questions that need to be answered. This dissertation is aimed at addressing three major questions. First of all, how was the ultrafast nonradiative decay channel built into today’s natural DNA bases? The reasons that caused the precursors of modern DNA base to gain this property are a mystery. Secondly, why do long-lived excited states exist in single- and double-stranded DNA and what is the origin of these states? It is interesting that nature provided DNA bases with ultrafast excited state decay channels but allowed a slower decay channel in longer DNA oligomer. Lastly, does DNA have the ability to repair photodamage by itself? If the answer is yes, then what is the mechanism of it? Background information for each question is described below.

Adenine multimers were studied intensively as a model single-stranded DNA system because of the strong π-π stacking between adenine bases.15,16,18,19,21 In this system the excited states have two different lifetimes: an ultrafast decay similar to that of the adenine monomer and a hundred picosecond decay. The assignment of this long-lived state has been controversial. Some studies assigned it to excimer/exciplex states formed between

π–stacked bases15,16 while others assigned it to excitonic states delocalized over at least

2 two bases.18 More evidence is needed to end the debate on the different previous assignments of the long-lived state. Experiments carried out on several adenine dinucleosides with different backbones at different temperature could answer the following questions: (1) is the DNA sugar and phosphate backbone necessary for generating the long-lived excited state; (2) can the long-lived states exist at higher temperature when base stacking is disrupted.

Photochemistry is a possible decay pathway of DNA excited state. A variety of photoproducts such as cis-syn cyclobutane pyrimidine dimers (CPD), pyrimidine (6-4) pyrimidone photoadducts and photohydrates can be formed in DNA under UV

24 irradiation. The most prevalent photoproduct formed by the dinucleotide TpT and longer single strand or duplex systems is the cis-syn thymine dimer. Recently thymine dimers were shown to form within 1 ps after UV excitation.25 The low quantum yield for dimerization of single strand and duplex systems possessing only A-T base pairs has been attributed to the low probability that adjacent thymines have ground state conformations appropriate for dimerization coupled with the very short lifetime of the thymine excited state.25-27 However, dimer yields are determined not only by the formation reaction, but also by the light-induced cleavage or so called photoreversal of dimers back to the original bases. Recent studies have fueled interest in whether the DNA bases can themselves photosensitize CPD repair.28-30 Electron transfer from neighboring base was proposed to be the mechanism of DNA self-photoreversal. Experimental results on trinucleotides with a purine adjacent to a thymine dimer will provide direct evidence as to the mechanism of the proposed DNA self-photoreversal process.

The ultrafast excited-state deactivation observed in isolated DNA bases is a photostability mechanism that may have led to the selection of today’s nucleobases overless photostable alternatives on the prebiotic Earth.31 It is recognized that the early

Earth was exposed to very high levels of UV radiation due to the lack of an ozone layer.32

It is obvious that the most photostable molecules would have had a better chance to become the precursors of modern DNA bases. Various lines of thought have suggested that today’s canonical nucleobases may have been preceded by alternative bases during the chemical evolution that took place before the RNA world.33-37 However, photostablility of these possible prebiotic nucleobases have not been extensively studied.

A study on photostability of alternative nucleobases will help people to better understand the general physical and chemical principles that determine photostability in nucleic acids.

This dissertation is organized into five chapters. Chapter 2 describes experimental methods and setups. In Chapter 3, steady-state and transient absorption measurements on purine-containing trinucleotides are described. Transient absorption signals of pure cis- syn thymine dimer and purnine-containing trinucleotides are recorded and the mechanism of photoreversal in this model system is discussed. Chapter 4 contains experimental results of the photostablity of hypoxanthine and four other methylxanthines.

Subpicosecond excited-state lifetimes of those five compounds are reported for the first time. Steady-state and transient absorption studies of five dinucleosides with different backbones are presented in Chapter 5. Long-lived states are observed in spite of the very different CD spectra of these compounds in aqueous solutions. No long-lived state is presented in methanol-water solutions where base stacking is denatured. Long-lived

4 states are observed at high temperatures indicating that base stacking still exists at high temperatures.

References

(1) American Cancer Society. Skin Cancer. Cancer.org. 2010. 30 Jan 2011.

(2) Crespo-Hernández, C. E.; Cohen, B.; Hare, P. M.; Kohler, B. Chemical Reviews

2004, 104, 1977.

(3) Middleton, C. T.; de La Harpe, K.; Su, C.; Law, Y. K.; Crespo-Hernández,

C. E.; Kohler, B. Annual Review of Physical Chemistry 2009, 60, 217.

(4) Pecourt, J.-M. L.; Peon, J.; Kohler, B. Journal of the American Chemical Society

2000, 122, 9348.

(5) Pecourt, J.-M. L.; Peon, J.; Kohler, B. Journal of the American Chemical Society

2001, 123,10370.

(6) Onidas, D.; Markovitsi, D.; Marguet, S.; Sharonov, A.; Gustavsson, T. Journal of

Physical Chemistry B 2002, 106, 11367.

(7) Gustavsson, T.; Sharonov, A.; Onidas, D.; Markovitsi, D. Chemical Physics

Letters 2002, 356, 49.

(8) Gustavsson, T.; Sharonov, A.; Markovitsi, D. Chemical Physics Letters 2002,

351, 195.

(9) Sharonov, A.; Gustavsson, T.; Carré, V.; Renault, E.; Markovitsi, D.

Chemical Physics Letters 2003, 380, 173.

(10) Matsika, S.Journal of Physical Chemistry A., 2004, 108, 7584.

(11) Serrano-Andrés, L.; Merchan, M. Journal of Photochemistry and Photobiology C.

2009, 10,21.

(12) M. Barbatti, A. J. A. Aquino, J. J. Szymczak, D. Nachtigallova, P. Hobza and H.

Lischka, Proceedings of the National Academy of Sciences of the United States of

America, 2010, 107, 21453.

(13) Hare, P. M.; Crespo-Hernández, C. E.; Kohler, B., Journal of Physical Chemistry

B 2006, 110,18641.

(14) Hare, P. M.; Crespo-Hernández, C. E.; Kohler, B., Proceedings of the National

Academy of Sciences of the United States of America 2007, 104, 435.

(15) Crespo-Hernández, C. E.; Kohler, B. Journal of Physical Chemistry B 2004, 108,

11182.

(16) Crespo-Hernández, C. E.; Cohen, B.; Kohler, B. Nature 2005, 436, 1141

(17) Kuimova, M. K.; Dyer, J.; George, M. W.; Grills, D. C.; Kelly, J. M.;Matousek,

P.; Parker, A. W.; Sun, X. Z.; Towrie, M.; Whelan, A. M. Chemical Communication

2005, 1182.

(18) Buchvarov, I.; Wang, Q.; Raytchev, M.; Trifonov, A.; Fiebig, T. Proceedings of the National Academy of Sciences of the United States of America 2007, 104, 4794.

(19) Schwalb, N. K.; Temps, F. Science 2008, 322, 243.

(20) Cohen, B.; Larson, M. H.; Kohler, B. Chemical Physics 2008, 350, 165.

(21) Kwok, W. M.; Ma, C. S.; Phillips, D. L. Journal of Physical Chemistry B 2009,

113,11527.

(22) de La Harpe, K.; Crespo-Hernández, C. E.; Kohler, B. Journal of the American

Chemical Society 2009, 131, 17557.

(23) Váya, I.; Changenet-Barret, P.; Gustavsson, T.; Zikich, D.; Kotlyar, A. B.;

Markovitsi, D. Photochemical & Photobiological Sciences 2010, 9, 1193.

(24) Cadet, J.; Sage, E.; Douki, T., Mutation Research-Fundamental and Molecular mechanisms of Mutagenesis 2005, 571, 3.

(25) Schreier, W. J.; Schrader, T. E.; Koller, F. O.; Gilch, P.; Crespo-Hernández, C. E.;

Swaminathan, V. N.; Carell, T.; Zinth, W.; Kohler, B. Science 2007, 315, 625.

(26) Law, Y. K.; Azadi, J.; Crespo-Hernández, C. E.; Olmon, E.; Kohler, B.

Biophysics Journal 2008, 94, 3590.

(27) Schreier, W. J.; Kubon, J.; Regner, N.; Haiser, K.; Schrader, T. E.; Zinth, W.;

Clivio, P.; Gilch, P. Journal of the American Chemical Society 2009, 131, 5038.

(28) Chinnapen, D. J. F.; Sen, D. Proceedings of the National Academy of Sciences of the United States of America 2004, 101, 65.

(29) Holman, M. R.; Ito, T.; Rokita, S. E. Journal of the American Chemical Society

2007, 129, 6.

(30) Nguyen, K. V.; Burrows, C. J. Journal of the American Chemical Society 2011,

133, 14586.

(31) A. L. Sobolewski and W. Domcke, Europhysics News 2006, 37, 20.

(32) Margulis, L.; Walker, J. C. G.; Rambler, M. Nature 1976, 264, 620.

(33) S. L. Miller, Advances in Chemical Physics 1984, 55, 85.

(34) R. Shapiro, Origins of Life and Evolution of Biospheres 1995, 25, 83.

(35) M. Levy, S. L. Miller, Proceedings of the National Academy of Sciences of the

United States of America, 1998, 95, 7933.

(36) L. E. Orgel, Critical Reviews in Biochemistry and Molecular Biology 2004, 39,

99.

(37) P. Ehrenfreund, S. Rasmussen, J. Cleaves, L. H. Chen, Astrobiology 2006, 6,490.

O NH2 N N N HN

N N H N N N H 2 H Adenine Guanine

O NH2 CH HN 3 N

O N O N H H Cytosine Thymine

Figure 1.1 Structures of four DNA bases.

CHAPTER 2

METHODS

2.1. Transient Absorption Spectrometers

The fundamental of the pump-probe transient absorption experiments was generated by an amplified Titanium-Sapphire femtosecond laser system as shown in Figure 2.1

(Libra, Coherent Inc., Santa Clara, CA). This system consists of a mode-locked femtosecond Titanium-Sapphire oscillator (Vitesse, Coherent Inc., Santa Clara, CA) which provides low-energy femtosecond seed pulses centered at 795 nm with a high repetition rate (76 MHz). The seed pulses are stretched in time by a stretcher with a grating before being sent into a regenerative amplifier. The regenerative amplifier uses two Pockels cells and a waveplate to control a seed pulse to overlap with a high energy nanosecond 532 nm pump pulse (Evolution, Coherent Inc., Santa Clara, CA) in a

Titanium-Sapphire rod. The amplified pulses are sent into a grating compressor to recover the femtosecond pulse duration. The fundamental output pulses have pulse energies of 3.6 mJ and a FWHM of 85 fs. The repetition rate is 1 kHz.

2.1.1. UV-pump/UV-Visible-probe Measurements

The experimental setups for UV-pump/UV-Visible-probe measurements are shown in

Figure 2.1. The fundamental output pulses are split by an 80/20 beamsplitter. 80% of the power is directed to pump three optical parametric amplifiers (OPA) (two OperAsolo and one TOPAS, Coherent Inc., Santa Clara, CA) and the other 20% of the power is used for third harmonic generation (THG).

UV-pump pulses at 267 nm are generated by the third harmonic generation of the fundamental 800 nm output pulses. First of all, the 800 nm pulses are reduced in beam size by a telescope which consists of two lenses (f1=250 mm and f2=-100 mm). Then the beam is split by a 75/25 beamsplitter with 25% of the power being sent to a manual delay stage for temporal overlap in a THG crystal. The other 75% of the power is sent in to a 1- mm type-I beta barium borate (BBO) crystal (27°) to generate 400 nm pulses with ~30% efficiency. The delayed 800 nm pulses with 90° polarization change by a half waveplate and the 400 nm pulses are now mixed in time and space in a1-mm type-I BBO crystal

(45°) to generate the 267 nm pulses via a sum frequency process.

UV-Visible-probe pulses are generated via an OPA (OperA solo, Coherent Inc., Santa

Clara, CA). Based on different harmonic packages the output wavelength of OPA can be tuned from 240 nm to 2000 nm. The initial processes in the OPA are the same for different output wavelengths. Briefly, a white light continuum is first generated from a sapphire plate as the seed beam. This seed beam is mixed with a small amount of 800 nm beam in a BBO crystal to generate pre-amplified signal pulses. The signal pulses are mixed with a high energy 800 nm pulse in a second BBO crystal to generate amplified

12 signal and idler pulses. The amplified signal and idler pulses can be frequency doubled or mixed with a fresh 800 nm pulse in a third BBO crystal to generate photons in the ultraviolet and visible region.

The pump pulses are chopped at 333 Hz by a mechanical chopper (New Focus, Santa

Clara, CA) and then pass through a half waveplate and a polarizer. The angle between the linear polarizations of pump and probe pulses was set to the magic angle (54.7°) via the polarizer. The pump power is controlled via the combination of the half waveplate and the polarizer. Typically the pump power is set to be 1mW (unchopped).The pump beam is focused at the sample with a spot size of 1 mm (FWHM) in the spin cell used in thymine dimer study. In the other two studies the pump spot size was chosen to be ~0.48 mm (FWHM) in the flow cell because lower pump energy was used in order to avoid photodamage of DNA samples. After passing through the sample the pump beam is blocked, preventing scattering from the pump beam from entering the detector.

A motorized IMS600PP optical delay stage (Newport Inc., Irvine, CA) with a maximum time delay of 4 ns is used to delay the probe pulses. The probe beam is also focused at the sample with a ~0.15 mm (FWHM) spot size. After the sample, the UV probe pulses are re-collimated and spectrally isolated by a monochromator and then detected by a photomultiplier tube. The visible probe pulses are detected by a joule-meter

(Molectron, Portland, OR) after passing through a 10 nm bandpass filter. Detector signals are sent to a SR830 DSP lock-in amplifier (Stanford Research Systems Inc., Sunnyvale,

CA) that referenced to the mechanical chopper. Amplified signals are retrieved through a

GPIB interface from the lock-in amplifier and recorded by LabVIEW based data collection software in a computer. The instrument response is ~200 fs.

2.2. Steady-state Measurements

2.2.1. UV/Vis Absorption Measurements

Steady-state UV/Vis absorption spectra were measured on a Lambda 25 spectrometer

(Perkin-Elmer, Wellesley, MA) at room temperature. Samples were held in a 1 mm quartz cuvette (Starna cells, Atascadero, CA). For spin cell experiments, samples were held in a home-built cell with 1 mm path length. All spectra were blanked with pure solvent.

2.2.2. Circular Dichroism Measurements

Steady-state circular dichroism (CD) spectra were recorded on a Model J-815 CD spectropolarimeter (JASCO, Easton, MD). Samples were held in the same 1 mm quartz cuvette used for UV/Vis measurements at room temperature. The solvent background spectra were corrected. For temperature-controlled measurements, samples were held in a

1 mm quartz cuvette with a stopper to prevent solvent evaporation. The CD spectra were recorded from 15 to 85℃ with 10 ℃ interval.

2.3. Sample Handling

2.3.1. Materials

In the thymine dimer study, cis-syn thymine dimer T<>T and the cis-syn dimer- containing trinucleotides were provided by the Prof. Frederick D. Lewis group at

Northwestern University.1

In the DNA base analog study hypoxanthine, paraxanthine, caffeine, theophylline, and theobromine were purchased from Sigma-Aldrich (Sigma-Aldrich Co., St Louis, MO) and used as received.

In the modified backbone study, ammonium salts of dApdA, dAp(abasic)pdA and dApc3pdA dinucleotides were purchased from Midland Certified reagent Company

(Midland, TX) as lyophilized powders. Ap4A (P1,P4-Di(adenosine-5’) tetraphosphate ammonium salt) and Ap5A (P1,P5-Di(adenosine-5’) pentaphosphate ammonium salt) were purchased from Sigma-Aldrich (Sigma-Aldrich Co., St Louis, MO) and used as received. Single capped hairpin and double capped dumbbell oligonucleotides were provide by Lewis group at Northwestern University.2

Samples were prepared in aqueous solution and in 50 mM phosphate buffer (pH 6.8).

The phosphate buffer was prepared from 25 mM Na2HPO4 and 25 mM KH2PO4 dissolved in ultrapure water (Millipore, Billerica, MA).

2.3.2. Spinning cell and Temperature-Controlled Flow Cell

Home built spinning cells designed by Middleton3 and de La Harpe4 were used in the thymine dimer study. Briefly, these cells use PTFE (Sigma-Aldrich Co., St Louis, MO)

15 and Gylon (Fluidol, Columbus, OH) spacers with thicknesses ranging from ＜100 μm to

1.2 mm.

For other studies samples were circulated between two CaF2 windows separated by 1 mm Teflon spacer in a TFC-M25-3 temperature-controlled flow cell (Harrick Scientific

Products, Pleasantville, NY) from a reservoir via a Masterflex7524-00 peristaltic pump

(Cole-Parmer, Vernon Hills, IL) as shown in Figure 2.3 and 2.4. This flow cell refreshed the sample between each set of pump and probe pulses and avoided re-excitation the sample. Sample temperature was controlled by a SC 150 circulating water bath (Thermo

Fisher Scientific, Waltham, MA) and measured with the K-type thermocouple built into the flow cell. Temperature was read out by a Meterman 38XR digitital multimeter

(Wavetek, San Diego, CA). Due to the heat gain or loss during water circulating outside of the water bath, sample temperature read out from the thermocouple was different than that of the setting temperature at the water bath. The differences are listed in Table 2.3.

2.4. Data Treatment

Transient signals were fit to the sums of exponentials in the form of ∑ Ai exp(−t /τ i ) i convoluted with a Gaussian instrument response function using the IGOR Pro 6.2 program (Wavemetrics Inc., Portland, OR). For UV-probe measurements, the full width at half-maximum of the instrument response function was determined by fitting the coherent two-photon signal of a pure solvent scan. The instrument response is 250 fs.

Reported uncertainties were twice the standard error. The coherent spike in the UV probe around t=0 was excluded from the fitting region. 16

Solvated electrons are generated by two-photon ionization of the water molecules and have absorption at 570 nm. The correction procedure was described by Crespo-

Hernández et al.5 Transient absorption signals were recorded in back-to-back measurements for the sample, an equal absorbance solution of adenine 5’-monophosphate

(AMP) and neat water. The AMP signal at delay times greater than several ps arises solely from solvated electrons produced by the two-photon ionization of water.6 In the equal absorbance solutions of AMP and the sample, the water solvent experienced the same amount of pump photons and thus resulted in the same amount of two photon ionization. These conditions result in equal populations of solvated electrons if the reasonable assumption is made that one-photonionization of the solute (AMP or the sample) can be neglected. The neat water signal was scaled to the long time signal of

AMP and then subtracted from the sample signal.

References

(1) Pan, Z. Z.; McCullagh, M.; Schatz, G. C.; Lewis, F. D. Journal of Physical

Chemistry Letters 2011,2, 1432.

(2) McCullagh, M.; Zhang, L.; Andrew, K.; Zhu, H.; Schatz, G. C.; Lewis, F. D.J.

Phys.Chem. B 2008, 112, 11415.

(3) Middleton, C.T. Dissertation, The Ohio State University, 2008.

(4) de La Harpe, K.D. Dissertation, The Ohio State University, 2011.

(5) Crespo-Hernández, C. E.; Cohen, B.; Hare, P. M.; Kohler, B. Chemical Reviews

2004, 104, 1977.

(6) Crespo-Hernández, C. E.; Kohler, B. Journal of Physical Chemistry B 2004, 108,

11182.

Figure 2.1 Coherent Libra femtosecond laser system. Output parameters of the regenerative amplifier are 3.6 mJ/pulse, FWHM 85fs, 800 nm at 1 kHz repetition rate.

Figure 2.2 UV-pump / UV-Visible-probe transient absorption setup on table. The solid purple line stands for 266 nm as the pump and the dashed purple line stands up for 266 nm as the probe. (Special thanks to Dr. Wolfgang Schreier for designing this new setup)

Figure 2.3 Harrick Scientific TFC-M25-3 temperature-controlled flow cell.

Figure 2.4 Temperature-controlled flow cell when working.

Table 2.1 Description of components of Harrick Scientific TFC-M25-3 temperature- controlled flow cell as labeled in Figure 2.3.

Item Description

a. Harrick Scientific TFC-M25-3 cell body

b. Rear window 25.0 mm diameter 2 mm thickness CaF2

c. Teflon spacer 1 mm thickness

d. Front window 25.0 mm diameter 1 mm thickness CaF2

e. Compression ring

f. Retaining nut

Table 2.2 Description of items in temperature-controlled flow cell at working condition as labeled in Figure 2.4.

Item Description

a. Harrick Scientific TFC-M25-3 flow cell

b. Masterflex7524-00 peristaltic pump

c. Meterman 38XR digital multimeter

d. Sample reservoir ( need to put it into item e when working)

e. Thermo-Fisher SC 150 circulating water bath

Table 2.3 Temperature differences between the water bath setting and the thermocouple reading.

