EXPLORING STRUCTURAL DIVERSITY IN NUCLEOSIDE AND NUCLEIC ACID DRUG DESIGN
A Dissertation Presented to The Academic Faculty
By
Peter Ivo O’Daniel
In Partial Fulfillment Of the Requirements for the Degree Doctor of Philosophy in Chemistry
Georgia Institute of Technology
December 2005
Copyright © Peter Ivo O’Daniel 2005 EXPLORING STRUCTURAL DIVERSITY IN NUCLEOSIDE AND NUCLEIC ACID DRUG DESIGN
Approved by:
Dr. Katherine L. Seley, Advisor Dr. E. Kent Barefield Department of Chemistry and School of Chemistry and Biochemistry Biochemistry Georgia Institute of Technology University of Maryland, Baltimore County Dr. Marcus Weck Dr. Donald F. Doyle School of Chemistry and Biochemistry School of Chemistry and Biochemistry Georgia Institute of Technology Georgia Institute of Technology
Dr. Haskell W. Beckham Date Approved: August 24, 2005 School of Polymer, Textile and Fiber Engineering School of Chemistry and Biochemistry, Adjunct Georgia Institute of Technology
To my mom, the one who stood behind me with loving support no matter what I decided
to do and for the inspiration that no matter what happens you will always survive.
Thank you.
ACKNOWLEDGEMENTS
I would like to thank my Advisor Professor Katherine L. (Seley) Radtke for the guidance and mentoring all these years. You have been an inspiration for me and to those who have not followed the traditional path in getting their education.
I am thankful to my thesis committee members: Professors Donald F. Doyle, E.
Kent Barefield, Haskell W. Beckham, and Marcus Weck for their precious time and support.
I am thankful to Professor Olaf Wiest for his time, support, guidance and the continued use of the supercomputer. It was a pleasure meeting and working with you.
I am grateful to Dr. Jean-Christophe Ducom in the Science Computing Facility at
Notre Dame for his help with problems on the supercomputer.
I am grateful to Dr. Lauren Schwimmer for the training and her patience and guidance with the enzyme molecular modeling.
I am grateful to Chris Harrison and Lauren O’Neal for their time, guidance and training with the ab initio and molecular dynamics simulations.
I would also like to thank Dr. Les Gelbaum for NMR assistance as well as David
Bostwick and Sarah Shealy for mass spectrometry assistance.
I would like to give many special thanks to the scientific community at the
University of Notre Dame for their hospitality and excellent resources and the Weist group for all of the good times while I was there.
iv I would like to give many special thanks to the scientific community at the
University of Maryland, Baltimore County for their excellent resources and amazing support throughout the last couple of years of graduate school.
It gives me great pleasure to thank the Seley Research Group members, past and present, Dr. Liang Zhang, Dr. Asmerom M. Hagos, Dr. Samer Salim, Dr. Brian Bakke
Sylvester L. Mosley, Joshua M. Sadler, Naresh K. Sunkara, Lizette Riverea and William
Motel for all their helpful advice, assistance and comradery during my years in the group.
I will always look back on our friendship within the group with fond memories. I would also like to thank the group and Mettachit Navamal for their help with editing and proof reading.
And, finally, I would like to thank my friends and family members for their love and support through out these many years.
v TABLE OF CONTENTS
Acknowledgements iv
List of Tables viii
List of Figures xiii
List of Schemes xxii
List of Abbreviations xxiii
Summary xxix
Chapter 1 INTRODUCTION 1
Chapter 2 CONCEPTS OF DRUG DESIGN 3
Chapter 3 MOLECULAR MODELING OF ISOADENOSINE DERIVATIVES AS INHIBITORS OF SAHASE 39
Background and Significance 39
Discussion 45
Conclusions and Future Directions 59
Experimental methods 60
Chapter 4 MOLECULAR MODELING OF FLEXIMERS AS BIOPROBES FOR FLEXIBLE ENZYMES 61
Background and significance 61
Discussion 68
Conclusion 78
Experimental methods 79
vi Chapter 5 MOLECULAR DYNAMICS SIMULATIONS OF EXTENDED BASES IN DNA 81
Background and significance 81
Modeling Results 116
Conclusion 129
Experimental methods 130
Chapter 6 CHLORINATED ADENOSINE DERIVATIVES 134
Background and Significance 134
Synthesis 145
Experimental 154
Appendix A: IsoA Data 163
Appendix B: Fleximer Data 227
References 263
Vita 295
vii LIST OF TABLES
Table 3.1 Relative energies of IsoA analogues 58
Table 4.1 Results of the enzyme essays with SAHase, parent nucleosides and the distal fleximers 70
Table 4.2 Change in dihedral angle 78
Table 4.3 Relative energies of Fleximer analogues 79
Table 5.1 Average structural parameters of A, B and Z DNA 90
Table 5.2 Modified DNA results 126
Table 6.1 Some genes involved in cell proliferation and methylated in cancer 138
Table 6.2 Difference in TFA volumes vs. product yields 147
Table A.1 Modeling results for Borchardt’s inhibitor, 4’,5’-enyl-3- deazaadenosine 165
Table A.2 Modeling results for Adenosine 166
Table A.3 Modeling results for Ari 1 167
Table A.4 Modeling results for Ari 2 168
Table A.5 Modeling results for NpcA 169
Table A.6 Modeling results for 3-deaza NpcA 170
Table A.7 Modeling results for 4’,5’-enyl-adenosine 171
Table A.8 Modeling results for Isoadenosine 172
Table A.9 Modeling results for IsoA#1 173
Table A.10 Modeling results for IsoA#2 174
Table A.11 Modeling results for IsoA#3 175
Table A.12 Modeling results for IsoA#4 176
viii Table A.13 Modeling results for IsoA#5 177
Table A.14 Modeling results for IsoA#6 178
Table A.15 Modeling results for IsoA#7 179
Table A.16 Modeling results for IsoA#8 180
Table A.17 Modeling results for IsoA#9 181
Table A.18 Modeling results for IsoA#10 182
Table A.19 Modeling results for IsoA#11 183
Table A.20 Modeling results for IsoA#12 184
Table A.21 Modeling results for IsoA#13 185
Table A.22 Modeling results for IsoA#14 186
Table A.23 Modeling results for IsoA#15 187
Table A.24 Modeling results for IsoA#16 188
Table A.25 Modeling results for IsoA#17 189
Table A.26 Modeling results for IsoA#18 190
Table A.27 Modeling results for IsoA#19 191
Table A.28 Modeling results for IsoA#20 192
Table A.29 Modeling results for IsoA#21 193
Table A.30 Modeling results for IsoA#22 194
Table A.31 Modeling results for IsoA#23 195
Table A.32 Modeling results for IsoA#24 196
Table A.33 Modeling results for IsoA#25 197
Table A.34 Modeling results for IsoA#26 198
Table A.35 Modeling results for IsoA#27 199
ix Table A.36 Modeling results for IsoA#28 200
Table A.37 Modeling results for IsoA#29 201
Table A.38 Modeling results for IsoA#30 202
Table A.39 Modeling results for IsoA#31 203
Table A.40 Modeling results for IsoA#32 204
Table A.41 Modeling results for IsoA#33 205
Table A.42 Modeling results for IsoA#34 206
Table A.43 Modeling results for IsoA#35 207
Table A.44 Modeling results for IsoA#36 208
Table A.45 Modeling results for IsoA#37 209
Table A.46 Modeling results for IsoA#38 210
Table A.47 Modeling results for IsoA#39 211
Table A.48 Modeling results for IsoA#40 212
Table A.49 Modeling results for IsoA#41 213
Table A.50 Modeling results for IsoA#42 214
Table A.51 Modeling results for IsoA#43 215
Table A.52 Modeling results for IsoA#44 216
Table A.53 Modeling results for IsoA#45 217
Table A.54 Modeling results for IsoA#46 218
Table A.55 Modeling results for IsoA#47 219
Table A.56 Modeling results for IsoA#48 220
Table A.57 Modeling results for IsoA#49 221
Table A.58 Modeling results for IsoA#50 222
x Table A.59 Modeling results for IsoA#51 223
Table A.60 Modeling results for IsoA#52 224
Table A.61 Modeling results for IsoA#53 225
Table A.62 Modeling results for IsoA#54 226
Table B.1 Modeling results for Flex#1 229
Table B.2 Modeling results for Flex#2 230
Table B.3 Modeling results for Flex#3 231
Table B.4 Modeling results for Flex#4 232
Table B.5 Modeling results for Flex#5 233
Table B.6 Modeling results for Flex#6 234
Table B.7 Modeling results for Flex#7 235
Table B.8 Modeling results for Flex#8 236
Table B.9 Modeling results for Flex#9 237
Table B.10 Modeling results for Flex#10 238
Table B.11 Modeling results for Flex#11 239
Table B.12 Modeling results for Flex#12 240
Table B.13 Modeling results for Flex#13 241
Table B.14 Modeling results for Flex#14 242
Table B.15 Modeling results for Flex#15 243
Table B.16 Modeling results for Flex#16 244
Table B.17 Modeling results for Flex#17 245
Table B.18 Modeling results for Flex#18 246
Table B.19 Modeling results for Flex#19 247
xi Table B.20 Modeling results for Flex#20 248
Table B.21 Modeling results for Flex#21 249
Table B.22 Modeling results for Flex#22 250
Table B.23 Modeling results for Flex#23 251
Table B.24 Modeling results for Flex#24 252
Table B.25 Modeling results for Flex#25 253
Table B.26 Modeling results for Flex#26 254
Table B.27 Modeling results for Flex#27 255
Table B.28 Modeling results for Flex#28 256
Table B.29 Modeling results for Flex#29 257
Table B.30 Modeling results for Flex#30 258
Table B.31 Modeling results for Flex#31 259
Table B.32 Modeling results for Flex#32 260
Table B.33 Modeling results for Flex#33 261
Table B.34 Modeling results for Flex#34 262
xii LIST OF FIGURES
Figure 2.1 Homologation 5
Figure 2.2 Bioisosteres 6
Figure 2.3 Lock and Key 7
Figure 2.4 Induced fit 8
Figure 2.5 Leonard’s benzo-separated adenine analogues. 9
Figure 2.6 Computer-assisted structure-based drug design 10
Figure 2.7 Purine nucleoside phosphorylases 15
Figure 2.8 PNP mechanism 16
Figure 2.9 S-adenosylhomocysteine hydrolase 18
Figure 2.10 Monomer binding domains 19
Figure 2.11 SAHase open and closed conformation 20
Figure 2.12 NAD+ and NADH 21
Figure 2.13 SAHase mechanism of action 22
Figure 2.14 Five most common naturally occurring nucleosides 24
Figure 2.15 Antiviral nucleosides 25
Figure 2.16 Potent HIV RT inhibitors 26
Figure 2.17 Purine and Pyrimidine Numbering 27
Figure 2.18 Deaza nucleosides 28
Figure 2.19 Aza Nucleosides 29
Figure 2.20 Other aza and aza-deaza nucleosides 30
Figure 2.21 IsoA and phosphorylated analogues 31
xiii Figure 2.22 N-3 to N-9 migration 31
Figure 2.23 C-nucleosides 32
Figure 2.24 Heteroatom modifications 33
Figure 2.25 Typical sugar modifications 34
Figure 2.26 Ring size modifications 34
Figure 2.27 Naturally occurring carbocyclic analogues (Ari and NpcA) and their synthetic analogues (4’,5’-tetrahydro and 4’,5’-unsaturated) 35
Figure 2.28 Potent antiviral carbocyclic nucleosides 36
Figure 2.29 SAHase adenosine probes 37
Figure 2.30 Nucleoside inhibitors of S-Adenosylhomocysteine Hydrolase 38
Figure 2.31 4’,5’-enyl-IsoA 38
Figure 3.1 Inhibitors of CTP synthetase 45
Figure 3.2 Borchardt’s inhibitor 46
Figure 3.3 Initial parent nucleosides as compared to Borchardt’s 47
Figure 3.4 IsoA nucleoside as compared to Borchardt’s 48
Figure 3.5 3-Deaza-IsoA nucleosides as compared to Borchardt’s 48
Figure 3.6 Ari and NpcA analogues as compared to Borchardt’s 49
Figure 3.7 Tetrahydro and enyl IsoA analogues as compared to Borchardt’s 50
Figure 3.8 IsoG and IsoX analogues as compared to Borchardt’s 52
Figure 3.9 Thio substituted Ari and NpcA analogues as compared to Borchardt’s 53
Figure 3.10 Thio 4’,5’-tetrahydro and 4’,5’-enyl IsoA analogues as compared to Borchardt’s 54
Figure 3.11 Thio IsoG analogues as compared to Borchardt’s 56
Figure 3.12 Thio IsoX analogues as compared to Borchardt’s 57
xiv Figure 3.13 Cytosine analogues as compared to Borchardt’s 57
Figure 4.1 Seley Fleximers 61
Figure 4.2 Fleximer base numbering for dist and prox 62
Figure 4.3 Flexible Reverse Transcriptase Inhibitors 63
Figure 4.4 Uridine and cytidine flexible analogues 64
Figure 4.5 DNA triplex probe 65
Figure 4.6 Parent distal fleximers as compared to Borchardt’s 69
Figure 4.7 Flex#4 guanosine Fleximer 71
Figure 4.8 Distal N1H and N3H isoguanosines, xanthine and 2-aminoadenosine fleximers as compared to Borchardt’s 72
Figure 4.9 Distal analogues as compared to Borchardt’s 73
Figure 4.10 Proximal IsoGs and distal IsoG N3H as compared to Borchardt’s 74
Figure 4.11 Proximal parent analogues as compared to Borchardt’s 75
Figure 4.12 Sugar moiety alignment of the distal and proximal guanosine and isoguanosine analogues 76
Figure 4.13 Amino acid residue alignment of the distal and proximal guanosine and isoguanosine analogues 77
Figure 4.14 Dihedral structures for Table DC 79
Figure 5.1 Expanded Tricyclic adenosine and guanosine analogues 81
Figure 5.2 Watson and Crick model of DNA 83
Figure 5.3 DNA double helix major and minor grooves 84
Figure 5.4 Watson-Crick and Hoogsteen hydrogen bonding 85
Figure 5.5 Tautomers of A and T 86
Figure 5.6 DNA backbone and sugar torsional angles 87
Figure 5.7 Sugar pseudorotation 89
xv Figure 5.8 DNA forms 91
Figure 5.9 Anti and Syn conformations of purine nucleosides 92
Figure 5.10 DNA base rotational movements 94
Figure 5.11 DNA base translational movements 96
Figure 5.12 Examples of modified hydrogen-bonding nucleosides 98
Figure 5.13 Modified nucleosides 98
Figure 5.14 Minor groove probes 99
Figure 5.15 Increased hydrogen bonding analogues and G Clamps 100
Figure 5.16 Examples of hydrophobic bases 102
Figure 5.17 Bipyridyl and biphenyl C-nucleosides 103
Figure 5.18 Incorporation of lin-benzoadenosine 104
Figure 5.19 Expanded and extended nucleosides 104
Figure 5.20 Tethered naphthalene and tethered adenine analogues 105
Figure 5.21 Minor groove binding drugs 108
Figure 5.22 Evolution of Molecular Dynamics simulations 114
Figure 5.23 Base polarization and electrostatic surface potentials 117
Figure 5.24 Base overlap of inserted tricyclic adenosines 119
Figure 5.25 Modeled 10-mer DNA strands 120
Figure 5.26 Tricyclic dangling ends 121
Figure 5.27 Partial flipped base 123
Figure 5.28 Base intercalation by dGNHC 124
Figure 5.29 Modeled 20-mer DNA strands 125
Figure 5.30 Expanded base pair stacking 127
xvi Figure 5.31 Normal DNA stacking 129
Figure 5.32 DNA energy stability check 131
Figure 5.33 Root mean square distance movement of backbone from dbGSC 133
Figure 6.1 5’ Cap structure of m-RNA 136
Figure 6.2 Unmethylated and methylated cytosine residue 137
Figure 6.3 Epigenetic deactivation process 140
Figure 6.4 Methyltransferase mechanism 141
Figure 6.5 SAHase inhibition mechanism 143
Figure 6.6 Current cancer chemotherapeutic drugs 144
Figure 6.7 Some halogenated nucleosides 144
Figure 6.8 Chlorinated adenine analogues 145
Figure 6.9 Chlorinated 3-deazaadenine analogues 146
Figure 6.10 Chlorination via epoxide mechanism 153
Figure A.1 Isoadenosine modeling numbering scheme 163
Figure A.2 Borchardt’s inhibitor, 4’,5’-enyl-3-deazaadenosine 165
Figure A.3 Adenosine 166
Figure A.4 Ari-1 167
Figure A.5 Ari-2 168
Figure A.6 NpcA 169
Figure A.7 3-Deaza NpcA 170
Figure A.8 4’,5’-enyl NpcA 171
Figure A.9 Isoadenosine 172
Figure A.10 IsoA#1 173
xvii Figure A.11 IsoA#2 174
Figure A.12 IsoA#3 175
Figure A.13 IsoA#4 176
Figure A.14 IsoA#5 177
Figure A.15 IsoA#6 178
Figure A.16 IsoA#7 179
Figure A.17 IsoA#8 180
Figure A.18 IsoA#9 181
Figure A.19 IsoA#10 182
Figure A.20 IsoA#11 183
Figure A.21 IsoA#12 184
Figure A.22 IsoA#13 185
Figure A.23 IsoA#14 186
Figure A.24 IsoA#15 187
Figure A.25 IsoA#16 188
Figure A.26 IsoA#17 189
Figure A.27 IsoA#18 190
Figure A.28 IsoA#19 191
Figure A.29 IsoA#20 192
Figure A.30 IsoA#21 193
Figure A.31 IsoA#22 194
Figure A.32 IsoA#23 195
Figure A.33 IsoA#24 196
xviii Figure A.34 IsoA#25 197
Figure A.35 IsoA#26 198
Figure A.36 IsoA#27 199
Figure A.37 IsoA#28 200
Figure A.38 IsoA#29 201
Figure A.39 IsoA#30 202
Figure A.40 IsoA#31 203
Figure A.41 IsoA#32 204
Figure A.42 IsoA#33 205
Figure A.43 IsoA#34 206
Figure A.44 IsoA#35 207
Figure A.45 IsoA#36 208
Figure A.46 IsoA#37 209
Figure A.47 IsoA#38 210
Figure A.48 IsoA#39 211
Figure A.49 IsoA#40 212
Figure A.50 IsoA#41 213
Figure A.51 IsoA#42 214
Figure A.52 IsoA#43 215
Figure A.53 IsoA#44 216
Figure A.54 IsoA#45 217
Figure A.55 IsoA#46 218
Figure A.56 IsoA#47 219
xix Figure A.57 IsoA#48 220
Figure A.58 IsoA#49 221
Figure A.59 IsoA#50 222
Figure A.60 IsoA#51 223
Figure A.61 IsoA#52 224
Figure A.62 IsoA#53 225
Figure A.63 IsoA#54 226
Figure B.1 Fleximer modeling numbering scheme 227
Figure B.2 Flex#1 229
Figure B.3 Flex#2 230
Figure B.4 Flex#3 231
Figure B.5 Flex#4 232
Figure B.6 Flex#5 233
Figure B.7 Flex#6 234
Figure B.8 Flex#7 235
Figure B.9 Flex#8 236
Figure B.10 Flex#9 237
Figure B.11 Flex#10 238
Figure B.12 Flex#11 239
Figure B.13 Flex#12 240
Figure B.14 Flex#13 241
Figure B.15 Flex#14 242
Figure B.16 Flex#15 243
xx Figure B.17 Flex#16 244
Figure B.18 Flex#17 245
Figure B.19 Flex#18 246
Figure B.20 Flex#19 247
Figure B.21 Flex#20 248
Figure B.22 Flex#21 249
Figure B.23 Flex#22 250
Figure B.24 Flex#23 251
Figure B.25 Flex#24 252
Figure B.26 Flex#25 253
Figure B.27 Flex#26 254
Figure B.28 Flex#27 255
Figure B.29 Flex#28 256
Figure B.30 Flex#29 257
Figure B.31 Flex#30 258
Figure B.32 Flex#31 259
Figure B.33 Flex#32 260
Figure B.34 Flex#33 261
Figure B.35 Flex#34 262
xxi LIST OF SCHEMES
Scheme 6.1 Dichloro to N-oxide 147
Scheme 6.2 N-oxide to amine 147
Scheme 6.3 Amine to diamine 148
Scheme 6.4 Diamine to 8-oxo 149
Scheme 6.5 Diamine to ring closed amine 150
Scheme 6.6 First 3 chloro bases 151
Scheme 6.7 3-Chloro base 152
Scheme 6.8 Alternative route 154
xxii LIST OF ABBREVIATIONS
3-deaza-A...………………………………………………….…………3-Deaza-adenosine
3-deaza-Ari…………………………………………………...…….3-Deaza-aristeromycin
3-deaza-NpcA……….……………………………………………...…3-Deaza-neplanocin
3TC………………………………………………..…….3’-thiaribofuranosyl-β-L-cytosine
5-FU………………………………………………………………………….5-fluorouracil
A………………………………………………………………………………….adenosine
A-………………………………………………………………………………..…A-DNA
ADA……………………………………………………………….…adenosine deaminase
Ado……………………………………………………………………………….adenosine
AIDS………………………………………………acquired immune deficiency syndrome
AMBER……………………………...Assisted Model Building using Energy Refinement
AMP……………………………………………………..……..adenosine monophosphate
ANH………………………………………..……….Tricyclic pyrroloadenosine nucleoside
AO…………………………………………………...Tricyclic furanoadenosine nucleoside
APC……………………………………………………...……adenomatous polyposis coli
Ara-A……………………………………………...……………...…adenosine arabinoside
Ara-C………………………………………………...……1-β-D-arabinofuranosylcytidine
Ara-T………………………………………………….…..1-β-D-arabinofuranosylthymine
Ari……………………………………………………………………..……..aristeromycin
ASFV…………………………………………………………….African swine fever virus
Asn……………………………………………………………………...……….asparagine
xxiii Asp…………………………………………………………………………….aspartic acid
At………………………………………………………Tricyclic thioadenosine nucleoside
ATP…………………………………………………………...…....adenosine triphosphate
AZT……………………………………………….………….3’-azido-3’-deoxythymidine
B-…………………………………………………………………………..………B-DNA bp…………………………………………………………………………………base pairs
BPTI…………………………………………….……Bovine Pancreatic Trypsin Inhibitor
BRCA1……………………..…………………………………...……Breast Cancer 1 gene
C………………………………………………………………………...………….cytosine
CA……………………………………………..……………..cytosine-adenosine base pair
CADD……………………………………………………...computer-assisted drug design
CC…………………………………………………………….……..cytosine-cytosine pair
CFF……………………………………………………………...…...Consistent Forcefield
CG……………………………………………………..……..cytosine-guanosine base pair
CHARMm…………………………..Chemistry as HARvard, Macromolecular mechanics
CpG…………………………………….………cytosine-guanosine dinucleotide base pair
CT………………………………………………………….…..cytosine-thymine base pair
CTP……………………………………………………....cytosine triphosphate synthetase
CVFF……………………………………………….………...Constant Valence Forcefield
Cys……………………………………………………………………………..…..cysteine d4T…………………………………………..…..2’,3’-didehydro-2’,3’-dideoxythymidine ddC…………………………………………………………………..2’,3’-dideoxycytidine ddI…………………………………………………………………….2’,3’-dideoxyinosine
xxiv dist………………………….…………………………………………………………distal
DMA……………………………………………………………..……..dimethylacetamide
DMF……………………………………………………………...N,N-dimethylformamide
DMSO………………………………………………………..…………dimethyl sulfoxide
DNA………………………………………………………………...deoxyribonucleic acid
FAP………………………………………………...….....Familial adenomatous polyposis
fs………………………………………………………………….………….femtoseconds
G………………………………………………………………...……………….guanosine
g+…………………………………………………………………..……………….gauche+
g-………………………………………………………………………...………….gauche-
Gln……………………………………………………………………………….glutamine
Glu……………………………………………………………….…………..glutamic acid
Gly…………………………………………………………………………………..glycine
GNH……………………………………..………….Tricyclic pyrrologuanosine nucleoside
GO…………………………………………...……...Tricyclic furanoguanosine nucleoside
Gt………………………………………………………Tricyclic thioguanosine nucleoside
HBV……………………………………………………………..……….Hepatitis B virus
HCMV……………………………………………………………Human cytomegalovirus
Hcy………………………………………………………………………..L-Homocysteine
HF……………………………………………………………………………Hartree-Fock
His………………………………………………………………………….……...histidine
HIV…………………………………………….………….human immunodeficiency virus
HMG-1………………………………………...…………....high mobility group D protein
xxv HMG-D……………..………………………………………high mobility group D protein
HSV1…………………………………………………………..herpes simplex virus type 1
IC50………………………………………….inhibition of 50% of the enzyme being tested
IHF………………………………………………………….integration host factor protein
Ile…………………………………………………………………...……………isoleucine
Imid…………………………………………………………………..…………..imidazole
IsoA………………………………………………………………………N-3 isoadenosine
IsoG………………………………………………………………………N-3 isoguanosine
IsoX………………………………………………………………...……N-3 isoxanthosine
IUdR……………………………….……………………………...5’-iodo-2’-deoxyuridine
Ki…………………………………………………………………….……rate of inhibition
LEF-1…………………………………………….…..lymphoid enhancer-binding factor 1
Leu…………………………………………………………………………………..leucine lin……………………………………………………………………………..………linear
Lys…………………………………………………………………………………....lysine m-CPBA……………………………………………….….meta-chloroperoxybenzoic acid
MC……………………………………………………………………...……. Monte Carlo
MD…………………………………………………………………….molecular dynamics
Met………………………………………………………………………..…….methionine
MeTase………………………………………………………...………...methyltransferase
MM2…………………………………………Molecular Mechanics, Allinger force field 2
MM……………………………………………………...…………....molecular mechanics
MMC…………………………………………………….………..Metropolis Monte Carlo
xxvi MMFF……………………………………………………….Merck Molecular Force Field
MMR……………………………………………………………..………..mismatch repair
m-RNA………………………………………………………...messenger ribonucleic acid
MSI…………………………………………………..…………....microsatellite instability
N……………………………………………………………………………...……….north
NAD+…………………………………..Nicotinamide adenine dinucleotide oxidized form
NADH………………………..…………Nicotinamide adenine dinucleotide reduced form
NpcA…………………………………………………………….…………..Neplanocin-A
NMR…………………………………………………………..nuclear magnetic resonance
P………………………………………………….…………….pseudorotation phase angle
Phe……………………………………………………………..…………….phenylalanine
PNP………………………………………………………purine nucleoside phosphorylase
ppm………………………………………………………………..………parts per million
prox………………………………………………………………………….…….proximal
ps………………………………………………………………………………picoseconds
PurR………………………...……………purine nucleotide biosynthesis regulator protein
pyr……………………………………………………………….………………pyrimidine
QSAR…………………………………………quantitative structure-activity relationships
RB……………………………………………………………….………….retinoblastoma
RNA…………………………………………………………..…………..Ribonucleic acid
RT……………………………………………………………………..reverse transcriptase
S………………………………………………………………………………………south
SA………………………………………………………….………….simulated annealing
xxvii SAH………………………………………………..……………..S-adenosylhomocysteine
SAHase……………………………………………….S-adenosylhomocysteine Hydrolase
SAM…………………………………………………..…………….S-adenosylmethionine
SAR…………………………………………….…………...structure-activity relationship
Sb-1……………………………………………...…………..staphylococcal bacteriophage
Ser…………………………………………………………………………………….serine
SRY……………………………………………………...………...sex determining protein
T……………………………………………………………………………………thymine
TBP…………………………………………………………………TATA binding protein
TFA…………………………………………………...………………...trifluoroacetic acid
Thr…………………………………………………………………..…………….threonine
TMS………………………………………………………….…………...tetramethylsilane
Tyr………………………………………………………………………………….tyrosine
VHL………………………………….………...von Hippel-Lindau tumor suppressor gene
VSV………………………………………………………………vesicular stomatitis virus
ASFV…………………………………………………………….African swine fever virus
VZV…………………………………………………………..………varicella zoster virus
Z-…………………………………………………………………...………………Z-DNA
xxviii SUMMARY
The design and optimization of chemotherapeutic molecules through molecular
modeling is a rapidly growing aspect of chemotherapeutic drug design. The recent increase in computer power and accompanying decrease in the cost of hardware has led to the wide use of computational chemistry in the development of new drugs. In addition, virtual screening of libraries of compounds also aids in the rapid development of new drugs. In that regard, three computational drug design projects plus a project involving the synthesis of potential inhibitors compile the research presented herein.
In the first project, molecular mechanics simulations were used to model a series of potential inhibitors in the binding site of S-adenosylhomocysteine hydrolase
(SAHase), a biologically significant enzyme that is an essential part in the methylation pathway. The modeled compounds are isomers of adenosine possessing the glycosidic linkage at the N3-position instead of the N9-position found in normal purine nucleosides.
All binding energies obtained for the SAHase enzyme-ligand complexes were compared to the most potent inhibitor reported to date, carbocyclic 4’,5’-enyl-3-deazaadenosine.
In the second project, another series of molecules was modeled using molecular mechanics simulations, and as before, these molecules were also modeled in the binding site of SAHase. This series focuses on using flexible nucleosides as bioprobes to investigate the implications of flexibility on the enzyme. These analogues have their purine bases separated into their imidazole and pyrimidine components, but are connected together by a single carbon-carbon bond. This allows rotation between the two pieces of the nucleobase thereby increasing the overall flexibility of the molecule.
xxix In the third project, molecular dynamics calculations were performed on expanded purine nucleotides in a modified DNA strand. The expanded adenosine and guanosine bases possess either a furan, pyrrole or thiophene spacer ring in between the normal imidazole and pyrimidine rings of the purine scaffold. The calculations were performed on 10- and 20-mer strands in order to probe the stability and parameters of the strands.
The final project focuses on the synthesis of a series of chlorinated 3- deazaadenine analogues that were designed as potential inhibitors of SAHase. This SAR study systematically placed various combinations of chlorine atoms on a 3-deazaadenine base in the 2, 6 and 8 positions to determine the effect of chlorination on a known inhibitor.
xxx CHAPTER 1
INTRODUCTION
One aspect that medicinal chemistry encompasses is the structural modification of
compounds that have known physiological or pharmacological effects with the hope that
biochemical rationale for drug discovery may be found.1 Medicinal chemistry has been practiced for several thousand years2 beginning with the search for cures for diseases by
chewing herbs, barks, roots and berries. The earliest written records are from the Chinese,
Indian, South American and Mediterranean cultures. Approximately 5100 years ago,
Chinese Emperor Shen Nung, in his book called Pentsao, described two of the earliest
medicines. One was derived from the Dichroa febrifuga root, which was later shown to
contain alkaloids, it was used for fevers and is still in use today to treat malaria. The
other, Ephedra sinca, was used as a heart stimulant, a diaphoretic agent and to relieve
coughing. It was later shown to contain ephedrine, which raises blood pressure and
alleviates bronchial spasms.
The “prescientific” period of drug discovery did not progress much further until
the knowledge of anatomy and physiology reached more sophisticated levels.1 In the late
1700’s, an extract from the foxglove plant was used to treat dropsy and this is thought of as the beginning of a more modern approach to the therapeutic treatment of diseases.2
The active compounds digitoxin and digoxin from Digitalis purpurea and Digitalis lanata respectively, were used in the treatment of heart failure. Today, all cardiac glycosides are referred to by the general term “digitalis” and are manufactured by the extraction of foxglove and related plants.
1 In the last 160 years, modern medicinal chemistry advanced rapidly with the purification of the active components from natural sources.2 In addition, the study and classification of diseases in the first half of the 19th century helped in identifying ineffective remedies.1 Today, natural products in their purest form make up only a small percentage of the drugs sold2 and most have been chemically modified to improve their therapeutic properties. In the first half of the 20th century, Paul Ehrlich made several discoveries with the use of dyes and the modification of active compounds to develop antibacterial, antimalarial and antiparasitic agents.1 He also postulated theories that enhanced the development of drug discovery. Since the 1940s, both the discovery of deoxyribonucleic acid (DNA)3 and advances in synthesis and characterization techniques, have led to a more rational approach to drug discovery using chemical design.2 The work present herein is specific to nucleic acid drug design.
2 CHAPTER 2
CONCEPTS OF DRUG DESIGN
Methods of Drug Discovery
Drug discovery is a very expensive and time-consuming process.4 It takes approximately 12-15 years and about $800 million to create a drug for the consumer. For every 10,000 compounds that are tested, 10 may make it to human clinical trials and only
1 usually makes it to market. Successful drug discovery and development can be accomplished by several approaches. For example, random broad-screen testing of chemical libraries, to a more focused approach involving rational design modifications to a lead compound. The first step in understanding the latter approach requires knowledge of the mechanism of action and structural features of an enzymatic target. 1
In contrast, random screening involves no intellectualization; the compounds are
assayed against a variety of diseases or enzyme targets without any regard to their
structure.4 This is widely used today, especially when there is little or no information
known regarding the structure of the target. One of the more recent cases for this type of
research is the “war on cancer” where large numbers of synthetic compounds and natural
products were screened by high throughput methodology.4
A slightly more focused approach is referred to as nonrandom screening. This is
used when the compound has a vague resemblance to a weakly active compound found in
random screening.4 If a compound is found to be active, then further modifications are
undertaken to investigate the source of the activity.
3 Some compounds show biological activity against more than one type of disease
and often times these are discovered during observations during clinical trials.4 For instance, bupropion hydrochloride better known as Wellbutrin® was originally developed
as an antidepressant but was found to help patients stop smoking during the clinical trials.
Another example involves the well-known drug sildenafil citrate (Viagra®), which is now
used for erectile dysfunction. Viagra® was originally designed to treat angina and
hypertension, but the obvious side effect was noted during the clinical trials.
A more rational approach to drug discovery involves the development of a lead compound already known to exhibit activity against a particular disease or biological system.4 Once a lead compound is identified, the mechanism of action for the lead
compound must be understood. Once this is known, the interaction of the lead with the
enzyme is studied to determine the specific atoms or functional groups that are
responsible for the biological activity.4 These atoms are collectively known as the pharmacophore. The remainder of the molecule is known as the auxophore. The parts of the auxophore that are not critical can be removed or modified to help improve binding to the enzyme in hopes of increasing activity, decreasing toxicity or lowering unwanted side effects. These systematic structural modifications, which include homologation, chain branching, ring chain transformations and bioisosterism are critical to the fundamental
design process known as a structure-activity relationship (SAR) study.
SAR studies help reveal which modifications to the lead are important for
recognition by the target enzyme or receptor.4 The process can start with homologation,
which involves a group of compounds that differ only by a constant unit, generally a CH2 group. These types of modifications show predictable regularities of increase or decrease
4 in biological properties and correspond with the overall lipophilicity of the compound.
Chain branching can also alter the potency of a compound because a branched chain is less lipophilic than a straight chain due to the larger molar volumes and different shapes of branched compounds (Figure 2.1). The modification of alkyl substituents into cyclic rings has also led to important pharmacokinetic effects such as increasing lipophilicity or decreasing metabolism.
Figure 2.1. Homologation.4
Another typical modification involves bioisosteres. Bioisosteres are substituents or groups that have chemical or physical similarities and produce similar biological effects. There are two classes of bioisosteres, classical and nonclassical (Figure 2.2 on the following page). Classical isosteres are defined as atoms, ions or molecules that contain peripheral layer of electrons that can be considered identical. A molecule that does not have the same number of atoms and does not fit the steric and electronic rules but produces similar biological activity defines the nonclassical type.
5
Figure 2.2. Bioisosteres.4
Over the years there have been many different hypotheses for how enzymes
catalyze reactions, but one common thread for all is that the reaction is initiated by the
formation of an enzyme-substrate complex.4 The formation of the complex was first
described by Fischer5 as a “lock and key” scenario where the enzyme is considered the
“lock” and the substrate the “key” as seen in Figure 2.3. This type of description accounts for the high degree of specificity of enzymes but does not explain certain phenomena that have been observed.4 Compounds that contained the same
pharmacophore with less bulky substituents often failed to be good substrates even
though they should have fit, while compounds with bulkier substituents exhibited a
greater binding affinity than the original substrate. If the lock and key concept were
6 universal, the reverse should have been true. In addition, if the active site was rigid, it would not be capable of adjusting to both the substrate and the products of a reversible reaction in an optimum fashion. The lack of a reasonable explanation for these issues with the lock and key theory then led to the “induced fit” hypothesis by Koshland.6
Figure 2.3. Lock and Key.7
The induced fit hypothesis proposes that the enzyme does not necessarily exist
100% of the time in the correct conformation to bind the substrate but rather, as the substrate approaches the enzyme, a conformational change is induced that orients the catalytic sites that are necessary for binding to the correct conformation.4 The induced fit theory requires a flexible active site to accommodate the different binding modes and conformational changes needed to accept different rigid substrates (Figure 2.4). The enzyme was suggested to be elastic and would return to its original conformation upon releasing the product. Another related theory was suggested that the substrate would undergo a conformational change even if it were energetically unfavorable.
7
Figure 2.4. Induced fit.7
Prior to the development of sophisticated instrumentation, structural information
such as highly resolved X-ray crystal structures or high field solution NMR structures of
enzymes and enzyme-ligand binding complexes was not available. As a result, scientists
were forced to elucidate the parameters of active sites through exhaustive and systematic
structural modification of substrates to probe the dimensional parameters of an enzyme or
receptor.1
One of the first examples of using dimensional probes were the expanded purine
nucleosides introduced in the early 70’s by Nelson Leonard. These novel nucleosides
contained spacer ring(s) separating the purine’s pyrimidine and imidazole rings, most
notably the linear (lin-), distal (dist-), and proximal (prox-)benzo-separated compounds
(Figure 2.5).8-11 These expanded nucleoside and nucleobase analogues were used as
bioprobes to explore the flexibility of numerous enzyme and coenzyme binding sites to
help elucidate structural information about enzymes. These and other shape-modified
nucleosides led to the development of a variety of shape-modified bases and nucleosides
designed to reveal critical information about enzyme structure.9-11
8
Figure 2.5. Leonard’s benzo-separated adenine analogues.8-11
Leonard’s expanded nucleosides spawned a myriad of research in the field. His
compounds were tested against a wide variety of enzymes such as pyruvate kinase,
phosphofructokinase, hexokinase, 3- phosphoglycerate kinase acetate kinase, adenosine
deaminase (ADA), xanthine oxidase and adenylate kinase.11,12 While in many cases these
analogues were recognized, they were rather limited in their utility due to their inherent
rigidity.12
Although this approach to drug design was rational, it remained somewhat
random, as well as tedious and of course, expensive. With the advent of faster computers
in the 1980s and the development of methodologies to calculate molecular properties, computer-assisted structure-based drug design (CADD) has emerged as a new tool in medicinal chemistry. The use of X-ray crystallography to elucidate three-dimensional structures of enzyme substrate complexes, and computers to calculate and display the information has also accelerated the drug design process. Today, the crystal structures for many enzymes are readily available through several databases, and this has facilitated a much more sophisticated approach to the design of new drugs.
9 The typical drug design approach used today is shown in Figure 2.6 and begins
with creating a computer model of an enzyme-binding site based on known
crystallographic data. Next, potential inhibitors or ligands are then docked into the
binding site; the computer model of the enzyme-substrate complex is then subjected to
energy minimization using any of a wide variety of innovative modeling programs
presently available. Based on the computational analysis of structural interactions
between the enzyme and several different potential substrates, the most promising inhibitors are identified then actually synthesized.
Figure 2.6. Computer-assisted structure-based drug design.
Each of the targets is then assayed to measure the levels of binding affinity (Ki) for the enzyme and the inhibitory activity against the enzyme (IC50). Once these values are determined, the inhibitors are either co-crystallized or “soaked” into a preformed
10 crystal of the enzyme and the enzyme-substrate complex subjected to X-ray
crystallography techniques. The data is evaluated and then fed back into the algorithm to
be used to design a second generation of inhibitors.
Computer-Assisted Drug Design
Molecular modeling arose from the concepts of molecular mechanics (MM) with the use of molecular bonding and van der Waals forces.13 The basic idea can be traced to
the 1930s. However, the first recorded attempts at MM were in 1946 to calculate the racemization rates of biphenyl derivatives.14 Due to the lack of computers the field
showed little growth until the early 1960s.13 Several independent research groups worked on the development of force fields based on the physical parameters of small compounds and functional groups. In the early 1970s, the first molecular dynamics work on polar molecules was reported and by the late 1970s molecular mechanics force fields were used to refine crystal structures and led to the development and use of simulated annealing techniques.
In the 1980s, molecular modeling expanded from a research activity at universities to an important tool for rational drug design.1 This occurred from the
decrease of computing power and the development of powerful graphics display
hardware and powerful modeling software. Molecular modeling is generally comprised
from quantitative structure-activity relationships (QSAR). These involve use of molecular
graphics, computational chemistry and statistical modeling, along with chemical data/information management. This latter component involves using databases for searching and organizing properties to provide strategies and statistical analysis.
11 Molecular mechanics calculations are the fastest approach to obtaining data in
computational modeling.1 Each atom is treated as a mass proportional to its atomic mass
and each bond formed is treated as a mechanical spring that has a specific force constant
associated with it. The bonds have restoring forces to preserve the optimal geometry of
the molecule. These forces add up to the total potential energy for each atom and bond in
the molecule and follow the formula below in which E represents the energy attributed to
that particular bond distortion and Σ represents the sum of all of a particular bond distortion.
