INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

H ie quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

A Bell & Howell information Company 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 313/761-4700 800/521-0600 Synthesis, Evaluation and Mechanistic Studies of Halogenated Psoralen and Acridine Photosensitizers

A Dissertation

Presented in Partial Fufillment of Requirements for

the Degree of Doctor of Philosophy

in the Graduate School of the Ohio State University

By

Tongqian Chen

* * * i|i 4c

The Ohio State University

1995

Dissertation Committee: Approved by

Dr. Matthew S. Platz

Dr. James A. Cowan

Dr. Gideon A. Fraenkel UMI Number: 9526010

UMI Microform 9526010 Copyright 1995, by DMI Company. All rights reserved.

This microform edition is protected against unauthorized copying under Title 17, United States Code.

UMI 300 North Zeeb Road Ann Arbor, MI 48103 To my beloved parents, Yicheng and Caimei, whose life-long suffering has paved my way.

To my beautiful wife, Lingxia,

who has suffered bravely and selflessly.

and

To my lovely dauther, Shannon, who reminds me that life is not suffering only. ACKNOWLEDGMENTS

I wish to express my sincere gratitude to Dr. Matthew Platz for being an excellent academic advisor as well as a life mentor.

Many thanks are extended to the past and present members of Platz’s research group who are too numerous to mention individually for their coorperation, help and friendship. Of particular help, Dr. Sang-Chul Park provided me with certain sensitizers of his own creation wich were used in this project. I also thank him for his great frendship and encouragement. I am grateful to Dr. Chanderica Kasturi who imparted to me many biotechniques. I appreciate Dr. Jacek Michalak who cooperated with me in the studies of azido psoralens and taught me the laser flash photolysis technique. I also enjoy Dr. John Toscano’s help in using laser flash photolysis instruments and being a great consultant.

I also acknowledge Cryopharm Corporation for financial support.

Above all, I am deeply in debt to my wife, Lingxia, who gave up her comfortable life at home and came with me abroad. During my whole graduate study here, her constant care and support was invaluble. Without her, everything would be impossible. VITA

Sept. 19, 1964 ...... Bom - Yuhuang, Zhejiang, China.

1986 ...... B. S.

Ocean University of Qingdao

Qingdao, Shandong, China

1986-1989 ...... Graduate Research Associate

Shanghai Institute of Organic Chemistry,

Chinese Academy of Sciences

Shanghai, China

1989 ...... M.S.

Shanghai Institute of Organic Chemistry,

Chinese Academy of Sciences

Shanghai, China

1990-1992 ...... Graduate Teaching Associate

The Ohio State University

Columbus, OH

1993-1994 ...... Graduate Research Associate

The Ohio State University

Columbus, OH

iv PUBLICATIONS

Biqi Wu, Tongqian Chen. Guojian Ou, and Lixin Dai. “A New Method to Synthesis of

Acetylenic Amides”, Youji Huaxue(Chinese J. Org. Chem.) 1989,9,431.

Matthew S. Platz, Tongqian Chen, et al. “The Design and Development of Selective

Viral Inactivation Sensitizers For Sterilization of Blood Products” , In “Prevention of

Transfusion Mediated Diseases” . Fritz Sieber Ed. CRC Press, in press.

Tongqian Chen. Jacek Michalak, and Matthew S. Platz. “Exploratory Photochemistry of 5-Azido*8-Alkoxy Substituted Psoralens Free and Bound toDNA”. Submitted to

Photochemistry & Photobiology.

Tongqian Chen. Eric Voelk, and Matthew S. Platz “Dramatic Improvements in Viral

Inactivation with a Halogenated Intercalator 2. 3-Amino-6-iodoacridine”. Submitted to

Photochemistry & Photobiology.

FIELD O F STUDY

Major Field: Chemistry

v TABLE OF CONTENTS

DEDICATION...... ii

ACKNOWLEDGEMENTS...... iii

VITA...... iv

LIST OF FIGURES...... ix

LIST OF SCHEMES...... xi

LIST OF TABLES...... xiii

LIST OF PLATES...... xiv

LIST OF SPECTRA...... xv

ABBREVIATIONS...... xix

CHAPTER I. INTRODUCTION...... 1

1.1 Problem Definition ...... 1

1.2 Drug Design Strategies ...... 2

1.3 Scope and Objective ...... 4

CHAPTER II. LITERATURE OVERVIEW...... 6

II. 1 Drug-DNA Interaction ...... 6

II. 1.1 Formation of Covalent Bonds between Small Molecules and

DNA...... 7

II. 1.2 Groove Binding via Non-Covalent Bond Forces ...... 15

II. 1.3 Interaction by Planar Aromatic Molecules between Base

Pairs of DNA ...... 26

II.2 Photoinduced Electron Transfer Reactions Relative to DNA...... 30

11.2.1 Theoretical Aspects of Electron Transfer Reactions ...... 31

11.2.2 Radical Reactions of DNA Bases ...... 41 vi II.2.3 Reagents for DNA Cleavage ...... 50

II.3 Overview of Psoralens ...... 52

CHAPTER III. RESULTS AND DISCUSSION...... 62

III. 1 Sythesis ...... 63

111.2 Solubilities in Aqueous Media ...... 71

111.3 Binding Constants with DNA ...... 71

111.4 Photophysics of Sensitizers ...... 78

111.5 Photochemistry of Sensitizers ...... 89

III.5.1 Photochemistry of Acridines ...... 89

111.5.2 Photochemistry of Halogenated Psoralen Derivatives 92

111.5.3 Photochemistry of Azido Psoralens ...... 95

111.6 Sensitized Unwinding and Cleavage of Plasmid DNA ...... 98

111.6.1 DNA Photocleavage with Acridines ...... 99

111.6.2 DNA Photocleavage with Psoralen Derivatives ...... 100

111.7 Linght Induced Viral Inactivation ...... 105

111.8 Possible Reaction Mechanisms of Sensitizers with DNA ...... 115

111.9 Conclusion ...... 120

CHAPTER IV. EXPERIMENTAL...... 123

IV. 1 Materials...... 123

IV.2 General Procedure ...... 124

IV.3 Synthesis ...... 125

IV.4 Preparative Photolysis ...... 139

IV.4.1 Photolysis of Psoralen Azide ...... 139

IV.4.2 Photocycloaddition of Psoralen derivatives with

Tetramethylethylene ...... 139

IV.5 Measurement of the Water Solubility of a Compound ...... 140

IV.6 Measurement of Binding Constants with DNA ...... 140 vii IV.7 Measurement of Relative Fluorescence and Triplet-Triplet Absorption

Intensities...... 141

IV.8 DNA Photocleavage ...... 142

IV.8.1 Preparation of Buffers ...... 142

IV.8.2 PHotolysis of Supercoiled DNA ...... 143

IV.8.3 Gel Electrophoresis ...... 144

IV.9 Viral Inactivation ...... 145

IV.9.1 Medium Preparation ...... 145

IV.9.2 Growth and Maintenance of E. coli Bacteria and X Phage... 148

IV.9.1 Viral Inactivation ...... 150

REFERENCES...... 152

APPENDIX A. Selected Gel Electrophoresis Plates ...... 173

APPENDIX B. Selected Absorption Spectra ...... 183

APPENDIX C. Selected IR ,1HNMR and 13CNMR Spectra ...... 204 LIST OF FIGURES

Figure Page

1. A shematic structure of an ideal drug ...... 3

2. Chemical structures of some widely used DNA alkylating agents ...... 8

3. Structures of some PBD family of antibiotics and an analogue ...... 12

4. Chemical structures of (+)-CC-1065, duocarmycin A and Pyrindamycin 13

5. The patter of hydrogen bond acceptor, a, and donor, d, of AT and GC base pair 16

6. Chemical structures of netropsin, distamycin A and Hoescht 33258 ...... 18

7. Chemical structures of and its Fe(II) complex ...... 20

8. Chemical structures of molecules ...... 25

9. Chemical structures of daunomycin, adriamycin and the general formula of

2,6-(co-aminoalkanamido)-9,10-anthracenediones ...... 27

10. Chemical structures of some DNA intercalating agents ...... 29

11. A comparison of the effect of solvent polarity on ET and energy transfer 38

12. Schematic drawing to show the hydrogen abstraction by the radical -TH from

sugar residues ...... 42

13. Chemical structure of psoralen ...... 53

14. A possible energy diagram of 8-MOP in different solvents ...... 55

15. Correlation between A.uvmax with Reichardt’s Ej(30) solvent scale and Taft’s

solvatochromic parameters 7t* and a ...... 82

16. Correlation between Xfmax with Taft’s solvatochromic parameters rc* and a.... 84

17. The effect of water content on fluorescence intensity of 8-MOP in solvent

mixtures...... 86 ix 18. Stern-Volmer plots of the fluorescence of Acr-I quenched by calf thymus

DNA, CpG, GMP and CMP ...... 91

19. Action spectra of A, phage viral inactivation with acridine sensitizers on

photolysis with 350nm light ...... 107

20. Action spectra of A. phage viral inactivation with acridine sensitizers on

photolysis with 400~440nm light ...... 110

21. Action spectra of A. phage viral inactivation with HEP, CEP, BEP, and IEP.... 113

22. Action spectra of A. phage viral inactivation with TBP, BCP, BPP, and AMT.. 114

23. Representation of the intercalation of psoralen into DNA ...... 119

x LIST OF SCHEMES

Scheme Page

1. Chemical mechanisms of DNA alkylation by different reagents ...... 9

2. Mechanistic scheme for DNA crosslinking by mitomycin C ...... 10

3. Mechanism of DNA strand scission induced by blemycin ...... 22

4. Mechanisms of electron transfer reaction ...... 32

5.Schematic representation of the Coulombic interactions on the photoinduced

electron transfer between sensitized benzophenone derivatives and leuco

crystal violet (LCV) ...... 40

6. Mechanism of hydroxyl radical reacting with adenosine ...... 45

7. Acid and base equilibrations of guanosine and its radicals ...... 46

8. Mechanism for the oxidation of guanosine and the following decomposition

of its radical ...... 47

9. Mechanism of one-electron reduction of adenosine ...... 49

10. [2+2]-cycloaddition reactions of psoralen with thymine residues of DNA 57

11. The synthetic route for 3-amino-6-iodoacridine (Acr-1) ...... 63

12. The synthetic route of 5-azido-8-methoxypsoralen (AMP) and 5-azido-

8-(3-diethylaminopropyloxy)psoralen ...... 65

13. Scheme for the syntheses of some halogenated psoralen derivatives ...... 67

14. Synthetic route to tribromopsoralen derivatives ...... 69

15. Synthetic route of 4’-triethylammoniomethyl-4,5\8-trimethylpsoralen 70

16. Reaction equation of photolysis of 5-bromo-8-methoxypsoralen in the

presence of TME in ethanol ...... 93

17. Possible mechanism of photochemical reaction of 8-MOP or 5-bromo- xi 8-methoxypsoralen with TME ...... 94

18. Reactions of phenyl nitrene ...... 96

19. Reaction of AMP upon irradiation with 35()nm light ...... 97

20. Reaction mechanism of DNA-bound Acr-I initiated by photoactivation 116

21. Possible reaction mechanism for DNA-bound halogenated psoralen

derivatives upon UVA irradiation ...... 117

xii LIST OF TABLES

Table Page

1. Hydrogen bonding patterns of DNA major and minor grooves ...... 17

2. The intrinsic binding constants of some psoralen derivatives with calf thymus

DNA in Tris HC1 buffer at room temperature ...... 76

3. The maximum wavelength of UV-vis absorption of Acr-NH 2 and Acr-I in

various solvents ...... 79

4. Collection of Xuvni.,x and A.rmax of 8-MOP in different solvents, Reichardts

Et(30) solvent polarity index, dielectric constants(e), and Taft’s 7i* a

solvatochromic parameters ...... 81

5. Photophysical data of some psoralen derivatives ...... 87

6. Data of reduction potentials and singlet state energy of Acr-NH 2 , Acr-I, 8-MOP,

and 5-bromo-8-methoxypsoralen ...... 90

7. Slopes(kqx M-1) of the Stern- Volmer fluorescence quenching plots ...... 90

8. The minimum concentrations of psoralen derivatives to completely unwind

and nick DNA...... 102 9. Logs of viral inactivation of X phage resulted from 350nm light photolysis

in the presence of acridine sensitizers ...... 106

10. Data of X phage viral inactivation of with acridine derivatives

and 350nm light ...... 108

11. Data of X phage viral inactivation of with acridine derivatives

and 400~440nm light ...... 108

12. Data of the viral inactivation of X phage with our new psoralen derivatives

and 350nm light ...... 112 xiii LIST OF PLATES

Plate Page

I. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with acridines and 350nm light ...... 174

II. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with acridines and 400~440nm light ...... 175

III. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with HEP and 350nm light ...... 176

IV. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with CEP and 350nm light ...... 176

V. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with BEP and 350nm light ...... 177

VI. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with IEP and 350nm light ...... 178

VII. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with TBP and 35()nm light ...... 179

VIII. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with BCP and 350nm light ...... 180

IX. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with BPP and 350nm light ...... 181

X. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with AMT and 350nm light ...... 182

xiv LIST OF SPECTRA

Spectrum Page

1. UV-vis absorption of Acr-NH 2 in 1% NHAc(aq), pH=5.5 ...... 184

2. UV-vis absorption of Acr-NH 2 in DNA solution ...... 185

3. UV-vis absorption of Acr-I in 1% NUAc(iU|), pH=5.5 ...... 186

4. UV-vis absorption of Acr-I in DNA solution ...... 187

5. UV-vis absorption of AMP in Tris HC1 buffer...... 188

6. UV-vis absorption of the product of AMP after photolysis with 350nm light

in EPA at 77K...... 189

7. UV-vis absorption of ADP in Tris HC1 buffer ...... 190

8. Transient aborption spectra produced by LFP of ADP in the presence and

the absence of DNA ...... 191

9. UV-vis absorption of HEP in Tris HC1 buffer ...... 192

10. UV-vis absorption of CEP in Tris HC1 buffer ...... 193

11. UV-vis absorption of BEP in Tris HC1 buffer ...... 194

12. UV-vis absorption of IEP in Tris MCI buffer ...... 195

13. UV-vis absorption of TBP in Tris HC1 buffer ...... 196

14. UV-vis absorption of BCP in Tris MCI buffer ...... 197

15. UV-vis absorption of BPP in Tris HC1 buffer ...... 198

16. Triplet-triplet absorption of HEP and CEP in Tris HC1 buffer ...... 199

17. Triplet-triplet absorption of BEP and IEP in Tris HC1 buffer ...... 200

18. Triplet-triplet absorption of TBP and AMT in Tris HC1 buffer ...... 201

19. UV-vis absorption of H PoTM E in ethanol ...... 202

20. UV-vis absorption of HPoTM E in 1,4-dioxane ...... 203 xv 21. JHNMR of 3-acetamido-6-aminoacridine ...... 205

22 .13CNMR of 3-acetamido-6-aminoacridine ...... 206

23. Mass spectrum of 3-acetamido-6-aminoacridine ...... 207

24. *HNMR of Acr-I ...... 208

2 5 .13CNMR of Acr-1 ...... 209

26. Mass spectrum of Acr-I ...... 210

27. IR of 8-methoxy-5-nitropsoralen ...... 211

28. *HNMR of 8-methoxy-5-nitropsoralen ...... 212

29. ,3CNMR of 8-methoxy-5-nitropsoralen ...... 213

30. Mass spectrum of 8-methoxy-5-nitropsoralen ...... 214

31. IR of 5-amino-8-methoxypsoralen ...... 215

32. !HNMR of 5-amino-8-methoxypsoralen ...... 216

33. 13CNMR of 5-amino-8-methoxypsoralen ...... 217

34. Mass spectrum of 5-amino-8-methoxypsoralen ...... 218

35. IR of 5-azido-8-methoxupsoraIen ...... 219

36. !HNMR of 5-azido-8-methoxupsoralen ...... 220

37. 13CNMR of 5-azido-8-methoxupsoralen ...... 221

38. Mass spectrum of 5-azido-8-methoxupsoralen ...... 222

39. JHNMR of 8-(3-bromopropyloxy)psoralen ...... 223

40. IR of 8-(3-bromopropyloxy)-5-nitropsoralen ...... 224

41. ]HNMR of 8-(3-bromopropyloxy)-5-nitropsoralen ...... 225

4 2 .13CNMR of 8-(3-bromopropyloxy)-5-nitropsoraIen ...... 226

43. Mass spectrum of 8-(3-bromopropyloxy)-5-nitropsoralen ...... 227

44. IR of 5-amino-8-(3-bromopropyloxy)psoralen ...... 228

45. !HNMR of 5-amino-8-(3-bromopropyloxy)psoralen ...... 229

46. ,3CNMR of 5-amino-8-(3-bromopropyloxy)psoralen ...... 230

47. Mass spectrum of 5-amino-8-(3-bromopropyloxy)psoralen ...... 231 xvi 48. IR of 5-amino-8-(3-diethylaminopropyloxy)psoralen ...... 232

49. *HNMR of 5-amino-8-(3-diethylaminopropyloxy)psoralen ...... 233

5 0 .13CNMR of 5-amino-8-(3-diethylaminopropyloxy)psoralen ...... 234

51. Mass spectrum of 5-amino-8-(3-diethylaminopropyIoxy)psoralen ...... 235

52. IR of 5-azido-8-(3-diethylaniinopropyloxy)psoralen ...... 236

53. *HNMR of 5-azido-8-(3-diethylaminopropyloxy)psoralen ...... 237

5 4 .13CNMR of 5-azido-8-(3-diethylaminopropyloxy)psoralen ...... 238

55. Mass spectrum of 5-azido-8-(3-diethylaminopropyIoxy)psoralen ...... 239

56. !HNMR of 5-iodo-8-methoxypsoralen ...... 240

57. 'HNMR of 8-(3-bromopropyloxy)-5-chloropsoraIen ...... 241

5 8 .13CNMR of 8-(3-bromopropyloxy)-5-chloropsoralen ...... 242

59. Mass spectrum of 8-(3-bromopropyloxy)-5-chloropsoralen ...... 243

60. 'HNMR of CEP ...... 244

61. !HNMRofBCP ...... 245

6 2 .13CNMR of. BCP ...... 246

63. Mass spectrum of BCP ...... 247

64. iHNMRofBPP ...... 248

65. 13CNMR of BPP ...... 249

66. Mass spectrum of BPP ...... 250

67. ]HNMR of 4,-chloromethyl-4,5\8-trimethylpsoralen ...... 251

68. 'HNMR of 4,-(triethylamonio)methyl-4,5\8-trimethylpsoralen ...... 252

6 9 .13CNMR of 4’-(triethylamonio)meihyl-4,5\8-trimethylpsoralen ...... 253

70. ^ - ^ C 2-dimentional NMR of 4,-(triethylamonio)methyl-

4,5’,8-trimethylpsoraleii ...... 254

71. Mass spectrum of 4’-(triethy!amonio)methyl-4,5\8-trimethylpsoralen 255

72. ^N M R of 4\5,5’-tribromo-4\5,-dihydro-8-methoxypsoralen ...... 256

73. ,3CNMR of 4\5,5,-tribromo-4\5’-dihydro-8-methoxypsoralen ...... 257 xvii 74. Mass spectrum of 4\5,5’-tribromo-4\5’-dihydro-8-methoxypsoralen ...... 258

75. !HNMR of 4\5,5,-tribromo-8-hydroxypsoralen ...... 259

76. Mass spectrum of 4\5,5’-tribromo-8-hydroxypsoralen ...... 260

77. !HNMR of 4’,5,5’-tribromo-8-(3-bromopropyloxy)psoralen ...... 261

78. Mass spectrum of 4\5,5,-tribromo-8-(3-bromopropyloxy)psoralen ...... 262

79. *HNMR of TBP ...... 263

80. Mass spectrum of TBP ...... 264

81. *HNMR of HPoTM E ...... 265

82. Mass spectrum of HPoTM E ...... 266

83. ’HNMR of BPoTM E ...... 267

84. Mass spectrum of BPoTM E ...... 268

xviii ABBREVIATIONS

A adenine DNA deoxyribonucleic acid

Acr-I 3-amino-6-iodoacridine DTT dithiothreitol

Acr-NH2 3,6-diaminoacridine EDTA ethylenediamineietraacetate

ADR adriamycin EPA diethyl ether:pentane:ethanol (5:2:5)

AIDS acquired immune deficiency ESR electron spin resonance syndrome ET electron transfer

AMSA amsacrine Et ethidium

AMT 4’-aminomethyl-4,5\8- fs femtosecond trimethylpsoralen G guanine

ATP adenosine triphosphate GMP guanosine-5’-monophosphate

BLM bleomycin HIV human immunodeficiency virus

BP bcnzophenone IR infrared

C cytosine LCV leuco crystal violet

CIP contact ion pair LFP laser flash photolysis

DAU daunomycin x-MOP x-methoxypsoralen

DBU l,8-diazabicyclo[5.4.0]undec-7-ene NBS N-bromosuccinimide

DDQ 2,3-dichloro-5,6- NC dicyanobenzoquinone NCS N-chlorosuccinimide

DMSO dimethylsulfoxide

xix NIAID National Institute of Allergy and Py pyrimidine base

Infectious Diseases RNA ribonucleic acid

NMR nuclear magnetic resonance SM suspension medium

ns nanosecond SSIP solvent-separated ion pair

OMA optical multichannel analyzer T thymine

PBD pyrrolo|2,l-c]|l,4]benzodiazepine TMP

PDT photodynamic therapy UV ultraviolet

ps picosecond

Pu purine base

H,N N NH,

Acr-NH2 Acr-I

CH

CH,

8-MOP TMP

H-.C

OCH3

H P o T M E B P oT M E

xx H Cl

0(C H 2)3N+(C2H5)3Br 0(C H 2)3N+(C2H5)3Br*

HEP CEP

Br I

0(C H 2)3N+(C2H5)3Br' 0(C H 2)3N+(C2H5)3Br-

BEP IEP

OCH2CH2CH2N+(C2H5)3Br 0(C H 2)3N+(C2H5)2B f

xxi OCH2CH2CH2N(C2H5)2

ADP Chapter I

INTRODUCTION

1.1 Problem Definition

The human immunodeficiency virus (HIV), the causative agent of the infamous acquired immune deficiency syndrome (AIDS) is still spreading all over world. The most recent estimate places the number of people infected with this fatal virus at 15 million1. Despite desperate efforts, the prospect for an effective vaccine or a definitive cure remains rather grim. Last June, the U. S. National Institute of Allergy and

Infectious Diseases (NIAID) decided to postpone the first real-world AIDS vaccine tests after a day-long debate. To date, there are only a few anti-retroviral nucleoside drugs available, and they are inefficiently treating those already infected. The most effective strategy to combat the HIV/AIDS pandemic remains risk behavior change2.

The latest international AIDS conference, held last August in Japan, called for the return to more basic research on HIV.

On the other hand, there are still finite risks of viral transmission through transfusion medicine or the use of blood components, despite great improvements in testing. For example, the current probability of acquiring HIV ranges from 1:40,000 to

1:150,000 per unit of blood transfused, depending on geographic location. The risk of hepatitis virus transmission is estimated to be much higher3. The exclusion of high risk donor groups and screening tests, the methods currently adopted to fend off virus 1 transmission, arc largely but not completely effective. Unfortunately, window periods

exist during which an infected blood sample may be tested as negative for the presence

of HIV but can still transmit disease4. Furthermore, screening assays can not test for the

presence of a new, not yet identified viral agent, which may be hazardous. As far as

large amounts of blood and blood products are concerned5, it is of practical importance

to develop a positive method to safeguard the blood supply in place of the passive

screening tests. A positive method would be to administer a drug to kill any pathogenic

contaminant present in blood without compromising the integrity of its various

components. An example of such a positive method is to treat plasma with solvent detergent, which dissolves the lipid membranes of enveloped viruses6. This procedure does not damage blood clotting proteins but kills platelets and red blood cells.

Unfortunately, the solvent detergent method does not inactivate non-enveloped pathogenic viruses, which include certain types of hepatitus viruses.

L2 Drug Design Strategies

With very few exceptions, drugs affect cell functions by interacting with three types of cellular components: (1) biological membranes; (2) various proteins, such as enzymes, receptors, transport proteins, etc.', (3) nucleic acids, including ribonucleic acids (RNA) and deoxyribonucleic acids (DNA). Theoretically, one can choose any one of the three possible targets in designing a drug.

For the sterilization of blood products, in fact, a huge body of membrane targeted sensitizers have been extensively studied. Among these are Photofrin II7, methylene blue8, and merocyanine9, just to name a few. Like all other drugs used in photodynamic therapy (PDT), these sensitizers try to exploit alleged, quantitative differences in the binding of the sensitizers to pathogens relative to blood components, or differences in the sensitivity of viral and cellular membranes to the reactive oxygen species generated by the sensitizers. The precise chemical factors responsible for this strategy, however, are essentially unknown. To date, theoretical guidance is lacking to guide the design of better membrane targeted drugs. At this point, the development of new compounds must be evaluated by trial and error.

Specific chemical selectivity is possible regarding drugs designed to target proteins. In principle, such a drug will usually inactivate only a single kind of virus or even only a single strain of virus. Another deficiency of this strategy is that such a drug may create evolutionary pressure on the virus and hasten its mutation into a new strain.

In fact the so-called “antibiotic resistance” has raised tremendous concern in the biotechnology community10.

* / Water Solual /o/cnlial Reactive DNA Binding Part .Part / /Fjfnction

Possible Functional Overlaps

Figure 1. A schematic structure of an ideal drug.

Nucleic acids have been chosen to be the targets of sensitizers developed in our laboratory. This strategy is based on the fact that red cells, platelets, and plasma proteins do not contain any genomic nucleic acid, while all pathogens (with the exception of prions) in the blood supply do. To be an effective DNA-targeted drug for the sterilization of blood products, the drug must incorporate several crucial features, as shown in Figure 1. Firstly, it must interact selectively with nucleic acids relative to biological membranes and any other cellular components present. The higher the selectivity is, the better the drug. Secondly, the drug must bind to nucleic acids so

tightly that it stops nucleic acids from replication. Alternatively, once an inert drug

binds to a nucleic acid, it must be capable of triggered activation to damage or destroy

the nucleic acids thereby inactivating the virus. Thirdly, in order for the drug to enjoy

these two traits, it must be water soluble, to a certain extent, to be compatible with

biological systems. Last but not least, the toxicity of the drug must be acceptable. To

achieve this goal, all the decomposition products, as well as the drug itself, must be

benign in all metabolic aspects. All of the aforementioned characters of a successful

drug are not totally independent of one another. One characteristic may compromise

other properties. The enhancement of the water solubility of a drug, for example, may

lessen its affinity for nucleic acid and biomembrane due to their hydrophilic moieties. It

may also alter the binding selectivity between these two classes of biomolecules.

1.3 Scope and Objective

As far as our drug molecule is concerned, planar aromatic DNA intercalators

have been chosen for the binding of DNA. Photochemistry will be employed to trigger

the production of free radicals in situ to cleave DNA. Quaternary ammonium salts will

be used to increase the water solubility of drugs and to achieve their binding selectivity

toward DNA. Psoralens have been selected to be DNA intercalators because of their minimal biological toxicity11. Fortunately, several sensitizers have been synthesized in this laboratory and have demonstrated potent viral inactivating ability toward model viruses12. Some of them have been successfully utilized for the sterilization of blood products13 and are currently undergoing further tests for clinical applications. Since the strategy adopted here is the cleavage of viral nucleic acids, the possibility that the virus might mutate is remote.

In this dissertation, it will be reported that several promising new sensitizers for viral inactivation have been synthesized and assessed. In addition, a series of experiments have been performed to investigate the mechanism of action of these drugs. A systematic description of the mechanism of action of these drugs will be proposed based on the experimental results. Although the exact mechanisms are far from clear, some conclusions can be drawn and they are useful in the design of the next generation of experiments. Chapter II

LITERATURE OVERVIEW

To rationally develop an effective drug calls for the perfect unification of

multidisciplinary knowledge. It is helpful to explore the advances in some related fields

to gain a proper perspective. Hence, this chapter will be devoted to a review of the

literature on fields such as drug-DNA interactions, photoinduced electron transfer, and

the photophysics and photochemistry of psoralens.

II. 1 Drug-DNA Interaction

DNA has taken a central position in understanding the conundrum of life, since

its discovery as genetic material14 and the elucidation of its physical structure15. The

past two decades have been marked by an explosion of research effort directed toward

the isolation and evaluation of naturally occurring nucleic acid targeted drugs and

toward the design and synthesis of model compounds that can specifically recognize

and cleave DNA16- 17. This is mainly a result of an increasing awareness among

chemists that DNA actually contains two levels of information. The first is the familiar

one embodied by successions of base pair triplet codons, used primarily for storing

information as to the amino acid sequence of the derived polypeptide chains. The readout of this information is realized by extremely complicated machinery during reproductive processes. The other level of information, however, is more

straightforward, as far as readout is concerned. Many control proteins such as 6 repressors, polymerases and other enzymes bind specifically to certain DNA base sequences. Some naturally occurring small compounds such as metallobleomycin18 and neocarzinostatin19, likewise, clearly have the ability to recognize and bind to specific sequences in DNA. In such cases, the recognition process is more or less through the interactions of these small molecules with DNA by a combination of shape, charge, polarity, or pattern of hydrogen bonding. Presently, there are generally three categories which describe how small molecules bind to DNA20: (1) Formation of covalent bonds between small molecules and DNA; (2) Groove binding via noncovalent interactions;

(3) Intercalation by small planar hydrophobic molecules between base pair planes. It is important to notice, however, that this classification is not applicable in many cases.

Since many drug molecules interact with DNA in multiple ways, the classification of these drugs are often arbitrary.

II.l.l Formation of Covalent Bonds between Small Molecules and DNA

Among the compounds first studied in regard to their interactions with DNA were nitrogen mustards, aziridines, chloroethylnitrosoureas and some related small synthetic DNA alkylating agents. Amazingly, the first drugs clinically employed to treat human cancer, in fact, are DNA cross-linking nitrogen mustards21. The structures of several of the most extensively studied drugs are presented in Figure 2. The biological effectiveness of these agents lies in their high reactivity towards DNA bases as substrates in nucleophilic reactions. DNA bases are susceptible to alkylation under physiological conditions, if these chemicals are available nearby. Calculations have clearly demonstrated that N7 of guanine is the most nucleophilic site of purine bases, followed by 06 of guanine and N1 and N3 of adenine22. Indeed, only guanine-N7 adducts with some alkylating reagents have been isolated in any significant amount23.

Recently, a host of modified base products were isolated from DNA treated with chloroethylnitrosoureas24. The relative biological significance of the different alkylated 8 base products is still an open question, It is clear that their biological significance appears to be unrelated to their chemical yields25.

C1CH2CH2 c ic h 2c h 2 /= = x ; n - c h 3 /V ch 2ch2co2h C1CH2CH2 c i c h 2c h 2 '—'

Mechlorethamine Chlorambucil

° - pP ^ •IS C1CH2CH2- n 'N_ y v N :f 'N - 7 c ic h 2cr>u h 2 H

Cyclophosphamide Thiotriethylenephophamide

O C1CH2CH2 . xyA 1}T NHCH2CH2Cl NO

Diaziquone Bichloroethylnitrosourea

Figure 2. Chemical structures of some widely used DNA alkylating agents

The chemical mechanisms of the action of various alkylating agents are straightforward as shown in Scheme l26. Both nitrogen mustards and aziridines can undergo the formation of aziridinium ion intermediates followed by the nucleophilic attack of DNA, or direct DNA attack, depending on the reaction conditions and the nature of the R group attached on the nitrogen atom. It is obvious, according to these chloroethylnitrosoureas undergo decomposition to form reactive chloroethyldiazohydroxide intermediates, which chloroethylate nucleophilic sites of

DNA bases. Additionally, it has to be pointed out that all useful DNA alkylating agents are multifunctional, so that they are capable of cross-linking DNA. It is believed that

DNA cross-linking is essential for potent cytotoxicity, while monoalkylation products are genotoxic and the probable cause of mutagenic and carcinogenic events27.

CH2CH2C1

CH2CH2C1 CH2CH2Nu +Nu* R-N R— N \ -Cl* \ CH2CH2C1 c h 2c h 2ci

RNHCH,CH,Nu

O

ClCHoCH- C1CH2CH2N=N0H

NO

Scheme 1. Chemical mechanisms of DNA alkylation by different reagents. 10 The clinical applications of these drugs are largely hampered by their lack of

affinity towards DNA. Much of the drug is lost as a result of hydrolysis and by alkylation of nucleophilic sites of cell components other than DNA. To enhance the drug’s selectivity for DNA, recent work has centered on attaching these agents to strong

DNA-binding carriers, such as acridine28, polyamines29, and distamycin A30.

mitomycin C

Nu Nu Nu N T !, Nu

HO. HO.

O’ CH CH CH

Scheme 2. Mechanistic scheme for DNA crosslinking by mitomycin C.

On the other hand, nature has also evolved many even more elaborate DNA alkylating agents, most of which bear the character of specificity for attacking sites.

Some of these agents are now in the repertory of clinical drugs.

Mitomycin C is a well characterized, bifunctional electrophilic DNA cross- linking reagent. It is reactive only after reductive activation. The presently accepted 11 mechanism of action of mitomycin and its analogs is described in Scheme 2. Reductive activation of mitomycin C results in the formation of the reduced pyrroloquinone, followed by the sequential departure of leaving groups in company with the entry of nucleophilic groups from nucleic acids31. Mitomycin C and its related pyrrolo-derived compounds exhibit preferential cross-linking at 5’-d(CG) over 5’-d(GC). Unlike small synthetic DNA alkylating agents, the N2 atom of guanine bases is the nucleophile instead of the N7 atom of guanine bases32.

Pyrrolo[2,l-c][l,4]benzodiazepines (PBDs) are a family of antitumor antibiotics with possible therapeutic applications as cancer drugs, as selective and-infective agents, and as probes and tools in molecular biology. Since the discovery of anthramycin some

30 years ago33, nine naturally occurring PBDs have been uncovered and a variety of synthetic analogs (Figure 3) have been prepared. Invariably, the PBD antibiotic family consists of an aromatic A ring, a middle seven-membered B ring of imine or carbinolamine methyl ether at the N10-C11 positions, and a pyrrolo C ring which is either saturated or unsaturated. The carbinolamine methyl ether, imine, and carbinolamine on the B ring are supposed to be capable of interconversion34 and the imine is the species responsible for the alkylation of DNA due to its electrophilic character. On the whole, PBD interferace with DNA processing is believed to proceed by first non-covalent binding to the minor groove of DNA, followed by the formation of an adduct between the imine of PBD and the N2 atom of a guanine base35, with an

(S)-configuradon of the C ll linkage36. Interestingly, all of the known, natural PBDs have an (S)-configuration at the chiral Cl la position, which forces the molecule into a right-handed twist when viewed from the C ring towards the A ring. The isohelicity of

PBDs with B-form DNA leads to a snug fit at the binding site. Indeed, a synthetic PBD with the (R)-configuration at Cl la is devoid of DNA binding affinity or cytotoxicity.

Moreover, fifteen dilactam PBD analogues have been synthesized to investigate the non-covalent interactions of PBDs with DNA37. The study demonstrated that the 12 presence of hydroxyl or acetoxyl groups at the C2 and the C8 positions enforced the

PBD binding ability for DNA, which vanished with a change of stereochemistry at C2 from the (R) to the (S) configuration. Additionally, DNA footprinting studies38 and the exonuclease III stop assay39 have shown that PBDs prefer to bond to 5’-PuGPu sequences, while S’-PyGPu and 5’-PuGpy are of intermediate sequence preference and

5’-PyGPy sequences are least favored.

^ CONH

Anthramycin Tomaymycin

CH3NH

Sibiromycin

Pyrrolo(l, 4)benzodiazepine- Chicamycin 5, lO-dione(dilactam)

Figure 3. Structures of some PBD family of antibiotics and an analogue. 13

(+)-CC-106540, duocarmycin41, and pyrindamycin42 arc still another class of structurally related natural occurring members of exceptionally potent antitumor antibiotics (Figure 4). Physically, like PBDs, (+)-CC-l065 is a right-handed twisted,

H,C OH OCHOH OCH O (+)-CC-1065

8 OCH, O

Duocarmycin A

8

Pyrindamycin A

Figure 4. Chemical structures of (+)-CC-1065, duocarmycin A and Pyrindamycin. 14 banana-shaped molecule optimally fitting within the minor groove of DNA. But it is

selective for AT-rich sites rather than for GC-rich regions. Chemically, it alkylates

DNA by the stereoelectronically-controlled addition of adenine at the N3 atom to the

least substituted carbon of the activated cyclopropane43. A direct relationship between the reactivity of an agent and its in vitro cytotoxicity were arrived at earlier from studies of simple derivatives of the (+)-CC-1065 alkylation subunit44. However, results which do not support this interpretation have been recently reported. Specifically, a linear relationship between solvolytic chemical stability and biological potency was observed45. This means that the chemically more stable agents may be expected to constitute the biologically more potent agents, provided that the agents possess sufficient reactivity to effectively alkylate DNA. Presumably, this may be attributed to the more effective delivery of the more stable agents to their intracellular targets.

Cis-diamminedichloroplatinum(II) (cisplatin), another DNA cross-linking antitumor antibiotic, has been the object of extensive studies, since the discovery of its antitumor properties in the mid 1960s46. It was approved for clinical trials in 1971, and today cisplatin-based chemotherapy achieves a cure rate of about 95% for testis cancer47. Mechanistically, once cisplatin reaches DNA, the platinum atom is coordinated by the N7 atom of two guanine bases in the major groove, preferably cross- linking DNA at purine-rich regions, particularly at the segment of multiple guanine residues48. The coordination between the platinum atom and the N7 atom of guanine bases is so strong that even cyanide ions replace the DNA ligand only very slowly49.

This tempted the authors to conclude that ligand substitution of the bifunctional DNA adducts by other intracellular nucleophiles probably does not happen at all. On the other hand, the two chlorine ligands in cisplatin can be easily substituted by intracellular nucleophiles, including water molecules, during passive diffusion into cells. This means that only a small portion of the platinum atoms bind to DNA50. However, most of these ligand substitution reactions are reversible, except some strong platinum ligands such as 15 the sulfur atoms available in glutathione and cysteine and methionine amino acid side- chains. This kind of unyielding coordination has been proposed to be responsible for the side effects of platinum drugs51. In addition, studies in vitro produced similar results for both cisplatin and its trans counterpart. However, the latter compound turned out to have no biological effect in vivo. Current views have attributed the differential effects of these two isomers to DNA repair mechanisms. One theory, dubbed “differential repair”, suggested that repair proteins can recognize the DNA distortion caused by rraw-diamminedichloroplatinum(II) adducts leading to the restoration of DNA, but the same proteins fail to discern and repair the cisplatin adducts52.

In summary, all the compounds discussed above can somehow alkylate DNA thus forming one or two covalent bonds. They can bind either at the major groove or at the minor groove of DNA, depending on their chemical structures. At the same time, most of them have a propensity for sequence selectivity. By and large, the atoms of

DNA which form covalent bonds are the nitrogen atoms of purine bases, particularly those on guanine residues. This can be attributed to their nucleophilicity. Once anchored on DNA, the resulting adducts are stable and thus can block the processing of

DNA replication. Finally, the biological side effects of these drugs are largely due to their indiscriminate alkylation of other intracellular components encountered on their way to the DNA.

II.1.2 Groove Binding via Non-Covalent Bond Forces

It is clear that the chemical environment of the DNA major groove is radically different from that of the DNA minor groove. Specifically, as indicated in Table 1 and

Figure 5, the major groove has an asymmetric four-symbol recognition code for an AT base pair and a three-symbol recognition code for a GC base pair. Nevertheless, the minor groove possesses a symmetric two-symbol recognition code for an AT base pair 16 and a three-symbol recognition code for a GC base pair. Therefore, the major groove

carries a larger volume of information than does the minor groove.

CH

H" N-

\ H'

H

Adenine Thymine Guanine Cytosine

Figure 5. The pattern of hydrogen bond acceptor, a, and donor, d, of AT and GC base pair. Proteins and drugs "read" the base sequence by sensing the pattern of hydrogen bond acceptor and donor on the DNA groove floor.

Generally, proteins, e. g. promoters and repressors, utilize the wide major groove for faithful replication, because of the greater information content, as well as ready accomodation to the bulky protein molecules, while the narrow minor groove is normally unoccupied and represents a vulnerable site for extraneous attack. This presumably gives an opportunity for the evolution of natural antibiotics which invade the DNA in competing organisms. Drugs and other small molecules tend to bind in the minor groove of DNA. The stability of binding interactions between DNA and its groove binders depends on numerous factors: (1) hydrogen bonds formed between the hydrogen bond acceptors and donors on the DNA groove edges and the binding molecule; (2) intermolecular van der Waals contacts between the ligands and the walls 17 of the DNA groove resulting from the overall shape matchup; (3) electrostatic interactions between the charges on ligands and the DNA phosphate ions.

Table 1. Hydrogen bonding patterns of DNA major and minor grooves.

Base pair Major groove Minor groove

AT a d a (CH3) a a

TA (CH3) a d a a a

GC a ad a d a

CG d a a a d a

There are mainly two kinds of non-covalent DNA groove binders, in regard to their influence on DNA. One class interacts with the DNA polymer in an equilibrium manner and does not leave any lasting record of its temporary binding to DNA. The other class can produce DNA strand scission after binding to DNA.

Members of the non-strand-breaking class are netropsin, distamycin A and their related synthetic analogues such as Iexitropsins and Hoescht 33258 pyrrole-amidine antibiotics53 (Figure 6). None of these antibiotics have to date made their way to routine clinical practice, largely due to their exteme fragility under physiological conditions.

However they currently serve as a paradigm of minor groove binding.

Both netropsin and distamycin A bind only to double-stranded B-DNA firmly and are selective for AT-rich segments54. It has been determined that the free energy change for the binding of each carboxamide unit of distamycin A to AT polymers is

2.0kcal/mol, which is comparable to the formation of a normal hydrogen bond, while that for the binding of a carboxamide to a GC unit is only 0.95kcal/mol55. In contrast to the alternating AT polymer, the binding of netropsin to the homopolymer 18 poly(dA)poly(dT) has been reported to be dominated by the entropy change (AS) rather than the enthalpy change (AH)56. This was interpreted to mean that the homopolymer is more heavily hydrated and that the disruption of hydration accounts for the entropy increase of binding. Furthermore, it must be pointed out that the hydrogen bonds

NHo NH- N NH2

V .C H 3

CH

H- jC NH CH3

Netropsin Distamycin A

Hoescht 33258

Figure 6. Chemical srtuctures of netropsin, distamycin A and Hoescht 33258. 19 associating netrposin and B-DNA are three-centered bonds57. These hydrogen bonds are formed between the amide NH groups of a drug and two atoms of the base edge nitrogen or oxygen atoms of the DNA. The formation of hydrogen bonds occurs when the drug approaches the groove floor, replaces the water molecules there, and fixes itself into the groove. However, when netropsin binds to the CGCGATATCGCG DNA oligomer, only normal hydrogen bonds are developed between the individual amide

NH’s of the drug and the 0-2 atom or N-3 atom of thymine and adenine58. This is also true for some lexitropsins, rationally designed synthetic analogs of netropsin/distamycin, which have an even stronger binding ability for DNA59. Finally, netropsin, distamycin A and their analogs with n amides cover a binding site of n+1 base pairs60. For example, with three amide linkages, netropsin covers four base pairs of DNA.

The specificity for the AT base pair displayed by netropsin and distamycin A is achieved by means of nonbonded contacts between the drug’s methylene or pyrrole CH groups and adenine C-2 hydrogen atoms on the floor of the minor groove. In contrast, the protruded 2-amino group of guanine interupts this close fit of the drug to the floor of the DNA minor groove54. To test the validity of this interpretation, a host of lexitropsins have been prepared and assessed61. In each of these lexitropsins, one of the naturally occurring pyrrole rings was substituted by a thiazole ring. It turned out that the lexitropsins bearing the sulfur atom directed towards the floor of the minor groove of the DNA exhibited pronounced selectivity for the AT base pair. This was ascribed to the unfavorable interactions between the sterically more demanding sulfur atom and the

C-2 amine groups of the guanine bases. On the other hand, the lexitropsins with the nitrogen atom oriented towards the floor of the DNA minor groove (with the thiazole ring installed opposite to that found in the first kind of lexitropsin) showed the ability to bind, to some extent, to the regions containing GC base pairs. Hoescht 33258 also locates its piperazine ring in the GC region sometimes62. In addition, making the 20 lexitropsin molecules bigger, particularly by inserting methylene groups, improves the drug’s tendency to bind to GC segments63.

Two families of antitumor antibiotics, represented by bleomycin (BLM)18 and neocarzinostatin (NC)19, can cause DNA strand cleavage. BLM, the prototypical DNA

H VH2 CH2CH(NH2)CONH2

n h 2c o * V " c o n h 2

N N CH, „ O

h,nVV° IjP>

> 0 H

0 H NH2

Bleomycin CH2CH(NH2)CONH2

h 2n c o nh AT>s A .N H \ NHjT >0 A-S

H2NCO n'/ \ \_| I / / H\

\—\ 02 O ^ Sugar

Fe(II)-Bleomycin complex

Figure 7. Chemical structures of bleomycin and its Fe(II) complex 21 cleaving agent, is a glycopeptide derivative effective only when complexing with certain transition metal ions, generally with Fe(II)64 (Figure 7). Structurally, BLM can be divided into three parts65: (a) two sugars: glucose and mannose, responsible for the selective distribution of the drug in some cancer cells; (b) metal ion chelators, consisting of pyrimidine, P-aminoalanine, and P-hydroxyimidazole moieties; (c) the bithiazole moiety and a positively charged tail, required for DNA binding and sequence-specific recognition.

To date, it is known that the biological activity of BLM is due to the oxidative damage to DNA induced by its complex with Fe(II) and oxygen. The decomposition of this complex produces reactive oxygen species such as hydroxyl radical and superoxide radical anion, which ultimately generate sugar-centered radicals. These reactions proceed preferentially to generate the radical center at C-4' of deoxyribose66. According to a recent theoretical study, the susceptibility of all the hydrogen atoms of the deoxyribose to radical abstraction reactions follow from energetic considerations: the rate constants for hydrogen atom abstraction from C’-l and C’-3 are approximately two-times larger than that for the hydrogen abstraction of C’-4, which in turn is another four-fold larger than that of C’-2. When steric factors are included, however, the rate constant of hydrogen atom abstraction from C’-4 becomes the largest67. The sugar radical intermediate initiates the collapse of the sugar ring, yielding two types of monomeric products from DNA (Scheme 3)68: free nucleic acid bases and base propenals. In the presence of an additional oxygen molecule, the radical intermediate resulting from the loss of a hydrogen atom at C-4' of deoxyribose combines with an oxygen molecule to form a peroxy radical, whose further decay yields the base propenal product. On the other hand, in the absence of oxygen, the free base is released along with an oxidatively damaged sugar in the intact DNA strand, which is alkali-labile.

Strand scission occurs when the DNA is treated with a base69. In addition to the specificity of homolysis for the hydrogen atom at C-4’, the drugs of the BLM family o 00 ?_0-v n Base R0-P-0^\ 0.,Base _ RO-jj-O.^Q-Base ■ x * v?-v?H O H ^ :B 0 r ? O^P-O- 0 = P-0- OR' UK 0 = P-0- » H 0 1 OR’ \ H y^—Base RO-P-On 0 ,Base W ' 6 - p O H< O O 9 R O -p - o ^ o - + 0*P-0- O- OR' o* Base 0 = P-0* OR’ RO-P-O Base RO-P-O °H' O _ o - —«-ro- p -o . „ 9 H 6- 1 Vh + o.P-or o c\ = y H o r 1

OR'

Scheme 3. Mechanism of DNA strand scission induced by bleomycin. 23 also display sequence selectivity for bases. For linear DNA, the single-stranded scission

at GpC and GpT sequence is preferred70, while the observed preference for the double­

stranded scission occurs at the sequence of Py-G-C-PuPy-C-G-Pu71.