Thermo-Fisher SC 150 K-type thermocouple

temperature setting (ºC) reading (ºC)

5 7

14 15

25.5 25

37.5 35

48.5 45

59.5 55

70 65

81.5 75

92.5 85

CHAPTER 3

THYMINE DIMER PHOTOREVERSAL IN PURINE-CONTAINING

TRINUCLEOTIDES*

3.1. Introduction

UV excitation single-stranded and duplex DNA will lead to the formation of mixtures of photoproducts, such as cyclobutane pyrimidine dimers (CPDs), (6-4) photoadducts, and oxidative photoproducts.1 These photoproducts, if not repaired, will lead to various types of mutations which could cause the formation of skin cancer lesions

2 eventually. The major product formed by the dinucleotide TpT and longer single strand or duplex systems is the cis-syn (2+2) dimer (T<>T, reaction 1 in Scheme 1), one of a class of CPDs. CPDs are believed to form by the concerted cycloaddition reaction of adjacent- stacked thymine bases.3,4 Because of the low probability that adjacent thymines have ground state conformations appropriate for dimerization as well as the very short lifetime of the thymine excited state the quantum yield for dimerization of single strand and duplex systems is very low.5,6 Recently a growing number of studies reveal significant conformational effects on dimer yields.2,3,7,8

It is worth noting that dimer yields are determined not only by the formation reaction, but also by the light-induced cleavage of a dimer back to the original bases. This process,

*Adapted with permission from Pan,Z.; Chen, J.; Schreier, W.J.; Lewis, F.D.; Kohler, D. J. Phys. Chem. B 2012, 114,698. Copyright 2012 American Chemical Society. 26 known as photoreversal, occurs by two mechanisms (Scheme 1, reactions 2 and 3).9 A dimer can be directly excited by deep UV light, creating an excited state that dissociates with high quantum yield.9,10 Alternatively, the excited state of a secondary chromophore can sensitize cleavage by one-electron oxidation or reduction of the dimer.

Photosensitized dimer reversal is the mechanism used by the enzyme photolyase to repair

CPDs. This enzyme, which is found in all three kingdoms of life, uses visible and near

UV light to initiate electron transfer from a flavin cofactor to a dimer.12-14 Cleavage of the dimer radical anion ensues, followed by back electron transfer to the oxidized electron donor.12 Photosensitized cleavage has also been studied in photolyase-inspired model systems in which an excited-state electron donor molecule is covalently tethered to a

CPD.15-20

Recent studies have fueled interest in whether the DNA bases can themselves photosensitize CPD repair.21-23 It was reported that in a DNA sequence an incorporated thymine dimer could be catalytically repaired by in vitro selection when exposed to 305 nm light.21 This self-repairing deoxyribozyme was proposed to function in its folded state by donating one or more electrons from a guanine-quadruplex to a nearby thymine dimer in a mechanism analogous to that used by the photolyase repair enzymes. In a later study,

Rokita et al. reported that the amount of thymine dimers induced in double-stranded

DNA 18-mers by 254 nm light is reduced when the dimer site is adjacent to a guanine.22

They suggested that electron transfer from a proximal guanine to the thymine dimer reduces photoproduct yields under photostationary conditions. The efficiency of duplex

TT dimerization depends upon the identity of the flanking base pairs, and is lower for

27 purine vs. pyrimidine flanking bases, for guanine vs. adenine bases, and for 5' vs. 3'- guanine.22,24-26 Charge transfer from neighboring purine bases has been proposed as a mechanism for self repair of thymine dimers in DNA.21,22 In this process electron transfer from photoexcited A or G to T<>T is expected to be exergonic as the singlet energies of both purines are higher (ca. 4.5 eV) than the energy needed to form the corresponding radical ion pairs. Oxidation potentials for G and A (i.e. E0(G⋅+/G) and E0(A⋅+/A)) in acetonitrile solution are 1.49 and 1.96 V vs. NHE, respectively.27 The reduction potential of T<>T is reported to be -2.2 V vs. SCE,28 corresponding to –1.95 V vs. NHE.

Therefore reduction of T<>T is thermodynamically slightly more facile than reduction of thymine as E0(T/T⋅–) = –2.14 V vs. NHE in acetonitrile. These facts make electron transfer repair by photo-excited guanine a plausible hypothesis. Additionally, it has been proposed that singlet states in DNA decay to charge transfer states, especially when A is stacked with T.29,31,32

To determine whether the electron transfer repair mechanism is present in DNA, steady-state and transient absorption measurements were carried out on pure cis-syn thymine dimer first and then on trinucleotide model systems containing a single cis-syn dimer that is adjacent to a purine base (A or G) on the 3’ or the 5’ position. Structures of the trinucleotides investigated are shown in Figure 3.1. The results indicate that photoreversal in these single-stranded model systems proceeds overwhelmingly via direct electronic excitation of T<>T with negligible repair upon electronic excitation of the neighboring purine base.

3.2. Results

3.2.1. Steady-State UV Absorption Measurements

Normalized UV absorption spectra of pure cis-syn thymine dimer and TpT are shown in Figure 3.2. This data matches well with previous study.30The extinction coefficient of thymine dimer is much smaller than that of TpT as summarized in Table 3.1. The decreasing of 267 nm absorption band (maximum absorption band of thymine) indicates the formation of thymine dimer. In our study we use the increasing of this band as the signal of thymine dimer photoreversal.

The change of UV absorption spectra of T<>T solution under 240 nm irradiation are clearly shown in Figure 3.3 a. The initial absorption spectrum (red curve) matches that of the thymine dimer. The absorbance near 267 nm increases dramatically with increasing irradiation time, and a broad band develops as expected when thymine dimers are cleaved to two separate thymine residues. After approximately one hour irradiation the T<>T solution held in the spin cell reached a photo-equilibrium state. UV absorption spectra of this state are shown in Figure 3.3 b. It is worth to point out that here instead of the total solution in the spinning cell only the amount of solution under the ring excited by the pump pulse reached the photo-stationary state as discussed in Chapter 2.3.2. The reappearance of the thymine absorption band indicates very efficient thymine dimmers photoreversal. Irradiation experiment has carried out on T<>TA to demonstrate that the purine-containing trinucleoties would have similar result when exposed to UV irradiation and the data are shown in Figure 3.4. Similar as the pure thymine dimer irradiation results increase of 267 nm absorption band is observed in T<>TA under 254 nm irradiation.

UV absorption spectra of four pure-containing trinucleotides under 266 nm irradiation are shown in Figure 3.5 and Figure 3.6. Unlike pure thymine dimer all four trinucleotides are highly photostable and irradiation of all four trinucleotides for over an hour caused virtually almost no change in the UV absorption spectra.

3.2.2. Transient Absorption Measurements of cis-syn Thymine Dimer

Figure 3.7 shows the transient absorption signal of cis-syn thymine dimer at a probe wavelength of 266 nm obtained by averaging five individual scans recorded between 0 and 37.5 minutes of irradiation time. Because the probe transmission steadily decreases with irradiation time due to dimer cleavage, the probe signal level used to calculate ΔA was the average value measured before and after each scan with the pump pulse blocked.

After the positive spike due to simultaneous absorption of pump and probe photons near time zero, a negative ΔA signal is observed at earliest times, but this signal rapidly increases to yield a positive offset that then increases much more slowly out to the longest delay times studied. Signal averaging was necessary to obtain a reasonable signal-to-noise ratio, but lengthy averaging is not feasible in these experiments because continued photolysis by the pump pulses continuously changes the solution composition.

Despite the noise in individual scans, the positive signal amplitude seen at long delay times very clearly decreases from scan to scan. Averaging the signal between 10 and 500 ps produces average ΔA values for the first five scans of 0.38, 0.26, 0.21, 0.14, and 0.12 mOD. The positive signal decreases steadily with increasing irradiation time until after approximately 1 h of total irradiation time the signal is negative at all delay times (black

30 circles in Figure 3.7). The solution after 50 minutes of irradiation time consists of an approximately photostationary mixture of thymine dimers and cleaved dimers. The signal recorded from the approximately photostationary mixture is indistinguishable from the signal recorded from a solution of the mononucleotide TMP as shown in Figure 3.8.

3.2.3. Transient Absorption Measurements of purine-containing Trinucleotides

Pump-probe experiments were carried out on four purine-containing trinucleotides using a pump wavelength of 266 nm to excite mainly the purine base. The molar absorbance coefficient data in the last two rows of Table 3.1 suggests that 97 – 98% of all photons at this wavelength are absorbed by adenine or guanine. Transient absorption signals are independent of irradiation time within experimental uncertainty and are very similar to the signals observed using the same pump and probe wavelengths for a solution of AMP or GMP (Figure 3.9 and Figure 3.11). Similar results were observed for and for

T<>TA in Figure 3.10 and T<>TG in Figure 3.12.

3.3. Discussion

3.3.1. Direct Photoreversal of thymine dimer under UV irradiation

The steady-state experiments showed that under 240 nm irradiation direct photoreversal of thymine dimer is very efficient. The cleavage quantum yield, Φ , for

T<>T at 240 nm is 0.41 ± 0.04.33 This value is similar to that reported by Sztumpf and

Shugar for 254 nm irradiation,34 but lower than two previously reported values for 240 nm irradiation.27,32 Because paper chromatography was used for product analysis and

31 secondary actinometers were used to measure light intensity in these earlier studies, our value should be more reliable. This value is also consistent with cleavage occurring via a conical intersection which partitions approximately equally between product and starting material.35

It is reported that positioning a purine base next to the thymine dimer suppresses the quantum yield of cleavage in every case, but yields do not depend significantly on whether the purine base is on the 5’ or the 3’ side of the dimer.33 The magnitude of suppression, however, does depend sensitively on wavelength. Cleavage yields are lower by a factor of between four and six at 240 nm. At 280 nm, yields are ~50 and ~100 times lower for the A- and, G-containing trinucleotides, respectively, compared to the T<>T cleavage quantum yield of 0.41 ± 0.04 measured at 240 nm. The cleavage quantum yield,

Φ , is about 30% larger for the G- vs. A-containing trinucleotides at 240 nm. At 280 nm, however, yields are about twice as high for AT<>T and T<>TA as for the G-containing trinucleotides. These patterns are difficult to explain by the electron-donating ability of G, a property that should depend weakly, if at all, on excitation wavelength, but they are fully consistent with cleavage resulting from direct excitation of T<>T.

The probability that an excitation is formed initially on T<>T is shown in parentheses in the final two rows of Table 3.1, assuming that molar absorption coefficients of the purine and T<>T are additive and neglecting energy transfer and the formation of delocalized excitations (excitons). These approximations are likely to hold for these dimer-containing trinucleotides which like T<>T-containing duplexes are expected to be poorly stacked.36 In particular, hypochromism, which would cause the true molar

32 absorption coefficients to be smaller than those estimated from summing coefficients for individual chromophores, should be small.

3.3.2. Conditional Quantum Yield

In photochemistry, a quantum yield,Φ, can be defined for any event triggered by light absorption, and is given by the number of times the outcome of interest is observed, such as the emission of a photon or the formation of a photoproduct, divided by the total number of absorbed photons. When two or more kinds of absorbers are present, as in a molecular mixture or a multichromophoric molecule like DNA, the quantum yield measures only the overall probability of reaction etc. and does not provide a measure of how efficiently the event of interest is initiated by excitations associated with the various kinds of absorbers. Here we seek the probability of observing the outcome of interest given excitation of a particular kind of absorber. This is described in probability theory by the concept of conditional probability37 and we introduce the term conditional quantum yield for the quantum yield of a process when excitation is restricted to a particular absorber.

Consider a photoreaction that occurs in a system with two kinds of absorbers, X and

Y. The conditional quantum yield of reaction given that X is excited is written as ΦX .

Similarly, the conditional quantum yield of reaction given that Y is excited is written as

ΦY . The conditional quantum yield can be defined as,

number of events induced by X Φ = . (3.1) X number of photons absorbed by X

A complication arises in situations where it may be difficult to know how to partition excitation between X and Y as in the case of excitonic states where both X and Y can be excited at the same time. A further complication is when energy transfer can occur from

X to Y or vice versa. Ignoring these possibilities, we assume henceforth that excitation can be assigned either to X with probability P(X), or to Y with probability P(Y). If X and

Y are the only two absorbers, then the assumption is made that,

P(Y) =1− P(X). (3.2)

Because conditional quantum yields are conditional probabilities, they obey Bayes’ formula,37

Φ P(X)+ Φ P(Y) =Φ. (3.3) X Y Although it is straightforward to count all events of interest, it is generally difficult to determine the subset of events caused, for example, by excitation of X (i.e. it is difficult to determine the numerator in eqn. 3.1). Because the conditional quantum yield is difficult to measure directly, a useful estimate is given by the (unconditional) quantum yield, Φ, divided by the probability of forming the excited state of interest. Thus, an

estimate, ΦX , for the conditional quantum yield of cleavage given that X is excited is given by,

Φ Φ = . (3.4) X P(X)

Combining eqns. 3.2 through 3.4 yields,

1− P(X) Φ +Φ = Φ . (3.5) X Y P(X) X

Because the second term on the left in eqn. 3.5 is always > 0, ΦX is an upper bound on

the conditional quantum yield, ΦX , and,

ΦX ≤ ΦX. (3.6)

The equality sign in eqn. 3.6 holds either when ΦY = 0 (i.e. excitation of Y cannot produce the event of interest) or when excitation is restricted exclusively to X (i.e. P(X) =

1). Similarly,

1− P(Y) Φ +Φ = Φ (3.7) X P(Y) Y Y where,

Φ Φ = . (3.8) Y P(Y)

Adapting the above formalism to the problem of dimer cleavage, we seek the conditional quantum yield of cleaving a dimer given that the purine base in the trinucleotides described in the main paper absorbs a photon. In addition, we seek the conditional quantum yield of dimer cleavage given that T<>T is excited. These quantities are estimated from eqns. 3.9 and 3.10, respectively, where X now stands for the purine base adjacent to the dimer.

Φ Φ = (3.9) X P(X)

Φ Φ = (3.10) T<>T P(T<>T)

The estimated conditional quantum yields ΦX and ΦT<>T are tabulated in Table 3.2.

All ΦT<>T values in Table 3.2 are comparable in size to the splitting quantum yield of

0.41 measured for T<>T. This indicates that direct photoreversal is the dominant cleavage mechanism because efficient electron transfer repair from the purine would

otherwise cause ΦT<>T values to exceed unity. ΦT<>T is twice as large for the G- containing than for the A-containing trinucleotides at both 240 and 280 nm (Table 3.2).

This could indicate that direct photoreversal is less efficient when A vs. G is attached to the dimer. However, the molar absorption coefficients of the constituent subunits are uncertain, as is the procedure for properly partitioning excitation probabilities in multichromophoric molecules like DNA. Trends in the conditional quantum yields should therefore be interpreted cautiously.

According to Table 3.2, the conditional quantum yield for T<>T cleavage by electron transfer from a neighboring purine is less than 0.008 at 280 nm, and less than about 0.1 at

240 nm. Although ΦX increases ten to thirty times from 280 to 240 nm, it is unlikely that

the true value of ΦX increases in this way. Because ΦT<>T is large, the difference between

ΦX and ΦX increases with decreasing wavelength as the probability of exciting the purine decreases (see eqn. 3.5). If the reasonable assumption is made that cleavage efficiency is independent of wavelength, whether cleavage occurs by direct excitation of

T<>T, or following excitation of the purine, then ΦX should be no greater than the smaller of the two values, or no greater than about 0.008. We conclude that there is negligible repair of thymine dimer due to excitation of the purine base at both 280 and

240 nm. To further identify the possible mechanism of photoreversal of thymine dimer

36 under 240 nm irradiation pump-probe experiments were carried out to look for dynamical signatures of electron transfer.

3.3.3. Transient absorption signals from aqueous cis-syn thymine dimer solution

Experiments were first carried out on pure cis-syn thymine dimer in order to first establish the kinetics of direct cleavage. Due to the competing cleavage and dimerization reactions an approximate equilibrium or photostationary state is established in our sample.

Transient absorption signals of thymine dimer and trinucleotides were recorded from a solution of the sample held between two 25.4 mm diameter CaF2 windows separated by a spacer with a thickness of 1.2 mm. The pump and probe spots were overlapped 8.0 mm from the center of rotation of the spinning cell and the cell was spun by an electrical motor at several hundred rpm with the axis of rotation perpendicular to the windows and approximately parallel to the direction of the crossed laser beams. As the cell rotates, the overlapped beams sample a circular ring of circumference 2 × 8 mm or ~50 mm. For a pump spot diameter of ~1 mm, this means that the longest time before previously excited solution will again be located at the position of pump-probe overlap is 0.15 s, which is the time for 50 pump pulses. Because voltage readings from the lock-in amplifier are made for approximately 7.5 s at each delay setting, the transient absorption measurements do not probe a solution of pure cis-syn dimers, but instead interrogate the induced absorbance change in a solution that is continuously changing in composition. Control experiments showed that the rate of diffusion of unexcited solution into the ring exposed to the pump beam occurred at a slow rate of several tens of minutes. Consequently, the

37 time needed to reach photoequilibrium is determined by the amount of solute in the ring excited by the pump pulse and not by the total solute in the spinning cell. The rapid approach to the photostationary state during the course of our measurements causes the transient absorption signals to change with irradiation time as shown in Figure 3.7.

Our quantity of purified dimer limited us to working with no more than 1.0 ml of solution at a time. Given a pump beam radius of 0.5 mm, a sample path length of 1 mm, and the pump pulse repetition rate after chopping of 1000 Hz / 3, nearly 16 ml of solution would be needed every minute if every pump pulse were to excite fresh solution. Because

5 to 10 minutes of signal averaging are required to obtain a reasonable signal-to-noise ratio using our spectrometer for a transient absorption scan, it was impossible to avoid re- exciting previously irradiated solution with the 0.3 mL-volume spinning cell used in these measurements. This results in continuous photolysis and the establishment of photostationarity in under an hour. In fact, a true photostationary state is not reached for

TpT because of the formation of non-reversible products like the (6-4) photoproduct.

However, because the quantum yields of formation of irreversible products like the (6-4) photoproduct are about an order of magnitude lower than for the thymine dimer, the solution is approximately photostationary under our experimental conditions.

The evolution from a fresh sample of pure T<>T to one that is approximately photostationary yields the transient absorption signals shown in Figure 3.7. Nonlinear least-squares fits to the data (see Table 3.3) confirm that the time-dependent signals decay identically within experimental uncertainty, but the signals recorded before the

38 photostationary state have a substantial positive offset that steadily decreases with irradiation time.

A reversible photoreaction that has attained photostationary concentrations of reactant and product will always have a transient absorption signal of zero at delay times long enough that all excited states and other intermediates have decayed away. In the case of

T<>T, photostationarity means that the number of dimers cleaved by each pump pulse exactly equals the number of new dimers formed. The transient absorption signal from the photostationary solution is shown by the black curve in Figure 3.7. The signal does approach zero, as expected at photostationarity, but only several hundred ps after excitation. At shorter times, the signal agrees fully with the signal recorded from a solution of the monomer TMP as shown in Figure 3.8. The TMP signal has fast and slow kinetics with time constants of 2.1 ps and 130 ps (Table 3.3). The fast kinetics are due to vibrational cooling following ultrafast conversion, while the slower kinetics reflect the lifetime of 1nπ* excited states that relax more slowly.38 It was shown earlier that UV/UV pump-probe signals for TpT differ insignificantly from those of the mononucleotide

TMP,32 and this is the reason for the excellent agreement in Figure 3.8.

The photostationary concentrations depend sensitively on the irradiation wavelength

(Figure 3.13) because reactant (TT) and product (T<>T) differ so greatly in their absorption spectra. With 254 nm irradiation from a germicidal lamp, only about 25% of

TpT can be converted to dimer products, but irradiation at 280 nm converts 90% of TpT to dimer.27 Despite this variation, the ratio of the number of excited states of TpT to the number of excited states of T<>T formed by each pump is the same at all pump

39 wavelengths for the photostationary solution, when the reasonable assumption is made that the quantum yields of the forward and reverse photoreactions are independent of wavelength. Given that the quantum yield of cleavage of T<>T (0.4133) is 40 times higher than the quantum yield of dimer formation in TpT (~0.0142), then 40/41 or 98% of excited states formed by the pump pulse are localized on TpT under photostationary conditions regardless of the pump wavelength. This is the reason why the kinetics seen from the photostationary solution are identical within uncertainty to kinetics seen from

TMP.

At times before photostationarity is established, transient absorption signals display a positive offset, but no new time constants are detected. These results are consistent with prompt (i.e. subpicosecond) dissociation of excited T<>T, a reasonable result for a unimolecular photoreaction characterized by a high quantum yield.

3.3.4. Transient absorption signals from purine-containing trinucleotides

In the pump-probe experiments on purine-containing trinucleotides, the 266 nm pump wavelength is expected to excite the purine base 97 – 98% of the time (estimated from the molar absorption coefficients shown in Table 3.1). The presence of the purine base could alter the time needed for reaching photostationarity, but should not alter the ultimate levels of dimer vs. non-dimerized TT. The low steady-state quantum yields for dimer cleavage at 280 nm indicate that attainment of the photostationary state at the 266 nm pump wavelength will happen much more slowly than for the T<>T solution.33 This was indeed the case and transient signals shown in Figures 3.5 and Figure 3.6 were

40 invariant with irradiation time since the start of measurements. No positive offset could be seen at long delay times that would indicate the dimer cleavage that is responsible for the positive signals seen in Figure 3.7.