Etotal = ΣEstretch + ΣEbend + ΣEtorsion + ΣvdW + ΣECoulomb (2.1)
The distortions are separated according to a their atomic contributions, where the
stretch contribution is the stretching or compression of a bond, the bending is the
movement of a bond angle away from the ideal angle and the torsional is the rigidity of
the twisting around a dihedral angle.1 The last two parameters, van der Waals and
Coulombic, are nonbonded interactions and describe the steric and electrostatic or charge interactions respectively and are the result of attraction or repulsion of partially or fully
charged species.
The bonding parameters are the dominant contribution to the formula. 1 The
bonding contributions are derived from Hooke’s law for harmonic motion, where k is the
force constant, r is the bond distance and r0 is the ideal bond distance and the values for k
and r0 are specific to the atom and bond types. Although the energy to compress a bond is
greater than stretching a bond, Hooke’s law works reasonably well for short distances.
12 (2.2)
The nonbonding terms are only involved with atoms that are not bonded together and are present in all molecules. The importance of the nonbonded terms is dependent on the distance between the interacting atoms. The Coulombic interaction between two-point charges follows the equation where q is the charge on atom i or j, D is the dielectric constant between the two charges and rij is the distance between the atoms.
(2.3)
The Van der Waals term is the more complicated equation.1 At close distances
where atomic orbitals overlap, there is a strong repulsive force because two objects
cannot occupy the same space at the same time. At great distances there is are weak
London dispersion forces. The function that relates this information is from the Lennard-
Jones potential shown below. Where ε is the minimum energy at rmin, rmin is the distance
between atoms at the minimum energy, and r is the actual distance between atoms.
(2.4)
Since these equations are based on classical models, computational time is
minimal compared to quantum mechanical methods and simulations on systems with
13 more than 10,000 atoms can be calculated relatively quickly.1 These simulations can be
corrected for temperature and pressure, so reliable trajectories can be computed as well as
vibrational modes. These simulations can also provide information about binding modes
and sterics of pharamacophores, however more advanced molecular mechanics
techniques must be used to determine the binding free energy for an enzyme-ligand
system. This makes these methods useful for investigating molecule conformational
energies and for studies on macromolecules like proteins and polymers.
That said, there are limitations and disadvantages to this method.1 Electronic
properties such as electron density are not calculated since this type of method is only
concerned with the nuclei of the molecule. Force constants of the energy terms must be
well characterized and be used appropriately for the molecule under investigation. To
overcome the disadvantage, several different types of forcefields have been developed.
These include Assisted Model Building using Energy Refinement (AMBER), Chemistry
at HARvard, Macromolecular mechanics (CHARMm), Constant Valence Forcefield
(CVFF), Consistent Forcefield (CFF), Molecular Mechanics, Allinger force field 2
(MM2) and Merck Molecular Force Field (MMFF).1,15 These forcefields and others15 streamlined the computational process and enabled modeling of larger structures such as enzymes and DNA which are used to help predict molecule interaction and structure.
Enzyme Theory and Molecular Modeling
With the greater availability of multiple crystal structures for a particular enzyme, it has been discovered that enzymes are more flexible than once thought and are often able to accept a wide variety of substrates. The inability to sample full movement of the
14 enzyme in the energy calculation causes inaccuracies when using traditional structure
based drug design. Even though a “best guess” compound may appear to fit well
computationally, in reality, it is often a poor inhibitor.16,17
An excellent example of this problem was seen with purine nucleoside
phosphorylase (PNP),18,19 which is a biologically significant enzyme involved in the
purine salvage pathway. PNP is a homotrimer of approximately 100 kDa subunits possessing three identical active sites containing phosphate ions (Figure 2.7).
Figure 2.7. Purine nucleoside phosphorylase.20
15 The enzyme catalyzes the reaction by distorting the substrate to favor the
transition state by forming an activated C1’ carbon. This activated glycosidic linkage is
then nucleophilically attacked by the bound phosphate ion, cleaving the nucleoside into a
its purine base and a phosphorylated sugar (Figure 2.8).21 The free purine moiety is then
either metabolized or salvaged via other biological pathways.
Since many nucleoside-based anticancer and antiviral agents are synthetic mimics
of the natural purine nucleosides, they are often recognized and cleaved by PNP. As a result, the design of potent PNP inhibitors that could be administered with these drugs to aid in their stability is an important goal.21 Furthermore, PNP is also needed for the
proper function of T-cells. Since excessive T-cell activity has been implicated in a variety
of autoimmune disorders such as multiple sclerosis, psoriasis, rheumatoid arthritis,
systemic lupus erythematosus, and insulin-dependent diabetes, it is clear that successful
inhibition of PNP will also aid in the suppression of T-cell activity, thereby may help to
control these diseases.
Figure 2.8. PNP mechanism.21
In that regard, researchers at Squibb Research Institute set about to design
inhibitors of PNP that, if administered with a nucleoside drug, would block the binding
16 site, thereby allowing the drug to successfully reach its chemotherapeutic target.19
Computer modeling of several inhibitors in the binding site of the enzyme indicated great affinity for the enzyme. The compounds were synthesized and subsequently tested, but unfortunately, the compounds were completely inactive.
After an exhaustive investigation involving the crystallization of several inhibitors in the binding site of PNP, it was discovered that the enzyme changed shape upon binding. More significantly, it possessed a short chain of amino acids that acted as a swinging gate, that would open to allow access to the active site, but once the substrate was bound, would then close down over the top of the substrate.19 This essentially
allowed a variety of different types of inhibitors to fit in the binding site giving the
appearance of a perfect fit, when in reality, it was not.
Most new modeling programs on the market today have the ability to take ligand
flexibility into account. However, very few programs are capable of accurately and quickly predicting the flexibility of enzyme when modeling inhibitors in their binding sites. At best, most only allow limited flexibility of the enzyme.22-24 Given that it is now
common knowledge that enzymes are capable of adapting to a variety of conformations,
it is clear that crystal structures of enzymes are only a “snapshot” of a brief moment in
time and do not represent the full mobility of the enzyme. As a result, what crystallizes
out may not be the structure that is responsible for the activity.25-30 With this in mind, it is
necessary to develop new approaches to computer-assisted drug design. With the
capabilities of computational power increasing exponentially everyday, the development
of better modeling programs and experimental techniques is becoming a more reachable goal.
17 Modified Nucleosides as Enzyme Inhibitors in Drug Design
One of the primary focuses for ongoing research in the Seley group is centered around the design and synthesis of inhibitors of S-adenosylhomocysteine hydrolase
(SAHase). SAHase is an important chemotherapeutic target since the enzyme is involved
in the pathways of many viruses. The targeting of the enzyme with inhibitors began in the
late 1970s and early 1980s.31 Most of the early inhibitors were naturally occurring
analogues of adenosine. SAHase has been studied extensively and recently it has been
shown to be flexible.32
Figure 2.9. S-Adenosylhomocysteine hydrolase.33
Mammalian SAHase is a homotetramer of ~48,000 Mr subunits Figure 2.9. The
tetramer forms via an inter-subunit interaction between two α helices in each symmetrical
18 monomer unit.33 The tetramer also involves interactions between catalytic domains on the
periphery to form the active site of SAHase.
Figure 2.10. Monomer binding domains.33
Each monomer contains 432 amino acid residues, a binding pocket, and one
molecule of nicotinamide adenine dinucleotide (NAD+).33 The monomer is folded to
afford two-domains and is structurally similar to NAD+-dependent dehydrogenases. The
coenzyme-binding and catalytic domains both exhibit a core α/β structure with the NAD+ and a substrate-binding site located in the cleft between the two domains Figure 2.10.
There is, however, a difference between SAHase and NAD+-dependent dehydrogenases.
In the NAD+ binding site of SAHase, the amino acid residues tyrosine (Tyr) 430 and
19 lysine (Lys) 426 of one monomer residue interact with the cofactor NAD+ predominately bound to another monomer.33
Upon binding of a substrate by SAHase, there is a 17º rigid body movement between the two domains to form a closed structure Figure 2.11.32 The movement occurs through a kink within an α helix, which serves as a hinge for motion between the two
domains.33 In the closed complex, the amino acid side chain, phenylalanine (Phe) 302,
obstructs the binding site and is also likely involved in hydrophobic interactions with the
aliphatic chain of the homocysteine moiety.
Figure 2.11. SAHase open and closed conformation.32
20 The crystal structure of SAHase is a dimer,33 which has a unique NAD+-binding domain. This domain involves an interaction between the two-monomer subunits, which requires contributions from both dimer subunits. As previously mentioned Lys 426 and
Tyr 430 of one monomer interacts with the NAD+ predominantly bound to the other monomer. This interaction is essential for the activity of the enzyme.33
SAHase Mechanism of Action
The catalytic mechanism of SAHase starts with the oxidation of the substrate’s 3’-
OH group, most likely by Lys 186. This is due to hydrogen bonding to the Lys residue by
the oxygens of glutamic acid (Glu) 156, asparagine (Asn) 191 and Asn 181 and makes
Lys 186 more nucleophilic.33,34 The stacking of the enzyme bound NAD+ ring and the
cyclopentyl ring of the inhibitor suggests a hydride transfer from the C3’H of the sugar to
the C4 of the nicotinamide to afford nicotinamide adenine dinucleotide (NADH) (Figure
2.12) and the 3’-keto AdoHcy intermediate (Figure 2.13 on the following page).33,34
Figure 2.12. NAD+ and NADH.
In the second step, the C4’ proton is abstracted by a base, which is most likely a water molecule. Homocysteine is then subsequently eliminated to form 3’-keto-4’-
21 dehydro-adenosine. The resulting exocyclic double bond undergoes a Michael-type
addition of water to give rise to the 5’-hydroxyl moiety. Finally, to afford adenosine,
NADH reduces the 3’ -ketoadenosine.33
Figure 2.13. SAHase mechanism.33
In most cases, SAHase inhibitors have been modified adenosine analogues,
however recently a guanosine analogue was found to be active against the enzyme.35
Modified nucleosides are important chemotherapeutic agents and have spawned a plethora of research in the field of drug design and discovery.
22
Nucleosides as Chemotherapeutics
DNA and its counterpart ribonucleic acid (RNA) are biopolymers, which are the
initial precursors in the formation of proteins that are responsible for many biological and
physiological processes in the body. These nucleic acids are composed of nucleosides
held together by phosphodiester linkages.1 Since the human body is dependent on
nucleosides for many cellular processes, it is logical to postulate that they could be used
as potential targets for drug design.
Levene and Jacobs proposed the term “nucleosides” in 1909 to describe sugar-
based analogues of purines and pyrimidines.36 Thee naturally-occurring nucleosides
contain a sugar moiety of 2’-deoxyribose-β-D-ribofuranose or β-D-ribofuranose
respectively coupled to a purine or pyrimidine heterocyclic base by a glycosidic bond.
The purine bases, adenine and guanine, are coupled at the N-9 position whereas the pyrimidines, cytosine, thymine and uracil, are attached at the N-1 position to the C1’ of the sugar (Figure 2.14 on the following page). Adenine, guanine and cytosine are found in both DNA and RNA while thymine is only found in DNA and uracil only in RNA.
These compounds are also important to many processes in the body and are components of biologically significant molecules. One important example is adenosine triphosphate
(ATP), which is an important energy source that is used throughout the body for processes such as biosynthetic reactions, ion transport and cell movement.3
23
Figure 2.14. Five most common naturally occurring nucleosides.
In the 1960s, modified nucleosides emerged from their primary use in cancer research to serve as potent antiviral agents.37 Compounds like 5’-iodo-2’-deoxyuridine
(IUdR) were found to be effective against herpes keratitis, and adenosine arabinoside
(ara-A) (Figure 2.15 on the following page) was used to treat encephalitis and neonatal
herpes. Although these compounds were shown to be very effective against herpes,
unfortunately they were not selective, which resulted in toxic side effects. Despite these shortcomings, other nucleoside analogues were synthesized to increase selectivity. This led to the design and development of biologically active acyclic nucleosides, most notably Acyclovir, which is still used to treat herpes simplex virus (HSV). Acyclovir is phosphorylated by a virus-encoded enzyme and incorporated into the viral DNA. Since acyclovir lacks a 3’-hydroxyl, it acts as a chain terminator and extension of the strand is halted. Other potent nucleoside analogues such as 1-β-D-arabinofuranosylcytidine (ara-
C), and 1-β-D-arabinofuranosylthymine (ara-T),38 have been used to treat varicella zoster
24 virus (VZV) and herpes simplex virus I (HSV1). These early findings point to the use of
modified nucleoside analogues as viable chemotherapeutic agents.
Figure 2.15. Antiviral nucleosides.37,38
The importance of nucleoside analogues was more recently confirmed with the
use of 3’-azido-3’-deoxythimidine (AZT). Acquired immune deficiency syndrome
(AIDS) is caused by a retrovirus, and37 upon infection of a host cell, all retroviruses are
reverse transcribed from viral RNA genome into a double-stranded DNA copy. Inhibiting
this process halts replication of the virus. The first drug approved for the treatment of human immunodeficiency virus (HIV) was AZT. AZT developed almost 27 years ago initially as an anticancer agent, and inhibits HIV-1 reverse transcriptase (RT) by acting as a chain terminator. Several other modified nucleosides are also involved in the inhibition of HIV-1 RT, namely 2’,3’-dideoxycytidine (ddC), and 2’,3’-dideoxyinosine (ddI), which
25 were the first to be used in HIV combinatorial drug therapy. More recently 3’-
thiaribofuranosyl-β-L-cytosine (3TC), and 2’,3’-didehydro-2’,3’-dideoxythymidine (d4T)
have proven to be even more potent against HIV (Figure 2.16).37
Figure 2.16. Potent HIV RT inhibitors.37
Although nucleosides have been successful as enzyme inhibitors, there have also
been some drawbacks. First, many of the processes and replication pathways are shared
by both the virus and healthy host cells.37 Second, the clinical use of these drugs is limited due to several factors, including toxicity, a wide range of side effects, problems
with lipophilicity, transportation across cell membranes, and the susceptibility to
enzymatic cleavage.39 Despite these limitations, researchers continue to investigate
modified nucleosides as potent chemotherapeutic agents.
26 Several basic structural modifications to purine and pyrimidine heterocyclic bases
have been explored. Figure 2.17 shows the numbering system commonly used for the
nucleosides.
Figure 2.17. Purine and Pyrimidine Numbering.
Structural modifications to nucleosides can be made to either the base moiety or the sugar moiety. Typical base modifications often involve the removal of nitrogen from a ring and replacing it with a methine to form deaza nucleosides, such as 1-deaza, 3- deaza, 7-deaza and 9-deaza for the purine ring system, or 1-deaza and 3-deaza for the pyrimidine system. These types of modifications have resulted in increased activity for nucleoside anticancer and antiviral agents.
For example, 1-deazaadenosine (as depicted in Figure 2.18 on the following page) has shown a broad spectrum of activity40 including inhibition of blood platelet
aggregation,41 as well as serving as an agonist of adenosine receptors.42 3-
Deazaadenosine, (Figure 2.18) has also exhibited potent biological activity and does not
undergo phosphorylation at the 5' position.43 In addition, 3-deazaadenosine has been
shown to be a potent inhibitor of SAHase,44 which is widely recognized as a desirable enzymatic target for antiviral, antiparasitic and anticancer agents.
27
Figure 2.18. Deaza nucleosides.40,43,45-49
The 7-deazapurines (Figure 2.18) have also gained wide attention since a broad
spectrum of activity has been found for such inhibitors as 7-iodotubercidin, a potent inhibitor of adenosine kinase.45 Tubercidin itself is an active inhibitor of a wide range of
viruses and leukemia L-1210 cells.50 9-Deazaadenosine has also been found to be a
potent inhibitor of human leukemia cell lines and a growth inhibitor of nine human solid
tumors in vitro.51,52 9-Deazainosine has been used for the treatment of African
trypanosomiasis (sleeping sickness)46,47 and has shown promise against Pneumocystis
carinii pneumonia.53
In contrast, fewer studies of pyrimidine analogues have been reported. Of those,
3-deazauridine and 3-deazacytidine (Figure 2.18) have exhibited significant growth
inhibition against L-1210 leukemia in vitro.48,49
28 An alternative approach involves the removal of a methine group and addition of
a nitrogen to form an aza nucleoside. Common aza purine modifications are 2-aza and 8-
aza, and for pyrimidines, the 5-aza nucleosides. The purine nucleoside 2-azaadenosine
has shown activity against ADA.54 8-Azainosine (Figure 2.19) has shown activity against leukemia L-1210 and 8-azaadenosine (Figure 2.19) is a substrate for adenosine kinase and has shown antileukemic activity.55 The modified pyrimidine analogues of 5-
azacytidine (Figure 2.19) have shown activity against prostate cancer, leukemia L-1210
and are known to inhibit DNA methylation.56-58
Figure 2.19. Aza Nucleosides. 55-58
Several other aza/deaza modifications have also been investigated. For instance,
8-aza-1-deaza-6-chloro purines have shown activity against the Coxs viruses and are
inhibitors of ADA (Figure 2.20 on the following page).59 The 8-aza-immucillins are potent inhibitors of PNP and nucleoside hydrolases.60 Furthermore, 2’-deoxy-8-aza-3-
deazaadenosine and 8-aza-3-deazaadenosine have shown activity against African swine
fever virus (ASFV) and vesicular stomatitis virus (VSV).61
29
Figure 2.20. Aza-deaza nucleosides.59
Another type of modification is the formation of structural isomers, isoadenosine
(IsoA) is a structural isomer of adenosine where the base is attached at the N-3 rather
than the N-9 to the anomeric carbon of ribose (Figure 2.21 on the following page).62
These types of compounds were first synthesized in 196363 and were phosphorylated in
1965.8 Some of these analogues were found to compete for enzymes in the pyrimidine
biosynthetic pathway.64,65 These analogues were also found to possess anticancer and antiviral properties.66-68 IsoA was tested against a variety of enzymes and found that it
was recognized and substituted for adenosine in coenzymes functions, enzymatic
conversions and the synthesis of polynucleotides.69-71 The phosphorylated compounds
IsoA, isoguanosine (IsoG) and isoxanthosine (IsoX) (Figure 2.21) were also readily
incorporated by enzymes into short template directed oligomers72,73 and were shown to
hydrogen bond through modified Hoogsteen base pairing.74
One inherent problem of the IsoA nucleosides was the instability of the glycosidic
bond. These types of compounds are susceptible to cleavage by acids and bases and
through a 1,3-migration, will form the natural nucleosides (Figure 2.22 on the following
page). This migration is highly favorable, as it restores the aromaticity of the ring. One
30 way of eliminating this problem would be to replace the N-3 nitrogen atom with a carbon forming a carbon-carbon bond. These nucleosides are known as “C-nucleosides”. This change imparts stability against acid and base cleavage, as well as by purine nucleoside phosphorylase and other enzymes.
Figure 2.21. IsoA and phosphorylated analogues.72,73
Figure 2.22. N-3 to N-9 migration.
31 Several C-nucleosides, such as pseudouridine and pyrazofurin, are naturally
occurring and are potent anticancer and moderately active antiviral agents (Figure
2.23).75,76 One of the most potent C-nucleosides is tiazofurin,77 which is active against
IMP-dehydrogenase and found to induce cell differentiation in neuroblastoma, leukemia
and melanoma cells.78-81 Furthermore, C-nucleosides have also shown antitumor and antiviral activity.82,83
Figure 2.23. C-nucleosides. 75-77
Other heterocyclic modifications to nucleosides have involved the isosteric
replacement of nitrogen or carbon atoms with other heteroatoms such as sulfur or oxygen, thereby allowing the heterocyclic ring to retain its aromaticity and provide an additional electron pair donor. Several studies on thio-containing purines have been investigated.84-
93 These modified purines displayed broad anti-microbial activity,86-89 molluscicidal93 and
anti-human cytomegalovirus (HCMV) activity92 while a thio-containing guanosine
analogue (Figure 2.24 on the following page) exhibited extremely potent immunoactivity
by inducing interferon production.91
Related to this, the IsoA nucleosides can also be considered as 5,6-disubstituted
pyrimidines, such as oxazolopyrimidines, thiazolopyrimidines and Thienopyrimidines
32 shown in Figure 2.24.94-100 Some of these compounds have shown moderate activity
against leukemia L1210 cell lines.94
Figure 2.24. Heteroatom modifications.94-100
Sugar modification is another typical alteration that focuses on replacement of the
hydroxyl groups with hydrogen or other functional groups. Also, removal of the furanose
oxygen and replacement with a methine (CH2), an aza (NH) or thio (S) as well as
changing ring size are all common sugar modifications Figure 2.25.
Ring size adjustments have also proven to be a beneficial modification. The 3- membered ring carbocyclic of acyclovir is (Figure 2.26) 20 times more potent than acyclovir itself against HSV-1 and ten times more against VZV and the four membered ring oxetanocin G (Figure 2.26) displays significant activity against HIV.101
33
Figure 2.25. Sugar modifications.
The replacement of the furanose oxygen with a methine has been extensively
employed to form carbocyclic nucleosides.102 As with C-nucleosides, removal of the
furanose oxygen imparts stability to the nucleoside. This modification transforms the
glycosidic bond from an unstable hemiaminal to a more stable tertiary amine, which
endows the nucleoside with the ability to resist cleavage by a variety of enzymes,
including phosphorylases, as well as to increase the overall lipophilicity of the
nucleoside103.
Figure 2.26. Ring size modifications.101
The replacement of the furanose oxygen with a methylene in adenosine forms the naturally occurring carbocyclic nucleoside aristeromycin (Ari). This nucleoside, along with the naturally-occurring neplanocin A (NpcA), an unsaturated analogue of Ari
34 (Figure 2.27), were the first carbocyclic nucleosides to display significant activity against SAHase.104,105 Although they were very active they were also significantly toxic since they were readily phosphorylated to their triphosphate form and resembled adenosine triphosphate, which is used throughout the body in numerous biological processes.
The 4’-unsubstituted analogues of Ari and NpcA were synthesized to alleviate the formation of the phosphorylated metabolites (Figure 2.27) and have shown significant biological activity against both SAHase and DNA methyltransferase (MeTase), both important enzymatic targets for antiviral, antiparasitic and anticancer agents.105-109
Figure 2.27. Naturally occurring carbocyclic analogues (Ari and NpcA) and their synthetic analogues (4’,5’-saturated and 4’,5’-unsaturated).105-109
Although the enantiospecific synthesis of carbocyclic nucleosides is
challenging,101,110-114 antiviral nucleosides such as carbovir, abacavir and lobucavir have
been made and shown to be extremely potent against HIV (Figure 2.28 on the following
page).111,115-120 In short, it is clear that the judicious use of carefully chosen structural modifications to nucleosides can result in meaningful biological activity.
35
Figure 2.28. Potent antiviral carbocyclic nucleosides.111,115-120
Nucleoside Inhibitors of S-Adenosylhomocysteine Hydrolase
As mentioned previously, SAHase is a widely recognized enzymatic target for antiviral, antiparasitic and anticancer agents. Many different types of nucleoside analogues have been shown to inhibit SAHase.44 The inhibitors can be placed into two categories, type I and type II. Type I are mechanism-based inhibitors, which are oxidized by the enzyme bound NAD+ to give the inactive enzyme-bound NADH.33 In contrast, type II starts with the enzyme activating the inhibitor, which then irreversibly inactivates the enzyme by formation of a covalent bond.
As a class, carbocyclic nucleosides are the most potent inhibitors of SAHase, and most are considered to be type I inhibitors.33 A study on the formation of S- adenosylhomocysteine (SAH) product analogues from many of the Ado analogue precursors showed that the SAH product is dependent on structural modifications introduced into the purine moiety (Figure 2.29 on the following page).121,122 Substrate activity was partially retained when the 6-amino group of Ado was replaced with a hydroxyl group or a hydrogen atom, but only mildly tolerated when replaced with a methyl group. However, larger groups such as benzyl or phenyl groups showed no activity.
36
Figure 2.29. SAHase adenosine probes.121,122
The first demonstration of potent inhibition of SAHase was by 3-deazaadenosine
(3-deaza-A) (Figure 2.30 on the following page).123 The compound was converted to S-
3-deazaadenosylhomocysteine, was resistant to deamination44 and was also found to
serve as a good substrate and a potent inhibitor.124-126 After broad screen testing, Ari was
later found to be more potent than 3-deazaadenosine.127 Based on the effective inhibition
of both 3-deazaadenosine and Ari, 3-deazaaristeromycin (3-deaza-Ari) was synthesized
and discovered to be more potent than the naturally occurring Ari. Capitalizing on this lead, 3-deazaneplanocinA (3-deaza-NpcA) was synthesized and proved to be the most
potent and least cytotoxic of the series.44,128 One interesting, but as yet unexplained,
phenomenon associated with the 3-deaza nucleosides is that they do not undergo
phosphorylation at the 5' position, despite possessing a 5’ hydroxy methyl group.43
More strikingly, 3-deaza nucleosides have shown extraordinary inhibitory effects against a variety of viruses and cancers including HSV-1, HIV, oncogenic DNA virus,
leukemia L1210, Hepatitis B virus (HBV) and several other viruses, parasites and
cancers.128-136
37
Figure 2.30. Potent SAHase inhibitors.
Using the various structural leads outlined herein, a series of carbocyclic isoadenosine analogues were proposed as potential dual inhibitors of SAHase, and indirectly, DNA MeTase. By combining the connectivity of IsoA with C-nucleosides, the stability and migration issues of IsoA would be eliminated. The addition of the sulfur to form a thiazole would hopefully increase substrate/active site interactions since the oxidative/reductive mechanism of SAHase involves water molecules bound in the active site. The extra electron pair donor of sulfur could stabilize the water molecules. In addition, this would also allow retention of the aromaticity. The synthesis of 4’,5’-enyl-
IsoA (Figure 2.31) and several other IsoA analogues have been completed and biological testing is underway. The computational efforts for this project are presented next.
Figure 2.31. 4’,5’-enyl-IsoA.
38 CHAPTER 3
MOLECULAR MODELING OF ISOADENOSINE DERIVATIVES AS INHIBITORS OF S-ADENOSYLHOMOCYSTEINE HYDROLASE
Background and Significance
® The modeling in the next two chapters utilized Affinity , a module of Insight II,
which is part of the Accelrys software package. The program is designed to use a two-
step calculation process via Monte Carlo simulation and simulated annealing to generate
possible binding modes of a ligand in the binding site of an enzyme. The program was
chosen for the calculations because it allows flexibility of both the binding site of the
enzyme and the ligand. Furthermore, it uses a forcefield instead of a scoring function for the simulation, which gives more reliable results.
The initial forcefield employed was CVFF but later efforts shifted to the use of the CFF forcefield since it contains a higher level of accuracy. CFF is an all atom forcefield with a level of accuracy between that of the Hartree-Fock (HF) and the B3LYP levels. The forcefield was built from the predecessors of CFF91 and CFF95.
CVFF was initially developed from a diagonal forcefield with a large number of cross terms with insufficient reparameterization, which results in low predictability.15
Furthermore, it contains parameters that are acquired from fitting gas and crystal structures to small organic molecules.
CFF is well parameterized and has been validated through a comparison of the theoretically calculated values to the experimental vibrational frequencies as well as the experimental conformational and internal rotation barrier energy differences. However, it does not contain sufficient halogen parameterization and is therefore not recommended
39 for molecules involving halogens.15 CFF is a class II forcefield developed from ab initio
calculations only using the HF level of theory and the 6-31G* basis set.137-139 This forcefield is greatly enhanced with the addition of the cubic and quartic polynomial terms to the stretching and diagonal energy formulas as well as addition of two and three fold
Fourier periodic terms to the torsional energy. The expansion is necessary to avoid complex problems regarding potential energy surfaces. The forcefield uses quantum calculations to determine the parameters for energy functions in the same fashion thereby establishing greater accuracy and consistency.140
CFF contains twelve types of energy terms: (1) bond stretching, (2) angle
bending, (3) dihedral rotation, (4) out of plane deformation, and eight cross terms, which
include the Coulombic electrostatic and a van der Waals energy terms (see equation 3.1
on the next page). The van der Waals term uses an inverse 9th power term for the
repulsive behavior of the atom instead of the traditional 12th power term, and the extra
cross terms allow it to more accurately reproduce vibrational spectra.141 The forcefield
was validated for use with carbohydrates, hydrocarbons, lipids, nucleic acids, proteins,
combinations of the above and complexes such as nucleic acids with proteins and counter ions.137,138,142-149
No hydrogen bonding terms are added to the forcefield since it was determined
that electrostatic and van der Waals calculations accurately capture hydrogen bonding
interactions141 both partial and formal charges are assigned based on CFF’s table of atom
types and associated charges.150 The functional form of CFF is represented by the following equation(s)148:
40 2 3 4 Epot = Σ [Kb,2(b – bo) + Kb,3(b – bo) + Kb,4(b – bo) ] bonds + 2 3 4 Σ [Kθ,2(θ – θo) + Kθ,3(θ – θo) + Kθ,4(θ – θo) ] angles + Σ [KΦ,1(1 – cosΦ) + KΦ,2(1 – cos2Φ) + KΦ,3(1 – cos3Φ)] dihedrals + 2 Σ Kχχ impropers + Σ Σ Kbb’(b – bo)(b’ – bo’) + Σ Σ Kbb’(θ – θo)( θ’– θo’) (3.1) bonds bonds’ angles angles’ + Σ Σ Kbb’(b – bo)(θ – θo) bonds angles + Σ Σ Kbb’(b – bo)[KΦ,b1 cosΦ + KΦ, b2 cos2Φ + KΦ, b3 cos3Φ] bonds dihedrals + Σ Σ Kbb’(b’ – bo’)[KΦ,b’1 cosΦ + KΦ, b’2 cos2Φ + KΦ, b’3 cos3Φ] bonds’ dihedrals + Σ Σ Kbb’(θ – θo)[KΦ, θ1 cosΦ + KΦ, θ2 cos2Φ + KΦ, θ3 cos3Φ] angles dihedrals + Σ Σ Σ(θ – θo)(θ’ – θo’)cosΦ angles angles’ dihedrals + + * 9 * 6 Σ qiqj / rij Σ ε [ 2(r ij / r ij) – 3(r ij / r ij)] Coulombic van der Waals
The SAHase crystal structure was acquired from the Protein Databank with PDB
ID: 1A7A. The file contains a dimer of the SAHase structure. The A structure was used exclusively in these studies for consistency and accuracy. The A structure along with its
NADH, inhibitor and associated waters were unmerged/separated from the total structure.
Since the X-ray diffraction process does not visualize hydrogen atoms, the atoms were added to all of the molecules using a pH of 7.0, which was used to determine the charge state of the amino acid side chains. InsightII compares the assigned pH with a template
41 and assigns the amino acid charge state and related pKa’s of each perspective amino acid.
If the pH matches the template, the charge state and appropriate hydrogens are added.
Otherwise, the amino acid is assigned a neutral charge. The associated waters were then oriented through energy minimization using steepest descents and conjugate gradient algorithms. The protein was then “soaked” in a 5 Å layer of water so that the simulation would be performed in explicit solvent. The soaking was achieved by centering the protein in a box of equilibrated water and then removing water molecules that either contact the protein crystal structure or are greater than 5 Å away from the structure.
The inhibitor, located within the binding pocket, was then modified and the atom potentials were set using the CFF forcefield. The atom types were assigned according to the hybridization, and functional group or amino acid configuration. Partial and formal charges were assigned based on CFF’s atom types and table of charges. The amino acids within 6 Å of the inhibitor were defined as the binding site and were allowed to freely rotate during the docking process. Hydrogen bond donors and acceptor atoms were defined for both the inhibitor and the atoms within 5 Å of the inhibitor. The torsion angles for the inhibitor were then defined for the bonds that allow rotation of heavy atoms around a single bond.
A Monte Carlo simulation was used to generate the possible binding modes of the inhibitor-protein complex. The inhibitor was “energy minimized” using the conjugate gradient method and then non-bonding interactions were calculated using van der Waals interactions with the Coulombic interactions set to zero. This function was used for the initial search since it (i) minimized the possibility that Discover would terminate if there were many undesirable contacts and (ii) allowed for the approximate placement of the
42 inhibitor in the binding pocket. After the initial minimization, the inhibitor was randomly rotated and translated in the x, y, and z directions and the inhibitor torsion angles were rotated to generate a new inhibitor orientation.
The resulting structure was then minimized and the energy of the new structure was compared to the previous structure. If the energy of the new structure was lower than the energy of the previous structure, it was accepted but only if the energy was within a predefined energy range or the Boltzmann factor was greater than a random number between 0 and 1. The Boltzmann factor was calculated using the equation:
N N exp[(-(Vnew(r ) – Vold(r )) / kBT)] (3.2)
where V(rN) is the potential energy (V) of each position (rN) for N particles in the system, kB is the Boltzmann constant and T is the temperature. When the Boltzmann factor is used in the criteria for structure acceptance within a Monte Carlo simulation, it is known as the
Metropolis Monte Carlo (MMC).13,151 This calculation allows minute energy-well uphill moves in which the smaller the uphill move; the greater the probability the move will be accepted. Up to 20 structures were generated and the ten lowest energy structures were accepted for simulated annealing (SA).
SA was used to refine the ten best/lowest energy structures, and requires heating the complex to 500 K and then slowly cooling the system down to 300 K in 4 K intervals.
At each interval, the system was permitted to reach thermal equilibrium through the addition of small potential energies, which allowed the system to overcome small energy barriers in the energy well of the complex and to reach a global minimum.
43 The electrostatic interactions within the SA were calculated using the cell
multipole method.13,152-155 This method uses a box that contains all of the atoms in the
system and is divided into uniform cubic cells. The multipole moment is calculated for
each cell and is summed over all of the moments for all atoms in the cell. In the next step,
the interactions between each cell and the 26 cells surrounding each cell are calculated using Coulomb’s Law. The cells outside the initial 27 are calculated using the Taylor series multipole expansion. For efficiency, cells are grouped into larger and larger cells as the distance increases from the initial 27 cells. This allows the time required for the calculation to scale linearly with N rather than N2 with the standard Ewald method.13
The final step was calculation of the inhibitor-receptor interaction energies for each structure, which was accepted and minimized. The total energy of the inhibitor- protein complex resulted from calculating the interaction energy of the protein and the inhibitor separately and then determining the interaction energy between the two moieties. The interaction energy was then separated into the van der Waals and electrostatic contributions, which allows the energies to be compared to the different structures. Although the energies are comparable, the forcefield does not involve important desolvation terms and that the entire system is not allowed to minimize therefore the true minimum may never be reached.
IsoA Modeling
By incorporating the aspects of the carbocyclic sugar along with the N3 isoadenosine and replacement of either the N7 or the N9 heteroatoms with sulfur it was hoped a synergistic effect would result, thereby increasing the potency of the IsoA
44 nucleoside analogues. Recently discovered literature reported potent activity against
CTPase for several carbocyclic pyrimidines (Figure 3.1).156,157 The connection of the
base at the N-3 position enables the IsoA nucleoside analogues to act as both N-3
ribosylated purine and a 5,6-disubstituted pyrimidine. Given their structural resemblance to the 5,6-disubstituted pyrimidines and their intermediates, it was hoped the carbocyclic
IsoA targets would also prove to be inhibitors of SAHase and CTPase, thereby attacking two different enzymes in two different, but related enzymatic pathways. This would greatly increase the chances for complete shut down of viral or tumor cell replication against SAHase and DNA MeTase. Although this enzyme has not been associated with virus-specific enzymes, inhibition of the enzyme should suppress RNA synthesis, which would lead to antiviral and anticancer cytotoxicity.
Figure 3.1. Inhibitors of CTP synthetase.157
Discussion
To form a basis for comparison, Borchardt’s inhibitor, 4’,5’-unsaturated analogue
of 3-deaza-NpcA (Figure 3.2, Figure A.2, Table A.1 on the following page), which is
the most potent inhibitor of SAHase known to date was modeled. The energies of all
other nucleoside analogues, unless otherwise stated, were compared to it by setting the
45 energy for this inhibitor to 0.000 kcal/mol. Table 3.1 on page 58 contains the relative energies for all molecules in this chapter.
Figure 3.2. Borchardt’s inhibitor.
Beginning with the simulation of the natural substrate adenosine, a free energy
that was obtained was 17.735 kcal/mol lower than the standard (Figure 3.3 on the
following page). Simulations of the potent inhibitors Ari, NpcA and 3-deaza NpcA were
also performed. The nucleoside Ari had two free energies that were within 1 kcal/mol of
each other and were 1.288 and 1.431 kcal/mol higher in energy than Borchardt’s inhibitor
while NpcA and its 3-deaza counterpart exhibited a free energy that was 3.373 kcal/mol
and 0.575 kcal/mol lower than Borchardt’s inhibitor (Figure 3.3). Next, the unsaturated
nucleoside 4’,5’-enyl-adenosine was modeled. The free energy for this nucleoside was
0.685 kcal/mol higher than Borchardt’s inhibitor (Figure 3.3). These nucleosides were
modeled to provide a basis of comparison since these initial nucleosides have been
extensively tested and are known for being potent inhibitors of SAHase.
46
Figure 3.3. Initial parent nucleosides as compared to Borchardt’s.
After these initial modeling efforts, attention turned to the IsoA parent nucleoside.
Although, as previously stated, testing showed it to be unstable in acids and bases, it was still modeled to obtain a general idea of its free energy. Surprisingly, the IsoA nucleoside
(Figure 3.4 on the following page) was lower in free energy by 4.314 kcal/mol, indicating that it should be a better inhibitor than Borchardt’s nucleoside. Since this nucleoside would most likely be a substrate of the enzyme (since adenosine is the natural substrate and IsoA is an isomer of adenosine), these results were not particularly useful.
The modeling did however indicate that IsoA was higher in energy than the natural substrate adenosine and therefore should not bind as well as adenosine.
47
Figure 3.4. IsoA nucleoside as compared to Borchardt’s.
As was stated previously, to eliminate the inherent instability of the IsoA nucleoside, the N3 nitrogen was removed to form a “C” nucleoside. The simulation of the
3-deaza-IsoA nucleoside would have to consider the two possible tautomers, which would result in the protonation of the N7 or the N9 atoms (Figure 3.5). The N7H
analogue (IsoA#1) was 7.070 kcal/mol lower in energy than the standard while the NH9
analogue produced two structures (IsoA#2 and IsoA#3) that contained energies that were
relatively close to one another. Both of the N9H structures were higher in energy by
6.118 and 6.768 kcal/mol from Borchardt’s inhibitor.
Figure 3.5. 3-Deaza-IsoA nucleosides as compared to Borchardt’s.
48 Next, the carbocyclic analogue of IsoA was modeled (IsoA#4). The free energy of this analogue was 4.389 kcal/mol lower than the standard (Figure 3.6). The Ari form of the 3-deaza-IsoAwas then modeled and analogous to the former 3-deaza nucleoside, this also contained N7H (IsoA#5) and N9H (IsoA#6) tautomers. These nucleosides exhibited free energies that were 1.925 and 3.143 kcal/mol lower than the standard respectively
(Figure 3.6).
The next step in the process was the simulation of NpcA analogues. The IsoA form of NpcA (IsoA#7) displayed an energy that was 1.074 kcal/mol lower in energy from the standard, while the two 3-deaza analogues, N7H (IsoA#8) and N9H (IsoA#9), displayed free energies that were 8.177 and 3.264 kcal/mol lower (Figure 3.6).
Figure 3.6. Ari and NpcA analogues as compared to Borchardt’s.
49 After running simulations on the Ari and NpcA analogues, the hydroxy methyl
group was removed from the 4’ position of the carbocyclic moiety to form the 4’,5’
tetrahydro analogue (IsoA#10). The initial inhibitor displayed an energy 1.504 kcal/mol
higher than Borchardt’s inhibitor (Figure 3.7). Results from the 3-deaza analogues
indicated that the theoretical free energies for these two inhibitors were higher than the
standard by 2.208 and 8.589 kcal/mol for the N7H (IsoA#11) and N9H (IsoA#12)
respectively (Figure 3.7).
Finally, the unsaturated 4’,5’-enyl IsoA analogue (IsoA#13) was investigated.
The free energy was lower than the standard by 5.142 kcal/mol and the corresponding 3-
deaza analogues were lower by 2.032 and 1.311 kcal/mol for the N7H (IsoA#14) and the
N9H (IsoA#15) molecules respectively (Figure 3.7).
Figure 3.7. Tetrahydro and enyl IsoA analogues as compared to Borchardt’s.
50 Once the adenosine derivatives were completed, a few simulations on the IsoG
and IsoX analogues were performed to help provide a basis of comparison for the
heteroatom substitution in the bases. Since these IsoG and IsoX analogues also exist as
tautomers, there were additional inhibitor structures that had to be simulated for each
target.
Of the IsoG analogues simulated, the first two inhibitors were the 4’,5’-tetrahydro
IsoGs, N7H and N9H. The N7H (IsoA#16) inhibitor had a lower free energy by 1.005
kcal/mol and the N9H (IsoA#17) had a higher free energy by 1.581 kcal/mol than
Borchardt’s inhibitor (Figure 3.8). The remaining three IsoG inhibitors were the 4’,5’-
enyl analogues, one N7H (IsoA#18) and two N9H structures (IsoA#19 and IsoA#20) with the N9H structures having energies within 1 kcal/mol of each other. All of these inhibitors have lower energies than the standard by 10.728, 7.338 and 7.995 kcal/mol respectively (Figure 3.8).
The IsoX 4’,5’-tetrahydro IsoXs, N7H and N9H were also modeled. The N7H
(IsoA#21) inhibitor had a lower free energy by 2.913 kcal/mol and the N9H (IsoA#22) had a higher free energy by 2.090 kcal/mol than Borchardt’s inhibitor (Figure 3.8 on the following page). The remaining two IsoX inhibitors were the 4’,5’-enyl analogues, the
N7H (IsoA#23) and N9H (IsoA#24) structures. These two inhibitors had lower theoretical energies than the standard by 9.164 and 24.357 kcal/mol respectively (Figure
3.8). Interestingly, the lowest free energy of all the potential inhibitors in this study was observed for IsoA#24, which is remarkable since the enzyme should only recognize
“adenosine like” molecules.
