University of California, San Diego

RNA Affinity and Specificity of ModifiedAminoglycosides, Metal Complexes, and IntercalatingAgents That Target the HIV-1 Rev Response Element

A dissertation submitted in partial satisfaction of the

requirements for the degree

Doctor of Philosophy

in Chemistry

by

Nathan W. Luedtke

Committee in charge:

Professor Yitzhak Tor, ChairProfessor Murray GoodmanProfessor Simpson JosephProfessor Roger TsienProfessor Flossie Wong-Staal

2003

ii

Copyright

by

Nathan W. Luedtke, 2003All rights reserved

iii

The dissertation of Nathan W. Luedtke is approved,and is acceptable in quality and form

for publication.

University of California, San Diego

2003

iv

Dedicated to Ross Cortese, Charles Liu, and Heidi Thomsen

Table of Contents

Signature Page.................................................................................................iii

Dedication........................................................................................................ iv

Table of Contents ............................................................................................. v

List of Figures ..................................................................................................vii

List of Tables ..................................................................................................xiii

List of Abbreviations ........................................................................................xv

Acknowledgments...........................................................................................xvi

Vita & Publications......................................................................................... xvii

Abstract ...........................................................................................................xx

1.0 Introduction ............................................................................................ 1

1.1 Translation ..................................................................................... 2

1.2 RNA Viruses .................................................................................. 3

1.3 Small Molecules That Modulate RNA Activity ................................ 4

1.4 Magnesium (II) ............................................................................... 5

1.5 Aminoglycosides............................................................................ 7

1.6 Ligand Specificity ......................................................................... 10

1.7 Goals ........................................................................................... 11

2.0 Background: The HIV Lifecycle ............................................................. 13

2.1 The Rev-RRE Interaction and HIV's Late Replication Phase...... 14

2.2 Minimized RRE Constructs ......................................................... 18

2.3 Other RNA Constructs ................................................................ 19

v

vi

3.0 Methods for Measuring RNA-Ligand Affinity and Specificity ................. 24

3.1 Fluorescence Polarization Anisotropy......................................... 24

3.2 Solid-Phase Peptide Displacement Assay .................................. 31

3.3 Ethidium Bromide Displacement................................................. 40

3.4 Direct Observation of Small Molecule-Nucleic AcidAssociation: Monitoring Photophysical Changesof the Small Molecule.................................................................. 42

3.5 Direct Observation of Small Molecule-Nucleic AcidAssociation: Monitoring Photophysical Changesof the Nucleic Acid ...................................................................... 45

3.6 Analysis of Binding Data ............................................................. 53

4.0 New RRE Ligands: [Ru(bpy)2Eilatin]2+................................................... 60

5.0 Ethidium Bromide: Biological Activities and Nucleic Acid Binding ......... 86

5.1 Electronic Properties of Ethidium and its Derivatives.................. 90

5.2 Synthesis of Ethidium Bromide Derivatives .............................. 118

5.3 Acid-Base Properties of Ethidium and itsGuanidino Derivatives............................................................... 124

5.4 Preliminary Evaluation of the Nucleic Acid Affinity andSpecificity of Ethidium Derivatives ............................................ 127

6.0 Aminoglycoside-Based Ligands .......................................................... 135

6.1 Neomycin-Acridine Conjugates................................................. 136

6.2 Other Aminoglycoside-Acridine Conjugates.............................. 145

6.3 Dimeric Aminoglycosides.......................................................... 147

6.4 Guanidinoglycosides................................................................. 154

6.5 Acridine-Containing Guanidinoglycosides................................. 165

6.6 Platinum-Containing Aminoglycosides andGuanidinoglycosides................................................................. 167

vii

6.7 BODIPY- and Fluorescein-Linked Glycosidesas RNA Probes ......................................................................... 175

6.8 Cellular Uptakes of Aminoglycosides andGuanidinoglycosides ................................................................ 179

6.9 Conclusions .............................................................................. 188

E 1.0 Compound sources, references, and all known IC50

values for HIV-1 Inhibition......................................................... 189

E 2.0 Experimental procedures for RNA synthesisand quantification ..................................................................... 193

E 2.1 Experimental procedures for Rev peptide synthesis................. 196

E 3.0 Experimental procedures and conditions for binding assays.... 201

E 4.0 Raw spectal data for Λ- and ∆-[Ru(bpy)2Eilatin]2+ upontitration of nucleic acids............................................................. 213

E 5.0 Synthesis and Characterization of Ethidium Derivatives .......... 217

E 6.0 Synthesis and Characterization of AminoglycosideDerivatives................................................................................ 247

References ................................................................................................... 294

List of Figures

Figure 1.0 Molecular recognition of RNA and the centraldogma of biology........................................................................ 1

Figure 1.1 Deactivation of mRNA translation by small-molecule binding .... 5

Figure 1.2 Representative aminoglycosides ............................................... 8

Figure 1.3 Summary of HH16 inhibition by deoxy-tobramycin derivatives .. 9

Figure 2.0 The HIV-1 lifecycle................................................................... 13

viii

Figure 2.1 HIV's late replication phase and the Rev-RRE interaction.. ..... 14

Figure 2.2 A domain map of the Rev protein and the secondarystructure of the 351 ntd. RRE................................................... 15

Figure 2.3 RRE sequence alignments and the Rev binding site ............... 17

Figure 2.4 Minimized RRE constructs ....................................................... 18

Figure 2.5 RRE JW mutants "3I RREJW" and "dRREJW" ........................ 19

Figure 2.6 Secondary structures of "other" RNAs used in these studies... 20

Figure 3.0 Fluorescence polarization anisotropy....................................... 25

Figure 3.1 Relationships between the concentration of RevFl and theemission intensity and anisotropy of the sample........................ 25

Figure 3.2 Binding of RevFl to three different RRE constructs.................. 27

Figure 3.3 Linear relationship between Gibbs free energy and thesize of each RRE construct...................................................... 28

Figure 3.4 Fluorescence anisotropy displacement experiments ............... 30

Figure 3.5 Assembly of the solid-phase immobilized Rev-RRE complex.. 32

Figure 3.6 Measuring the relative RRE affinity and specificity of aligand using the solid-phase assay .......................................... 34

Figure 3.7 Known small-molecule inhibitors of Rev-RRE binding ............. 36

Figure 3.8 Raw data from solid-phase displacement experiments............ 37

Figure 3.9 Native gel-shift electrophoresis proves the efficacyof the phase assay................................................................... 38

Figure 3.10 Octahedral metal complexes evaluated for RRE affinityand specificity using the solid-phase assay ............................. 39

Figure 3.11 Structures of two classic intercalating agents thatbind the RRE............................................................................ 40

Figure 3.12 Raw emission data from ethidium displacement experiments.. 41

ix

Figure 3.13 Binding of eilatin to the RRE .................................................... 44

Figure 3.14 Binding of ethidium bromide to the RRE .................................. 44

Figure 3.15 Examples of thermal denaturation experiments ....................... 46

Figure 3.16 Adenosine-uracil and 2AP-uracil base pairs ............................ 47

Figure 3.17 Enzyme-substrate association and cleavage of the HH16....... 49

Figure 3.18 Enzyme-substrate association of the HH16 ribozymeat three different ionic strengths............................................... 49

Figure 3.19 Real-time cleavage of the HH16 as evidenced from theincrease in fluorescence emission of 2AP ............................... 50

Figure 3.20 Binding of neomycin B to the fluorescent HH16 complex ........ 51

Figure 3.21 Three hypothetical binding isotherms for different ratiosof the observable's concentration versus affinity...................... 54

Figure 3.22 Three theortical binding isothers showing positive,negative, and non-cooperative binding .................................... 55

Figure 3.23 Scatchard plot for ethidium bromide binding to the RRE JW ... 57

Figure 4.0 Structures of eilatin and eilatin-containing metal complexes.... 60

Figure 4.1 CD spectra of eilatin-containing metal complexes 9 and 10 .... 61

Figure 4.2 Anti-HIV activities of eilatin-containing metal complexes ......... 62

Figure 4.3 Viscosity of changes of DNA solutions upon addition ofethidium bromide or eilatin-containing metal complexes.......... 67

Figure 4.4 UV-vis absorption spectra of ∆-[Ru(bpy)2Eilatin]2+(10)upon addition of DNA............................................................... 70

Figure 4.5 Binding isotherm for poly r(U) at a high concentration of 10 .... 74

Figure 4.6 Binding of poly r(U) at three concetrations of 10 ...................... 75

Figure 4.7 Thermal denaturation of a poly r(U) - 10 complex.................... 76

Figure 4.8 van't Hoff plot for the hermal denaturation ofa poly r(U) - 10 complex .......................................................... 77

x

Figure 4.9 Scatchard analysis for the binding of 10 to RRE JW,TAR31, and tRNAPhe ................................................................ 78

Figure 4.10 Changes in the emission specrtum of eilatin (14)upon addition of DNA............................................................... 80

Figure 4.11 Direct binding of 14 to the RRE versus its ability todisplace RevFl from the RRE................................................... 81

Figure 5.1 New ethidium derivatives and a summary oftheir spectral properties ........................................................... 89

Figure 5.2 1H NMR spectra of 3,8-diamino-6-phenylphenanthridineand ethidium chloride (12) ....................................................... 96

Figure 5.3 1H NMR spectra of ethidium chloride (12), and its cbzprotected derivatives 16, 17, and 18 ........................................ 98

Figure 5.4 13C DEPT NMR spectra of ethidium chloride (12) .................. 100

Figure 5.5 13C-1H HETCOR NMR spectrum of ethidium chloride (12) .... 101

Figure 5.6 HMBC NMR spectrum of ethidium chloride (12) .................... 102

Figure 5.7 13C NMR spectra of ethidium chloride (12), and its cbzprotected derivatives 16 and 17 ............................................. 103

Figure 5.8 Bond lengths for three recent crystal structures ofphenanthridine and phenanthridinium-containing molecules . 105

Figure 5.9 All possible Clar and Kekulé depictions of ethidium (12) ...... 106

Figure 5.10 A summary of phenanthridine's π-bond order ........................ 106

Figure 5.11 Bond lengths for three crystal structures of ethidium (12)...... 108

Figure 5.12 Bond lengths observed in the crystal structures of themono-cbz ethidium derivatives 16 and 17.............................. 111

Figure 5.13 Charge delocalization to the 3-amine of ethidium .................. 112

Figure 5.14 Resonance structures showing charge separation in 12........ 113

Figure 5.15 A summary of partial positive and negative charges in 12 ..... 115

xi

Figure 5.16 Protection of ethidium bromide with benzyl chloroformate..... 119

Figure 5.17 Synthesis of 3-guanidino ethidium chloride (19) ................... 120

Figure 5.18 Unsuccessful attempts at synthesizing 3-urea ethidium ........ 121

Figure 5.19 Typical synthetic route for urea-ethidium derivatives ............. 122

Figure 5.20 Examples of ethidium-urea-conjugates.................................. 123

Figure 5.21 Synthesis of 3,8-bis pyrole ethidium acetate (27) .................. 124

Figure 5.22 Changes in the maximum absorbance wavelength ofethidium (12) as a function of pH ........................................... 125

Figure 5.23 Changes in the maximum absorbance wavelength of theguanidino ethidium derivatives 19, 20, and 21as a function of pH. ................................................................ 125

Figure 5.24 A summary of the pKa values for 12, 19, 20, and 21 ............. 127

Figure 6.0 Secondary structure of the RRE 66 showing theneomycin B footprinting site and potentialhigh-affinity intercalation sites................................................ 136

Figure 6.1 Structures of three neomycin-acridine conjugates ................. 137

Figure 6.2 Gel-shift electrophoresis and footprinting site forneo-S-acridine and Rev-IA on the RRE 66 ............................ 138

Figure 6.3 Structures of acridine conjugates of tobramycin andkanamycin A (41 and 42) ....................................................... 146

Figure 6.4 Structures of the dimeric aminoglycosides43, 44, 45, 46, and 47 ............................................................ 149

Figure 6.5 Structures of the aminoglycosides 1 - 5 andof the guanidinoglycosides 48 - 52......................................... 155

Figure 6.6 The anti-HIV activities of 3, 5, 50, and 52 .............................. 161

Figure 6.7 Structures of the tobramycin-Arg conjugates 53 and 54 ........ 162

Figure 6.8 Structures of the kanamycin A - amino acid conjugates53 and 54 ............................................................................... 164

xii

Figure 6.9 Structures of guanidinylated tobra-N-acridine andneo-N-acridine compounds 58 - 61........................................ 166

Figure 6.10 Structures of the cisplatin-neomycin conjugates 62 and 63 ... 168

Figure 6.11 DPAGE of RRE JW showing multiple adducts withthe cisplatin-neomycin conjugates 62 and 63 ........................ 169

Figure 6.12 MALDI TOF MS of RRE JW, its mutants, andthe cisplatin-neomycin conjugate 62 ...................................... 171

Figure 6.13 MALDI TOF MS of enzymatic digest reactions containingthe RRE JW and the neomycin conjugates 62 and 63........... 172

Figure 6.14 A summary of the enzymatic stop sites on the RRE JWdue to cross-linking of 62 and 63 ........................................... 173

Figure 6.15 Native gel-shift electrophoresis showing displacementof Rev by either neomycin B or Neo-Pt.................................. 174

Figure 6.16 New fluorescent aminoglycosides based upon BODIPYor fluorescein conjugation ...................................................... 176

Figure 6.17 Examples of binding experiments using the BODIPY-containing guanidinoglycoside ............................................... 177

Figure 6.18 FACS histograms showing the cellular uptakes of BODIPY-containing aminoglycosides and guanidinoglycosides ........... 182

Figure 6.19 FACS histograms showing the cellular uptakes of BODIPY-containing guanidino-neomycin and a poly Arg peptide......... 184

Figure 6.20 Fluorescence emission spectroscopy of cells exposed tofluorescein-containing aminoglycosides,guanidinoglycosides, and the poly-Arg peptide...................... 185

Figure 6.21 Thiol-reactive neomycin and guanidino-neomycinderivatives as cellular transport substrates ............................ 187

Figure E2.0 UV-vis absorption spectrum of RRE66 beforeand after hydrolysis................................................................ 194

Figure E2.1 Sequences of modified Rev34-50 peptides............................... 196

xiii

Figure E2.2 Synthesis of N-terminus and C-terminus modifiedRev peptides.......................................................................... 197

Figure E3.1 Synthesis of biotinylated RRE66............................................ 204

Figure E5.0 ORTEP drawing of ethidium chloride (12).............................. 217

Figure E5.1 ORTEP drawing of 3-cbz ethidium chloride (16) .................... 219

Figure E5.2 ORTEP drawing of 8-cbz ethidium chloride (17) .................... 220

Figure E6.1 1H NMR spectrum of neo-neo before andafter dimerization (43) ............................................................ 259

List of Tables

Table 3.0 RevFl binding constants and free energies for tRNAPhe,CT DNA, TAR 31, and three RRE constructs........................... 27

Table 3.1 IC50 values for Rev-RRE inhibition by twofamilies of known Rev-RRE inhibitors ...................................... 35

Table 3.2 IC50 values for Rev-RRE inhibition by the eilatin-containing metal complexes 9 and 10 ...................................... 39

Table 4.1 IC50 values for HIV-1 inhibition by the eilatin-containingmetal complexes 9 and 10 and for RevFl displacementfrom RRE 66 and TAR31 by the Ru complexes 11 and 15 ...... 63

Table 4.2 IC50 values for ethidium displcement bb 9, 10, and 15using 8 different nucleic acids.................................................. 65

Table 4.3 C50 values for the binding of 9, 10, 12, and 14 to17 different polymeric nucleic acids ........................................ 72

Table 4.4 Summary of binding stoichiometries and affinities between9 and 10 for the RRE JW, tRNAPhe, and TAR 31 ..................... 78

Table 4.5 Estimated Kd values for the binding of 9, 10, 12, and 14 to17 different polymeric nucleic acids ........................................ 82

Table 5.1 Summary of 13C assignments and chemical shifts of

xiv

ethdium (12) and 3,8-bis cbz ethidium chloride (18) ............. 117

Table 5.2 Summary of IC50 values for RevFl displacement and C50

values for CT DNA binding for 14 differentethidium derivatives ............................................................... 129

Table 5.3 Summary of the changes in spectral properties ofethidium bromide derivatives upon binding CT DNA.............. 131

Table 5.4 C50 values for binding of the deoxyneamine-ethidiumconjugate 37 to six different nucleic acids.............................. 133

Table 6.0 IC50 and Ki values for Rev-IA, neo-S-acridine, neo-neo,and neomycin B for binding to the RRE 66 ............................ 139

Table 6.1 IC50 values and RRE selectivities of aminoglycosides,aminoglycoside-acridine conjugates, aminoglycosidedimers, and Rev-IA, according to the solid-phase assay ...... 142

Table 6.2 RevFl IC50 values, RRE 66 Ki values, and ethidiumdisplacement IC50 values for aminoglycosdies,aminoglycoside-acridine conjugates and aminoglycosidedimers using CT DNA and poly r(A) - r(U) ............................. 144

Table 6.3 RevFl IC50 values for aminoglycosides andguanidinoglycosides at 10 nM and 100 nM of RRE 66 .......... 156

Table 6.4 IC50 values and RRE selectivities of the aminoglycosidesand guanidinoglycosides according to the solid-phase assay........................................................................... 157

Table 6.5 IC50 values and approximate RRE 66 Ki values foraminoglycoside - amino acid conjugates 53 - 57 ................... 163

Table 6.6 Preliminary RRE 66 affinitites of the guanidinylatedtobra-N-acridine and neo-N-acridine compounds 58 - 61 ...... 166

Table 6.7 C50 values for the neomycin-BODIPY conjugates 67 and68 for binding to the RRE JW and two RRE JW mutants....... 177

Table 6.8 IC50 values for the displacement of the neomycin-BODIPYconjugates 67 and 68 from the RRE JW................................ 179

Table 6.9 Qualitative assessment of the cellular uptakes of acridine-containing aminoglycosides and guanidinoglycosides ........... 180

xv

Table 6.10 Quantum yields of acridine- and BODIPY-containingaminoglycosides and guanidinoglycosides ............................ 181

Table 6.11 Summary of the mean fluorescence intensities of individualcells after exposure to BODIPY-containingaminoglycosides and guanidinoglycosides ............................ 183

Table E1.0 Compound source index ........................................................ 189

Table E1.1 Summary of all IC50 values for HIV-1 inhibition ..................... 192

Table E2.0 Extinction coefficients used for RNA and DNA quantification. 195

Table E5.1 Crystallographic parameters for ethidium chloride ................. 218

Table E5.2 Crystallographic parameters for 3-cbz ethidium chloride ....... 219

Table E5.3 Crystallographic parameters for 8-cbz ethidium chloride ....... 221

List of Abbreviations

Ar argon gasBoc tert-butyloxycarbonylBODIPY N-(4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene)bpy 2,2'-bipyridinecbz benzyl chloroformateCD circular dichroismDMAP 4-(dimethylamino)pyridineDMF N,N-dimethylformamideDMSO dimethylsulfoxideDNA deoxyribonucleic acidDPAGE denaturing polyacrylamine gel electrophoresisESI MS electrospray ionization mass spectrumFAB MS fast atom bombardment mass spectrumHPLC high performance liquid chromatographyMALDI TOF matrix-assisted laser desorption/ionization time-of-flightNMR nuclear magnetic resonancentd nucleotidephen 1,10-phenanthrolineRNA ribonucleic acidTEA triethylamineTFA trifluoroacetic acidTm thermal melting temperatureUV-Vis ultraviolet-visible

xvi

Acknowledgments

Many people have made essential contributions to this work. I would like to thank

my mentor Dr. Yitzhak Tor for his constant support, wisdom, and for planting the

seeds from which this work has grown. I am also very grateful to the many

talented chemists who have contributed compounds and/or data to this thesis,

especially Mr. Qi (Charles) Liu, Ms. Judy S. Hwang, Dr. Sarah R. Kirk, Dr. Hai

Wang, Dr. Tracy J. Baker, Ms. Phoebe C. Glazer, Mr. Peter Carmichael, and Dr.

Jürgen Boer. I thank all of my science teachers, mentors, and other “scientific

heros”, especially Mr. S. Koepp, Dr. P. L. Fischhaber, Dr. P. B. Hopkins, Dr. E.

O. Wilson, Dr. J. E. Kyte, Dr. J. S. Siegel, and Dr. M. Goodman. I also want to

thank all of my family for giving me the love, freedom, and means to pursue my

own interests, especially my parents, William D. Luedtke, Lorna L. Luedtke, my

brother David P. Luedtke, my grandmother Louise Rigelman, and my uncles and

aunts, especially Ross Cortese and Ellie Cortese. Last, and certainly not least,

thank you Heidi W. Thomsen for all of your support, encouragement, and

understanding during this long adventure.

xvii

Vita

Nathan W. Luedtke

Education

2003 University of California, San DiegoDoctor of Philosophy, Chemistry

1999 University of California, San DiegoMaster of Science, Chemistry

1997 University of WashingtonBachelor of Science, Chemistry

Awards and Fellowships

The Traylor Award for Outstanding Graduate Research, UCSD, 2001

The Award for Outstanding Teaching Assistant, UCSD, 2000

Award for Excellent Poster Presentation, Gordon Research Conference forBioorganic Chemistry, July 2001

2nd Place Award for Excellent Poster Presentation, Maria Goeppert-MayerInterdisciplinary Symposium, 2002

The Universitywide AIDS Research Program, Doctoral Fellowship, 2001

Research Publications

N. W. Luedtke, Y. Tor, A Novel Solid-Phase Assembly for Identifying Potent andSelective RNA Ligands. Angew. Chem. Intl. Ed. 2000, 39, 1788-1790.

N. W. Luedtke, T. J. Baker, M. Goodman, Y. Tor, Guanidinoglycosides: A NovelFamily of RNA Ligands. J. Am. Chem. Soc. 2000, 122, 12035-12036.

N. W. Luedtke, J. S. Hwang, E. C. Glazer, D. Gut, M. Kol, Y. Tor, Eilatin Ru(II)Complexes Display Anti-HIV Activity and Enantiomeric Diversity in the Binding ofRNA. Chembiochem, 2002, 3, 766-771.

xviii

Research Publications (cont.)

N. W. Luedtke, Y. Tor, Targeting HIV RNA with Small Molecules. In: DNA andRNA Binders Vol.1, M. Demeunynck, C. Bailly, W. D. Wilson (Eds.), 2002, Wiley-VCH, 18-40.

N. W. Luedtke, Y. Tor, Fluorescence-Based Methods for Evaluating the RNAAffinity and Specificity of HIV-1 Rev-RRE Inhibitors. Biopolymers, In Press

N. W. Luedtke, Q. Liu, Y. Tor, The Electronic Structure of Ethidium BromideReveals Charge Separation in an Organic Cation. In preparation

N. W. Luedtke, Y. Tor, Synthesis, Nucleic Acid Binding, and PhotophysicalProperties of Phenanthridine Derivatives Based Upon Ethidium Bromide. Inpreparation

N. W. Luedtke, J. S. Hwang, E. C. Glazer, D. Gut, M. Kol, Y. Tor, Nucleic AcidSpecificity and Binding Mode of Specificities of Eilatin Ru(II) Complexes. Inpreparation

N. W. Luedtke, P. Carmichael, Y. Tor, Impressive Cellular Translocation ofGuanidine-Modified Natural Products, In preparation

N. W. Luedtke, Q. Liu, Y. Tor, Affinity and Specificity of Aminoglycoside Dimersand Acridine Conjugates That Bind the HIV-1 Rev Response Element . Inpreparation

T. J. Baker, N. W. Luedtke, Y. Tor, M. Goodman, Synthesis and Anti-HIV Activityof Guanidinoglycosides. J. Org. Chem., 2000, 65, 9054-9058.

S. R. Kirk, N. W. Luedtke, Y. Tor, Neomycin-Acridine Conjugate: A PotentInhibitor of Rev-RRE Binding, J. Am. Chem. Soc. 2000, 122, 980-981.

J. Boer, N. W. Luedtke, Y. Tor, RNA-Selective Covalent Modification by Neo-platin (A Neomycin-Cisplatin Conjugate), In preparation.

Q. Liu, N. W. Luedtke, Y. Tor, A Simple Conversion of Amines intoMonosubstituted Ureas in Organic and Aqueous Solvents. Tetrahedron Lett.2001, 42, 1445-1447.

S. R. Kirk, N. W. Luedtke, Y. Tor, 2-Aminopurine as a Real-Time Probe ofEnzymatic Cleavage and Inhibition of Hammerhead Ribozymes. Bioorg. Med.Chem. 2001, 9, 2295-2301.

xix

Research Publications (cont.)

J. T. Millard, N. W. Luedtke, R. J. Spencer, The 5'-GNC Preference for MustardCross-Linking is Preserved in a Restriction Fragment. Anti-Cancer Drug Des.,1996, 11, 485-492.

A. Friedler, D. Friedler, N. W. Luedtke, Y. Tor, A. Loyter, C. Gilon, Developmentof a Functional Backbone Cyclic Mimetic of the HIV-1 Tat Arginine-rich Motife. J.Biol. Chem. 2000, 275, 23783-23789.

K. Fineberg, T. Finberg, A. Graessmann, N. W. Luedtke, Y. Tor, R. Lixin, D. A.Jans, A. Loyter, Inhibition of Nuclear Import Mediated by the Rev-Arginine RichMotif by RNA Molecules, Biochemistry, 2003, 42, 2625-2633.

xx

Abstract

RNA Affinity and Specificity of Modified Aminoglycosides, Metal Complexes, and

Intercalating Agents That Target the HIV-1 Rev Response Element

by

Nathan W. Luedtke

Doctor of Philosophy in Chemistry

University of California, San Diego, 2003

Professor Yitzhak Tor, Chair

RNA plays a pivotal role in the replication of all organisms, including viral and

bacterial pathogens. The development of small molecules that can selectively

interfere with undesired RNA activity is a promising new direction for drug design.

Systematic studies of the binding interactions between small molecules and RNA

are essential for deciphering the parameters that govern RNA recognition. The

synthesis of new ligands with high RNA affinity is relatively easy compared to

discovering new ligands with high specificity for the desired RNA target. The

RRE affinity and specificity of every small molecule is not necessarily

proportional to its anti-HIV activity. Issues related to cellular uptake, localization,

and the non-specific binding of other cellular components can dramatically affect

the biological activities of small molecules.

1

1.0 Introduction

The central dogma of biology states that from DNA, RNA is transcribed to serve

as an information messenger that is translated into proteins.1 Messenger RNA

(mRNA) accounts, however, for only a small fraction of the RNA present within a

cell. Far from being a passive carrier of genetic code, RNA exhibits vast

structural and functional diversity and is intimately involved in a wide range of

biological activities, including information storage and chemical catalysis.

Figure 1.0: Molecular recognition of RNA often precedes catalytic events that areessential to a wide range of cellular activities including: (a) initiation of DNAreplication,2 (b) extension of the telomeric regions of chromosomes,3 (c) splicingof pre-mRNA,4 and iron chelation.5 In addition, RNA serves as the primarygenome of most pathogenic viruses.6

A more “modern” interpretation of the central dogma of biology is that RNA has

structural and functional characteristics that are, in many ways, similar to both

2

DNA and proteins. RNA is, therefore, an intermediate between DNA and proteins

in more than one respect.

1.1 Translation

Gene expression relies upon an interplay between recognition events and

catalytic activities that are mediated by RNA-protein complexes. Ribosomal RNA

(rRNA) accounts for the vast majority of total cellular RNA (80%) and provides

both the molecular scaffold and enzymatic activities needed for protein

translation.7 The key step of translation occurs in the ribosome’s A-site, where

codon-anticodon recognition decodes mRNA. Upon a correct codon-anticodon

match between mRNA and the anticodon loop of tRNA, the ribosome’s peptidyl

transferase activity catalyzes the formation of a new peptide bond between the

amino acid-charged tRNA in the A-site and the growing protein chain on the

tRNA in the P-site.7 Studies have shown that prokaryotic ribosomes that are

stripped of protein are still capable of limited peptidyl transferase activity.8 In

accordance with this result, recent crystal structures show that the peptidyl

transferase active site is composed entirely of rRNA.9 A single, unusually basic

adenosine may be the key player in the mechanism of peptidyl transfer.10

Transfer RNAs, at 15% of total cellular RNA, are the most common type of

“soluble” RNA (i.e. lacking any associated proteins). The binding of tRNA to the

ribosomal A-site is mediated by extensive RNA-RNA interactions (including

rRNA-tRNA, and mRNA-tRNA binding). Through these, and other important

3

interactions, the ribosome amplifies the relatively small energetic differences

between cognate and non-cognate codon-anticodon pairing to achieve an

astounding 99.9% accuracy in its translation of mRNA.7,11

The transport, translation efficiency, and stability of individual messenger RNAs

is controlled by numerous protein-RNA, ribonucleoprotein-RNA, and RNA-RNA

interactions.12 Upon transcription from DNA, ribonucleoprotein complexes called

splisosomes excise the introns from pre-mRNA and from other heterogeneous

RNAs (Figure 1.0). Some organisms are capable of intron excision (splicing)

without protein assistance, and have provided the first examples of RNA

enzymes (or ribozymes).13 The translation efficiency of individual mRNAs is

regulated at many levels, including the binding of the 5' and 3' untranslated

regions (UTRs) of the mRNA by proteins,14 microRNAs,15 and by small

molecules.16

1.2 RNA Viruses

Viral epidemics have accounted for more human deaths than all known wars and

famine combined. About 65% of the known families of viruses use RNA for a

primary genome and cause many modern-day plagues including AIDS, cancer,

hepatitis, smallpox, ebola, and influenza.6 Most viruses are, however, benign.

Interestingly, approximately 42% of the human genome is composed of

transposable elements that multiply by reverse transcription, using an RNA

intermediate similar to that of a retrovirus.17 In general, reverse transcription is a

4

highly error-prone process allowing viral elements to evolve rapidly under

selective pressures (such as anti-viral drugs). An additional 8% of the human

genome is composed of repetitive genomic elements known as “retrovirus-like

elements”.17 Their structures very closely resemble those of retroviruses, carrying

the open reading frames common to all retroviruses (Gag, Pol, Env), flanked by

5' and 3' long terminal repeats. Overall, the human genome is composed of

approximately 50% self-repeating parasitic sequences. Compare this with the

unique (non-repeated) genes, representing only ~5% of the human genome!17

1.3 Small Molecules That Modulate RNA Activity

The ability of RNA to facilitate the essential biochemical activities needed for

information storage, signal transduction, replication, and enzymatic catalysis has

distinguished it as a candidate for being the central biomolecule in a prebiotic

world.18 If such an “RNA world” ever did exist, then small molecule-RNA

interactions certainly played a key role in the regulation of RNA replication,

processing, as well as other enzymatic and regulatory activities.19

The conceptual proof demonstrating the ability of small organic molecules to

regulate gene expression was first revealed in the context of an artificial gene

construct.16a An RNA aptamer (see endnote [20]) located in the 5' untranslated

region (UTR) of an mRNA, was shown to inactivate the translation of a down-

stream reporter gene upon binding to its cognate small molecule (Figure 1.1).

5

The mechanism proposed for the small molecule-dependent translation

inactivation involves a structural rearrangement of the 5'-UTR into a rigid

complex that cannot be scanned by the ribosomal pre-initiation machinery.

Recent studies have shown that natural systems use small molecule-RNA

binding (accompanied by RNA structural rearrangements) to directly modulate

mRNA translational efficiencies.16 b,c

Figure 1.1: The mature mRNA of an artificial gene construct is actively translatedin the absence of small-molecule binding (Top). Upon binding the 5'-UTR by itscognate small molecule, the translation of the gene is deactivated (Bottom).16a

Recent studies indicate that similar mRNA-small molecule control mechanismsoccur in vivo and appear, therefore, to represent a normal aspect ofmetabolism16b,c

1.4 Magnesium (II)

Much like proteins, the primary sequence of an RNA directs its folding into a

unique 3-D structure.21 Correct RNA folding, however, typically relies upon the

binding of divalent metal ions (especially Mg2+). The Mg2+ induced folding of the

6

Tertrahymena thermophila group I intron has become an important paradigm for

RNA folding.22 In the absence of Mg2+, it occupies an ensemble of highly dynamic

secondary structures that are dominated by duplex regions interrupted by internal

bulges and stem loops. The group 1 intron secondary structure can be predicted

from its nucleotide sequence using base-pairing and nearest neighbor rules.23

Upon Mg2+ binding it collapses into a more rigid, enzymatically active, tertiary

structure with fewer conformations available. In at least one region of the group 1

intron, Mg2+ binding induces a rearrangement of the RNA secondary structure

itself.24 These cation-mediated “higher-order” folding interactions remain a major

obstacle in the prediction of a 3-dimensional RNA structure given only its primary

sequence.

Mg2+ exhibits a low to moderate affinity to many unrelated RNAs. Mg2+ binding

affinities (Kd) range from 0.01 mM through 10 mM in the presence of 0.1 – 0.2 M

of monovalent ions.25 Given the 3-dimensional structure of an RNA, an

electrostatic contour map can be calculated, allowing for the theoretical

prediction of Mg2+ binding sites.26 The accuracy of such predictions is

complicated by issues related to induced fit and by the limited understanding of

the characteristics of the cations themselves. Crystal structures of tRNAPhe, for

example, indicate that different cations bind at different RNA sites, depending

upon the identity of the ion.27 Few of the metal cation binding sites overlap with

one another (as would be predicted by electrostatic contour mapping).27

Tremendous diversity in the position, size, and affinity provided by RNA

7

coordination sites, suggests that metal ion-RNA complexes may have exhibited

diverse catalytic activities in a prebiotic “RNA world”.

RNA-cation binding interactions are essential for the proper folding and catalytic

function of RNA.22 There are some cases, however, where small molecules other

than metal cations can be used to facilitate the folding and enzymatic activity of

RNA. Linear polyamines (like spermine) and aminoglycosides (Figure 1.2) can

displace Mg2+ from RNA, and have been shown to directly facilitate the

enzymatic activities of the hairpin and hammerhead ribozymes even in the

absence of divalent metal ions.28 Structurally complex and semi-rigid polycations

may have once served as RNA scaffolds, similar to the ribosomal proteins of

today.

1.5 Aminoglycosides

Aminoglycoside antibiotics are a diverse family of natural products that interfere

with prokaryotic protein biosynthesis (Figure 1.2). Their ability to non-specifically

bind to RNA through electrostatic interactions was described over 20 years

ago.29 The aminoglycosides are also capable, however, of site-specific

recognition of prokaryotic rRNA. Early footprinting experiments indicated that

aminoglycosides bind to discrete locations within the ribosome.30 Later

experiments showed that aminoglycosides increase the affinity of tRNA to the

30S ribosomal A-site,31 thus providing an attractive mechanism to explain their

ability to selectively decrease the fidelity of prokaryotic translation.32 A recent

8

crystal structure of three aminoglycosides (streptomycin, paromomycin, and

spectinomycin) bound to the Thermus thermophilus 30S ribosomal subunit

confirms the location of the aminoglycoside binding sites and provides a high-

resolution picture of how RNA-small molecule recognition occurs within a

ribonucleoprotein complex.33 This type of structural information will prove

indispensable for the structure-based design of aminoglycoside derivatives that

have an improved “fit” within their ribosomal binding pockets. Structural

information by itself cannot, however, answer basic questions related to the

energetics involved in the binding of small molecules to RNA. Equilibrium binding

constants must be measured in order to establish the actual energetic values

associated with RNA-small molecule recognition.

OHO

HO

NH2

H2N

O

NH2

OH

OH2N

O HN

OH

OH

HO

OHO

H2N

NH2

OHH2NHO

NH2

O

HO

OO

NH2

HOO

HOO

HO

H2N

OMe

HO

NH2

H2N

O

NH2O

OH

MeHN

OH2N

HO

HNHO OH

NHCH3HOHO

HOO

NH

OO

OHOH3C

H2NNH

HN

NH2OH

O H

Amikacin

Sisomycin

Neomycin B

StreptomycinO

H2N

HO

O

OH

O

HOHO

H2NNH2

NH2

OH2N

HO

OH

O

HNO

OH Me

Apramycin

Figure 1.2: Representative aminoglycosides. Five out of the six amino groups ofneomycin B have pKa values over 7.0, giving it a highly positive charge underphysiological conditions.34

9

Despite the structural details provided by aminoglycoside-RNA complexes, the

energetic contributions made by the pendant hydroxyl groups of the

aminoglycosides remain unclear. Hydrogen bonding between these groups and

RNA are apparent in some structures,33 but the energetic contributions made by

hydrogen bonding in aqueous media is still debated.35 In an attempt to determine

their role in RNA affinity, the hydroxyl groups of tobramycin were systematically

removed (Figure 1.3). RNA binding was then tested by measuring the HH16

ribozyme inhibitory activity of each tobramycin derivative.36

Figure 1.3: Summary of hammerhead inhibition by deoxy-tobramycinderivatives.36 A lower relative rate suggests better RNA binding.

Interestingly, the removal of hydroxyls at positions R2, R3, and R4 (Figure 1.3)

lead to aminoglycosides with better RNA cleavage inhibition. The current

explanation for this “unexpected” result is that certain hydroxyl groups decrease

the basicity of neighboring amines. Therefore, removing hydroxyls typically

increases the overall positive charge, and hence the RNA affinity of the deoxy-

O

NH2O

OR2

H2N

H2NO

HO

H2NNH2

R3

R4

R1

OH

OH

2''-Deoxytobramycin

OH

Aminoglycoside

OH

R3

OH

R2

H OH6''-Deoxytobramycin

Tobramycin

H

OH

OH

R1

4'-Deoxytobramycin OH

4''-Deoxytobramycin H

OH

R4

OH

H

0.17

OH

OH

OH OH

OH

Relative RibozymeCleavage Rate

1.0

0.33

1.4

0.33

10

derivatives. These de-hydroxylated tobramycin derivatives have not yet been

tested for the binding of other RNAs, so the roles of the hydroxyls in RNA

specificity remain unclear.

1.6 Ligand Specificity

For the purposes of this thesis, specificity will be defined as the binding affinity

(Keq) of a small molecule to a particular RNA site, divided by its average affinity to

“all” other potential binding sites:

specificity = Keq(interaction of interest)average Keq(other sites of interaction)

Specificity is proportional to occupancy of the “desired” RNA site, versus the

occupancy of all other potential binding sites. For practical reasons, specificity is

a relative term, where the affinity between a small molecule and its RNA “target”

is weighted by its affinity to “other” nucleic acids. The reported specificity is,

therefore, always dependent on the selection of the competitor or “non-specific”

nucleic acids used for the comparison.

High specificity is a prerequisite for the effective modulation of RNA activity in

vivo. Since non-specific binding sites are typically present at much higher

concentrations as compared to the desired target, the bioavailability of a small

molecule may suffer even if it has a moderate affinity to “other” sites. The binding

11

of the small molecules to tRNA, rRNA, DNA, proteins, phospholipids, etc., may

also cause undesired biological “side effects” including toxicity and mutagenicity.

Aminoglycosides, for example, are not ideal antibiotics. Their promiscuous

binding of RNA and/or membrane components may be related to the multiple

therapeutic side effects and the low-moderate bacteriacidal potency exhibited by

these compounds.37,38 Aminoglycosides bind to and inhibit the function of a wide

range of unrelated RNAs with moderate activities (IC50 = 0.1 – 100 µM).19,28c,39

Aminoglycosdies, therefore, exhibit a low specificity for most of these RNA sites.

Aminoglycosides do, however, show excellent specificity for RNA over DNA

(Section 6.0). For this reason, we have used aminoglycosides as “scaffolds” for

the synthesis of new small molecules targeted towards specific RNA sites. These

derivatives are found to exhibit dramatically different RNA specificities and

altered biological activities when compared to their aminoglycoside precursors.

1.7 Goals

There are still no "rules" for the structure-based design of small molecules that

are targeted to a specific RNA tertiary fold. One obstacle is that there are still

very few examples of small molecules that bind to natural RNA structures with

high specificity. There are other potential reasons as well. For example, RNA is a

highly dynamic molecule known to occupy multiple conformations. The structural

details of an RNA do not typically entail the potential structural changes it can

12

adopt upon ligand binding. This adds additional complexity to the structure-based

design of RNA ligands.40

Despite recent progress in the understanding of how small molecules recognize

RNA,41 the following fundamental questions remain largely unanswered:

1. How do electrostatic interactions affect the RNA affinity and specificity ofaminoglycoside-based ligands?

2. How does one design small molecules that exhibit high affinity and highspecificity for a pre-determined RNA target?

To help answer these questions, we have addressed a number of goals:

1. Design and synthesize new small molecules that are targeted to a pre-determined RNA site.

2. Rapidly characterize the affinity and specificity of new RNA ligands usingfluorescence-based methodologies.

3. Conduct experiments in a systematic fashion so that trends in RNA-smallmolecule recognition can be identified.

To evaluate the “higher-order” biological impacts of RNA binding, we have

chosen the HIV-1 Rev-RRE interaction as our model system. This way, new RNA

ligands that show promising activities may eventually prove themselves as future

antiviral agents. Our work, along with the efforts by many other groups,

contributes to the growing body of knowledge that will aid in the future design,

synthesis, and application of small molecules directed to RNA.

13

2.0 Background: The HIV Lifecycle

HIV, the virus responsible for AIDS, relies upon the transcription and translation

machinery of its host cell.6 There are, however, a number of protein-RNA

interactions that are unique and essential to the HIV lifecycle (Figure 2.0).42

Inhibition of these key binding interactions may eventually provide a new class of

therapeutic compounds for the treatment of AIDS.

Figure 2.0: The HIV-1 lifecycle begins with the CD4-dependent invasion of thehost cell (a). The viral particle "tricks" a CD4+ host cell with the gp120 domain ofthe "Env" protein, which mimics a normal component of the human immunesystem (the major histocompatibility complex). Binding of CD4 to gp120 initiatesendocytosis and fusion of the viral and host membranes. Fusion is mediated bygp41 (the membrane-bound component of the Env protein). Upon fusion, theviral capsid is partially degraded and reverse transcriptase (Pol), transcribes the9 kilo-base single-stranded viral RNA genome into double-stranded DNA (b). Thedouble-stranded DNA copy of the HIV genetic code is then imported into thehost's nucleus, where it is permanently integrated into the host's genome (c). Atthis point, viral activity can remain dormant for years. Eventually, an RNA copy ofthe HIV DNA will be transcribed (d). Early in the replication cycle, this viral mRNAbecomes highly spliced and is translated into three regulatory proteins: Rev, Tat,and Nef (e). The Tat protein binds to its cognate RNA site termed TAR andfacilitates the transcription of full-length HIV RNA (d); since Tat itself is aneventual product of this transcription, Tat’s biosynthesis is in a positive feedbackloop, leading to a rapid accumulation of viral RNA as well as the Tat, Rev, andNef proteins. Once the concentration of Rev is sufficient, it polymerizes along theHIV RNA at a site termed the Rev response element (RRE), and initiates HIV’s“late replication” phase.

14

Figure 2.1: HIV’s “late” replication phase is initiated by the Rev-RRE interactionwhich prevents the splicing of the HIV RNA (a). From the unspliced and singlyspliced HIV RNA, the proteins needed for viroid construction are translated (b).Unspliced RNA is then packaged into outgoing viral particles to serve as theprimary genome (c).

2.1 The Rev-RRE interaction and HIV’s Late Replication Phase

The Rev response element (RRE) is an HIV-1 RNA structure essential to viral

replication.43 The Rev protein binds to the RRE and facilitates the export of HIV

RNA out of the host nucleus, while protecting it from the cell’s splicing machinery

(Figure 2.1). The importance of the "underspliced" HIV RNA is two-fold. First, it

serves as an open reading frame for the proteins that are essential to the

construction of new viral particles (Gag, Pol, and Env). Second, the unspliced

HIV RNA is packaged into the outgoing viroid, serving as its primary genome.

Successful inhibition of Rev-RRE binding, therefore, prevents the production of

15

new viral particles in two ways: the proteins essential for viroid construction are

never translated, and the future HIV genomic RNA becomes highly spliced.

Through the efforts of a number of research groups, many details regarding the

sequences, structures, and dynamics of the Rev-RRE interaction have been

revealed.44a-e The 116 amino acid Rev protein has at least three functional

domains, including an arginine-rich motif spanning amino acids #34-50 (Figure

2.2a). This positively charged domain binds to a single high-affinity site on the

RRE (bolded in Figure 2.2b) with high affinity (Kd = ~1 nM) and high specificity

(average Kd to a yeast tRNA mixture = ~1,000 nM).

Figure 2.2: (a) A domain-map of Rev protein is shown with the sequence of itsRRE binding domain (amino acids 34-50). This “arginine-rich motif” also servesas Rev’s nuclear localization signal, allowing Rev to shuttle continuously betweenthe cytoplasm and nucleus.45 The peptide RevFl, used for fluorescence-basedassays, is succinylated at its N-terminus, amidated at its C-terminus, andcontains a four alanine spacer to fluorescein (a). These modifications have beenshown to increase the helicity of this peptide, resulting in a greater RRE affinityand specificity.46 (b) The RRE, as defined by Mann,47 spans a total of 351 basesand contains a single high-affinity Rev binding site in the “Stem II” of the RRE(bolded bases between #97-162).44d

16

The RRE serves as the high-affinity Rev binding site and is part of the highly

conserved open reading frame of the gp41 “fusion domain” of the Env protein

(Figure 2.0).48 This dual function is likely responsible for the unusually low

mutation rate found in the RRE (Figure 2.3 A & B).49 The bases directly involved

in Rev binding are almost invariable, even for genetically diverse groups of HIV-1

isolates (Figure 2.3 A & B). The Rev binding site within the RRE is, therefore, a

highly attractive target for small molecule therapeutics, since small molecules

that are targeted to this site should possess consistent activities for diverse HIV-1

groups. In addition, the mutations that can lead to drug resistance should be

strongly impeded by the RRE’s dual function.

Following association of Rev with the high-affinity RRE “nucleation site” (bold

Figure 2.3 B), additional Rev molecules (approximately 10) can polymerize along

the length of the RRE in a step-wise fashion by utilization of both protein-protein

and protein-RNA interactions.47 Other groups have shown, however, that Rev

can first polymerize, then bind to the RRE in its polymeric form.44e Despite this

ambiguity, minimal mutations of the RRE within the high-affinity Rev binding site

(bold Figure 2.3 B) prevent polymerization of Rev along the entire length of

RRE.47 Without high-affinity Rev-RRE binding, the “late-phase” HIV proteins are

not expressed and the viral replication cycle is broken.44a For these reasons, we

have chosen the high-affinity Rev binding site on the RRE as our target for small

molecules (bold Figure 2.3 B).

17

Figure 2.3 A: Sequence homology for Stem II of the RRE from 5 representativeisolates from HIV-1 group M (SF2, HXB2, MAL, ELI, HIVU455), and 2 from groupO (HIVANT70, MVP5180). Sequence alignments were adopted from J-H Chen etal.50 The high affinity Rev binding site is shown in blue.

Figure 2.3 B: Secondary structure of the 66 nucleotide RRE Stem II fragment“RRE 66” from the HIV strain HXB3. The bases essential for binding the firstequivalent of Rev are shown in bold.43 A summary of homology alignment is

shown: (*) indicates a highly mutable position, (*) indicates conserved orinfrequent mutations, no symbol indicates no significant variation. The secondequivalent of Rev is proposed to bind, in vivo, to Stem IIA.51 It is interesting thatboth of the Rev binding sites within Stem II are essentially immutable.

U

GA

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

G

C

U

A

AC A

AUG

A

*

** *

****

*

*

**

*****

**

***

**

Stem IIA

Stem IIB

Stem IIC

GCACUAUGGGCGCAGCG UCAAUGACGCUGACGGUACAGGCCAGACAAUUAUUGUCUGGUAUAGUGCGCACUAUGGGCGCAGUG.UCAUUGACGCUGACGGUACAGGCCAGACAAUUAUUGUCUGGUAUAGUGC

GCACUAUGGGCGCAGCC.UCAAUGACGCUGACGGUACAGGCCAGACAAUUAUUGUCUGGUAUAGUGC

GCACGAUGGGCGCAGCG.UCACUAACGCUGACGGUACAGGCCAGACAGUUACUGUCUGGUAUAGUGCGCACGAUGGGCGCA.CGGUCAGUGACGCUGACGGUACAGGCCAGACAAUUAAUGUCUGGUAUAGUGCGCACAAUGGGCGCGGCG.UCAAUAACGCUGACGGUACAGGCCAGACAAUUAUUGUCUGGUAUAGUGCGCACUAUGGGCGCAGCG.GCAACAACGCUGGCGGUACAGACCCACACUUUGCUGAAGGGUAUAGUGC

HXB3SF2

HXB2

MALELIHIVU455HIVANT70MVP5180 GCACUAUGGGCGCAGCG.GCAACAGCGCUGACGGUACGGACCCACAGUGUACUGAAGGGUAUAGUGC

18

2.2 Minimized RRE Constructs

To facilitate biophysical analysis of the Rev-RRE interaction, numerous groups

have minimized the Rev response element into smaller, more manageable

fragments (figure 2.4).

Figure 2.4: Removal of stem loop IIA from the 66 ntd. RRE Stem II fragmentresults in the 47 ntd. construct “RRE IIB”.52 Additional minimization of RRE IIBproduces a construct small enough (34 ntd.) to be used for NMR structuredetermination (RRE JW).53 A duplex version of the RRE allows for thermaldenaturation studies (RRE DW).54

We have found that the “core element” of the RRE (bold Figure 2.4) contributes

the same Rev-peptide binding energy for different RRE constructs, including:

RRE 66, RRE IIB, and the RRE JW (Section 3.1).

U

GA

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

G

C

U

A

AC A

AUG

A

RRE 66

U

A

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

A CA

A

G

G

C

G

UG

C

G

C

RRE JW

U

A

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

G

C

U

A

AC A

AUG

G

AG

UG

C

G

C

U

A

C

G

G

C

G

C

RRE IIB

U

A

G

C

G GG

CG

C

A GU

C

G U

G

C

C

G

AG

C

U

A

G

CG

C

RRE DW

Stem IIB

Stem IIC

Stem IIA

19

2.3 Other RNA Constructs

Mutation of the minimized construct “RRE JW” can reveal which bases of the

RRE core element are essential for ligand binding. The G-rich bulge of RRE JW

has been substituted with inosine to generate “3I RRE JW” (Figure 2.5). A deoxy

version of the RRE “dRRE JW” allows a direct comparison between RNA and

DNA stem-loop recognition (Figure 2.5).

U

A C

G

CG

C

A GU A

C

G U

G

C

C

G

A C

AA

G

G

C

G

UG

C

G

C

3I RRE JW

A

G

C

G GG

CG

C

A GA

C

G

G

C

C

G

A C

AA

G

G

C

G

TG

C

G

C

dRRE JW

I I I

TT

T

Figure 2.5: RRE JW mutants used in these studies.

To evaluate the RRE specificity of a small molecule, its affinity to many different

nucleic acids must be evaluated (Section 1.6). A mixture of tRNAs (Sigma type-

X) provides a convenient and biologically relevant mixture of RNA (over 30

different mature and pre-tRNAs). Evaluation of such mixtures, however, is often

problematic (the binding constant of a small molecule to a mixture of nucleic

acids cannot be calculated). For this reason, a number of well-defined and

biologically relevant RNA constructs have also been used to examine small

molecule specificity (Figure 2.6).

20

Figure 2.6: Secondary structures of “other” RNAs used in these studies,including the 31 ntd. “TAR31” and the 27 ntd. prokaryotic “A-site27” constructs.

2.3.1 TAR31

“TAR31” is a model of the TransActivation Control Region of HIV-1.55 The Tat-

TAR interaction is necessary for the transcription of full-length HIV RNA (Figure

2.0). Inhibition of this RNA-protein interaction is another important target for drug

design.56 Both the TAR and the RRE recognize a homologous arginine-rich

domain within their respective binding partners (Tat and Rev, respectively). Both

RNAs do, in fact, have similar affinities for the Rev peptide “RevFl” (section 3.1,

Table 3.0). The RRE and TAR, however, are known to exhibit different types of

binding interactions with small molecules. The TAR has, for example, a high

affinity cation binding pocket that upon binding to guanidinium or Ca2+, facilitates

an alternate fold of the TAR.57,58 There is no evidence that such a binding pocket

G

G

G

G

GGG

G G

G

G

G

G

G

G G

G G

3'

3'

5'

5'

cleavage site

"Substrate" (S)

"Enzyme" (E) C

U

A

CC

C

C

CC

C

C

CA A A

A

A

A AA

AA

A

U

U

U

UU

C

U

UC

C

C CCG

HH16 Ribozyme

G G

GG

G G G

GG

G

C C

C C

C

C

C

CC

A A A

AU U

U U U

U CG

TAR31

C

G

G

G

GG

G GG G

CC

C

CC

C C

C C C

AU

AA

A A

A

A

A A

A

A

UU

U

U

U

UU U

U

UU

DD

G

GG

G

G GG

G

G

Cm

GmY

Ψ

T

m5C

A AAA

A

A

m5C

m7GΨ

m1A

C

C

CC

22m G

G

tRNAPhe

A-Site27

G G

G GC C

C

G G

CU A A

U

U

U GG

G C

AC

C C U

CA

21

exists on the RRE.59,60 No direct comparison of aminoglycoside affinities for the

TAR versus the RRE has been published; preliminary evaluations in our lab,

however, indicate that the trends in aminoglycoside affinity are similar for the

TAR and the RRE (as well as the HH16 Ribozyme).

2.3.2 The HH16 Ribozyme

RNA enzymes (ribozymes) are intriguing biopolymers capable of catalyzing a

variety of transformations.61 Their functional diversity can be attributed to the

complex three-dimensional folds that RNA can assume via a multitude of

secondary structures and tertiary interactions.62 Ribozymes, therefore, provide a

unique opportunity to correlate RNA structure and function. The hammerhead

ribozyme (HH16) is among the best characterized RNA enzymes.63,64 Following

association and proper folding of the enzyme and substrate strands,

phosphodiester hydrolysis is catalyzed at a specific location (Figure 2.6). Small

molecule ligands, including aminoglycosides, have been shown to inhibit the

enzymatic activity of hammerhead ribozymes.36,65,66 Two models are currently

proposed for the mechanism by which aminoglycosides inhibit enzymatic

cleavage: 1) direct displacement of enzymatically essential Mg+2 by the

ammonium groups of the aminoglycoside,65,67 2) upon binding the

aminoglycoside, RNA structural perturbations prevent the ribozyme from

adopting an enzymatically active conformation. Since Mg+2 is needed for proper

folding and cleavage of HH16,64 these mechanisms may be intimately linked.

22

2.3.3 A-site27

Many aminoglycosides exert their anti-bacterial activity by binding to the

decoding region of the 30S ribosomal subunit.30 “A-site27” is a hairpin model of

the prokaryotic ribosomal A-site, and has been shown to form a well-defined

complex with paromomycin.68 The NMR structure of parmomycin bound by the A-

site27 provided one of the first high-resolution views of a biologically relevant

RNA-small molecule complex.68 Recent crystal structures of the ribosome

indicate, however, that this region of the A-site is significantly distorted when

compared to NMR structure of A-site27.33 In addition, the binding of the A-site by

aminoglycosides may be affected by interactions mediated by tRNA residing in

the A-site.33 These “higher-order” interactions can complicate the relationships

observed in RNA-small molecule studies when minimized RNA constructs are

used. Indeed, Wong’s data demonstrates that there is no correlation between the

A-site27 affinity and the antibacterial activities of a large library of aminoglycoside

derivatives.69 Despite this, the A-site27 hairpin continues to serve as a model for

the development of small molecules with antibiotic activities.

2.3.4 tRNAPhe

tRNAPhe (Figure 2.6) has only recently been shown to bind to aminoglycosides.70

While tRNAPhe may not be an important target of drug design, it provides perhaps

one of the most extensively studied RNAs ever.71 Efficient commercial production

has allowed the widespread study of tRNAPhe for over 30 years. These studies

23

indicate that tRNAPhe has diverse structural features that provide well-defined

binding sites for small molecules.27

2.3.5 Duplex DNA and RNA

Simple duplex DNA and RNA also serve as convenient sources of “other” nucleic

acids as well. These commercially available duplexes are enzymatically

synthesized and will contain only a small fraction of hairpins and other structural

imperfections. These polymers will typically extend, unbroken, for thousands of

bases. Examples of the nomenclature used for these duplexes is as follows:

Name Type Structure

Poly r[A] – r[U] RNA duplex

Poly d[AT] – d[AT] DNA duplex

Poly r[A] RNA single strand

Oligo d[T]15 DNA single strand

By evaluating the binding affinity of a small molecule to many different nucleic

acids, the RRE specificity of new ligands can be determined, and the overall

nucleic acid selectivity of the ligand becomes apparent.

AU

A A A AU U U U

.

...

.

...

AT

T A T AA T A T

.

...

.

...

A A A A A. .. .

T (T)13T

24

3.0 Methods for Measuring RNA-Ligand Affinity and Specificity

Numerous methods are available for evaluating a small molecule’s ability to

inhibit the binding of two macromolecules. Traditional techniques such as gel-

shift electrophoresis and analytical ultracentrifugation are limited by their inability

to rapidly screen large numbers of potential inhibitors. Newer methodologies

such as mass spectrometry and surface plasmon resonance are powerful tools,

but hardware costs and data analysis can be vexing. Fluorescence-based

techniques have proven to be highly sensitive, versatile, and relatively

inexpensive for the examination of RNA-ligand binding interactions.72 Five

fluorescence-based techniques are described for the evaluation of both affinity

and specificity of RNA ligands.

3.1 Fluorescence Polarization Anisotropy

Organic fluorophores possess a transition dipole moment that is related to the

differences between the dipole moments of the ground state versus the excited

state.73 In order for a fluorophore to absorb energy from light, the oscillating

electronic vector of the photon must be parallel to the transition dipole of the

molecule. Following excitation, the photon that can be emitted from the excited

state will also be parallel to the transition dipole of excitation. Plane polarized

light that is absorbed by a fluorophore immobilized into a solid state will,

therefore, retain its polarization in the emitted photon. For fluorophores in

solution, the tumbling rate of the fluorophore is proportional to the degree of

depolarization. Polarization anisotropy is measured by comparing the

25

fluorescence emission intensity that passes through parallel versus perpendicular

polarizing filters (relative to the plane polarized excitation source (Figure 3.0)).