There is also no evidence from AT<>T and GT<>T signals (Figures 3.9 and Figure

3.10) of electron transfer from the purine to the thymine dimer. The decays are unchanged compared within experimental uncertainty compared to what is observed from

AMP or GMP alone (Table 3.4). T<>TA and T<>TG also have similar transient absorption data shown in Figure 3.11 and Figure 3.12. These data therefore rules out the possibility of efficient electron transfer from the excited purine to T<>T followed by efficient back electron transfer before cleave can occur.

The results show that electron transfer from G or A to T<>T does not occur to a significant extent in single-stranded trinucleotides. Through-space or through-bond electron transfer in unstacked systems such as our trinucleotides may be too slow to compete with rapid non-radiative decay of the purine excited state. Singlet excited state of all native nucleobases can decay via ultrafast internal conversion with a time constant of less than 1 ps.29 1nπ* states with lifetimes of several tens of picoseconds have been observed in pyrimidine, but not in purine bases.38 It is interesting to note that the rate of forward electron transfer in photolyase takes place with biexponential kinetics and lifetimes of 60 and 335 ps,39 or much more slowly than excited-state relaxation by native nucleobases. In addition, although definitive kinetic measurements are lacking, it has been suggested that electron transfer takes place on the ns timescale in model compounds in which a sensitizer is tethered to a thymine dimer.40,41

3.4 Conclusions

Thymine dimer photoreversal was studied in single strand model trinucleotides in order to understand how it affects the observed dimer yields in UV-irradiated DNA.

Direct excitation of T<>T by UV light induces cleavage with high quantum efficiency, but is of limited biological significance because the high singlet energy of T<>T virtually precludes excitation by the lower-energy UV radiation incident on earth’s surface. In the trinucleotide substrates studied, no evidence showed that the natural purines adenine and guanine, the best electron donors among the natural bases, could repair T<>T by a photolyase-like electron transfer mechanism within experimental uncertainty.

References

(1) Cadet, J.; Vigny, P. In Bioorganic Photochemistry; Morrison, H., Ed.; Wiley: New

York, 1990; Vol. 1, p 1.

(2) Murphy, G. M. British Journal of Dermatology 2009, 161, 90-95.

(3) McCullagh, M.; Hariharan, M.; Lewis, F. D.; Markovitsi, D.; Douki, T.; Schatz, G.

C. Journal of Physical Chemistry B 2010, 114, 5215.

(4) Law, Y. K.; Azadi, J.; Crespo-Hernández, C. E.; Olmon, E.; Kohler, B. Biophysics

Journal 2008, 94, 3590.

(5) Schreier, W. J.; Schrader, T. E.; Koller, F. O.; Gilch, P.; Crespo-Hernández, C. E.;

Swaminathan, V. N.; Carell, T.; Zinth, W.; Kohler, B. Science 2007, 315, 625.

(6) Schreier, W. J.; Kubon, J.; Regner, N.; Haiser, K.; Schrader, T. E.; Zinth, W.;

Clivio, P.; Gilch, P. Journal of the American Chemical Society 2009, 131, 5038.

(7) Hariharan, M.; McCullagh, M.; Schatz, G. C.; Lewis, F. D. Journal of the

American Chemical Society 2010, 132, 12856.

(8) Pan, Z. Z.; McCullagh, M.; Schatz, G. C.; Lewis, F. D. Journal of Physical

Chemistry Letters 2011, 2, 1432.

(9) Fisher, G. J.; Johns, H. E. In Photochemistry and Photobiology of Nucleic Acids;

Wang, S. Y., Ed.; Academic Press: New York, 1976; Vol. 1, p 225.

(10) Johns, H. E.; Rapaport, S. A.; Delbruck, M. Journal of Molecular Biology 1962,

4, 104.

(11) Garces, F.; Davila, C. A. Photochemistry and Photobiology 1982, 35, 9.

(12) Sancar, A. Chemical Reviews 2003, 103, 2203.

(13) Kao, Y.-T.; Saxena, C.; Wang, L.; Sancar, A.; Zhong, D. Cell Biochemistry and

Biophysics 2007, 48, 32.

(14) Carell, T.; Burgdorf, L. T.; Kundu, L. M.; Cichon, M. Current Opinion in

Chemical Biology 2001, 5, 491.

(15) Van Camp, J. R.; Young, T.; Hartman, R. F.; Rose, S. D. Photochemistry and

Photobiology 1987, 45, 365.

(16) Young, T.; Kim, S. T.; Van Camp, J. R.; Hartman, R. F.; Rose, S. D.

Photochemistry and Photobiology 1988, 48, 635.

(17) Kim, S. T.; Hartman, R. F.; Rose, S. D. Photochemistry and Photobiology 1990,

52, 789.

(18) Hartzfeld, D. G.; Rose, S. D. Journal of the American Chemical Society 1993,

115, 850.

(19) Tang, W. J.; Guo, Q. X.; Song, Q. H. Journal of Physical Chemistry B 2009, 113,

7205.

(20) Wu, Q. Q.; Song, Q. H. Journal of Physical Chemistry B 2010, 114, 9827.

(21) Chinnapen, D. J. F.; Sen, D. Proceedings of the National Academy of Sciences of the United States of America 2004, 101, 65.

(22) Holman, M. R.; Ito, T.; Rokita, S. E. Journal of the American Chemical Society

2007, 129, 6.

(23) Nguyen, K. V.; Burrows, C. J. Journal of the American Chemical Society 2011,

133, 14586.

(24) Bourre, F.; Renault, G.; Seawell, P. C.; Sarasin, A. Biochimie 1985, 67, 293.

(25) Kundu, L. M.; Linne, U.; Marahiel, M.; Carell, T. Chemistry - A European

Journal 2004, 10, 5697.

(26) Cannistraro, V. J.; Taylor, J. S. Journal of Molecular Biology 2009, 392, 1145.

(27) Seidel, C. A. M.; Schulz, A.; Sauer, M. H. M. Journal of Physical Chemistry

1996, 100, 5541.

(28) Scannell, M. P.; Prakash, G.; Falvey, D. E. Journal of Physical Chemistry A 1997,

101, 4332.

(29) Takaya, T.; Su, C.; de La Harpe, K.; Crespo-Hernández, C. E.; Kohler, B.

Proceedings of the National Academy of Sciences of the United States of America 2008,

105, 10285.

(30) Fenick, D. J., Carr, H. S., and Falvey, D. E. The Journal of Organic Chemistry

1995, 60, 624.

(31) Middleton, C. T.; de La Harpe, K.; Su, C.; Law, Y. K.; Crespo-Hernández, C. E.;

Kohler, B. Annual Review of Physical Chemistry 2009, 60, 217.

(32) Crespo-Hernández, C. E.; Cohen, B.; Kohler, B. Nature 2005, 436, 1141.

(33) Pan,Z.; Chen, J.; Schreier, W.J.; Lewis, F.D.; Kohler, D. Journal of Physical

Chemistry B 2012, 114,698.

(34) Sztumpf, E.; Shugar, D. Biochimica et Biophysica Acta, Specialized Section on

Nucleic Acids and Related Subjects 1962, 61, 555.

(35) Boggio-Pasqua, M.; Groenhof, G.; Schäfer, L. V.; Grubmüller, H.; Robb, M. A.

Journal of the American Chemical Society 2007, 129, 10996.

(36) Park, H.; Zhang, K.; Ren, Y.; Nadji, S.; Sinha, N.; Taylor, J. S.; Kang, C.

Proceedings of the National Academy of Sciences of the United States of America 2002,

99, 15965.

(37) Ross, S. A First Course in Probability; Macmillan Publishing Co., Inc.: New

York, 1976.

(38) Hare, P. M.; Crespo-Hernández, C. E.; Kohler, B. Proceedings of the National

Academy of Sciences of the United States of America 2007, 104, 435.

(39) Kao, Y. T.; Saxena, C.; Wang, L. J.; Sancar, A.; Zhong, D. P. Proceedings of the

National Academy of Sciences of the United States of America 2005, 102, 16128.

(40) Yeh, S. R.; Falvey, D. E. Journal of the American Chemical Society 1992, 114,

7313.

(41) Chatgilialoglu, C.; Guerra, M.; Kaloudis, P.; Houee-Levin, C.; Marignier, J. L.;

Swaminathan, V. N.; Carell, T. Chemistry - A European Journal 2007, 13, 8979.

(42) Johns, H. E.; Pearson, M. L.; LeBlanc, J. C.; Helleiner, C. W. Journal of

Molecular Biology 1964, 9, 503.

(1) T + T + hν → T<>T dimer formation

(2) T<>T + hν → T<>T* → T + T direct photoreversal

(3) T<>T + S + hν → T<>T + S* → T + T + S indirect photoreversal

Scheme 3.1 Thymine dimer (T<>T) formation and cleavage photoreactions. S is a photosensitizer and electronic excitation is denoted by an asterisk.

O O O O MeMe MeMe HN NH HN NH

O N H H N O 240 nm O N H H N O 280 nm O O O O RO O RO O OR' OR' O P O O P O O O (2+2) cis-syn thymine dimer

TT: R = H, R' = H T<>T: R = H, R' = H ATT: R = dAMP, R' = H AT<>T: R = dAMP, R' = H TTA: R = H, R' = dAMP T<>TA: R = H, R' = dAMP GTT: R = dGMP, R' = H GT<>T: R = dGMP, R' = H TTG: R = H, R' = dGMP T<>TG: R = H, R' = dGMP

Figure 3.1 Structure and nomenclature of the trinucleotides investigated.

1.0 T<>T 0.8 TpT

0.6

0.4 Absorbance 0.2

0.0 220 240 260 280 300 320 Wavelength (nm)

Figure 3.2 Normalized UV/vis spectra of the cis-syn thymine dimer (T<>T) solution and

TpT solution.

2.0 a) 0 min 1.5 7.5 min 15 min 22.5 min 1.0 37.5 min Absorbance 0.5

0.0 220 240 260 280 300 320 Wavelength (nm) 1.4 b) 1.2 0 min 1.0 50 min 65 min 0.8 0.6

Absorbance 0.4 0.2 0.0 220 240 260 280 300 320 Wavelength (nm)

Figure 3.3 (a) UV/vis spectra of the T<>T solution after the indicated irradiation times at

240 nm; (b) T<>T solution reached photo-equilibrium in the spin cell ring excited by the pump pulse.

0.30 0 min 0.25 1 min 2 min 0.20 8 min 0.15 20 min

Absorbance 0.10

0.05

0.00 200 220 240 260 280 300 320 Wavelength (nm)

Figure 3.4 UV/vis spectra of the T<>TA after the indicated irradiation times at 254 nm.

a) 0.7 0.6 0.5 0 min 18 min 0.4 54 min 0.3 72 min Absorbance 0.2 0.1

200 220 240 260 280 300 320 Wavelength (nm)

0.7 b) 0.6 0 min 9 min 0.5 36 min 0.4 72 min 0.3

Absorbance 0.2 0.1

200 220 240 260 280 300 320 Wavelength (nm)

Figure 3.5 UV/vis spectra of AT<>T (a) and T<>TA (b) after the indicated irradiation times during pump-probe experiments with 266 nm pump and 250 nm probe pulses.

a) 0.6

0.5 0 min 0.4 18 min 54 min 0.3 72 min

Absorbance 0.2

0.1

200 220 240 260 280 300 320 Wavelength (nm) b) 0.8 0 min 0.6 9 min 36 min 0.4 63 min Absorbance 0.2

0.0 200 220 240 260 280 300 320 Wavelength (nm)

Figure 3.6 UV/vis spectra of GT<>T (a) and T<>TG (b) after the indicated irradiation times during pump-probe experiments with 266 nm pump and 250 nm probe pulses.

3.0 2.0 1.0 -3 0.2

A / 10 0.0 Δ -0.2

-0.4 0 10 20 100 Time delay (ps)

Figure 3.7 Average transient absorption signal (240 nmpump;266 nm probe) recorded between 0 and 38 minutes of 240 nm excitation (green triangles) and the transient absorption signal recorded between 50 and 65 minutes irradiation time (black circles).

The solid curves are best-fit curves to the experimental points.

0.1

0.0 -3 -0.1 A / 10

Δ -0.2

-0.3

0 10 20 100 Time delay (ps)

Figure 3.8 Transient absorption signal from a reference solution of TMP (240 nm pump;

266nm probe).

0.00

-0.05 -3 -0.10

A / 10 -0.15 Δ -0.20

-0.25 0 10 20 30 40 50

Time delay (ps)

Figure 3.9 Transient absorption signals (266 nm pump; 250 nm probe) from aqueous solutions of AMP (triangles) and AT<>T (circles).

0.0

-3 -0.1

-0.2 A / 10 Δ

-0.3

0 10 20 30 40 50

Time delay (ps)

Figure 3.10 Transient absorption signals (266 nm pump; 250 nm probe) from aqueous solutions of GMP (triangles) and GT<>T (circles).

0.0

-0.5 -3 -1.0

A / 10 -1.5 Δ

-2.0

0 10 20 30 40 50

Time delay (ps)

Figure 3.11 Transient absorption signals (266 nm pump; 250 nm probe) from aqueous solutions of T<>TA.

0.00

-3 -0.05

-0.10

A / 10 -0.15 Δ -0.20 -0.25 0 10 20 30 40 50

Time delay (ps)

Figure 3.12 Transient absorption signals (266 nm pump; 250 nm probe) from aqueous solutions of T<>TG.

1.0

0.8

0.6

0.4 φTT / φT<>T fraction T<>T 10 20 0.2 40

0.0 220 240 260 280 300 λ / nm

Figure 3.13 Dimer level at equilibrium predicted for the indicated values of the quantum

33 yield ratio, φTT /φT<>T . The red filled circles indicate the photostationary levels of dimer

(includes both cis-syn and trans-syn dimers) measured experimentally by Johns et al. for

TpT in 1964 (see Fig. 9 of H. E. Johns et al.Journal of Molecular Biology1964, 9, 503-

24).

Table 3.1 Molar absorption coefficients at wavelengths used in this study.a

Molar absorption coefficient (M-1 cm-1)

Sample 240 nm 266 nm 280 nm

T 2,681 9,472 6,237

TTb 5,362 18,944 12,474

A 5,878 13,101 2,073

G 9,859 10,460 8,129

ATT,c TTAc 11,240 32,045 14,547

GTT,c TTGc 15,221 29,404 20,603

d T<>T 1,840 294 107

c c AT<>T, T<>TA 7,718 (0.24) 13,395 (0.022) 2,180 (0.049) GT<>T,c T<>TGc 11,699 (0.16) 10,754 (0.027) 8,236 (0.013)

a Values for the mononucleotides T, A, and G are from Cavaluzzi, M. J.;

Borer, P. N. Nucleic Acids Research2004, 32, e13/1-9. The fraction of excitations localized on T<>T in the dimer-containing trinucleotides is shown in parentheses in the last two rows. b TT values are twice the T values. c Values are assumed to be additive (TT or T<>T + purine). d Determined from fit to data in H. E. Johns et al., Journal of Molecular Biology1964, 9,

503.

Table 3.2 Cleavage quantum yields, Φ , and conditional cleavage quantum yields, ΦX

and ΦT<>T , for thymine dimers in the listed substrates at irradiation wavelength, λirrad.*

a b Sample λirrad Φ ΦX ΦT<>T

T<>T 240 0.41 ± 0.04 - -

AT<>T 240 0.065 ± 0.004 0.085 0.27

280 0.00806 0.0085 0.16

T<>TA 240 0.074 ± 0.007 0.097 0.31

280 0.00796 0.0084 0.16

GT<>T 240 0.097 ± 0.01 0.12 0.62

280 0.00418 0.0042 0.32

T<>TG 240 0.091 ± 0.009 0.11 0.58

280 0.00393 0.0040 0.30

*This table is Adapted from ref. 33. a ΦX =Φ/ P(X) , where the probability of exciting the purine base X, P(X), is estimated from the molar absorption coefficients in Table 3.1. b ΦT<>T =Φ/ P(T<>T) , where the probability of exciting the dimer T<>T, P(T<>T), is estimated from the molar absorption coefficients in Table 3.1.

Table 3.3 Global Fitting parameter of TMP, T<>T and T<>T at the photostationary state

(PSS).

Sample τ1 (ps) A1(%) τ2 (ps) A2(%) A3(%)

TMP 2.1±0.2 -90 130 -7 -3

T<>T 2.1±0.2 -56 130 -7 36

T<>T(PSS) 2.1±0.2 -87 130 -9 -4

Table 3.4 Best-fit parameters for transient absorption signals in Figure 3.9 and 3.11.

Sample τ1 (ps) A1 (%) A2 (%)

AMP 2.2 ± 0.4 -96 -4

AT<>T 2.2 ± 0.4 -88 -12

GMP 2.6 ± 0.6 -98 -2

GT<>T 2.6 ± 0.5 -86 -14

CHAPTER 4

UV HARDINESS OF ALTERNATE NUCLEOBASES*

4.1 Introduction

Excited states of nucleic acid base monomers present in solution and in the gas phase generally have subpicosecond lifetimes due to ultrafast nonradiative decay.1, 2 These ultrafast lifetimes make the building blocks of life highly resistant to damage caused by

UV radiation. It is still a mystery how is this UV damaged resistant property built into the nucleobases. It is widely recognized that the early Earth was exposed to very high levels of UV radiation due to the absence of an ozone layer 3,4,5 and the radiation characteristics of the young sun. Molecules with this UV damage resistant property would have greater chances to take part in early chemical evolution and could become the precursor of today’s life on Earth. A long time goal of this project is to elucidate the general physical and chemical principles that determine photostablity (UV hardiness) in nucleic acids and to determine the consequences of these properties on the prebiotic Earth.

Rapid excited state deactivation is the mechanism that may have led to the selection of today’s nucleobases over less photostable alternatives on the prebiotic earth.1,6 Notably, various lines of thought suggest that the today’s canonical nucleobases may have been preceded by alternative bases during the chemical evolution that took place before the

RNA world.7-11 Post-transcriptionally modified RNA bases12 are attractive candidates for prebiotic building blocks given their base pairing properties .However, it is hard to identify which one from these various molecules possess sufficient photostability to accumulate to any significant degree under the intense UV radiation incident on the surface of the early earth.13

Recently, hypoxanthine was detected alongside adenine and guanine among the reaction products from UV-irradiated formamide.14 Xanthine and hypoxanthine can be formed from ammonium cyanide in plausible prebiotic syntheses.15 Both compounds have been detected in carbonaceous meteorites, suggesting that they could have been available organic compounds in the early solar system.15 Because of its flexible base pairing possibilities, xanthine has been proposed to be a critical building block of all- purine nucleic acids, which may have preceded RNA.16 For these reasons above hypoxanthine and xanthine (structures showed in Figure 4.1) are chosen to be the first system to study in this project. The parent compound xanthine is poorly soluble in water at neutral pH22 and was not included in this study. Instead, the four methylxanthines shown in Figure 4.2 were studied: theophylline (1,3-dimethylxanthine), paraxanthine

(1,7-dimethylxanthine), theobromine (3,7-dimethylxanthine), and caffeine (1,3,7- trimethylxanthine).

An important motivation of this study is to determine whether these modified purine bases have efficient nonradiative decay pathways that complement their favorable molecular recognition properties, making them well adapted to a prebiotic world rich in

UV radiation. Past studies have shown that chemical substitution can dramatically modify nucleobase photophysics.17-21 A further motivation is thus to understand how nonradiative decay in hypoxanthine differs from that of the closely related purine bases, adenine and guanine.

4.2 Results

4.2.1 Steady-state UV Absorption Measurements

To simulate the early ocean environment four xanthine derivatives are solved in pure water and UV/vis spectra of the aqueous solutions are shown in Figure4.3(a).

Theobromine and the parent compound xanthine are poorly soluble in water and were not studied. The observed long wavelength absorption maxima (λmax) closely match literature values (Table 4.1). The absorption spectrum of caffeine, which lacks an ionizable proton, is independent of pH between 2and 14.23 On the other hand, each of the dimethylxanthine compounds has a pKa of between 8 and 10, and a past report indicates that λmax shifts to longer wavelength as the pH is raised above 7.24 The good agreement between our observed λmax values and the literature values therefore indicates that our unbuffered solutions are composed overwhelmingly of neutral solute molecules.

The λmax of hypoxanthine occurs at a shorter wavelength than other methylxanthines as well as other canonical bases. Guanine (2-aminohypoxanthine) has two distinct long- wavelength absorption bands. The higher energy of the two bands has approximately the same λmax as hypoxanthine. The red shift in λmax for the methylxanthines compared to hypoxanthine has been attributed to the presence of the second oxo group.25 The greatest

67 red shift is seen for methylation at N3. Consistent with these earlier observations, we note a sizable red shift when N3 of paraxanthine is methylated to yield caffeine.