There remain several points open for question about BLM, despite the

tremendous progress on the elucidation of the working mechanism of this drug. First,

the exact chemical structure of the BLM-02-Fe(II) complex is not settled.72 There is no

question about its binding to the secondary amine, pyrimidine, and imidazole, but the

roles of the P-hydroxyhistidine amide and the carbamoyl group of mannose in the

coordination chemistry are still controversial. Furthermore, the observation that BLM

exquisitely mediates specific C-4' hydrogen bond cleavage leaves little doubt that the

BLM-mediated attack proceeds within the minor groove of DNA. However the binding

mode of the drug to DNA is not yet clarified17. Besides the major role of the tripeptide

in binding to DNA, there is also evidence for the participation of the bithiazole moiety

by intercalation. These two kinds of binding modes cannot be discerned so far. Finally, the exclusive abstraction of a hydrogen atom from C-4’ of ribose has recently been challenged. Duff et a/73 reported products envisioned as a result of hydrogen atom abstraction at the C -l’ position, in addition to those of the hydrogen abstraction of C-4’, have been isolated for the oligonucleotides of d(CGCT 3 A 3 GCG) and d(GCGTAGCG) subjected to the complex of Fe(II)-BLM.

The Co(III) complexes of BLM, despite their strong binding to DNA, do not inflict any damage to DNA under normal conditions, because the electronic structure

(low-spin, d6) of Co(III) renders it kinetically inert 74. However, Co(III)-BLMs do cleave DNA in a way which is independent of oxygen, when illuminated with light75.

The mechanism of the DNA cleavage is proposed to be photo-induced electron transfer, which will be discussed in detail later, leading to the formation of hydroxyl radicals in close proxomity to the DNA helix.76 24 Neocarzinostaiin(NC) was the first antibiotic uncovered containing a highly strained enediyne motif and later was joined by several other members, e.g., and esperamicin (Figure 8). NC itself consists of an apoprotein that non- covalently binds to and protects a labile bicyclic enediyne chromophore77. The secondary structure of the apoprotein is almost entirely P-sheets. The chromophore is the active ingredient of the biological toxicity and unstable under physiological conditions when separated from the apoprotein. But the addition of methanol can stablize the NC chromophore. In fact, the storage of the chromophore has been achieved in pure methanol at -70°C78.

The NC chromophore binds to DNA in the minor groove by two modes: the intercalation of the naphthoate moiety and the electrostatic interaction and hydrogen bond of the amino sugar moiety with the DNA backbone. Chemically, like all other enediyne antibiotics, the NC chromophore undergoes electronic rearrangement to form a bicyclic diradical species, analogous to the Bergman rearrangement in the formation of 1,4-dehydroaromatic intermediates79, when the chromophore is activated by thiols.

After that, the succeeding chemical processes which cleave DNA are quite similar to that of BLM. However the abstraction of the hydrogen atom occurs selectively at the C-

4' of deoxyribose as displayed by BLM, whereas it mainly proceeds at the C-5' position of deoxyribose for the NC chromophore in the case of single-stranded DNA scission80 and, to a lesser extent, on C -l’ and C-4', as well as on C-5', of deoxyribose in the case of double-stranded DNA scission. It has also been shown that the NC chromophore induced single-stranded DNA cleavage exhibits base selectivity in the order

T>A »C>G and little sequence preference81. Bis-strand DNA damage induced by NC is minor and prefers A G I-A d sites (at the underlined residues), if it happens at all82. 25

CH

OH

CH

NHCH-

NHCH,CHNeocarzinostatin (Chromophore) CH- jSSS OH C X ^N H

OH OCH Calicheamicin Yi HO OCH CH-iO.

NH

CH OH CH NHCH,CH

HO

Esperamicin Aj

Figure 8. Chemical stryctures of enediyne molecules. 26

II.1.3 Intercalation by Planar Aromatic Molecules between Base Pairs of DNA

Lerman has been associated with DNA intercalation ever since he proposed the intercalation model to explain fiber diffraction and solution phase observations of DNA acridine complexes in 196183. This simple yet elegant intercalation model postulates that the planar intercalators slide between adjacent base pairs within the DNA helix by extension of the DNA backbone. This non-covalent binding is reversible. The driving force for this kind of binding derives from k-k stacking and dipole-dipole interactions of the hydrophobic aromatic moieties of the intercalators with that of nucleic acid bases. The side chains and positive charges on the intercalators futher stabilize the complexes formed with hydrogen bonds and electrostatic interactions. It is now well documented84 that the following physical effects and characteristic changes to DNA structure should ensue as a result of the intercalation: (1) an extension, unwinding, and stiffening of the DNA helix; (2) the tc-tc stacking and dipole-dipole interactions between the sandwiched intercalating faces; and (3) the rigidity and orientation of the intercalator within the helix due to a substantial structure overlap between the bases and the intercalator. Accordingly, a four-point minimal set of experimental criteria have been recommended in order to establish intercalation. They are DNA structural changes, electronic interaction, molecular orientation and rigidity, and structural ramifications85. Additionally, x-ray crystallography has helped to visualize the intercalative interaction of some drug-oligonucleotide complexes86.

Daunomycin (daunorubicin, DAU) and adriamycin (doxorubicin, ADR) (Figure

9), two prominent representative anthracycline antibiotics, intercalate into DNA. Their aglycone moieties both contain an anthraquinone type structure. The structure of the 2:1 complex between DAU and d(CGTACG) obtained by x-ray diffraction analysis87 showed that the chromophore of the drug was sandwiched by CpG steps at both ends of a distorted B-DNA double helix, with the D ring protruding into the major groove and 27 the amino sugar lying in the minor groove. Presumably, the two hydrogen bonds formed between the hydroxyl group 09 and the ether 07 of DAU and the N3 and N2 positions of the guanine base further stabilize the complex and provide the specificity of 5’-(A/T)CG for DAU-DNA intercalation88. Anthracycline derivatives wich lack the

09 group are inactive.

Daunomycin, R=H; Adriamycin, R=OH.

O

2,6-Bis(G)-aminoalkanamido)-9,10-anthracenediones.

Figure 9. Chemical structures of daunomycin, adriamycin and the general formula of 2,6-bis(co-aminoalkanamido)-9,10-anthracenediones.

Clinical applications of both DAU and ADR are severely limited by their toxicity to the heart89. The cardiotoxicity of these drugs are believed to be related to the 28 free radical species produced by the reduction of the quinone moity90. Thus, recent attention has been focussed on the development of aminoalkyl-functionalized anthraquinones91. For example, aminoalkanamido-substituted anthracene-9,10-diones

(Figure 9) were designed with the aid of computer modeling for the substitution of

DAU and ADR, as anthracenedione is less readily reduced than anthraquinone.

Echinomycin and other related quinoxalines antibiotics are examples of molecules which display unique bifunctional intercalation. At low ionic strength, echinomycin yields twice the unwinding and elongation of the DNA helix as does ethidium92. Each natural quinoxaline antibiotic contains two quinoxaline chromophores cross-linked by a peptide ring. These antibiotics only differ in respect of parts of the bridging peptide rings, as shown in Figure 10. The molecule of the exquinoxaline antibiotic is believed to have a rigid U-shaped appearance which can snuggly bis- intercalate into the DNA helix so as to sandwich two base pairs between the two quinoxaline chromophores. Indeed, x-ray crystallography has confirmed the validity of this structure93. However, this neat picture has been challenged by recent work, which suggests a suprising degree of conformational heterogeneity of quinoxaline antibiotics in solution94. In fact, this conformational heterogeneity has been evoked to explain a long-standing puzzle of this kind of antibiotics i. e., the magnitude of the unwinding and extenstion of the DNA helix associated with the binding of quinoxaline antibiotics is diminished almost in parallel with that of ethidium as the medium ranges from moderate to high ionic strength91. Finally, like another peptide ring bearing antibiotic actinomycin D, quinoxaline antibiotics generally exhibit a broad GC base pair preference in binding to DNA with different binding constants95.

Historically, it was generally accepted that the primary pharmacological target of DNA intercalating drugs was indeed DNA, although important information on the mechanism of drug action was still lacking. Later, more and more observations |—L-Ala—N(CH3)jTHCO Y-

|j ^ ^ N*r'D-Scr 2 X 0

U-JN L I—Y ---- N(CH3)CHCOl —L-Ala—1 I

General structure of quinoxaline antibiotics Echinomycin: X=-CH2-S-CH(SCH3)-, Y=L-N-methyl-valine.

L-Pro L-Pro L-N-MeVal L-N-McVal D-Val D-Val

L-Thr L-Thr

CH30 . ^ n h s o 2c h 3

XJ

Actinomycin D m-Amasacrinc

OCH, J ^ n h s o 2c h 3

c o 2h

CH2 CH

o-Amasacrine Norfloxcin: a quinolone antibiotic

Figure 10. Chemical structures of some DNA intercalating agents. 30 indicated that the mechanisms of the action of these drugs are more complicated than previously thought. First, it became clear that strong DNA binding ability alone was not sufficient for a drug to have powerful biological acitivity. A striking example of this is the comparison of the properties of m-amsacrine (mAMSA) and o-amsacrine

(oAMSA). The small change of the methoxy group position (Figure 10) does not vary the DNA binding characteristics of these two compounds. Nonetheless, only mAMSA exhibits the biological effects. Further experiments demonstrated that ADR could exercise its cytotoxic effects without entering the cells96. Finally, it was found that in many cases DNA topoisomerase II was associated with DNA breaks in cells treated with DNA intercalating drugs97. Now, it is clear that many DNA intercalating agents attack DNA topoisomerase II or the complex of DNA and DNA topoisomerase II instead of DNA itself. Recently, a cooperative quinolone-DNA binding model has been proposed to explain the working mechanism of quinolone drugs98. According to this model, topoisomerase bound to DNA induces a specific quinolone binding site in the relaxed DNA substrate in the presence of ATP. The drug molecules bind this single­ stranded DNA pocket produced by the enzyme so as to block enzyme processing.

II.2 Photoinduced Electron Transfer

All the DNA drugs that have been discussed hitherto can somehow bind with

DNA causing the disruption of DNA processing. In most cases, except for the drugs of the BLM and NC families, the DNA molecules as chemical structures remain intact.

After binding to DNA, and unlike other drugs, members of these families can produce reactive oxygen species, which lead to the breakage of the nucleotide polymer chains.

Actually, electron transfer (ET) reactions, as well as radiation, are known to yield radicals and hence may cause DNA damage as well. 31 With the elucidation of the complex structure of photosynthetic apparatus and the acceptance and experimental confirmation of Marcus theory, research on photoinduced ET has reached its culmination over the past few years. At present, the structures of the molecular electronics of photoreaction centers are known to atomic resolution". The information of these structures is serving as an invaluable model in studying ET in both biological and non-biological systems. To date, photoinduced long- lived charge separation has been achieved in synthetic molecular devices100. In addition, a photoconverter using dye-sensitized colloidal TiC >2 films has also been developed to generate electricity with good efficiency101. In the following sections, some fundamental aspects of ET will be briefly stated, followed by a survey of recent advances in radical reactions related to nucleic acids.

II.2.1 Theoretical Aspects of Electron Transfer Reactions

Photoinduced ET is a process in which an excited-state species is quenched by an electron jumping from one reactant (the donor) to an unoccupied or a half-occupied orbital of the other reactant (the acceptor). The excited-state species can be either the electron donor or acceptor. Photoinduced ET between two uncharged species leads to the formation of a radical ion pair or charge-transfer complex. Unlike energy transfer quenching, the range of effectiveness of ET is usually limited to a distance of less than

10 A102.

A complete ET process includes orbital jump or resonant charge transfer and nuclear vibrational fluctuations. The former has a time span on the order of femtoseconds (fs, 10*15s), and the latter takes -0.1-1 picosecond (ps, 1 0 12s), and thus is the rate-determining step in ET. On the other hand, a typical collision between two uncharged small organic molecules has a duration of about 0.1-1 nanosecond (ns, 10'

9s). Consequently, during the lifetime of a typical collision, quenching may take place within the “encounter complex”, depending on the driving force of the ET. The product 32 of this collisional quenching, if it happens, is a contact ion pair (CIP) or a solvent- separated ion pair (SSIP, Scheme 4). The latter occurs only if the two reactants of the encounter complex is slightly separated by solvent molecules during the ET process.

Following the ET, the CIP may be rapidly stabilized by solvent molecules to form a

SSIP, which may further separate to generate free radical ions103. But all these processes are reversible. The outcome of these successive events are subject to the regulation of the polarity of solvent and many other variables (vide infra).

D + A -r- D*/A • [D*A- DtAT] Encounter complex ^ Exciplex

CIP SSIP Free radical ions

Sheme 4. Mechanisms of electron transfer reactions. For simplicity, D is treated as the exclusive electron donor and A the exclusive electron acceptor, and only D is excited during the whole reaction span.

If the interactions between the donor and the acceptor are strong, an encounter complex can be swiftly assembled to give a relatively stable intermediate, termed an exciplex (Scheme 4), which is characterized by strong binding energy (~5-20kcal/mol), partial charge transfer character and a large dipole moment. Planar organic molecules are particularly capable of forming exciplexes, with essentially unit efficiency in most cases104. The stability of exciplexes has been interpreted as the result of the overlap between the lowest antibonding orbitals or between the highest bonding orbitals of the reactants105. A quantum mechanical description of the exciplex is given as the linear combination of all possible wave functions:

¥ = c,\j/(D*A) + c 2\j/(D A * )+ c3y(D +#A -) + c4\|/(D -A +#) + c5y(DA) (1)

where the coefficients pertain to the relative contribution of each state. The first two terms indicate locally excited states, the third and fourth radical ion pairs, and the fifth the ground-state complex. When A is an exclusive electron acceptor and ground-state interactions are negligible, eq 1 simplifies to

¥ = c , y (D* A) + c 2 V(D A*) + c 3\j/(D +*A“ * ) (2)

If cj or C2 » c3, the ET does not proceed, and the formation of the exciplex only facilitates energy transfer. In such a case, the exciplex tends to emit light or the two components of the exciplexes simply drift apart and return to the independent ground states, thereby converting the excess energy into heat. Alternatively, if c3 » C 2 and cj, the exciplex has pronounced charge-transfer character and is likely to dissociate into a radical ion pair, especially in polar solvents. Sometimes, a CIP, whose fate has been discussed earlier, is a more usefull picture to describe the structure of the exciplex. In extreme cases, it is possible that the exciplex is bypassed in a polar solvent and ET directly forms an SSIP, i. e. collision quenching. The SSIP then rapidly collapses to inversely generate the exciplex106.

The proceeding discussion is valid only if the ET process between an excited- state sensitizer and a quencher is thermodynamically feasible. For a bimolecular ET in solution, the standard free energy change (AG°) is given by 34

where IPd is the ionization potential of a donor, EAa the electron affinity of an

acceptor, hv the excitation energy of a sensitizer (either donor or acceptor), e the

electron charge, tq and ta are the radii of the donor and acceptor respectively, e is the

solvent dielectric constant, and d is the distance between the radical ions formed. The

fourth item in eq 3 represents solvent stabilization effects and the fifth the Coulombic

interactions of the radical ion pair. Replacing the ionization potential of the donor and

the electron affinity of the acceptor with their redox potentials in solution, E(d+/d> and

E(a/a')» respectively gives the Weller equation107:

AG°(kcal / mol) = 23.06[E - e 2/(ed)]-hv ■(D+’/D) E (A/A-) (4)

In polar solvents, the Coulombic energy term can be neglected. For example, the

Coulombic energy in acetonitrile is less than 1.3 kcal/mol at separation distances of ion

pairs exceeding i k .

Traditional transition-state theory describes the kinetics of a chemical reaction

in terms of the atomic nuclei of reactants moving along a potential energy surface to the

transition state (an energy maximum) in the most economical way, followed by the

spontaneous collapse of the transition state into the products. In contrast to such an essentially adiabatic process, the probability of ET process between distant, weakly coupled donor and acceptor is dictated by the nonadiabatic description of Fermi’s

Golden Rule108:

k« = 4,i% |vr |2fcwd (5) 35

where kct is the rate of ET, h Planck’s constant, |VR| the coupling matrix element of

reactant and product electronic wavefunctions, and FCWD the Franck-Condon

weighted density of states. |VR| depends on the spatial overlap of donor and acceptor orbitals, which falls

off exponentially with the edge-to-edge distance (R) of the donor and acceptor

|VR|2 =|V 0fe x p (-p R ) (6)

In this case, |V0| is the maximum electronic coupling, i. e. the electronic coupling

when the two partners are at the van de Waals contact distance. The damping factor, P,

reflects the nature of the intervening medium. The magnitude of P has been calculated

to be ~2.8A** for the through-space electron tunneling mechanism. Experimental

measurements give a P value ranging from 0.9-2.0A-1 for ET via superexchange

through orbitals of the spacer109. ET in proteins has a fairly uniform p value of 1.4 A '1

110. In addressing the fact that a simple n-alkyl bridge mediates ET through-bonds more

efficiently than a polynorbornyl bridge does, two mechanisms, “cross-talk” and “cross­

links”, have recently been proposed to be responsible for the destructive interference to

electronic coupling in the polynorbornyl dienes111.

On the other hand, the FCWD term in eq 5 contains the dependence of kct on

the free energy (AG°) and the reorganization energy (A.), which, in turn, is the sum of

the reactant (A^) and solvent (A,s) inner sphere reorganization energies,

FCWD = (47tA.kT)”°'5 exp[-(A G ° + A.) 2/(47tA.kT)] (7)

where k is Boltzman’s constant, and T is the temperature. The value of A,j may be calculated from the force constants for all the molecular vibrations in both the reactant 36 and product, while X$ can be determined by application of the dielectric continuum

model of a solvent112. The influence of the solvent on the rate of the ET depends

strongly on the distance (R) between the donor and acceptor in the intermediate

distance regime, as well as both the optical (eop) and static (es) dielectric properties of

the medium, as shown in eq 8:

(8)

where Ae is the charge transferred, and rp and ta are the radii of the donor and acceptor,

respectively. According to eqs 7 and 8, the overall distance dependence of ?i* results in

two different regions for ET in polar media, resting on the relative magnitudes of the

free energy change and the reorganization energy. One is the normal region, where X >

-AG°, in which a monotonic exponential decrease is expected in the rate of ET as a

function of increasing distance. The other is the so-called “inverted region”, where X <

-AG°, in which the rate of ET increases weakly at short distances and is succeeded by

an exponential decrease as the distance continues to increase after passing the

maximum point113. Indeed, the inverted region for ET has been observed experimentally114.

At present, complete quenching of an excited sensitizer is by no means difficult.

Compared to the ground state reactants, however, the products formed from ET reactions possess great free energy and are thus likely to undergo reverse reactions. If this happens, the net effect of these processes, like many other means for deactivating an excited-state sensitizer, will end up with the transformation of light into heat or another form of light, an all too common occurrence. To circumvent reverse ET is, therefore, of pivotal importance for the utilization of the ET products, as well as for the utilization of solar energy. 37 As far as preventing the thermodynamically favored back ET, two quantum

mechanical contributions which kinetically retard this process come into consideration

immediately. They are electron spin and the rate of ET115. Since electron spin is

conserved during the ET process, and if the excited state is a triplet, basic quantum

mechanics guarantees that the reverse ET to the ground state is forbidden until the spins

dephase. The time for electron spin dephasing depends on the mechanism of coupling

to other sources of angular momentum and on the strength of the interaction of the two

electron spins. The spin coupling is strongly distance dependent. The longer lifetime of

the triplet state allows the reactants time to diffuse apart and hence enhances the yields

of products. On the other hand, the rate of ET via both electron superexchange and

electron tunneling decreases exponentially with the distance between the donor and

acceptor, as discussed above. Upon bringing the reactants closer, the rate of forward ET

becomes faster, as does the reverse ET, which can even surpass the rate of forward ET. * Explicitly, there will be an optimum distance for the yields of the products in ET reactions. Several strategies have been advanced and exercised to deter the back ET115.

Based on the mechanisms of the ET process presented earlier, it is evident that the nature of the medium can have great influence on the quantum yield of photoinduced ET, as well as on its rate. Solvent polarity affects energy transfer much less than ET and hence the ionic products tend to be more likely formed in polar solvents (Figure 11). As a matter of fact, it is for this reason that the majority of photoinduced ET studies have been performed in polar solvents, such as acetonitrile. As a rule of thumb, an exciplex can dissociate into SSIP in solvents with e>7, provided that the stabilization energy of the exciplex is small116. Finally, it is worth mentioning that another important character of solvents, viscosity, does not exert any impact over the yield of ET, given that there is an absence of external ionic species117. 38

D t+ AT

D* + A

D t+ AT

Solvent Polarity

Figure 11. A comparison of the effect of solvent polarity on ET and energy transfer.

A straightforward way to circumvent the back electron transfer utilizes a fast electron relay device, in which a succession of electron transfer steps swiftly transport the electron far away from its hole. Each transmitting step relays the electron forward at the expense of an energy loss (AEj). According to the energy gap rule, the rate of return to the ground state is disfavored by the Boltzman factor, exp(-IAE;/kT)118, in addition to any applicable distance, orbital and symmetry effects. In principle, it is not required for each middle electron acceptor to have a more positive potential than its electron donor. Nonetheless, this kind of design is seldom adopted practically, as a faster back

ET of this step can greatly impede the desired electron flow. Because of the exponential dependence of the ET rate on distance, direct ET can be essentially impossible without the intermediate transmitting steps regarding the distance between the final locations of the electron and the hole. Due to the energy requirement for driving electron flow forward, a serious limitation of this strategy arises from the conversion efficiency of light to chemical potential119. The photosynthetic reaction center is exemplary in taking 39 on these stepwise ETs120. An artificial pentad assembled according to this strategy is

able to produce charge separation with a lifetime of 55p.s100a. Similarly, attaching dye-

sensitized transparent polycrystalline film to a modified semiconductor electrode

affords long-lived charge separation100b.

In order for an ET reaction to proceed effectively, another strategy is to separate

the resulting ion pair from each other to deter back ET. The rate constant for the

dissociation of exciplex into radical ions ranges from 5x108s*1 in polar solvents to less

than 106s*1 in nonpolar solvents. The Coulombic effect can be exploited to hasten the dissociation of the ion pairs and elongate their lifetimes121. For example, benzophenone

(BP) as a triplet oxidizing sensitizer is capable of photo-oxidizing leuco crystal violet

(LCV) with a quantum yield (l c v + 0 of only 0.026. Introducing a quaternary ammonium ion to the para position of BP can raise the quantum yield as high as 0.420 for the oxidation of LCV. As depicted in Scheme 5, the Coulombic attraction between the BP radical anion and the LCV radical cation favors the back electron transfer. On the other hand, the Coulombic attraction between the BP radical anion and the quaternary ammonium ion stabilizes the radical anion. However the Coulombic repulsion between the LCV radical cation and the quaternary ammonium ion of the benzophenone repels the radical cation. When a similar experiment was carried out on anthracene or pyrene derivatives (quaternary ammonium and sulfate salts) and methylviologen (MV2+), it demonstrated that the Coulombic interactions between the ammonium ion and the reactants and products retard both forward and backward electron transfers, with the latter process suppressed to a greater extent122. 40

------► Products

Scheme 5. Schematic representation of the Coulombic interactions on the photoinduced electron transfer between sensitized benzophenone derivatives and leuco crystal violet (LCV). Dotted arrow represents attraction and the solid arrow repulsion.

Another effective strategy used to increase the yield of ET is to destroy the charge formed immediately following ET. For example, either protonation of the radical anion and/or deprotonation of the radical cation can competently suppress the back ET. One reaction of importance in organic synthesis and one of the most extensively investigated photoinduced ET reactions utilizes a tertiary amine as electron donor123. This sequence is effective not only because tertiary amines generally have low ionization potentials, but more importantly because proton transfer competes effectively with back ET after the tertiary amine loses an electron in the primary ET step124. In fact, back ET is the exclusive pathway after tertiary amines wich lack a- protons quench excited states. Ethylenediaminetetraacetate (EDTA) and diethylglycine are more effective electron donors with respect to other tertiary amines. The following fragmentation reaction is assumed to be responsible125.

R2N-CH2C02" + A* ► R2N-CH2C02- + AT ► R2N-CH2 + C02 + AT 41 The use of interfacial reaction systems such as micelles, vesicles, membranes

and colloidal particles to separate the oxidation and reduction sites has also been proved

to be effective, but these examples will not be discussed here.

II.2.2 Radical Reactions of DNA Bases

DNA damage can originate at one of the two possible sites: nucleic bases or

sugar residues. Due to the electronic structures of the nucleic acid bases, they are likely

to be subject to both one-electron oxidation and reduction. On the other hand, damage

initiated at a sugar moiety is generally caused by free radical abstraction of hydrogen

atoms, particularly at the C’-4 position as mentioned earlier67. It is important to keep in

mind that nucleic acid bases are by far more susceptible to primary attack o f radicals

than are sugars. It is often believed, especially in the PDT community, that the

hydroxyl radical, the most destructive agent towards DNA, owes its potency to its

ability to abstract hydrogen atoms from sugar residues which results in DNA strand

cleavage. In fact, hydroxyl radical damage to DNA occurs largely because it acts as a

superb oxidation reagent for DNA bases (see below in detail) rather than because it

withdraws hydrogen atoms from sugar residues. The radicals resulting from the reaction

of hydroxyl radicals with base residues are responsible for DNA chain

fragmentation126. For example, 93% of hydroxyl radicals generated in the presence of polyuracil add to the uracil residues of poly(U) and only 7% abstract hydrogen atoms from the sugar residues127. Therefore, it is more appropriate to assume that the mechanism of hydroxyl radical damage of DNA is such that the radicals of the base residues transfer the radical centers to the neighboring sugar residues through the hydrogen atom abstraction and the sugar radicals then direct the polymer chain break­ down following the mechanisms of Scheme 3 as presented in section II. 1.2. In Figure 42

12 it is shown how a base radical (-TH, the radical of the protonated thymine anion

radical O'*")) transfer the radical center to the nearby sugar residue.

O

Figure 12. Schematic drawing to show the hydrogen abstraction by the radical *TH from sugar residue.

It was noticed as early as 1950 that the fluorescence of acriflavine can be quenched by DNA128. This ET process between the excited singlet state of a dye and the ground state of nucleic acid bases has been extended to many dyes responsible for fluorescence quenching since the original observation129. The effect of this ET process on DNA was generally thought to be trivial because the fast back ET reaction outpaces many other possible secondary reactions. As a result, even though dye-sensitized DNA cleavage has long been known, it is often incorrectly attributed to the effects of reactive oxygen species, and the results obtained under the deoxygenated conditions were left unexplained130. Actually, the primary one-electron oxidation and reduction products of nucleic bases undergo deprotonation or protonation very quickly. These reactions can, 43 but often inefficiently, compete with back ET in many cases. For example, the

photosensitization of strand breaks in DNA by methylene blue has a maximum

quantum yield of 3xl0‘7 ,30d. To inhibit the back ET, it has been reported recently that

using ethidium bromide (Et) as an intercalator and MV2+ as an externally bound

cosensitizer, irradiation pBR322 DNA with 488nm light elevated the rate of single­

strand breakage by 10 fold compared to that using Et alone131. The selective cleavage at

guanosine residues backed the proposed reaction mechanism, /. e. photoinduced ET

creates guanine radical cation species, which in turn delivered the radical center to

DNA backbone sugar causing single-strand breakage of DNA.

In one-electron oxidation reaction purines react faster than pyrimidines due to

the electron-rich imidazole ring. The electron-donating groups on the purine ring, such

as the amine and hydroxyl groups, enhance the reaction rates even further. For example,

adenine and guanine react with SCV- with second-order rate constants of 2.7xl09M-1S'1

and 4.1xl09M-,s*1 at pH 7 respectively132. These results are in accord with their oxidation potentials133. In addition, the gas-phase IP of guanine is significantly lower than that of adenine (ca. 7kcal/mol) and both purine bases have substantially lower IPs than the two pyrimidine bases134. Accordingly, one-electron oxidation damage to the non-guanine bases is quickly repaired by ET from guanine. This is followed by losing a proton from the resulting guanine radical cation (G-+) to form deprotonated guanine radical (-G(-H)). In dideoxynucleotide GpA, the intramolecular ET from the guanine to the adenine radical cation (A-+) takes less than 50ns, corresponding to the rate constant

> 2xl07s_1. The initial 70% of G + and 30% of A-+ from the photoionization of GpA forms 95% of -G(-H) radical135. Since the driving force for ET from purine bases to pyrimidine bases is even larger, the authors further speculated that adenine can serve as a relay for ET from guanine to pyrimidine radical cations, if a direct ET pathway is not available. Similarly, it has been reported that G + accounts for more than 90% of radical cations in y-irradiated DNA136. 44

The S04- reaction with adenine leads to the very short-lived radical cation A*+

in aqueous solution. This short lifetime of A*+ is due to its high Br0nsted acidity leading

to a rapid deprotonation reaction. The pKa value of A-+ is estimated to be less than 1132.

Compared to the pKa (>13.75) of the parent compound, the driving force, free energy

change, for the deprotonation reaction is larger than 18kcal/mol! Only if the chemical

structure makes the deprotonation reaction impossible, does the corresponding radical

cation become quite stable, exhibiting a small tendency for hydration (kclO V 1) in the

absence of an electron donor137. The product of the deprotonation reaction of A-+,

deprotonated adenine radical (-A(-H)), centers its unpaired electron on the N-6 atom of

the adenine, resulting in an anilino type radical. When both hydrogens on the N-6 atom

of adenine are replaced by methyl groups, deprotonation takes place at the N-9 atom.

The anilino radical is a very strong one-electron oxidant and can react with antioxidants

readily, wherein the adenine sysytem is repaired. On the other hand, when a hydroxyl

radical acts as an oxidant, adenine produces two types of radicals: -A(-H) and a ring-

opening radical, via an addition-elimination mechanism138 (Scheme 6). The majority of

the hydroxyl radicals add to the adenine residues at the most elelctron-rich C-4 double

bond to give a radcial, which is represented by the four-resonance structures of Scheme

6. This radical undergoes dehydration to generate -A(-H). The same radical is also produced from the deprotonation of the primary product of other one-electron oxidation. The dehydration reaction can, however, be slowed by increased acidity of the medium. Additionally, the hydroxyl radical also adds itself to adenine at the C-8 atom producing an isomeric radical, which, in the absence of extra oxidant or reductant, can undertake a ring-opening transformation under the catalysis of either acid or base.

Guanine is the most easily oxidized of the nucleic acid bases and has been studied exhaustively. Like adenine, the primary product of one-electron oxidation of the guanine residue, G +, deprotonates to offer the corresponding neutral radical -G(-H) with the unpaired electron centered on the oxygen of C-6. Under physiological 45

R * A(-H)

+

NH NR NH. NH.

OH N OH N OH

R R

+H+ or OH'; -H20

NH .OH Minor

i R Ring-opening radical

Scheme 6. Mechanism of hydroxyl radical reacting with adenosine. 46 conditions, -G(-H) can be reduced to reform guanine by a variety of reducing reagents such as phenol and thiolates126a. Moreover, -G(-H) can behave either as a Brpnsted acid or a base139 (Scheme 7). Finally, the major products of the decomposition of a guanosine model compound have been successfully isolated and characterized as depicted in Scheme 8140.

0 ? T

n n n n m n H2N A N X N>^ h 2N ^ N ^ nI >'T H2N -^N'^- N n r RR R GH+(pKa=2.4) G(pKa=9.4) G(-H)

-e -e 11 O* 0 ‘

HN I • RR R • RR R R

GKpKa=3.9) • G(-H)(pKa=10.8) G(-2H)T

Scheme 7. Acid and base equilibriations of guanosine and its radicals.

Pyrimidine cation radicals are relatively more difficult to generate compared to their purine counterparts. Even if produced, these species may get electrons back from purine base residues rapidly (vide supra). Consequently, they are of relatively little significance in biological systems. Model studies in the absence of purine bases have shown that the pyrimidine cation radicals are very short-lived. Thus no direct evidence has been obtained for their existence. Nevertheless, optical and conductance detection 47 of radical ions are feasible, when all the hydrogen atoms on heteroatoms are replaced by alkyl groups141. The final decomposition product for the oxidation of both thymine142 and cytosine143 have also been isolated and characterized.

HN

^ Jto* +e,+e. HH+ N

R

HNAj.°NHH -CO, H N ° v ” v H -HCONHj^VyO I JL .A TxA.AAu n=C NH R R I

H20 H2N^ 0>y 0 OH- NlNH _ w n _ ^

n h h 2n^ " nh2 I

Scheme 8. Mechanism for the oxidation of guanosine and the following decomposition of its radical.

On the other hand, the electron-deficient pyrimidine endows the bases of nucleic acids with high intrinsic electron affinity. The added electron-rich imidazole ring, as well as electron-donating amine groups, hardly affects its properties in reaction with e* aq. Unlike the one-electron oxidation reaction, one-electron reduction reactions of 48 purine and pyrimidine bases and nucleosides thus proceed at approximately the same

rate constants (reacting with e‘aq with second-order rate constants of about lO^M^s*1),

which are virtually independent of their individual structure. However, the negative charges of the phosphate residues of nucleotides do reduce the reaction rates in a

multiplicative way, due to electrostatic repulsion. Hence DNA reacts with e'aq which a rate constant as low as 108M‘1s‘1 per nucleotide unit.

As pointed out above, the guanine residue is the dominant choice for the cationic site in the oxidation reactions. But both cytosine and thymine have been proposed as anionic sites in reduction reactions. Thermodynamically, thymine has greater electron affinity than does cytosine in the gas phase. Water has a leveling effect on their electronic affinities due to the higher basicity of the cytosine anion radical (C-*, pK>13) relative to that of T-* (pK=6.9). The reduction potentials of these two bases are thus the same (E=-1.1V/NHE)144. Analysis of the spectra obtained from y-irradiated

DNA in D2 O indicated that the primary anionic sites were a mixture of the cytosine anion ( ca . 77%) and the thymine anion ( ca . 23%)145. In addition, y-irradiation gave a majority of cytosine anions in hydrated double-stranded DNA, whereas a near random distribution of the anion on cytosine and thymine has been found in hydrated single­ stranded DNA146. The authors interpreted these observations as a result of interstrand base pairing and base stacking effects. Base pairing and stacking in double-stranded

DNA facilitate the ET in both intra- and interstrands. However, they are partially disrupted, especially for the adjacent pyrimidines in single-stranded DNA. As a consequence, ion localization occurs in a more random fashion in single-stranded DNA.

The reaction mechanism of purine reacting with e"aq (represented by adenosine in Scheme 9) has now been established essentially in great detail132. Since nitrogen has a greater electron affinity than does carbon, the adenine moiety first captures solvated electrons (which are primarily localized on the nitrogen atoms) at a rate close to the 49

Radical anion

NH. +OH\ -H20; NH. N or +H2Q, -OH~ HN

101.4x10V 1 N I R R

N-Protonated radical

NH. NH.

H + H

R R

C-Protonated radical

Scheme 9. Mechanism of one-electron reduction of adenosine.

upper limit of e‘aq reactions. The radical anion thus formed, depicted in the three resonant structures of Scheme 9, are then subject to rapid protonation (k>1.4xl08s-1) to give an N-protonated radical which can shift among the three tautomeric forms. The N- protonated radical is a strong one-electron reductant. It is likely for it to donate an electron thereby resulting in the repair of the nucleotide, provided the existence of an electron accepting agent in the neigborhood. Alternatively, the N-protonated radical can 50 slowly rearrange to a thermodynamically more stable C-protonated species

(ko'l^xlO'V1) through a 1,2-hydrogen shift mechanism. The conversion of the N- protonated radical to the C-protonated radical was found to be subject to general base catalysis and the rate reaches a plateau value of 3.6xl06s-1 at pH 13. These two types of radicals can be distinguished spectrophotometrically (the N-protonated one with a

Xmax=320nm in its absorbance spectrum, and the C-protonated one with a

^max=355nrn)147.

With respect to the other nucleic bases, guanine is unique in its one-electron reduction reactions. It is expected to be the last base to gain an electron, but the isomerization to the C-protonated radical from the resulting N-protonated radical should be the fastest and irreversible. In fact, this type of reaction for guanine in a neutral unbuffered medium is about 100 times faster than that for adenine and about

1000 times faster than that for thymine148. Despite the relatively rapid C-protonation reaction for guanosine, a C- or N-protonated electron adduct of guanine has never been detected in irradiated DNA. It is believed that intra- and interstrand ET from purine radical ions proceeds much faster than the irreversible protonation reactions.

II.2.3 Reagents for DNA cleavage

The most commonly employed agent to initiate DNA cleavage is the hydroxyl radical generated chemically or photochemically in situ. Redox-active metal ions such as iron and copper can catalyze the reduction of hydrogen peroxide generating hydroxide ion and the highly reactive hydroxyl radical.

Mn+ + H20 2 -» M++ OH- + -OH

The resulting hydroxyl radical reacts either with nucleic bases or deoxyribose of DNA leading to the fragmentation of the DNA helix. The aforementioned BLMFe(II) is an 51 elaborately evolved reagent which works by this mechanism. This metal ion mediated reduction of hydrogen peroxide, generally termed the Fenton reaction, is believed to play a central role in oxygen toxicity149.

EDTA-Fe(II) is an efficient mediator of the Fenton reaction150. It does not form a complex with DNA in its scission reaction and therefore the DNA cleavage is not selective. On the other hand, its compact structure restricts its reactivity with the hydroxyl radical that it generates. It cuts evenly at all sequence positions as long as the substrates are solvent accessible. Hence, it has been widely used for “hydroxyl radical footprinting”. The lack of sequence selectivity has been overcome by the innovative work of Dervan and his colleagues by tethering EDTA-Fe(II) to some known molecular recognition carriers151.

Prigodich and Martin152 studied the reaction of hydroxyl radical generated with

EDTA-Fe(II) complex with single stranded oligodeoxyribonucleotides. They observed damage to the DNA which could not be visualized by electrophoresis until piperidine treatment, implying that the hydroxyl radicals react extensively with nucleic bases of

DNA.

Another extensively investigated DNA cleavage reagent of this kind is the 1,10- phenanthroline-copper complex153. The 2:1 1,10-phenanthroline-cuprous complex binds first to DNA in part by intercalation. The DNA-bound complex takes an orientation in which the copper ion is accessible to the C’-l hydrogen of the deoxyribose in the DNA minor groove. The second step is the oxidation of the DNA- bound cuprous ion by hydrogen peroxide resulting in a copper-oxo species. The copper oxene or a copper-coodinated hydroxyl radical as the reactive species is directly responsible for initiating the cleavage reaction. B-DNA is assumed to be the most susceptible secondary strucutre to this copper complex. Similarly, The 2:1 1,10- phenanthroline-cuprous complex has also been linked to a variety of DNA recognizing side chains to achieve sequence specificity. 52

In addition to EDTAFe(II) and 1,10-phenanthroline-cuprous complexes, other

chelates capable of cleaving DNA under physiologic conditions include

metalloporphyrins, uranyl acetate, and many octahedral metal complexes154. These

reagents initiate DNA scission by serving either as photosensitizers for the formation of

singlet oxygen or as agents for hydrogen atom abstraction155. Recently. Schuster and

coworkers130®*156 have applied cationic anthraquinone derivatives as DNA cleaving

reagents. It has been proposed that electron transfer from a nearby base and hydrogen

abstraction from the deoxyribose residue of DNA by excited-state quinone is

responsible for the DNA cleavage.

II.3 Overview of Psoralens

Psoralens, a class of linear-fused furocoumarins (Figure 13), have been used to

treat human skin diseases since ancient times157. Musajo first reported that psoralens

reacted with DNA in 1965158. Soon, he and his coworkers isolated and characterized

the products formed by irradiation of frozen solutons of psoralen and the DNA

components thymine and cytosine159. In the ensuing years, psoralens have attracted

great attention both in basic reseach and in clinical studies160. To date, it is well established that psoralen derivatives bind to DNA by intercalative interactions. Upon irradiation with UVA, psoralen molecules undergo consecutive [2+2] cycloaddition reactions. Cross-links of psoralens with complementary strands of DNA thereby retard cell division. Due to its ablity to modify nucleic acids, psoralens has proved useful in the elucidation of nucleic acid structure and function. As a matter of fact, three members of this family of compounds, 8-methoxypsoralen (8-MOP), 5- methoxypsoralen (5-MOP), and 4,5’,8-trimethylpsoralen (TMP), are now in clinical practice for the photochemotherapy of psoriasis and vitiligo161. Moreover, many psoralen compounds have been tested to be powerful agents for inactivating various bacteria and viruses160, including HIV162. 53

6(3')

7(2

Figure 13. Chemical structure of psoralen

As photomedicines, psoralens are usually ingested orally 2~3hr prior to irradiation with near ultraviolet (UVA) light. Studies have shown that only 1~2% of the ingested drug arrives in the patient’s blood at the time of irradiation treatment163, and more than 95% of the ingested drug is excreted within 24hr164. The biological effectiveness of psoralens have generally been attributed to the formation of psoralen photoadducts (both monoadducts and crosslinkings) with thymine residues of DNA165.

However, other cellular components such as lipid membranes and proteins have also been suggested as alternative or co-targets166.

Psoralens interact with DNA in the dark by intercalation of the planar molecules into the double helix of DNA. This mode of physical interaction has been substantiated by the cis-syn stereochemistry of the isolated DNA photoadducts167, fluorescence anisotropy measurement168, and nuclear magnetic resonance (NMR) studies169. The intercalation occurs preferentially in regions having an alternate sequence of Pu and Py, particularly in the alternate sequence of adenine and thymine166 17°. Like all DNA intercalating drugs, 8-MOP at high concentration has demonstrated an ability, albeit a poor one, to inhibit DNA synthesis in the dark171.

Psoralens react slowly upon photolysis with free thymine in aqueous solution to give a [2+2]-cycloadduct. In line with the spectroscopic studies, the reactions produce the product, via the psoralen triplet excited states, from the cycloaddition of the C=C 54 bond of the pyrone moiety of psoralens to the 5,6 C=C bond of thymine172. To support the triplet excited state mechanism, oxygen and paramagnetic ions have been shown to quench the photochemical reaction173. The UV-vis spectra of psoralens have a band at about 300nm and a shoulder at 340-360 depending on the solvent. The shoulder at longer wavelength becomes more obvious in a less polar solvent. Both bands have been attributed to (n, n*) absorptions based on their lifetimes, polarization studies, triplet- triplet absorptions, as well as theoretical calculations174. The singlet excited states of psoralens are generally extremely short-lived and emit a structureless fluorescence at

380-520nm, depending on both the properties of solvent and substituents. The phosphorescence of psoralen derivative occurs at 450-600nm. The frequencies of the 0-

0 phosphorescence bands of psoralen derivatives are essentially independent of the substituents or the extent of conjugation, implying that the Tj(7t, tc*) state is localized in the C=C bond of the pyrone moiety. Indeed, the experimental results have been correlated with the electronic characteristics of excited psoralens.

Like many nitrogen heterocyclic and aromatic carbonyl compounds, psoralens possess a lowest energy (n, 7t*) state that is close in energy to the lowest (7t, n*) state.

The close proximity of these two states leads to very efficient radiationless decay from the lower of the two excited states. This proximity effect (pseudo-Jahn-Teller effect175) is believed to be a consequence of vibronic interaction between the two closely spaced states. The vibronic interaction between the two interacting states offers the fast radiationless decay of the Si state to two lower states, Tj and S„ Consequently, these compounds generally have a relatively high ratio of phosphorescence to fluorescence yield (p/<|>f). Additionally, theoretical investigations indicate that the proximity effect is greater the larger the energy gap for the radiationless transition, i. e. since the Si —>S0 electronic energy gap is much greater than that of Si-»Ti intersystem crossing. Thus, the Sj —»S0 radiationless transition can become the dominant pathway for the decay of excited state. Futhermore, the out-of-plane bending interaction between the two 55 interacting states can be thermally excited in a condensed phase. Often the fluorescence and phosphorescence emissions of these compounds are too weak to be observable in hydrocarbon solvents at elevated temperature.

T(7C, 7t*)

Solvent Polarity

Figure 14. A possible energy diagram of 8-MOP in different solvents. Solid anrows represent absorptions and emissions, dotted arrows internal conversions

Based on the observation of large Stoke’s shifts, the relatively big value of

p/f, and the extremely short lifetimes of the singlet excited states, it has been suggested that an S 2 (n, n*) state is only slightly above the S\(n, 7t*) state. The energy level ordering of the T 2 (n, k*) and Si(7t, 7t*) states is reversible in different solvents, i.e

ET(n, ji*)>Es(ji, it*) in polar solvent; ET(n.n*)

Therefore, the proximity effect plays a substantial role in the photophysics and photochemistry of psoralens. The absorption spectrum of the So-S 2 (n, n*) transition, however, is probably buried under the (it, 7t*) absorptions and has never been observed. 56 The intercalation complexes of psoralens with DNA are converted into covalently bound complexes through [2+2]-cycloaddition reactions upon UVA irradiation177. Unlike the reactions of psoralen derivatives with free pyrimidines, the addition of pyrimidine to the 4 \5 ’-double bond of the furan ring of psoralen derivatives generally dominates over that of pyrimidine to the 3,4-double bond on the pyrone ring.

This has been interpreted as follows. Intercalation brings psoralens not only into close association with DNA bases, but also in a face-to-face orientation that is perfect for cycloaddition. The cycloaddition reaction thus can happen at the furan moiety of psoralens via the singlet state178, despite its very short life span179. Evidence has led to the suggestion that the formation of a fluorescent 4’,5’-monoadduct proceeds via a singlet state reaction. Laser flash photolysis (LFP) studies of 4’-aminomethyl-4,5’,8- trimethylpsoralen (AMT) demonstrated that no triplet-triplet absorption was observed in the presence of calf thymus DNA180. There are two possible reasons for the inability to detect the AMT triplet-state transient in the presence of DNA: (1) it is possible that triplet state formation is suppressed by fast photoreaction via the excited singlet state, when AMT is complexed with DNA, or (2) that the reaction does indeed occur via the triplet state, but that the AMT triplet bound to DNA is quenched too rapidly to be observed by nanosecond flash photolysis. As the resultant 4 \5 ’-monoadduct retains a strong absorption of light above 320nm, an interstrand cross-linked diadduct can be generated through another [2+2]-cycloaddition of the 3,4-double bond of psoralen residue with a pyrimidine base by the absorption of a second photon, given that the two reaction components are appropriately positioned159 (Scheme 10).

On the other hand, the addition between pyrimidine base and the 3,4-C=C bond of psoralens forms a non-fluorescent monoadduct. This 3,4-monoadduct does not absorb light above 320nm and thus further reactions which cross-link DNA by psoralens are impossible (Scheme 10). In fact, even if it absorbs a photon of short wavelength, the 3,4-monoadduct is predicted to be unreactive with respect to a second 57

HN

HN HN Sugar hv

Sugar Sugar

H <0

Sugar Sugar y ° M \ Sugar O HN

Sugar

Scheme 10. [2+2]-Cycloaddition reactions of psoralen with thymine residues of DNA.

cycloaddition181. Under normal reaction conditions, the ratio of the cross-linking to monoaddition is subject to the steric effects of substituents at the relevant C=C bonds.