Figure 3.0: Fluorescence polarization anisotropy. The extent of polarization (the“anisotropy” value) is inversely proportional to the tumbling rate of thefluorophore in solution.74 The tumbling rate of the fluorophore is proportional tothe total size and shape of the fluorophore, as well as to the temperature andviscosity of the solvent.74

Figure 3.1: (A) Linear relationship between the concentration of the fluorescentpeptide “RevFl” and the total fluorescence intensity. (B) Fluorescence anisotropyof the fluorescent peptide “RevFl” as a function of the total emission intensity ofthe sample. Excitation is at 490 nm. Error bars for anisotropy reflect the standarddeviation for 10 anisotropy measurements made on the same day using thesame sample reflecting, therefore, only those errors related to the instrumentitself (Perkin Elmer LS-50B Luminescence Spectrometer). Systematic errorsassociated with macromolecular and small molecule quantifications are typicallythe largest sources of errors for the experiments described within this thesis(typically ± 30%). Errors associated with volume and weight measurements arerelatively small and random (typically ± 15%).

I

I

Anisotropy =I

I

I

I

-

+2

Emission Intensity (530nm)

0 200 400 600 800

Ani

sotr

opy

ofR

evF

l

0.070

0.075

0.080

0.085

0.090

0.095

0.100B)

Concentration of RevFl (nM)

0 5 10 15 20 25 30 35

Em

issi

onIn

tens

ity(5

30nm

)

0

200

400

600

800A)

26

In theory, the fluorescence anisotropy of a fluorophore is an intensive property,

and therefore not dependent of the total fluorescence intensity of the sample.74 In

practice, however, a significant perturbation is found at low emission intensities

(Figure 3.1 B).75 This effect is due in part to instrumental limitations, but is easily

avoided by keeping emission intensities above 100 (arbitrary units) using a

Perkin Elmer LS-50B Luminescence Spectrometer with maximum slit widths.

Fluorescence anisotropy provides a useful tool to follow the real-time association

and subsequent inhibition of the Rev-RRE interaction.76 The tumbling rate of the

fluorescein labeled Rev peptide “RevFl” decreases upon binding to the RRE (see

Figure 2.2 and Section E 2.1 for the sequence and structure of RevFl). As the

RRE is added to a 10 nM solution of RevFl, their association is observed by an

increased anisotropy value (Figure 3.2 A). Three of the RRE constructs shown in

Figure 2.4 (RRE 66, RRE IIB, and RRE JW) were evaluated for RevFl affinity.

Since the binding stoichiometry between Rev34-50 and RRE is known to be 1:1

(for all three of these RRE constructs), the association curves can be used to

calculate binding constants for each RNA (Figure 3.2 B, Table 3.0). The largest

RRE construct (RRE 66) has a 4-fold higher affinity for RevFl than RRE IIB and a

12-fold higher affinity than RRE JW (Table 3.0). Detailed kinetic analysis of Rev-

RRE binding reveals the association rate for 7 different RRE constructs is very

similar (near diffusion-limited at 106 M-1 s-1); the differences in affinity between

these different RRE constructs were attributed to different dissociation rates.77

27

Figure 3.2: (A) Binding of RevFl to three different RRE constructs, according tothe changes in the fluorescence anisotropy of RevFl (See section E 3.1 – E 3.2for experimental details). Notice how the magnitude of the anisotropy change atsaturation is roughly proportional to the size of each RNA construct (Figure 2.4).(B) Since these binding interactions have been shown to be simple two-statesystems, the change in fluorescence anisotropy is directly proportional to thefraction of RevFl bound by the RRE, allowing for simple analysis of the bindingdata (Section 3.6).

Table 3.0: RevFl Binding Constants (Kd) and Free Energies (Kcal/mol) at 22 °C.

RRE 66* RRE IIB* RRE JW* tRNAPhe* C.T. DNA** TAR 31*

Kd 3 ± .5nM 10 ± 3nM 35 ± 5nM 2.4 ± .4µM 25.6 /- 4µM 35 ± 5nM

∆G -11.6 ± .2 -10.8 ± .2 -10.0 ± .2 -7.5 ± .1 -6.18 ± .1 -10.0 ± .2

* Kd calculated per construct** Kd calculated per helical repeat (20 bases)*** ∆G = -RT ln(1/Kd)

A)

RNA Concentration (nM)

0 200 400 600

Flu

ores

cenc

eA

niso

trop

yof

Rev

Fl

0.075

0.080

0.085

0.090

0.095

0.100

0.105

0.110

0.115

RRE 66RRE IIBRRE JW

B)

Concentration of RNA (nM)

1 10 100 1000 10000

Fra

ctio

nof

Rev

FlB

ound

0.0

0.2

0.4

0.6

0.8

1.0

RRE 66RRE IIBRRE JW

28

Larger RNA constructs possess a greater total negative charge. It is not

surprising, therefore, that RRE 66 shows the highest affinity for the polycationic

Rev peptide (Table 3.0). By comparing the Rev binding energies to the size of

each RRE construct, an interesting trend is revealed (Figure 3.3). A plot of ∆G

versus the size of each RRE (in nucleotides) reveals a straight line that allows

one to extrapolate the binding energy of the RRE when the size of the RRE = 0

nucleotides (Figure 3.3). This indicates that the energetic contribution of the RRE

“core element” for all three constructs is ∆G = -8.4 Kcals / mole. The slope of this

line indicates that ∆G = -0.05 Kcal/ mole*nucleotide. This value may represent

the intra-complex electrostatic “advantage” provided by each additional base and

may be related to the magnitude of non-contact or “medium-range” electrostatic

interactions in this system. This type of binding data may assist future

computational modeling of electrostatically-driven RNA-ligand binding

interactions.78

Figure 3.3: A linear fit of ∆G (per nucleotide) versus RRE size (in nucleotides)has a Y-intercept of 8.4 Kcal/mole and a slope of –0.05 Kcal/mole*nucleotide.

0 10 20 30 40 50 60 70 80 90-13

-12

-11

-10

-9

-8

-7

-6

Size of RRE Construct (Nucleotides)

∆∆ ∆∆G

(Kca

l/mol

e)

29

The affinity of RevFl to a number of other nucleic acids has been evaluated. Both

tRNAPhe and a mixture of yeast tRNAs (Sigma type-X) have similar binding

isotherms that indicate a 1,000-fold lower affinity for RevFl as compared to a

similarly sized RRE (RRE 66, Table 3.0). Each helical repeat of calf thymus DNA

has approximately a 10,000-fold lower affinity to RevFl than the RRE (assuming

one peptide binding site per helical repeat of DNA). Rev’s low affinity to most

other nucleic acids indicates a high specificity for the RRE (Section 1.6). RevFl

does, however, bind to TAR31 with high affinity (Table 3.0), this finding has also

been reported by other groups.79 Since both the TAR and the RRE bind to

homologous arginine-rich domains within their respective proteins, the high

affinity of RevFl for TAR31 is not surprising. The full-sized Rev protein, however,

is reported to effectively discriminate between the TAR and the RRE, suggesting

that the Rev-protein has a higher RRE specificity as compared to RevFl.77 The

Rev protein cannot, however, easily be studied using fluorescence anisotropy

due to its tendency to self-associate and to bind the RRE in greater than 1:1

stoichiometry.80 In addition, significant changes in fluorescence anisotropy are

typically only observed when a small fluorescent ligand binds to a much larger

macromolecule (like the RevFl- RRE respectively). The mass of the Rev-protein

is on the same order as the 66 nt. stem II of the RRE; therefore upon binding the

RRE66, only a small change in the anisotropy of a fluorescent Rev protein would

be expected.

30

Fluorescence anisotropy provides a fast and accurate method for the evaluation

of molecules that can bind to the RRE and displace the Rev peptide (Figure 3.4).

Figure 3.4: Fluorescence anisotropy displacement experiments. Once the RevFl-RRE complex is formed (top), an inhibitor "X" can be added (bottom).Competitive inhibitors will bind to the free RRE, and shift the 3-way bindingequilbria towards release of the free RevFl peptide (causing a decreasedanisotropy of RevFl). If a single inhibitor binding site on the RRE displaces Rev,the absolute affinity of the RRE – “X” binding interaction can be calculated (Ki isequivalent to Kd). In most cases, however, the binding stoichiometry of theinhibitor is not known, so its affinity to the RRE is evaluated by measuring theconcentration of inhibitor needed to displace ½ of the Rev peptide from the RRE(defined as its IC50 value). See Section E 3.1 for experimental details.

To evaluate what impact, if any, the addition of fluorescein to the Rev peptide

has on RRE binding, a Rev peptide without fluorescein (“Rev-IA”) was used to

displace RevFl (see Section E 2.1 for the sequence and synthesis of each

peptide). Using fluorescence anisotropy, a Ki of 2 ± 1 nM was measured for Rev-

IA (see Section E 3.1 for details). This indicates that only a small perturbation

0 2 4 6 8 10 12 14 16

0 .0 80

0 .0 82

0 .0 84

0 .0 86

0 .0 88

0 .0 90

0 .0 92

0 .0 94

0 .0 96

0 .0 98

Concentration of "X"

0 50 10 0 15 0 200

0.0 80

0.0 85

0.0 90

0.0 95

0.1 00

0.1 05

0.1 10

Concentration ofRRE (nM)

0 .115

X

X

+Kd

Kd = 2 nM

+

Ki

Ki

RRE RevFl Rev-RRE Complex

=

=

[RRE] [X]

[X - RRE]

[RevFl]

Kd [Rev-RRE Complex] [X]

[X - RRE]A

niso

trop

yA

niso

trop

y

Kd =[RRE] [RevFl]

[Complex]

IC50 = 1 µM

Kd = 3 nM

31

(less than 2-fold) is introduced by the addition of fluorescein to the Rev peptide.

Native polyacrylamide gel-shift experiments also indicate that only a small

perturbation in affinity is introduced by the addition of fluorescein to the C-

terminus (data not shown). Compared to RevFl, a 5-fold lower affinity is observed

when fluorescein is attached to the N-terminus of the Rev peptide (see section E

2.1 and reference 76 for “FlRev”). The differences between the N- and C- termini

can be rationalized by examining the NMR structure of the Rev-peptide RRE

complex.53 The N-terminus of the α-helical Rev peptide is buried in the major

groove of the RRE, while the C-terminus is free in solution.53 Modification of

Rev’s N-terminus, therefore, interferes with RRE binding.

3.2 Solid-Phase Peptide Displacement Assay

A novel biophysical approach has been developed to characterize the affinity and

specificity of small molecules for the RRE. The assay is based on the competition

between potential RRE binders and RevFl for binding to an immobilized RRE 66

construct (Figure 3.5 – 3.6). Ligands that bind to the Rev binding site on the

RRE will displace RevFl into solution. The emission intensity of the solution is

then measured to evaluate the RRE affinity of the ligand.

The components of the assay are schematically shown in Figure 3.5 and include:

(a) insoluble agarose beads covalently modified with streptavidin, (b) a

biotinylated RRE fragment, and (c) the fluorescein-labeled Rev peptide “RevFl”.

Assembly of these components forms an immobilized “ternary complex”, where

32

the biotinylated RRE binds to streptavidin, and RevFl binds to the RRE with a

dissociation constant that is the same as measured in solution.

Figure 3.5: Assembly of the complex used for the solid-phase assay. Seesection E 3.2 for technical details regarding the solid-phase assay.

Calculating the binding constant for RevFl and the RRE66 on the solid-phase is

different than in solution, since the “surface volume” of free RRE is the same as

the surface volume of Rev-RRE complex. The equation for Kd is simplified:

Kd (on a surface) = [RRE][Rev-Fl] = (moles of free RRE)[Rev-Fl][complex] (moles of complex)

The concentration of RevFl in the supernatant “[Rev-Fl]” is calculated by

comparing the fluorescence intensity of the supernatant to a calibration curve of

known concentration versus fluorescence emission intensity (Figure 3.1 A). The

total moles of RNA on the solid support “(moles RRE)” is determined during the

immobilization of the biotinylated RRE: the UV absorbance of the supernatant is

used to calculate the amount of RNA absorbed by the agarose/streptavidin gel.

33

The moles of RRE-bound by RevFl “(moles complex)” is determined by stripping

RevFl from the gel with 8M guanidinium HCl, then quantifying the supernatant in

a buffered solution (see section E 3.2 for additional details).

The “solid-phase” Kd for RevFl-RRE66 is found to be 3 ± 2 nM, and is

independent of the effective loading of the RRE onto the agarose beads (from

0.1 – 2.0 nmoles RRE/mL agarose), suggesting that the immobilized RRE

complexes are non-interacting over this loading range.

As in solution, RevFl binds to the RRE with high specificity. The immobilized

Rev-RRE complex is not significantly inhibited by the addition of large excesses

(200-fold) of tRNA or DNA (low µM concentrations). Titration of the RRE or Rev-

IA, however, efficiently displaces RevFl from the solid-support (IC50 = 90 nM and

35 nM, respectively). Importantly, RevFl shows no binding of

argarose/strapavidin gel that lacks immobilized RRE. See section E 3.2 for

technical details regarding the solid-phase assay.

Following formation of the immobilized Rev-RRE complex, small molecule

inhibitors are added, and the fraction of RevFl in solution is monitored as a

function of inhibitor concentration (Figure 3.6 A). The concentration of inhibitor

needed to displace ½ of the peptide (the IC50 value) is measured in the presence

and absence of an excess of competitor nucleic acids (Figure 3.6 A and B).

34

Figure 3.6: Measuring the relative RRE affinity and specificity of a ligand usingthe solid-phase assay.

The ratio of these IC50 values reflects the affinity of the small molecule to the

competitor nucleic acids relative to the RRE (Figure 3.6 A and B). To evaluate

the RRE specificity of a small molecule, numerous mixtures of competing nucleic

acids are used (tRNAMIX, C.T. DNA, etc.). Weighting the RRE affinity of the small

molecule by its affinity to a large number of other potential binding sites allows

the RRE specificity of a small molecule to be established (Section 1.6). Indeed,

the RRE specificity, as we have defined it, is proportional to the Rev-RRE

35

inhibitory activity of a ligand in the presence of “all” other potential nucleic acids

targets (Section 1.6). Most biophysical methods (including fluorescence

anisotropy, surface plasmon resonance, NMR) are not amenable to the analysis

of complex mixtures of biomolecules, making the examination of specificity an

arduous task.

To examine the effectiveness and dependability of this novel assay, two families

of known Rev-RRE inhibitors have been evaluated: the aminoglycoside

antibiotics and the aromatic diamidines (Figure 3.7). An example of the raw

displacement data is shown in Figure 3.8, where increasing concentrations of

neomycin B displace the RevFl peptide. Table 3.1 summarizes the IC50 values

obtained by the solid-phase assay in comparison to fluorescence anisotropy

measurements. An excellent agreement between the two assays is observed for

the aminoglycosides.

Table 3.1: IC50 Values (µM) for Rev-RRE inhibition*

Compound Anisotropy** Solid-Phase Solid-Phase Solid-Phasewith DNA with tRNA

Neomycin B 7 7 8 20Tobramycin 47 45 50 100Kanamycin B 80 90 90 170Kanamycin A 780 750 750 1200

"DB340" 0.3 0.1 1.3 0.7"DB182" 1.5 0.4 4 0.8"A132" 10*** 1.2 10 3

* Standard deviation of all measurements are within ± 25% of reported value.** Using 10 nM RevFl and 100 nM RRE66.*** Significant fluorescence quenching during course of titration.

36

OHOH2N

OH2N NH2

O

O

OH

R1

NH2

OH

HTobramycin (3)

R2OH

R1

HO

Kanamycin A (1)

HO

OHKanamycin B (2)

R2

OH

NH2

NH2

R O R

R OR NH R

HN

N

R

NH2

"A132" (6)

"BD182" (8)

"DB340" (7)

R =NH (CH2)3NH(CH3)2

(A)

(B)

O

NH2

H2NO

HOHO

OO

NH2

OH

R

OHO

HOO

HO

HO

NH2

H2N

Neomycin B (5)

OHR1

Paromomycin (4)

NH2

Figure 3.7: Known small-molecule inhibitors of Rev-RRE binding: theaminoglycosides (A)82 and the “aromatic diamidines” (B).83

The aromatic diamidines are more potent Rev-RRE inhibitors than the

aminoglycosides, and their previously reported trends in RRE affinity are again

accurately reproduced (DB340 > BD182 > A132).83a A 10-fold increase in IC50

values in the presence of competing DNA indicates that these compounds also

have a high affinity for double-stranded DNA (Table 3.1). In addition, these

compounds have a moderate affinity for a mixture of tRNA. Taken together, this

indicates that these compounds are only moderately selective for the RRE.

DB340, however, forms a well-defined complex with the RRE that has been

characterized by NMR.84 Interestingly, this compound binds as a dimer in the

37

minor groove of the RRE.84 Since Rev binds in the major groove of the RRE,53

the displacement of RevFl by DB340 must involve an allosteric mechanism.

Figure 3.8: Raw data from solid-phase displacement experiments showing thefluorescence emission of the supernatant as a function of neomycin Bconcentration. To reduce light scattering at 500 nm, excitation of RevFl is at470nm. Excitation at 490nm also yields the same IC50 values. See E 3.2 foradditional details.

Systematic discrepancies between IC50 values determined from the solid-phase

assay versus fluorescence anisotropy (Table 3.1) reflect a general problem

associated with applying fluorescence anisotropy to binders that are either

fluorescent or fluorescence quenchers. Fluorescent inhibitors will themselves

have a fluorescence anisotropy value. In addition, significant quenching of the

RevFl peptide can affect the measured anisotropy (Figure 3.1 B). The solid-

phase assay easily adapts to these problems: the supernatant (containing

excess inhibitor and displaced RevFl) can be removed, and the fluorescence

500.0 520 540 560 580 600 620 640 660 680 700.5

0.0

5

10

15

20

25

30

35

40

45

7.1In

tens

ity(a

rbitr

ary

units

)

Wavelength (nm)

Control (no neomycin )

4.3 µM

10 µM

21 µM

43 µM

100 µM

2000 µMControl (8M guanidinium HCl)

38

remaining on the solid support can be quantified. After removing the supernatant,

isopropyl alcohol is used to wash the support (this effectively “locks” the

remaining RevFl onto the immobilized RRE, due to salt-bridge formation). The

peptide remaining on the support is quantified by “stripping” the gel with 8M

guanidinium HCl and quantifying the supernatant in buffer. Gel mobility shift

assays were conducted to confirm that for inhibitors that interfere with RevFl

fluorescence, IC50 values obtained from the solid-phase assay are, indeed, more

accurate than the IC50 values derived from fluorescence anisotropy (Figure 3.9).

Figure 3.9: Native gel-shift electrophoresis is used to visualize formation andsubsequent inhibition of the Rev-RRE complex, thereby providing an orthogonalmethod to measure the approximate IC50 values of Rev-RRE inhibitors. Thenumerical values indicate the concentration of each ligand (µM). Interestingly, akinetically stable complex is formed between DB340 and the RRE66, resulting ina small shift compared to the free RRE. See E 3.2.3 for experimental details.

The solid-phase assay is used to discover and characterize new RRE

ligands.81,85 For example, the enantiomeric octahedral metal complexes ∆- and

Λ- [Ru (bpy)2 Eilatin]+2 are found to be potent inhibitors of the Rev-RRE complex

and exhibit very interesting selectivities (see Figure 3.10 and Table 3.2). Both

fluorescence anisotropy and the solid phase assay indicate that each enantiomer

has approximately the same RRE affinity. In the presence of DNA, however, the

Λ enantiomer loses its potency by a factor of 25, and the ∆ enantiomer by a

RRE

RRE+

Rev

Neomycin

100101

"A132" "DB340" "BD182" Ru(bpy)3 EilatinRu(bpy)20.1 1 10 0.1 1 10 0.1 101 10,0001,000100 0.1 1 10

39

factor of 5. This indicates that the Λ enantiomer has a higher affinity for double-

stranded DNA as compared to the ∆ enantiomer. In the presence of tRNAMIX, the

potency of neither compound is affected. This indicates that these metal

complexes have a high specificity for the RRE as compared to other RNAs, but

also exhibit high affinity for DNA. Eilatin-containing metal complexes also

possess anti-HIV activities and some very surprising trends in nucleic acid

specificity (Section 4).

FIgure 3.10: Octahedral metal complexes evaluated by the solid-phase assay.

Table 3.2: IC50 Values (µM) for Rev-RRE inhibition*

Compound Anisotropy Solid-Phase Solid-Phase Solid-Phase+DNA +tRNA

∆-[Ru(bpy)2Eilatin]2+ 1 2 10 2Λ-[Ru(bpy)2Eilatin]2+ 1 2 50 2[Ru(bpy)3]

2+ 170** >10,000 n.d. n.d.

* Standard deviation of all measurements are within ± 30% of reported value.**Significant fluorescence quenching during course of titration.

Both affinity and specificity are crucial for the design and evaluation of RNA-

protein inhibitors. The assembly of an immobilized Rev-RRE complex has

generated a simple, rapid, and quantitative assay that has facilitated the

∆-[Ru(bpy)2Eilatin]2+ (10)Λ-[Ru(bpy)2Eilatin]2+ (9) rac-[Ru(bpy)3]2+ (11)

Ru

N

N

N

N

N

N

N

N

Ru

N

N

N

N

N

N

2+

Ru

N

N

N

N

N

N

N

N

2+ 2+

40

discovery and characterization of new RNA ligands. Critical evaluation of the

assay itself has shown that the expected inhibitory activities of diverse and

potentially problematic families of molecules are faithfully reproduced. Since the

solid-phase assay can be conducted in complex mixtures of biomolecules, it is a

powerful tool for the characterization of an inhibitor's specificity. This type of

assembly should be useful in other nucleic-acid protein systems and amenable to

high throughput automation.

3.3 Ethidium Bromide Displacement

The fluorescence intensity of ethidium bromide (Figure 3.11) typically increases

upon the binding of nucleic acids.86 Subsequent displacement of ethidium is

apparent by a decrease in emission intensity upon addition of a competitive

inhibitor (Figure 3.12). Since it binds to most DNAs and RNAs, ethidium

displacement experiments can be conducted using a wide range of nucleic

acids.72b

Figure 3.11: Two classic intercalating agents that bind to the RRE. 9-aminoacridine also carries a positive charge under neutral conditions (pKa = 8.1).88

BrN

NH2H2N

ethidium bromide (12)

N

NH2

9-amino acridine (13)

- +

41

Figure 3.12: (A) Raw emission data of 1.25 µM ethidium bromide (uponexcitation at 546 nm in buffer only) (black square), and upon addition of 0.88 µMbase pairs of C.T. DNA (black circles). (10) is then titrated from 0.077 µM through7.1 µM. (B) The fluorescence intensity of the ethidium-DNA complex decreaseswith increasing concentrations of (10). The dotted line indicates the fluorescenceintensity of 1.25 µM ethidium in buffer only. (C) Assuming a linear relationshipbetween the change in fluorescence intensity and the fraction of ethidiumdisplaced, classic S-shaped binding isotherms are obtained. This allows for thedetermination of the concentration of each inhibitor needed to displace ½ of theethidium from C.T. DNA. A small amount of direct quenching of free ethidium bythese metal complexes results in final intensities that are slightly lower than thatof free ethidium (B). This is reflected in the secondary linear element in eachbinding isotherm starting at approximately 5 µM (C).

Wavelength / nm

570 580 590 600 610 620 630 640 650 660 670 680 690 700

Inte

nsity

(arb

itrar

yun

its)

0

10

20

30

40

50

60

70

80

90

100 Ethidium+DNA0.077µM (10)0.21µM (10)0.46µM (10)1.43µM (10)3.2µM (10)7.1µM (10)Ethidium

Concentration of (10) (µM)0 1 2 3 4 5 6 7 8

Inte

nsity

at60

5nm

0102030405060708090

Concentration (µM)0.01 0.1 1 10

Fra

ctio

nD

ispl

aced

0.00.20.40.60.81.01.2

(9)(10)(15)

A

B

C

42

Ethidium is known to have variable binding stoichiometries and a wide range of

affinities to different nucleic acids.87 The IC50 values for a given inhibitor,

therefore, cannot be compared between different nucleic acids (unless Ki values

are calculated, requiring a rigorous assessment of both the ethidium-nucleic acid

affinity, and the binding stoichiometries involved in displacement). Comparing the

IC50 values for ethidium displacement from the same nucleic acid, however, is

useful for comparing the relative affinity of each ligand. Representative examples

of ethidium displacement experiments are shown in Figure 3.12. In agreement

with the solid-phase assay, ethidium displacement indicates that the Λ

enantiomer of [Ru(bpy)2Eilatin]2+ (9) has a significantly higher affinity to calf

thymus DNA as compared to the ∆ enantiomer (10).

3.4 Direct Observation of Small Molecule - Nucleic Acid Association:Monitoring Photophysical Changes of the Small Molecule

A number of Rev-RRE inhibitors exhibit changes in their photophysical properties

upon binding nucleic acids (some examples of “direct binding studies” using

fluorescence anisotropy have already been presented in Section 3.1). In addition

to fluorescence anisotropy, UV-vis absorbance and/or fluorescence emission

spectroscopy are routinely used to probe RNA or DNA binding of small

molecules. The sensitivity of each technique is, however, significantly different.

Fluorescence emission measurements are often 100 – 1,000 fold more sensitive

than absorbance spectroscopy. By measuring the spectral changes of the ligand

as a function of nucleic acid concentration, binding isotherms can be generated

43

and used to calculate affinities. Direct binding analysis offers some advantages

over the displacement assays described above. Direct binding isotherms can be

used to calculate absolute binding affinities (Kd), which are more

thermodynamically meaningful than activity measures (IC50 values). Direct

observation of small molecule - RRE binding cannot, however, determine the

functional inhibition of an RNA or DNA. Some small molecules, for example, can

bind the RRE with high affinity, but not displace the Rev peptide. For example

eilatin (14) binds to the RRE with sub-µM affinity (as evidenced from its changes

in fluorescence anisotropy), but it does not displace the Rev peptide at these

concentrations (Figure 3.13). This indicates that eilatin’s high-affinity binding

site(s) on the RRE are not disruptive to Rev binding. Neomycin B is another

example of a ligand that has a high-affinity binding site on the RRE, but

association to this site is not sufficient to displace Rev.89

Some small molecules, including ethidium bromide (12), show more consistency

between direct binding and Rev displacement isotherms (Figure 3.14). It

appears, therefore, that ethidium’s highest affinity binding site(s) on the RRE are

capable of displacing Rev.

44

Figure 3.13: Binding of eilatin (200 nM) to the RRE as determined by monitoringthe fluorescence intensity of eilatin (14) as a function of RRE66 concentration(●), or by using eilatin to displace RevFl from the Rev-RRE complex (10 nMeach) (●).

Figure 3.14: Binding of ethidium bromide (600 nM) to the RRE as determined bymonitoring ethidium bromide’s fluorescence intensity as a function of RREconcentration (●), or by using ethidium bromide (12) to displace RevFl from theRev-RRE66 complex (●).

10-3 10-2 10-1 100 101 102

0.0

0.2

0.4

0.6

0.8

1.0 RRE TitrationRev-Fl Displacement

10-3 10-2 10-1 100 101 102

0.0

0.2

0.4

0.6

0.8

1.0 RRE TitrationRev-Fl Displacement

Concentration (µM)

Concentration (µM)

Fra

ctio

n alC

hang

eF

ract

ion a

lCha

nge

45

The interpretation of direct binding data is critically dependent on the

experimental conditions, especially the concentration of the observable species

relative to the affinity of the RNA-ligand interaction. Most small molecules will

have more than one binding site on any given RNA construct, and each of these

binding sites will typically have a different affinity for the small molecule.89 If the

concentration of the observable ligand is low compared to the concentration of

nucleic acid added to bind ½ of the ligand, the binding isotherm will reflect the

ligand’s association to its highest affinity binding site. At higher ligand

concentrations, the binding isotherm may also reflect binding to some of its

lower-affinity binding sites as well. If the concentration of the ligand is very high

compared to its nucleic acid affinity, the binding isotherm will reflect its binding to

all potential sites. The information provided by each these three situations is very

different, and the binding isotherms must be interpreted accordingly (Section

3.6).

3.5 Direct Observation of Small Molecule - Nucleic Acid Association:Monitoring Photophysical Changes of the Nucleic Acid

3.5.1 Thermal Denaturation

The bases of RNA and DNA are active chromophores. The thermal denaturation

of nucleic acid can be monitored by measuring the change in UV absorbance of

a nucleic acid solution as a function of temperature. A significant hyperchromicity

(a higher molar extinction coefficient) of the nucleic acid at 260 nm is observed

upon the melting of duplex regions into single stands (usually a 5 – 15%

46

increase, depending on the specific base composition and initial structure). The

temperature of melting (Tm) is defined as the temperature where ½ of the DNA or

RNA becomes denatured; for duplex nucleic acids this transition is highly

cooperative and appears as an inflection point in the plot of absorbance versus

temperature (Figure 3.15).

Figure 3.15: Thermal denaturation of calf thymus DNA (20 µM of bp) (A). The Tm

appears as an inflection point at 88.5 °C. Upon addition of 10 µM of ∆- [Ru (bpy)2

Eilatin]+2 a ∆Tm = +4.3 °C increase is observed (B), or upon addition of 10 µM ofΛ- [Ru (bpy)2 Eilatin]+2 a ∆Tm = >+9 °C increase is observed (C). See E 3.5 forexperimental details.

Temperature (oC)

50 60 70 80 90 100

Cha

nge

inA

bsor

banc

e(2

60nm

)

0.00

0.05

0.10

0.15

0.20Tm = 92.8 oC

50 60 70 80 90 100

Cha

nge

inA

bsor

banc

e(2

60nm

)

0.00

0.05

0.10

0.15

0.20

Tm > 97.6 oC

Tm = 88.5oC

50 60 70 80 90 1000.00

0.05

0.10

0.15

0.20

A

B

C

Cha

nge

inA

bsor

banc

e(2

60nm

)

47

The Tm can be monitored in the presence of small molecule ligands to examine

relative binding energies (Figure 3.15). In most cases, the small molecule binds

to the duplex nucleic acid with a higher affinity than the single-stranded form,

thus stabilizing the duplex form. Stabilization shifts the Tm to higher

temperatures.90 The change in the Tm (∆Tm ) is proportional to the ligand’s affinity

to the folded form of nucleic acid weighed by its affinity to the melted form.

Consistent with ethidium displacement and the solid-phase assay, Tm

experiments indicate that Λ- [Ru (bpy)2 Elatin]+2 (9) has a higher affinity for

duplex DNA as compared to ∆- [Ru (bpy)2 Elatin]+2 (10) (Figure 3.15).

3.5.1 Fluorescent Nucleoside Mimics

A recent and significant improvement in the direct spectroscopic observation of

nucleic acids has proven to be the incorporation of fluorescent base analogs into

nucleic acids.91 Currently, 2-aminopurine (2-AP) is unmatched in its desirable

fluorescence properties along with its minimal perturbation of nucleic acid

structure.92 2-AP is capable of base-pairing with both uracil and thymidine and

introduces little or no disruption of RNA and DNA duplexes (Figure 3.16).92,93

Figure 3.16: Adensine-uracil base pair and 2AP-uracil base pair.

NN H

O

R O

R

HNH

NN

N

N

AU

NN H

O

R O

RNN

N

N

H NH

U 2AP

48

2-AP’s fluorescence emission intensity is highly sensitive to environmental

changes, making it an attractive probe for nucleic acid structure and dynamics.92

A number of interesting studies have been made possible by site-specific

incorporation of 2-AP into RNA, including: the analysis of RNA-RNA binding,94

folding and catalysis,95 and Mg+2 binding.96 When 2-AP was incorporated into

the RRE IIB, important new insights were gained regarding the mechanism

responsible for Rev peptide displacement from the RRE by neomycin B.89

In our continuing investigation into the mechanism(s) responsible for the

inhibition of the enzymatic activity of the HH16 ribozyme (section 2.2 and refs

35,64), we have incorporated 2-AP into the substrate strand of HH16 ribozyme

(Figure 3.17). By placing 2-AP immediately adjacent to the cleavage site, it was

hoped that the cleavage chemistry of HH16 could be monitored in real time.97

Both enzyme-substrate association and enzymatic cleavage can be followed by

monitoring the fluorescence of emission of 2-AP (Figure 3.18 and 3.19). Upon

titration of the enzyme strand, the emission intensity of the 2-AP-containing

substrate is significantly quenched, allowing for the measurement of enzyme-

substrate affinity at three different salt concentrations (Figure 3.18).

49

Figure 3.17: Enzyme-substrate association, followed by enzymatic cleavage forthe hammerhead HH16 ribozyme (M = G, N = C) and for the fluorescent HH16ribozyme (M = 2-AP, N = U).

Figure 3.18: Conducted in the absence of Mg2+, the fluorescence intensity of the2-AP modified HH16 substrate is monitored as a function of enzymeconcentration. At 50 mM NaCl, Kd = 17 ± 2 nM, at 200 mM NaCl Kd = 5.7 ± 1 nM,and at 500 mM NaCl Kd = 1 ± 0.7 nM. Curve fitting was conducted with Prism 3.0(using exponential decay fit). See section E 3.5.1 for experimental conditions.

1 10 100 10000

10

2050mM NaCl200mM NaCl500 mM NaCl

Enzyme Concentration (nM)

Em

issi

onIn

ten

sity

G

G

G

G

GGG

G G

G

G

G

G

G

G G

G G

3'

3'

5'

5'

cleavage siteC

U

A

CC

C

C

CC

C

C

CA A A

A

A

A AA

A

A

A

U

U

U

UU

N

U

UC

C

C CCM

G

G

G GG

GG

G G

G

G

G

G

GG G

G G

3'

3'

5'

5'

C

U

A

CC

C

C

CC

C

C

CA

A AA

A

A AA

A

A

A

UU U

UU

N

U

UC

C

C CCM

E•S

Product (P2)

Product (P1)

E•P1P2

k2

"Substrate" (S)"Enzyme" (E)

+G

G

G

G

GGG

G G G

GG

G

G

G GG G

3'

3'

5'

5'

C

U

A

CC

C

C

CC

C

C

CA A A

A

A

A AA

AA

A

U

U

UUU

N

U

UC

CC CCM

Keq

50

As expected, the enzyme-substrate affinity increases at higher ionic strengths.

Interestingly, the fluorescence intensity at the end-point of titration is different for

each ionic strength. This suggests that at higher NaCl concentrations, 2-AP

stacks better with neighboring bases (Figure 3.18).

Following enzyme-substrate association, cleavage chemistry can be initiated by

the addition of Mg2+ (Figure 3.19). Substrate cleavage is evident by an

approximate 3-fold increase fluorescence intensity of 2-AP (Figure 3.19 black

circles). A large excess of the enzyme relative to the substrate is used, allowing

for pseudo first-order kinetic analysis of the cleavage data, and yields a cleavage

rate of 0.23 ± 0.01 min-1 for the fluorescent HH16 ribozyme. This value is

identical to the rate constant measured for the non-fluorescent HH16 ribozyme,

as determined by polyacrylamide gel electrophoresis.97

Figure 3.19: Real-time cleavage of an ensemble of fluorescent HH16 ribozymesin the presence of 200 mM NaCl and10 mM MgCl2, as evidenced by the increasein emission intensity of the 2-AP containing substrate (black circles). Increasingconcentrations of neomycin decreases the cleavage rate (red = 100 µM, blue =250 µM, green = 500 µM, pink = 1000 µM). See Section E 3.5.1 forexperimental details.

0

10

20

30

40

50

0 5 10 15 20

0

Time (min)

Cha

nge

inem

issi

onin

tens

ity

51

At physiologically relevant ionic strengths, neomycin B is a poor inhibitor of HH16

cleavage. Approximately 250 – 500 µM of neomycin is needed to decrease the

HH16 cleavage rate by 50% (at 200 mM NaCl and 10 mM Mg2+).97 At lower ionic

strengths (50 mM NaCl), neomycin is 5-10 fold more effective.66 This is

consistent with its electrostatically-driven binding of RNA.36

Figure 3.20: Binding of neomycin to the fluorescent HH16 complex in theabsence of Mg2+ (1.8 µM of substrate and 10 µM of enzyme in 200 mM NaCl). Athird neomycin binding isotherm becomes obvious at higher concentrations (notshown). See E 3.5.1 for experimental details.

Interestingly, neomycin binds to at least one site on the HH16 ribozyme with a

much higher affinity as compared to the concentrations needed for functional

inhibition of enzymatic cleavage. To prevent cleavage chemistry, the fluorescent

HH16 substrate-enzyme complex is formed in the absence of Mg2+, and upon

titration of neomycin two different responses from the fluorescent ribozyme are

observed (Figure 3.20). At low neomycin concentrations, the emission intensity

decreases (Figure 3.20 Left); as more neomycin is added, the emission intensity

recovers back to its original level (Figure 3.20 Right).

0 500 1000 1500 200025

30

35

40

45

Neomycin concentration (µµµµM)

Em

issi

onIn

ten

sity

(arb

)

0 10 20 30 40 50 6025

30

35

40

45

Em

issi

on

Inte

nsi

ty(a

rb.)

Neomycin concentration (µµµµM)

52

Assuming that the first binding isotherm (Figure 3.20 Left) represents association

to a single neomycin binding site, the neomycin affinity (Kd) of this site is 4.3 µM

(at 200 mM NaCl) and Kd = 2.2 µM (at 50 mM NaCl). Interestingly, no inhibition of

enzymatic cleavage is observed at low neomycin concentrations (through 10

µM). A low concentration of neomycin (2 µM) may slightly increase the cleavage

rate of the HH16 ribozyme to 0.25 ± 0.02 min-1 (at 200 mM NaCl). These results

indicate that neomycin binds to the HH16 at one or more non-inhibitory binding

sites.

If the second binding isotherm (Figure 3.20 Right) represents another

independent binding site, the affinity of neomycin to this site is Kd = 400 µM (at

200 mM NaCl), and Kd = 60 µM (at 50 mM NaCl). This represents a 100-fold

lower affinity compared to neomycin’s higher affinity binding site(s). A third

binding isotherm is sometimes observed at high neomycin concentrations,

suggesting either non-specific binding or aggregation can occur at concentrations

over 1 mM (not shown). The second binding isotherm of neomycin (Figure 3.20

Right) correlates with its HH16 inhibitory activity, where the affinity of the second

binding site is approximately equal to the concentration of neomycin needed to

slow ribozyme cleavage by 50% (IEC50).

Neomycin exhibits a similar behavior with the RRE, where its high-affinity site is

not capable of displacing the Rev peptide.89 This may represent a general feature

53

for aminoglycosides, where many potential binding sites are available and

multiple equivalents of the small molecule are needed for functional inhibition.

3.6 Analysis of Binding Data

Analysis of direct binding data must take into account the concentration of the

observable species relative to the affinity of the small molecule - RNA interaction

being studied. Three extreme theoretical cases illustrate how the “concentration

versus affinity ratio” dictates the relative sensitivity of each binding isotherm to

affinity versus stoichiometry (Figure 3.21). For most titrations, the concentration

of the observable species is kept constant, while the non-observable species is

titrated in. If the concentration of the observable species is much lower than the

Kd of the interaction, then the binding isotherm is sensitive only to the highest

affinity binding site. In this case, the binding stoichiometry can be assumed to be

1:1, and the Kd at this site is equal to the concentration of the non-observable

species needed to bind ½ of the observable species (defined as the C50 value)

(Figure 3.21 A). On the other hand, if the concentration of observable species is

much higher than the Kd of the interaction, the binding isotherm is sensitive only

to the maximum binding stoichiometry of the interaction (Figure 3.21 B). The

stoichiometry is determined by taking a ratio of concentrations at the intersection

of the two linear elements apparent within the binding isotherm (Figure 3.21 B). If

the concentration of the observable ligand is approximately equal to the Kd of the

interaction, then the binding isotherm reflects both the affinity and stoichiometry

of the interaction (Figure 3.21 C).

54

Figure 3.21: Three hypothetical binding isotherms for the addition of RNA into asolution containing an observable ligand. If the ligand concentration is very lowcompared to its affinity for the RNA, the binding curve is hyperbolic, and theconcentration of RNA needed to bind ½ of the ligand is equal to its affinity (A). Ifthe ligand concentration is very high compared to the affinity of the interaction,then the binding isotherm is composed of two linear elements. The point ofintersection indicates the binding stoichiometry (B). If the ligand concentration isapproximately equal to the affinity, then the binding isotherm reflects both thestoichiometry and affinity of the interaction.

0 1 2 3 4 50.00

0.25

0.50

0.75

1.00

RNA Concentration

Fra

ctio

nB

ou

nd

0 1 2 3 4 50.00

0.25

0.50

0.75

1.00

RNA Concentration

Fra

ctio

nB

oun

d

0 1 2 3 4 50.00

0.25

0.50

0.75

1.00

RNA Concentration

Fra

ctio

nB

oun

d

A

B

C

Kd

Stoichiometry =RNA concentration

ligand concentration

55

In many cases, a nucleic acid will have multiple binding sites for a small molecule

with similar affinity. The occupancy of one binding site can increase or decrease

the apparent affinity of the neighboring binding sites. By plotting the binding

isotherm on a semi-log scale, the “Hill coefficient” (a measure of cooperativity)

can be determined (Figure 3.22).

Figure 3.22: Three theoretical binding curves for the addition of a ligand into asolution of an observable nucleic acid, showing positive cooperatvity (n = 2), non-cooperativity (n = 1), and negative cooperativity (n = 0.5). The Hill coefficient “n”is apparent as the slope of each plot on a semi-log scale. The Hill coefficient for a1:1 association is always n = 1 (Figure 3.2 B and 3.18); but a Hill coefficient of n= 1 does not necessarily mean the stoichiometry is 1:1.

The “n” value of a binding isotherm can be determined by fitting the data to the

simplified Hill equation: Z = ( Y / EC50 )n / ( 1 + ( Y / EC50 )n )

Z = fraction of ligand bound, n = Hill coefficient, EC50 = concentration of added

nucleic acid or small molecule where the fraction bound = 0.5, and Y is the total

concentration of the added species. Notice that the concentration of the

observable species is not included in the equation. The slope of the Hill plot can,

0.001 0.01 0.1 1 10 100 1000 10000

0.0

0.2

0.4

0.6

0.8

1.0

n = 2

n = 0.5n = 1

Concentration of Small Molecule

Fra

ctio

nof

Site

sB

ound

56

therefore, be affected by the concentration of observable species (especially if

the concentration is high compared to the affinity of the interaction being

studied). If the nucleic acid is the observable species, the lower its concentration

compared to the affinity of the interaction, the better. When the small molecule is

the observable species, the situation is inherently more difficult, and the best

conditions for observing cooperativity is when the concentration of the small

molecule is approximately equal to the affinity of the interaction(s) (Kd). In this

case, titrations must be conducted at different concentrations of the small

molecule to observe what dependence on concentration is observed (if any).

3.6.1 Linear Analysis

If the binding interaction is non-cooperative (Hill coefficient n = 1), and if the

isotherm is sensitive to both affinity and stoichiometry (3.21 C), direct binding

data can be analyzed using a Scatchard plot. This plot can simultaneously

determine both binding stoichiometry and the average affinity of each site.

The Scatchard equation: r/Cf = Keq (N-r)

where r = concentration of bound small moleculetotal concentration of nucleic acid

Cf = concentration of unbound small molecule, and N = number of binding sites.

A Scatchard plot is made by plotting r versus r/Cf. The slope of this plot will equal

Keq, and the X-intercept will equal the number of binding sites. The direct binding

data from Figure 3.14 can be incorporated into a Scatchard plot (Figure 3.23).

57

Figure 3.23: Scatchard plot for ethidium bromide (600 nM) binding to RRE JWsuggests the presence of two similar binding sites with an average affinity of Kd =160 nM.

Large errors with Scatchard analysis are often encountered.98 Since the

concentration of observable species can determine the number of binding sites

reflected in the isotherm (Figure 3.21), the apparent stoichiometry can change

based upon the concentration of the observable species. In addition, the errors

associated with assigning spectral properties of the 100% “free“ versus the 100%

“bound” become amplified in all the data points, since the fraction bound at each

data point is calculated from these two extremes. The data points for the 100%

free and the 100% bound states are, therefore, “weighed” much more heavily

than the points in the middle of the titration. Non-linear analysis of binding data

can help reduce the errors associated with quantifying the spectral properties of

these “extreme” (and often inaccurate) data points.

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.0×10 -00

5.0×10 06

1.0×10 07

1.5×10 07

r

r/C

f

58

3.6.2 Non-Linear Analysis

Line shape analysis can eliminate much of the error associated with quantifying

the spectral properties of the 100% “free“ versus the 100% “bound” states of the

observable species. Non-linear analysis typically weighs all data points equally

and fits all the points to a theoretical curve.

In the case where the concentration of ligand is much lower than the Kd (Figure

3.21A), a hyperbolic curve is used to fit both the affinity and end-point of the

titration: Y = (Bmax*X) / (Kd + X)

Where X = nucleic acid concentration, Y = change in the spectroscopic signal of

the ligand, Kd = EC50, and Bmax = the end-point of the titration (maximum change

in the spectroscopic signal of the ligand).

In cases where binding stoichiometry is experimentally determined to be 1:1, raw

spectroscopic data can be fit with the following equation to determine affinity

(Kd):99,100

Where “RevFl” is the observable species that has an initial anisotropy or

fluorescence value of A0, [RNA] is the concentration of added species, and ∆A =

the total change upon saturation.

A = A0 + ∆A([RNA]total+[RevFl]total+Kd) - ([RNA]total+[RevFl]total+Kd)2 - 4[RNA]total [RevFl]total

2[RevFl]total

59

In many cases, binding isotherms are complex and do not fit any of these simple

theoretical models. Issues such as small molecule - small molecule association,

multiple binding sites, etc. can complicate analysis such that binding constants

cannot be calculated with confidence. In some cases, more advanced models

can assist with analysis.98 In other cases, however, no model can be applied with

any degree of confidence, and trends in activity (IC50 or C50 values) must be used

to evaluate relative binding affinities.

Fluorescence-based assays are important new tools to measure macromolecular

association and inhibition. Potential problems arise, however, when evaluating

Rev-RRE inhibitors that are themselves fluorescent or quenchers of

fluorescence. These problems can be overcome by using two or more orthogonal

methods when evaluating new ligands, including displacement of ethidium and

(when possible) the direct spectroscopic observations of the small molecule or

nucleic acid. Such “direct binding” studies can elucidate the RNA affinity and

stoichiometry of a ligand, but cannot be used to examine the ability of a ligand to

interfere with the function of the RRE. Rev displacement experiments are

essential, therefore, for evaluating the functional inhibition of the RRE.

60

4.0 New RRE Ligands: [Ru(bpy)2Eilatin]2+

Eilatin (14) is a fused, heptacyclic aromatic alkaloid that was isolated from the

Red Sea tunicate Eudistoma sp.101 It is reported to posses cytotoxic and

antiproliferation activities in a broad range of tissue cultures.102-104 Eilatin is

potentially a bi-facial metal chelator. Upon incorporation into octahedral metal

complexes of the type [Ru(L)2eilatin]2+ (where L= bpy, phen, etc.), only the less

hindered face of eilatin binds to the metal ion (Figure 4.0).105,106 The

enantiomerically pure metal complexes Λ-[Ru(bpy)2eilatin]2+ (9) and ∆-

[Ru(bpy)2eilatin]2+ (10) were synthesized from chiral precursors.106 To confirm

their correct assignment as Λ and ∆, respectively, circular dichroism

spectroscopy was employed (Figure 4.1).

Figure 4.0: Eilatin and eilatin-containing metal complexes. Unlike “free” eilatin(14), the chloride salts of 9, 10, 11, and 15 are freely soluble in water. 15 is asynthetic pre-cursor of rac-9/10 and provides an analog where the eilatin moietyis non-planar.107

Eilatin (14)

NN

NN

2+

Ru

N

N

N

N

N

N

N

N

∆-[Ru(bpy)2Eilatin]2+ (10)Λ-[Ru(bpy)2Eilatin]2+ (9)

rac-[Ru(bpy)3]2+ (11) rac-[Ru(bpy)2Pre-Eilatin]2+ (15)

Ru

N

N

N

N

N

N

N

N

Ru

N

N

N

N

N

N

2+

Ru

N

N

N

N

N

N

N

N

2+2+

61

Figure 4.1: CD spectra of the dichloride salts of Λ-[Ru(bpy)2eilatin]2+ (9) and ∆-[Ru(bpy)2eilatin]2+ (10) in 50 mM sodium phosphate pH 7.5. See section E 4.0.1for further experimental details. The dominant CD transitions (between 260 and300 nm) for 9 and 10 match those predicted by the exciton theory for the correctassignment of the absolute configuration.108

Λ-[Ru(bpy)2eilatin]2+ (9) and ∆-[Ru(bpy)2eilatin]2+ (10) bind to the RRE with

moderate affinity and high specificity (relative to a mixture of tRNAs, see Section

3.2 and reference 81). Fluorescence anisotropy, native gel-shift electrophoresis,

and the solid-phase binding assay all indicate that 9 and 10 bind to the RRE and

displace a Rev protein fragment with approximately 10-times greater activities as

compared to the widely studied RNA ligand neomycin B (Section 3.2 – 3.3). The

ability of 9 and 10 to inhibit Rev-RRE binding implicated their potential use as

anti-HIV agents. To explore this possibility, the anti-HIV activities of 9, 10, 11,

and 15 were measured in cell cultures, using HIV-1 infected CD4+ HeLa cells.

The dose-dependent decrease in plaques (syncytia) is shown in Figure 4.2 and

IC50 values are summarized in Table 4.1. Plaques are multinucleated cellular

Wavelength (nm)

250 300 350 400 450 500 550 600 650 700

[Θ]x

10-3

-300

-200

-100

0

100

200

300

(9)(10)

62

masses that are initiated by HIV-infected cells due to the surface expression of

the ENV protein, which allows CD4+ cells to fuse with infected cells, resulting in

large multinucleated cell masses. Plaque formation may be the primary cause of

T-cell death in vivo.109

Figure 4.2: The anti-HIV activity of compounds 9, 10, 11, and 15 as evidencedby the fractional decrease in plaque formation in CD4+ HeLa cells. Assays wereconducted at the Center for AIDS Research, San Diego. See experimentalsection E 4.0.2 and reference 110 for details. Free eilatin (14) was not evaluatedfor anti-HIV activity due to its poor solubility in DMSO/water and its high toxicityto cell cultures. 9 was also evaluated for anti-HIV activity in human peripheralblood monocytes (PBMC) by measuring the dose-dependent decrease in HIV-1p24 expression using ELISA (as described in E 4.0.3 and reference 110); thisassay also yielded an IC50 = 1 µM.

Compounds 9 and 10 have anti-HIV activities that are from 15 to over 100-fold

greater than those of 11 and 15. This clearly illustrates the significance of the

eilatin moiety to the anti-HIV activities of 9 and 10. To date, only one other family

Concentration (µM)

0.1 1 10 100

Fra

ctio

nalD

ecre

ase

ofP

laqu

es

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

(9)(10)rac-(15)rac-(11)

63

of polypyridine-containing metal complexes has been shown to inhibit HIV

replication.111 These compounds, however, are chemically reactive, create

covalent cross-links to DNA, and are highly toxic.111a In contrast, compounds 9

and 10 are chemically inert, and show no sign of toxicity to HeLa cells (through

100 µM).

Table 4.1: IC50 values (µM) for HIV inhibition and for RevFl displacementexperiments from RRE66 and TAR31 (determined by fluorescence anisotropy).

Compound HIV-1a RRE66b TAR31b

9 0.8 0.9 5.3

10 2.0 1.1 3.5

rac-9/10 0.9 1.0 3.9

rac-15 30 20 30

rac-11 >100 >170c >86c

a Concentration needed to decrease HIV-1 activity by 50% in HeLa plaqueassay, standard deviation is less than ± 40% of reported value.b Concentration needed to displace 50% of RevFl from 100 nM of eachRNA construct, standard deviation is less than ± 20% of reported value.c Only limits can be determined with this method due to fluorescenceinterference from Ru(bpy)3.

To investigate the ability of 9, 10, 11, and 15 to interfere with the function of the

RRE66 and the TAR31, fluorescence anisotropy (Section 3.1) was used to

monitor the ability of each compound to displace RevFl from each RNA. A

summary of RevFl displacement data is given in Table 4.1. Overall, a correlation

exists between the anti-HIV activity of each compound and its affinity for HIV

RNA. Compared to 9 and 10, the non-planar analog [Ru(bpy)2pre-eilatin]2+ (15)

has approximately a 20-fold lower RNA affinity and lower anti-HIV activity (Table

64

4.1). The poor nucleic acid affinity of [Ru(bpy)3]2+ (11) appears to be a general

phenomenon.112 Gel shift electrophoresis confirms that [Ru(bpy)3]2+ shows no

inhibition of the Rev-RRE complex through 10 mM (Figure 3.9). This indicates

that complementary electrostatic interactions are not the primary energetic

driving force for the binding of eilatin-containing metal complexes (9 and 10) to

the RRE. The poor anti-HIV activities and poor HIV RNA affinities of 11 and 14

support a correlation between the nucleic acid affinity and anti-HIV activity of

these octahedral metal complexes (Table 4.1). There are, of course, other

potential mechanisms that explain the anti-HIV activities of 9 and 10. In

conjunction with Immusol Inc. (Sorrento Valley, CA) the mechanism(s) that

govern the anti-HIV activities of 9 and 10 are currently being examined in vivo.

Preferential binding of both 9 and 10 to the RRE66 as compared to the TAR31, is

suggested by the 3-5 fold lower IC50 values observed for RevFl peptide

displacement for RRE66 versus TAR31 (Table 4.1). Compared to TAR31, the

RRE66 has a higher affinity to RevFl (Table 3.0), so it should be “more difficult”

to displace RevFl from the RRE66 than TAR31. Differences in the binding

stoichiometries could, in principle, make the trends in IC50 values inconsistent

with the trends in affinity (Kd). Direct binding studies have been utilized to confirm

the higher affinity of 9 and 10 for the RRE versus TAR (presented later).

To confirm the trends indicated by peptide displacement and to expand the study

to include simple duplex RNAs and DNAs, ethidium bromide displacement

65

assays were conducted using the RRE66, TAR31, and 6 other nucleic acids (see

Section 3.3 and Figure 3.12 for examples of ethidium displacement experiments

using 9, 10, and 15). Table 4.2 summarizes the IC50 values for ethidium

displacement from the RRE66, TAR31, poly r[I] – poly r[C], poly r[A] – poly r[U],

calf thymus (C.T.) DNA, poly d[AT] – poly d[AT], and poly d[GC] – poly d[GC].

Since ethidium has variable binding stoichiometries and a broad range of

affinities to different nucleic acids, the IC50 values in Table 4.2 are not directly

comparable between different nucleic acids. Rather a comparison of each

compound for each nucleic acid is more appropriate (comparisons should be

made vertically, not horizontally).

Table 4.2: IC50 values (µM) for ethidium displacement.a See E 3.3 for experimental details.

Compound RRE66b TAR31bpoly r[I] -poly r[C]c

Poly r[I] -poly r[C]d

poly r[A]-poly r[U]d

C.T.DNAd

polyd[AT]d

polyd[GC]d

9 0.25 1.8 10.5 1.5 2.0 0.1 0.07 0.25

10 0.4 1.6 7.7 1.4 5.0 0.4 0.2 0.4

rac-15 10 9.0 22 4.0 23 7.0 n.d.e n.d.e

a Errors associated with measurement within ± 25% of reported value.b IC50 values measured using 50 nM of RNA construct.c IC50 values measured using 11 µM of duplex base pairs.d IC50 values measured using 0.88 µM of duplex base pairs.e Not determined.

Consistent with RevFl displacement experiments, ethidium displacement

suggests that the RRE66 has a small preferential binding of 9 over 10, while the

TAR31 has a slight preference for the other enantiomer (Table 4.2). Simple

duplex RNAs also exhibit variable selectivity between 9 and 10. Poly r[A] – poly

r[U] shows better binding of 9 over 10, while poly r[I] – poly r[C] shows a small,

66

yet consistent, selectivity for 10 over 9. To our knowledge these are the first

reported examples of octahedral metal complexes that exhibit variable chiral

discrimination between simple duplex RNAs. Consistent with conclusions from

the solid-phase assay (Section 3.2), all three of the simple duplex DNAs exhibit

preferential binding of 9 over 10 (Table 4.2). Interestingly, this is the opposite

enantiomeric selectivity as reported for most other metal complex-DNA

interactions to date.115a-e

Ethidium displacement experiments indicate that, in general, 9 and 10 have

higher affinity to all nucleic acids as compared to 15. The absence of a single

carbon-carbon bond in [Ru(bpy)2pre-eilatin]2+ (15) should impart a fluctuating

dihedral twist averaging 25° between the two tricyclic ring systems in the pre-

eilatin ligand.113 Prior to these studies, it was unknown how this twisting would

affect nucleic acid affinity.114 A crystal structure of rac-(9/10) shows the RMS

deviation of eilatin from planarity (including hydrogen atoms) is 0.130 Å.106

Without hydrogen atoms, the RMS deviation from planarity is 0.098 Å.106 The

deviation of eilatin itself (14) from planarity is 0.044 Å (without H’s).101 Eilatin is,

therefore, nearly a planar ligand. These studies indicate that the planarity of the

eilatin ligand is necessary for the anti-HIV activities and the nucleic acid binding

exhibited by 9 and 10.

67

Octahedral metal complexes can bind to nucleic acids through various modes,

including: intercalation, electrostatic, and groove binding.115a-k Given the apparent

“planarity requirement” for the high-affinity binding of 9 and 10 to nucleic acids,

we investigated the possibility that 9 and 10 bind by intercalation. The viscosity of

a DNA solution was measured with increasing concentrations of rac- 9/10 (Figure

4.3). In accordance with the hypothesis proposed by Lerman,116 the viscosity of a

DNA solution increases upon the addition of an intercalating agent.117 Upon

intercalation of a small molecule, the axial length of DNA increases and it

becomes more rigid. Both factors increase the frictional coefficient, and hence,

the viscosity of the DNA in solution.117

Figure 4.3: Changes in the viscosity of a solution of calf thymus DNA (500 µMb.p.) with increasing concentrations of ligand. The viscosity of a DNA solution isdetermined by measuring the time needed for it to flow through a capillaryviscometer. η is the viscosity of the solution in the presence of a ligand, ηo is theviscosity of the DNA-only solution. Random errors for this experiment areapproximately the size of the symbols used for each point. A low-salt buffer (50mM sodium phosphate pH 7.5) was used. See Section E 4.0.4 for additionalexperimental details.

ratio [ligand / b.p.DNA]

0.0 0.2 0.4 0.6 0.8 1.0

η /η o

0

1

2

3

4

5Ethidium Bromiderac-(9/10)

ratio [ligand / b.p.DNA]

0.0 0.2 0.4 0.6 0.8 1.0

(η/η

ο)1/

3

0.9

1.0

1.1

1.2

1.3

1.4

1.5

1.6

1.7Ethidium Bromiderac-(9/10)

68

Since the change in viscosity should be proportional to the change in the axial

length of a rigid DNA molecule to the third power, viscosity results are often

reported as (η/ηo)1/3 versus molar fraction of the ligand (Figure 4.3 Right).117

Theory predicts that for an “ideal” intercalating agent, this plot will be linear with a

slope of 1.0.117 Interestingly, rac-(9/10) has a slope of 1.0 and ethidium has a

slope of 0.7. Other groups have also observed a slope of less than 1.0 for

ethidium.115i At low ionic strengths, ethidium can bind to DNA by both

intercalation and by surface-binding via electrostatic interactions.86 Compounds

that bind to duplexes through electrostatic interactions, such as tobramycin (3),

can decrease the viscosity of the nucleic acid solution.118 This may explain why,

under these conditions, viscometric analysis of the ethidium-DNA complex yields

a slope significantly lower than 1.0 (Figure 4.3 Right).