Xanthine is structurally equivalent to uracil with a fused imidazole ring and the red shifts upon methylation parallel observations for uracil. Thus, methyl substitution atN1 of uracil leads to a 1,000 cm-1 red shift, while methylation at N3 does not significantly perturb the spectrum.26 Note that N1 in uracil and other pyrimidines corresponds to N3in purines such as xanthine and vice versa.

UV/vis spectra in acetonitrile are shown in Figure 4.3(b) for the xanthine derivatives which have adequate solubility in this solvent. Compared to water, λmax exhibits a small red shift of ~1 nm. In contrast, blue shifts were observed for uracil and several of its methylated derivatives upon changing the solvent from water to acetonitrile.26

4.2.2 Transient Absorption Measurements on hypoxanthine and four methylxanthine derivatives

Transient absorption signals were recorded of hypoxanthine, paraxanthine, theophylline, and caffeine in aqueous solutions with both visible and UV probe. As observed for the canonical nucleobases, transient absorption signals at visible probe wavelengths are very weak and accompanied by two-photon ionization of the solvent.27

The signals showed in Figure 4.4 are corrected with solvated electron signal as described in Chapter 2.4. The corrected signals of all the samples decay to the baseline within 1 or

2 ps. The time constants in water are all less than 600 fs as shown in Table 4.2, indicating ultrafast depopulation of the initial excited state.

Transient absorption signals from paraxanthine, theobromine, theophylline, and caffeine in acetonitrile solution are shown in Figure 4.5. No solvent correction is needed here because there is negligible two-photon ionization of this solvent. The weak residual signal seen at delay times greater than 8 ps is likely due to a small amount of ionization of the solute by the intense pump pulses. The best-fit lifetimes from monoexponential fits are shown in Table 4.2 and are seen to be about three times longer than for the same compound in aqueous solution.

Transient absorption signals with UV probe are shown in Figure 4.6and Figure 4.7.

At 250 nm probe hypoxanthine shows a negative or bleach signal at 250 nm while the other three compounds exhibit positive signals in aqueous solutions. The monoexponential decay time of 2.3 ps for hypoxanthine is very similar to the time constant seen for adenine at the same wavelength.28 The methylxanthines studied all yield positive ΔA signals at 250 nm. At 290 nm all the compounds show positive ΔA signals again. Both probe wavelengths is close to a minimum in absorbance for these compounds and the positive signal suggests that hot ground-state molecules absorb more strongly than the thermally equilibrated species. At 250 nm the lifetimes in acetonitrile are once again about three times longer than in aqueous solution, consistent with slower vibrational cooling dynamics in this solvent.29 At 290 nm one more time constant is need to describe the slow signal build up near time zero. Similar dynamics were seen in 9- methyladenine in aqueous solution and assigned to hot ground-state absorption by the high-energy absorption band after ultrafast relaxation to the ground state.29

In order to seek the negative bleach recovery signals transient absorption measurements at 280 nm probe were also carried out on paraxanthine, caffeine and theophylline. As shown in Figure 4.8 three compounds first show negative bleach signals with the lifetimes around ~1.3 ps. The signals recover to positive after the first bleach.

The lifetimes of the two positive components are found to be ~2.4 ps and ~0.7 ps. Similar dynamics were seen in TMP at the same probe wavelength.28 The first ~1.3 ps lifetime should be assign to vibrational cooling of the hot 1ππ* state. The second ~2.4 ps time constant matches well with that of hypoxanthine at 250 nm probe and it should be assign to vibrational cooling from the hot ground state.

4.3 Discussion

4.3.1. Ultrafast lifetimes observed in hypoxanthine and xanthine derivatives

Hypoxanthine has a lifetime of 130 ± 20 fs. To the best of our knowledge, this is the shortest lifetime reported on a natural or modified nucleobase with a single-exponential fit. It is shorter than the shortest lifetime component of 180 ± 30fs previously reported for adenine in aqueous solution.18 Kwok et al reported adenosine has a lifetime of 130 fs for the first time component when using two-exponential fit. However, when use a single- exponential fit to the same signal they yielded a lifetime of0.23 ± 0.02 ps.30 Calculation by Barbatti et al. have suggested that the shorter of the two experimental time constants represents relaxation within 1ππ* excited states rather than nonadiabatic passage to the electronic ground state.31

Almost at the same time with this study Villabona-Monsalve et al. reported fluorescence lifetimes measured by the femtosecond upconversion technique for hypoxanthine and its nucleoside and nucleotide at several pH values.32 Their instrument- limited lifetime of < 200 fs for hypoxanthine at pH 5 agrees with the lifetime of 130 ± 20 fs determined in our transient absorption experiment at 570 nm. Rӧttger et al. also reported transient absorption study results on hypoxanthine and its nucleoside inosine and nucleotides inosine monophosphate at pH 7 recently.33 They reported the lifetime of the excited-state of these three compounds to be 0.21 ± 0.08 ps with their broadband detection at 370 – 660 nm following the 260 nm excitation. Ground state bleach recovery signals at 242 nm were also recorded in their experiments and the lifetime was found to be 1.8 ± 0.4 ps. Their results matches nicely with our 130 ± 20 fs excited-state lifetime at

570 nm and 2.0 ± 0.3 ps ground state bleach recovery lifetime at 250 nm.

The complete recovery of the bleach signal from hypoxanthine at 250 nm (Figure4.6) is strong evidence that hypoxanthine, like adenine and guanine, decays to the electronic ground state without significant population trapping in a longer-lived 1nπ* state.34For paraxanthine, caffeine and theophylline the 280 nm probe data also clearly demonstrate that 100% of the excited state population recovers to the electronic ground state on the several picosecond time scale. The chance of a long-lived 1nπ* state existed in the methylxanthine is unlikely because the lack of any slow component in the UV where a

1nπ* state is expected to absorb.34

All of the compounds studied here have ultrashort lifetimes in both aqueous solution and acetonitrile, suggesting that these molecules are quite hardy against deep UV(UVC)

71 radiation. In another word, all the compounds could be the candidates of possible precursor of today’s canonical nucleobases.

4.3.2. Comparison of excited-state decay mechanism between canonical and alternate nucleobases.

Xanthine and hypoxanthine are closely related structurally to adenine and guanine as shown in Figure 4.1. Adenine in aqueous solution consists of 22% 7H-aminotautomer and 78% 9H-amino tautomer.18 Transient absorption of adenine is distinctly biexponential because the 7H tautomer has a nearly 50-fold longer lifetime than the

9Htautomer (8.8 ps vs. 0.18 ps).18Like adenine, hypoxanthine has 7H and 9H tautomers, which co-exist in aqueous solution. 13C NMR experiments indicate that the 9H- oxotautomer of hypoxanthine exists in aqueous solution with an approximately equal population of the 7H-oxo tautomer.35Our transient absorption study on hypoxanthine shows no evidence of biexponential decays, strongly suggesting that the two tautomers should have similar ultrashort lifetimes. Rӧttger et al. also have pointed out that these two tautomers should have similar deactivation dynamics because of the ultrafast monoexponential decay observed in their transient absorption experiments.33 Calculations by

Matsika and coworkers suggest that the same C2 pyramidalization important for radiationless decay in adenine and guanine is also active in hypoxanthine.32

It is worth noting that all N7-substituted compounds studied here have ultrashort lifetimes. This result contrasts with the excited-state behavior of N7-substituted derivatives of adenine and guanine. 7-methyladenine and the 7H-tautomer of adenine

72 have much longer lifetimes than adenine.18 Additional computational study is needed to understand why radiationless decay in these purines is relatively insensitive to N7- substitution.

It is interesting to speculate on the nuclear motions required to reach the deactivating conical intersections for the methylxanthines. These compounds are equivalent to uracil,

1-methyluracil, or 1,3-dimethyluracil with a fused imidazole ring. Like uracil, all of the methylxanthines have just a single π bond in the six-membered ring. Because they lack the C2=N3 double bond of hypoxanthine, adenine, and guanine, decay via C2 pyramidalization is not expected. Torsion about the C5=C6 double bond in uracil is key to its excited statedeactivation,31, 36, 37 but the equivalent motion (corresponding to twisting of the C4=C5double bond in xanthine) may be strongly inhibited by the attached imidazole ring. Calculations by Yamazaki et al.38 suggest that excited xanthine decays via out-of-plane deformation in the five-membered ring (imidazole ring )in contrast to adenine and guanine where puckering in the six-membered pyrimidine ring drives radiationless decay.31 Further calculations are warranted to see whether the key conical intersection in all methylxanthines is due to out-of-plane motion in the imidazole ring and to explain the relative insensitivity of the dynamics to methylation.

Our results show that relaxation occurs about three times more slowly in acetonitrile than in water. Solvent effects on nucleobase lifetimes are difficult to rationalize from experimental data alone. Computational studies have shown that the solvent can subtly change the energy separation between the various electronic states.39 This can affect both vibronic coupling between states as well as change the height of barriers that restrict

73 access to key conical intersections. Interestingly, there is a rough correlation between

λmax and excited state lifetime for the compounds in this study. Thus, the shortest lifetimes are generally observed for compounds with blue-shifted steady-state absorption spectra. This pattern also holds for the substituted uracils studied by Gustavsson et al. by femtosecond fluorescence up-conversion.26

4.4 Conclusions

Five xanthine derivatives were studied by ultrafast transient absorption in two solvents. The results show that hypoxanthine and other four methylxanthines have subpicosecond excited-state lifetimes. The efficient nonradiative deactivation of hypoxanthine observed in our study is consistent with its high photostablility property demonstrated by Guan et al.40All the compounds studied here could be the candidates of possible precursor of today’s canonical nucleobases. Our study will inspire more calculation study on methylxanthines excited-state decay mechanism.

References

(1) C. E. Crespo-Hernández, B. Cohen, P. M. Hare, B. Kohler, Chemical Reviews

2004, 104, 1977.

(2) C. T. Middleton, K. de La Harpe, C. Su, Y. K. Law, C. E. Crespo-Hernández, B.

Kohler, Annual Review of Physical Chemistry 2009, 60, 217.

(3) Sagan, C. Journal of Theoretical Biology 1973, 39, 195.

(4) Cockell, C. S.; Raven, J. A. Philosophical Transactions of the Royal Society of

London. Series A 2007, 365, 1889.

(5) Canuto, V. M.; Levine, J. S.; Augustsson, T. R.; Imhoff, C. L. Precambrian

Research 1983, 20, 109.

(6) A. L. Sobolewski and W. Domcke, Europhysics News 2006, 37, 20.

(7) S. L. Miller, Adv. Chemical Physics 1984, 55, 85.

(8) R. Shapiro, Origins of Life and Evolution of Biospheres 1995, 25, 83.

(9) M. Levy, S. L. Miller, Proceedings of the National Academy of Sciences of the

United States of America 1998 95, 7933.

(10) L. E. Orgel, Critical Reviews in Biochemistry and Molecular Biology 2004, 39,

99.

(11) P. Ehrenfreund, S. Rasmussen, J. Cleaves, L. H. Chen, Astrobiology 2006, 6,490.

(12) P. A. Limbach, P. F. Crain, J. A. McCloskey, Nucleic Acids Research 1994,

22,2183.

(13) C. S. Cockell, G. Horneck, Photochemistry and Photobiology 2001, 73, 447.

(14) H. L. Barks, R. Buckley, G. A. Grieves, E. Di Mauro, N. V. Hud, T. M.

Orlando, ChemBioChem 2010, 11, 1240.

(15) M. P. Callahan, K. E. Smith, H. J. Cleaves, J. Ruzicka, J. C. Stern, D. P. Glavin,

C., H. House, J. P. Dworkin, Proceedings of the National Academy of Sciences of the

United States of America 2011, 108, 13995.

(16) G. Wachtershauser, Proceedings of the National Academy of Sciences of the

United States of America 1988, 85, 1134.

(17) R. J. Malone, A. M. Miller, B. Kohler, Photochemistry and Photobiology 2003,

77, 158.

(18) B. Cohen, P. M. Hare, B. Kohler, Journal of the American Chemical Society,

2003, 125, 13594.

(19) K. A. Kistler, S. Matsika, Journal of Physical Chemistry A 2007, 111, 8708.

(20) E. Mburu, S. Matsika, Journal of Physical Chemistry A 2008, 112, 12485.

(21) S. Matsika,P. Krause, Annual Review of Physical Chemistry 2011, 62, 621.

(22) M. M. Stimson, M. A. Reuter, Journal of the American Chemical Society 1943,

65, 153.

(23) L. F. Cavalieri, J. J. Fox, A. Stone and N. Chang, Journal of the American

Chemical Society 1954, 76,1119.

(24) F. Bergmann, S. Dikstein, Journal of the American Chemical Society 1955, 77,

691.

(25) H. M. Kalckar, Journal of Biological Chemistry 1947, 167, 429.

(26) T. Gustavsson, N. Sarkar, E. Lazzarotto, D. Markovitsi and R. Improta, Chemical

Physics Letters 2006, 429, 551.

(27) J.-M. L. Pecourt, J. Peon, B. Kohler, Journal of the American Chemical Society

2001, 123, 10370.

(28) C. E. Crespo-Hernández, B. Cohen, B. Kohler, Nature 2005, 436, 1141.

(29) C. T. Middleton, B. Cohen, B. Kohler, Journal of Physical Chemistry A 2007,

111, 10460.

(30) W.-M. Kwok, C. Ma, D. L. Phillips, Journal of the American Chemical Society

2006, 128, 11894.

(31) M. Barbatti, A. J. A. Aquino, J. J. Szymczak, D. Nachtigallova, P. Hobza, H.

Lischka, Proceedings of the National Academy of Sciences of the United States of

America 2010, 107, 21453.

(32) J. P. Villabona-Monsalve, R. Noria, S. Matsika, J. Peon, Journal of the American

Chemical Society 2012, in press.

(33) K. Rӧttger, R. Siewertsen, F. Temps, Chemical Physics Letters 2012, 536, 140.

(34) P. M. Hare, C. E. Crespo-Hernández, B. Kohler, Proceedings of the National

Academy of Sciences of the United States of America 2007, 104, 435.

(35) M. M. Karim, C. W. Jeon, H. S. Lee, S. M. Alam, S. H. Lee, J. H. Choi, S. O. Jin,

A. K. Das, Journal of Fluorescence 2006, 16, 713.

(36) S. Matsika, Journal of Physical Chemistry A 2004, 108, 7584.

(37) D. Nachtigallova, A. J. A. Aquino, J. J. Szymczak, M. Barbatti, P. Hobza,

H.Lischka, Journal of Physical Chemistry A 2011, 115, 5247.

(38) S. Yamazaki, A. L. Sobolewski, W. Domcke, Physical Chemistry Chemical

Physics 2009, 11, 10165.

(39) R. Improta, V. Barone, A. Lami, F. Santoro, Journal of Physical Chemistry B,

2009 113,14491.

(40) Y. Y. Guan, N. Fray, P. Coll, F. Macari, D. Chaput, F. Raulin, H. Cottin,

Planetary and Space Science 2010, 58, 1327.

Adenine Guanine

Hypoxanthine Xanthine

Figure 4.1 Structures of adenine, guanine, hypoxanthine and xanthine.

Theophylline Paraxanthine

Theobromine Caffeine

Figure 4.2 Structures of four xanthine derivatives in this study.

(a) hypoxanthine paraxanthine theophylline caffeine Normalized Absorbance 220 240 260 280 300 320 340 Wavelength (nm)

(b) paraxanthine theobromine theophylline caffeine Normalized Absorbance 220 240 260 280 300 320 340

Wavelength (nm)

Figure 4.3 UV/Vis absorption spectra of xanthine derivatives studied in a) aqueous solutions and b) in acetonitrile solutions.

theophylline A

Δ caffeine

paraxanthine Normalized Normalized hypoxanthine

-1 0 1 2 3 4 5

Time Delay (ps)

Figure 4.4 Transient absorption signal recorded at 570 nm of four xanthine derivatives in aqueous solutions. These signals were corrected for solvated electrons using the procedure described in Chapter 2.4.

theophylline A

Δ theobromine

caffeine Normalized paraxanthine

-2 0 2 4 6 8 10

Time Delay (ps)

Figure 4.5 Transient absorption signal recorded at 570 nm of four xanthine derivatives in acetonitrile solutions.

10 (a) hypoxanthine paraxanthine 5 theophylline caffeine A

Δ 1.0

0.5

0.0

Normalized -0.5

-1.0 0 2 4 6 8 10 Time Delay (ps) 10 (b) 5 paraxanthine theobromine

A theophylline

Δ 1.2 caffeine

0.8

Normalized 0.4

0.0 0 5 10 15 20 25

Time Delay (ps)

Figure 4.6 Transient absorption signal recorded at 250 nm of xanthine derivatives studied in (a) aqueous solutions and (b) in acetonitrile solutions.

2.15 hypoxanthine paraxanthine

A 1.2 theophylline Δ caffeine 0.8

Normalized 0.4

0.0

0 5 10 15 20

Time Delay (ps)

Figure 4.7 Transient absorption signal recorded at 290 nm of four xanthine derivatives in aqueous solutions.

-3 3x10 paraxanthine

2 A Δ

0 5 10 15 20 25 Time Delay (ps)

-2

A caffeine Δ -4

-6

-3 -8x10 0 5 10 15 20 25 Time Delay (ps)

1 0 -1 A

Δ -2 theophylline -3 -4

-3 -5x10 0 5 10 15 20 25 Time Delay (ps)

Figure 4.8 Transient absorption signal recorded at 280 nm of paraxanthine, caffeine and theophylline in aqueous solutions.

Table 4.1 Lamda max values of xanthine derivatives in aqueous and acetonitrile solutions.

hypoxanthine paraxanthine theophylline theobromine caffeine

H2O, this 250 nm 269 nm 271 nm - 273 nm

study

23 24 24 H2O, 250 nm 268 nm 272 nm - 273 nm

literature

CH3CN - 270 nm 272 nm 275 nm 274 nm

Table 4.2 Lifetimes of xanthine derivatives in aqueous and acetonitrile solutions.

hypoxanthine paraxanthine theobromine theophylline caffeine

570 nm in 130±20 220±28 - 540±60 480±45

H2O τ(fs)

570 nm in - 0.8±0.1 2.0 ±0.2 1.7 ±0.1 1.3 ±0.1

CH3CN

τ(ps)

250nm in 2.3 ±0.3 1.0 ±0.2 - 1.3 ±0.2 1.7 ±0.2

H2O τ(ps)

250 nm in - 3.5 ±0.6 6.3 ±2 3.5 ±0.4 5.2 ±0.6

CH3CN

τ(ps)

Table 4.3 Lifetime and amplitude of xanthine derivatives in aqueous solutions at 290 nm probe.

τ1(ps) A1 τ2(ps) A2 A3

hypoxanthine 0.4±0.2 -0.1 1.4±0.4 0.4 0.5

paraxanthine 1.3±0.8 -0.3 2.3±0.8 0.4 0.3

theophylline 1.4±0.9 -0.3 2.8±1 0.4 0.3

caffeine 1.9±0.9 -0.3 3.7±1 0.4 0.3

Table 4.4 Lifetime and amplitude of xanthine derivatives in aqueous solutions at 280 nm probe.

τ1(ps) A1 τ2(ps) A2 τ3(ps) A3 paraxanthine 1.3±0.1 -0.5 2.4±0.1 0.3 0.6±0.1 0.2

theophylline 1.4±0.1 -0.6 2.6±0.1 0.1 0.8±0.1 0.3

caffeine 1.3±0.1 -0.5 2.3±0.1 0.2 0.7±0.1 0.3

CHAPTER 5

LONG-LIVED EXCITED STATES IN ADENINE DINUCLEOSIDES WITH

DIFFERENT BACKBONES AND GLYCOL-LINKED DNA DUMBBELL AND

HAIRPIN CONJUGATES

5.1 Introduction

Excited electronic states in DNA, which are generated by absorption of a UV photon, have been studied for decades.1-3 Great attention has been directed to this area during the last decade due to the advantages of ultrafast transient absorption technique.2-3 Excited states of single DNA bases decay back to ground state within several hundred femtoseconds.2-5 Meanwhile, excited states of DNA oligomers and polymers can decay orders of magnitude more slowly. Base stacking is proposed to be the reason that causes these long-lived states in single- and double-stranded DNA in previous studies.6-12

In one earlier study from our group, Takaya et al. proposed a two state model for short-lived and long-lived signals observed in a series of dinucleosides.10 In this model π- stacked bases could lead to electronic coupling producing an exciton state. The initial exciton decays to an exciplex, or a charge-separated state which has a longer lifetime.

When bases are in a poorly stacked conformation only a local excited state is generated upon UV excitation and it decays to a hot ground state with a short lifetime like single bases. However, de La Harpe et al. showed that there is still a large amplitude short-lived decay in a G·C rich genomic oligonucleotide despite the high degree of base stacking present in this double stranded system.13 It is possible that the short-lived component is caused by base pairing in the double stranded system. On the other hand, it is also

90 possible that the short-lived decay channel could still exist in a system which has a highly stacked conformation. The latter explanation is inconsistent with the two state model proposed by Takaya et al. because in a highly stacked conformation only a long-lived state is expected.