Introducing a methyl group into the C-4 of 8-MOP reduced the yield of pyrone monoadduct from 19% to 2%, while the presence of a methyl group at C-5’ hardly affected the product distribution182. Consequently, it is not suprising that TMP and 58 AMT which have a methyl group at C-4 exhibite higher potency for cross-linking than

8-MOP and 5-MOP183. As a useful method to acquire pure monoadduct, irradiation of

AMT in the presence of DNA at 390nm has been reported to produce the monoadducts

of the psoralen with DNA in high yield184. Subsequent irradiation of the monoadduct

obtained using short wavelength light converted about half of the monoadduct into a

cross-linked diadduct. This means that the long-wavelength irradiation does not

selectively activate any specific C=C bond.

Besides reacting with pyrimidine base, psoralens also react with other nucleic

acid bases. In contrast to psoralen-thymine photocycloaddition, TMP reacts with

adenosine in dry film to form covalent bonds between the C4 atom of the pyrone

moiety of psoralen and the C’l, C’4, or C’5 atom of the ribose ring of the

nucleoside185. The mechanism for this reaction is not yet clear, but it is believed to

proceed via an ET process at the beginning of the reaction. To date, all the reactions of

psoralens with adenosine have been performed in dry film. No direct evidence has been obtained to prove these reactions really happen in a DNA-psoralen complex, not to

mention in cells or viruses. Moreover, compared to adenosine, guanosine is a better electron donor (see section II.2.2), but it has not been possible to acquire any product

from the reaction of guanosine with psoralens even in dry films185. Finally, the involvement of radicals in the reactions of psoralens with DNA has recently led the following proposal of a new reaction mechanism in regard to the cycloaddition between psoralens and thymine moieties of DNA186,

*Ps + A —> [exciplex] —> Ps” + A+' Ps" + T —> Ps—T' Ps—T' + A+- —> P s o T + A A+- + T —> T+- + A T+> + Ps” —> P so T 59

where Ps is a psoralen derivative, A and T are the adenosine and thymidine nucleosides, respectively. P so T represents the cycloaddition products of psoralen derivative with thymine residues.

Several amino acids, particularly tryptophan, can quench the excited states of psoralens. As the quenching with nucleic acid bases, the quenching mechanism of the amino acids is most likely to be ET, for these quenchers have higher energy levels of their first triplet states than that of psoralens. Though a few covalently linked psoralen- protein products have been observed, the products of psoralens reacting with individual amino acids have not been isolated or identified166. The umerous biological side effects of psoralen photochemotherapy have been ascribed to be the results of the covalent connection of psoralens with proteins166, including the recent attribution of the skin phototoxicity of psoralens to the DNA-protein cross-linking by psoralens187.

The ET reactions of psoralens with nucleic acid bases and amino acids have also stimulated the investigations of the primary reaction products, the radical anion intermediates188 The absorption spectra of the radical anions of psoralen derivatives are remarkably similar to that of their triplet state absoiptions in water. A triplet exciplex,

3(psoralen8--H205+), has been proposed to represent the structure of tthe riplet excited state of psoralens in water. Unfortunately, the similarity of the psoralen triplet and radical anion UV-Vis absorption spectra in water has caused great complication in photochemical investigations. The most reliable method to study the radical anions is pulse radiolysis. There are also a few observations of radical anion intermediates produced by LFP of psoralens in the presence of a quencher such as tryptophan189 and

L-dopa190.

By definition191, the aforementioned reactions are often referred to as type I photoreactions, for the excited sensitizers react with substrates directly. Alternatively, the excited triplet state of a sensitizer can transfer its excited-state energy to a nearby 60 ground state triplet oxygen molecule to generate singlet oxygen, which in turn leads to the generation of other reactive oxygen species. The reactions of these reactive oxygen species with cellular components are often called type II photoreactions. The mechanisms of processes involving reactive oxygen species, particularly hydroxyl radical, which lead to biological effects has been discussed earlier and will not be elaborated further. Because of the relatively high yields of psoralen triplet states, it is possible for type II reactions to be involved in psoralen phototherapy192. Early measurements of singlet oxygen yields produced by various psoralens are largely unreliable due to the inappropriate use of singlet oxygen trapping agents193. Probably, the most reliable measurment of the singlet oxygen yields was made by the direct detection of the near infrared luminescence of singlet oxygen194. According to this study, the order of singlet oxygen quantum yields was 3-carboxypsoralen > TMP > 5-

MOP > psoralen > 8-MOP. Since only a triplet state sensitizer can give rise to singlet oxygen, the yield of singlet oxygen should be proportional to the quantum yield of the triplet excited state of a sensitizer192. As mentioned earlier, LFP of psoralens intercalated between DNA base pairs did not give any detectable triplet-triplet absorption. Thus the singlet oxygen damage to DNA sensitized by psoralens must be the result of excitation of free psoralens in solution195. As a result, the physical binding ability of a psoralen with DNA plays an important role in determining the singlet oxygen effects.

The magnitude of the binding constant measures the amount of the sensitizer intercalated into DNA. There are many methods available to determine the binding constant of a sensitizer. By and large, the computation of binding constants is conceptually evident, however the treatment of the experimental data is less than straightforward. The interpretation of the data is often suceptible to bias. Consequently, the measurement of DNA binding constants is fraught with uncertainty, and the data can only be used for comparative purposes. Introducing hydrophobic groups onto the 61 aromatic rings of psoralen reduces its water solubility and increases its binding ability,

because of the hydrophobic interior of DNA helix. Polar, charged groups on the

aromatic rings of psoralens interfere with the intercalative interactions between

psoralens and DNA for the same reason196. The introduction of positively charged side

chains, e. g. ammonium salts, on the psoralen aromatic ring can, however, not only dramatically raises the water solubility of the psoralens, but also increases their binding constants with DNA, presumably because of hydrogen bonding and Coulombic interactions between the ammonium ion and the phosphate of the DNA backbone170*.

The effects of side-chain length on the binding constant has been examined. It was found that an interval of 3 to 5 atoms gives the best results197. Our laboratory has obtained similar results. In contrast, a negatively charged side chain increases water solubility but reduces the DNA binding constant due to repulsive electrostatic interactions. Lastly, DNA conformations affect the binding constants of psoralen derivatives. Generally speaking, psoralens bind better to B-form DNA than to Z-form

DNA.

In conclusion, the close spacing of the energy levels of different electronic states greatly complicates the photophysics and photochemistry of psoralens. In spite of extensive investigations, the subject is still left with many open questions. However, it is clear that, for strongly DNA-binding psoralens, type II photoreaction is not important in damaging DNA. Chapter III

RESULTS AND DISCUSSION

It is possible to damage DNA with free radicals. When a radical center is transferred to a sugar residue of DNA by the abstraction of a hydrogen atom, the resulting radical can initiate a series of fragmentation reactions leading to DNA strand breaks. There exist a variety of means, including ET, to generate radicals, but few ET processes have been able used to cleavage DNA. It occurred to us that a free radical can be generated by light by installing one or more functional groups, particularly bromine and iodine, on a sensitizer complexed with DNA. Such a sensitizer could be an efficient agent with which to cleave DNA strands.

Voelk has isolated some products resulting from an aromatic radical upon the irradiation of an iodide derivative of acridine. He also observed sensitized photocleavage of CpG when this compound was intercalated into a (CpG )2 mini duplex198. As expected, we found that the acridine iodide compound cleaved DNA and inactivated X phage viruses more efficiently than its non-iodinated counterpart.

Although this compound will not be capable of any practical use in blood banks due to the high toxicity of acridine compounds towards biological systems, its study provided a basis for evaluating and improving the design of new DNA cleaving agents. Psoralen azides were then prepared and evaluated for DNA nicking ability. Unfortunately, it soon became evident that psoralen azides did not possess the capability of either 62 63 labelling DNA or causing DNA strand scission. Efforts were then turned to the synthesis of a series of halogenated psoralen derivatives. The results from these halogenated compounds, especially from those with bromine atoms on the aromatic ring, were extrmely encouraging. Some of these compounds are much better than AMT, the best psoralen derivative commercially available, both in DNA nicking and in viral inactivation of X phage. In addition, it will also be presented in this chapter some interesting results from mechanistic investigations of these sensitizers.

III.l Synthesis

3-Amino-6-iodoacridine (Acr-I) has been synthesized by Martin199 et al employing the procedures shown in Scheme 11. One of the two amino groups of 3,6- diaminoacridine (Acr-NH 2 , proflavine) was protected by mcnoacetylation with acetic anhydrous in acetic acid. The monoacetylated compound was then diazotizated with nitrous acid produced in situ, which was followed by nucleophilic displacement with potassium iodide. Due to the strongly acidic conditions, deprotection of the amino group occurred under the reaction conditions to yield the required compound, Acr-I.

AC2O, AcOH NHAc

Acr-NH2

l)NaN02. HC1

Acr-I

Scheme 11: The synthetic route for 3-amino-6-iodoacridine (Acr-I). 64 Following literature procedures, we were able to prepare Acr-I. However the intermediate 3-acetamido-6-aminoacridine was always obtained with impurities of starting material and diacetylation product. With this impure intermediate, the second step of the reaction proceeded in very low yield. Pure Acr-I was attained by extensive purification at the expense of product yield. Actually, the fact that there is not any analytical data for 3-acetamido-6-aminoacridine in literature implies that this intermediate was impure. Upon analysis of the spectra of Acr-I in Voelk’s thesis 198 it was determined that the final product Acr-I prepared by Voelk was not pure. Although the character of our project limited us from intensive investigation into synthesis, some efforts were made to improve the yield. Different protecting groups were tried in the first step: a) mono- and diethoxyformylated Acr-NH 2 could not be separated on thin layer chromatography (TLC); b) Poor solubility in typical solvents limited the use of phthalimide-protected Acr-NH 2 - In the second step, using trifluoroacetic acid as diazotization solvent in place of hydrochloric acid did not improve the reaction outcome as it does in other cases200. Finally, successful purification of 3-acetamido-6- aminoacridine was achieved by carefully choosing solvent mixture (ethanol: ethyl acetate: triethylamine=6:3:l) as flash chromatography eluent. Pure 3-acetamido-6- aminoacridine gave a much improved yield of the second step of the reaction. A careful multistep recrystallization afforded pure Acr-I with an acceptable yield of 37%.

The synthetic route used to obtain 5-azido-8-methoxypsoralen (AMP) and 5- azido-8-(3-diethylaminopropyloxy)psoralen (ADP) is shown in Scheme 12. The demethylation of 8 -MOP was carried with tribromoboron in freshly distilled 1,2- dichloroethane201. The resulting 8 -hydroxylpsoralen was refluxed with at least 5 equivalents of 1,3-dibromopropane in acetone in the presence of excessive anhydrous potassium carbonate to give 8-(3-bromopropyloxy)psoralen202. The resulting compound, as well as 8 -MOP196>203, can be easily nitrated with fuming nitric acid in OCH2R o c h 2r o c h 2r a, b R=H- R=CH2CH2Br —

R=CH2 CH2Br 1 R=CH2 CH2N(CH2CH3)yH

o c h 2r

AMP, R=H; ADP, R=CH2 CH 2 N(C2H5 )2

Key: a) BBr3, C1CH 2CH2 C1, 87%; b) BrCCH^Br, K 2 C 0 3, CH3COCH3, 85%: c) HN03,

HOAc, R=H, 87%, R= CH 2 CH2 Br, 82%; d) SnCl2, Zn, HC1, R=H, 64%, R= CH 2CH2 Br,

84%; e) HN(CH 2CH3)2 , CH3CN, 6 8 %; 0 N aN 03, CF3C 0 2H. 0-5°C; NaN3, R=H, 72%, R= CH2CH2N(CH2CH3)2,64%.

Scheme 12. The synthetic route of 5-azido-8-methoxypsoralen (AMP) and 5-azido-8- (to-diethylaminopropyloxy)psoralen (ADP).

acetic acid and reduced to amines with stanneous chloride in 95% ethanol at room temperature. The low solubility of these psoralen nitrates in various solvents limited the reduction reactions to small scales and required a large excess of reducing agents. In an attempt to facilitate the reduction step of the nitropsoralen, nitration of 8-(3- diethylaminopropyloxy)psoralen under the same conditions was attempted but unsuccessful. A possible expanation of this failure is due to the oxidation of amine by nitric acid. Nitration of 8-(3-bromopropyloxy)psoralen followed by an Sn 2 reaction 66

with either diethylamine or triethylamine gave dark inseparable mixtures. 5-Amino-8-

(3 -bromopropyloxy)psoralen was converted to 5-amino-8-(3-bromopropyloxy)psoralen

by simply refluxing it with diethylamine in acetonitrile under an argon atmosphere. The required azide compounds, AMP and ADP, were finally obtained from their amine derivatives by a standard diazotization reaction followed reaction with sodium azide200.

In Scheme 13 are described the syntheses of some halogenated psoralen derivatives. Of the final products, 8-[3-(triethylammonio)propyloxy]psoralen bromide salt (HEP), 5-bromo-8-[3-(triethylammonio)propyloxy]psoralen bromide salt (BEP) and 8-[3-(triethylammonio)propyloxy]-5-iodopsoralen bromide salt (IEP) were kindly provided by Dr. Park of Cryopharm Corporation.

Several different methods are available to obtain monobrominated 8 -

MOP203*204. Among them are direct bromination with elemental bromine in various solvents and refluxing in the presence of N-bromosuccimide (NBS) in carbon tetrachloride. The NBS method provides a quantitative yield of product without concurrent multibromination. Iodination was achieved with elemental iodine and silver acetate, using a method that is slightly different from the literature procedure20413'205.

Alternatively, iodination with monoiodochloride gave a mixture of mono- and multiiodo products. The related chloride derivative could not be obtained by chlorination with N-chlorosuccimide (NCS) either with or without the addition of concentrated hydrochloric acid as catalyst, but the chlorination of 8-(3- bromopropyloxy)psoralen with NCS in benzene in the presence of hydrochloric acid successfully afforded 8-(3-bromopropyloxy)-5-chloropsoralen. 8-(3-Bromopropyloxy)-

5-halo-psoralens, except for 8-(3-bromopropyloxy)-5-chloropsoralen, were prepared from their corresponding 8-methoxy-5-halo-psoralens by the same procedure for making 8-(3-bromopropyloxy)psoralen as given in Scheme 12. The final water soluble products were readily prepared by reacting appropriate nitrogen nucleophiles with the 67

1) BBr3, C1CH2CH2 C1;

2) Br(CH2 )3 Br, KC03,* Acetone. OCH, OCH2 CH 2CH2Br

NBS -► X=Br x ^ n csjt x=Cii (PhC0)20 X=H “ X=Br, h X=I. ► X =I CH3C 0 2Ag

NR!R2 R3

0(C H 2 )3 N+R 1 R2 R3 Br'

HEP, X=H, R 1=R2 =R3=C2 H5;

CEP, X=C1, R 1=R2 =R3=C2H5;

BEP, X=Br, R 1=R2 =R3 =C2 H5;

IEP, X=I, R 1=R 2=R3 =C2 H5; BPP, X=Br, R ^R ^ C02Me

1) (C2H 5 )2NH, C2H 5OH;

0 (CH2 )3 If (C 2 H5)2 Br-

BCP

Scheme 13. Scheme for the syntheses of some halogenated psoralen derivatives. 68

appropriate bromides produced in the previous step. However, the reaction of 5-bromo-

8-(3-bromopropyloxy)psoralen with cinnamyldiethylamine failed to give the desired

quaternary ammonium salt BCP. The problem was circumvented by reacting 5-bromo-

8-(3-diethylaminopropyloxy)psoralen with cinnamyl bromide.

With monohalogenated water-soluble psoralen derivatives in hand, the next

synthetic goal was to prepare multihalogenated analogues. It has recently been reported

that direct bromination of 5-MOP with elemental bromine gave a mixture of di- and

tribromosubstituted psoralen products2048. Unlike 5-MOP, it is well documented 203 that bromination of 8 -MOP with NBS yields 5-bromo-8-methoxypsoralen and direct bromination of 8 -MOP in chloroform gives, depending upon the amount of bromine used, either 5-bromo-8-methoxypsoralen or 4’,5,5’-tribromo-4,,5’-dihydropsoralen.

The latter compound is not very stable and can be easily converted back to monobromo product simply either by refluxing in a solvent or by adding a reducing agent.

To prepare multibrominated psoralen derivatives, efforts to eliminate one molecule of hydrobromide from 4,,5,5’-tribromo-4’,5’-dihydropsoralen were always unsuccessful. These methods involved heating 4’,5,5’-tribromo-4’,5,-dihydropsoralen in various solvents or adding bases such as sodium acetate, potassium carbonate, and

l,8-diazabicyclo[5.4.0]undec-7-ene (DBU). Additionally, dehydrogenation of 4\5,5’- tribromo-4’,5’-dihydropsoralen with 2,3-dichloro-5,6-dicyanobenzoquinone (DDQ) also failed to give the expected tribromo compound. Furthermore, to our surprise, no bromine atom could be introduced onto the pyrone ring by using either an excess of elemental bromine or strong bromination reagents under various conditions. Such a reaction can easily be accomplished with coumarine itself in the presence of elemental bromine206. At last, the introduction of more than one bromine atom onto 8 - methoxypsoralen ring was realized by direct bromination in the presence of anhydrous aluminum chloride. This reaction not only replaced three hydrogen atoms with bromine atoms, but also hydrolyzes the methyl ether, directly providing the demethylated 69 4\5,5’-tribrorno-8-hydroxylpsoralen. To guarantee the water solubility of drugs, one or two water soluble groups were then installed by the same procedures employed with monobromopsoralens. However caution has to be exercised to avoid decomposition of the tribromopsoralen in a reducing environment at high temperature. The temperature of the reactions was always set as low as possible. A schematic representation for the syntheses of some tribrominated psoralen compounds is depicted in Scheme 14.

OCH3 OH

OCH2 CH2 CH2 N+R 1R2 R3Br

TBP, R 1 =R2=R3=CH2 CH3;

R ^ R ^ C H * R 3=CH2 CH2N(CH3 )2 ; R ^R ^C H s, R3=CH2 CH2OH; r ’= r 2= c h 3, R3=CH2 CH2CH2 OH.

Scheme 14. Synthetic route to tribromopsoralen derivatives.

As psoralen derivatives with quaternary ammonium ion side chain have demonstrated better binding selectivity for DNA over other cellular components than psoralens with other kinds of ammonium ion side chains13- 207, we prepared 4’- triethylammoniomethyl-4,5’,8-trimethylpsoralen, following a straightforward route 70 (Scheme 15). Chloromethylation of TMP with chloromethyl methyl ether yielded 4’- chloromethyl-4,5\8-trimethylpsoralen208, which was heated with triethylamine in toluene gave the required quaternary ammonium salt. Interestingly, the 13CNMR spectrum of this compound exhibited only 15 different kinds of carbon atoms, while there are 17 different kinds of them in the molecule. The question was answered by two dimentional NMR. 1H-13C correlation spectra demonstrated that two different methylene groups coincidentally have the same chemical shifts in the 13CNMR, as does one of the methyl groups on the psoralen ring and the methyl group of triethylamine.

CH3 9 H 3 CH2 C1

CH3 0C H 2 C1

CH 3 C 0 2H 0 ‘

c h 3 c h 3

TMP

CH 3 CH 2 N+(C2 H5 ) 3 Cl'

N(C2 H5 )3 Toluene, A

c h 3

Scheme 15. Synthetic route of 4'-triethylammoniomethyl-4,5',8-trimethylpsoralen. 71 111.2 Solubility in Aqueous Medium

Water solubility of a drug is necessary for it to be compatable with biological systems. A good DNA-binding drug must have a certain degree of water solubility to be useful.

Acr-NH2 , which has two hydrophilic amino groups, is quite soluble in 1 %

NRtAcfaq), pH=5.5. The substitution of one of the amino groups in Acr-Nffc with hydrophobic iodide reduces the solubility of Acr-I in water. Although Acr-I has limited solubility in 1% NRjAqaq), it is still sufficiently soluble for most of the studies described in the following sections. Acr-I was first dissolved in small amount of DMSO and then diluted with 1 % NH4 Ac(aq), whenever it became necessary.

The azido compound AMP, having a highly hydrophobic azido group, is the least soluble of all of the compounds investigated. The solubility of AMP is merely

43.7|iM in Tris HC1 buffer (lOmM Tris. 0.2M NaCl, pH=7.4), assuming it has the same extinction coefficient of ADP. Replacement of the methoxy group with a 3- diethylaminopropyloxy group on side chain gave ADP an elevated solubility of

8.80mM in the same medium.

8 -MOP has a solubility of 0.18mM in water208. Halogenation results in further diminished solubility. However, the side chain of quaternary ammonium salts greatly improves the water solubility of psoralen derivatives. Most of these compounds are very soluble in polar aprotic solvents and water. Even the tribromo compound TBP has a solubility as high as 2.25M in Tris HC1 buffer. Suprisingly, 4’- triethylammoniomethyl-4,5’,8-trimethylpsoralen chloride barely dissolves in water. The stock solution thus was prepared in DMSO.

111.3 Binding Constants with DNA

The magnitude of the binding constant of a sensitizer with DNA determines the amount of the sensitizer that can interact with DNA and ultimately the efficiency with 72 which the sensitizer can cleave DNA. In the case of psoralen derivatives, the magnitude of the binding constant also influences the extent of singlet oxygen involvement in

DNA damage 195 Many physicochemical methods are available to derive the binding constants of weak interactions between DNA and its ligands. Spectrophotometric methods, which detect changes in absorption or fluorescence spectra as a result of ligand interaction with DNA, are widely employed to assess the binding ability of a ligand to intercalate into DNA. Although the physical concepts for many methods are relatively evident in measuring binding constants, the actual numerical treatment is often liable to random errors. It is not uncommon that the binding constant of a particular compound with DNA is reported by different sources with a discrepancy of several orders of magnitude. Thus data obtained from the literature must be treated with caution and should only be used for comparative purposes.

We first tried to determine the binding constants (K) of our sensitizers with

DNA by employing Crother’s method using UV-vis absorption spectroscopy209. In principle, the interior of DNA helices is dramatically different from the aqueous medium wherein DNA exists. As a result, many photophysical properties of a sensitizer are changed accordingly when the sensitizer molecules move from aqueous solution into the two layers of hydrophobic DNA base pairs. In this method, the intrinsic binding constant of an intercalator is derived from the difference in the extinction coefficient of free intercalater with that of bound intercalator. According to eq 9, plotting [DNA]/(eap-

Ef) vs [DNA] will give a straight line with a slope of l/(£b-£f)and an intercept of 1/[K

(Eb-Ef)], where [DNA] is the concentration of DNA base , £ap, £f ,and £b apparent, free, and bound extinction coefficients, respectively. The value of £ap is calculated from

Aap/[compd] according Beer’s law.

[DNA]/(Eap-Er)=[DNA]/(£b-£f)+l/[K(£b-Ef)] (9) 73

A binding constant of 2. 1x 104M '1 was obtained for Acr-NH 2 , when the UV-vis

absorbance method was employed. This result is comparable with literature values

(V.ZxK^M *1 by the same method but under slightly different conditions, S.SxK^M 1 for

apparent binding constant using the dialysis method)207. Unfortunately, this same

method failed to work in the case of the Acr-I and psoralen derivatives, because the

changes of their extinction coefficients as a function of environment are not sufficiently

large to allow accurate measurement.

The ethidium bromide fluorescence displacement method was then adopted210.

In this method, ethidium bromide and the sensitizer in question compete for

hydrophobic intercalation sites on the DNA double helix. Ethidium bromide fluoresces

very weakly in aqueous medium, due to the efficient quenching of its excited state by

fast proton transfer from ethidium bromide amino groups to water. As this quenching

mechanism does not exist in hydrophobic media or in polar solvents incapable of

accepting protons, ethidium bromide fluoresces intensely211. The displacement of

intercalated ethidium bromide molecules by sensitizer molecules in DNA helices thus

leads to a decrease of the fluorescence intensity of ethidium. The binding constant of

the sensitizer can be determined by the change of ethidium fluorescence intensity. As

far as the experimental procedure is concerned, a series of sample solutions were

prepared such that the concentrations of DNA and ethidium bromide are kept constant,

while the concentration of the sensitizer increases steadly. After allowing for equilibration, the ethidium fluorescence of the series of solutions is recorded. The

fluorescence intensity thus obtained is plotted against the concentration of the sensitizer to determine the Cso% value. Cso% is the concentration of the sensitizer, with which the fluorescence intensity of ethidium bromide is half of that measured in absence of the sensitizer. The intrinsic binding constant is then calculated according the following simplified equation. 74

K=KEt[Et]C5o% (10)

where Kei=1.5x104 M - 1 is the intrinsic binding constant of ethidium bromide and K is

the intrinsic binding constant relative to calf thymus DNA212.

The plot of the fluorescence intensity of ethidium bromide versus the concentration of the sensitizer is curved, and C 5095, is obtained by extrapolation of this curve. It is thus difficult to accurately determine the Cso% value by this method. The problem becomes particularly serious when the sensitizer has a small binding constant with DNA or tends to aggregate in aqueous solution. According to our experience, the aggregation of sensitizer prevents many of the compounds tested from accurate measurement of ethidium fluorescence near the C 509&, concentration. This led us to develop a different numerical treatment of the fluorescence displacement data. The mathematical basis of this new procedure follows

KEt DNA + Ethidium . DNA-Ethidium

[DNA-Et] K0-[DNAp] 0I> where KEt is the binding constant of ethidium bromide with DNA, [DNA-Et], [DNA], and [Et] are the equilibrium concentrations of the complex of ethidium with DNA,

DNA and free ethidium bromide, respectively. Similarly,

^-Sens DNA + Sensitizer DNA-Sens

_ [DNA Sens] S8ns [DNA][Sens] U ' 75 where, accordingly, KsCns is the binding constant of a sensitizer with DNA,

[DNA-Sens], and [Sens] are the equilibrium concentrations of the complex of the sensitizer with DNA, and free sensitizer. Combining equations (11) qnd (12) gives

K DNA-Ethidium + Sensitizer — DNA-Sens + Ethidium

K _ KSans _ fDNA Sens][Et] KEt [DNAEt][Sens] U’

where K is the competitive equilibrium of sensitizer with ethidium bromide binding to

DNA. The fluorescence intensity (I) of ethidium bromide is proportional to the concentration of the ethidium-DNA complex, when other components in the solution do not fluoresce at the selected wavelength, Thus,

[DNA-Sens] _ l0 - I [DNA-Et] I K *

Furthermore,

[Sens] = CSens - |DN A . Sens] = CS9ns - ([Et] - |Et]0) = CSe„s - [Et]+ [Et)0 (15)

where I0 and I are the fluorescence intensity of ethidium bromide without and with the presence of the sensitizer, respectively. The quantity Cscns is the total concentration of the sensitizer added. [Et] 0 is the equilibrium concentration of ethidium bromide without the competitive binding from the sensitizer. Combining (13) with (14) and (15) gives

•o _ Ksens CSens+[Et]0 -[Et) (16) t - - ^ r x m 76 If the sensitizer only displaces a small portion of ethidium from the complex with DNA,

i. e. [Et]«[Et]0, eq 16 can be simplified to

Upon plotting Iq /I versus CsCns» the magnitude of Kscns can be calculated from the slope

of the resulting straight line.

Employing the new mathematical treatment for the ethidium bromide fluorescence displacement method, the intrinsic binding constant of Acr-NH 2 was determined to be 1.8xl0 4 M_1. This result is somewhat smaller than the earlier values obtained by other methods. But it should be kept in mind that the intrinsic binding constant is generally smaller than the apparent binding constant212. The binding constant of Acr-I with double helical DNA (V.SxK^M-1) was then determined to be slightly larger than that of Acr-NH 2 , apparently due to the hydrophobicity of the iodine atom substituent.

Table 2. The intrinsic binding constants of some psoralen derivatives with calf thymus ______DNA in Tris HC1 buffer at room temperature. ______

Compound ADP HEP CEP BEP IEP BCP BPP TBP AMT

Binding Constant 5 9 0 2 g 3 2 4 ] 4 2 3 2 4 Q 2? 5 Q

(xlQ 3 M '1),______

Table 2 shows the intrinsic binding constants of some psoralen derivatives with calf thymus DNA, measured by our method in Tris HC1 buffer at room temperature.

The binding constant of 8 -MOP with DNA is too small to be determined by this method. The replacement of the methoxy group with a 3-triethylammoniopropyloxy side chain increases the binding constant to 2.8xl0 3 M‘1. The enhancement of binding 77 ability most probably results from the attractive Coulombic interaction between the quaternary ammonium ion of the psoralen derivative side chain and the negatively charged phosphate on the DNA backbone 171 a. Halogenation of the psoralen aromatic ring increases the hydrophobicity of the psoralen core, which should enforce the tendency of the compound to intercalate into hydrophobic DNA base pairs. Indeed, it turns out that the binding constants of all of the monohalogenated psoralen derivatives are at least 10 times larger than that of their hydrogen counterparts. As expected from their hydrophobicity, the binding constant follows the order H < Cl < Br < I < N 3 . The substitution of one of the ethyl groups of the ammonium ion side chain of BEP with a cinnamyl group (BCP) slightly decreases the binding ability of the compound with

DNA. This may be the result of a diminished attractive electrostatic interaction between the ammonium ion and phosphate, as the bulky cinnamyl group keeps the two ions at longer distance. Another possible reason is that the more hydrophobic cinnamyl group is less coordinated with phosphate in an aqueous environment. On the other hand, the substitution of the triethylamine of BEP with a methyl isonicotinate (BPP) has little effect on the DNA binding ability. The binding constants of all of these halogenated psoralen derivatives are slightly smaller than that of AMT, except the azide derivative

ADP. The binding constant of another azide derivative, AMP, with DNA can not be measured conveniently by the ethidium bromide fluorescence displacement method due to its low solubility in aqueous media. The tribromopsoralen derivative has a binding constant which is about 7 times bigger than that of its monobromide counterpart and more than 5 times of that of AMT, as its hydrophobicity further increases. Finally, the substitution of an ethyl group with 2 -hydroxyethyl or 2 -(dimethylamino)ethyl group at the end of side chain, as shown in Scheme 13, approximately doubles the binding constant of TBP. 78 In conclusion, the binding constant of a psoralen derivative with DNA increases as the psoralen ring becomes more hydrophobic, and as positively charged ions are attached on its side chain.

III.4 Photophysics of Sensitizers

A progressive red shift of the UV-vis absorbance spectrum is a common characteristic of DNA intercalative agents in the presence of increasing amounts of double helical DNA. The spectral changes are neither a result of simple aggregation nor that of the polarity change around the intercalator. Explanations involving molecular orbital interactions have emerged, though they are quite complicated213. The longest

UV-vis absorption maximum of Acr-NH 2 has a wavelength of 444nm with an extinction coefficient of 4.30xl04M '1cn r1 in 1% NH 4 AC(aq), pH=5.5214. This peak can be shifted to a maximum of 462nm in calf thymus DNA in Tris HC1 buffer, pH=7.4

(Table 3). Spectral changes of Acr-NH 2 in DNA solutions have long been noticed and taken as evidence of intercalative binding of Acr-NH 2 with DNA. The substitution of one of the amine groups by iodide breaks the symmetry of Acr-NH 2 - For this reason probably, a new absorption band of Xmax 376nm in 1% N H^Ac^), pH=5.5 emerges, in addition to the red shift of the 444nm peak to 458nm. The later compound was determined to have an extinction coefficient of 1.32xl04M*1cm 1. Like Acr-NH 2 , the

UV-vis spectrum of Acr-I displays a similar pattern of solvent dependence and a red shift in DNA solution, as indicated in Table 3. Therefore, Acr-I is assumed to bind with

DNA in the same way as does Acr-NH 2 .

It is worth noticing that the variation of the UV-vis spectra of these two compounds as a function of the solvent polarity exhibits the characteristics of a (n, k*) transition, i. e. there is a red shift as the polarity of the solvent increases. The large extinction coefficients of the absorption peaks also support the 7t, 7t* assignment of these absorption bands. Generally speaking, the absorption maximum of an (n, ti*) 79 transition should be nearto but have a longer wavelength absorption maxima than the lowest (n, 7t*) transition. Nonetheless, there is no distinguishable (n, 7t*) absorption peak in both the Acr-NH 2 and the Acr-I UV-vis spectra. One can only surmise that the energy difference between the (n, ji*) transition and the lowest ( 7t, n*) transition must be small. As a result, the absorption peak of the (n, 7t*) transition is buried under that of the lowest (tc, k*) transition, because the extinction coefficient of a (rc, 7t*) transition is far larger than that of an (n, n*) transition.

Table 3. The maximum wavelength of UV-vis absorption of Acr-NH 2 and Acr-I in various solvents.

Compound Medium ^•maxCnm) eCxK^M^cm*1)

benzene 398

chloroform 398

Acr-NH2 methanol 456

1% NH4 AC(ac)), pH=5.5 444 4.30a

DNA in Tris HC1 buffer, pH=7.4 462

benzene 364,412

chloroform 364,412

Acr-I methanol 362,436

1% NH4 Ac(aq), pH=5.5 376,458 I.32(@ 458nm)

DNA in Tris HC1 buffer, pH=7.4 382,468 a. This datum is from reference 214.

Both Acr-NH 2 and Acr-I fluoresce in 1% NH4 Ac(aq), pH=5.5. However the quantum efficiency of Acr-NH 2 was 33 times that of Acr-I, as they were measured on samples of identical absorbance at 450nm in 1% NH4 Ac(aq), pH=5.5 (Xcm=525nm). 80 Additionally, the fluorescence quantum efficiencies of Acr-I varied when Acr-I was

excited at two points with identical UV-vis absorbance of different bands. In 1%

NH4 AC(aq), pH=5.5, when it was excited at 446nm, the Acr-I fluorescence intensity was

2.8 times greater than when it was excited at 376nm. In benzene, however, the

fluorescence intensity ratio of exciting at 369 and 523nm is 1.5. The fluorescence

lifetime of Acr-NH 2 was determined to be 4.9ns in water but the fluorescence of Acr-I

was too weak to allow accurate measurement of its fluorescence lifetime215.

As important biological sensitizers, the photophysics and photochemistry of

psoralen derivatives have undergone extensive investigations. However, there remain

many open questions in this extremely complicated subject. The UV-vis spectra of

psoralen derivatives have an absorption peak around 300nm. On the long wavelength

side of this band, there is a shoulder tailing as far as 380nm depending on the individual

psoralen structure and solvent. The shoulder becomes more prominant in a less polar

solvent. Both the shoulder and the peak around 300nm are believed to result from ( 7t,

7t*) transitions. The (n, 7t*) transition is buried under (tc, n*) transitions and can not be

observed164. It is well-known that all spectra of psoralen derivatives are strongly

solvent-dependent. We started by measuring the UV-vis spectra of 8 -MOP in a variety

of solvents and hoped to discover some relationships between the excitation energy of a

psoralen and solvent (data shown in Table 4).

Since the maximum wavelength (Xmax) of a transition is inversely proportional

to the transition energy, which can be affected by the environment of the molecule, it is reasonable to expect some correlation between A.max and the properties of solvents. In addition, it has been reported that there is a linear solvation energy relationship with the triplet yield of psoralen216. When the A.max data of 8 -MOP are plotted against

Reichardt’s E t (30) solvent polarity indexes, no obvious correlation was found (Figure

15a). Neither was there any correlationship between the Xmax data and dielectric constants of solvents. However, when the kmax data of 8 -MOP were plotted against Taft’s solvatochromic parameter n*, a good linear relationship was produced with a correlation efficient (Ra2) of 0.92. The addition of another solvatochromic parameter a improved the correlation efficient to 0.95, when the weighting coefficient c was between 0.15 and 0.18 (Figure 15b). Along with the shift of A.max in different solvents, there were small, irregular fluctuations in extinction coefficient.

Table 4. Collection of ^ uvmax and X.fmax of 8 -MOP in different solvents, Reichardts Ej(30) solvent polarity index217, dielectric constants(e), and Taft’s 7t* a ______solvatochromic parameters218. ______

Solvent k uvmax(nm) ^max(nm) Et(30) ea it* a

hexane 293 30.9 1.89020 -0.08 0 . 0 0 cyclohexane 294 440b 31.2 2.02320 0 . 0 0 0 . 0 0

1,4-dioxane 297 470b 36.0 2.20925 0.55 0 . 0 0

benzene 298 460b 34.5 2.28420 0.59 0 . 0 0

DMSO 301 45.0 4.720 1 . 0 0 0 . 0 0

ethanol 297 475c 51.9 24.3025 0.54 0.83

acetonitrile 300 482b 46.0 37.520 0.75 0.19

methanol 300 498b 55.5 32.6325 0.60 0.93

water 304 512C 63.1 80.3720 1.09 1.17

acetone 452b 42.2 20.7020 0.71 0.08

n-butanol 488b 15.820 0.47 0.79 a. The right superscriptions indicate temperature, b. data from reference 220. c. data from reference 175.

Reichardt’s Et (30) scale is constructed from the molar transition energy Ej, expressed in kcal/mol, for the longest wavelength solvatochromic absorption band of pyridinium-N-phenoxide betaine dye 30 in the given solvent. It is a general solvent polarity scale. Dielectric constant is a macroscopic polarity scale. Nonetheless, Taft’s 82

304 - G a 302 ■ E c □ '•g 300 - G G J 298 - G G G 296 -

294 - G □ oood 1.... ” "1... i ■ i 30 4 0 5 0 60 &r(30)

305 ✓—s E 303 - s»/c s 301 - J 299 -

297 -

295 -

293 08 0.22 0 .52 0.82

Figure 15. Correlation between with Reichardt's Bj<30) solvent scale and Taft's

solvatochromic parameters 7t* and a. a. There is no good correlationship

between Xmax vs. ET(30); b. The plot of vs. 7t* (a ) gives a good correlationship with Ra2=0.92, which is improved to 0.95 for the plot of Xmax vs. 7t*+ca(o,c=0.15~0.18).

7t* is only an index of polarity and polarizability of the solvent. It is used with a correction term 8 when applied to aromatic solvents (since there was only one such 83 solvent in our data, this correction was neglected in our data analysis process). A good

correlation of the Xmax data of 8 -MOP with n* rather than Et(30) probably means that

only the polarity and polarizability of solvents play important roles in the transition

energy of 8 -MOP. The a parameter is the index of hydrogen-bonding donor ability of

the solvent. The fact that the participation of a with n* betters the linear correlationship

indicates that the ground state of 8 -MOP can, to a certain degree, form hydrogen bonds

with certain solvents. Since the time scale for the completion of the excitation process

is on the order of femtoseconds, it is reasonable to believe that the hydrogen bond

existing between ground-state 8 -MOP and solvent molecules will not change during the

period of the excitation process. Given that the 8 -MOP molecule contains polar

functional groups, it is not unusual for some 8 -MOP molecules in the ground state to

form hydrogen bonds with suitable solvent molecules. Actually, Ishikawa has reported one-to-one hydrogen bond formation between trichloroacetic acid and ground state

TMP in cyclohexane219. The equilibrium constant for this hydrogen bond forming reaction was measured to be 2300M'1 in hexane at low concentration of trichloroacetic

acid. When ethanol replaces trichloroacetic acid to serve as the hydrogen-bonding donor, the equilibrium constant diminishes to 2.4M'1.

A similar red shift in the UV-vis absorption of 8 -MOP occurs in dioxane-water mixtures as the content of water increases. Again, no correlationship is found between the maximum wavelength data and Reichardt’s Ex(30) or dielectric constant e.

Checking for the correlation of the maximum wavelength with Taft’s solvatochromic parameters is impossible, due to the lack of the parameters for solvent mixtures.

The emission spectra of psoralen derivatives have also been explored in detail.

The fluorescence quantum yields of psoralen derivatives increase, as the solvent becomes more polar and hydroxylic, wherein the energy levels of (n, rc) and (tc, tc*) states are better separated. This phenomenon has been attributed to the reduced proximity effect which affects the rate of S i —>S 0 internal conversion. The large Stokes 84 shift has been suggested to be the result of intermolecular charge transfer between an excited singlet psoralen molecule and a solvent molecule220. The fluorescence decay does not fit to a single exponential. This has been explained in terms of the existence of multiple excited species such as free singlet state and hydrogen-bonded singlet state.

Solvents have little effect on the phosphorescence of psoralens. This has been taken as the evidence of the energy localization of triplet state on the 3,4 C=C bond of the pyrone moiety.

520 - 520- 510 - □ 510 • a S-s b 500 - □ E 500 - a c 490 - □ 490 : d / a 480 _ E 480 - □ / □ r

Figure 16. Correlation between A.fmax with Taft's solvatochromic parameters n* and a. r (a)Plotting V max vs. 7t* gives poor correlation; (b)The correlation coefficient

is improved to 0.843 when plotting Xfmax vs. 7t* +coc(c=0.99~1.06).

Sukigara et al 176 reported a good correlationship between the maximum wavelength of 8 -MOP fluorescence and the dielectric constants of solvents, but there were only five solvents included. We found that this correlation did not exist, when it was extended to include more solvents. Like UV-vis absorption, we tried to plot the maximum wavelength data of 8 -MOP fluorescence versus Taft’s solvatochromic parameters n* and a. A plot of Xfmax vs. n* gave a poor linear relationship, but ^fmax 85

vs. 7t*+ca (c=0.99~1.06) improved the correlation coefficient to 0.84 (Figure 16). The

larger weighting coefficient c in this case is much bigger, compared to Xuvmax vs. 7t*+c.

It is not surprising that more 8 -MOP molecules at the first excited state form hydrogen

bonds with solvent molecules, with respect to the more polar character of (n, n*) excited-state molecules than ground-state molecules. Actually, protons have been reported to be capable of quenching the fluorescence of 8 -MOP221.

The quantum yields of fluorescence and phosphorescence of psoralen itself were reported to increase linearly against both Reichardts Et(30) solvent polarity index, and

Taft’s solvatochromic parameters 7t* a. However, the quantum yields of fluorescence and phosphorescence of 5-MOP in dioxane-water mixtures first increase as the concentration of water increases and reach their maximum values at a water concentration of 8 M. Thereafter, the further addition of water causes a decrease of the emission quantum yields222. The authors attributed the enhancement of emission quantum yields for the region of increasing water content to the result of the weakening of the proximity effect between the two closely spaced singlet excited states. In the second region, the authors argued, the proximity effect no longer functions. The addition of water in this region only increases the rate of internal conversion through hydrogen bond formation of the excited state. We measured the fluorescence intensity of 8 -MOP in a series of dioxane-water and ethanol-water mixtures. Similarly, the fluorescence intensity of 8 -MOP peaked at 75% water (by volume) in dioxane-water mixtures and 70% water (by volume) in ethanol-water mixtures (Figure 17). 86

30- a 25* B ■ “ D 20- a a a a £ . 5 - • a 1 10- a □ 5 " □ a ii n n 0 10 20 30 40 50 60 70 80 90 100

H 20%(in volume)

5.0

4.5-

&S? 3.0-3’5 "

1.0 v — r ■> i "i " i i i ■ i ■ r--i i i— i— i— i— i— i— r— 0 10 20 30 40 50 60 70 80 90 100

H20%(in volume)

Figure 17. The effect of water content on fluorescence intensity of 8 -MOP in (a)dioxane- water and (b)ethanol-water mixtures. I and I are the fluorescence intensities of 8 -MOP in pure dioxane and solvent mixtures, respectively.

Table 5 includes the UV-vis absorption properties and relative emission yields of some new psoralen derivatives. The replacement of hydrogen with halogen results in 87 red shifts of the maximum wavelength in UV-vis absorption in the order of HEP < CEP

< BEP < 1EP < ADP. The results are in accord with the electron donating ability of the

substituents. These substituents raise the energy level of the ground state of psoralen

and thus reduce the transition energy. The exception is that Xmax of TBP is even smaller

than that of BEP. This is probably a result of an aggregation effect of TBP due to its

high hydrophobicity. In other words, ground-state molecules of TBP probably form

complexes or micelles in aqueous solution. As the maximum wavelength red-shifts, so

does the extinction efficient increase slightly for every compound. The changes on the

side chain of these compounds do not affect the maximum wavelength (BCT and BPP),

as expected. Accordingly, their extinction coefficients are assumed to be the same as

BEP.

Table 5. Photophysical data of some psoralen derivatives.

Compound ^■max(nm)a E(xl03M-,cm-,)a (fob Tc

ADP 320 1.40

HEP 304 1.16 1 1

CEP 308 1.21 1.16 1.0

BEP 312 1.34 0.74 2.0

DEP 316 1.30 0.17 0.3

TBP 309 1.43 0.58 1.0

BCP 312 1.34 0.71 0

BPP 312 1.34 0.27 0

AMT 14.4 >1.3 a. Data obtained in Tris HCI buffer, b. Relative fluorescence intensily against HEP measured excited at 330nm with same asorbanccs. c. Relative triplci-iriplci absorbance against HEP measured by LFP, XeCl, 308nm, 55mJ, 17ns with A3Qg=1.0 in Tris HCI buffer. The relative triplet yield equals to the relative triplet-lriplei absorption on the condition that the effects of halogen substituents on the extinction coefficient of the triplcl-lriplci absorptions arc negligible. 8 8 Halogenation dramatically changes the quantum yields of the photophysical processes of psoralen excited states, as demonstrated by their relative fluorescence and triplet-triplet absorption intensities. Chlorination should affect the energy level of a n electron more than that of a nonbonding electron on the oxygen atom of a carbonyl group. A slightly larger separation between (n, n) and (it, it*) states occurs and thus increases the quantum yield of fluorescence slightly due to the corresponding weakening of the proximity effect. The triplet-triplet absorption intensity of CEP is the same as HEP. Bromination reduces the quantum yield (x) of formation of the triplet excited state, as is evident by comparing the data of BEP and HEP. The reduction of the quantum yield of fluorescence and the enhancement of the quantum yield of triplet formation are a consequence of the heavy atom effect, which raises the intersystem crossing rate.

However both the f and <|>t of iodocompound IEP are diminished dramatically relative to HEP. The former effect can be due to the heavy atom effect again; The latter observation can be explained by photolytic dissociation of the C-I bond. The singlet state energy of psoralen derivatives is around 73.8kcal/mol. The bond dissociation energy of an aromatic C-I bond is about 64kcal/mol223. Thus, the excitation of a psoralen derivative provides enough energy for breaking the C-I bond. Similarly, the

TBP data can be explained in the same way but with a far slower rate of C-Br bond dissociation, due to the similarity of the bond dissociation energy of an aromatic C-Br

(74kcal/mol223) bond and the singlet state energy of a psoralen derivative.

The singlet state energy of styrene is 99kcal/mol223, which is larger than that of psoralen. The cinnamyl moiety on the BCP side chain can not quench the singlet excited state of the psoralen derivative. On the other hand, the triplet state energy of

(E)-p-methylstyrene (60kcal/mol) is lower than that of psoralen (~63kcal/mol)223. The cinnamyl moiety of the BCP side chain can quench the triplet excited state of the psoralen derivative by an energy transfer mechanism. Consequently, BCP has 89 approximately the same <|>r as that of BEP but no observable triplet-triplet absorption.

The pyridinium moiety, however, can quench both singlet and triplet excited states of the brominated psoralen derivative by ET process. BPP thus has only a small quantum yield of fluorescence and no triplet-triplet absorption.

Finally, it is of interest that the substituents on the psoralen ring of AMT increase the <(>f of AMT more than 14 times relative to that of HEP. The large quantum yield of fluorescence means slower internal conversion rates. AMT hence can be more efficient in singlet state photochemical reactions with DNA bases.

III.5 Photochemistry of Sensitizers

Acridine, psoralen, and their derivatives have long been known to react with

DNA, proteins, and their building blocks, the nucleic acid bases and amino acids, as exemplified by fluorescence and phosphorescence quenching studies129. The excited state quenching processes in these cases are believed to be ET rather than energy transfer processes with respect to the excited state energy levels of the relevant reactants.