Other interesting differences between rac-9/10 versus ethidium (12) are also

apparent in the viscosity titrations (Figure 4.3). The ratio at which the DNA

becomes saturated is significantly different for each compound. The curve for

ethidium reaches saturation at ~3 ethidium molecules / 10 base pairs. This value

is consistent with the reported binding stoichiometry of ethidium to C.T. DNA.87

This is also the same binding stoichiometry reported by Chaires for

[Ru(phen)2DPPZ]2+ (a metal complex very similar to 9 and 10).115c For rac-9/10,

however, saturation is reached at much higher ratios (Figure 4.5), suggesting a

maximum binding stoichiometry of ~ 6 equivalents of ligand for every 10 base-

pairs of DNA. This is an unusually high binding stoichiometry for such a large

69

metal complex! Only one other example of such a high binding stoichiometry for

a metal complex is found in the literature.119 Barton and Lippard report ~6.7

equivalents of [(terpy)Pt(HET)]1+ per every 10 base-pairs of poly r(A) – poly

r(U).119 This metal complex is, however, square planar, and the intercalation of

this compound into duplexes should be much less sterically hindered as

compared to 9 and 10.

To better examine the nucleic acid sequence and structure specificity of 9 and

10, direct binding studies have been conducted by monitoring the UV

absorbance spectrum of either 9 or 10 as a function of nucleic acid concentration

(Figure 4.4). Compounds 9 and 10 exhibit multiple electronic transitions in their

absorbance spectra that are sensitive to nucleic acid binding. The presence of

multiple isosbestic points, and saturation at high DNA concentrations, suggests a

simple two-state transition exists between the “bound” versus “free” states of the

ligand upon titration of nucleic acid (Figure 4.4). The concentration of nucleic acid

needed to bind ½ of each compound (defined as the C50 value) is determined by

assuming a linear relationship between the spectral changes and the fraction of

ligand bound. A lower C50 value suggests a higher binding affinity and/or more

binding sites.

Titrations were conducted using many different nucleic acids by observing the

UV-vis absorption changes of either 9, 10, or 12 as a function of nucleic acid

concentration (Table 4.3). For all duplex nucleic acids tested, the UV-vis

70

absorbance spectra of 9 and 10 show changes similar to those observed upon

titration of Calf Thymus (CT) DNA (Figure 4.4). The absolute change in intensity

and degree of red-shifting are slightly different for each nucleic acid tested, but all

titrations exhibit simple two state transitions (see section E 4.0.5 for the “raw”

spectral data of 9 and 10 upon titration of each nucleic acid). Each nucleic acid

endows unique spectral changes to 9 and 10, but no obvious correlation is found

between the composition of the nucleic acid and the spectral changes exhibited

by ligand (see section E 4.0.5).

Figure 4.4: Uv-vis absorbance spectrum of 10 (8 µM) as a function of calfthymus DNA (concentration in base-pairs). See E 4.0.5 for additionalexperimental details and the raw data for all polymeric nucleic acids tested.

Consistent with the results from the solid-phase assay (Section 3.3), thermal

denaturation (Section 3.5), and ethidium bromide displacement experiments

(Table 4.2), the direct titration experiments indicate that 9 has a higher affinity for

CT DNA as compared to 10 (Table 4.3). Interestingly, both 9 and 10 exhibit

300 350 400 450 500 550 600 650 700 7500.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

0.22∆-Eilatin Ru (bpy)2

+2 µM C.T. DNA+4 µM C.T. DNA+10 µM C.T. DNA+20 µM C.T. DNA+40 µM C.T. DNA+60 µM C.T. DNA+100µM C.T. DNA

Wavelength (nm)

Abs

orba

nce

71

sequence selectivity in the binding double-stranded DNA and RNA. For example,

10 has over a 5-fold higher affinity for poly d(G) – d(C) as compared to poly

d(GC) – d(GC), and both 9 and 10 have a higher affinity to poly d(AT) – d(AT) as

compared to poly d(A) – d(T) (Table 4.3).

The classic intercalating agent ethidium bromide (12) has been evaluated for

nucleic acid selectivity by monitoring its changes in UV-vis absorption spectrum,

and the eilatin “free ligand” (14) has been evaluated using fluorescence emission

spectroscopy (presented below). All four compounds have a much higher affinity

to poly r(A) – r(U) as compared to both poly r(I) – r(C), and r(G) – r(C). This may

indicate that some duplex nucleic acids have, in general, a low affinity for “any”

intercalating agent. Some nucleic acids, on the other hand, are highly sensitive to

the identity of the intercalating agent. For example, poly d(A) – d(T), has a very

low affinity for ethidium (12), a modest affinity for the metal complexes (9 and

10), and a very high affinity for the free eilatin ligand (14) (Table 4.3) All four

compounds exhibit a moderate selectivity for duplex DNA over duplex RNA.

Each compound, for example, has approximately a 10-fold higher affinity for poly

d(G) – d(C) versus poly r(G) – r(C) (Table 4.3). We conclude that, for the most

part, the trends in the nucleic acid affinity for 9, 10, and 12 are similar for all the

duplex nucleic acids tested. We interpret this as additional evidence that 9 and

10 bind to duplex nucleic acids via intercalation.

72

Table 4.3: C50 values (µM)* for 8 µM of Λ−[Ru(bpy)2eilatin]2+ (9), 8 µM∆−[Ru(bpy)2eilatin]2+ (10), 8 µM ethidium bromide (12), and 0.1 µM eilatin (14).

Nucleic Acid Λ−[Ru(bpy)2eilatin]2+

(9)∆−[Ru(bpy)2eilatin]2+

(10)ethidium bromide

(12)**eilatin(14)***

C.T. DNA 18 27 20 3

Poly d(A) –Poly d(T)

13 23 105 1

Poly d(AT) –Poly d(AT)

9 8 13 5

Poly d(G) –Poly d(C)

15 13 22 5

Poly d(GC) –Poly d(GC)

18 55 16 8

Poly r(A) –Poly r(U) 13 22 18 12

Poly r(G) –Poly r(C)

125 85 400 20

Poly r(I) –Poly r(C)

200 180 170 150

Poly r(U)8 8 4000 >2000

Poly r(C)52 75 8000 >1000

Poly r(A)14 13 1500 70

Poly r(G)18 20 240 4

Poly r(I)13 10 950 0.2

Oligo r(I)15 68 53 n.d. n.d.

Poly d(A) 20 ~200 n.d. n.d.

Oligo d(A)15 ~200 70 n.d. n.d.

Oligo d(T)15 21 21 n.d. n.d

* C50 values for duplexes are reported per base pair, the values for single-stranded polymers are reported per base. The approximate experimentaldeviation for all values is less than or equal to ±35% of the reported C50 value.** Changes in UV-vis absorption monitored for duplex nucleic acids, and changesin emission monitored for single-stranded nucleic acids.*** Compared to 9, 10, 12, a much lower concentration of 14 was used forspectroscopic measurements (8.0 µM versus 0.1µM, respectively), the C50

values for 14, cannot, therefore, be directly compared to those of 9, 10, and 12.

73

The eilatin-containing metal complexes 9 and 10 do, however, demonstrate

some interesting differences when compared to ethidium bromide (12); the most

surprising revelation is their high affinity for single-stranded nucleic acids (Table

4.3). Classic intercalating agents, like ethidium bromide, exhibit a relatively low

affinity for single-stranded RNAs and DNAs. Ethidium has a 1,000-fold lower

affinity for most single-stranded nucleic acids as compared to duplex nucleic

acids (Table 4.3). 9 and 10, however, exhibit a higher apparent affinity for both

poly r(U), and poly r(I) as compared to most duplex nucleic acids. In terms of

RNA selectivity, 9 and 10 have a higher apparent affinity for the single-stranded

versus double-stranded polymeric RNA. Since most natural RNAs, including the

RRE, have bulged regions and other features with single-strand characteristics, it

is possible that the single-stranded regions bind, with high affinity, to 9 and 10.

Differences in binding stoichiometry could, however, make the C50 values in

Table 4.3 disproportionate to the actual affinities (Kd). To determine the single

stranded binding stoichiometeries of 9 and 10, the UV-vis absorbance spectrum

of a high concentration of 10 (40 µM) was monitored upon titration poly r(U)

(Figure 4.5). Interestingly, a very high stoichiometry is indicated, where

approximately 5 equivalents of 9 or 10 are bound per every 10 bases of poly r(U).

This is very similar to the stoichiometry observed for CT DNA, where 6

equivalents of rac-9/10 bind for every 10 base pairs (Figure 4.5). Another duplex

DNA, poly d(AT) – poly d(AT), also exhibits the same binding stoichiometry for

both 9 and 10 (at 5 equivalents per every 10 base pairs) (not shown).

74

Figure 4.5: At high concentrations of 10 (40 µM), the binding isotherm for polyr(U) is composed of two linear elements, the intersection of these lines (at 80 µMof poly r(U)) indicates the maximum binding stoichiometry for 10 is 1 equivalentfor every 2 bases of poly r(U). The other enantiomer 9 yields the same result.The same experiment was conducted with duplex DNA and also yielded abinding stoichiometry of 1 equivalent of either 9 or 10 for every 2 base pairs ofpoly d(AT) – d(AT).

Titrations using poly r(U) were also conducted with either 4 µM or 8 µM of 10

(Figure 4.6). At 4 µM and 8 µM of 10, the binding data fit very well to Hill

coefficients of n =2 (Figure 4.6). This suggests positive cooperativity for these

binding interactions (Section 3.6). Duplex nucleic acids can also show positive

cooperativity in their binding of 9 and 10. For example, both poly d(AT) – d(AT)

and poly d(A) – d(T) have Hill coefficients of n = 2 for binding to both 9 and 10.

Indeed, all of the “higher affinity” interactions (C50 < 30 µM, Table 4.3) show some

degree of positive cooperativity in their binding of 9 and 10 (Hill coefficients range

from 1.5 to 2.2).

0 100 200 300 400-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Concentration of poly r(U) (ntds.)

Fra

ctio

nof

10bo

und

75

Figure 4.6: Association of poly r(U) with 10 at three different concentrations. At 4µM and 8 µM a good fit is found for a Hill coefficient of n = 2. As expected, a poorfit is found at 40 µM.

To further investigate the high affinity and cooperative binding of 9 and 10 to

single-stranded poly r(U), thermal denaturation experiments were conducted

(Figure 4.7). By themselves, the UV adsorption of 9, 10, and poly r(U) (260 nm)

show only small non-cooperative changes as a function of temperature (Figure

4.7). The complex formed between 10 and poly r(U), however, shows a highly

cooperative melting behavior reminiscent of duplex denaturation (Figure 4.7 solid

line). The complex formed between 9 and poly r(U), however, shows multiple,

non-reversible transitions (Figure 4.7 dotted line). It is possible that under these

conditions, some non-specific aggregation between 9 and poly r(U) occurs. The

complex between 10 and poly r(U), however, appears well-behaved (Figure 4.7).

Concentration of poly r(U) (µM ntds.)

0.1 1 10 100 1000

Fra

ctio

nof

10B

ound

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.24 µM 108 µM 1040 µM 10

76

Figure 4.7: Change in adsorption of poly r(U) (30 µM of bases) with or without, 9or 10 (15 µM). See section E 4.0.6 for experimental details.

The thermodynamic parameters of the melting transition (70 – 90 °C) for the poly

r(U)-10 complex can be fit to a van’t Hoff plot (Figure 4.8). This analysis assumes

that the melting of the complex represents a simple two-state transition. Based

upon their known binding stoichiometry (Figure 4.5), a 2:1 ratio of poly r(U) bases

to 10 was used for the Tm experiment so that excess ligand would not drive

complex formation. From the van’t Hoff plot (Figure 4.8), a large enthalpy of

formation (∆H0 = -84 kcal / mole) and a modest entropy of formation (∆S0 = -240

cal / mole K-1) are calculated. These values represent the energetic parameters

for the complex as a whole and do not reflect the binding of individual molecules

of 10 to poly r(U). These ∆H0 to ∆S0 values are similar to those for the melting of

Temperature (oC)

20 30 40 50 60 70 80 90 100

Cha

nge

inad

sorp

tion

(260

nm)

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.209 only10 onlypoly r(U) only9 + poly r(U)10 + poly r(U)

77

long polymeric nucleic acids. For example, CT DNA, under these same

conditions has enthalpy of formation (∆H0 = -82 kcal / mole) and a entropy of

formation (∆S0 = -280 kcal / mole K-1) (data not shown). Compared to the poly

r(U)-10 complex and to CT DNA, shorter duplex nucleic acids have smaller ∆S0

terms and, therefore, exhibit less cooperative melting transitions. 120

Figure 4.8: van’t Hoff plot of the thermal melting of the poly r(U)-10 complex. Keq

is taken as the fraction of folded complex relative to unfolded complex. ∆H0 = -(R*slope) and ∆S0 = ∆H0 / Tm (°K).

The melting transition of the poly r(U)-10 complex provides additional evidence

that the binding interaction between 10 and poly r(U) is highly cooperative and

large binding energies (∆Gcomplex = 8 – 12 kcal / mole) are involved. This raises

the possibility that the poly r(U)-10 complex is a well-ordered supermolecular

structure suitable for crystallization and subsequent analysis by X-ray

crystallography.

0.0027 0.0028 0.0029 0.0030 0.0031

-5.0

-2.5

0.0

2.5

5.0

ln(K

eq)

1 / T (K-1)

78

Interestingly, the non-polymeric nucleic acids (RRE JW, TAR 31, and tRNAPhe)

do not show cooperativity in their binding of 9 and 10 (Hill coefficients range from

1.1 – 1.4). Scatchard plots can, therefore, be used to establish the average

affinity and stoichiometry of 9 and 10 for binding these RNA transcripts (Figure

4.9 and Table 4.4).

Figure 4.9: Scatchard analysis of the interaction between 9 and the RRE JW,tRNAPhe, and TAR31. The slope of the graph = 1/Kd (ave), and the r value at r / Cf= 0 is the number of apparent binding sites.

Table 4.4: Summary of the binding stoichiometries and average affinity of 9 and10 for RRE JW, tRNAPhe and TAR 31.Nucleic Acid Binding sites

for 9Average Kd

(µM) for 9Binding sites

for 10Average Kd

(µM) for 10RRE JW 2.4 1.9 3.8 1

tRNAPhe 9 5.6 10 5.6

TAR 31 3 4.6 4 9

Non-whole numbers for the binding stoichiometry are obtained for some of the

interactions (Table 4.4). This suggests that some RNAs have multiple,

nonequivalent binding sites, and therefore, the reported affinity (Kd) represents

0 1 2 3 4 5 6 7 8 9

250000

500000

750000

1000000

1250000

1500000

RRE JWtRNAPheTAR31

r

r/

Cf

79

the average affinity to all of the “active” binding sites. Consistent with the Rev

peptide and ethidium displacement experiments as well as the solid-phase

assay, 9 and 10 have a higher affinity for the RRE as compared to both TAR31

and tRNAPhe. Unfortunately, due to the limited sensitivity of monitoring changes in

UV-vis absorption, it is difficult to probe the highest affinity binding site(s) on the

RRE. Since 9 and 10 are non-emissive, low µM concentrations are needed to

observe the spectral changes of these compounds. For example, even if the

RRE possesses a 10 nM binding site, it cannot be characterized under these

conditions since the metal complex is at a much higher concentration compared

to the affinity. This limitation is also reflected in some of the C50 values reported

in Table 4.3. Since the concentration of 9 and 10 used for these titrations is 8 µM,

and the binding stoichiometry is approximately 1 ligand per every 2 bases (or

every 2 base-pairs), the theoretical minimum C50 value for all polymeric nucleic

acids is 8 µM (Table 4.3). We can only report, therefore, that the interaction

between poly r(U) and both 9 and 10 exhibit affinities of Kd < 0.5 µM (Table 4.5).

Given the unusual nucleic acid selectivities of 9 and 10, we decided to

investigate the nucleic acid affinity and specificity of “free” eilatin (14). Unlike the

eilatin-containing metal complexes (9 and 10), eilatin itself (14) is highly emissive

and shows changes in its emission spectrum upon addition of nucleic acids

(Figure 4.11). For all nucleic acids tested, a decrease in emission intensity is

observed, allowing for C50 values to be established (Figure 4.10 and Table 4.3).

80

Figure 4.10: Changes in the fluorescence emission spectrum of 0.1 µM eilatin(14) upon addition of CT DNA. Nearly identical changes are seen for all othernucleic acids tested. Excitation is at 417 nm, see Section E 4.0.7 for additionalexperimental details.

The C50 values measured for eilatin (14) are not directly comparable with those

determined for 9, 10, and 12 (Table 4.3). Due to detection limits of 9 and 10, and

the limited water solubility of 14, the concentration of eilatin (14) used for the

fluorescence experiments is 80-fold lower than the concentrations needed for the

UV-vis titrations of 9, 10, and 12 (Table 4.3). Binding constants for all of the

compounds can be estimated from the C50 values (summarized in Table 4.5). For

eilatin (14), the Kd ≈ (C50 / 2)-0.05 (this assumes that binding is non-cooperative

and that the neighbor exclusion limit of 14 is one molecule per 2 base pairs). Kd

values have also been estimated for the eilatin-containing metal complexes (9

and 10), and for ethidium bromide (12) (Table 4.5). All estimates are equal to the

C50 value divided by the binding stoichiometry, minus half the concentration of

the ligand. Issues related to the cooperative binding of 9 and 10 do, however,

Wavelength (nm)

450 500 550 600

Inte

nsity

(arb

.)

0

100

200

300

400100 nM Eilatin (14)1 µM CT DNA3 µM CT DNA7 µM CT DNA15 µM CT DNA30 µM CT DNA60 µM CT DNA120 µM CT DNA240 µM CT DNA480 µM CT DNA

81

complicate this estimate. Nonetheless, these Kd estimates allow for a more direct

comparison of the binding affinities of all four compounds (Table 4.5).

Unlike the eilatin-containing metal complexes (9 and 10), eilatin (14) is not very

soluble in water (maximum concentration of 14 at pH 7.5 = 3 µM). Limited water

solubility prevents Rev peptide and ethidium displacement experiments from

reaching saturation. The IC50 value of 14 for Rev-RRE inhibition, for example, is

>3 µM (Figure 4.11). This represents, however, at least a 3-fold lower Rev-RRE

inhibition activity as compared to 9 and 10. Eilatin (14) does, however, bind to the

RRE with high affinity (Figure 4.11). Fluorescence binding studies indicate that

eilatin (14) binds to one or more sites on the RREJW with a Kd of approximately

0.5 µM, but cannot displace Rev by binding to these site(s).

Figure 4.11: A comparisons between the “direct” binding of eilatin (14) to theRRE JW and its ability to inhibit Rev-RRE66 binding. For the direct bindingexperiment, the fluorescence intensity of eilatin was monitored as a function ofRRE JW concentration. The direct binding data is fit to a C50 value of 0.5 µM withno cooperativity (Hill coefficient = 1.0). For the RevFl displacement experiment,fluorescence anisotropy was used to monitor the fraction of RevFl bound to theRRE66 (10 nM of each).

Concentration (µM)

10-3 10-2 10-1 100 101 102 103 104 105

Frac

tiona

lCha

nge

0.0

0.2

0.4

0.6

0.8

1.0

Direct Binding to RRERevFl Displacement

82

Table 4.5: Approximate Kd values (µM)* of Λ−[Ru(bpy)2eilatin]2+ (9),∆[Ru(bpy)2eilatin]2+ (10), ethidium bromide (12), and eilatin (14) as calculatedfrom C50 values (Table 4.3).*

Nucleic Acid Λ−[Ru(bpy)2eilatin]2+

(9)∆−[Ru(bpy)2eilatin]2+

(10)ethidium bromide

(12)eilatin(14)

C.T. DNA 5 9.5 2.7 1.5

Poly d(A) –Poly d(T)

2.5 7.5 31 0.5

Poly d(AT) –Poly d(AT)

0.5 < 0.5 0.33 2.5

Poly d(G) –Poly d(C)

3.5 2.5 3.3 2.5

Poly d(GC) –Poly d(GC)

5 24 1.3 4

Poly r(A) –Poly r(U) 2.5 7 2 6

Poly r(G) –Poly r(C)

59 39 130 10

Poly r(I) –Poly r(C)

96 86 53 75

Poly r(U)< 0.5 < 0.5 1,300 > 1,000

Poly r(C)22 34 2,700 > 500

Poly r(A)3 2.5 500 35

Poly r(G) 5 6 76 2

Poly r(I)2.5 1 310 0.1

Oligo r(I)15 30 23 n.d. n.d.

Poly d(A)6 96 n.d. n.d.

Oligo d(A)15 96 31 n.d. n.d.

Oligo d(T)15 6.5 6.5 n.d. n.d

* The Kd values for 14 are estimated to = ((C50 / 2)-0.05). Consistent with earlierstudies, the Kd values for 12 are (Kd ≅ ((C50 / 3) – 4)).87 For 9 and 10, the Kd

values are approximately ≅ ((C50 / 2) – 4). These values serve as estimates only,as differences in binding stoichiometries and cooperativity will affect thecalculated Kd values.

83

Tables 4.4 and 4.5 represent, to the best of our knowledge, the most

comprehensive study of ethidium bromide and its affinity to different nucleic acids

to date. Bresloff and Crothers used equilibrium dialysis to study the binding of

ethidium to 12 different nucleic acids, five of which appear in Table 4.5.87 These

studies were conducted at 1M NaNO3, and our studies are conducted at

physiological ionic strength (see section E 4.0.2 for experimental conditions).

Despite the different conditions, we report similar trends in ethidium affinity for all

five nucleic acids.87

Tables 4.4 and 4.5 highlight how the nucleic acid specificity of an intercalating

ligand can change upon incorporation into an octahedral metal complex. For

most, but not all duplex nucleic acids, free eilatin (14) has a slightly higher DNA

and RNA affinity compared to 9 and 10 (Table 4.5). A notable exception is poly

d(AT) – poly d(AT). The eilatin-containing metal complexes (9 and 10) have a

much higher affinity for poly d(AT) – poly d(AT) as compared to poly d(A) – poly

d(T), while eilatin (14) has the opposite selectivity (Table 4.5).

The trends observed for single-stranded versus double-stranded nucleic acids

are strikingly different for 9, 10, 11, and 14 (Table 4.5). Ethidium (12) is a

“classic” intercalating agent and, hence, has a low affinity for all the single-

stranded nucleic acids evaluated (Table 4.5). Molecules with extended aromatic

surfaces can, however, show appreciable affinity for single-stranded nucleic

acids.121,122 Both eilatin (14) and the eilatin metal complexes bind to single-

84

stranded nucleic acids, but with very different trends in affinity. Eilatin (14)

exhibits a high degree of differentiation among the three homopurine single-

stranded nucleic acids (poly r(I) > poly r(G) > poly r(A)). This suggests a

preference for electron-poor purines. Both eilatin (14) and ethidium (12) have a

much lower affinity for the homopyrimidines (poly r(U) and poly r(C)) as

compared to the homopurines (poly r(I), poly r(G), poly r(A)). In contrast, no clear

preference of purines versus pyrimidines is seen for 9 and 10 (Table 4.5).

The eilatin metal complexes (9 and 10) have, in general, a much higher affinity to

single-stranded nucleic acids as compared to both eilatin (14) and ethidium (12).

Compounds 9 and 10 bind to poly r(U) with sub-µM affinities. Only one other

small molecule, to our knowledge, has been reported to possess high affinity for

poly r(U).121 Given the preferential binding of 9 and 10 to single-stranded nucleic

acids, we examined the possibility that the eilatin-containing metal complexes

recognize the internal bulge of the RRE (Figure 2.4). This bulge is known to

possess single-stranded characteristics in the absence of Rev binding.123 We

have conducted titrations using an inosine-substituted RRE JW (Figure 2.5). This

mutant RRE binds to the RevFl peptide with the approximately same affinity (2

nM), but it has over a 2-fold higher affinity for 9 as compared to the RRE JW.

This is consistent with these metal complexes exhibiting a higher affinity for

single-stranded r(I) versus r(G) (Table 4.5), and suggests that the internal bulge

of the RRE is the preferred binding site of these complexes.

85

To the best of our knowledge, only one other group has examined the nucleic

acid affinity of eilatin (14), and they reported very different results as compared to

ours.104 Ethidium bromide displacement experiments were conducted using CT

DNA, but only a low DNA affinity was suggested (IC50 > 100 µM).104 The limited

solubility of eilatin (solubility in water < 3 µM) may explain this result. The same

report also used fluorescence experiments to monitor the change in “emission” of

a solution of 14 as a function of CT DNA concentration.104 They showed an

increase in the fluorescence intensity of eilatin upon addition of CT DNA. We find

a decrease for all nucleic acids tested (Figure 4.10). In their case, however, the

sample was both excited and monitored at 520 nm.104 Eilatin shows no

absorption at this wavelength.101 Since the sample was excited and monitored at

the same wavelength, the increase in “emission” was actually the increased light

scattering by the solution upon titration of DNA.104

Eilatin-containing metal complexes are new anti-HIV agents with unusual nucleic

acid specificities. The eilatin ligand is essential for nucleic acid binding and the

anti-HIV activity of these compounds, but the free ligand exhibits drastically

different trends in nucleic acid affinity as compared to the eilatin-containing metal

complexes. Unusual trends in enantiomeric selectivity, single-stranded nucleic

acid binding, and sequence specificity make the eilatin-containing metal

complexes two of the most unusual, and best characterized, RRE ligands to

date.

86

5.0 Ethidium Bromide: Biological Activities and Nucleic Acid Binding

Ethidium bromide is the common name for 3,8-diamino-5-ethyl-6-

phenylphenanthridinium bromide (12). First reported in 1952, it was developed by

the Boots company as an anti-trypanosomal agent.124 Homidium bromide

(ethidium) is prepared in large quantities by alkylation of 3,8-dinitro-6-phenyl

phenanthridine with ethyl-p-toluene sulfonate, followed with reduction by Fe, and

precipitation using ammonium bromide.125 Ethidium is a common stain for

double-stranded DNA and RNA.126 It is also reported to possess significant anti-

cancer,127 and anti-parasitic activities.124 Its potential applications in human

treatment have been prevented, however, due to its mutagenic and carcinogenic

activities in model systems.128-129 For example, Frog larvae (Xenpous laevis)

exposed to near toxic concentrations of ethidium develop gross malformations of

all major organ systems.128 Upon metabolic activation with liver homogenate,

ethidium causes frame shift mutations in the Ames assay.129 From these, and

similar studies, it has been concluded that long-term exposure to ethidium may

cause inheritable genetic damage and/or cancer in humans. Despite this,

ethidium is still marketed by Laprovet as a safe and inexpensive treatment for

cattle suffering from trypanosomosis.130

A possible relationship between ethidium’s nucleic acid binding and its biological

activities has been examined both in vivo and in vitro. Experiments by Nass

indicated that the growth of both mouse fibroblasts and hamster kidney cells is

inhibited by 0.3 – 13 µM of ethidium, and that mitochondrial, not nuclear, DNA

87

synthesis is inhibited by ethidium.131 Other studies showed that ethidium

accumulates in isolated rat mitochodrion and interferes with normal respiration.132

Ethidium is, however, only moderately toxic to mammals, with an LD50 in mice of

~300 µM (100 mg/kg, subcutaneous);133 and it is an effective trypanocide in

cattle at ~ 3 µM (1mg/kg, intravenous).124,130

The in vitro study of ethidium-nucleic acid binding can be conducted by

monitoring the photophysical changes of ethidium upon addition of nucleic

acids.134-135 Ethidium binds to DNA and RNA duplexes with moderate affinity (Kd

ranges from 0.2 – 200 µM) and variable stoichiometry (3 – 6 equivalents of

ethidium per helical repeat) depending on the sequence of the duplex.87 It was

proposed that ethidium binds to nucleic acids via two distinct modes: electrostatic

surface binding (at low ionic strengths) and intercalation (at higher,

physiologically relevant ionic strengths).135 Crystal structures and NMR studies of

tRNA bound to ethidium have confirmed its ability to both intercalate and to bind

the surface of nucleic acids.136-138 Ethidium is reported to bind to duplexes that

contain bulged bases with a higher affinity as compared to the same duplexes

without a bulge.139 This could potentially explain ethidium’s high affinity for the

RRE. Ethidium has two high-affinity binding sites on the RRE (with an average Kd

= 0.2 µM), and it displaces the Rev peptide with similar activity (Section 3.4 and

Figure 3.14). In addition, we have found ethidium to be a potent inhibitor of HIV-1

replication. Using the plaque-formation assay described in figure 4.1, an IC50

value of 0.2 µM is measured for ethidium. This activity is only ~20-fold lower than

88

the well-know HIV inhibitor AZT (see E 1.1 for a summary of anti-HIV IC50

values). The HeLa cells used in these assays, however, detach from their plates

after a three-day exposure to 3 – 10 µM of ethidium bromide. This is indicative of

toxicity and is not observed for the other compounds evaluated in this thesis. In

an attempt to simultaneously decrease its toxic/mutagentic properties while

increasing its anti-HIV activity, a number of ethidium derivatives have been

synthesized and are currently being evaluated for nucleic acid affinity and

biological activities (Figure 5.1).

To date, only a small number of ethidium derivatives have been reported.125,140-

146 Early modifications included variation of the alkyl chain with groups other than

ethyl (methyl, propyl, etc).125 Compared to ethidium, these derivatives have the

same DNA affinity,140 but are significantly more toxic.124 The 6-phenyl ring of

ethidium has been substituted with other groups (hydrogen, 4-amino phenyl, 4-

nitro phenyl, methyl, napthyl, etc).141 Again, similar DNA affinities were measured

for these derivatives.140 Ethidium’s exocyclic amines have been converted to

azido (N3),142 leading to highly reactive photo-crosslinking agents.143 These

compounds are also reported to have similar DNA affinity as compared to

ethidium,142 and are highly mutagenic.144 Amino acids have been conjugated to

ethidium through its exocyclic amines and through its phenyl ring, but the nucleic

acid affinities of these derivatives have not been reported.145,146

89

Figure 5.1. New ethidium derivatives and a summary of the wavelengths ofmaximum adsorption (λ-abs), emission maximum (λ-em), and the approximatequantum efficiency (emission intensity) relative to ethidium (φrel). All valuesdetermined at 10 µM of each compound in 50 mM sodium phosphate (pH = 7.5).

N

R2R1

R1 R2

N

N N

X-

H2N

O

NH

NH2

O

NH

H2N

O

NH

NH2

O

NH

H2N

NH2

H2N

Number Compound

N NH2

NH2H2N

NN

H2N

NH

NH

NH2

NH

NH

H2N

NH

NH

NH2

NH

NH

NH2

H2N

O Ph

O

HN

OPh

O

HN

(25)

8-pyrole ethidium

3,8-bis pyrole ethidium

3-pyrole ethidium

3-urea ethidium

8-urea ethidium

3,8-bis urea ethidium

λ-abs (nm)

(26)

(27)

(22)

(23)

(24) 434

455

466

(12) ethidium 486

(28) tetramethyl ethidium 534

3-guanidino ethidium

8-guanidino ethidium

3,8-bis guanidino ethidium

(19)

(20)

(21)

443

455

398

428

458

464

OPh

O

HN

NH2(16)

(17)

(18)

3-cbz ethidium

8-cbz ethidium

3,8-bis cbz ethidium

468

473

430

+

O Ph

O

HNH2N

φrel

57

0.8

0.09

1.0

0.08

0.5

0.08

3.1

1.2

1.0

0.2

1.0

0.2

57

λ-em (nm)

521

594

607

606

643

593

606

504

504

589

603

589

603

510

90

Surprisingly, the exocyclic amines of ethidium have not, until now, been

systematically substituted with other functional groups. Modification of one, or

both, of these amines provides a “modular” approach for introducing new

chemical diversity into ethidium (Figure 5.1). These modifications are found to

dramatically affect the electronic structure, nucleic acid affinity, and the biological

activities of the resulting derivatives.

5.1 Electronic Properties of Ethidium and its Derivatives

The charge distribution in ethidium is believed to be important for its high affinity

intercalation into nucleic acids.142, 147 Significant efforts have been invested in

constructing electrostatic models that can simulate the stacking interactions of

ethidium with base pairs.148-150 These models are based upon quantum

mechanical calculations and have not, to date, been able to predict or rationalize

the sequence specificity of ethidium bromide (Table 4.5).

π-stacking interactions are sometimes modeled as having partial covalent

character, where the interactions between the LUMO of the intercalator and the

HOMO base-pair are used to predict stacking energies.150 The LUMO of the

intercalator is believed, by some, to be electrostatically relevant, as the ability to

accept electrons should be related to the electron density at each position.149

Some groups have interpreted the spectral changes of ethidium, upon binding

nucleic acids, as having some of the characteristics of a charge transfer complex;

and they have suggested that HOMO-LUMO stabilization may be involved.150

91

This would provide evidence that Mulliken theory is applicable for evaluating the

aromatic stacking interactions in ethidium-base pair complexes.151

The theoretical electrostatic potential around an intercalator or a base pair can be

calculated by evaluating the interactions between individual atoms upon fitting

the molecule in an electrostatic force field.149,152,153 These calculations are useful

for examining the potential electronic charge of each atom, but do not reflect the

actual electrostatic field emanating from the molecule itself. Consequently, the

polarizability and the induced dipoles of the base stacking interaction are

weighted much more heavily than any complementation between the

electrostatic field of the intercalator with the base pair. As a result of these, and

similar theoretical studies, some groups believe that dispersion forces (induced

dipole-dipole interactions) dominate nucleic acid intercalation and have

concluded: “there is no need to consider any out of plane πcharges (“sandwich”

model) to account for aromatic intercalator-base pair stacking”.149

Dispersion forces depend, mostly, on the total shared surface area of the

interaction and on the polarizability of each surface. These effects may, in fact,

be a dominant driving force for intercalation.149 They cannot, however, explain

the sequence specificity of ethidium bromide. Crystal structures of ethidium

bound by different dinucleotides, show that about the same surface contacts are

made between ethidium and different intercalation sites.138 In addition, the larger,

92

more polarizable base pairs (G-C) do not exhibit the highest binding affinity for

ethidium (Table 4.5).

Ethidium binds to different duplex nucleic acids with a free energy range of 5.3 –

8.8 kcal/mole, depending on the structure and sequence of the nucleic acid

(Table 4.5). It is possible that this sequence selectivity is related to electrostatic

complementation between the unique distribution of π-electron charge density of

ethidium’s phenanthridinium “core” with the different electronic features of each

intercalation site. Electrostatic complementation is known to be an important

component of π-stacking interactions.151 For phenyl-phenyl stacking, electron-

withdrawing groups lower the Coulombic repulsion between the πelectrons of

each ring. 154,155 Coulombic attraction is often observed when an electron-rich

ring and an electron-poor ring are stacked face-to-face.151,156 This type of

electrostatic attraction may involve interactions between the opposite

quadrupolar moments of each ring, and, unlike charge transfer complexes, does

not necessarily involve direct orbital-orbital overlap between each of the π

systems.156 To date, most of the systems that have been used to examine π-

stacking interactions involve molecules that are formally neutral.151,153 Ethidium

has a formal charge of +1, so Coulombic interactions, not charge transfer, may

dominate its π-stacking interactions. We believe that current modeling efforts,

and hence the rational design of new intercalating agents, can be significantly

improved by a more detailed understanding of ethidium’s electronic structure.

93

We have accumulated data using crystallography, NMR, UV-vis absorption, and

fluorescence spectroscopy that have allowed us to probe the electrostatic

features of ethidium’s ground-state. In contrast to previous theoretical

models,149,156 we find the amine at ethidium’s 3 position (Figure 5.1

(-R1)) is significantly more electropositive than the amine at the 8 position (Figure

5.1 (-R2)). In addition, an unprecedented magnitude of charge separation in the

ground-state of ethidium is revealed. Using X-ray crystallography, we have

determined the bond order, and hence positive charge on each of ethidium’s

nitrogen atoms. We have related these partial formal charges to the π-electron

charge densities on each of ethidium’s “core” carbon atoms by using 13C NMR.

These results have been used to generate the first experimentally determined

electrostatic map of an intercalating agent.

The photophysical properties of ethidium are dramatically altered upon

modification of its exocyclic amines (Figure 5.1). Ethidium has an internal charge-

transfer band that is sensitive to solvent polarity.147 Similar to DNA intercalation,

the wavelength of maximum absorbance increases as the polarity of solvent

decreases (λabs = 480 – 540 nm).147 For all derivatives that “block” electron

donation by the exocyclic amines, a blue-shift in this charge-transfer band is

observed (λabs 398 – 473 nm) (Figure 5.1). Interestingly, modification of the 3

position consistently leads to a larger blue shift as compared to modification of

the 8 position (compare 16 to 17, 19 to 20, 22 to 23, and 25 to 26, Figure 5.1).

The degree of blue-shifting is also sensitive to the identity of the functional group,

94

where electron withdrawing functional groups lead to larger blue shifts as

compared to functional groups that merely “block” electron donation (compare

the guanidinium derivatives 19, 20 and 21, to the urea derivatives 22, 23, 24).

The only derivative that augments electron donation (28) exhibits a red-shifted

wavelength of maximum absorbance (Figure 5.1).

These derivatives also show interesting trends in their fluorescence properties.

Ethidium has a weak emission band (quantum yield < 0.01) that is sensitive to

solvent polarity (λem 610 – 645 nm, longer wavelengths in less polar solvents).147

As before, all derivatives of ethidium that effectively “block” electron donation by

its exocyclic amines, shift the emission maxima to the blue. Again, the blue-shift

is more pronounced when the 3 position is modified as compared to the 8

position; but unlike the absorbance maxima, the changes in emission maxima

are not very sensitive to the identity of the functional group. For all 3-modified

derivatives λem = 591 ± 3 nM; for all 8-modified derivatives λem = 605 ± 2 nM.

Trends in the relative quantum efficiencies of each compound also indicate

significant electronic differences between the 3 and 8 positions. The modification

of the 8 position leads to a significant decrease (3 – 15 fold) in emission intensity

(relative to ethidium). Modification of the 3 position, however, shows little to no

effect (Figure 5.1). For the bis-substituted derivatives that block electron donation

from the nitrogens (18 and 24), significantly higher quantum yields are measured

(57-fold greater compared to ethidium). The only derivative that augments

95

electron donation (28) exhibits a much lower quantum efficiency as compared to

ethidium (12). These observations suggest that electron donation by ethidium’s

exocyclic amines quenches its excited state through an internal charge transfer.

The differences observed between the amines at the 3 and 8 positions suggest

that electron donation from the 3 position amine lowers the energy of excitation

and emission, but also lowers the quantum efficiency of emission (much more so

than the 8 position). Additional interpretations of these photophysical trends are

complicated by the likelihood that the energetic levels of both the ground-state(s)

and excited state(s) are changed upon derivatization. Other groups have also

observed spectral differences between 3 and 8 modified positions, but have

attributed it to the differences in LUMO energy levels of the two positions.147 This

is an inadequate explanation, as we have found significant differences in the

ground-state electronic properties of these two amines that present some

explanations for the spectral differences at these positions. To probe the

structural and electronic properties of ethidium’s ground-state, NMR

spectroscopy and X-ray crystallography have been used.

It is clear, from 1H-NMR, that the chemical environments of ethidium’s exocyclic

amines are significantly different (Figure 5.2). As compared to ethidium, the

proton NMR spectrum of ethidium’s uncharged analog (3,8-diamino-6-

phenylphenanthridine) shows much higher symmetry across the phananthridine

core where similar chemical environments are seen for each of its exocyclic

amines and for protons a1 & a2, b1 & b2, and c1 & c2 (Figure 5.2 A). Interestingly,

96

ethylation of 3,8-diamino-6-phenylphenanthridine introduces significant electronic

asymmetry (Figure 5.2 B). Upon ethylation, most protons are downfield shifted

by 0.2 – 0.5 ppm (Figure 5.2 B). There are two notable exceptions: proton a2 is

upfield shifted by 0.7 ppm, and the protons on one of the two exocyclic amines

are downfield shifted by 1.0 ppm (compare A and B Figure 5.2).

Figure 5.2. 1H NMR of 3,8-diamino-6-phenylphenanthridine (A) and of ethidiumchloride (B) in DMSO-d6. A partial assignment of ethidium (B) has previouslybeen made,157,158 but assignments for the exocyclic amines have not yet beenreported.

The upfield shift of proton a2 is highly sensitive to modifications of the 8 position,

but not sensitive to modifications of the 3 position. The conversion of the 8-amine

into any group that blocks electron donation (including compounds 17, 18, 20,

21, 23, 24, 26, 27) results in a downfield shift of a2 to a more “normal” aromatic

4.55.05.56.06.57.07.58.08.59.09.5

N

NH2H2N

a1 a2

b2c2c1b1

Phd d

N

NH2H2N

a1 a2

b2c2c1b1

Ph

c1 & c2

c1 & c2

Ph

Ph

b2

a2

a1 & a2b1 & b2

a1

b1

NH2 NH2

NH2NH2

d

A)

B)

4.55.05.56.06.57.07.58.08.59.09.5

+

97

frequency (δ = 7 – 7.4 ppm). The modification of the 8-amine with groups that

augment electron donation, results in an even further upfield shift of a2 (∆δ = -0.2

ppm for 28). This suggests that the 8-amine donates electron density directly

onto the carbon attached to the a2 proton. 13C NMR experiments support this

conclusion (presented below). Previous groups have assigned ethidium’s 1H

NMR spectrum based upon the highly shielded proton a2.157-158 It has always

been assumed that the phenyl ring at the 6 position shields the a2 proton via a

ring current effect.145, 157-159 This is not a sufficient explanation, as ethidium’s un-

charged analog (3,8-diamino-6-phenylphenanthridine) shows no unusual

shielding of a2 (Figure 5.2A). 160 Our results indicate that the presence of the

quarternary amine and the electron donation from the 8-amine are largely

responsible for the unusually high electron density at a2.13C NMR experiments

(presented below) confirm this conclusion.

The differences in the 1H chemical shifts of ethidium’s two exocyclic amines

indicate that one amine is significantly more electropositive than the other (Figure

5.2 B). The identity of each of ethidium’s exocyclic amines is apparent by

comparing the proton NMR spectra of the cbz derivatives (16-18) (Figure 5.3). By

comparing the chemical shifts of the unblocked amine present in 8-cbz ethidium

chloride (Figure 5.3 A) to that of 3-cbz ethidium chloride (Figure 5.3 B), it is clear

that the 3 position of ethidium is more electron poor than the 8 position. All other

chemical shifts (such as protons e1 versus e2, and f1 versus f2) are consistent

with this conclusion (Figure 5.3 A – C). The 1H NMR assignment of ethidium

98

was confirmed using 1H-13C Heteronuclear Multiple Bond Correlation (HMBC)

experiments (presented below).

Figure 5.3: 1H NMR of (A): 8-cbz ethidium chloride (17), (B): 3-cbz ethidiumchloride (16), (C): 3,8-bis cbz ethidium chloride (16), (D): ethidium chloride (12) ind6-DMSO. Sample (A) is doped with 5% of sample (16) to serve as an internalreference.

The 3 and 8 positions of ethidium are not equivalent, but electronic

communication between the positions is apparent. The e1 and e2 protons are

shielded in both of the mono-cbz derivatives relative to bis-cbz ethidium

(compare Figures 5.3 A & B to C). Conversely, the amines at the 3 and 8

4.55.05.56.06.57.07.58.08.59.09.510.010.511.0

4.55.05.56.06.57.07.58.08.59.09.510.010.511.0

4.55.05.56.06.57.07.58.08.59.09.510.010.511.0

4.55.05.56.06.57.07.58.08.59.09.510.010.511.0

N

NN

a1 a2

b2c2c1b1

Phd d

O

O

O

Oe1 e2

f1f1 f2f2

PhPh

N

NH2N

a1 a2

b2c2c1b1

Phd d

O

Oe2

f2f2

Ph

N

NH2N

a1 a2

b2c2c1b1

Phd d

O

Oe1

f1f1

Ph

c1 & c2

c1 & c2

c1 & c2

c1 & c2

3-NH2

8-NH2

8-NH23-NH2

d

d

d

d

a2

a1

f2

f1

Pha2

Ph

Phb1

Ph

Ph

Ph

f2f1

e2

e1

e1

e2

Ph

a2

b1

b2

b2

b2 b1

b2

N

NH2H2N

a1 a2

b2c2c1b1

Phd d

A)

B)

C)

D)

a1

a1b1

a2a1

+

+

+

+

99

positions are deshielded upon cbz modification of the other position (compare

Figures 5.3 A & B to D). This indicates that upon blocking electron donation of

one amine, the other amine can “compensate” by donating more of its electron

density into the phenanthridinium core. This conclusion is confirmed by

comparing the nitrogen-carbon bond lengths in the X-ray crystal structures of 12,

17, and 18 (presented below).

Importantly, protons a1 & a2 are much more sensitive to cbz-modification as

compared to both b1 & b2 and c1 & c2 (Figure 5.3). Both the a and b protons are

ortho- to each amine, however, upon cbz modification the a proton is shifted

downfield by 1.4 ppm, and the b proton is shifted downfield by 0.7 ppm (compare

Figures 5.3 A & B to D). This is true for both the 3 and 8 amines, and indicates

that electron density from each amine is preferentially drawn towards the positive

center, thus shielding the a protons more so than the b protons. This conclusion

is supported by 13C NMR (presented below).

As for its 1H NMR spectrum, ethidium’s 13C spectrum shows an unusually wide

disbursal of peaks in the aromatic region (Figure 5.4, Bottom). We have fully

assigned ethidium’s 13C spectrum by using Distortionless Enhancement by

Polarization Transfer (DEPT), Carbon-Hydrogen Correlation spectroscopy

(HETCOR), and Heteronuclear Multiple Bond Correlation (HMBC) experiments.

100

Figure 5.4: 13C NMR spectra of ethidium chloride in DMSO. DEPT was used toidentify the primary, secondary, tertiary, and quarternary carbons.

DEPT experiments reveal the number of hydrogens attached to each carbon

(Figure 5.4). Interestingly, five of ethidium’s tertiary core carbons “13C-(1H)” are

upfield relative to typical phenyl carbons (Figure 5.4). The δ for benzene, for

comparison, is 129 ppm. This suggests that, despite ethidium’s positive charge,

these carbons possess high electron densities. All of the carbons that are

electron poor, relative to benzene, are found to be quarternary (Figure 5.4).

Full 13C spectrum

All 13C - 1H

13C - (1H)

13C - (1H)2

13C - (1H)3

101

Carbon-Hydrogen Correlation spectroscopy (HETCOR) has been used to assign

the identity of each carbon that is directly bound to a single hydrogen (Figure

5.5). We find that the most electron-rich carbons are C4 and C7 (see Figure 5.7

for the numbering scheme and a summary of assignments). Assignments of the

carbons at the 1, 2, 9, and 10 positions were also determined by HETCOR, and

confirmed by HMBC. These assignments are consistent with those determined

by NOE experiments.161

Figure 5.5: 13C-1H HETCOR NMR spectrum of ethidium chloride in DMSO withthe 1H spectrum (left) and the “13C – (1H)” DEPT spectra added for reference.

HMBC experiments have been used to identify all the quarternary carbons and to

confirm the assignment of the 1H NMR spectrum of ethidium (Figure 5.6). For

ppm8090100110120130140150160170

ppm

5.05.25.45.65.86.06.26.46.66.87.07.27.47.67.88.08.28.48.68.89.09.29.49.69.8

102

HMBC, cross peaks are not observed between carbons and hydrogens that are

directly bound to each other; rather, HMBC gives cross peaks between a

hydrogen and the carbons that are ortho- or meta- to the atom that is bound to

that hydrogen. In our case, the ortho- carbons have been decoupled so that all

cross peaks traverse exactly three bonds. Three cross peaks are observed for

each 1H-C type hydrogen; two are meta- relative to the 1H-C and one is on an

adjacent ring. Two cross peaks are observed for each 1H-N type hydrogen, both

from the carbons meta- to each amine.

Figure 5.6: HMBC spectrum of ethidium chloride in DMSO with 13C and 1Hspectra added for reference. Coupling between the “d” hydrogens of the ethylgroup and C6 allows unambiguous assignment of C6 at 157.5 ppm (highfieldregion not shown).

103

The most striking features of ethidium’s 13C spectrum are the highly shielded

carbons at positions 4 and 7. At 97 and 107 ppm, respectively, these carbons are

in the same chemical shift region observed for aromatic carboanions.162

Consistent with the trends observed in the 1H NMR spectra, the modification of

ethidium’s 8-amino position with cbz results in a large downfield shift of C7 (∆δ =

+10 ppm). The modification of ethidium’s 3-amino position results in the same

downfield shift of C4 (∆δ = +10 ppm) (Figure 5.7 A and C). This indicates that the

high electron density at these positions depends largely on electron donation

from the neighboring exocyclic amines.

Figure 5.7: 13C NMR spectra of (A): ethidium chloride (12), (B): 8-cbz ethidiumchloride (17), (C): 3-cbz ethidium chloride (16), each in d6-DMSO. See Table 5.1for the full 13C assignment of 12.

5060708090100110120130140150160170

5060708090100110120130140150160170

5060708090100110120130140150160170

A)

B)

C)

C3

C2 C1

C12

C13C4

C11

C14

C6N

C10 C9

C8

C7

NH2H2N

PhC15C16

+

O

OCZ

C3

C2 C1

C12

C13C4

C11

C14

C6N

C10 C9

C8

C7

NH

H2N

PhC15C16

+

Ph

O

O

CZ

C3

C2 C1

C12

C13C4

C11

C14

C6N

C10 C9

C8

C7

NH2NH

P hC15C16

+

Ph

C6 C3

C8

C4 C15C7C12C2

C14C1C13

Ph

C10

104

The only carbons in ethidium with chemical shifts that are significantly downfield

shifted, relative to benzene, are the carbons that are bound directly to nitrogens

(Figure 5.7 A). C3 and C8 have chemical shifts similar to the nitrogen-bound

carbon of aniline (148 ppm).162 Interestingly, C6 is significantly more deshielded,

and C13 is significantly more shielded as compared to both C3 and C8 (Figure

5.7A).

To relate ethidium’s electronic features to the bond orders between its C and N

atoms, X-ray crystallography was used to solve the 3-D structure of ethidium

chloride and of the mono-cbz derivatives 16 and 17. The bond lengths observed

in the phenanthridinium core of ethidium chloride are similar to those of two

recent structures of phenanthridine, and 5,6-dimethyl phenylphenanthridinium

chloride (Figure 5.8 A – C).163-164 For all three compounds, the carbon-carbon

bond lengths in the phenyl rings of phenanthridine alternate between relatively

short and relatively long (for comparison, benzene’s C-C bond lengths are 1.39

Å). The patterns of bond orders that contribute to ethidium’s phenanthridinium

core are summarized in Figure 5.9. Another group has recently calculated the

bond orders for phenanthridine (Figure 5.10).164 Significant double-bond

character exists between C3-C4 and C7-C8 carbons, while some single-bond

character is observed at C2-C3 and C8-C9. This may, in part, explain the greater

electron donation from the exocyclic amines to the 4 and 7 positions versus the 2

and 9 positions.

105

A) Phenanthridine (Brett, Rademacher, and Boese)163

B) 5,6-dimethyl phenylphenanthridinium chloride (Kiralj, Kojic-Pordic, Zinic,Alihodzic, and Trinajstic)164

C) Ethidium chloride (Luedtke, Liu, Tor)165

Figure 5.8. Bond lengths (Å) observed for three recent crystal structures ofphenanthridine and phenanthridinium-containing molecules. Errors in bondlengths are approximately ± 0.003 Å for (A) and (C) and ± 0.005 Å for (B). See E5.1.1 for the ORTEP drawing and crystallographic parameters for ethidiumchloride (C).

N 1.296

1.377

1.407

1.374

1.446

1.407

1.418

1.41

6

1.37

3

1.401

1.381

1.40

3

1.408

1.41

31.

396

1.432

N

1.5011.322

1.370

1.416

1.374

1.440

1.415

1.409

1.41

2

1.35

1

1.391

1.357

1.38

9

1.394

1.40

91.

422

1.410

1.49

0

N

NH2H2N

1.517 1.488

1.337

1.48

9

1.371 1.381

1.365

1.402

1.369

1.431

1.412

1.417 1.41

4

1.38

4

1.417

1.378

1.386

1.38

41.38

4

1.394

1.389

1.378

1.41

6

1.417

1.41

41.

406

1.416

106

Figure 5.9. All possible Clar and Kekulé depictions of ethidium. Theexperimentally determined bond lengths in ethidium indicate that somedepictions contribute more to the actual structure (near Bottom), while otherscontribute much less (near Top).

Figure 5.10. A summary of phenanthridine’s π-bond orders as determined byKiralj.163 Pauling bond orders were calculated by analyzing the bond lengths of ahigh-resolution crystal structure of phenanthridine (Figure 5.5 A).162 Ethidium hasapproximately the same bond lengths as phenanthridine and, therefore, similarbond orders (Figure 5.5 C).

Crystallographic data indicates that both phenanthridine and 5,6-dimethyl

phenanthridinum are perfectly planar.163,164 In contrast, ethidium is found to have

a 4o twist in its phenanathridinium core (the angle between two vectors defined

N .8

.2.2

.2

.4.4

.4.4

.6

.4

.6

.4 .4

.6

.4

.6

N

NH2H2N

N

NH2H2N

N

NH2H2N

+

+

+

N

NH2H2N

+

N

NH2H2N

+

N

NH2H2N

+

N

NH2H2N

+

ContributesLess

ContributesMore

107

by protons b1c1 and b2c2, relative to the central axis of the biphenyl “core” of

ethidium). Taken together this suggests that some sp3 character is introduced

into the phenanthridinium core of ethidium by the exocyclic amines. Significant

differences are observed in the C-N bond lengths of ethidium’s two exocyclic

amines. A shorter bond length is found for the amine at the 3 position, compared

to the amine at the 8 position (Figure 5.8 C and 5.11 C). This indicates that more

C=N double bond character, and hence positive charge, is at the 3 position

versus the 8 position. Both positions, however, have significantly shorter N-C

bond lengths as compared to aniline (1.40 Å), and therefore, both amines carry

partial positive charges. This is consistent with the chemical shifts of these

amines in the 1H NMR spectrum of ethidium (Figure 5.3 D).

Two other groups have also solved the crystal structure of ethidium. The first,

and most cited structure, was published by Hospital et. al. in 1969.165 It shows no

differences in the C-N bond lengths of its two exocyclic amines, and predicts the

phenanthridine core of ethidium to be perfectly planar (Figure 5.11 A).165 Unlike

the modern crystal structures of phenanthirdine-containing compounds (Figure

5.8), this structure has significant asymmetry in the C-C bond length/bond orders

in the biphenyl core of phenanthidinium. The deviation from symmetry for the two

phenyl rings (with respect to each other) is ca. 0.011 Å for the Hospital structure

compared to 0.004 Å for all the recent phenanthridine-containing compounds.

108

A) Courseille, Busetta, & Hospital (1969, 1974)165

B) Subramanian, Trotter, & Bugg (1971)166

C) Luedtke, Liu, & Tor (2002)167

Figure 5.11: Bond lengths (Å) observed for three crystal structures of ethidium.The crystal system is the same for all three structures (monoclinic), even thoughdifferent solvent molecules and anions are co-crystalized in each.

N

NH2H2N

1.500 1.499

1.338

1.48

0

1.364 1.395

1.353

1.416

1.362

1.424

1.407

1.397

1.41

7

1.36

9

1.395

1.377

1.370

1.35

91.36

8

1.379

1.394

1.367

1.40

0

1.413

1.42

61.

409

1.417

N

NH2H2N

1.500 1.497

1.333

1.50

1

1.378 1.375

1.340

1.395

1.362

1.431

1.421

1.4051.

415

1.38

1

1.415

1.378

1.383

1.39

51.36

9

1.393

1.369

1.377

1.41

5

1.4151.

413

1.41

11.406

N

NH2H2N

1.517 1.488

1.337

1.48

9

1.371 1.381

1.365

1.402

1.369

1.431

1.412

1.417 1.41

4

1.38

4

1.417

1.378

1.386

1.38

41.38

4

1.394

1.389

1.378

1.41

6

1.417

1.41

41.

406

1.416

109

The second crystal structure of ethidium, by Subramanian, Trotter & Bugg, is

much more consistent with ours (Figure 5.11 B). It shows a shorter N-C bond and

hence, greater positive charge at the 3 position as compared to the 8 position. In

addition, the phenanathidinium core in this structure is twisted out of planarity by

4o. Interestingly, the Subramanian crystal structure shows a different

hybridization at each of its exocyclic amines. The 3- and 8-amino groups are

predicted to be sp2 and sp3, respectively.166 We also find unusual hybridizations

of the exocyclic amines. Electron difference maps indicate that both the 3- and 8-

amino groups are intermediate between sp2 and sp3 hybridized.

Both of the older crystal structures of ethidium (Figure 5.11 A & B) unexpectedly

show C-C bond lengths in the 6-phenyl ring that deviate substantially from that of

benzene (1.39 Å). The average deviation in bond lengths of the 6-phenyl ring,

compared to benzene, is 0.013 Å for the Subramanian and Courseille structures,

and 0.006 Å for our structure (Figure 5.11). Both 1H and 13C NMR indicate that, in

solution, the 6-phenyl of ethidium has the typical chemical shifts of a mono

substituted phenyl ring, and therefore, no unusual bond orders are expected.

Ethidium is an electronically and structurally plastic molecule. Evidence for

electronic plasticity can be found in the crystal structures of 3-cbz ethidium

chloride (16) and 8-cbz ethidium chloride (17) (Figure 5.12). For both mono-cbz

derivatives, the cbz group merely “blocks” electron donation of the exocyclic

amine without making it an electron withdrawing group, as the C-N bond length

110

for both the 3 and 8 cbz-protected amine is 1.399 Å (Figure 5.12). This is the

same C-N bond length for aniline. The C-N bond lengths for the “unblocked”

amines are shorter compared to the same bond lengths in the parent structure

(compare Figure 5.11 C and Figure 5.12). This indicates that upon “blocking”

electron donation from one amine (by cbz protection) the other side can

compensate by donating more of its electron density into the phenanthridinium

core. This is consistent with the changes in 1H chemical shifts of the amines of

compound 12 as compared to 16 and 17 (Figure 5.3 A, B, D). It is possible that

this increased electron donation from the 3 position leads to lower quantum yield

for the 8 cbz derivative (Figure 5.1). The increased electron donation from the 8

position does not, however, change the quantum yield of the 3 cbz derivative

relative to ethidium (Figure 5.1). Both NMR and the changes in bond order

suggest that the magnitude of increased electron donation is approximately equal

for both derivatives (upon “blocking” the other side with cbz). It is possible that

the charge transfer observed between the 3 position and the quaternary amine is

related to the differences in quantum yield upon cbz modification (Figre 5.13).