In order to better understand the relationship between long-lived excited states and base stacking five dinucleosides with different backbone linkers (structures shown in

Figure 5.1) were studied. dApdA (2’-deoxyadenylyl (3’-5’) 2’-deoxyadenosine) is the smallest base-stacked unit in this study. dAp(abasic)pdA has normal DNA backbone with an abasic (no base) site in the middle. The abasic site is a tetrahydrofuran derivative, in which a methylene group occupies the 1 position of 2’-deoxyribose. We chose to study it because the distance between the two adenine bases in this sample could be longer than that in dApdA. It is also possible that these two bases may not stack at all.48 dApc3pdA is a dinucleoside with a three carbon linker. The three carbon linker is very flexible and we wanted to study whether this backbone could lead to different stacking fraction or different geometries. Ap4A (P1,P4-tetraphosphate di(5’-adenosine) ) and Ap5A (P1,P5- pentaphosphate di(5’-adenosine)) were recently studied with NMR and reported to have base stacking even with the long phosphate backbones.26 It is important to point out that dApdA, dAp(abasic)pdA and dApc3pdA have DNA sugars while the other two compounds have RNA sugars. Transient absorption experiments were carried out on five dinucleosides in both aqueous and methanol-water solutions. The results suggest that the two adenine bases prefer to stack with each other in aqueous solution even with long linkers.

In a further study, glycol-linked DNA hairpin and dumbbell conjugate A·T systems

(structures shown in Figure 5.2) were studied with the same technique. These samples were synthesized by the group of Prof. Frederick D. Lewis at Northwestern University and have been shown to adopt B-DNA base pairing structures.17 AKT-12 has ten A·T base pairs and is capped with hexa(ethylene glycol) linker at both end. AKT-13 also has ten A·T base pairs but is capped only at one end with hexa(ethylene glycol) linker. AKT-

14 and AKT-15 are the single stranded dT10 and dA10, respectively. The ethylene glycol linker could prevent the end-fraying and produce a more rigid base stacking system.

Study on these highly stacked model systems will provide new insight into how base stacking will affect DNA excited-state dynamics. Experimental results from these samples indicate that there is still a fast decay channel in these rigid double stranded systems.

5.2 Results

5.2.1 Steady-state UV-Vis Absorption Measurements

UV-Vis absorption spectra of five dinucleosides, dAMP (2'-deoxyadenosine 5'- monophosphate), and ATP (adenosine 5'-triphosphate) were recorded in phosphate buffer solution (Figure 5.3). The dApdA, dAp(abasic)pdA, and dApc3pdA show a red tail of the

260 nm absorption band compared with dAMP. dApdA has the most intense red tail which may indicate that it has the strongest exciton coupling.14 Stronger absorption by dApdA around 230 nm could be due to an increase in the oscillator strength of the 230 nm 1nπ* transition.15 Ap4A and Ap5A have very similar absorption spectra to that of

ATP in buffer solution.

Figure 5.4 shows the absorption spectra of five dinucleosides in methanol-water solution. No red tail is seen in the 260 nm absorption band for dApdA, dAp(abasic)pdA, and dApc3pdA which indicates that methanol could effectively disrupt exciton coupling in dinucleosides. However, there is a small shoulder that shows up at 320 nm in the dApdA solution. The reason behind this new band is under investigation.

Figure 5.5 shows UV-Vis absorption spectra of glycol-linked DNA hairpin and dumbbell conjugate A·T systems (see structures at Figure 5.2). Because the ethylene glycol linker (Figure 5.2) has no absorption in the wavelength range we studied the UV absorption spectra of single stranded AKT-14 (dT12) and AKT-15 (dA12) match well with our group’s previous results on single strand dT18 and dA18. AKT-14 shows similar

7 absorption spectra as that of dT18 , while the absorption maximum of AKT-15 blue shifts by 2 nm and tails to the red compared with dAMP. After forming a double strand, the

93 dumbbell AKT-12 and hairpin AKT-13 have almost identical UV absorption. The absorption maximum shifts 2 nm to the red and the red tail matches that seen for AKT -

14. The results of the double stranded systems are similar to our previous reported UV

7 absorption spectrum of dA18·dT18 double strand.

5.2.2 Steady-state Circular Dichroism (CD) Measurements

CD spectra of the five dinucleosides, dAMP and ATP in phosphate buffer solution are shown in Figures 5.6. dApdA, dAp(abasic)pdA, and dApc3pdA all show CD maxima near 220 nm and a shoulder at 230 nm as dAMP. dApdA shows a pair of negative- positive conservative band from 240 nm to 290 nm. Instead of showing a conservative band at long wavelengths, dAp(abasic)pdA and dApc3pdA show CD minimum at 265 nm and 255 nm, respectively. Ap4A and Ap5A show minimum CD signal near 220 nm like ATP. These two compounds also show a conservative band around 240 nm to 290 nm. However, the sign of the conservative band is reversed compared with that in dApdA.

CD spectra of all five dinucleosides in methanol-water solution are shown in Figure

5.7. All dinucleosides studied, except for dApdA, show only monomer-like CD signals

(dAMP or ATP) in methanol-water solutions. Although dApdA still shows a weak conservative band around 240 nm to 290 nm, the intensity of the CD signal is at least three times smaller than that of dApdA in buffer solution. These changes in CD spectra suggest that in methanol-water solutions base-base interaction among all five dinucleosides have been effectively disrupted.

5.2.3 Temperature Dependent Circular Dichroism Measurements

Temperature-dependent CD measurements were carried out on all five dinucleosides as well as glycol-linked DNA dumbbell and hairpin conjugate A·T systems and the results are shown in Figures 5.9 to 5.15. To record the reference spectrum of the mononucleotide, an experiment was first carried out on dAMP and the result is shown in

Figure 5.8. The CD spectrum of dAMP shown in Figure 5.8 barely has any temperature dependence in the temperature range from 15°C to 85°C. On the other hand, the CD signals of dApdA, dApc3pdA, Ap4A, and Ap5Asystematically decrease in amplitude as the temperature is increased. dApdA is the only sample which still shows a conservative band at 85°C, consistent with earlier studies on dApdA thermal denaturation.16 The CD minimum band of dApc3dA shifts from 255 nm to 260 nm as temperature increases

(Figure 5.11). The conservative band in Ap4A changes into a single negative band centered around 270 nm at temperatures higher than 65 °C (Figure 5.12). Similar change is also observed in Ap5A when temperatures were higher than 55 °C (Figure 5.13). At

85 °C the CD spectra of dApc3pdA, Ap4A, and Ap5A resemble the monomer CD signal in shape. It is worth noting that unlike the other four dinucleosides, dAp(abasic)pdA shows very similar results to dAMP in the temperature dependent CD experiment as shown in Figure 5.10. It is very difficult to identify the reason for the different thermal properties of these five dinucleosides solely from CD data.

Temperature-dependent CD spectra and UV melting curves recorded for AKT-12 and

AKT-13 are shown in Figures 5.14 and 5.15. The maxima of the CD signal appears near

220 and 280 nm and the minimum near 250 nm matches well with earlier studies17,

95 suggesting that these samples have the well-defined structures reported by McCullagh et al.17 Melting temperature determined from the melting curve of AKT-13 is about 55 °C.

The melting temperature of AKT-12 is much higher (~ 80 °C17) than AKT-13 due to the absence of end fraying in the dumbbell structure. That means adenine and thymine bases in AKT-12 should have larger possibility to keep base pairing and base stacking at 55 °C and we expect to observe different transient absorption signals due to their structural differences.

5.2.4 Transient Absorption (TA) Measurements

Figure 5.16 compares the pump-probe signals recorded at a probe wavelength of 250 nm with 266 nm excitation from dApdA, dAp(abasic)pdA, and dApc3pdA. Figure 5.16 (a) shows back-to-back measurements of AMP and dApdA in buffer. The bleach signal of dAMP recovers with a time constant of 2.2 ± 0.2 ps. At the same time, the signal of dApdA has a 2.6 ± 0.2 ps short-lived decay and a 170 ± 18 ps long-lived decay. These results are consistent with ones from the previous study by Takaya et al.10 Despite very different CD spectra, dApdA, dAp(abasic)pdA, and dApc3pdA show identical bleach recovery signals in transient absorption experiments as shown in Figure 5.16 (b).

Bleach recovery signals from ATP, Ap4A, and Ap5A are shown in Figure 5.17. Same as dAMP, ATP also shows a 2.2 ± 0.1 ps lifetime in bleach recovery signal which suggests that the extra two phosphate groups in ATP do not contribute to the transient absorption signal (see Figure 5.17 a). The bleach signals from Ap4A and Ap5A in buffer are identical. The long-lived states in Ap4A and Ap5A also have a 170 ± 20 ps lifetimes

96 when the data are globally fit, but the amplitude of this component is reduced compared to the other three dinucleosides discussed above.

Pump-probe signals of five dinucleosides in methanol-water solutions are shown in

Figures 5.16 (c) and 5.17 (c). There is no long-lived state in methanol-water solutions for any of the dinucleosides. Bleach recovery signals for dApdA, dAp(abasic)pdA, and dApc3pdA are globally fitted and the lifetime is 3.7 ± 0.2 ps. Global fitting of bleach recovery signals for Ap4A and Ap5A yields a 3.3±0.3 ps lifetime.

Transient absorption results of single- and double-strand conjugate A·T systems are shown in Figure 5.18. Double strand AKT-12 and AKT-13 show identical dynamics at

250 nm at room temperature as listed in Table 5.4. The signals from AKT-12 and AKT-

13 increased slightly in the first several hundred femtoseconds after time zero and a subpicosecond decay component with positive amplitude is needed to fit the data. The second decay component has a lifetime of 4.7 ps which is believed to reflect a fast channel of the excited state to recover to the ground state, although the lifetime is more than twice that of the ground state recovery in dAMP. The third decay component is a relatively long-lived state with a lifetime of 62 ps. However, this lifetime is shorter than the lifetime of the long-lived state observed in dA18 single strand and dA18·dT18 double strand.7 Single strand AKT-14 and AKT-15 data are not fitted but compared with our

7 group’s previous results on dA18 and dT18 in Figure 5.18. The shape of these decay traces matches with each other very nicely as shown in Figure 5.18.

5.2.5 Temperature Dependent Transient Absorption Measurement

Both dApdA and dAp(abasic)pdA in aqueous buffer solution show a systematic decrease in the amplitude of the long-lived component with an increase in temperature from 15 to 75ºC as shown in Figure 5.19. However, the long-lived component is still present even at 75ºC in both dinucleosides. TA data at different temperatures were first fitted individually and the lifetimes of short-lived and long-lived components were very similar. Global fitting of the TA data yields time constants of 2.6 ± 0.1 ps and 170 ± 20 ps for the short-lived and long-lived components, respectively. The amplitudes of the two components are different at different temperatures as listed in Table 5.3.

Transient absorption measurements at 55 °C and 75 °C were carried out on AKT-12 and AKT-13. 55 °C was chosen because this is the melting temperature of AKT-13 and the steady-state UV and CD spectra of AKT-12, shown in Figure 5.14, are little affected at this temperature. As shown in Figure 5.20 AKT-12 has almost identical decay dynamics at room temperature and 55 °C. At the same time, the sub-picosecond decay component in AKT-13 disappears at 55 °C and the long-lived component has a longer lifetime than that at room temperature (Table 5.4). These results show that the vibrational cooling dynamics are different for AKT-13 at room temperature and the melting temperature. At 75 °C, AKT-12 still shows a similar decay trace as at room temperature.

The lifetime of the short-lived component in AKT-13 decreases from 4.7 ps to 3.5 ps at

75 °C while the long-lived component shows the same lifetime as that at 55 °C.

5.3 Discussion

5.3.1 Different base stacking conformations in five dinucleosides could lead to different CD signals but similar TA signal

In our group’s earlier studies, Crespo-Hernández et al. observed a hundred picoseconds decay in single- and double-stranded A-T oligonucleotides and assigned this decay to excimers formed in stacked bases with high yields.7 Kwok et al. later reported a

8 long-lived excimer state in (dA)20. Long-lived states have been reported in dinucleoside monophosphates, which are the smallest base stacked units, and assigned to a charge- transfer state localized on two stacked bases.10 These previous results suggest strongly that base stacking is the origin of the long-lived states.

Despite the very different CD spectra, long-lived states are observed in all five dinucleosides in our transient absorption experiments. Similar bleach recovery signals from dApdA, dAp(abasic)pdA, and dApc3pdA provide strong evidence that base stacking exists in these dinucleosides because of the long-lived states. Now the question that needs an answer is why do dAp(abasic)pdA and dApc3pdA show monomer like CD spectra while base stacking exists in these two compounds?

DNA and RNA monomers and oligomers have been studied intensively with CD spectroscopy since the 1960s. Previous CD studies on dAMP and AMP have shown that both 2’-deoxy-D-ribose (no oxygen atom at position 2 of D(-)-ribose) and D(-)-ribose have no CD signal above 190 nm and CD signal from dAMP and AMP do not involve the phosphate group.19 The CD band at 260 nm has been assigned to a combination of two 1ππ* transitions (centered at 259 nm and 267 nm) and one 1nπ* transition. The 230

99 nm and 220 nm CD bands were assigned to a second 1nπ* transition and the third 1ππ* transition respectively.15,20 As shown in Figure 5.6, these two bands have a positive sign in dAMP which turn into a negative sign in ATP. dApdA, dAp(abasic)pdA, and dApc3pdA have DNA sugars and they all show positive 230 nm and 220 nm bands like dAMP. Meanwhile, Ap4A and Ap5A have RNA sugars and negative bands at 230 nm and 220 nm like ATP. In dApdA, Ap4A, and Ap5A a new pair of conservative bands show up between 240 and 280 nm. The positive-negative pair of conservative band was

1 22-23 assigned to the coupling of B2u ππ* transition of the two stacked bases. The opposite sign of the conservative band in dApdA and Ap4A are due to the nucleobases being in α or β positions (nucleobase up or down relative to C5´ of the sugar ring as illustrated in

Scheme 2 in ref. 19) at C1´ on the sugar ring.19 The lack of a conservative band in dAp(abasic)pdA and dApc3pdA suggests that either there is no coupling between the two adenine bases, or the coupling is cancelled out for some reason. To find the answer we need to consider the differences in base stacking conformation between dApdA and dAp(abasic)pdA or dApc3pdA.

Scott et al. studied optical properties of ApA (adenylyl 3’-5’ adenosine), AppA

(P1,P2-diphosphate di(5’-adenosine)), and Ap4A (P1,P4-tetraphosphate di(5’-adenosine) ) and reported the difference in CD spectra between them.23 In that study, they proposed to designate the two sides of the adenine base as α- and β-faces. They defined α-face to be the one facing the C2’ and C3’ hydrogen atom in sugar when the C8 of adenine is near the C5’ in the sugar (Figure 5.21 a).23 This definition was widely used in later studies.24-26

With their definition they found the stacked conformation in ApA is a α-β stacking

100

(Figure 5.21 b). Due to the longer and more flexible backbone they proposed AppA would prefer β-β stacking and Ap4A could have a α-α stacking geometry (Figure 5.21 c).23 With this model they suggested that (1) there is base stacking present in ApA, AppA, and Ap4A; (2) the transition contributing the most to the CD signal is different between

ApA, AppA, and Ap4A; (3) the different stacking conformations account for the opposite sign of the conservative band in ApA and Ap4A.23

Pettegrew and coworkers also studied ApA and AppA with CD spectroscopy. They suggested that there is only one stacked conformation present in ApA while a multi- conformation equilibrium of stacked conformations could exist in AppA because of the

1 longer and more flexible backbone. They assigned the B2u ππ* coupling to this unique stacked conformation in ApA which is represented by the conservative band in CD spectra.24 Thornton and coworkers have systematically studied the conformational state of

ApA and AppA by semiempirical energy calculations.25 In their study, 20 separate energy minimizations on AppA were performed and 13 stacked structures were identified. They discovered that the stacked structures would obtained lower energies than the unstacked ones which means that the stacked structures are more energetically favorable.

Furthermore they showed that α-α, β-β, and α-β stacking could all exist in AppA while only the α-β parallel stacking is exhibited in ApA. Also, the energy levels of all the stacked structures were very similar no matter which stack conformation they have, indicating that these structures could covert to each other very easily because the energy barrier is very low.

101

In a recent study Stern et al. confirmed the AppA stacking conformation results from

Thornton et al. by NMR and MD simulations.26 They also suggested that AppA, Ap4A and Ap5A have a multi-conformation equilibrium between three stacking modes. Cantor and co-workers studied linear and cyclic deoxythymidine di- and trinucleotides by CD spectroscopy.27 Their CD results suggested that the bases in cyclic dithymidylic acid have a stacked structure. However, the orientation of the bases in this cyclic deoxythymidine dimer is quite different than the linear dTpdT. All these previous studies suggested that

DNA bases have the ability to adopt different stacking structures when their backbones are modified. Therefore, several different base stacking conformations could exist in dAp(abasic)pdA and dApc3pdA in this study because of their longer and more flexible backbones. Next, we will discuss how the different stacking conformations affect the CD signals of the dinucleosides.

The positive-negative conservative CD band of ApA was predicted by Tinoco using exciton coupled model.28 Following Tinoco’s idea theoretical analysis of the CD spectrum of ApA has been done by Van Holde and co-workers.39 For the transitions polarized in the adenine base plane the rotational strengths of the conservative band was calculated by equations below:

πZ 2 πZ 2 R1 = − μ sinα and R2 = + μ sinα 2λ0 2λ0

Here μ is the magnitude of the transition moment and α is the angle between the given transition moment in one base and the corresponding moment in the other base. Z is the vertical distance between base planes and λ0 is the wavelength of the transition in the

102 monomer base. From the equations above it is clear that the orientation (α) between the two transition moments in different stacking conformations could control the intensities and the signs of the CD signals if we assume that the distance Z is always the same. For example, Soctt et al.23 suggested that ApA has α-β stacking at room temperature and the predominate 259 nm transition dipole moments in two adenine bases were related 30° to

45°(these angles was predicted based on previous calculation and X-ray diffraction results)23. This stacking conformation will give rise to a pair of positive-negative conservative band. With longer and more flexible backbones dAp(abasic)pdA and dApc3pdA could have α-β, α-α, and β-β stacking and the angle between the 259 nm transition dipole moments could vary from -180° to 180° in different possible stacking conformations. This means different stacking conformations would have different intensities and signs of the CD signals. Ts’o and co-workers51gave a simple experession for modeling the long-wavelenth exciton-coupled CD of a dinucleoside like ApA:

2 2 μ10a μ20a [θ]CD ∝ 2 sinθ cosθ R12

Here [θ]CD is the molar ellipticity of ApA, μ10a and μ20a are the transition dipole moments in the two adenine bases, R12 is the distance between the two transition dipole moments and θ is the angle between μ10a and μ20a. This expression also suggests that the CD signals of ApA is very sensitive to the angle between the two transition dipoles. CD signals vanish when θ is at 0, 90, and 180° and reach the maximum when θ is near 45°. CD signals can also change sign with different θ values. The CD spectra of compounds with longer backbones are actually a combination of CD signals from all the stacking 103 conformations. There could be a range of stacked conformations that lead to nearly complete cancellation of the conservative band in CD because different stacking conformations have CD signals with opposite signs. This is the reason that dAp(abasic)pdA and dApc3pdA only showed monomer like CD spectra.

Evidence from studies discussed above suggests that our dAp(abasic)pdA, dApc3pdA,

Ap4A and Ap5A could adopt different stacking conformations (adenine bases could be stacked with α-α, β-β, or α-β faces and the transition dipole moments that couple iwht one another could adopt different angles) in buffer solutions due to their longer backbones. Different stacking conformations give rise to CD signals with opposite signs and lead to nearly complete cancellation of the conservative band in CD spectra of dAp(abasic)pdA and dApc3pdA.39,51 Meanwhile, our transient absorption (TA) results of dApdA, dAp(abasic)pdA, dApc3pdA, Ap4A and Ap5A in buffer solution suggest that base stacking is required for observing long-lived states in TA measurements. However, the actual lifetime of the long-lived state is apparently independent of the precise faces of adenine that are overlapped and is independent on different angular seeting between the stacked bases. This finding is consistent with our group’s earlier results.29,52 de La Harpe et al29 showed that the excited-state dynamics of a GC DNA duplex were identical within experimental uncertainty in B- and Z-form structures despite the different stacked base conformations in B- vs Z-DNA. Another study on single-stranded adenine oligomers showed that the DNA form and RNA form oligomers have the same excited-states lifetimes even though the DNA oligomers would have B- form structure and RNA oligomers would have A-form structure in solution.52 Because of the insensitivity of TA 104 signals to different types of base stacking conformation, transient absorption technique becomes a better tool to study how thermal energy denatures the base stacking structures of dinucleosides as will be discussed below.