51
Figure 3.8. IsoG and IsoX analogues as compared to Borchardt’s.
Once the initial study was completed, focus shifted to the targets containing
heteroatom substitution in the bases, replacing either the N7 or N9 of 3-deazaAri with a
sulfur atom. Three structures were produced from the simulation, the first two were the
S7 (IsoA#25 and IsoA#26) structure, which had relatively the same energies and were
both higher in energy than the standard by 4.923 and 4.949 kcal/mol (Figure 3.9). The S9
structure (IsoA#27) was also higher in free energy by 4.338 kcal/mol. Comparison of the
“natural” analogues to the thio analogue correspond to when the sulfur is in the 7 position of the nucleoside the NH in the “natural” analogue is also in the 7 position. Comparison of the thio-substituted analogues with the unsubstituted nucleosides showed that the thio analogues were higher energy than their normal counterparts. The energy difference
between IsoA#5 and IsoA#25 was 6.848 kcal/mol and the difference between IsoA#6
52 and IsoA#27 was 7.481 kcal/mol. This implies that these thio analogues would not be
better than their unsubstituted equivalent.
The next step was to simulate the NpcA analogues. The S7 (IsoA#28) and S9
(IsoA#29) were both lower in energy than the standard by 5.205 and 2.062 kcal/mol
respectively (Figure 3.9) and they were only slightly higher in energy than IsoA#8 and
IsoA#9 by 2.972 and 1.202 kcal/mol.
Figure 3.9. Thio substituted Ari and NpcA analogues as compared to Borchardt’s.
Attention turned then to both the 4’,5’-tetrahydro and the 4’,5’-enyl analogues.
The 4’,5’-tetrahydro analogues had four structures, three of which are S9 possessed energies within 1 kcal/mol. The differences in free energies for the S7 and the S9 from the standard were 6.028 kcal/mol higher for the S7 (IsoA#30) and 11.903, 12.053 and
53 12.683 kcal/mol higher for the S9 (IsoA#31 #32 and #33) (Figure 3.10). These analogues were also higher in energy compared to the analogues IsoA#11 and IsoA#12 by 3.820 and 3.314 kcal/mol for IsoA#30 and IsoA#31 respectively. The 4’,5’-enyl analogues were both higher in energy than the standard with the S7 (IsoA#34) being 0.932 kcal/mol
higher and the S9 (IsoA#35) was 4.142 kcal/mol higher (Figure 3.10). These analogues
were also higher in free energy compared to the analogues IsoA#14 and IsoA#15 by
2.946 and 5.453 kcal/mol.
Figure 3.10. Thio 4’,5’-tetrahydro and 4’,5’-enyl IsoA analogues as compared to Borchardt’s.
Next, the IsoG analogues were modeled starting with the Ari analogue form.
These analogues displayed energies of 5.307 kcal/mol below for the S7 (IsoA#36) and
9.190 kcal/mol below the standard free energy for the S9 (IsoA#37) (Figure 3.11 on page
54 56). The NpcA analogues both displayed free energies that were lower than the standard.
The S7 (IsoA#38) was 10.916 kcal/mol lower and the S9 (IsoA#39) was 13.237 kcal/mol lower (Figure 3.11). As before, the 4’,5’-tetrahydro and the 4’,5’-enyl analogues were next. The 4’,5’-tetrahydro analogues gave free energies that were either above or below
Borchardt’s inhibitor energy; the S7 (IsoA#40) being 0.282 kcal/mol lower and the S9
(IsoA#41) being 1.703 higher (Figure 3.11). The comparison between these analogues and the non-heteroatom substituted IsoGs indicated that thio analogues were only slightly higher in energy than IsoA#16 and IsoA#17 by 0.223 and 0.122 kcal/mol. In the 4’,5’- enyl analogues, the S7 (IsoA#42) was 1.906 kcal/mol higher and the S9 (IsoA#43) being
9.673 kcal/mol higher (Figure 3.11 on the next page). These analogues were significantly higher in energy than IsoA#18 and IsoA#20 equivalents by 21.634 and 17.668 kcal/mol.
Finally, simulations were performed on the xanthine heteroatom substituted analogues starting with the Ari analogues. The S7 (IsoA#44) analogue had a free energy that was lower than the standard by 1.855 kcal/mol whereas the S9 (IsoA#45) had a free energy that was higher than the standard by 2.204 kcal/mol (Figure 3.12) of the three
NpcA analogues, one for S7 and two for S9, the free energy of the S7 (IsoA#46) was
higher than Borchardt’s inhibitor by 3.594 kcal/mol while the two S9 (IsoA#47 and
IsoA#48) analogues were both lower by 7.316 and 7.676 kcal/mol (Figure 3.12).
The next analogues were of the 4’,5’-tetrahydro and the 4’,5’-enyl analogues.
Following the same pattern, the 4’,5’-tetrahydro analogues displayed a higher free energy for the S7 (IsoA#49) than the standard by 1.492 kcal/mol and a lower free energy for the
S9 (IsoA#50) by 2.528 kcal/mol (Figure 3.12). Upon further investigation, the S7 analogue was higher in energy by 4.405 kcal/mol than IsoA#21 as seen with other thio
55 analogues, however, the S9 analogue was actually lower by 4.618 kcal/mol than
IsoA#22. The 4’,5’-enyl analogues were both lower in energy than the standard by 2.218
kcal/mol for the S7 (IsoA#51) and 1.711 kcal/mol for the S9 (IsoA#52) (Figure 3.12).
Similar to other thio analogues, these were also higher in energy than the non-substituted
equivalent compounds. IsoA#51 was higher by 6.946 kcal/mol than IsoA#23 and
IsoA#52 was considerably higher by 22.646 kcal/mol than IsoA#24.
Figure 3.11. Thio IsoG analogues as compared to Borchardt’s.
The analogues 4’,5’-tetrahydrocytosine and the 4’,5’-enylcytosine were also looked at due to the finding in the literature of CTP synthetase activity for similar nucleosides. These analogues, although not purine nucleosides, are carbocyclic based and therefore maybe potent inhibitors of SAHase. The results indicated both nucleosides were
56 lower in free energy compared to the standard with the 4’,5’-tetrahydro (IsoA#53) 9.256 kcal/mol lower and the 4’,5’-enyl (IsoA#54) 4.903 kcal/mol lower (Figure 3.13).
Figure 3.12. Thio IsoX analogues as compared to Borchardt’s.
Figure 3.13. Cytosine analogues as compared to Borchardt’s.
57 Table 3.1. Relative energies of IsoA analogues. Molecule Energy Molecule Energy Name/number (kcal/mol) Name/number (kcal/mol) Borchardt's inhibitor 0.000 IsoA#24 -24.357 Ado -17.735 IsoA#25 4.923 Ari 1 0.173 IsoA#26 4.949 Ari 2 0.856 IsoA#27 4.338 NpcA -3.373 IsoA#28 -5.205 3-Deaza-NpcA -0.575 IsoA#29 -2.062 4',5'-Enyl-adenosine 0.685 IsoA#30 6.028 IsoA -4.314 IsoA#31 11.903 IsoA#1 -7.070 IsoA#32 12.053 IsoA#2 6.118 IsoA#33 12.683 IsoA#3 6.768 IsoA#34 0.932 IsoA#4 -4.389 IsoA#35 4.142 IsoA#5 -1.925 IsoA#36 -5.307 IsoA#6 -3.143 IsoA#37 -9.190 IsoA#7 -1.074 IsoA#38 -10.916 IsoA#8 -8.177 IsoA#39 -13.237 IsoA#9 -3.264 IsoA#40 -0.282 IsoA#10 1.504 IsoA#41 1.703 IsoA#11 2.208 IsoA#42 1.906 IsoA#12 8.589 IsoA#43 9.673 IsoA#13 -5.142 IsoA#44 -1.855 IsoA#14 -2.032 IsoA#45 2.204 IsoA#15 -1.311 IsoA#46 3.594 IsoA#16 -1.005 IsoA#47 -7.316 IsoA#17 1.581 IsoA#48 -7.676 IsoA#18 -10.728 IsoA#49 1.492 IsoA#19 -7.338 IsoA#50 -2.528 IsoA#20 -7.995 IsoA#51 -2.218 IsoA#21 -2.913 IsoA#52 -1.711 IsoA#22 2.090 IsoA#53 -9.256 IsoA#23 -9.164 IsoA#54 -4.903
58 Conclusion and Future Directions
As shown in the majority of the nucleoside analogues modeled, the presence of an
exocyclic amine in the C6 position and proton donor at the N7, lowered the binding free
energy hence an increase in binding. However, when a proton donor was not present, the
binding free energy was higher, resulting in poor binding. When a carbonyl was in the C6
position, the reverse was true. For the thiazole analogues, the amine in the C6 position
and the sulfur in the 7 position of the purine lowered the binding energy. This was
probably due to an interaction between the hydrogen bond donor amine and the hydrogen
bond acceptor N7, which would lower the overall hydrogen bond donating/accepting
strength of those heteroatoms. Since the sulfur has lower electron density than the
nitrogen, it would therefore have less interaction with the exocyclic amine.
Comparison of the Ari, NpcA and 3-deaza NpcA binding energies to the actual
potency of the analogues (on page 37) showed no indications that the free energy was
directly related to the potency of the analogues. However, one must note that there are
more factors involved than just free energy to binding. These factors include hydrophobic
interactions, spatial arrangement and steric factors. Also, these simulations do no allow
free movement for the entire enzyme, which is essential for accurate simulations and
evaluation. Due to the cost of protracted simulation time, these models must be restricted
in their movement. This restriction limits the accuracy of the modeling however, making
the evaluation of the data more subjective. Also, all forcefields contain limitations to their accuracy. Examples of these errors can be found in reference 15. Forcefields are usually
designed for specific functional groups and classes of compounds and are parameterized
59 accordingly. Therefore, a full understanding of forcefield limits should be understood before applying it to a calculation.
These theoretical studies indicated that all of the thio analogues except for the S9-
xanthine analogue should not be better inhibitors than their unsubstituted equivalent
nucleosides, however, experimental data may indeed contradict these findings.
Experimental methods
General. All potential inhibitors were docked into the SAHase crystal structure
(PDB 1A7A) and calculations performed on a Silicone Graphics Octane2 workstation running the IRIX 64 Release 6.5 operating system using Accelrys® Inc. InsightII (version
2000) and dynamics calculations were performed with the Discover (version 9.2X) program using the CFF forcefield, which is derived from ab initio calculations on the
Hartree-Fock level of theory using the 6-31G* basis set. The inhibitors were optimized using steepest descent calculations. Next, using Monte Carlo techniques, up to twenty spatial conformers (or more if necessary) were generated for each inhibitor. The ten lowest energy conformations were then minimized using simulated annealing techniques.
The final structures were analyzed for the best ligand structure with the lowest energy value. The raw data is shown in Appendix A.
60 CHAPTER 4
MOLECULAR MODELING OF FLEXIMERS AS BIOPROBES FOR FLEXIBLE ENZYMES
Background and Significance
During the course of the previous investigation it was discovered that SAHase is fairly flexible and exhibits a large difference between the “open” and “closed” conformations,158 thereby representing a dilemma when considering molecular modeling.
As a possible solution to this problem, a series of novel flexible nucleosides termed fleximers were designed (Figure 4.1).159-161
Figure 4.1. Seley Fleximers.159-161
61 A single carbon-carbon bond separates the pyrimidine and imidazole rings of
these compounds allowing for torsional flexibility and additional degrees of freedom in
the base. The labeling followed a similar as was used for Leonard’s tricyclic nucleosides,
and the numbering can be seen in Figure 4.2. Where the imidazole C-5 is attached to the
pyrimidine C-6, the fleximer is termed as a distal (dist-) fleximer and when the imidazole
C-4 is attached to the pyrimidine C-5, it is termed as a proximal (prox-) fleximer. This
added flexibility would allow the enzyme and the nucleoside to each adjust to the lowest
energy conformation possible, therefore (theoretically) providing the optimal conformation of the substrate and enzyme complex.
Figure 4.2. Fleximer base numbering for dist- and prox-.
It has also been reported that flexibility is extremely important to recognition and
helps to lower the entropic barrier towards complex formation.162,163 Structural studies
have shown that inhibitor flexibility helps analogues adapt to changes in the binding
pocket and enables the compounds (Figure 4.3) to “wiggle” and “jiggle”.164 This alteration appears to be critical for potency against wild type and mutant drug resistant
HIV-1 reverse transcriptase enzymes. 164 Recently the nucleoside tenofovir (Figure 4.3) has been shown to inhibit nucleoside reverse transcriptase.165 The flexibility of the nucleoside analogue allows the molecule to be “slippery” and elude the two general
62 pathways of drug resistance (i) discrimination at incorporation and (ii) excision from the primer terminus after incorporation.165
Figure 4.3. Flexible Reverse Transcriptase Inhibitors.164,165
The binding of any substrate to an enzyme consists of many enthalpic and entropic contributions.162 The entropic terms consist of desolvation and rigidification of residues in a ligand-enzyme complex. There is also loss of global rotational and translational degrees of freedom in the biomolecular association to form the complex. If there is an overall difference between binding energies, it can most likely be attributed to the entropy associated with internal bond rotations. In a test of flexible inhibitors, the binding affinities increased as the conformational flexibility decreased indicating that flexibility assists in increasing binding affinity.162
Flexible Nucleosides
To date, only a few flexible nucleosides have been synthesized and investigated.
One of the earliest were 5-substituted cytidine and uridine carbocyclic nucleosides designed by Gronowitz and coworkers (Figure 4.4).166 These nucleoside analogues displayed low antiviral activity towards HSV-1. Herdewijn and coworkers designed some
63 analogous 5-substituted ribose analogues related to the carbocyclics reported by
Gronowitz (Figure 4.4). These nucleosides also exhibited potent activity against
HSV.167,168 Computational studies on these inhibitors were also carried out and the findings correlated with the experimental results.168,169
Figure 4.4. Uridine and cytidine flexible analogues. 166,168,169
Flexible nucleosides have also been used to probe triplex formation with DNA
through Hoogsteen hydrogen bonding (Figure 4.5).170-172 Triplex formation is mostly
restricted to pyrimidine bases that hydrogen bond to a purine base tract of a double helix.
Any interruption by a pyrimidine base in the double helix disrupts the stability of triplex
formation. The flexible nucleosides were designed to extend recognition to all four base
pairs by giving greater access to both bases.
64 Despite the successes of these early examples, it was notable that none of the
flexible nucleosides reported to date really resembled purine nucleosides. As an answer
to this, a series of flexible nucleosides were designed and synthesized that more closely resembled the normal purine scaffold as shown in Figure 4.1 on page 61.
Figure 4.5. DNA triplex probe.170-172
The distal adenosine, inosine and guanosine fleximers were successfully
synthesized and (Figure 4.1), were assayed against SAHase.35 The results showed that
the adenosine and inosine analogues were poor substrates for the enzyme whereas the
guanosine analogue unexpectedly showed inhibitory activity. This was quite surprising
since no guanosine analogue has ever been reported to have shown activity against
SAHase, an adenosine metabolizing enzyme.
Preliminary gas phase ab initio computational studies on the fleximers indicated
that there was minimal hindrance for rotation to occur between the imidazole-pyrimidine
moieties.160,173 The investigation showed that the bases formed a planar arrangement
when modeled without the ribose sugar. However when modeled with the sugar, the
rotation barrier was reduced and the bases preferred a non-planar orientation with respect to each purine moiety. Another observation was the orientation of the base around the
65 glycosidic bond; the calculations showed a strong tendency towards the anti
conformation. In addition, there was also a shift in sugar puckering from south to more
north.174 Another study also indicated that the unrestricted distal fleximers prefer the anti
conformation but the guanosine fleximer formed an intramolecular hydrogen bond
between the O2’ and the pyrimidine exocyclic amine locking the base in the anti conformation.35
Computing Flexibility
Recently, more powerful and faster computers have become available to address
the issue of the protracted computational time needed to calculate protein flexibility,
although there are still problems dealing with the number of compounds to be tested.175
Many proteins and enzymes are flexible and therefore cannot be described by a single crystal structure and it has been found that many active sites contain regions that have low structural stability, which translates into high flexibility.
The energy landscape of most proteins resembles a folding funnel where many unfavorable states collapse from multiple routes into several favorable states, and these states are condition-dependent.23 Conditions such as pH, temperature, or introduction of a
new ligand can shift the minima and change the most populated conformational state.
Since enzymes exist in a range of conformations, a ligand will preferentially bind to one
of those states or even multiple states thereby increasing the need to predict all of those
states. Therefore, a single crystal structure may not be sufficient to describe the activity
of the binding complex.
66 Initial attempts at flexible modeling were through the use of soft docking. To
achieve soft docking, multiple crystal structures of the same enzyme with different enzyme-inhibitor complexes are overlaid to create a composite binding site.23 This describes the most available space to receptor molecules, which are docked and evaluated. The process is simple and quick and leads to potent inhibitors.23 In a
subsequent study, free rotational movement of hydrogen bond donors and acceptors was permitted thereby decreasing the rigidity of some side chains.23 Although the free
rotations were limited to mostly torsional, the computation was too intensive for the
movement by all side chains.
The best methods for quality Monte Carlo (MC) or Molecular Dynamics (MD)
simulations use unrestricted energy calculations, however they are significantly slower
than previous calculations.23 To reduce the time required for these calculations, most of
the molecule, with the exception of the binding site, is held rigid, however this still limits
the conformational sampling.
In a study of protein flexibility, amino acids were checked for extreme
conformational changes, which can be described as angular movements greater than 60
degrees upon binding.176 The study found that the probability of these movements varied
from 1% for phenylalanine to 38% for lysine and that for eight other residues had a
greater than 10% chance of movement. Comparison of different crystal structures of the
same protein showed that 10% of buried side chains moved by 11 degrees or more.
Furthermore, these studies revealed that 40% of all side chains had a 10-40% chance of
large side chain movement, thereby strongly suggesting that a single crystal structure is
an inadequate representation of enzyme structure.
67 With that thought in mind, a series of purine fleximers were modeled in the
binding site of SAHase. As Seley put it,
“The design of better medicinal agents relies heavily on the understanding of structural interactions between enzymes and their substrates or cofactors. Purines are the most ubiquitous heterocyclic ring systems in nature; they are contained in numerous biologically significant molecules and therefore provide an excellent scaffold for constructing bioprobes.”159
The heterocyclic base was designed to be flexible so that both the nucleoside and the
enzyme could achieve the tightest binding lowest energy conformation possible.
The fleximers were modeled in the binding site of SAHase, since as previously
stated, this biologically significant enzyme is of particular interest because it is an important chemotherapeutic target. It is commercially available from several sources, easy to purify and crystallize, and as stated before, was recently shown to be flexible.
Discussion
The initial modeling on the parent fleximer nucleosides (Figure 4.1 on page 61) was carried out to form a basis for comparison. Borchardt’s inhibitor, 4’,5’-unsaturated analogue of neplanocin A (Figure 3.2, Figure A.2, Table A.1 on pages 46, 174, 174
respectively), which is known to be the most potent inhibitor of SAHase was also
modeled. For comparison, the energy of Borchardt’s nucleoside was set to zero so that it
would be possible to compare the energies of all other nucleoside analogues to it, unless
otherwise stated. Table 4.3 on page 78 contains the relative energies for all molecules in this chapter.
The parent distal adenosine (Flex#1), inosine (Flex#2) and guanosine (Flex#3 and
Flex#4) fleximers (Figure 4.6 on the following page) all exhibited lower energy values
68 than the standard (Borchardt’s). Surprisingly, the inosine nucleoside exhibited the lowest
energy of –22.461 kcal/mol while the guanosine fleximer was second lowest, –18.380
kcal/mol. This was unexpected since the enzyme normally recognizes only adenosine-
based nucleosides. The adenosine fleximer was lower only by 3.739 kcal/mol from the
standard and comparison of Flex#1 to the modeled adenosine showed that the fleximer
was higher in energy by 13.996 kcal/mol than the normal parent nucleoside, and therefore should not bind as effectively as the normal nucleoside.
Figure 4.6. Parent distal fleximers as compared to Borchardt’s.
After synthesis and subsequent screening of the fleximers,35,159,160 it was found
that the adenosine fleximer was a reasonable substrate for the enzyme, albeit as expected
worse than adenosine itself Table 4.1. The inosine was also a poor substrate but also
exhibited weak inhibitory activity Table 4.1. It was also discovered that the guanosine
fleximer displayed inhibitory activity against SAHase in both the synthetic and hydrolytic
35 activity directions Table 4.1. The Ki values of 217 and 81 µM, respectively, are not
outstanding since aristeromycin and neplanocin A have Ki values in the nanomolar range
against SAHase. They are, however, notable since, to our knowledge, no guanosine
nucleosides have ever been reported to exhibit activity against SAHase.
69 From the modeling results, it was realized that the pyrimidine ring of dist-
guanosine twisted thereby reversing the substituents on the ring. This could give rise to a more “adenosine-like” nucleoside, which would then be recognized by SAHase (Figure
4.7). Although the theoretical energy of Flex#4 was higher in energy by 9.967 kcal/mol
than Flex#3, it was theorized that this change in conformation could be the most
plausible solution to the activity exhibited by this nucleoside analogue. Studies are
currently underway with crystallization of the guanosine fleximer in SAHase.
Table 4.1. Results of the enzyme essays with SAHase, parent nucleosides and the distal fleximers.35 -1 Nucleoside Km (µM) Kcat (m ) Ki (µM) Synthesis Direction Adenosine 0.82 ± 0.004 91.2 ± 0.2 nd Adenosine Fleximer 54.4 ± 1.1 0.56 ± 0.04 nd Inosine 2.5 ± 0.03 44.2 ± 0.7 925 ± 21 Inosine Fleximer 421 ± 5.7 0.005 ± 0.003 422 ± 16 Guanosine 844 ± 6.6 0.06 ± 0.02 1472 ± 32 Guanosine Fleximer ns ns 217 ± 13 Hydrolysis Direction: G-fleximera 6.6 ± 0.2 28.3 ± 1.2 128 ± 18 G-fleximerb 7.9 ± 0.2 72 81 ± 13 a Rabbit SAHase, sigma; b Human SAHase, Howell laboratories; nd- not determined; ns- not a substrate
With this result in mind, it was then theorized that isoguanosine, a tautomer of
guanosine, might bind more favorably than guanosine, since the exocyclic amine would
already be in the correct position for interaction with the enzyme. Since there are two
isomers of isoguanosine, the N3H (Flex#5) and the N1H (Flex#6), simulations were
performed on both. The results indicated that Flex#5 was bound more effectively than the
Flex#6 by 18.850 kcal/mol (Figure 4.8 on page 72). This could be due to resonance
70 stabilization since the exocyclic amine is able to delocalize an electron pair into the ring while the N1 nitrogen can abstract a proton from either water or the enzyme to stabilize the negative charge. Unfortunately, the modeling results showed that the pyrimidine ring of the IsoG was inverted therefore positioning the exocyclic groups in the orientation of the normal guanosine. Despite this, it is still possible that the inverted orientation of the guanosine is the reason for the unusual activity observed. Synthesis of the IsoG is underway.
Figure 4.7. Flex#4 guanosine fleximer.
Next, in order to complete the theoretical SAR study, simulation of the distal xanthine fleximer (Flex#7) was initiated, and results showed that it was only slightly lower in energy than the standard, -1.508 kcal/mol (Figure 4.8). This was not surprising though since the xanthine analogue does not contain an exocyclic amine and therefore should not be readily recognized by the enzyme. Comparison of the inosine (Flex#2) and
71 the xanthine analogue (Flex#7), showed that the inosine analogue was much lower in energy than the xanthine analogue. The 2-aminoadenosine derivative (Flex#8) was modeled and found to be 6.123 kcal/mol lower in energy than Borchardt’s inhibitor
(Figure 4.8). This series of simulations indicated that the pyrimidine exocyclic amine group is critical for proper binding and activity while a carbonyl is beneficial in the normal pyrimidine 2 position only. The differences in binding could be due to the increased sterics of the nucleoside or the difference in charge distribution of the
pyrimidine moiety, but stacking interactions showed also be taken into consideration. As
a result, it will be interesting to see the actual binding energies of these inhibitors once
the biological studies have been completed.
Figure 4.8. Distal N1H and N3H isoguanosines, xanthine and 2-aminoadenosine fleximers as compared to Borchardt’s.
72 The next inhibitor modeled was the 3-deazaadenosine fleximer (Flex#9). The
nucleoside exhibited a theoretical free binding energy 8.016 kcal/mol below the standard
(Figure 4.9). Attention then turned to the carbocyclic analogues. The fleximer distal Ari
analogue (Flex#10) had a free binding energy 21.627 kcal/mol lower than the standard
Figure 4.9 and was 22.483 kcal/mol lower in free energy than the modeled Ari
nucleoside (IsoA#3).
Figure 4.9. Distal analogues as compared to Borchardt’s.
The distal fleximer form of 4’,5’-enyl-adenosine (Flex#11) was 22.348 kcal/mol lower than the standard Figure 4.9 and was 23.033 kcal/mol lower than 4’,5’-enyl-
adenosine (IsoA#7). The distal carbocyclic form of guanosine (Flex#12) was also
modeled. This inhibitor displayed a free energy that was 24.745 kcal/mol lower than the
73 standard (Figure 4.9) and was 6.365 kcal/mol and 16.332 kcal/mol lower than
guanosines Flex#3 and Flex#4 (Figure 4.6) respectively.
Following on the distal leads, the corresponding proximal fleximers were then synthesized and modeled.161 These also displayed theoretical lower energies than the
standard. The lowest of this series was a proximal isoguanosine and as before, both the
N3H (Flex#13) and the N1H (Flex#14) tautomers were investigated. The lowest energy isomer provided to be Flex#14 (Figure 4.10). and was 1.382 kcal/mol lower than the corresponding lowest energy distal counterpart Flex#5 (Figure 4.10).
Figure 4.10. Proximal IsoGs and distal IsoG N3H as compared to Borchardt’s.161
In some cases, however, the proximal fleximers were more favorable in free
energy than their distal counterparts. The proximal adenosine (Flex#15) was 5.804
kcal/mol lower in free energy (Figure 4.11) than the standard and 2.065 kcal/mol lower
in energy than Flex#1 (Figure 4.6). Also, a comparison against the normal adenosine
nucleoside proved that Flex#15 was higher in energy by 11.931 kcal/mol. As a result,
Flex#15 can theoretically be expected to bind less favorably to SAHase than the normal adenosine nucleoside, but not as poorly as its distal isomer Flex#1 (Figure 4.6). The
74 proximal xanthine (Flex#16) was 16.470 kcal/mol lower in energy than the standard
(Figure 4.11) and was also 14.962 kcal/mol lower than its distal isomer Flex#7 (Figure
3.8). The proximal guanosine (Flex#17) was 17.686 kcal/mol lower in energy than the standard (Figure 4.11) but was higher in energy than Flex#3 (Figure 4.6) by 0.694 kcal/mol but lower than Flex#4 (Figure 4.6) by 9.273 kcal/mol. The proximal 2,6- diamino provided two structures that were almost identical in energy; Flex#18 and
Flex#19 were 9.872 and 9.878 kcal/mol lower in free energy than the standard (Figure
4.11) and were 3.749 kcal/mol lower than distal 2-aminoadenosine Flex#8 (Figure 4.8).
Figure 4.11. Proximal parent analogues as compared to Borchardt’s.161
75 The proximal carbocyclic guanosine (Flex#20) was also modeled (Figure 4.11).
This molecule was 13.772 kcal/mol lower in free energy than the standard, but was
10.973 kcal/mol higher in energy compared to Flex#12 its corresponding distal isomer
(Figure 4.9 on page 73).
Figure 4.12. Sugar moiety alignment of the distal and proximal guanosine and isoguanosine analogues.
All of the fleximers can obtain a variety of conformations; however, the
conformation is dependent on the sterics and donor/acceptor properties of the functional
groups as well as the stacking and hydrogen binding interactions with the enzyme. The
range of conformations for the fleximers can be seen from the overlap of the C5’, C3’
and O3’ atoms for the distal and proximal guanosine and isoguanosine analogues (Figure
76 4.12). These analogues also exhibit a variety of conformations in the binding site of the
enzyme. The distal and proximal guanosine and isoguanosine analogues have been oriented according to their interaction with specific amino acid residues (Figure 4.13).
The amino acid residues are removed for clarity. The aligned residues are Glu 57, Thr60,
Asp131, Thr157 and Asn346. The lowest energy conformers have been highlighted for clarity. Results from experimental studies are anticipated, which should elucidate the actual conformation adopted by the inhibitors in the enzyme.
Figure 4.13. Amino acid residue alignment of the distal and proximal guanosine and isoguanosine analogues.
77 After the initial simulations, the distal analogues were revisited. Upon splitting the imidazole and pyrimidine moieties of the purine ring system, two new points of attachment for additional functional groups were created. Introduction of either hydrogen bond acceptors or donors, electron donating or withdrawing groups was done to determine whether it would be possible to “tune” the dihedral angle (Appendix B,
Flex#21 through Flex#34 pages 248 through 261). This study was designed to investigate the change in the dihedral angles between the imidazole and pyrimidine rings.
As observed in Figure 4.14 and Table 4.2, a full range of dihedral angles rotating from
180º to -180º could be achieved. This shows that the rotation between imidazole and pyrimidine can indeed be “tuned”.
Table 4.2. Change in dihedral angle. Nucleoside Dihedral angle Flex#21 A, X = NH2, Y = H. 106.82 Flex#22 A, X = NH2, Y = H. 126.16 Flex#23 A, X = H, Y = NH2. -129.90 Flex#24 A, X = CHO, Y = H. -109.49 Flex#25 A, X = H, Y = CHO. 92.57 Flex#26 A, X = CHO, Y = CHO. 133.08 Flex#27 A, X = OH, Y = H. 154.54 Flex#28 A, X = H, Y = OH. -88.82 Flex#29 A, X = CH3, Y = H. 85.85 Flex#30 A, X = H, Y = CH3. 59.26 Flex#31 A, X = CH3, Y = CH3 -116.83 Flex#32 G, X = H, Y = OH , Z = O. 27.80 Flex#33 GT, X = H, Y = OH. -14.67 Flex#34 G, X = H, Y = OH, Z = CH2. -29.75
78
Figure 4.14. Dihedral structures for Table 4.2.
Table 4.3. Relative energies of Fleximer analogues. Molecule Energy Molecule Energy Name/number (kcal/mol) Name/number (kcal/mol) Borchardt's inhibitor 0.000 Flex#11 -22.348 Flex#1 -3.739 Flex#12 -24.745 Flex#2 -22.461 Flex#13 -7.290 Flex#3 -18.380 Flex#14 -30.381 Flex#4 -8.413 Flex#15 -5.804 Flex#5 -28.999 Flex#16 -17.686 Flex#6 -10.908 Flex#17 -16.407 Flex#7 -1.508 Flex#18 -9.872 Flex#8 -6.123 Flex#19 -9.878 Flex#9 -8.016 Flex#20 -13.772 Flex#10 -21.627
Conclusion
The modeling efforts have shown that these nucleoside analogues are extremely
flexible in their ability to adjust to a variety of conformations. These new analogues, as
shown through several studies, have low barriers to rotation and are able to adopt lower
energy conformations and orient themselves to adopt new hydrogen bonding patterns.
Even though these analogues theoretically show free energies lower than that
exhibited by Borchardt’s inhibitor, it must be understood that the lowest energy conformation may not necessarily be the species that is responsible for activity. Overall,
79 the fleximers were designed to probe enzyme-ligand recognition, but more importantly
provide the capability for overcoming resistance mechanisms for potential inhibitors. As
results of the biological testing are obtained, it will be very interesting to see how the
results of these computational studies compare to reality.
Experimental
General. All potential inhibitors were docked into the SAHase crystal structure
(PDB 1A7A) and calculations performed on a Silicone Graphics Octane2 workstation running the IRIX 64 Release 6.5 operating system using Accelrys® Inc. InsightII (version
2000) and dynamics calculations were performed with the Discover (version 9.2X) program using the CFF forcefield, which is derived from ab initio calculations on the
Hartree-Fock level of theory using the 6-311G* basis set. The inhibitors were optimized using steepest descent calculations. Then using Monte Carlo techniques, up to twenty spatial conformers or more if necessary were generated for each inhibitor. The ten lowest energy conformations were then minimized using simulated annealing techniques. The final structures were analyzed for the best ligand structure with the lowest energy value.
80 CHAPTER 5
MOLECULAR DYNAMICS SIMULATIONS OF EXTENDED BASES IN DNA
Background and Significance
The focus for the ongoing research in our laboratories has involved the design and synthesis of structurally modified nucleosides and nucleotides to study structure and function in biologically significant systems. The work described in this chapter describes the computational investigation of a series of expanded nucleosides where a five- membered heterocycle is inserted between the imidazole and pyrimidine moieties of a purine base (Figure 5.1). This structural modification was incorporated to allow the nucleosides to retain their basic motif while providing an additional hydrogen bond donor or hydrogen bond acceptor to the purine scaffold. It was our hypothesis that the inclusion of a heteroaromatic spacer ring would endow the purine scaffold with several significant advantages over other modified nucleosides previously studied in DNA.
Figure 5.1. Expanded Tricyclic adenine and guanine analogues.
The furan, thiophene or pyrrole spacers provide an additional H-bond donor or acceptor, which would prove advantageous for interactions with cations or water in the
81 spine of hydration of the DNA helix. In addition, the heterocyclic spacer ring would
increase the polarizability of the nucleobases as well as the aromatic π system, thereby
significantly increasing the forces contributing to stacking and stability. The results of the
molecular dynamics simulations are outlined herein.
As previously stated, DNA is the carrier of vital genetic information and is
involved in the replication of living cells as well as viruses. In order for cells to transfer
genetic information to “new cells”, proper DNA replication is critical. The expression of
genetic information is facilitated by RNA, which is transcribed from a template DNA
strand. Following their syntheses, messenger RNAs are translated into proteins and
cellular enzymes. This close relationship between the transfer of genetic information and
the nucleic acids establish them as potential therapeutic targets for drug design.
DNA Structure
Watson and Crick first proposed the model for DNA structure in 1953 (Figure
5.2).177 This model consists of a double-strand helix of complementary consisting of
varying patterns of nucleotides. The nucleotides vary only in their heterocyclic bases; the
base moieties are two purines, adenine (A) and guanine (G), and two pyrimidines,
cytosine (C) and thymine (T) located perpendicular to the central axis of the helix. The nucleotides pair in a very specific manner; A with T and G with C. The complementary strands of DNA are “held” together by hydrogen bonding between the base pairs; two between A-T and three between C-G (Figure 5.2). The bases have interplanar distances
of 3.4 Å and a rotation of approximately 36° between adjacent pairs.2 The hydrogen
82 bonds, along with London dispersion forces and hydrophobic interactions, between the stacked bases help stabilize the structure of the double helix.
Figure 5.2. Watson and Crick model of DNA.177
Each nucleobase is each attached to a deoxyribose sugar with a phosphodiester bond joining the 5’ hydroxyl group of one sugar to the 3’ hydroxyl group of an adjacent sugar to form a polymeric chain (Figure 5.2).2 The “backbone” polymer chains are aligned antiparallel along the periphery of the DNA strand forming a very concise
83 pattern. The glycosidic bond connecting the sugar and the base moieties are not directly aligned with its complementary strand partner; therefore the double helix is not equally spaced along its axis. As a result, the spatial relationship between the two strands in the helix forms a major (wide) groove and a minor (narrow) groove (Figure 5.3) with 10
bases per turn for B-DNA.
Figure 5.3. DNA double helix major and minor grooves.
84 The base pair oxygen and nitrogen atoms that project into the major groove face
toward the center of the DNA strand while the base pair oxygen and nitrogen atoms that
project into the minor groove face outward from the sugar-phosphate backbones towards
the outer edge of the DNA strand. The specific sequence of bases constitutes the genetic
code of each gene.
Based on the initial structure of DNA, it was thought that hydrogen bonding,
whether it was Watson-Crick or Hoogsteen type hydrogen bonding (Figure 5.4), was the
key contributing factor to its stability, recognition and replication. More recently, it has
been found that factors involved in the stability of DNA is much more complicated than
once realized. Researchers believe that other factors such as π stacking, shape, sterics,
electronics, and hydrophobic interactions, as well as solvent effects and functional
groups, may be as, or even more important than hydrogen bonding.178
Figure 5.4. Watson-Crick and Hoogsteen hydrogen bonding.
Regardless, since hydrogen bonding is involved in maintaining the structure of
DNA, the actual tautomeric forms of the bases are important.2 As demonstrated in Figure
5.5, the preferred tautomeric form for the amino substituent in the bases A, C, and G is
85 the amino and not the imino while the exocyclic oxygen atoms in G and T prefer the keto form rather than the enol form.
Figure 5.5. Tautomers of A and T.
The three-dimensional structure of DNA is also dependent on the torsion angles of the backbone and the glycosidic bond.179 Ultimately, there are seven torsional angles,
α, β, γ, δ, ε, and ζ that describe DNA movement; six for the phosphate sugar backbone
(Figure 5.6-A) and a seventh for the orientation of the base relative to the sugar (Figure
5.6-B). The torsion of the backbone atoms is described in the 5’-3’ direction. The atoms,
starting from the O3’(-1) of the previous sugar, or O3’(-1)-P-O5’-C5’, define the α
torsion angle, and the β torsion angle continues with P-O5’-C5’-C4’. The angles γ, ε, δ, and ζ follow the same method of labeling. The seventh torsional angle, χ, characterizes the orientation of the base relative to the sugar and is defined by the atoms O4’-C1’-N9-
C4 for purines and O4’-C1’-N1-C2 for pyrimidines. A positive angle refers to a clockwise rotation of the dihedral angle. Backbone atoms in the same plane with a rotation of 0˚ are in a cis (c) conformation, while atoms in the same plane with a rotation
86 of 180˚ are in a trans (t) conformation. There are two gauche forms, gauche positive (g+)
and gauche negative (g-), in which the dihedral angles are 60˚ and -60˚, respectively. The dihedral angles of the backbone vary widely, but DNA with stacked bases has a tendency towards the following angles, α(g-), β(t), γ(g+), ε(t), ζ(g-). The δ dihedral angle is related
to the sugar pucker where as the χ torsional angle involves the conformation of the base- sugar glycosidic bond.
Figure 4.6. DNA backbone and sugar torsional angles. 180
As previously mentioned, both RNA and DNA contain the 5-membered ribose
and deoxyribose sugars respectively, which can also be described by specific dihedral
179 angles. The torsion angle, υ0, is formed with the atoms C4’, O4’, C1’ and C2’ of the
sugar moiety. The angle, υ1, continues with the atoms O4’, C1’, C2’ and C3’. The torsion
angles υ2, υ3 and υ4 are labeled in a clockwise direction as shown in Figure 5.6-B. In
87 addition, the torsion angle υ3 also describes the backbone torsion angle δ, although different atoms are used to describe the angle than are used to determine δ.
A planar five-membered ring is sterically and energetically unfavorable, therefore
an atom or atoms prefer to reside out of the plane to form a sugar pucker. The
conformation of the atom(s) out of the plane denotes the specific type of sugar pucker.
For example, when the 3’ carbon is out of the C4’-O4’-C1’-C2’ plane on the same side as
the base, it is referred as a 3’-endo conformation. If the 3’ carbon is located on the
opposite side relative to the base, it is termed 3’-exo. Both 3’-endo and 3’-exo puckering
result in an envelope shape. In contrast, when two adjacent atoms such as the C2’ and
C3’ are out of the plane, a twisted conformation is formed. The two out of plane atoms are referred to as endo and exo.179
The pucker conformations result in a pseudorotation phase angle P (Figure 5.7), which describes the torsional movement around the ring. The radius describes the degree of puckering from 0 to 45 degrees. As depicted in Figure 5.7, every 18˚ of puckering forms alternate envelope or twist conformations. A 180˚-rotation of the sugar pucker gives its mirror image. For clarity, the conformers are labeled North or South. North (N) conformers are primarily found in A-DNA and have phase angles closet to 0˚, whereas south (S) conformers have phase angles closest to 180˚ and are mainly found in B-
DNA.179
There are three major structural forms of DNA A-, B-, and Z-forms (depicted in
Figure 5.8 on page 91).2 The main differences are that A-DNA is wide and compact,
while Z-DNA is considered thin and B-DNA falls somewhere between the two.179 The A and B forms have a right-handed twist or clockwise rotation moving down the helix axis,
88 whereas the Z form is a left-handed helix with a counterclockwise rotation. Specific
structural parameters for A- B-and Z-DNA are listed in Table 5.1.
Figure 5.7. Sugar pseudorotation.180
The twist in the DNA helix is caused by stacked bases that are rotated with respect to each other.179 Both A and B-DNA contain one base pair before rotating; A-
DNA is rotated approximately 33˚, while B-DNA averages a 36˚ rotation. Z-DNA has two base pairs per rotation with an average twist angle of 60˚ shared unequally between the G-C pair step of -50˚ and the C-G step of -10˚. The rotation results in 11, 10 and 12 bases per turn for A-, B- and Z-DNA respectively. Along with the twist angle, the actual
89 distance between adjacent base pairs differs and is referred to as the rise. A-DNA has a
rise of 2.9 Å per base pair, B-DNA has a rise of 3.4 Å, and Z-DNA has a rise of –3.9 Å for the C-G step and –3.5 Å for the G-C step. The rise, along with the number of base pairs per turn, gives the distance for the helix pitch of 23 Å, 34 Å and 45 Å for A-, B-, and Z-DNA, respectively. The base pairs are also inclined at a large angle in the A form, but are relatively perpendicular to the helix axis in the B and Z forms.