The relative planarity of each derivative may also be a factor.

Structural flexibility of the phanathidinium ion can be inferred by comparing the

crystal structures of 8-cbz and 3-cbz ethidium chloride. As for the parent

compound, the phenanthridinium cores of these molecules are not planar. The

twist angle is, however, different for each cbz derivative. The crystal structure of

3-cbz ethidium chloride has the same twist as the parent compound (4°). 8-cbz

111

ethidium, however, shows a greater twist of the phenenthiridine core (8°). This is

consistent with more sp3 character residing on the phenanthridinium core when

the 8 position is blocked. This may, in part, explain the differences in quantum

yield of emission for these derivatives (Figure 5.1). Blocking the 8 position with

cbz increases the electron donation by the 3-amine and induces a greater twist of

the phenanathridinium core (relative to ethidium), while modification of the 3

position changes neither the twist nor the quantum yield of the cbz derivative

(relative to ethidium) (Figure 5.1).

Figure 5.12: Bond lengths (Å) found in the crystal structures of 3-cbz ethidium(16) (Top) and 8-cbz ethidium chloride (17) (Bottom). Errors in bond lengths are+/- 0.002 Å. See section E 5.1 – E 5.2 for ORTEP drawings and crystallographicparameters for 16 and 17.

N

NH2NH

O

O

1.513 1.497

1.333

1.49

21.399 1.364

1.371

1.397

1.369

1.437

1.412

1.415 1.41

3

1.38

0

1.4131.385

1.385

1.38

91.39

1

1.393

1.395

1.393

1.41

9

1.419

1.41

31.

405

1.429

1.3541.348

1.21

2

1.4551.503

1.391

1.3861.392

1.3921.385

1.38

4

N

NH

H2N

O

O

1.525 1.491

1.333

1.49

7

1.356 1.399

1.366

1.394

1.371

1.429

1.416

1.413

1.42

1

1.39

3

1.418

1.383

1.382

1.39

31.39

2

1.391

1.392

1.374

1.41

9

1.414

1.41

91.

412

1.419

1.365

1.207

1.3511.457 1.500

1.38

4 1.384

1.394

1.374

1.367

1.385

112

The crystal structure of 8-cbz ethidium chloride (17) shows that the C-N bond

length at the 3 position is shorter compared to the C-N bond at the 8 position

length for 3-cbz ethidium chloride 16 (Figure 5.12). This is consistent with 1H

NMR experiments (Figure 5.3), and indicates that even for these mono-cbz

derivatives, a greater total positive charge is on the 3-amine as compared to the

8-amine (for the 8-cbz and 3-cbz derivatives, respectively). 13C NMR indicates,

however, that the 3- and 8-amines donate the same amount of electron density

onto their neighboring carbons (C4 and C7, respectively) (Figure 5.7).

We believe the electronic differences between ethidium’s 3 and 8 positions are

consistent with simple resonance theory. The positive charge on ethidium’s

quarternary amine can delocalize to the 3 position but not to the 8 position

(Figure 5.13). No matter which Kekulé structure is used (Figure 5.9), it is not

possible to push the lone pair of electrons from the 8-amine directly onto the

quarternary amine.

Figure 5.13. Resonance structures showing charge transfer between thequarternary amine and the 3- amino.

Both of the ethidium’s exocyclic amines, however, carry a partial positive charge.

We have found that both of these amines donate electron density onto the

carbon atoms of the phenanthridinium core of ethidium (especially C4 and C7).

N

NH2H2N+

N

NH2H2N

+

113

We propose, therefore, that charge-separated states make significant

contributions to the overall electronic structure of ethidium (Figure 5.14).

Figure 5.14. Two examples of charge separation that contribute to ethidium’sground-state electronic structure.

For most uncharged molecules, charge-separated states contribute very little to

the overall electronic structure. It appears, however, that for ethidium, charge

separation is an essential component of its ground-state electronic structure.

Using data from crystallography, NMR, and computation we have estimated the

partial positive and partial negative charges on each of the carbon and nitrogen

atoms on ethidium’s phenanthridine core. A summary of the charges are

presented in Figure 5.15. The partial formal charges at each of ethidium’s

nitrogens are calculated from the N-C bond lengths obtained from the high-

resolution X-ray crystal structure (Figure 5.8 C). A linear correlation between the

bond length and bond order has already been proven for aromatic nitrogen-

carbon bonds of phenanthiride-containing molecules.164 As a standard for the

exocyclic amines, the N-C bond length for the cbz-blocked amines (1.40 Å) is

taken as 0 (no charge separation) (Figure 5.9). This is the same bond length as

N

NH2HN

+N

NH2H2N

+

_+

N

NH2H2N

+N

NH2H2N

+_

+

114

aniline, and, therefore, the charges on each amine are relative to aniline. The

N=C bond length of dimetyl iminium chloride and phenanthridine (both 1.30 Å)

are taken as standards to equal +1 (a full double bond). For the endocyclic

amine, the N5-C6 bond length is used to assign its charge, taking a bond length

of 1.47 Å as a single bond and 1.30 Å as a double bond (based upon published

crystal structures of similar compounds). Using these bond lengths as standards,

a linear correlation is used to assign the magnitude of positive charge on each

nitrogen (Figure 5.12 A). The partial formal positive charge on each carbon

(Figure 5.12 A) was assigned by a summation of all the nitrogen partial charges

and subsequent distribution of the remaining positive charge to each carbon

atom according to a Penderson-Parr calculation of ethidium. According to both

13C and 1H NMR the 8-amine and 3-amine donate the same magnitude of

electron density onto the carbon atoms of ethidium. The difference in charge

between each amine is, therefore, the total magnitude of charge delocalization

between the 3-amine and quarternary amine (Figure 5.13). This value (+0.10) is

added to the partial charge on the quarternary amine (+0.60) and subtracted from

(+1.0) to obtain the total positive charge on all of the ethidium’s carbon atoms

(+0.30). This assumes that all of the electron density from the 8 position is

transferred onto carbon atoms, and that the total charge of the molecule (ignoring

the charge separation) equals (+1.0). The total positive charge on the carbon

atoms (+0.30) was then distributed to each carbon atom using the theoretical

partial charge on each atom as previously determined by a 18 π-electron

Penderson-Parr calculation of ethidium bromide.156

115

Figure 5.15: Estimates of the distribution of positive charge (A), negative charge(B), and the sum showing resulting electrostatic map of ethidium (C).

The negative charge density at each carbon atom, arising from charge

separation, was determined by comparing the 13C NMR spectrum of ethidium

chloride to that of 3,8-bis cbz ethidium chloride (Table 5.1). Upon blocking both of

-0.036

-0.079-0.070

-0.030

-0.009 -0.002

-0.014 -0.015-0.017 -0.011

+0.76

+0.18+0.28

N

NH2H2N

-0.045N

NH2H2N

+0.062

+0.022

+0.016+0.011

+0.007 +0.008

+0.008

N

NH2H2N

+0.76+0.02

0 -0.07

-0.03

+0.01-0.07

-0.02

+0.18+0.28

A)

B)

C)

+0.003 +0.005

+0.0060

0

+0.002

-0.017 -0.016

-0.01

-0.01-0.01 -0.01

-0.01 0

116

its exocyclic amines with cbz, ethidium’s carbon atoms shift by a total of +47.5

ppm. This total change in chemical shifts was correlated to the partial charges on

the amines by adding the total charges on the exocyclic amines (+0.45) and

subtracting the contribution by charge delocalization at the 3 position (+0.10).

The total amount of π-electron charge density provided by the amines is -0.36

formal units. A linear correlation between the π-electron charge density and the

change in 13C chemical shift has already been established.162 Interestingly, our

finding that ∆δ 48 ppm = 0.36 charge units is the same relationship (per carbon

atom) as previously reported for a series of simple anionic and cationic aromatic

rings.162 The total negative charge from charge separation (-0.36) was distributed

over the phenanathridinium core of ethidium by multiplying this value by the

fractional change in chemical shift for each carbon upon blocking the electron

donation from the exocyclic amines (Table 5.1 and Figure 5.15 B). The C3 and C8

carbons are excluded from this analysis, as the changes in their chemical shifts

are dominated by the substitution of the nitrogen atom to which they are

attached. These positions are not, however, expected to receive electron density

from the amine to which they are directly attached. Instead, C3 receives electron

density from the 8-amine, and C8 receives electron density from the 3-amine

(Figure 5.3 A – C). We have used, therefore, the chemical shifts of the mono-cbz

derivatives (relative to ethidium) to estimate the electron donation to these two

positions (Table 5.1).

117

Table 5.1: Summary of 13C assignments and chemical shifts of ethidium (12) and3,8-bis cbz ethidium chloride (18). HETCOR and HMBC experiments were usedto assign all carbon atoms in the phenanthridinium core of 12 and 18.

Carbon 12 δ (ppm) 18 δ (ppm) ∆∆∆∆ δ (ppm) π-electron density

1 123.7 125.5 1.8 -0.0142 119.2 121.4 2.2 -0.0173 150.5 142.3 2.2* -0.0174 97.7 107.0 9.3 -0.0706 157.5 163.4 5.9 -0.0457 107.1 117.5 10.4 -0.0798 147.2 140.3 2.1** -0.0169 127.1 128.6 1.5 -0.011

10 121.8 123.8 2 -0.01511 126.6 130.5 3.9 -0.03012 116.6 121.4 4.8 -0.03613 133.3 134.5 1.2 -0.00914 125.4 125.7 0.3 -0.002

*Change is the difference between 8-cbz ethidium and ethidium chloride.**Change is the difference between 3-cbz ethidium and ethidium chloride.

We have combined the contributions of cation distribution (Figure 5.15 A) with

charge separation (Figure 5.15 B) to create the first experimentally determined

electrostatic map of ethidium (Figure 5.15 C). This map illustrates the π-electron

distribution of partial positive and negative charges over the phenanthridinium

core of ethidium. As expected, the most electron-poor carbons are ortho- and

para- relative to the quarternary amine (C6, C11 and C13). Interestingly, these

carbons are directly bound by the four most electron-rich carbons (C4, C7, C12,

and C14). The electron density of the exocyclic amines, therefore, appears to be

directed onto carbons that can diminish the electrostatic repulsion between these

positively charged positions.

118

Crystal structures of ethidium intercalated into nucleotide diphosphates show that

the C6 and the quarternary amine are not involved in base stacking

interactions.138 We can consider, therefore, only the biphenyl “core” of ethidium

for modeling stacking interactions with base pairs. It is interesting that formal

negative charges dominate most of the carbons in the phenthidinium core of

ethidium. These electron-rich positions may base-stack with the electron poor

positions of the base pairs (the exocyclic amines). In collaboration with Dr. Kim

Baldridge, we are comparing possible electrostatic complementation between

ethidium and different intercalation sites. Previous electrostatic models of

ethidium neglect the significant electronic differences between the 3 and 8

positions and vastly underestimate the contributions made by “directed” charge

separation.149,156 We find that the differences in partial positive charges on the

exocyclic amines correlates with the differences in reactivity and basicicity of

these positions (next section).

5.2 Synthesis of Ethidium Bromide Derivatives

The exocyclic amines of ethidium, especially the 3 position, are poor

nucleophiles and weakly basic (pKa3 = 0.8, pKa8 = 2) (Section 5.3). The electron

withdrawing effect of phenanthridinium requires the use of highly reactive

electrophiles to modify ethidium’s exocyclic amines. Until now, no general

method for the systematic protection and “modular” modification of ethidium’s

exocyclic amines has been reported. The exocyclic amines of ethidium can be

protected by a reaction with benzyl choroformate (Figure 5.16). By using one

119

equivalent of cbz-chloride, a mixture of products is obtained that can be

separated using standard silica gel chromatography on gram-scale quantities

(see E 5.2 for synthetic details). The main products are 8-cbz ethidium chloride

(50%) and 3-cbz ethidium chloride (10%). The remainder is 3,8-bis cbz ethidium

chloride (2%) and unreacted starting material (Figure 5.16).

Figure 5.16: Protection of ethidium bromide with benzyl chloroformate. Seesection E 5.2.1 for synthetic details. The assignment of each isomer (16) and(17) was confirmed by X-ray crystallography (Section 5.2).

Other less electrophilic reagents, including Boc anhydride, are capable of

reacting only with the 8 position and give none of the 3 protected or bis-protected

products (even using a large excess of Boc anhydride and with catalytic DMAP).

Other research groups have also observed a greater reactivity of the 8 position

Br

N

NHH2N

Cl

O

N

NH2HN

Cl

N

NHHN

Cl

N

NH2H2N

ethidium bromide (12)

8-cbz ethidium (17)(50%)

3-cbz ethidium (16)(10%)

3,8-bis cbz ethidium (18)( 2%)

-

-

-

-

+

+

+

+

PhCH2OCOCl

O

O

O

O

O

O

O

120

relative to the 3 position and have attributed the differences to steric constraints

imposed by dimerization of ethdium in solution.145 It is more likely, however, that

the inherent electronic differences between the these two exocyclic amines are

responsible for the differences in reactivity (Section 5.1).

Figure 5.17: Synthesis of 3-guanidino ethidium chloride (19). Isolated yields arereported. See E 5.2.2 for experimental details and characterization.

N

HNH2N

Cl

O

O

S

NHNN

HNHN

Cl

O

ONH

N

OO

O

O

N

NH2HN

NH2

N

Cl

Cl

H

H

HgCl2(70%)

6M HCl / MeOH

O

O

HNO

O

NH

NS

O

OFFF

N

NHN

NN

NHNHN

Cl

O

ONH

N

OO

O

O(20%)

Cl

(70%)

8-cbz ethidium (17)

No reaction

OO

OO

OO

OO

3-guanidino ethidium dichloride (19)

3-diBoc guanidino8-cbz ethidium chloride

8-cbz ethidium (17)

8-cbz ethidium (17)

3-diBoc guanidino8-cbz ethidium chloride

+-

N

HNH2N

Cl

O

O

+-

N

HNH2N

Cl

O

O

+-

+

+-

-

-

-

121

Guanidinylation of ethidium was conducted using three different methods (Figure

5.17). The best reagent for guanidinylation of ethidium proved to be N,N’-bis-

Boc-S-methyl-isothiourea (activated with mercury dichloride) to afford a 70%

isolated yield of the protected product (3-diBoc-guanidino 8-cbz ethidium

chloride) (Figure 5.17). Removal of both the cbz and Boc groups was performed

simultaneously by refluxing 3-diBoc-guanidino 8-cbz ethidium chloride in 6M HCl

/ MeOH (1 hr). Reversed-phase chromatography was used to purify the desired

product 19 that was obtained in 70% yield. The other guanidino-ethidium

derivatives 20 and 21 were also synthesized by this method with similar yields

(see Section E 5.2.3 and E 5.2.4 for details and characterization).

Figure 5.18: Unsuccessful attempts at synthesizing urea derivatives of ethidium.

Urea is a non-charged isostructural analog of guanidine, and provides an

important comparison for guanidinium-containing ethidium derivatives. Once

again, the limited reactivity of ethidium’s electron-poor exocyclic amines rendered

some urea-forming reagents ineffective (Figure 5.18). By using phenyl

No reaction

OHN

O

O2N

NC

O

Multiple products

NHNH2N

Cl

O

O8-cbz ethidium (17)

+-

NHNH2N

Cl

O

O8-cbz ethidium (17)

+-

122

chloroformate, however, the exocyclic amine can be successfully activated for

urea formation (Figure 5.19). This two-step approach allows for facile synthesis

of substituted and unsubstituted ureas (Figure 5.19 and 5.20).

Figure 5.19: Typical synthetic route for urea derivatives. Amines other thanammonia produce substituted ureas. Harsh deprotection conditions (refluxing 6M HCl) result in partial urea hydrolysis (last step), but the phenanthridinium“core” of ethidium is susceptible to reduction with the standard method of cbzremoval (H2/Pd).

Various amines have been conjugated to ethidium through urea linkages (Figure

5.20). All of these derivatives have photophysical properties similar to their

“simple” urea ethidium counterparts (Figure 5.1), but these compounds exhibit

dramatically different nucleic acid affinities and specificities (next section).

O

ClO

H2PO4

O

O

NH3 / MeOH

(90%)

(70%)

3-phenoxy carbamate8-cbz ethidium phosphate

acetone/phosphate buffer pH 6.6

6M HCl / MeOH

(20%)

3-urea ethidium chloride (22)

Cl

ethidium chloride (12)

N

HNH2N

Cl

O

O

8-cbz ethidium (17)

+-

N

HNN

HO

O

+

(not isolated)

H2PO4

N

HNN

HO

O

+

-

-

O

H2N

N

NH2NH

+O

H2N

-

ClN

NH2H2N

+-

123

Figure 5.20: Examples of ethidium-urea conjugates and isolated yields (seesection E 5.2.11 – E 5.2.16 for synthesis and characterization). A “*” indicatessynthesized and characterized by Charles Liu.

Pyrrole-containing ethidium derivatives have been prepared using dimethoxy-

tetrahydrofuran (Figure 5.21). As for urea and guanidine, pyrrole is a flat

functional group that should be capable of intercalating into nucleic acids. Pyrrole

modification may facilitate stacking interactions, while the possibility of hydrogen-

bonding interactions is eliminated. The lone pair of electrons on the nitrogen of

the pyrrole are part of the aromatic sextet of the pyrrole ring system and are,

N

NH2NH

O

NH2H2NHO

HOO

O

H2N OHNH

N

NH

H2N

H2NNH2

OHOH

O

O

NH2HO

HN

O

N

NH

NH

OO

NN

N

NH

O

NH

H2NHO

HOOH

NH

O

NH

NH2OH

OHHO

3,8-bisurea pyrrolidineethidium (30) (100%)

3,8-bisurea 2-DOSethidium (33) (25%)

3-urea-ethidium-6'-3'deoxyneamine (37)(20%)

HO

N

NH

NH

OO

NH

NH

H2N NH2

3,8-bisurea ethylene diamineethidium (31) (91%)

N

NH

NH

OO

NH

NH

OHOHN

NH

H2NO OH

HN

NH

NH2

3,8-bisurea arginineethidium (32) (31%)

+ +

++

+

+

N

NH2NH

O

NH2H2NHO

HOO

O

H2N OHNH

3-urea-ethidium-6'-neamineethidium (35)*

HO

+

N

NHH2N

O

H2N NH2OHOH

O

O

NH2HO

HN

8-urea-ethidium-6'-3'deoxyneamine (36)(20%)

+

8-urea-ethidium-6'-neamine (34)*

124

therefore, not available for electron donation into the phenanthridinium system or

for hydrogen bonding with nucleic acids. Interestingly, the UV absorbance

maximum of 27 is very similar to both 3, 8-bis cbz ethidium (18) and to 3, 8-bis

cbz ethidium (22), but it is approximately 50-fold less emissive compared to

these compounds (Figure 5.1). This suggests that charge transfer from pyrrole

into phenanthridinium may occur in the excited state of 27, but electron donation

into phenanthridinium does not occur in the ground-state.

Figure 5.21: Synthesis of 3,8-bis pyrrole ethidium acetate. The mono pyrrolederivatives 25 and 26 were synthesized under similar conditions (using the monocbz derivatives as starting materials). See section E 5.2.9 for full synthetic detailsand characterization).

5.3 Acid-Base Properties of Ethidium and its Guanidino Derivatives

The UV-vis spectrum of ethidium changes as a function of pH. The wavelength of

its visible adsorption band (λabs) is blue-shifted upon addition of acid (Figure

5.22). A plot of λabs versus pH reveals a fully reversible, biphasic curve between

pH 0.5 and 3.0. Since the 3 position of ethidium is electron poor and not very

reactive (relative to the 8 position) we have assigned the transition at pH = 0.8 to

protonation of the 3 position, and the transition at pH = 2.0 to protonation of the 8

position. This assumption is supported by measuring the pKa values of 3-

N

NH2H2N

BrN

N N

O OOCH3

H3C

ethidium bromide (12) 3,8-bis pyrrole ethidium (27)

acetic acid, 130oC(90%)

OAc+--

125

guanidino ethidium (19) and 8-guanidino ethidium (20) by using the same

method (Figure 5.23, Left).

Figure 5.22: Changes in the λabs of ethidium as a function of pH; no changes inλabs are observed between pH 5.0 – 12.8. By taking the pH at each inflectionpoint as equal to the pKa, a pKa of 0.8 and 2.0 are measured for 3 and 8 amines,respectively. These values are similar to those reported by Zimmerman (at 0.7and 2.4, respectively).168

Figure 5.23: Changes in the λabs of 3-guanidino ethidium (19) and 8-guanidinoethidium (20) as a function of pH (left). Changes in the λabs of 3,8-bis guanidinoethidium (21) and 3, 8-bis urea ethidium (24) as a function of pH (right). Theerrors for each point are approximately ±1 nm and ± 0.05 pH units. See SectionE 5.3 for additional experimental details.

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0425

450

475

500

525

pKa1

pKa2

pH

λλ λλa

bs

(nm

)

0.0 2.5 5.0 7.5 10.0 12.5390

400

410

420

430

440

450

(21)(24)

pH

λλ λλa

bs

(nm

)

6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0440

445

450

455

460

465

(19)(20)

pH

λλ λλa

bs

(nm

)

126

Both 19 and 20 exhibit monophasic titration curves (Figure 5.23, Left). By taking

the pH of each inflection point as equal to the pKa, a pKa of 9.2 and 9.9 are

measured for 3-guanidino ethidium (19) and 8-guanidino ethidium (20),

respectively. 8-guanidino ethidium (20) is, therefore, significantly more basic than

3-guanidino ethidium (19). This is consistent with the 8 position of ethidium

having more electron density and greater reactivity as compared to the 3 position

(Sections 5.1 – 5.2). 3,8-bis guanidino ethidium (21) has a biphasic titration curve

(Figure 5.23, Right). Since 8-guanidino ethidium (20) is more basic than 3-

guanidino ethidium (19), the inflection points at pH 6.2 and pH 9.5 are assigned

as the 3 and 8 positions of 21, respectively. These results are consistent with the

pKa values of the mono-guanidino derivatives 19 and 20, since protonation of the

8 position should significantly lower the pKa of the 3 position. As a control 3, 8-

bis urea ethidium (24) was also monitored as a function of pH, and shows no

changes in its absorption spectrum over this pH range (Figure 5.23, Right).

These studies indicate that at pH 7.5 all three of the guanidino derivatives 19, 20,

and 24 exist as dications and ethidium (12) exists as a monocation. A summary

of the pKa values for each compound is presented in Figure 5.24.

127

Figure 5.24: A summary of pKa values as determined by UV-vis absorption. Thecounterions for 12 are biphthalate, and the counterions for 19, 20, and 21 arephosphate.

5.4 Preliminary Evaluation of the Nucleic Acid Affinity and Specificity ofEthidium Derivatives

A-G and G-G base pairs in the internal bulge of the RRE may present unique

recognition elements for small-molecule binding (Figure 2.3 B). By increasing the

dimensions of an intercalating agent, it may become too large to intercalate into

“normal” duplex regions but still “fit” into the unusually wide purine-purine base

pairs of the RRE. This may lead to intercalating agents with improved RRE

selectivity. Importantly, by decreasing its affinity to “nonspecific” nucleic acids,

the mutagenic activities of ethidium-based compounds may be minimized.

N

NH

H2N

N

NH2NH

N

NH3H3N

N

NH

NHH2N

N

H

H

pKa = 0.8pKa = 9.2

pKa = 9.8pKa = 9.5pKa = 6.2

pKa = 2.0

H2N

N

H

H

NH2

N

H

H

NH2

N

H

H

3-guanidino ethidium (19)

8-guanidino ethidium (20)3,8-bis guanidino ethidium (21)

ethidium (12)

+

++

+

+ +

+

+

+

128

The positive charge afforded by ethidium’s quarternary amine is essential for its

high-affinity binding of the RRE. 3,8-diamino-6-phenylphenanthridine, an

uncharged analog of ethidium, has over a 20-fold lower affinity to the RRE

compared to ethidium. It was hypothesized, therefore, that conversion of the

amino groups on ethidium into guanidinium would increase the positive charge of

ethidium and, therefore, increase its affinity to the RRE. The guanidino

derivatives 19, 20, and 21 are 105 – 108 times more basic compared to ethidium,

and each compound exists as a dication at pH 7.5. The binding of these

compounds to the RRE was evaluated by using a Rev peptide displacement

assay (Section E 3.1), and the results are summarized in Table 5.2

According to fluorescence anisotropy RevFl peptide displacement experiments

(Section 3.1), the guanidino derivatives (19, 20, and 21) have approximately 20 –

60 fold lower RRE affinities as compared to ethidium (Table 5.2). It is possible,

that the introduction of guanidino groups lowers RRE affinity due to unfavorable

steric interactions. To test this, the urea derivatives 22, 23, and 24 were

evaluated for RRE affinity (Table 5.2). Interestingly, the urea derivatives have

much better RRE affinities as compared to the guanidino derivatives. It appears

that the introduction of an additional charge actually decreases RRE affinity. This

is the opposite trend as observed for compounds that bind in the grooves of RNA

(next section).

129

Table 5.2: Summary of RevFl displacement experiments (by anisotropy) and ofdirect binding experiments with calf thymus (CT) DNA.

Compound Rev-RREIC50 (µM)I

CT DNAC50 (µM)II

ApproximateCT DNAKd (µM)V

SpecificityRatioVI

Ethidium (12) 0.2 12 3.1 0.065

3-guanidino ethidium (19) 4.1 30 8.6 0.48

8-guanidino ethidium (20) 8.1 120 36 0.23

3,8-bis guanidino ethidium (21) 11 70 21 0.52

3-urea ethidium (22) 0.4 120 36 0.011

8-urea ethidium (23) 4.0 60 18 0.22

3,8-bisurea ethidium (24) >1III 350 106 n.d.

3-pyrrole ethidium (25) 0.6 40 12 0.05

8-pyrrole ethidium (26) 0.4 20 5.6 0.071

3,8-bispyrrole ethidium (27) >4III n.d.IV n.d. n.d.

3,8-bis urea-2-DOS ethidium (33) 0.2 10 2.5 0.08

8-urea-ethidium-6'-neamine (34) 0.016 10 2.5 0.0064

3-urea-ethdium-6'-neamine (35) 0.010 20 5.6 0.0017

8-urea-ethidium-6'-3'deoxyneamine (36) 0.0014 10 2.5 0.00056

3-urea-ethidium-6'-3'deoxyneamine (37) 0.00080 18 5.0 0.00016

Neo-S-acridine (39) 0.016 1.2 < 0.1 >0.16

I 10 nM each RevFl and RRE66II Concentration of CT DNA (in b.p.) needed to bind ½ of a 1 µM solution of each ligand asmonitored by the fluorescence intensity of the ligand.III Fluorescence interference with RevFl allows only a limit to be reported.IV No significant changes in fluorescence intensity upon binding DNA.V Approximate Kd = ((C50 / 3.3) – 0.5)VI Ratio of (Rev-RRE IC50 / CT DNA C50)

The introduction of an additional positive charge onto ethidium may disrupt the

charge distribution on the core of ethidium (Section 5.1), resulting in less

favorable electrostatic complementation with its intercalation site(s) on the RRE.

130

Or, the desolvation of the guanidinium group (upon intercalation) may impose a

significant energetic penalty such that the guanidino derivatives may actually

surface-bind to nucleic acids. The changes in spectral properties of 3,8-bis

guanidino ethidium upon saturation with CT DNA, however, are very similar to

those for both the bis-pyrrole and bis-urea derivatives, suggesting a common

binding mode for all three derivatives (Table 5.3).

To evaluate the affinities of these compounds to “non-specific” nucleic acids, C50

values were measured by monitoring the fluorescence of each ethidium

derivative upon titration of CT DNA. Most derivatives show significant changes in

their spectral features upon binding DNA (Table 5.3). The concentration of CT

DNA needed to bind ½ of each derivative (C50 value) is summarized in Table 5.2.

Compounds that have a high specificity for the RRE will have a low affinity to CT

DNA and a high Rev-RRE inhibition activity. Among the “simple” ethidium

derivatives (Figure 5.1), 3-urea ethidium has the best RRE specificity (relative to

CT DNA). By taking the ratio of Rev-RRE IC50 divided by CT DNA C50 for each

compound, a “specificity ratio” can be calculated for each compound (Table 5.2).

A lower “specificity ratio” indicates a higher RRE specificity (relative to CT DNA).

The specificity ratio for 3-urea ethidium is 5-fold better than ethidium, due to its

low affinity for CT DNA (Table 5.2).

The conjugation of ethidium to neamine and to 3'deoxyneamine produces

compounds with extremely high RRE affinity and specificity. Compounds 34 – 37

131

(Figure 5.20) have 13 – 250 fold higher Rev-RRE inhibition activities as

compared to ethidium (Table 5.2). Upon saturation with CT DNA, the spectral

changes of 34 – 37, especially for emission intensity, are almost identical to the

changes exhibited by their respective “simple” urea derivatives 22 and 23 (Table

5.3). This suggests that the ethidium moiety in compounds 34 – 37 does, in fact,

intercalate upon binding nucleic acids.

Table 5.3: Summary of the spectral changes upon saturation with CT DNA.Compare with Figure 5.1 to obtain absolute values.

Compound Change in λ-abs (nm)

Change in λ-em (nm)

Change inemission intensity

Ethidium (12) +34 -7 +700%

3-guanidino ethidium (19) +27 -4 +180%

8-guanidino ethidium (20) +32 0 +200%

3,8-bis guanidino ethidium (21) +20 0 -30%

3-urea ethidium (22) +32 -16 +360%

8-urea ethidium (23) +29 -7 +740%

3,8-bisurea ethidium (24) +11 0 -400%

3-pyrrole ethidium (25) +35 -16 +230%

8-pyrrole ethidium (26) +28 -12 +730%

3,8-bis pyrrole ethidium (27) +18 0 0%

8-urea-ethidium-6'-neamine (34) +28 -17 +630%

3-urea-ethidium-6'-neamine (35) +26 -17 +340%

8-urea-ethidium-6'-3'deoxyneamine (36) +26 -18 +690%

3- urea-ethidium-6'-3'deoxyneamine (37) +26 -14 +370%

132

Interestingly, conjugation at the 3 position of ethidium is more favorable than

conjugation at the 8 position (compare 34 to 35 and 36 to 37). This is the same

trend as observed with the “simple” urea-derivatives 22 and 23 (Table 5.2).

The deoxyneamine derivatives have about a 10-fold higher RRE affinity

compared to the neamine conjugates (compare 34 to 36 and 35 to 37). This is

the same trend as observed for deoxytobramycin derivatives, where the removal

of hydroxyl groups increases the HH16 affinity of most derivatives.36

Importantly, compounds 34 – 37 have a much higher affinity to the RRE

compared to ethidium, but the same or lower affinity for CT DNA. These

compounds, therefore, have a high RRE specificity, relative to CT DNA.

Included in Table 5.2 is the fully characterized RRE ligand “Neo-S-acridine”

(presented in the next section).169 This compound is typical of most glycoside-

intercalator conjugates. Neo-acridine has a high RRE affinity, but also binds to all

“non-specific” nucleic acids with very high affinity, and therefore exhibits a low

RRE specificity. Compared to neo-acridine, 3-urea-deoxy-neamine ethidium (37)

has, at least, a 1,000-fold higher RRE specificity (relative to CT DNA). 37 has

also been evaluated for its affinity to poly r(A) – r(U), poly r(I) – r(C), tRNAMIX, and

tRNAPhe (Table 5.3). It has about the same or an even lower affinity to all of these

“non-specific” RNAs as compared to CT DNA (Table 5.3). The C50 value for

RRE66 (0.25 µM) represents the minimum theoretical C50 value if 2 high-affinity

binding sites for 37 are on the RRE (since 1 µM of 37 was used). When a 5-fold

133

lower concentration of 37 is used (200 nM), the C50 for RRE66 is 50 nM. This

confirms that 37 has 2 high-affinity binding sites (Kd < 20 nM) on the RRE66. This

represents a least a 10,000-fold higher affinity for the RRE as compared to

tRNAPhe (Table 5.4). In collaboration with Q. Liu (Tor Lab), we are continuing the

investigation of the RRE affinity and specificity of compounds 34 – 37.

Table 5.4: C50 values for 1 µM of 37.Nucleic Acid C50 (µM)

CT DNA 18*

Poly r(A) – r(U) 10*

Poly r(I) – r(C) 65*

tRNAMIX 17**

tRNAPhe 250***

RRE66 0.25***

* Reported in base-pairs** Reported in bases*** Reported in strands

New derivatives of ethidium bromide have provided a fascinating collection of

small molecules. The electronic structure of ethidium has, for the first time, been

thoroughly characterized and has provided some surprising revelations. The

presence of carbons atoms with partial anionic charges in the phenanthridinium

core of ethidium may be an important determinant of base stacking interactions.

Other groups have shown that ethidium thermally stabilizes duplex RNA, more so

than all other “classic” intercalating agents.170 Ethidium’s unusual electronic

features may eventually explain this finding. The substitution of ethidium’s

134

exocyclic amines with other functional groups significantly perturbs its electronic

structure. Most of the new derivatives have much lower nucleic acid affinities and

may, therefore, prove less mutagenic as compared to ethidium. The conjugation

of neamine and deoxyneamine to ethidium, especially through the 3 amine,

however, provides compounds with truly impressive RRE affinities and

specificity. The anti-HIV activity of these conjugates and of the “simple” urea

derivatives 23 and 24 are currently being evaluated at the Center for AIDS

Research, UCSD.

Much has been learned about ethidium, but some mysteries still remain. For

example, upon binding DNA, the visible absorbance spectrum of ethidium red-

shifts and the emission spectrum blue-shifts (Table 5.3). This suggests that upon

intercalation, the energetic gap between the ground-state and the first excited (or

“Frank Condon”) state becomes smaller, but the energy difference between the

emissive state and the ground state becomes larger. The quantum yield, upon

intercalation, increases by 700% (Table 5.3). There is still no accepted

explanation for all three of these photophysical changes. The new derivatives of

ethidium show diverse spectral changes upon binding CT DNA (Table 5.3).

Trends are apparent between the 3 and 8 positions, and differences between

functional groups are also observed. An astute reader may, one day, use this

data to construct a model that can fully explain the spectral changes of ethidium

upon intercalation.

135

6.0 Aminoglycoside-Based Ligands

Aminoglycosides are highly selective in their preferential binding of RNA over

DNA, but show much less discrimination between different RNA molecules.

Aminoglycosides bind to a wide range of unrelated RNAs including simple duplex

RNA,118 16S and 18S rRNAs,30,171 mRNA transcripts,172 tRNA,70 and a large

number of autocatalytic RNAs.173 This general affinity for RNA is related to the

ability of aminoglycosides to bind the major groove of RNA through electrostatic

interactions mediated by ammonium groups.36,118,174-175,176 Aminoglycosides are

selective for A-form over B-form type duplexes, and can facilitate the

conformational change of calf thymus (CT) DNA from a B-form into an A-form

double helix.176

In 1993, an important demonstration of viral RNA targeting was established for

the aminoglycoside family of antibiotics.82 Zapp et al. reported that certain

aminoglycosides bind to the RRE and displace Rev with moderate activities (1 µM

<IC50<100 µM).82 Later studies, by Hendrix et al., showed that the most active

aminoglycoside, neomycin B, binds non-specifically to multiple sites on the

RRE.175 Unfortunately, adverse side effects and poor anti-HIV activities have

prevented the use of aminoglycosides as anti-viral agents. In attempts to improve

these properties, modifications to aminoglycosides have been made. The

resulting compounds exhibit dramatically enhanced RRE affinity, interesting

trends in nucleic acid specificity, and some derivatives possess highly potent anti-

HIV activities.

136

6.1 Neomycin-Acridine Conjugates

Neomycin B was reported to bind near the purine-rich bulge of the RRE.82

Flanking this internal bulge is a single adenosine bulge (Figure 6.0). Intercalating

agents are known to preferentially bind to duplex regions containing single-base

bulges.139 Indeed, both 9-aminoacridine and ethidium bromide bind, with high

affinity to the RRE (Section 3.4 Figures 3.13 & 3.14). In an attempt to create an

RRE ligand that will simultaneously recognize both the internal bulge and the

single-base bulge of the RRE, 9-aminoacridine was conjugated to neomycin B

(Figure 6.1). In collaboration with Dr. Sarah Kirk, neo-S-acridine (39) was the first

aminoglycoside – acridine conjugate to be synthesized and fully characterized for

RNA binding.177

Figure 6.0: High-affinity Rev binding site (bold),43 enzymatic foot-printing ofneomycin B (*),82 and potential high-affinity intercalation sites (arrows).139

U

GA

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

G

C

U

A

AC A

AUG

A

* **

*

*Neomycin B Footprinting

Potential High-Affinity Intercalation Sites

137

Figure 6.1: Structures of three Neo-acridine conjugates. See section E 6.1 forthe synthesis and characterization of compounds 38 and 40.

Native gel-shift electrophoresis and enzymatic footprinting experiments suggest

that neo-S-acridine binds to a single high-affinity site on the RRE and displaces

the Rev peptide with a 10-fold higher activity compared to neomycin B (Figure

6.2 A & B). Conducted in the presence of a large excess of tRNA, native gel-shift

experiments show that Neo-S-acridine displaces ½ of the Rev peptide at about

0.75 µM (lane 16, Figure 6.2 A) and the IC50 value for neomycin B is

approximately 10 µM (lane 9 Figure 6.2 A). These results are similar to the IC50

values measured by the solid-phase assay when RevFl displacement

experiments are conducted in the presence of a large excess of tRNA (Table

6.1).

SNH

NNH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

HN

N

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

S NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

NH

N

Neo-S-acridine (39)

Neo-N-acridine (38)

Neo-C-acridine (40)

S

138

(a)

(b)

Figure 6.2: (a): Conducted in the presence of a large excess of tRNAMIX (50 µg /mL), gel-shift mobility assays have been employed to qualitatively study theinteractions of neo-S-acridine (39) with the RRE 66. All lanes contained 25 nM ofRRE66 with a trace of 5'-32P labeled RRE, lanes 2–19 contain 2 µM Rev-IA.Lane 1, RRE only; lane 2 and 19, RRE + Rev only; lanes 3–10, RRE + Rev with0.5 µM, 1 µM, 1.25 µM, 2.5 µM, 3.75 µM, 5 µM, 7.5 µM, 10 µM neomycin B,respectively; lanes 11–18 RRE + Rev with 25 nM, 50 nM, 100 nM, 200 nM, 500nM, 750 nM, 1.25 µM, 2.5 µM neo–acridine, respectively. While not observed forneomycin B, a binary neo–acridine-RRE complex is formed at 2.5 µM (lane 18).(b) Summary of enzymatic footprinting experiments conducted at 2.5 µM of 39and in the presence of a large excess of tRNA. Three ribonucleases (RNase T1,RNase A, and RNase V1) were used in this study.177 A second equivalent ofNeo-S-acridine can bind to the RRE at higher concentrations (10 µM) andenzymatic footprinting indicates that this second binding site is in the sameregion as the first binding site.177

139

Native gel shift electrophoresis suggests that a single equivalent of neo-S-

acridine is sufficient to displace RevFl (Figure 6.2 A). Upon complete

displacement of Rev, a complex is formed between RRE66 and neo-S-acridine,

having ½ the mobility of the Rev-RRE complex (Figure 6.2 A, lane 18). Since

Neo-acridine has approximately ½ the charge and mass of Rev-IA, we interpret

this as a 1:1 complex of Neo-S-acridine and RRE 66. Three different

ribonucleases were utilized to generate a footprint of this binary complex (results

summarized in Figure 6.2 B). As expected, neo-S-acridine binds to the RRE in

the same region as the Rev peptide.

Table 6.0: IC50 and Ki values for RevFl displacement at differentRRE66 concentrations (by fluorescence anisotropy, Section 3.1).

Compound / Value 10 nM RRE66 120 nM RRE66

Rev IA / IC50 14 ± 4 nM 210 nM ± 30 nM

Rev IA / Ki*

2.0 ± 1.0 nM 2.1 nM ± 0.6 nM

Neo-S-Acridine / IC50 17 ± 3 nM 270 ± 50 nM

Neo-S-Acridine / Ki*

2.7 ± 0.8 nM 4.5 ± 1.0 nM

Neo-Neo / IC50 17 ± 3 nM 440 ± 30 nM

Neo-Neo / Ki*

2.7 ± 0.8 nM 7.3 ± 0.6 nM

Neomycin B / IC50 0.9 ± 0.3 µM 7.0 ± 1 µM

Neomycin B / Ki*

0.22 ± 0.07 µM 0.18 ± 0.03 µM

* Ki calculated from IC50, assuming a single binding site displaces RevFland that the RevFl-RRE66 affinity is Kd = 3 nM

In the absence of competitor nucleic acids, neo-S-acridine binds to the RRE with

about the same affinity as the unlabeled Rev peptide “Rev-IA” (Table 6.0 and

Table 6.1). Since a single Neo-S-acridine binding site can displace Rev, Ki

values can be calculated from Rev-RRE IC50 values (as described in Figure 3.4

and Section E 6.1.2). Both neo-S-acridine and the Rev peptide bind to the RRE

140

66 with high affinity (Table 6.0). Ki values calculated for neo-S-acridine and Rev-

IA are similar even at different concentrations of RRE 66 (Table 6.0). A Ki = 3 ± 2

nM is calculated for neo-S-acridine (39), and a Ki = 2 ± 1 nM is calculated for the

unlabeled Rev peptide Rev-IA.

Based on fluorescence anisotropy measurements, neomycin B has a slightly

lower Ki value when calculated at 120 nM of RRE compared to 10 nM of RRE

(Table 6.0). This effect is much more pronounced for tobramycin (Table 6.2).

This suggests that at higher RRE concentrations, the “regular” aminoglycosides

have additional (lower affinity) binding sites on the RRE that are capable of

displacing the Rev peptide. Indeed, other groups have concluded that neomycin

B has multiple binding sites and that more than one equivalent of neomycin is

necessary to displace Rev from the RRE.89,175 As opposed to the unmodified

aminoglycosides, neo-S-acridine has a slightly higher Ki value at higher RRE

concentrations (Table 6.0). This effect is also seen for the other two neomycin-

acridine conjugates 38 and 40 (Table 6.2) and suggests that neo-S-acridine has

one or more non-inhibitory binding sites on the RRE that become more apparent

at higher RRE concentrations. Interestingly, this effect is much more pronounced

for the neomycin B dimer “neo-neo” (presented later). In contrast, the Ki value for

the unlabeled Rev peptide (Rev-IA) is consistent at both RNA concentrations

(Table 6.0).

141

At the time of publication, the binding between Neo-S-acridine and the RRE66

was the highest affinity interaction between a natural RNA and synthetic ligand

ever reported.177 Two other groups have also reported aminoglycoside –

intercalator conjugates, but neither group has disclosed compounds with such a

high RNA affinity.178,179 Gel shift electrophoresis and enzymatic footprinting

experiments indicate that neo-S-acridine forms a well-defined complex with the

RRE (Figure 6.2), and its affinity for the RRE is similar to that of the Rev peptide

(Table 6.0). It was surprising, therefore, to discover that neo-S-acridine has a

very low specificity for the RRE.

To evaluate the RRE specificity of neo-S-acridine, the solid-phase assay (Section

6.2) was used to measure IC50 values in the presence or absence of other

nucleic acids (Table 6.1). By dividing the average IC50 values measured in the

presence of competing nucleic acids by the IC50 values measured in the absence

of competing nucleic acids, a “selectivity ratio” is determined (Table 6.1). The

lower this value is, the more selective the ligand is for the RRE. The absolute

RRE specificity of the compound, as we have defined it (Section 1.6), is a

combination of both the RRE selectivity and the absolute affinity for the RRE.

Consistent with fluorescence anisotropy, the solid-phase assay indicates that

both neo-S-acridine and the unlabeled Rev peptide have approximately the same

affinities for the RRE (Table 6.1). Upon addition of CT DNA or tRNAMIX, however,

the activity of neo-S-acridine drops by 10 – 40 fold. Under the same conditions,

142

the IC50 values for the unlabeled Rev peptide drop by only 2 – 3 fold (Table 6.1).

This indicates that neo-S-acridine has a much higher affinity to non-specific

RNAs and DNAs as compared to the Rev peptide. Indeed, by monitoring the

fluorescence of either the RevFl peptide or neo-S-acridine as a function of DNA

concentration, a low affinity between RevFl and CT DNA is measured (Kd = ~250

µM, Table 3.1), while a very high affinity between neo-S-acridine and CT DNA is

seen (Kd < 100 nM, Table 5.2).

Table 6.1: IC50 Values(µM) according to solid-phase assay.*

Compound IC50 IC50 withC.T. DNA**

IC50 withtRNAmix***

RRESelectivityRatio****

Neo-N-acridine (38) 0.040 0.15 1.2 17

Neo-S-acridine (39) 0.040 0.45 1.6 26

Neo-C-acridine (40) 0.040 1.7 2.3 50

Tobra-N-acridine (41) 0.24 4.0 2.8 14

KanaA-N-acridine (42) 0.61 2.5 2.2 3.9

Neo-Neo (43) 0.050 0.11 4.8 49

Tobra-Tobra (45) 0.13 0.20 2.6 11

KanaA-KanaA (47) 0.24 0.60 2.5 6.5

Rev-IA 0.035 0.080 0.085 2.3

* Approximate error for each value +/- 35% of IC50 value.** 230 µM (bases) of a tRNA mixture (Sigma type X) included.*** 115 µM (base pairs) of CT DNA included.**** Selectivity ratio = (average IC50 in the presence of DNA and tRNAMIX)

(IC50 in the absence of other nucleic acids)

The RRE specificity of neo-S-acridine can be improved by decreasing the linker

length between its two moieties. Neo-N-acridine (38) has approximately the

same RRE affinity, but compared to neo-S-acridine, it retains more of its activity

in the presence of a vast excess of DNA and tRNAMIX (Table 6.1). The three neo-

143

acridine conjugates 38 – 40 each have the same RRE affinity, so the differences

in selectivity are directly proportional to the differences in RRE specificity (Table

6.1). An obvious trend in RRE specificity is observed for these three conjugates

(Figure 6.1). Neo-N-acridine has the shortest linker and an improved RRE

specificity as compared to neo-S-acridine, and neo-C-acridine has the longest

linker and the worst RRE specificity (Table 6.1). The linker length, therefore, is

inversely proportional to RRE specificity exhibited by these compounds. The

short linker length provided by neo-N-acridine may provide fewer degrees of

freedom to the conjugate and, therefore, fewer nucleic acids can bind to it with

high affinity.

Ethidium bromide displacement experiments (as described in Section 3.3)

confirm that both neo-C-acridine and neo-S-acridine have much higher affinities

to “non-specific” competitors (including calf thymus DNA and poly r(A) – r(U)) as

compared to neo-N-acridine (Table 6.2). One interesting difference does exist,

however, between tRNAMIX and poly r(A) – r(U) (compare Table 6.1 and 6.2).

Neo-S-acridine has the “optimal” linker length for binding to poly r(A) – r(U)

(Table 6.2). It has 3- and 13-fold higher affinities to poly r(A) – r(U) as compared

to neo-C-acridine and neo-N-acridine, respectively (Table 6.2). This may indicate

that neo-N-acridine has too short of a linker for both the acridine and the

neomycin moieties to bind the duplex simultaneously. On the other hand, the

extremely long linker of neo-C-acridine may impose an entropic disadvantage for

the simultaneous binding of both moieties to poly r(A) – r(U).

144

Table 6.2: RevFl-RRE66 IC50 values (µM) and approximate RRE affinity (Ki), andthe IC50 values for ethidium displacement from RNA and DNA.

Compound

RevFl(10nM)

+ RRE66(100nM)

IC50

RRE66Affinity**

(Ki)(100 nM)

RevFl(10nM)

+ RRE66(10nM)

IC50

RRE66Affinity***

(Ki)(10 nM)

Ethidium+ poly

r(A)-r(U)IC50****

Ethidium+ C.T. DNA

IC50****

Kanamycin A (1) 750 1.9x10-05 100 2.5x10-05 450 >15,000

Tobramycin (3) 45 1.1x10-06 10 2.5x10-06 25 >800

Neomycin B (5) 6.7 1.7x10-07 0.9 2.2x10-07 6.0 60

Neo-N-acridine (38) 0.23 3.5x10-09 0.016 2.5x10-09 0.65 0.85

Neo-S-acridine (39) 0.25 4.0x10-09 0.017 2.8x10-09 0.05 0.14

Neo-C-acridine (40) 0.27 4.5x10-09 0.017 2.8x10-09 0.13 0.08

Tobra-N-acridine (41) 0.33 6.0x10-09 0.030 6.0x10-09 1.3 0.49

KanaA-N-acridine (42) 0.48 9.8x10-09 0.045 9.8x10-09 4.4 2.1

Neo-Neo (43) 0.40 7.8x10-09 0.017 2.8x10-09 0.067 0.65

Neo-N-Neo (44) 0.26 4.3x10-09 0.015 2.3x10-09 0.035 0.45

Tobra-Tobra (45) 0.64 1.4x10-08 0.040 8.5x10-09 0.61 1.0

Tobra-N-Tobra (46) 0.80 1.8x10-08 0.043 9.3x10-09 0.47 1.5

KanaA-KanaA (47) 1.2 2.8x10-08 0.050 1.1x10-08 6.0 12.0

* Random errors are less than or equal to ± 35% of each IC50 value.** Calculated from Rev-RRE IC50 values at 10 nM of RRE, these estimates assume that 1equivalent of the small molecule is sufficient to displace RevFl.*** Calculated from Rev-RRE IC50 values at 100 nM of RRE, these estimates assume that 1equivalent of the small molecule is sufficient to displace RevFl.**** 1.3 µM of ethidium bromide + 300 nM (base pairs) of the duplex nucleic acid.

As indicated by the solid-phase assay, ethidium displacement experiments show

that neo-C-acridine has the highest affinity for CT DNA, followed by neo-S-

acridine, and neo-N-acridine has the lowest affinity (among the neomycin-

acridine conjugates). The longer linker provided by neo-C-acridine may give DNA

better “access” to the intercalating agent, as neomycin B is an RNA-selective

molecule and acridine is essentially a non-specific intercalating agent.

145

Another group has also synthesized intercalator-glycoside conjugates with

variable linkers.178 Their approach, however, involves incorporation of a variable-

length methylene spacer to link the 6" amine of paromomycin to either pyrene or

thiazole orange. Subsequent evaluation of these compounds as A-site

antagonists indicated that as the number of atoms in the linker increases, the A-

site affinity decreases.178 This is a very different trend as observed with our

conjugates. All of the neo-acridine conjugates have, within error, the same RRE

affinities (Table 6.1 and Table 6.2). In addition, the neo-acridine conjugates with

a longer linker have a higher DNA affinity. It is possible that highly flexible,

hydrophobic linkers (including poly methylene chains) introduce a significant

penalty for RNA binding. Indeed, other groups have found that linker composition

is an important determinant for RNA binding.69b Ethylene glycol and thioether

linkers are well-hydrated flexible linkages that should not, by themselves, exhibit

significant favorable (or unfavorable) interactions with RNA.

6.2 Other Aminoglycoside-Acridine Conjugates

We have conjugated aminoacridine to the 6'' hydroxyls of both tobramycin and

kanamycin A (Figure 6.3). RevFl displacement experiments indicate that tobra-N-

acridine has a 300-fold higher affinity, and kanaA-N-acridine has over a 2,200-

fold higher affinity to the RRE, as compared to their “parent” aminoglycosides

(Table 6.2). In comparison, the conjugation of acridine to neomycin B only

increases its affinity to the RRE by a factor of 65 (Table 6.2).

146

Figure 6.3: Structures of kanamycin A and tobramycin acridine-conjugates. Seesection E 6.2.1 for synthesis and characterization of 41.

KanaA-N-acridine (42) has a slightly lower RRE affinity, but exhibits a much

better RRE selectivity ratio as compared to the other aminoglycoside-acridine

conjugates (Table 6.1). Ethidium displacement experiments confirm that,

compared to the other acridine conjugates, kanaA-N-acridine has a much lower

affinity to other nucleic acids, including CT DNA and poly r(A) – r(U) (Table 6.2).

Compared to neo-S-acridine, kanaA-N-acridine has about a 4-fold lower affinity

for the RRE, but a 90-fold lower affinity to poly r(A) – r(U) and a 15-fold lower

affinity to CT DNA (Table 6.2). This illustrates why kanaA-N-acridine is

significantly more selective for the RRE as compared to the neo-acridine

conjugates (Table 6.1).

Tobra-N-acridine (41) has an RRE affinity and selectivity that is intermediate

between the neo-acridine conjugates and kanaA-N-acridine (Table 6.1 and Table

6.2). It appears, therefore, that there is an inverse relationship between the RRE

affinity and the RRE selectivity of these compounds: as the affinity for the RRE

O

NH2

HO

H2N

H2NO

HOHOO

O

H2N OH

NH2

NH

Tobra-N-acridine (41)

O

NH2

HO

H2N

H2NO

HOHOO

O

HO OH

NH2

NH

N

KanaA-N-acridine (42)

HO

N

147

increases, the selectivity for the RRE typically decreases. This trend is observed

in two other families of compounds and may, therefore, reflect a general trend

(presented later).

Interestingly, the calculated RRE66 affinities (Ki) for kanaA-N-acridine and tobra-

N-acridine are identical at 10 nM and 100 nM of RRE66 (Table 6.2). This

suggests that these compounds, unlike all three neo-acridine conjugates, do not

have multiple non-inhibitory binding sites on the RRE over this concentration

range. This is fully consistent with the improved RRE selectivity of these

compounds as compared to the neo-acridine conjugates (Table 6.1), and

illustrates how the non-inhibitory binding sites on the RRE can exert the same

effect as the addition of competitor nucleic acids. These results illustrate the

importance of varying the nucleic acid concentration when evaluating the affinity

and specificity of small molecule - RNA binding interactions.

6.3 Dimeric Aminoglycosides

The synthetic dimerization of aminoglycosides was first reported by Wang and

Tor.180 Neomycin B , kanamycin, and tobramycin dimers, like the parent

compounds, bind to a wide range of different RNAs, including the HH16

ribozyme,180 tRNAPhe,70 a dimerized A-site,181 and the Tetrahymena ribozyme.182

No systematic studies, until now, have examined the RNA specificity of these

compounds.

148

The RRE contains multiple binding sites for neomycin B (Section 6.1). 89,175

Neomycin dimers could, in principal, simultaneously occupy multiple binding

sites, resulting in entropic and enthapic advantages relative to the monomeric

species. Indeed, the RRE affinity of neomycin B is enhanced 20 – 80 fold upon

dimerization (Table 6.0). As for the neo-acridine conjugates, neo-neo (43)

exhibits a similar RRE affinity as the unlabeled Rev peptide (Table 6.0). A range

in affinity enhancement is reported for all the aminoglycoside dimers because a

lower apparent affinity (Ki) is measured at 100 nM versus 10 nM of RRE (Table

6.2). This suggests that the dimers, like the neo-acridine conjugates, have

multiple non-inhibitory binding sites on the RRE that become more apparent at

higher RRE concentrations. In some ways, these non-inhibitory binding sites act

as competing nucleic acids and, at higher concentrations of RRE, serve to

decrease the Rev-RRE inhibition activities of these compounds. The Ki values

measured at 10 nM of RRE, therefore, better reflect the actual affinity of these

compounds for the Rev binding site (Table 6.2).

Unlike the neo-acridine conjugates, neo-neo (Figure 6.4) maintains most of its

activity even in the presence of a vast excess of CT DNA (Table 6.1). A relatively

low affinity for CT DNA is confirmed by ethidium displacement experiments

(Table 6.2). Neo-neo, however, compared to the neo-acridine conjugates, loses

much more of its Rev-RRE inhibition activity in the presence of tRNAMIX (Table

6.1).

149

Figure 6.4: Dimeric aminoglycosides evaluated for RRE affinity and specificity.See Section E 6.3.1 for the synthesis and characterization of neo-N-neo (44).

Overall, neo-neo exhibits the worst RRE selectivity (relative to tRNAMIX) as

compared to all other high-affinity RRE affinity ligands (Table 6.1). Its high affinity

for non-specific RNA may explain why the differences in Ki values at 10 nM

SS

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

SS

H2N

NH2O

OO

NH2

HO

OH O

HO O

OH

H2N

OOHOH

H2N

H2N

SO

H2N

NH2O

OO

NH2

HO

OH O

HO O

OH

H2N

OOHOH

H2N

H2N

SO

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

S S

O

NH2

HOH2N

H2NO

HOHO

O

O

H2N OH

NH2

O

H2N

OHNH2

NH2O

OH HOO

O

NH2HO

H2N

O

NH2

HOH2N

H2NO

HOHO

O

O

H2N OH

NH2

O

H2N

OHNH2

NH2O

OH HO

O

O

NH2HO

H2N

O

NH2

HOH2N

H2NO HOHO

O

O

HO OH

NH2

O

H2N

OHNH2

NH2O

OH HO

O

O

OHHO

H2N

HOOH

Neo-Neo (43)

Neo-N-Neo (44)

Tobra-Tobra (45)

Tobra-N-Tobra (46)

KanaA-KanaA (47)

SO S

OS S

SSS S

SO S

OS S

150

versus 100 nM of RRE are much more pronounced when compared to the neo-

acridine conjugates (Table 6.1).

Dimerization effectively doubles the total charge and hence, increases the affinity

of neomycin to all nucleic acids (tested so far). Ethidium displacement

experiments indicate that, compared to neomycin B (5), neo-neo (43) has a 90-

fold higher affinity for both duplex DNA and RNA (Table 6.2). Since ethidium

bromide has approximately the same affinity and the same number of binding

sites on both CT DNA and poly r(A) – r(U),87 the ethidium displacement values

are, in this case, comparable between CT DNA and poly r(A) – r(U). The ratios of

IC50 values for ethidium displacement from poly r(A)-r(U) and CT DNA are 1 : 10

for both neo-neo and for neomycin B (Table 6.2). This indicates that, indeed, the

RNA over DNA selectivity of neomycin is maintained upon dimerization.

The 90-fold higher affinity to simple duplex nucleic acids may explain why the

dimerization of neomycin B dramatically lowers its RRE selectivity (Table 6.1).

This increase is larger than the affinity enhancement observed for the RRE66 (20

– 80 fold). Other aminoglycoside dimers do, however, show much better RRE

selectivities, as compared to neo-neo.