5.3.2 Dinucleoside structure in methanol-water solutions

Solvent denaturation of DNA has been well studied during the 1960-1970s. Organic solvents could effectively denature the DNA double helix, leading to a decreased CD signals, hypochromicity, and melting temperature. Methanol is one of the most common organic solvents used in previous DNA solvent denaturation experiments.27, 30-33 Cantor et al. studied linear and cyclic deoxythymidine dimers in methanol solutions.27 They found that the conservative band observed for the linear dimer in water solution vanished and showed the same CD spectrum as the monomer. Hall et al. showed that the poly(dG- dC) double helix undergoes a B to Z transition in methanol-water solution and reduced the rotational strength of the CD signal.32 These results indicate that methanol is an effective denaturing solvent of DNA. Therefore, we expect that base stacking structures of dinucleosides to be disrupted in methanol-water solutions. As shown in Figure 5.7 CD spectra from all dinucleosides studied in methanol-water solution resemble the monomer’s CD signal. Our transient absorption data in Figure 5.16 and 5.17 also confirm that there is no long-lived state that exists in methanol-water solutions for all five dinucleosides. That means the base stacking conformation does not exist in methanol- water solutions and the bases in dinucleosides would be hydrated like a free base.

105

To the best of our knowledge we have not seen any theoretical study that reports on dinucleoside conformations in methanol-water solutions. However, MD simulation studies on flavin adenine dinucleotide (FAD) suggest that the distance between the adenine and isoalloxzine rings of FAD is 3.4 Å in aqueous solutions, indicating that they are well stacked even though the backbone linking them together is very long (two phosphate and five carbon). Compared with our dinucleosides the backbone of FAD may be equally flexible. In the methanol-water solution, FAD favors a fully extended structure and there is no base stacking at all.35-37 These results suggest that the dinucleosides we studied here would also prefer extended structures in the methanol-water solution rather than stacked structures. Our transient absorption data confirmed this structural prediction as no long-lived state is observed in any of the five dinucleosides in methanol-water solutions. These results also confirm the earlier assignment of the long-lived state to base-stacking in dinucleosides.10

A longer lifetime of vibrational cooling dynamics in methanol-water solutions observed in our transient absorption experiments is expected. In a previous study,

Middleton et al. suggested that the ability of hydrogen bonds to promote rapid energy transfer from the hot solute to molecules in the first solvent shell is the key to fast vibrational cooling.18 Fewer hydrogen bonding in methanol-water solutions will cause a longer lifetime in vibrational cooling dynamics.

Thermal and solvent denaturation were the major methods used to study of DNA stacking properties.30-34,38-50 Increasing the temperature or the percentage of organic solvent would cause the signal of the conservative band in CD to decrease. As shown in

106

Figure 5.7, we can clearly see that the CD spectra of our dinucleosides in methanol-water solutions are very similar to that of the monomer, except for dApdA. Meanwhile,

Figures 5.8 to 5.13 demonstrate that the CD spectra of all compounds but dApdA in buffer solutions resemble the monomer CD signals at high temperatures. Based on CD spectra alone, it seems that a high concentration of methanol and increasing temperature have the same ability on denaturating our dinucleosides. However, our temperature dependent TA results demonstrate that these two methods actually have very different effects. As discussed above the dinucleosides we studied here would prefer fully extended structures in methanol-water solutions that have no base stacking at all. This is the reason for the total vanishing of the conservative band in CD and the long-lived states in TA. In Figure 5.19 we can see that large amplitudes of the long-lived states are still present in dApdA and dAp(abasic)pdA even at the highest temperature we studied. This means base stacking still exists in dApdA and dAp(abasic)pdA at high temperatures.

Instead of unstacking the bases all the time, thermal energy provides the possibility to either populate different base stacking conformations because the energies of different stacking conformations are very similar or change the angle between the transition dipole moments in the stacked bases .25-26 These changes would cause a vanishing of the conservative band in the CD spectra as we have discussed in 5.3.1. Since the two bases are still stacked, long-lived states will still be seen in transient absorption measurements.

Observations in our study are consistent with previous NMR studies of ApA at 90 ºC by

Davis and co-workers.16 They concluded that ApA in solution at high temperature maintains conformations that hold the two bases close to each other with no solvent in

107 between most of the time. Another NMR study on dinucleosides by Ts’o et al. also suggested that ApA responds to a temperature change by rotation of the unrestricted O-

5’-P-5 bond and the O-3’-P-3 bond leading to the transformation of right-handed stacking to left-handed stacking.53 Although they did not point out the particular stacking structure after the rotation, the chance of the two adenine bases to keep the identical stacking structure as room temperature is very small. Thus, it is possible the CD signals could be cancelled at high temperatures.

5.3.3 Thermodynamic analysis of temperature-dependent experiments on dinucleosides

Brahms et al. proposed a two-state model to explain their CD study of ApA thermal denaturation in aqueous solution.38-39 In this model only stacked and unstacked conformations exist in equilibrium. At low temperatures, the optically active, stacked configuration is favored, while at high temperatures only the unstacked form is proposed to be present. This model seemed to explain the temperature dependent CD data of ApA very well at first. Using this two-state model base stacking properties of dApdA and ApA were intensively studied from the 1960s to the 1990s by a variety of experimental methods.38-50 However, Davis et al. have pointed out that the two-state model is too simple because ApA could have a range of stacked and unstacked conformations, which exist at the same time in solution.16 Watts and co-workers argued that there are more than two states for ApA stacking and it is difficult to use the two-state model to explain their

108

NMR results.44 Recent quantum mechanics calculations and MD simulation studies have suggested that the stacking-unstacking process is a multistate process for ApA.53-56

In order to calculate the equilibrium constant K of the stack-unstack reaction using the two-state model, equation 5.1 was widely used in those earlier studies:

X − X K = 0 T (5.1) X − X T n

Here X0 stands for the temperature dependent property (like rotational strength or absorbance) of the fully stacked state at the low temperature limit and Xn stands for the same property of the completely unstacked state at the high temperature limit. XT is the property measured at different temperatures. After obtaining K at different temperatures, a Van’t Hoff plot was made to calculate ΔH and ΔS.

The rotational strength at 266 nm was used to calculate ΔH and ΔS in this study. The

38,39 high temperature limit Xn was assumed to be zero. The low temperature limit has been estimated by using the fact that X0=2XT when temperature is at the melting point

38,39 provided that Xn=0. By differentiating the melting curves shown in Figure 5.22 (a) with respect to the temperature we can determine the inflection points and those are the melting temperature of the dinucleosides. The thermodynamic parameters calculated from temperature dependent CD data of the five dinucleosides are listed in Table 5.5 and the Van’t Hoff plot is shown in Figure 5.22 (b). The values of ΔH and ΔS for dApdA calculated from the CD data in this study are -6.9 ± 0.2 kcal mol-1 and -23.1 ± 1.3 cal mol-

1 K-1, respectively. Since these values agree with values reported in earlier studies within

39 experimental error the estimation of X0 is considered to be good. Van Holde et al. have 109 pointed out that when they vary X0 to an extreme value (the highest rotational strength that might be considered reasonable) in their study only about 10% difference was observed for the enthalpy and entropy changes. Since we used the same method as they did to calculate the ΔH and ΔS, the uncertainty of our results should also be ± 10%.

To calculate ΔH and ΔS from the transient absorption data, the percentages of the long-lived state amplitude at different temperatures were used because these values won’t change during normalization. Figure 5.23 (a) shows how the long-lived state amplitude of dApdA and dAp(abasic)pdA change as temperature increases. The melting temperatures indicated from this plot are much higher than the values we estimated from CD. ΔH and

ΔS of dApdA and dAp(abasic)pdA calculated by using the melting temperatures determined from TA experiments are listed in Table 5.5 and the Van’t Hoff plot is shown in Figure 5.23 (b). The values of ΔH and ΔS for dApdA and dAp(abasic)pdA agree with each other within experimental uncertainty. The assignment of these thermodynamic parameters will be discussed below.

In previous studies,38-50 ΔH and ΔS were assigned to the unstacking reaction.

However, if unstacking in dApdA is indeed a two-state process then we should obtain the same values for ΔH and ΔS using any experimental method that gives different signals for the stacked and unstacked states. Unfortunately, thermodynamic parameters reported in earlier studies are not self-consistent (summarized in Table 5.6). We can see from

Table 5.6 that CD and ORD (optical rotatory dispersion) measurements seem to give the largest absolute values of ΔH and ΔS while NMR (nuclear magnetic resonance) and calorimetry studies give the smallest. Davis et al. suggested that the two-state model is 110 too simple to describe the actual unstacking process and could cause the difference when calculate ΔH and ΔS are determined by different experimental techniques.16 Kondo and co-workers discovered that the interaction between adenine bases in dinucleoside monophosphates has different response to NMR, CD, and UV absorption methods.51

Their results suggest that different experimental methods should obtain different values of ΔH and ΔS. With the advantage of the TA technique as discussed in 5.3.1, we can make further progress in understanding the resons why different methods yield different thermodynamic parameters.

In this study, the ΔH and ΔS values for dApdA obtained from temperature-dependent

CD measurements are two times higher than those obtained from temperature dependent

TA experiments. On the other hand, dAp(abasic)pdA has almost the same ΔH and ΔS values measured from temperature-dependent CD and TA data (Table 5.5). Although the choice of the low temperature limit could vary from one analysis to another it is hard to reconcile a two-fold variation in ΔH and ΔS. This means that the two techniques we used in this study respond differently to the processes that take place in dinucleosides as temperature is increased. We have argued that the vanishing of the conservative band in the CD spectra is because at higher temperatures dinucleosides could populate different base stacking conformations (different adenine faces or different angles between the transition dipole moments) which have different signs for the CD signal. This suggests that CD signals are sensitive to the conformation changes in the dinucleosides caused by increasing temperature. Therefore, ΔH and ΔS values calculated from the temperature- dependent CD data could be due to the enthalpy and entropy changes needed to convert

111 from one base stacking conformation to another without disrupting base stacking. The energy barrier for changing the angle between two bases that are stacked could differ for different backbones. Therefore, values of ΔH and ΔS calculated from CD data of different dinucleosides could vary a lot as observed experimentally. Because of the need to alter the rotational setting of a larger number of single bonds in adenine dimers with longer and more flexible backbones, we expect that more energy is needed for the conformation change in dAp(abasic)pdA, dApc3pdA, Ap4A, and Ap5A. Indeed, ΔG calculated at 298 K of dinucleosides with longer backbones are at least ten times higher in magnitude than that of dApdA as listed in Table 5.5.

On the other hand, our temperature-dependent transient absorption results as well as some earlier NMR studies16,44,47,53 indicate that the two adenine bases in dApdA (or ApA) remain in the stacked conformation even at very high temperatures. TA results also suggest that TA signal is insensitive to different types of base stacking conformations.

Therefore, transient absorption signals are directly linked to the base stacking fraction in the dinucleosides and it is actually seeing the unstacking reaction of the dinucleosides.

The high melting temperatures of dApdA and dAp(abasic)pdA observed here are consistent with previous studies that concluded base stacking still exists in dinucleosides at high temperatures.16,23,26,53 ΔH and ΔS values calculated from the temperature- dependent TA data agree with those reported in NMR and calorimetry studies.44,47, 60

Watts et al. suggested that CD signals depend more on different stacking geometry while

NMR is more sensitive to the nature of the stacked states.44 That means NMR is also probing the actual unstacking reaction and that is the reason for the ΔH and ΔS values to

112 match with those in our TA study. Since calorimetry directly measures the heat of the unstacking reaction of ApA, it should be independent of the base stacking conformations and ΔH should also be assigned to the actual unstacking reaction. Because ΔH and ΔS values of dApdA and dAp(abasic)pdA calculated from temperature TA data match well with NMR and calorimetry studies, they should be assigned to the unstacking processes of these two dinucleoside. Our results indicate that the enthalpy change of the unstacking process in dApdA should be -3.0 kcal mol-1.Because different techniques could differ in sensitivity of the conformation changes one should not expect agreement of the thermodynamic parameters obtained from different techniques. This could explain thermodynamic parameters reported in earlier studies do not agree with each other.

Without further experiments, it is difficult to answer what other techniques used in the past (UV, ORD etc.) actually monitor. However, we can conclude that these techniques may not monitor the unstacking reaction since the ΔH and ΔS values do not agree with these calculated from our temperature-dependent TA data.

5.3.4 The fast decay channel is still present in well-stacked DNA oligonucleotides

In one earlier study from our group, Takaya et al. proposed a two state model for short-lived and long-lived TA signals observed in a series of dinucleosides.10 If the two- state decay model of DNA excited-states is valid, then there should only be long-lived excited states in a DNA double strands in which every base is fully stacked. However, a

7 fast decay component was observed in the dA18·dT18 double strand and in the genomic oligonucleotide with a disordered base sequence.13 Calculation from Santoro et al.

113 suggest that in A-T double strand systems the excited state favors to localize on T strand and can follow the same decay pathway as that found in the isolated monomer.57

Calculations have also suggested out that the steric hindrance of the double strand can increase the energy barrier of the monomer-like decay pathway and cause a slower ground state recovery.57,58

In the previous studies on dA18·dT18 double strand and in the genomic oligonucleotide,7,13 the double stranded DNA samples were prepared by mixing identical volumes of equal-molar solutions of DNA single strands. NMR study has shown that the terminal AT base pair has higher proton exchange rate with solvent than that of the terminal GC base pair in a double stranded DNA system.59 Therefore, the terminal AT base pairs in dA18·dT18 are more likely to be open and act like monomers. Thus, an unstacked and unpaired single base in these oligonucleotides could contribute to the fast decay component. The AKT-12 dumbbell used in this study has a lower possibility of end fraying because of the ethylene glycol linkers at both ends.17 This property makes it a better system for studying the double stranded DNA excited state decay mechanism. On the other hand, the AKT-13 hairpin only has one ethylene glycol linker and can be melted as normal double stranded DNA. A study on AKT-13 will reveal how the excited state dynamics change when the double strand structure of DNA is disturbed.

The 4.7 ps decay component in AKT-12 or AKT-13 at room temperature has the largest amplitude reflecting the bleach recovery of the ground state in the first 10 ps. This lifetime is longer than that of the single stranded DNA samples we studied before.7,13

Middleton et al. suggested that vibrational cooling is sensitive to hydrogen bonding.18

114

Therefore, in the rigid base pairing system we expect to observed slower vibrational cooling dynamics because of the lower hydration possibility. In the temperature dependent experiment AKT-12 retains the 4.7 ps lifetime even at 75 ºC. However, the lifetime changes from 4.7 ps to 3.5 ps in AKT-13 when the temperature reached 75 ºC.

Since the melting temperature of AKT-13 is ~55 ºC, 80% of this sample should have been melted into a single strand at 75 ºC. Because of more hydrogen bonding with water, the vibrational cooling dynamics are accelerated and a short lifetime is observed. Another possible explanation of the 4.7 ps lifetimes is base pairing actually slows down the out- of-plane deformations that lead to ultrafast nonradiative decay. When the double stranded system is melted at high temperature, there is no limit of the out-of-plane motion and the lifetime is shorter.

It is worth pointing out that the signals from AKT-12 actually increased slightly in the first several hundred femtoseconds after time zero and a 0.8 ps decay component with positive amplitude is needed to fit the data. Similar results were shown in 9- methyladenine18 and the genomic oligonucleotide13 at the same probe wavelength. This decay component was assigned to hot ground-state absorption by a higher energy absorption band. This means the ground state is repopulated in less than 1 ps. Our results here are consistent with the calculations done by Santoro et al.57 and suggest that the fast monomer-like excited state decay pathway does exist in double stranded DNA systems.

115

5.4 Conclusions

Five dinucleosides with different backbones were studied by time-resolved and steady-state electronic spectroscopy in order to better understand how secondary structure affects the excited-state dynamics in DNA. Long-lived states have been observed in all the dinucleosides in aqueous solutions suggesting that the two adenine bases in the dinucleosides studied here have a stacked conformation. The vanishing of the conservative band in the CD spectra of some dinucleosides is caused by the fast multi- conformation conversion between different base stacking conformations. Transient absorption results of all the dinucleosides in methanol-water solutions confirm that the origin of the long-lived states is base stacking that is presented in all the dinucleosides.

Organic solvent is proved to have different denaturation effects on the dinucleosides compared with the thermal energy. Temperature-dependent transient absorption studies reveal that base stacking is still presented in all the dinucleosides even at high temperatures. More meaningful enthalpy and entropy changes assigned to the unstacking reaction were reported because we believe temperature dependent transient absorption can monitor the actual unstacking process.

Both short-lived and long-lived states were observed in DNA hairpin and dumbbell conjugate A·T systems. Because of the rigid structure of our samples, we conclude that a fast monomer-like decay channel of DNA excited states is present in double stranded AT systems. Temperature dependent transient absorption data reveals that the longer vibrational cooling lifetime of the short-lived state in AT systems is due to the lower

116 possibility of hydrogen bonding between DNA bases and the solvent in the double stranded structures.

References

(1) P.R. Callis, Annual Review of Physical Chemistry 1983, 34, 329.

(2) C.E. Crespo-Hernández, B. Cohen, P.M. Hare, B. Kohler, Chemical

Reviews 2004, 104, 1977.

(3) Middleton, C. T.; de La Harpe, K.; Su, C.; Law, Y. K.; Crespo-Hernández,

C. E.; Kohler, B. Annual Review of Physical Chemistry 2009, 60, 217.

(4) Pecourt, J.-M. L.; Peon, J.; Kohler, B. Journal of the American Chemical Society

2000,122, 9348.

(5) Pecourt, J.-M. L.; Peon, J.; Kohler, B. Journal of the American Chemical Society

2001,123, 10370.

(6) Crespo-Hernández, C. E.; Kohler, B. Journal of Physical Chemistry B 2004, 108,

11182..

(7) Crespo-Hernández, C. E.; Cohen, B.; Kohler, B. Nature 2005, 436, 1141.

(8) Kwok, W.-M.; Ma, C.; Phillips, D. L. Journal of the American Chemical Society

2006, 128, 11894.

(9) Buchvarov, I.; Wang, Q.; Raytchev, M.; Trifonov, A.; Fiebig, T. Proceedings of the National Academy of Sciences of the United States of America 2007, 104, 4794.

117

(10) Takaya, T.; Su, C; de La Harpe, K.; Crespo-Herna´ndez, C. E.; Kohler, B.

Proceedings of the National Academy of Sciences of the United States of America 2008,

105, 10285.

(11) Crespo-Hernández, C. E.; de La Harpe, K.; Kohler, B. Journal of the American

Chemical Society 2008,130,10844.

(12) de La Harpe, K; Crespo-Hernández, C. E.; Kohler, B. Journal of the American

Chemical Society 2009,131,17557.

(13) de La Harpe, K; Kohler, B. Journal of Physical Chemistry Lett., 2011, 2 , 133.

(14) Kononov A.I.; Bakulev V.M.; Rapoport V.L. Journal of Photochemistry and

Photobiology B: Biology 1993,19,139.

(15) Kelly, G.R.; Kurucsev, T. Biopolymers 1974, 13, 769.

(16) Davis, R.; Tinoco, I. Biopolymers 1968, 6, 223.

(17) McCullagh, M.; Zhang, L.; Karaba, A.; Zhu, H.; Schatz, G.; Lewis, F.D. Journal of Physical Chemistry B 2008, 112,11415.

(18) Middleton, C. T.; Cohen, B.; Kohler, B. Journal of Physical Chemistry A 2007,

111,10460.

(19) Nielsen, S.B.;Chakraborty, T.;Hoffmann,S.V. ChemPhysChem 2005,6, 2619.

(20) Sprecher C.A.; Johnson W.C. Biopolymers1 977, 16, 2243.

(21) Holm et al. Journal of Photochemistry and Photobiology A: Chemistry 2007, 187,

293.

(22) Hug, W.;Tinoco, I., Jr. Journal of the American Chemical Society1973, 95,2803.

118

(23) Scott, J.F.; Zamecnik, P.C. Proceedings of the National Academy of Sciences of the United States of America 1969, 64, 1308.

(24) Pettegrew, J.W.; Miles, D.W.; Eyring H. Proceedings of the National Academy of

Sciences of the United States of America 1977, 74, 1785.

(25) Thornton, J.M.; Bayley, P.M. Biopolymers 1976, 15, 955.

(26) Stern, N.; Major, D.T.; Gottlieb, H.E.; Weizman, D.; Fischer, B. Organic &

Biomolecular Chemistry 2010, 8, 4637.

(27) Cantor, C.; Fairclough, R.; Newmark R.A. Biochemistry 1969, 8, 3610.

(28) Tinoco, I. Journal of the American Chemical Society 1964, 86, 297.

(29) de La Harpe, K; Crespo-Hernández, C. E.; Kohler, B. ChemPhysChem 2009, 10,

1421.

(30) Herskovits, T.; Singer S.J. Archives of Biochemistry and Biophysics 1961, 94, 99.

(31) Brahms, J.; Mommaerts, W.F.H.M. Journal of Molecular Biology 1964, 10, 73.

(32) Sinanoglu, O.; Abdulnur S. Photochemistry and Photobiology 1964, 3, 333.

(33) Levine, L.; Gordon, J.; Jencks W.P. Biochemistry. 1963, 2, 168.

(34) Hall, K.B.; Maestre, M.F. Biopolymers 1984, 23, 2127.

(35) van den Berg, P.A.W. et al. Journal of Physical Chemistry B 2002, 106, 8858.