Table 5.1. A-, B- and Z-DNA Structural Parameters.179 A-DNA B-DNA Z-DNA Helix handedness Right Right Left bp/repeating unit 1 1 2 bp/turn 11 10 12 Helix twist, (º) 32.7 36 -10, -50 Rise/bp, (Å) 2.9 3.4 -3.9, -3.5 Helix pitch, (Å) 32 34 45 Base pair inclination, (º) 12 2.4 -6.2 P distance from helix axis, (Å) 9.5 9.4 6.2, 7.7 X displacement from bp to helix axis, (Å) -4.1 0.8 3 Glycosidic bond orientation anti anti anti, syn Sugar conformation* C3'-endo C2'-endo C3'-endo, C2'-endo Major groove depth 13.5 8.5 convex width (Å) 2.7 11.7 Minor groove depth 2.8 7.5 9 width (Å) 11 5.7 4 Helix diameter (Å) 26 20 18
* For B-DNA possesses range of conformations, however the preferred is stated
The distance of the base pairs from the helix axis of DNA is a major component.
The bases in B-DNA are on the helix axis and the distance from the axis averages 0.8 Å,
90
A-DNA B-DNA Z-DNA
Figure 5.8. DNA forms.181
91 whereas, the A-DNA base-helix distance is approximately –4 Å, and Z-DNA’s averages
3 Å.179 Variable base to helix axis distances help regulate the differences in the major and minor groove. A-DNA’s major groove dimensions are 2.7 Å wide and 13.5 Å deep, while the minor groove dimensions are 11.0 Å wide and 2.8 Å deep. In B-DNA, the major and minor grooves have similar depths of 8.5 Å for the major and 7.5 Å for the minor groove; however, the widths are 11.7 Å and 5.7 Å for the major and minor groove, respectively.
In Z-DNA there is no depth to the major groove since it is convex, however the minor groove is 9 Å deep and 4 Å wide and is lined with phosphate groups.
Figure 5.9. Anti and Syn conformations of purine nucleosides.
Within the three DNA forms, the orientation of glycosidic bond and the sugar conformations also vary. In A- and B-DNA, the bases are in an anti orientation to the sugar (Figure 5.9) while the Z-DNA contains cytosines that are anti and guanosines that are syn.179 Furthermore, the sugar pucker is different for each nucleoside. A-DNA sugar pucker is C3’-endo, B-DNA is generally C2’-endo, while Z-DNA is C2’-endo for cytosine and C3’-endo for guanosine. It has also been shown that Z-DNA is more rigid than A or B forms. In addition, B-DNA can bend easily by collapsing the major groove.
Although there are many differences between the three strands, the relative flexibility of
92 each component remains the same. The local movement of phosphates is always greater
than the sugar, which is greater than the bases.179
DNA Movement
Even though π stacking and hydrogen bonding maintain the stability of DNA,
there is some variability with regards to the internal movement of the DNA bases thereby causing distortions to arise.182 This is often observed in the right-handed A- and B- forms of DNA. However, due to the rigidity of the backbone in the Z-form, the base stacking pattern is fixed.179 The two main types of base pair movements are rotational and translational, which vary depending on the base sequence and stacking pattern of the base pairs. The relative degree of the displacement also depends on the energetics of the base stacking.
A visual model of the various base rotational movements is illustrated in Figure
5.10 on the following page.179 The rectangular block represents the base pair and the base
corners in the +y, -x and the -y, -x direction are the points of attachment of the glycosidic bond to the sugar. Strand I of the DNA double helix is conceptually on the left going from 3’ to 5’ in the -z direction while strand II is on the right going from 3’ to 5’ in the +z direction. The base motions in the top row of Figure 5.10 are coordinated; the motions of the middle row are opposed to each other; and the motions in the bottom row are two- base-pair rotations. The left, center and right columns describe motion in the z, y and x direction respectively.
These are the definitions of base pair rotational movements.179
93 • Positive tip rotates the base towards the major groove when viewed along the +y-
axis direction.
• Positive inclination rotates the base pair clockwise when looking towards the +x-
axis direction and is relative to a plane that is perpendicular to the helix axis. This
angle varies from 9˚ to 22˚ in A-DNA and about 2.4˚ in B-DNA.
Figure 5.10. DNA base rotational movements.179
• Base opening occurs through counter rotation of the bases about the z-axis and is
positive when opening towards the +x-axis direction.
• Positive propeller twist rotates the base attached to strand II clockwise relative to
the base attached to strand I when viewed along the -y-axis. Variances of
94 propeller twist have been observed between 6˚ and 16˚ for A-DNA and 13˚ to -18˚
in B-DNA but are insignificant in Z-DNA.
• Buckling in all three forms of DNA is normally only a few degrees and is positive
when the bases bend in the center of the strand towards the -z-axis direction
relative to the strand ends.
• Twist is the angle needed to rotate a base pair to match the next base pair in the
helix axis. In B-DNA, a twist of 24˚ to 51˚ is required for base pair alignment and
in A-DNA a twist of 23˚ to 44˚ is required.
• Positive roll between base pairs widens the angle on the side towards the minor
groove. Angles between 6˚ and 9˚ have been observed for A-DNA but B-DNA
angles are almost zero although bending does result in base pair rolling.
• Positive tilt widens the angle towards strand I along the +y-axis.
The visual description of translational base movements can be seen in Figure
5.11 on the following page. The columns, from the left to the right, describe the motions in the z, x and y directions respectively. The first two rows describe the various movements involving two base pairs. Coordinated motion between the bases is shown in the first row, while the second row is composed of opposed motions and the third row consists of motions between successive base pairs.
These are the definitions of base pair rotational movements.183
• Displacements in the x and y directions are the movements of the base pairs
relative to the helix axis with positive movements in +x-axis and +y-axis
direction.
95 • Stagger is the counter movement of the individual bases where one base rises and
the other falls relative to the helix z-axis.
• Stretch is where the bases shift away from each other in the y-axis.
• Shear motion is the counter movement of the base pair along the x-axis.
• Rise is the distance between adjacent base pairs along the helix axis or z-axis.
Figure 4.11. DNA base translational movements.179
• Slide is when one base pair moves along its C6-C8 vector midpoint and parallel to
the y-axis.
96 • Shift is the relative movement of the vector midpoints, atom C6 of pyrimidines
and C8 of purines (Figure 2.17 on page 27), of adjacent base pairs perpendicular
to the y-axis of the base pairs.
Previous Research
With this general understanding of DNA structure and movement, researchers
designed nonstandard base pairs and nucleotides as potential inhibitors of DNA
replication, as well as to expand the genetic alphabet. As stated previously, it was once
thought hydrogen bonding was the primary contributing factor for the stability of DNA.
However, with the implementation of more advanced scientific techniques and
equipment, scientific views are changing.184 Several recent studies have focused on
probing the true importance of hydrogen bonding versus stacking effects.178,185-194 Others have focused on the stability of modified DNA, as well as to investigate factors involved with recognition by polymerases.192,195-202
Recently, ab initio calculations were performed on base pairing to confirm that
electrostatic interactions stabilize H-bonded base pairs.203 Theoretical studies have
indicated that the individual strengths of the hydrogen bonds between AT and GC bases
cannot be merged into averaged strengths of AT and GC base pairs, respectively.204 The central and major groove hydrogen bonds in GC base pairs are approximately two times stronger than the hydrogen bonds in AT pairs. While there is no hydrogen bonding pair in the minor groove for an AT pair, there is a positive interaction; this interaction is five times weaker than that of the minor groove hydrogen bonding of the GC pair. This in turn
97 shows that the contribution to stability of the bases through hydrogen bonding is at least two times greater for a GC pair than an AT pair.
Experimentally, Benner et al has carried out several extensive studies on
unnatural hydrogen bonding patterns. Cytosine and guanosine nucleobases were modified
to produce new Watson-Crick hydrogen bonding patterns (Figure 5.12).205-208 The modified nucleosides were shown to pair with each other, thereby increasing the structural diversity of nucleic acids.209
Figure 5.12. Examples of modified hydrogen-bonding nucleosides.205,206,209
Figure 5.13. Modified nucleosides.201
98 Related to this, Switzer and colleagues synthesized and incorporated isocytosine
and isoguanosine (Figure 5.13 on the previous page) into DNA in an effort to
demonstrate the structural diversity of nucleic acids.201 These bases also preferentially
paired with each other,201 although, they were later shown to pair with other bases.210
These and other theoretical studies showed that isoguanosine can form an alternative tautomeric form in order to pair with thymidine. They also showed that the isomers are recognized by DNA and RNA polymerases.211-213
The stability of DNA through minor groove modifications was studied by Bailly
and colleagues by the addition of an amino group to the 2-position of guanosine to form
2,6-diaminopurine nucleoside (Figure 5.12 on the previous page).214-216 These findings
showed that the extension of a hydrogen bond donor into the minor groove disrupts minor
groove width and hydrogen bond bridges with water, which is critical for stability.217
The importance of minor groove hydrogen bonding interactions was further demonstrated by removal of the minor groove hydrogen bond acceptors of adenosine and thymine to form 3-deazaadenosine and 3-methyl-2-pyridone (Figure 5.14).218 These acceptor atoms formed hydrogen bond bridges with water to create a “spine of hydration” that stabilizes B-DNA. The study showed that loss of these hydrogen bond atoms destabilizes the B-DNA structure and forces the adoption of the more stable A- form.219
Figure 5.14. Minor groove probes.218
99
To further illustrate the importance of hydrogen bonding in DNA structural
stability, nucleosides with additional hydrogen bonding moieties resembling natural bases
have been synthesized (Figure 5.15).191,193,220 The investigations have shown that the
additional hydrogen bond donor or acceptor increases overall stability of the DNA strand.
Although modifications to hydrogen bond acceptors and donors have shown promise,
removal of these elements has also shown that these features are not necessarily needed for recognition and incorporation into DNA.195
Figure 5.15. Increased hydrogen bonding analogues and G-clamps. 191,193,220
Another factor that has been shown to increase DNA stability is base stacking.
Base stacking refers to the interaction of the π orbitals of adjacent bases and has been
shown to increase stability by approximately 0.4 to 3.6 kcal/mol.179 There are two types of base stacking, intra- and inter-strand stacking. Several factors that enhance intra- and
inter-strand stacking are (i) increased surface area221 (ii) incorporation of a dangling base
100 at the end of the strand,188,199,222 and (iii) addition of substituents that increase
polarizability.200,221
It has been proposed that the increased stacking of the bases correlates with
surface area and hydrophobicity of the larger modified nucleosides.223 Studies have
shown that nonpolar nucleosides, when paired with polar nucleosides, destabilize the
DNA strand, but when both pairs are nonpolar the DNA strand is only slightly
destabilized.186,224 Hydrophobic analogues have also been shown to selectively pair with their hydrophobic partners over normal bases.224,225 An interesting aspect is that hydrophobicity stabilizes the DNA double helix only when hydrophobic base(s) are placed at the terminal end of a DNA strand as pairs or as a dangling end. If the hydrophobic base(s) is placed in the center, the same base(s) can destabilize the helix.184,185,224
The dangling ends can contribute to the DNA duplex stability as much as a
Watson-Crick A-T base pair.222 However, the stability is dependent on the ending base
pair. For instance, adenine-dangling ends are always more stabilizing than other dangling
nucleotides. Moreover, 5’ dangling ends are more stabilizing than 3’ in DNA, while the
reverse is true for RNA.188,226 For the natural bases, adenine has the greatest stacking
ability followed by guanine, which is slightly greater than thymine and cytosine, which
are equivalent.221 Most studies on dangling ends were focused on single dangling bases as
opposed to multiple on a single end. However, a recent study on long dangling ends
indicated that there is a quantitative increase in stability with additional bases199 and that
the increase in polarizability of the dangling end also increases stability.200
101 It was then proposed that van der Waals attractions are the most important factors
in base stacking.200 The effects of polarizability on molecules through induced dipoles or
from electrostatic interactions from permanent dipoles have been shown to contribute
between one-third to one-half to stacking effects which enhances helix stability.190,197,200
Studies on fluorinated benzene nucleosides have shown that dispersive van der Waals interactions are among the most important of these forces in DNA stacking.190
Romesberg et al has probed DNA with the use of self-pairing hydrophobic
nucleobases.197 These nucleobases, which contain sulfur or oxygen heteroatoms and are
incapable of tautomerization, have shown increased stability that can thermodynamically
compensate for the Watson-Crick hydrogen bonds found in natural bases.198,227 The nucleobases have the ability to form base pairs in DNA duplexes. Several of these self- pairing hydrophobic nucleobases have also been efficiently incorporated into DNA by polymerases at rates approaching those for the natural bases.189,228-230 The best
incorporation of these are the pairing between 7-azaindole (7AI) and isocarbostyril (ICS)
and the self pair of 7AI Figure 5.16.202,228,231
Figure 5.16. Examples of hydrophobic bases.202,228,231
102 Other hydrophobic bases such as bipyridyl and biphenyl C-nucleosides have also
been incorporated into DNA (Figure 5.17).232 The use of both bipyridyl and biphenyl C- nucleosides introduced the possibility of intercalation of extended bases. These modified
nucleosides form a zipper-like configuration.232 These bases also show an increase in
stability due to increased entropy from the release of water upon forming a double strand,
which is expected with a nonpolar base in a polar solvent. The bases were found to form
a stacked and highly ordered state with a planar arrangement in the center of a DNA
helix.196
Figure 5.17. Bipyridyl and biphenyl C-nucleosides.232
Base-expanded, tricyclic nucleosides have also been used to probe the stability of
DNA. Early studies by Nelson Leonard demonstrated that tricyclic expanded adenosine
nucleosides could be incorporated into DNA by polymerases although at low frequency.
They also act as chain terminators since the 3’-hydroxyl is mispositioned thereby
preventing the linking of the α-5’-phosphonate of the next incoming nucleotide (Figure
5.18). 10,233,234
103
Figure 5.18. Incorporation of lin-benzoadenosine.233
Figure 5.19. Expanded and extended nucleosides.
104 More recently, Leonard’s lin-benzo separated adenosine11,12,235 and
guanosine236,237 nucleosides have been incorporated into DNA238-241 along with an
extended thymine241-243and cytosine analogue238,241 Figure 5.19 on the previous page.
These analogues were used to form an “expanded” DNA chain by incorporating the
extended bases opposite the natural bases in a DNA strand.
These analogues increase the length of the natural nucleosides by 2.4 Å and still retain the normal functionalities necessary for hydrogen bonding. When incorporated into expanded helices, the helix shows increased stability from (i) increased stacking of the bases from within its own strand or (ii) the overlap of the bases from the opposite strand when paired with their complementary bases in DNA. Although these nucleosides show promise, the hydrophobicity of the benzene ring leaves much to question as to whether they will be recognized by nucleoside and nucleotide metabolizing enzymes.
Figure 5.20. Tethered naphthalene and tethered adenine analogues.194
Related to this, studies have indicated that two tethered naphthalene residues
(Figure 5.20) in aqueous solution do not interact with each other whereas two tethered adenines as well as a tethered adenine and naphthalene showed interactions. Furthermore, the study showed that polarizability is more important than hydrophobicity with regards to stacking.194 A similar study on tricyclic pyrimidine nucleosides indicated that
increased π-π overlap of adjacent bases increases stability and is probably due to induced
105 dipole effects.244,245 Hecht and coworkers have shown that generally, the greater the
degree of overlap of the heteroatom aromatic surfaces, the greater the stability introduced
by base-pair stacking.182 Although the benzene-expanded nucleoside analogues contain
the normal nucleoside hydrogen bonding aspects and their hydrophobic portions have
shown to increase DNA stability, most of these analogues have yet to be incorporated into DNA by polymerases. Other than the expanded tricyclic adenosine analogue it is not known whether polymerase enzymes will incorporate them into DNA and continue replication.
Recognition of DNA by Enzymes and Inhibitors
Recognition of DNA by enzymes is important for replication, modification and repair.215,246-248 The interactions of most proteins with DNA occur in the major groove
rather than the minor groove. The major groove (i) offers greater accessibility than the
minor groove and (ii) displays characteristic pattern of H-bond donors and acceptors that
enables the four base pairs to be distinguished.215 It has recently become apparent
however that the minor groove does serve as a receptor for several proteins (SRY, LEF-1,
PurR, TBP, HMG-1 and IHF).215 The H-bond acceptors of all four base pairs are
arranged in the minor groove almost identically which implies that they confer little
sequence discrimination.
DNA polymerases have the ability to hydrogen bond to the minor grove and thus
the minor groove is important to the replication process.249 It was found that the
interactions necessary for extension, rather than insertion, of DNA were more important
and furthermore, they were stronger for the primer terminus than the template.192 The
106 minor groove interaction in the primer has the probability to make a 300-fold difference in extension efficiency. The polymerase may be even more sensitive to base pairing geometry in extension than for the insertion of a base.249 Formation of a hydrogen bond between the primer terminus and the polymerase may be necessary to form the correct geometry to efficiently continue extension of DNA.192 Studies have also indicated that minor groove binding drugs need hydrogen bond acceptors in the minor groove in order to recognize of DNA.250,251
Recognition of hydrogen bond acceptors by AT and GC specific minor groove binding drugs such as distamycin252 and mithramycin253,254 is critical for their mechanism of action (Figure 5.21). It is also essential for high mobility group protein D (HMG-D) one of the Drosophila melanogaster counterparts of the chromosomal protein HMG-1 that binds primarily in the minor groove and is vital in developing embryos.215 HMG-D is involved in bending DNA and binds selectively to DNA structures that are both bent and underwound.255
DNA bending is an essential part of gene regulation. Spontaneous bending of
DNA strands is normal.256 Sharp DNA bending allows proteins to have a synergistic effect by acting on more than one site on DNA. The bends occur almost exclusively along the base pair axis, via the roll motion, towards the major groove257 and at purine- pyrimidine steps since there is less overlap by the base rings. The difference in bendability of base pairs is important for minor groove recognition and is both sequence dependent and facultative. Researchers studying synthetically bent DNA found that
HMG-D possessed a 15 fold greater affinity to the bent DNA than for the natural straight strand.258
107 Related to this, studies to determine the requirements needed for minor groove
recognition by enzymes have also been pursued.184 One of these studies was with 2,6-
diaminopurine (Figure 5.13 on page 98), which was used to alter the minor groove width
and the spine of hydration.217 Molecules such as these were used with the intent to expand the genetic alphabet, as well as to investigate the elasticity of DNA.
Figure 5.21. Minor groove binding drugs.252,253
It is well established that that both helix flexibility and bendability contribute
significantly to the process of identifying the sequences and molecular features of
108 DNA.214,259 The precise recognition of a nucleotide sequence in DNA by a given protein, necessitates an optimal complementary shape between the interacting species. The mutual conformational adjustment between the two components leads to increased hydrogen bonding, electrostatic contacts and van der Waals interactions thereby affording increased sequence specificity. These binding interactions normally involve small and, in some cases large, deformations in B-DNA.260 The reinforcement of base pairing by
increasing the quantity of hydrogen bonds reduces the flexibility of DNA and in general
reduces the extent of protein binding for small proteins.217 Even though DNA flexibility
is crucial for the binding of enzymes to B-DNA, the possibility that only DNA bending
would allow sufficient access to the bases for modification in the double helix was
difficult to comprehend until the discovery of base flipping.
Base flipping is a phenomenon that affects DNA methylation, mismatch repair,
nucleotide-excision repair and base-excision.261-263 Furthermore, it has also been
proposed as an essential step involved in the early events of transcription and replication
processes264 as well as methylation.265 Methylated bases are not incorporated into DNA
by replication machinery but rather, are modified after replication.265 Base flipping was
initially observed with 5-methylcytosine in DNA methyltransferase (M.HhaI)266,
(M.HaeIII)267, human uracil DNA glycosylase268,269 and T4 endonuclease.270 Base flipping has been ascribed260,271 for several other enzymes, however, their mechanisms of
action have yet to be fully elucidated. Nevertheless, three potential pathways have been proposed to explain this phenomenon. The first one is considered “passive”, where the base spontaneously “flips out” of the helix structure and is then acted upon by the enzyme.261,272 The other two processes are termed as “active” and are achieved through
109 several enzymatic pathways by pushing the target base out of the DNA helix and
subsequently drawing it into the binding pocket of the enzyme. This is achieved by either
compression of the opposite strand, which forces the base to flip out of the helix into the
major groove 261 or, through the destabilization of the Watson-Crick pairs by the enzyme thereby allowing the excised base to achieve a more stable conformation.264,272,273 Both
active pathways stabilize the DNA strand by insertion of amino acid residue(s) through
the minor and/or major groove(s) to offset the gap created by the excised base, thereby
stabilizing the “orphaned” base through hydrogen bonding.261,264,272,273 It has also been
established that enzymes which excise bases contain very specific “lock and key” binding
sites and can recognize differences between hydrogen, methyl or hydroxyl groups.274
Theoretical studies in solution have shown that base flipping of cytosine residues can occur through either the major or minor groove of DNA since both active pathways have similar energy barriers.264,275,276 Recently, MD simulations have indicated that base flipping for HhaI methyltransferase occurs through the major groove and is facilitated by the rotation of the sugar pucker (of the excised base) from the south to the north conformation halfway through the flipping process and is essential for stability.277 The
proposed mechanism begins with an initial DNA-enzyme complex which is opened to
display a collection of flipped out bases until a more compact complex is formed when
the target residue is locked into the binding site of the enzyme.278 It may be possible that
base flipping can be impeded either by increased stacking interactions of bases or through
natural DNA bending. If base flipping can indeed be inhibited, it could induce inhibition
of cellular processes.
110 One approach is to increase stacking either through the hydrophobic effect or
through increased polarizability. We have designed a series of expanded bases that
increase the polarizability of the bases. As stated previously, these analogues contain a
heteroaromatic spacer ring (Figure 5.1 page 81) that can increase the polarizability of the
base. In order to further examine stacking and bonding interactions, molecular dynamics
simulations were performed on our target compounds in DNA.
Molecular Dynamics
MD simulations are computational tools used to estimate equilibrium and dynamic properties of complex systems that cannot be calculated mathematically by hand.13 The static view of a biomolecule, as obtained from X-ray crystallography, is insufficient for understanding a wide range of biological activities since it only provides an average “snapshot” image of a complex system. By following the motions of a molecular system in space and time, we can obtain a rich amount of information about the structural and dynamic properties of a system.
The critical information provided from molecular dynamics simulations includes molecular geometries and energies, mean atomic fluctuations, local fluctuations, rates of configurational changes, enzyme/substrate binding, free energies and the nature of various types of concerted motions and large scale deformations.13 This information is obtained through the use of Newton’s equations of motion for one or more molecules over time.1 This results in a series of predicted coordinates that trace the movement of
atoms within a molecule and may be displayed as a movie by rapidly moving through
individual frames.
111 The motion of the molecules’ nuclei follows the laws of classical physics;1 the molecules vibrate according to the individual masses of the atoms and the force constants of the bonds between them. The force for each atom is expressed using Newton’s second law of motion.
(5.1)
However, this equation is limited in that it only describes a force during one brief
point in time and therefore must be reformulated to include the motion of the molecule
based on the coordinates and potential energy of the atom.1 The differential equation
below describes this motion where the change, δ, in the coordinates of the atom ri are
described using the change in time t.
(5.2)
Unfortunately when there are more than two atoms, a solution cannot be
determined, however using a Taylor series expansion, as seen below, a solution can be
approximated.1 The expression states that over a short time interval ∆t, the position of the
atom can be calculated if the coordinates of the atom are known at that time r(t). In
addition, the velocity of the atom can be formulated from the first derivative (δr/δt), the
acceleration from the second derivative (δr2/δt2) and the approximation of higher order terms from the remainder of the formula. Each atom’s behavior is integrated over time
112 and evaluation of the average velocity of each atom during that time period over all the
atoms of the molecule provides the movement for that time period. Finally, by updating
the coordinates, the next step can be calculated.
(5.3)
Unfortunately, these equations make assumptions that still limit the accuracy of
the method.1 The most important is the time interval; ∆t must be small enough to simulate
the motions of the atoms and it must be able to evaluate the motion of the fastest
vibrating atoms otherwise the equation will fail.1 Since hydrogen is considered the fastest
moving atom and the C-H stretch periodicity is about 10-14 seconds, a sampling of at least
10 shots is needed to capture the motions accurately. Therefore, this requires the time interval (∆t) to be a minimum of 10-15 seconds or 1 femptosecond (fs). Heavier atom
movements such as carbon, nitrogen and oxygen occur on the picosecond (ps) time scale, or 10-12 seconds. Therefore, a minimum of 10 ps must be simulated to acquire enough
information on the heavy atom movements for analysis.1 Although longer simulation
times would be more beneficial, it is necessary to consider the computational time
required is prohibitive, but with the development of faster computers, larger and more
complicated simulations are now being attempted.
Some of the milestones of modern day molecular dynamics simulations are depicted in Figure 5.22. The field began to make strides in the early 1970s with the
simulation of a dinucleoside CpG pair.13 At the time, the simulation proved challenging,
113 however, with intelligent strategies and constraints the correct conformation was found.
In the late 1970s the Bovine Pancreatic Trypsin Inhibitor (BPTI) protein was simulated and atomic fluctuations on the picosecond timescale were visualized.13 DNA was initially simulated in a vacuum in the early 1980s, but the initial simulations had numerous problems with strand stability, which were corrected in the following decade.13 A simulation of the protein myoglobin for 300 ps was the longest simulation of that time and it appeared that many thermodynamic properties were beginning to converge.13
Figure 5.22. Evolution of Molecular Dynamics simulations.13
114 In 1992, HIV protease was simulated in solution and a bending motion in the
binding site was visualized.13 In the late 1990s, significant strides in simulations were
accomplished. As seen with the estrogen/DNA complex, DNA simulation using long
range electrostatics, peptide simulation, villin-headpiece and the solvated protein bc1 embedded in a phospholipid bilayer suggested the passage of a proton through a water channel.13 The longest published simulation to date is of a 16-residue β-hairpin from
protein G with a time of 38 µs and was reported in 2001.13 In the early stages of
molecular dynamics calculations the molecules were simulated in a vacuum but by the
early 1990s simulations were run in a solvent with the hopes of producing more accurate
structures. Some proposals for future molecular dynamics simulations by Duan et al.,
suggest that if computer power increases by a factor of 10 every 3-4 years, the simulation
of one second of time of medium-sized proteins by 2020 and the simulation of an entire
life cycle, 20 minutes, of an E. Coli cell may be feasible by 2055.279
In the 1990s, the first generation of biomolecular forcefields was developed which revitalized the use of molecular dynamics.13 The development of solvent force fields by
Berendsen and coworkers280 and by Jorgensen and coworkers281 provided the
groundwork for simulations in solution. Peter Kollman and coworkers282 then developed
force fields designed for studying enzyme catalysis and protein/ligands. The use of free
energy and the combined quantum/molecular mechanics methods could then be used to
simulate important biomolecular and medicinal problems. These methods have opened
the doors to many new applications, although there remain some limitations to the
process.
115 The limitations to molecular dynamics are similar to molecular mechanics.13
Unfortunately, the simulations ignore the electronic motion and quantum effects thus only taking into account the movement of an atom’s nuclei. Although this type of simulation is useful for a wide range of simulations, it is unsuitable for reactions involving bond formation and cleavage, polarization, and chemical bonding of metal ions. Moreover, it is also unsuitable for low temperatures due to the formation of large energy gaps as dictated by quantum physics, which is much larger than the thermal energy available to the system. The system is confined to one or a few discrete low energy states under the low temperature conditions. The previous described limitations to the molecular dynamics simulations were not of concern since the research described herein focused on the stability and the parameters of expanded bases in DNA.
Modeling Results
Insertion of either a furan, pyrrole, or thiophene ring between the imidazole and pyrimidine rings of the purine scaffold extends the overall length of the bases by ~1.42
Å. The bases curve approximately 30º, 33º and 43º respectively towards the minor groove. As previously mentioned, it was our hypothesis that inclusion of these spacer rings would provide several advantages to the base pairs that should result in a significant increase in stability for the modified helices. A series of oligonucleotide strands were modeled with various sequences containing the extended bases to test the stability, the increased stacking and the effect of base curvature and purine expansion on the DNA backbone.
116
Figure 5.23. Base polarization and electrostatic surface potentials.
In that regard, it is well known that increasing the base polarizability in DNA contributes significantly to stacking and is crucial for recognition and replication by polymerases.197,200 Using Spartan, the polarizabilities and the electrostatic surface
potentials of each base were calculated using the AM1 level of theory. The results are
shown in Figure 5.23. (Note: The sugar moieties were removed and replaced with methyl
117 groups since our initial calculations indicated that the sugars do not affect the polarizability or dipole moments of the bases.) According to our data, the polarizability of the expanded furan, pyrrole, and thiophene nucleosides was increased compared to the parent adenosine and guanosine. This should translate into greater dispersion forces and, along with the extended π-system, should therefore increase the stacking contributions and overall increase duplex stability.
The electrostatic surface potentials indicate the normal hydrogen-bonding base moieties remain untouched for all of the expanded nucleosides. For the furano tricyclics, an additional electronegative/hydrogen bond acceptor (yellow to red) surface can be seen on the top center of the bases. With the pyrrolo tricyclics, a more electropositive/hydrogen bond donor (green to blue) surface is observed. The thiazole tricyclic is slightly electropositive (green), which is likely due to the increased size and dispersive effect of the sulfur atom. However, like the oxygen of the furan system, there are two pairs of electrons from the sulfur available for donation and bonding. The additional hydrogen bond acceptor or donor atoms should also increase recognition of the modified nucleotide in the major groove. This also should increase the chances of forming triplex and tetraplex DNA strands, which have been shown to help in regulating site specific DNA cleavage, protein binding and gene expression.283 These systems have been exploited and studied by several groups since their discovery in 1987.170-172,251,283-287
The initial molecular dynamics simulation of the expanded DNA was performed on a 10-mer composed of expanded thiophene adenosine (AT) bases with the sequence dATTATATTATTTATT. The bases were incorporated into the strand with no additional modifications except for adjusting the dihedral angle of the glycosidic bond in order to
118 place the extended At bases in the same plane as the normal thymine base. This forced an
extreme overlap of atoms between the bases on opposite strands (Figure 5.24).
As expected, during the initial minimization most of the bases did not hydrogen bond with their complementary base on the opposite strand. This was most likely due to the close proximity of the complementary bases, the steric and van der Waals repulsions,
as well as the repulsions from the interaction of the electron donor-to-donor and accepter-
to-accepter moieties. However, the bases did remain stacked and a few ultimately paired
correctly. Following this initial test, the dihedral angles of the phosphate backbones in the
remainder of the simulations were adjusted to align the complementary hydrogen
bonding partners and minimize electronic repulsions.
Figure 5.24. Base overlap of inserted tricyclic adenosines.
The next sets of strands studied were also AXT and were composed of the same
sequence of the initial ATT strand (Figure 5.25). Another set of DNA strands were then
119 investigated using GXC sequences; finally, the extended bases AXT or GXC were sequentially placed on the same strand (Figure 5.25). All DNA oligomers exhibited an increased DNA diameter but, due to the curvature of the extended bases, the minor groove widths remained similar to normal B-DNA although the widths increased or decreased moderately depending on the specific base used.
All of the expanded oligomers exhibited an inherent tip towards the major groove.
This varied, although only slightly, the stacking heights between each base, which were also affected by differences in distance from the backbone. The base rise distances were greatest nearer to the backbone, by as much as 4.1 Å, and closest at the ends of the bases near the hydrogen bond in the major groove, by as small as 3.1 Å. It should be noted that the backbone charge was neutralized with sodium ions, the increased distance at the backbone may be due to the repulsions by the negative charges on the phosphates.
Despite the greater distance at the backbone, the strands still remained stacked.
Figure 5.25. Modeled 10-mer DNA strands.
The simulations of AOT, AST, GOC, GNHC and GSC did not display any
anomalies. However, the simulation of ANHT contained a sodium ion that moved into the
120 minor groove. Interestingly the ion coordinated to the carbonyl on the terminal and penultimate thymine on opposite strands as well as the oxygen on the third furanose ring and the N3 of the third tricyclicadenosine. These interactions minimized the minor groove repulsions and collapsed the minor groove. This type of interaction was previously reported in normal B-DNA MD simulations.288 After approximately 250 ps however, the sodium ion escaped the minor groove and the backbone returned to the normal width.
Figure 5.26. Tricyclic dangling ends.
121 After the initial simulations were completed, we performed calculations on
strands with extended bases only on one strand with structure labels of dAXT or dGXC.
With the exception of dGOC all of these extended strands shifted by one base pair leaving
a dangling base on each end as seen in Figure 5.26. The dangling tricyclic bases either
remained stacked beneath the adjacent base or overlapped both bases while the thymine
bases freely rotated in and out of the DNA helix structure throughout the simulation time.
One strand (dAST) initially contained only one dangling thymine, which resulted in an
internal unpaired tricyclic A base. This allowed the DNA strand pairings to shift
throughout the simulation.
The unpaired tricyclic A was the sixth base from the 5’-end in the dAST strand, and remained unpaired for the initial part of the simulation but a subsequent base shift by the fourth thymine to fill the void, caused a shift of the fourth and fifth thymines to pair with the fifth and sixth tricyclic bases, then leaving the fourth tricyclic A unpaired. The third tricyclic A then underwent a propeller twist, which then caused a rotation of the terminal and penultimate tricyclic A bases, thereby pushing the fourth tricyclic A out towards the minor groove (Figure 5.27 on the following page). After the collapse of the three 5’-tricyclic bases towards the major groove, another base shift occurred leaving the
5’-terminal tricyclic A as a dangling end and the third thymine unpaired, but still stacked with the thymines above and below. This also resulted in the collapse of the top three tricyclic A nucleosides into the major groove, thereby preventing the fourth tricyclic A from reentering the DNA strand due to sterics. Although the fourth base was pushed towards the minor groove, the dislodged base showed stacking interactions between
122 either the imidazole ring and the fifth tricyclic A imidazole moiety, or the thiazole ring and the third tricyclic A pyrimidine ring.
Figure 5.27. Partial flipped base.
Two intercalations of the nonstandard bases were observed in the simulations of dGNHC and dGSC, and involved the 5’-terminal tricyclic base inserting in-between the 3’- terminal cytosine and the penultimate cytosine/Tricyclic G base pair. Surprisingly, although the results were identical, the process by which each base intercalated was different. The dGNHC intercalation was initiated by a slide of the base pair in the +y-axis direction, pushing the terminal cytosine partially out of the stacking arrangement over the penultimate cytosine base. This enabled the terminal tricyclic G to stack over both penultimate bases as shown in Figure 5.28 on the following page. The terminal cytosine then buckled towards the +z-axis and subsequently shifted to form a staggered
123 conformation with the terminal tricyclic G. The bases then contracted with the terminal
cytosine, moving in the –x-axis direction and the tricyclic G in the +x-axis direction,
thereby completing the intercalation (Figure 5.28). IN contrast, in the dGSC intercalation,
the bases remained stacked and the terminal and penultimate base pairs formed a
staggered conformation with the cytosine bases moving away from the helix strand in the
+z-axis direction. The penultimate cytosine then dropped to reestablish hydrogen bonding
with the penultimate tricyclic G. This allowed the terminal tricyclic G to slip into the void
created by the vacating penultimate cytosine and intercalate between the cytosine bases.
Figure 5.28. Base intercalation by dGNHC.
Once the 10-mer MD simulations were completed, the data from the strands was
abstracted. It was soon realized that due to the curvature of the DNA, several of the
parameters such as helix pitch and major groove width could not be resolved without
longer DNA strands. It was estimated that the minimum length for one turn of the extended DNA was approximately 14 base pairs. However, since the terminal ends of the
124 DNA strands in the simulations are unstable and are inappropriate for data extraction, it
was determined that a minimum length of 16 base pairs was necessary for the simulation.
In addition, incorporation of a larger buffer would allow the extraction of more data
points, so 20-mer strands of AXT and GXC were simulated. The base sequences are shown in Figure 5.29. The simulations produced DNA strands that contained 15 base
pairs per turn for most of the strands, and all strands were stable for the complete 4 ns
simulation.
As shown by the data in Table 5.2 on page 126, the visualized measurement of
bases per turn is at a minimum of 13 for the pyrrole extended purines while the others are
15. In order to validate this finding, the bases per turn value was recalculated from the
helix twist where one rotation is 360 degrees. The same value was observed suggesting
that these values are within reason. The bases per turn was also calculated using the helix
pitch per base rise, however the value obtained was considerably less than the other two measurements. Since the base inclination was not considered for that calculation, we
believe addition of that value may allow the results to be more similar with the others.
Figure 5.29. Modeled 20-mer DNA strands.
125 Table 5.2. Modified DNA results. dbA Tstd devdbA Tstd devdbAT std dev dbG Cstd devdbG Cstd devdbGCstd dev B-DNA O NH S O NH S Helix handedness Right Right Right Right Right Right Right bp/repeating unit 1 1 1 1 1 1 1 bp/turn 10 15 1 13 1 15 1 15 1 13 1 15 1 Helix twist, (º) 36 24.53 6.00 25.68 6.08 24.45 3.10 26.66 4.15 30.09 10.61 25.61 2.83 Rise/bp, (Å) 3.4 3.60 0.19 3.62 0.19 3.60 0.31 3.49 0.15 3.53 0.14 3.45 0.08 Helix pitch, (Å) 34 42.55 2.88 38.66 3.19 44.64 2.39 38.08 2.64 45.41 3.02 41.15 2.28 Base pair inclination, (º) 2.4 14.51 4.95 13.89 5.35 11.39 3.44 16.54 3.98 10.11 6.31 14.03 4.52 P distance from helix axis, (Å) 9.4 12.86 1.52 12.97 1.17 13.76 0.82 13.58 0.88 11.81 0.98 13.65 0.95 X displacement from bp to helix axis, (Å) 0.8 -5.53 1.22 -5.67 0.88 -6.33 0.72 -6.75 0.63 -4.76 1.14 -6.92 0.49 Glycosidic bond orientation anti anti anti anti anti anti anti Sugar conformation* C2'-endo C2'-endo O4'-endo C2'-endo C4'-exo C2'-endo C4'-exo C1'-exo C1'-exo C1'-exo C1'-exo C3'-endo Major groove depth 8.5 10.31 1.07 10.20 0.99 10.04 0.97 10.24 1.10 9.89 1.21 9.97 1.10 width (Å) 11.7 18.47 2.85 16.84 1.97 22.00 1.80 15.81 1.87 17.99 2.30 18.01 1.88 Minor groove depth 7.5 4.39 0.43 4.37 0.39 4.09 0.38 5.76 1.09 5.61 0.98 5.39 0.97 width (Å) 5.7 6.82 1.03 6.99 0.92 6.24 0.67 7.81 0.72 7.50 1.19 7.14 0.85 C1'-C1' distance (Å) 10.7 10.70 0.17 10.74 0.19 10.49 0.13 11.05 0.08 11.05 0.08 10.80 0.07 Diameter P-P (Å) 18.40 26.61 1.51 26.43 1.24 27.59 1.02 27.80 1.06 23.78 1.26 28.22 1.00
* Range of conformations but majority conformation is stated
126 The base inclinations of the expanded DNAs are much steeper compared to
natural B-DNA and the inclination is closer to that of the A-DNA form, which is 12º. The
greater inclination can be attributed to the increase in base length and possibly to the curvature of the bases themselves. Also, it has been shown that an “A-DNA” formation
can be attributed to a negative slide and a positive roll movement of the bases in the
strand.289 Interestingly, this increase in inclination has not been seen in previous extended
bases.240 The extension of the bases also increased the diameter of the helix. The increase
in diameter from P-P ranges from 5.4 to 9.8 Å wider than B-DNA and is dependent on
the heteroatom in the ring. The increased diameter along with the curvature of the bases,
displaced the bases from the x-axis from -4.75 to almost -7 Å. This displacement is
greater than the displacement for A-DNA, which is -4.1 Å.
The sugar pucker of the ring for the natural B-DNA is the C2’endo envelope
configuration although other configurations are seen. For the extended DNA strands,
several sugar conformations were observed but the majority proved to be the C2’endo-
C1’exo twist conformation. This is similar to the puckering observed in B-DNA.
The base pairs of the extended DNA all align correctly with respect to each other.
There is almost complete overlap of the pyrimidine rings from adjacent bases (Figure
5.30 on the following page) when the tricyclic base on the 3’ to 5’ strand is above the
tricyclic base on the 5’ to 3’ strand. When reversed, the tricyclic bases stack fairly well
with their adjacent bases. This overlap is not observed in normal DNA (Figure 5.31 on page 128) thus it is likely this overlap will impart greater stability to the helix.
127
Figure 5.30. Expanded base pair stacking.
As expected, the major groove widths and depths increased, however the depth of the major groove only increased by 1.4- 1.8 Å, which is reasonable since the extension of the bases was within that range. The width of the major groove increased dramatically however, ranging from 4.1 to 10.3 Å. This large change can be attributed to the curvature of the bases, which moved the phosphate backbone away from the major groove and closer to the minor groove thereby increasing the width. The minor groove width
128 expanded by a nominal amount, 0.5 to 2.1 Å, while the depth decreased by 1.7 to 3.4 Å.
The decrease in depth was not surprising and is likely due to the curvature of the tricyclic
bases towards the major groove.
Figure 5.31. Normal DNA stacking.
The distance from the C1’ atoms of base pairs across the minor groove was also
measured and it was found that the distances averaged between 10.49 to 11.05 Å, which is well within the range of the B-DNA distance of 10.7 Å. In contrast, the minor groove depths and widths are well within reason, and as such, minor groove recognition and
129 binding by enzymes should be plausible. In addition, the increase in the width and depth
of the major groove should enable major groove binders and enzymes greater access to
the major groove. This, along with the additional hydrogen bonding capabilities of the heteroatoms in the spacer rings should provide greater recognition.