Tobra-tobra (45) (Figure 6.4) has a 80-300 fold higher affinity to the RRE as

compared to tobramycin (3) (Table 6.2). As observed for neo-neo, tobra-tobra

retains most of its Rev-RRE inhibition activity in the presence of excess CT DNA,

151

but still loses activity in the presence of other RNAs (Table 6.1). The RRE

selectivity of tobra-tobra is, however, much better than neo-neo (Table 6.1). It is

interesting that compared to neo-neo, tobra-tobra has about a 5-fold lower RRE

affinity, but it is more active for Rev-RRE inhibition in the presence of tRNAMIX

(Table 6.1). This indicates that tobra-tobra has a higher specificity for the RRE as

compared to neo-neo. Ethidium displacement experiments confirm that upon

dimerization of tobramycin, only a 40-fold increase in affinity for simple duplex

RNA is seen (Table 6.2). This is much smaller than the 80 – 300 fold increase for

the RRE, and may explain the improved RRE selectivity of tobra-tobra, as

compared to neo-neo (Table 6.1).

The kanamycin A dimer (47), as compared to kanamycin A (1) has a 700 – 2,200

fold higher affinity for the RRE, but only a 75-fold higher affinity for poly r(A) –

r(U) (Table 6.2). Compared to all the aminoglycoside dimers evaluated, the

kanaA-kanaA dimer has the best RRE selectivity (Table 6.1). Ethidium

displacement experiments confirm that, compared to all of the high-affinity RRE

ligands (Kd < 10 nM), the kanaA-kanaA dimer has the lowest affinity to simple

duplex DNA and RNA (Table 6.2). These trends are similar to those observed for

the aminoglycoside-acridine conjugates, where the dimers with lower RRE

affinity exhibit much better RRE selectivity. The kanaA-kanaA dimer exhibits a

higher Rev-RRE inhibition activity in the presence of a tRNA mixture, as

compared to some compounds that have 2-5 folder higher affinities for the RRE

(Table 6.1). This illustrates the inherent difficulty of measuring the RNA affinity of

152

a small molecule in the presence of non-specific competitor nucleic acids, as

these measurements reflect more the specificity of the interaction as opposed to

the actual affinity.

Another group has recently published the binding of neo-neo and tobra-tobra to

RREIIB.183 Tok reports that the dimerization of neomycin increases its RRE

affinity by 17-fold (compared to the 80-fold increase reported here). There are

many potential reasons for this difference. Tok uses the displacement of a

fluorescent aminoglycoside to determine RREIIB affinities of the dimeric and

monomeric aminoglycosides. 600 nM of the RREIIB and curve-fitting was used to

measure an RRE affinity (Ki) of 10 nM for neo-neo. This should not, theoretically,

be possible. Examination of Tok’s raw data suggests that 10 nM of neo-neo

binds to 300 nM of RRE IIB.183 In our experiments, we use 10 nM of the RRE66

to measure an RRE affinity (Ki) of 2.5 nM (Table 6.2). At higher RRE66

concentrations (100 nM), the apparent RRE affinity is significantly higher (Ki = 7.8

nM, Table 6.2). Tok’s data provides additional evidence that at high RRE

concentrations, the low-affinity non-specific binding sites exhibit more influence

on the displacement isotherms, and in extreme cases can cause aggregation.

Tok measures an affinity for tobra-tobra that is approximately the same as

monomeric tobramycin (Ki = 3.9 µM).183 Our experiments, however, indicate that

tobra-tobra binds to the RRE with a Ki = 8.5 nM (Table 6.2). There is no obvious

explanation for this 500-fold difference. It is possible, of course, that tobra-tobra

does not bind to the same sites on the RREIIB versus the RRE 66. This does

153

not, however, seem to be a plausible explaination, as Tok reports the same RRE

affinities for each of the monomeric species (neomycin B and tobramycin) as we

have measured (Table 6.2).183

Linker lengths have been shown to play a major role in the RRE specificity of the

Neo-acridine conjugates, where shorter linkers provide higher RRE specificity

(Section 6.1). Neomycin B and tobramycin dimers were, therefore, synthesized

with shorter linker lengths as compared to neo-neo and tobra-tobra (Figure 6.4).

Compared to neo-neo, neo-N-neo has approximately a 2-fold higher RRE affinity,

but only a slightly higher affinity for CT DNA. This suggests that compared to CT

DNA, neo-N-neo is more selective for the RRE. Neo-N-neo also has, however, a

2-fold higher affinity for the simple duplex RNA, as compared to neo-neo (Table

6.2). This suggests that, relative to poly r(A) – r(U) , neo-N-neo has the same

RRE selectivity as neo-neo. Surprisingly, no significant differences are observed

between tobra-tobra and tobra-N-tobra (Table 6.2).

The dimeric aminoglycosides have, in general, a lower RRE affinity as compared

to the corresponding acridine conjugates (Table 6.2). This suggests that, under

these conditions, the acridine moiety contributes more binding energy to the

conjugate as compared to the aminoglycoside moiety. This is somewhat

surprising, as 9-amino acridine appears to have higher RRE affinity as compared

to neomycin B (not shown). It is possible that the combination of binding modes

(intercalation and electrostatic groove binding) is the reason for the unusually

154

high nucleic acid affinity of the acridine conjugates. The acridine moiety provides

significant binding energy to the conjugate, but it also confers DNA affinity to the

aminoglycoside moiety. Bringing these moieties closer together partially reverses

this trend and restores the RNA over DNA selectivity of the conjugate.

In summary, new high-affinity RRE ligands have been synthesized by the

conjugation of acridine or dimerization of aminoglycosides. The thorough study of

these ligands with a variety of nucleic acids has revealed interesting trends in

affinity and specificity. Neomycin-based ligands exhibit the highest RRE affinities,

but display high affinities to many other nucleic acids as well. These neomycin-

based ligands, therefore, exhibit the lowest RRE selectivity. Among all of the

“high affinity” small molecule RRE ligands presented, the kanamycinA-based

derivatives have the lowest RRE affinities, but exhibit the highest RRE

selectivities. The RRE affinities and specificities of the tobramycin derivatives are

intermediate between those of the neomycin and kanamycin A derivatives. This

indicates that, for these compounds, an inverse relationship exists between the

RNA affinity and selectivity of each small molecule. This trend highlights the true

challenge of this field: it is relatively difficult to synthesize a new ligand that has

both high affinity and high selectivity for the desired RNA target.

6.4 Guanidinoglycosides

The guanidinium group plays a key role at many RNA–protein binding interfaces,

including the complexes formed between transcriptional elongation factors with

155

mRNA, tRNA synthetases with tRNAs, ribosomal proteins with rRNA, and viral

regulatory proteins with their cognate RNA binding sites.184 In contrast to amines,

guanidines are highly basic, planar, and exhibit directionality in their H-bonding

interactions. Effective recognition between the Rev and the RRE is mediated,

largely, by the guanidinium groups present in the Rev’s arginine-rich domain.46,53

Substitution of arginine at position 35, 38, 39, or 44 with lysine leads to an over

20-fold loss of RRE binding in vivo.185 Since many competitor nucleic acids are

present in vivo, this represents a large decrease in RRE specificity. We

speculated, therefore, that the transformation of the amines on the

aminoglycosides into guanidinium groups would increase both the affinity and

specificity of these compounds for the RRE.

Figure 6.5: Structures of some aminoglycosides and guanidinoglycosides.

In collaboration with Prof. Murray Goodman and Dr. Tracy Baker, we

guanidinylated five aminoglycosides and used fluorescence anisotropy to

evaluate the RRE affinity of the resulting “guanidinoglycosides” (Figure 6.5 and

Table 6.3).186 IC50 values were measured at both 10 nM and 100 nM RRE

OHOR3

OR3 R3O

OR1

R3OH

R2OH

HO HO

OH

HTobramycin (3)

R1

Kanamycin A (1)

OHKanamycin B (2)

R2

OH

NH2

NH2

R3

OHGuanidino-kanamycin A (48) OH NH(C=NH)NH2

OHGuanidino-kanamycin B (49) NH(C=NH)NH2

H NH(C=NH)NH2Guanidino-tobramycin (50) NH(C=NH)NH2

NH(C=NH)NH2

NH2

NH2

NH2

Glycoside

O

R2

R2 O

HOHO

OO

R2

OH

R1

OHO

HOO

HO

HO

R2

R2

OH

R1

Paromomycin (4)

Neomycin B (5)

R2

OHGuanidino-paromomycin (51)

Guanidino-neomycin B (52) NH(C=NH)NH2

NH2

NH(C=NH)NH2

NH(C=NH)NH2

NH2

NH2

Glycoside

156

(consistently using 10 nM of RevFl) (Table 6.3). By taking the ratio of these IC50

values, some information regarding the small molecule – RRE binding

stoichiometries can be gleaned (Table 6.3). A lower ratio suggests that at higher

RRE concentrations there are more binding sites capable of displacing the

peptide.

Table 6.3: IC50 Values (µM)* for RevFl displacement by anisotropy.

Glycoside 10 nM RRE 100 nM RRERatio

(100nM/10nM)

Kanamycin A (1) 100 750 7.5

Guanidino-kanamycin A (48) 8 65 8.1

Kanamycin B (2) 15 80 5.3

Guanidino-kanamycin B (49) 0.7 3.5 5.0

Tobramycin (3) 10 45 4.5

Guanidino-tobramycin (50) 0.8 3.8 4.7

Paromomycin (4) 12 65 5.4

Guanidino-paromomycin (51) 3.4 18 5.3

Neomycin B (5) 0.9 6.7 7.4

Guanidino-neomycin B (52) 0.18 1.3 7.2

* Random errors are less than or equal to ± 30% of each IC50 value.

Surprisingly, the changes in binding stoichiometries for peptide displacement are

significantly different for molecules with similar sizes (including kanamycin A

versus kanamycin B, and paromomycin versus neomycin B). This ratio, however,

does not change for each compound upon guanidinylation, indicating that the

number and relative importance of the binding sites on the RRE are similar for

each guanidinoglycoside and its aminoglycoside precursor. The differences in

157

Rev-RRE IC50 values, upon guanidinylation, are therefore directly proportional to

the change in RRE affinity. Upon guanidinylation of neomycin B and

paromomycin, a 3 – 5 fold increase in RRE affinity is obtained. Upon

guanidinylation of the kanamycin family of glycosides, a 12 – 22 fold increase in

affinity is gained (Table 6.3).

Table 6.4: IC50 Values (µM) for RevFl Displacement by solid-phase assay.*

Glycoside IC50

IC50

with DNA**IC50

with RNA**RRE

SelectivityRatio***

Kanamycin A (1) 700 750 1300 1.5

Guanidino-kanamycin A (48) 55 55 60 1.1

Kanamycin B (2) 90 90 170 1.5

Guanidino-kanamycin B (49) 3 3 4.7 1.3

Tobramycin (3) 45 50 130 2.0

Guanidino-tobramycin (50) 3 3 4.7 1.3

Paromomycin (4) 55 60 110 1.6

Guanidino-paromomycin (51) 15 18 18 1.2

Neomycin B (5) 7 8 16 1.8

Guanidino-neomycin B (52) 1 1.2 4.5 2.9

* Random errors are equal to or less than 30% of each reported value.** 115 µM (base pairs) of either plasmid DNA or poly r(A) – r(U) included, representing a 100-foldmolar excess of bases relative to the RRE. Similar results were also obtained using CT DNA anda tRNA mixture.*** Selectivity ratio = (average IC50 in the presence of DNA and RNA)

(IC50 in the absence of other nucleic acids)

The solid-phase assay (Section 3.2) indicates that guanidinylation also improves

RRE specificity (Table 6.4). By dividing the average IC50 values obtained in the

presence of competitor RNA and DNA by the IC50 values obtained in the absence

158

of other nucleic acids, a selectivity ratio is calculated (Table 6.4). A lower ratio

indicates better RRE selectivity. Upon guanidinylation, this ratio decreases for all

the glycosides (except for neomycin B), showing that, in general, guanidinylation

improves the RRE specificity of the aminoglycosides. Importantly, guanidinylation

does not affect the RNA over DNA selectivity of these glycosides (Table 6.4).

The IC50 values for RevFl displacement do not change significantly upon addition

of an excess of DNA (Table 6.4). This is, to our knowledge, the first indication

that the identity of the basic group is not an important factor for the ability of

aminoglycosides to preferentially bind RNA over DNA.

Analysis of the activity trends, apparent for both the aminoglycosides and the

guanidinoglycosides (Table 6.4), suggests some fundamental relationships

relating the total charge of a ligand to its RNA affinity and specificity. By

comparing either the aminoglycosides or guanidinoglycosides with other

compounds in the same family, it is clear that the total number of basic groups is

roughly proportional to the RRE affinity, and the higher this affinity, the less likely

it is to be selective for the RRE. These same trends are also exhibited by the

dimeric aminoglycosides and aminoglycoside-acridine conjugates and may,

therefore, reflect a general rule. Unlike the dimeric aminoglycosides and

aminoglycoside-acridine conjugates, however, most of the guanidinoglycosides

show improved RRE affinity and specificity as compared to the parent

aminoglycosides. Guanidino-neomycin is the only exception as it is significantly

less selective for the RRE as compared to regular neomycin (Table 6.4). This is

159

consistent with the trends observed for the dimeric aminoglycosides and

aminoglycoside-acridine conjugates (Section 6.3). We conclude, therefore, that

neomycin-based compounds, in general, have a low specificity for the RRE, and

that all modifications, to date, further decrease its RRE selectivity.

Other interesting trends in RRE affinity and selectivity are revealed by the solid-

phase assay (Table 6.4). By comparing aminoglycosides that possess the same

number of basic groups (like kanamycin B, tobramycin and, paromomycin),

additional structure-activity relationships can be examined. For example,

tobramycin and kanamycin B differ only by a single hydroxyl (Figure 6.3), but

tobramycin is found to have a higher RRE affinity as compared to kanamycin B.

This finding is consistent with the ability of hydroxyls to decrease the basicity of

their neighboring amines (giving tobramycin a higher overall charge and a lower

RRE specificity as compared to kanamycin B).36 Upon guanidinylation, the

differences between kanamycin B and tobramycin are lost, because the pKa

values of aliphatic guanidinium groups are significantly less variable than those

of amines.

Guanidino-tobramycin (50) and guanidino-paromomycin (51) each have five

guanidinium groups, but guanidino-tobramycin has approximately a 5-fold higher

affinity for the RRE. It appears, therefore, that the RRE is more selective for

guanidinoglycosides from the kanamycin family as compared to the

guanidinoglycosides from the neomycin family. This may explain why the

160

guanidinylation of the kanamycin family provides a much larger increase in RRE

affinity as compared to the guanidinylation of the neomycin family of glycosides

(Table 6.3 and Table 6.4). It is possible that, compared to guanidino-

paromomycin, the 3-dimensional arrangements of guanidinium groups presented

by guanidino-tobramycin and guanidino-kanamycin better mimic the Rev peptide.

Another possibility is that the guanidinium groups on these compounds are closer

together in space, and therefore exhibit a higher charge density as compared to

guanidino-paromomycin.

Guanidinium groups are highly basic, capable of defined multi-directional

hydrogen bonding, and have planar surfaces that may facilitate stacking

interactions. The higher basicity of the guanidinoglycosides, as compared to the

aminoglycosides, should impart a higher total charge, and therefore higher RNA

affinity as compared to the aminoglycosides. Indeed, the aminoglycosides are

known to be only partially protonated at neutral pH.34 The higher overall positive

charge of the guanidinoglycosides cannot, however, explain their improved RRE

selectivity (this would oppose the general relationships stated above). To

understand why the guanidinoglycosides show improved RRE affinity and

specificity (relative to the aminoglycosides), a number of urea-containing

glycosides were synthesized.188 Urea is an uncharged structural analog of

guanidine. Since both urea and guanidinium solutions can denature nucleic

acids, the urea-glycosides may have bound to RNAs that possess single-

stranded characteristics (such as the purine-rich bulge of the RRE). The urea

161

glycosides are, however, unable to inhibit Rev-RRE binding (through 1 mM). This

indicates that electrostatic interactions are an essential driving force for

guanidinoglycoside – RRE binding. Guanidinoglycoside – RNA binding has,

however, been shown (for the HH16 ribozyme) to be much less dependent on

the total ionic strength of the media, when compared to the aminoglycosides.97

Taken together this suggests that stacking and/or other hydrophobic

interaction(s) mediate binding, but that electrostatic forces are still essential for

guanidinoglycoside-RNA RNA binding.

Figure 6.6: The anti-HIV activities of aminoglycosides and guanidinoglycosidesas evidenced by fractional decrease in plaque formation (as described in Figure4.1). Anti-HIV activities were also evaluated by using ELISA to monitor thedecrease in HIV-1 p24 expression in human PBMC and yielded similar results.Both assays were conducted as described.110 It is possible that at lowconcentrations of the aminoglycosides, the replication of HIV is slightly better.Linear polyamines, like polybrene, are often added to these assays to increasethe infectivity of the HIV particles.110 For these assays, however, no polybrene isadded. This does not affect the IC50 value measured for the positive control (AZTIC50 = ~10 nM).

0.01 0.1 1 10 100

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Guanidino-tobramycinGuanidino-neomycin BTobramycinNeomycin B

Concentration (µM)

Fra

ctio

nalD

ecre

ase

inP

laqu

es

162

The guanidinoglycosides are found to possess significantly improved RRE

affinity, specificity, and anti-HIV activities when compared to the aminoglycosides

(Figure 6.6).187 The guanidinoglycosides exhibit 100-fold improved anti-HIV

activities when compared to the parent aminoglycosides (Figure 6.6). It is

possible that the enhanced anti-HIV activities of these compounds are related to

their improved RRE affinity and specificity. The guanidinoglycosides are shown,

however, to be transported into eukaryotic cell cultures much more efficiently

than aminoglycosides (presented in Section 6.8). In addition, the

guanidinoglycosides bind to other HIV viral sites (including the TAR) with good to

moderate affinities (not shown). The exact mechanism(s) responsible for the anti-

HIV activities of the guanidinoglycosides are currently under investigation in

collaboration with Immusol Inc.

Other groups have conjugated amino acids to aminoglycosides in order to mimic

the Tat protein arginine-rich motif, with similar reported results.189 Conjugation of

arginine to kanamycin B is reported to increase its affinity for the HIV-1 TAR

RNA.189 This synthetic modification effectively doubles the number of basic

groups present.

Figure 6.7: Structures of tobramycin-arginine conjugates (see section E 6.4.1 –E 6.4.2 for synthesis and characterization).

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

OH

NHO

HNHN NH2

HN

O

NH NH

NH2

HN

ONH NH

NH2HN

O

HNHN NH2

NH

O

HN NH

NH2

O

OO

O

O

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

OH

H2NO

HNHN NH2

H2N

O

NH NH

NH2

NH2

ONH NH

NH2NH2

O

HNHN NH2

H2N

O

HN NH

NH2

ace-Tobra-Arg (54)Tobra-Arg (53)

163

Their findings suggest, however, that the guanidinium-containing side chains do

not play the dominant role in RNA binding.189 The α-amines of the amino acids,

in close proximity to the “core” glycoside, are reported to be essential for TAR

binding.189 We have made similar observations with tobramycin-arginine

conjugates in the Rev-RRE system (Figure 6.7). The conjugate “Tobra-Arg” has

the same number of basic groups and a similar RRE affinity as the tobramycin

dimer “Tobra-Tobra” (compare Table 6.2 and Table 6.5). If the α-amines of the

arginine-tobramycin conjugate are blocked by acetylation, the resulting

compound “ace-Tobra-Arg“ has the same number of basic groups as guanidino-

tobramycin (50). Both guanidino-tobramycin and the aceylated tobramycin-

arginine conjugate possess 5 guanidinium groups; their comparison allows us to

establish the contributions made by the flexible, hydrophobic, methylene linkages

present in the arginine-conjugates. We find that the aceylated tobramycin

arginine conjugate (54) has a 30-fold lower RRE affinity as compared to

guanidino-tobramycin (50) (Table 6.5 and 6.3).

Table 6.5: IC50 values and approximate RRE66 affinities (Ki)according to fluorescence anisotropy.

Compound RevFl-RRE66 IC50* RRE66 Ki**

Tobra-Arg (53) 15 nM 2.3 x 10-09

Ace-Tobra-Arg (54) 24 µM 6.0 x 10-06

Kana A-Dab (55) 1.5 µM 3.7 x 10-07

Kana A-Lys (56) 1.7 µM 4.2 x 10-07

γ-Guanidino Kana A-Dab (57) 350 µM 8.6 x 10-08

* IC50 values are for 10 nM of RevFl and 10 nM of RRE66, approximaterandom errors are ± 30% of each reported value.** Calculated from Rev-RRE IC50 values at 10 nM of RRE, these estimatesassume that 1 equivalent of the small molecule is sufficient to displace RevFl.

164

This indicates that a significant energetic penalty is introduced by the flexible

methylene linkages of the amino acids. Both the density and rigidity of the five

guanidino groups presented by 50 should both be greater than those presented

by 54. This may, in part, explain the energetic penalty introduced by the flexible

methylene chains in 54. To address the effects of charge density, a number of

kanamycinA-amino acid conjugates have been synthesized and evaluated for

their RRE affinity (Figure 6.8 and Table 6.5).

Figure 6.8: Structures of kanamycin A-amino acid conjugates (see sections E6.4.1 – E 6.4.3 for the synthesis and characterization of 53-57).

Both kana A-Dab (55) and kana A-Lys (56) have 8 amines, but they exhibit a

much lower RRE affinity as compared to the kanaA-kanaA dimer (Table 6.2 and

Table 6.5). This suggests that no simple relationship exists between the density

of the 8 amines and the RRE affinity exhibited by each compound, as the

kanamycin dimer should have a lower density of its 8 amines. Compared to

kanaA-Dab, the longer methylene linkers provided by kanaA-Lys were expected

to give this compound a lower overall charge density (Figure 6.8). Both

O

HN

HOHN

HN

OHOHO

O

O

HO OHNH

OH

H2NO

NH2

ONH2

NH2

O

H2N

O

Kana A-Dab (55)

HO

H2N

NH2

NH2

O

HN

HOHN

HN

OHOHO

O

O

HO OHNH

OH

H2NO

NH2

ONH

NH2

O

H2N

O

γ-Guanidino Kana A-Dab (57)

HO

HN

NH

NHNH

H2N

HN NH2

HNNH2

NH

NH2

O

HN

HOHN

HN

OHOHO

O

O

HO OHNH

OH

H2NO

NH2

ONH2

O

H2N

O

HO

NH2

H2N

H2N

NH2

Kana A-Lys (56)

165

compounds, however, have approximately the same RRE affinity (Table 6.5). It is

possible that, due to the plasticity of the pKa values of amines, kanaA-Lys and

the kanaA-kanaA dimer may have higher total charges, but the same or lower

charge densities as compared to kanaA-Dab. These effects may for kanaA-Lys

and kanaA-Dab cancel out, explaining perhaps, the similar RRE affinities of

these compounds. Upon guanidinylation of the γ-amines of kanaA-Dab, a 5-fold

increase in RRE affinity is realized (Table 6.5). This is a much smaller increase

as compared to the 12-fold increase upon guanidinylation of kanamycinA (Table

6.3 and Table 6.4). It appears therefore, that guanidinylation of the amines closer

to the “core” of the glycoside is more advantageous as compared to

guanidinlyation of the amines distal to the core. This is consistent with the proven

importance of the α-amines of tobra-Arg (Table 6.3) and the kanamycinA-Arg

conjugates.189

6.5 Acridine-Containing Guanidinoglycosides

In an attempt to increase the RRE affinity and selectivity of neo-N-acridine and

tobra-N-acridine (Section 6.1 – 6.2), these compounds were guanidinylated using

N,N'-diBoc-N''-triflylguanidine (Section E 6.5.1). The main side products for both

reactions are the “n-1” guanidinylated products 59 and 61. Interestingly, these

compounds have a higher RRE affinities compared to the fully guanidinylated

compounds 58 and 60. Fluorescence anisotropy was used to evaluate the RRE

affinity of compounds 58 – 61 (Table 6.6). Limits must be reported for 60 and 61,

as these compounds have higher RRE affinities compared to RevFl.

166

Figure 6.9: Structures of guanidininylated acridine-containing glycosides. SeeSection E 6.5.1 for synthesis and characterization. *For compounds 59 and 61,the non-guanidinylated site has not, unambiguously, been assigned. COSY andROESY NMR experiments of 59, have proven the nonguanidinylated site is onthe 2-DOS ring. Similarly, for 61, the least basic amine is likely the non-guanidinylated position.34

Table 6.6: Preliminary RRE66 affinities, according to RevFldisplacement experiments using fluorescence anisotropy.

Compound RRE66 Ki

58* 6.5 ± 1 x 10-9

59* 4.8 ± 2 x 10-9

60** < 1.1 ± 0.9 x 10-9

61** < 0.7 ± 0.5 x 10-9

* Calculated from RevFl-RRE66 (10 nM / 100 nM, respectively) IC50 values.** Calculated from RevFl-RRE66 (10 nM each) IC50 values.

O

HN

HO HN

H2NO

HOHOO

O

NH

OHNH

NH

H2N

NHNH2

NH

NH2

NH

H2N

NH

N

HN

N

HN

HNO

OO

HN

OH

HOO

HOO

HO

HN

OHO

HO

NH

NH

H2NNH

HNNH2

NH

NH2

NH

NH2NH

H2N

HNNH2

Guanidino4 Tobra-N-acridine (59)*

Guanidino6 Neo-N-acridine (60)

O

HN

HO HN

HN

OHO

HOO

O

NH

OHNH

NH

H2N

NH

H2NNH

NH2

NH

NH2

NH

H2N

NH

N

Guanidino5 Tobra-N-acridine (58)

HN

N

HN

HNO

OO

H2N

OH

HOO

HOO

HO

HN

OHO

HO

NH

NH

H2NNH

NH

NH2

NH

NH2NH

H2N

HNNH2

Guanidino5 Neo-N-acridine (61)*

167

Compounds 58 – 60 have been evaluated for anti-HIV activities; guanidino5

tobra-N-acridine (58) has an IC50 value of 0.2 µM in both HeLa cells and human

HMBC (summarized in Table E 1.2). This represents a modest (5-fold)

improvement over guanidino-tobramycin (50). Interestingly, guanidino4 tobra-N-

acridine (59) has, according to the PBMC assay, little or no anti-HIV activity at all

(through 10 µM, Table E 1.2). This result can be understood by the differences in

cellular uptakes of these two compounds (presented in Section 6.8).

6.6 Platinum-Containing Aminoglycosides and Guanidinoglycosides

Binding of the anticancer drug cisplatin (cis-Pt) to nucleic acids is believed to

favor DNA over RNA.190 The bifunctional covalent modification of DNA by cis-Pt

produces intrastrand cross-links (at GpG, ApG, and GpNpC),191 and interstrand

cross-links (at GpC).192 Cisplatin also reacts with other cellular components

including: proteins,193 phospholipids,194 and small molecules.195 To enhance the

DNA targeting of cis-Pt, numerous conjugates have been prepared, where the

cis-Pt moiety has been tethered to high-affinity nucleic acid ligands (e.g.,

acridine, ethidium, and short oligonucleotides).196 In an attempt to “reverse” the

nucleic acid selectivity of cis-Pt, and to covalently inactivate the RRE, the

cisplatin-glycoside conjugates Neo-platin (Neo-Pt) and Guanidino Neo-platin (G.

Neo-Pt) were synthesized in good yields and fully characterized by Dr. Jürgen

Boer (Figure 6.10).197

168

Figure 6.10: Structures of cis-platin, neoplatin, and guanidino-neoplatin.

The non-covalent inhibition of the Rev-RRE interaction may result in the splicing

of newly transcribed HIV RNA into smaller fragments. From these RNA

fragments, the proteins Rev, Tat, and Nef are translated (Section 2.0 – 2.1).

Inhibition of Rev-RRE binding may, therefore, actually increase Rev production,

leading to a higher cellular concentration Rev. This would make the continued

non-covalent inhibition of the Rev-RRE interaction increasingly difficult. Covalent

modification of the RRE may provide a mechanism to irreversibly inactivate the

RRE, and thus permanently eliminate its abilities to bind Rev or to be translated

into protein.

To explore the ability of Neo-Pt to form covalent cross-links to the RRE JW,

denaturing polyacrylamide gel-shift electrophoresis (DPAGE) and matrix assisted

laser desorption ionization time of flight mass spectrometry (MALDI TOF MS)

have been utilized. Upon titration of Neo-platin with RRE JW, DPAGE indicates

R

NH2

HN

NH

NH2

Compound

Neo-platin (62)

Guanidino-neo-platin (63)

NH2

NH2

Pt

Cl

Cl

Cisplatin (64)

N

O

R

RO

HOHO

OO

R

OH

HOO

HOO

HO

R

R

R

NH2 O

H2N

Pt

Cl

Cl

H

169

that multiple equivalents of Neo-platin bind to the RRE, while the addition of

neomycin B produces no effect (Figure 6.11 A). Cross-linking is rapid (less than

30 min) and kinetically stable adducts are formed (persist greater than 20 h at 25

°C). Consistent with earlier studies, G. Neo-Pt appears to be five times more

active for cross-linking the RRE as compared to Neo-Pt (Figure 6.11B).

Figure 6.11: DPAGE results with 25 nM of 32P 5'-labeled RRE JW. A) lane 1:RNA only, lanes 2 – 7: Neo-Pt at 12.5 nM, 125 nM, 625 nM, 1.25 µM, 12.5 µM,125 µM, respectively, lanes 8 – 11: neomycin B at 625 nM, 1.25 µM, 12.5 µM,125 µM, respectively, lanes 12 – 17: 625 nM of Neo-Pt at 0.5 hr, 1 hr, 2 hr, 3 hr,4 hr, 5 hr, respectively. B) lane 1: RNA only, lanes 2 –9: Neo-Pt at 50 nM, 100nM, 500 nM, 1 µM, 2.5 µM, 5 µM, 7.5 µM, 10 µM, respectively, lane 10: RRE JWonly, lanes 11 – 16: G. Neo-Pt at 50 nM, 100 nM, 500 nM, 1 µM, 2.5 µM, 5 µM,respectively. C) lane 1: RNA only, lanes 2 – 8: 1 µM of G. Neo-Pt with CT DNA at8.5 mM, 4.25 mM, 850 µM, 85 µM, 8.5 µM, and 0 µM b.p., respectively, lanes 9and 17 RRE JW only, lanes 10 – 16: 1 µM of Neo-Pt with CT DNA at 8.5 mM,4.25 mM, 850 µM, 85 µM, 8.5 µM, 0 µM, respectively.

A)

B)

C)

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

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

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

170

Neo-Pt cross-links to the RRE even in the presence of a vast excess of calf

thymus (C.T) DNA (Figure 6.11 C). Approximately a 5,000-fold excess of C.T.

DNA (relative to the number of bases in RRE JW) must be added to inhibit cross-

linking of the RRE (Figure 6.11C). This suggests that the RNA over DNA

selectivity of these compounds is dominated by the neomycin moiety. Indeed, a

CN− deactivated Neo-Pt shows approximately the same RRE affinity and

selectivity as unmodified neomycin B (presented in Section 6.7).

To further explore the RRE cross-linking specificity of Neo-Pt and G. Neo-Pt,

another strongly denaturing technique MALDI TOF MS has been used in

conjunction with two RRE JW mutants (Figure 2.5). MALDI confirms the

formation of the covalent adducts between the RRE JW and Neo-Pt as well as G.

Neo-Pt (Figure 6.12 A and B, respectively). After the same 1 hr incubations, both

an inosine-substituted RRE (3I RREJW) and a DNA version of the RRE JW

(dRREJW) show little to no cross-linking of Neo-Pt (compare Figure 6.12A to

6.12C and 6.12D). At a longer incubation time (18 hrs), significant cross-link

formation is observed between Neo-Pt and dRREJW (similar magnitude as 6.12

A). Both dRREJW and 3I RREJW have much lower affinities to neomycin as

compared to the RRE JW (presented next section). This may explain the slower

cross-linking of these constructs to Neo-Platin. G. Neo-Pt, however, has a high

affinity for all three RRE constructs (next section) and shows approximately the

same cross-linking efficiencies to all three RRE constructs (similar to 6.12 B).

This confirms the decrease in RRE selectivity upon guanidinylation of neomycin

171

B (Section 6.4). Neo-Pt, however, exhibits a clear selectivity for RREJW over

both the 3I RREJW and the dRREJW constructs.

Figure 6.12: MALDI TOF MS of 1 µM of RRE JW treated with 0.75 µM of either:Neo-Pt (A) or G. Neo-Pt (B), and 1 µM of 3I RRE JW (C) or dRRE JW (D) treatedwith 0.75 µM of Neo-Pt and incubated for 1 hr. cis-Pt (64) shows no reactivitytowards any of the RRE constructs under these conditions (even after overnightwater activation and 15 hr incubations with the nucleic acids). See section E6.6.1 for reaction conditions and sample preparation.

To determine the cross-linking site(s) for Neo-Pt and G. Neo-Pt on the RRE JW,

enzymatic digestions of 1:1 glycoside-RRE adducts were performed using either

Snake Venom Phosphodiesterase (3’! 5’ exonuclease) or Bovine Spleen

Phosphodiesterase (5’! 3’ exonuclease). The Neo-Pt adducts are found to

generate stop-sites for these enzymes, resulting in RRE fragments that can be

characterized by MALDI TOF MS (Figure 6.13 and 6.14).

Neo-platin + RRE JW

4 000 6 000 8 000 10 000 12 000 14 000 16 000 18 000 20000

Co

unts

0

2000

4000

6000

Guanidino-Neo-platin+ RRE JW

Co

unts

0

4000

8000

12000

M+2

M+2

11,103

12,046

11,104

12,380

4000 6000 8000 10000 12000 14000 16000 18000 20000

M/Z

4 000 6000 8000 10000 12000 14 000 16000 18 000 20 000

Neo-platin +3I RRE JW

4 000 6000 8000 10000 12000 14 000 16000 18 000 20 000

Co

unts

0

10000

20000

Neo-platin +dRRE JW

4000 6000 8000 10000 12000 1 4000 16000 18000 2 0000

Co

unts

2000

4000

6000

8000

M/Z

M+2

M+2

11,024

10,582

A)

B)

C)

D)

172

Figure 6.13: Snake Venom Phosphodiesterase digestion of RRE JW control (A),Neo-Pt cross-linked (B), or G. Neo-Pt cross-linked (C) RRE JW. Bovine SerumPhosphodiesterase digestion of RRE JW control (D), Neo-Pt cross-linked (E), orG. Neo-Pt cross-linked (F) RRE JW. See E 6.6.1 for reaction conditions andsample preparation.

Interestingly, the nuclease-resistant fragments for both Neo-Pt and G. Neo-Pt

adducts are identical for both enzymes (Figure 6.13). This indicates that the Neo-

Pt and G. Neo-Pt adducts are absent in the nuclease-resistant RNA fragments.

The stop sites for each phosphodiesterase are summarized (Figure 6.14 Left).

The initial site of covalent cross-linking is near the loop region of the RRE JW for

both Neo-Pt and G. Neo-Pt. This data suggest that a single G-G cross-link is

responsible for the pattern of enzymatic stop sites. The proposed cross-linking

BSP Control Digestion

2000 4000 6000 8000 10000 12000

Co

unts

10000

20000

30000

Neo-platin / RRE JWBSP Digestion

2000 4000 6000 8000 10000 12000

Co

unts

5000

10000

15000

Guanidino-Neo-platin / RRE JWBSP Digestion

M / Z

2000 4000 6000 8000 10000 12000

Co

unts

3000

6000

9000

8,57

68,

2 30

7,8 8

37,

576

6,26

55,

923

5,61

55,

267

2,86

92,

527

2 ,21

5

5,75

06,

108

6 ,46

8

5,38

7

6,4 8

3

6,1 1

05,72

35 ,35

1

Guanidino-Neo-platin / RRE JWSVP Digestion

M / Z

2000 4000 6000 8000 10000 12000

Co

unts

3000

6000

90005,044

SVP Control Digestion

Co

unts

10000

20000

1,10

21,

470

1,83

82,

213

2,58

5

5,08

05,

458

Neo-platin / RRE JWSVP Digestion

Co

unts

10000

20000

30000 5,040

5,138

M+2

A)

B)

C)

D)

E)

F)

173

site (Figure 6.14, Right) is a typical 5’GpC3’ interstrand-like cross-link that has

been well characterized for cis-Pt and DNA duplexes.192 It is interesting that both

enzymes appear to “read through” the first modified G, but stop before the

second G of the cross-link. It is possible that the enzymes initiate cleavage in the

loop of the RRE, and therefore bypass the “first” cross-linked G.

Figure 6.14: Enzymatic stop sites for both Neo-Pt and G. Neo-Pt (Left). Theopen arrow indicates the stop site for Snake Venom Phosphodiesterase; theclosed arrows indicate position and relative intensity of stop sites for BovineSpleen Phosphodiesterase. Proposed site of cross-linking (Right).

Consistent with earlier studies,89 multiple equivalents of Neo-Pt are needed to

displace the Rev peptide from the RRE. Native gel-shift electrophoresis was

used to monitor the displacement of Rev-IA from the RRE JW (Figure 6.15). At

low concentrations, Neo-Pt can cross-link the Rev-RRE complex without

displacing the peptide. It appears that three equivalents of Neo-Pt are necessary

for peptide displacement (Figure 6.15). Marino has proposed that two equivalents

of neomycin B are sufficient to displace the Rev peptide from the RRE.89 No Rev-

RRE-neomycin B tertiary complex is observed by electrophoresis and may

indicate the poor kinetic stability of such a complex (Figure 6.15). The IC50 values

for both Neo-Pt and neomycin B are approximately 2.5 µM, but Neo-Pt covalently

CA

A

U U

A

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

A G

G

C

G

G

C

G

C

RRE JW

CA

A

U U

A

G

C

G GG

CG

C

A GU A

C

G U

G

C

C

G

A G

G

C

G

G

C

G

C

RRE JW

174

modifies the RRE and retards its mobility even after the Rev peptide has

dissociated (Figure 6.15).

Figure 6.15: Native gel-shift electrophoresis. 25 nM of RRE JW is mixed with 2µM of “Rev-IA“, then Neomycin B is titrated (0.5 µM, 1.0 µM, 1.25 µM, 2.5 µM,3.75 µM, 5.0 µM, 7.5 µM, 10 µM) or Neo-Pt is titrated (0.25 µM, 0.5 µM, 0.75 µM,1.0 µM, 1.25 µM, 2.5 µM, 5.0 µM, 10 µM). On a denaturing gel, RREJW appearsas a single band (Figure 6.11), on the native gel it consistently appears as adouble band, this suggests two alternate folds of the RRE JW are possible.

Neo-platin (62), guanidino-neoplatin (63), and cis-platin (64) have each been

evaluated for their ability to inhibit HIV replication in HeLa cells (using a syncytia-

formation assay).110 These experiments suggest that all three compounds are

highly potent inhibitors of HIV-1 replication with IC50 values similar to AZT (10 –

100 nM) (Table E 1.1). Experiments using different methodologies (conducted at

Immusol Inc.) suggest, however, that these compounds are not very active.

Clearly, more tested is needed to establish the biological activities of 62 and 63.

Neo-platin and G. Neo-platin are found to create site-specific, RNA-selective

covalent cross-links to the RRE. These molecules are, to the best of our

knowledge, the first examples of small molecules that covalently modify RNA

RRERev+RRE

Neomycin B Neo-Pt

175

over DNA. These results also serve as an important precedent for the selective

covalent modification of therapeutically relevant RNA targets by small molecules.

6.7 BODIPY- and Fluorescein-Linked Glycosides as RNA Probes

Fluorescent dyes have been attached to aminoglycosdies to facilitate the

analysis of RNA binding interactions.72a,76a,183 In all previous cases, the amino

positions of aminoglycosides were used for one-step conjugation reactions,

resulting in an amide-bond at that position. The amines of the aminoglycosides

are, however, essential for RNA binding. 36,118,174-175 We use, therefore, the

primary alcohols of aminoglycosdies for conjugation reactions. Indeed, the

hydroxyl groups of tobramycin have been shown to be non-essential for RNA-

binding.36 Both BODIPY and fluorescein have been conjugated to the 6"

positions of tobramycin, guanidino-tobramycin, neomycin B, and guanidino-

neomycin B (Figure 6.16). BODIPY is an uncharged, non-intercalative dye that

has proven itself an ideal probe for the study of RNA-aminoglycoside binding.

To further explore the RRE affinity and specificity of neomycin and guanidino-

neomycin, the fluorescence anisotropy of A. Neo-BODIPY (67) and G. Neo-

BODIPY (68) were monitored as a function of increasing RREJW, 3IRREJW, or

dRREJW concentrations (Figure 2.4 and Figure 2.5). The fluorescence

anisotropy of the BODIPY-conjugates increases upon binding of each RRE

construct (see Figure 6.17 for an example). These association curves are used to

determine C50 values (defined as the concentration of RNA needed to bind ½ of

176

the BODIPY conjugate). The C50 value is approximately equal to the Kd if the C50

value is significantly higher than the concentration of the fluorescent probe (10

nM). C50 values that are in the same order as the concentration of fluorescent

probe can also be used to calculate Kd values, but a robust assessment of

binding stoichiometery should be taken into consideration (Section 3.6). Table

6.7 summarizes the C50 values for A. Neo-BODIPY and G. Neo-BODIPY for

binding to the RRE JW, the 3I RRE JW, and the dRRE JW.

Figure 6.16: New fluorescent aminoglycosides based upon BODIPY (65 – 68)and fluorescein (69 – 71). See Section 6.7.1 – 6.7.7 for the synthesis andcharacterization of these compounds.

O

R

RO

HOHO

OO

R

OH

R

OHO

HOO

S

HO

R

R

OS

Guanidino Neo-Fl (71)NH

NH

NH2

R Compound

OHOR

OR R

O

ORR

OH

HO HO

S

NNB

FF

OS

NH

O

Amino Tobra-BODIPY (65)

Guanidino Tobra-BODIPY (66)NH

NH

NH2

NH2

R Compound

O CO2H

HO

O

NH

O

OHOR

OR R

O

ORR

OH

HO HO

S

OS

Amino Tobra-Fl (69)

Guanidino Tobra-Fl (70)HN

NH

NH2

NH2

R Compound

O CO2H

HO

O

NH

O

O

R

RO

HOHO

OO

R

OH

R

OHO

HOO

S

HO

R

RNN

BFF

OS

NH

O

Amino Neo-BODIPY (67)

Guanidino Neo-BODIPY (68)NH

NH

NH2

NH2

R Compound

177

Figure 6.17: Changes in fluorescence anisotropy of G. Neo-BODIPY (68) as afunction of RRE concentration.

Consistent with the peptide displacement experiments using the unlabeled

glycosides (Section 6.4), G. BODIPY-Neo has about a 5-fold higher RRE affinity

compared to A. BODIPY-Neo (Table 6.7). Interestingly, Guanidino BODIPY-Neo

has a much higher affinity to the mutant RRE constructs as compared to Amino

BODIPY-Neo (Table 6.7). This is also consistent with the decrease in RRE

specificity of neomycin B, upon guanidinylation (as indicated by the solid phase

assay, Table 6.4).

Table 6.7: C50 values (nM) according to fluorescence anisotropy.*RRE Construct Amino Neo-BODIPY Guanidino Neo-BODIPY

RRE JW 60 153I RRE JW 800 17dRRE JW 12,000 250

*Estimated error ± 30% of each reported C50 value.

Concentration of RRE (nM)

0 1000 2000 3000 4000 5000 6000

Ave

rage

Ani

sotr

opy

Val

ue(u

nitle

ss)

0.02

0.04

0.06

0.08

0.10

RRE JWdRRE JW

178

Since the size, and hence negative charge, for all three of these RREJW

constructs is the same, the differences in affinities likely reflect differences in the

BODIPY neomycin off-rates from each construct. Differences in off rates are also

the most likely explanation for the different cross-linking rates for Neo-Pt and the

three RREJW constructs (Section 6.8).

To probe how Pt-modification affects the RRE affinity of neomycin; Neo-Pt, G.

Neo-Pt, neomycin B, and guanidino-neomycin B, were used to displace the

BODIPY-Neo compounds 67 and 68 from the RRE. To deactivate the covalent

cross-linking abilities of Neo-Pt and G. Neo-Pt, 10 mM stocks of these

compounds were prepared in 40 mM KCN. DPAGE experiments confirm that

these stocks are not capable of forming covalent cross-links with the RRE (not

shown). Displacement experiments are conducted by mixing 10 nM of either A.

Neo-BODIPY or G. Neo-BODIPY with 50 nM of RRE JW. By monitoring the

fluorescence anisotropy of the BODIPY conjugate, its displacement from the

RREJW becomes apparent. The IC50 values for each compound are summarized

in Table 6.8. Since multiple equivalents of neomycin can bind to the RRE, the Ki

values estimated in Table 6.8 are for comparison purposes only and may not

reflect the actual affinities. The calculated Ki value for the displacement of A.

Neo-BODIPY from the RREJW by neomycin B is, however, nearly the same

value as determined by the displacement of RevFl from the RRE66 (Table 6.0

and Table 6.8).

179

Table 6.8: Summary of IC50 values (nM) for displacement from RRE JW.Compound Compound

Displaced IC50* Ki

**

Neomycin B (67) 400 240

Neo-Pt*** (67) 120 60

Guanidino Neomycin B (68) 155 22

Guanidino Neo-Pt*** (68) 120 15

* Estimated error ± 30% of reported value.** Calculated assuming a single binding site on RRE JW for both the Neo-BODIPY probe andfor each compound evaluated, and assuming the Kd for Amino Neo-BODIPY / RRE JW = 55 nMand the Kd for Guanidino Neo-BODIPY / RRE JW = 10 nM (Table 6.7).*** CN− treated (see above).

These values suggest that the addition of BODIPY to Neomycin introduces a

small (2 – 7 fold) increase in affinity for RRE JW. The conjugation of Pt to

neomycin, likewise, introduces a very small (2 – 4 fold) increase in affinity for the

RRE JW. These differences are small enough to be explained by potential

differences in displacement stoichiometries.

In summary, the conjugation of BODIPY to neomycin B and guanidino-neomycin

B is found to introduce only a small perturbation for RNA binding. These

conjugates should, therefore, be useful probes for the future study of other RNA

-aminoglycoside and RNA-guanidinoglycoside binding interactions.

6.8 Cellular Uptake of Aminoglycosides and Guanidinoglycosides

Despite their high affinity for the RRE, neither tobra-N-acridine (41), nor neo-N-

acridine (38) possess anti-HIV activity. The guanidino forms of these compounds,

however, are active for HIV inhibition (compounds 58 and 60, Table E 1.2).

180

Differences in cellular uptakes may, potentially, explain the different anti-HIV

activities of these compounds. Fluorescence microscopy was used to monitor the

uptake of these compounds into COS cells (Table 6.9). These qualitative

observations suggest that mammalian cells show significant cellular uptake of the

guanidinoglycoside-acridine conjugates but not the aminoglycoside-acridine

conjugates (Table 6.9).

Table 6.9: Qualitative and preliminary assessment of cellular uptake.*

Compound Cellular Uptake into COS Cells?

RevFl Yes

9-amino acridine Yes

Guanidino5 Tobra-N-acridine Yes

Neo-N-acridine No

Guanidino4 Tobra-N-acridine No

* Based upon fluorescence microscopy of COS cells after a2 hr exposure to 5 µM of each compound (1 µM of RevFl).

Acridine is not, however, an ideal fluorescent probe for cellular uptake studies. A

short wavelength of excitation (440 nm) causes significant background

fluorescence. In addition, the acridine-conjugates themselves have a wide range

of quantum yields (Table 6.10). Indeed, the guanidino-containing conjugates

have 2 – 5 fold higher quantum yields compared to the aminoglycoside-acridine

conjugates (Table 6.10). In addition, the quantum efficiencies of the acridine

conjugates change upon binding cellular components (including nucleic acids).

For these reasons, the BODIPY fluorophore was used for quantitative

assessments of the cellular uptakes of amino- and guanidinoglycosides.

181

Table 6.10: Quantum yields of acridine- and BODIPY-containingglycosides in 100 mM sodium phosphate pH 6.7.*

Compound Quantum Yield (φ)

Neo-S-acridine (39) 0.053**

Neo-N-acridine (38) 0.089**

Guanidino6 Neo-N-acridine (60) 0.24**

Tobra-N-acridine (41) 0.094**

Guanidino5 Tobra-N-acridine (58) 0.17**

RevFl 0.39***

Amino BODIPY-Tobra (65) 1.0***

Guanidino BODIPY-Tobra (66) 1.0***

Amino BODIPY-Neo (67) 1.0***

Guanidino BODIPY-Neo (68) 1.0***

* Determined by comparisons with standards. φ = φstd.(Iunk./Aunk.)(Istd../Astd.), whereI = area of emission (in wave numbers), and A = absorption.

** Relative to 9-aminoacridine (φ = 0.96), 100 mM sodium phosphate pH 6.7.*** Relative to fluorescein (φ = 0.90), 0.1 M NaOH.

All four of the BODIPY containing glycosides 65 – 68 are found to have quantum

yields equal to 1.0 (Table 6.10). This allows for simple interpretation of cellular

uptake data, where fluorescence intensity is directly proportional to the

concentration of each compound.

The uptake of BODIPY-containing glycosides into eukaryotic cells is significantly

enhanced upon guanidinylation. Fluorescence activated cell sorting (FACS) was

used to quantify the fluorescence intensity of 10,000 cells that were treated with

0.5 µM of compounds 65 – 69. Cell cultures were treated for 1 hr, washed two

182

times with buffer, cleaved with trypsin, and quantified for fluorescence. Upon

guanidinylation, the cellular uptake of tobramycin is enhanced, on average, by 4-

fold. The enhancement for neomycin is over 7-fold (Figure 6.18 and Table 6.11).

Interestingly, both A. Tobra-BODIPY (65) and A. Neo-BODIPY (67) have

approximately the same cellular uptakes, but G. Neo-BODIPY (68) exhibits 2-fold

better uptake as compared to G. Tobra-BODIPY (66) (Figure 6.18 and Table

6.11).

A number of poly-arginine containing proteins, peptides, and peptoids have

recently been found to exhibit highly efficient uptake into a wide range of

mammalian cell types.198 We were interested, therefore, to compare the uptake

of a poly-Arg peptide to the uptake of the guanidinoglycosides. The BODIPY-

containing poly (Arg)9 peptide “BODIPY-CR9” (72) was synthesized (see E 6.8.1)

and evaluated for cellular uptake efficiency (Figure 6.19 A and Table 6.11).

Figure 6.18: FACS histograms showing the fluorescence intensity and cell countfor 10,000 individual10T1/2 cells following a 1 hr incubation with 0.5 µM of: A)Amino Tobra-BODIPY (65) (Red) and Guanidino Tobra-BODIPY (66) (White). B)Amino Neo-BODIPY (67) (Red) and Guanidino Neo-BODIPY (68) (White).Similar results were also observed for HeLa cells (not shown).

101 104102 103100

Fluorescence Intensity (530 nm)

Num

ber

ofC

ells

128

64

101 104102 103100

Fluorescence Intensity (530 nm )

Num

ber

ofC

ells

128

64

A) B)

183

Table 6.11: Summary of the cellular uptake efficiencies of 0.5 µMof each BODIPY-containing compound.

Compound Mean fluorescenceintensity (FACS)

A. Tobra-BODIPY (65) 61

G. Tobra-BODIPY (66) 241

A. Neo-BODIPY (67) 58

G. Neo-BODIPY (68) 431

BODIPY-CR9 (72) 284

BODIPY-CR9 (72) + 10 µM G. Neo (52) 105

BODIPY-CR9 (72) + 50 µM G. Neo (52) 88

BODIPY-CR9 (72) + 200 µM G. Neo (52) 69

Guanidino tobramycin-BODIPY (66), compared to BODIPY-CR9 (72), has 4 fewer

guanidinium groups but still shows the same transport efficiency (Table 6.11).

This indicates that the flexible, amphipathic properties provided by the methylene

chain of Arginine residues are not needed for effective membrane transport.

Importantly, guanidino Neo-BODIPY (68) has a better cellular uptake as

compared to the poly-Arg peptide BODIPY-CR9 (72) (Figure 6.19, Table 6.11).

This suggests that the semi-rigid pre-organization of the guanidinium groups on

the glycoside core may better facilitate translocation across the cell membrane.

Indeed, other groups have found that for peptides secondary structure formation

increases transport efficiency.199 Other groups have found that poly Arginine

exhibits better uptake properties as compared to poly Lysine.200 Taken together,

it appears that no matter what type of scaffold is used, amines exhibit poor

transport properties into mammalian tissue cultures, as compared to the

corresponding guanidino compounds.

184

Figure 6.19: FACS histograms showing the fluorescence intensity and cell countfor 10,000 individual10T1/2 cells following a 1 hr incubation with: A) 0.5 µM ofBODIPY CR9 (72) (Red), or Guanidino Neo-BODIPY (68) (White). B) 0.5 µM ofBODIPY CR9 transport inhibited by guanidino-neomycin B (52), at 0 µM (red), 10µM (black), 200 µM (green).

To examine the mechanism by which guanidinoglycosides enter eukaryotic cells,

a competition experiment was conducted with unlabeled guanidino-neomycin B

(52) and BODIPY-CR9 (72). FACS analysis shows that guanidino-neomycin B

(52) effectively inhibits the transport of BODIPY-CR9 (Figure 6.16 B and Table

6.11). This may indicate that guanidinoglycosides competitively bind to the same

cellular receptor(s) responsible for the transport of poly-Arg peptides, suggesting

a common pathway responsible for the uptake of both compounds.

In collaboration with Peter Carmichael, fluorescence microscopy has been used

to examine the cellular localization of compounds (65) – (73). All of the

guanidinium-containing BODIPY-conjugates (66, 68, 72) and the fluorescein

versions of these molecules (70, 71, 73) are found, under these conditions, to

have diffuse, cytoplasmic and nuclear distributions (Figure 6.19 A, B, D). A

101 104102 103100

Fluorescence Intensity (530 nm)

Num

ber

ofC

ells

128

64

A)

101 104102 103100

Fluorescence Intensity (530 nm)

Num

ber

ofC

ells

128

64

B)

185

phase-contrasted reflected light image allows for detailed visualization of all cells

present in each field (Figure 6.20 E – H).

Figure 6.20: (A – D): Fluorescence emission microscopy of 10T1/2 cellsexposed to: (A) Guanidino Neo-Fl, (B) Guanidino Tobra-Fl, (C) Amino Tobra-Fl,and (D) Fl-CR9. For image (C) the sensitivity of detection is increased to allowimaging of the cellular localization of Amino Tobra-Fl. (A – D): Phase contrastmicroscopy of (E) Guanidino Neo-Fl, (F) Guanidino Tobra-Fl, (G) Amino Tobra-Fl, and (H) Fl-CR9. These particular images are made with the fluoresceinversions of these compounds (69 – 73), as binding of the BODIPY compounds tothe culture dishes made for poor fluorescence contrast.

Interestingly, the fluorescent aminoglycoside-conjugates appear to exhibit a

different cellular localization as compared to the guanidino-conjugates. The

fluorescent aminoglycoside-conjugates (65, 67, and 69) have punctate,

perinuclear localizations, which may correspond to acidic vesicles (Figure 6.20 C

& G). Confocal microscopy has been used to take optical cross sections through

cells that have been exposed to the guanidinium-containing BODIPY-conjugates

(66, 68, 72) and the fluorescein versions of these molecules (70, 71, 73); this

process confirms that these compounds are localized inside of cells (not shown).

A) B) C) D)

E) F) G) H)

186

The vesicular trafficking activities of these cells were also visually inspected over

time, and indicate that these compounds are accumulated by healthy living cells

(not shown). Similar experiments with other cell types, however, (including COS

and HeLa) suggest that the cellular localization, vesicular trafficking, and even

the toxicity of these compounds may change as a function of incubation time, cell

type, media, and the identity and concentration of the fluorescent transport

substrate used (not shown).

In summary, we have found that guanidinoglycosides show highly efficient

uptake into eukaryotic cell cultures via a similar mechanism as poly-Arginine.

Guanidino-neomycin exhibits a more efficient transport than the poly-Arginine

peptide BODIPY-CR9. Guanidinoglycosides may also provide other advantages

as transport substrates, including resistance to proteases, and serving as tools

for the elucidation of the currently unknown poly Arg transport mechanism(s).

Interestingly, a report on aminoglycoside-cholesterol conjugates recently

described the incorporation of these compounds into liposomal phospholipid

complexes capable of the delivery of DNA into living cells. They report, however,

that the guanidinylation of a kanamycin A-cholesterol conjugate actually

decreased its efficiency as a DNA delivery agent. Our preliminary results indicate

that 5 guanidinium groups are the minimum number needed to stimulate efficient

cellular uptake of the tobramycin – acridine conjugates (Table 6.9).

187

The conjugation of poly-arginine peptides to large molecules can facilitate the

transport of peptide, protein, and nucleic acid conjugates into cells.202 We were

interested, therefore, in synthesizing thiol-reactive guanidinoglycosides that can

be used for one-step conjugation reactions with various “cargo” molecules. The

pyridyl-disulfide activated compounds “Amino Neo-Pyridyl Disulfide” (74) and

“Guanidino Neo-Pyridyl Disulfide” (75) have recently been synthesized and

characterized (Figure 6.21). These compounds should, under aqueous or

methanolic conditions, readily form disulfide bonds with potential transport

substrates. Upon successful translocation of the transport conjugate, reductive

conditions inside of the cell should reduce the disulfide bond, thus releasing the

“cargo” molecule into the cytoplasm and nucleus of the cell.

Figure 6.21: Thiol-reactive amino- and guanidino-neomycin B derivatives showgood solubility in both water and methanol (see Sections E 6.8.3 and E 6.8.4 forsynthesis and characterization of 74 and 75).

O

R

RO

HOHO

OO

R

OH

R

OHO

HOO

S

HO

R

R

OS

Amino Neo-Pyridyl Disulfide (74)

Guanidino Neo-Pyridyl Disulfide (75)NH

NH

NH2

NH2

R Compound

N

S

188

6.9 Conclusions

RNA plays a pivotal role in the replication of all organisms, including viral and

bacterial pathogens. The development of small molecules that can selectively

interfere with undesired RNA activity is a promising new direction for drug

development. Continuing progress in the understanding of small molecule – RNA

recognition may provide future scientists a means for selective targeting of a pre-

determined RNA regulatory element.

Systematic studies of the binding interactions between small molecules and RNA

are essential for deciphering the parameters that govern RNA recognition. The

RRE affinity, specificity, and anti-viral activities of aminoglycosides, intercalating

agents, and metal complexes can be dramatically altered through synthetic

modifications. The synthesis of new ligands with high RNA affinity is relatively

easy compared to discovering new ligands with high specificity for the desired

target. The RRE affinity and specificity of every small molecule is not, however,

always proportional to its anti-HIV activity. Issues related to cellular uptake,

localization, and non-specific binding to other cellular components (especially

membranes) can dramatically affect the biological activities of small molecules.