III.5.1 Photochemistry of Acridines

The Weller equation (eq 4 in section II.2.1) predicts facile ET from GMP to the excited Acr-NH 2 and Acr-I, according to the data in Table 6 . Furthermore, halogenated deirvatives are more easily reduced than their hydrogen counterparts. But Table 7 shows that all the kqx values of Acr-I are slightly smaller than the corresponding data of

Acr-NH2 . If the ET rate constants in quenching processes for theses two compounds are the same, the smaller kqx values of Acr-I implies that the excited state of Acr-I has a shorter lifetime than that of Acr-NH 2 . The shorter lifetime of Acr-I is a consequence of a faster intersystem crossing process, which competes with ET and fluorescence. 90

Table 6 . Data of reduction potentials and singlet state energy of Acr-NH 2 , Acr-I, ______8 -MOP, and 5-bromo-8-methoxypsoralen. ______

Compound Acr-NH 2 Acr-I 8 -MOP 5-Bromo-8-methoxypsoralen

EredCeV)8 -2.02 -1.40,-1.70 -1.82b -1.50b

Es(kcal/mol)c 60.4 ______57JJ______718 ______718 ______a. Data obtained by cyclic voltammctry vs. Ag/AgCI with 0.1M TBAP in DMSO. b. Data from the quartly report of Jean Nicolas Acbischcr c. Data calculated for the starting points of the fluorescence of the corresponding compounds.

Table 7. Slopes(kqX M-1) of the Stem-Volmer fluorescence quenching plots8.

Acr-NH2 Acr-I HEP BEP

GMP 172 119 43.9b 60.8b

CpG 295 266

DNA 1.57X104 5.34x10 3

Tryptophan 83.2 63.6 6.42 6.55

Phenol 27.5 25.9 4.77 4.78

Sodium azide 13.58

DABCO 2.17 2.18 a. Slock solution of acridinc compounds were made in 1% ammonium acetate, pH=5.5. Slock solution of psoralen compounds were made in Tris HCI, ph=7.4 buffer. In the case of DNA as a quencher, a scries of solution of appropriate concentration were left for equalisation for at least 4hr before their fluorescence data were recorded, b. Data of Kasturi, C. D issertation^,

In a manner consistant with this prediction, the fluorescence of both Acr-NH 2 and Acr-I is quenched very efficiently by GMP. Very weak fluorescence quenching was observed with AMP and there was no discernible fluorescence quenching with

CMP and TMP. These results are in line with the ability of the nucleic acid bases to donate an electron, i. e. their oxidation potentials. Interestingly, the kqT values of fluorescence quenching with CpG increase by a factor of about 2, compared to the 91

2.50 • GMP ♦ CpG 2.25 - o DNA A CMP

e* 2 .0 0 -

1.75-

1.50-

1.25-

1.00 I i A |...f-T—A— | i Q i | it 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

[Quencher] x 103M

Figure 18. Stenr-Volmer plots of the fluorescence of Acr-I quenched by calf thymus DNA, CpG, GMP, and CMP. CMP essentially does quench, while DNA, CpG and GMP quench with kqxs of 5.34x103, 266 and 119M'1, respectively. The DNA data is calculated in terms of the concentration of base pairs present.

quenching with GMP (Table 7), This, along with the red shift of the maximum wavelength of UV-vis absorption, is in line with Voelk’s conclusion that these acridine compounds bind with (CpG> 2 196. Intercalation into the (CpG ) 2 mini duplex brings acridine molecules into close proximity with the guanosine quenching residue of CpG, which facilitates the ET process. The fluorescence quenching is further enhanced by a factor of 5~10 on a per base pair basis when calf thymus DNA is the quencher. Figure

18 shows the Stern-Volmer plot of Acr-I fluorescence quenching with some nucleotides. Additionally, the fluorescence of Acr-NH 2 and Acr-1 can be quenched by 92 several amino acids. Tryptophan is the most reactive amino acid in regard to giving up

an electron. It quenches the fluorescence of both Acr-NH 2 and Acr-I easily. Tyrosine

and cystine are another two amino acids that can quench the fliorescence of acridines.

Due to their poor water solubility, however, the kqT value of cystine could not be

determined, and that of tyrosine was modeled by phenol.

III.5.2 Photochemistry of Halogenated Psoralen Derivatives

The fluorescence quenching of psoralen derivatives by various quenchers

generally produces small kqx values, due to their extremely short-lived singlet excited

states. For example, the lifetime of TMP in cyclohexane was reported to be 14ps, that of hydrogen-bonded TMP 0.36ns219. Similarly, the singlet lifetimes of 5-MOP and 8 -

MOP in various solvents are smaller than 0.5ns222. As a result, free psoralen molecules often react via their triplet states. Halogenation of psoralen facilitates fluorescence quenching. To determine if an ET process is involved in the reactions of DNA with psoralen derivatives, particularly halogenated psoralens, we tried to observe the production of the radical anions of psoralen derivatives produced by LFP in the presence of an excited state quencher. A host of different quenchers in a variety of solvents have been examined in these experiments. For unknown reasons, we could only see the quenching of the triplet state of the psoralen derivative, but never the radical anion resulting from the ET quenching even under the same conditions reported in the literature189* 19°. These experiments are difficult in aqueous solution, where the triplet state of 8 -MOP and its radical anion have the same absorption spectra.

We then turned our efforts to product analysis. We hoped to observe some reaction products which might implicate the ET reaction mechanism. Several simple olefins have been used as models for biological molecules to react with psoralen deirvatives225. Tetramethylethylene (TME) was chosen to react with 5-bromo-8- methoxypsoralen because of its symmetry, as well as its electron-rich character. 93 Interestingly, irradiation of 5-bromo-8-methoxypsoralen in the presence of excess TME

in ethanol with 350nm light produces a debrominated cycloaddition product

(HPoTM E), the expected normal cycloaddition product (BPoTM E), as well as other

side products (Scheme 16). Both cycloaddition products have been separated and

confirmed by !HNMR, and mass spectrometry. Possible reaction mechanisms of the

reactions of 8-MOP and 5-bromo-8-methoxypsoralen with TME are described in

Scheme 17.

hv (350nm)

H ethanol, 4°C OCH,

+ + Other Products

B P oT M E H P o T M E

Scheme 16. Reactrion equation of photolysis of 5-bromo-8-methoxypsoralen in the presence of TME in ethanol.

5-Bromo-8-methoxypsoralen is first promoted to the singlet excited state, a small portion of which undergoes intersystem crossing to the triplet state. Singlet states as reactive species cannot be ruled out, since olefins do quench the fluorescence of 8-

MOP225c. A triplet mechanism is more likely however, due to the low quantum yield of R-H ------1 A -R- X 1 OCH3 i * cOCH-i

X=H or Br CQ o o II X CO OCH-i - H

+ o OCH OCH3

+

BPoTME + HPoTME

Exciplex

Scheme 17. Possible mechanism of photochemical reaction of 8-MOP or 5-bromo-8-methoxypsoralen with TME.

vo 95 fluorescence and the short lifetime of the singlet excited state of psoralen derivatives.

More importantly, the quantum yield of photocycloaddition of 8-MOP with TME by benzophenone sensitization is about 5 times that determined by direct irradiation2250.

Furthermore, oxygen and paramagnetic ions quench the photocycloaddition173. The triplet state of psoralen derivatives could interact with a ground-state TME to form a triplet state exciplex. The exciplex might then proceed to form a biradical intermediate.

Coupling of the two radical centers of the biradical affords the cycloadduct. However this route can only give the cycloadduct BPoTM E, when 5-bromo-8-methoxypsoralen is the substrate. Since we found that the cycloproduct BPoTM E, like its hydrogen analogue HPoTM E, does not absorb light over 320nm, it is unlikely that BPoTM E will photochemically (exposed to 350nm light) lose the bromine atom. Futhermore, the cycloadduct of 8-MOP with an alkene is photochemically inert, even if excited with light of short wavelength181. We would then expect that the cycloadduct B PoTM E has similar properties.

Alternatively, the excited triplet state of 5-bromo-8-methoxypsoralen can capture an electron from TME to form a radical anion. The resulting radical anion forms an aromatic psoralen radical by losing bromide ion, competing with a back ET process. The aromatic psoralen radical can abstract a hydrogen atom from solvent to become 8-MOP. 5-Bromo-8-methoxypsoralen and the resultant 8-MOP can then yield the cycloadducts BPoTM E and HPoTM E respectively by photochemically reacting with TME via radical ion mechanism.

III.5.3 Photochemistry of Axido Psoralens

The photochemistry of phenyl azide is now well understood226 (Scheme 18).

Phenyl azide, upon the absorption of a photon, releases singlet phenylnitrene. The singlet nitrene undergoes intersystem crossing to form a triplet nitrene at temperatures below 150K or rearranges to form a ketenimine at temperatures above 170K. 96

Ni N hv T > 170K a -N2 a O Ketenimine

T < 150K HN(C2H5)2

N 3 M C 2U5)2 a

Azepine

P h x N =N Vh Azobenzene

Scheme 18. Reactions of phenyl nitrene.

LFP of AMP in a variety of solvents produced transient absorption spectra with a sharp band around 410nm and a weak broad band ranging from 450 to 550nm. The transient species could not be the triplet excited state of AMP, since the presence of either isoprene or molecular oxygen did not affect the lifetime of the transient. LFP of

AMP in a glassy diethyl ether:pentane: ethanol (5:2:5, EPA) at 77K gave rise to a similar spectrum but with reduced intensity of the broad band. Furthermore, the transient UV-vis absorption spectrum observed in benzene was reminiscent in its general features of that of triplet phenylnitrene recorded at 77K (Appendix B, spectrum

6). The sharp band thus was assigned to the transient absorption of the triplet nitrene. 97 This assignment was further supported by the observation of an electron paramagnatic

resonance (EPR) spectrum recorded in EPA at 77K. A resonance signal around 6210

Gauss, a characteristic value for a polycyclic aromatic nitrene, was observed after

photolysis o f AMP in a cold matrix.

lN

ISC -N2 o OCH

AMP

NH

+ Other Products O OCH, OCH3

Major

Scheme 19. Reaction of AMP upon irradiation with 350nm light.

It has been reported by Li et al227 that electron donating groups, including the

methoxy group, located at the para position of phenyl azides reduces the yields of ketenimines produced upon LFP. Consistent with this report, we found that addition of diethylamine only led to a reduction of the broad band (450-550nm) in the transient absorption spectrum and had little effect on the sharp band (~410nm) upon LFP of

AMP. This indicated that the minor transient species responsible for the broad band was a ketenimine. Photolysis of AMP in pure diethylamine at 3°C yields as a major product

5-amino-8-methoxypsoralen rather than an azepine (Scheme 19). However photolysis 98 of AMP performed in other solvents gave products which could not be identified.

Therefore, it is tempting to conclude that the major intermediate produced on photolysis

of AMP is a triplet nitrene under all of the conditions examined here.

In the transient absorption spectrum produced by the LFP of ADP in the presence of calf thymus DNA in phosphate buffered saline (PBS), however, the band at

525-550nm has almost the same intensity as the band at 408nm (Appendix B, Spectrum

8). The band at 408nm is again believed to be due to the absorption of a triplet nitrene,

which shows little decay 800ps after the laser pulse. On the other hand, the 525-550nm band, which has a lifetime of 2.5|is, is probably a ketenimine associated with DNA. The lifedme of the DNA-bound ketenimine is unaffected by the addition of diethylamine, an excellent scavenger of ketenimine.

III.6 Sensitized Unwinding and Cleaving Plasmid DNA

pBR322 is a double-stranded closed circular E. coli plasmid DNA, containing

4363 base pairs228. It exists normally in a supercoiled form and is very sensitive to damage by a variety of physical and chemical agents. This plasmid DNA, therefore, has been widely used as a target molecule in investigating the interactions between chemicals and DNA.

For closed circular DNA, its topology can be described by the linking number a, defined by a=P+x, where P is the number of helical turns of the DNA duplex and x is the number of superhelical turns229. The linking number remains constant as long as the

DNA strand does not break. A compound which unwinds DNA reduces the number of helical turns p. To keep the linking number a constant, the number of superhelical turns x, usually a negative number for native DNA, must become less negative. The net effect is that the superhelical density decreases, i. e. DNA unwinds. After the number of helical turns P reaches zero, additional bound intercalaters shall result in positive supercoiled DNA230. On the other hand, if DNA helical scission occurs, it induces 99 supercoiled DNA to relax and yields a still double-stranded, but open circular DNA. It is assumed that further cleavage does not cause any dramatic change in the physical properties of DNA, unless the breakage occurs in opposite strands within about five base pairs of one another231. If that occurs, a linear DNA molecule will be generated.

All of these changes plus the deformation due to bending can be revealed by the changes of DNA’s migration rates on agarose gel through electrophoresis. Gel electrophoresis is usually a reliable technique which reveals these changes232, except for certain extreme conditions233.

III.6.1 DNA Photocleavage with Acridines

Samples of supercoiled plasmid DNA pBR322 in the presence of 50p.M of Acr-

NH2 or Acr-I were irradiated with 350nm ultraviolet light with an intensity of

7.7mW/cm2 for specific periods of time. The samples were then developed on an agrose gel by electrophoresis. No discernible damage to plasmid DNA occurred after 60min of photolysis in the absence of sensitizer. Only a small portion of supercoiled DNA was converted to the circular form upon 40, 50, and 60min of photolysis in the presence of

50|iM Acr-NH2 - On the other hand, substantial cleavage was observed in the presence of 50p.M Acr-I as a sensitizer after lOmin of irradiation, and all of the supercoiled form of DNA was transformed to a circular form in 25min. Additionally, the plasmid DNA nicking experiment was also performed at a variety of concentrations of these acridine sensitizers with 400-440nm light with a dosage of 9.2kJ/m2. We found that the minimum concentration of Acr-I needed to completly convert supercoiled DNA to circular DNA was 50|iM under these conditions. However a concentration as high as

0.1 mM of Acr-NH 2 could only cause a small conversion of plasmid DNA under the same conditions. Thus Acr-I is more effective agent at nicking plasmid DNA than Acr-

NH2. 100 In both cases, there was no discernible differences in the DNA nicking

experiments carried out in the presence and the absence of oxygen. These results

indicated that the production of singlet oxygen, superoxide ion or hydroperoxyl radicals

was not responsible for the DNA nicking. Nonetheless, the addition of dithiothreitol

(DTT) with a concentration of lager than lOmM protected DNA from cleavage by the

sensitizer Acr-I. This means that the DNA damage induced by the sensitizers can be repaired by DTT.

III.6.2 DNA Photocleavage with Psoralen Derivatives

As discussed earlier in section II.3, psoralen derivatives can react with DNA upon photoactivation to form monocycloadduct and cross-linked diadduct, as well as other type I and type II photoreactions. As a result of these photochemical processes,

DNA bending, unwindng, and nicking often arise. First, DNA cross-linking forces

DNA helix bending. It has been suggested from the X-ray crystallographic analysis of an 8-MOP-thymine monoadduct that psoralen cross-links could bend DNA by as much as 70°234. The combination of molecular modeling and energy minimization techniques235, as well as two-dimensional JHNMR studies236, estimated a 40-50° bending towards the DNA major groove at the psoralen cross-linking site. As a result of bending, DNA migrates more slowly by gel electrophoresis, which has been confirmed by both experimental237 and theoretical studies238. Suprisingly, in the case of psoralen derivatives, however, gel electrophoresis demonstrated that psoralen-DNA interstrand cross-linking did not change the DNA mobility directly. This result infers that DNA cross-linking by psoralens does not produce bending on the DNA helix239. These contradictory results remain to be resolved. In this dissertation, we will ignore the effect of DNA bending on the migration rate of DNA on agarose gels. Second, both monocycloaddition and cross-linking dicycloaddition reactions lead to unwinding of supercoiled DNA. It is estimated that the average unwinding angle of a monoadduct of 101 psoralen derivatives with DNA is about 18°, while the average value of that for a diadduct is close to 38° 240. The unwinding degree observed on agrose gels is an average value which reflects the ratio of cross-linking diadduct to monoadduct. Third,

DNA unwinding and DNA nicking are two completely different concepts, though they both cause changes in the migration rate of DNA on gel electrophoresis. Unwinding results from the distortion of DNA helices, but it does not change the linking number of

DNA. Nicking, however, means the cleavage of DNA backbone, causing supercoiled

DNA to change to circular DNA or circular DNA to change to linear DNA. Apparently, cycloaddition is not capable of causing DNA cleavage. Reactive oxygen species cannot be solely responsible for DNA nicking, because DNA nicking is observed under anaerobic conditions. We are not aware of any explanation which describes DNA photocleavage sensitized by psoralen derivatives. Herein we will present the results of

DNA photocleavage with our psoralen derivatives and our own interpretations.

In order for the results to be comparable, all experiments were performed with different concentrations of psoralen derivatives under the same photolytic conditions.

All psoralen derivatives assessed did not change the mobility of plasmid DNA without

UVA irradiation, neither did UVA treatment in the absence of a sensitizer. Like 8-

MOP, psoralen azide ADP and 4’-triethylammoniomethyl-4,5\8-trimethylpsoralen chloride failed to label or nick DNA appreciably. Thus, the data obtained with these compounds will not be presented and discussed.

In all cases, the mobility of DNA after photolysis was reduced and the band broadened more or less as the concentrations of psoralen derivatives increased. For each compound, there was a concentration at which the migration rates of DNA reached a minimum (Table 8). This pattern of change has been observed for many other DNA intercalators as well, e. g. ethidium bromide241 and platinum anticancer drugs218. The minimum concentration, designated as Cu, is the concentration required for a compound to completely unwind a DNA helix, provided that the bending of the DNA helix caused 102 by monoaddition and cross-linking does not affect the mobility of the DNA on agarose

gel, if any occurs. As the concentrations of psoralen derivatives continue to increase,

the open circular form of DNA increased in yield and finally reached a point where all

supercoiled DNA completely disappeared. Accordingly, this concentration is designated

as Q, to represent the minium concentration of a compound to completely convert DNA

from supercoiled form to circular from under constant (49.2kJ/mol) light dose.

Table 8. The minimum concentrations of psoralen derivatives to completely unwind and nick DNA under 20min irradiation with 350nm light of an intensity of ______4.1mW /cnr2 at 3°Ca.______

Compd. HEP CEP BEP IEP TBP BCP BPP AMT

Cu(p.M) 10 20 6 15 4 15 50 30

Cn(fiM) >600 300 40 15 4 50 50 70

Monoaddition unwinds DNA with a smaller angle than does cross-linking bicycloaddition. The degree of DNA unwinding observed on agarose gel thus depends on both the amount of cycloadducts and the ratio of cross-linking diaddition to monoaddition reactions. Thus the following factors can affect the magnitude of the Cu of a compound. First, substituents on C-4 positions, as pointed out earlier, increase the ratio of cross-linking diadduct to monoadduct due to the steric effect on the cycloaddition. Psoralens bearing substituents at this position will display larger DNA unwinding angles, if all molecules react with DNA by cycloaddition. Second, the quantum yields of the formation of cycloadducts can affect the results, and so does the different ability of these psoralen derivatives to absorb light, /. e. extinction coefficients.

With limited light influx, the structure of a compound and all of the avenues available for the decay of the excited state should exert an influence. Third, the ratio of cross- linking diadduct to monoadduct may vary with the radiation influx and the base 103 sequence at the site of the adduct. These two factors should not have an impact on our results, as all the experiments were performed with DNA under the same conditions. On the other hand, the magnitude of C„ of a compound should depend on all of the processes yielding radicals, which then lead to the scission of DNA helix.

To our surprise, all of our psoralen derivatives, except for BPP, showed relatively large efficiency in sensitizing DNA unwinding relative to AMT. With respect to the high fluorescence yield of AMT, we can only speculate that the three methyl groups and the one amminomethyl group reduce the quantum yield of the cycloaddition reactions of AMT with the pyrimidine residues of DNA. AMT thus is not as efficient as our psoralen derivatives in inducing DNA unwinding.

The cinnamyl group reduced, to a small degree, the efficiency of the psoralen derivative BCP in sensitizing DNA unwinding. The pyridinium ion of BPP, however, reduced to a much larger extent the efficiency of BPP induced DNA unwinding. The cinnamyl group in BCP quenches the excited triplet sate of psoralen but not the excited singlet state. As a result, BCP displays no observable triplet-triplet absorption but approximately the same fluorescence quantum yield as does BEP. On the other hand, the pyridinium ion of BPP quenches the excited singlet state of psoralen. BPP thus has a low fluorescence quantum yield compared to BEP. These different photophysical and photochemical properties should be responsible for the difference in DNA unwinding displayed by these compounds. THe reduced efficiency of BPP induced DNA unwinding may well ascribed to that the pyridinium ion quenches part of the cycloddition reactions of the psoralen ring with the pyrimidine residues of DNA via reversible electron transfer, while cinnamyl group does not. This explanation is in accord with the claim that the excited singlet state of a psoralen derivative, rather than the excited triplet state, is responsible for the cycloaddition reactions when the psoralen derivative intercalates into DNA. 104 As far as the cleavage of DNA helix is concerned, HEP did not completely cleave DNA even at a concentration as high as 0.6mM, although it binds to DNA more

strongly than does 8-MOP. Chlorinated derivative CEP showed improvement to some extent. These results implied that HEP and CEP under the experimental conditions are inefficient in sensitizing the production of reactive oxygen species or generating radical species capable of cleaving DNA in any way. AMT showed dramatic improvement in

DNA cleavage relative to HEP in the presence oxygen. Brominated and iodinated derivatives, BEP, IEP, and TBP, made the psoralens even better DNA nicking reagents than AMT. These compounds were so effective in nicking DNA that the unwinding process overlapped with the nicking process i.e, the extraordinarily small Cu values probably resulted from the dominance of unwinding processes over nicking processes.

Interestingly, the built-in excited state quenching devices in BCP and BPP only exhibited a very small negative effect on DNA photocleavage. A possible reason for the inability to prevent the DNA photocleavage by these quenchers is that the intercalation of the psoralen ring into the hydrophobic interior protects them from the hydrophilic quenchers.

In an attempt to unveil some mechanistic aspects of DNA photocleavage, the nicking experiment with AMT was examined under deoxygenated conditions. The results showed no appreciable differences relative to those obtained under aerated conditions. Unfortunately, a prolonged deoxygenating process caused severe damage to supercoiled DNA.

Additional DNA photocleavage experiments were performed in the existence of the radical scavenger DTT. Two main reasons made us select DTT over other singlet oxygen quenchers. First, DTT does not quench the fluorescence of psoralen derivatives, whereas other singlet oxygen quenchers such as sodium azide, DABCO and P-carotene quench, to some extent, the fluorescence of psoralen derivatives according to our fluorescence quenching experiments, as well as a literature report242. The fluorescence 105 quenching should result in diminished efficiency of cycloaddition reactions of psoralen derivatives with the pyrimidine residues of nuclei acids. Second, it is well established that mercaptan compounds act as good repair agents for DNA damage243. As expected, we found that DTT in low concentration rendered little protection of DNA from nicking. Surprisingly, a concentration as high as 0.1M, DTT demonstrated protection only to a small extent. This result implied that radical species must be produced in the

DNA interior, so that they are out of the reach of DTT molecules in the bulk solution.

Reactive oxygen species sensitized by psoralen derivatives cannot be major sources of

DNA cleavage.

Finally, the lifetime of singlet oxyger> is 13 times longer in deuterium oxide than that in water244. If singlet oxygen is responsible for the DNA scission, the sensitizer should induce the DNA scission more efficiently in deuterium oxide. This solvent isotope effect has been successfully demonstrated in DNA nicking245 and the inactivation of yeast cell processes246, in which singlet oxygen is involved. Thus, DNA photocleavage experiments were also performed in a medium containing about 90% deuterium oxide. However we again failed to observe any meaningful isotope effect on the DNA photocleavage. This result further supports that oxygen effect in these experiments is too small to be detectable by gel electrophoresis.

III.6 Light Induced Viral Inactivation

Musajo era/153 were the first to demonstrate photochemical inactivation of

DNA viruses with psoralen in 1965. By culturing cells in a medium containing radioactive thymine, Cole165 then demonstrated in 1969 that interstrand cross-links were formed by TMP with X phage DNA in vitro , E. coli bacterial DNA in intact cells, and mammalian cell (mouse leukemia-lymphoma L5178Y cells) DNA in vivo.

However, it is still controversial to date whether cross-linking adduct or monoadduct is responsible for the biological effects of psoralen derivatives. It seems that both cross­ 106 linking addition and monoaddition, as well as the damage to other cellular components,

altogether contribute to the biological effects159.

With good results obtained from DNA photocleavage experiments, these new

sensitizers were tested for viral inactivation. X phage was chosen to be our model virus

because of its ease of manipulation. It is a protein coated double stranded DNA

bacteriophage.

The efficacy of viral inactivation induced by a sensitizer was determined by the

plaque forming assay. Specifically, X phage in the presence of a sensitizer was

irradiated for a specified period of time. The irradiated X phage was then allowed to

infect E. coli bacteria. The survival rate of X phage was deduced from plaque formation

of the virus on a solid medium. As demonstrated by comparative experiments, in all

cases, the activity of X phage was not reduced by the presence of a sensitizer in the

dark, and irradiation alone did not affect the activity of X phage.

Table 9. Logs of viral inactivation (VI) of X phage resulted from 350nm light ______photolysis in the presence of acridine sensitizers (21.6kJ/m2). ______

Sensitizer Acr-NH2 Acr-NH2 Acr-I Acr-I Acr-I Acr-Ia Acr-I Concentra­ tion (|iM) 500 50 500 50 10 10 5.0

Logs of VIb 3.7 0.3 7.5 7.5 3.9 4.2 1.9 а. Experiment was performed under dcoxygcnatcd conditions, b. The initial viral litre was б.0x lO^pfu/mL. Logs of VI were ±0.5log. This is the same for all of other VI experiments.

We first performed the viral inactivation experiments using acridine derivatives

as sensitizers with 350nm light at a fixed dosage of 21.6kJ/m2. Under these conditions,

a 0.31og and a 3.71og reduction o f viral activity was observed in the presence of 50 and

500^iM of Acr-NH 2 , respectively. On the other hand, 50jiM of Acr-I produced

complete inactivation of X phage under the same conditions. The log reduction of viral 107 activity diminished accordingly as the concentration of Acr-I was lowered below 50pM

(Table 9). Notice that the same reductions in viral titre were achieved with 10(iM Acr-I

in the presence and the absence of oxygen with 350nm light.

0 c #o -1 L.0 3 b. -2 w> c -3 ’> • Acr-NH2 1 a -4 o Acr-I Cfl W> -5 3 6

7

•8 0 2 4 6 8 10 1 2 1 4 1 6 L ig h t D o sag e(k J/m 2)

Figure 19. Action spectra of X-viral inactivation with Acr-I and Acr-NH2 on photolysis with 350nm light.

We also performed the viral inactivation experiments at a fixed concentration of sesitizers with varying light dosage. According to the viral inactivation data of Figure

19 and Table 10, complete inactivation (7.51ogs) of X phage was achieved in the presence of 50pM of Acr-I, upon photolysis by 350nm light with a dose of 13.5kJ/m2.

Nonetheless, there was no appreciable reduction of viral activity, when X phage was photolyzed in the presence of 50pM of Acr-NH 2 under the same conditions. In line with the results of DNA photocleavage experiments, these results shows that Acr-I is a Table 10. Data of viral inactivation of lambda phage with 3-amino-6-icxioacridine(ACT-I) and proflavine(Acr-NH 2) plus light(350nm).

foradiation time(min) 2 4 ( 8 10 12 14 16 18 20

Light dosage(kJ/m2) 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 13.5 15.0

-Alog(Acr-I+hv) 0.63 1.28 1.74 2.24 3.32 3.72 4.96 6.32 7.50

-Alog(Acr-NH 2 +hv) 0.16 0.08 0.16 0.26 0.30 0.19 0.34 0.35 0.34 0.35

Table 11. Data of viral inactivation of lambda phage with 3-amino-6-iodoacridine(Acr-I) and proflavine(Acr-NH 2) plus light(400~440nm) under different conditions.

Irradiation time(min) 1 2 3 4 5 6 7 8 9 10 15 20 25 30 35 40

Light dosage(J/m2) 15.6 31.2 46.8 62.4 78.0 93.6 109 125 141 156 234 312 390 468 546 624

-Alog(hv) 0.21 0.24 0.24 0.28 0.24 0.21 0.21 0.21 0.17 0.21 0.12 0.14 0.12 0.12 0.14 0.17

-Alog(Acr-I+hv) 0.92 2.08 3.46 5.08 6.10 7.40

-Alog(Acr-NH 2 +hv) 0.24 0.46 0.72 1.38 1.76 2.15 2.58 3.00 3.61 4.23

-Alog(Acr-I-02+hv) 0 0.09 0.13 0.15 0.38 2.74 3.30 4.96 5.56 7.45

-Alog(Acr-NH2-02+hv) 0.02 0.03 0.04 0.06 0.04 0.04 0.06 0.07 0.06 0.70 1.33 2.09 3.39 3.89 4.29

-Alog(Acr-I+DTT+hv) 0.05 0.29 0.84 0.97 1.27 1.72 1.86 2.18

-Alog(Acr-NH 2+DTT+hv) 0.14 0.51 0.87 1.48 1.72 2.18 2.60 2.97 109 far more powerful sensitizer of viral inactivation than its non-halogenated counterpart

Acr-NH* However, it is important to note that Acr-I absorbs 350nm light more strongly than does Acr-NH 2 , which may contribute to these results.

Table 11 and Figure 20 shows the action spectra of X phage inactivation with

Acr-NH2 and Acr-I on photolysis with 400~440nm light (where both acridines have comparable absorbances) under a variety of different conditions. Both Acr-NH 2 and

Acr-I demonstrated efficient sensitization of X phage inactivation. However, the addition of 0.1M DTT protected the virus from inactivation to some extent. The absence of oxygen dramatically reduced the ability of both Acr-NH 2 and Acr-I to inactivate X phage, although Acr-I still inactivated the virus slightly better than Acr-

NH2 . These results are in contradition with the results of DNA photocleavage. This implies that these sensitizers probably target the virus protein capsid instead of viral nucleic acids. The excited states of these sensitizers are quenched by DNA. The lifetimes of their excited states are so short that there is not enough time for molecular oxygen to react with the excited-state acridines, when they are bound to DNA. In viral inactivation experiments, the viral protein capsid is the first hydrophobic region encountered by the sensitizers. Photolysis of the sensitizers in the protein capsid generates their triplet states, which can react with oxygen to make singlet oxygen. The singlet oxygen in turn may lead to the production of other reactive oxygen species.

Together, the active oxygen species are responsible for the viral inactivation under aerobic conditions.

Recall that only those psoralen derivatives which bind DNA relatively poorly exhibit oxygen effects both on DNA photocleavage and cell inactivation243- 244 and those psoralen derivatives which bind strongly to DNA exhibit little oxygen effect193*.

We believe that the strong dependence of viral inactivation ability on oxygen is a statement that these acridine sensitizers bind the viral protein capsid, as well as the viral

DNA. As a result, the excitation of Acr-I leads to the competitive production of the 110

reactive oxygen species and the free radical Acr- in the viral protein capsid. The free

radical damage hence may be directed at the viral protein capsid, the viral DNA, or both

targets.

e o - 2 - ‘■Su «u - 3 - c > -4 - u> C /3 -5 - Light OC5B Acr-I J -6- Acr-NH2 Acr-l-02 Acr-NH2-C 2 -7 - Acr-I+DTT Acr-NH2+[ f

0 100 200 3 0 0 4 0 0 5 0 0 6 0 0 Light Dosage(J/m2)

Figure 20. Action spectra of X-viral inactivation with Acr-I and Acr-NH2 on photolysis with 400~440nm light under different conditions.

Another important aspect of these experiments is that, when photolyzed with different lights sources, these acridine derivatives exhibited the striking differences in the efficacy of viral inactivation and their oxygen dependencies. With the same concentration of sensitizers (50(iM), Acr-NH 2 showed little X phage viral inactivation with a dosage of 21.6kJ/m2 for 350nm light but considerable viral inactivation with only a dosage of 156J/m2 for 400-420nm light. Similarly, for Acr-I to completely inactivate X phage took a dosage of 21.6kJ/m2 for 350nm light but only that of 93.6J/m2 for 400-420nm light. Both compounds absorb light more strongly around 420nm than at I ll 350nm. However the different ability of light absorption can not solely account the tremendous difference in the viral inactivation efficacy, particularly in the case of Acr-

I, which has only a small difference in light absortion at these two wavelength. We reported in the section of photophysics of this chapter that Acr-I fluoresces much more intensely when excited at the band of short wavelength than it does when excited at the band of long wavelength. The difference in the fluorescence quantum yields also varies with the change of solvent as well. Kasha-Valivov’s rule247 states that radiative processes always proceed from the lowest electronic state of a given spin multiplicity, independent of the energy of the eletronic state initially excited. Consequently, the fluorescence quantum yield is usually independent of the energy of the electronic state excited initially. This rule clearly does not apply for this compound. Therefore, the different photophysical behavior of different absorption bands of these two acridine derivatives may contribute to the greate differences of their viral inactivation efficacy with different lights.

The ability of new psoralen derivatives to inactivate X phage is shown in Figure

21 and Figure 22. As a reference compound, the X phage inactivation by AMT was also included. All action spectra were obtained by photolysis with 350nm light using the same batch of phage in the presence of a sensitizer at a concentration of 50(iM.

Consistent with the results of DNA unwinding, HEP demonstrated a limited viral inactivation ability. Chlorinated psoralen CEP showed an enhanced inactivation ability over HEP. Nonetheless, both compounds were less efficient in viral inactivation than AMT. Considering these results, as well as those of DNA unwinding and nicking, it seems that both the cycloaddition reactions and the reactions leading to DNA scission, if the psoralen derivatives target the viral DNA, contribute to viral inactivation. Brominated psoralen BEP turned out to be a better sensitizer than AMT under these conditions. Iodide compound IEP only had a comparable ability to 112 inactivate X phage with CEP. Tribrominated psoralen TBP was found to be the most

powerful sensitizer for viral inactivation of X phage.

Table 12. Data of the viral inactivation of A. phage with our new psoralen derivatives ______and 350nm light. ______

Light dosage (kJ/m2) 0.75 1.50 3.00 4.50 6.00 7.50 9.00 10.5 12.0

-Alog (HEP) 0.18 0.70 1.40 2.03 2.91 3.85 4.51

-Alog (HEP-02) 0.11 1.00 1.96 3.11 3.81 4.23 4.66

-Alog (HEP+DTT) 0.18 0.61 1.40 1.66 2.96 3.35 4.13 4.35 5.26

-Alog (CEP) 0.96 1.93 4.00 5.14 6.19 7.50

-Alog (CEP+DTT) 0.44 0.74 1.65 2.85 3.63 4.70 5.63 6.23

-Alog (BEP) 1.15 3.85 7.23

-Alog (BEP-O 2 ) 1.37 3.37 7.30

-Alog (BEP+DTT) 1.16 2.25 4.41 7.20

-Alog (BEP in D20 ) 1.20 3.94 7.20

-Alog (IEP) 0.58 1.49 3.96 4.41 5.81 7.11

-Alog (IEP+DTT) 0.76 1.24 2.04 2.67 3.29 4.29 5.72 7.50

-Alog (TBP) 5.16 7.50

-Alog (TBP+DTT) 3.62 7.50

-Alog (BCP) 1.21 3.09 7.50

-Alog (BCP+DTT) 0.81 2.88 4.27 7.11

-Alog (BPP) 0.53 0.89 2.18 2.21 3.07 3.70 4.20

-Alog (BPP+DTT) 0.42 0.62 1.16 1.45 1.71 1.91 2.34 2.70 3.00

-Alog (AMT) 0.45 1.05 2.81 4.40 7.11

-Alog (AMT+DTT) 0.61 1.48 2.73 4.11 5.22 7.26 Figure 21. Action spectra of viral inactivation of of inactivation viral of spectra Action 21. Figure Alog reduction of viral survival -5.5 -4.5- -3.5- -2.5 -1.5- -0.5- -7- -4- -3- - 2 - 0 0 light with HEP, CEP, BEP tyid IEP. tyid BEP CEP, HEP, with light 2 1 4 2 6 3 BEP-02 inBEP D20 BEP BEP+DTT HEP+DTT • HEP-02 « 8 4 ih oae( / 2) j/m (k dosage Light 10

12 5 - -7- - -5- 6 6 X - - phage upon photolysis of 350nm 350nm of photolysis upon phage IEP+DTT IEP C&> CEP+DTT 10 10 113 Figure 22. Action spectra of viral inactivation of of inactivation viral of spectra Action 22. Figure Alog reduction of viral survival -4.0 -4.5 -3.5 -3.0- -2.5 - -1.5 - -0.5 2 1.0 0.0 T . -7- - 0 •5- -4- -3- - 6 2 0.0 - 0 - - light with TBP. BCP, BPP and AMT. and BPP BCP, TBP. with light TT 2 0.4 TT 4 0.8 6 BPP+DTT • a 8 BPP 1.2 TBP+DTT TBP ih oae( / 2) j/m (k dosage Light 10

12 1.6 -5- -4- •3- - -7- • -5- -4- -3- - X 2 6 2 phage upon photolysis of 350nm 350nm of photolysis upon phage - 1 0 - - 0 1 2345678 2 3 * n AMT+DTT AMT BCP+DTT BCP 4 5 114 115 The cinnamyl group on the side chain of BCP hardly affected the viral inactivation ability of psoralen derivative. A pyrimidinium side chain of BPP, however, dramatically reduced viral inactivation efficacy. This similar viral inactivation efficacy showed by BEP and BCP but reduced that displayed by BPP is again in line with the results of DNA unwinding and nicking. In other words, the pyridinium ion of BPP quenches part of cycloaddition reactions of psoralen with the pyrimidine residues of

DNA. The parallel results obtained from DNA unwinding and nicking experiments and viral inactivation experiments indicate that these psoralen derivatives indeed target X phage viral DNA in the viral inactivation process.

It is important to notice, however, that unlike the acridine sensitizers, all of these psoralen derivatives showed no obvious difference in viral inactivation efficiency in the presence and the absence of oxygen (only the data of HEP and BEP were shown in the action spectra). The quencher DTT at a concentration as high as 0.1 M could only slightly protect virus from inactivation. Futhermore, viral inactivation was also examined with BEP or AMT in about 85% deuterium oxide. Again, no appreciable difference was observed. All of these results are in accord with the DNA photocleavage experiments. These results are consistant with our prediction that these new psoralen derivatives target the nucleic acids of X phage, but do not prove this assertion.

III.8 Possible Reaction Mechanisms of Sensitizers with DNA

Thermodynamically, the singlet state energy of Acr-I apparently is not sufficiently large to directly cause the cleavage of the C-I bond of this compound.

However, Voelk198 has clearly demonstrated that photolyzing Acr-I in a variety of solvents with 350nm light affords products derived from the aromatic radical Acr-. A reasonable mechanism for the production the radical Acr- is the single elctron transfer reaction248 depicted in Scheme 20. when a ground-state of Acr-I molecule is excited, the resulting singlet state of Acr-I can acquire an electron from purine bases, 116 particularly from guanine residues, of DNA to form a radical anion. This primary ET product forms a secondary product, the aromatic radical Acr-, when the C-I bond of the radical anion breaks. This radical is very reactive and is able to abstract a hydrogen atom from various molecules, including the sugar moiety of DNA, to become 3- aminoacridine. The sugar radical can initiate the fragmentation of the DNA backbone according to the reaction mechanism of Scheme 3. Moreover, the other primary ET product, the radical cation of a purine base, can also lead to the scission of DNA helix

(vide infra). This mechanism is corroborated by fluorescence quenching experiments.

* Donor -Donor t

R-X

Acr*

Scheme 20. Reaction mechanism of DNA-bound Acr-I initiated by photoactivation.

Based on our experimental results and literature reports, we believe that our psoralen derivatives also react with DNA by the single electron transfer mechanism as presented in Scheme 21, though the establishment of the detailed mechanism will require more investigation. 117

Ps-X ’[Ps-X]* ISC>» 3[Ps-X]* Ps-X* + Base ^ [Exciplex] Ps-X * + Baset

Ps-X7 + T ------► Ps-X — T *

Ps-X — T * + B a s e t ► Ps-X <> T + Base

Baset + G ------► Base+ G t

G t ------► ’G(-H) + H +

*G(-H) + Sugar ------► G + ’ Sugar(-H)

Ps-X7 ► Ps’ + X" Ps# + R -H ► Ps-H + R*

Scheme 21. Possible reaction mechanism for DNA-bound halogenated psoralen derivatives upon UVA irradiation. Ps-X and Base represent the halogenated psoralen derivatives and any nucleic acid bases. T and G are thymine and guanine residues of nucleic acids. Ps-X oT is the cycloadducts of psoralen derivatives and thymine.

A DNA-bound psoralen derivative molecule Ps-X is first excited to its singlet excited state upon the absorption of a UV photon. The psoralen singlet excited state can proceed to its triplet state by intersystem crossing (ISC), if it has a long enough lifetime. However, as mentioned earlier, it is more likely that the excited singlet molecule interacts with a nucleic acid base to form an exciplex, which in turn forms a psoralen radical anion and base radical cation. The resulting psoralen radical anion produced by single electron transfer then adds to a thymine residue to form a psoralen- thymine adduct radical anion. This radical anion can donate an electron back to the base radical cation yielding the psoralen-thymine cycloadduct Ps-X oT. If this cycloaddition occurs at the furan moiety of psoralen, the monocycloadduct is capable of absorbing 118 another photon. When appropriately positioned, the excited monocycloadduct can form

a cross-linked diadduct by a similar mechanism. The formation of these monoadducts

and cross-linked diadduct is the commonly accepted mechanism responsible for the

biological activities of psoralen derivatives. These cycloaddition reactions are manifested as DNA unwinding.

According to the reduction potentials of nucleic acids, the nucleic acid base radical cation resulting from single ET is most likely to be a guanine radical cation. If the original radical cation center appears on an adenine residue due to a proximity effect, it can acquire an electron from a nearby guanine residue as discussed in Section

II.2.2. The guanine radical cation can deprotonate quickly to form radical -G(-H), which ultimately transfers the radical center to a sugar residue of the nucleic acid. The sugar radical then initiates fragmentation of the DNA backbone following the mechanism described in Scheme 3. We believe that this is another cause for the biological effectiveness of psoralen derivatives. This mechanism also explains the observation of

DNA photocleavage induced by psoralen derivatives, which occurs even under deoxygenated conditions.

It must be kept in mind that the reverse of the electron transfer process, i. e. the back electron transfer process generally is very fast. It is probably one of the reasons that there is no specific chemical explanation of DNA photocleavage induced with psoralen derivatives. The efficiency of DNA unwinding and cleaving depends on the competition among the possible cycloaddition reactions, back ET process between the radical anion and radical cation, and all other deactivation avenues of the excited psoralen derivatives. For our halogenated psoralen derivatives, the scission of the carbon-halogen bond competes with back ET. Once this occurs, it blocks back ET and greatly increases the possibility of DNA helix scission. 119

C K = Purines

0 = Pyrimidines

P= Phosphate Ester Linkage

S= Deoxyribose Sugar

Figure 23. Representation of the intercalation of psoralen into DNA.

In addition to the radicals derived from nucleic acid bases, another possible species responsible for the scission of DNA helix is the psoralen radical resulting from the breakage of the carbon-halogen bond. However, it is not likely to be an important one. According to the DNA intercalation model of psoralen derivatives depicted in

Figure 23, this aromatic radical center is located in the middle of the major or minor groove of the DNA helices. The distance of this radical center prevents it from abstracting a hydrogen atom from a sugar residue of DNA. The long lifetime and the poor ability of psoralen azide ADP to either cleave or label DNA supports this geometric explanation. The relatively weak viral inactivation efficacy of the iodopsoralen IEP also corroborates this explanation. Therefore, halogenation of 120 psoralen facilitates single electron transfer reaction and deters the competitive back electron transfer process. The resulting psoralen radical of single electron transfer reaction itself does not play an important role in the scission of DNA helix.

III.8 Conclusion

Consistent with our design principles, Acr-I is a more powerful reagent than

Acr-NH2 both in DNA photocleavage and X phage viral inactivation experiments. The absence of oxygen effects in DNA photocleavage means that free sensitizer molecules do not exist in large amount in DNA solutions and thus do not play an important role in

DNA cleavage. The effectiveness of Acr-I over Acr-NH 2 in cleaving DNA hence can be attributed to the facility of Acr-I reduction via a single ET reaction. In contrast, the prominent oxygen effects observed in the X phage viral inactivation experiments indicate that these acridine derivatives not only bind to DNA, but also other cellular components, when they are applied to real biological systems. Consequently, the free radical damage may directed at both the viral protein capsid and the viral DNA.

Furthermore, the different efficiencies in viral inactivation with different light sources displayed by Acr-I are a consequence of its violation of Kasha-Wavili’s rule, as well, as its different extinction coefficients.

In regard to the photophysical properties of psoralen derivatives, neither the maximum wavelength of UV-vis absorption nor the fluorescence of 8-MOP correlates with the general solvent polarity scale, Et (30), nor with a macroscopic solvent polarity scale such as dielectric constant. However, there do exist good linear solvation energy relationships of the maximum wavelength of UV-vis absorption and fluorescence of 8-

MOP with Taft’s combined solvatochromic parameters, rt* + ca. These results indicate that the ability of solvents to form hydrogen bonds, as well as its polarity and polarizability, affects the ground-state and the excited-state energy levels of 8-MOP.

Specifically, polar solvents stabilize polar 8-MOP molecules in both its ground state 121 and excited state. 8-MOP molecules either in the ground state or in the excited state can form hydrogen bonds with solvent molecules if the solvent molecules can serve as hydrogen donors in the formation of hydrogen bonds. The greater weighting coefficient c in the case of fluorescence, compared to that in the case of UV-vis absorption, indicates that excited 8-MOP molecules are more likely to form hydrgen bonds than molecules in the ground state. Such a result is in agreement with the characteristic of a

(re, 7t*) transition. The existence of hydrogen-bonded singlet excited 8-MOP species, in addition to that of free singlet excited 8-MOP, is also in accord with the observation of non-single exponential fluorescence decay displayed by psoralen derivatives.

Moreover, the formation of hydrogen bond, together with the proximity effect, explains the fact that the fluorescence quantum yield of 8-MOP diminishes as a result of the sequential increase of the water content in the dioxane-water mixture after a maximum point and that protons quench the fluorescence of psoralen derivatives.

Attaching an ammonium ion side chain on a psoralen ring not only greatly increases the solubility of psoralen derivatives in aqueous medium, but also enhances their binding affinity with DNA. The latter effect is believed to be a consequence of

Coulombic interactions betweeen the two oppositively charged ions, ammonium and phosphate171®. Earlier work in this laboratory has demonstrated that psoralen derivatives with quaternary ammonium ion side chains penetrate model biological membranes more readily than those with other kinds of ammonium ion side chains224.

Moreover, psoralen derivatives with quaternary ammonium ion side chains, unlike those with other kinds of ammonium ion side chains, also exhibit preferential binding with nuclei acids over model biological membranes207 224.

Halogenation of the aromatic rings of psoralen dramatically changes both the photophysical and photochemical behavior of psoralen derivatives. First, it enhances the binding ability of psoralen derivatives with DNA due to an increased hydrophobicity. Consequently, the stronger DNA binding ability reduces the possible 122 involvment of singlet oxygen effects in viral inactivation, when these psoralen derivatives serve as sensitizers. Second, halogen atoms on the psoralen ring shift the maximum wavelength of UV-vis absorption to longer wavelength and increase their extinction coefficients according to their conjugative ability. Third, due to the heavy atom effect, halogen atoms accelerate the intersystem crossing process from excited singlet state to triplet state in the order of their atomic weights. Four, halogenation facilitates reductive single electron transfer reaction. The rate of carbon-halogen bond scission reaction of the halogenated radical anions should proceed in the order of the strength the carbon-halogen bonds. This reaction will compete with back electron transfer process, after the forward ET occurs.