Conclusion
The computational studies carried out to date have suggested that theoretically, as predicted, these systems display increased levels of stacking, polarizability, and stability, as well as increased helical width and greater base overlap as compared to B-DNA.
Intercalation of the bases, or use as dangling ends are also intriguing possibilities for these expanded base pairs that could prove useful. The increased polarizability should help stabilize the bases in both cases, and as a result, stabilize the overall helix.
Experimental results are presently underway, and will be reported elsewhere, however, it is hoped that those results will validate the theoretical predictions described herein.
Experimental Methods
All tricyclic base potentials were calculated using Gaussian03 Revision C.01 at the B3LYP/6-31G** level. All single point energies were corrected for zero-point energies from harmonic frequency analysis of the optimized structures. The DNA strands were built and modified in InsightII. All Watson-Crick hydrogen-bonding pairs were
aligned prior to steepest decent optimization except for the initial ATT strand. All
modified base single point energies were added to the AMBER base recognition file. The
bonds, bond angles and dihedral angles for the furan, pyrrole and thiazole rings were
130 added to the AMBER forcefield parameter file from predetermined ab initio
calculations.290,291 Additional bond, bond angle and dihedral angle terms were added to
the parameter file using data from ab initio calculations along with parameters existing in
the parameter file that were modified for different atom types. The modified DNA strands were entered into AMBER and the program added adenosine or guanosine bases were removed.
Green = EKTOT, Red = ETOT, Black = EPTOT
Figure 5.32. DNA energy stability check.
After checking for parameter errors, sodium counter ions were added to neutralize
the phosphate backbone charge and the strand was surrounded with an 8-Å layer of
131 water. The complex was refined using restrained minimizations followed by 400 step free minimizations. The process was allowed to continue until restrained minimizations ran for no longer than 2000 steps. All MD simulations were run with a 1-fs time step for simulation times no shorter that 4 ns. The potential, kinetic and total energies and other variables were checked for stability. The plot of the potential (EPTOT), kinetic (EKTOT) and total (ETOT) energy is displayed in Figure 5.32.
The root mean square deviation of the backbone was then abstracted from the coordinate file. Once backbone stability was determined, for the 10-mers and 20-mers in this study usually about 1.5 ns and 2 ns respectively, DNA structures were averaged together where the backbone movements were equal to or less than 1 Å difference Figure
5.33 on the following page. Measurements were then taken from the averaged structures and the data was averaged to produce the results found in Table 5.2 on page 126.
132
Figure 5.33. Root mean square distance movement of backbone from dbGSC.
133 CHAPTER 6
CHLORINATED ADENOSINE DERIVATIVES
Halogenated substitutions on purine and pyrimidine nucleosides and the removal
of nitrogen to form deaza nucleosides have been successfully used as treatments for
various forms of cancer. The nucleoside analogue 8-chloroadenosine has been shown to
inhibit cell growth in a variety of cancer cell lines and 3-deazaadenosine is an extremely
potent inhibitor of SAHase. By combining the potent inhibitory properties of chlorinated
adenine analogues and 3-deazaadenine it was hoped that a synergistic effect would be produced. The results of these efforts are presented in this chapter.
Background and Significance
Cancer is presently the second greatest killer of human beings, second only to heart disease.292 Cancer is a disease in which cells replicate without limitations and with
indifference to positive growth signals, as well as disregard for growth inhibitory
factors.293 Cancer cells contain all of the “genetic” information of normal cells but exhibit abnormal gene expression.2 The cells evade programmed cell death (apoptosis), sustain angiogenesis, and have the ability to invade normal tissue and metastasize. Due to the similarity to normal cells, drugs that interact with cancer cells also interact with normal cells, therefore creating significant problems with selectivity.2
The first known written record of cancer was documented over 4000 years ago by
the Egyptians.294,295 Their writings mention benign and malignant tumors and also
treatments and surgeries. After the fall of the Egyptian empire, it was not until the ancient
134 Greek philosophers, Hippocrates and Galen, “transformed” medicine from magic, superstition and religion to a more scientific outlook. Hippocrates is credited with naming the disease “Karkinoma” (carcinoma), which is Greek for crab, since tumors resembled crabs.295 In the middle ages, Europe acquired the Greek writings and it was believed that
cancer was the result of excess black bile and was curable only during its early stages.
This theory continued until the 18th and 19th centuries when a better understanding of
cancer was known.
Today it is known that there are over 100 different types of cancers, and in all
forms, abnormal methylation of nucleotide bases have been found.
DNA Methylation
Methylations are essential processes that occur throughout many biological systems, such as phospholipids, proteins, DNA, RNA, and small molecules.44
Methylation of proteins is also correlated to either cell mortality, repair functions or stress adaptations.44 In addition methylation can provide protection against retroviral elements, transcriptional noise44,296 and for recognition of m-RNA by enzymes (Figure
6.1).122,297,298
Methylation is also used to control gene expression in DNA.299 Ultimately, these
functions are involved in maintenance and defense of the genome and could compromise
the effectiveness of gene therapy.296 Although DNA methylation plays an essential role in
many systems, abnormal methylation patterns have been observed in cancers.293
Increasing evidence has shown that methylation of the promoter regions of several genes resulted in the subsequent failure of the genes to express their proteins.293
135
Figure 6.1. 5’ Cap structure of m-RNA.122,297,298
Most organisms, from bacteria to humans, have some heterocyclic bases that are
methylated.265 For example, in bacteria and some lower eukaryotes, both 6-
methyladenine and 5-methylcytosine have been found,265 however, in higher eukaryotes
and vertebrates only 5-methylcytosine has been observed in DNA.265 Most methylated
cytosines exist in a CG dinucleotide sequence, however methylation in some methylated
sequences CC, CT and CA have also been reported.265 Until recently, methylated
sequences other than CG were poorly understood, but findings have suggested that they
are concentrated at the origins of replication.265 Most of the research on DNA
methylation has been carried out on CG dinucleotide pairs since they play a vital role in
cancer. Nonmethylated CGs are distributed in a gene- and tissue- specific manner,
forming specific methylation patterns.300 It is well established that methylation regulates gene expression, parental imprinting and tissue specific gene expression.265
Methylated bases are not incorporated into DNA by replication machinery but rather, are modified after replication.265 DNA methylation is an enzyme-induced process,
in which methyltransferases control the conversion of cytosine to 5-methylcytosine. 5-
136 Methylcytosine possesses a methyl group (–CH3), covalently bonded to the C-5-position on a cytosine base (Figure 6.2).293 The methyl group is provided by S-
adenosylmethionine (SAM), which is subsequently converted to S-adenosylhomocysteine
(SAH). Biological methylation of DNA only occurs on cytosine bases and also primarily
to those linked directly to guanine by a phosphodiester linker forming a CpG dinucleotide
pair.293
Figure 6.2. Unmethylated and methylated cytosine residue.
Studies have indicated that over 70% of cytosine (C) bases in the body which are
covalently linked to guanosine (G) bases by a phosphodiester linker are methylated.265,296
It has been found that certain nonmethylated CG pairs result in gene expression, whereas methylated CG pairs give rise to genes that are not expressed.265 Once methylated, the
cytosine base can be readily deaminated to thymine.296 Consequently, defective repair of
this mistake can lead to carcinogenic lesions.
The CpG base pairs are commonly found in discrete “islands”, 0.5-2 kb long, which are found in the 5’-region of approximately 60% of “housekeeping” genes.293 The term “CpG islands” describes a region containing more than 197 bp with the cytosine/guanosine content above 0.5 and an observed CpG content above 0.6.293
137 Statistically, the prevalence of the CG combination is theoretically expected to be 6%
(1/16), however, it is experimentally found to be close to 1%.293 Localized concentrations
of CpG repeat sequences ranging from several hundred to a few thousand have been seen
in the promoter regions of many common genes, and in particular, ones associated with
tumor suppression.293 Substantial evidence of localized hypermethylation has been linked
with gene inactivation in several studies (Table 6.1).293
Table 6.1. Some genes involved in cell proliferation and methylated in cancer.293 Gene Function Cancer p16 (CDKN2A) Cell cycle control Esophagus, gastric, colorectal, pancreas, lung, bladder, ovary, breast, melanoma p15 (CDKN2B) Cell cycle control Leukemia MlH1 (HNPCC) Mismatch repair Gastric, colorectal, endometrium, ovary THBS1 Angiogenesis Colorectal (Thrombospondin-1) inhibition CDH-1 (E-Cadherin) Metastasis inhibition Breast, thyroid TIMP-3 (Tissue inhibitor Metastasis inhibition Kidney, Brain, breast, colon, lung MP3) ER (estrogen receptor) Growth suppression Colorectal, breast, lung, leukemia, prostrate AR (Androgen receptor) Growth suppression Prostrate
Methylation of the promoter region of mismatch repair gene MLH1 results in the
failure to produce a functional protein and impairs the cells’ ability to repair mismatches
that occur during proliferation.293 The methylation is due to microsatellite instability,
which is caused by a short polymorphic repeating segment of DNA between 1 to 4 base
pairs and is distributed across the genome. Microsatellite instability has been noted in
13% of all sporadic cases of colorectal cancer and has been found in all cases of
hereditary colon cancer,293 however, no mutation abnormality has been shown, but
138 hypermethylation and loss of MLH1 expression does occur. It is important to note that
hypermethylation of the p16 promoter region is the most widely reported epigenetic event
to occur in the development of human cancers.293
The expression of genetic information within an individual cell dictates cellular
behavior.293 Modifications at the molecular level can influence the function of a cell thereby changing its cellular processes and ultimately may fail to express functional proteins. Methylation can suppress gene expression even in small amounts as long as they are in the promoter region of a gene.293 Increased occurrence of methylation of genes in aging normal tissue suggests that methylation may be an early event in the formation of cancer, albeit not the only event.293 Researchers have shown that epigenetic events also
occur but it takes two deactivations, termed as “hits”, one on each chromosome at their
genetic copy to inactivate a gene.293 The hits can be both epigenetic or methylation or a
combination of both to suppress the expression of the gene or to produce an inactive
protein (Figure 6.3). This has been observed in several cancers in the past which include
RB and retinoblastoma, APC and colorectal cancer, VHL (von Hippel-Lindau) and renal
cancers and BRCA1 and ovarian and breast cancer.293
Altered methylation patterns have been observed in the “hit” DNA of cancer
cells.293 There are two types of methylation; hypomethylation, which consists of wide
areas along the genome, and hypermethylation which is localized at specific sites within
the gene promoter regions known as CpG islands.293 In hypomethylation, which is
essentially a decrease in methylations, a normally silenced gene may be expressed such
as a protooncogene, which can induce cell proliferation events, ultimately leading to
cancer. Although to date hypomethylation has not been implicated in cancer, the
139 possibility still exists. Alternatively, in hypermethylation, a normally expressed gene can
be silenced. If the methylation occurs in the promoter region of a tumor suppressor gene,
the gene can be silenced and therefore the cell will be unable to suppress proliferation.
Figure 6.3. Epigenetic deactivation process.293
Abnormal methylations can also lead to point mutations, as well as decreased chromosomal stability, causing the synthesis of dysfunctional proteins.293 For example, methylated cytosines have a greater chance of undergoing spontaneous deamination to afford thymine, thereby forming a point mutation. This results in the loss of control over cell proliferation, and has been observed in a majority of mutated p53 genes. In contrast, when conversion from cytosine to thymidine occurs followed by a transition of the complementary base from guanosine to adenosine, the series of changes is consistent with a methylation-induced mutational event.
140 Unfortunately, it remains unclear what induces hypermethylation of DNA in cells,
but researches have postulated 293 it is due to either, (i) over activation of a methylating
factor, or (ii) the loss of a demethylating factor, which causes excessive methylations.
Methylation Process
Figure 6.4. Methyltransferase mechanism.301
There are several methyl donors involved in a variety of biological pathways in
humans, however SAM is the most prolific methyl donor and is considered the most
important biologically.44,297 Methylations by SAM occur by way of various
methyltransferases (MeTase) by the mechanism seen in Figure 6.4.301 Enzymatic attack at the C6 carbon of a cytosine residue, followed by resonance stabilization by the
141 enzyme, initiates a nucleophilic attack by the C5 carbon of the cytosine residue.
Subsequent transfer of a methyl group from SAM then affords a cytosine tetrahedral
intermediate and a byproduct, SAH. Subsequent deprotonation and release of the enzyme
gives the methylated C5 cytosine residue.
As previously mentioned, the byproduct of all SAM methylations is SAH.
Concentrations of SAM and SAH are regulated such that SAH must be removed in order
for additional methylations to continue. As a result, SAH is considered a potent
competitive inhibitor of all methylation reactions dependent upon SAM as the methyl
donor. Studies have shown that cells are unable to remove SAH as a whole molecule
through the plasma membrane, therefore it must be removed by another process.121 In mammals, SAHase is the only known enzyme to break down SAH into adenosine (Ado) and homocysteine (Hcy).297 The hydrolysis of SAH by SAHase requires assistance from
NAD+ an enzyme bound cofactor.127 The reaction proceeds in the hydrolytic direction because Ado and Hcy go on to be used in other biological pathways; Ado can be deaminated to inosine by adenosine deaminase or phosphorylated by adenosine kinase to
AMP. Hcy is either remethylated to methionine or converted to cystathionine by condensation with serine. Inhibition of SAHase is essentially a cofactor depletion mechanism, since the inhibitor competes with the normal substrate Ado for the NAD+.
Since there is a dependence on this enzyme for SAH removal, inhibitors of SAHase can indirectly inhibit DNA methyltransferase, thereby initiating a biofeedback mechanism that suppresses the methylation process Figure 6.5.302-304
142
Figure 6.5. Mechanism of SAHase inhibition.
Although methylations are needed for normal cellular processes, it has been observed that toxicity to normal cells due to inhibition of DNA MeTase does not occur,300 however, it does affect tumor cell growth. One possible explanation for this is that p21, which is a cell growth regulator, is not expressed in normal growing cells during the S phase and as a result, inhibition of DNA MeTase will not affect those cells. In contrast, in cancer cells p21 is expressed, but the growth inhibitory affect of p21 is over ridden with the affect of DNA MeTase.300
Nucleoside activity against cancer.
Halogenated substitutions on nucleosides have been successfully used as
treatments for various forms of cancer. Current cancer treatments include the use of
surgery, radiation and/or chemotherapy using 5-fluorouracil (5-FU) (Figure 6.7) and
leucovorin, or 5-FU plus levamisole (Figure 6.6).305
143
Figure 6.6. Current cancer chemotherapeutic drugs.305
Analogues such as fludarabine, gemcitabine and 5-FU in Figure 5.7 are all potent inhibitors. However, these nucleoside analogues have exhibited cytotoxic effects due to conversion to their 5’-phosphoralated derivatives.306
Figure 6.7. Some halogenated nucleosides.306
It has been reported that the nucleoside analogue 8-Cl-adenosine inhibits cell growth in a variety of cancer cell lines, including breast, ovarian, pancreatic, and colon.307,308 It has also shown potential for use against multiple myelomas and
leukemias.306,309 Related to this, the nucleoside 2-chloroadenosine is a potent inhibitor of
hairy cell leukemia and active against lymphoid neoplasms.310,311 This suggests that
chlorination of adenosine could be significant for inhibition of transformed cell growth;
144 however it is unclear if additional chlorinated substituents on the heterocyclic moiety would further enhance the anti-proliferative properties of the nucleoside.
As stated previously, 3-deaza nucleosides have exhibited extremely potent inhibitors of SAHase and these modified nucleosides are active against a variety of viruses and cancers. They also do not undergo phosphorylation and are resistant to deamination.
Based on these observations, sights were set on the design and synthesis of a series of chlorinated 3-deazaadenine analogues with the goal of producing a synergistic biological effect. The analogues would follow the chlorination pattern of compounds 1 through 7 as shown in Figure 6.8.
Figure 6.8. Chlorinated adenine analogues.
Synthesis
By combining the potent inhibitory properties of 3-deazaadenine and chlorinated adenine analogues (Compounds 1 through 7 Figure 6.8) it was hoped that a synergistic
effect would be produced. The target bases are shown below in Figure 6.9.
145
Figure 6.9. Chlorinated 3-deazaadenine analogues.
Starting with commercially available 2,6-dichloropyridine (15), the N-oxide (16)
was synthesized using 35% H2O2 and refluxing trifluoroacetic acid (TFA) (Scheme 6.1).
Initially, a procedure as described by Rousseau and Robbins 312 was followed, but upon further investigation it was discovered that increasing the reaction time to 6.5 hours increased the yield by at least 20%. Surprisingly however, any greater increase in time over
6.5 hours did not increase the yield. Since TFA is expensive and the amounts required for macroscale synthesis were significant, the possibility of decreasing the quantities was explored (Table 6.2). The results showed that a 50% decrease in acid provided a 20-25%
loss in yield, whereas a 25% decrease in TFA resulted in a 10-15% decrease in yield, which
proved to be an acceptable compromise.
The N-oxide (16) was then subjected to nitration to form 17,313 which was
subsequently reduced to the amine (18) with refluxing glacial acetic acid and iron powder
Scheme 6.2.
146 Scheme 6.1. Dichloro to N-oxide.
Reaction conditions: a) TFA, H2O2 30%.
Table 6.2. Difference in TFA volumes and product yields. TFA volume (mL) % total TFA reaction time (hrs) % yield % difference 485 100 6.5 88 0 485 100 6.5 86 2 485 100 6.5 79 10 485 100 6.5 76 14 485 100 8.0 86 2 364 75 8.25 81 8 364 75 6.5 74 16 363 75 6.5 74 16 267 55 6.5 67 24 245 51 6.5 80 9 245 51 6.5 65 26 244 50 6.5 60 32 243 50 6.5 49 44 245 51 7.0 61 31 243 50 7.0 60 32
Scheme 6.2. N-oxide to amine.
Reaction Conditions: a) HNO3, H2SO4, 160 ºC; b) Fe, acetic acid, reflux.
147 The amine was then converted to the nitrosoamine (19) using standard conditions,
which, following treatment of 19 with acid resulted in migration of the nitro group to give
the nitro amine (20) in quantitative yields.312 Reduction of the “migrated” nitro group by
the same conditions313 as were used for the reduction of 17, was only moderately successful. The byproduct formed, albeit in moderate quantities, proved to be the acetylated diamine (22) Scheme 6.3. Attempts at deprotection via concentrated ammonium hydroxide were unsuccessful, however, NaOH 30% (w/v) proved fruitful, and conversion of 22 back to the desired diamine (21) was accomplished quantitatively.
Scheme 6.3. Amine to diamine.
Reaction conditions: a) HNO3, H2SO4; b) H2SO4, 100 °C; c) Fe, acetic acid, reflux; d) Fe, HCl, EtOH, H2O, 95 °C; e) NaOH 30%, reflux.
Unfortunately, this was not the only problem encountered with the reduction of 17;
the work-up proved tedious due to the heavy emulsions consistently formed during the
extraction process, hence a search for an alternative method was undertaken. It was
discovered that using iron, ethanol, and aqueous HCl314 under refluxing conditions gave
148 equivalent yields as had been previously realized, but with a much more facile workup. No formation of emulsions or the unwanted acetylated byproduct were observed. As a result, the procedures were subsequently changed for both 17 and 20 to employ the alternative procedure.
Ring closure of the diamine (21) was accomplished with refluxing formic acid to afford 2,6-dichloro-3-deazaadenine, (11). However, poor to moderate yields, in addition to the formation of a diamide (23) as a major byproduct (Scheme 6.4) made this method unattractive. All attempts to convert 23 back to 21 were unsuccessful, but upon treatment with potassium hydroxide, a new heterocyclic base 24 was formed, which, to the best of our knowledge, has not been previously reported. Ring closure of 21 was finally achieved using triethylorthoformate and acetic anhydride in a 1:1 ratio to give 11 in excellent yield with no evidence of the byproduct 23.312
Scheme 6.4. Diamine to 8-oxo.
Reaction Conditions: a) formic acid, reflux; b) KOH, reflux.
149 Next, 2,6-dichloro-3-deazaadenine, (11) was selectively converted to 2-chloro-3- deazaadenine (9) with methanolic ammonia at 160 oC (Scheme 6.5), which was then
subjected to standard hydrogenation conditions to produce the 3-deazaadenine base (25).43
Scheme 6.5. Diamine to ring closed amine.
Reaction Conditions: a) Triethylorthoformate/acetic anhydride 1:1, reflux; b) NH3, MeOH, 160 °C; c) Pd/C, NaOH 10%, H2O, H2 34 psi, 20 h.
Several different standard chlorination methods were then attempted in an effort to
convert 25 to 8-chloro-3-deazaadenine (8), as well as to convert 9 to the desired 2,8-
dichloro base 10, and 11 to the trichloro base 14 (Scheme 6.6).315-317 Unfortunately, none of the conditions employed provided 14. Conversion of 25 to 8-chloro-3-deazaadenine (8) and 9 to 2,8-dichloro-3-deaza (10) was poor at best. Many attempts at modifying reaction conditions failed to produce any improvement in the yields. Furthermore, in several cases, an intractable mixture of products formed for which the isolation and purification of the
desired product could not be achieved. The conditions that afforded the highest yield and
most facile purification proved to be t-butyl hypochlorite in DMSO, producing a yield of
31%.317
150 Scheme 6.6. First 3 chloro bases.
Reaction Conditions: a) HCl·DMA, m-CPBA; b) n-BuLi, diisopropylamine, p-TsCl; c) DMSO, t-butyl hypochlorite.
When 3-deazaadenine (25) was treated with m-CPBA, a mixture of products
formed, including 8-chloro- and 2-chloro-3-deazaadenine (8 and 9, respectively) neither of
which were recovered in significant quantities.315 In addition, another chlorinated base was
isolated Scheme 6.7, which proved to be 3-chloro-3-deazaadenine (26). To our knowledge,
this is the first report of this particular chlorinated base, and is likely to be the result of the
formation of an epoxide by the m-CPBA at the C-2, C-3 double bond of 3-deaaadenine.
The epoxide, which is susceptible to nucleophilic attack by chlorine from either side, can
then undergo a facile elimination of water to reestablish the aromaticity of the ring system
(Figure 6.10 on page 153).
151 Scheme 6.7. 3-Chloro base.
Reaction Conditions: a) DMA, HCl·DMA, m-CPBA
Comparison of the 1H NMR spectra of 26 with 8-chloro-3-deazaadenine (8) and
2-chloro-3-deazaadenine (9) showed a significant difference; 8-chloro-3-deazaadenine
(8) exhibits two doublets centered at 7.82 ppm and 6.98 ppm for the C-2 and C-3 protons
and a singlet at 6.63 ppm for the amine protons. The doublets are coupled to each other
with J-values of 5.7 Hz, and integrated to 1:1:2 respectively. For 2-chloro-3-deazaadenine
(9), literature values showed singlets at 6.58 ppm for the amine protons, 6.99 or 7.03 ppm
for the C-3 proton and 8.37 or 8.30 ppm for the C-8 proton, and our spectra agreed with these values. In contrast, the 1H NMR spectrum for 3-chloro-3-deazaadenine (26)
exhibited signals at 8.16 ppm for the C-2 proton 7.63 ppm for the C-8 proton and 6.32
ppm for the amine protons with an integration of 1:1:2. The C-2 proton for 26 is shifted
downfield in comparison to the C-2 proton for 8-chloro-3-deaza-adenine due to the
presence of the C-3 chlorine and the proximity of the pyridine nitrogen. The C-3 proton
in the 2-chloro-3-deazaadenine is not shifted as far as the 3-chloro since the proton is
further from the pyridine nitrogen.
152
Figure 6.10. Proposed Mechanism of chlorination via epoxide formation.
Next, in an effort to utilize the byproduct 24 produced earlier in Scheme 6.4, an alternative route to the trichloro base (14) was pursued, but all attempts to chlorinate 24
318 319 using either POCl3 or SOCl2 proved unsuccessful (Scheme 6.8). Selective conversion
of the 6-chloro substituent of 24 to form the amine (27), was then tried with sights set on
transformation to 10 with methanolic ammonia, followed by chlorination. Unfortunately,
standard ammonolysis conditions were unsuccessful even after 150 hours at 140 ˚C; only
starting material was recovered, therefore this route was also abandoned.
At this point a report in the literature was found,320 which postulated that electron-
donating groups must be present on the base in order to activate the 8-position towards
chlorination. This would explain many of the difficulties, since loss of the N-3 nitrogen for
the ring system would render the base much less reactive to substitution than is normally
seen with the parent adenine. This theory is substantiated by two observations; one, that in
153 all cases, the identical reaction conditions described herein have proven successful for the
parent adenine ring system. Also, while the chlorination of the C-8 from the amino-
substituted 3-deaza precursors was successful, albeit in low yields, there was a complete
lack of reactivity for the more deactivated dichloro precursors. Due to the low yields and
inability to chlorinate half of the product bases, further research on this project was
discontinued.
Scheme 6.8. Alternative route.
Reaction conditions a) tetraethylammonium chloride, POCl3; b) SOCl2, DMF; c) methanolic ammonia, 140ºC.
Experimental:
General. Melting points were recorded on a Meltemp II melting point apparatus
and are uncorrected. Combustion analyses were performed by Atlantic Microlabs, Inc.,
Atlanta, GA. 1H and 13C NMR spectra were recorded on a Varian 300 spectrometer
154 (operated at 300 and 75 MHz, respectively) all referenced to internal tetramethylsilane
(TMS) at 0.0 ppm. The spin multiplicities are indicated by the symbols s (singlet), d
(doublet), t (triplet), m (multiplet) and b (broad). Reactions were monitored by thin-layer
chromatography (TLC) using 0.25-mm Whatman Diamond silica gel 60-F254 precoated
plates with visualization by irradiation with a Mineralight UVGL-25 lamp. Column
chromatography was performed on Whatman silica, 200-400 mesh, 60 Å, and elution
with the indicated solvent system. HPLC purification was carried out on a Hewlet-
Packard 1090 liquid chromatograph. Yields refer to chromatographically and
spectroscopically (1H and 13C NMR) homogeneous materials.
Preparation of 2,6-dichloropyridine-N-oxide (16): A mixture of trifluoroacetic acid
(485.0 mL, 6.3 mol), 2,6-dichloropyridine (48.0 g, 324.0 mmol) and 35% hydrogen
peroxide (85.0 mL, 971.0 mmol) were heated on a stream bath for 6.5 hours. After the
solution had cooled to room temperature, 2.4 L of water was added. The flask was stored at
0 ºC overnight. The resulting precipitate (unreacted 2,6-dichloropyridine) was removed by
filtration and the yellow-orange filtrate was evaporated to a reduced volume. Chloroform
(500 mL) was added and the solution was treated with anhydrous potassium carbonate until
carbon dioxide evolution had ceased. After filtration, the filtrate was evaporated to dryness and dried under vacuum for 3 days to give the product as light yellow crystals (46.6 g,
1 13 88%). H NMR (DMSO-d6), 7.81 (d, 2H, H3, H5, J=8.1Hz), 7.32 (t, 1H, H4, J=8.1Hz); C
NMR (d6-DMSO) 158.58, 158.08, 141.96, 125.94, 125.52.
155 Preparation of 2,6-dichloro-4-nitropyridine (17): To 2,6-dichloropyridine-N-oxide (39.1
g, 238.0 mmol) was added 95% fuming nitric acid (63.0 mL, 1.5 mol) and 98%
concentrated sulfuric acid (153.7 mL, 2.9 mol). The mixture was heated at 148 oC for 2
hours and then the temperature was increased to 165 oC and heated until the evolution of
nitrogen dioxide had ceased and a dark green/black solution remained. After cooling to
room temperature, the mixture was poured onto crushed ice (300 g) and concentrated
aqueous ammonia (28%) was added with stirring, while maintaining the mixture
temperature below 40 oC (dry ice/acetone). After the solution had reached a pH of 6, the
yellow solid that precipitated was filtered washed with water until light yellow and dried to
1 13 produce the product (33.2 g, 72%). H NMR (DMSO-d6), 8.34 (s, 2H, H3, H5); C NMR
(DMSO-d6) 156.91, 150.85, 117.64.
Preparation of 4-amino-2,6-dichloropyridine (18): Iron powder (3.03 g, 54.18 mmol)
was added portion wise to a solution of 2,6-dichloro-4-nitropyridine (10.4 g, 53.9 mmol) in
glacial acetic acid (90.0 mL, 1.6 mmol) and the mixture was refluxed for 2.5 hours. After
cooling to room temperature, the reaction mixture was neutralized with 35% (w/v)
potassium hydroxide solution. The product was extracted with ethyl acetate (3 x 100mL).
The combined extracts were dried (MgSO4) and evaporated to produce the product as a
1 yellow solid (8.5 g, 97%). H NMR (DMSO-d6), 8.36 (s, 2H, H3, H5), 6.76 (s, 2H, NH2),
13 6.48 (s, 2 H); C NMR (DMSO-d6) 158.62, 150.84, 149.20, 117.70, 106.48.
Preparation of 2,6-dichloro-4-nitraminopyridine (19): Concentrated H2SO4 (83.5 mL,
1.6 mol) was carefully added to 4-amino-2,6-dichloropyridine (16.67 g, 102.27 mmol). The
156 mixture was cooled to 0 oC and 90% nitric acid (33.5 mL, 798.0 mmol) was added
dropwise while the inside temperature was maintained below 10 oC with an acetone/dry ice
bath. The solution was stirred for 2 hours at room temperature and then poured onto
crushed ice (335 g) and allowed to cool at 0 oC for 1.5 hours. The cream–colored
precipitate was removed by filtration, washed well with ice water and dried to yield the
1 product as a tan solid (21.3 g, quantitative). H NMR (DMSO-d6) 8.36 (s, 1H, NH), 7.47 (s,
13 2H, ); C NMR (DMSO-d6) 150.84, 149.91, 147.63, 117.73, 111.36.
Preparation of 4-amino-2,6-dichloro-3-nitropyridine (20): Concentrated H2SO4 (228.2 mL, 4.3 mol), was carefully added to 2,6-dichloro-4-nitraminopyridine (29.04 g, 139.63 mmol) and then heated on a steam bath for 2.5 hours. The solution was cooled to room temperature and poured onto ice (600 g) at which time a creamy brown precipitate was noted. The mixture was neutralized with concentrated ammonium hydroxide solution
(28%) while the temperature was maintained below 20 oC with an acetone/dry ice bath. The
mixture was cooled for 2 hours at 0 oC, filtered, and the product was washed well with ice
1 water to yield the product as a brown solid (27.95 g, 96%). H NMR (DMSO-d6), 8.36 (s,
13 1H), 7.65 (br s, 2H, NH2), 6.86 (s, 1H, H5); C NMR (DMSO-d6) 150.84, 150.70, 148.79,
142.08, 130.38, 117.72, 117.66, 110.19, 110.08, 99.44.
Preparation of 3,4-diamino-2,6-dichloropyridine (21): Iron powder (19.91 g, 356.58
mmol), water (65.8 mL, 3.6 mol), and HCl (14.1 mL, 464.1 mmol) were added
consecutively to a solution of 4-amino-2,6-dichloro-3-nitropyridine (14.82 g, 71.27 mmol)
in ethanol (350 mL). After stirring at 95 ºC for 16 hours, the reaction mixture was allowed
157 to cool to room temperature, neutralized, filtered, and then evaporated. Water (300 mL) and
ethyl acetate (300 mL) were added to the crude product. The water layer was extracted with
ethyl acetate (3 x 300 mL) and the organic layer was dried (MgSO4) and evaporated to
1 produce the product (11.9 g, 94%). H NMR (DMSO-d6), 9.20 (s, 1H), 7.62 (s, 1H), 6.49
13 (s, 2H), 6.44 (s, 1H); C NMR (DMSO-d6) 158.62, 158.53, 149.20, 117.70, 106.48.
Preparation of 3(4)-acetoamide-4(3)-amino-2,6-dichloropyridine (22): To 4-amino-2,6-
dichloro-3-nitropyridine (5.00 g, 24.04 mmol) in glacial acetic acid (70.0 mL, 1.21 mol) was added iron powder (7.09 g, 127.02 mmol). The mixture was refluxed for 2 hours. After cooling to room temperature, the mixture was neutralized with 35% (w/v) potassium hydroxide solution and filtered through a celite pad. The pad was rinsed with glacial acetic acid (50 mL) and water (100 mL) and the resulting filtrate was combined with the previous filtrate and neutralized again. The product was extracted with ethyl acetate (4 x 100mL) and the combined extracts dried (MgSO4), filtered and evaporated. The crude solid was
purified via column chromatography eluting with EtOAc:hexane (2:1) to give the product
o 1 as brown solid (2.2 g, 51%). mp 234.1-237.3 C. H NMR (DMSO-d6), 9.30 (s, 1H, H5),
13 6.65 (s, 2H, NH), 6.60 (s, 1H, NH), 2.00 (s, 3H, COCH3); C NMR (DMSO-d6) 168.95,
155.50, 148.18, 146.30, 115.33, 107.09, 22.77. Anal. Calc. for C7H7N3Cl2O: C 38.21, H
3.21, Cl 32.22, N 19.10, Found: C 38.55, H 3.25, Cl 31.89, N 18.91.
Preparation of 4,6-dichloroimidazo [4,5-c] pyridine (11): Formic acid (30.2 mL, 800
mmol) was added to 3,4-diamino-2,6-dichloropyridine (109.6 mg, 615 µmol) and the
mixture was refluxed for 19 hours. The mixture was cooled and evaporated under reduced
158 pressure to produce a brown solid. The compound was purified via column
chromatography eluting with EtOAc:MeOH (95:5) to produce the product as a light brown
solid (63 mg, 54%) whose mp was in agreement with literature value. 1H NMR (DMSO-
13 d6), 8.52 (s, 1H, H2), 7.73 (s, 1H, H7); C NMR (DMSO-d6) 167.85, 149.85, 146.37,
139.73, 110.84, 108.29.
Preparation of 3,4-diamido-2,6-dichloropyridine (23): Formic acid (75.6 mL, 2.0 mol)
was added to 3,4-diamino-2,6-dichloropyridine (2.06 g, 11.54 mol) and the whole mixture was refluxed for 72 hours. The mixture was cooled, evaporated under reduced pressure and purified via column chromatography eluting with EtOAc:MeOH (95:5) to afford the
1 product as a light brown solid (1.02 g, 38 %). H NMR (DMSO-d6), 8.84 (s, 2H, CHO),
13 6.93 (s, 1H, H2), 6.83 (s, 2H, NH2); C NMR (DMSO-d6) 168.74, 155.12, 150.40, 141.51,
141.16, 132.19, 123.05, 106.43, 96.88.
Preparation of 2,6-dichloro-8-oxo-3-deazaadenine (24): Aqueous KOH 35% w/v (15 mL) was added to a crude mixture of 3,4-diamido-2,6-dichloropyridine (1.02 g, 4.36 mmol), refluxed overnight and then allowed to cool. The mixture was neutralized with glacial acetic acid, extracted with EtOAc (3 x 200 mL) and evaporated. The crude solid was purified via column chromatography eluting with CH2Cl2:MeOH (95:5) to produce the
o 1 product as a light brown solid (500 mg, 49%). mp 176.5 –181.6 C; H NMR (DMSO-d6),
13 6.84 (s, 1H, H2), 6.74 (s, 1H, H7), 6.48 (s, 1H, H9); C NMR (DMSO-d6) 168.02, 158.61,
149.85, 149.20, 110.86, 106.54. HRMS Calc. for C6H3N3Cl2O: 202.96534, Found
202.96516.
159
Preparation of 2,6-dichloro-3-deazaadenine (11): To a mixture of triethylortho-
formate:acetic anhydride (1:1 volume, 50.0 mL) was added 3,4-diamino-2,6-dichloro
pyridine (4.64 g, 26.09 mmol)312. The solution was refluxed for 4.5 hours. The excess
reagents were removed in vacuo and the residue was dissolved in 10% NaOH (66 mL) and
warmed on a steam bath for 40 min. The resulting solution was cooled to room
temperature, neutralized to pH 6 with glacial acetic acid, and was cooled for 2 hours at 0
ºC. The resulting precipitate was removed by filtration and dried to produce the product as
a light brown solid (4.69 g, 96%); mp was in agreement with literature value. 1H NMR
13 (DMSO-d6), 8.52 (s, 1H, H2), 7.73 (s, 1H, H7); C NMR (DMSO-d6) 167.85, 149.85,
146.37, 139.73, 110.84, 108.29.
Preparation of 2-chloro-3-deazaadenine (9): To saturated methanolic ammonia stirring at
–78 oC (100 mL) was added 2,6-dichloro-3-deazaadenine (3.20 g, 17.04 mmol) and was
heated in a Parr bomb at 160 oC for 90 hours. The solution was evaporated to dryness and the crude solid was purified on a silica gel column eluting with EtOAc:hexane (2:1) to give
1 the product as a light brown solid (1.40 g, 49%). mp (dec.) 266 ºC. H NMR (CD3OD- d4),
13 8.05 (s, 1H, H2), 6.79 (s, 1H, H7), 4.92 (s, 2H, NH2); C NMR (DMSO-d6) 151.96,
142.67, 142.12, 115.56, 114.98, 98.48.
Preparation of 3-deazaadenine (25): To a solution of crude 2-chloro-3-deazaadenine
(1.38 g, 8.18 mmol) in water (239.0 mL) was added 10% NaOH (8.18 mL, 20.45 mmol) and 10% palladium on carbon catalyst (497.7 mg). The mixture was hydrogenated at 34 psi
160 for 20 hours. The solution was filtered through a celite pad and the pad was washed with
boiling H2O (200 mL). The combined filtrate was evaporated in vacuo to produce a crude
solid, which was dissolved in EtOH and filtered to yield the pure product as a yellowish
brown solid (790.3 mg, 72%) mp was in agreement with literature value. 1H NMR
(DMSO-d6), 8.09 (s, 1H, H8), 7.61 (d, 1H, H2, J = 5.7 Hz), 6.77 (d, 1H, H3, J = 5.7 Hz),
13 6.23 (s, 2H, NH2); C NMR (DMSO-d6) 151.25, 140.38, 139.59, 124.91, 99.43.
Preparation of 3-chloro-3-deazaadenine (26): DMA (10.0 mL) containing 3-
deazaadenine (268.5 mg, 2.0 mmol) was evaporated to dryness. Another 10 mL aliquot of
DMA was added and reduced to a light brown oil and cooled to room temperature. The oil
was then dissolved in 0.5 M HCl in DMA (4.0 mL, 2.0 mmol) and stirred for 7 minutes, at
which point m-CPBA (345.5 mg, 2.0 mmol) in DMA (2.0 mL) was added. The solution was stirred for an additional 6 minutes, and an additional amount of 0.5 M HCl in DMA
(4.0 mL, 2.0 mmol) was added. The solution was allowed to stir at room temperature for 25
minutes. The brown reaction mixture was evaporated and remaining traces of DMA were
removed by co-evaporation from EtOH:xylene (1:2, 2 x 8 mL). The gummy residue was
dissolved in MeOH (50.0 mL) and H2O was added. A tan precipitate formed, which was
filtered and washed with H2O. The combined filtrate and washings were extracted with diethyl ether (2 x 50 mL), the aqueous layer was neutralized 10% NaOH solution, and reduced under vacuum. The crude mixture was purified using column chromatography eluting with CHCl3:MeOH (97:3 to 95:5) to yield the product as a light tan solid (20.8 mg,
1 13 6.2%). H NMR (DMSO-d6), 8.16 (s, 1H, H3), 7.63 (s, 1H, H2), 6.32 (s, 2H, NH2); C
161 NMR (DMSO-d6) 151.00, 140.70, 137.50, 135.80, 127.05, 103.00; HRMS Calc. For
C6H5N4CL2: 201.98130, Found: 201.98131.
Preparation of 2,8-dichloro-3-deazaadenine (10): In a manner analogous to the method
used with (26), 2-chloro-3-deazaadenine (337.8 mg, 2.0 mmol) afforded the product as a
o 1 light tan solid (152.6 mg, 38%); mp 325.2 C (dec). H NMR (DMSO-d6), 8.18 (s, 1H, H3),
13 6.80 (s, 2H, NH2); C NMR (DMSO-d6) 133.37, 132.94, 132.74, 130.68, 128.86, 127.95
+ +2 +4 MS (EI) 201.9 (M , 100), 204 (M , 73), 206 (M , 17); HRMS Calc. for C6H4N4Cl2:
201.98130, Found: 201.98131.
Preparation of 8-chloro-3-deazaadenine (8): To a solution of diisopropylamine (1.5 mL,
11.0 mmol) in THF (14 mL) at 0 oC under Ar, was added dropwise n-BuLi (1.6 M in
hexanes, 6.3 mL, 10.0 mmol). The resulting solution was stirred for 15 min. To this a
solution of 3-deazaadenine (269.6 mg, 2.0 mmol) in THF (20 mL) was added slowly at –70
oC. After the mixture was stirred for 2 hours, a solution of p-toluene sulfonylchloride
(TsCl, 1.53 g, 8.01 mmol) in THF (4.0 mL) was added while maintaining the temperature below –70 oC. The reaction mixture was stirred for 1.5 hours, quenched with AcOH (460
µL, 8.0 mmol), and evaporated to dryness. The residue was placed on a silica gel column and eluted with (CH2Cl2 : EtOH, 97:3) to give the product (64.8 mg, 19.1% as a yellow
1 brown solid), which could not be separated from p-TsOH. H NMR (DMSO-d6), 7.82 (d,
13 1H, H3, J = 5.7 Hz), 6.98 (d, 1H, H2, J = 5.7 Hz), 6.63 (s, 2H, NH); C NMR (DMSO-d6)
139.44, 135.81, 133.43, 130.68, 127.36, 125.75; MS (EI) 168 (M+, 100), 170 (M+2, 34);
HRMS Calc. for C6H5N4Cl: 168.02027, Found 168.02057.