We have found that guanidinoglycosides exhibit impressive cellular uptake and

desirable cellular localization. These compounds may, one day, prove

indispensable for the delivery of therapeutic agents.

189

Table E 1.0 Compound Source Index

Number Name(s) Source / Reference #

- RevFl (peptide) This thesis, Section E 2.1

- Rev-IA (peptide) This thesis, Section E 2.1

- Rev-TR (peptide) This thesis, Section E 2.1

- Fl-Rev (peptide) This thesis, Section E 2.1

- NH2-Rev (peptide) This thesis, Section E 2.1

(1) – (5) Aminoglycosides Commercially Available

(6) – (8) “Diphenyl furan” or “Aromatic diamindines” Profs. W.D. Wilson and D.W. Boykin83

(9) Λ-[Ru(bpy)2Eilatin]2+ Dr. Moshe Kol and Dalia Gut106

(10) ∆-[Ru(bpy)2Eilatin]2+ Dr. Moshe Kol and Dalia Gut106

(11) rac-[Ru(bpy)3]2+ Commercially Available

(12) ethidium bromide Commercially Available

(13) 9-amino acridine Commercially Available

(14) eilatin Moshe Kol101

(15) rac-[Ru(bpy)2pre-Eilatin]2+ P. Glazer (Tor lab)107

(16) 3-cbz ethidium This thesis, Section E 5.2.1

(17) 8-cbz ethidium This thesis, Section E 5.2.1

(18) 3,8-bis cbz ethidium This thesis, Section E 5.2.1

(19) 3-guanidino ethidium This thesis, Section E 5.2.2

(20) 8-guanidino ethidium This thesis, Section E 5.2.3

(21) 3,8-bis guanidino ethidium This thesis, Section E 5.2.4

(22) 3-urea ethidium This thesis, Section E 5.2.5

(23) 8-urea ethidium This thesis, Section E 5.2.6

(24) 3,8-bis urea ethidium This thesis, Section E 5.2.7

(25) 3-pyrrole ethidium This thesis, Section E 5.2.8

190

Table E 1.0 Compound Source Index (cont.)

(26) 8-pyrrole ethidium This thesis, Section E 5.2.8

(27) 3,8-bis pyrrole ethidium This thesis, Section E 5.2.9

(28) tetramethyl ethidium This thesis, Section E 5.2.10

(29) 3,8-diamino-6-phenylphenanthridine Commercially available; must desaltbefore use.

(30) 3,8-bis urea-pyrrolidine ethidium This thesis, Section E 5.2.11

(31) 3,8-bis-urea-ethylene diamine ethidium This thesis, Section E 5.2.12

(32) 3,8-bisurea-arginine ethidium This thesis, Section E 5.2.13

(33) 3,8-bisurea-2-DOS ethidium This thesis, Section E 5.2.14

(34) 8-urea-ethidium-6'-neamine Charles Liu (Tor lab)*

(35) 3-urea-ethidium-6'-neamine Charles Liu (Tor lab)*

(36) 8-urea-ethidium-6'-3'deoxyneamine This thesis, Section E 5.2.15

(37) 3-urea-ethidium-6'-3'deoxyneamine This thesis, Section E 5.2.16

(38) Neo-N-acridine This thesis, Section E 6.1.1

(39) Neo-S-acridine Sarah Kirk (Tor lab)177

(40) Neo-C-acridine This thesis, Section E 6.1.2

(41) Tobra-N-acridine This thesis, Section E 6.2.1

(42) KanaA-N-acridine Charles Liu (Tor lab)*

(43) Neo-Neo Hai Wang (Tor lab)180 and E 6.3.1

(44) Neo-N-Neo This thesis, Section E 6.3.2

(45) Tobra-Tobra Hai Wang (Tor lab)180

(46) Tobra-N-Tobra Charles Liu (Tor lab)*

(47) KanaA-KanaA Hai Wang (Tor lab)180

(48) – (52) Guanidinoglycosides Tracy Baker (Goodman lab)187

(53) Tobra-Arg This thesis, Section E 6.4.1

(54) ace-Tobra-Arg This thesis, Section E 6.4.2

(55) Kana-A-Dab This thesis, Section E 6.4.3

191

Table E 1.0 Compound Source Index (cont.)

(56) Kana-A-Lys This thesis, Section E 6.4.4

(57) γ-guanidino- Kana-A-Dab This thesis, Section E 6.4.5

(58) guanidino5-Tobra-N-acridine This thesis, Section E 6.5.1

(59) guanidino4-Tobra-N-acridine This thesis, Section E 6.5.1

(60) guanidino6-Neo-N-acridine This thesis, Section E 6.5.2

(61) guanidino5-Neo-N-acridine This thesis, Section E 6.5.2

(62) amino-Neo-Pt Jürgen Boer (Tor lab)197

(63) guanidino-Neo-Pt Jürgen Boer (Tor lab)197

(64) cisplatin Commercially Available

(65) amino-Tobra-BODIPY This thesis, Section E 6.7.1

(66) guanidino-Tobra-BODIPY This thesis, Section E 6.7.2

(67) amino-Neo-BODIPY This thesis, Section E 6.7.3

(68) guanidino-Neo-BODIPY This thesis, Section E 6.7.4

(69) amino-Tobra-Fluorescein This thesis, Section E 6.7.5

(70) guanidino-Tobra-Fluorescein This thesis, Section E 6.7.6

(71) guanidino-Neo-Fluorescein This thesis, Section E 6.7.7

(72) BODIPY-CR9 This thesis, Section E 6.8.1

(73) Fluorescein-CR9 This thesis, Section E 6.8.2

(74) amino-Neo-pyridyl disulfide This thesis, Section E 6.8.3

(75) guanidino-Neo-pyridyl disulfide This thesis, Section E 6.8.4

*Manuscript in preparation.

192

Table E 1.1 Summary of IC50 values (µµµµM)* for HIV-1 inhibition (x794LAI)Compound HeLa Plaque Assay** PBMC / p24**

tobramycin (3) >30 >10

neomycin B (5) >30 n.d.***

Λ-[Ru(bpy)2Eilatin]2+ (9) 0.8 1.4

∆-[Ru(bpy)2Eilatin]2+ (10) 2.0 n.d.***

rac-[Ru(bpy)3]2+ (11) >100 n.d.***

ethidium bromide (12) 0.2 n.d.***

rac-[Ru(bpy)2pre-Eilatin]2+ (15) 30 n.d.***

3-urea ethidium (22) 1.5 n.d.***

8-urea ethidium (23) 3.0 n.d.***

3,8-bis urea ethidium (24) 1.5 n.d.***

8-pyrrole ethidium (25) 1.5 n.d.***

8-urea-ethidium-6'-neamine (34) >10 n.d.***

3-urea-ethidium-6'-neamine (35) >10 n.d.***

neo-N-acridine (38) >10 n.d.***

tobra-N-acridine (41) >10 n.d.***

guanidino-tobramycin (50) 1 2

guanidino-neomycin B (52) 1 n.d.***

guanidino5-Tobra-N-acridine (58) 0.2 0.4

guanidino4-Tobra-N-acridine (59) n.d.*** >10

guanidino6-Neo-N-acridine (60) 1 2

amino-Neo-Pt (62) 0.01 n.d.***

guanidino-Neo-Pt (63) 0.1 n.d.***

cisplatin (64) 0.01 n.d.***

AZT 0.01 <0.001

* Errors do not exceed ± 70% of each reported value.** Conducted as described,110 at CFAR, UCSD.*** n.d. = not determined

193

E 2.0 Synthesis and Quantification of RRE Constructs

For the RRE66, T7 RNA polymerase was utilized for "run-off" in vitro transcription

with a complementary, 83-nt DNA template as described.203 Synthesis of the 83-

nt DNA template was carried out using standard phosphoramidite chemistry.

The DNA oligonucleotide was purified using denaturing polyacrylamide gel

electrophoresis (PAGE), followed by extraction into 200 mM potassium acetate /

10 mM EDTA and two rounds of ethanol precipitation (3 volumes of ethanol).

The sequence and homogeneity of the DNA template was confirmed using di-

deoxy sequencing techniques. Transcription products were also purified using

denaturing PAGE, extraction and multiple rounds of ethanol precipitation. The

expected sequence of the 66-nt RNA transcript was verified by 5' end labeling

with 32P and subsequent enzymatic digestions (RNase T1, A, and U2).177 The

molecular extinction coefficient (260 nm) for the RNA transcripts are calculated

by a summation of the number of each nucleotide multiplied by the number of

times it appears in the sequence (A = 15,400 cm-1 M-1, G = 11,700 cm-1 M-1, C =

7,300 cm-1 M-1 , U = 10,100 cm-1 M-1, T = 8,700 cm-1 M-1, I (inosine) = 7,200 cm-1

M-1, X (modified base of unknown extinction coefficient) = 11,125 cm-1 M-1) .

These values are the extinction coefficients for each nucleotide at 260 nm in

neutral water. In order for the calculated extinction coefficient for each transcript

to be accurate, the transcript must be first hydrolyzed into free nucleotides before

quantification. Procedure for hydrolysis: 1 µl of RNA solution is added to 10 µl of

1 M NaOH and heated 90 °C for 10 min. The reaction is quenched in 10 µl of 1 M

HCl and diluted into 500 µl of 50 mM sodium phosphate pH 7.5 for quantification

194

by spectrophotometry. By comparing the UV-vis absorbance spectrum of RRE 66

both before and after base hydrolysis, the extinction coefficient of the non-

hydrolyzed transcript can be calculated at 560,000 cm-1 M-1 (Figure E 2.0 and

Table E 2.0). Shorter RNA constructs (including RREJW and RREIIB) have been

purchased (Dharmacon Research Inc.), purified as above, and quantified with the

extinction coefficients in Table E 2.0. All RNAs were folded in neutral buffer by

heating at 90 °C for 3 min, then cooling to R.T. over 10 min. MALDI TOF is used

to confirm the expected masses of all RNA transcripts. Duplex nucleic acids are

purchased from either Sigma-Aldrich or Amersham-Pharmacia, dissolved in 10

mM Tris pH 7.5, 1 mM EDTA and quantified in their native form using the

reported extinction coefficients (summarized in reference #87 and the 2002

Amersham-Pharmacia catalog).

Figure E 2.0: UV-vis absorption spectra of RRE 66 before and after basehydrolysis, both spectra are taken in 50 mM sodium phosphate pH 7.5.

Wavelength (nm)

150 200 250 300 350 400 450

Abs

orba

nce

(uni

tless

)

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30RRE 66 before hydrolysisRRE 66 after base hydrolysis

195

Table E 2.0: Extinction coefficients (ε) used for quantification.

Nucleic Acid Base-hydrolyzedε (cm-1 M-1)

Not-hydrolyzedε (cm-1 M-1)

RRE 66 741,400 560,000

RRE IIB 515,000 331,000

RRE JW 371,000 267,000

3I RRE JW 358,000 267,000

deoxy-RRE JW n.d. 267,000

TAR 31 323,900 264,000

A-site 27 276,000 n.d.

tRNAPhe 859,400 733,000

tRNAMIX 11,125* 9,640*

Calf Thymus DNA n.d. 13,100**

Poly (dA) – Poly (dT) n.d. 12,200**

Poly (dG) – Poly (dC) n.d. 14,800**

Poly (dAT) – Poly (dAT) n.d. 13,300**

Poly (dGC) – Poly (dGC) n.d. 16,800**

Poly (rA) – Poly (rU) n.d. 14,280**

Poly (rG) – Poly (rC) n.d. 14,800**

Poly (rI) – Poly (rC) n.d. 10,000**

Poly (rA) n.d. 10,500*

Poly (rC) n.d. 6,200*

Poly (rU) n.d. 9,500*

Poly (dA) n.d. 8,600*

Poly (dT) n.d. 8,520*

Poly (rI) n.d. 10,200*

* = per base** = per base pair

196

Section E 2.1 Design and Synthesis of Rev Peptides

Both N-terminus and C-terminus modified Rev34-50 peptides were synthesized

and studied. Succinylation of the N-terminus, amidation of the C-terminus, and

addition of four alanines has been shown to increase the α helicity of the Rev

peptide, and thereby increase the affinity and specificity to the RRE.46 In our

case, these four alanines also serve as a spacer for subsequent fluorescein

modification.

N-terminus modified: NH2AAAATRQARRNRRRRWRERQRam

C-terminus modified: sucTRQARRNRRRRWRERQRAAAACam

Figure E 2.1: Modified Rev peptides based upon Rev34-50 (bold).

The N-terminus free amine of the "N-terminus modified" peptide is used for an

on-resin conjugation to fluorescein, resulting in “Fl-Rev” (Figure E 2.2). The "C-

terminus modified" peptides are first cleaved from the resin, purified, then the

Cys residue is used for a conjugation to fluorescein (resulting in “Rev-Fl”), or

conjugation to Texas Red (resulting in “Rev-TR”). For the non-fluorescent C-

terminus modified Rev peptide (“Rev-IA”), the C-terminal Cys is capped as a

thioether acetamide derivative (SCH2CONH2) to prevent dimerization (Figure E

2.2).

197

Figure E 2.2: Synthetic scheme for N-terminus and C-terminus modified Rev 34-50

peptides.

N-terminus modification:NH -AAAA-Rev34-50

Rink Amide ResinO

O2C

O O N

O

"Fl-Rev"

NH-AAAA-TRQARRNRRRRWRERQR-am

2

NH -AAAA-TRQARRNRRRRWRERQR-am

"NH -Rev"

cleavage from resinand purification

2

2

DMSO, 1hfollowed by cleavage

from resin and purification

OO

O

O

O2C

O

OO

OO O

O2C

N O

I

H

C-terminus modifications:

suc-Rev34-50AAAAC-amsynthesis, succinylation,

cleavage fronm resin, and purification

suc-TRQARRNRRRRWRERQRAAAAC-am

S

O OO

CO2

NOH

N O

I

H

Aq. Buffer(pH 8.0)

DMF

SH

suc-TRQARRNRRRRWRERQRAAAAC-am

S

H

"Rev-IA"

"Rev-Fl"

Aq. Buffer(pH 8.0)DMSO

NOH

H

“Rev-TR”

suc-TRQARRNRRRRWRERQRAAAAC-am

S

O N

SO3-

SO

O

HN

NHO

S

+N

198

Peptide synthesis was carried out using standard Fmoc/HBTU chemistry on ABI

Applied Biosystems 431A peptide synthesizer using protected amino acids

purchased from Calbiochem. Protected amino acids used in the synthesis were:

Fmoc-Arg-(Pbf)-OH, Fmoc-Asn-(Trt)-OH, Fmoc-Gln-(Trt)-OH, Fmoc-Trp-(Boc)-

OH, Fmoc-Thr-(tBu)-OH, Fmoc-Glu-(OtBu), and Fmoc-Cys-(Trt)-OH. For the C-

terminus modified peptide, the first peptide coupling was done manually to avoid

isomerization problems associated with Cys. Fmoc-Cys-(Trt)-OH was converted

into its symmetric anhydride by reaction with 0.45 equivalents of DCC in dry

methylene chloride for 10 min. The anhydride was separated from the DCU

precipitate by vacuum filtration and methylene chloride was subsequently

removed in vacuo. The solid anhydride was then dissolved in 30% N-

methylpyrrolidone (NMP) in DMF and added to deprotected Rink Amide MBHA

resin. The coupling lasted 2.5 hours at 25 °C and was monitored by ninhydrin

(Kaiser Test). Upon completion, the resin was washed with methylene chloride

followed by NMP, and loaded onto the ABI synthesizer. The remainder of the

peptide synthesis was carried out using standard HOBt/HBTU chemistry and

monitored with a built-in conductivity meter. For both peptides, the first round of

automated coupling reactions was done as a "double coupling" to ensure

efficiency. Upon completion of the automated synthesis, the N-terminus modified

resin was washed (with methylene chloride) and coupled to fluorescein by

reaction with ~100 equivalents of 5-carboxyfluorescein succimidyl ester

(Molecular Probes) in DMSO for 1 h at room temperature. The C-terminus

modified resin was washed with DMF and reacted with 0.3 M succinic anhydride,

199

0.3 M 1-hydroxybenztrizole hydrate, and 0.03 M DMAP in DMF, for 1 h at room

temperature. Both reactions were monitored using the Kaiser Test. Upon

completion of each reaction, the resin was washed (in methylene chloride) and

cleaved from the resin at room temperature for 2.5 h in a "cleavage cocktail"

containing 88% TFA, 5% water, 5% phenol, and 2% triisopropylsilane (v/v). The

filtrate was then mixed with 15 volumes of 2% acetic acid, and extracted 4 times

with diethyl ether. Crude peptides were purified on a C-18 semiprep HPLC

column with an isocratic mixture of either 19% or 14% acetonitrile (0.1% TFA) in

water (0.1% TFA) (for the fluorescein and non-fluorescein containing peptides

respectively). The purified peptides were lyophilized and the C-terminus

modified peptide was reacted with 100 equivalents of 5-iodoacetamidofluorescein

(or Texas Red C5 bromoacetamide) in 100 mM sodium phosphate (pH 8.0), 2

mM EDTA, and 30% (v/v) DMSO at room temperature in the dark for 2 h. Or,

separately, the purified C-terminus modified peptide was reacted with

iodoacetamide (85 mM in 75% DMF/water (v/v) and 20 mM HEPES, pH 8.0) at

25 °C in the dark for 2 h. The resulting peptides ("Rev-Fl", “Rev-TR”, and "Rev-

IA", respectively) were then purified on a C-18 semiprep HPLC column with an

mixture of acetonitrile (0.1% TFA) and water (0.1% TFA); isocratic conditions

were 19% acetonitrile for Rev-Fl, 30% acetonitrile for Rev-TR, and 14%

acetonitrile for Rev-IA. Electrospray mass spectrometry confirmed masses for

Fl-Rev (3,079 amu), NH2-Rev (2,721 amu), RevFl (3,312 amu), Rev-TR (3,656

amu), and Rev-IA (2,982 amu). Analytical HPLC confirmed a greater than 98%

purity for each peptide. Molecular extinction coefficients for the fluorescein

200

labeled peptides were taken as 73,500 cm-1 M-1 for Fl-Rev (498 nm), 77,000 cm-1

M-1 for Rev-Fl (498 nm), 115,000 for Rev-TR (586 nm) (based upon the extinction

coefficients of the free dyes). The extinction coefficients for un-labeled peptides

have been taken as 5,600 cm-1 M-1 for both NH2-Rev and Rev-IA (280 nm) (based

upon the presence of a single Trp residue.46 Amino acid analysis of Rev-IA

(complete acid hydrolysis, followed by ninhydrin modification, and HPLC

quantification of each amino acid compared the HPLC integration of known

concentration of a standard) suggests, however, that the extinction coefficient of

Rev-IA is closer to 9,600 cm-1 M-1. The average values of amino acid hydrolysis

and UV-absorbance (7,600 cm-1 M-1) have been used for these studies.

Section E 2.3 Evaluation of Rev Peptides

The C-terminus modified peptides are found to be superior to the N-terminus

peptides. Gel shift mobility assays indicate that perturbation of RRE affinity is

significantly worse when the N-terminus modified peptide is conjugated to

fluorescein compared to the addition of fluorescein to the C-terminus modified

peptide (not shown). Furthermore, upon binding the RRE, fluorescence

quenching of "Fl-Rev" is much more significant (~30%) than quenching of "Rev-

Fl" (~10%) (this makes Rev-Fl better suited for fluorescence anisotropy

experiments). These differences can be rationalized by examining the solution

structure of the Rev-RRE complex which shows the N-terminus of the Rev

peptide buried in the major groove of the RRE and the C-terminus free in

solution.53

201

Section E 3.0 Conditions for Peptide Displacement Binding Assays

All RevFl peptide displacement experiments (including those using the solid-

phase assay and fluorescence anisotropy) were conducted at 22 °C in a buffer

containing 30 mM HEPES (pH 7.5), KCl (100 mM), sodium phosphate (10 mM),

NH4OAc (10 or 20 mM), guanidinium HCl (10 or 20 mM), MgCl2 (2 mM), NaCl (20

mM), EDTA (0.5 mM), and Nonidet P-40 (0.001%). This complex mixture of

cations and anions is found to minimize the non-specific binding of ligands to the

Rev-RRE complex and to maximize the reversibility of the Rev-RRE interaction

(as evident in both anisotropy and solid-phase assays).

Section E 3.1 Fluorescence Anisotropy

In a Perkin Elmer LS-50B luminometer equipped with polarizing filters, 10 nM of

RevFl in buffer, is excited at 490 nm and its emission is monitored at 530 nm.

Maximum slit widths and a Gf value of 1.006 are used for taking, at least, 6

anisotropy values for each data point. A RevFl-RNA complex is then formed by

adding 10 nM, 100 nM, or 120 nM of the pre-folded RNA. Upon binding the RRE

or TAR, only minor changes in the emission spectrum of fluorescein were seen

(about 10% quenching of RevFl). We have, therefore, taken the change in

anisotropy as being directly proportional to the fraction of RevFl bound. Upon

titration of the inhibitor, sufficient mixing time must be provided to allow for

equilibrium to be reached. For compounds that bind through electrostatic

interactions, equilibrium is reached almost immediately. For compounds that

202

bind, primarily, through hydrophobic interactions, 1-2 minutes is sometimes

needed to reach equilibrium.

Ki values have been calculated, from IC50 values by considering the following

equation:

Upon mixing 10 nM of RRE66 with 10 nM of RevFl, the concentration of RevFl-

RRE complex [RevFl-RRE] = 5.8 nM, and the concentrations of free RRE and

free RevFl are each 4.2 nM. Upon reaching the IC50 of the titration, the

concentration of the complex [RevFl-RRE] = 2.9 nM, [I-RRE] = 5 nM, [RevFl] =

7.1 nM and the concentration of free inhibitor [I] = (IC50 value – [I-RRE]).

Therefore, at 10 nM of RRE, Ki = ((IC50 – 5 nM) / 4)). Upon mixing 100 nM of

RRE66 with 10 nM of RevFl, the concentration of complex [RevFl-RRE] = 9.6

nM, the concentration of free RRE = 90 nM and free RevFl is 0.4 nM . Upon

reaching the IC50 of the titration, the concentration of the complex [RevFl-RRE] =

4.8 nM, [I-RRE] = 94 nM, [RevFl] = 5.2 nM and the concentration of free inhibitor

[I] = (IC50 value – [I-RRE]). Therefore, at 100 nM of RRE, Ki = ((IC50 – 94 nM) /

34)). Upon mixing 120 nM of RRE66 with 10 nM of RevFl, the concentration of

+

[ RRE ] + [ RevFl ] [ RevFl - RRE ]

[ I ]

Ki

[ I - RRE ]

[ RevFl ]

Kd [ RevFl-RRE ]

[ I - RRE ]Ki =

[ I ]

Kd = nM

203

complex [RevFl-RRE] = 9.7 nM, the concentration of free RRE = 110 nM and free

RevFl is 0.3 nM . Upon reaching the IC50 of the titration, the concentration of the

complex [RevFl-RRE] = 4.9 nM, [I-RRE] = 114 nM, [RevFl] = 5.1 nM and the

concentration of free inhibitor [I] = IC50 value – [I-RRE]. Therefore, at 120 nM of

RRE, Ki = ((IC50 – 114) / 40)).

E 3.2 The Solid-Phase Assay

E 3.2.1 Synthesis of biotinylated 66-nt RRE

See Figure E 3.1 for a flow chart of biotin-RRE synthesis. The synthesis of the 5'

biotinylated RRE transcripts is a modification of the procedure described.175 To

generate RNA transcripts with a phosphorothioate 5'-end, 8 mM guanosine

monophosphorothioate (purchased from USB) was included in the transcription

reaction (along with 2 mM of each NTP), this approach was found to be more

efficient than the previously described method.175 T7 RNA polymerase was

utilized for "run-off" in vitro transcription with a complementary, 83-nt DNA

template as described (Section E 2.0). Transcripts with a phosphorothioate 5’

end were reacted with Iodoacetyl-LC-Biotin (purchased from Pierce) as

described,175 purified and quantified as described in E 2.0. Before use, RRE

transcripts were folded in a buffer containing 30 mM HEPES (pH 7.5), KCl (100

mM), MgCl2 (2 mM), and EDTA (0.5 mM) by heating to 90 °C and cooling to

room temperature.

204

Figure E 3.1: Synthesis of biotinylated RRE66.

E 3.2.1 Assembly of the Solid-Phase Immobilized Rev-RRE Complex

“ImmunoPure” beaded agarose covalently modified with stretavidin was

purchased from Pierce. Immediately before use, the resin is washed with 3

volumes of buffer (Section E 3.0) with gentle mixing (by inversion of the tube).

After each wash (4 total) the slurry of solid-support and buffer is gently

centrifuged (1,000 rpm 30 s) to settle the agarose beads, and the buffer removed

by pipette. Following the last wash, 500 µl of dilute biotinylated RRE (400 nM in

buffer) is added to 1 mL of a 40% slurry (400 µl of beads and 600 µl of buffer)

and immediately inverted to mix. The slurry is incubated at room temperature for

1.5 h with constant inversion for mixing. By monitoring absorbance of the

HN NH

S

O

O

HN

NH

I

O

HN NH

SO

HN

NH

SO

OP

O

O

SP

O

OO

O

DNA Template

5'3'

T7 Promoter (18 mer)

5'

RNA Transcript3'

GMPS, NTPsT7 RNA Polymerase

5'

5'3'

Aq. buffer (pH 8.0)DMF

Biotinylated RNA Transcript

205

supernatant at 260 nm, the efficiency of RRE immobilization is calculated. The

loading efficiency is >95% for biotinylated RRE and is <1% for non-biotinylated

RRE. Following RRE immobilization, 1 equivalent of RevFl is added, incubated

for 1 h, and the supernatant quantified for fluorescence intensity. Under these

conditions, non-specific binding of RevFl to the solid-support is not observed (as

determined by addition of RevFl to streptavidin-agarose which lacks immobilized

RRE). Since a linear relationship between fluorescence intensity and

concentration of RevFl can be established by a calibration of UV absorbance

versus fluorescence emission (Figure 3.1), the concentration of RevFl in the

supernatant (following incubation with the immobilized RRE) can be calculated

from its fluorescence intensity. It is found that under the above loading

conditions, ~85% of RevFl is bound by the immobilized RRE. The Kd of the

RevFl-RRE interaction on solid support is then calculated by assuming that 85%

of the RRE is bound by RevFl (which is confirmed by mixing a known volume of

gel with 8M guanidinum HCl and quantifying the amount of RevFl present on the

gel itself).

E 3.2.2 Conducting the Solid-Phase Assay

In 0.75 mL siliconized tubes, 200 µL of buffer (see Section E 3.0) is added,

followed by 30 µL of a 30% Rev-RRE slurry (at 0.5 pmole RRE/µL of gel). A

small volume of concentrated inhibitor and/or competitor is added, mixed gently,

and incubated at room temperature for 90 minutes with constant inversion for

mixing. The tube is then gently centrifuged (1,000 rpm for 30 s) to settle the

206

beads. 200 µL of supernatant is then added to 800 µL of “quantification buffer”

(4M guanidinium HCl, 250 mM Tris/HCl (pH 9.0), 250 mM KCl, 50 mM MgCl2,

0.01% Nonidet P-40), and quantified for fluorescence intensity (Perkin Elmer LS-

50B fluorimeter, maximum slit widths, Excitation 490 nm). For the maximum

signal control (Figure 3.8), 200 µL of 8M guanidinium HCl is added instead of

buffer. For inhibitors that interfere with emission of RevFl, the quantity of RevFl

remaining on the agarose beads is quantified by removing the supernatant (as

above) then washing the beads 2 times with 200 µL of isopropanol. After each

wash the mixture is gently centrifuged (1,000 rpm, 30 s) to settle the agarose

beads, and isopropanol removed by pipette. 200 µL of 8M guanidinium HCl is

then added to the beads, mixed, heated to 55 °C, mixed again, then centrifuged

to settle the agarose beads. 200 µL of supernatant is then added to 800 µL of

“quantification buffer” (see above) and quantified for fluorescence intensity (as

above).

E 3.2.3 Gel Shift Experiments

To address the discrepancies between IC50 values derived from the solid-phase

assay and those from fluorescence anisotropy, gel shift experiments were

conducted (Figure 3.9). A Rev-RRE complex was formed by mixing 10 nM RRE

(a fraction 5'-end labeled with 32P) with 80 nM “Rev-IA” (see above) in buffer

(Section E 3.0). This mixture was incubated at room temperature for 15 min then

9 µL aliquots were added to tubes containing 1 µL of a inhibitor solution, the

mixture was incubated for an additional 15 min at RT and loaded onto a native

207

17% polyacryamide gel. The running buffer and gel (1:29 bis:acrylamide) were

both at 1X TBE. The gel was run overnight at a constant 300 V and exposed to a

Molecular Dynamics Phosphor Screen and Phosphorimager™ 445 SI and

analyzed with Imagequant™ software (Molecular Dynamics).

E 3.3 Ethidium Bromide Displacement

A 1.25 µM solution of ethidium bromide is excited at 546 nm, and its fluorescence

emission is monitored at 605 nm before and after the addition of nucleic acids.

For the RRE and TAR constructs, 50 nM of each stand is used, for the polymeric

nucleic acids, 0.3 µM, 0.88 µM, or 11 µM of base pairs is used. Under these

conditions, only a small fraction of the ethidium bromide is bound (less than

20%). Inhibitors are then titrated until the fluorescence decrease reaches

saturation (Figure 3.12).

E 3.4 Direct Observation of Small Molecule Nucleic Acid Association : 9-Amino Acridine Fluorescence Anisotropy

Similar to E 3.1, a 20 nM solution of 9-aminoacridine, in buffer (Section E 3.0),

was excited at 400 nm and emission monitored 431 nm with maximum slit widths

and Gf set to 1.000. RRE 66 was then titrated. The fluorescence intensity of 9-

aminoacridine remains constant (111 arb. units) through out the titration.

Interestingly, under these conditions, the acridine-glycoside conjugates (38 – 42)

show significant quenching upon binding nucleic acids. The “direct” binding

isotherms for these compounds, therefore, are determined by monitoring the

changes in fluorescence intensity.

208

E 3.5 Direct Observation of Small Molecule Nucleic Acid Association :Thermal Denaturation:

Tm experiments are conducted with a Varian Cary 1E UV-vis spectrophotometer

equipped with a temperature control unit. 23 µM (base pairs) of Calf Thymus, in

buffer, is denatured at 0.5 °C/min and absorbance is monitored at 260 nm in

either the presence or absence of 10 µM of either 9 or 10. The temperature of

melting is taken at the inflection point of the transition.

E 3.5.1 Direct Observation of Small Molecule Nucleic Acid Association: 2-AP and the HH16 Ribozyme:

2-AP containing substrates were purchased from Dharmacon. All deprotected

RNA sequences were purified by preparative 20% denaturing polyacrylamide gel

electrophoresis and quantified as described (Section E 2.0). The following molar

extinction coefficient of 2-AP was taken as 1,000 cm-1M-1. RNAs were quantified

in 8 M urea instead of using base-hydrolysis. All HH16 experiments were

conducted at 37 °C in a buffer containing 50 mM Tris-HCl (pH 6.5–8.5) and NaCl

(0–500 mM), cleavage reactions contain 10 mM MgCl2. Fluorescence emission

measurements were taken by exciting the samples at 305 nm and monitoring

emission at 370 nm. Enzyme-substrate association experiments were conducted

in 3 mL solution of buffer (with no MgCl2) containing 20 nM of the 2-AP-

containing substrate. The fluorescence emission intensity of the substrate was

monitored upon titration of the 38-mer enzyme. A 5 minute equilibration time was

allowed after each addition of enzyme, before quantifying the emission intensity

of the substrate (Figure 3.18). Cleavage experiments were conducted in a

masked micro fluorescence cuvette thermoregulated at 37 °C. Buffered solutions

209

(35 µl) of the substrate and enzyme were denatured separately by heating to 90

°C for 90 sec and cooling to room temperature over 10 min to allow for refolding.

MgCl2 (10 mM) and inhibitors (when appropriate) were added to both the enzyme

and substrate and allowed to equilibrate at 37 °C for 5 min. The cleavage

reaction was then initiated by manually mixing the substrate with the enzyme in a

heat block or a fluorescence cuvette preheated to 37 °C. For all cleavage

experiments the final concentration of substrate and enzyme is 1.85 µM and 10.3

µM, respectively. Since cleavage is conducted under single-turnover conditions,

the pseudo first order rate constant k2 can be easily extracted from

measurements of the ribozyme’s initial rate of cleavage. Since the cleavage

reaction occurs within the hammerhead complex, it can be treated as a first order

intramolecular reaction. Rate constants (k2) were calculated as the slope of ln(1–

S/S0) vs. time, where S/S0 is the fraction of cleaved substrate. For experiments

utilizing a radioactively labeled substrate, S/S0 was determined by dividing the

amount of cleaved substrate by the sum of the full length and cleaved substrate.

The initial point (y0) of the fluorescently-monitored reactions was derived by fitting

the raw data to an exponential curve fit y = y0 + a(1 – e–bx), where y =

fluorescence intensity, a = the rise, b = the rate and x = time. The end point (f)

was separately determined by measuring the fluorescence of individually

synthesized cleavage products (P1 and P2) annealed to the enzyme (E) under

the same reaction conditions; hence S/S0 (fraction cleaved) = (y – y0)/(f – y0).

Using this method the percent cleavage after 20 min was essentially the same for

the radioactive- and fluorescence-generated data; both reactions were found to

210

reach 75% of completion. The IEC50 values (concentration of inhibitor leading to

a 50% reduction of ribozyme cleavage rate) were determined by measuring the

concentration of inhibitor needed to decrease the initial slope (between 30

seconds and two minutes) by 50%. The IEC50 values are used only for

comparative purposes and are not to be confused with actual k2 values.

E 4.0.1 Circular Dichroism of Metal Complexes

CD spectra were collected on an AVIV model 303A Circular Dichroism

Spectrometer using a 1cm pathlength quartz cell containing 50 mM sodium

phosphate pH 7.5. Spectra were collected with 16 µM of either 9 or 10 and these

spectra were corrected by subtraction of a “buffer-only” spectrum. The raw data

(ellipticity in mdeg) was converted into molar ellipticity ([Θ]) with the following

formula: [Θ] = Θ (mdeg) / (10*C*l)

Where C = concentration and l = path length in cm.

The units for [Θ] are deg * cm2 * decimole-1

E 4.0.2 HeLa Plaque Assay.

Two independent sets of duplicate points were collected at the CFAR, UCSD as

described.110 HT-6C cells were grown and assayed for plaque forming units

(PFUs), in Dulbecco's Modified Eagle's Medium containing: 10% fetal calf serum

(FCS), 2 mM glutamine, and 100 µg/mL penicillin and streptomycin. 9 and 10

were tested in the presence and absence of Polybrene© and the same IC50

values were measured. Cells were seeded in 24-well Falcon plates at 2.5 × 104

211

cells/well and incubated overnight (37 oC in the presence of CO2). The HIV-1

strain LAI X794 was then added such that 70 ± 10 PFUs/well were apparent after

an additional 3 d incubation. Inhibitors were added 2 h following the addition of

HIV-1 and incubated (as above) for 3 d. Cells were then washed in MeOH and

stained with Crystal Violet (0.5% in H2O). Cell density and PFUs were counted

and compared to no-inhibitor controls. As a control, AZT is included in each

round of testing and consistently shows an IC50 value of approximately 0.01 µM.

E 4.0.3 PBMC Assay

Experiments were conducted at the CFAR, UCSD as described.110 Peripheral

blood monoocytes (PBMCs) were obtained from an HIV- negative human donor

and maintained in RPMI 1640 (Bio Wittaker Corp.) containing: 10% FCS,

100µg/ml pen/strep, 10 units/ml IL-2 and were stimulated with 3µg/ml PHA lectin

(Sigma) from 48 to 72 hours following isolation. The HIV-1 strain X794 LAI, is

then added to a 1/1000 multiplicity of infection, incubated for two hours at 37 °C

(in the presence of CO2), washed two times in PBS, resuspended in media

(above), and added to 96-well plates at 2x105 cells per well. Inhibitor is then

added (pre-suspendended in media and filtered). Plates are then incubated an

additional 4 days and the concentration of p24 is measured for each sample and

compared to a no-drug controls. As a control, AZT was included and showed an

IC50 value of less than 0.001 µM.

212

E 4.0.4 Viscosity Experiments

A Cannon-Manning semi-micro size 75 viscometer was submerged in a glass

beaker and thermoregulated at 25.9 ± 0.1 °C. Concentrated stocks (10 mM) of

either ethidium (12) or rac-(9/10) were titrated into 0.5 mL buffered solutions (50

mM sodium phosphate pH 7.5) of CT DNA (0.5 mM bases). Flow times ranged

between 126.4 and 160.5 seconds depending the concentration of small

molecule. The η is the viscosity of the solution in the presence of a ligand, ηo is

the viscosity of the DNA-only solution. Viscosity is calculated as η = t – to where t

= time for the sample to flow through the viscometer and to is the time needed for

the buffer only.

E 4.0.5 UV-Vis Titrations

UV-vis spectra were collected using a Hewlett Packard 8452A diode array

spectrophotometer. 8 µM of either 9 or 10 in buffer (Section E 3.0) were

monitored as small volumes of concentrated nucleic acids were added. See E

2.0 for source and quantification of nucleic acids. The fractional change at

multiple wavelengths were averaged to calculate each IC50 value. Upon titration

of TAR31, RREJW, and tRNAPhe both 9 and 10 exhibited the same spectral

changes as observed for CT DNA (Figure 4.4). Significant diversity, however, is

observed in the magnitudes of red-shifting and the changes in epsilon upon

addition of the polymeric nucleic acids (especially for the single stranded

polymers):

213

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[A]–r[U] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[A]–r[U]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[G]–r[C] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[G]–r[C]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly d[AT]–d[AT] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly d[AT]–d[AT]

214

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly d[GC]–d[GC] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly d[GC]–d[GC]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly d[G]–d[C] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly d[G]–d[C]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly d[AT]–d[AT] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly d[AT]–d[AT]

215

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[I]–r[C] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[I]–r[C]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[A] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[A]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[U] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[U]

216

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[C] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[C]

ΛΛΛΛ-[Ru(bpy)2Eilatin]2+ / poly r[G] ∆∆∆∆-[Ru(bpy)2Eilatin]2+ / poly r[G]

E 4.0.6 Thermal Denaturation Experiments

A Beckman Coulter DU 640 spectrophotometer equipped with a Peltier was used

to monitor buffered solutions (10 mM sodium phosphate pH 7.5, 1 mM MgCl2, 50

mM KCl) containing 30 µM (bases) of poly r(U), and/or 15 µM of either 9 or 10.

Heating was at the rate of 0.5 °C/min and all samples were monitored at 260 nm

(Figure 4.7).

217

E 4.0.7 Eilatin Titrations

In a Perkin Elmer LS-50B luminometer a dilute solution of eilatin (14) (0.1 µM) in

buffer (Section E3.0) was excited at 417 nm using maximum slit widths. The

emission of the solution was the monitored as small volumes of highly

concentrated nucleic acids are titrated (Figure 4.10).

E 5.1.1 Ethidium Chloride Crystal Structure

Ethidium chloride was crystallized by vapor infusion. In a test tube, ethidium was

mixed with chloroform, then methanol was added drop-wise until the ethidium

was fully dissolved. The tube was then sealed in a container containing diethyl

ether. Crystal growth was apparent after approximately 5-7 days. A trace of

bromide counter ion (13 %) was observed in place of chloride (87%).

Figure E 5.0: ORTEP drawing of ethidium chloride.

218

Table E 5.1: Crystallographic parameters for ethidium chloride.

Formula weight 389.34

Temperature 100(2) K

Wavelength 0.71073 Å

Crystal system Monoclinic

Space group P2(1)/n

Unit cell dimensions a = 9.6706(9) Å α= 90°.

b = 10.4323(10) Å β= 100.786(2)°.

c = 19.6144(19) Å γ = 90°.

Volume 1943.9(3) Å3

Z 4

Density (calculated) 1.330 Mg/m3

Absorption coefficient 0.535 mm-1

F(000) 820

Crystal size 0.644 x 0.199 x 0.156 mm3

Theta range for data collection 2.11 to 27.54°.

Index ranges -12<=h<=12, -12<=k<=13, -25<=l<=25

Reflections collected 16036

Independent reflections 4381 [R(int) = 0.0275]

Completeness to theta = 27.54° 97.6 %

Absorption correction None

Refinement method Full-matrix least-squares on F2

Data / restraints / parameters 4381 / 0 / 266

Goodness-of-fit on F2 1.051

Final R indices [I>2sigma(I)] R1 = 0.0375, wR2 = 0.0970

R indices (all data) R1 = 0.0432, wR2 = 0.1004

Largest diff. peak and hole 0.483 and -0.228 e.Å-3

E 5.1.2 3-Cbz Ethidium Chloride Crystal Structure

3-Cbz ethidium chloride (16) was crystallized by vapor infusion. In a test tube, it

was mixed with chloroform, then a few drops of methanol were added until 16

219

was fully dissolved. The tube was then sealed in a container containing diethyl

ether. Crystal growth was apparent after approximately 1-2 days.

Figure E 5.1: ORTEP drawing of 3-cbz ethidium chloride.

Figure E 5.1 ORTEP drawing of 3-cbz ethidium chloride.

Table E 5.2: Crystallographic parameters for 3-cbz ethidium chloride.

Formula weight 530.05

Temperature 100(2) K

Wavelength 0.71073 Å

Crystal system Monoclinic

Space group P2(1)/n

Unit cell dimensions a = 13.2854(17) Å α= 90°.

b = 10.5945(13) Å β= 90.567(2)°.

c = 18.976(2) Å γ = 90°.

Volume 2670.8(6) Å3

Z 4

Density (calculated) 1.318 Mg/m3

Absorption coefficient 0.181 mm-1

F(000) 1120

Crystal size 0.73 x 0.17 x 0.17 mm3

Theta range for data collection 1.86 to 27.50°.

Index ranges -17<=h<=17, -13<=k<=13, -24<=l<=23

Reflections collected 21640

Independent reflections 5982 [R(int) = 0.0315]

220

Completeness to theta = 27.50° 97.6 %

Absorption correction None

Max. and min. transmission 0.9700 and 0.8787

Refinement method Full-matrix least-squares on F2

Data / restraints / parameters 5982 / 0 / 348

Goodness-of-fit on F2 1.045

Final R indices [I>2sigma(I)] R1 = 0.0463, wR2 = 0.1265

R indices (all data) R1 = 0.0536, wR2 = 0.1333

Largest diff. peak and hole 1.047 and -0.387 e.Å-3

E 5.1.3 8-Cbz Ethidium Chloride Crystal Structure

8-Cbz ethidium chloride (17) was crystallized by vapor infusion. In a test tube, it

was mixed with chloroform, then a few drops of ethanol were added until 17 was

fully dissolved. The tube was then sealed in a container containing diethyl ether.

Crystal growth was apparent after approximately 1-2 days.

Figure E 5.2: ORTEP drawing of 8-cbz ethidium chloride.

221

Table E 5.2: Crystallographic parameters for 8-cbz ethidium chloride.Formula weight 603.35

Temperature 100(2) K

Wavelength 0.71073 Å

Crystal system Monoclinic

Space group P2(1)/c

Unit cell dimensions a = 12.6040(8) Å α= 90°.

b = 13.3162(9) Å β= 107.8060(10)°.

c = 17.4523(11) Å γ = 90°.

Volume 2788.8(3) Å3

Z 4

Density (calculated) 1.437 Mg/m3

Absorption coefficient 0.459 mm-1

F(000) 1248

Crystal size 0.47 x 0.11 x 0.10 mm3

Theta range for data collection 1.70 to 27.52°.

Index ranges -16<=h<=16, -17<=k<=17, -21<=l<=22

Reflections collected 23292

Independent reflections 6310 [R(int) = 0.0252]

Completeness to theta = 27.52° 98.3 %

Absorption correction None

Refinement method Full-matrix least-squares on F2

Data / restraints / parameters 6310 / 0 / 354

Goodness-of-fit on F2 0.951

Final R indices [I>2sigma(I)] R1 = 0.0418, wR2 = 0.1094

R indices (all data) R1 = 0.0464, wR2 = 0.1129

Extinction coefficient 0.0000(3)

Largest diff. peak and hole 0.538 and -0.364 e.Å-3

222

E 5.2.1 Cbz Protection of Ethidium:

8-Cbz ethidium · Cl (17), 3-cbz ethidium · Cl (16), and 3,8-bis cbz ethidium ·

Cl (18). Ethidium bromide containing 8% water (4.13 g, 9.64 mmoles) was

dissolved in 0.2 M sodium phosphate pH 6.6 (100 mL), acetone (80 mL) and

warmed to 32 °C. To this, a solution containing 1.43 mL of cbz chloride (10

mmoles, 1 equiv) in 20 mL of acetone was slowly added and the reaction

warmed to 40 °C for 20 min. AG1-X4 (Cl-) exchange resin was then added (20 g,

70 mmoles, 7 mequiv) and stirred 5 min, 40 °C. The slurry was loaded onto a

column containing another 20 g (7 mequiv) of AG1-X4 (Cl-) exchange resin and

the eluent collected. The resin was washed with 30 mL of 1:1 water/acetone, and

the eluents were combined and reduced to a solid under reduced pressure. The

products were separated on silica gel using three consecutive columns (8 – 10%

Br

N

HNH2N

ClO

Cl O

O

N

NH2HN

Cl

O

O

N

HN

HN

Cl

O

OO

O

ON

NH2H2N acetone/phosphate bufferpH 6.6

(1 equiv.)

ethidium bromide (12)

8-cbz ethidium chloride (17)

3-cbz ethidium chloride (16)

3,8-bis cbz ethidium chloride (18)

(50%)

(10%)

(4%)

+

+

+

+

-

-

-

-

223

MeOH / CH2Cl2, 5 – 12% MeOH / CH2Cl2, and 10% MeOH / CH2Cl2). The pure

fractions from each column were combined to yield : 2.3 g of the purple solid, 8-

cbz ethidium · Cl (17) (50%). Rf = 0.38 (20% MeOH / CHCl3).1H NMR (400

MHz, d6-DMSO, 25 °C): δ 10.31 (s, 1H), δ 8.83 (d, J = 9.6 Hz, 1H), δ 8.78 (d, J =

9.2 Hz, 1H), δ 8.11 (d,d J1 = 9.6 Hz, J2 = 1.6 Hz, 1H), δ 7.71-7.77 (m, 5H), δ 7.64

(d, J = 1.6 Hz, 1H), 7.44 (s, 1H), δ 7.36-7.42 (m, 6H), 6.66 (s, 2H), 5.08 (s, 2H), δ

4.49 (q, J = 7.2 Hz, 2H), δ 1.42 (t, J = 7.2 Hz, 3H). ESI MS calculated for

C29H26N3O2: 448.2, found 448.3 [M]+. UV-vis (50 mM sodium phosphate pH 7.5):

λmax (nm) and ε (cm-1M-1): 214 (3.4x104), 286 (4.4x104), 460 (4.5x103). 0.48 g of

the orange-red solid, 3-cbz ethidium · Cl (16) (10%). Rf = 0.5 (20% MeOH /

CHCl3).1H NMR (400 MHz, d6-DMSO, 25 °C): δ 10.61 (s, 1H), δ 8.92 (d, J = 9.2

Hz, 1H), δ 8.76 (d, J = 9.2 Hz, 1H), δ 8.67 (s, 1H), δ 7.97 (d, J = 9.2 Hz, 1H), δ

7.72-7.79 (m, 5H), δ 7.58 (d,d J1 = 9.2 Hz, J2 = 2.2 Hz, 1H), δ 7.36-7.48 (m, 5H),

δ 6.38 (d, J = 2.2 Hz, 1H), 6.22 (s, 2H), 5.24 (s, 2H), δ 4.54 (q, J = 7.0 Hz, 2H), δ

1.45 (t, J = 7.0 Hz, 3H). ESI MS calculated for C29H26N3O2: 448.2, found 448.3

[M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 212

(4.1x104), 284 (5.6x104), 454 (4.9x103). 0.23 g of the yellow solid, 3,8-bis cbz

ethidium · Cl (18) (4%). Rf = 0.67 (20% MeOH / CHCl3).1H NMR (400 MHz, d6-

DMSO, 25 °C): δ 10.76 (s, 1H), δ 10.45 (s, 1H), δ 8.85 (d, J = 9.2 Hz, 1H), δ 9.03

(d, J = 9.2 Hz, 1H), δ 8.78 (s, 1H), δ 8.28 (d,d J1 = 9.2 Hz, J2 = 2.0 Hz, 1H), δ

8.10 (d, J = 9.2 Hz, 1H), δ 7.74-7.81 (m, 6H), δ 7.33-7.49 (m, 10H), δ 5.26 (s,

224

2H), 5.10 (s, 2H), δ 4.62 (q, J = 7.0 Hz, 2H), δ 1.49 (t, J = 7.2 Hz, 3H). ESI MS

calculated for C37H32N3O4 : 582, found 582 [M]+.

E 5.2.2 Synthesis of 3-Guanidino Ethidium Dichloride (19):

3-diBoc guanidino 8-Cbz ethidium · Cl. 8-cbz ethidium · Cl (17) (48 mg, 100

µmoles, DMF (4 mL), N,N’ bis BOC-S-methyl-isothiourea (145 mg, 500 µmoles, 5

equiv), and mercury dichloride (227 mg, 837 µmoles, 8.4 equiv) were combined

and sonicated. 2,4,6 collidine (177 µL, 1.34 mmoles, 13.3 equiv) was added

drop-wise and the reaction was stirred RT, 15 min with occasional sonication.

The reaction was then moved into CHCl3 (150 mL) and washed with 0.1 M citric

acid (4x40 mL), brine (40 mL), dried over sodium sulfate and concentrated to a

solid under reduced pressure. The product was purified on a short (2 in) silica gel

column (20 mL) using a gradient (0-5% MeOH/CHCl3) to yield 47 mg of a solid

S

HNN NHNHN

Cl

O

ONH

N

OO

OO

N

NH2HN

NH2

N

ClCl

H

H

DMF, HgCl2, collidine (70%)

6M HCl / MeOH (99%)

OO

OO

3-guanidino ethidium dichloride (19)

NHNH2N

Cl

O

O8-cbz ethidium chloride (17) 3-diBoc guanidino

8-cbz ethidium chloride

+ +- -

+-

225

yellow product (70%). 1H NMR (400 MHz, d6-DMSO, 25 °C): δ 11.32 (s, 1H), δ

10.51 (s, 1H), δ 10.44 (s, 1H), δ 9.23 (s, 1H), δ 9.06-9.09 (m, 2H), δ 8.26 (d, J =

9.2 Hz, 1H), δ 8.17 (d, J = 8.4 Hz, 1H), δ 7.77-7.80 (m, 7H), δ 7.35-7.38 (m, 5H),

δ 5.10 (s, 2H), δ 4.68 (q, J = 7.2 Hz, 2H), δ 1.33-1.56 (m, 19H).

3-Guanidino ethidium · 2HCl (19). 3-diBoc guanidino 8-cbz ethidium · Cl. (20

mg, 29 µmoles) was dissolved in methanol (2 mL), saturated HCl (2 mL) and

heated at 120 °C for 40 min. The reaction was then moved onto ice and 2 M

NaOH was added drop-wise until the yellow solution started to turn orange. The

solution was loaded directly onto an activated Water’s “Sep-pack” C-18 reversed

phase column (activated with 10 mL acetonitrile, 10 mL water) and washed with

water (5 mL), and the product was eluted with 20-30% acetonitrile/water (0.01M

HCl) and lyophilized to 13 mg (99%) of an orange solid. 1H NMR (300 MHz, d6-

DMSO, 25 °C): δ 10.61 (s, 1H), δ 9.01 (d, J = 9.0 Hz, 1H), δ 8.85 (d, J = 9.3 Hz,

1H), δ 8.34 (s, 1H), δ 7.74-7.91 (m, 10H), δ 7.62 (d,d J1 = 9.0 Hz, J2 = 2.1 Hz,

1H), δ 6.43 (d, J = 2.1 Hz, 1H), δ 6.35 (br s, 2H), δ 4.66 (q, J = 6.9 Hz, 2H), δ 1.43

(t, J = 7.1 Hz, 3H). FAB MS calculated for C22H22N5 : 356.1876, found 356.1892

[M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 214

(3.3x104), 284 (4.8x104), 444 (4.1x103).

226

E 5.2.3 Synthesis of 8-Guanidino Ethidium Dichloride (20):

8- diBoc guanidino 3-cbz ethidium · Cl. 3-cbz ethidium · Cl (16) (37 mg, 76

µmoles, DMF (4 mL), N,N’ bis BOC-S-methyl-isothiurea (112 mg, 290 µmoles,

3.8 equiv), and mercury dichloride (175 mg, 645 µmoles, 8.5 equiv) were

combined, sonicated, and 2,4,6 collidine (136 µL, 1.03 mmoles, 13.5 equiv) was

added drop-wise and the reaction was stirred at RT for 30 min with occasional

sonication. The reaction was them moved into CHCl3 (200 mL) and washed with

0.1 M citric acid (3x50 mL), brine (50 mL), dried over sodium sulfate and

concentrated to a solid under reduced pressure. The product was purified on a

short (2 in) silica gel column (20 mL) using a gradient (0-5% MeOH/CHCl3) to

yield 40 mg of a solid yellow product (76%). 1H NMR (300 MHz, d6-DMSO, 25

°C): δ 10.92 (s, 1H), δ 10.74 (s, 1H), δ 10.14 (s, 1H), δ 9.13 (d, J = 9.3 Hz, 1H), δ

9.04 (d, J = 9.3 Hz, 1H), δ 8.78 (s, 1H), δ 8.31 (d,d J1 = 9.0 Hz, J2 = 2.1 Hz, 1H),

S

HNNN

HN

HN

Cl

N

HNH2N

Cl

DMF, HgCl2, collidine (76%)

OO

OON

NH2HN

Cl

O

O

O

O HN

N

OO

O

O

H2N

N

Cl

H

H

6M HCl (95%)

8-guanidino ethidium dichloride (20)

8-diBoc guanidino3-cbz ethidium3-cbz ethidium (16)

+

+

+

-

-

-

227

δ 8.08 (d, J = 8.7 Hz, 1H), δ 7.85 (d, J = 2.1 Hz, 1H), δ 7.75-7.77 (m, 5H), δ 7.37-

7.49 (m, 5H), δ 5.26 (s, 2H), δ 4.59 (q, J = 6.3 Hz, 2H), δ 1.31-1.51 (m, 19H).

8-Guanidino ethidium · 2HCl (20). 8-diBoc guanidino 3-cbz ethidium · Cl. (6 mg,

8.7 µmoles) was dissolved in 6M HCl (2 mL) and heated to 100 °C for 1 hr. The

reaction was then moved onto ice and NaHCO3 was added until the yellow

solution turned orange. The solution was loaded directly onto an activated

Water’s “Sep-pack” C-18 reversed phase column (activated with 10 mL

acetonitrile, 10 mL water). The column was washed with 1M NaCl (5 mL), water

(5 mL) and the product eluted with 25% acetonitrile/water (0.01M HCl) and

lyophilized to 3.5 mg (95%) of an orange solid. 1H NMR (300 MHz, d6-DMSO, 20

°C): δ 10.28 (s, 1H), δ 8.91 (d, J = 9.6 Hz, 1H), δ 8.88 (d, J = 9.3 Hz, 1H), δ 8.02

(d,d J1 = 9.3 Hz, J2 = 2.1 Hz, 1H), δ 7.72-7.85 (m, 9H), δ 7.47 (s, 1H), δ 7.43 (d, J

= 8.7 Hz, 1H), 7.02 (d, J = 2.1 Hz, 1H), δ 6.78 (br s, 2H), δ 4.52 (q, J = 6.9 Hz,

2H), δ 1.43 (t, J = 6.9 Hz, 3H). FAB MS calculated for C22H22N5: 356.1876, found

356.1862 [M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-

1M-1): 213 (2.8x104), 242 (1.5x104), 288 (3.6x104), 453 (4.4x103).

228

E 5.2.4 Synthesis of 3,8-bis Guanidino Ethidium Dichloride (21):

3,8-bis Guanidino ethidium · 2HCl (21). Ethidium bromide / 8% water (12) (30

mg, 70 µmoles) was dissolved in DMF (3 mL), brought to 0 °C, and N, N’ bis

BOC-S-methyl-isothiourea (180 mg, 620 mmoles, 4.4 mequiv), mercury

dichloride (282 mg, 1.04 mmoles, 7.4 mequiv) and 2,4,6 collidine (220 µL, 1.67

mmoles, 12 mequiv) were added. The reaction was slowly warmed to RT, stirred

for an additional 12 h., then diluted into 150 mL CHCl3 and washed with 0.1 M

citric acid (3x50 mL), 50 g/L of EDTA (40 mL), brine (40 mL) dried over sodium

sulfate and concentrated to a solid under reduced pressure. Silica gel (40 mL)

was used to purify the BOC-protected product (1.5 – 5% MeOH / CHCl3) to yield

40 mg of a yellow solid. All of the product was then deprotected by adding TFA (4

mL, containing 2.5 % (v/v) of triisopropyl silane) and mixing for 1 hr at RT, it was

then diluted into 150 mL water and washed with diethyl ether (3x40 mL) and

CHCl3 (3x40 mL). The aqueous phase was concentrated to a solid, then

dissolved in water (5 mL) and treated with 1.5 g of AG1-X4 (Cl-) exchange resin

(20 g, 5.2 mmoles, 74 mequiv) for 5 min RT. and filtered over an activated

Water’s “Sep-pack” C-18 reversed phase column (activated with 10 mL

acetonitrile, 10 mL water), the remainder of the product eluted from the column

S

HNNN

HN

HN

Cl

DMF, HgCl2, collidineOO

OO

3,8-bis guanidino ethidium dichloride (21)

N

NH2H2NH2N

N

Cl

H

H

Br

NH2

N

H

2) TFA/TIPS (76% 2 steps)

1)

Ethidium bromide (12)

++

- --

+

229

using 20% acetonitrile / water and lyophilized to 25 mg of a yellow solid (76%).

1H NMR (300 MHz, D2O, 25 °C): δ 8.90 (d, J = 9.0 Hz, 1H), δ 8.83 (d, J = 9.3 Hz,

1H), δ 8.22 (d, J = 1.8 Hz, 1H), δ 7.96 (d,d J1 = 8.7 Hz, J2 = 2.1 Hz, 1H), δ 7.84

(d,d J1 = 9.0 Hz, J2 = 1.8 Hz, 1H), δ 7.57-7.65 (m, 3H), δ 7.42-7.46 (m, 2H), 7.32

(d, J = 2.4 Hz, 1H), δ 4.73 (q, J = 7.2 Hz, 2H), δ 1.39 (t, J = 7.2 Hz, 3H). ESI MS

calculated for C23H24N7: 398.2, found 398.3 [M]+. UV-vis (50 mM sodium

phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 213 (3.6x104), 278 (5.4x104), 398

(5.0x103).