The absence of oxygen effects with our new psoralen derivatives in both DNA photocleavage and X phage viral inactivation experiments is a result of their strong and selective binding ability with DNA. Two reasons may contribute to the exclusion of oxygen effects in these experiments. First, due to the proximity effect, the excited states of these DNA-bound psoralen derivaitves can be quenched by nucleic bases rapidly through single electron transfer process, which leaves little time for the involvement of molecular oxygen. Second, intercalated psoralen derivatives are protected by DNA from the access of molecular oxygen disolved in the aqueous reservoir249.

In summary, we have synthesized and evaluated a series of sensitizers. To our best knowledge, some of our psoralen derivarives are the most efficient drugs to specifically target viral nucleic acids. Based on our experimental results , we have also proposed the mechanisms of the action of these drugs in viral inactivation and, for the first time, explained DNA unwinding and nicking with psoralen derivatives on a molecular level. Chapter IV

EXPERIMENTAL

IV.I Materials

Acr-NH2 was purchased as the hydrochloride from Aldrich Chemical Co.

(Milwaukee, WI) and was purified by neutralization with potassium carbonate and recrystallization from ethanol. AMT was obtained from HRI Inc.(Concord, CA). Calf thymus DNA (Type I, highly polymerized), CpG, GMP, Tris base, ethidium bromide and agarose (Type II, Medium EEO) were the products of Sigma Chemical Co. (St.

Louis, MO). Supercoiled plasmid pBR322 DNA (0.25mg/mL) and DNA molecular weight marker II (0.25mg/mL) were purchased in Tris-EDTA (TE) buffer from

Boehringer Mannheim (Indianapolis, IN). Deuterated solvents were obtained from

Isotec Inc. (Miamisburg, OH). All other chemicals were the products of Aldrich

Chemical Co. and were used directly without purification unless otherwise specified.

TLC plates were from EM science (Gibbstown, NJ).

E. coli bacterial (#23724) and X phage (#23724-B2) strains were obtained from

American Type Culture Collection (ATCC, Rockville, MD). Bacto-tryptone, Bacto- yeast extract and Bacto-agar were from DIFCO Laboratories (Detroit, MI). Polystyrene disposable petri dishes were purchased from VWR Scientific (West Chester, PA).

Acrodisc syringe filter (0.2|im) was the product of Gelman Sciences (Ann Arbor, MI).

123 124 Tris HC1 buffer (Tris-HCl, 10mM,; NaCl, 0.2M; pH=7.4) and phosphate buffered saline (PBS, pH=7.4) were made by the Reagent Laboratory of The Ohio State

University (Columbus, OH).

The photolysis reactors were built at the Machine Shop of the Department of

Chemistry, The Ohio State University. The lamps required for photolysis were obtained from the Southern New England Ultraviolet Company (Branford, CT).

IV.2 General Procedures

UV-vis spectra were taken on a Spectronic 3000 Array spectrophometer equipped with RapidScan software of the Milton Roy Company. Infrared spectra were recorded with a Perkin Elmer 1600 Series FTIR spectrophotometer. *HNMR and

13CNMR spectra were recorded with a Bruker AM-200 (200MHz) or Varian AM-250

(250MHz) instrument. Mass spectra were obtained on a Kratos MS-30 mass spectrometer by the operator at the Campus Chemical Instrumentation Center at The

Ohio State University. Fluorescence spectra and fluorescence quenching experiments were performed on a Perkin Elmer LS 5 fluorescence spectrometer. EPR measurements were made with a Varian E-l 12 X-band EPR spectrometer. Melting points were taken with an Electrothermal melting point apparatus.

Reduction potentials were measured on OMNI 90 Analog Potentiostat (Cypress

Systems, Inc., Lawrence, KA) cyclic voltammetry vs. Ag/AgCl with 0.1M tetrabutylammonium perchlorate (TBAP) in DMSO.

The laser flash photolysis system consists of a Lumonics TE-861-4 excimer laser (XeF, 351nm, 17ns, 50mJ) and a Lambda Physik LPX-100 excimer laser (XeCl,

308nm, 17ns, 150mJ). A monitoring beam of Xe arc lamp was set at a right angle to the excimer laser pulse. Signals were collected with a photomultiplier and converted to a digitized signal by a Tektronix 7912 transient digitizer. Transient absption spectra were 125 obtained with an EG and G Princeton Allied Research Model 1460 Optical

Multichannel Analyzer (OMA).

Incubation of bacteria and virus was carried out in a Gravity Convection

Incubator (Precision Scientific, Inc. Model 4) equipped with an orbit shaker (Lab-line

Instruments, Inc.). Centrifugation was performed with an international centrifuge

(International Equipment Co.).

IV.3 Synthesis

3-acetamido-6-aminoacridinel99a A solution of 3.3mL (33mmol) of acetic anhydride and 25mL of glacial acetic acid were slowly added to a mixture of 6.27g (30.0mmol) of

3,6-diaminoacridine in lOOmL of glacial acetic acid at 100°C. After stirring for lhr, the reaction mixture was poured onto about 500mL of ice water and neutralized with aqueous NaOH. The yellow precipitate was collected by suction filtration in quantitative yield. The crude product was purified by flash chromatography (eluent: hot acetone until most of the diacetylated byproduct was eluted, then ethanolrethyl acetate: triethylamine=6:3:l) to yield 4.23g (56%) of crimson product, m.p. 183-5°C (dec), some yellow diacetylated product, and their mixture. The mixture was combined with the next batch of crude product for purification. IR (KBr, film): 3402, 3312, 3205,

3034, 1691, 1634, 1537, 1483, 1371, 1264, 1174, 806cm1. !HNMR (DMSO-d6, ppm);

10.25 (br, 1H), 8.54 (s, 1H), 8.31 (s, 1H), 7.85 (d, J=9.0Hz, 1H), 7.75 (d, J=9.0Hz, 1H),

7.45 (dd, Ji=9.0Hz, J 2 =1.9Hz, 1H), 7.00 (dd, Ji=9.0Hz, J 2 =1.9Hz, 1H), 6.87 (d,

J=1.9Hz, 1H), 6.06 (br, 2H), 2.13(s, 3H). 13CNMR (DMSO-d6, ppm); 168.3, 150.7,

150.4,147.3, 139.8, 134.3, 128.9, 128.4, 120.2, 119.6, 119.5, 117.3, 113.0, 102.1,23.4.

High resolution MS (m/e, intensity%): calcd for C 15H 13N3 O (M+) 251.1059, found

251.1057 (89.30), 252 (M+1+, 16.84), 209 (100). 126

3 -amino- 6 -iodoacridine,99b l.OOg (4.00mmol) of 3-acetamido-6-aminoacridine was suspended in 200mL of 6 N hydrochloric acid and cooled with an ice-salt bath. A solution of 0.33g (4.8mmol) of sodium nitrite in 3mL of water was added dropwise to the above suspension. 5mL of 20% aqueous urea solution was then introduced to remove the extra sodium nitrite 5 min later, which was followed by slowly adding the aqueous solution of 0.8 lg (4.8mmol) of potassium iodide in 8 mL of water (approximate

30min). The reaction mixture was then refluxed for lhr and stirred at room temperature overnight. The brown solid was collected by filtration and recrystallized with ethanol (a little bit of activated charcoal was used to decolorize the product). The harvested orange-red crystals were dissolved again in boiling ethanol and neutralized with dilute sodium hydroxide aqueous solution. After collection by filtration, the product was once again recrystallized from ethanol-benzene to finally yield 0.537g (37%) of russet solid, m.p. 220-2°C (lit199b. 224-225°C). IR (KBr, film): 3380, 3154,1634, 1600, 1452,1405,

1215, 1157, 912, 794cm-1. !HNMR (DMSO-d6, ppm); 8 . 6 6 (s, 1H), 8.31 (s, 1H), 7.80

(d, J=9.0Hz, 1H), 7.72 (d, J=8.7Hz, 1H), 7.58 (dd, J,=8.7Hz, J 2 =1.6Hz, 1H), 7.10 (dd,

J!=9.0Hz, J 2 =2.0Hz, 1H), 6.89 (d, J=2.0Hz, 1H), 6.20 (br, 2H). 13CNMR (DMSO-d6, ppm); 151.1, 150.9, 148.8, 135.5, 135.0, 130.4, 129.6, 129.0, 122.1, 120.9, 120.7,

102.0,96.7. High resolution MS (m/e, intensity%): calcd for C 13H9 N2 I (M+) 319.9811, found 319.9844 (100), 321 (M+1+, 15.57), 193 (34.40), 192 (18.54).

8 -Hydroxypsoralen 201 l.OOg (4.60mmol) of 8 -methoxypsoralen was dissolved in

20mL methylene chloride (distilled frashly from P 2 Os) in a 50mL three-neck round bottom flask purged with an argon flow. lOmL (lOmmol) of 1M boron tribromide solution in hexane was added to the above solution in one portion. The mixture was stirred for 6 hrs at room temperature, and then left overnight. A small amount of water was added carefully to the reaction mixture to destroy the extra BBr 3 . Then another

30mL of water was added and the mixture was stirred for 5hrs. The crude, grey product 127 was collected by suction filtration, and recrystallized from acetonitrile to afford 0.584g

(62%, lit201. 54%) of yellowish needles, m.p. 250-2°C (lit . 201 247-9°C). !HNMR

(acetone-d6, ppm): 9.36 (br, 1H), 8.00 (d, J=9.6Hz, 1H), 7.90 (d, J=2.2Hz,lH), 7.41 (s.

1H) 6.70 (d, J=2.2Hz, 1H), 6.31 (d, J=9.6Hz, 1H).

8-(3-Bromopropyloxy)psoralen202 To a solution of 0.50g of 8 -hydroxypsoralen and

6 g of 1,3-dibromopropane in 50mL acetone, 2.5g of anhydrous K 2 CO3 was added. The mixture was refluxed for 24hrs. After the solid was filtered from the cooled mixture and washed with acetone, the combined acetone filtrate was evaporated to yield a red liquid, to which 20mL hexane was added. Soon a solid yellow precipitate was formed. The crude product was collected and recrystallized from benzene-hexane to give 0.603g

(75%, lit. 202 48%) of white solid, mp. 104-5°C (lit .202 1 33°C). JHNMR (acetone-d6, ppm): 8.01 (d, J=9.6Hz, 1H), 7.93 (d, J=2.2Hz, 1H), 7.61 (s. 1H) 7.02 (d, J=2.2Hz,

1H), 6.33 (d, J=9.6Hz, 1H), 4.61 (t, J=6.4Hz, 2H), 3.84 (t, J=6.4Hz, 2H), 2.39 (tt, j 1=j 2 =6.4Hz, 2H).

8-Methoxy-5-nitropsoralen196 203 4.32g (20.0mmol) of 8 -methoxypsoralen was dissolved in 50mL of glacial acetic acid with warming, then cooled to room temperature. 30mL of fuming nitric acid (90%, gr. 1.53) was added, dropwise, with strong stirring. The addition of nitric acid proceeded at such a rate that the reaction temperature was kept below 15°C with the help of an ice and water bath. As nitric acid was introduced, a yellow solid precipitated immediately. After the addition was complete, the reaction mixture was stirred for 2 0 min at room temperature, and then was poured onto ice. The amorphous yellow solid was collected by suction filtration and washed with a large amount of water. After drying, the crude product was refluxed in 150mL of absolute ethanol for one hour and collected after cooling to give 4.54g

(87%, lit196-. 96%) of final product, mp. 234-7°C (lit196-. 235-8°C). IR (KBr film, cm' 128

»): 3146, 2964, 1741, 1574, 1510, 1318, 1124, 836, 788. »HNMR (CDCI 3 , ppm): 8.80

(d, J=10.3Hz, 1H), 7.87 (d, J=2.3Hz, 1H), 7.44 (d, J=2.3Hz, 1H), 6.64 (d, J=10.3Hz,

1H), 4.50 (s, 1H). ,3CNMR (DMSO-d6, ppm): 157.4,150.9,138.9,136.3,123.2, 117.7,

115.0, 111.3, 106.0, 60.8. High resolution MS (m/e, intensity%): calcd for C 1 2 H 7 NO6

261.0273, found 261.0272 (100), 262 (13.75, M++1), 215 (29.14), 187 (13.43).

8-(3-BromopropyloxyI)-5-nitropsoraIen This compound was prepared following the same procedure used to prepare 8-methoxy-5-nitropsoralen. 3.23g (lO.Ommol) of 8-(3- bromopropyloxy)psoralen produced 3.03g (82%) of bright yellow crystal, mp. 114-5°C.

IR (KBr film, cm*1): 3159, 3128, 2974, 1733, 1589, 1497, 1317, 1164, 830, 784.

1HNMR (DMSO-d6, ppm): 8.60 (d, J=10.2Hz, 1H), 8.43 (d, J=2.2Hz, 1H), 7.37 (d,

J=2.2Hz, 1H), 6.75 (d, J=10.2Hz, 1H), 4.77 (t, J=6.0Hz, 2H), 3.75 (t, J=6.5Hz, 2H),

2.34 (quaintet., J=6.2Hz, 2H). 13CNMR (CDCI 3 , ppm): 158.1, 149.5, 145.2, 142.4,

139.1, 136.5, 124.2, 118.5, 112.2, 107.4, 71.8, 32.8, 29.5. High resolution MS (m/e, intensity%): calcd for Ci 4 H i0 BrNO6 366.9691, found 366.9692 (30.05), 370 (10.21),

369 (30.71), 368 (10.21), 247 (100), 201 (64.93).

5-Amino-8-methoxypsoralen 196 3.61g (16.0mmol) of stannous chloride dihydrate and

2.0g of zinc metal was added at once to a suspension of 1.04g (4.00mmol) of 8 - methoxy-5-nitropsoralen in 5mL of concentrated hydrochloric acid and 20mL of ethanol. The resulting reaction mixture was stirred for one day. The yellow solid then was collected by suction filtration and washed with saturated aqueous sodium bicarbonate solution. The crude product was recrystallized from ethanol to give 0.590g

(64%, lit196 53%) thick yellow needles, mp. 238-45°C (dec., lit 196 244-5°C). IR (KBr film, cm '1): 3479, 3385, 3106, 2957, 1708, 1484, 1138, 808, 742. *HNMR (DMSO-d 6 , ppm): 8.32 (d, J=9.8Hz, 1H), 7.85 (d, J=2.2Hz, 1H), 7.19 (d, J=2.2Hz, 1H), 6.51 (br.,

2H), 6.12 (d, J=9.8Hz, 1H), 3.89 (s, 1H). ,3CNMR (DMSO-d6, ppm): 159.7, 148.9, 129 144.0, 143.7, 140.5, 135.6, 122.4, 110.7, 107.6, 105.1, 99.5, 60.7. High resolution MS

(m/e, intensity%): calcd for C 12H 9 NO4 231.0532, found 231.0533 (75.11), 232 (10.35,

M++1), 216 (100), 188 (12.89).

5-Amino-8-(3-bromopropyloxy)psoralen To a mixture of 0.552g (1.50mmol) of 8 -

(3-bromopropyloxy)-5-nitropsoralen, 2mL of concentrated hydrochloric acid, and 8 mL of ethanol was added at once 1.35g (6.00mmol) of stannous chloride dihydrate. The raising of the temperature and the dissolution of the solid mixture indicated the reaction progress. The reaction mixture was then stirred for 30min at 50°C. After cooling, the yellow solid was collected by suction filtration and recrystallized from ethanol to give

0.426g (84%) of yellow-green crystal, mp. 168-70°C. IR (KBr film, cm"1): 3446, 3364,

3251, 3128, 2964, 1713, 1692, 1646, 1619, 1474, 1446, 1386, 1256, 1135, 827, 742.

!HNMR (DMSO-d6, ppm): 8.34 (d, J=9.8Hz, 1H), 7.88 (d, J=2.2Hz, 1H), 7.21 (d,

J=2.2Hz, 1H), 6.56 (br, 2H), 6.13 (d, J=9.8Hz, 1H), 4.21 (t, J=5.9Hz, 2H), 3.78 (t,

J= 6 .6 Hz, 2H), 2.21 (quintet., J=6.2Hz, 2H). 13CNMR (DMSO-d6, ppm): 159.6, 149.1,

144.3, 143.7, 140.5, 135.8, 120.9, 110.6, 107.6, 105.1, 99.5, 70.9, 32.2, 30.7. High resolution MS (m/e, intensity%): calcd for Ci 4 H i2 BrN0 4 336.9950, found 336.9944

(29.16), 339 (29.69), 217 (100), 216 (95.61).

5-Amino-8-(3-diethylaminopropyloxy)psoralen The mixture of 0.676g (2.00mmol) of 5-amino-8-(3-bromopropyloxyl)psoralen and lmL of diethyl amine in lOmL of acetonitrile was refluxed under an Argon atmosphere for 6 h. To the cooled mixture,

50mL of chloroform was added and the resulting solution was washed with water

(3xl0mL) and dried over anhydrous Na 2 SC>4 . The residue obtained by evaporating acetonitrile was recrystallized from benzene to yield 0.452g ( 6 8 %) of yellow product, mp. 139-40°C. IR (KBr film, cm"1): 3445, 3374, 3081, 2967, 1719, 1713, 1646, 1585,

1477, 1136, 826, 744. JHNMR (CDCI 3 , ppm): 7.82 (d, J=9.8Hz, 1H), 7.58 (d, J=2.3Hz, 130 1H), 6.75 (d, J=2.3Hz, 1H), 6.21 (d, J=9.8Hz, 1H), 4.31 (t, J=6.4Hz, 2H), 4.29 (br.,

2H), 2.73 (t, J=7.4Hz, 2H), 2.56 (quartet, J=6.1Hz, 4H), 1.96 (quintet, J= 6 .8 Hz, 2H),

1.03 (t, J=7.1Hz, 6 H). 13CNMR (CDCI3 , ppm): 160.9, 150.0, 144.9, 144.5, 138.6,

132.4, 124.6, 112.9, 110.7, 103.9, 102.3, 73.1, 49.4, 46.9, 27.6, 11.7. High resolution

MS (m/e, intensity%): calcd for C 18 H 2 2 N2 O4 330.1579, found 330.1569 (3.20), 315

(3.60), 301 (.45), 245 (.74), 216 (6.70), 114 (32.56), 8 6 (100), 72 (13.04), 58 (9.24).

5-Azido-8-methoxypsoralen 1.85g (8.00) of 5-amino-8-methoxypsoralen was dissolved in 15mL of trifluoroacetic acid in 50mL flask cooled with an ice-salt bath. A solution of 0.69g (lOmmol) of sodium nitrite in 5mL of water was slowly added to the above mixture. To the reaction mixture, which was stirred for lOmin after the completion of the addition of sodium nitrite solution, the solution of 0.65g (lOmmol) of sodium azide in 5mL of water was added dropwise. After another 30min, the reaction mixture was transferred to a separatory funnel containing 80mL of chloroform, followed by washing with water until neutral and dried over anhydrous Na 2 SC>4 ,

Rotatory evaporation produced crude brown product, which was purified by flash chromatography on silica gel (eluent: chloroform) to yield 1.48g (72%) of yellow crystals, mp. 125-9°C (dec.). This azide product turned pink when exposed to light or after storage. IR (KBr film, cn r1): 3154, 3129, 2955, 2113, 1732, 1621, 1584, 1469,

1158, 827,748. JHNMR (CDCI 3 , ppm): 8.02 (d, J=10.0Hz, 1H), 7.71 (d, J=2.4Hz, 1H),

7.15 (d, J=2.4Hz, 1H), 6.32 (d, J=10.0Hz, 1H), 4.21 (s, 3H). 13CNMR (CDCI3 , ppm):

159.7, 148.4, 146.2, 143.5, 138.9, 130.1, 121.9, 118.6, 113.9, 108.4, 102.6, 61.5. High resolution MS (m/e, intensity%): calcd for C 1 2 H 7 N3 O4 257.0437, found 257.0432

(11.68), 229 (100), 214 (16.71), 201 (14.76), 158 (24.02), 130 (31.70), 120 (40.57).

5-Azido-8-(3-diethylaminopropyloxy)psoralen This compound was prepared following the exact same procedure applied to the synthesis of 5-azido-8- 131 methoxypsoralen except that the flash chromatography eluent was a mixture of

benzene, diethyl amine, and petroleum ether (8:1:1). 0.330g (l.OOmmol) of 5-amino-8-

(3-diethylaminopropanoxy)psoralen gave 0.228g (64%) of gray solid, mp. 53-4°C. IR

(KBr film, cm‘»): 3157, 3081, 2965, 2111, 1736, 1589, 1467, 1164. 839, 748. !HNMR

(CDC13, ppm): 8.03 (d, J=9.9Hz, 1H), 7.69 (d, J=2.4Hz, 1H), 7.14 (d, J=2.4Hz, 1H),

6.31 (d, J=9.9Hz, 1H), 4.43 (t, J=6.3Hz, 2H), 2.72 (t, J=7.4Hz, 2H), 2.55 (quartet,

J=7.1Hz, 4H), 1.97 (quintet, J= 6 .8 Hz, 2H), 1.02 (t, J=7.1Hz, 6 H). »3CNMR (CDCI3 ,

ppm): 159.9, 146.3, 139.1, 129.4, 122.1, 118.6, 102.8, 73.0,49.2,46.9, 27.7, 11.6. High

resolution MS (m/e, intensity%): calcd for C 1 8 H2 0 N4 O4 356.1485, found 356.1485

(2.80), 341 (1.63), 328 (2.91), 315 (2.29), 217 (2.96), 114 (28.32), 8 6 (100), 72 (9.72),

58(14.13).

S-Bromo-S-methoxypsoralen 2 03- 204 To a suspension of 4.32g (20.0mmol) of 8 -MOP

in 50mL of tetrahydrofuran, was added dropwise, with stirring 4.0g (25mmol) of

elemental bromine in 20 min at room temperature. The reaction mixture turned red

when bromine was added. Precipitate appeared in 5 min. Extra bromine was consumed

by solvent THF in an extended 2hr-stirring or by the addition of 10% sodium

thiosulfate solution. The crude product was collected by filtration and was purified with recrystallization from ethanol to afford 5.38g (91%) of white solid, mp. 178-80°C

(lit.203 185-186°C). •HNMR (DMSO-d6, ppm): 8.22 (d, J=2.2Hz, 1H), 8.11 (d,

J=9.9Hz, 1H), 7.01 (d, J=2.2Hz, 1H), 6.53 (d, J=9.9Hz, 1H), 4.17 (s, 3H).

Alternatively, 5mmol of 8 -MOP and 6 mmol of NBS was refluxed in lOOmL of carbon tetrachloride for 4hr. The solvent was evaporated and the residue was recrystallized from ethanol to afford the product in quantitative yield.

5-Iodo-8-methoxypsoralen 204b' 205 To a mixture of 1.08g (5.00mmol) of 8 -MOP with

1.02g (6.00mmol) of silver acetate in 50mL of dry chloroform, was added dropwise, 132 with vigrous stirring a solution of 1.52g (6.00mmol) of iodine in lOOmL of dry chloroform in the dark. The reaction mixture was stirred for another 3hr after the complete addition of iodine. The precipitated silver iodide was removed by filtration.

The filtrate was evaporated to give a crude product, which was easily purified by recrystallization from chloroform/methanol. The yield was 1.37g (80%) of pale yellow solid, mp. 189-91°C (.lit . 205 191°C) »HNMR (CDCI3 , ppm): 8.01 (d, J=9.9Hz, 1H),

7.74 (d, J=2.3Hz, 1H), 6.79 (d, J=2.3Hz, 1H), 6.37 (d, J=9.9Hz, 1H), 4.30 (s, 3H).

5-Bromo-8-hydroxypsoralen204b This compound was prepared in a yield of 62% using the same procedure as that utilized in the synthesis of 8 -hydroxypsoralen. mp.

254-6°C. *HNMR (DMSO-d6, ppm): 10.99 (br, 1H), 8.19 (d, J=2.2Hz, 1 H), 8.12 (d,

J=9.9Hz, 1H), 6.97 (d, J=2.2Hz, 1H), 6.52 (d, J=9.9Hz, 1H).

5-Bromo-8-(3-bromopropyloxy)psoralen204b 2.00g (7.12mmol) of 5-bromo-8- hydroxylpsoralen , 7mL of 1,3-dibromopropane, and 5g of anhydrous potassium carbonate were mixed in 120mL of acetonitrile. The resulting mixture was refluxed with stirring overnight. The potassium carbonate residue was removed by filtration and washed with hot acetonitrile. The filtrate was concentrated under reduced pressure to about 15mL. Then 80mL of ethyl ether was introduced to the filtrate. The resulting precipitate was collected by filtration to afford 2.37g (83%) of pure product, mp. 96-

8 °C. !HNMR (DMSO-d6, ppm): 8.26 (d, J=2.2Hz, 1H), 8.17 (d, J=9.9Hz, 1 H), 7.07 (d,

J=2.2Hz, 1H), 6.57 (d, J=9.9Hz, 1H), 4.52 (t, J=6.0Hz, 2H), 3.77 (t, J= 6 .6 Hz, 2H), 2.29

(quint, J=6.3Hz, 2H).

5-Chloro-8-(3-bromopropyloxy)psoralen A mixture of 0.650g (2.5mmol) of 8-(3- bromopropyloxy)psoralen with 0.40g (3.0mmol) of NCS in 30mL of benzene was refluxed for 24hr in the presence of 2mL of concentrated hydrochloric acid. The crude 133 product was obtained after evaporation of benzene. Flash chromatography of the crude product with benzene as eluent gave 0.702g (95%) of yellowish solid, mp. 98-100°C.

JHNMR (CDCI3 , ppm): 8.16 (d, J=9.9Hz, 1H), 7.73 (d, J=2.2Hz, 1H), 6.93 (d, J=2.2Hz,

1H), 6.47 (d, J=9.9Hz, 1H), 4.60 (t, J=5.8Hz, 2H), 3.76 (t, J=6.4Hz, 2H), 2.39 (tt,

Ji=6.4Hz, J 2 =5.8Hz, 2H). ,3CNMR (CDCI 3 , ppm): 159.4, 147.2, 146.9, 143.7, 140.0,

130.6, 125.4, 116.5, 115.3, 114.0, 105.6, 71.5, 33.0, 29.8. High resolution MS (m/e, intensity%): calcd for Ci 4 HjoBrC 1 0 4 355.9451, found 355.9466 (16.40), 360 (M+4,

5.45), 358 (M+2, 21.58), 238 (39.72), 236 (100), 210 (9.21), 208 (27.92).

5-Chloro-8-[(3-triethylammonio)propyloxy]psoralen bromide (CEP) 0.500g

(1.70mmol) of 5-chloro-8-(3-bromopropyloxy)psoralen and 5mL of triethylamine were mixed in lOmL of acetonitrile. The mixture was stirred under argon at 50°C for 48hr.

Extra triethylamine and solvent were evaporated under reduced pressure. The resulting residue was recrystallized from ethanol and benzene to afford 0.48 Ig (72%) of white crystals, mp. 203-8°C (dec.). !HNMR (DMSO-d6, ppm): 8.26 (d, J=2.2Hz, 1H), 8.18

(d, J=9.9Hz, 1H), 7.59-7..55 (m, 2H), 7.42-7.28 (m, 3H), 7.10 (d, J=2.2Hz, 1H), 6.99

(d, J=15.7Hz, 1H), 6.59 (d, J=9.9Hz, 1H), 6.47 (dt, J,=15.7Hz, J 2 =7.2Hz, 1H), 4.54 (t,

J=5.6Hz, 2H), 4.13 (d, J=7.2Hz), 3.60-3.52 (m, 2H), 3.36 (q, J=7.1Hz, 6 H), 2.30-2.15

(m, 2H), 1.33 (t, J=7.1Hz, 9H).

5-Bromo-8-[(3-cinnamyldiethylammonio)propyloxy]psoralen bromide (BCP)

0.804g of 5-bromo8-(3-bromopropyloxy)psoralen was mixed with 5mL of diethyl amine in 15mL of absolute ethanol. The mixture was stirred under argon at 50°C for

48hr. Excess diethyl amine and solvent were then evavorated under reduced pressure.

To the resulting residue, were added in sequence 20mL of acetonitrile, 2.0mL of cinnamyl bromide, and 0.5g of anhydrous potassium carbonate. Again, the resulting reaction mixture was stirred under argon at 50°C for 48hr. The potassium carbonate 134 residue was removed by filtration and washed with hot acetonitrile. The combined

filtrate was then evaporated under reduced pressure to give a white solid. The crude

product was recrystallized from ethanol and ethyl acetate to yield 0.674g (57%) of

white solid. >HNMR (DMSO-d6, ppm): 8.29 (d, J=2.3Hz, 1H), 8.27 (d, J=9.7Hz, 1H),

7.20 (d, J=2.3Hz, 1H), 6.62 (d, J=9.7Hz, 1H), 4.51 (t, J=5.6Hz, 2H), 3.52-3.42 (m, 2H),

3.32 (q, J=7.1Hz, 6 H), 2.23-2.13 (m, 2H), 1.24 (t, J=7.1Hz, 9H). MS (m/e, intensity%):

513 (18.46, M+3+), 512 ( 6 6 .8 6 , M+2+), 511 (19.96, M+1+), 510 (64.19, M+), 394

(28.40) 392 (25.44), 117(100).

5-Bromo-8{3-[(4-methoxycarbonyl)pyridinio]propyloxy}psoralen bromide (BPP)

0.804g (2.00mmol) of 5-bromo8-(3-bromopropyloxy)psoralen was dissolved in 5mL of distilled methyl isonicotinate. The mixture was stirred under argon at 50°C for 18hr. A grey pricipitate was formed and 20mL of ethyl acetate was added to ensure complete precipitation of the product. The grey solid was separated by filtration and recrystallized from ethanol and ethyl acetate to afford 0.887g (82%) of yellow crystals, mp. 192-5°C (dec.). *HNMR (DMSO-d6, ppm): 9.45 (d, J=J=6.5Hz, 2H), 8.53 (d,

J=6.5Hz, 2H), 8.29 (d, J=2.2Hz, 1H), 8.17 (d, J=9.9Hz, 1H), 7.08 (d, J=2.2Hz, 1H),

6.58 (d, J=9.9Hz, 1H), 5.04 (t, J=6.4Hz, 2H), 4.58 (t, J=5.5Hz, 2H), 3.99 (s, 3H), 2.59-

2.52 (m, 2H). 13CNMR (DMSO-d6, ppm): 162.6, 159.0, 148.8, 146.7, 145.7, 143.8,

143.0,142.4, 130.0,127.4,127.1, 116.0, 115.1, 107.0, 105.3,70.7, 59.2, 53.8, 30.8. MS

(m/e, intensity%): 461 (31.04, M+3+), 460 (100, M+2+), 459 (31.27, M+1+), 458

(99.79, M+), 178 (28.11), 151 (37.95).

4’-Chloromethyl-4,5’,8-trimethylpsoralen 208 1.14g (5.00mmol) of TMP was dissolved in 150mL of glacial acetic acid with gentle heating. lOmL of chloromethyl methyl ether was added to the resulting solution at once. The reaction mixture stood at room temperature for 24hr, then another lOmL of chloromethyl methyl ether was added 135 to the reaction mixture. The mixture was stoppered for another 48hr and then placed in a refrigerator. Precipitation was collected to give 1.06g (77%, lit208. 62.5%) white solid product, mp. 214-6°C (lit208. 215-7°C ). JHNMR (CDC13, ppm): 7.60 (s, 1 H), 6.27 (s,

1H), 4.75 (s, 2H), 2.58 (s, 3H), 2.53 (s, 3H), 2.52 (s, 3H).

4’-Triethylammoniomethyl-4,5’,8-trimethylpsoraIen lOOmg of 4’-chloromethyl-

4,5’,8-trimethylpsoralen and 2mL of triethylamine were mixed in 5mL of toluene. The mixture was stirred at 80-85°C for 4 days. The precipitate was collected and washed with chloroform. This white solid product weighed 104mg (76%). mp. 218-22°C (dec.).

!HNMR (DMSO-d6, ppm): 8.04 (s, 1H), 6.37 (s, 1H), 4.82 (s, 2H), 3.40 (q, J=5.8Hz,

6 H), 2.66 (s, 3H), 2.54 (s, 3H), 2.48 (s, 3H), 1.34 (t, J= 6 .8 Hz, 9H). 13CNMR (DMSO- d6 ,ppm): 160.8, 159.8, 153.9, 153.5, 148.6, 124.9, 116.2, 113.2, 112.5, 108.0, 104. 6 ,

53.0, 18.8, 13.3, 8.2. Note: Two dimentional NMR shows that two methylene carbons are superimposed at 53.0ppm, and one methyl group on psoralen ring has the same chemical shift (8.2) as the methyl goups on triethylammio group in the , 3 CNMR. MS

(m/e, intensity%): 317 (5.14), 276 (52.42), 241(100), 213 (17.59), 8 6 (25.78).

4’,5,5’-Tribromo-4’,5’-dihydro-8-methoxypsoralen 0.432g (2.00mmol) of 8 -MOP and 1.96g (6.00mmol) of mercuric acetate were dissolved in lOmL of glacial acetic acid. To the above solution, 0.960g (6.00mmol) of bromine solution in 3mL of glacial acetic acid was added dropwise in 20min. The red color of bromine disappeared quickly as a white solid precipitated. The solid was collected by filtration and recrystallized from methanol to give 0.784g ( 8 6 %) of product, mp. 122-7°C (dec.). *HNMR (CDCI 3 , ppm): 7.98 (d, J=10.0Hz, 1H), 6.97 (s, 1H), 6.42 (d, J=10.0Hz, 1H), 5.63 (s, 1H), 4.11

(s, 3H). ,3CNMR (CDCI3 , ppm): 158.8, 141.7, 133.3, 125.5, 116.0, 115.6, 111.8, 89.3,

61.4, 52.7. High resolution MS (m/e, intensity%): calcd for Cj 2 H 7 Br3 0 4 451.7894, found 451.7944 (0.02), 456 (M++4, 0.14), 454 (M++2, 0.18), 376 (1.70), 374 (3.22), 136 372 (1.84), 296 (100), 294 (99.26), 281 (33.29), 279 (33.04), 253 (42.77), 251 (43.37),

215 (78.44).

4\5,5’-Tribromo-8-hydroxypsoralen A solution of 4.32g (20.0mmol) of 8 -MOP in

60mL of 1,2-dichloroethane was added dropwise over 30min to a mixture of aluminum chloride (13.4g, lOOmmol) and 40mL of 1,2-dichloroethane, which was previously purged with nitrogen, cooled with an ice-salt bath. This was followed by the addition of bromine (9.6g, 60mmol) over lh. After stirring at 0°C for 2h, the reaction was then allowed to stand overnight at room temperature. The reaction mixture was poured onto a mixture of ice (200g) and concentrated hydrochloric acid (30mL), and stirred over 3h.

The solid was collected by vacuum filtration, washed with acetone, and then transferred into 500mL of acetone, which consequently was refluxed over 2h. After cooling, the gray solid was collected, yield 8.74g (100%), mp 305-10°C. !HNMR (DMSO-d 6 , ppm):

11.40 (br, 1H), 8.12 (d, J=9.2Hz, 1H), 6.55 (d, J=9.2Hz, 1H). 13CNMR: the spectrum was not taken due to the lack of suitable solvent. HRMS mass calcd for CnH 3 Br3 C>4

435.7581, found 435.7561, major fragments m/e (relative intensity %): 442 (0.93,

M++ 6 ), 440 (2.98, M++4), 438 (3.14, M++2), 436 (0.95, M+), 396 (20.76), 394 (27.69),

317 (11.70), 315 (45.52), 313 (34.56), 271 (23.45), 269 (35.28), 82 (98.56), 80 (100),

79(41.99).

4’,5,5’-Tribromo-8-(3-broinopropyloxy)psoralen To a mixture of 4.39g (lO.Ommol) of 4,,5,5’-tribromo-8-hydroxypsoralen and 10.lg (50mmol) of 1,3-dibromopropane in

80mL of DMSO, was added 7.0g of K 2 CO3 . After stirring at 60°C for 2 days, the reaction mixture then was poured into 250mL of ethyl acetate, which was followed by washing 3 times with 150mL of water, dried over Na 2 SC>4 and concentrated by rotatory evaporation to about 50mL. About the same amount of petroleum ether was added to the residue to obtain a precipitate, which was collected by filtration to yield 3.70g 137

(6 6 %) of yellow solid, mp 116-8°C. 'HNMR (DM S0-d6, ppm): 8.13 (d.d., Ji=10Hz,

J 2 =2Hz, 1H), 6.59 (d, J=10Hz, 1H), 4.43 (t., J=5.9Hz, 2H), 3.74 (t., J=6.5Hz, 2H), 2.27

(m. 2H). HRMS mass calcd for C i4 HsBr4 0 4 555.7156, found 555.7178, major fragments m/e(relative intensity %): 564 (0.43, M++8 ), 562 (1.56, M++6 ), 560 (3.02,

M++4), 558 (1.34, M++2), 556 (0.36, M+), 518 (14.90), 516 (23.75), 514 (18.28), 474

(11.00), 472 (34.13), 470 (36.94), 468 (13.86), 396 (44.12), 394 (60.20), 352 (47.04),

349 (100), 348 (60.91), 271 (33.81), 269 (54.70).

8-[(3-Triethylammonio)propyIoxy]-4’,5,5’-tribromopsoralen bromide(TBP) 2.0mL of triethylamine was added to a mixture of 0.561g(1.00mmol) of 4’,5,5’-tribromo-8-(3- bromopropyloxy)psoralen and 8 mL of dimethyl sulfone. After stirring at 50-60°C for 2 days, the reaction mixture was poured into approximately 30mL of water, the precipitate so produced was then collected by filtration, recrystallized from ethanol and ethyl acetate to give 0.435g ( 6 6 %) yellow solid, mp 224-6°C (dec.). *HNMR (DMSOd.

6, ppm): 8.21 (d d, J,=10.0Hz, J 2 =2.0Hz, 1H), 6.64 (d, J=10.0Hz, 1H), 4.46 (t, J=4.8Hz,

2H), 3.47-3.26 (m, 8 H), 2.15 (m, 2H),1.23 (t, J=7.0Hz, 2H), Major fragments m/e

(relative intensity %): 584 (6.20, M+-Br+ 6 ), 582 (18.33, M+-Br+4), 580 (18.55, M+-

Br+2), 578 (6.16, M+-Br), 540 (12.19), 538 (67.50), 536 (100), 534 (40.74), 494

(42.11), 492 (90.03), 490 (51.77), 114 (24.04).

3(8)-(4’,5,5’-Tribromopsoralen)oxy]propyl dimethyl 2-dimethylamminoethyl ammonium bromide 2.0mL of N.N.N’jN’-tetramethylethylenediamine was added to a mixture of 1.12g (2.00mmol) of 4’,5,5’-tribromo-8-(3-bromopropyloxy)psoralen and

20mL of DMSO. After stirring at 50-60°C for 2 days, the reaction mixture was poured into about lOOmL of water, and the precipitate produced was collected by filtration, recrystallized from ethanol and ethyl acetate to give 0.700g yellow solid, mp 189-92°C

(dec.). JHNMR (CDC13, ppm): 8.18, 8.12 (d, J=10.0Hz, 1H), 6.45 (d, J=10.()Hz, 1H), 138

4.49 (t, J=6.2Hz, 2H), 4.12 (t, J=8.4Hz, 2H), 3.88 (t, J=5.7Hz, 2H), 3.55 (s, 6 H), 2.89 (t,

J=5.4Hz, 2H), 2.53-2.41 (m, J,=6.3Hz, 2H), ,3CNMR (CDCI3 , ppm): 159.7, 147.1,

146.9, 144.1, 142.5, 131.5, 127.8, 115.7, 115.6, 107.4, 105.6, 72.7, 72.6, 71.3, 70.6,

70.3,61.6,49.1,46.9,42.7,27.7, 11.6.

3(8)-(4’,5,5’-Tribromopsoralen)oxypropyl 2-hydroxyethyl dimethylammonium brom ide A mixture of 0.261mg (0.500mmol) of 4\5,5’-tribromo-8-(3- bromopropyloxy)psoralen and 0.5mL of the required amine in 5mL of isopropanol was refluxed for 24hrs. After cooling, the mixture was poured into lOmL of water. The yellow solid was collected by filtration and recrystallized from ethanol-ethyl acetate.

The yield was 0.255g (79%), mp 225-8°C (dec.). 'HNMR (DMSO-d6, ppm): 8.27 (dd,

Ji=10.0Hz, J2 =2.0Hz, 1H), 6.67 (d, J]=10.0Hz, J2 =1.3Hz, 1H), 4.44 (t, J=5.4Hz, 2H),

3.62-3.35 (m, 6 H), 3.10 (s, 6 H), 2.28-2.19 (m, 2H), Major fragments m/e (relative intensity %): 572 (4.82, M+-Br+ 6 ), 570 (14.73, M+-Br+4), 568 (14.73, M+-Br+2), 566

(5.06, M+-Br), 526 (65.66), 524 (93.91), 522 (40.87), 482 (42.05), 480 (100), 478

(59.49), 136(12.84).

3-[8-(4\5,5’-Tribromopsoralen)oxy]propyI 3-hydroxypropyl dimethylammonium bromide The procedure described above was applied. 0.261g starting material produced 0.22lg (67%) of yellow product, mp 225-8°C (dec.). !HNMR (DMSO-d 6 , ppm): 8.27 (dd, Ji=10.0Hz, J 2 =2.0Hz, 1H), 6.67 (d, Jj=10.0Hz, J2 =1.3Hz, 1H), 4.48 (t,

J=5.7Hz, 2H), 3.66-3.58 (m, 2H), 3.50 (t, J=6.2Hz, 2H), 3.43-3.35 (m, 3H), 3.11 (s,

6 H), 2.28-2.20 (m, 2H), 1.93-1.84 (m, 2H), Major fragments m/e(relative intensity %):

586(5.02, M+-Br+ 6 ), 584(15.34, M+-Br+4), 582(15.37, M+-Br+2), 580(5.29, M+-Br),

542(13.32), 540(67.27), 538(100), 536(42.73), 498(6.83), 496(43.81), 494(98.11),

492(58.41), 144(9.05), 136(7.07). 139 IV.4 Preparative Photolysis

IV.4.1 Photolysis of Psoralen Azide

50mg of AMP was dissolved in 20mL of diethylamine. The resulting solution was photolyzed with 350nm light (3.5mW/cm2) at 3°C for 30min. Diethylamine was then evaporated to give a brown oily residue. The residue was washed with benzene and dissolved in ethanol. The residue was then applied onto a preparative TLC plate, which was developed by ethyl acetate.ethanol (8:2). A yellow product was recovered and was determined by !HNMR to be 5-amino-8-methoxypsoralen (25mg, 55.6%). The identity of the product was further supported by gas chromatography.

Similar experiments were performed in other solvents such as methanol and

DMSO. The products could not be identified.

IV.4.2 Photocycloaddition of Psoralen Derivatives with Tetramethylethylene

30mg of 5-bromo-8-methoxypsoralen was dissolved with slight warming in

150mL of ethanol in a pyrex flask. 4mL of TME was added to the resulting solution.

The mixture was degassed with Argon for 5min and photolyzed with 350nm

(3.0mW/cm2) at 3°C for 24hr. After ethanol was evaporated, the products were separated by preparative TLC and repeatedly developed with benzene to give two fluorescing products. The products were identified as the brominated [2+2]- cycloaddition product and the debrominated [ 2 +2 ]-cycloaddition product.

Brominated [2+2]-Cycloaddition Product (BPoTM E) 1 HNMR (CDCI3 , ppm):

7.62 (d, J=2.2Hz, 1H), 6.75 (d, J=2.2Hz, 1H), 4.20(s, 3H), 3.91 (d, J=10.0Hz, 1H), 3.09

(d, J=10.0Hz), 1.38 (s, 3H), 1.33 (s, 3H), 1.07 (s, 3H), 0.80 (s, 3H). HRMS mass calcd for C i 8 Hj9Br0 4 378.0467, found 378.0464. Major fragments m/e(relative intensity %):

380 (M++2, 10.57), 378 (M \ 12.57), 296 (99.17), 294 (100), 215 (23.70), 162(25.70),

149 (7.70), 84 (69.13), 69 (69.25). 140

Debrominated [2+2]-Cycloaddition Product (HPoTM E) *HNMR (CDCI 3 , ppm):

7.59 (d, J=2.2Hz, 1H), 6.84 (s, 1H), 6 . 6 8 (d, J=2.2Hz, 1H), 4.20(s, 3H), 3.52 (d,

J=10.0Hz, 1H), 3.17 (d, J=10.0Hz), 1.36 (s, 3H), 1.23 (s, 3H), 1.05 (s, 3H), 0.74 (s,

3H). HRMS mass calcd for C 18 H 20 O4 300.,1362 found 300.1366. Major fragments m/e(relative intensity %): 300 (M+, 9.53), 216 (100), 201 (8.80), 173 (7.49), 149 (6.29),

84 (19.24), 69 (24.75).

IV.5 Measurement of the W ater Solubility of a Compound

The solubility of a compound is determined by measuring the UV-vis absorbance of its saturated solution at the wavelength of maximum absorption and calculating the concentration of the saturated solution by applying the extinction coefficients obtained by Lambert-Beer’s Law. Excess solid compound was suspended in the appropriate medium. The container of the suspension was then placed on a shaker, in the dark if necessary, and allowed to equilibrate for at least 24h at 25°C.

Filtrating off any solid residue gave a saturated solution of the compound.

IV . 6 Measurement of Binding Constants of Sensitizers to DNA

The solutions of acridine compounds were made with 1% ammonium acetate, pH=5.5 aqueous solution, while the solutions of psoralen derivatives were made with

Tris-HCl buffer. Ethidium bromide solutions were prepared with Tris-HCl buffer. All solutions were then sterilized by filtration using Acrodisc syringe filters and stored at

4°C. Calf thymus DNA was suspended in Tris-HCl buffer. The suspension was left in a refrigerator for at least 2 days before the DNA was completely dissolved. The concentrations of reagents were determined spectrophotometrically by using the following extinction coefficients210: ethdium bromide, 6 4 8 0 = SSbOM^cnv1; calf thymus

DNA, £2 6 0 = (i^OOM-'cnv1; and the extinction coefficients listed in Table 3 and Table 5. 141 Using the above stock solutions, a series of 2mL solutions were made with constant concentrations of calf thymus DNA and ethidium bromide at about 5~10 p.M, and varying concentrations of the compound in question. The biggest concentration of the compound in question was chosen carefully so that the fluorescence of ethidium bromide was displaced by less than 30%. The serial solutions were left in the dark to allow equilibration at room temperature for at least 4h before the ethidium bromide fluorescence of the serial solutions (Xcx=545nm, X,.m=595nm) were recorded. After correcting for the backgroud fluorescence, plotting the relative fluorescence intensities

(I0 /I) against the concentration of the compound in question resulted a straight line. The intrinsic binding constant of the compound was then calculated from the slope of the acquired line by using cq 17.