162 APPENDIX A
IsoA Data
Guide to the modeling data for IsoA
The data described herein is for the Isoadenosine project. The molecules are not placed in alphabetical order but rather; the first 7 molecules are named and listed according to their appearance in Chapter 2. The rest are also referenced according to their appearance but are numbered starting with IsoA#1. Each molecule image and data table for that image is placed on the same page for clarity. The molecular binding free energies in the tables are unadjusted but can be compared to each other. For the description of hydrogen bonding, the molecule naming and atom numbering scheme for
the isoadenosine molecules are as described in Figure A.1. The molecule is colored for
clarity.
Figure A.1. Isoadenosine modeling numbering scheme.
163 The ligand hydrogen bonding residues are labeled according to the molecule
attachment. For example base C6NH describes the nitrogen atom attached to the C6
carbon atom in the IsoA base. The bold letter indicates the atom that is hydrogen bonding
to the amino acid residue of the enzyme or water. Dihedral angles are measured from
atom numbers 2-3-1’-5’ for the base to sugar angle and the 5’-4’-6’-O6’ for the hydroxyl
angle.
In some cases, more than one structure is represented for a molecule. When a
computed molecule contains more than one free energy value within 1 kcal/mol, those
structures are considered of equal importance and therefore both are represented.
The molecules begin on the next page.
164
Figure A.2. Borchardt’s inhibitor, 4’,5’-enyl-3-deazaadenosine.
Table A.1. Modeling results for Borchardt’s inhibitor, 4’,5’-enyl-3-deazaadenosine. Relative energy (kcal/mol) -24.788 Dihedral angles (º) base to sugar -69.20 Hydrogen bond distance (Å) inhibitor/enzyme Lys186 Asp190 Ile 299 Asn346 His353 Met358 Ser361 water base N1 3.32 base C6NH 2.45 1.93 base N7 2.72 2.60 2'-OH 1.80 2.01 3'-OH 2.02
165
Figure A.3. Adenosine.
Table A.2. Modeling results for Adenosine. Relative energy (kcal/mol) -42.523 Dihedral angles (º) imidazole to pyrimidine 154.52 5' hydroxyl -165.99 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Glu59 Asp131 Glu156 His353 base C6NH 1.81 2.18 base N7 1.97 2.98 2.95 2'-OH 1.81 3'-OH 1.92 3'-OH 2.04 2.02 5'-OH 2.05
166
Figure A.4. Ari-1.
Table A.3. Modeling results for Ari 1. Relative energy (kcal/mol) -23.932 Dihedral angles (º) imidazole to pyrimidine -4.15 5' hydroxyl -176.12 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Glu59 Thr60 Asp131 Thr157 Asp190 base N3 1.88 2.77 2.78 base C6NH 1.86 1.84 1.96 2'-OH 1.82 2'-OH 1.85 3'-OH 1.86 5'-OH 1.90 1.91
167
Figure A.5. Ari-2.
Table A.4. Modeling results for Ari 2. Relative energy (kcal/mol) -24.075 Dihedral angles (º) base to sugar -14.58 5' hydroxyl -35.54 Hydrogen bond distance (Å) Ile inhibitor/enzyme His55 Thr60 Asp131 Thr157 299 Asn346 base N3 2.86 2.87 base C6NH 1.96 2'-OH 2.02 3'-OH 1.92 2.05 5'-OH 1.97 5'-OH 2.17
168
Figure A.6. NpcA.
Table A.5. Modeling results for NpcA. Relative energy (kcal/mol) -28.161 Dihedral angles (º) imidazole to pyrimidine -43.39 5' hydroxyl -62.99 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Glu156 Asn346 Leu347 water base C6NH 1.84 1.85 base N7 2.86 2'-OH 1.84 2.01 3'-OH 2.01 5'-OH 1.87
169
Figure A.7. 3-Deaza NpcA.
Table A.6. Modeling results for 3-deaza NpcA. Relative energy (kcal/mol) -24.561 Dihedral angles (º) base to sugar -68.18 5' hydroxyl -93.57 Hydrogen bond distance (Å) inhibitor/enzyme His55 Cys79 Asn 80 Ser83 Asp131 Glu156 Asp190 Asn346 base N1 1.89 2.97 1.96 base C6NH 1.98 1.81 base N7 1.96 1.94 2.67 2'OH 2.12 3'OH 2.01 1.90 5'OH 2.10 5'OH 1.84
170
Figure A.8. 4’,5’-enyl NpcA.
Table A.7. Modeling results for 4’,5’-enyl-adenosine. Relative energy (kcal/mol) -24.103 Dihedral angles (º) base to sugar 61.49 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Asp131 Asp190 His353 base N1 1.90 2.59 base C6NH 1.92 base N7 1.84 2'-OH 1.85 1.93 3'-OH 2.01
171
Figure A.9. Isoadenosine.
Table A.8. Modeling results for Isoadenosine. Relative energy (kcal/mol) -29.102 Dihedral angles (º) imidazole to pyrimidine 58.13 5' hydroxyl 169.22 Hydrogen bond distance (Å) inhibitor/ enzyme Leu54 His55 Thr57 Glu59 Asp131 Glu156 Asp190 His353 Met358 base N1 1.87 base C6NH 1.87 base C6NH 1.93 1.89 base N7 1.90 2.77 2.78 base N9 1.89 3.21 3'-OH 1.80 3'-OH 1.85 2.03 5'-OH 1.94 1.91
172
Figure A.10. IsoA#1.
Table A.9. Modeling results for IsoA#1. Relative energy (kcal/mol) -31.858 Dihedral angles (º) imidazole to pyrimidine 42.35 5' hydroxyl -167.07 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 Asp131 Thr157 Asn346 His353 water base N1 1.87 base C6NH 2.13 base C6NH 1.88 base N7H 1.87 3'-OH 2.06 5'-OH 1.95 1.80
173
Figure A.11. IsoA#2.
Table A.10. Modeling results for IsoA#2. Relative energy (kcal/mol) -18.670 Dihedral angles (º) base to sugar 56.50 5' hydroxyl -172.67 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu59 Asn346 Met351 base C6NH 1.97 1.92 base N9H 1.90 3'-OH 1.92
174
Figure A.12. IsoA#3.
Table A.11. Modeling results for IsoA#3. Relative energy (kcal/mol) -18.020 Dihedral angles (º) imidazole to pyrimidine -25.45 5' hydroxyl 149.91 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Thr157 Asp190 Asn346 Met358 base C6NH 1.88 base N7 3.46 base N9H 1.99 2.03 3'-OH 1.79 3'-OH 1.86 1.93 5'-OH 1.88 5'-OH 1.93 2.02
175
Figure A.13. IsoA#4.
Table A.12. Modeling results for IsoA#4. Relative energy (kcal/mol) -29.177 Dihedral angles (º) base to sugar -152.99 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu59 Asp131 Glu156 Thr157 His353 water base N1 2.15 2.77 2.80 1.81 2.70 base C6NH 2.19 base N9 2.15 2'-OH 1.91 2'-OH 1.95 3'-OH 1.73 5'-OH 2.03 5'-OH 1.89
176
Figure A.14. IsoA#5.
Table A.13. Modeling results for IsoA#5. Relative energy (kcal/mol) -26.713 Dihedral angles (º) base to sugar 159.54 5' hydroxyl 156.90 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Asp190 Asn191 Ser198 Met358 base N1 3.56 base C6NH 1.70 3'-OH 1.77 3'-OH 1.86 1.90 5'-OH 1.95
177
Figure A.15. IsoA#6.
Table A.14. Modeling results for IsoA#6. Relative energy (kcal/mol) -27.931 Dihedral angles (º) base to sugar 134.94 5' hydroxyl 60.63 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Asn346 base C6NH 1.64 base C6NH 1.91 base N9H 2.08 2.23 3'-OH 1.88 5'-OH 2.04
178
Figure A.16. IsoA#7.
Table A.15. Modeling results for IsoA#7. Relative energy (kcal/mol) -25.862 Dihedral angles (º) imidazole to pyrimidine 51.03 5' hydroxyl -159.97 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Glu156 Lys186 Ser198 Asn346 His353 water base N1 2.58 base C6NH 2.22 base C6NH 1.94 base N7 1.98 1.95 2.65 2'-OH 2.13 3'-OH 1.81 3'-OH 1.88 5'-OH 1.86
179
Figure A.17. IsoA#8.
Table A.16. Modeling results for IsoA#8. Relative energy (kcal/mol) -32.965 Dihedral angles (º) imidazole to pyrimidine 166.01 5' hydroxyl 76.47 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu59 Glu156 Lys186 Asp190 Asn191 Met358 base N1 3.67 base NH2 1.99 base NH 2.06 1.93 2'-OH 1.98 2'-OH 1.81 1.84 3'-OH 2.01 2.06 3'-OH 1.84 5'-OH 1.85 5'-OH 1.99 1.92
180
Figure A.18. IsoA#9.
Table A.17. Modeling results for IsoA#9. Relative energy (kcal/mol) -28.052 Dihedral angles (º) imidazole to pyrimidine -106.12 5' hydroxyl 25.66 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu59 Thr157 Asp190 Ser198 Asn346 base C6NH 2.06 2.20 base N7 1.95 2.97 2.65 base N9H 1.85 3'-OH 1.90 5'-OH 1.98 5'-OH
181
Figure A.19. IsoA#10.
Table A.18. Modeling results for IsoA#10. Relative energy (kcal/mol) -23.284 Dihedral angles (º) base to sugar 36.24 Hydrogen bond distance (Å) inhibitor/enzyme Glu156 Lys186 Asn346 His353 water base N1 2.72 2.03 base N7 2.03 2.67 base N9 2.05 2'-OH 1.98 2.03 3'-OH 1.87 3'-OH 1.96
182
Figure A.20. IsoA#11.
Table A.19. Modeling results for IsoA#11. Relative energy (kcal/mol) -22.580 Dihedral angles (º) base to sugar 96.02 Hydrogen bond distance (Å) inhibitor/enzyme Asn 80 Asp131 Ile 299 Gly300 Asn346 base N1 1.92 base C6NH 1.89 1.87 2.29 base N7H 1.96 1.88 base N9 1.95 2'-OH 1.72 2.00 3'-OH 1.84
183
Figure A.21. IsoA#12.
Table A.20. Modeling results for IsoA#12. Relative energy (kcal/mol) -16.199 Dihedral angles (º) base to sugar -1.63 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Thr157 Asn346 Leu347 water base N1 2.00 base C6NH 2.14 base C6NH 2.04 base N7 1.90 base N9H 1.95 1.84 2'-OH 1.89 3'-OH 1.87
184
Figure A.22. IsoA#13.
Table A.21. Modeling results for IsoA#13. Relative energy (kcal/mol) -29.930 Dihedral angles (º) base to sugar -115.94 Hydrogen bond distance (Å) inhibitor/enzyme Asp190 Asn191 Ser361 2'-OH 2.17 2.02 2.05 3'-OH 1.79 1.96 3'-OH 1.92 1.95
185
Figure A.23. IsoA#14.
Table A.22. Modeling results for IsoA#14. Relative energy (kcal/mol) -26.820 Dihedral angles (º) base to sugar 76.34 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Glu156 Thr157 Asn346 water base N1 2.89 2.19 2.83 base C6NH 2.29 base C6NH 1.94 1.81 1.90 1.94 1.97 base N7H 2.00 2.02 2'-OH 1.85 2'-OH 1.82 3'-OH 1.95
186
Figure A.24. IsoA#15.
Table A.23. Modeling results for IsoA#15. Relative energy (kcal/mol) -26.099 Dihedral angles (º) base to sugar 8.60 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Asn346 Gly352 His353 base N1 2.20 2.06 base C6NH 2.17 2.27 base N7 2.07 2.97
187
Figure A.25. IsoA#16.
Table A.24. Modeling results for IsoA#16. Relative energy (kcal/mol) -25.793 Dihedral angles (º) base to sugar -123.09 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Glu59 Asp131 Glu156 Lys186 Asn346 His353 base N1 2.99 base C6NH 1.9 1.93 base N7H 1.93 2.45 base N9 2.08 2'-OH 1.82 3'-OH 1.91 3'-OH 1.93 1.95
188
Figure A.26. IsoA#17.
Table A.25. Modeling results for IsoA#17. Relative energy (kcal/mol) -23.207 Dihedral angles (º) base to sugar -157.92 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Ser83 Glu156 Lys186 Asp190 Asn191 base C6NH 2.40 base N7 2.95 1.88 2.02 2.76 2.75 2'-OH 1.98 2.08 2'-OH 1.91 1.93 3'-OH 1.72 1.99
189
Figure A.27. IsoA#18.
Table A.26. Modeling results for IsoA#18. Relative energy (kcal/mol) -35.516 Dihedral angles (º) base to sugar 18.68 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Glu59 Asp131 Glu156 His353 base N7H 1.99 1.87 2'-OH 2.45 2'-OH 1.90 3'-OH 1.77 3'-OH 1.89
190
Figure A.28. IsoA#19.
Table A.27. Modeling results for IsoA#19. Relative energy (kcal/mol) -32.783 Dihedral angles (º) base to sugar -98.32 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Asp131 Lys186 Asn346 His353 water base N1 2.15 1.93 base N7 2.14 base N9 2.38 base N9H 2.10 2.34 2'-OH 1.76 3'-OH 1.85 3'-OH 1.90 1.96
191
Figure A.29. IsoA#20.
Table A.28. Modeling results for IsoA#20. Relative energy (kcal/mol) -32.126 Dihedral angles (º) base to sugar -88.44 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Thr157 Asp190 Asn346 water base N1 1.97 base N9H 2.10 2'-OH 1.81 1.90 3'-OH 2.04 3'-OH 1.94 1.96
192
Figure A.30. IsoA#21.
Table A.29. Modeling results for IsoA#21. Relative energy (kcal/mol) -27.701 Dihedral angles (º) base to sugar -4.91 Hydrogen bond distance (Å) inhibitor/enzyme Glu156 Gly352 His353 base N1H 1.90 base C6O 1.86 1.73 3'-OH 1.87 1.88
193
Figure A.31. IsoA#22.
Table A.30. Modeling results for IsoA#22. Relative energy (kcal/mol) -22.698 Dihedral angles (º) base to sugar -124.40 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Asp131 Glu156 Thr157 base C2O 2.74 2.87 base C6O 1.99 1.86 1.83 2'-OH 3'-OH 3'-OH 2.07 inhibitor/enzyme Lys186 Asp190 Asn191 water base C2O 2.19 base C6O 2'-OH 1.84 3'-OH 1.78 2.20 3'-OH 1.92
194
Figure A.32. IsoA#23.
Table A.31. Modeling results for IsoA#23. Relative energy (kcal/mol) -33.952 Dihedral angles (º) base to sugar 19.17 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 Glu156 Asp190 Asn191 Ser361 base N7H 1.90 2'-OH 1.94 3'-OH 2.09 2.09 3'-OH 1.81 1.74
195
Figure A.33. IsoA#24.
Table A.32. Modeling results for IsoA#24. Relative energy (kcal/mol) -49.145 Dihedral angles (º) base to sugar -27.42 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Thr157 Lys186 Asn346 His353 base N1H 2.12 base C6O 2.07 2'-OH 1.88 2.07 2'-OH 1.94 3'-OH 2.01
196
Figure A.34. IsoA#25.
Table A.33. Modeling results for IsoA#25. Relative energy (kcal/mol) -19.865 Dihedral angles (º) base to sugar 163.83 5' hydroxyl -169.45 Hydrogen bond distance (Å) inhibitor/enzyme Asp190 His353 Ser361 base N9 1.98 5'-OH 1.79 5'-OH 1.84 1.89
197
Figure A.35. IsoA#26.
Table A.34. Modeling results for IsoA#26. Relative energy (kcal/mol) -19.839 Dihedral angles (º) base to sugar 67.16 5' hydroxyl 116.58 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Asp190 Ser198 His301 His353 base N1 2.08 2.62 base C6NH 2.41 base S7 2.92 3'-OH 1.89 3'-OH 1.75 1.96 5'-OH 2.28
198
Figure A.36. IsoA#27.
Table A.35. Modeling results for IsoA#27. Relative energy (kcal/mol) -20.400 Dihedral angles (º) base to sugar 63.36 5' hydroxyl -62.30 Hydrogen bond distance (Å) inhibitor/enzyme His55 3'-OH 2.46
199
Figure A.37. IsoA#28.
Table A.36. Modeling results for IsoA#28. Relative energy (kcal/mol) -29.993 Dihedral angles (º) base to sugar 71.38 5' hydroxyl 131.66 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Glu59 Asp131 Glu156 His301 Asn346 His353 base N1 2.75 base C6NH 2.03 1.86 2.12 2.00 base S7 3.31 2'-OH 1.85 1.89 3'-OH 1.85 1.86 5'-OH 1.56 5'-OH 1.98 2.02
200
Figure A.38. IsoA#29.
Table A.37. Modeling results for IsoA#29. Relative energy (kcal/mol) -26.850 Dihedral angles (º) base to sugar 76.68 5' hydroxyl 163.84 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 Glu85 Gly194 base C6NH 1.94 2'-OH 2.01 3'-OH 1.89 2.03
201
Figure A.39. IsoA#30.
Table A.38. Modeling results for IsoA#30. Relative energy (kcal/mol) -18.760 Dihedral angles (º) base to sugar -48.80 Hydrogen bond distance (Å) inhibitor/enzyme Asn 80Asp131 His301 Asn346 water base N1 1.72 2.92 base C6NH 1.7 base C6NH 1.97 base N9 2.11 3.00 2'-OH 2.06 3'-OH 1.89 1.96 3'-OH 1.85
202
Figure A.40. IsoA#31.
Table A.39. Modeling results for IsoA#31. Relative energy (kcal/mol) -12.885 Dihedral angles (º) base to sugar -15.22 Hydrogen bond distance (Å) inhibitor/enzyme Ser78 Cys79 Asn 80 Thr157 water base N1 2.00 base C6NH 2.06 base N7 1.89 1.93 2.81 2.74 2'-OH 1.90 1.86 3'-OH 1.74
203
Figure A.41. IsoA#32.
Table A.40. Modeling results for IsoA#32. Relative energy (kcal/mol) -12.735 Dihedral angles (º) base to sugar -105.14 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 water 2'-OH 1.95 2'-OH 2.14 3'-OH 1.90 3'-OH 1.72 1.90
204
Figure A.42. IsoA#33.
Table A.41. Modeling results for IsoA#33. Relative energy (kcal/mol) -12.105 Dihedral angles (º) base to sugar 1.73 Hydrogen bond distance (Å) inhibitor/enzyme His55 Ser78 Cys79 His301 base N1 1.89 base S9 3.40 3'-OH 1.91 1.96 3'-OH 1.95
205
Figure A.43. IsoA#34.
Table A.42. Modeling results for IsoA#34. Relative energy (kcal/mol) -23.856 Dihedral angles (º) base to sugar 118.87 Hydrogen bond distance (Å) inhibitor/enzyme Lys186 Ser198 Gly352 His353 base C6NH 1.98 2'-OH 2.04 3'-OH 1.93 1.89 3'-OH 1.83
206
Figure A.44. IsoA#35.
Table A.43. Modeling results for IsoA#35. Relative energy (kcal/mol) -20.646 Dihedral angles (º) base to sugar 36.49 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Asp131 Thr157 His353 Met358 base N1 1.77 3.20 base C6NH 1.92 Base N7 1.88 2.56 2.71 2'-OH 2.00 3'-OH 1.95 3'-OH 1.85 1.91 1.96
207
Figure A.45. IsoA#36.
Table A.44. Modeling results for IsoA#36. Relative energy (kcal/mol) -30.095 Dihedral angles (º) base to sugar 49.77 5' hydroxyl -60.71 Hydrogen bond distance (Å) inhibitor/enzyme Glu156 Thr157 Asp190 Asn346 His353 Ser361 water base N1 2.85 2.60 base N3 2.09 base C6NH 1.81 1.75 2.04 2.01 base N9 2.00 2.00 2'-OH 1.89 2'-OH 2.14 3'-OH 2.00 3'-OH 1.89 5'-OH 1.93 5'-OH 1.77
208
Figure A.46. IsoA#37.
Table A.45. Modeling results for IsoA#37. Relative energy (kcal/mol) -33.978 Dihedral angles (º) base to sugar 70.12 5' hydroxyl -176.94 Hydrogen bond distance (Å) inhibitor/enzyme His301 His353 Met358 base C6NH 1.32 2.43 N7 3.46 5'-OH 2.19
209
Figure A.47. IsoA#38.
Table A.46. Modeling results for IsoA#38. Relative energy (kcal/mol) -35.704 Dihedral angles (º) base to sugar -137.78 5' hydroxyl -46.15 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 Glu197 Ser198 Asn346 His353 base C6NH 2.06 base C6NH 2.38 2.81 2'-OH 1.92 3'-OH 2.79 3'-OH 1.98 2.00 5'-OH 1.86 5'-OH 1.85
210
Figure A.48. IsoA#39.
Table A.47. Modeling results for IsoA#39. Relative energy (kcal/mol) -38.025 Dihedral angles (º) base to sugar 83.49 5' hydroxyl -75.99 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 Cys195 His353 Ser355 Ser359 base C6NH 2.28 2'-OH 2.08 2'-OH 1.73 3'-OH 1.96 3'-OH 1.96 5'-OH 1.86
211
Figure A.49. IsoA#40.
Table A.48. Modeling results for IsoA#40. Relative energy (kcal/mol) -25.070 Dihedral angles (º) base to sugar 101.08 Hydrogen bond distance (Å) inhibitor/enzyme Thr157 Asp190 Asn346 Gly352 base C2O 2.26 base C6NH 2.02 1.93 1.93 base N9 1.93 2'-OH 2.07 2'-OH 1.88 2.23 3'-OH 1.98 3'-OH 1.95
212
Figure A.50. IsoA#41.
Table A.49. Modeling results for IsoA#41. Relative energy (kcal/mol) -23.085 Dihedral angles (º) base to sugar -124.67 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu156 Thr157 Asp190 Ser198 Asn346 Leu347 base N1 2.77 base C2O 2.01 base C6NH 2.09 base C6NH 1.89 base N7 1.81 base S9 3.23 2'-OH 1.93 2.09 3'-OH 1.82 3'-OH 1.85 2.03
213
Figure A.51. IsoA#42.
Table A.50. Modeling results for IsoA#42. Relative energy (kcal/mol) -22.882 Dihedral angles (º) base to sugar 70.61 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 Thr157 Lys186 Asn346 His353 water base N1 2.65 2.20 2.43 base C6NH 2.40 base N9 1.84 2.93 2'-OH 1.97 3'-OH 1.89 3'-OH 1.82
214
Figure A.52. IsoA#43.
Table A.51. Modeling results for IsoA#43. Relative energy (kcal/mol) -15.115 Dihedral angles (º) base to sugar -111.79 Hydrogen bond distance (Å) inhibitor/enzyme Glu156 Ser198 Asn346 Met351 base C6NH 2.29 base C6NH 1.89 1.84 base N7 2.63 2'-OH 1.88 3'-OH 1.80 1.81
215
Figure A.53.IsoA#44.
Table A.52. Modeling results for IsoA#44. Relative energy (kcal/mol) -26.643 Dihedral angles (º) base to sugar 114.27 5' hydroxyl -3.27 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Glu156 Asn346 S7 2.16 2'-OH 2.06 3'-OH 1.92 1.95 5'-OH 1.81
216
Figure A.54. IsoA#45.
Table A.53. Modeling results for IsoA#45. Relative energy (kcal/mol) -22.584 Dihedral angles (º) base to sugar -50.11 5' hydroxyl -173.33 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Asp131 Glu156 Asn346 His353 base N1H 1.95 base C2O 1.82 base C6O 1.83 3'-OH 1.88 2.04 5'-OH 1.81 2.23
217
Figure A.55. IsoA#46.
Table A.54. Modeling results for IsoA#46. Relative energy (kcal/mol) -21.194 Dihedral angles (º) base to sugar -134.20 5' hydroxyl 88.83 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Asp131 Glu156 base N1H 1.95 base C2O 1.96 2.07 2'-OH 1.89 1.96 5'-OH 1.93 5'-OH 1.94
218
Figure A.56. IsoA#47.
Table A.55. Modeling results for IsoA#47. Relative energy (kcal/mol) -32.464 Dihedral angles (º) base to sugar -167.61 5' hydroxyl -148.03 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Asp131 Thr157 Lys186 base N1H 2.12 1.98 base C6O 2.10 2.13 1.83 3'-OH 1.85 1.97 5'-OH 1.96 1.82 5'-OH 1.95
219
Figure A.57. IsoA#48.
Table A.56. Modeling results for IsoA#48. Relative energy (kcal/mol) -32.104 Dihedral angles (º) base to sugar -155.49 5' hydroxyl 137.59 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Glu156 Lys186 Ile 299 His301 Asn346 base N1H 2.07 base C2O 1.99 base C6O 1.80 2'-OH 1.74 1.88 3'-OH 1.93 3'-OH 1.86 5'-OH 1.78 2.04
220
Figure A.58. IsoA#49.
Table A.57. Modeling results for IsoA#49. Relative energy (kcal/mol) -23.296 Dihedral angles (º) base to sugar -97.11 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu156 Thr157 Ser198 base C6O 1.94 2'-OH 1.95 2'-OH 1.91 1.95 3'-OH 2.03
221
Figure A.59. IsoA#50.
Table A.58. Modeling results for IsoA#50. Relative energy (kcal/mol) -27.316 Dihedral angles (º) base to sugar 23.43 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Glu156 Thr157 Asn346 Met351 Gly352 His353 water base N1H 2.17 2.38 base C2O 1.93 base C6O 2.13 2.01 base N7 2.02 2.84 3'-OH 2.26 1.90 1.79
222
Figure A.60. IsoA#51.
Table A.59. Modeling results for IsoA#51. Relative energy (kcal/mol) -27.006 Dihedral angles (º) base to sugar 159.49 Hydrogen bond distance (Å) inhibitor/enzyme Glu59 His353 Met358 Ser355 base N1H 1.96 3'-OH 1.91 2.02 3'-OH 1.89
223
Figure A.61. IsoA#52.
Table A.60. Modeling results for IsoA#52. Relative energy (kcal/mol) -26.499 Dihedral angles (º) base to sugar -54.70 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Asp131 Thr157 Asn346 His353 base N1H 2.04 base N3 2.27 base C6O 1.72 base N7 2.05 2.85 2'-OH 1.87 1.91 3'-OH 1.96 2.13
224
Figure A.62. IsoA#53.
Table A.61. Modeling results for IsoA#53. Relative energy (kcal/mol) -34.044 Dihedral angles (º) base to sugar -79.77 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Ser198 Asn346 His353 base C6NH 2.03 2'-OH 1.92 2.03 2'-OH 1.88
225
Figure A.63. IsoA#54.
Table A.62. Modeling results for IsoA#54. Relative energy (kcal/mol) -29.691 Dihedral angles (º) base to sugar -109.50 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Asn346 base C6NH 1.83 2'-OH 1.90
226 APPENDIX B
Fleximer Data
Guide to fleximer modeling data
The data described herein is for the fleximer project. The molecules are not placed in alphabetical order but are placed according to the appearance in Chapter 3 and are numbered starting with Flex#1. Each molecule image and data table for that image is placed on the same page for clarity. The molecular binding free energies in the tables are unadjusted but can be compared to each other. For a description of hydrogen bonding, the molecule naming and atom numbering scheme for the molecules are described in Figure
B.1. The molecule is colored for clarity.
Figure B.1. Fleximer modeling numbering scheme.
The inhibitor hydrogen bonding residues are labeled according to the molecule attachment. For example pyr C2NH describes the nitrogen atom attached to the C2 carbon atom in the pyrimidine ring of the fleximer. The bold letter indicates the atom that
227 is hydrogen bonding to the amino acid residue of the enzyme or water. Dihedral angles
are measured from atom numbers 2-1-1’-O4’ for the imidazole to sugar angle and the
O4’-4’-5’-O5’ for the hydroxyl angle. The dihedral angles between the imidazole and
pyrimidine rings are measured from 4-5-4-5 for the distal fleximers and 5-4-5-4 for the proximal fleximers.
In some cases, more than one structure is represented for a molecule. When a computed molecule exhibits more than one energy value within 1 Kcal/mol, those structures are considered of equal importance and therefore both are represented.
The data begins on the next page.
228
Figure B.2. Flex#1.
Table B.1. Modeling results for Flex#1. Relative energy (kcal/mol) -28.527 Dihedral angles (º) imidazole to pyrimidine 50.63 imidazole to sugar -35.82 5' hydroxyl 75.05 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Lys186 Asp190 Asn346 His353 Leu347 water pyr N1 1.91 2.71 imid N3 2.27 2.82 2.46 2.63 2'-OH 1.96 2.06
229
Figure B.3. Flex#2.
Table B.2. Modeling results for Flex#2. Relative energy (kcal/mol) -47.249 Dihedral angles (º) imidazole to pyrimidine -115.89 imidazole to sugar 159.40 5' hydroxyl 71.37 Hydrogen bond distance (Å) inhibitor/ enzyme His55 Thr57 Glu59 Asp131 Glu156 Thr157 Asn346 His353 Met358 pyr N1H 2.26 pyr N3 2.04 3.10 pyr C6O 1.93 imid N1 1.97 imid N3 1.90 2.02 2'-OH 2.88 2'-OH 2.04 1.99 3'-OH 2.98 3'-OH 1.88 2.06 5'-OH 1.95
230
Figure B.4. Flex#3.
Table B.3. Modeling results for Flex#3. Relative energy (kcal/ mol) -43.168 Dihedral angles (º) imidazole to pyrimidine -13.38 imidazole to sugar 41.73 5' hydroxyl 99.63 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Glu156 Asp190 Asn346 pyr C2NH 1.99 2.06 imid N3 2.84 2'-OH 1.79 3'-OH 2.34 5'-OH 1.99
231
Figure B.5. Flex#4.
Table B.4. Modeling results for Flex#4. Relative energy (kcal/mol) -33.201 Dihedral angles (º) imidazole to pyrimidine 84.71 imidazole to sugar -86.16 5' hydroxyl -69.48 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Glu59 Thr60 Asp131 Glu156 Thr157 Asp190 Asn346 pyr N1H 1.93 imid N3 2.60 3.00 2.78 2'-OH 2.02 1.82 3'-OH 1.95 1.93 2.00 5'-OH 1.94
232
Figure B.6. Flex#5.
Table B.5. Modeling results for Flex#5. Relative energy (kcal/mol) -53.787 Dihedral angles (º) imidazole to pyrimidine -123.13 imidazole to sugar -82.07 5' hydroxyl -64.65 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Glu59 Thr60 Thr157 Asn346 water pyr C2O 1.79 1.91 2.20 pyr N3H 1.87 imid N3 2.62 2'-OH 1.87 1.85 5'-OH 1.92
233
Figure B.7. Flex#6.
Table B.6. Modeling results for Flex#6. Relative energy (kcal/mol) -34.937 Dihedral angles (º) imidazole to pyrimidine 61.55 imidazole to sugar -74.66 5' hydroxyl 45.14 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Glu59 Thr60 Asp131 Glu156 Asn346 Leu347 pyr N1H 2.03 1.88 pyr C2O 1.96 2.03 1.83 imid N1 2.20 imid N3 1.99 2.19 2'-OH 2.15 2'-OH 1.77 3'-OH 1.72 1.80 5'-OH 1.98
234
Figure B.8. Flex#7.
Table B.7. Modeling results for Flex#7. Relative energy (kcal/mol) -26.296 Dihedral angles (º) imidazole to pyrimidine 63.34 imidazole to sugar -53.51 5' hydroxyl 48.27 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 pyr N1H 2.07 pyr C2O 2.15 pyr C6O 2.82 2'-OH 2.33 2'-OH 1.82 2.04 3'-OH 1.96 3'-OH 1.92 2.05
235
Figure B.9. Flex#8.
Table B.8. Modeling results for Flex#8. Relative energy (kcal/mol) -30.911 Dihedral angles (º) imidazole to pyrimidine 92.64 imidazole to sugar -106.18 5' hydroxyl 75.13 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu59 Asp131 Glu156 Asp190 His353 Ser361 pyr C2NH 1.90 1.76 pyr C6NH 2.08 imid N3 1.92 2'-OH 1.92 1.89 3'-OH 1.99 3'-OH 1.81 1.82 5'-OH 1.85 5'-OH 1.69
236
Figure B.10. Flex#9.
Table B.9. Modeling results for Flex#9. Relative energy (kcal/mol) -32.804 Dihedral angles (º) imidazole to pyrimidine 114.02 imidazole to sugar -52.81 5' hydroxyl -50.35 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Thr60 Glu156 Ser198 Asn346 His353 Ser359 pyr C6NH 1.94 imid N3 2.00 2.89 2.88 2'-OH 2.17 2'-OH 2.14 3'-OH 2.06 3'-OH 1.89 5'-OH 1.96
237
Figure B.11. Flex#10.
Table B.10 Modeling results for Flex#10. Relative energy (kcal/mol) -46.415 Dihedral angles (º) imidazole to pyrimidine -120.44 imidazole to sugar -160.03 5' hydroxyl 77.11 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Asp131 Glu156 Lys186 Asn346 His353 pyr N1 1.82 2.78 pyr N3 1.97 pyr C6NH 1.99 2.06 imid N3 1.76 2.85 2'-OH 1.8 3'-OH 1.87 1.82 5'-OH 1.95
238
Figure B.12. Flex#11.
Table B.11. Modeling results for Flex#11. Relative energy (kcal/mol) -47.136 Dihedral angles (º) imidazole to pyrimidine 155.86 imidazole to sugar 7.46 5' hydroxyl none Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Glu156 Asp190 Met358 pyr N1 2.83 pyr C6NH 1.97 2.17 2'-OH 1.92 2.00 3'-OH 1.84 3'-OH 2.06 1.95
239
Figure B.13. Flex#12.
Table B.12. Modeling results for Flex#12. Relative energy (kcal/mol) -49.533 Dihedral angles (º) imidazole to pyrimidine -85.02 imidazole to sugar 67.90 5' hydroxyl 111.18 Hydrogen bond distance (Å) inhibitor/enzyme Asp131 Glu156 Thr157 Lys186 Asp190 pyr N1H 1.98 pyr C2NH 1.97 2'-OH 2.10 2'-OH 1.96 2.29 3'-OH 1.97 1.95 5'-OH 1.85 1.96
240
Figure B.14. Flex#13.
Table B.13. Modeling results for Flex#13. Relative energy (kcal/mol) -32.078 Dihedral angles (º) imidazole to pyrimidine -34.49 imidazole to sugar -115.36 5' hydroxyl 153.36 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Glu59 Glu156 Asp190 Gly352 His353 Ser361 pyr N1 2.01 2.90 2.70 2.99 pyr C2O 2.12 2.01 pyr C6NH 1.99 2'-OH 1.93 2'-OH 1.81 1.81 5'-OH 1.90
241
Figure B.15. Flex#14.
Table B.14. Modeling results for Flex#14. Relative energy (kcal/mol) -55.169 Dihedral angles (º) imidazole to pyrimidine -121.78 imidazole to sugar -141.15 5' hydroxyl 159.03 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr60 Asp131 pyr N1 2.36 pyr C6NH 2.41 3'-OH 1.89 2.21 5'-OH 2.43
242
Figure B.16. Flex#15.
Table B.15. Modeling results for Flex#15. Relative energy (kcal/mol) -30.592 Dihedral angles (º) imidazole to pyrimidine -150.04 imidazole to sugar 1.74 5' hydroxyl 67.00 Hydrogen bond distance (Å) inhibitor/ enzyme His55 Thr57 Thr60 Lys186 Asp190 Asn191 Gly300 Asn346 Met358 water pyr N1 1.94 pyr N3 2.79 2.66 1.93 2.08 2.89 imid N3 3.41 2'-OH 1.88 2.50 2.11 1.94 2'-OH 1.88 1.75 3'-OH 2.02 5'-OH 1.78 5'-OH 1.87 ring O 2.03
243
Figure B.17. Flex#16.
Table B.16. Modeling results for Flex#16. Relative energy (kcal/mol) -41.258 Dihedral angles (º) imidazole to pyrimidine -50.76 imidazole to sugar 167.71 5' hydroxyl 59.40 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Glu59 Glu156 Thr157 Asp190 His353 water pyr N1H 1.86 1.95 pyr C2O 2.90 pyr N3H 1.92 1.979 2.06 pyr C6O 1.79 2'-OH 1.92 2'-OH 1.86 3'-OH 1.91 5'-OH 1.89
244
Figure B.18. Flex#17.
Table B.17. Modeling results for Flex#17. Relative energy (kcal/mol) -42.474 Dihedral angles (º) imidazole to pyrimidine 93.43 imidazole to sugar 167.30 5' hydroxyl -60.19 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Asp131 Glu156 Thr157 Asn346 pyr N1H 1.91 2.08 pyr C2NH 1.76 1.91 2.08 pyr C6O 1.98 imid N3 1.79 3'-OH 1.95 1.92 3'-OH 1.74 5'-OH 2.08
245
Figure B.19. Flex#18.
Table B.18. Modeling results for Flex#18. Relative energy (kcal/mol) -34.666 Dihedral angles (º) imidazole to pyrimidine 61.57 imidazole to sugar -84.40 5' hydroxyl -147.23 Hydrogen bond distance (Å) inhibitor/enzyme Cys79 Ser83 Asp131 Thr157 Lys186 water pyr N1 3.40 2.22 pyr C2NH 2.15 2'-OH 1.82 3'-OH 1.79 2.45 1.68 3'-OH 1.88 5'-OH 1.92 1.93
246
Figure B.20. Flex#19.
Table B.19. Modeling results for Flex#19. Relative energy (kcal/mol) -34.660 Dihedral angles (º) imidazole to pyrimidine -126.15 imidazole to sugar 64.82 5' hydroxyl 27.08 Hydrogen bond distance (Å) Ile inhibitor/enzyme His55 Asp131 Gly194 299 His353 Met358 pyr N1 3.08 pyr C2NH 2.24 pyr C6NH 2.49 imid N1 1.96 imid N3 1.96 3'-OH 1.92 5'-OH 1.83 1.91
247
Figure B.21. Flex#20.
Table B.20. Modeling results for Flex#20. Relative energy (kcal/mol) -38.560 Dihedral angles (º) imidazole to pyrimidine -158.73 imidazole to sugar 121.76 5' hydroxyl -169.51 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu59 Glu156 Asn346 His353 Ser361 pyr N1H 1.86 1.94 pyr C2NH 1.97 imid N1 1.99 imid N3 2.08 1.93 2'-OH 1.84 2'-OH 1.90 2.03 3'-OH 1.82 1.97 5'-OH 1.96
248
Figure B.22. Flex#21.
Table B.21. Modeling results for Flex#21. Relative energy (kcal/mol) -39.884 Dihedral angles (º) imidazole to pyrimidine 106.82 imidazole to sugar 95.47 5' hydroxyl 93.34 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr60 Asp131 Glu156 Thr157 Lys186 pyr C5NH 1.88 pyr C5NH 1.88 1.86 2.09 pyr C6NH 1.91 2.36 1.95 imid N3 2.11 2.18
2.96
3'-OH 1.82
249
Figure B.23. Flex#22.
Table B.22. Modeling results for Flex#22. Relative energy (kcal/mol) -40.577 Dihedral angles (º) imidazole to pyrimidine 126.16 imidazole to sugar 37.46 5' hydroxyl -57.15 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Thr60 Glu156 Lys186 Asp190 pyr N3 2.88 pyr C6NH 1.85 1.94 imid C4NH 2.04 2'-OH 2.00 2'-OH 1.81 3'-OH 2.07 3'-OH 1.80 2.03
250
Figure B.24. Flex#23.
Table B.23. Modeling results for Flex#23. Relative energy (kcal/mol) -50.235 Dihedral angles (º) imidazole to pyrimidine -129.90 imidazole to sugar -87.91 5' hydroxyl 71.54 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr60 Glu156 Thr157 Ser198 Asn346 Met358 pyr N1 2.95 pyr C5NH 1.93 imid N3 1.97 imid C4NH 2.22 imid C4NH 1.93 2'-OH 1.85 2.47 1.98 5'-OH 2.24 1.76
251
Figure B.25. Flex#24.
Table B.24. Modeling results for Flex#24. Relative energy (kcal/mol) -50.822 Dihedral angles (º) imidazole to pyrimidine -109.49 imidazole to sugar -96.30 5' hydroxyl 77.96 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Thr157 Lys186 Asp190 Cys195 water pyr C6NH 1.97 imid N3 3.39 2'-OH 2.07 2'-OH 1.84 1.99 3'-OH 1.79 furanose O 1.72 2.76
252
Figure B.26. Flex#25.
Table B.25. Modeling results for Flex#25. Relative energy (kcal/mol) -44.936 Dihedral angles (º) imidazole to pyrimidine 92.57 imidazole to sugar -145.13 5' hydroxyl 52.18 Hydrogen bond distance (Å) inhibitor/enzyme Glu156 Asn346 His353 Met358 Ser361 water pyr C5CO 2.00 pyr C6NH 1.99 pyr C6NH 1.94 1.95 imid N3 1.90 2'-OH 1.96 3'-OH 1.84 5'-OH 1.80 1.87
253
Figure B.27. Flex#26.