E 5.2.5 Synthesis of 3-Urea Ethidium Chloride (22)

Ethidium Activation for Urea Formation:

3-Phenoxycarbamate 8-cbz ethidium · H2PO4. 8-cbz ethidium · Cl (17) (180

mg, 372 µmoles), acetone (12 mL), and 500 mM sodium phosphate pH 6.6 (4

mL) were combined and phenyl chloroformate (200 µL, 1.58 mmoles, 4.2

mequiv, diluted into 2 mL of acetone) was added drop-wise and stirred for 2h at

RT. Water (5 mL) was then added drop-wise and the precipitate was collected by

vacuum filtration. The precipitate was then washed with water (10 mL), 3:1 water

/ acetone (10 mL) and eluted from the filter with methanol (200 mL). The

N

HNH2N

Cl

O

O

O

ClO

N

HNN

H

H2PO4

O

O

O

O

8-cbz ethidium (17)3-phenoxy carbamate8-cbz ethidium phosphate

acetone/ phosphatebuffer pH 6.6 (90%)

+

+

-

-

230

methanolic fraction was then concentrated to a solid under reduced pressure to

211 mg of a yellow solid (90%). 1H NMR (400 MHz, d6-DMSO, 20 °C): δ 11.20

(s, 1H), δ 10.46 (s, 1H), δ 9.13 (d, J = 8.8 Hz, 1H), δ 9.05 (d, J = 9.2 Hz, 1H), δ

8.82 (d, J = 1.2 Hz, 1H), δ 8.29 (d,d J1 = 9.2 Hz, J2 = 2.0 Hz, 1H), δ 8.14 (d,d J1 =

8.8 Hz, J2 = 1.2 Hz, 1H), δ 7.76-7.81 (m, 6H), δ 7.45-7.49 (m, 2H), δ 7.30-7.39

(m, 8H), δ 5.10 (s, 2H), δ 4.62 (q, J = 7.2 Hz, 2H), 1.49 (t, J = 7.6 Hz, 3H). ESI

MS calculated for C36H30N3O4: 568.2, found 568.3 [M]+.

3-Urea ethidium · Cl (22). In a 15 mL pressure tube, 3-phenoxycarbamate 8-

cbz ethidium · H2PO4 (40 mg, 60 µmoles) and methanol (8 mL) were mixed and

brought to -78 °C whereupon approximately 3 mL of ammonia was added, the

pressure tube sealed and allowed to warm to RT. The reaction was then heated

at 76 °C for 2 hr, cooled back to -78 °C, the tube opened and the ammonia was

out-gassed by passing argon into the solution as it slowly warmed to RT. All

volatiles were then removed under reduced pressure. The solid yellow product

N

HNN

H

H2PO4

O

O

O

O N

HNN

HO

O

H2N

O

H2PO4

N

NH2NH

H2N

ON

NH2HN

ClCl

NH3 / MeOH

70 % (2 steps)

3-phenoxy carbamate8-cbz ethidium phosphate

6M HCl / MeOH

(not characterized)

20%

3-urea ethidium chloride (22)ethidium chloride (12)

++

++

--

- -

231

was then deprotected with of 1:1 methanol / saturated HCl (8 mL) by heating at

96 °C for 1 hr. The reaction was then concentrated to a solid under reduced

pressure and purified by reversed-phase chromatography (C-18 silica gel 60).

The column was conditioned with methanol, water, and the crude product was

then loaded in water (0.01M HCl) and a methanol gradient (0-10% methanol /

water (0.01M HCl)) was used to separate ethidium chloride (elutes first) from the

desired product (elutes second) to yield 16.5 mg of an orange solid (70%). 1H

NMR (400 MHz, d6-DMSO, 20 °C): δ 9.58 (s, 1H), δ 8.83 (d, J = 9.2 Hz, 1H), δ

8.75 (d, J = 1.2 Hz, 1H), δ 8.73 (d, J = 9.2 Hz, 1H), δ 7.85 (d,d J1 = 9.2 Hz, J2 =

2.0 Hz, 1H), δ 7.73-7.78 (m, 5H), δ 7.57 (d,d J1 = 8.8 Hz, J2 = 2.4 Hz, 1H), δ 6.35

(d, J = 2.0 Hz, 1H), 6.29 (br s, 2H), 6.04 (br s, 2H), δ 4.53 (q, J = 7.6 Hz, 2H), δ

1.45 (t, J = 7.0 Hz, 3H). ESI MS calculated for C22H21N4O: 357.2, found 357.3

[M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 214

(2.4x104), 284 (3.5x104), 458 (3.1x103).

E 5.2.6 Synthesis of 8-Urea Ethidium Chloride (23):

Ethidium Activation for Urea Formation:

O

ClO

acetone/ phosphatebuffer pH 6.6 (79%)

N

NH2HN

Cl

O

O

3-cbz ethidium (16)

N

NH

HNO

O

O

O

3-cbz 8-phenoxycarbamateethidium phosphate

H2PO4 ++-

-

232

3-Cbz 8-phenoxycarbamate ethidium · H2PO4. 3-cbz Ethidium · Cl (16) (180

mg, 372 µmoles), acetone (10 mL), and 500 mM sodium phosphate pH 6.6 (4

mL) were combined and phenyl chloroformate (200 µL, 1.58 mmoles, 4.2

mequiv, diluted into 2 mL of acetone) was added drop-wise and stirred for 30 min

at RT. Water was then added (6 mL) drop-wise and the precipitate collected by

vacuum filtration and washed with water (20 mL). The precipitate was dried

under reduced pressure for 30 min to yield 195 mg (79%) of a yellow solid. 1H

NMR (300 MHz, d6-DMSO, 20 °C): δ 10.88 (s, 1H), δ 10.74 (s, 1H), δ 9.11 (d, J =

9.3 Hz, 1H), δ 9.07 (d, J = 9.0 Hz, 1H), δ 8.29 (d,d J1 = 9.0 Hz, J2 = 1.8 Hz, 1H), δ

8.09 (d, J = 8.4 Hz, 1H), δ 7.74-7.76 (m, 5H), δ 7.37-7.49 (m, 7H), δ 7.16-7.29 (m,

3H), δ 5.26 (s, 2H), δ 4.62 (q, J = 7.5 Hz, 2H), 1.49 (t, J = 7.2 Hz, 1H). ESI MS

calculated for C36H30N3O4: 568.2, found 568.3 [M]+.

8-Urea ethidium · Cl (23). In a 15 mL pressure tube, 3-cbz 8-phenoxy-

carbamate ethidium · H2PO4 (45 mg, 67 µmoles) and methanol (8 mL) were

N

HNN

H

H2PO4

O

O

O N

HNN

HNH2

O

O

H2PO4

N

HNH2N

N

NH2HN

ClCl

NH3 / MeOH

80 %

3-cbz 8-phenoxy carbamateethidium phosphate

6M HCl / MeOH

(not characterized)

10%

8- urea ethidium chloride (23)ethidium chloride (12)

O O

NH2

O

++

++

--

- -

233

mixed and brought to -78 °C whereupon approximately 2 mL of ammonia was

added, the pressure tube sealed and allowed to warm to RT. The reaction was

then heated at 80 °C for 1 hr, cooled back to -78 °C, the tube opened and the

ammonia was out-gassed by passing argon into the solution as it slowly warmed

to RT. All volatiles were then removed under reduced pressure. The solid yellow

product was then deprotected with of 10 mL of 1:1 methanol / saturated HCl and

heating to 96 °C for 1 hr. The reaction was then concentrated to a solid under

reduced pressure and purified by reversed-phase chromatography (C-18 silica

gel 60). The column was conditioned with methanol, water, and the crude

product was loaded in 2% methanol / water (0.01M HCl) and a methanol gradient

(2-10% methanol / water (0.01M HCl)) was used to separate ethidium chloride

(elutes first) from the desired product (elutes second), to afford 21 mg of an

orange solid (80%). 1H NMR (300 MHz, d6-DMSO, 20 °C): δ 9.12 (s, 1H), δ 8.76

(d, J = 9.0 Hz, 2H), δ 8.22 (d,d J1 = 9.0 Hz, J2 = 2.1 Hz, 1H), δ 7.71-7.77 (m, 5H),

δ 7.33-7.39 (m, 5H), δ 6.55 (br s, 2H), 5.97 (br s, 2H), δ 4.48 (q, J = 7.2 Hz, 2H),

δ 1.41 (t, J = 6.9 Hz, 3H). ESI MS calculated for C22H21N4O: 357.2, found 357.3

[M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 286

(4.9x104), 464 (4.7x103).

234

E 5.2.7 Synthesis of 3,8-bis Urea Ethidium Chloride (24)

Ethidium Activation for Urea Formation:

3,8-bis Phenoxycarbamate ethidium · H2PO4. Ethidium bromide / 8% water

(200 mg, 466 µmoles), 500 mM sodium phosphate pH 6.6 (5 mL), and acetone (8

mL) were combined and phenyl chloroformate (587 µL, 4.66 mmoles, 5 mequiv,

pre-dissolved in 2.5 mL acetone) was added drop-wise. The reaction preceded

10 min. RT. then cooled to -80 °C, and vacuum filtered to collect the precipitate.

The precipitate was washed with 20% acetone / water (10 mL), 100% acetone

(-80 °C, 10 mL), and dried under reduced pressure to yield 300 mg (98%) of a

yellow solid. 1H NMR (400 MHz, d6-DMSO, 20 °C): δ 11.36 (s, 1H), δ 10.98 (s,

1H), δ 9.17 (d, J = 9.2 Hz, 1H), δ 9.12 (d, J = 8.8 Hz, 1H), δ 8.87 (s, 1H), δ 8.36

(d,d J1 = 9.2 Hz, J2 = 2.0 Hz, 1H), δ 8.22 (d, J = 9.2 Hz, 1H), δ 7.85 (d, J = 2.4

Hz, 1H), δ 7.77 (s, 5H), δ 7.38-7.50 (m, 4H), δ 7.24-7.43 (m, 4H), δ 7.15-7.19 (m,

2H), δ 4.64 (q, J = 7.6 Hz, 2H), 1.48 (t, J = 7.2 Hz, 3H). ESI MS calculated for

C36H30N3O4: 554.2, found 554.3 [M]+.

O

ClON

NH2HN

Br

acetone/ phosphatebuffer pH 6.6 (98%)

ethidium bromide (12)

N

HN

HN

H2PO4

O

O

O

O

3,8-bis phenoxycarbamateethidium phosphate

-+

+-

235

3,8-bis urea ethidium · Cl (23). In a 15 mL pressure tube, 3,8-bis

phenoxycarbamate ethidium · H2PO4 (48 mg, 74 µmoles) and methanol (10 mL)

were mixed and brought to -78 °C whereupon approximately 2 mL of ammonia

was added, the pressure tube sealed and allowed to warm to RT. The reaction

was then heated at 80 °C for 1 hr, cooled back to -78 °C. The tube opened and

the ammonia was out-gassed by passing argon into the solution as it slowly

warmed to RT. All volatiles were then removed under reduced pressure. The

solid product was washed in diethyl ether (2x20 mL) then dissolved in 20%

acetonitrile / water and treated with AG1-X4 (Cl-) exchange resin (1 g, 3.5

mmoles, 47 mequiv) for 5 min. RT. The resin was removed by filtration, and the

solution lyophilized to yield 32 mg (99%) of a yellow solid. 1H NMR (300 MHz, d6-

DMSO, 20 °C): δ 9.76 (s, 1H), δ 9.34 (s, 1H), δ 8.97 (d, J = 9.3 Hz, 1H), δ 8.92

(d, J = 9.3 Hz, 1H), δ 8.85 (d, J = 1.2 Hz, 1H), δ 8.30 (d,d J1 = 9.3 Hz, J2 = 2.4

Hz, 1H), δ 7.93 (d,d J1 = 9.3 Hz, J2 = 1.2 Hz, 1H), δ 7.74-7.78 (m, 5H), δ 7.55 (d,

J = 2.1 Hz, 1H), δ 6.36 (s, 2H), 6.05 (s, 2H), δ 4.58 (q, J = 7.8 Hz, 2H), δ 1.48 (t, J

= 6.9 Hz, 3H). ESI MS calculated for C23H22N5O2: 400.2, found 400.3 [M]+. UV-vis

N

HN

Cl

(99%)3,8-bis urea ethidium chloride (23)

NH2

O

+

-

N

HN

HN

H2PO4

O

O

O

O

3,8-bis phenoxycarbamateethidium phosphate

-+

HNH2N

O

NH3 / MeOH

236

(50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 280 (4.3x104),

434 (3.6x103).

E 5.2.8 Synthesis of the Mono-Pyrrole Ethidium Derivatives 25 & 26:

3-Pyrrole ethidium · TFA (25), and 8- pyrrole ethidium · TFA (26). A mixture

(5:1 respectively) of 8-cbz ethidium · Cl (17) and 3-cbz ethidium · Cl (16) (90 mg,

186 µmoles) was added to glacial acetic acid (4 mL), heated to 120 °C, and 3

additions (15 minutes apart) of dimethoxy-tetrahydrofuran (3x15 µL, 348 µmoles

total, 1.87 mequiv) were added over 30 min and the reaction was kept at 120 °C

for an additional 45 min then cooled to RT, diluted into CHCl3 (100 mL), and

washed with saturated sodium bicarbonate (3x50 mL), brine (50 mL), dried over

sodium sulfate, and concentrated to a solid under reduced pressure. The mixture

of cbz-protected products was purified using a neutral alumina column using pure

acetone as an eluent and concentrated to a yellow solid. This mixture was

carried over, directly to the next step. Deprotection was conducted in a 3 : 1 mix

N

HNH2N

Cl

O

O

N

NH2HN

Cl

O

O

8-cbz ethidium (17)

3-cbz ethidium (16)

O OO

acetic acid, 100 oC

H2 / Pdo

N

NH2N

N

NH2N

TFA

TFA

3-pyrrole ethidium (25)

8-pyrrole ethidium (26)

+

+

+

+

-

--

-

237

of methanol/acetic acid (4 mL), with Pd black (30 mg), and rigorously stirring

under 1 atm of H2 for 3 hr RT. The catalyst was removed by centrifugation, and

the solution concentrated to an orange solid under reduced pressure. The

products separated using a reversed phase C-18 semi-prep column using 38%

acetonitrile / water (0.1% TFA) to yield 14.6 mg (18%) of 3-pyrrole ethidium ·

TFA (25) (Rt = 13.8 min). 1H NMR (400 MHz, D2O, 20 °C): δ 8.43 (d, J = 9.2 Hz,

1H), δ 8.31 (d, J = 9.2 Hz, 1H), δ 7.87 (d, J = 1.6 Hz, 1H), δ 7.70 (d,d J1 = 9.2 Hz,

J2 = 1.6 Hz, 1H), δ 7.60-7.67 (m, 3H), δ 7.42 (d,d J1 = 9.2 Hz, J2 = 2.4 Hz, 1H), δ

7.30-7.32 (m, 2H), δ 7.19 (ab quart, Japp = 2.0 Hz, 2H), δ 6.52 (d, J = 2.4 Hz, 1H),

δ 6.29 (ab quart, Japp = 2.0 Hz, 2H), δ 4.61 (q, J = 7.2 Hz, 2H), δ 1.30 (t, J = 7.0

Hz, 3H). FAB MS calculated for C25H22N3: 364.1814, found 364.1823 [M]+. UV-vis

(50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 223 (2.8x104),

287 (4.7x104), 453 (4.0x103). 8-pyrrole ethidium · TFA (26) (3.3 mg (4%)), (RT

= 18.1 min). 1H NMR (400 MHz, D2O, 20 °C): δ 8.42 (d, J = 9.2 Hz, 1H), δ 8.40

(d, J = 8.8 Hz, 1H), δ 7.86 (d,d J1 = 9.2 Hz, J2 = 2.0 Hz, 1H), δ 7.61-7.71 (m, 3H),

δ 7.36-7.37 (m, 2H), δ 7.29 (d, J = 1.2 Hz, 1H), δ 7.24 (d,d J1 = 8.8 Hz, J2 = 1.2

Hz, 1H), δ 6.99 (d, J = 2.4 Hz, 1H), δ 6.84 (ab quart, Japp = 2.0 Hz, 2H), δ 6.13 (ab

quart, Japp = 2.0 Hz, 2H), δ 4.49 (q, J = 7.6 Hz, 2H), δ 1.30 (t, J = 7.6 Hz, 3H).

FAB MS calculated for C25H22N3: 364.1814, found 364.1822 [M]+. UV-vis (50 mM

sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 223 (2.1x104), 239

(1.4x104), 289 (4.8x104), 466 (4.1x103).

238

E 5.2.9 Synthesis of 3,8-bis Pyrrole Ethidium Acetate (27):

3,8-bis pyrrole ethidium · Ac (27). Ethidium bromide / 4% water (264 mg, 670

µmoles) and glacial acetic acid (10 mL) were brought to 130 °C and 2 additions

(15 minutes apart) of dimethoxy tetrahydrofuran (2x110 µL, 1.65 mmoles total,

2.5 equiv) were added and the reaction was kept under reflux (at 130 °C) for an

additional 1h and cooled to RT. All volatiles were then removed under reduced

pressure, and the solid was dissolved in methanol (~40 mL) and filtered over a

plug of silica gel (~30 mL), the gel was washed with methanol (~60 mL), and the

methanolic fractions combined and concentrated to 280 mg (90%) of a yellow

solid under reduced pressure. The product can be purified further using a

reversed phase C-18 semi-prep column with a 50 – 80% acetonitrile / water

(0.1% TFA) gradient over 20 min. 1H NMR (400 MHz, d6-acetone, 20 °C): δ 9.38

(d, J = 9.0 Hz, 1H), δ 9.32 (d, J = 9.0 Hz, 1H), δ 8.73 (d, J = 2.1 Hz, 1H), δ 8.64

(d,d J1 = 9.0 Hz, J2 = 2.4 Hz, 1H), δ 8.49 (d,d J1 = 9.0 Hz, J2 = 2.1 Hz, 1H), δ

7.91-8.01 (m, 5H), δ 7.72 (ab quart, Japp = 2.1 Hz, 2H), δ 7.54 (d, J = 2.4 Hz, 1H),

δ 7.24 (ab quart, Japp = 2.1 Hz, 2H), δ 6.46 (ab quart, Japp = 2.1 Hz, 2H), δ 6.34

(ab quart, Japp = 2.1 Hz, 2H), δ 5.24 (q, J = 7.2 Hz, 2H), δ 1.58 (t, J = 7.2 Hz, 3H).

FAB MS calculated for C29H24N3: 414.1970, found 414.1951 [M]+. UV-vis (50 mM

N

NH2H2N

BrN

N N

O OOCH3

H3C

ethidium bromide (12) 3,8-bis pyrrole ethidium acetate (27)

acetic acid, 130oC(90%)

OAc ++-

-

239

sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 302 (4.6x104), 428

(5.2x103).

E 5.2.10 Synthesis of Tetramethyl Ethidium Chloride (28):

Tetramethyl ethidium chloride (28). Ethidium bromide / 4 % water (50 mg, 122

µmoles) was combined with trimethyl phosphate (9 mL) and heated, under argon,

to 165 °C for 20 hrs. The trimethyl phosphate was removed under reduced

pressure (80 °C). The purple solid was then dissolved in 10% methanol / CHCl2,

loaded onto a plug of silica gel (15 mL), washed with 10% methanol / CHCl2 (70

mL), eluted using 20% methanol / CHCl2 (50 mL), and concentrated to a solid

under reduced pressure. The product was dissolved into 3 mL of 5% acetonitrile /

water, treated with AG1-X4 (Cl-) exchange resin (0.6 g, 2.1 mmoles, 17 mequiv)

for 5 min at RT, filtered, and lyophilized to yield 48 mg (97%) of a purple solid. 1H

NMR (400 MHz, d6-DMSO, 20 °C): δ 8.85 (d, J = 9.2 Hz, 1H), δ 8.79 (d, J = 9.6

Hz, 1H), δ 7.82 (d,d J1 = 9.6 Hz, J2 = 2.4 Hz, 1H), δ 7.77 (br s, 5H), δ 7.59 (d, J =

8.4 Hz, 1H), δ 6.09 (d, J = 2.4 Hz, 1H), δ 4.70 (q, J = 7.2 Hz, 2H), δ 3.20 (s, 6H),

δ 2.82 (s, 6H), 1.44 (t, J = 7.2 Hz, 3H). FAB MS calculated for C25H28N3:

N

NH2H2N

Br

ethidium bromide (13)

N

NN

tetramethyl ethidium chloride (28)

P

O

OCH3H3CO

H3COCl+ +

- -

240

370.2283, found 370.2279 [M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax

(nm) and ε (cm-1M-1): 224 (2.1x104), 302 (3.7x103), 538 (3.8x104).

E 5.2.11 Synthesis of 3,8-bisurea Pyrrolidine Ethidium Trifluoroacetate (30):

3, 8-bisurea Pyrrolidine ethidium · TFAc (30). 3,8-bis Phenoxycarbamate

ethidium · H2PO4 (12 mg, 19 µmoles) (Section E 5.2.7), DMSO (1 mL) and

pyrrolidine (40 µL, 460 µmoles, 27 equiv) were combined and heated for 5 min

(90 °C). The reaction was then diluted into water (9 mL, 0.01% TFA) and loaded

onto an activated Water’s “Sep-pack” C-18 reversed phase column (activated

with 10 mL acetonitrile, 10 mL water), the column was washed with water (10

mL, 0.01% TFA), then 10% acetonitrile (10 mL, in water with 0.01% TFA). The

product was then eluted with 35% acetonitrile (10 mL, in water with 0.01% TFA)

and lyophilized to yield a yellow solid (12 mg, 100%). 1H NMR (400 MHz, d6-

DMSO, 21 °C): δ 9.01 (d, J = 9.6 Hz, 1H), δ 8.98 (s, 1H), δ 8.94 (d, J = 9.6 Hz,

1H), δ 8.89 (d, J = 2.0 Hz, 1H), δ 8.72 (s, 1H), δ 8.40 (d,d J1 = 9.2 Hz, J2 = 2.4 Hz,

1H), δ 8.26 (d,d J1 = 9.2 Hz, J2 = 1.6 Hz, 1H), δ 7.34-7.80 (m, 6H), δ 4.56 (q, J =

6.4 Hz, 2H), δ 3.46 (t, J = 6.4 Hz, 2H), δ 3.15 (t, J = 6.6 Hz, 2H), δ 1.91 (t, J = 6.4

N

HN

HN

H2PO4

O

O

O

O3,8-bis phenoxycarbamateethidium phosphate

N

HN

HN N

OO

3,8-bisurea pyrrolidine ethidiumtrifluoroacetate (30)

HN

+ +

-

NDMSO, 90 oC

(100%)

TFAc-

241

Hz, 2H), δ 2.81 (t, J = 6.6 Hz, 2H), δ 1.48 (t, J = 7.2 Hz, 3H). ESI MS calculated

for C31H34N5O2: 508, found 508 [M]+. UV-vis (50 mM sodium phosphate pH 7.5):

λmax (nm) and ε (cm-1M-1): 288 (4.7x104), 438 (4.5x104).

E 5.2.12 Synthesis of 3,8-bisurea Ethylene Diamine Trifluoroacetate (31):

3,8-bisurea Ethylene diamine ethidium · TFA3 (31). 3,8-bis phenoxycarbamate

ethidium · H2PO4 (9 mg, 13.8 µmoles) (Section E 5.2.7), DMSO (300 µL), and

ethylene diamine (100 µL) were heated at 85 °C for 30 min then cooled to RT.

500 mM sodium phosphate pH 6.5 (0.6 mL) was added and the reaction was

diluted into water (5 mL, 0.1% TFA) and loaded onto an activated Water’s “Sep-

pack” C-18 reversed phase column (activated with 10 mL acetonitrile, 10 mL

water). The column was washed with water / 0.1% TFA (5 mL), the product

eluted with 25% acetonitrile/water (0.1% TFA) and lyophilized to yield 10.5 mg

(91%) of a yellow solid. 1H NMR (300 MHz, D2O, 20 °C): δ 8.58 (d, J = 8.4 Hz,

1H), δ 8.50 (d, J = 8.4 Hz, 1H), δ 8.43 (s, 1H), δ 7.76 (d, J = 9.0 Hz, 1H), δ 7.60-

7.69 (m, 4H), δ 7.50 (s, 1H), δ 7.40-7.43 (m, 2H), δ 4.60 (q, J = 7.5 Hz, 2H),

δ 3.40 (t, J = 6.0 Hz, 2H), δ 3.26 (t, J = 5.7 Hz, 2H), δ 3.03 (t, J = 5.7 Hz, 2H),

N

HN

HN

H2PO4

O

O

O

O3,8-bis phenoxycarbamateethidium phosphate

N

HN

HN

HN

OO

3,8-bisurea ethylene diamineethidium chloride (31)

++

-

HN

DMSO, 85 oC(91%) NH2H2N

3 TFAH2N

NH2

242

δ 2.92 (t, J = 5.7 Hz, 2H), δ 1.37 (t, J = 7.2 Hz, 3H). ESI MS calculated for

C27H32N7O2: 486, found 486 [M]+. UV-vis (50 mM sodium phosphate pH 7.5):

λmax (nm) and ε (cm-1M-1): 216 (3.8x104), 286 (6.6x103), 444 (6.5x104).

E 5.2.13 Synthesis of 3,8-bisurea Arginine Ethidium Trifluoroacetate (32):

3,8-bisurea arginine ethidium · TFA3 (32). 3,8-bis phenoxycarbamate ethidium

· H2PO4 (10 mg, 13.8 µmoles), DMSO (400 µL), water (100 µL), L-Arg · HCl (50

mg, 237 µmoles, 8.6 mequiv) and 2,4,6 collidine (84 µL, 711 µmoles, 26 mequiv)

were heated for at 90 °C for 1hr then cooled to RT and quenched 500 mM

sodium phosphate pH 6.5 (0.6 mL). The reaction was then diluted into 5 mL

water (0.1 % TFA) and loaded onto an activated Water’s “Sep-pack” C-18

reversed phase column (activated with 10 mL acetonitrile, 10 mL water). The

column was washed with water / 0.1% TFA (5 mL), the product eluted with 25%

acetonitrile/water (0.1% TFA) and lyophilized. The product was further purified

using a reversed phase C-18 semi-prep HPLC column using 15% acetonitrile /

water (0.1% TFA) (RT = 6.3 min) to yield 4.5 mg (31%) of a yellow solid. 1H NMR

(400 MHz, D2O, 20 °C): δ 8.37 (d, J = 8.8 Hz, 1H), δ 8.31 (d, J = 8.8 Hz, 1H), δ

8.30 (d, J = 8.8 Hz, 1H), δ 7.67-7.76 (m, 4H), δ 7.53 (d, J = 9.2 Hz, 1H), δ 7.39-

NHN

HN

H2PO4

O

O

O

O

NHN

HN

HN

O

HN

O CO2HHO2CNH

3- 8- bis phenoxycarbamateethidium phosphate

NH

NH

H2N

NH

NH2

3,8-bisurea arginine ethidiumthrifluoroacetate (32)

3 TFA-+ +

H2N

O2CNH

NH2

NH2-

DMSO, collidine,90 oC (31%)

243

7.45 (m, 3H), δ 4.59 (q, J = 6.8 Hz, 2H), δ 4.20 (t, J = 5.6 Hz, 1H), δ 4.08 (t, J =

5.2 Hz, 1H), δ 3.15 (t, J = 6.6 Hz, 2H), δ 3.10 (t, J = 6.8 Hz, 2H), δ 1.52-1.83 (m,

8H), δ 1.33 (t, J = 7.0 Hz, 3H). ESI MS calculated for C35H44N11O6: 714, found

715 [M+H]+, UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-1

M-1): 216 (3.8x104), 288 (6.8x104), 444 (5.7x103).

E 5.2.14 Synthesis of 3,8-bisurea 2-DOS Ethidium Trifluoroacetate (33):

3,8-bisurea 2-DOS ethidium · TFA3 (33). 3,8-bis phenoxycarbamate ethidium ·

H2PO4 (20 mg, 31 µmoles), DMSO (1.5 mL), phenol (1.5 g), Na2CO3 (50 mg, 472

µmoles, 15 equiv), 2-deoxy streptamine · 2HCl (110 mg, 277 µmoles, 8.9 equiv)

pre-dissolved in water (0.7 mL), were heated at 85 °C for 45 min. The reaction

was diluted into water (80 mL) and washed with CH2Cl2 (2x40 mL), CHCl3 (2x40

mL), and ethyl acetate (40 mL). The aqueous phase was then concentrated to a

solid and purified by reversed-phase chromatography (C-18 silica gel 60). The

column was conditioned with pure acetonitirile, pure water, and the crude product

was loaded in water (0.01 % TFA) and an acetonitrile gradient (0-8% acetonitrile

/ water (0.01% TFA) was used to elute the product. Fractions were collected and

lyophilized to yield 8 mg of a yellow solid (25%). 1H NMR (400 MHz, D2O, 20

N

HN

HN

H2PO4

O

O

O

O

N

HN

HN

HN

O

HN

O

H2N NH2

3,8-bis phenoxycarbamateethidium phosphate

OH OHOH

HOHOHO

3,8-bisurea 2-DOS ethidiumtrifluoroacetate (33)

3TFA-

+H2N NH2

OHOH

HO

DMSO, phenol,Na2CO3, 85 oC(25%)

+

244

°C): δ 8.55 (d, J = 8.4 Hz, 1H), δ 8.47 (d, J = 8.4 Hz, 1H), δ 8.44 (s, 1H), δ 7.62-

7.74 (m, 6H), δ 7.50 (s, 1H), δ 7.39 (s, 1H), δ 7.38 (d, J = 8.4 Hz, 1H), δ 4.59 (q, J

= 6.4 Hz, 2H), δ 3.70 (m, 1H), δ 3.52 (m, 1H), δ 3.08-3.38 (m, 8H), δ 2.21 (t,d J1 =

12.4 Hz, J2 = 4.0 Hz, 1H), δ 2.08 (t,d J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ 1.56 (q, J =

12.4 Hz, 1H), δ 1.44 (q, J = 12.4 Hz, 1H), δ 1.37 (t, J = 6.4 Hz, 3H). ESI MS

calculated for C35H44N7O8: 690, found 690 [M]+, UV-vis (50 mM sodium

phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 216 (3.8x104), 288 (6.8x104), 444

(5.7x103).

E 5.2.15 Synthesis of 8-Urea-Ethidium-6'-3'Deoxyneamine Trifluoroacetate(36):

Tobramycin has 5 amines that could, potentially react with the activated ethidium

derivative, resulting in multiple isomers. Indeed, at least other 2 isomers are

observed from the above reaction. The major product, however, is the only

isomer to have in its 1H NMR spectrum (d6-DMSO) a triplet for the splitting of the

highlighted N-H (above). The other isomers have a doublet for this hydrogen.

The non-hydrolyzed tobramycin-containing conjugates have also been prepared

using this method. To avoid hydrolysis, cbz is removed by using H2 / Pd black for

N

NH

HNO

O

O

O

3-cbz 8-phenoxycarbamateethidium phosphate

H2PO4 +- Tobramycin-TFA5 (3)

DMSO, Na2CO3

6M HCl/MeOH

20% (2steps)

N

NH

H2N

O+

H2NNH2

OHOH

O

O

NH2HO

NH

4 TFA

8-urea-ethidium 6'-3'deoxyneaminetrifluoroacetate (36)

245

a short (2.5 h) deprotection that results in some partial deprotection and some

degradation the ethidium, but can yield significant quantities of desired product(s)

(not shown).

8-urea-ethidium-6'-3'Deoxyneamine Trifluoroacetate (36). 3-cbz 8-phenoxy-

carbamate ethidium · H2PO4 (20 mg, 30 µmoles), tobramycin·TFA5 (160 mg, 154

µmoles), DMSO (2mL), and Na2CO3 (80 mg, 750 µmoles) were combined and

heated at 90 °C for 15 min. The reaction was then filtered, diluted into water (9

mL, 0.01% TFA), and loaded onto an activated Water’s “Sep-pack” C-18

reversed phase column (activated with 10 mL acetonitrile, 10 mL water), the

column was washed with water (10 mL), and the product was then eluted with

25% acetonitrile (10 mL, in water with 0.01% TFA) and lyophilized to yield a

yellow solid (40 mg) that was carried over directly to the next step. Deprotection

of cbz and the partial hydrolysis of the tobramycin-ethidium conjugate was

accomplished by refluxing the product in 6M HCl / MeOH for 1 hr (110 °C). The

reaction was then concentrated to a solid and purified using a reversed phase C-

18 semi-prep HPLC column with 13% acetonitrile/water (0.1% TFA) at 3.5 mL /

min (Rt = 18.6 min, the third and last major peak) to yield 7.4 mg (20%, 2 steps)

of an orange solid. 1H NMR (400 MHz, d6-DMSO, 20 °C): δ 9.35 (s, 1H), δ 8.74-

8.78 (m, 2H), δ 7.50 (br s, 6H), δ 8.07 (d,d J1 = 12.4 Hz, J2 = 2.8 Hz, 1H), δ 7.85

(br s, 3H), δ 7.71-7.77 (m, 5H), δ 7.52 (d, J = 2.8 Hz, 1H), δ 7.40 (s, 1H), δ 7.38

(d, J = 12.8 Hz, 1H), δ 6.56 (br s, 2H), δ 6.46 (br t, J = 7.0 Hz, 1H), δ 6.05 (br s,

1H), δ 5.76 (br s, 1H), δ 5.45 (d, J = 2.8 Hz, 1H), δ 5.32 (br s, 1H), δ 4.49 (q, J =

246

8.8 Hz, 2H), δ 3.07-3.66 (m, 10H), δ 2.27 (t,d J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ

2.01 (t,d J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ 1.81 (q, J = 12.4 Hz, 1H), δ 1.62 (q, J =

12.4 Hz, 1H), δ 1.41 (t, J = 9.2 Hz, 3H). ESI MS calculated for C34H44N7O6: 646,

found 646 [M]+. λmax (nm) and ε (cm-1M-1): 216 (3.2x104), 288 (5.0x104), 464

(4.5x103).

E 5.2.16 Synthesis of 3-urea-Ethidium-6'-3'Deoxyneamine Trifluoroacetate(37):

Tobramycin has 5 amines that could, potentially react with the activated ethidium,

resulting in multiple isomers. Indeed, at least 3 isomers are observed for this

reaction. The major product, however, is the only isomer to have in its 1H NMR

spectrum (d6-DMSO) a triplet for the splitting of the highlighted N-H (above). The

other isomers are found to have a doublet for this hydrogen.

3-urea-ethidium-6'-3'Deoxyneamine trifluoroacetate (37). 8-cbz 3-phenoxy-

carbamate ethidium · H2PO4 (20 mg, 30 µmoles), tobramycin·TFA5 (160 mg, 154

µmoles), DMSO (2mL), and Na2CO3 (80 mg, 750 µmoles) were combined and

heated at 80 °C for 15 min. The reaction was then filtered, diluted into water (9

mL, 0.01% TFA), and loaded onto an activated Water’s “Sep-pack” C-18

N

HN

HN

8-cbz 3-phenoxycarbamateethidium phosphate

H2PO4 +

- Tobramycin-TFA5 (3)

DMSO, Na2CO3

6M HCl/MeOH

20% (2 steps)

3-urea-ethidium-6'-3'deoxyneaminetrifluoroacetate (37)

O

NH2H2NHO

HOO

O

H2N OH

NH

N

NH2HN

+4 TFA

O

O

O

O

247

reversed phase column (activated with 10 mL acetonitrile, 10 mL water), the

column was washed with water (10 mL), and the product was eluted using 25%

acetonitrile (10 mL, in water with 0.01% TFA) and lyophilized to yield a yellow

solid (35 mg) that was carried over directly to the next step. Deprotection of cbz

and the partial hydrolysis of the tobramycin-ethidium conjugate was

accomplished by refluxing the product in 6M HCl / MeOH for 1 hr (105 °C). The

reaction was then concentrated to a solid and purified using a reversed phase C-

18 semi-prep HPLC column with 13% acetonitrile / water (0.1% TFA) at 3.5 mL /

min (RT = 17.5 min, the fourth and last major peak) to yield 6.7 mg (18%, 2 steps)

of an orange solid. 1H NMR (300 MHz, d6-DMSO, 20 °C): δ 9.90 (s, 1H), δ 8.85

(d, J = 9.3 Hz, 1H), δ 8.81 (s, 1H), δ 8.73 (d, J = 9.3 Hz, 1H), δ 8.25 (br s, 3H), δ

8.15 (br s, 3H), δ 7.91 (br s, 3H), δ 7.73-7.78 (m, 6H), δ 7.57 (d,d J1 = 9.0 Hz, J2 =

2.4 Hz, 1H), δ 6.87 (br t, J = 4.5 Hz, 1H), δ 6.36 (d, J = 2.4 Hz, 1H), δ 6.06 (br s,

1H), δ 5.80 (br s, 1H), δ 5.52 (d, J = 3.0 Hz, 1H), δ 5.40 (br s, 1H), δ 4.53 (q, J =

6.6 Hz, 2H), δ 3.28-3.78 (m, 10H), δ 2.28 (t,d J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ

2.03 (t,d J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ 1.86 (q, J = 12.4 Hz, 1H), δ 1.63 (q, J =

12.4 Hz, 1H), δ 1.43 (t, J = 6.9 Hz, 3H). ESI MS calculated for C34H44N7O6: 646,

found 646 [M]+. UV-vis (50 mM sodium phosphate pH 7.5): λmax (nm) and ε (cm-

1M-1): 216 (3.0x104), 288 (6.0x104), 458 (5.2x103).

248

E 6.1.1 Synthesis of Neo-N-acridine (38):

5''-amino-5''-deoxy-Boc6-Neomycin B. Synthesized as described (Hai Wang

Ph. D. Thesis, University of California, San Diego, 1998).

9-phenoxyacridine. Prepared according to Dupre and Robinson (J. Chem. Soc.

1945, 549-551).

5''-aminoacridine-5''-deoxy-Boc6-Neomycin B. 5''-amino-5''-deoxy-Boc6-

Neomycin B (12.7 mg, 10.4 µmoles) was combined with phenol (200 mg), DMSO

(350 µL), and 9-phenoxy acridine (8.2 mg, 30 µmoles, 2.9 equiv.) and heated,

under argon, at 70 °C for 2 h. The reaction was loaded directly onto a silica

column and purified using flash chromatography (5-10% MeOH / CH2Cl2) to

afford 6.7 mg of a yellow product (46%). Rf = .28 (10% MeOH / CH2Cl2). This

product was carried over, without further characterization to the next step.

Neo-N-acridine · TFA6 (38). 5''-amino-5''-deoxy-Boc6-Neomycin B (6.7 mg, 4.8

µmoles) was dissolved in a “deprotection cocktail” (4.5 mL trifluoroacetic acid,

HN

N

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2NH2N

O

NHBoc

BocHNO

HOHO

OOBocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

N

O NH

O

NHBoc

BocHNO

HOHO

OOBocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

N

Neo-N-acridine (38)5''-amino-5''-deoxy-Boc6 Neomycin B

Phenol, 65 oC

(46%)

TFA, phenol,TIPS, H2O (99%)

5''-aminoacridine-5''-deoxyBoc6 Neomycin B

249

0.215 mL phenol, 0.215 mL water, and 87 µL of triisopropyl silane) and mixed for

30 min at RT. The reaction was then diluted into 2% acetic acid/water (30 mL)

and washed with diethyl ether (4x15 mL). The aqueous phase was then

concentrated to a solid under reduced pressure. The crude product was HPLC

purified on C-18 semiprep column using isocratic conditions 11% acetonitrile in

water and 0.1% TFA (Rt = 7.6 min) and lyophilized to yield 7.1 mg of a yellow

solid (99% assuming 6 TFA counter ions). 1H NMR (300 MHz, D2O, 25 °C): δ

8.36 (d, J = 8.7 Hz, 2H), δ 7.86 (t, J = 7.8 Hz, 2H), δ 7.73 (d, J = 8.4 Hz, 2H), δ

7.46 (t, J = 7.8 Hz, 2H), δ 5.76 (d, J = 3.9 Hz, 1H), δ 5.29 (d, J = 0.9 Hz, 1H), δ

5.14 (d, J = 1.2 Hz, 1H), δ 4.48-4.55 (m, 2H), δ 4.37 (d,d J1 = 4.2 Hz, J2 = 1.5 Hz,

1H), δ 4.25 (d,d J1 = 13.8 Hz, J2 = 8.1 Hz, 1H), δ 4.1 (t, J = 4.2 Hz, 1H), δ 4.03 (t,

J = 3.0 Hz, 1H), δ 3.74-3.85 (m, 2H), δ 3.65 (d, J = 3.0 Hz, 1H), δ 3.51-3.59 (m,

1H), δ 3.40-3.47 (m, 3H), δ 3.27-3.35 (m, 1H), δ 3.10-3.20 (m, 3H), δ 2.94 (d,d J1

= 13.5 Hz, J2 = 3.6 Hz, 1H), δ 2.18-2.33 (m, 3H), δ 1.95 (d,d J1 = 10.2 Hz, J2 =

3.9 Hz, 1H), δ 1.66 (q, J = 12.6 Hz, 1H). MALDI TOF MS calculated for

C36H54N8O12: 790.4, found 813.3 [M+Na]+, found 829.3 [M+K]+. UV-vis (50 mM

sodium phosphate pH 7.5): λmax (nm) and ε (cm-1M-1): 222 (2.0x104), 266

(5.6x104), 412 (10.3x103), 434 (8.3x103).

250

E 6.1.2 Synthesis of Neo-C-acridine (40):

5'' TIPS Boc6 Neomycin B. Synthesized as described (Hai Wang Ph. D. Thesis,

University of California, San Diego, 1998).

5'' Ethanedithiol Boc6 Neomycin. 5'' TIPS Boc6 Neomycin B (38 mg, 25.6

µmoles), cesium carbonate (176 mg, 540 µmoles, 20 equiv), DMF (2 mL) and 1,

2 ethane dithiol (230 µL, 2.7 mmoles, 107 equiv), were stirred under argon for 4 h

at RT, and diluted into ethyl acetate (180 mL) and washed with 1M NaH2PO4

(4x80 mL), brine (80 mL) and dried for 24 h. under reduced pressure. The crude

product was purified using standard flash chromatography (4-5% MeOH /

CH2Cl2) to yield 22 mg of a white solid (67%). Rf = 0.5 (10% MeOH / CH2Cl2) (the

same Rf as the starting material). 1H-NMR (400 MHz, MeOD) δ 5.41 (s, 1H), δ

S NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

NH

N

Neo-C-acridine (40)

S

O

O

NHBoc

BocHNO

HOHO

OO

BocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

S

O

O

HS SH

DMF, Cs2CO3 (67%)

S

O

NHBoc

BocHNO

HOHO

OO

BocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

HS

H2NBr

EtOH, NaOEtN

Cl

Phenol, NaH

O

NHBoc

BocHNO

HOHO

OO

BocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

SNH

NS

TFA cocktail

(22%, 3 steps)

5'' TIPS Boc6 Neomycin B 5'' Ethanedithiol Boc6 Neomycin B

Neo-C-acridine Boc6

251

5.13 (s, 1H), δ 4.92 (s, 1H), δ 4.21-4.24 (m, 2H), δ 4.08 (q, J = 5.2 Hz, 1H), δ

3.85-3.89 (m, 2H), δ 3.70-3.76 (m, 2H), δ 3.17-3.57 (m, 12H), δ 2.86-2.95 (m,

5H), δ 2.72 (t, J = 7.2 Hz, 2H), δ 1.92-1.98 (m, 1H), δ 1.28-1.46 (m, 56H).

Neo-C-acridine · TFA6 (40). Under argon, sodium ethoxide (14 mg, 206 µmoles)

argon-sparged ethanol (2 mL), and 1-bromo 2-ammonium ethane chloride (14

mg, 68 µmoles) were stirred for 5 min at RT. To this solution, 5'' ethanedithiol

Boc6 neomycin B (5 mg, 3.8 mmoles) (pre-dissolved in 1 mL of argon-sparged

ethanol) was added and allowed to react for 2.5 hr at RT. The reaction was then

diluted into ethyl acetate (150 mL) and washed with 0.1M citric acid (4x40 mL),

0.1M Na2HPO4 (50 mL), and brine (50 mL), dried over sodium sulfate and

concentrated to a solid under reduced pressure. All of this crude product (Rf =

0.1 (10 %MeOH / CH2Cl2)) was taken, directly, into the next step. 9-phenoxy

acridine was generated in situ by melting phenol (1 g), under argon, at 70 °C and

adding sodium hydride (10 mg, 60% dispersion in wax, 250 µmoles), followed by

9-chloroacridine (20 mg, 47 µmoles) and mixing, under argon for 5 min. The

neomycin-containing starting material (pre-dissolved in 1 mL of DMSO) was then

added and the reaction was kept at 70 °C under argon for 14 hr, diluted into ethyl

acetate (150 mL) and washed with 1M Na2CO3 (4x40 mL), and brine (50 mL),

then dried over sodium sulfate and concentrated to an oil under reduced

pressure. The crude product was then purified on silica gel, using standard flash

chromatography (0-9% MeOH / CH2Cl2) to afford a yellow solid Rf = 0.2 (10%

MeOH / CH2Cl2). All of this product was then taken into the next step without

252

further characterization. Deprotection was conducted by dissolving the product in

a deprotection cocktail (5 mL of trifluoroacetic acid, containing 1% ea (v/v) of

triisopropyl silane, and 1,2-ethanedithiol) and stirring, RT, for 20 min. The

reaction was then diluted into water (100 mL) and washed with CHCl3 (2x50 mL),

diethyl ether (2x50 mL), and concentrated to a yellow solid under reduced

pressure. The crude product was HPLC purified on C-18 semiprep column using

isocratic conditions (15% acetonitrile in water and 0.1% TFA) (Rt = 7.4 min) and

lyophilized to yield 1.3 mg of a yellow solid (22% three steps). 1H NMR (400

MHz, D2O, 25 °C): δ 8.22 (d, J = 8.8 Hz, 2H), δ 7.78 (t, J = 7.6 Hz, 2H), δ 7.63

(d, J = 8.8 Hz, 2H), δ 7.39 (t, J = 7.2 Hz, 2H), δ 5.86 (d, J = 4.0 Hz, 1H), δ 5.22 (s,

1H), δ 5.08 (s, 1H), δ 4.23 (t, J = 6.0 Hz, 1H), δ 4.18 (s, 1H), δ 4.12-4.16 (m, 1H),

δ 4.07 (t, J = 4.8 Hz, 1H), δ 4.03 (t, J = 2.4 Hz, 1H), δ 3.91 (t, J = 10 Hz, 1H), δ

3.82 (t, J = 9.6 Hz, 1H), δ 3.73 (t, J = 8.6 Hz, 2H), δ 3.62 (s, 1H), δ 3.52 (t, J = 10

Hz, 1H), δ 3.16-3.40 (m, 8H), δ 3.04 (d,d J1 = 14 Hz, J2 = 6.8 Hz, 1H), δ 2.98 (t, J

= 6.4 Hz, 2H), δ 2.90 (d,d J1 = 14 Hz, J2 = 4.4 Hz, 1H), δ 2.61-2.67 (m, 6H), δ

2.31 (d t, J1 = 12.4 Hz, J2 = 4.4 Hz, 1H), δ 1.73 (q, J = 12 Hz, 1H). MALDI TOF

MS calculated for C40H62N8O12S2: 910.4, found 911.3 [M+H]+, found 933.3

[M+Na]+, found 949.3 [M+K]+. UV-vis (50 mM sodium phoshate pH 7.5): λmax

(nm) and ε (cm-1M-1): 222 (1.95x104), 266 (5.0x104), 412 (9.85x103), 434

(8.3x103).

253

E 6.2.1 Synthesis and Characterization of Tobra-N-acridine (41):

6"-amino 6"-deoxy Boc5 Tobramycin. Synthesized as described (Hai Wang

Ph. D. Thesis, University of California, San Diego, 1998).

Boc5 Tobra-N-acridine. 6"-amino 6"-deoxy Boc5 Tobramycin (38 mg, 40

µmoles), 9-phenoxyacridine (50 mg, 190 µmoles), phenol (0.5 g), and DMSO (1

mL), were heated under Ar for 2 hr at 70 °C. Volatiles were then removed, under

vacuum, for 14 hr. The crude product was then purified on ~100 mL of silca gel

using standard flash chromatography and a 8 – 13% methanol/CH2Cl2 gradient,

yielding a yellow solid (41 mg, 91%). Rf = .23 (10% MeOH / CH2Cl2).1H NMR

(300 MHz, d6-MeOD, 25 °C): δ 8.55 (d, J = 8.7 Hz, 2H), δ 7.98 (t, J = 7.5 Hz,

2H), δ 7.81 (d, J = 7.8 Hz, 2H), δ 7.63 (t, J = 7.5 Hz, 2H), δ 5.13 (d, J = 3.3 Hz,

1H), δ 4.94 (d, J = 3.3 Hz, 1H), δ 4.60-4.67 (m, 3H), δ 4.09-4.17 (m, 1H), δ 3.30-

3.78 (m, 18H), δ 1.85-1.98 (m, 2H), δ 1.15-1.62 (m, 47H). MALDI FTMS

calculated for C56H85N7O8: 1143.60, found 1144.60 [M+H]+.

O

NH2

HO

H2N

H2NO

HOHOO

O

H2N OH

NH2

NH

Tobra-N-acridine (41)

N

O

NH-Boc

HO

Boc-HN

Boc-HNO

HOHOO

O

Boc-HN OH

NH-Boc

NH

Boc5 Tobra-N-acridine

N

O

NH-Boc

HO

Boc-HN

Boc-HNO

HOHOO

O

Boc-HN OH

NH-Boc

NH2

6''-amino 6''-deoxy Boc5 Tobramycin

9-phenoxyacridine,phenol, DMSO, 70 oC

(91%)

TFA, TIPS,water, phenol

(86%)

254

Tobra-N-acridine · TFA5 (41). Boc5 Tobra-N-acridine (11 mg, 9.6 µmoles) was

dissolved in a “deprotection cocktail” (9 mL trifluoroacetic acid, 0.4 mL phenol,

0.4 mL water, and 300 µL of triisopropyl silane) and mixed for 30 min at RT. The

reaction was then diluted into 2% acetic acid/water (30 mL) and washed with

diethyl ether (4x15 mL). The aqueous phase was concentrated to a solid under

reduced pressure and HPLC purified on C-18 semiprep column under isocratic

conditions (11% acetonitrile in water and 0.1% TFA) (Rt = 13.3 min) and

lyophilized to yield 10 mg of a yellow solid (86% assuming 5 TFA counter ions).

1H NMR (400 MHz, D2O, 25 °C): δ 8.29 (d, J = 8.8 Hz, 2H), δ 7.85 (t, J = 7.8 Hz,

2H), δ 7.70 (d, J = 8.8 Hz, 2H), δ 7.43 (t, J = 7.8 Hz, 2H), δ 5.47 (d, J = 3.6 Hz,

1H), δ 5.09 (d, J = 3.6 Hz, 1H), δ 4.46-4.50 (m, 1H), δ 4.24-4.28 (m, 2H), δ 3.94

(d,d J1 = 10.4 Hz, J2 = 3.2 Hz, 1H), δ 3.87 (t, J = 9.4 Hz, 1H), δ 3.56-3.75 (m,

4H), δ 3.36-3.94 (m, 3H), δ 3.30 (d,d J1 = 13.6 Hz, J2 = 4.0 Hz, 1H), δ 3.22 (d,t J1

= 12.0 Hz, J2 = 2.0 Hz, 1H), δ 3.09 (d,d J1 = 13.6 Hz, J2 = 7.6 Hz, 1H), ), δ 2.41

(d,t J1 = 12.8 Hz, J2 = 4.4 Hz, 1H), δ 2.08 (d,t J1 = 13.2 Hz, J2 = 4.8 Hz, 1H), δ

1.83 (q, J = 12.8 Hz, 1H), δ 1.81 (q, J = 12.8 Hz, 1H). MALDI FTMS calculated

for C31H45N7O8: 643.34, found 644.34 [M+H]+. (50 mM Tris·HCl pH 7.5): λmax

(nm) and ε (cm-1M-1): 222 (1.7x104), 266 (5.0x104), 412 (8.9x103), 434 (7.4x103).

255

E 6.3.1 Synthesis and Characterization of Neo-Neo (43):

Neo-neo was first synthesized by Hai Wang.180 It has more recently been

synthesized using a simplified method:

5'' ββββ-Mercaptoethyl Ether Neomycin · TFA6: 5'' TIPS Boc6 Neomycin B (40 mg,

27 µmoles) was dissolved in DMF (1.5 mL) and treated with Cs2CO3 (100 mg,

307 µmoles), and 2-mercaptoethyl ether (125 µL, 1 mmoles, 37 equiv). The

reaction was kept under argon for 7h at 30 °C, diluted into ethyl acetate (150 mL),

washed with 0.1 M citric acid (50 mL), water (3x50 mL), brine (50 mL) and dried

over sodium sulfate. The organic layer was then concentrated under reduced

pressure, and kept under a high vacuum over night. The crude product was

SO

H2N

NH2O

OO

NH2

HO

OH O

HO O

OH

H2N

OOH

HO

H2N

H2N

SO

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

S S

Neo-Neo (43)

SO

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

HS

O

O

NHBoc

BocHNO

HOHO

OO

BocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

S

O

O

5'' TIPS Boc6 Neomycin B5'' β-Mercaptoethyl EtherNeomycin B

1) β-mercaptoethyl ether,DMF, Cs2CO3

2) TFA cocktail (79%, 2 steps)

I2 / MeOH

(77%)

256

dissolved in CH2Cl2 (4 mL), and treated with 1,2 ethanedithiol (20 µL), and

triisopropy silane (20 µL). and trifluoroacetic acid (5 mL) for 15 min. at RT. The

reaction was diluted into toluene (50 mL) and concentrated under vacuum at

50 °C (2x) and kept under high-vacuum for 6 h, the solid was then dissolved in

0.1% TFA in water (3 mL), filtered through glass wool and lyophilized to afford 30

mg of a white solid (79% yield, two steps). 1H-NMR (400 MHz, D2O) δ 5.88 (d, J

= 4.0 Hz, 1H), δ 5.23 (d, J = 3.2 Hz, 1H), δ 5.12 (s, 1H), δ 4.21-4.24 (m, 2H), δ

4.13-4.18 (m, 2H), δ 4.04 (s, 1H), δ 3.91 (t, J = 10 Hz, 1H), δ 3.82 (t, J = 9.6 Hz,

1H), δ 3.71-3.75 (m, 2H), δ 3.64 (s, 1H), δ 3.48-3.57 (m, 4H), δ 3.17-3.41 (m, 8H),

δ 3.11 (d,d J1 = 13.6 Hz, J2 = 9.6 Hz, 1H), δ 2.97 (d,d J1 = 13.2 Hz, J2 = 3.6 Hz,

1H), δ 2.65-2.74 (m, 4H), δ 2.55 (t, J = 6.0 Hz, 2H), δ 2.30 (d, t J1 = 12.4 Hz, J2 =

4.4 Hz, 1H), δ 1.71 (q, J = 12.4 Hz, 1H). ESI MS calculated for C27H54N6O13S2:

734.3, found 735.3 [M+H]+.

Neo-Neo · TFA12 (43). 5'' β-Mercaptoethyl Ether Neomycin · TFA6 (9 mg, 6.4

µmoles) was dissolved in methanol (3 mL) and, while stirring, a dilute solution of

I2 (in methanol) was added drop-wise until two drops past the end-point (where

the reaction no longer decolorized the I2 solution) and stirred and additional 15

min at RT. The methanol was then removed under reduced pressure and the

solid was washed with acetonitrile (0.1 % TFA) (3x2 mL) to remove the remaining

traces of I2. The white solid was then lyophilized from water (0.1% TFA) to afford

6.9 mg of a white solid (77%). 1H-NMR (300 MHz, D2O) δ 5.90 (d, J = 4.0 Hz,

2H), δ 5.25 (d, J = 3.2 Hz, 2H), δ 5.15 (s, 2H), δ 4.25-4.29 (m, 4H), δ 4.16-4.17

257

(m, 4H), δ 4.06 (t, J = 2.8 Hz, 2H), δ 3.95 (t, J = 9.8 Hz, 2H), δ 3.85 (t, J = 9.9 Hz,

2H), δ 3.73-3.79 (m, 4H), δ 3.52-3.68 (m, 10H), δ 3.20-3.43 (m, 16H), δ 3.14 (d,d

J1 = 13.5 Hz, J2 = 6.9 Hz, 2H), δ 2.97 (d,d J1 = 13.5 Hz, J2 = 3.6 Hz, 2H), δ 2.68-

2.79 (m, 12H), δ 2.32 (d,t J1 = 12.9 Hz, J2 = 4.2 Hz, 2H), δ 1.73 (q, J = 12.6 Hz,

2H). MALDI TOF MS calculated for C54H106N12O26S4: 1466.6, observed: 1489.7

[M+Na]+.

E 6.3.2 Synthesis and Characterization of Neo-N-Neo (44):

Boc6 Neo-N-Neo. 5'' Ethanedithiol Boc6 Neomycin (Section E 6.1.2) (9 mg, 7

µmoles) was dissolved in 1:1 CHCl3/methanol (4 mL) and triethylamine (10 µL,

70 µmoles, 10 equiv). A dilute I2 solution (~3mg/mL in CHCl3) was titrated until

two drops past the end-point (where the reaction no longer decolorized the I2

SS

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

H2N

SS

H2N

NH2O

OO

NH2

HO

OH O

HO O

OH

H2N

OOH

HO

H2N

H2N

Neo-N-Neo (44)

S

O

NHBoc

BocHNO

HOHO

OOBocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

HS

5'' Ethanedithiol Boc6 Neomycin B

S

O

NHBoc

BocHNO

HOHO

OO

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

SS

O

BocHN

NHBocO

OHHO

OO

NHBoc

HO

BocHN

OH O

HO O

OH

BocHN

S

Boc6 Neo-N-Neo

NHBoc

NHBoc

I2 / MeOH

(99%)

TFA cocktail

(99%)

258

solution) and stirred an additional 45 min at RT. The reaction was then diluted

into CHCl3 (40 mL) and washed with 1M NaH2PO4 (2x15 mL), 1M Na2CO3, (2x15

mL), brine (20 mL) dried over sodium sulfate, and concentrated to a white solid

under reduced pressure (9 mg, 99%). Rf = 0.45 (10 % MeOH / CH2Cl2).1H-NMR

(300 MHz, MeOD) δ 5.38 (s, 1H), δ 5.15 (s, 1H), δ 4.95 (s, 1H), δ 4.24-4.28 (m,

2H), δ 4.07-4.12 (m, 1H), δ 3.89-3.41 (m, 2H), δ 3.72-3.78 (m, 2H), δ 3.17-3.57

(m, 12H), δ 2.89-2.95 (m, 7H), δ 1.93-1.99 (m, 1H), δ 1.27-1.48 (m, 56H).