(36) Radoszkowicz, L.; Huppert, D.; Nachliel, E.; Gutman, M. Journal of Physical

Chemistry A 2010, 114,1017.

(37) Radoszkowicz, L.; Presiado, I.; Erez, Y.; Nachliel, E.;Huppert, D.; Gutman, M.

Physical Chemistry Chemical Physics 2011, 13, 12058.

119

(38) Brahms, J.; Michelson, A.M.; Van Holde, K.E. Journal of Molecular Biology

1966, 15,467.

(39) Van Holde, K.E.;Brahms, J.; Michelson, A.M. Journal of Molecular Biology

1965, 12,726.

(40) Olsthoorn, C. S. M.; Bostelarr, L. J.; Vanboom, J. H.; Altona, C. European

Journal of Biochemistry 1980, 112, 95.

(41) Olsthoorn, C. S. M.; Bostelarr, L. J.; Derooij, J. F. M.; Vanboom, J. H.; Altona, C.

European Journal of Biochemistry 1981, 115, 309.

(42) Olsthoorn, C. S. M.; Haasnoot, C.; Altona, C. European Journal of Biochemistry

1980, 106, 85.

(43) Johnson, N.P; Schleich, T. Biochemistry 1974, 13,981.

(44) Watts, M. T.; Tinoco, I. Biochemistry 1978,17,2455.

(45) Powell, J. T.; Richards, E. G.; Gratzer, W.B. Biopolymers 1972, 11, 235.

(46) Fornasiero, D.; Kurucsev, T. European Journal of Biochemistry 1984, 143,1.

(47) Lee, C. H.; Tinoco, I. Biochemistry 1977, 16, 5403.

(48) Lowe, M. J.; Schellma, J. Journal of Molcular Biology 1972, 65, 91.

(49) Ogasawara, N.; Inoue, Y. Journal of the American Chemical Society 1976, 98,

7048.

(50) Kang, H.; Chou, P.; Johnson, W.C.; Weller, D.; Huang, S.; Summerton, J.

Biopolymers 1992,32, 1351.

(51) Kondo, N.; Holmes, H.; Stempel, L.; Ts’o P. Biochemistry 1970, 9, 3479.

(52) Su, C. Dissertation, The Ohio State University, 2011.

120

(53) Ts’o, P.; Kondo, N.; Schweizer, M.; Hollis, D. Biochemistry 1969, 8, 997.

(54) Norberg, J.; Nilsson, L. Journal of Physical Chemistry 1995, 99, 13056.

(55) Florián, J.; Šponer, J.; Warshel, A.; Journal of Physical Chemistry B 1999, 103,

884.

(56) Jafilan, S.; Klein, L.; Hyun, C.; Florián, J. Journal of Physical Chemistry B 2012,

116, 3613.

(57) Santoro, F.; Barone, V.; Improta, R. Journal of the American Chemical Society

2009,131,15232.

(58) Conti, I.; Alto_e, P.; Stenta, M.; Garavelli, M.; Orlandi, G. Physical Chemistry

Chemical Physics 2010, 12, 5016.

(59) Nonin, S.; Leroy, J.; Guéron, M. Biochemistry 1995, 34, 10652.

(60) Breslauer, K.; Charko, G.; Hrabar, D.; Oken, C.; Biophysical Chemistry 1978, 8,

393.

121

NH2

N N

N N NH2

HO O N N H H

H H N O H N

O P O O

O- H H H H OH H

(a)

NH2

N N

HO O H H

H H O H NH2

O P O O N N O- H H

H H N O H N

O P O O

O- H H

H H OH H (b)

Figure 5.1 Structures of the dinucleosides investigated (dinucleotide a-c are in DNA form and d-e are in RNA form): (a) dApdA (2’-deoxyadenylyl (3’-5’) 2’- deoxyadenosine), (b) dAp(abasic)pdA, (c) dApc3pdA, (d) Ap4A (P1,P4-tetraphosphate di(5’-adenosine) ) and (e) Ap5A Ap5A (P1,P5-pentaphosphate di(5’-adenosine)).

Continued

122

Figure 5.1 Continued

NH2

N N

HO O H H

H H O H

O P OH

NH2

N N

N O N

O P O O

O- H H H H OH H

NH2 (c) N N

N O- N

O P O O H H O H H OH OH O P O-

NH2

O N N O P O-

N O N

O P O O

O- H H H H OH H

(d)

Continued 123

Figure 5.1 Continued

NH2

N N

N O- N

O P O O H H O H H OH OH O P O-

O P O-

O NH2

N O P O- N

N O N

O P O O

O- H H H H OH H

(e)

124

Figure 5.2 Structures of DNA hairpin and dumbbell conjugate A·T systems (synthesis

detail of these compounds can be found in reference 17)

125

a) 2.0 dApdA dAp(abasic)pdA 1.5 dApc3pdA dAMP 1.0

0.5 Normalized Absorbance Normalized

0.0 200 220 240 260 280 300 320 340 Wavelength (nm) b)

1.6 Ap4A 1.4 Ap5A ATP 1.2 1.0 0.8 0.6 0.4 Normalized Absorbance 0.2 0.0 200 220 240 260 280 300 320 340

Wavelength (nm)

Figure 5.3 UV-Vis absorption spectra of (a) dAMP, dApdA, dAp(abasic)pdA and dApc3pdA in buffer solution (pH=6.8); (b) ATP, Ap4A and Ap5A in buffer solution

(pH=6.8).

126

dApdA 1.5 dAp(abasic)pdA dApc3pdA

1.0

0.5 Normalized Absorbance Normalized

0.0 200 220 240 260 280 300 320 340 Wavelength (nm)

1.6 b) 1.4 Ap4A Ap5A 1.2 1.0 0.8 0.6 0.4 Normalized Absorbance Normalized 0.2 0.0 200 220 240 260 280 300 320 340 Wavelength (nm)

Figure 5.4 UV-Vis absorption spectra of (a) dApdA, dAp(abasic)pdA and dApc3pdA in methanol-water solution; (b) Ap4A and Ap5A in methanol-water solution.

127

1.0 AKT-12 AKT-13 AKT-14 0.8 AKT-15 0.6

0.4 Absorbance 0.2

0.0 200 220 240 260 280 300 320 340 Wavelength (nm)

Figure 5.5 UV-Vis absorption spectra of DNA hairpin and dumbbell conjugate AT systems ((pH=7.05).

128

1.0 a) -7

) 0.5 -1

0.0 ·dmol 2

-0.5 dApdA dAp(abasic)pdA dApc3pdA

(mdeg·cm -1.0 dAMP Residue ellipticity x 10 x ellipticity Residue -1.5 200 220 240 260 280 300 Wavelength (nm)

1.0 b) Ap4A

-7 Ap5A

) 0.5 ATP -1 0.0 ·dmol 2 -0.5

-1.0 (mdeg·cm Residue ellipticity x 10 -1.5 200 220 240 260 280 300 320

Wavelength (nm)

Figure 5.6 Circular dichroism spectra of (a) dAMP, dApdA, dAp(abasic)pdA and dApc3pdA in buffer solution (pH=6.8); (b) ATP, Ap4A and Ap5A in buffer solution

(pH=6.8).

129

1.0 a) -7 dApdA )

-1 0.5 dAp(abasic)pdA dApc3pdA

·dmol 0.0 2

-0.5

-1.0 (mdeg·cm Residue ellipticity x 10 -1.5 dAMP in methanol-water

200 220 240 260 280 300 Wavelength (nm)

-7 1.0 b)

) Ap4A

-1 Ap5A 0.5 ·dmol

2 0.0

-0.5

-1.0 (mdeg·cm

Residue ellipticity x 10 -1.5 ATP in methanol-water

200 220 240 260 280 300 320

Wavelength (nm)

Figure 5.7 Circular dichroism spectra of (a) dAMP, dApdA, dAp(abasic)pdA and dApc3pdA in methanol-water solution; (b) ATP, Ap4A and Ap5A in methanol-water solution.

130

1.0

-7 0.5 ) -1 0.0

·dmol 15 °C 2 -0.5 25 °C 35 °C 45 °C -1.0 55 °C

(mdeg·cm 65 °C

Residue ellipticity x 10 -1.5 75 °C 85 °C

220 240 260 280 300

Wavelength (nm)

Figure 5.8 Temperature dependent circular dichroism spectra of dAMP (pH=6.8).

131

-7 1.0 ) -1 0.5

·dmol 0.0 2 15°C -0.5 25°C 35°C 45°C -1.0 55°C 65°C

(mdeg·cm 74°C -1.5 84°C Residue ellipticity x 10 ellipticity Residue

200 220 240 260 280 300

Wavelength (nm)

Figure 5.9 Temperature dependent circular dichroism spectra of dApdA (pH=6.8).

Inserted lower panel shows dAMP circular dichroism spectra in buffer solution (pH=6.8).

132

1.0 -7 )

-1 0.5

0.0

·dmol 15°C 2 -0.5 25°C 35°C 45°C -1.0 55°C 65°C -1.5 75°C (mdeg·cm 85°C Residue ellipticity x 10

220 240 260 280 300

Wavelength (nm)

Figure 5.10 Temperature dependent circular dichroism spectra of dAp(abasic)pdA

(pH=6.8). Inserted lower panel shows dAMP circular dichroism spectra in buffer solution

(pH=6.8).

133

1.0 -7

) 0.5 -1

0.0

·dmol 15 °C 2 -0.5 25 °C 35 °C 45 °C -1.0 55 °C 65 °C

(mdeg·cm -1.5 75 °C

Residue ellipticity x 10 ellipticity Residue 85 °C

220 240 260 280 300

Wavelength (nm)

Figure 5.11 Temperature dependent circular dichroism spectra of dApc3pdA (pH=6.8).

Inserted lower panel shows dAMP circular dichroism spectra in buffer solution (pH=6.8).

134

1.0 -7 )

-1 0.5

0.0 ·dmol 2 15 °C -0.5 25 °C 35 °C 45 °C -1.0 55 °C

(mdeg·cm 65 °C -1.5 75 °C Residue ellipticity x 10 x ellipticity Residue 85 °C

200 220 240 260 280 300 320

Wavelength (nm)

Figure 5.12 Temperature dependent circular dichroism spectra of Ap4A (pH=6.8).

Inserted lower panel shows ATP circular dichroism spectra in buffer solution (pH=6.8).

135

1.0 -7

) 0.5 -1

0.0 ·dmol

2 15 °C -0.5 25 °C 35 °C -1.0 45 °C 55 °C 65 °C (mdeg·cm -1.5 75 °C

Residue ellipticity x 10 Residue 85 °C

200 220 240 260 280 300 320

Wavelength (nm)

Figure 5.13 Temperature dependent circular dichroism spectra of Ap5A (pH=6.8).

Inserted lower panel shows ATP circular dichroism spectra in buffer solution (pH=6.8).

136

a) 20

0 20 °C 30 °C -10 40 °C Ellipticity (mdeg) 50 °C 60 °C -20 70 °C 80 °C

200 220 240 260 280 300 320 Wavelength (nm)

b) 0.64

0.62 AKT-12 at 250 nm 0.60

0.58

0.56 Absorbance

0.54

0.52

20 30 40 50 60 70 80

Temperature (°C)

Figure 5.14 (a) Temperature dependent circular dichroism spectra (pH=7.05) and (b) UV melting curve of dumbbell AKT-12.

137

a) 20

0 20 °C 30 °C

Ellipticity (mdeg) -10 40 °C 50 °C 60 °C 70 °C -20 80 °C

200 220 240 260 280 300 320 Wavelength (nm)

b) 0.70 AKT-13 at 250 nm 0.68 0.66 0.64 0.62

Absorbance 0.60 0.58 0.56 20 30 40 50 60 70 80

Temperature (°C)

Figure 5.15 (a) Temperature dependent circular dichroism spectra (pH=7.05) and (b) UV melting curve of dumbbell AKT-13.

138

a) A Δ

dAMP dApdA Normalized

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

A b) Δ

dApdA dAp(abasic)pdA dApc3pdA Normalized

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

A c) Δ

dApdA dAp(abasic)pdA dApc3pdA Normalized

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

Figure 5.16 Transient absorption signals at 250 nm following excitation at 266nm of

(a)AMP (black square) and dApdA (red square) in buffer solution (pH=6.8); (b) dApdA(red square), dAp(abasic)pdA(green circle) and dApc3pdA (black triangle) in buffer solution (pH=6.8); (c) dApdA(red square), dAp(abasic)pdA(green circle) and dApc3pdA (black triangle)in methanol-water solution. Solid curves are from least- squares fits. 139

a A Δ

dAMP ATP Normalized Normalized

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

A b Δ

Ap4A Ap5A

Normalized ATP

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

c A Δ

Ap4A Ap5A Normalized Normalized

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

Figure 5.17 Transient absorption signals from (a) dAMP (red circle) and ATP (black circle); (b) ATP (black circle), Ap4A (red square) and Ap5A (green triangle) in buffer solution (pH=6.8); (c) Ap4A(red square)and Ap5A (green triangle) in methanol-water solution.

140

a) 0

-1

AKT-12 A

Δ -2 AKT-13

-3

-3 -4x10 2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps) b) 0.0

-0.5

-1.0 AKT-14

A dT18

Δ -1.5

-2.0

-3 -2.5x10

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps) c) 0 -2 -4 -6 A

Δ AKT-15 -8 dA18 -10 -12 -3 -14x10 2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

Figure 5.18 Transient absorption signals from DNA hairpin and dumbbell conjugate A·T systems in buffer solution (pH=7.05).

141

a) A Δ 15°C 35°C 55°C 75°C Normalized Normalized

2 4 6 8 2 4 6 8 0 4 8 10 100 1000 Time Delay (ps)

b) A Δ

15°C 35°C

Normalized 55°C 71°C

2 4 6 8 2 4 6 8 0 4 8 10 100 1000

Time Delay (ps)

Figure 5.19 Temperature dependent transient absorption signals from (a)dApdA and (b) dAp(abasic)pdA in buffer solution (pH=6.8) at different temperature indicated.

142

0.0 -0.5 -1.0 AKT-12 at 75 °C -1.5 AKT-12 at 55 °C A

Δ AKT-12 at 22 °C -2.0 -2.5 -3 -3.0x10

2 4 6 8 2 4 6 8 0 5 10 15 20 100 1000 Time Delay (ps)

0.0 -0.5 -1.0 AKT-13 at 75 °C A -1.5 Δ AKT-13 at 55 °C -2.0 AKT-13 at 22 °C -2.5

-3 -3.0x10 2 4 6 8 2 4 6 8 0 5 10 15 20 100 1000 Time Delay (ps)

Figure 5.20 Temperature dependent transient absorption signals from (a) AKT-12 and (b)

AKT-13 in buffer solution (pH=7.05) at different temperatures indicated.

143

(a)

(b) (c)

Figure 5.21 Different base stacking conformation suggested by previous studies.23-26 (a)

α-face of adenine base facing reader; (b) α-β stacking; (c) α-α or β-β stacking.

144

a) 20 dApdA Tm= 25 °C dAp(abasic)pdA T = 45 °C 15 m dApc3pdA Tm= 35 °C 10 Ap4A Tm= 45 °C Ap5A Tm= 45 °C 5 0 -5 -10 Rotational strengthRotational (mdeg)

20 30 40 50 60 70 80 Temperature (°C) b)

2.0 dApdA 1.5 dAp(abasic)pdA dApc3pdA 1.0 Ap4A Ap5A 0.5

ln K 0.0 -0.5 -1.0 -1.5 -3 2.8 2.9 3.0 3.1 3.2 3.3 3.4x10 1/T

Figure 5.22 (a) Rotational strength change vs temperature; (b) Van’t Hoff plot of five dinucleosides calculated from temperature dependent TA data.

145

0.20 a) dApdA Tm= 75 °C dA(abasic)dA Tm= 60 °C 0.15

0.10

0.05 statelong-lived amplitude 0.00 20 40 60 80 temperature (°C) b) 0.2 dApdA dA(abasic)dA 0.0 -0.2 -0.4 ln K -0.6 -0.8 -1.0

-3 2.8 3.0 3.2 3.4x10 1/T

Figure 5.23 (a) long-lived state amplitude change vs temperature; (b) Van’t Hoff plot of dApdA and dAp(abasic)pdA calculated from temperature dependent TA data.

146

Table 5.1 Global fitting parameters of five dinucleosides in buffer solutions

τ1(ps) A1 (%) τ2 (ps) A2 (%) A3 (%)

dApdA in buffer 2.6 ± 0.2 81 170 ± 20 13 3

dAp(abasic)pdA in 2.6 ± 0.2 84 170 ± 20 14 2

buffer dApc3pdA in buffer 2.6 ± 0.2 83 170 ± 20 14 3

Ap4A in buffer 2.5 ± 0.2 90 170 ± 20 8 2

Ap5A in buffer 2.5 ± 0.2 90 170 ± 20 8 2

147

Table 5.2 Global fitting parameters of five dinucleosides in methanol-water solutions

τ1(ps) A1 (%) A2 (%)

dApdA in methanol 3.7 ± 0.2 99 1

dAp(abasic)pdA in 3.7 ± 0.2 99 1

methanol

dApc3pdA in methanol 3.7 ± 0.2 99 1

Ap4A in methanol 3.3 ± 0.3 99 1

Ap5A in methanol 3.3 ± 0.3 99 1

148

Table 5.3 Global fitting parameters of temperature dependent TA data of dinucleosides

τ1(ps) A1 (%) τ2 (ps) A2 (%) A3 (%)

dApdA 15°C 2.6 ± 0.1 81.6 170 ± 20 13.6 2.2

dApdA 35°C 2.6 ± 0.1 82.2 170 ± 20 13.3 1.6

dApdA 55°C 2.6 ± 0.1 83.4 170 ± 20 11.1 1.6

dApdA 75°C 2.6 ± 0.1 86.3 170 ± 20 9.4 1.4 dAp(abasic)pdA 15°C 2.6 ± 0.1 82.4 170 ± 20 13.3 1.8 dAp(abasic)pdA 35°C 2.6 ± 0.1 82.5 170 ± 20 10.4 2.7 dAp(abasic)pdA 55°C 2.6 ± 0.1 82.7 170 ± 20 9.2 2.8 dAp(abasic)pdA 71°C 2.6 ± 0.1 84.1 170 ± 20 7.7 2.8

149

Table 5.4 Global fitting parameters of temperature dependent TA data of hairpin and

dumbbell conjugate A·T systems

τ1(ps) A1 (%) τ2 (ps) A2 (%) τ3 (ps) A3 (%) A4 (%)

AKT-12 22°C 0.8 ± 0.3 24 4.7 ± 0.5 -58 62 ± 17 -14 -4

AKT-12 55°C 0.8 ± 0.3 20 4.7 ± 0.5 -62 62 ± 17 -14 -4

AKT-12 75°C 0.8 ± 0.3 20 4.7 ± 0.5 -63 62 ± 17 -14 -3

AKT-13 22°C 0.8 ± 0.3 19 4.7 ± 0.5 -63 62 ± 17 -15 -3

AKT-13 55°C - - 4.6 ± 0.5 -72 108 ± 25 -24 -4

AKT-13 75°C - - 3.5 ± 0.4 -75 108 ± 25 -22 -3

150

Table 5.5 Thermodynamics parameters for the five dinucleosides studied (transition from unstacked state to stacked state).

Sample Method ΔH(kcal mol-1) ΔS(cal mol-1 K-1) ΔG(kcal mol-1) c

dApdA CD a -6.9 ± 0.2 -23.1 ± 1.3 -0.02

dApdA TA b -3.1 ± 0.4 -8.8 ± 1.5 -0.5

dAp(abasic)pdA CD a -2.6 ± 0.2 -8.2 ± 0.7 -0.2

dAp(abasic)pdA TA b -3.4 ± 0.5 -10.2 ± 1.6 -0.4

dApc3pdA CD a -3.7 ± 0.5 -11.8 ± 1.6 -0.2

Ap4A CD a -7.2 ± 1.2 -22.4 ± 3.2 -0.5

Ap5A CD a -7.0 ± 0.7 -21.8 ± 2.8 -0.5 a. Melting temperature indicated in Figure 5.22 b. Melting temperature indicated in Figure 5.23 c. at 298 K

151

Table 5.6 Thermodynamics parameters of dApdA and ApA in earlier studies*

Sample Method a ΔH(kcal mol-1) ΔS(cal mol-1 K-1) Ref.

dApdA CD -6.7 -20.7 40

dApdA CD -7.3 ± 0.4 -22.7 ±1.4 41

dApdA CD -7.4 ± 0.2 -19.9 ± 0.5 46

dApdA CD -4.8 ± 0.1 -14.9 ± 0.1 50

dApdA UV-Vis -6.7 ± 0.7 -19.3 ± 0.7 46

ApA CD -8.5 ± 0.3 -29.9 46

ApA CD -6.5 -23 44

ApA CD -8.0 ± 1.6 -28 ± 5.6 38

ApA CD -5.5 ± 0.2 -19.1 ± 0.7 46

ApA CD -7.0 ± 0.4 -23.9 ± 1.3 42

ApA CD -9.0 ± 0.5 -30 ± 1.4 45

Continued

152

Table 5.5 Continued

ApA CD -6.2 ± 0.1 -22.0 ± 0.1 50

ApA NMR -3 ± 2 -13 ± 5 44

ApA NMR -3.5 -12.5 47

ApA ORD -5.3 -19 16

ApA ORD -6.6 -23 ± 4.6 48

ApA ORD -9.8 ± 0.8 -33 ± 2.4 45

ApA UV-Vis -4.8 ± 0.5 -17.9 ± 0.7 46

ApA UV-Vis -8.5 ± 0.2 -28.3 ± 0.6 45

ApA UV-Vis -5.5 ± 0.3 -18 ± 0.9 49

ApA UV-Vis -7 -20 44

ApA CAL 2.7 ± 0.2 60

* Part of the data shown in this table has been summarized in ref. 52. a. CD: circular dichroism; UV-Vis: absorption spectroscopy; NMR: nuclear magnetic resonance; ORD: optical rotator dispersion; CAL: calorimetry.