IV.7 Measurement of Relative Fluorescence and Triplet-Triplet Absorption

Intensities

Three samples, two of Acr-NH 2 and one of Acr-I, were prepared in 1%

NH4 Ac(aq), pH=5.5 solution. The first Acr-NH 2 sample and the Acr-I one had identical

UV-vis absorbances at 450nm. The concentration of the second Acr-NH 2 sample was smaller than the first one and used as a reference. The fluorescence intensities of the two Acr-NH 2 samples at the maximum wavelength were taken by exciting the samples at 450nm with a certain set of excitation and emission slits on the fluorimeter. The fluorescence intensity of the first Acr-NH 2 sample devided by that of the second one gave the relative fluorescence intensities. The excitation and emission slits of the fluorimeter were then adjusted to smaller values. At this time, the fluorescence intensities of the reference Acr-NH 2 sample and the Acr-I were recorded at their maximum wavelength when excited at 450nm and the relative fluorescence intensity of the reference Acr-NH2 sample over Acr-I was calculated. The product of the two 142 relative fluorescence intensities is the relative fluorescence intensity of Acr-NH 2 versus

Acr-I 1% NH4 Ac(aq), pH=5.5 medium.

In other cases, the relative fluorescence intensity was obtained by monitoring the fluorescence intensities at the maximum wavelength, when two samples were excited at a selected wavelength, where they had identical optical density.

The relative triplet-triplet absorbance intensities were measured by LFP.

Samples were prepared in the appropriate medium with an optical density of 1.0 at

308nm. The samples were deoxygenated by purging with argon for at least 5min in quartz cuvettes before LFP. The resulting tansient absorption was monitored with

OMA.

IV . 8 DNA Photocleavage

IV.8.1 Preparation of Buffers

TAE Electrophoresis Buffer The recipe for 50x stock solution of this buffer is 242g of tris(hydroxymethyl)aminomethane (Tris base), 57.1mL of glacial acetic acid, and

37.2g o f Na 2 EDTA-2 H2 0 were mixed. The volume of the resulting solution was made up to 1L with doubly distilled water. The solution was then sterilized by filtration through 0.2|im Acrodisc syringe filters.

The working solution was obtained by diluting the above stock solution by 50 fold. The working solution had a pH of 8.5 with the concentrations of Tris acetate and

Na2 EDTA at 40mM and 2mM, respectively. The buffer was desired for both making agrose gel and running the gel.

Tris-EDTA (TE) Buffer 121g of Tris base, with the addition of hydrochloric acid, was made up to 1L of solution with a pH of 7.4. Similarly, 186g of Na 2 EDTA-2 H2 0 was made up to 1L of solution with a pH of 8.0, adjusted with sodium hydroxide 143 solution. lOmL of the above Tris buffer was then mixed with 2.0mL of the EDTA buffer, and the volume of the mixture was adjusted to 1L with doubly distilled water.

The resulting solution was sterilized by filtration through 0.2pm Acrodisc syringe filters. The TE buffer thus obtained was composed of lOmM Tris-HCl and l.OmM

EDTA. It was used to prepare irradiated DNA samples.

Tracking Dye lOx stock solution was prepared as follows: 0.10g of orange G sodium salt, 0.50g of xylene cyanole FF, and 0.50g of bromophenol blue sodium salt were dissolved with shaking in 4mL of double distilled water. After the mixture became clear, 6.0g of glycerol was added to the solution. For practical use, l.OmL of the stock solution was diluted with 3.0mL of water and 6.0mL of glycerol. This tracking dye helped to monitor the progress of gel electrophoresis.

DNA Molecular Marker Sample To 10|iL of DNA molecular weight marker II solution (0.25pg/pL) were added 40pL of TE buffer and lOpL of tracking dye. The concentration of the resultant mixture was about 40|ig/mL.

IV.8.2 Photolysis of Supercoiled Plasmid DNA

All stock solutions of the sensitizers to be examined were prepared as in the experiment measuring the binding constant. The calculated amounts of sensitizer solution and supercoiled DNA were then mixed. The mixture was then deluted to the required volume with Tris-HCl buffer. The final DNA concentration was 40pg/mL. The volume of final sample usually was 50pL for the experiments designed to have a fixed time of irradiation, and it was lOOpL for the experiments in which aliquots of sample was taken out at the appropriate time period. The solution was left to equilibrate in the dark for at least 2 h before irradiation took place. 144 ' The solutions prepared for photolysis were put in 300|iL Pyrex HPLC sample injection vials (35x3mm), each of which in turn was placed in a home made cylindrical

Pyrex sample holder (2.5x6.5cm). The purpose to use these Pyrex materials was to cut off short wavelength lights from lamps if any. For deoxygenated experiments, sample holders with air-tight joints were used. The deoxygenated condition was obtained by very gently blowing argon flow into the HPLC injection vial for at least 5min. All irradiations were conducted in a Rayonet reactor equipped with lamps emitting desired wavelength light in a cold room at 3°C. The lamps were arranged in a circular array around the samples to ensure uniform irradiation. The light intensity was controlled by adding or subtracting the number of lamps. Aliquots (10|iL) of samples were removed at the desired time intervals and resolved for gel electrophoresis.

IV.8.3 Gel Electrophoresis

The mold used for preparing agrose gel was a rectangular one of the Bio-rad mini sub gel apparatus, having a size of 9 x 6 x 0.75cm. The mold, along with its twocombs, was sterilized with ethanol. The ends of the mold were sealed with Scotch tape and the two combs were set at desired positions.

0.30g of Agrose was dissolved in about 32mL of TAE gel electrophoresis buffer by heating the suspension in a microwave oven. It was important to ensure that all agrose particle were completely melted and not much of water was lost during the heating process. After cooled slightly, the resulting ~1% agrose solution was poured onto the prevoiusly set mold. Once the agrose gel solidified, it was covered with TAE gel electrophoresis buffer. The combs were carefully removed to avoid destroying the formed wells. According to our experience, leaving the gel covered by TAE buffer for a while before removing the combs could reduced the chance of gel damage.

After the agrose gel was ready, to each of 10|iL photolyzed DNA solutions were added 15|iL of TE buffer and 5|iL of tracking dye. Each of the resulting mixtures 145 (5~7pL) was carefully loaded into a well in the agrose gel. DNA molecular weight marker II sample was also loaded in proper wells serving as reference later. The loaded agrose gel was carefully placed in a gel electrophoresis chamber, which was previously filled with TAE buffer, and the gel was then developed with a electric voltage of

2.5V/cm at room temperature. The developing process generally took 4~5hr. After the electrophoresis completed, the agrose gel was stained in a water bath containing about

0.5mg/mL ethidium bromide for at least 30min and then destained in distilled water for at least lhr. The gel thus obtained was visualized with a Fotodyne model 3-4400

Transilluminator and photographed with a Polaroid MP-4 land camera. An orange filter was equipped on the camera to achieve a desirable film image of light transmitted by

DNA. Polaroid P667 film was used.

IV.9 Viral Inactivation

For X phage inactivation experiment, E. coli bacteria (ATCC #23724) have to be prepared and used as the host to grow X phage (ATCC #23724-B2). The harvested X phage are then administered with drug to be tested to examine the effects of the drug on the X phage activity. Several media are required for this process and have to be ready before the experiment can start.

IV.9.1 Medium Preparation

Luria Bertani (LB) medium This medium sometimes is reffered to as Luria or Lenox broth. It is a liquid rich medium desired for the growth of E. coli bacteria. There are many different recipes for LB medium that differs only in the pH value of the medium.

We decided the pH of this medium to be 7.5. Also, it is of notice that the buffer capacity of this medium is limited and its pH value drops as the growing of E. coli bacteria in it nears saturarion. 146 lOg of Bacto-tryptone, 5g of Bacto-yeast extract, and lOg of sodium chloride were mixed with about 900mL of double distilled water in a 2-Liter Erlenmeyer flask.

The pH of the mixture was adjusted to 7.5 with sodium hydroxide solution. The volume of the mixture was then made up to 1L with double distilled water. The flask was stoppered with a peace of sponge, on which was covered with a peace of aluminum foil.

The mediu was then autoclaved for sterilization at 120°C for 20min. The autoclaved medium was poured in a pre-autoclaved bottle. The bottle was then sealed and stored in refrigerator for later use.

LB Plating Medium This solid rich mediumwas required to obtain single colonies of

E. coli bacteria. It was made in the same way as LB medium, except 15g of bacto-agar was added before it was autoclaved. This autoclaved medium can stay liquid indefinitely at temperature above 50°C, but it rapidly solidifies if its temperature falls below 45°C. So the autoclaved medium should be poured into sterile disposable petri dishes in a sterilized environment and allow to solidify, before it cools to 50°C.

Freshly poured plates are wet and unable to absorb liquid spread onto them.

Moreover, the moisture in wet plates can cause the freshly streaked bacteria to float away. The amount of moisture in a plate is proportional to the amount of medium poured in. Hence, excess of medium poured in one plate does not only wast material, but also cause potential problem for the experiment. In a typical handling, each plate is poured about 30mL of medium, thus about 30 plates are produced from one liter medium. To prevent the moisture in a plate from floating bacterial streak, newly poured plates are generally left on the bench for at least 3 to 4 hr before they are wrapped upside down in the origenal bags used to pack the empty plates. Thus packed plates are stored at 4°C for later use. 147 LB Top Agar This top agar is used to distribute X phage evenly in a thin layer over the surface of a LB plate. LB top agar contains less agar than LB plates. It can stay liquid for days when it is kept at 45°C to 50°C. The recipe for LB top agar is the same as LB plating medium, except using 7g of agar instead of 15g. The autoclaved medium are stored separatedly in pre-sterilized 100-mL bottles at 4°C for later use.

Before use, LB top agar should be molten in a microwave oven and then cool to and hold at 45-50°C. It is critical that no solid remains and the temperature is not too high before the top agar is applied onto the surface of LB plates. Unmolten or solidified top agar will interfere the examination of the acitivity of X phage later. Bacteria and X phage can be killed by even brief exposure to agar that is hotter than 65°C.

Storage Medium (SM) This medium is also called suspension medium. It is desired for harvesting and storing X phage, as well as for performing serial dilution. 5.8g of sodium chloride, 2g of magnisium sulfate dihydrate, and 5omL of 1M Tris HC1 buffer

(pH=7.5) were mixed in a 2-liter Erlenmeyer flask. The volume of the mixture was made up to 1L with double distilled water. O.lg of Bacto-gelatin was added to the above mixture. This could also be done by adding 5mL of 2% gelatin solution, which was prepared by heating a suspension of 2g of Bacto-gelatin in lOOmL of double distilled water. The final concentration of gelatin was 0.01%. The medium was then sterilized by autoclaving at 120°C for 20min and stored in a sealed bottle at 4°C for later use.

Other Medium Required

20% maltose aqueous solution is dsired for the growth of E. coli bacteria.

Maltose induces the E. coli bacteria grown in it the production of the X receptor (lam B protein), which is necessary for maltose transport. Without the addition of maltose in 148 the medium wherein E. coli bacteria grow, X phage will not be capable of infecting the

E. coli bacteria.

0.1M magnisium sulfate aqueous solution is the medium in which harvested E. coli bacteria are stored. The presence of Mg2+ ions aid X phage adsorbtion on the host bacteria.

Like all menia used in X phage inactivation experiment, these two solotions should also be sterilized before use.

IV.9.2 Growth and Maintenance of E. coli Bacteria and X Phage

All the materials used in this experiment have to be sterilized. The experimental environment should be free of contamination. Due to the limitation of our equipment, we achieved this by bleaching the environment with Clorox. All manipulation of bacteria and virus was conducted near the flame of a burner.

Growth and Maintenance of E. coli Bacteria

The lyophilized E. coli bacterial pellet obtained from ATCC was devided into three parts. Two of them were stored in a small sealed virals for use in futhure. To the other part was added l.OmL of LB medium. The suspension was vortexed untill the pellet well suspended. A platinum inoculating loop was sterilized by heating it on the flame of the buener until red and cooled by buried it in the lawn of a previously prepared LB petri dish. The sterilized inoculating loop was used immediately to pick up a small portion of the bacterial suspension and streaked the bacteria across one side of the lawn of LB petri dish. Repeat sterilizing inoculating loop and streaking the bacteria for another two times to cover most part of the plate. More plates could be obtained by repeating the process if necessary. The leftover of bacterial suspension was disposed by discarding it into bleach solution. The streaked plates were then left in an incubator for at least 24hr at 37°C to obtain isolated bacterial colonies. 149 Bacteria grown from single colonies ensure that each cell in a population is descended from a single founder cell and thus has the same genetic makeup. A single colony of bacteria was picked up with a sterile platinum inoculating loop and washed in a 150-mL sterile Erlenmeymer flask with a screwing cap, which was filled with 50mL of LB medium and 0.5mL of 20% maltose solution. Any of the single colonies may be saved for further use by storing the plates at 4°C wrapped in parafilm. The flask inoculated with bacteria was then left on a roller drum in a incubator at 37°C overnight.

To ensure proper aeration, the rate of the roller was set ~120rpm and the flask cap was loosed. Freshly saturated culture of E. coli bacteria was slightly opaque with a typical density of 1 to 2xl0 9pfu. According to our experience, sometimes it needed more than

24hr to reach saturation.

The saturated bacterial suspension was transferred into a sterile tube and centrifuged at SOOOrpm for about lOmin. The supernatant was discarded into a bleach solution. The resulting bacterial pellet was resuspended in about 50mL of 0.1M magnisium sulfate solution and stored in a refrigerator.

Growth and Maintenance of X Phage

The lyophilized X phage pellet obtained from ATCC was reconsitituted with

0.3mL of LB medium in a small sterile vial and vortexed. IOOjiL of this viral suspension was taken for ten-fold serial delutions.

Six sterile culture tubes were line up in a row on a tube holder. To each of the six tubes was introduced 900|iL of SM. Using a pipet, lOOpL of viral suspension was transferred from the small vial into the first tube of SM. The pipet tip was washed in the

SM. The tube was then mildly shaked or vortexed to get well mixed. Using a new pipet tip, lOO^L of suspension from the the first tube was transferred into the second tube.

Repeat the above procedure until six times of dilution was finished. At this time, lOOpL of each of the prepared viral dilutions was placed into another row of six sterile 150 culture tube respectively with new pipet tips. To each culture tube of the second row was added IOOjiL of E. coli bacterial suspension made earlier. If necessary, more samples of bacteria and virus mixture can be made in the same as the second row of culture tubes. The first row of culture tubes were then disposed in bleach solution. The second row of culture tubes were incubated at 37°C for about 30min to let X phage infect E. coli bacteria.

After allowed time for infection, about 3mL of previously prepared 45-50°C molten LB top agar was introduced to each culture tube. The content in the culture tubes was mildly agitated and then poured evenly on the lawn of previously prepared

LB plates. It is important that each plate was properly marked. Once the top agar hardened, the plates were placed in the incubator at 37°C overnight.

The plates of a very large number of clear plaques were isolated. To each of them was added 3mL of SM. The plates were left in a refrigerator overnight. The supernatant was isolated and placed in a sterile tube. The tube was then centrifuged at

5000rpm for 15min. It is helpful for the precipitation of bacteria that two or three drops of chloroform were added before the centrifuge. The viral supernatant was separated and stored at 4°C for viral inactivation experiment.

The viral titre was obtained plaque forming assay. lOO^L of viral suoernatant was performed six times of serial ten-fold dilution and obtained viral plaques as described above. By counting the plaque number on countable plates, the viral titre was determined to be No. of plaques xlOlimcordolulion+1.

IV.9.3 Viral Inactivation

Stock solutions of the sensitizers to be examined were prepared as in the experiment of measuring binding constant. Calculated amounts of sensitizer solution and X phage was mixed in a Pyrex petri dish (a home-made air tight with an inlet and oulet one was used for experiments under deoxygenated conditions). The mixture was then diluted with Tris-HCl buffer to 3mL and left in the dark at 4°C for about lhr. The final viral density was 6xl07pfu and the concentration of sensitizer was generally

50|lM. The sample solution was placed under horizontally arranged bulbs on a rotatory shaker and photolyzed with light of desired wavelength and intensity. IOOjiL of sample was taken away from the petri dish after the desired irradiation time reached. The irradiated samples were performed plaque forming assay to determine the viral survival fractions. REFERENCES

[1] “The current global situation of the HIV/AIDS pandemic” World Health Organization/Global Programme on AIDS, Geneva, Switzerland, 1994.

[2] Aggleton, P.; O’Reilly, K.; Slutkin, G. & Davies, P., “Risking everything? risk behavior, behavior change, and AIDS”, Science, 1994, 265, 341-345.

[3] Cohen, N. D.; Munoz, A.; Reitz et at., “Transmission of retroviruses by transfusion of screened blood in patients undergoing cardiac surgery”, N. Engl. J. Med., 1989 ,320, 1172-1176.

[4] a. Cummings, P. D.; Wallace, E. L.; Schorr, J. B.; & Dodd, R. Y., “Exposure of patients to human immunodeficiency virus through the transfusion of blood components that test antibody-negative”, N. Engl. J. Med., 1989, 321,941-946. b. Ward, J. W.; Holmberg, S. D.; Allen, J. R. etal., “Transmission of human immunodeficincy virus(HIV) by blood trasnfusions screened as negative for HIV antibody”, ibid, 1988, 318, 473-478.

[5] Wallace, E. L.; Surgenor. D. M.; Hao, H. S.; An, J.; Chapman, R. H.; & Churchill, W. H., “Collection and transfusion of blood components in the United States”, Transfusion, 1993 ,33, 139-144.

[6] a. Horowitz, M. S.; Horowitz, B.; Rooks et al., “Virus safety of solvent/detergent treated antihaemophilic factorconcentrate”, Lancet, ii, 1988,186-188. b. Horowitz, B.; Wiebe, M. E.; Lippin, A. etal., “Inactivation of viruses in labile blood derivatives 1: Disruption of lipid-enveloped viruses by tri(n-butyl)phosphate /detergent combinations”, Transfusion, 1985,25, 516-522.

[7] Matthews, J. L.; Newman, J. T.; Sugandares-Bernal, F.; Judy, M. M.; Skiles, H.; Levenson, J. E.; Marengo-Rowe, A. J.; & Chanh, T. C., “Photodynamic therapy of viral contaminents with potential for blood banking applications”, Transfusion, 1988,25,81-83.

[8] Lambrecht, B.; Mohr, H.; Knuver-Hopf, J; & Schmitt, H„ “Photoinactivation of viruses in human frewsh plasma by phenothiazine dyes in combination with visible light”, Vox Sang, 1991, 60, 207-213.

[9] a. O ’Brien, J. M.; Montgomery, R. R.; Burns, W. H.; Goffney, D. K.; & Sieber, F., “ Evaluation of merocyanine-540 sensitized photoirradiation as a means to inactivate enveloped virus in blood products”, J. Lab. Clin. Med., 1990, 116, 439-447. 152 153 b. Dodd, R. Y.; Moroff, G.; & Wagner, S., “ Inactivation of viruses in platelet suspensions that retain their in vitro characteristics: Comparison of psoralen- ultraviolet A and merocyanine 540-visible light methods”, Transfusion , 1991, 31, 483-490.

[10] “Frontier in biotechnology: Antibiotic resistance”, Special report, Science , 1994, 264 ,360-393.

[11] Platz, M. S.; Park, S. C.; Kasturi, C.; Rai, S.; Aebischer, J.-N.; Chen, T.; Kulaga, L.; Goodrich, R. P.; Sowemimo-Coker, S. O.; Yerram, N. R.; Tay-Goodrich, B. H.; Forster, P.; Crandall, S.; & Jones, D., in “Prevention of transfusion mediated diseases”, Sieber, Ed., “The design and Development of selective viral inactivation sensitizers for sterilization of blood products”, CRC press, 1994, in press.

[12] Rai, S.; Kasturi, C.; Grayzar, J.; Platz, M. S.; Goodrich, R. P.; Yerram, N. R.; Wong, V.; & Tay-Goodrich, B. H., “Dramatic improvements in viral inactivation with brominated psoralens, naphthalenes and anthracenes” Photochem. Photobiol., 1993 ,5 8 ,59-65.

[13] Goodrich, R. P.; Yerram, N. R.; Tay-Goodrich, B. H.; Forster, P.; Platz, M. S.; Kasturi, C.; Park, S. G.; Aebischer, J. N.; Rai, S.; & Kulagl, L. “Selective inactivation of viruses in the presence of human platelets: UV sensitization with psoralen derivatives”, Proc. Natl. Acad. Sci. USA, 1994, 91, 5552-5556.

[14] Avery, O. T.; Macleod, C. M.; & McCarty, M., “Studies on the chemical nature of the substance inducing transformation of pneumococcal types. I. Induction of transformation by a deoxyribonucleic acid fraction isolated from pneumococcus type III”, J. Exp. Med. 1944, 79, 137.

[15] Watson, J. D. & Crick, F. H. C., “Molecular structure of nucleic acids: A structure for deoxyribose nucleic acid”, Nature, 1953, 171,737-738.

[16] a. Neidle, S. & Waring, M., Ed., “Molecular Aspects of Anticancer Drug-DNA interactions”, CRC Press, 1993, Vol. /. b. Propst, C. L. & Perun, T. J., Ed., “Nucleic Acid Targeted Drug Design”, Marcel Dekker, Inc. New York, 1992.

[17] a. Dervan, P. B., “Design of sequence-specific DNA-binding molecules”, Science, 1986, 232,464-471. b. Hurley, L. H., “DNA and associated targets for drug design”, J. Med. Chem., 1989, 32, 2027-2033. c. Turston, D. E. & Thompson, A. S., “The molecular recognition of DNA”, Chem. Brit., 1990, 26, 767-772.

[18] Stubee, J. & Kozarich, J. W. “Mechanisms of bleomycin-induced DNA degradation”, Chem. Rev., 1987, 87, 1107-1136.

[19] Goldberg, I. H., “Mechanism of neocarzinostatin action: Role of DNA microstructure in determination of chemistry of bistranded oxidative damage”, Acc. Chem. Res., 1991 ,24, 191-196.

[20] a. Guschlbauer, W. & Saenger, W., Ed., “DNA-ligand interactions: From drugs to proteins”, NATO ASI Series, Plenum Press, New York and London, 1986. 154 b. Zimmer, Ch. & Wahnert, U., “Nonintercalating DNA-binding ligands: Specificity of the interaction and their use as tools in biophysical, biochemical and biological investigations of the genetic material”, Prog. Biophys. Molec. Biol., 1986, 47, 31-112.

[21] Haskel, C. M., Ed., “Cancer treatment”, 2nd ed. W. B. Saunders, New York, 1985.

[22] Pullman, A. & Pullman, B., “Molecular electrostatic potential of the nucleic acids”, Quart. Rev. Biophys., 1981, 14, 289-380.

[23] a. Tong, W. P. & Ludlum, D. B., “Crosslinking of DNA by busulphan formation of diguanyl derivatives”, Biochim. Biophys. Acta, 1980, 608,174-181. b. Mehta, J. R.; Przybylski, M.; & Ludlum, D. B., “Alkylation of guanosine and deoxyguanosine by phosphoramide mustard”, Cancer Res., 1980, 40,4183-4186.

[24] Ludlum, D. B., “DNA alkylation by the haloethylnitrosoureas: Nature of modificatins produced and their enzymatic repair or removal”, Mutation Res., 1990,233, 117-126.

[25] D’Incalci, M.; Citti, L.; Taverna, P.; & Catapano, C. V., “Importance of DNA repair enzyme 06-alkyl guanine alkyltransferase (AT) in cancer chemotherapy”, Cancer Treatment Rep., 1988, 15, 279-292.

[26] Wilman, D. E. V., Ed., “Chemistry of antitumour agents”, Blackie, 1990.

[27] Brendel, M. & Ruhland, A., “Relationship between functionality and genetic toxicology of selected DNA-damaging agents”, Mutation Res., 1984, 133, 51-85.

[28] a. Valu, K. K.; Gourdie, T. A.; et al. & Denny, W. A., “DNA-directed alkylating agents 3. Structure-activity relationships for acridine-linked anline mustards: Consequences of varying the length of the linker chain”, J. Med. Chem. 1990, 33, 3014-3019, and references therein. b. Prakash, A. S.; Denny, W. A.; Gourdie, T. A.; Valu, K. K.; Woodgate, P. D.; & Wakelin, L. P. G., “DNA-directed alkylating ligands as potential antitumor agents: sequence specificity of alkylation by intercalating anline mustards”, Biochemistry, 1990 ,2 9 ,9799-9807.

[29] Cohen, G. M.; Cullis, P. M.; Hartley, J. A.; Mather, A.; Symons, M. C. R.; & Wheelhouse, R. T., “Targeting of cytotoxic agents by polyamines: Synthesis of a chlorambucil-spermidine conjugate”, J. Chem. Soc., Chem. Commun., 1992,298- 300.

[30] Lee, M.; Rhodes, A. L.; Wyatt, M. D.; D’Incalci, M.; Forrow, S.; & Hartley, J. A., “In vitro cytotoxicity of GC sequence directed alkylating agents related to distamycin”, J. Med. Chem. 1993, 36, 863-870.

[31] Moore, H. W. & Czerniak, R. Med. Res. Rev., 1981, 1 ,249.

[32] Woo, J.; Sigurdsson, S. Th.; & Hopkins, P. B., “DNA interstrand crosslinking reactions of pyrrole-derived bifunctional electrophiles: Evidence for a common target site in DNA”, J. Am. Chem. Soc., 1993, 115, 3407-3415. b. Millard, J. T.; Weidner, M. F.; Raucher, S.; & Hopkins, P. B., “Determination of the DNA cross-linking sequence specifity of reductively activated mitomycin C at 155 single-nucleotide resolution: Deoxyguanosine residues at CpG are cross-linked preferentially”, ibid, 1990, 112, 3637-3641.

[33] Leimgruber, W.; Batcho, A. D.; & Schenker, F., “The structure antramycin”, J. Am. Chem. Soc., 1965, 87, 5793-5795. b. Leimgruber, W.; Stefanovic, V.; Schenker, F.; Karr, A.; & Berger, J, “Isolation and characterization of antramycin, a new antitumor antibiotic”, .ibid, 1965, 87, 5791-5793.

[34] Bose, D. S.; Jones, G. B.; & Thurston, D. E., “New approaches to pyrrolo[2,l- c][l,4]benzodiazepines: Synthesis, DNA-binding and cytotoxicity of DC-81”, Tetrahedron, 1992,48,751-758.

[35] Remers, W. A., “Pyrrolo(l,4)benzodiazepine”, in “The chemistry of antitumor antibiotics”, Wiley, New York, 1988, Vol. 2 , 28-92.

[36] a. Petrusek, R. L.; Anderson, G. L; Gamer, T. F.; Fannin, Q. L.; Kaplan, D. J.; Zimmer, S. G.; & Hurley, L. H., “Pyrrolo(l,4)benzodiazepine antibiotics”, Biochemistry, 1981, 2 0 ,1111-1119. b. Krugh, T. R.; Graves, D. E.; & Stone, M. P., “Two-dimensional NMR studies on the anthramycin-d(ATGCAT )2 adduct”, ibid, 1989, 28,9988-9994. c. Boyd, F. L.; Cheatham, S. F.; Remers, W.; Hill, G. C.; & Hurley, L. H., “Characterization of structure of the anthramycin-d(ATGCAT ) 2 adduct by NMR and molecular modeling studies”, J. Am. Chem. Soc., 1990, 112, 3279-3289.

[37] Jones, G. B.; Davey, C. L.; Jenkins, T. C.; Kamal, A.; Kneale, G. G.; Neidle, S.; Webster, G. D.; & Thurston, D. E., “The non-covalent interaction of pyrrolo[2,l- c][l,4]benzodiazepine-5,l 1-diones with DNA”, Anti-Cancer Drug Design, 1990,5, 249-264.

[38] Hertzberg, R.; Hecht, S.; Reynolds, V. L.; Molineux, I. J.; & Hurley, L. H., “DNA sequence specificity of the pyrrolo[l,4]benzodiazepine antitumor antibiotics. Methidiumpropyl-EDTA-iron(II) footprinting analysis of DNA binding sites for anthramycin and related drugs”, Biochemistry, 1986,25, 1249-1258.

[39] Hurley, L. H.; Reck, T.; Thurston, D. E.; Holden, K. G.; Hertzberg, R. P.; Hoover, J. R. E.;, Gallagher, G. Jr.; Faucette, L. F.; Mong, S.-M.; & Johnson, R. K., “Pyrrolo(l,4)benzodiazepine antitumor antibiotics. Relationship of DNA binding and sequence specificity to biological activity of natural and synthetic compound” Chem. Res. Toxicol., 1988, 1 ,258.

[40] Hanka, L. J.; Dietz, S. A.; Gerpheide, S. A.; Kuentzel, S. L.; & Martin, D. G., “CC- 1065 (NSC-298223), a new antitumor antibiotic: Production, in vitro biological activity, microbiological assays and taxonomy of the producin microorganism”, J. Antibiot., 1978,5/, 1211-1217.

[41] Ichimura, M.; Ogawa, T.; Takahashi, K.; Kobayashi, E.; Kawanoto, I.; Yasuzawa, T.; Takahashi, I.; & Nakano, H., “Duocarmycin SA, A new antitumor antibiotic from Streptomyces SP",J. Antibiot., 1990, 43, 1037-1038.

[42] Sugiyama, H.; Hosoda, M.; & Saito, I., “Covalent alkylation of DNA with duocarmycin A. Identification of abasic site structure”, Tetrahedron Lett., 1990, 57,7197-7200. [43] Boger, D. L.; Johnson, D. S.; & Wrasdlo, W., Bioorg. Med. Chem. Lett., 1994,4, 631.

[44] Warpehoski, M. A.; Gebhard, I.; Kelly, R. C. etal., “Stereoelectronic factors influencing the biological activity and DNA interaction of synthetic antitumor agents modeled on CC-1065”, J. Med. Chem., 1988, 31, 590-603.

[45] Boger, D. L. & Yun, W., “Role of the CC-1065 and duocarmycin N2 substituent: Validation of a direct relationship between solvolysis chemical stability and in vitro biolobical potency”, J. Am. Chem. Soc., 1994, 116, 5523-5524.

[46] Rosenberg, B.; Van Camp, L.; & Krigas, T„ “Inhibition of cell division in Escherichia coli by electrolysis products from platinum electrode”, Nature, 1965, 205,698-699.

[47] Fox, E. P. & Loehrer, P. J., “Chemotherapy for advanced testicular cancer”, Hematology!ontology clinis of North America, 1991, J , 1173-1187.

[48] Pinto, A. L. & Lippard, S. J., “Binding of the antitumor drug cis- diamminedichloroplatinuni (II) (cisplatin) to DNA”, Biochem. Biophys, Acta, 1985, 780, 167-180.

[49] Jones, M. M. & Beaty, J. A., “Rate of displacement of guanosine by cyanide from c/s-diamminebis (guanosine) platinum (II) chloride determined by *3CNMR”, Inorg, Chem., 1991,30, 1584-1587.

[50] Plooy, A. C. M.; van Dijk, M. & Lohman, P. H. M., “Induction and repair of DNA cross-links in Chinese hamster ovary cells treated with various platinum coordination compounds in relation to platinum binding to DNA, cytotoxicity, mutagenicity, and antitumor activity”, Cancer Res., 1984,44, 2043-2051.

[51] Bodenner, D. L.; Dedon, P. C.; Keng, P. C.; Katz, J. C.; & Borch, R. F., “Selective protection against d.v-diamminedichloroplatinum (II) -induced toxicity in kidney, gut, and bone marrow by diethyldithiocarbamate rescue”, Cancer Res., 1986, 46, 2751-2756.

[52] Ciccarelli, R. B.; Solomon, M. J.; Varshavsky, A.; & Lippard, S. J., “In vivo effects of cis- and trans-diamminedichloroplatinum(II) on SV40 chromosomes: Differenrtial repair, DNA-protein cross-linking, and inhibition of replication”, Biochemistry, 1985,24,7533-7540.

[53] Zimmer, Ch. & Wahnert, U., “Nonintercalating DNA-binding ligands: Specificity of the interaction and their use as tools in biophysical, biochemical and biological investigations of the genetic material”, Prog. Biophys. Mol. Biol., 1986,47, 31- U2.

[54] Wartell, R. M.; Larson, J. E.; & Wells, R. D., “Netropsin: a specific probe for A-T regions of duplex deoxyribonucleic acid”, J. Biol. Chem., 1974,249,6719-6731.

[55] Krylov, A. S.; Grokhovsky, S. L.; Zasedatelev, A. S.; Zhuse, A. L.; Gursky, G. V.; & Gottikh, B. P., “Quantitative estimation of the contribution of pyrrolocarboxamide groups of the antibiotic distamysin A into specificity of its binding to DNA AT pairs”, Nucl. Acids Res., 1979, 6, 289-304. 157 [56] Marky, L. A. & Breslauer, K. J., “Origins of netropsin binding affinity and specificity: Correlations of thermodynamic and structural data”, Proc. Natl. Acad. Sci. USA, 1987, 84,4359-4363.

[57] Taylor, R.; Kennard, O.; & Versichel, W., “Geometry of the N—H—0=C hydrogen bond 2. Three-center(“bifurcated”) and four-center(“trifurcated”) bonds”, J. Am. Chem. Soc., 1984, 106, 244-248.

[58] Coll, M.; Aymami, J.; van der Marel, G. A.; van Boom, J. H.; Rich, A.; & Wang, A. H.-J., “Molecular structure of the netropsin-d(CGCGATATCGCG )2 complex: DNA conformation in an alternating AT segment”, Biochemistry, 1989, 28, 310- 320.

[59] a. Mrksich, M. & Dervan, P. B. “Enhanced sequence specific recognition in the minor groove of DN by covalent peptide dimers: Bis(pyridine-2- carboxamidonetropsin)(CH 2 )3 -6”, J. Am. Chem. Soc., 1993, 115,9892-9899. b. Chen, Y.-H. & Lown, J. W. “A new DNA minor groove binding motif: Cross- linked lexitropsins”, ibid, 1994, 116, 6995-7005. c. Singh, M. P.; Plouvier, B.; Hill, C. G.; Gueck, J.; Pon, R. T.; & Lown, J. W., “Ioshelicity and strand selectivity in the minor groove binding of chiral (1R, 2R)- and (IS, 2S)-bis(netropsin)-l,2-cyclopropanedicarboxamide ligands to duplex DNA”, ibid, 1994, 116, 7006-7020.

[60] Youngquist, R. S & Dervan, P. B., “Sequence specific recognition of B-DNA by oligo-N-methylpyrrolocarboxamides”, Proc. Natl. Acad. Sci. USA, 1985, 82, 2565- 2569.

[61] Rao. K. E.; Shea, R. G.; Yadagiri, B.; & Lown, J. W.; “Molecular recognition between oligopeptides and nucleic acids: DNA sequence specificity and binding properties of thiazole-lexitropsins incorporaring the concepts of base site acceptance and avoidance”, Anti-cancer Drug Design, 1990, 5, 3-20.

[62] Pjura, P. E.; Grzeskowiak, K, & Dickerson, R. E., “Binding of Hoescht 33258 to the minor groove of B-DNA”, J. Mol. Biol., 1987, 197, 257-271.

[63] a. Wahnert, U.; Zimmer, Ch.; Luck, G.; 8c Pitra, C., “(dAdT)-dependent inactivation of DNA template properties by interaction with netropsin and distamycin A”, Nucl. Acids Res., 1975,2, 391-404. b. Zimmer, Ch.; Luck, G.; Birch-Hirshfeld, E.; Weiss, R.; Arcamone, F.; & Guschlbauer, “Chain length dependent association of distamycin-type oligopeptides with AT and GC pairs in polydeoxynucleotide duplexes” Biochim. Biophys. Acta, 1983, 741, 15-22.

[64] Sausville, E. A.; Peisach, J.; & Horwitz, S. B., “A role for ferrous ion and oxygen in the degradation of DNA by bleomycin”, Biochem. Biophys. Res. Commun., 1976, 73, 814-822.

[65] a. Takita, T.; Umezawa, Y.; Saito, S. etal., “Total synthesis of bleomycin”, Tetrahedron Lett., 1982, 23, 521-524. b. Aoyagi, Y.; Suguna, H.; Katano, K.; Primeau, J.; Chang, L. H.; Hecht, S. M., “Total synthesis of bleomycin”, J. Am. Chem. Soc., 1982, 104, 5537-5538. 158 [66] a. Kozarich, J. W.; Worth, L. Jr.; Frank, B. L.; Christner, D. F.; Vanderwall, D. E.; & Stybbe, J. A., “Sequence-specific isotope effects on the cleavage of DNA by bleomycin”, Science, 1989, 245,1396-1399. b. Worth, L. Jr.; Frank, B. L.; Christner, D. F.; Avsalon, M. J.; Stubbe, J. A.; & Kozarich, J. W., “Isotope effects on the cleavage of DNA by bleomycin: mechanism and modulation”, Biochemistry, 1993, 32, 2601-2609.

[67] Miaskiewicz, K. & Osman, R., “Theoretical study on the deoxyribose radicals formed by hydrogen abstraction”, J. Am. Chem. Soc., 1994, 116, 232-238.

[68] a. Giloni, L.; Takeshita, M.; Johnson, F.; Iden, C.; & Grollman, A. P., “Bleomycin- induced strand-scission of DNA: Mechanism of deoxyribose cleavage”, J. Biol. Chem., 1981, 256, 8608-8615. b. Burger, R. M.; Berkowitz, A. R.; Peisach, J.; Horwitz, S. B., “Origin of malondialdehyde from DNA degraded by Fe(II)-bleomucin”, ibid, 1980, 255, 11832-11838.

[69] Hecht, S. M.; “The chemistry of activated bleomycin”, Acc. Chem. Res., 1986, 19, 383-391.

[70] a. Takeshita, M.; Grollman, A. P.; Ohtsubo, E.; & Ohtsubo, H., “Interaction of bleomycin with DNA”, Proc. Natl. Acad. Sci. USA, 1978, 75,5983-5987. b. D’Andrea, A. D. & Haseltine, W. A., “Sequence specific cleavage of DNA by the antitumor antibiotics neocarzinostatin and bleomycin”, ibid, 1978, 75, 3608-3612.

[71] Mirabelli, C. K.; Ting, A.; Huang, C. H.; Mong, S.; & Crooke, S. T., “Bleomycin and talisomycin sequence-specific strand scission of DNA: A mechanism of double-strand cleavage”, Cancer Res., 1982, 42, 277-2785.

[72] Dabrowiak, J. C., “Bleomycin”, Adv. Inorg. Biochem. 1982, 4, 69-113.

[73] Duff, R. J.; de Vroom, E.; Geluk, A.; & Hecht, S. M., “Evidence for C -l’ Hydrogen abstraction from modified oligonucleotides by Fe-Bleomycin”, J. Am. Chem. Soc., 1993, 115, 3350-3351.

[74] DeRiemer, L. H.; Meares, C. F.; Goodwin, D. A.; & Diamanti, C. I., “BLEDTA: Tumor localization by a bleomycin analogue containing a metal-chelating group”, J. Med. Chem., 1979,22, 1019-1023.

[75] a. Chang, C.-H. & Meares, C. F., “Light-induced nicking of deoxyribonucleic acid by cobalt(III) ” Biochemistry, 1982, 21, 6332-6334. b. Saito, I.; Morii, T.; Sugiyama, H.; Matsuura, T.; Meares, C. F.; & Hecht, S. M., “Photoinduced DNA strand scission by cobalt bleomycin green complex”, J. Am. Chem. Soc., 1989, 111, 2307-2308.

[76] Tan, J. D.; Hudson, S. E.; Brown, S. J.; Olmstead, M. M.; & Mascharak, P. K., “Syntheses, structures, and reactivities of synthetic analogues of the three forms of Co(III)-bleomycin: Proposed mode of light-induced DNA damage by the Co(III) chelate of the drug”, J. Am. Chem. Soc., 1992, 114, 3841-3853. 159 [77] Napier, M. A.; Holmquist, B.; Strydom, D. J.; & Goldberg, I. H., "Neocarzinostatin: Spectral characterization and separation of the nonprotein chromophore", Biochem. Biophys. Res. Commun., 1979, 89,635-642.

[78] Kappen, L. S. & Goldberg, I. H., "Stablilization of neocarzinostatin nonprotein chromophore activity by interaction with apoprotein and with HeLa cells", Biochemistry, 1980, 19,4786-4790.

[79] Bergman, R. C., "Reactive 1,4-dehydroaromatics", Acc. Chem. Res., 1973, 6 , 25- 31.

[80] Kappen, L. S. & Goldberg, I. H., "Deoxyribonucleic acid damage by neocarzinostatin chromophore: Strand breaks generated by selective oxydation of C-5' of deoxyribose", Biochemistry, 1983,22,4872-4878.

[81] a.Hatayama, T.; Goldberg, I. H.; Takeshita, M.; & Grollman, A. P., "Nucleotide specificity in DNA scission by neocarzinostatin", Proc. Natl. Acad. Sci. USA, 1978, 75, 3603-3607. b. Lee, S. H. & Goldberg, 1. H., "Sequence-specific, strand-selective, and directional binding of neocarzinostatin chromophore to oligodeoxyribonucleotides", Biochemistry, 1989,25, 1019-1026.

[82] Dedon, P. C. & Goldberg, I. H., "Sequence-specific double strand brteakage of DNA by neocarzinostatin involves different chemical mechanisms within a staggered cleavage site", J. Biol. Chem., 1990, 265, 14713-14716.

[83] Lerman, L. S., “Structural considerations in the intercalation of DNA acridines”, J. Mol. Biol., 1961,5,18-30.

[84] Waring, M. J., “DNA modification and cancer”, Ann. Rev. Biochem., 1981, 50, 159-192.

[85] Long, E. C. & Barton J. K., “On demonstrating DNA intercalation”, Acc. Chem. Res., 1990,25,271-273.

[86] Kennard, O. & Hunter, W. N., “Oligonucleotide structure : a decade of results from single crystal X-ray diffraction studies”, Q. Rev. Biophys., 1989,22, 327-379.

[87] Wang, A. H.-J.; Ughetto, G.; Quigley, G. J.; & Rich, A., “Interactions between an anthracycline antibiotic and DNA: Molecular structure of daunomycin complexed to d(CGTACG )2 at 1.2A resolution”, Biochemistry, 1987, 26, 1152-1163.

[88] Chaires, J. B.; Fox, K. R.; Herrera, J. E.; Britt, M.; & Waring, M. J., “Site and sequence selective binding of daunomycin-DNA interaction”, Biochemistry, 1987, 26, 8227-8236.

[89] Young, R. C.; Ozols, R. F.; & Myers, C. E„ “The anthracycline antineoplastic drugs”, New Eng. J. Med., 1981,505, 139-153.

[90] Lown, J. W., “The mechanism of action of quinone antibiotics”, Mol. Cell. Biochem., 1983,55, 17-40. 160 [91] Cheng, C. C. & Zee-Cheng, R. K.-Y., “The design, synthesis and development of a new class of potential antineoplastic anthraquinones”, Prog. Med. Chem., 1983, 20, 83-118.

[92] Wakelin, L. P. G. & Waring, M. J., “The binding of echinomycin to DNA”, Biochem. J., 1976, 157,721-740.

[93] Viswamitra, M. A., Kennard, O. et al. & Olsen, R. K., “The structure of TANDEM, a quinoxaline antibiotic analogue, and its implication for bifunctional intercalation into DNA”, Nature, 1981, 289,817-819.

[94]Alfredson, T. V.; Maki, A. H.; & Waring, M. J., “Conformational heterogeneity of quinoxaline peptides in solution”, Biopolymers, 1991, 3 1 ,1689-1708.

[95] Lee, J. S. & Waring, M. J., “Bifunctional intercalation and sequence specificity in the binding of quinomycin and triostin antibiotics to DNA”, Biochem. J., 1978, 173, 115-128.

[96] Tritton, T. R. & Lee, G., “The anticancer agent adriamycin can be actively cytotoxic without entering cells”, Science, 1982,2/7,248-250.

[97] Tewey, K. M.; Chen, G. L.; Nelson, E. M.; & Liu, L. F., “Intercalative antitumor drugs interfere with the breakage-reunion reaction of mammalian DNA topoisomeras II”, J. Biol. Chem., 1984, 259,9182-9187.

[98] a. Mitscher, L. A. & Shen, L. L., “A cooperative quinolone-DNA binding model for DNA gyrase inhibition: Implications in drug design”, in “Nucleic acid targeted drug design”, Ed. Propst, C. L. & Perun, T. J., Marcel Dekker, Inc., New York, New York, 1992,423-474. b. Shen, L. L.; Mitscher, L. A.; Sharma, P. N.; O’Donnell, T. J.; Chu, D. W. T.; & Cooper, C. S., “Mechanism of inhibition of DNA gyrase by quinolone antibacterials: A cooperative drug-DNA binding model”, Biochemistry, 1989, 28, 3886-3894.

[99] a. Deisenhofer, J. P.; Epp, O.; Miki, K.; Huber, R.; & Michel, H., “X-ray structure analysis of a membrane protein complex”, J. Mol. Biol., 1984, 180, 385-398. b. Deisenhofer, J. P.; Epp, O.; Miki, K.; Huber, R.; & Michel, H., “Structure of the protein subunits in the photosynthetic reaction center of Rhodopseudomonas viridis at 3A resolution”, Nature, 1985, 318,618-624.

[100] a. Gust, D.: Moore, T. A.; Moore, A. L., et al., “Efficient multistep photoinduced electron transfer in a molecular pentad”, Science, 1990, 248, 199-201. b. Marguerettaz, X. & Fitzmaurice, D., “Heterosupramolecular chemistry: Long- lived charge trapping by vectorial electron flow in a heterotriad”, J. Am. Chem. Soc., 1 9 9 4 ,/ 16, 5017-5018.

[101] O’Regan, B. & Gratzel, M., “A low-cost, high-efficiency solar cell based on dye- sensitized colloidal Ti02 films”, Nature, 1991, 253,737-740.

[102] Turro, N. J., “Modem molecular photochemistry”, Benjamin/Cummings: Menlo Park, CA, 1978; Chapter 9. 161 [103] Kavamos, G. J. & Turro, N. J., “Photosensitization by reversible electron transfer: theories, experimental evidence, and examples”, Chem. Rev., 1986, 86, 401-449.

[104] Gould, I. R.; Young, R. H.; Mueller, L. J.; & Farid, S., “Mechanisms of exciplex formation. Roles of superexchange, solvent polarity, and driving force for electron transfer”, J. Am. Chem. Soc., 1994, 116, 8176-8187.

[105] A. Weller, “Singlet- and triplet-state exciplexes”, In “ The exciplex", Gordon, M. A.; Ware, W. R., Eds., Academic: London, 1975, pp 23-36.

[106] Mataga, N., “Photoinduced electron transfer and multiple states mechanisms”, Pure. Appl. Chem., 1993,(55, 1605-1610.

[107] Rehm,, D. & Weller, A., “Kinetics of fluorescence quenching by electron and H- atom transfer”, Isr. J. Chem., 1970, 8, 259-271.

[108] a. Devault, D., “Quantum mechanical tunnelling in biological systems”, Q. Rev. Biophys. 1980, 13, 387-564. b. Wasielewski, M. R., “Distance dependencies of electron transfer reactions”, In Photoinduced electron transfer”, Fox, M. A. & Chanon, M. Eds., Elsevier: Amsterdam, 1988, Part A, 161-206. c. Wasielewski, M. R., “Photoinduced electron transfer in supramolecular systems for artificial photosynthesis”, Chem. Rev., 1992, 92,435-461.

[109] Newton, M. & Sutin, N., “Electron transfer reaction in condensed phase”, Ann. Rev. Phys. Chem. 1984, 35, 437-480.

[110] Moser, C. C.; Keske, J. M.; Wamcke, K.; Farid, R. S.; & Dutton, P. L., “Nature of biological electron transfer”, Nature, 1992, 355,796-802.

[111] Shephard, M. J.; Paddon-Row, M. N.; & Jordan, K. D., “Why is a simple n-alkyl bridge more efficient than a polynorbornyl bridge at mediating through-bond coupling?”, J. Am. Chem. Soc., 1994, 116, 5328-5333.