Table B.26. Modeling results for Flex#26. Relative energy (kcal/mol) -51.581 Dihedral angles (º) imidazole to pyrimidine 133.08 imidazole to sugar 172.98 5' hydroxyl 49.77 Hydrogen bond distance (Å) inhibitor/enzyme Thr57 Asp131 Glu156 Asp190 Ser198 Asn346 His353 water pyr C6NH 1.88 imid N3 1.90 imid C4CO 1.85 2'-OH 1.84 3'-OH 1.95 3'-OH 2.08 5'-OH 1.94 2.11
254
Figure B.28. Flex#27.
Table B.27. Modeling results for Flex#27. Relative energy (kcal/mol) -52.283 Dihedral angles (º) imidazole to pyrimidine 154.54 imidazole to sugar 26.77 5' hydroxyl 176.74 Hydrogen bond distance (Å) inhibitor/ enzyme His55 Thr57 Glu59 Asp131 Glu156 Lys186 Asp190 Gly352 His353 pyr N1 1.89 pyr C2NH 1.85 imid N3 2.92 1.91 imid C4OH 1.88 imid C4OH 1.88 2'-OH 1.71 3'-OH 1.84 3'-OH 1.91 1.96 5'-OH 1.86
255
Figure B.29. Flex#28.
Table B.28. Modeling results for Flex#28. Relative energy (kcal/mol) -31.206 Dihedral angles (º) imidazole to pyrimidine -88.82 imidazole to sugar -138.48 5' hydroxyl -77.67 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr57 Thr60 Asp131 Glu156 Asp190 Asn346 His353 pyr C5OH 1.84 pyr C5OH 2.01 pyr C6NH 1.94 2.28 imid N3 1.84 2.18 2.83 2'-OH 1.83 1.95 3'-OH 1.83 1.83 5'-OH 2.03 2.98 1.97
256
Figure B.30. Flex#29.
Table B.29. Modeling results for Flex#29. Relative energy (kcal/mol) -39.600 Dihedral angles (º) imidazole to pyrimidine 85.85 imidazole to sugar 131.24 5' hydroxyl -165.95 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Glu156 Thr157 Asp190 Cys195 Ser198 pyr N1 2.83 pyr C6NH 2.06 pyr C6NH 2.31 1.82 2'-OH 2.01 2.05 2'-OH 1.87 2.07 3'-OH 1.82 1.99 5'-OH 1.94
257
Figure B.31. Flex#30.
Table B.30. Modeling results for Flex#30. Relative energy (kcal/mol) -39.692 Dihedral angles (º) imidazole to pyrimidine 59.26 imidazole to sugar -128.16 5' hydroxyl 82.63 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Glu156 Lys186 His353 pyr N1 1.74 2'-OH 1.92 2'-OH 1.95 3'-OH 2.36 3'-OH 1.84 1.86 5'-OH 1.88 furanose O 2.10
258
Figure B.32. Flex#31.
Table B.31. Modeling results for Flex#31. Relative energy (kcal/mol) -48.121 Dihedral angles (º) imidazole to pyrimidine -116.83 imidazole to sugar -55.96 5' hydroxyl 118.97 Hydrogen bond distance (Å) inhibitor/enzyme Glu156 Lys186 Asp190 Asn191 Asn346 Gln365 pyr C6NH 1.89 pyr C6NH 1.96 2.19 imid N3 1.88 2'-OH 1.85 3'-OH 1.91 1.84 3'-OH 1.88 1.94 5'-OH 1.94
259
Figure B.33. Flex#32.
Table B.32. Modeling results for Flex#32. Relative energy (kcal/mol) -38.657 Dihedral angles (º) imidazole to pyrimidine 27.80 imidazole to sugar -135.91 5' hydroxyl -76.70 Hydrogen bond distance (Å) inhibitor/enzyme His55 Glu156 Lys186 Asp190 Met358 pyr N1 1.93 2.03 pyr C2NH 2.05 1.88 2.11 pyr N3 3.29 2'-OH 1.85 2'-OH 1.87 3'-OH 1.88 1.95 5'-OH 1.93 5'-OH 1.80
260
Figure B.34. Flex#33.
Table B.33. Modeling results for Flex#33. Relative energy (kcal/mol) -39.565 Dihedral angles (º) imidazole to pyrimidine -14.67 imidazole to sugar 109.47 5' hydroxyl -177.21 Hydrogen bond distance (Å) inhibitor/enzyme His55 Thr60 Asp190 Asn346 His353 water pyr C5OH 1.94 2.14 pyr C6OH 1.89 1.73 pyr C2NH 2.42 imid N3 1.97 3'-OH 1.87 5'-OH 1.96
261
Figure B.35. Flex#34.
Table B.34. Modeling results for Flex#34. Relative energy (kcal/mol) -43.287 Dihedral angles (º) imidazole to pyrimidine -29.75 imidazole to sugar -123.77 5' hydroxyl 136.85 Hydrogen bond distance (Å) inhibitor/enzyme His55 Asp131 Glu156 Thr157 Ser198 pyr C2NH 1.73 imid N3 2.02 2.88 2'-OH 1.81 2.01 3'-OH 1.84 1.89 2.38
262 REFERENCES
(1) Foye, W. O.; Lemke, T. L.; Williams, D. A. Principles of Medicinal Chemistry; 4th ed.; Williams and Wilkins: Baltimore, 1995; 995.
(2) Silverman, R. B. The Organic Chemistry of Drug Design and Drug Action; Academic Press: Evanston, 1992.
(3) Voet, D.; Voet, J. G. Biochemistry; 2nd ed.; John Wiley & Sons: Somerset, 1995.
(4) Silverman, R. B. The Organic Chemistry of Drug Design and Drug Action; 2nd ed.; Elsevier: Amsterdam, 2004; 617.
(5) Fischer, E. Ber. Dtsch. Chem. Ges 1894, 27, 2985.
(6) Koshland, D. E., Jr.; Neet, K. E. Annual Reviews in Biochemistry 1968, 37, 359.
(7) Berg, J. M.; Tymoczko, J. L.; Stryer, L. Biochemistry; 5th ed.; W.H. Freeman and Co., 2002.
(8) Leonard, N. J.; Laursen, R. A. Synthesis of 3-β-D-Ribofuranosyladenine and (3-β- D-Ribofuranosyladenine)-5'-phosphate. Biochemistry 1965, 4, 354-365.
(9) Leonard, N. J.; Morrice, A. G.; Sprecker, M. A. Linear Benzoadenine. A Stretched-Out Analog of Adenine. Journal of Organic Chemistry 1975, 40, 356- 366.
(10) Leonard, N. J.; Hiremath, S. P. Dimensional Probes of Binding and Activity. Tetrahedron 1986, 42, 1917-1961.
(11) Leonard, N. J. Dimensional Probes of Enzyme-Coenzyme Binding Sites. Accounts of Chemical Research 1982, 15, 128-135.
(12) Leonard, N. J.; Scopes, D. I. C.; VanDerLijn, P.; Barrio, J. R. Dimensional Probes of the Enzyme Binding Sites of Adenine Nucleotides. Biological Effects of Widening the Adenine Ring by 2.4 Å. Biochemistry 1978, 17, 3677-3685.
263 (13) Schlick, T. Molecular Modeling and Simulation; Springer: New York, 2002; 634.
(14) Westheimer, F. H.; Mayer, J. E. Journal of Chemical Physics 1946, 37, 733.
(15) Bultinck, P.; De Winter, H.; Langenaeker, W.; Tollenaere, J. P. Computational Medicinal Chemistry for Drug Discovery; Marcel Dekker, Inc.: New York, 2004; 794.
(16) Montgomery, J. A. Purine Nucleoside Phosphorylase: a Target for Drug Design. Medicinal Research Reviews 1993, 13, 209-228.
(17) Montgomery, J. A.; Niwas, S.; Rose, J. D.; Secrist, J. A., III; Babu, Y. S. et al. Structure-Based Design of Inhibitors of Purine Nucleoside Phosphorylase. 1. 9- (arylmethyl) Derivatives of 9-Deazaguanine. Journal of Medicinal Chemistry 1993, 36, 55-69.
(18) Ealick, S. E.; Babu, Y. S.; Bugg, C. E.; Erion, M. D.; Guida, W. C. et al. Application of Crystallographic and Modeling Methods in the Design of Purine Nucleoside Phosphorylase Inhibitors. Proceedings of the National Academy of Sciences, U.S.A. 1991, 88, 11540-11544.
(19) Bugg, C. E.; Carson, W. M.; Montgomery, J. A. Drugs by Design. Scientific American 1993, 269, 92-98.
(20) Ealick, S. E. Personal communication; O'Daniel, P. I. Ed., 2005.
(21) Federov, A.; Shi, W.; Kicska, G.; Federov, E.; Tyler, P. C. et al. Transition State Structure of Purine Nucleoside Phosphorylase and Principles of Atomic Motion in Enzymatic Catalysis. Biochemistry 2001, 40, 853-860.
(22) Meyer, E. F.; Swanson, S. M.; Williams, J. A. Molecular Modeling and Drug Design. Pharmacol. Ther. 2000, 85, 113-121.
(23) Carlson, H. A.; McCammon, J. A. Accommodating Protein Flexibility in Computational Drug Design. Molecular Pharmacology 2000, 57, 213-218.
(24) Joseph-McCarthy, D. Computational Approaches to Structure-based Ligand Design. Pharmacol. Ther. 1999, 84, 179-191.
264 (25) Das, K.; Clark, A. D., Jr.; Lewi, P. J.; Heeres, J.; de Jonge, M. R. et al. Roles of Conformational and Positional Adaptability in Structure-Based Design of TMC125-R165335 (Etravirine) and Related Non-Nucleoside Reverse Transcriptase Inhibitors that are Highly Potent and Effective Against Wild-Type and Drug-Resistant HIV-1 Varients. Journal of Medicinal Chemistry 2004, 47, 3550-2560.
(26) Lewis, P. J.; de Jonge, M.; Frits, D.; Koymans, L.; Vinkers, M. et al. On the Detection of Multiple-Binding Modes of Ligand to Proteins, From Biological, Structural, and Modeling Data. Journal of Computer-Aided Molecular Design 2003, 17, 129-134.
(27) Wilson, E. K. Dealing With Flexible Receptors. Chemical & Engineering News 2004, 46-47.
(28) Henry, C. M. Clues for Overcoming HIV Drug Resistance. Chemical & Engineering News 2004, 40-41.
(29) Jimenez, R.; Salazar, G.; Yin, J.; Joo, T.; Romesberg, F. E. Protein Dynamics and the Immunological Evolution of Molecular Recognition. Proceedings of the National Academy of Sciences, U.S.A. 2004, 101, 3803-3808.
(30) Jimenez, R.; Salazar, G.; Baldridge, K. K.; Romesberg, F. E. Flexibility and Molecular Recognition in the Immune System. Proceedings of the National Academy of Sciences, U.S.A. 2002, 100, 92-97.
(31) Yuan, C.-S.; Liu, S.; Wnuk, S. F.; Robins, M. J.; Borchardt, R. T. Design and Synthesis of S-Adenosylhomocysteine Hydrolase Inhibitors as Broad-Spectrum Antiviral Agents. Advances in Antiviral Drug Design; JAI Press, Inc.: Greenwich, CT, 1996; 41-88.
(32) Howell, P. L. Open and Closed SAHase Conformation; Seley, K. L. Ed., 2004.
(33) Turner, M. A.; Yuan, C.-S.; Borchardt, R. T.; Hershfield, M. S.; Smith, D. G. et al. Structure Determination of Selenomethionyl S-Adenosylhomocysteine Hydrolase Using Data at a Single Wavelength. Nature Structural Biology 1998, 5, 369-376.
(34) Palmer, J. L.; Abeles, R. H. The Mechanism of Action of S- Adenosylhomocysteine. Journal of Biological Chemistry 1979, 254, 1217-1226.
265 (35) Seley, K. L.; Quirk, S.; Salim, S.; Zhang, L.; Hagos, A. Unexpected Inhibition of S-Adenosyl-L-homocysteine Hydrolase by a Guanosine Nucleoside. Bioorganic & Medicinal Chemistry Letters 2003, 13, 1985-1988.
(36) Levene, P. A.; Jacobs, W. A. Ber. 1909, 42.
(37) Blair, E.; Darby, G.; Gough, G.; Littler, E.; Rowlands, D. et al. Antiviral Therapy; 1st ed.; Springer: Verlag, 1998; 161.
(38) Chu, C. K.; Baker, D. C. Nucleosides and Nucleotides as Antitumor and Antiviral Agents; Plenum Press: New York, 1993.
(39) Simons, C. Nucleoside Mimetics: Their Chemistry and Biological Properties; Gordon and Breach: Amsterdam, 2001.
(40) Cristalli, G.; Franchetti, P.; Grifantini, M.; Sauro, V.; Bordoni, T. et al. Inmproved Synthesis and Antitumor Activity of 1-Deazaadenosine. Journal of Medicinal Chemistry 1987, 30, 1686-1688.
(41) Antonini, I.; Cristalli, G.; Franchetti, P.; Grifantini, M.; Martelli, S. et al. Journal of Pharmaceutical Sciences 1984, 1984, 366-369.
(42) Cristalli, G.; Grifantini, M.; Vittori, S.; Balduini, W.; Cattabeni, F. Nucleosides & Nucleotides 1985, 4, 625-639.
(43) Montgomery, J. A.; Shortnacy, A. T.; Clayton, S. D. A Comparison of Two Methods for the Preparation of 3-Deazapurine Ribonucleosides. Journal of Heterocyclic Chemistry 1977, 14, 195-197.
(44) Chiang, P. K. Biological Effects of Inhibitors of S-Adenosylhomocysteine Hydrolase. Pharmacol. Ther. 1998, 77, 115-134.
(45) Seela, F.; Peng, X. Regioselective Synthesis of 7-Halogenated 7-Deazapurine Nucleosides Related to 2-Amino-7-deaza-2'-deoxyadenosine and 7-Deaza-2'- deoxyisoguanosine. Synthesis 2004, 8, 1203-1210.
(46) Fish, W. R.; Marr, J., Joseph; Berens, R. L.; Looker, D. L.; Nelson, D. J. et al. Inosine Analogs as Chemotherapeutic Agents for African Trypanosomes:
266 Metabolism in Trypanosomes and Efficacy in Tissue Culture. Antimicrobial Agents and Chemotherapy 1985, 27, 33-36.
(47) Bacchi, C. J.; Berens, R. L.; Nathan, H. C.; Klein, R. S.; Elegbe, I. A. et al. Synergism Between 9-Deazainosine and DL-α-Difluoromethylornithine in Treatment of Experimental African Tryanosomiasis. Antimicrobial Agents and Chemotherapy 1987, 31, 1406-1413.
(48) Copp, R. R.; Marquez, V. E. Syntheisi of Two Cyclopentenyl-3-deazapyrimidine Carbocyclic Nucleosides Related to Cytidine and Uridine. Journal of Medicinal Chemistry 1991, 34, 208-212.
(49) Li, Z. R.; Campbell, J.; Rustum, Y. M. Effect of 3-deazauridine on the Metabolism, Toxicity, and Antitumor Activity of Azacitidine in Mice Bearing L1210 Leukemia Sensitive and Resistant to Cytarabine. Cancer Treatment Reports 1982, 67, 547-554.
(50) Bergstrom, D. E.; Brattesani, A. J.; Ogawa, M. K.; Reddy, P. A.; Schweickert, M. J. et al. Antiviral Activity of C-5 Substituted Tubercidin Analogues. Journal of Medicinal Chemistry 1984, 27, 285-292.
(51) Glazer, R. I.; Hartman, K. D.; Knode, M. C. 9-Deazadenosine: Cytocidal Activity and Effects on Nucleic Acids and Protein Synthesis in Human Colon Carcinoma Cells in Culture. Molecular Pharmacology 1983, 24, 309-315.
(52) Chu, M. Y.; Zuckerman, L. B.; Sato, S.; Crabtree, G. W.; Bogden, A. E. et al. 9- Deazaadenosine-A New Potent Antitumor Agent. Biochemical Pharmacology 1984, 33, 1229-1234.
(53) Smith, J. W.; Bartlett, M. S.; Queener, S. F.; Durkin, M. M.; Jay, M. A. et al. Pneumocystis carinii Pneumonia Therapy with 9-Deazainosine in Rats. Mycology and Parasitology 1987, 7, 113-118.
(54) Bennett, L. L., Jr.; Allan, P. W.; Carpenter, J. W.; Hill, D. L. Nucleosides of 2- Aza-purines Cytotoxicities and Activities as Substrates for Enzymes Metabolizing Purine Nucleosides. Biochemical Pharmacology 1976, 25, 517-521.
(55) Montgomery, J. A.; Elliott, R. D. Analogues of 8-Azainosine. Journal of Medicinal Chemistry 1977, 20, 116-120.
267 (56) Thibault, A.; Figg, W. D.; Bergan, R. C.; Lush, R. M.; Myers, C. E. et al. A Phase II Study of 5-Aza-2'-deoxycytidine (decitabine) in Horomone Independent Metastatic (D2) Prostrate Cancer. Tumori 1998, 84, 87-89.
(57) Wilson, V. L.; Jones, P. A.; Momparler, R. L. Inhibition of DNA Methylation in L1210 Leukemic cells by 5-Aza-2'-deoxycytidine as a Possible Mechanism of Chemotherapeutic Action. Cancer Research 1983, 43, 3493-3496.
(58) Cheng, J. C.; Weisenberger, D. J.; Gonzales, F. A.; Liang, G.; Xu, G.-L. et al. Continuous Zebularine Treatment Effectively Sustains Demethylation in Human Bladder Cancer Cells. Molecular and Cellular Biology 2004, 24, 1270-1278.
(59) Franchetti, P.; Messini, L.; Cappellacci, L.; Sheikha, G. A.; Grifantini, M. et al. 8- Aza-1-Deazapurine Nucleosides as Antiviral Agents. Nucleosides & Nucleotides 1994, 13, 1739-1755.
(60) Evans, G. B.; Furneaux, R. H.; Gainsford, G. J.; Hanson, J. C.; Kicska, G. A. et al. 8-Aza-immucillins as Transation-State Analogue Inhibitors of Purine Nucleoside Phosphorylase and Nucleoside Hydrolases. Journal of Medicinal Chemistry 2003, 46, 155-160.
(61) Franchetti, P.; Messini, L.; Cappellacci, L.; Grifantini, M.; Nocentini, G. et al. 8- Aza Derivatives of 3-Deazapurine Nucleosides. Synthesis and Invitro Evaluation of Antiviral and Antitumor Activity. Antiviral Chemistry and Chemotherapy 1993, 4, 341-352.
(62) Kumar, S.; Wilson, S. R.; Leonard, N. J. Structure of 3-Isoadenosine. Acta Crystallographica. Section C. Crystal Structure Communications 1988, 44, 508- 510.
(63) Leonard, N. J.; Laursen, R. A. The Synthesis of 3-β-D-Ribofuranosyladenine. Journal of the American Chemical Society 1963, 85, 2026-2028.
(64) Krenitsky, T. A.; Elion, G. B.; Strelitz, R. A.; Hitchings, G. H. Ribonucleosides of Allopurinol and Oxoallopurinol. Isolation from Human Urine, Enzymatic Synthesis and Characterization. Journal of Biological Chemistry 1967, 242, 2675- 2682.
(65) Beardmore, T. D.; Kelley, W. N. Mechanism of Allopurinol-Mediated Inhibition of Pyrimidine Biosynthesis. J. Lab. Clin. Med. 1971, 78, 696-704.
268 (66) Fujii, T.; Walker, G. C.; Leonard, N. J.; DeLong, D. C.; Gerzon, K. 3-Substituted Adenines. In Vitro Enzyme Inhibition and Antiviral Activity. Journal of Medicinal Chemistry 1979, 22, 125-129.
(67) Nair, V.; Buenger, G. S.; Leonard, N. J.; Balzarini, J.; De Clercq, E. Synthesis of 2',3'-Dideoxy-3-isoadenosine: A New Structural Analogue of the Anti-HIV Active Compound, 2',3'-Dideoxyadenosine. Journal of the Chemical Society, Chemical Communications 1991, 1650-1651.
(68) Bzowska, A.; Kulikowska, E.; Poopeiko, N. E.; Shugar, D. Kinetics of Phosphorolysis of 3-(β-D-Ribofuranosyl)adenine and 3-(β-D- Ribofuranosyl)hypoxanthine, Non-conventional Subatrates of Purine-Nucleoside Phosphorylase. European Journal of Biochemistry 1996, 239, 229-234.
(69) Gerzon, K.; Johnson, I. S.; Boder, G. B.; Cline, J. C.; Simpson, P. J. et al. Biological Activities of 3-Isoadenosine. Biochimica et Biophysica Acta 1966, 119, 445-461.
(70) Wolfenden, R.; Sharpless, T. K.; Ragade, I. S.; Leonard, N. J. Enzymatic and Chemical Deamination of 3-(β-D-Ribofuranosyl)adenine. Journal of the American Chemical Society 1966, 88, 185-186.
(71) Michelson, A. M.; Monny, C.; Laursen, R. A.; Leonard, N. J. Polynucleotide Analogues VIII. Poly 3-Isoadenylic Acid. Biochimica et Biophysica Acta 1966, 119, 258-267.
(72) Hill, A. R., Jr.; Kumar, S.; Leonard, N. J.; Orgel, L. E. Template-directed Oligomerization of 3-Isoadenosine 5'-Phosphate. J. Mol. Evol. 1988, 27, 91-95.
(73) Hill, A. R., Jr.; Kumar, S.; Patil, V. D.; Leonard, N. J.; Orgel, L. E. Which 3- Ribofuranosyl-Substituted Purine 5'-Phosphates Undergo Template-Directed Oligomerization? J. Mol. Evol. 1991, 32, 447-453.
(74) Bhat, B.; Neelima; Leonard, N. J.; Robinson, H.; Wang, A. H.-J. 2'-Deoxy-3- isoadenosine Forms Hoogsteen-Type Base Pairs with Thymidine in the d(CG[iA]TCG)2 Duplex. Journal of the American Chemical Society 1996, 118, 3065-3066.
269 (75) Gutowski, G. E.; Sweeney, M. J.; DeLong, D. C.; Hamill, R. L.; Gerzon, K. et al. Biochemistry and Biological Effects of the Pyrazofurins (Pyrazomycins): Initial Clinical Trials. Ann. N. Y. Acad. Sci. 1975, 255, 544-551.
(76) Sweeney, M. J.; Davis, F. A.; Gutowski, G. E.; Hamill, R. L.; Huffman, D. H. et al. Experimental Antitumor Activity of Pyrazomycin. Cancer Research 1973, 33, 2619-2623.
(77) Srivastava, P. C.; Pickering, M. V.; Allen, L. B.; Streeter, D. G.; Campbell, M. T. et al. Synthesis and Antiviral Activity of Certain Thiazole C-Nucleosides. Journal of Medicinal Chemistry 1977, 20, 256-262.
(78) Olah, E.; Natsumeda, Y.; Ikegami, T.; Kote, Z.; Horanyi, M. et al. Induction of Erythroid Differentiation and Modulation of Gene Expression by Tiazofurin in K- 562 Leukemia Cells. Proceedings of the National Academy of Sciences, U.S.A. 1988, 85, 6533-6537.
(79) Kiguchi, K.; Collart, F. R.; Henning-Chibb, C.; Huberman, E. Cell Differentiation and Altered IMP Dehydrogenase Expression Induced in Human T- Lymphoblastoid Leukemia Cells by Mycophenolic Acid and Tiazofurin. Exp. Cell Res. 1990, 187, 47-53.
(80) Pillwein, K.; Schuchter, K.; Ressmann, G.; Gharehbaghi, K.; Knoflach, A. et al. Cytotoxicity, Differentiating Activity and Metabolism of Tiazofurin in Human Neuroblastoma Cells. International Journal of Cancer 1993, 55, 92-95.
(81) Kiguchi, K.; Collart, F. R.; Henning-Chibb, C.; Huberman, E. Induction of Cell Differentiation in Melanoma Cells by Inhibitors of IMP Dehydrogenase: Altered Patterns of IMP Dehydrogenase Expression and Activity. Cell Growth Differ. 1990, 1, 259-270.
(82) Srivastava, P. C.; Robins, R. K. Synthesis and Antitumor Activity of 2-β-D- Ribofuranosylselenazole-4-carboxamide and Related Derivatives. Journal of Medicinal Chemistry 1983, 26, 445-448.
(83) Witkowski, J. T.; Robins, R. K.; Sidwell, R. W.; Simon, L. N. Design, Synthesis and Broad-Spectrum Antiviral Activity of 1-β-D-ribofuranosyl-1,2,4-triazole-3- carboxamide and Related Nucleosides. Journal of Medicinal Chemistry 1972, 15, 1150-1154.
270 (84) Liaw, Y.-C.; Wang, A. H.-J.; Lin, G.-S.; Chern, J.-W. 3-β-D-Ribofuranosyl-6,7- dihydro-9H-thiazolo[3,2-a]purin-9-one Hydrate. Acta Crystallographica 1994, C50, 734-736.
(85) Schmidt, C. L.; Rusho, W. J.; Townsend, L. B. The Synthesis of Bicyclic Nucleosides Related to Uridine, 4-β-D-Ribofuranosyl)thiazolo[5,4-d]pyrimidines. Journal of the Chemical Society, Chemical Communications 1971, 1515-1516.
(86) Badawey, E. S. A. M.; Rida, S. M.; Hazza, A. A.; Fahmy, H. T. Y.; Gohar, Y. M. Potential Anti-Microbials. I. Synthesis and Structure-Activity Studies of Some New Thiazolo[4,5-d]pyrimidine Derivatives. European Journal of Medicinal Chemistry 1993, 28, 91-96.
(87) Badawey, E. S. A. M.; Rida, S. M.; Hazza, A. A.; Fahmy, H. T. Y.; Gohar, Y. M. Potential Anti-Microbials. II. Synthesis and in vitro Anti-Microbial Evaluation of Some Thiazolo[4,5-d]pyrimidines. European Journal of Medicinal Chemistry 1993, 28, 97-101.
(88) Habib, N. S.; Rida, S. M.; Badawey, E. S. A. M.; Fahmy, H. T. Y. Condensed Thiazoles, I. Synthesis of 5,7-Disubstituted Thiazolo[4,5-d]pyrimidines as Possible Anti-HIV, Anticancer, and Antimicrobial Agents. Monatschefte für Chemie 1996, 127, 1203-1207.
(89) Habib, N. S.; Rida, S. M.; Badawey, E. S. A. M.; Fahmy, H. T. Y. Condensed Thiazoles, II: Synthesis of 7-Substituted Thiazolo[4,5-d]pyrimidines as Possible Anti-HIV, Anticancer, and Antimicrobial Agents. Monatschefte für Chemie 1996, 127, 1209-1214.
(90) Nagahara, K.; Sekine, M.; Takada, A. Study of Thiazolo[4,5-d]pyrimidines: The Synthesis of Thiazolo[4,5-d]pyrimidine-2,7-diones and Novel Ring Opening to 2,4-Thiazolindinedione. Heterocycles 1993, 36, 923-927.
(91) Nagahara, K.; Anderson, J. D.; Kini, G. D.; Dalley, N. K.; Larson, S. B. et al. Thiazolo[4,5-d]pyrimidine Nucleosides. The Synthesis of Certain 3-β-D- Ribofuranosylthiazolo[4,5-d]pyrimdines as Potential Immunotherapeutic Agents. Journal of Medicinal Chemistry 1990, 33, 407-415.
(92) Lewis, A. F.; Revenkar, G. R.; Fennewald, S. M.; Huffman, J. H.; Rando, R. F. Thiazolo[4,5-d]pyrimidines. Part 1. Synthesis and Anti-Human Cytomegalovirus
271 (HCMV) Activity in vitro of Certain Alkyl Derivatives. Journal of Heterocyclic Chemistry 1995, 32, 547-556.
(93) El-Bayouki, K. A. M.; Basyouni, W. M. New Thiazolo[5,4-d]pyrimidines with Molluscicidal Properties. Bull. Chem. Soc. Jpn. 1988, 61, 3794-3796.
(94) Patil, V. D.; Wise, D. S.; Townsend, L. B. Synthesis and Biological Activity of Selected 2-Substituted 6-(β-D-Ribofuranosyl)oxazolo[5,4-d]pyrimidin-7-ones. Journal of Medicinal Chemistry 1974, 17, 1282-1285.
(95) Schmidt, C. L.; Townsend, L. B. Bicyclic Nucleosides Related to Pyrimidine Nucleosides. IV. Synthesis of 4- and 6-Ribofuranosylthiazolo[5,4-d]pyrimidines and 4-Arabinofuranosylthiazolo[5,4-d]pyrimidines. Journal of Organic Chemistry 1975, 40, 2476-2481.
(96) Goldman, I. M. A Novel Thiazole Synthesis. 4,5,6,7-Tetrahydrothiazolo[4,5- d]pyrimidine-5,7-diones. Journal of Organic Chemistry 1969, 34, 3285-3289.
(97) Tindall, C. G.; Robins, R. K.; Tolman, R. L.; Hutzenlaub, W. Directed Glycosylation of 8-Bromoadenine. Synthesis and Reactions of 8-Substituted 3- Glycosyladenine Derivatives. Journal of Organic Chemistry 1972, 37, 3985-3989.
(98) Rajeev, K.; Broom, A. D. 5,6-Diaminocytidine, a Versatile Synthon for Pyrimidine-Based Bicyclic Nucleosides. Organic Letters 2000, 2, 3595-3598.
(99) Mizuno, Y.; Watanabe, Y.; Ikeda, K. Synthesis of a Potential Antitumor Agent: 4- (β-D-Ribofuranosyl)-4,5,6,7-tetrahydrothiazolo[4,5-d]pyrimidine-5,7-dione (Thioanalog of 3-Isoxanthosine). Chemical & Pharmaceutical Bulletin 1974, 22, 1198-1200.
(100) Athmani, S.; Iddon, B. Azoles. Part 10. Thiazolo[4',5':4,5]thieno[3,2- d]pyrimidine, a New Heterocyclic Ring System. Tetrahedron 1992, 48, 7689- 7702.
(101) Zhu, X.-F. The Latest Progress in the Synthesis of Carbocyclic Nucleosides. Nucleosides, Nucleotides and Nucleic Acids 2000, 19, 651-690.
(102) Crimmins, M. T. New Developments in the Enantioselective Synthesis of Cyclopentyl Carbocyclic Nucleosides. Tetrahedron 1998, 54, 9229-9272.
272 (103) Marquez, V. E. Carbocyclic Nucleosides. Advances in Antiviral Drug Design; JAI Press: Greenwich, 1996; pp 89-146.
(104) Glazer, R. I.; Knode, M. C. Neplanocin A. A Cyclopentyl Analog of Adenosine with Specificity for Inhibiting RNA Methylation. Journal of Biological Chemistry 1984, 259, 12964-12969.
(105) Borchardt, R. T.; Keller, B. T.; Patel-Thombre, U. Neplanocin A. A Potent Inhibitor of S-Adenosylhomocysteine Hydrolase and of Vaccinia Virus Multiplication in Mouse L929 Cells. Journal of Biological Chemistry 1984, 259, 4353-4358.
(106) De Clercq, E. S-Adenosylhomocysteine Hydrolase Inhibitors as Broad-Spectrum Antiviral Agents. Biochemical Pharmacology 1987, 36, 2567-2575.
(107) Chiang, P. K.; Miura, G. A. S-Adenosylhomocysteine Hydrolase. Biological Methylation and Drug Design; Humana Press: Clifton, NJ, 1986; pp 239-251.
(108) Wolfe, M. S.; Lee, Y.; Bartlett, W. J.; Borcherding, D. R.; Borchardt, R. T. 4'- Modified Analogues of Aristeromycin and Neplanocin A: Synthesis and Inhibitory Activity toward S-Adenosyl-L-homocysteine Hydrolase. Journal of Medicinal Chemistry 1992, 35, 1782-1791.
(109) Narayanan, S. R.; Keller, B. T.; Borcherding, D. R.; Scholtz, S. A.; Borchardt, R. T. 9-(trans-2', trans-3'-Dihydroxycyclopent-4'-enyl) Derivatives of Adenine and 3-Deazaadenine: Potent Inhibitors of Bovine Liver S-Adenosylhomocysteine Hydrolase. Journal of Medicinal Chemistry 1988, 31, 500-503.
(110) Hong, J. H.; Shim, M. J.; Ro, B. O.; Ko, O. H. An efficient Synthesis of Novel Carbocyclic Nucleosides with Use of Ring-Closing Metathesis from D-Lactose. Journal of Organic Chemistry 2002, 67, 6837-6840.
(111) Ludek, O. R.; Meier, C. New Convergent Synthesis of Carbocyclic Nucleoside Analogues. Synthesis 2003, 13, 2101-2109.
(112) Siddiqi, S. M.; Schneller, S. W.; Ikeda, S.; Snoeck, R.; Andrei, G. et al. S- Adenosyl-L-homocysteine Hydrolase Inhibitors as Antiviral Agents: 5'- Deoxyaristeromycin. Nucleosides & Nucleotides 1993, 12, 185-198.
273 (113) Song, G. Y.; Paul, V.; Choo, H.; Morrey, J.; Sidwell, R. W. et al. Enantiomeric Synthesis of D- and L-Cyclopental Nucleosides and Their Antiviral Activity Against HIV and West Nile Virus. Journal of Medicinal Chemistry 2001, 44, 3985-3993.
(114) Ryong, M. H.; Choi, W. J.; Kim, H. O.; Jeong, L. S. Improved and Alternative Synthesis of D- and L-Cyclopentenone Derivatives, teh Versatile Intermediates for the Synthesis of Carbocyclic Nucleosides. Tetrahedron: Asymmetry 2002, 13, 1189-1193.
(115) Genovesi, E. V.; Lamb, L.; Medina, I.; Taylor, D.; Seifer, M. et al. Antiviral Efficacy of Lobucavir (BMS-180194), a Cyclobutyl-guanosine Nucleoside Analogue, in the Woodchuck (Marmotamonax) Model of Chronic Hepatitus B Virus (HBV) Infection. Antiviral Research 2000, 48, 197-203.
(116) Tenney, D. J.; Yamanaka, G.; Voss, S. M.; Cianci, C. W.; Tuomari, A. V. et al. Lobucavir is Phosphorylated in Human Cytomegalovirus-infected and -uninfected Cells and Inhibits the Viral DNA Polymerase. Antimicrobial Agents and Chemotherapy 1997, 41, 2680-2685.
(117) Vince, R. Synthesis and Anti-HIV Activity of Carbovir and Related Carbocyclic Nucleosides. Nucleic Acids Symposium Series, 1991; pp 193-194.
(118) Yeom, Y. H.; Remmel, R. P.; Huang, S. H.; Hua, M.; Vince, R. et al. Pharmacokinetics and Bioavailability of Carbovir, a Carbocyclic Nucleoside Active Against Human Immunodeficiency Virus, in Rats. Antimicrobial Agents and Chemotherapy 1989, 33, 171-175.
(119) Vince, R.; Hua, M.; Brownell, J.; Daluge, S.; Lee, F. et al. Potent and Selective Activity of a New Carbocyclic Nucleoside Analog (Carbovir: NSC 614846) Against Human Immunodeficiency Virus in vitro. Biochemical and Biophysical Research Communications 1988, 1516, 1046-1053.
(120) Carter, S. G.; Kessler, J. A.; Rankin, C. D. Activities of (-)-Carbovir and 3'- Azido-3'-deoxythymidine Against Human Immunodeficiency Virus in vitro. Antimicrobial Agents and Chemotherapy 1990, 34, 1297-1300.
(121) Zimmerman, T. P.; Wolberg, G.; Duncan, G. S.; Elion, G. B. Adenosine Analogues as Substrates and Inhibitors of S-Adenosylhomocysteine Hydrolase in Intact Lymphocytes. Biochemistry 1980, 19, 2252-2259.
274 (122) Pugh, C. S. G.; Borchardt, R. T. Effects of S-Adenosylhomocysteine Analogues on Vaccinia Viral Messenger Ribonucleic Acid Synthesis and Methylation. Biochemistry 1982, 21, 1535-1541.
(123) Chiang, P. K.; Richards, H. H.; Cantoni, G. L. S-Adenosyl-L-homocysteine hydrolase: Analogs of S-Adenosyl-L-homocysteine as Potential Inhibitors. Molecular Pharmacology 1977, 13, 939-947.
(124) Svardal, A.; Djurhuus, R.; Ueland, P. M. Disposition of Homocysteine and S-3- Deazaadenosylhomocysteine in Cells Exposed to 3-Deazaadenosine. Molecular Pharmacology 1986, 30, 154-158.
(125) Bennett, L. L., Jr.; Brockman, R. W.; Allan, P. W.; Rose, L. M.; Shaddix, S. C. Alterations in Nucleotide Pools Induced by 3-Deazaadenosine and Related Compounds. Biochemical Pharmacology 1988, 37, 1233-1244.
(126) Chiang, P. K.; Im, Y. S.; Cantoni, G. L. Phospholipids Biosynthesis by Methylations and Choline Incorporation: Effect of 3-Deazaadenosine. Biochemical and Biophysical Research Communications 1980, 94, 174-181.
(127) Guranowski, A.; Montgomery, J. A.; Cantoni, G. L.; Chaing, P. K. Adenosine Analogues as Subetrates and Inhibitors of S-Adenosylhomocysteine Hydrolase. Biochemistry 1981, 20, 110-115.
(128) Gordon, R. K.; Ginalski, K.; Rudnicki, W. R.; Rychlewski, L.; Pankaskie, M. C. et al. Anti-HIV-1 Activity of 3-Deaza-adenosine Analogues: Inhibition of S- Adenosylhomocysteine Hydrolase and Nucleotide Congeners. European Journal of Biochemistry 2003, 270, 3507-3517.
(129) Liu, M.-C.; Luo, M.-Z.; Mozdziesz, D. E.; Lin, T.-S.; Dutschman, G. E. et al. Synthesis of Halogen-substituted 3-Deazaadenosine and 3-Deazaguanosine Analogues as Potential Antitumor/Antiviral agents. Nucleosides, Nucleotides and Nucleic Acids 2001, 20, 1975-2000.
(130) Bodner, A. J.; Cantoni, G. L.; Chaing, P. K. Biochemical and Biophysical Research Communications 1981, 98, 476-481.
(131) Flexner, C. W.; Hildreth, J. E.; Kunci, R. W.; Drachman, D. B. Lancet 1992, 339, 438.
275 (132) Bader, J. P.; Brown, N. R.; Chiang, P. K.; Cantoni, G. L. 3-Deazaadenosine, an Inhibitor of Adenosylhomocysteine Hydrolase, Inhibits Reproduction of Rous Sarcoma Virus and Transformation of Chick Embryo Cells. Virology 1978, 89, 494-505.
(133) Chaing, P. K.; Cantoni, G. L.; Bader, J. P.; Shannon, W. M.; Thomas, H. J. et al. Biochemical and Biophysical Research Communications 1978, 82, 417-423.
(134) Chiang, P. K. Conversion of 3T3-L1 Fibroblasts to Fat Cells by an Inhibitor of Methylation: Effect of 3-Deazaadenosine. Science 1981, 211, 1164-1166.
(135) Glazer, R. I.; Hartman, K. O.; Knode, M. C.; Richard, M. M.; Chiang, P. K. et al. 3-Deazaneplanocin: A New and Potent Inhibitor of S-Adenosylhomocysteine Hydrolase and its Effects on Human Promyelocytic Leukemia Cell Line HL-60. Biochemical and Biophysical Research Communications 1986, 135, 688-694.
(136) Houston, D. M.; Dolence, E. K.; Keller, B. T.; Patel-Thombre, U.; Borchardt, R. T. Potential Inhibitors of S-Adenosylmethionine-Dependent Methyltransferases. 8. Molecular Dissections of Carbocyclic 3-Deazaadenosine as Inhibitors of S- Adenosylhomocysteine Hydrolase. Journal of Medicinal Chemistry 1985, 28, 467-471.
(137) Hwang, M. J.; Stockfisch, T. P.; Hagler, A. T. Derivation of Class II Force Fields. 2. Derivation and Characterization of a CLass II Force Field,CFF93, for the Alkyl Functional Group and Alkane Molecules. Journal of the American Chemical Society 1994, 116, 2515-2525.
(138) Hagler, A. T.; Huler, E.; Lifson, S. Energy Functions for Peptides and Proteins. I. Derivation of a Consistent Force Field Including the Hydrogen Bond from Amide Crystals. Journal of the American Chemical Society 1974, 96, 5319-5327.
(139) Francl, M. M.; Pietro, W. J.; Hehre, W. J.; Binkley, J. S.; Gordon, M. S. et al. Self-consistent Molecular Orbital Methods. XXIII. A Polarization-Type Basis Set for Second-Row Elements. Journal of Chemical Physics 1982, 77, 3654-3665.
(140) Kaapro, A.; Ojanen, J. Protein Docking. 2002, http://www.lce.hut.fi/teaching/S- 114.500/k2002/Protdock.pdf.
276 (141) Varma, C. K. Molecular Mechanics Force Fields: Review and Critical Analysis of Modern Day Force Fields with Application to Protein and Nucleic Acid Structure. Biochemistry 218: Stanford University, 2001; 11.
(142) Sun, H.; Mumby, S. J.; Maple, J. R.; Hagler, A. T. An ab Initio CFF93 All-Atom Force Field for Polycarbonates. Journal of the American Chemical Society 1994, 7, 2978-2987.
(143) Weiner, S. J.; Kollman, P. A.; Case, D. A.; Singh, U. C.; Ghio, C. et al. A New Force Field for Molecular Mechanical Simulation of Nucleic Acids and Proteins. Journal of the American Chemical Society 1984, 106, 765-784.
(144) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M. J. et al. A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids and Organic Molecules. Journal of the American Chemical Society 1995, 117, 5179- 5197.