153

BIBLIOGRAPHY

(1) American Cancer Society. Skin Cancer. Cancer.org. 2010. 30 Jan 2011.

(2) Crespo-Hernández, C. E.; Cohen, B.; Hare, P. M.; Kohler, B. Chemical Reviews

2004, 104, 1977.

(3) Middleton, C. T.; de La Harpe, K.; Su, C.; Law, Y. K.; Crespo-Hernández,

C. E.; Kohler, B. Annual Review of Physical Chemistry 2009, 60, 217.

(4) Pecourt, J.-M. L.; Peon, J.; Kohler, B. Journal of the American Chemical Society

2000, 122, 9348.

(5) Pecourt, J.-M. L.; Peon, J.; Kohler, B. Journal of the American Chemical Society

2001, 123,10370.

(6) Onidas, D.; Markovitsi, D.; Marguet, S.; Sharonov, A.; Gustavsson, T. J.

Phys. Chem. B 2002, 106, 11367.

(7) Gustavsson, T.; Sharonov, A.; Onidas, D.; Markovitsi, D. Chemical Physics

Lett. 2002, 356, 49.

(8) Gustavsson, T.; Sharonov, A.; Markovitsi, D. Chemical Physics Letters 2002,

351, 195.

(9) Sharonov, A.; Gustavsson, T.; Carré, V.; Renault, E.; Markovitsi, D.

Chemical Physics Letters 2003, 380, 173.

(10) Matsika, S.Journal of Physical Chemistry A 2004, 108, 7584.

154

(11) Serrano-Andrés, L.; Merchan, M. Journal of Photochemistry and Photobiology C.

2009, 10,21.

(12) M. Barbatti, A. J. A. Aquino, J. J. Szymczak, D. Nachtigallova, P. Hobza and H.

Lischka, Proceedings of the National Academy of Sciences of the United States of

America, 2010, 107, 21453.

(13) Hare, P. M.; Crespo-Hernández, C. E.; Kohler, B., Journal of Physical Chemistry

B 2006, 110,18641.

(14) Hare, P. M.; Crespo-Hernández, C. E.; Kohler, B., Proceedings of the National

Academy of Sciences of the United States of America 2007, 104, 435.

(15) Crespo-Hernández, C. E.; Kohler, B. Journal of Physical Chemistry B 2004, 108,

11182.

(16) Crespo-Hernández, C. E.; Cohen, B.; Kohler, B. Nature 2005, 436, 1141

(17) Kuimova, M. K.; Dyer, J.; George, M. W.; Grills, D. C.; Kelly, J. M.;Matousek, P.;

Parker, A. W.; Sun, X. Z.; Towrie, M.; Whelan, A. M. Chemical Communication 2005,

1182.

(18) Buchvarov, I.; Wang, Q.; Raytchev, M.; Trifonov, A.; Fiebig, T. Proc Proceedings of the National Academy of Sciences of the United States of America 2007, 104, 4794.

(19) Schwalb, N. K.; Temps, F. Science 2008, 322, 243.

(20) Cohen, B.; Larson, M. H.; Kohler, B. Chemical Physics 2008, 350, 165.

(21) Kwok, W. M.; Ma, C. S.; Phillips, D. L. Journal of Physical Chemistry B 2009,

113,11527.

155

(22) de La Harpe, K.; Crespo-Hernández, C. E.; Kohler, B. Journal of the American

Chemical Society 2009, 131, 17557.

(23) Váya, I.; Changenet-Barret, P.; Gustavsson, T.; Zikich, D.; Kotlyar, A. B.;

Markovitsi, D. Photochemical & Photobiological Sciences 2010, 9, 1193.

(24) Cadet, J.; Sage, E.; Douki, T., Mutation Research-Fundamental and Molecular mechanisms of Mutagenesis 2005, 571, 3.

(25) Schreier, W. J.; Schrader, T. E.; Koller, F. O.; Gilch, P.; Crespo-Hernández, C. E.;

Swaminathan, V. N.; Carell, T.; Zinth, W.; Kohler, B. Science 2007, 315, 625.

(26) Law, Y. K.; Azadi, J.; Crespo-Hernández, C. E.; Olmon, E.; Kohler, B. Biophysics

Journal 2008, 94, 3590.

(27) Schreier, W. J.; Kubon, J.; Regner, N.; Haiser, K.; Schrader, T. E.; Zinth, W.;

Clivio, P.; Gilch, P. Journal of the American Chemical Society 2009, 131, 5038.

(28) Chinnapen, D. J. F.; Sen, D. Proceedings of the National Academy of Sciences of the United States of America 2004, 101, 65.

(29) Holman, M. R.; Ito, T.; Rokita, S. E. Journal of the American Chemical Society

2007, 129, 6.

(30) Nguyen, K. V.; Burrows, C. J. Journal of the American Chemical Society 2011,

133, 14586.

(31) A. L. Sobolewski and W. Domcke, Europhysics News, 2006, 37, 20.

(32) Margulis, L.; Walker, J. C. G.; Rambler, M. Nature 1976, 264, 620.

(33) S. L. Miller, Advances in Chemical Physics 1984, 55, 85.

(34) R. Shapiro, Origins of Life and Evolution of Biospheres 1995, 25, 83.

156

(35) M. Levy, S. L. Miller, Proceedings of the National Academy of Sciences of the

United States of America 1998, 95, 7933.

(36) L. E. Orgel, Critical Reviews in Biochemistry and Molecular Biology 2004, 39, 99.

(37) P. Ehrenfreund, S. Rasmussen, J. Cleaves, L. H. Chen, Astrobiology 2006, 6,490.

(38) Pan, Z. Z.; McCullagh, M.; Schatz, G. C.; Lewis, F. D. Journal of Physical

Chemistry Letters 2011,2, 1432.

(39) McCullagh, M.; Zhang, L.; Andrew, K.; Zhu, H.; Schatz, G. C.; Lewis, F. D.J.

Phys.Chem. B 2008, 112, 11415.

(40) Middleton, C.T. Dissertation, The Ohio State University, 2008.

(41) de La Harpe, K.D. Dissertation, The Ohio State University, 2011.

(42) Cadet, J.; Vigny, P. In Bioorganic Photochemistry; Morrison, H., Ed.; Wiley: New

York, 1990; Vol. 1, p 1.

(43) Murphy, G. M. British Journal of Dermatology 2009, 161, 90-95.

(44) McCullagh, M.; Hariharan, M.; Lewis, F. D.; Markovitsi, D.; Douki, T.; Schatz, G.

C. Journal of Physical Chemistry B 2010, 114, 5215.

(46) Hariharan, M.; McCullagh, M.; Schatz, G. C.; Lewis, F. D. Journal of the

American Chemical Society 2010, 132, 12856.

(47) Fisher, G. J.; Johns, H. E. In Photochemistry and Photobiology of Nucleic Acids;

Wang, S. Y., Ed.; Academic Press: New York, 1976; Vol. 1, p 225.

(48) Johns, H. E.; Rapaport, S. A.; Delbruck, M. Journal of Molecular Biology 1962, 4,

104.

(49) Garces, F.; Davila, C. A. Photochemistry and Photobiology 1982, 35, 9.

157

(50) Sancar, A. Chemical Reviews 2003, 103, 2203.

(51) Kao, Y.-T.; Saxena, C.; Wang, L.; Sancar, A.; Zhong, D. Cell Biochemistry and

Biophysics 2007, 48, 32.

(52) Carell, T.; Burgdorf, L. T.; Kundu, L. M.; Cichon, M. Current Opinion in

Chemical Biology 2001, 5, 491.

(53) Van Camp, J. R.; Young, T.; Hartman, R. F.; Rose, S. D. Photochemistry and

Photobiology 1987, 45, 365.

(54) Young, T.; Kim, S. T.; Van Camp, J. R.; Hartman, R. F.; Rose, S. D.

Photochemistry and Photobiology 1988, 48, 635.

(55) Kim, S. T.; Hartman, R. F.; Rose, S. D. Photochemistry and Photobiology 1990,

52, 789.

(56) Hartzfeld, D. G.; Rose, S. D. Journal of the American Chemical Society 1993, 115,

850.

(57) Tang, W. J.; Guo, Q. X.; Song, Q. H. Journal of Physical Chemistry B 2009, 113,

7205.

(58) Wu, Q. Q.; Song, Q. H. Journal of Physical Chemistry B 2010, 114, 9827.

(59) Bourre, F.; Renault, G.; Seawell, P. C.; Sarasin, A. Biochimie 1985, 67, 293.

(60) Kundu, L. M.; Linne, U.; Marahiel, M.; Carell, T. Chemistry - A European Journa l2004, 10, 5697.

(61) Cannistraro, V. J.; Taylor, J. S. Journal of Molecular Biology 2009, 392, 1145.

(62) Seidel, C. A. M.; Schulz, A.; Sauer, M. H. M. Journal of Physical Chemistry 1996,

100, 5541.

158

(63) Scannell, M. P.; Prakash, G.; Falvey, D. E. Journal of Physical Chemistry A 1997,

101, 4332.

(64) Takaya, T.; Su, C.; de La Harpe, K.; Crespo-Hernández, C. E.; Kohler, B.

Proceedings of the National Academy of Sciences of the United States of America 2008,

105, 10285.

(65) Fenick, D. J., Carr, H. S., and Falvey, D. E. Journal of Organic Chemistry 1995,

60, 624.

(66) Pan,Z.; Chen, J.; Schreier, W.J.; Lewis, F.D.; Kohler, D. Journal of Physical

Chemistry B 2012, 114,698.

(67) Sztumpf, E.; Shugar, D. Biochimica et Biophysica Acta, Specialized Section on

Nucleic Acids and Related Subjects 1962, 61, 555.

(68) Boggio-Pasqua, M.; Groenhof, G.; Schäfer, L. V.; Grubmüller, H.; Robb, M. A.

Journal of the American Chemical Society 2007, 129, 10996.

(69) Kao, Y. T.; Saxena, C.; Wang, L. J.; Sancar, A.; Zhong, D. P. Proceedings of the

National Academy of Sciences of the United States of America 2005, 102, 16128.

(70) Yeh, S. R.; Falvey, D. E. Journal of the American Chemical Society 1992, 114,

7313.

(71) Chatgilialoglu, C.; Guerra, M.; Kaloudis, P.; Houee-Levin, C.; Marignier, J. L.;

Swaminathan, V. N.; Carell, T. Chemistry - A European Journal 2007, 13, 8979.

(72) Johns, H. E.; Pearson, M. L.; LeBlanc, J. C.; Helleiner, C. W. Journal of

Molecular Biology 1964, 9, 503.

(73) Sagan, C. Journal of Theoretical Biology 1973, 39, 195.

159

(74) Cockell, C. S.; Raven, J. A. Philosophical Transactions of the Royal Society of

London. Series A 2007, 365, 1889.

(75) Canuto, V. M.; Levine, J. S.; Augustsson, T. R.; Imhoff, C. L. Precambrian

Research 1983, 20, 109.

(76) Park, H.; Zhang, K.; Ren, Y.; Nadji, S.; Sinha, N.; Taylor, J. S.; Kang, C.

Proceedings of the National Academy of Sciences of the United States of America 2002,

99, 15965.

(77) Ross, S. A First Course in Probability; Macmillan Publishing Co., Inc.: New York,

1976.

(78) P. A. Limbach, P. F. Crain, J. A. McCloskey, Nucleic Acids Research 1994,

22,2183.

(79) C. S. Cockell, G. Horneck, Photochemistry and Photobiology 2001, 73, 447.

(80) H. L. Barks, R. Buckley, G. A. Grieves, E. Di Mauro, N. V. Hud, T. M.

Orlando, ChemBioChem 2010, 11, 1240.

(81) M. P. Callahan, K. E. Smith, H. J. Cleaves, J. Ruzicka, J. C. Stern, D. P. Glavin, C.,

H. House, J. P. Dworkin, Proceedings of the National Academy of Sciences of the United

States of America 2011, 108, 13995.

(82) G. Wachtershauser, Proceedings of the National Academy of Sciences of the

United States of America 1988, 85, 1134.

(83) R. J. Malone, A. M. Miller, B. Kohler, Photochemistry and Photobiology 2003, 77,

158.

160

(84) B. Cohen, P. M. Hare, B. Kohler, Journal of the American Chemical Society 2003,

125, 13594.

(85) K. A. Kistler, S. Matsika, Journal of Physical Chemistry A 2007, 111, 8708.

(86) E. Mburu, S. Matsika, Journal of Physical Chemistry A 2008, 112, 12485.

(87) S. Matsika,P. Krause, Annual Review of Physical Chemistry 2011, 62, 621.

(88) M. M. Stimson, M. A. Reuter, Journal of the American Chemical Society 1943, 65,

153.

(89) L. F. Cavalieri, J. J. Fox, A. Stone and N. Chang, Journal of the American

Chemical Society 1954, 76,1119.

(90) F. Bergmann, S. Dikstein, Journal of the American Chemical Society 1955, 77,

691.

(91) H. M. Kalckar, Journal of Biological Chemistry 1947, 167, 429.

(92) T. Gustavsson, N. Sarkar, E. Lazzarotto, D. Markovitsi and R. Improta, Chemical

Physics Letter 2006, 429, 551.

(93) J. P. Villabona-Monsalve, R. Noria, S. Matsika, J. Peon, Journal of the American

Chemical Society 2012, in press.

(94) K. Rӧttger, R. Siewertsen, F. Temps, Chemical Physics Letters 2012, 536, 140.

(95) M. M. Karim, C. W. Jeon, H. S. Lee, S. M. Alam, S. H. Lee, J. H. Choi, S. O. Jin,

A. K. Das, Journal of Fluorescence 2006, 16, 713.

(96) D. Nachtigallova, A. J. A. Aquino, J. J. Szymczak, M. Barbatti, P. Hobza,

H.Lischka, Journal of Physical Chemistry A 2011, 115, 5247.

161

(97) S. Yamazaki, A. L. Sobolewski, W. Domcke, Physical Chemistry Chemical

Physics 2009, 11, 10165.

(98) R. Improta, V. Barone, A. Lami, F. Santoro, Journal of Physical Chemistry B 2009,

113,14491.

(99) Y. Y. Guan, N. Fray, P. Coll, F. Macari, D. Chaput, F. Raulin, H. Cottin, Planetary and Space Science 2010, 58, 1327.

(100) P.R. Callis, Annual Review of Physical Chemistry 1983, 34, 329.

(101) Crespo-Hernández, C. E.; de La Harpe, K.; Kohler, B. Journal of the American

Chemical Society 2008,130,10844.

(102) de La Harpe, K; Kohler, B. Journal of Physical Chemistry Letters 2011, 2 , 133.

(103) Kononov A.I.; Bakulev V.M.; Rapoport V.L. Journal of Photochemistry and

Photobiology B: Biology 1993,19,139.

(104) Kelly, G.R.; Kurucsev, T. Biopolymers 1974, 13, 769.

(105) Davis, R.; Tinoco, I. Biopolymers 1968, 6, 223.

(106) McCullagh, M.; Zhang, L.; Karaba, A.; Zhu, H.; Schatz, G.; Lewis, F.D. Journal of Physical Chemistry B 2008, 112,11415.

(107) Nielsen, S.B.;Chakraborty, T.;Hoffmann,S.V. ChemPhysChem 2005,6, 2619.

(108) Sprecher C.A.; Johnson W.C. Biopolymers 1977, 16, 2243.

(109) Holm et al. Journal of Photochemistry and Photobiology A: Chemistry 2007, 187,

293.

(110) Hug, W.;Tinoco, I., Jr. Journal of the American Chemical Society 1973, 95,2803.

162

(111) Scott, J.F.; Zamecnik, P.C. Proceedings of the National Academy of Sciences of the United States of America 1969, 64, 1308.

(112) Pettegrew, J.W.; Miles, D.W.; Eyring H. Proceedings of the National Academy of

Sciences of the United States of America 1977, 74, 1785.

(113) Thornton, J.M.; Bayley, P.M. Biopolymers 1976, 15, 955.

(114) Stern, N.; Major, D.T.; Gottlieb, H.E.; Weizman, D.; Fischer, B. Organic &

Biomolecular Chemistry 2010, 8, 4637.

(115) Cantor, C.; Fairclough, R.; Newmark R.A. Biochemistry 1969, 8, 3610.

(116) Herskovits, T.; Singer S.J. Archives of Biochemistry and Biophysics 1961, 94, 99.

(117) Brahms, J.; Mommaerts, W.F.H.M. Journal of Molecular Biology 1964, 10, 73.

(118) Sinanoglu, O.; Abdulnur S. Photochemistry and Photobiology 1964, 3, 333.

(119) Levine, L.; Gordon, J.; Jencks W.P. Biochemistry 1963, 2, 168.

(120) Hall, K.B.; Maestre, M.F. Biopolymers 1984, 23, 2127.

(121) van den Berg, P.A.W. et al. Journal of Physical Chemistry B 2002, 106, 8858.

(122) Radoszkowicz, L.; Huppert, D.; Nachliel, E.; Gutman, M. Journal of Physical

Chemistry A 2010, 114, 1017.

(123) Radoszkowicz, L.; Presiado, I.; Erez, Y.; Nachliel, E.;Huppert, D.; Gutman, M.

Physical Chemistry Chemical Physics 2011, 13, 12058.

(124) Brahms, J.; Michelson, A.M.; Van Holde, K.E. Journal of Molecular Biology

1966, 15,467.

(125) Van Holde, K.E.;Brahms, J.; Michelson, A.M. Journal of Molecular Biology

1965, 12,726.

163

(126) Olsthoorn, C. S. M.; Bostelarr, L. J.; Vanboom, J. H.; Altona, C. European

Journal of Biochemistry 1980, 112, 95.

(127) Olsthoorn, C. S. M.; Bostelarr, L. J.; Derooij, J. F. M.; Vanboom, J. H.; Altona, C.

European Journal of Biochemistry 1981, 115, 309.

(128) Olsthoorn, C. S. M.; Haasnoot, C.; Altona, C. European Journal of Biochemistry

1980, 106, 85.

(129) Johnson, N.P; Schleich, T. Biochemistry 1974, 13,981.

(130) Watts, M. T.; Tinoco, I. Biochemistry 1978,17,2455.

(131) Powell, J. T.; Richards, E. G.; Gratzer, W.B. Biopolymers 1972, 11, 235.

(132) Fornasiero, D.; Kurucsev, T. European Journal of Biochemistry 1984, 143,1.

(133) Lee, C. H.; Tinoco, I. Biochemistry 1977, 16, 5403.

(134) Lowe, M. J.; Schellma, J. Journal of Molecular Biology 1972, 65, 91.

(135) Ogasawara, N.; Inoue, Y. Journal of the American Chemical Society 1976, 98,

7048.

(136) Kang, H.; Chou, P.; Johnson, W.C.; Weller, D.; Huang, S.; Summerton, J.

Biopolymers 1992,32, 1351.

(137) Kondo, N.; Holmes, H.; Stempel, L.; Ts’o P. Biochemistry 1970, 9, 3479.

(138) Su, C. Dissertation, The Ohio State University, 2011.

(139) Ts’o, P.; Kondo, N.; Schweizer, M.; Hollis, D. Biochemistry 1969, 8, 997.

(140) Norberg, J.; Nilsson, L. Journal of Physical Chemistry 1995, 99, 13056.

(141) Florián, J.; Šponer, J.; Warshel, A.; Journal of Physical Chemistry B 1999, 103,

884.

164

(142) Jafilan, S.; Klein, L.; Hyun, C.; Florián, J. Journal of Physical Chemistry B 2012,

116, 3613.

(143) Santoro, F.; Barone, V.; Improta, R. Journal of the American Chemical Society

2009,131,15232.

(144) Conti, I.; Alto_e, P.; Stenta, M.; Garavelli, M.; Orlandi, G. Physical Chemistry

Chemical Physics 2010, 12, 5016.

(145) Nonin, S.; Leroy, J.; Guéron, M. Biochemistry 1995, 34, 10652.

(146) Breslauer, K.; Charko, G.; Hrabar, D.; Oken, C.; Biophysical Chemistry 1978, 8,

393.

(147) de La Harpe, K; Crespo-Hernández, C. E.; Kohler, B. ChemPhysChem 2009, 10,

1421.

165