[112] Marcus, R. A., “On the theory of electron transfer reactions VI. Unified treatment for homogeneous and electrode reactions”, J. Chem. Phys., 1965, 43, 679-701.

[113] Brunschewing, B. S.; Ehrenson, S.; & Sutin, N., “Distance dependence of electron-transfer reactions: Rate maximum and rapid rates at large reactant”, J. Am. Chem. Soc., 1984, 106, 6858-6859.

[114] a. Miller, J. R.; Calcaterra, L. T.; & Closs, G. L., “Intramolecular long-distance electron transfer in radical anions. The effects of free energy and solvent on the reaction rates”, J. Am. Chem. Soc., 1984, 106, 3047-3049. b. Closs, G. L. & Miller, J. R., “Intramolecular long-distance electron transfer in organic molecules”, Science, 1988, 240,440.

[115] Mauzerall, D. C., “Circumventing reverse electron transfer”, In “ Photoinduced electron transfer”, Fox, M. A. & Chanon, M. Eds., Elsevier: Amsterdam, 1988, Part A, pp 228-244. 162 [116] Beecroft, R. A.; Davidson, R. S.; Goodwin, D.; & Pratt, J. E., “Photo-induced electron transfer reactions-Pitfalls and opportunity”, Pure Appl. Chem , 1982, 54, 1605-1621.

[117] Mauzerall, D. & Ballard, S. G., “Ionization in solution by photoactivated electron tansfer”, Ann. Rev. Phys. Chem., 1982, 33, 377-407.

[118] Sutin, N., “Light-induced electron transfer reactions”, J. Photochem. 1979, 10, 19-40.

[119] Ross, R. T., “Bounds on rate constants and relative potentials in electron transport chains”, Biochim. biophys. Acta, 1977, 459, 321-324.

[120] Barber, J. & Andersson, B., “Revealing the blueprint of photosynthesis”, Nature, 1994,570,31-34.

[121] Tazuke, S.; Kitamura, N.; & Kawanishi, Y., “Problems of back electron transfer in electron transfer sensitization”, J. Photochem. 1985, 2 9 ,123-138.

[122] Tazuke, S.; Takasaki, R.; Iwaya, Y.; & Suzuki, Y., “Polymer-bonded electron- transfer sensitizers”, In “Materials for microlithography: Radiation-sensitive polymers ”, Thompson, L. F.; Willson, C. G.; 8c Frechet, J. M. J., Eds., ACS Symp. ser., 1984 ,266, 187-206.

[123] Yoon, U. C. & Mariano, P. S., “Mechanistic and Synthetic aspects of amine- enone single electron transfer photochemistry”, Acc. Chem. Res., 1992,25,233- 240.

[124] Lewis, F. D., “Proton-transfer reactions of photogenerated radical ion pairs”, Acc. Chem. Res., 1986, 19,401-405.

[125] a. Mauzerall, “The photoreduction on phorphyrins and oxidation of amines by photo-excited dyes”, J. Am. Chem. Soc., 1960 ,82, 1832-1833. b. Goodspeed, F. C.; Scott, B. L.; & Burr, J. G., “Photooxidation of tertiary nitrogen compounds by methylene blue”, J. Phys. Chem., 1965, 6 9 ,1149-1153.

[126] a. Symons, M. C. R. “Application of electron spin resonance spectroscopy to the study of the effects of ionising radiation on DNA and DNA complexes”, J. Chem. Soc., Faraday Trans. 1, 1987, 83, 1-11. b. Lemaire, D. G. E.; Bothe, E.; & Schulte-Frohlinde, D., “Yields of radiation- induced main chain scission of poly U in aqueous solution: strand break formation via base radicals”, Int. J. Radiat. Biol., 1984, 45 351-358.

[127] Deeble, D. J.; Schulz, D.; & von Sonntag, C., “Reactions of OH radicals with poly(U) in deoxygenated solutions: Sites of OH radical attack and the kinetics of base release”, Int. J. Radiat. Biol., 1986, 49,915-926.

[128] Oster, G. “Fluorescence quenching by nucleic acids”, Trans. Faraday Soc. 1950, 46 , 660-666.

[129] a. Georghiou, S., “Interaction of acridine drugs with DNA and nucleotides”, Photochem. Photobiol. 1977, 26, 59-68. b. Kubota, Y. and Fujisaki, Y., “Fluorescence of 9-aminoacridine bound to polynucleotides”, Bull. Chem. Soc. Japan. 1977, 50, 297-299. 163 c. Tubbs, K. R.; Ditmars, Jr. W. E.; & van Winkler, Q., “Heterogeneity of the interaction of DNA with acriflavine”, J. Mol. Biol. 1964,9,545-557.

[130] For example, a. Koch, T.; Ropp, J. D.; Sligar, S. G.; & Schuster, G. B.; “Photocleavage of DNA: Irradiation of quinone-containing reagents converts supercoiled to linear DNA”, Photochem. Photobiol., 1993, 58 ,554-558. b. Kagan, J.; Chen, X.; Wang, T. P.; & Forlot, P., “Psoralens cleave pBR322 DNA under ultraviolet radiation”, ibid, 1992 ,5 6 , 185-193. c. Nakamura, Y. “Photocrosslinking and the cleavage of DNA by iron complex- substituted psoralen”, Chem. Lett., 1989,51-54. d. Blau, W.; Croke, D. T.; Kelly, J. M., et al, “Base-specific photocleavage of DNA induced by nanosecond UV pulsed laser radiation or methylene blue sensitisation”, J. Chem. Soc., Chem. Commun., 1987,751-752.

[131] Dunn, D. A., Lin, V. H.; & Kochevar, I. E., “Base-selective oxidation and cleavage of DNA by photochemical cosensitized electron transfer”, Biochemistry, 1992,37,11620-11625.

[132] Steenken, S., “Purine bases, nucleosides, and nucleotides: Aqueous solution redox chemistry and transformation reactions of their radical cations and e- and OH adducts”, Chem. Rev. 1989, 89,503-520.

[133] a. Jovanovic S. V. 8c Simic, M. G., “One-electron redox potentials of purines and pyrimidine”, J. Phys. Chem., 1986, 90,974-978. b. Kittler, L.; Lober, G.; Gollmick, F. A.; & Berg, H. “Redox processes during photodynamic damage of DNA III. Redox mechanism of photosensitization and radical reaction”, Bioelectrochem. Bioenerg. 1980,7,503-511.

[134] McGlynn, S. P.; Doughterty, D.; Mathers, T.; & Abdulner, S., In “Excited states in Organic chemistry and biology", Pullman, B. & Goldblum, N., Eds., D. Reidel: Dordrecht, Holland, 1977, pp 247-256.

[135] Candeias, L., P. & Steenken, S., “Electron transfer in di(deoxy)nucleoside phosphates in aqueous solution: Rapid migration of oxidative damage (via adenine) to guanine”, J. Am. Chem. Soc, 1993, 115,2437-2440.

[136] Sevilla, M. D.; Becker, D.; Yan, M.; & Summerfield, S. R., “Relative abundances of primary ion radicals in g-irradiated DNA: Cytosine vs thymine anions and guanine vs adenine cations”, J. Phys. Chem., 1991, 95, 3409-3415.

[137] Vieira, A. J. S. C. & Steenken, S., “Pattern of OH radical reaction with N6,N6- dimethyladenosine. Production of three isomeric OH adducts and their dehydration and ring-opening reactions”, J. Am. Chem. Soc, 1987, 109,7441- 7448.

[138] Vieira, A. J. S. C. & Steenken, S., “Pattern of OH radical reaction with adenine and its nucleosides and nucleotides. Characterization of two types of isomeric OH adducts and their unimolecular transformation reactions”, J. Am. Chem. Soc, 1990, 112,6986-6994.

[139] Candeias, L., P. & Steenken, S., “Structure and acid-base properties of one- electron-oxidized deoxyguanosine, guanosine, and 1-methylguanosine”,./. Am. Chem. Soc, 1989, 111, 1094-1099. 164

[140] Cadet, J.; Berger, M.; Buchko, G. W.; Joshi, P. C.; Raoul, S.; & Ravanat, J.-L., “2,2-Diamino-4-[(3,5-di-0-acetyl-2-deoxy-{i-D-erythro-pentofuranosyl)amino]-5- (2H)-oxazolone: A novel and predominant radical oxidation product of 3’,5’-di-0- acetyl-2’-deoxyguanosine”, J. Am. Chem.. Soc, 1994, 116, 7403-7404.

[141] Deeble, D. J.; Schuchmann, M. N.; Steenken, & S.; von Sonntag, C., “Direct evidence for the formation of thymine radical cations from the reaction of S 0 4 *- with thymine derivatives: A pulse radiolysis study with optical and conductance detection”, J. Phys. Chem. 1990, 96, 8186-8192.

[142] Cadet, J.; Berger, M.; Decarroz, C.; Mouret, J. F.; van Lier, J. E.; & Wagner, R. J., “Oxydations radicalaires photo- et radio-induites bases purines et tyrimidiniques des aides nucleiques”, J. Chim. Phys., 1991, 88, 1021-1042.

[143] Wagner, J. E.; van Lier, J. E.; Decarroz, C.; Berger, M.; & Cadet, J., “Photodynamic methods for oxy radical-induced DNA damage”, Methods Enzymol., 1990, 186, 502-511.

[144] Steenken, S.; Telo, J. P.; Novais, H. M.; & Candeias, L. P., “One-electron- reduction potentials of pyrimidine bases, nucleosides, and nucleotides in aqueous solution. Consequences for DNA redox chemistry”, J. Am. Chem. Soc., 1992, 114, 4701-4709.

[145] Sevilla, M. D.; Becker, D.; Yan, M.; & Summerfield, S. R., “Relative abundances of primary ion radicals in y-irradiated DNA: Cytosine vs thymine anions and guanine vs adenine cations”, J. Phys. Chem., 1991, 95, 3409-3415.

[146] Yan, M.; Becker, D.; Summerfield, S.; Renke, P.; & Sevilla, M. D., “Relative abundance and reactivity of primary ion radicals in y-irradiated DNA at low temperatures. 2. Single- vs double-stranded DNA”,}. Phys. Chem., 1992, 96, 1983-1989.

[147] a. Hissung, A.; von Sonntag, C.; Veltwisch, D.; & Asmus, K.-D., “The reactions of the 2’-deoxyadenosine electron adduct in aqueous solution. The effects of the radiosensitizer p-nitroacetophenone. A pulse spectroscopic and pulse conductometric study”, Int. J. Radiat. Biol., 1981, 39,63-71. b. Visscher, K. J.; de Haas, M. P.; Loman, H.; Vojnovic, B.; & Warman, J. M., “Fast protonation of adenosine and of its radical anion formed by hydrated electron attack: A nanosecond optical and dc-conductivity pulse radiolysis study" ibid, 1987,52,745-753.

[148]Candeias, L. P.; Wolf, P.; O’Neil, P.; & Steenken, S., “Reaction of hydrated electrons with guanine nucleosides: Fast protonation on carbon of the electron adduct”,/. Phys. Chem., 1992, 96, 10302-10307.

[149] Fridovich, 1., “The biology of oxygen radical”, Science, 1978, 201, 875-880.

[150] a. Henner, W. D.; Grunberg, S. M.; & Haseltine, W. A., “Enzyme action at 3’- termini of ionizing radiation-induced DNA strand breaks”, J. Biol. Chem., 1983, 258,15198-15205. b. Wu, J. G; Kozarich, J. W.; & Stubbe, J., “The mechanism of free base formation from DNA by bleomycin”, ibid, 1983, 258,4694-4697. 165

[151] a. Schultz, P. G.; Taylor, J. S.; & Dervan, P. B., “Design and synthesis of a sequence-specific DNA cleaving molecule. (Distamycin-EDTA)iron(II)”, J. Am. Chem. Soc., 1982, 104,6861-6863. b. Dervan, P, B., “Characterization of protein-DNA complexes by affinity cleaving”, Methods Enzymol., 1991, 208,497-515.

[152] Prigodich, R. V. 8c Martin C. T., “Reaction of single-stranded DNA with hydroxyl radical generated by iron(II)-ethylenediaminetetraacetic acid”. Biochemistry , 1990,29, 8017-8019.

[153] Sigman, D. S.; Bruice, T. W.; Mazumder, A.; & Sutton, C., “Targeted chemical nucleases”, Acc. Chem. Res., 1993, 26, 98-104.

[154] Sigman, D. S.; Mazumder, A.; & Perrin, D. M., “Chemical nucleases”, Chem. Rev., 1993,93,2295-2316.

[155] a. Mei, H. Y. & Barton, J. K., “Tris(tetramethylphenathroline)ruthenium(II): A chiral probe that cleaves A-DNA conformations”, Proc. Natl. Acad. Sci. USA, 1988, 85, 1339-1343. b. Uchida, K.; Pyle, A. M.; Morii, T.; & Barton, J. K., “High resolution footprinting of EcflRl and distamycin with Rh(phi) 2 (bpy)3+, a new photofootprinting reagent”, Nucleic Acids Res., 1988, 17,10259-10279.

[156] Armitage, B.; Yu, C.; Devadoss, & Schuster, G. B., “Cationic anthraquinone derivatives as catalytic DNA photonucleases: Mechanisms for DNA damage and quinone recycling”, j. Am. Chem. Soc., 1994, 116, 9847-9859.

[157] Pathak, M. A.; Kramer, D. M.; & Fitzpatrick, T. B., Sunlight and Man, University of Tokyo Press, 1973.

[158] Musajo, L.; Rodighiero, G.; Colombo, I.; Torlane, & Dall’acqua, F., “Photosensitizing furocoumarins interactions with DNA and photoinactivation of DNA containing viruses”, Experientia, 1965, 21,22-24.

[159] a. Musaio, L.; Bordin, G.; Caporale, S.; Maraiani, & Rigatti, G., “Photoreactions at 3655A between pyrimidine bases and skin-photosensitizing furocoumarins” Photochem. photobiol., 1967, 6 , 711-719. b. Musajo, L.; Bordin, G;; & Bevilacqua, R., “Photoreactions at 3655A linking the 3-4 double bond of furocoumarins with pyrimidine bases”, ibid, 1967, 6 , 927-931.

[160] a. Gasparro, F. P, Ed., “Psoralen DNA photobiology", CRC Press, Inc., Boca Raton, FL, 1989. b. Cimino, G. D.; Gamper, H. B.; Isaacs, S. T.; & Hearst, J. E., “Psoralens as photoactive probes of nucleic acid structure and function: Organic chemistry, photochemistry, and biochemistry”, Ann. Rev. Biochem., 1985, 5 4 ,1151-1193.

[161] Regan, J. D. & Parrish, J. A., The Science of Photomedicine, Plenum Press, New York, 1982.

[162] Zmudzka, B. Z.; Strickland, A. G.; Miller, S. A.; Valerie, K.; Dall’Acqua, F.; & Beer, J. Z., “Action of the human immunodeficiency virus promoter by UVA 166 radiation in combination with psoralens or angelicins”, Photochem. photobiol., 1993, 58,226-232. [163] Ljungren, B.; Carter, D. M.; Albert, J.; & Reid, T., “Plasma levels of 8- methoxypsoralen determined by high-pressure liquid chromatography in psoriatic patients ingesting drug from two manufacturers”, J. Invest. Dermatol., 1980, 74, 59-62.

[164] Brickl, R.; Schmid, J.; & Koss, F. W., “Clinical pharmacology of oral psoralen drugs”, Photodermatology, 1984, 1, 174.

[165] a. Cole, R. S., “Light-induced cross-linking of DNA in the presence of a furocoumarin (psoralen): Studies with phage X, Escherichia coli, and mouse leukemia cells”, Biochim. Biophys. Acta 1970,2/7, 30-39. b. Cole, R. S. & Zusman, D., “Sedimentation properties of phage DNA molecules containing light-induced psoralen cross-links”, ibid 1970, 224,660-662.

[166] Midden, W. R., “Chemical mechanisms of the bioeffects of furocoumarins: The role of reactions with proteins, lipids, and other cellular constituents”, In “Psoralen DNA photobiology ”, Gaspairo, F. P, Ed., CRC Press, Inc., Boca Raton, FL, 1989, Vol. II, pp 1-50.

[167] Straub, K.; Kanne, D.; Hearst, J. E.; & Rapoport, H., “Isolation and characterization of pyrimidine-psoralen photoadducts from DNA”, J. Am. Chem. Soc., 1981, 103,2347-2355.

[168] Dall’acqua, F.; Terbojevich, M.; Marciani, S.; Veddaldi, D.; & Recher, M., “Investigation on the dark interactionbetween furocoamarins and DNA”, Chem. Biol. Inter., 1978, 2 1 ,103.

[169] Romer, R. & Anders, A., “NMR study of the drug-base overlap geometry in the complex of 8-methoxypsoralen and d(pApT) 4 ”, Biochemistry, 1985, 24, 7450-7456.

[170] Dall’Acqua, F.; Marciani-Mango, S.; Zambon, F.; & Rodighiero, G., “Kinetic analysis of photoreaction(315nm) between psoralens and DNA", Photochem. photobiol., 1979, 29,489-495.

[171]a. Hansen, J. B., Bjerring, P.; Buchardt, O., etal, “Psoralenamines. III. Synthesis, pharmacological behavior, and DNA binding of 5-(aminomethyl)-8-methoxy- 5-[[(3-aminopropyl)oxy]methyl]-, and 8-[(3-aminopropyl)oxy]psoralen derivatives”,/. Med. Chem., 1985, 2 8 ,1001-1010. b. Kraemer, K. H.; Waters, H. L.; Elingson, O. L.; & Tarone, R. E., “Psoralen plus ultraviolet radiation-induced inhibition of DNA synthesis and viability in human lymphoid cells in vitro”, Photochem. photobiol., 1979, 30,263-270.

[172] Song, P.-S.; Harter, M. L.; Moore, T. A.; & Herndon, W. C., “Luminescence spectra and photocycloaddition of the excited coumarins to DNA bases”, Photochem. photobiol., 1971, 14,521-530.

[173] Bevilacqua, R. & Bordin, F., “Photo-C4-cycloaddition of psoralen and pyrimidine bases: Effect of oxygen and paramagnetic ions”, Photochem. photobiol., 1973, 17, 191-194. 167 [174] Mantulin, W. & Song, P.-S. “Excited states of skin-sensitizing coumarins and psoralens. Spectrum studies”,/. Am. Chem. Soc., 1973, 95,5122-5129.

[175] Urn, E. C., “Proximity effect in molecular photophysics: Dynamical consequences of pseudo-Jahn-Teller interaction”, J. Phys. Chem. 1986, 90,6770-6777.

[176] Sasaki, M; Sakata, T; & Sukigara, M. “Photochemical study on the photochemotherapy. I. Solvent effect on fluorescence spectrum of 8- methoxypsoralen”, Chem. Letters, 1977,701-704.

[177] a. Musajo, L.; Rodighiero, J.; Breccia, A.; Dall’acqua, F.; & Malesani, G., “Skin- photosensitizing furocoumarins: Photochemical interaction between DNA and -0 ,4CH3 bergapten (5-methoxypsoralen)”, Photochem. Photobiol., 1966,5,739- 745. b. Pathak, M. A.; & Kramer, D. M., “Photosensitization of skin in vivo by furocoumarins (psoralen)”, Biochim. Biophys. Acta 1969, 195,197-206.

[178] Song, P. S., “Photoreactive states of furocoumarins”, In “ Photobiologic, toxicologic, and pharmacologic aspects of psoralens", (NIH Publication #84- 2692), Natl. Cancer Inst. Monogr., 1984, 6 6 , 15-19.

[179] Beaumont, P. G ; Parsons, B. J.; Phillips, G. O.; & Allen, J. G , “A laser flash photolysis study of the reactivities os 8-methoxypsoralen and 4,5’,8- trimethylpsoralen with nucleic acid bases in solution”, Biochim. Biophys. Acta, 1979, 562, 214-221.

[180] Salet, G; Macedo de Sa E Melo, T.; Bensasson, R. V.; & Land, E. J., “Photophysical properties of aminomethylpsoralen in presence and absence of DNA”, Biochim. Biophys. Acta, 1980, 607, 379-383.

[181] Ben-Hur, E. & Song, P. S., Adv. Radiat. Biol., 1984, 11, 131-171.

[182] Kanne, D.; Rapoport, H.; & Hearst, J. E., “8-Methoxypsoralen-nucIeic acid photoreaction. Effect of methyl substitution on pyrone vs. futan photoaddition”, J. Med. Chem., 1984,27, 531-534.

[183] Gasparro, F. P., “Psoralen-DNA interactions: Thermodynamics and photochemistry”, In “ Psoralen DNA photobiology ”, Gasparro, F. P, Ed., CRC Press, Inc., Boca Raton, FL, 1989, pp 5-36.

[184] Chatterjee, P. K. & Cantor, C. R., “Photochemical production of psoralen-DNA monoadducts capable of subsequent photocrosslinking”, Nucleic Acids Res., 1978, 5, 3619-3633.

[185] Yun, M. H.; Choi, S. J.; & Shim, S. C., “A novel photoadduct of 4,5’,8- trimethylpsoralen and adenosine”, Photochem. Photobiol., 1992 ,5 5 ,457-460.

[186] Yun, M. H.; Choi, S. J.; & Shim, S. A., “Adenosine-mediated photoreaction of 4,5’,8-trimethylpsoralen with alcohols”, Photochem. Photobiol., 1993 ,58, 164- 168.

[187] Bordin, F.; Carlassare, F.; Busulini, L.; & Baccichetti, F., “Furocoumarin sensitization induces DNA-protein cross-links”, Photochem. Photobiol., 1993, 58, 133-136. 168

[188] a. Bensasson, R. V.; Chalvet, O.; Land, E. J.; & Ronfard-Haret, J. C., “Triplet, radical anion and radical cation spectra of furocoumarins”, Photochem. Photobiol., 1984, 39, 287-291. b. Bensasson, R. V.; Land, E. J.; & Truscott, T. G., “Flash photolysis and pulse radiolysis: Contributions to the chemistry of biology and medicine”, Pergamon Press, Oxford, England, 1983, pp 192-204.

[189] Bensasson, R. V.; Salet, C.; Land, E. J.; & Rushton, F. A. P., “Triplet excited state of the4*5*-photoadduct of psoralen and thymine”, Photochem. Photobiol., 1980,3/, 129-133.

[190] Craw, M.; Chedekel, M. R.; Truscott, T. G.; & Land, E. J., “The photochemical interaction of 8-methoxypsoralen and the melanin precursor L-3,4- dihydroxyphenylalanine”, Photochem. Photobiol., 1984, 39, 155-159.

[191] Foote, C. S., “Definition of type I and type II photosensitized oxidation”, Photochem. Photobiol., 1991, 54, 659.

[192] a.Poppe, W. & Grossweiner, L. I. Photochem. Photobiol., 1975,22, 217-219. b. Singh, H. & Vadsaz, J. A. ibid 1978, 28,539-545. c. Joshi, P. C. & Pathak, M. A., “Production of singlet oxygen and superoxide reactions by psoralens and their biological significance”, Biochem. Biophys. Res. Comm. 1983, 112,638-646.

[193] a. de Mol, N. J. & Beijersbergen van Henegouwen, G. M. J. “Formation of singlet oxygen by 8-methoxypsoralen”, Photochem. Photobiol., 1979, 30, 331-335. b. de Mol, N. J.; Beijersbergen van Henegouwen, G. M. J.; & vanBeele, B. “Singlet oxygen formation by sensitization of furocoumarins complexed with, or bound covalently to DNA”, ibid 1981, 34,661-666.

[194] Knox, C. N.; Land, E. J.; & Truscott, T. G., “Singlet oxygen generation by furocoumarin triplet states—I. Linear furocoumarins (psoralen)”, Photochem. Photobiol., 1986, 43, 359-363.

[195] Beaumont, P. C.; Rodgers, M. A. J.; Parsons, B. J.; & Glyn, O., “Singlet oxygen production by some furocoumarin derivatives in the presence of DNA: Timeresolved luminescence measurements”, Photochem. Photobiol., 1985, 42, 605-608.

[196] Isaacs, S. T. Chum, C. Hyde, J. E. Rapoport, H. and Hearst J. E. “A photochemical characterization of reactions of psoralen derivatives with DNA”. In “Trends in Photobiology ” Helene, C. Charlier, M. Montenay-Garastier, T. Lanstriat, G. Ed., Plenum Press, New York, 1982, pp. 279-294.

[197] Baccichetti, F.; Antonello, C.; Carlassare, F.; Gia, O.; Palumbo, M.; & Bordin, F., “Photobiological properties of some water-soluble 5-MOP and 8-MOP derivatives”, Photodermatology, 1985,2,221.

[198] Voelk, E., “The photolytic action of 3-amino-6-iodoacridine with nucleic acids”, Ph D. Dissertation. The Ohio State University, 1991.

[199] a.Martin, R. F. & Tong, J. H., “Synthesis of 3-aminoacridine from proflavine”, Aust. J. Chem., 1969,22, 487-489. 169 b. Martin, R. F. & Kelly, D. P., “Synthesis and NMR spectra of substituted aminoiodoacridines”, ibid, 1979, 32, 2637-46.

[200] Soundararajan & Platz, M. S., “Descriptive photochemistry of polyfluorinated azide derivatives of methyl benzoate”, J. Org. Chem. 1990,55,2034-2044.

[201] Isaacs, S. T. Rapoport, H. & Hearst J. E., “Synthesis of deuterium and tritium labelled psoralens. J. Labelled Compd. Radiopharm. 1981, 19,345-356.

[202] Antonello, C. Marciani Magno, S. Gia, O. Carlassare, F. Baccichetti, F. & Bordin, F., “Diethylaminoalkyloxycoumarin and -furocoumarin derivatives. Farmco. Ed. Sci. 1978 ,3 4 ,139-156.

[203] Brokke, M. E. & Christensen, B. E., “Psoralen I: certain reactions of xanthotoxin”, J. Org. Chem. 1958, 23, 589-596.

[204] a. Kadry, H.; Osman, A. N.; Al-Azizi, M.; & Ashour, M. S., “Synthesis of furo[3,2-g]coumarin derivatives likely to possess antimicrobial activity”, Rev. Roum. Chim., 1992, 37,915-920. b. Rai, S., “Viral inactivation using DNA intercalators and dyes for sterization of blood product”, Ph D. Dissertation. The Ohio State University, 1992.

[205] Foster, N. I.; Dammals, R.; Bums, D.; & Helndel, D. N., J. Radioanal. Chem., 1981,65,95.

[206] Fuson, R. C.; Kneisley, J. W.; & Kaiser, E. W., “Coumarilic acid”, Org. Synth. Coll. Ill, 1967,209-211.

[207] Kulaga, L. M., “Binding affinity studies of potential antiviral drugs for use in blood sterization”, M. S. Thesis, The Ohio State University, 1994.

[208] Isaacs, S. T.; Shen, C.-K, J.; Hearst, J. E.; & Rapoport, H., “Synthesis and characterization of new psoralen derivatives with superior photoreactivity with DNA and RNA”, Biochemistry, 1977, 16, 1058-1064.

[209] Schmechel, D.E. V. & Crothers.D. M., “Kinetic and hydrodynamic studies of the complex of proflavine with poly A.ploy U”, Biopolymers, 1971, 10,465-480.

[210] Hansen, J. B.; Koch, T.; Buchardt, O.; Nielsen, P. E.; Wirth, M.; & Norden, B., “Acridine-psoralen amines and their interaction with deoxyribonucleic acid”, Biochemistry, 1983,22,4878-4886.

[211] Olmsted, J. Ill & Kearns, D. R., “Mechanism of ethidium bromide fluorescence enhancement on binding to nucleic acids”, Biochemistry, 1977, 16, 3647-3654.

[212] Reinhardt, C. G. & Krugh, T. R., “A comparative study of ethidium bromide complexes with dinucleotides and DNA: Direct evidence for intercalation and nucleic acid sequence preferences”, Biochemistry, 1978, 17,4845-4854.

[213] Philpott, M. R., “Exciton theory of the electronic states of dye-polymer complexes. 1. Linear polymers”, ./. Chem. Phys., 1970,53,968-981.

[214] Muller, W. & Crothers, D. M., “Interactions of heteroaromatic compounds with nucleic acids”,Ear. J. Biochem., 1975, 54, 267-277. 170

[215] The author is grateful for the generous help of these people in determining the fluorescence lifetime of proflavine. Toscano, J.; Butterbaugh, J.; & Gustafson, T., unpublished research at the Ohio State University.

[216] Sa e Melo, T.; Bazin, M.; Ronfard-Haret, J. C.; & Santus, R., “Quantification of the solvent effects on the triplet quantum yield of psoralen by the linear solvation energy relationship”, Photochem. Photobiol., 1993, 58, 19-27.

[217] Reichardt, G , Solvents and solvent effects in organic chemistry, 2nd Ed., VCH, Weinheim, 1988.

[218] Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H., & Taft, R. W., “Linear solvation energy relationships. 23. A comprehensive collection of the solvatochromic parameters, n*, a, and (3, and some methods for simplifying the generalized solvatochromic equation”, J. Org. Chem., 1983, 48,2877-2887.

[219] Ishikawa, M., “solvation effect on the picosecond flyorescence lifetimes and anisotropy decays of a hydrogen-bounded biologically active furocoumarin derivative: 4,5’,8-trimethylpsoralen”, Chem. Phys. Letter, 1990, 173,76-82.

[220] Matsumoto, H. & Isobe, A., “Solvent effect on the emission spectra and fluorescence quantum yields of psoralens”, Chem. Pharm. Bull., 1981,29, 603- 608.

[221] Menger, E. L.; Bhattacharjee, H. R.; & Hammond, G. S., “Quenching of the fluorescence of 8-methoxypsoralen by protons”, Photochem. Photobiol., 1979, 29, 393-394.

[222] Sa e Melo, T.; Macanita, A.; Prieto, M.; Bazin, M.; Ronfard-Haret, J. C.; & Santus, R., “Photophysical behaviour of 5-methoxypsoralen in dioxane-water mixture”, Photochem. Photobiol., 1988, 48,429-437.

[223] Murov, S. L.; Carmichael, I.; & Hug, G. L., Handbook of photochemistry, Second Ed., Marcel Dekker, New York, 1993.

[224] Kasturi, C., “Photochemical inactivation of viruses using psoralen, naphthalene and porphyrin derivatives: Development of a protocol for the selective inactivation of viruses in blood products”, Ph D. Dissertation. The Ohio State University, 1993.

[225] a.Semin, D. J.; Winkler, P. C.; & Rowlen, K. L., “Psoralen-olefin photoproducts: First observation of a photo-ene reaction”, Photochem. Photobiol., 1994, 6 0 , 185- 195. b. Otsuki, T., “Photochemical reaction of furocoumarin (psoralen). The reaction site in its photoexcited state” Chem. Lett., 1987,453-454. c. Shim, S. C. & Kim, Y. Z„ “The photocycloaddition of 8-methoxypsoralen with olefins”, J. Photochem., 1983, 23, 83-92.

[226] Schuster, G. B. & Platz, M. S., “Photochemistry of phenyl azide”, Adv. Photochem., 1992,77,69-143. 171 [227] Li, Y-Z.; Kirby, J. P.; George, M. W.; Poliakoff, M.; & Schuster, G. B., “ 1,2- Didehydroazepines from the photolysis of substituted aryl azides: Analysis of their chemical and physical properties by time-resolved spectroscopic methods”, J. Am. Chem. Soc. 1988, 110, 8092-8098.

[228] Sutcliffe, J. G., “Complete nucleotide sequence of the Escherichia coli plasmid pBR322”, Cold Spring Harbor Symp. Quant. Biol., 1978, 43,77-90.

[229] Bauer, W. R., “Structure and reactions of closed duplex DNA”, Annu. Rev. Biophys. Bioeng. 1978, 7,287-313.

[230] Keck, M. V. & Lippard, S. J., “Unwinding of supercoiled DNA by platinum- ethidium and related complexes”, J. Am. Chem. Soc. 1992, 114, 3386-3390.

[231] Boon, P. J.; Cullis, P. M.; Symons, M. C. R.; & Wren, B. W., “Effects of ionizing radiation on deoxyribonucleic acid and related systems. Part 1. The role of oxygen”, J. Chem. Soc. Perkin Trans. II, 1984, 1393-1399.

[232] a. McDonell, M W.; Simon, M. N.; & Studier, F. W., “Analysis of restriction fragments of T7 DNA and determination of molecular weights by electrophoresis in neutral and alkaline gels”, J. Mol. Biol. 1977, 110,119-146. b. Southern, E., “Gel electrophoresis of restriction fragments”, Meth. Enzymol. 1979, 6 8 ,152-176.

[233] Mickel, S.; Arena, V. Jr.; & Bauer, W ., “Physical properties and gel electrophoresis behavior of R12-derived plasmid DNAs”, Nucleic Acid Res. 1977, 4, 1465-1482.

[234] Peckler, S.; Graves, B.; Kanne, D.; Rapoport, H.; Hearst, J. E.; & Kim, S.-H., “Structure of a psoralen-thymine monoadduct formed in photoreaction with DNA”, J. Mol. Biol. 1982, 162, 157-172.

[235] Pearlman, D. A.; Holbrook, S. R.; Pirkle, D. H.; & Kim, S.-H., “Molecular models for DNA damage by photoreaction”, Science, 1985,227, 1304-1308.

[236] Tomic, M. T.; Wemmer, D. E.; & Kim, S.-H., “Structure of a psoralen cross- linked DNA in solution by nuclear magnetic resonance”, Science, 1987, 238, 1722-1725.

[237] a. Marini, J. C.; Levene, S. D.; Crothers, D. M.; & Englund, P. T., “Bent helical structure in kinetoplast DNA”, Proc. Natl. Acad. Sci. USA, 1982, 79,7664-7668. b. Wu, H.-M. & Crothers, D. M., “The locus of sequence-directed and protein- induced DNA bending”, Nature, 1984, 308, 509-513.

[238] Lumpkin, O. J.;& Zimm, B. H., “Mobility of DNA in gel electrophoresis”, Biopolymers 1982, 21,2315-2316.

[239] Sinden, R. R.; & Hagerman, P. J., “Interstrand psoralen cross-links do not introduce appreciable bends in DNA”, Biochemistry 1984, 23, 6299-6303. b. Zhen, W.-P.; Dahl, O.; Buchardt, O.; & Nielsen, P. E., “On the bending by psoralen interstrand crosslinking. A gel electrophoresis study", Photochem. Photobiol., 1988, 48, 643-646. 172 [240] Isaac, S. T.; Wiesehahn, G.; & Hallick, L. M., “In vitro characterization of the reaction of four psoralen derivatives with DNA”, In “ Photobiologic, toxicologic , and pharmacologic aspects of psoralens”, (NIH Publication #84-2692), Natl. Cancer Inst. Monogr. 1984, 66, 21-30.

[241] a. Keller, W., “Mechanism of action of the pyruvate dehydrogenase multienzyme complex from Escherichia coli”, Proc. Nat. Acad. Sci. USA. 1975, 72,4876-4880. b. Wang, J. C., “The degree of unwinding of the DNA by ethidium I. Titration of twisted PM2 DNA molecules in alkaline cesium chloride density gradients”, J. Mol. Biol. 1974, 89,783-801.

[242] Bordin, F.; Conconi, M. T. and Capozzi, A., “Certain singlet oxygen quenchers affect the photoreaction between 8-MOP and DNA”, Photochem. Photobiol. 1987, 46, 301-304.

[243] a. Held, K. D.; Harrop, H. A.; & Michael, B. D., “Effects of oxygen and sulphydryl-containing compounds on irradiated transforming DNA Part I. Actions of dithiothreitol”, Int. J. Radiat. Biol. 1981, 40,613-622. b. Held, K. D.; Harrop, H. A.; & Michael, B. D.,“Effects of oxygen and sulphydryl-containing compounds on irradiated transforming DNA Part II. Glutathione, Cysteine and cysteamine”, ibid 1984, 45,615-626. c. Held, K. D.; Harrop, H. A.; & Michael, B. D.,“Effects of oxygen and sulphydryl-containing compounds on irradiated transforming DNA Part III. Reaction Rates”, ibid 1984, 45, 627-636.

[244] a. Rodgers, M. A. J. & Snowden, P. T., “Lifetime of O 2 (’Ag) in liquid water as determined by time-resolved infrared luminescence measurement”, J. Am. Chem. Soc. 1982, 104,5541-5543. b. Merkel, P.B. & Kearns, D., “Radiationless decay of singlet molecular oxygen in solution. An experimental and theoretical study of electronic-to vibrational energy transfer”, ibid 1972, 94,7244-7253.

[245] a. Blazek, E. R.; Peak, J. G. and Peak, M. J. “Singlet oxygen induces frank strand breaks as well as alkali- and piperidine-labile sites in supercoiled plasmid DNA", Photochem. Photobiol. 1989, 49, 607-613. b. Peak, M. J.; Peak, J. G. and Carnes, B. A., “Induction of direct and indirect single-strand breaks in human cell DNA by far- and near-ultraviolet radiations: Action spectrum and mechanisum”, ibid 1987, 45, 381-387.

[246] Stenstrom, A. G. K.; Moan, J.; Brunborg, G. and Eklund, T., “Photodynamic inactivation of yeast cells sensitized by hematopophyrin”, Photochem. Photobiol. 1980, 32, 349-352.

[247] Kasha, M., “Characterization of electronic transitions in complex molecules” Discuss. Faraday Soc., 1950, 9, 14-19.

[248] Freeman, P. K.; Jang, J.-S.; & Ramnath, N., “The photochemistry of polyhaloarenes. 10. The photochemistry of bromobiphenyl”, J. Org. Chem., 1991,56,6072-6079.

[249] Brun, A. M. & Harriman, A. “Energy- and electron transfer Processes involving palladium porphyrins bound to DNA”, ./. Am. Chem. Soc. 1994, 116, 10383-10393. APPENDIX A

Selected Gel Electrophoresis Plates

173 174

Plate I. Photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with Acr-NH2 (50|iM, bottom lanes) and Acr-I(50|iM, top lanes) and 350nm light

(I=7.7mw/cm2). First and last lanes, molecular marker, second lanes, DNA exposed to

UV light for 75min; lanes 3 to 13, DNA plus sensetizer exposed to light for 0, 5,10,15,

20, 25, 30,40, 50,60min respectively. 175 Plate II. photograph of agarose gel electrophoresis of the photocleavage of pBR322

DNA with Acr-NH2 and Acr-1 and 400-440nm light (9.2kJ/m2). Top: lanes 1 and 2 are DNA, DNA+hv respectively, lanes 3-8 are Acr-I with concentrations of 10, 20,50, 80,

100, and lOOjiM+DTT (0.1M), lane 9, molecular marker, lanes 10-15, Acr-NH 2 with concentrations of 10, 20,50, 80, 100, and 100|iM+DTT (0.1M); Below: lanes 2-6, Acr- I with concentrations of 10,20, 50, 80, and lOOpM under deoxygened conditions, lanes

7 and 8, DNA after deoxygenation process and molecular marker, lanes 9-12, Acr-NH 2 with concentrations of 20, 50, 80, and 100pM under deoxygened conditions Plate III. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with HEP and 350nm light (49.2kJ/m2). Lane 1 is molecular marker, lane 2 is DNA + 0.6mM sensitizer in the dark, lanes 3-15 are DNA irradiated with sensitizer of concentrations of 0, 2, 5, 10, 15, 20, 30, 60,200, 400, 500, 600 and 600|J.M + 0.2M DTT, respectively.

Plate IV. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with CEP and 350nm light (49.2kJ/m2). Lane 1 is molecular marker, lane 2 is DNA + 0.5mM sensitizer in the dark, lanes 3-15 are DNA irradiated with sensitizer of concentrations of 0, 2, 5, 10, 15. 20, 30, 50, 100, 200, 300, 500 and 500^tM + 0.1M DTT, respectively 177 Plate V. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with BEP and 350nm light (49.2kJ/m2). Lane 1 is molecular marker, lane 2-3 are DNA and DNA + 50pM sensitizer in the dark, lanes 4-15 are DNA irradiated with sensitizer of concentrations of 0, I, 2, 4, 6, 10, 15, 20, 30, 40, 50, and 50ftM + 0.1M DTT, respectively. 178

Plate VI. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with IEP and 350nm light (49.2kJ/m2). Lane 1 is molecular marker, lane 2-3 are DNA and DNA + 60pM sensitizer in the dark, lanes 4-15 are DNA irradiated with sensitizer of concentrations of 0, 2, 4, 6, 10, 15, 20, 30, 40, 50, 60, and 60|i.M + 0.1M DTT, respectively. 179 Plate VII. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with TBP and 350nm light (49.2kJ/m2). Lane 1 is molecular marker, lane 2-3 are DNA and DNA + 40|iM sensitizer in the dark, lanes 4-15 are DNA irradiated with sensitizer of concentrations of 0, 0.5, 1, 2, 4, 6, 10, 15, 20, 30, 40, and 40|i.M + 0.1M DTT, respectively. 180

Plate VIII. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with BCP and 350nm light (49.2kJ/m2). Lane 1 is molecular marker, lane 2-3 are DNA and DNA + 50|iM sensitizer in the dark, lanes 4-15 are DNA irradiated with sensitizer of concentrations of 0, 1, 2, 4, 6 , 10, 15, 20, 30, 40, 50, and 50|iM + 0.1M DTT, respectively. Plate IX. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with BPP and 350nni light (49.2kJ/m2). Lane 1 is molecular marker, lane 2-3 are DNA and DNA + 50|iM sensitizer in the dark, lanes 4-15 are DNA irradiated with sensitizer of concentrations of 0, 2, 5, 10, 15, 20, 30, 40, 50, 60,70, and 50p.M + 0.1M DTT, respectively. 182 Plate X. photograph of agarose gel electrophoresis of the photocleavage of pBR322 DNA with AMT and 350nni light (49.2kJ/m2). Lane 1 is molecular marker, lane 2-3 are DNA and DNA + IOOjiM sensitizer in the dark, lanes 4-15 are DNA irradiated with sensitizer of concentrations of 0, 5, 10, 15, 20, 30, 50, 60, 70, 80, 100, and 100)iM + 0.1M DTT, respectively. APPEDIX B

Selected Absorption Spectra

183 LASTACQU.ttRD (25G.G - 6GQ.G)

iP. N^it.Ac

303 350 •100 500 Nanometers

Spectrum 1. UV-Vis aborption of Acr-NH2 in 1% NH4Ac(aq), pH=5.5 . Nanometers

Spectrum 2. UV-Vis aborption of Acr-NH 2 in the solution of DNA Tris HC1 buffer(10mM Tris base, 0.2M NaCl, pH=7.4). [Acr-NH2]= 1.8 x 10-5M, [DNA]=1 .OxlO^M. LflSTACQU.flRD (250.0 - 600.0) 3-ftmino-6-iodoacridine A Sv -?4

p/V >3 [ ft .

a b s

0 r b a cVb n / " ’“ X c \ e \ v . / \ \

~T I i I j I I n I I I I |"l I I I I l~I r [ n T i " T T l ; j T T T" -[ n 11 m 11 r i 11 i 11 t j 3-X) 350 400 450 500 550 600

Nanometers

Spectrum 3. UV-Vis aborption of Acr-I in 1% NH4Ac(aq), pH=5.5 . LASTACQU.flRD (Z50.0 - 600.0)

2.406 260 ft IA b s 2- 0 r b a n e

u .r /) 0.222 4VG

\

i • i i i i i i i'ii. i j"i i f ~i• j• i r T T i- r r i : j i r i i r r~i~r t |t i~i i i i i r~ i ; ■ . i i V n ~H r vj 1 300 350 400 450 5G0 550 600

Nanoneters

Spectrum 4. UV-Vis aborption of Acr-I in the solution of DNA Tris HC1 buffer(10mM Tris base, 0.2M NaCl, pH=7.4). [Acr-I]= 1 .5x 10‘5M , [DNA]=1 .Ox lO^M. 2.053 203 2-1 A A/j

0 0 >\j

1-1 0.650 z n 0.523

A 320

x \

p — r *i~'TTTTrrr-rrrrnr r~ t~i-t I ' I i i i i ~r i i ; i l TI ; t—i-'t ;t 200 250 300 350 400 450 500 Manometers

Spectrum 5. UV-Vis aborption of AMP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). Spectrum 6.UV-Vis aborption of the product of AMP after Spectrum photolysis with350nm light in EPA at 77K.

Absorbance 0 1 5 40 5 50 5 60 5 70 5 800 750 700 650 600 550 500 450 400 350 408 5 485 458 Nanometers 586 OCH 637 in EPAat 77K EPAat in 1.0 0.909 236 0.762 275

O'

0.5

200 tCO 450 503

Nanometers

Spectrum 7. UV-Vis aborption of ADP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). 190 pcrmSpectrum Transient Absorption x 10'3 c) a) 8 . Transient aborption produced by LFP of ADP (2.9xl0 ADP of LFP by produced aborption . Transient with [DNA] a) 0, b) 2xl0 b) 0, a) [DNA] with fe h lsrple )wsrcre (satrtelsrple £ pulse. laser the after 3(is recorded was d) pulse, laser the after -40 -80 40 5 40 5 650 550 450 350 350 450 4 550 M, c) and d) 6xl0 d) and c) M, 650 Nanometers 750 4 M. a), b), and c) were recorded over a window of 400ns immediately immediately 400ns of window a over recorded were c) and b), a), M. b) d) 4 160 120 -40 M) in Tris HC1 buffer(10mM Tris base, 0.2M NaCl, pH=7.4). pH=7.4). NaCl, 0.2M base, Tris HC1 buffer(10mM Tris in M) 140 120 160 100 80- 40 -00 -60 -40 -20 40 0 5 40 5 650 550 450 350 - 350 550 650 750 1.870 216 1.665 A 247

206 256 356 596366

Nanometers

Spectrum 9. UV-Vis aborption of HEP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). 1.567 1.576 222 269 A b s o r b a n c e

289 388 358 508

Spectrum 10. UV-Vis aborption of CEP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). 1.551 221

1.154 1.196 253 26?

i.eee 312

258 306 358 588

Spectrum 11. UV-Vis aborption of BEP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). VO 2.816 225 A b s

1.374 1.361 256 a&9

206 256 396 359 596

Spectrum 12. UV-Vis aborption of IEP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). vo 2.119 272

1.569 Cf 226

1.G9G 309

2B6 250 300 506359 Manometers

Spectrum 13. UV-Vis aborption of TBP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). 1.336 210 1.866 253 r i Q G HCCv^l) J I C ^ v N C ^ i

259 390 359

Nanometers

Spectrum 14. UV-Vis aborption of BCP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). 2.596 223

1.818 268

200 250 300 350 400 450 Manometers

Spectrum 15. UV-Vis aborption of BPP in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4). t

'.vf soi

•i, w jitx e* - = - 1 iiJ -j ^ V . ‘tii jliu ■ 11 i : ri;M|*iFf' . 1 Mil n ;*] i B l I l I1 ! I, »|j?“’*1 J! • W' ! \V i if V' (j * ,4’

rk I i JP iif * •*! 'ij :v *’ H IH Hii i!• f L I # J ’ • f I.!’ i\ !! ;sc '59 Spectrum 16. Tripelt-triplet absorption produced by LFP (308nm) of HEP (above) and

CEP (below) in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4) with A308=l-0. 200

o* Source: Raw Curne 21 Cur so- : 22* , 370.3. > tu*d^ : 0. 3?75*2 ar 400 250 -j Vi

200 J

‘ Z’C O i

i : 1 m m S r* • r, r. Ji.- * .i I • | P 'V f f iV1'?-; If:I f

W ■ , jV I}1. fnUr ■ 1 , . . V v " I'-V■•' tf - ; 10;

250 350 450 650 750

Luri:-*: ;;c : ; .A, f icittr i tuds : £.Ciif174

i' L'Q 11!