(145) Hagler, A. T.; Lifson, S.; Dauber, P. Consistent Force Field Studies of Intermolecular Forces in Hydrogen-Bonded Crystals. 2. A Benchmark for the Objective Comparison of Alternative Force Fields. Journal of the American Chemical Society 1979, 101, 5122-5130.
(146) Hagler, A. T.; Dauber, P.; Lifson, S. Consistant Force Field Studies of Intermolecular Forces in Hydrogen-Bonded Crystals. 3. The C=O...H_O Hydrogen Bond and the Analysis of the Energetics and Packing of Carboxylic Acids. Journal of the American Chemical Society 1979, 101, 5131-5141.
(147) Dinur, U.; Hagler, A. T. Direct Evalutation of Nonbonding Interactions from ab Initio Calculations. Journal of the American Chemical Society 1989, 111, 5149- 5151.
(148) Maple, J. R.; Hwang, M. J.; Jalkanen, K. J.; Stockfisch, T. P.; Hagler, A. T. Derivation of Class II Force Fields: V. Quantum Force Field for Amides, Peptides, and Related Compounds. Journal of Computational Chemistry 1998, 19, 430-458.
(149) Hwang, M. J.; Ni, X.; Waldman, M.; Ewig, C. S.; Hagler, A. T. Derivation of Class II Force Fields. VI. Carbohydrate Compounds and Anomeric Effects. Biopolymers 1998, 45, 435-468.
277 (150) Accelrys CFF 2003, http://www.bio.unizh.ch/docu/acc_docs/doc/cff/cff95_1.html.
(151) Metropolis, N.; Rosenbluth, A.; Rosenbluth, M. N.; Teller, A. H.; Teller, E. Equation of State Calculations by Fast Computing Machines. The Journal of Chemical Physics 1953, 21, 1087-1092.
(152) Greengard, L.; Rokhlin, V. A Fast Algorithm for Particle Simulations. Journal of Computational Physics 1987, 73, 325-348.
(153) Ding, H.-Q.; Karasawa, N.; Goddard, W. A. I. Atomic Level Simulations of a Million Particles: The Cell Multipole Method for Coulomb and London Nonbond Interactions. Journal of Chemical Physics 1992, 97, 4309-4315.
(154) Ding, H.-Q.; Karasawa, N.; Goddard, W. A. I. The Reduced Cell Multipole Method for Coulomb Interactions in Periodic Systems with Million-Atom Unit Cells. Chemical Physics Letters 1992, 196, 6-10.
(155) Greengard, L. Fast Algorithms for Classical Physics. Science 1994, 265, 909-914.
(156) Goto, M.; Omi, R.; Nakagawa, N.; Miyahara, I.; Hirotsu, K. Crystal Structures of CTP Synthase Reveal ATP,UTP, and Glutamine Binding Sites. Structure 2004, 12, 1413-1423.
(157) Huang, M.; Wang, Y.; Collins, M.; Graves, L. M. CPEC Induces Erythroid Differentiation of Human Myeloid Leikemia K562 Cells Through CTP Depletion and p38 MAP Kinase. Leukemia 2004, 18, 1857-1863.
(158) Yin, D.; Yang, X.; Hu, Y.; Kuczera, K.; Schowen, R. L. et al. Substrate Binding Stabilizes S-Adenosylhomocysteine Hydrolase in a Closed Conformation. Biochemistry 2000, 39, 9811-9818.
(159) Seley, K. L.; Zhang, L.; Hagos, A. "Fleximers". Design and Synthesis of Two Novel Split Nucleosides. Organic Letters 2001, 3, 3209-3210.
(160) Seley, K. L.; Zhang, L.; Hagos, A.; Quirk, S. "Fleximers". Design and Synthesis of a New Class of Novel Shape-Modified Nucleosides. Journal of Organic Chemistry 2002, 67, 3365-3373.
278 (161) Seley, K. L.; Salim, S.; Zhang, L.; O'Daniel, P. I. "Molecular Camelions". Design and Synthesis of a Second Series of Flexible Nucleosides. Journal of Organic Chemistry 2005, 70, 1612-1619.
(162) Kahn, A. R.; Parrish, J. C.; Fraser, M. E.; Smith, W. W.; Bartlett, P. A. et al. Lowering the Entropic Barrier for Binding Conformationally Flexible Inhibitors to Enzymes. Biochemistry 1998, 37, 16839-16845.
(163) Wilson, E. K. Dealing with Flexible Receptors. In Chemical & Engeneering News, 2004; pp 46-47.
(164) Das, K.; Clark, A. D. J.; Lewi, P. J.; Heeres, J.; de Jonge, M. R. et al. Roles of Conformational and Positional Adaptability in Structure-Based Design of TMC125-R165335 (Etravirine and Related Non-nucleoside Reverse Transcriptase Inhibitors that are Highly Potent and Effective against Wild-Type and Drug- Resistant HIV-1 Variants. Journal of Medicinal Chemistry 2004, 47, 2550-2560.
(165) Tuske, S.; Sarafianos, S. G.; Clark, A. D. J.; Ding, J.; Naeger, L. K. et al. Structure of HIV-1 RT-DNA Complexes Before and After Incorperation of the Anti-AIDS Drug Tenofovir. Nature of Structural & Molecular Biology 2004, 11, 469-474.
(166) Popescu, A.; Hornfeldt, A.-B.; Gronowitz, S. Catalytic Osmylation and Antiviral Activity of Some Carbocyclic 5-substituted Uridine and Cytidine Analogues. Nucleosides & Nucleotides 1995, 14, 1639-1657.
(167) Herdewijn, P. 5-Substituted-2'-deoxyuridines as Anti-HSV-1 Agents: Synthesis and Structure Activity Relationship. Antiviral Chemistry and Chemotherapy 1994, 5, 131-146.
(168) De Winter, H.; Herdewijn, P. Understanding the Binding of 5-Substituted 2'- Deoxyuridine Substrates to Thymidine Kinase of Herpes Simplex Virus Type-1. Journal of Medicinal Chemistry 1996, 39, 4727-4737.
(169) Harris, D. G.; Shao, J.; Morrow, B. D.; Zimmerman, S. S. Molecular Modeling of the Binding of 5-Substituted 2'-Deoxyuridine Substrates to Thymidine Kinase of Herpes Simplex Virus Type-1. Nucleosides, Nucleotides and Nucleic Acids 2004, 23, 555-565.
279 (170) Purwanto, M. G. M.; Weisz, K. Binding of Imidazole-Derived Nucleosides to a CG Base Pair. Journal of Organic Chemistry 2004, 69, 195-197.
(171) Lengeler, D.; Weisz, K. New Nucleobase Analogs for the Extension of the Triple Helix Recognition code. Nucleosides & Nucleotides 1999, 18, 1657-1658.
(172) Wang, W.; Purwanto, M. G. M.; Weisz, K. GC Base Pair Recognition by Substituted Phenylimidazole Nucleosides. Organic and Biomolecular Chemistry 2004, 2, 1194-1198.
(173) Bardon, A. B.; Wetmore, S. D. How Flexible are Fleximer Nucleobases? A Computational Study. Journal of Physical Chemistry A 2005, 109, 262-272.
(174) Polak, M.; Seley, K. L.; Plavec, J. Conformational Properties of Shape Modified Nucleosides - Fleximers. Journal of the American Chemical Society 2004, 126, 8159-8166.
(175) Carlson, H. A. Protein Flexibility is an Important Component of Structure-Based Drug DEsign. Current Pharmaceutical Design 2002, 8, 1571-1578.
(176) Carlson, H. A. Protein Flexibility and Drug Design: How to Hit a Moving Target. Current Opinion in Chemical Biology 2002, 6, 447-452.
(177) Watson, J. D.; Crick, F. H. C. Genetical Implications of the Structure of Deoxyribonucleic Acid. Nature 1953, 171, 964-967.
(178) Kool, E. T. Hydrogen Bonding, Base Stacking, and Steric Effects in DNA Replication. Ann. Rev. Biophys. Biomol. Struc. 2001, 30, 1-22.
(179) Bloomfield, V. A.; Crothers, D. M.; Tinoco, J. I. Nucleic Acids: Structure, Properties, and Functions; University Science Books: Sausalito, CA, 2000; 794.
(180) Lescrinier, E.; Froeyen, M.; Herdewijn, P. Difference in Conformational Diversity Between Nucleic Acids with a Six-Membered 'Sugar' Unit and Natural 'Furanose' Nucleic Acids. Nucleic Acids Research 2003, 31, 2975-2989.
(181) Brown, T. A. Genomes; 2nd ed.; BIOS Scientific Publishers Ltd, 2005.
280 (182) Hecht, S. M. Bioorganic Chemistry: Nucleic Acids; Oxford University Press: New York, 1996; 500.
(183) Kennard, O.; Hunter, W. N. Oligonucleotide Structure: a Decade of Results from Single Crystal X-ray Diffraction Studies. Quarterly Reviews of Biophysics 1989, 22, 327-379.
(184) Kool, E. T.; Morales, J. C.; Guckian, K. M. Mimicking the Structure and Function of DNA: Insights into DNA Stability and Replication. Angewandte Chemie, International Edition In English 2000, 39, 990-1009.
(185) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils, C. J.; Paris, P. L. et al. Experimental Measurement of Aromatic Stacking Affinities in the Context of Duplex DNA. Journal of the American Chemical Society 1996, 118, 8182-8183.
(186) Matray, T. J.; Kool, E. T. Selective and Stable DNA Base Pairing Without Hydrogen Bonds. Journal of the American Chemical Society 1998, 120, 6191- 6192.
(187) Guckian, K. M.; Krugh, T. R.; Kool, E. T. Solution Structure of a Nonpolar, Non- Hydrogen-Bonded Base Pair Surrogate in DNA. Journal of the American Chemical Society 2000, 122, 6841-6847.
(188) Burkard, M. E.; Kierzek, R.; Turner, D. H. Thermodynamics of Unpaired Terminal Nucleotides on Shirt RNA Helixes Correlates with Stacking at Helix Termini in Larger RNAs. J. Mol. Biol. 1999, 290, 967-982.
(189) Henry, A. A.; Romesberg, F. E. Beyond A, C, G, and T: Augmenting Nature's Alphabet. Current Opinion in Chemical Biology 2003, 7, 727-733.
(190) Lai, J. S.; Qu, J.; Kool, E. T. Fluorinated DNA Bases as Probes of Electrostatic Effects in DNA Base Stacking. Angew. Chem., Int. Ed. Engl. 2003, 42, 5973- 5977.
(191) Lin, K.-Y.; Matteucci, M. D. A Cytosine Analogue Capable of Clamp-Like Binding to a Guanine in Helical Nucleic Acids. Journal of the American Chemical Society 1998, 120, 8531-8532.
281 (192) Matsuda, S.; Henry, A. A.; Schultz, P. G.; Romesberg, F. E. The Effect of Minor- Groove Hydrogen-Bond Acceptors and Donors on the Stability and Replication of Four Unnatural Base Pairs. Journal of the American Chemical Society 2003, 125, 6134-6139.
(193) Minakawa, N.; Kojima, N.; Hikishima, S.; Sasaki, T.; Kiyosue, A. et al. New Base Paring Motifs. The Synthesis and Thermal Stability of Oligodeoxynucleotides Containing Imidazopyridopyrimidine Nucleosides with the Ability to Form Four Hydrogen Bonds. J. Am. Chem. Soc. 2003, 125, 9970- 9982.
(194) Newcomb, L. F.; Gellman, S. H. Aromatic Stacking Interactions in Aqueous Solution: Evidence That Neither Classical Hydrophobic Effects nor Dispersion Forces are Important. Journal of the American Chemical Society 1994, 116, 4993- 4994.
(195) Berger, M.; Luzzi, S. D.; Henry, A. A.; Romesberg, F. E. Stability and Selectivity of Unnatural DNA with Five-Membered-Ring Nucleobase Analogues. Journal of the American Chemical Society 2002, 124, 1222-1226.
(196) Brotschi, C.; Haberli, A.; Leumann, C. J. A Stable DNA Dluplex Containing a Non-Hydrogen-Bonding and Non-Shape-Complenentary Base Couple: Interstrand Stacking as the Stability Determining Factor. Angew. Chem., Int. Ed. Engl. 2001, 40, 3012-3014.
(197) Henry, A. A.; Yu, C.; Romesberg, F. E. Determinants of Unnatural Nucleobase Stability and Polymerase Recognition. Journal of the American Chemical Society 2003, 125, 9638-9646.
(198) McMinn, D. L.; Ogawa, A. K.; Wu, Y.; Liu, J.; Schultz, P. G. et al. Efforts Toward Expansion of the Genetic Alphabet: DNA Polymerase Recognition of a Highly Stable, Self-Pairing Hydrophobic Base. Journal of the American Chemical Society 1999, 121, 11585-11586.
(199) Ohmichi, T.; Nakano, S.-i.; Miyoshi, D.; Sugimoto, N. Long RNA Dangling End Has Large Energetic Contribution to Duplex Stability. J. Am. Chem. Soc. 2002, 124, 10367-10372.
(200) Rosemeyer, H.; Seela, F. Modified Purine Nucleosides as Dangling Ends of DNA Duplexes: The Effect of the Nucleobase Polarizability on Stacking Interactions. J. Chem. Soc., Perkin Trans. 2 2002, 746-750.
282 (201) Switzer, C.; Moroney, S. E.; Benner, S. A. Enzymatic Incorporation of a New Base Pair into DNA and RNA. Journal of the American Chemical Society 1989, 111, 8322-8323.
(202) Tae, E. L.; Wu, Y.; Xia, G.; Schultz, P. G.; Romesberg, F. E. Efforts Toward Expansion of the Genetic Alphabet: Replication of DNA with Three Base Pairs. J. Am.Chem. Soc. 2001, 123, 7439-7440.
(203) Sponer, J.; Jurecka, P.; Hobza, P. Accurate Interaxtion Energies of Hydrogen- Bonded Nucleic Acid Base Pairs. Journal of the American Chemical Society 2004, 126, 10142-10151.
(204) Grunenberg, J. Direct Assessment of Interresidue Forces in Watson-Crick Base Pairs Using Theoretical Compliance Constants. Journal of the American Chemical Society 2004, 126, 16310-16311.
(205) Voegel, J. J.; Benner, S. A. Nonstandard Hydrogen Bonding in Duplex Oligonucleotides. The Base Pair Between and Acceptor-Donor-Donor Pyrimidine Analog and a Donor-Acceptor-Acceptor Purine Analog. Journal of the American Chemical Society 1994, 116, 6929-6930.
(206) Benner, S. A.; Battersby, T. R.; Eschgfaller, B.; Hutter, D.; Kodra, J. T. et al. Redesigning Nucleic Acids. Pure and Applied Chemistry 1998, 70, 263-266.
(207) Benner, S. A. Understanding Nucleic Acids Using Synthetic Chemistry. Acc. Chem. Res. 2004, 37, 784-797.
(208) Martinot, T. A.; Benner, S. A. Artifical Genetic Systems: Exploiting the "Aromaticity" Formalism to Improve the Tautomeric Ratio for Isoguanosine Derivatives. Journal of Organic Chemistry 2004, 69, 3972-3975.
(209) Hutter, D.; Benner, S. A. Expanding the Genetic Alphabet: Non-Epimerizing Nucleoside with the pyDDA Hydrogen Bonding Pattern. J. Org. Chem. 2003, 68, 9839-9842.
(210) Roberts, C.; Bandaru, R.; Switzer, C. Theoretical and Experimental Study of Isoguanine and Isocytosine: Base Pairing in an Expanded Genetic System. Journal of the American Chemical Society 1997, 119, 4640-4649.
283 (211) Robinson, H.; Gao, Y.-G.; Bauer, C.; Roberts, C.; Switzer, C. et al. 2'- Deoxyisoguanosine Adopts More than Tautomer to Form Base Pairs with Thymidine Observed by High-resolution Crystal Structure Analysis. Biochemistry 1998, 37, 10897-10905.
(212) Rice, K. P.; Chaput, J. C.; Cox, M. M.; Switzer, C. RecA Protein Promotes Strand Exchange with DNA Substrates Containing Isoguanine and 5-Methyl Isocytosine. Biochemistry 2000, 39, 10177-10188.
(213) Blas, J. R.; Luque, F. J.; Orozco, M. Unique Tautomeric Properties of Isoguanine. Journal of the American Chemical Society 2004, 126, 154-164.
(214) Bailly, C.; Waring, M. J.; Travers, A. A. Effect of Base Substitutions on the Binding of a DNA-Bending Protein. Journal of Molecular Biology 1995, 253, 1- 7.
(215) Bailly, C.; Payet, D.; Travers, A. A.; Waring, M. J. PCR-Based Development of DNA Substrates Containing Modified Bases: An Efficient System for Investigating the Role of the Exocyclic Groups in Chemical and Structural Recognition by Minor Groove Binding Drugs and Proteins. Proceedings of the National Academy of Sciences, U.S.A. 1996, 93, 13623-13628.
(216) Mollegaard, N. E.; Bailly, C.; Waring, M. J.; Nielsen, P. E. Effects of Diaminopurine and Inosine Substitutions on A-Tract Induced DNA Curvature. Importance of the 3'-A-Tract Junction. Nucleic Acids Research 1997, 25, 3497- 3502.
(217) Bailly, C.; Waring, M. J. The Use of Diaminopurine to Investigate Structural Properties of Nucleic Acids and Molecular Recognition Between Ligands and DNA. Nucleic Acids Research 1998, 26, 4309-4314.
(218) Lan, T.; McLaughlin, L. W. Minor Groove Hydration is Critical to the Stability of DNA Duplexes. Journal of the American Chemical Society 2000, 122, 6512-6513.
(219) Lan, T.; McLaughlin, L. W. Minor Groove Functional Groups Are Critical for the B-Form Conformation of Duplex DNA. Biochemistry 2001.
(220) Wilds, C. J.; Maier, M. A.; Manoharan, M.; Egli, M. Structural Basis for Recognition of Guanosine by a Synthetic Tricyclic Cytosine Analogue: Guanidinium G-Clamp. Helvetica Chimica Acta 2003, 86, 966-977.
284 (221) Guckian, K. M.; Schweitzer, B. A.; Ren, R. X.-F.; Sheils, C. J.; Tahmassebi, D. C. et al. Factors Contributing to Aromatic Stacking in Water: Evaluation in the Context of DNA. J. Am. Chem. Soc. 2000, 122, 2213-2222.
(222) Bommarito, S.; Peyret, N.; SantaLucia, J. J. Thermodynamic Parameters for DNA Swquences with Dangling Ends. Nucleic Acids Research 2000, 28, 1929-1934.
(223) Kim, T. W.; Kool, E. T. A Series of Nonpolar Thymidine Analogues of Increasing Size: DNA Base Pairing and Stacking Properties. Journal of Organic Chemistry 2005, 70, 2048-2053.
(224) Schweitzer, B. A.; Kool, E. T. Hydrophobic, Non-Hydrogen-Bonding Bases and Base Pairs in DNA. Journal of the American Chemical Society 1995, 117, 1863- 1872.
(225) Henry, A. A.; Romesberg, F. E. The Evolution of DNA Polymerases with Novel Activities. Current Opinion in Biotechnology 2005, 16, 1-8.
(226) Senior, M.; Jones, R. A.; Breslauer, K. J. Influence of Dangling Thymidine Residues on the Stability and Structure of Two DNA Duplexes. Biochemistry 1988, 27, 3879-3885.
(227) Ogawa, A. K.; Wu, Y.; McMinn, D. L.; Liu, J.; Schultz, P. G. et al. Efforts Towards the Expansion of the Genetic Alphabet: Information Storage and Replication with Unnatural Hydrophobic Base Pairs. Journal of the American Chemical Society 2000, 122, 3274-3287.
(228) Wu, Y.; Ogawa, A. K.; Berger, M.; McMinn, D. L.; Schultz, P. G. et al. Efforts Toward Expansion of the Genetic Alphabet: Optimization of Interbase Hydrophobic Interactions. J. Am. Chem. Soc. 2000, 122, 7621-7632.
(229) Wu, Y.; Fa, M.; Tae, E. L.; Schultz, P. G.; Romesberg, F. E. Enzymatic Phosphorylation of Unnatural Nucleosides. J. Am.Chem. Soc. 2002, 124, 14626- 14630.
(230) Ogawa, A. K.; Wu, Y.; Berger, M.; Schultz, P. G.; Romesberg, F. E. Rational Design of an Unnatural Base Pair with Increased Kinetic Selectivity. Journal of the American Chemical Society 2000, 122, 8803-8804.
285 (231) Romesberg, F. E.; Yu, C.; Matsuda, S.; Henry, A. A. Development of a Universial Nucleobase and Modified Nucleobases for Expanding the Genetic Code. Current Protocols in Nucleic Acid Chemistry; John Wiley and Sons, Inc, 2002; pp 1.5.1- 1.5.36.
(232) Brotschi, C.; Leumann, C. J. DNA with Hydrophobic Base Substitutes: A Stable, Zipperlike Recognition Motif Based On Interstrand-Stacking Interactions. Angew. Chem., Int. Ed. Engl. 2003, 42, 1655-1658.
(233) Leonard, N. J. Adenylates: Bound and Unbound. Biopolymers 1985, 24, 9-28.
(234) Lessor, R. A.; Gibson, K. J.; Leonard, N. J. Synthesis and Biochemical Evaluation of 2'-Deoxy-lin-benzoadenosine Phosphates. Biochemistry 1984, 23, 3868-3873.
(235) Leonard, N. J.; Sprecker, M. A.; Morrice, A. G. Defined Dimensional Changes in Enzyme Substrates and Cofactors. Synthesis of lin-Benzoadenosine and Enzymatic Evaluation of Derivatives of the Benzopurines. Journal of the American Chemical Society 1976, 98, 3987-3994.
(236) Leonard, N. J.; Petric, A.; Rykowski, A. Defined Dimensional Alterations in Enzyme Substrates. Birch Reduction of lin-Benzopurines. A Contribution to Information Concerning the Binding Sites of Adenosine Deaminase and Xanthine Oxidase. Journal of Organic Chemistry 1988, 53, 3873-3875.
(237) Leonard, N. J.; Kazmierczak, F.; Rykowski, A. A Convenient Synthesis of lin- Benzopurines through a Common Intermediate. Journal of Organic Chemistry 1987, 52, 2933-2935.
(238) Lee, A. H. F.; Kool, E. T. A New Four-Base Genetic Helix, yDNA, Composed of Widened Benzopyrimidine-Purine Pairs. J. Am.Chem. Soc. 2005, 127, 3332-3338.
(239) Liu, H.; Gao, J.; Lynch, S. R.; Saito, D.; Maynard, L. et al. A Four-Base Paired Genetic Helix with Expanded Size. Science 2003, 302, 868-871.
(240) Liu, H.; Lynch, S. R.; Kool, E. T. Solution Structure of xDNA: A Paired Genetic Helix with Increased Diameter. J. Am. Chem. Soc. 2004, 126, 6900-6905.
(241) Lee, A. H. F.; Kool, E. T. Novel Benzopyrimidines as Widened Analogues of DNA Bases. J. Org. Chem. 2005, 70, 132-140.
286 (242) Liu, H.; Gao, J.; Maynard, L.; Saito, D.; Kool, E. T. Toward a New Genetic System with Expanded Dimensions: Size-Expanded Analogues of Deoxyadenosine and Thymidine. J. Am. Chem. Soc. 2004, 126, 1102-1109.
(243) Liu, H.; Gao, J.; Kool, E. T. Helix-Forming Properties of Size-Expanded DNA, an Alternative Four-Base Genetic Form. J. Am. Chem. Soc. 2005, 127, 1396-1402.
(244) Lin, K.-Y.; Jones, R. J.; Matteucci, M. D. Tricyclic 2'-Deoxycytidine Analogs: Syntheses and Incorporation into Oligonucleotides Which Have Enhanced Binding to Complementary RNA. Journal of the American Chemical Society 1995, 117, 3873-3874.
(245) Matteucci, M. D.; von Krosigk, U. Hybridization Properties of Oligonucleotides Bearing a Tricyclic 2'-Deoxycytidine Analog Based on a Carbazole Ring System. Tetrahedron Letters 1996, 37, 5057-5060.
(246) Beard, W. A.; Osheroff, W. P.; Prasad, R.; Sawaya, M. R.; Jaju, M. et al. Enzyme-DNA Interactions Required for Efficient Nucleotide Incorperation and Discrimination in Human DNA Polymerase β. The Journal of Biological Chemistry 1996, 271, 12141-12144.
(247) Subramanya, H. S.; Doherty, A. J.; Ashford, S. R.; Wigley, D. B. Crystal Strucrture of an ATP-Dependent DNA Ligase from Bacteriophage T7. Cell 1996, 85, 607-615.
(248) Vassylyev, D. G.; Kashiwagi, T.; Mikami, Y.; Ariyoshi, M.; Iwai, S. et al. Atomic Model of a Pyrimidine Dimer Excision Repair Enzyme Complexed with a DNA Substrate: Structural Basis for Damaged DNA Recognition. Cell 1995, 83, 773- 782.
(249) Morales, J. C.; Kool, E. T. Minor Groove Interactions between Polymerase and DNA: More Essential to Replication than Watson-Crick Hydrogen Bonds? Journal of the American Chemical Society 1999, 121, 2323-2324.
(250) Wemmer, D. E.; Dervan, P. B. Targeting the Minor Groove of DNA. Current Opinion in Structural Biology 1997, 7, 355-361.
(251) Dervan, P. B. Molecular Recognition of DNA by Small Molecules. Bioorganic & Medicinal Chemistry 2001, 9, 2215-2235.
287 (252) Khalaf, A. I.; Waigh, R. D.; Drummond, A. J.; Pringle, B.; McGroarty, I. et al. Distamycin Analogues with Enhanced Lipophilicity: Synthesis and Antimicrobial Activity. Journal of Medicinal Chemistry 2004, 47, 2133-2156.
(253) Hou, M.-H.; Wang, A. H.-J. Mithramycin Forms a Stable Dimeric Complex by Chelating with Fe(II): DNA-interacting Characteristics Cellular Permeation and Cytotoxicity. Nucleic Acids Research 2005, 33, 1352-1361.
(254) Van Dyke, M. W.; Dervan, P. B. Chromomycin, Mithramycin, and Olivomycin Binding Sites on Heterogeneous Deoxyribonucleic Acid. Footprinting with (Methidiumpropyl-EDTA)iron(II). Biochemistry 1983, 22, 2373-2377.
(255) Churchill, M. E.; Jones, D. N.; Glaser, T.; Hefner, H.; Searles, M. A. et al. HMG- D is an Architecture-Specific Protein that Preferentially Binds to DNA Containing the Dinucleotide TG. The EMBO Journal 1995, 14, 1264-1275.
(256) Cloutier, T. E.; Windom, J. Spontaneous Sharp Bending of Double-Stranded DNA. Molecular Cell 2004, 14, 355-362.
(257) Dickerson, R. E. DNA Bending: The Prevalence of Kinkiness and the Virtues of Normality. Nucleic Acids Research 1998, 26, 1906-1926.
(258) Wolfe, S. A.; Ferentz, A. E.; Grantcharova, V.; Churchill, M. E.; Verdine, G. L. Modifying the Helical Structure of DNA by Design: Recruitment of an Architecture-Specific Protein to an Enforced DNA Bend. Chemistry & Biology 1995, 2, 213-221.
(259) VanWye, J. D.; Bronson, E. C.; Anderson, J. N. Species-Specific Patterns of DNA Bending and Sequence. Nucleic Acids Research 1991, 19, 5253-5261.
(260) Roberts, R. J.; Cheng, X. Base Flipping. Annual Reviews in Biochemistry 1998, 67, 181-198.
(261) Kunkel, T. A.; Wilson, S. H. DNA Repair: Push and Pull of Base Flipping. Nature 1996, 384, 25-26.
(262) Banerjee, A.; Yang, W.; Karplus, M.; Verdine, G. L. Structure of a Repair Enzyme Interrogating Undamaged DNA Elucidates Recognition of Damaged DNA. Nature 2005, 434, 612-618.
288 (263) Lloyd, R. S.; Cheng, X. Mechanistic Link Between DNA Methyltransferases and DNA Repair ENzymes by Base Flipping. Biopolymers 1997, 44, 139-151.
(264) Huang, N.; Banavali, N. K.; MacKerell, A. D. J. Protein-Facilitated Base Flipping in DNA by Cytosine-5-methyltransferase. PNAS 2003, 100, 68-73.
(265) Szyf, M. The DNA Methylation Machinery as a Target for Anticancer Therapy. Pharmacol. Ther. 1996, 70, 1-37.
(266) Klimasauskas, S.; Kumar, S.; Roberts, R. J.; Cheng, X. Cell 1994, 76, 357-369.
(267) Reinisch, K. M.; Chen, L.; Verdine, G. L.; Lipscomb, W. N. The Crystal Structure of HaeIII Methyltransferase Covalently Complexed to DNA: An Extrahelical Cytosine and Rearranged Base Pairing. Cell 1995, 82, 143-153.
(268) Slupphaug, G.; Mol, C. D.; Kavli, B.; Arvai, A. S.; Krokan, H. E. et al. Nature 1996, 384, 87-92.
(269) Vassylyev, D. G.; Morikawa, K. DNA-Repair Enzymes. Current Opinion in Structural Biology 1997, 7, 103-109.
(270) Vassylyev, D. G.; Kashiwagi, T.; Mikami, Y.; Ariyoshi, M.; Iwai, S. Structure 1995, 4, 1381-1385.
(271) Roberts, R. J. On Base Flipping. Cell 1995, 82, 9-12.
(272) Huang, N.; MacKerell, A. D. J. Atomistic View of Base Flipping in DNA. Phil. Trans. R. Soc. Lond. A 2004, 362, 1439-1460.
(273) Nelson, H., C. M.; Bestor, T. H. Base Eversion and Shuffling by DNA Methyltransferases. Chemistry & Biology 1996, 3, 419-423.
(274) Vassylyev, D. G.; Morikawa, K. Precluding Uracil from DNA. Structure 1996, 4, 1381-1385.
(275) Lavery, R.; Varnai, P. Base Flipping in DNA: Pathways and Energetics Studied with Molecular Dynamic Simulations. Journal of the American Chemical Society 2002, 124, 7272-7273.
289 (276) Lavery, R.; Giudice, E. Nucleic Acid Base Pair Dynamics: The Impact of Sequence and Structure Using Free-Energy Calculations. Journal of the American Chemical Society 2003, 125, 4998-4999.
(277) Horton, J. R.; Ratner, G.; Banavali, N. K.; Huang, N.; Choi, Y. et al. Caught in the Act: Visualization of an Intermediate in the DNA Base-Flipping Pathway Induced by HhaI Methyltransferase. Nucleic Acids Research 2004, 13, 3877- 3886.
(278) Wang, P.; Brank, A. S.; Banavali, N. K.; Nicklaus, M. C.; Marquez, V. E. et al. Use of Oligodeoxyribonucleotides with Conformationally Constrained Abasic Sugar Targets to Probe the Mechanism of Base Flipping by HhaI DNA (Cytosine C-5)-Methyltransferase. Journal of the American Chemical Society 2000, 122, 12422-12434.
(279) Duan, Y.; Kollman, P. A.; Harvey, S. C. Protein Folding and Beyond. Chemistry for the 21st Century; Wiley-VCH: Weinheim, Germany, 2000.
(280) Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. Numerical Intergration of the Cartesian Equations of Motion of a System with Constraints: Molecular Dynamics of n-Alkanes. Journal of Computational Physics 1977, 23, 327-341.
(281) Jorgensen, W. L.; Chandrasekar, J.; Madura, J.; Impey, R.; Klein, M. Comparison of Simple Potential Functions for Simulating Liquid Water. Journal of Chemical Physics 1983, 79, 926-935.
(282) Wang, W.; Donini, O.; Reyes, C. M.; Kollman, P. A. Biomolecular Simulations: Recent Developments in Force Fields, Simulations of Enzyme Catalysis, Protein- Ligand, Protein-Protein, and Protein-Nucleic Acid Noncovalent Interactions. Ann. Rev. Biophys. Biomol. Struc. 2001, 30.
(283) Stilz, H. U.; Dervan, P. B. Specific Recognition of CG Base Pairs by 2- Deoxynebularine within the Purine·Purine·Pyrimidine Triple-Helix Motif. Biochemistry 1993, 32, 2177-2185.
(284) Griffin, L. C.; Kiessling, L. L.; Beal, P. A.; Gillespie, P.; Dervan, P. B. Recognition of All Four Base Pairs of Double-Helical DNA by Triple-Helix Formatin: Design of Nonnatural Deoxyribonucleosides for Pyrimidine·Purine Base Pair Binding. Journal of the American Chemical Society 1992, 114, 7976- 7982.
290 (285) Horne, D. A.; Dervan, P. B. Recognition of Mixed-Sequence Duplex DNA by Alternate-Strand Triple-Helix Formation. Journal of the American Chemical Society 1990, 112, 2435-2437.
(286) Purwanto, M. G. M.; Lengeler, D.; Weisz, K. Nucleosides Derived from Urocanic Acid: Potential Ligands for GC Base Pairs. Tetrahedron Letters 2002, 43, 61-64.
(287) Spackova, N.; Cubero, E.; Sponer, J.; Orozco, M. Theoretical Study of the Guanine- 6-Tioguanine Substitution in Duplexes,Triplexes, and Tetraplexes. Journal of the American Chemical Society 2004, 126, 14642-14650.
(288) Hamelberg, D.; Williams, L. D.; Wilson, W. D. Effect of a Neutralized Phosphate Backbone on the Minor Groove of B-DNA: Molecular Dynamics Simulation Studies. Nucleic Acids Research 2002, 30, 3615-3623.
(289) Lu, X.-J.; Olson, W. K. 3DNA: A Software Package for the Analysis Rebuilding and Visualization of Three-Dimensional Nucleic Acid Structures. Nucleic Acids Research 2003, 31, 5108-5121.
(290) Tai, J. C.; Allinger, N. L. Molecular Mchanics Calculations on Conjugated Nitrogen-Containing Heterocycles. Journal of the American Chemical Society 1988, 110, 2050-2055.
(291) Tai, J. C.; Lii, J.-H.; Allinger, N. L. A Molecular Mechanics (MM2) Study of Furan, Thiophene, and Related Compounds. Journal of Computational Chemistry 1989, 10, 635-647.
(292) Anderson, R. N.; Smith, B. L. Deaths: Leading Causes for 2002. National Vital Statistics Report 2005, 53, 1-90.
(293) Wajed, S. A.; Laird, P. W.; DeMeester, T. R. DNA Methylation: An Alternative Pathway to Cancer. Annals of Surgery 2001, 234, 10-20.
(294) Young, J.; Fritz, A.; Gonghua, L.; Roffers, S. Cancer Historic Perspective 2005, Http://training.seer.cancer.gov/module_cancer_disease/unit1_historic_perspective .html.
291 (295) Cancer Timeline 2003, http://www.chemheritage.org/EducationalService/pharm/chemo/readings/timeline .htm.
(296) Jubb, A. M.; Bell, S. M.; Quirke, P. Methylation and Colorectal Cancer. Journal of Pathology 2001, 195, 111-134.
(297) Turner, M. A.; Yang, X.; Yin, D.; Kuczera, K.; Borchardt, R. T. et al. Structure and Function of S-Adenosylhomocysteine Hydrolase. Cell Biochemistry and Biophysics 2000, 33, 101-125.
(298) Backlund, P. S. J.; Carotti, D.; Cantoni, G. L. Effects of the S- Adenosylhomocysteine Hydrolase Inhibitors 3-Deazaadenosine and 3- Deazaaristeromycin on Methylation and Synthesis. European Journal of Biochemistry 1986, 160, 245-251.
(299) Wahnon, D. C.; Shier, V. K.; Benkovic, S. J. Mechanism-based Inhibition of an Essential Bacterial Adenine DNA Methyltransferase: Rationally Designed Antibiotics. Journal of the American Chemical Society 2001, 123, 976-977.
(300) Szyf, M. Targeting DNA Methyltransferase in Cancer. Cancer and Metastasis Reviews 1998, 17, 219-231.
(301) Verdine, G. L. Cell 1994, 76, 197-200.
(302) Borchardt, R. T. S-Adenosyl-L-Methionine-Dependent Macromolecule Methyltransferases: Potential Targets for the Design of Chemotherapeutic Agents. Journal of Medicinal Chemistry 1980, 23, 347-357.
(303) Borchardt, R. T.; Creveling, C. R.; Ueland, P. M. Biological Methylation and Drug Design; Humana Press: Clifton, NJ, 1986.
(304) Cantoni, G. The Centrality of S-Adenosylhomocysteinase in the Regulation of the Biological Utilization of S-Adenosylmethionine. Biological Methylation and Drug Design; Humana Press: Clifton, NJ, 1986; pp 227-238.
(305) Boland, C. R.; Sinicrope, F. A.; Brenner, D. E.; Carethers, J. M. Colorectal Cancer Prevention and Treatment. Gastro. 2000, 118, S115-S128.
292 (306) Chen, L. S.; Sheppard, T. L. Synthesis and Hybridization Properties of RNA Containing 8-Chloroadenosine. Nucleosides, Nucleotides and Nucleic Acids 2002, 21, 599-617.
(307) Carlson, C. C.; Chinery, R.; Burnham, L. L.; Dransfield, D. T. 8-Cl-Adenosine- Induced Inhibition of Colorectal Cancer Growth In Vitro and In Vivo. Neoplasia 2000, 2, 441-448.
(308) Taylor, C. W.; Yeoman, L. C. Inhibition of Colon Tumor Cell Growth by 8- Chloro-cAMP is Dependent Upon its Conversion to 8-Chloro-adenosine. Anti- Cancer Drugs 1992, 3, 485-491.
(309) Chen, L. S.; Bahr, M. H.; Sheppard, T. L. Effects of 8-Chlorodeoxyadenosine on DNA Synthesis by the Klenow Fragment of DNA Polymerase I. Bioorganic & Medicinal Chemistry Letters 2003, 13, 1509-1512.
(310) Lauria, F.; Rondelli, D.; Zinzani, P. L.; Bocchia, M.; Marotta, G. et al. Long- lasting Complete Remission in Patients with Hairy Cell Leukemia Treated with 2- CdA: a 5-year Survey. Leukemia 1997, 11, 629-632.
(311) Saven, A.; Carrera, C. J.; Carson, D. A.; Beutler, E.; Piro, L. D. 2- Chlorodeoxyadenosine: An Active Agent in the Treatement of Cutaneous T-Cell Lymphoma. Blood 1992, 80, 587-592.
(312) Rousseau, R. J.; Robins, R. K. The Synthesis of Various Chloroimidazo[4,5- c]pyridines and Related Derivatives. Journal of Heterocyclic Chemistry 1965, 2, 196-201.
(313) Cosstick, R.; Li, X.; Tuli, D. K.; Williams, D. M.; Connolly, B. A. et al. Molecular Recognition in the Minor Groove of the DNA Helix. Studies on the Synthesis of Oligonucleotides and Polynucleotides Containing 3-Deaza-2'- Deoxyadenosine. Nucleic Acids Research 1990, 18, 4771-4778.
(314) Merlic, C. A.; Motamed, S.; Quinn, B. Structure Determination and Synthesis of Fluoro Nissl Green: An RNA-Binding Fluorochrome. Journal of Organic Chemistry 1995, 60, 3365-3369.
(315) Ryu, E. K.; MacCoss, M. New Procedure for the Chlorination of Pyrimidine and Purine Nucleosides. Journal of Organic Chemistry 1981, 46, 2819-2823.
293 (316) Hayakawa, H.; Tanaka, H.; Haraguchi, K.; Mayumi, M.; Nakajima, M. Preparation of 8-Chloropurine Nucleosides Through the Reaction Between Their C-8 Lithiated Species and p-Toluene Sulfonyl Chloride. Nucleosides Nucleotides 1988, 7, 121-128.
(317) Ikehara, M.; Maruyama, T.; Miki, H.; Takatsuka, Y. Studies of Nucleosides and Nucleotides. LXXXV. Purine Cyclonucleosides. (35). Synthesis of Purine Nucleosides Having 2'-Azido and 2'-Amino Functions by Cleavage of Purine Cyclonucleosides. Chemical & Pharmaceutical Bulletin 1977, 25, 754-760.
(318) von Tilburg, E. W.; von Frijtag Drabbe Kunzel, J.; de Grotte, M.; Ijzerman, A. P. 2,5'-Disubstituted adenosine derivatives: Evaluation of selectivity and efficacy for the adenosine A1, A2A and A3 receptors. J. Med. Chem. 2002, 45, 420-429.
(319) Seyama, F.; Akahori, K.; Sakata, Y.; Misumi, S.; Aida, M. et al. Synthesis and properties of purinophanes: Relationship between the magnitude of hypochromism and stacking geometry of purine rings. J. Am.Chem. Soc. 1988, 110, 2192-2201.
(320) Lister, J. H. The Purines. Supplement 1; Wiley-Interscience: New York, 1996; 91- 116.
294 Vita
Peter Ivo O’Daniel, youngest son of Philip Scott O’Daniel Sr. and Mary Leona
O’Daniel, was born in Louisville, Kentucky on July 12, 1970. He graduated from
Concord High School in Concord, North Carolina, USA in 1988. After a short time at the
University of North Carolina at Charlotte, he left school and worked as a warehouse manager for the chemical company Auto-Chlor System. He returned to the University of
North Carolina at Charlotte in January 1996 and received his BS degree in Chemistry in
December 1998. Peter then entered graduate school in the school of chemistry and
Biochemistry at the Georgia Institute of Technology in August 1999. In November of that same year, he joined the Seley research group under the direction of Professor Katherine
L. Seley. In August 2003, the Seley group relocated to the University of Maryland,
Baltimore County, Baltimore, Maryland, USA where they reside today.
295