J> J

250 350 550 750

Spectrum 17. Tripelt-triplet absorption produced by LFP (308nm) of BEP (above) and

IEP (below) in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4) with A308=l-0. Spectrum 18. Tripelt-triplet absorption produced by LFP (308nm) of TBP (above) and

AMT (below) in Tris HC1 buffer (lOmM Tris base, 0.2M NaCl, pH=7.4) with A 308 =1.0. 1.5- 1.307 207 1.164 A 253 b s o r b

e.e 296 300 359 506 Manometers

Spectrum 19. UV-Vis aborption of HPoTM E in ethanol. 1.223 242

1.0 -

0.517 223 0.463 289 0.5

0 .0 —j—i 200 250 300 350 500

Nanonetcrs

Spectrum 20. UV-Vis aborption of BPoTM E in 1,4-dioxane. APPEDIX C

Selected IR, >HNMR, UCNMR, and Mass Spectra

204 f + u I M s T - i /»<1 lfi * ' 11 11 . /• ■* * ? 4 i i *•. . S 1 l i t / •*t • 2 9 0 .1 3 3 I * Sf 11 :.;-4 : . i :•«. 7 ST 8 3 .0 I s 11 *-•>*. 4 . 1 0 * 7 1 . m h o *. 01 9 2 8 4 .0 0 0 S! 3 2 7 6 8 TO 3 2 7 6 8 S* 3 4 9 6 .9 0 3 M 2/PT .2 1 3 « 4 .7 RO 0.0 AO 4 .6 8 6 RG 40 NS 66 n 3 0 3 F » 4400 02 90000.000 OP 3 0 1 PO

LB 0.0 68 0.0 CX 3 4 .0 0 CT 1 3 .0 0 f 1 1 1 .0 0 1 P F? - .5 0 0 ® H 7/C * 8 4 .6 0 4 Il'M/CM511 41 .3 3 8

Spectrum 21. !HNMR (DMSO-d6, ppm) of 3-acetamido-6-aminoacridine. 206 1 .1 1 4 ci*» 0 .0 3 00 9 .4 0 .0 3 4 .0 0 6 9 .8 9 6 i n nc»i»* 303 !ti'H4.000 7139 9 0 0 0 .0 0 0 1M400 14 7 0 3.689 ?mi ?mi 39766 39766 9 3 .0 tf m IF sn -3669.01 U.» m* LB GBCX r.t i i 0I .• .0 11//CM 3J3 7J4 PPM/CM 3 .1 4 7 10 f f PN on AO MGMS 400 S M //PT .6 9 0 S! Sf 01 SF OATF. 1 9 -3 -9 3 ?:.c ?:.c 6:

e o .o ppm :e:.: :e:.: :*:.o c. i<:.o :;?.o ?c. :3o.o : :cc.o 90.0 w n a r Spectrum of3-acetamido-6-aminoacridine. ppm) 22.13CNMR (DMSO-d6, 4 0 .

' " f l illL|lllk 1 | Ltu4i]fcuy«li Lj PlflSS CflCC) 40 80 80 100 120 140 180 1S0 200 220 240 280

Spectrum 23. Mass spectrum of 3-acetamido-6-aminoacridine. 207 f0M«4l*4 0411 4-9-9J or K 4 .I M ST m : Ot 92M.94C 91 MM to 31*19 9* 34M .903 M 2/»1 .1 1 3 4 .7 no 9 .9 40 4.M C M 49 N9 :*» tc i : j r« 4<:i 01 t : : : 9 . 9 c ; o» X . »0 tIt i 1.09 .4 CX )« .« 4 I #1CT 9.97J90 .9 4 - p r m JNi.Sr 208 Spectrum 24. *HNMR (DMSO-d6, ppm) of Acr-I. TONGQUN DATE 3 * 8 - 9 3 TIME 18: 09

ST 9 0 . 3 2 4 9E 0 9 0 . 3 2 0 SfO? 200.130 S> 9 0 . 0 01 9399.269 S I 3 2 7 6 8 TO 3 2 7 6 8 H Z/PT . 6 7 2

Pw 2 .C RO O.C AO 1 .1 4 7 RG 3 2 0 NS 2 9 0 9 IE 2 9 7

0 2 3 4 0 0 . COO HZ/CM 244.192 PPM/CM 4.692 SR 3 6 0 4 .6 2

Spectrum 25.13CNMR (DMSO-d6, ppm) of Acr-I. 209 -Wwut'iUaidrth A' Rm

Spectrum 26. Mass spectrum of Acr-I. Spectrum 27. IR (KBr, cm'1) of 8-methoxy-5-nitropsoralen. Spectrum 28. JHNMR(CDCI 28. Spectrum

lNTE*a»'. 5 4. .0 1 S . i 0 . 2 5 . 2 0 / 3 5 . 3 .0 4 .5 4 0 . 5 5 . 5 0 . 6 5 . 6 0 . 7 5 . 7 0 . 8 5 . 8 0 . 9 5 . 9 : - £ 5 c. 4 ■ 7 1 r o s r u c 3 5 5 4 1 3 9 4 4 1 3437 8* 34 8* 65ft9 2 2 5 6 «46 4 r« 9 ? 7 7 73V?

71 ?: I FPEOUENCT 195 9 .1 4 2 0 3 ? 1 4 . 4 ? 3 | ?7f, f 7 .? 7 0 4 I 1 4 t» j, 34fc j, 1 t» 4 1 5 .9 4 7 5 1 1 7 5 5 .6 17 f*. 7 ? ; n ; ? f*.7 1 0 .5 9 * 4 1 17 .6 5 5 7 1 8 1 .9 5 6 7 1 571 7 .5 1 0 9 766 6 .7 5 1 3 l 9 e . 2 , ppm) of ,8-methoxy-5-nitropsoralen.ppm) of 9 4 0 5 . 4 6166 6 1 .6 6 t 0 7 I. 2ft 9 *\ HfcV. *\ *. *1476 0134 3 1 .0 u7?C:*. C ? 7 .u fctSu . n p p 3 3 2 9 . 11 , . . J » » . i , , . 'I ■ 1 ' I ' ' ' 1 ■' ■ ■ ■ I ' . . , I , \ i t . ^ T » » 1 J I . . . J , , -1-1 NC SlTY iNrCM : .*.: .*.: : 5 1 - . 9 13 r . : ;• * m . ? ? . 675 . ? 977 7 .9 7 ■*.1 V(. 40 4 * 4 • * t. PPM J J 3-92 2 9 - -3 7 1 « A 0 3 K N 3 8 9 0 .0 0 I I 0 .0 0 9 8 3 N K P/M 294 4 9 .2 9 9 .8 8 9 PPM/CM HZ/CM R AO 0 2 .2 00 PM T 0 2/P 8 0 H 8 7 .0 2 0 3 4 7 3 0 I 3 S .1 0 0 ' .0 2 01 0 0 3 0 2 .1 0 0 ST 2 P02 S SFO 3 2 2 2 TIME 000 0 .0 0 0 0 0 9 2 0 NS O 79 .7 0 4 3 2 SO 8 6 7 2 3 TO TE F 6 132 2 3 .1 0 0 2 97 2 109 32 SSS S .S 4 .0 0 2 . 3 nj 10*639.001 DATE 2 0 - 1 2 - 9 3 TIME It: t4

S F 3 0 .3 2 3 SFO 30.320 SF02 200.130 S T 9 0 .0 Ot 9399.269 5 2 3 2 7 6 6 TO 3 2 7 6 8 KZ/PT .8 7 2 r- " O 2.0 m 0.0 A9 t . 1 4 7 3 20

2 9 7

02 3400.000 H2/CN 291.603 PPM/O' S.000 S 3 8 0 4 .6 2

i « . . . I 160 130 140 130 120 110 100 90 60 70 60 SO 40 30 20 10 PPH

Spectrum 29. ^CNMR (DMSO-d6, ppm) of 8-methoxy-5-nitropsoralen. nj oj 100. 30.

80. r i •'/ - „L . 70. ' A , / o ( ' ‘ > 00 .

DU. 40. 180 30. 130 84

103 1 0 - ?F. 51 a / ILJ1 i u i J t «. 80 88 180 180 140 180 180 J O

Spectrum 30. Mass spectrum of 8-methoxy-5-nitropsoralen. 98.53

4 4 .1 0 4000 3500 3000 2500 2000

Spectrum 31. IR (KBr, cm'1) of 5-amino-8-methoxypsoralen OO 5 . 9 SO.O

Spectrum 32. JHNMR(DMSO-d6, 32. ppm) of Spectrum t»re«*ui 5 -amino- 0 . 5 PPM 8 -methoxypsoralen. 0.0 132 2 3 .1 0 0 2 f S 3 9 - 2 1 - 1 2 OATE ZP . 0 2 .2 0 6 HZ/PT 7 2 3 0 3 I S .1 0 0 2 .0 0 0 3 0 2 .1 0 SY 0 2 2 0 F S SFO ? 0 : 0 2 TIME TONCOUN 6 0 o n P* 0 0 .0 0 4 7 3 01 ZC 61. 4 0 .0 1 6 0 0 .0 0 HZ/CM 0 0 0 3 2 0 MS AO R 41 .4 0 9 2 3 9 0 .3 SR PPM/CM O 0 6 7 2 3 TO TE 7 9 2 160 160 737 0.0 2 . 3 359 9 5 .3 4

TONGOIAN OATE 4 - 1 0 - 9 3 TIME 2 1 : 0 6

SF 5 0 .3 2 4 SFO 50.320 SF02 200.130 ST 5 0 . 0 01 9359.265 S t 3 2 7 6 6 TO 3 2 7 6 8 H 2 /P T .6 7 2 noPH o.o 2 .0 AQ 1 .1 4 7 R6 3 2 0 MS 2 2 2 2 IE 2 9 7

02 3.00.000 HZ/CM 251.603 PPM/CM 5.000 SR 3 6 0 4 .6 2

M l

150.0 140.0 130.0 9 0 . 0 6 0 . 0 7 0 . 0 PPM

Spectrum 33.13CNMR (DMSO-d6, ppm) of 5 -amino- 8 -methoxypsoralen. CO

«^>/llfy »*;i|. .iJilL.t^-t■ *jil*. . 40 80 80 108 120 149 160 180 209

Spectrum 34. Mass spectrum of 5 -amino- 8 -methoxypsoralen. cm-: * 3128.« 28.04 1620.0 30.56 1409.1 b.82 12?'. .0 25.68 1118.C 9.70 9fjo. C 25.53 0 3 0 . 4 32.9# l

Spectrum 35. IR (KBr, cm'1) of 5-azido-8-methoxypsoralen. JU vo T0NG0IAN OATE 4 - 1 0 - 6 3 H MC 1 9 :4 6

Sf 200.132 SrO 200.130 ST02 200.130 ST 2 0 0 .0 01 3740.000 S I 3 2 7 6 8 TO 3 2 / 6 8 1 H Z /P I .2 2 0 o r 3.2 ru> 0.0 AO 4 .5 5 9 M 6 0 Nf> too II. ;**.»/ 0? 50000.000

II2/CN 61.707 PPM/CM .3 0 9 Sfl 2341.71

■\_____ .JL

9 .5 6.0 5 . 5 3 .5 2 . 5 1.0 PPM

Spectrum 36. lHNMR (CDC13, ppm) of 5 -azido- 8 -methoxypsoralen. TO 3 8 7 6 0 H Z/PT .0 7 8 pn a.o PO 0 .0 AO 1 .1 4 7 0 6 3 8 0 MS 1 9 0 1 TC 8 9 7

02 3400.000

HZ/CM 891.603 PPM/CM 6.000 SA 3 3 2 1 .9 0

'—r— "i ■ 1 I 1 -—I—" ■—i—■ ■"I — '“T— ■I 1 6 0 .0 1 5 0 .0 1 4 0 .0 1 3 0 .0 110.0 100.0 e o .o . 0 6 0 .0 5 0 . 070 4 0 .0 10.0

Spectrum 37 .13CNMR (CDC13, ppm) of 5-azido-8-methoxypsoralen. 100- 229 9 0 . 8 0 .

MflSS CflCO 40 68 80 100 120 140 160 180 200 220 240 260

Spectrum 38. Mass spectrum of 5-azido-8-methoxypsoralen. ►I, -I, »

■ < (JM50t< 1 *fl IH 4 .M I Y I ’I'M 1H 1LN 5 1 I f

1 3 0 0 7 4 7 3 .2 0 1 0 . 2 4 0 6 2 .6 4 2 2 4 0 7 2 4 6 3 .4 9 1 0 . 2 0 6 0 2 . 9 6 5 Lc-. ec 3 8 1 7 2 3 6 9 .7 4 1 7 . 0 0 2 3 4 .3 7 0 4 i * I 8 2 5 2 3 6 3 . 7 3 9 7 . 6 7 5 6 4 .6 4 0 a 1 7 BO 2 1 3 4 . 2 7 3 7 . 1 1 ) 1 4 . 3 3 9 f. 1 7 9 0 2 1 3 2 . 1 0 3 7 .1 0 4 1 4 .1 2 1 2 6 7 9 1 9 3 3 .6 0 1 6 . 4 4 2 7 3 . 5 6 2 M M f.M I b 2 0 6 9 1 9 2 4 .0 0 7 6 . 4 1 0 0 3 . 6 3 5 Oa TC l > M 1 4 7 6 1 '* 1 4 2 S .4 S 3 4 . 7 4 9 4 6 . 5 9 3 1<) ST 290.134 4 9 J 7 1 3 0 9 ./.».(> 4 . 6 2 9 5 1 3 4 .0 1 5 ST 9 3 .0 I 5 5 5 1 1?3». .*#37 4 . 1 2 1 3 7 0 . 51 *j 01 9400.000 6 2 '. 1 SI 337M lO f.7 .0 0 4 • 5 5 7 0 6 7 0 0 10 J2700 J 1 #.7l»7 1 0 6 7 .#*1 « 5 4 1 1 5 1*4 .■ Sa 2904.01? 1 4 6:»1 1 0 V . .* 71 i 51 70 'I / • . / . HI/PI .102 i '• (• I4(, 1047 .4# II « .4 9 0 0 6 .O#..* Pm 7 .0 6 .V , 0 1 0 4 4 . 1 4«> i .4 7 0 9 5 . 751 00 0 .0 1 6 3 *,? 1 04 1 .# 4 1 < 4 777 5 * 4 6 1 M f.? 6 5 1 0 4 7 .9 6 4 i .4 7 5 0 5 . 7 7 7 s r

3000 0 .0 I X 00 I t 0.0 n »S:S< R !!:&• K2/CJ* T j i l o * ppn/c* .»■* so 4 1 9 4 .?;

i s s

I. \. . I .. V \> \ 3.C 2.a 2.0 1.9 l.o .9

Spectrum 39. !HNMR (DMSO-d6, ppm) of 8-(3-bromopropyloxy)psoralen. 8 95.86

o

m

ou

0.41 4000 3500 3000 2500 2000 1600 1000 224 Spectrum 40. IR (KBr, cm'1) of 8-(3-bromopropyloxy)-5-nitropsoralen. iNTttRM. Spectrum 41. *HNMR (CDCI *HNMR 41. Spectrum C\c* tJQi Oj u 3 , ppm) of 8-(3-bromopropyloxy)-5-nitropsoralen. of ,ppm) - ■* t 5 4 3 0 V :l i - *• I 1 4 • 1 2.5 2.0 1. 0 . 0 . 0 .5 .0 1 .5 1 0 . 2 5 . 2 0 1 14574 1112 10114 : : 1 11 t t : : 10140 7 B 0 0 1 L’.’F’ *• : . ••: ;* os.*:. H4JJ7 tn r r.-. /H04/ »‘M/ *.H/. ; 1 t '

. */. ft ;■? mi

/ -i 1 c.

* * 1 • 1*25. o:.o > 4 * » : * .H 4 1 1 ' * 5 1 / . . 3 . 7 . : 1 * » . * ! l . 0 4 4 1 4:m. ». v-n 7 4 9 . 6 2 3 2 3 2 . 9 7 9 497.tVH 503.074 4/9.0J64M5.4*M.9?r i t 4 744.475 3 7 9 9a?. 756.9bU 750.709 "• 11 n in's 1 025 2 .0 .?r: .*./. * 3 0 : / 1

1

1 '■ ' I'" 4241? 1 4 2 .4 2 2 6337 3 3 .6 1 2 * S Z 0 4 2 6 . 6 ? 3 1 3 ( 2 7824 2 8 .7 3 4.3644 4.8*30 2.45:.l 3 9 6 . 8 1 0 5 7 6 . 6 7311 1 3 .7 3 4 0 / 2 . 7 9 1 2 9 . 4 7 7 3 4 . 7 • I. 11 7 6 2 4 . 7 7.09 t,l 7.0050 008112.910 1 9 . 2 1 1 8 0 .0 .4037 . 17-M . II • 9:0 39 fhi*y 199 9 1 t.'t 1

1

l c i l l u l ...... 3 9 6 . 4 1 0 9 0 . 3 1 9 8 1 . 1 1 1599 9 5 .1 0 1 .337 3 3 . 1 1 8 4 . 2 1 6 7 6 . 1 1 3 1 7 . 2 1 6 / 2 . 8 1 3 3 9 . 6 2 7 8 . 3 1 9 6 . 8 9 4 0 . 2 4 0 3 . 6 IJ.U94 3 2 8 . 7 3 8 3 . 8 3 9 3 . 2 3 6 0 . 8 v././,4 11 I y ’1 AE 3 9 - 0 1 - 9 OATE ZP . 0 2 .2 0 0 6 HZ/PT 6 .0 7 2 0 3 4 7 3 0 I 3 S .1 0 0 2 .0 01 0 0 2 SY 2 0 F S 200.130sro f S S100 MS R6 AO RO Pit TOMGQIAM ZC 61. 4 0 .8 1 6 0 0 .0 0 0 HZ/CM 0 0 3 2 0 TE 8 6 7 2 3 TO I I 9: TIME A 83 .8 8 3 3 2 9 0 .3 SA PPM/CM 13? 3 .1 0 0 2 BQUKTR 7 9 2 80 2 . 3 0 . 0 555 5 .5 4 TONCOIAN OATE 9 - 1 0 - 9 3 TIME 9: 54

SF 5 0 . 3 2 3 SFO 5 0 . 3 2 0 SF02 200.130 SY 5 0 . 0 01 9359..265 S t 3 2 7 6 8 TO 3 2 7 6 8 rJOl H 2/PT .8 7 2

PW 2 ..0 n o 0 ..0 AO 1..1 4 7 n s 3 2 0 NS 2 0 0 1 TE 2 9 7

02 3400..000

j f , , , | i . . - j -,, . . ( . . . . t i t ' ■ « ■ | ' «■ , - . - 1 • . , - . • t . i 1 . , . . , . , 160.0 150.0 140.0 130.0 120.0 110.0 100.0 90.0 60.0 70.0 60.0 50.0 40.0 30.0 20.0 10.0 PPM

Spectrum 4 2 .13CNMR (CDC13, ppm) of 8-(3-bromopropyloxy)-5-nitropsora!en. 247 ™i

7 8 . 378

3SS 3S8 369 378 371

87

MASS (AGO

Spectrum 43. Mass spectrum of 8-(3-bromopropyloxy)-5-nitropsoralen. 227 4000 3500 3000 2500 2000 1500

Spectrum 44. IR (KBr, cm'1) of 5-amino-8-(3-bromopropyloxy)psoralen. 228 NJECS.OOt • CUftSOR rftCOUCNCV Pffl 1NTENS1 ! V DATETl* IkM 2S -S -S3 1 6 0 9 3 1673.702 6 .3 6 3 0 1.314 a r t M . m 2 6139 1663.901 U .3140 1.31.1 art 201.D0 1 .046 fft} MO. 1)0 3 6333 1377.4 36 7.8020 ST 2 0 0 .• 4 6542 15 7 3 .7 6 3 7 .071 1 1 .0 7 2 01 4300.000 5 7158 1444.737 7 .2 1 0 9 1 .017 SI 327M TO 227M 6 7147 1442.601 7.2082 1 .743 K2/PT .220 7 7737 1313.893 6.3611 1 .834 W 2 .2 8 8104 1732.601 6 .1 3 0 9 1.621 «D 0.0 9 8148 1 2 7 7 .0 1 8 6 .1 1 8 0 1 .371 AO 4 .0 3 9 18 VCIM I140.39J 4 .7 3 9 2 1 . 124 M 04 MS 100 1 1 vi im 1147.AM 4 .709!. '.•.394 TC 20? ir 99811 1)36.31 4 4 . i/VO 1.211 02 90000.000 13 1024? 763.780 3.8133 1 .003 14 1077? 756. 394 :./uo3 2 .4 1 4 K2/CM 0 1 .0 0 4 10307 730.037 3.74/7 1 . 1 MN/CM .2 0 0 If* 8A u o o . r o 16 10651 6 7 3 .7 0 8 3 .3 6 4 2 1 0 .0 6 9 17 11473 503.707 2.3173 1.173 111 1 1431 307.00!) 2.MMM i IV 1 1437 MW. 307 7 . 4999 1 .1 4 / 7« 1 1/.70 447.934 7.7M«? 1 .0 2 9 71 I 1 /IV. 44 1 .6 9 6 7 .2 0 / 0 1 .MM 1 1 /.M 4 V*. 4VO 1 //.o . •• / / ;\\ I3/J1I .O.V ’.0001 1. //»

J v J J \

Spectrum 45. ^HNMR (DMSO-d6, ppm) of 5-amino-8-(3-bromopropyloxy)psoralen. TOMGOIAN GATE 1 6 - 1 2 - 9 3 TIME 1 7 :4 9

SP 5 0 .3 2 4 SPO 50.320 SP02 200.130 SY 5 0 .0 01 9359.265 S I 3 2 7 6 8 TO 3 2 7 6 8 H Z/PT .8 7 2

PM 2.0 RO 0.0 AO 1 .1 4 7 n o 3 2 0 NS 300 1 TE 2 97

02 3400.000 HZ/C* 291.603 PPM/CM 5.000 s n 3 6 0 4 .6 2

90 90 l*PM

Spectrum 46 .13CNMR (DMSO-d6, ppm) of 5-amino-8-(3-bromopropyloxy)psoralen. 0 3 2 100—

j mss

3 0 .

£57 50 100 150 350

Spectrum 47. Mass spectrum of 5-amino-8-(3-bromopropyloxy)psoralen. 96.95- %T

00 o

m 743.6

CM

o

v in V o

I I

17.70 4000 3500 3000 2500 2000 1500 1000

Spectrum 48. IR (KBr, cm'1) of 5-amino-8-(3-diethylaminopropyloxy)psoralen. 1111 f * i< | I V TONGQIAN OATE 1 9 -1 2 * 9 3 TIME 2 1 : 2 3

Sf 200.132 SfO 200.130 SF02 200.130 ST 2 0 0 .0 OS 3 7 4 0 .0 0 0 S I 3 2 7 6 6 TO 3 2 7 6 6 H Z /P T .2 2 0

PM 3 . 2 AO 0 . 0 AO 4 .9 9 9 A6 100 NS 4 3 TE 2 9 7

02 90000.000

HZ/CM 61.604 PPM/CM .309 SO 2 3 4 0 .7 9

IK f

9 . 3 9 . 0 6 . 3 6 . 3 3 . 0 3 . 3 0.0 PPM

Spectrum 49. !HNMR (CDCI3, ppm) of 5 -amino- 8 -(3-diethylaminopropyloxy)psoralen. TOMGOlAN OATE 1 5 * 1 2 * 9 3 \( TIME 2 2 : 14 SF 50.323 SFO 50.320 SF02 200.130 ST 5 0 .0 01 9359.265 S I 3 2 7 6 6 WH. TO 3 2 7 6 6 HZ/PT .9 7 2

PN 2 . 0 n o o . o ^ r . j , > AO 1 .1 4 7 RG 3 2 0 NS 2 5 0 1 TE 2 9 7

02 3400.000

HZ/CM 251.603 PPM/CK 5.000 SR 3 3 1 6 .7 5

«MI

"~T,T' ’■’I ' 1, - r - I - r - * I ' i i ' I '« ’"T "' r - T - r r- T -r ,—r" r '~rr 160 130 110 90 60 70 60 40 30 20 PPM

Spectrum 50.13CNMR (CDC13, ppm) of 5-amino-8-(3-diethylaminopropyloxy)psoralen. 180. 90. 80. 70. 60.

oilpft 40. 114 38. 28. 10. 5*

o , Sf * ti i 338 182 1v. v . cb/ 8 . i^ki JUtl1 JlJk! . UtlmalLa** X J ~ 5i? 188 158 288 258

Spectrum 51. Mass spectrum of 5 -amino- 8 -(3 -diethylaminopropyloxy)psoralen. Cffi-1 cm-1 % Cfl)“l X 30.41 cm*l % 28.31 2965 .3 18.23 2794.3 24.27 2235.4 44 .40 2111 .4 3 .95 1735.7 2.84 1623.8 29.11 1 5 8 8 .9 12.51 1540 .1 40 .20 1466.7 10.6S 1430.4 25 .09 1379.3 24 .32 1357 .C 47.92 16.10 1318.7 20.35 1272.C 34 .24 1190.1 28.57 1164.3 5.99 1116.5 13.56 .2 33.14 1051.8 26.63 1001.5 ion 23.49 978.1 37.06 919.1 50.27 899.9 4 8 . 5 5 874 . 38 .45 839.9 24.73 798 .4 42.84 748.1 21.90 51 .15 698.1 49.06 665.1 56.1C

' -V ./I <*

— I— 4000 3500 3000 2500 2000 1500 1000 cmr\

Spectrum 52. IR (KBr, cm'1) of ADP. jo

O N ■ Spectrum 53.1HNMR (CDC13, 53.1HNMR ppm)ADP. of Spectrum

iNTCtAAL i z 7 ; * v. .v . t i -i.v.v . v . i - : : 7 : r ; i \:.vc*< * c v . : \ ; i . u t : K 7 i v r \:> 11 1C M ’ i - '

r r V ’■ 9 r »■ **• . r-- ic-„: c i . 2 iri-s-r : ° : r i * : t - j ) r : .‘ i v i .‘ : V57 5 IV 1 r ■ - r r i i i ; 4 4 4 . 7 4 5 ; ;•> i 1 : i *:* :* * iiv 2 /».*• : j HV.-.v. 2117? 049; 9 4 0 c e s ^ e 7 f 397-4 j 9 7 9 'V i.O . / 2 4 1 i 6 » y VO lV H •*.)« . - M ' * 1 *.11 . ♦*1.01*1 0 4 9 . 7 5 2 1 7B.3 4 7 t 2 . 6 J4 3 . 1 ?7*B 126 * M .H -.-V l4 017- 7 1 .0 7 6 2 1 I * I I •• • v •'w i • / . - . s i • M'1 1 .1 ' M : o . u m , 0 9 4 «U'H. /.*. .w * w . : . : i - t .'2 .6 5 1 5 * . 6 0 5 1 11..7 .r:.**' * * . : r . : *.*0 7» .». 1 1 r V I I / . V H 4 4 6 1 4 6 14**. . 1* I . / 4 9 * i*5. 5 * ii . / 9 1 11 *r. * * ' 4 . 4 .‘ r * 1117 . * » v . / * • • . 9 *.2 u.rr . 'u 4 / t I . B m . . . * *• * . i : 2 5 :>t : * " : M M • • •• * ; i*. 4 . 4 c : t . - : . > . i : * * - : * . * I 6 95.7 5 .9 2 t ‘ 8 2 . 6 »• . O * .; / ; . . . . •* 7 .1 7*--.‘ .1 7 : 4 * 1 . / : . '•< • .'*•!< . / : *.«.«• . *..’ . . i* *. 1* -•.‘•.tl t . • ‘ . • - 1 . *i ’ 1 iNM i i i M i.N . 1 ! • . . .o*-.-. . Tee. 7 . . ‘•1 .1*'. »\ \V .» 4**. 4**. r r 1 ?v • ?v 1 i i fl1' l f , / I . *7v;- v 7 * ••r »■ _ z 6 9 6 . 0 - • -.*.•• , K I •i / i ;• . n t v. .v t . V ... U '.-./.M ? 1 7 . 7 I -/ - .IU V -J.iO J .- / > 'O l '.O ■>. H2 1H . 7 5 0 4 . 7 7 9 0 . 2 -11/ * / 1 .-.1 1 1 • . i 1 1 jr. • . r j • -...*.-it;

/ ll* /- .

26 2 4 . .I**--• t r. 4 . . i i . . . .21 .6 •! 0 9 i x 4 n i . / i .. :iv ’>t

‘.A •••/ *• * r /> rUc. c•',) l t t . n / u i 1Ml II l M 11 1 M 1 u* It u* 1 M 1 1. 4 f I H j1 1. l A \ \ j 1 t s t . I I I n u m n i i i n i n n t t It: 1 111 M 1/ Ml .M U *11/ M It K J I I / t i l t i 1 l . l l t II l M i m / 11. M . I I I t u i u II

V

.1*1 .III // .1 .1 .t •IS r SI IS .IS 1 .IS 111 237 eortwn JON0OIAM DATE 1 5 -1 2 * 9 3 11 mi-: 1 * * 1

SF 5 0 .3 2 3 SF0 50.320 sro? 200.130 ST 5 0 . 0 01 9359.265 S I 3 2 7 6 0 TO 3 2 7 6 0 IIZ/P1 .072

PW 2 . 0 n o 0 . 0 AO 1 .1 4 7 (10 3 2 0 V"x>i MS 1 3 0 7 TE 2 9 7 0 2 3 4 0 0 .0 0 0

H2/CM 251.603 PPM/CM 5.000 SA 3 3 1 6 .7 5

160 160 140 130 120 110 100 90

Spectrum 54 .13CNMR (CDC13, ppm) of ADP. 00 i«*

/ 0

mss

10

II 'I

Spectrum 55. Mass spectrum of ADP. OATE 2 1 - 4 - 9 4 SF 2 0 0 .1 3 2 ST 2 0 0 .0 01 3 7 4 0 .0 0 0 S! 3 2 7 6 6 TO 3 2 7 6 6 SM 3 9 9 7 .1 2 2 H 2 /P T .2 2 0

PM 3 . 2 n o 0 . 0 AO 4 .S S S PS 6 0 MS 10 2 TE 2 9 7 FM 4 9 0 0 0 2 9 0 0 0 0 .0 0 0 OP 30L PO

LB 0 . 0 6 6 0 . 0 CX 3 4 .0 0 CT 9 . 0 0 FI 1 0 .0 0 9 P F2 - . 4 6 9 P HZ/CN 61.791 PPM/CM .301 30 2 3 3 9 .2 7

._____

2*0 1.0

Spectrum 56. *HNMR (CDC13, ppm) of 5-iodo-8-methoxypsoralen. O / 1 |V / ifr.'M .V ? ? /1 1 1 w » ;.’ . u v i i 6 ; n . u s : v»» 8 . 4 7 * ::-» 7 3 .5 4 4 ex j m v . : 1 7 1 4 . V .* i m ; . . i .n i : j 4 - SF 200.132 i: '*». SY 2 0 0 .0 01 3740.000 . . i / ' S I 3 2 7 6 0 TO 3 2 7 6 0 SM 3397.122 H 2 /P T .2 2 0

PM 3 . 2 RO 0 .0 AQ 4 . 5 3 5 R6 00 MS 1 03 TE 2 9 7

FM 4 3 0 0 02 30000.000 OP 3 0L PO

LB o . o 6 0 o . o CX 3 4 .0 0 CY 12.00 FI 1 0 .0 0 7 P F 2 - . 4 9 3 P HI/CM 6 1 .6 0 4 PPM/CM .3 0 9 sn 2 3 3 9 .4 0

B.O b.O b.O 4 . 0 ? . 0 0.0

Spectrum 57. fHNMR (CDCI3, ppm) of 8-(3-bromopropyloxy)-5-chloropsoralen. \(

cl

T0N6QXAM OATE 3-4-94 SF 90.323 SY SO.O 01 9399.269 SI 32760 to 32760 5 * 14209.714 M2/PI .872 Pit 2 .0 no 0 .0 AO t.1 4 7 R6 400 NS 1800 TE 297 Fit 17900 02 3400.000 OP 17H CPO LB 0 .0 6B 0 .0 CX 34.00 CY 12.00 F l 170.009P F2 .018P HZ/C* 231.603 PPM/CM 3.000 Sfl 3322.89

90 GO 70 GO SO 40 30 20 PPM

Spectrum 58.13CNMR (CDC13, ppm) of 8-(3-bromopropyloxy)-5-chloropsoralen. 242 96 .

i*Wu 100 S O 300

Spectrum 59. Mass spectrum of 8-(3-bromopropyloxy)-5-chloropsoralen. * •- . . : r • , . »*« . ** t .; \ 11- '•r»t T • • T ! : *<'■ . • . .. 4 ( S i . ' I Of. 7 *»;• V . 4. u . *..,.v M 9 i t . . i-ii e r r ,* ? . ’ 1 r> C,7~ * ! * . . 1 7 OATE 9 - 4 - 9 4 *> * *-v c. : : . ; 4 c _ .» - lk! ' i*Z A !■. - t SF 200.133 1 1 SY 2 0 0 . 0 1 T ; . * ! ;■* 1 i ; - . . ' ? 01 4SQ0.040 •‘•s . ’ S I 3 2 7 6 S i ■■ ! ' : TO 3 2 7 6 0 SM 3 3 0 7 .1 2 2 1* ■ * H 2 /P T .2 2 0 4 . •. PM 3 . 2 no 0 . 0 A0 4 .3 3 3 no 6 4 NS 1 1 6 TE 2 9 7

FM 4 3 0 0 02 30000.000 OP 30C PO

10 0 . 0 06 0.0 cx 3 4 .0 0 Ct 1 6 .0 0 FI 1 0 .0 1 9 P F2 -.4«0P H2/CM 6 1 .9 0 4 PPM/CM .3 0 9 SO 3 2 6 6 .9 6

A jJWk r

9.0 B.O 7.0 6.0 5.0 4.0 PPM

Spectrum 60. *HNMR (DMSO-d6, ppm) of CEP. T0N6QIAM OATE 31-3-94

4> SF 200.132 SY 2 0 0 .0 01 4300.000 S I 3 2 7 6 6 Oil TO 3 2 7 6 6 SM 3 3 9 7 .1 2 2 y-y\ H 2 /P T .2 2 0 P n 3 . 2 RO 0 . 0 AO 4 .3 S S RG 40 NS 190 TE 2 9 7 F tt 4 3 0 0 02 30000.000 OP 3 0 1 PO

LB 0 . 0 GB 0 . 0 CX 3 4 .0 0 CY 1 3 .0 0 FI 1 O .0 O 1 P F 2 - . 4 9 9 P HZ/C* 6 1 . 6 0 4 PPM/C* .3 0 9 SR 3 2 6 6 .3 4

9 . 0 e.o 7 .0 0 . 0 I».0 3 .0 2.0 1.0 0 .0 245 Spectrum 61. *HNMR (DMSO-d6, ppm) of BCP. tc J19S C)3 0B NHH AM-250-?

0 Wv [ACC v^x),

1 CO i

80- &(

512 I

60-

40-

394

2 0 - 118 188 I 136 174 !202\20‘ 230 281 314 378 410 432 482 528 548 590 628 663 681 I ' I I r~ ' f- I I I r JH, r. 1 .[L, i i ■ , i - ' r 1 I t- t-4- '"I"' 100 200 300 400 500 600 700

Spectrum 63. Mass spectrum of BCP. v]tf* BaClKER FFE'Jl/F.N^Y

"V. DATE 4-5-94 SF 200.132 SY 200.0 01 4500.000 £ c > SI 32768 TD 32768 SW 3597.122 HZ/PT .220 PW 3.2 RD 0.0 AQ 4.555 RG 40 NS 127 TE 297 FW 4500 02 50000.000 DP 30L P0 LB 0.0 GB 0.0 CX 20.00 CY 15.00 FI 10.024P F2 - . 476P HL,/CM 105.066 PPM/CM .525 SR 3285.22

3.0 2.0 1.0 0 . 0

Spectrum 64. !HNMR (DMSO-d6, ppm) of BPP. 248 r vl Y

T0NG0IAN o< OATE: 7-5-94 SF 50.323 V SY 50.0 01 9359.265 SI 32768 0 (C'U)t Vji-CO, fi

" i ■ ■ " T ■ "T" ■ ’ I ■ ' I 160 140 120 100 60 PPM Spectrum 65.13CNMR (DMSO-d6-, ppm) of BPP. Spectrum 6 6 . Mass spectrum of BPP. VLM.MI OATf 4 * | H l t l « « I t 44

a r i m . i u » l 1 9 0 .1 )8 s r « I M .I M 8T U .O SI *38— 10 STM _ M2/01 .Ml M m Sl«7* IXII MM IV 100 87 ic 103

\ . iiv » 03 0 .0 ‘/.• • M i H2/OI 73 .8 8 4 !*.•»*«» Ont/O' .384 . 1 • ' . -I . K P 4 * SH 3883.37 • 1* . - 1 1 .*•. /.’.♦V i • • . i f I *.

Spectrum 67. JHNMR (CDC13, ppm) of 4'-chloromethyl-4,5',8-trimethylpsoralen. DATt 2 -4 -9 3 n o t t r . ^ I f >00.141 SfO >00.110 If 02 200.110 IT 2 1 0 .0 01 0000.000 • I >2700 TQ 227M HZ/PI .230 > f 4 .0 ft) 0 .0 AO 4 .1 0 4

0? 00000.000 MZ/C* 92.002 P9W/CM .3 0 0 SO 4 7 0 6 .2 4

Spectrum 68. !HNMR (DMSO-d6, ppm) of 4'-(triethylammonio)methyl,4,5',8-trimethylpsoralen. TOCHCN DATE 2 * 4 * 9 3 TIK E 2 1 : 5 7

SF 5 0 .3 2 3 SFO 50.320 SF02 200.130 ST 5 0 . 0 01 9359.265 S I 3 2 7 6 0 TO 3 2 7 6 6 H 2 /P T .0 7 2 PM 2 .0 no 0 .0 AO 1 .1 4 7 A6 3 2 0 MS 9 6 0 0 TE 2 9 7 02 3400.000

H2/CM 254.553 PPM/CM 5.050 S n 3 5 7 9 .4 0

— i— —T— 130 90 60 20 10 PPM

Spectrum 69.13CNMR (DMSO-d6, ppm) of 4'-(triethylammonio)methyl,4,5',8-trimethylpsoralen. TCXCM.SMX n mm PILCF2.00I joco m . m OATE 1-4 -1 3 T t * lie 31 SOLVENT Ortl»rf« » R .B M SFO 12.900 s re * n o . m V 2 92.191 SFl 8 9 .1 3 0 f t 8 . 9 01 1449.999 S t 4008 sta 4 0 9 1 S t l 8 9 TO 4090 S» 10939.299 S>? 10939.299 SV| 949.949 MOO 3 HZ/NT 2 .9 0 ? vo o .o 0 .0 no 0 .0 AO

303 0€ 0 1 .3 ON 13 Oa 4 7 333M 02 3200.000 OP 20H 00 aOV2 0 ■Oat o U 0.0 M 0 .0 SS02 3 SSSI 2 ■C2 MC 0 ex 2 3 .0 0 c t 10.00 51 '3:85

f t - 2 .9 7 V «*> e o L im f t 0 .4 3 7 0 72 . 0 7 8 a t 0 .0 DC 9.000 M2/CN 423.324 PPM/CM 0 .7 0 2 M *37ll.OO M *3212.000 » t *1100.304 01 SI PI 00 .0000030 P4 03 .0037700 — 1— T" “T— P3 3 .0 0 160 130 29 04

Spectrum 70. 2-Dimentional NMR (DMSO-d6, ppm) of 4'-(triethylammonio)methyl,4,5',8-trimethylpsoralen. 2 4 1 1 0 0 ~i

80-

60- 276

2 0 - 213 212 368 43 57 106 121 251 317 369 I I128 3J1 169 183 222 286 301 348 391 |l W i . y L r . 1 l_ I 50 100 150 200 300 350 400250

Spectrum 71. Mass spectrum of 4'-(triethylamrnonio)methyl,4,5\8-trimethylpsoralen. Spectrum 72. *HNMR (CDC13, ppm) of 4',5,5'-tribromo-4’,5’-dihydro-8-methoxypsoralen. DATE 8 - 0 * 4 2 TIME 1 2 :2 8

t f 5 0 .3 2 3 SFO 50.320 SF02 200.130 SY 5 0 . 0 01 0359.285 SX 3 2 7 6 0 TO 3 2 7 6 0 H 2 /P T .0 7 2 Ptf 2.0 RO 0 .0 AO 1 .1 4 7 RG 3 2 0 MS 2 0 2 0 IE 2 9 7 0? 3 4 0 0 .0 0 0 HZ/CK 244.218 PPM/CM 4.053 SR 3 3 2 1 .9 0

w *w *0* fm

1 1 ' , - - - T,-.-" . 1 . J . . , . . . - , - - T l»J- , - - - - ...... | ■ . . . j—. /. ,1 . . . ( .. . . . r r t 1 , 1. I ry ■. . | i . . . ( . j t i r , f . 160 ISO 140 130 120 110 100 90 60 70 60 SO 40 30 20 10 PPM

Spectrum 73.13CNMR (CDC13, ppm) of ^.S'-tribromo-^S'-dihydro-S-methoxypsoralen. vj 338

90 452.9741 56. 453.7E8 315 50 70j

80 58 40 r n

3110 451 -53 -:■!

10

0 j , 386 358 •:m 400 450

Spectrum 74. Mass spectrum of 4',5,5,-tribromo-4',5,-dihydro-8-methoxypsoralen. Spectrum 75. JHNMR(DMSO-d6, 75. ppm)of 4',5,5'-tribromo-8-hydroxypsoralen. Spectrum

1 iMfCGAAl °n ‘ C J S ZP . 0 2 .2 PO 0 0 HZ/PT .0 0 0 3 4 t 0 O 3 .1 0 0 2 SFO 2 9 - 0 1 * 9 DATE f 3290. 3 .6 0 9 2 3 3 3 3 . 7 Sft 3 .6 0 7 I’PN/CM HZ/CM SOOOO.OOO 2 0 NS PG AQ PM 8 6 7 2 3 0 3 I S .1 0 0 2 .0 0 0 2 ST 2 0 F S S TE 8 6 7 2 3 TO T0NY3106 IE 3 0 : 3 1 TIME F 133 3 3 .1 0 0 2 7 9 2 174 20 0.0 2 . 3 SS5 S .S 4 100

80 141.7516

V< t t

*1 *

r . -iA-X'&isi. i *“ .* »*. .*

Spectrum 76. Mass spectrum of 4\5,5'-tribromo-8-hydroxypsoralen. Spectrum 77.1HNMR (DMSO-d6, 77.1HNMR of 4',5,5’-tribromo-8-(3-bromopropyloxy)psoralen.ppm) Spectrum

IMUWUI. ' - I ae D ? a e OG*nC**CHfc.8 OG*nC**CHfc.8 Be p r 17*4. *»5» . 2 . f 4 • 4.-9o 9 .;- '4 C r i s e :-t. rr*. :-t. e s : s : 1 c**c . " J I J " . 1c**c rc-uCNC'- f . at r F 200. 0 3 .1 0 0 2 SFO • 3 9 - 1 - 1 1 OATE F 139 9 3 .1 0 0 2 SF / 230 3 .2 no PH T 0 9 0 I/P 8 H 7 .0 2 0 3 4 7 3 0 3 I S .1 0 0 . 0 0 2 0 01 2 t S 2 0 F S TOMSQIAM NS AG AQ 9 :0 4 1 T2HE P/M 309 9 0 .3 4 0 .8 1 8 0 PPM/CM 0 .0 0 0 HZ/CM 0 0 9 2 0 8 6 7 2 3 TO O 79 .7 0 4 3 2 SO TE 7 9 2 100 100 191 0.0 0.0 3.2 998 8 9 .9 4 5 S 9 .7 J 1 4

350 100-1

S 60

L-

9 * 4

516

2 0 - 121

177 2. 1 3 2.1'S 203 ! 2 07 520 560 599 1--^ ^ 't 100 300 500 600

Spectrum 78. Mass spectrum of 4',5,5'-tribromo-8-(3-bromopropyloxy)psoralen. 262 10NGQUN i j f 4. DATE i l - 2 ? 9 3 TIME 1 7 :2 $

SP 200.132 SFO 200.130 L I •>. c •<. )_ rir1- SF02 200.130 SY 2 0 0 .0 31 4300.000 S t 3 2 7 6 0 TO 3 2 7 6 0 i2/*r .220 3* 3 . 2 uASOft rRcouCN Cr f’Pri 1NTCNS1IV 30 0.0 AO 4 .9 9 3 5 2 ? 4 1 6 1 0 .3 2 9 9 .2 4 6 ? I . 989 00 . 40 ? 2 P - •:s too i f - 5 .2 9 6 8 .2 3 4 0 2.32C- 2 9 7 53 1 9 l e - 0 . 3 3 9 3 .1 9 6 3 2 . 127 5 3 2 9 1 6 5 0 .2 6 9 0 .1 3 5 9 2 .1 4 9 38 90000.000 6714 1 2 2 4 .2 0 6 6 .6 6 6 6 2 .4 4 6 67 5 9 1 2 2 4 .2 9 1 6 .6 1 7 0 2 .2 6 7 H2/C0 01.004 970 4 5 * 6 .9 ? e s s H ' c n .3 0 9 4 .4 6 1 7 2 .0 3 6 * 3 3 2 9 0 .3 3 *?:? 5 * 2 .2 0 3 4 .45S f- 3 .5 9 9 575 0 5 5 7 .2 4 1 4 .4 3 3 3 2.wo2 9635 6*4.387 3.4696 i .990 96?:- 4 * 7 .3 9 5 3 .4 0 9 7 2. 742 ®722 e“2.922 3.3669 4 .2 3 4 37* 7 dsi.17) 3 .3 2 0 7 ?2. :-5t :!*.!*(. 3 . - 9 ; : i : . ::*i- 9?:*: t ' 2 . 4 9 2 3 .2 6 0 3 :-.* 4 o :? 5 2 l 4**.2® 5 2 .4 0 « 5 s. :■*: ::>c2± 4 5 :.4 4 9 2 .1 5 5 5 ; . * j £ 4 ^ . 4 4 1 2 .1 3 5 2 ; . * i 5 : : * : . 6 4 3 : . 5s24 *. : 7 S I I z "Z :- .r . s r r : . Z l ' l . : . ^ i : : 4 : . : * 2 5 r . : . a

j : ? : r 5 r’: * ! :!c e 5 c7: ! f

Spectrum 79. 'HNMR (DMSO-d6, ppm) of TBP. 263 100-1

4 92 4 92

80-

538

4 94 60- 490

990

40- ‘U,0J S0C 519 ..li i~ 500 550 114

580 2 0 - 136 154 160 214 282 328 356 412 598 637 687 761 806

200 400 600 800

Spectrum 80. Mass spectrum of TBP. 264 MX0N6.204 OATE 16-9*•94 s r 200..132 s r 200.0 0) 3740..000 s i 32768 TO 32768 s» 3597..122 HZ/PT .220

P* 3..2 no 0..0 AO 4 ,.393 RG 40 NS 80 TE 297 ©

FW 4300 o o 02 30000. OP 30L P0 LB 0..0 08 0..0 cx 34..00 CT 16 .00 Ft :o, .000® F2 » (.499® H2/CM 61..604 PPK/CK .3 0 9 SR <329..93

Spectrum 81. 'HNMR (CDC13, ppm) of HPoTME. 300 273 mss 250 300

Spectrum 82. Mass spectrum of HPoTME. 266 80 70 60 268 Spectrum 84. Mass spectrum of BPoTME.