Neo-N-Neo · TFA12 (44). Boc6 Neo-N-Neo (9 mg, 7 µmoles) was dissolved in

CHCl3 (1.5 mL), tiisopropyl silane (100 µL), TFA (1.5 mL), and stirred RT for 10

min. The reaction was then partitioned into water (40 mL) and diethyl ether (20

mL), the ether was discarded and the aqueous phase was washed with 1:30

methanol / diethyl ether (3x20 mL). The aqueous phase was then concentrated

to a solid under vacuum. The solid was dissolved in 4 mL 0.1% TFA / water and

lyophilized to yield a white solid (9 mg, 99%). 1H-NMR (400 MHz, D2O) δ 5.92 (d,

J = 4.0 Hz, 2H), δ 5.27 (d, J = 2.0 Hz, 2H), δ 5.17 (s, 2H), δ 4.27-4.31 (m, 4H), δ

4.21 (q, J = 4.0 Hz, 2H), δ 4.17 (t, J = 4.8 Hz, 2H), δ 4.08 (t, J = 2.4 Hz, 2H), δ

3.96 (t, J = 9.6 Hz, 2H), δ 3.88 (t, J = 10 Hz, 2H), δ 3.75-3.80 (m, 4H), δ 3.68 (d, J

= 2.4 Hz, 2H), δ 3.56 (t, J = 9.8 Hz, 2H), δ 3.18-3.46 (m, 14H), δ 3.11 (d,d J1 =

9.2 Hz, J2 = 7.6 Hz, 2H), δ 2.98 (d,d J1 = 13.6 Hz, J2 = 4.0 Hz, 2H), δ 2.84-3.02

(m, 10H), δ 2.73 (d,d J1 = 13.2 Hz, J2 = 6.0 Hz, 2H), δ 2.35 (d,t J1 = 12.8 Hz, J2 =

4.0 Hz, 2H), δ 1.76 (q, J = 12.4 Hz, 2H). MALDI TOF MS calculated for

C50H98N12O24S4:1378.6, observed: 1401.7 [M+Na]+.

259

Figure E 6.1: 1H NMR of Neo-Neo (43) before and after dimerization (A and B,respectively) in D20. Notice that the only significant changes are in the “linkerregion” at 2.6 and 3.6 ppm. 1H NMR of Neo-N-Neo (44) (C); again the onlysignificant differences between Neo-Neo and Neo-N-Neo are in the “linkerregion” at 2.8 and 3.6 ppm. The 1H NMR spectrum of neomycin B · TFA6 is alsoshown (D).

2.02.53.03.54.04.55.05.56.0

2.02.53.03.54.04.55.05.56.0

2.02.53.03.54.04.55.05.56.0

2.02.53.03.54.04.55.05.56.0

A) 5” β-Mercaptoethyl etherNeomycin B • TFA6

B) Neo-Neo • TFA12

C) Neo-N-Neo • TFA12

D) Neomycin B • TFA6

260

E 6.4.1 Tobra-Arginine (53):

Tobramycin-Arg(Pbf)-Fmoc. Tobramycin free base (96 mg, 0.206 mmoles) was

added to a solution containing: Arg(Pbf)Fmoc (1.34 g, 2.06 mmoles, 2 mequiv),

TBTU (1.33 g, 4,12 mmoles, 4 mequiv), DMF (15 mL, freshly purified over silica

gel) 2,4,6 collidine (8.24 mmoles, 8 mequiv) and heated at 35 °C for 16 hr then

diluted into ethyl acetate (150 mL) and washed with 0.1M citric acid (2x70 mL),

saturated sodium bicarbonate (2x70 mL), brine (50 mL), dried over sodium

sulfate and concentrated to a solid under reduced pressure. The product was

purified using silica gel and standard flash chromatography (0-7% MeOH /

CH2Cl2) to yield 237 mg (32%) of a white solid. Rf = 0.33 (10 % MeOH / CH2Cl2).

The product was used in subsequent steps without further characterization

O

NH2

HOH2N

H2NO

HOHO

O

O

H2N OH

NH2

HOH

NHOH

OO

O

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

HOH

NH-Fmoc

O

NHHN

Pbf-HN NH-FmocO

HNHN

NH-Pbf

Fmoc-HN

O

NH NH

NH-Pbf

NH-Fmoc

OHNHN

Pbf-HN

NH-Fmoc

O HN

HNNH-Pbf

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

OH

H2NO

HN

HN NH2

H2N

O

NH NH

NH2

NH2

ONH NH

NH2

NH2

O

HNHN NH2

H2N

O

HN

NH

NH2

TFA , water, TIPS (56%)

Tobra-Arg (53)

Tobramycin (3)

Tobramycin-Arg(Pbf)-Fmoc

Tobramycin-Arg(Pbf)-NH2

20% piperidine / DMF (96%)

HOBt, TBTU, 2,4,6collidine, DMF (32%)

NHNH

HNS

O

O

O

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

HOH

NH2

OHNHN

NH-PbfH2N

O

HN

HNNH-Pbf

H2N

O

NH NH

NH-Pbf

NH2

O

NHHN

Pbf-HN

NH2

OHN

NH

NH-Pbf

261

Tobramycin-Arg(Pbf)-NH2. Tobramycin-Arg(Pbf)-Fmoc (100 mg, 27 µmoles)

was dissolved in DMF (4 mL) and piperidine (1 mL) was added, mixed RT for 14

hr, then concentrated and dried under reduced pressure (40 °C) to yield 67 mg

(98%) of a white solid. MALDI TOF MS calculated for C113H177O29S5: 2528, found

2531 [M+Na]+.

Tobramycin-Arg (53). Tobramycin-Arg(Pbf)-NH2 (25 mg, 9.8 µmoles) was

dissolved in CH2Cl2 (7 mL), added to a TFA solution (7 mL, containing 5% water

and 1.5% triisopropyl silane) and mixed RT for 2hr, water (40 mL) was then

added and the reaction washed with diethyl ether (3x20 mL), the aqueous phase

was then concentrated to a solid under reduced pressure. The product was

HPLC purified on C-18 semiprep column using isocratic conditions (7%

acetonitrile in water and 0.1% TFA) (Rt = 9.9 min) and lyophilized to yield 14 mg

of a white solid (56% assuming 10 TFA counter ions). 1H NMR (400 MHz, D2O,

22 °C): δ 5.33 (d, J = 3.2 Hz, 1H), δ 4.93 (d, J = 4.0 Hz, 1H), δ 3.82-3.97 (m,

9H), δ 3.77 (t, J = 6.6 Hz, 1H), δ 3.63-3.70 (m, 2H), δ 3.29-3.58 (m, 7H), δ 3.03-

3.11 (m, 10H), δ 1.74-1.84 (m, 12H), δ 1.45-1.59 (m, 12H). MALDI TOF MS

calculated for C48H97N25O14: 1247.7, observed: 1248.7 [M+H]+.

262

E 6.4.2 ace-Tobra-Arginine (54):

ace-Tobra-Arginine (54). Tobramycin-Arg(Pbf)-NH2 (section E 6.4.1, 35 mg,

13.9 µmoles) was treated with DMF (4 mL), acetic anhydride (0.2 mL), and dry

pyridine (0.4 mL) for 1.5 hr at 44 °C, then concentrated under reduced pressure,

re-dissolved in DMF (7 mL), concentrated under reduced pressure (50 °C), the oil

was re-dissolved in 1:1 DMF / water and concentrated to a solid under reduced

pressure (50 °C). All of the resulting product was carried into the next step upon

addition of CH2Cl2 (13 mL) and a TFA solution (13 mL, containing 5% water and

1.5% triisopropyl silane) the reaction was mixed at RT for 2hr, then diluted into

water (40 mL) and washed with diethyl ether (3x20 mL), the aqueous phase was

then concentrated to a solid, under reduced pressure. The product was HPLC

purified on a C-18 semiprep column using isocratic conditions (7% acetonitrile in

O

HN

HOHN

HN

OHOHO

O

O

NH

OH

NH

OH

HN

O

HNHN NH2

HN

O

NH NH

NH2

NH

ONH NH

NH2HN

O

HNHN NH2

HN

O

HN NH

NH2

O

O

O

O

O

TFA , water, TIPS

ace-Tobra-Arg (54)

(Ac)2O, pyridine

(42%, 2 steps)

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

HOH

NH2

OHNHN

NH-PbfH2N

O

HN

HNNH-Pbf

H2N

O

NH NH

NH-Pbf

NH2

O

HNHN

Pbf-HN

NH2

OHN

NH

NH-Pbf

Tobramycin-Arg(Pbf)-NH2

263

water with 0.1% TFA) (Rt = 9.4 min), fractions were pooled and lyophilized to

yield 8.6 mg of a white solid (42% assuming 10 TFA counter ions). 1H NMR (400

MHz, D2O, 21 °C): δ 5.11 (d, J = 2.8 Hz, 1H), δ 5.08 (d, J = 4.0 Hz, 1H), δ 4.11-

4.16 (m, 3H), δ 3.79-4.08 (m, 6H), δ 3.56-3.63 (m, 5H), δ 3.31-3.41 (m, 6H), δ

3.03-3.08 (m, 10H), δ 1.81-1.90 (m, 14H), δ 1.45-1.71 (m, 25H). MALDI TOF MS

calculated for C58H107N25O19: 1457.8, observed: 1458.5 [M+H]+.

E 6.4.3 Kana-A-Dab (55):

Kana-A-Dab(Cbz)-Boc. Kanamycin A free base (100 mg, 206 µmoles), water (2

mL), 1,4 dioxane (6 mL), Boc-Dab(Z)-Osu (0.5 g, 1.1 mmoles, 1.35 mequiv),

triethyl amine (150 µL, 1.1 mmoles) were stirred for 14 hr at RT, then ethylene

diamine (0.2 mL) was added and allowed to react for 10 min. The reaction was

partitioned into ethyl acetate (150 mL) and 1 M NaH2PO4 (100 mL), the aqueous

layer was removed and the organic phase washed with 1 M NaH2PO4 (50 mL),

O

NH2

HOH2N

H2NO

HOHO

O

O

HO OH

NH2

OH

OH

NH

O

O

N

O

O

HN

O

O

O

O

NH

OHN

O

O OO

O

NH

HO HN

HN

OHOHO

O

O

HO OH

NH

OH

OH

NH

O

HN O

O O OHN

O

NHO

O

O

O

HN

O

NHO

O

O

O

NH2

O

H2N

O

NH

HO HN

HN

OHOHO

O

O

HO OHNH

OH

OH

NH2

O

NH2

H2N

O

H2N

H2N

O

H2N

Kanamycin A (1)

Kanamycin A -Dab(Cbz)-Boc

water, dioxane, TEA (27%)

Kanamycin A -Dab (55)

1) TFA2) H2 / Pd (99%, 2 steps)

264

saturated sodium bicarbonate (2x50 mL), brine (50 mL), then dried over sodium

sulfate and concentrated to a solid under reduced pressure. The crude product

was purified on silica gel (100 mL) using a 3 – 9% methanol/CH2Cl2 gradient (the

product elutes from 8 – 9% methanol). To yield a white solid (100 mg, 27%). Rf =

.25 (10% MeOH / CH2Cl2). The product was used further without additional

characterization.

Kana-A-Dab · TFA8 (55). Kana-A-Dab(Cbz)-Boc (30 mg, 16.5 µmoles), CH2Cl2

(2 mL), and TFA (2 mL) were stirred at RT for 15 min then concentrated to an oil

under reduced pressure. Anhydrous toluene (15 mL) was used to dissolve the oil

and was then removed under reduced pressure (50 °C) to yield white solid. All of

this product was taken to the next step without characterization. The white solid

was combined with methanol (3 mL), glacial acetic acid (2 mL) 10% Pd on

carbon (20 mg) and reacted with H2 (1.2 atm) for 14 hr at RT. The catalyst was

then removed, washed with water (2 mL), and the solvent removed under

reduced pressure. The resulting oil was then dissolved in water (2 mL, 0.1%

TFA), filtered through a 0.45 µm nylon filter, and lyophilized. The solid was then

re-dissolved in water (5 mL 0.1% TFA) and lyophilized to yield a white solid (31

mg, 99%, assuming 8 TFA counter ions). The product was analyzed using an

HPLC reverse phase C-18 column (3. 5mL/min 1 % acetonitrile in water and

0.1% TFA) and found to be >95% pure (detection at 215 nm, Rt = 6.9 min). 1H

NMR (400 MHz, D2O, 22 °C): δ 5.41 (d, J = 3.6 Hz, 1H), δ 4.99 (d, J = 3.6 Hz,

1H), δ 3.88-4.03 (m, 7H), δ 3.44-3.76 (m, 9H), δ 3.28-3.35 (m, 3H), δ 3.17 (t, J =

265

9.6 Hz, 1H), δ 2.93-3.03 (m, 8H), δ 2.10-2.17 (m, 8H), δ 1.89 (d,t J1 = 12.0 Hz, J2

= 3.8 Hz, 1H), δ 1.48 (q, J = 12.0 Hz, 1H). MALDI TOF MS calculated for

C34H68N12O15: 884.49, observed: 907.47 [M+Na]+.

E 6.4.4 Kana-A-Lys (56):

Kana-A-Lys(ClCbz)-Boc. Kanamycin A free base (40 mg, 82 µmoles), methanol

(8 mL), Boc-Lys(ClCbz)-Osu (0.5 g, 0.98 mmoles, 3 mequiv), triethyl amine (200

µL, 1.5 mmoles) were stirred for 24 hr at RT, then diluted into ethyl acetate (120

mL) and washed with 1 M Na2CO3 (2x50 mL), 1 M NaH2PO4 (2x50 mL), and

brine (50 mL). The organic phase was then dried over sodium sulfate and

reduced, under vacuum, to an oily solid that was purified on 80 mL of silica gel

using standard flash chromatography and a 3 – 10% methanol CH2Cl2 gradient to

yield a white solid (30 mg, 18%). Rf = .28 (10% MeOH / CH2Cl2). The product

was used further without additional characterization.

O

NH2

HOH2N

H2NO

HOHO

O

O

HO OH

NH2

OH

OH

NH

O

O

N

O

O

O

O

NH2

O

O

HN

HO HN

HN

OHOHO

O

O

HO OHNH

OH

OH

NH2

O

H2N

OH2N

O

Kanamycin A (1)

Kanamycin A -Lys(ClCbz)-Boc

methanol TEA (18%)

Kanamycin A -Lys (56)

1) TFA2) H2 / Pd (90%, 2 steps)

HN

O

OCl

NH2 H2N

H2NNH2

NH-Boc

O

O

HN

HO HN

HN

OHOHO

O

O

HO OH

NH

OH

OH

NH-Boc

O

Boc-HN

O

Boc-HN

O

NH-ClCbz

ClCbz-HN

NH-ClCbzNH-ClCbz

266

Kana-A-Lys · TFA8. Kana-A-Dab(ClCbz)-Boc (29 mg, 14 µmoles), methanol (5

mL), glacial acetic acid (1mL), 10% Pd on carbon (20 mg), were mixed under 1.2

atm of H2 for 14 hr at RT. The catalyst was then removed by centrifugation, and

the supernatant concentrated to a solid under reduced pressure. 20% of this

crude product was then carried over, directly, to the next step by adding CH2CL2

(2 mL), TFA (2 mL) and mixing at RT for 15 min. The reaction was then

concentrated to an oil under reduced pressure. Anhydrous toluene (15 mL) was

used to dissolve the oil and was then removed under reduced pressure (50 °C) to

yield white solid. A C-18 reversed phase column and HPLC was used for

purification (3. 5mL/min 2.5% acetonitrile in water and 0.1% TFA, detection at

215 nm, Rt = 6.5 min). To yield a white solid (4.8 mg, 90%). 1H NMR (400 MHz,

D2O, 21 °C): δ 5.37 (d, J = 3.6 Hz, 1H), δ 4.93 (d, J = 3.6 Hz, 1H), δ 3.59-3.99

(m, 13H), δ 3.26-3.49 (m, 6H), δ 3.08 (t, J = 9.8 Hz, 1H), δ 2.81-2.87 (m, 8H), δ

1.72-1.78 (m, 9H), δ 1.50-1.60 (m, 9H), δ 1.26-1.31 (m, 8H). MALDI TOF MS

calculated for C42H84N12O15: 996.62, observed: 977.71 [M+H]+, 1019.69 [M+Na]+.

267

E 6.4.5 γγγγ–guanidino Kana-A-Dab (57):

γγγγ–guanidino Kana-A-Dab · TFA8 (57). Kana-A-Dab(Cbz)-Boc (section 6.4.3) (27

mg, 14.8 µmoles), methanol (5 mL), 10% Pd on carbon (17 mg), glacial acetic

acids (0.2 mL) were stirred under 1.2 atm of H2 at RT for 16 h. The catalyst was

then removed by centrifugation, and the supernatant concentrated to an oil under

reduced pressure. Anhydrous toluene was then added and removed under

reduced pressure at 50 °C (repeated 2x total) to afford 18 mg of a white solid.

MALDI TOF MS calculated for C54H100N12O23: 1284.70, observed: 1285.63

[M+H]+, 1307.61 [M+Na]+. ½ of this product was carried over, without further

characterization, to the guanidinylation steps. Kana-A-Dab-Boc (9 mg, 7.0

µmoles), methanol (1 mL), CH2Cl2 (3 mL), N,N'-diBoc-N''-triflylguanidine (40 mg,

102 µmoles), triethyl amine (60 µL, 450 µmoles) were stirred at RT for 21 hr then

reduced to an oil under vacuum and purified using 20 mL of silica gel and flash

NH

OHN

O

O OO

O

NH

HO HN

HN

OHOHO

O

O

HO OH

NH

OH

OH

NH

O

HN O

O O OHN

O

NHO

O

O

O

HN

O

NHO

O

O

O

H2N

O

HN

O

HN

HOHN

HN

OHOHO

O

O

HO OH

NH

OH

OH

H2N

O

NH

H2N

O

HN

H2N O

HN

Kanamycin A -Dab(Cbz)-Boc

γ-guanidino Kanamycin A -Dab (57)

NH

OH2N

OO

O

NH

HO HN

HN

OHOHO

O

O

HO OH

NH

OH

OH

NH

O

H2N

O OHN

O

NH2O

O

HN

O

NH2O

O

Kanamycin A -Dab-Boc

H2 / Pd

HN

NH2

HN

NH2

NH

NH2NH

H2N

N-Tf

NH-BocBoc-HN

TFA50% (3 steps)

268

chromatography (4 – 6% methanol/CH2Cl2 gradient) to yield 8 mg of a white

solid. Rf = 0.33 (7% methanol/CH2Cl2). All of this product was deprotected by

adding CH2Cl2 (2 mL), triisopropyl silane (40 µL), and TFA (3 mL). The reaction

was carried out at RT for 4 hr then partitioned into 1% acetic acid in water (70

mL), and diethyl ether (70 mL). The ether was back-extracted one time with 30

mL of water. The combined aqueous fractions were then washed with diethyl

ether (2x40 mL) and concentrated to a solid under reduced pressure. The solid

was then dissolved in water (3 mL, 0.1% TFA), filtered trough a 0.45 mm nylon

filter, and lyophilized to a white solid (7.2 mg, 50%, 3 steps). The product was

analyzed using an HPLC reverse phase C-18 column (3. 5mL/min 0.5 – 4%

acetonitrile gradient in water and 0.1% TFA over 20 min) and found to be >90%

pure (detection at 215 nm, Rt = 16.5 min). For the RNA binding assays, a small

amount of this compound (3 mg) was purified to >99% using the same HPLC

conditions. 1H NMR (400 MHz, D2O, 21 °C): δ 5.39 (d, J = 3.2 Hz, 1H), δ 4.98 (d,

J = 3.2 Hz, 1H), δ 3.87-4.02 (m, 7H), δ 3.72 (t, J = 8.8 Hz, 1H), δ 3.16-3.63 (m,

20H), δ 2.02-2.07 (m, 8H), δ 1.87 (d,t J1 = 12.0 Hz, J2 = 3.8 Hz, 1H), δ 1.48 (q, J

= 12.0 Hz, 1H). MALDI TOF MS calculated for C38H76N20O15: 1052.58, observed:

1053.45 [M+H]+, 1075.43 [M+Na]+.

269

E 6.2.1 Synthesis and Characterization of guanidino5 Tobra-N-acridine (58)and guanidino4 Tobra-N-acridine (59):

Guanidino5 Tobra-N-acridine · TFA5 (58), and guanidino4 Tobra-N-acridine ·

TFA5 (59). Tobra-N-acridine · TFA5 (Section E 6.2.1) (13 mg, 11 µmoles), water

(0.45 mL), 1,4 dioxane (1.8 mL), N,N'-diBoc-N''-triflylguanidine (52 mg, 133

µmoles), triethyl amine (45 µL, 190 µmoles) were stirred at RT for 72 hr, then the

volatiles were removed under reduced pressure. The yellow oil was dissolved in

1,4 dioxane (1 mL) and triethyl amine (22 µL, 93 µmoles) and stirred at RT for 24

hr, then diluted into ethyl acetate (100 mL) and washed with water (3x50 mL),

brine (50 mL), dried over sodium sulfate, and concentrated to a solid under

reduced pressure. The Boc-protected products (Rf = .31, 10% methanol/CH2Cl2)

were purified on silica gel using a 5 – 12% methanol/CH2Cl2 gradient. The

mixture of the two Boc protected products were then deprotected in a

deprotection cocktail (10 mL of 88% TFA, 5% water, 5% phenol, and 2%

O

HN

HO HN

H2NO

HOHO

O

O

NH

OHNH

NH

H2N

NH

NH2

NH

NH2

NH

H2N

NH

N

Guanidino4 Tobra-N-acridine (59)

O

HN

HOHN

HN

OHOHO

O

O

NH

OHNH

NH

H2N

NH

H2NNH

NH2

NH

NH2

NH

H2N

NH

N

Guanidino5 Tobra-N-acridine (58)

O

NH2

HOH2N

H2NO

HOHO

O

O

H2N OHNH2

NH

N

N-Tf

NH-BocBoc-HN

TFA

Tobra-N-acridine (41)

(21%)

(12%)

270

triisopropylsilane (v/v)) at RT for 2 hr. The reaction was then was then diluted into

2% acetic acid (100 mL) and washed with diethyl ether (3x30 mL). The aqueous

phase was then reduced to a solid under reduced pressure and purified on a C-

18 reversed phase HPLC column with an isocratic mixture of 10% acetonitrile

(0.1% TFA) in water (0.1% TFA) (3 mL/min). The first peak to elute from the

column yielded 1.8 mg (12%) of guanidino4 tobra-N-acridine · TFA5 (59) (Rt =

18.6 min). 1H NMR (400 MHz, D2O, 22 °C): δ 8.26 (d, J = 8.0 Hz, 2H), δ 7.78 (t,

J = 7.8 Hz, 2H), δ 7.63 (d, J = 8.4 Hz, 2H), δ 7.38 (t, J = 7.6 Hz, 2H), δ 5.08 (d, J

= 3.6 Hz, 1H), δ 4.53-4.62 (m, 2H), δ 4.11 (d, J = 3.6 Hz, 1H), δ 3.76-3.83 (m,

1H), δ 3.65 (d,d J1 = 10.0 Hz, J2 = 4.0 Hz, 1H), δ 3.46-3.61 (m, 4H), δ 3.27-3.39

(m, 4H), δ 3.14-3.20 (m, 1H), δ 2.61 (d,t J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ 2.20

(d,t J1 = 12.8 Hz, J2 = 4.0 Hz, 1H), δ 1.78 (d,t J1 = 11.6 Hz, J2 = 4.0 Hz, 1H), δ

1.58 (q, J = 12.0 Hz, 1H), δ 1.35 (q, J = 12.0 Hz, 1H). MALDI FTMS calculated

for C35H53N15O8: 811.420, found 812.418 [M+H]+.

The second peak from the column yielded 3.2 mg (21%) of guanidino5 tobra-N-

acridine · TFA5 (58) (Rt = 19.9 min). 1H NMR (400 MHz, D2O, 22 °C): δ 8.27 (d, J

= 7.6 Hz, 2H), δ 7.79 (t, J = 7.8 Hz, 2H), δ 7.63 (d, J = 8.8 Hz, 2H), δ 7.40 (t, J =

7.6 Hz, 2H), δ 5.09 (d, J = 3.6 Hz, 1H), δ 4.70 (d, J = 1.2 Hz, 1H), δ 4.53-4.62 (m,

2H), δ 3.02-3.79 (m, 11H), δ 2.59 (d,t J1 = 12.4 Hz, J2 = 4.0 Hz, 1H), δ 2.07 (d,t J1

= 12.8 Hz, J2 = 4.0 Hz, 1H), δ 1.79 (d,t J1 = 11.6 Hz, J2 = 4.0 Hz, 1H), δ 1.45 (q, J

= 12.0 Hz, 1H), δ 1.21 (q, J = 12.0 Hz, 1H). MALDI FTMS calculated for

C36H55N17O8: 853.442, found 854.451 [M+H]+.

271

E 6.2.2 Synthesis and Characterization of guanidino6 Neo-N-acridine (60)and guanidino5 Neo-N-acridine (61):

Guanidino6 Neo-N-acridine · TFA6 (60), and guanidino5 Neo-N-acridine ·

TFA6 (61). Neo-N-acridine · TFA6 (Section E 6.1.1) (1 mg, 670 nmoles), water

(0.1 mL), 1,4 dioxane (0.4 mL), N,N'-diBoc-N''-triflylguanidine (15 mg, 38

µmoles), triethyl amine (45 µL, 190 µmoles) were incubated at RT for 72 hr, then

the volatiles were removed under reduced pressure. The yellow oil was dissolved

in 1,4 dioxane (0.3 mL) and kept at RT for 72 hr, then concentrated to an oil

under reduced pressure. The Boc-protected products were purified on silica gel

using 8% methanol/CH2Cl2. The mixture of the two Boc protected products were

then deprotected in a deprotection cocktail (4 mL of 88% TFA, 5% water, 5%

phenol, and 2% triisopropylsilane (v/v)) at RT for 2 hr. The reaction was then was

then diluted into 2% acetic acid (30 mL) and washed with diethyl ether (3x10

HN

N

HN

HNO

OO

HN

OH

HOO

HOO

HO

HN

OHO

HO

NH

NH

H2NNH

HNNH2

NH

NH2

NH

NH2NH

H2N

HNNH2

Guanidino6 Neo-N-acridine (60)

HN

N

HN

HNO

OO

H2N

OH

HOO

HOO

HO

HN

OHO

HO

NH

NH

H2NNH

NH

NH2

NH

NH2NH

H2N

HNNH2

Guanidino5 Neo-N-acridine (61)

HN

N

NH2

H2NO

OO

H2N

OH

HOO

HOO

HO

NH2

OHO

HO

NH2

NH2

Neo-N-acridine (38)

N-Tf

NH-BocBoc-HN

TFA

272

mL). The aqueous phase was then reduced to a solid under reduced pressure

and purified on a C-18 reversed phase HPLC column with an isocratic mixture of

11.5% acetonitrile (0.1% TFA) in water (0.1% TFA) (3 mL/min). The UV-vis

extinction coefficients for Neo-N-acridine (Section 6.1.1) were used to quantify

the products and yielded 9 nmoles (1.3%) of guanidino6 neo-N-acridine · TFA6

(60) (Rt = 12.2 min) MALDI FTMS calculated for C42H66N20O12: 1042.517, found

1043.524 [M+H]+. And, 5 nmoles (0.9%) of guanidino5 neo-N-acridine · TFA6 (61)

(Rt = 10.8 min) MALDI FTMS calculated for C41H64N18O12: 1000.495, found

1001.498 [M+H]+.

E 6.7.1 Synthesis and Characterization of Amino-Tobra-BODIPY (65):

6''-TIPS Boc5 Tobramycin. Synthesized as described (Hai Wang Ph. D. Thesis,

University of California, San Diego, 1998).

OHONH2

ONH2

NH2O

ONH2NH2

OH

HO HO

S

NNB

FFO

S

HN

O

Amino Tobra-BODIPY (65)

OHONH2

ONH2

NH2O

ONH2NH2

OH

HO HO

S

O

HS

OHOBoc-HN

OBoc-HN NH-Boc

O

O NH-BocOH

HO HO

O

S OO

β-mercapto-ethylether

TFA

(61%, 2 steps)

6''-TIPS Boc5 Tobramycin

6''-β-mercaptoethylether-Tobramycin

BODIPY C1-IA(47%)

Boc-HN

273

6"-ββββ-Mercaptoethyl Ether Tobramycin · TFA5: 6" TIPS Boc5 Tobramycin (40

mg, 32 µmoles) Cs2CO3 (21 mg, 64 µmoles), dry DMF (3 mL) and 2-mercapto-

ethyl ether (34 µl, 274 µmoles, 8.6 equiv) were stirred under argon for 2h at

30 οC, diluted into ethyl acetate (100 mL), washed with water (4x50 mL), brine

(50 mL) then dried over sodium sulfate. The organic layer was then concentrated

under to an oil reduced pressure, and kept under a high vacuum for 40 min. The

crude product was dissolved in CH2Cl2 (2 mL), 1,2 ethanedithiol (15 µL),

triisopropy silane (15 µL), trifluoroacetic acid (3 mL), and stirred at RT for 15

minn. The reaction was then diluted into toluene (50 mL) and concentrated to a

solid at 50 οC under reduced pressure (2x). The white solid was then dissolved

into water and (100 mL) and washed with CHCl3 (4x50 mL). The aqueous layer

was concentrated under reduced pressure and lyophilized (2x) from 0.1% TFA (3

mL in water) to yield 25 mg of a white powder (61% yield, two steps). 1H-NMR

(400 MHz, D2O) δ 5.59 (d, J = 3.6 Hz, 1H), δ 4.85 (d, J = 4.0 Hz, 1H), δ 3.3-3.8

(m, 17H), δ 3.05 (d,d J1 = 13.6 Hz, J2 = 7.2 Hz, 1H), δ 2.94 (d,d J1 = 11.6 Hz, J2 =

2.4 Hz, 1H), δ 2.65-2.71 (m, 3H), δ 2.57 (t, J = 6.4 Hz, 2H), δ 2.40 (d,t J1 = 12.8

Hz, J2 = 4.0 Hz, 1H), δ 2.14(d,t J1 = 12.4 Hz, J2 = 4.4 Hz, 1H), δ 1.87 (q, J = 11.2

Hz, 1H), δ 1.78 (q, J = 12.8 Hz, 1H). ESI MS calculated for C22H45N5O9S2: 587.3,

found 588.2 [M+H]+.

274

Amino Tobra-BODIPY · HCl (65): 6"-β-Mercaptoethyl ether tobramycin · TFA5 (5

mg, 4.3 µmoles) was dissolved in a degassed aqueous buffer (1 mL of 150 mM

NaCl, 10 mM sodium phosphate pH 7.5, Ar sparged). Separately, BODIPY C1-IA

(2.5 mg, 6 µmoles, 1.4 equiv., Molecular Probes) was dissolved in DMSO (0.75

mL), and added, dropwise, to the tobramycin solution. The reaction was kept in

the dark for 2h at RT then diluted into water (8 mL) and loaded onto an activated

C-18 reversed-phase cartridge (Waters, Sep-pack), the column was washed with

1M NaCl (5 mL), pure water (5 mL), then a 0 – 30% acetonitrile/water gradient

was applied, the fractions between 5 – 15% acetonitrile/water were lyophilized to

yield 2.2 mg (47%) of a red powder. All BODIPY-glycoside conjugates are slightly

to moderately hygroscopic, therefore the absorption at 502nm of each compound

(in methanol) is used to confirm the yield of the conjugation reaction (taking ε502

nm = 76,000 cm-1 M-1). 1H-NMR (400 MHz, D2O) δ 7.39 (s, 1H), δ 6.91 (d, J = 4.0

Hz, 1H), δ 6.30 (d, J = 4.0 Hz, 1H), δ 6.22 (s, 1H), δ 5.55 (d, J = 3.6 Hz, 1H), δ

4.94 (d, J = 4.0 Hz, 1H), δ 4.49 (s, 2H), δ 3.25-3.90 (m, 19H), δ 3.05 (d,d J1 =

13.6 Hz, J2 = 7.2 Hz, 1H), δ 2.93 (d,d J1 = 11.6 Hz, J2 = 2.4 Hz, 1H), δ 2.60-2.67

(m, 5H), δ 2.40 (s, 3H), δ 2.29 (d,t J1 = 12.4 Hz, J2 = 3.6 Hz, 1H), δ 2.10-2.14 (m,

4H), δ 1.84 (q, J = 11.6 Hz, 1H), δ 1.66 (q, J = 12.8 Hz, 1H). MALDI TOF MS

calculated for C36H59BF2N8O10S2: 876.3 found 877.4 [M+H]+, found 899.3

[M+Na]+ found 915.4 [M+K]+.

275

E 6.7.2 Synthesis and Characterization of Guanidino-Tobra-BODIPY (66):

Boc10 Guanidino5 6"-ββββ-Mercaptoethyl Ether Tobramycin: 6"-β-mercaptoethyl

ether tobramycin · TFA5 (Section 6.7.1) (70 mg, 60 µmoles), was dissolved in

methanol (4mL) and treated with N,N′-di-Boc-N′′ -trifylguanidine (420 mg, 1.08

mmoles, 17.9 equiv.), dithiothreitol (42 mg, 272 µmoles), and triethyl amine (210

µL, 1.5 mmoles, 25 eqiv.) for 26 h at RT under argon. The reaction was then

diluted into 150 mL of CHCl3 and washed with 0.1M citric acid (3x50 mL), brine

(50 mL) and dried over sodium sulfate. The organic layer was then concentrated

to a solid and purified on silica gel using flash chromatography and 0 – 2%

methanol in CH2Cl2 to afford 90 mg of an off-white solid (83% yield). 1H-NMR

(400 MHz, CDCl3) δ 11.50 (s, 1H), δ 11.47 (s, overlapping, 2H), δ 11.45 (s, 1H), δ

11.38 (s, 1H), δ 8.86 (d, J = 3.6 Hz, 1H), δ 8.55 (d, J = 9.0 Hz, 1H), δ 8.46 (t, J =

NNB

FF

HN

O

Guanidino Tobra-BODIPY (66)

OHONH2

ONH2

NH2O

ONH2NH2

OH

HO HO

S

O

HS

6''-β-mercaptoethylether-Tobramycin

BODIPY C1-IA (17%)

OHOHN

OHN

HN

O

ONHNH

OH

HO HO

S

O

HS

Boc10Guanidino5-6"-β-Mercaptoethyl-ether Tobramycin

NH-Boc

N-Boc

N-Boc

Boc-HN

NH-Boc

NH-Boc

Boc-N

NH-Boc

N-BocBoc-HN

OHOHN

OHN

HN

O

ONHNH

OH

HO HO

S

O

HS

NH2

NH

NH

H2N

NH2

NH

HN

NH2

NHH2N

Guanidino5-6"-β-MercaptoethyletherTobramycin

(73%)

(83%)

TFA

N-Tf

NH-BocBoc-HN

OHOHN

OHN

HN

O

ONHNH

OH

HO HO

SOS

NH2

NH

NH

H2N

NH2

NHHN

NH2

NHH2N

276

6.3 Hz, 1H), δ 8.17 (d, J = 8.7 Hz, 1H), δ 5.30-5.40 (m, 2H), δ 4.97 (d, J = 3.9 Hz,

1H), δ 4.02 (br d,d J1 = 12 Hz, J2 = 8.4 Hz, 1H), δ 3.78-3.94 (m, 2H), δ 3.31-3.72

(m, 10H), δ 3.18 (br d, J1 = 11.4 Hz, 2H), δ 2.95-3.03 (m, 2H), δ 2.61-2.73 (m,

8H), δ 2.36 (s, 1H), δ 2.02 (s, 1H), δ 1.93 (s, 1H), δ 1.62-1.69 (m, 18H), δ 1.42-

1.51 (m, 18H), δ 1.28 (s, 1H). ESI MS calculated for C77H135N15O29S2: 1797.8,

found 1798.3 [M+H]+, found 899.7 [M+2H]2+.

Guanidino5 6" ββββ-Mercaptoethyl Ether Tobramycin · TFA5: Boc10 guanidino5

6"-β-mercaptoethyelther tobramycin (41 mg, 23 µmoles) was dissolved in CHCl3

(1mL) and treated with triisopropy silane (30 µL, 146 µmoles), 1,2-ethanedithiol

(30 µL, 358 µmoles), and trifluoroacetic acid (1.5mL) for 3 h at RT. The reaction

was then diluted into water (100 mL) and washed with CHCl3 (2x30 mL) and

diethyl ether (2x30 mL). The aqueous layer was concentrated to a solid under

vacuum, then dissolved in 0.1% trifluoroacetic acid in water (2 mL) and

lyophilized to afford 22 mg of a white solid (73% yield). 1H-NMR (400 MHz, d6-

MeOD) δ 5.65 (d, J = 3.6 Hz, 1H), δ 5.06 (d, J = 3.6 Hz, 1H), δ 4.10(t, J = 6.4 Hz,

1H), δ 3.45-3.88 (m, 17H), δ 3.04 (d,d J1 = 13.6 Hz, J2 = 2.8 Hz, 1H), δ 2.62-2.78

(m, 5H), δ 2.11-2.19 (m, 2H), δ 1.68-1.78 (m, 2H). MALDI TOF MS calculated for

C27H55N15O9S2: 797.37, found 820.32 [M+Na]+.

277

Guanidino Tobramycin-BODIPY · HCl (66): Guanidino5 6" β-mercaptoethyl-

ether tobramycin · TFA5 (10 mg, 4.3 µmoles) was added to an aqueous

degassed buffer (2 mL of 50 mM sodium phosphate pH 7.5, Ar sparged),

separately BODIPY C1-IA (2.5 mg, 6 µmoles, 1.4 equiv., Molecular Probes) was

dissolved in DMSO (0.75 mL), and added, dropwise, to the tobramycin solution.

The resulting precipitation of guanidino-tobramycin was partially reversed upon

addition of NaCl (150 mM final concentration) the reaction was then allowed to

react in the dark for 2h at RT and diluted into 5% acetonitrile in water (15 mL,

containing 100 mM NaCl) and loaded onto an activated C-18 reversed-phase

cartridge (Waters, Sep-pack). The column was then washed with 5 mL of water

and the product was eluted with 25% acetonitrile/water and lyophilized to yield

1.3 mg (17% yield) of a red powder. All BODIPY-glycoside conjugates are slightly

to moderately hygroscopic, the absorption at 502nm (in methanol) was used to

calculate the yield of the conjugation reaction (taking ε502 nm = 76,000 cm-1 M-1).

The low yield of this particular reaction was attributed to the solubility problems of

the guanidino-tobramycin starting material in 50 mM phosphate/25% DMSO in

water (its solubility in 10 mM sodium phosphate pH 7.5, 250 mM NaCl, 25%

DMSO is, however, significantly better). 1H-NMR (400 MHz, D2O) δ 7.41 (s, 1H),

δ 6.92 (d, J = 3.6 Hz, 1H), δ 6.31 (d, J = 3.6 Hz, 1H), δ 6.24 (s, 1H), δ 5.40 (d, J =

3.6 Hz, 1H), δ 4.98 (s, 1H), δ 4.51 (s, 2H), δ 4.03 (t, J = 6.8 Hz, 1H), δ 3.79 (t, J =

8.4 Hz, 1H), δ 3.27-3.64 (m, 18H), δ 2.90 (d, J = 13.2 Hz, 1H), δ 2.55-2.67 (m,

5H), δ 2.41 (s, 3H), δ 2.02-2.15 (m, 5H), δ 1.54-1.62 (m, 2H). MALDI TOF

278

calculated for C41H69BF2N18O10S2:1086.49, found 1087.36 [M+H]+, found

1109.30 [M+Na]+.

E 6.7.3 Synthesis and Characterization of Amino-Neo-BODIPY (67):

Amino Neo-BODIPY · HCl (67) : 5" β-mercaptoethylether neomycin B · TFA6

(Section 6.3.1) (10 mg, 7 µmoles), was dissolved in an aqueous buffer (1.5 mL of

10 mM sodium phosphate, 150 mM NaCl, pH 7.5, Ar sparged), separately,

BODIPY C1-IA (1.8 mg, 4.3 µmoles, 0.61 equiv., Molecular Probes) was

dissolved in DMSO (1.5 mL), and added, dropwise, to the neomycin solution and

allowed to react in the dark for 2h at RT. The reaction was then diluted into water

(8 mL) and loaded onto an activated C-18 reversed-phase cartridge (Waters,

Sep-pack), the column was then washed with 5% acetonitrile (5 mL, containing

100 mM NaCl in water), then pure water (1 mL), and the product eluted between

0 – 15% acetonitrile (in water) and lyophilized to yield 2.9 mg (55% yield) of a red

powder. All BODIPY-glycoside conjugates are slightly to moderately hygroscopic,

therefore the absorption at 502nm of each compound (in methanol) is used to

confirm the yield of the conjugation reaction (taking ε502 nm = 76,000 cm-1 M-1).

O

NH2

NH2O

HOHO

OO

NH2

OH

NH2

OHO

HOO

S

HO

NH2

NH2

NNB

FF

OS

NH

O

Amino Neo-BODIPY (67)

O

NH2

NH2O

HOHO

OO

NH2

OH

NH2

OHO

HOO

S

HO

NH2

NH2

OHS

5''-β-mercaptoethylether-neomycin B

NNB

FF NH

OI

DMSO, water (pH 7.5)(55%)

279

1H-NMR (400 MHz, D2O) δ 7.41 (s, 1H), δ 6.92 (d, J = 4.0 Hz, 1H), δ 6.30 (d, J =

4.0 Hz, 1H), δ 6.24 (s, 1H), δ 5.89 (d, J = 4.0 Hz, 1H), δ 5.27 (d, J = 3.6 Hz, 1H),

δ 5.15 (s, 1H), δ 4.50 (s, 2H), δ 4.26-4.32 (m, 2H), δ 4.21 (p, J = 4.0 Hz, 1H), δ

4.15 (t, J = 4.4 Hz, 1H), δ 4.07 (t, J = 3.2 Hz, 1H), δ 3.83-3.88 (m, 3H), δ 3.77 (t, J

= 9.2 Hz, 1H), δ 3.67 (s, 1H), δ 3.53-3.59 (m, 4H), δ 3.45 (s, 1H), δ 3.40 (d,d J1 =

11.2 Hz, J2 = 4.0 Hz, 1H), δ 3.67 (s, 1H), δ 3.20-3.35 (m, 5H), δ 3.09 (d d, J1 =

13.2 Hz, J2 = 7.6 Hz, 1H), δ 2.98 (d d, J1 = 13.6 Hz, J2 = 4.4 Hz, 1H), δ 2.67-2.77

(m, 5H), δ 2.41 (s, 3H), δ 2.25 (d t, J1 = 12.4 Hz, J2 = 4.4 Hz, 1H), δ 2.15 (s, 3H),

δ 1.65 (q, J = 12.8 Hz, 1H). MALDI TOF MS calculated for C41H68BF2N9O14S2:

1023.44, observed 1024.42 [M+H]+, observed 1046.43 [M+Na]+, observed

1062.54 [M+K]+.

280

E 6.7.4 Synthesis and Characterization of Guanidino Neo-BODIPY (68):

Boc12 Guanidino6-5"-ββββ-Mercaptoethylether Neomycin: 5" β-Mercaptoethyl

ether Neomycin B · TFA6 (Section E 6.3.1) (90 mg, 63 µmoles), was dissolved in

methanol (5 mL), CHCl3 (3 mL), and treated with N,N′-di-Boc-N′′ -trifylguanidine

(530 mg, 1.35 mmoles, 21 equiv.), dithiothreitol (50 mg, 324 µmoles), and triethyl

amine (530 µL, 3.8 mmoles, 60 eqiv.) for 96 h at RT under argon. The reaction

was then diluted into CHCl3 (200 mL) and washed with 0.1M citric acid (2x100

mL), brine (50 mL) and dried over sodium sulfate. The organic layer was then

concentrated to a solid under reduced pressure and purified on silica gel using

flash chromatography (0-1% methanol in CHCl3 to afford 71 mg of an off-white

NNB

FF NH

O

Guanidino Neo-BODIPY (68)

O

NH2

NH2O

HOHO

OO

NH2

OH

NH2

OHO

HOO

S

HO

NH2

NH2

OHS

5''-β-mercaptoethylether-neomycin B

BODIPY C1-IA (56%)

Boc12Guanidino6-5"-β-Mercaptoethyl-ether Neomycin B

Guanidino6-6"-β-Mercaptoethylether Nwomycin B

(52%)

N-Tf

NH-BocBoc-HN

O

HN

HNO

HOHO

OO

NH

OH

NH

OHO

HOO

S

HO

HN

NH

OHS

NH2

NH

NH2

NH

NHH2N

NHH2N

NHH2N

H2NNH

O

HN

HNO

HOHO

OO

NH

OH

NH

OHO

HOO

S

HO

HN

NH

OHS

NH-Boc

N-Boc

NH-Boc

N-Boc

N-BocBoc-HN

N-BocBoc-HN

N-Boc

Boc-HN

Boc-HN

N-Boc

(60%)

TFA

O

HN

HNO

HOHO

OO

NH

OH

NH

OHO

HOO

S

HO

HN

NH

OS

NH2

NH

NH2

NH

NHH2N

NHH2N

NHH2N

H2NNH

281

solid (52% yield). 1H-NMR (300 MHz, CDCl3) δ 11.41 (s, 2H, overlapping), δ

11.40 (s, 1H), δ 11.38 (s, 1H), δ 11.33 (s, 1H), δ 11.30 (s, 1H), δ 9.34 (d, J = 8.1

Hz, 1H), δ 8.92 (d, J = 7.2 Hz, 1H), δ 8.47 (t, J = 5.4 Hz, 1H), δ 8.41 (t, J = 5.4

Hz, 1H), δ 8.34 (d, J = 6.6 Hz, 1H), δ 8.18 (d, J = 9.0 Hz, 1H), δ 5.94 (d, J = 4.4

Hz, 1H), δ 5.62 (d, J = 4.2 Hz, 1H), δ 5.01 (d, J = 4.8 Hz, 1H), δ 4.91-4.94 (m,

2H), δ 4.38-4.95 (m, 3H), 4.05-4.22 (m, 4H), δ 3.79-3.92 (m, 3H), δ 3.66-3.74 (m,

2H), δ 3.55-3.59 (m, 4H), δ 3.25-3.42 (m, 4H), δ 2.54-2.74 (m, 5H), δ 2.44 (d t, J1

= 12.4 Hz, J2 = 4.4 Hz, 1H), δ 1.22-1.60 (m, 109H). ESI MS calculated for

C94H163N17O37S2: 2186.1, found 1094.2 [M+2H]2+.

Guanidino6 5"-ββββ-Mercaptoethylether Neomycin B · TFA6. Boc12 Guanidino6-

5"-β-mercaptoethyl ether neomycin (65 mg, 30µmoles) was dissolved in CHCl3

(1.5 mL) and treated with triisopropy silane (80 µL, 390 µmoles), 1,2-

ethanedithiol (30 µL, 955 µmoles), and trifluoroacetic acid (3 mL) for 3 h at RT.

The reaction was then diluted into water (200 mL) and washed with CHCl3

(2x100 mL) and diethyl ether (2x50 mL). The aqueous layer was then

concentrated to a solid under reduced pressure, dissolved in 0.1% trifluoroacetic

acid in water (2 mL) and lyophilized to afford 30 mg of a white solid (60% yield).

1H-NMR (400 MHz, D2O) δ 5.83 (d, J = 3.2 Hz, 1H), δ 5.05 (s, 1H), δ 4.94 (s, 1H),

δ 4.23-4.26 (m, 2H), δ 3.93-4.03 (m, 3H), δ 3.25-3.69 (m, 20H), δ 2.87 (d,d J1 =

14.4 Hz, J2 = 4.4 Hz, 1H), δ 2.55-2.67 (m, 5H), δ 2.08 (d t, J1 = 12.0 Hz, J2 = 4.4

282

Hz, 1H), δ 1.71 (q, J = 12.0 Hz, 1H). MALDI TOF MS calculated for

C33H66N18O13S2: 986.45, found 987.49 [M+H]+.

Guanidino Neo-BODIPY · HCl (68). Guanidino6 5"-β-Mercaptoethylether

Neomycin B · TFA6 (9 mg, 5.4 µmoles) was dissolved in an aqueous buffer (3.0

mL of 10 mM sodium phosphate pH 7.5, 150 mM NaCl, Ar sparged), separately

BODIPY C1-IA (1.7 mg, 4.1 µmoles, 0.76 equiv., Molecular Probes) was

dissolved in DMSO (1.5 mL), and added, dropwise, to the neomycin solution and

allowed to react in the dark for 2h at RT. The reaction was then diluted into water

(8 mL) and loaded onto an activated C-18 reversed-phase cartridge (Waters,

Sep-pack), the column was washed with 5% acetonitrile (5 mL, containing 100

mM NaCl in water), then pure water (1 mL), the product eluted between 0 – 20%

acetonitrile (in water) and was lyophilized to yield 4.5 mg (56% yield) of a red

powder. All BODIPY-glycoside conjugates are slightly to moderately hygroscopic,

therefore the absorption at 502nm of each compound (in methanol) is used to

confirm the yield of the conjugation reaction (taking ε502 nm = 76,000 cm-1 M-1). 1H-

NMR (400 MHz, D2O) δ 7.40 (s, 1H), δ 6.92 (d, J = 4.0 Hz, 1H), δ 6.31 (d, J = 4.0

Hz, 1H), δ 6.24 (s, 1H), δ 5.84 (d, J = 3.6 Hz, 1H), δ 5.07 (s, 1H), δ 4.91 (s, 1H), δ

4.51 (s, 2H), δ 4.21-4.24 (m, 2H), δ 3.93-4.00 (m, 3H), δ 3.80 (t, J = 8.4 Hz, 1H),

δ 3.66-3.71 (m, 2H), δ 3.25-3.55 (m, 15H), δ 2.80 (d,d J1 = 14.4 Hz, J2 = 4.4 Hz,

1H), δ 2.50-2.67 (m, 6H), δ 2.41 (s, 3H), δ 2.15 (s, 3H), δ 2.08 (d t, J1 = 12.0 Hz,

J2 = 4.4 Hz, 1H), δ 1.57 (q, J = 12.0 Hz, 1H). MALDI TOF MS calculated for

C47H80BF2N21O14S2: 1275.57, found 1276.57 [M+H]+.

283

E 6.7.5 Synthesis and Characterization of Amino-Tobra-Fluorescein (69):

Amino Tobra-Fluorescein · TFA5 (69): 6"-β-mercaptoethyl ether tobramycin ·

TFA5 (Section 6.7.1) (3 mg, 2.6 µmoles) was dissolved in an aqueous degassed

buffer (2 mL of 400 mM NaCl, 25 mM sodium phosphate pH 7.5, Ar sparged),

separately 5-iodo-acetamido-fluorescein (5-IAF) (3.0 mg, 5.8 µmoles, 2.0 equiv.,

Molecular Probes) was dissolved in DMSO (1 mL), and added, dropwise, to the

tobramycin solution and allowed to react in the dark for 2h at RT then 0.1M HCl

was added until the solution turned from orange to yellow. The reaction was then

diluted into water (8 mL) and loaded onto an activated C-18 reversed-phase

cartridge (Waters, Sep-pack), the column was then washed with pure water (10

mL), and the product was eluted with 20% acetonitrile/water, lyophilized, and

found to be >95% pure by HLPC. The product was purified further using a C-18

reversed phase HPLC column with an isocratic mixture of 20% acetonitrile (0.1%

TFA) in water (0.1% TFA) (3 mL/min) (Rt = 8.5 min) to yield 3.3 mg (77%) of an

orange solid. All fluorescein-glycoside conjugates are slightly to moderately

hygroscopic, therefore the absorption at 496nm (in aqueous buffer pH 9.0) of

Amino Tobra-Fluorescein (69)

OHONH2

ONH2

NH2O

ONH2NH2

OH

HO HO

S

O

HS

6''-β-mercaptoethylether-Tobramycin

(77%)

OHONH2

ONH2

NH2O

ONH2NH2

OH

HO HO

S

OS

O CO2H

HO

O

NH

O

5-IAF

284

each compound is used to confirm the yield of the conjugation reaction (taking

ε502 nm = 77,000 cm-1 M-1). 1H-NMR (400 MHz, D2O) δ 8.13 (d, J = 2.0 Hz, 1H), δ

7.67 (d,d J1 = 8.4 Hz, J2 = 1.6 Hz, 1H), δ 7.15-7.22 (m, 3H), δ 6.93 (d, J = 2.4 Hz,

2H), 6.77-6.80 (m, 2H), δ 5.53 (d, J = 2.8 Hz, 1H), δ 4.85 (d, J = 3.6 Hz, 1H), δ

3.22-3.83 (m, 21H), δ 3.05 (d,d J1 = 13.2 Hz, J2 = 7.2 Hz, 1H), δ 2.87 (d,d J1 =

11.6 Hz, J2 = 2.0 Hz, 1H), δ 2.77 (t, J = 5.8 Hz, 2H), δ 2.57-2.65 (m, 3H), δ 2.36

(d,t J1 = 12.0 Hz, J2 = 3.6 Hz, 1H), δ 2.09 (d,t J1 = 12.4 Hz, J2 = 3.6 Hz, 1H), δ

1.83 (q, J = 11.6 Hz, 1H), δ 1.74 (q, J = 12.4 Hz, 1H). MALDI TOF MS calculated

for C44H58N6O15S2: 974.34 found 975.42 [M+H]+, found 997.43 [M+Na]+ found

1013.41 [M+K]+.

E 6.7.6 Synthesis and Characterization of GuanidinoTobra-Fluorescein (70):

Guanidino Tobra-Fluorescein · TFA5 (70): Boc10 guanidino5 6" β-mercaptoethyl

ether tobramycin (Section 6.7.2) (10 mg, 5.6 µmoles), DMF (3 mL), Cs2CO3 (30

mg), and 5-iodo-acetamido-fluorescein (5-IAF) (5 mg, 9.7 µmoles, 1.7 equiv.,

Molecular Probes) were stirred at RT in the dark for 2h then diluted into ethyl

Guanidino Tobra-Fluorescein (70)

OHOHN

OHN

HN

O

ONHNH

OH

HO HO

NH2

NH

NH

H2N

NH2

NHHN

NH2

NHH2N

S

OS

O CO2H

HO

O

NH

O

(29%, 2 steps)

5-IAFOHOHN

OHN

HN

O

ONHNH

OH

HO HO

S

O

HS

Boc10Guanidino5-6"-β-Mercaptoethyl-ether Tobramycin

NH-Boc

N-Boc

N-BocBoc-HN

NH-Boc

NH-Boc

Boc-N

NH-Boc

N-BocBoc-HN

TFA

285

acetate (150 mL) and washed with 1M Na2CO3 (2x50 mL), 0.1M citric acid (2x50

mL), brine (50 mL), dried over sodium sulfate then concentrated under reduced

pressure to a solid. All of this product was then deprotected in CHCl3 (1 mL)

triisopropyl silane (0.15 mL), and TFA (3 mL) for 2.5 h at RT. Excess anhydrous

toluene was then added and all volatiles were removed at 50 °C under reduced

pressure. The reaction was then diluted into water (8 mL) and loaded onto an

activated C-18 reversed-phase cartridge (Waters, Sep-pack), the column was

then washed with pure water (10 mL), and the product eluted at 20%

acetonitrile/water, lyophilized, and found to be >95% pure (by HLPC). The final

product was purified further using a C-18 reversed phase HPLC column with an

isocratic mixture of 20% acetonitrile (0.1% TFA) in water (0.1% TFA) (3 mL/min)

(Rt = 11.5 min) to yield 2.3 mg (29%, 2 steps) of an orange solid. All fluorescein-

glycoside conjugates are slightly to moderately hygroscopic, therefore the

absorption at 496nm (in aqueous buffer pH 9.0) was used to confirm the yield of

the conjugation reaction (taking ε502 nm = 77,000 cm-1 M-1). 1H-NMR (400 MHz,

D2O) δ 8.17 (d, J = 2.0 Hz, 1H), δ 7.70 (d,d J1 = 10 Hz, J2 = 1.6 Hz, 1H), δ 7.15-

7.20 (m, 3H), δ 6.97 (d, J = 1.6 Hz, 2H), 6.79-6.80 (m, 2H), δ 5.19 (d, J = 3.2 Hz,

1H), δ 4.91 (d, J = 2.8 Hz, 1H), δ 3.95 (t, J = 6.8 Hz, 1H), δ 3.20-3.60 (m, 23H), δ

2.80 (d,d J1 = 11.2 Hz, J2 = 2.4 Hz, 1H), δ 2.77 (t, J = 5.8 Hz, 2H), δ 2.51-2.60 (m,

4H), δ 1.97-2.07 (m, 2H), δ 1.48-1.53 (m, 2H). MALDI TOF MS calculated for

C49H68N16O15S2: 1184.45 found 1185.63 [M+H]+, found 1207.60 [M+Na]+.

286

E 6.7.7 Synthesis and Characterization of Guanidino-Neo-Fluorescein (71):

Guanidino-Neo-Fluorescein · TFA6 (71): Boc12 guanidino6-5"-β-mercapto-

ethylether neomycin (Section 6.7.4) (3 mg, 1.4 µmoles), DMF (0.5 mL), 5-iodo-

acetamido-fluorescein (5-IAF) (5 mg, 9.7 µmoles, 1.7 equiv., Molecular Probes),

and triethyl amine (20 µL) were stirred at RT in the dark for 2h then diluted into

ethyl acetate (150 mL) and washed with 1M Na2CO3 (4x50 mL), 0.1M citric acid

(2x50 mL), brine (50 mL), dried over sodium sulfate then concentrated under

reduced pressure to a solid. All of this product was then deprotected in CHCl3 (3

mL) triisopropyl silane (0.05 mL), and TFA (5 mL) for 3.5 h at RT. Excess

anhydrous toluene was then added and all volatiles were removed at 50 °C

under reduced pressure. The reaction was then diluted into water (8 mL, 300 mM

NaCl) and loaded onto an activated C-18 reversed-phase cartridge (Waters, Sep-

pack), the column was then washed with pure water (10 mL), and the product

eluted between 5-20% acetonitrile/water (0.001 M HCl), lyophilized, and found to

be >85% pure (by HLPC). The product was purified further using a C-18 reversed

phase HPLC column with an isocratic mixture of 20% acetonitrile (0.1% TFA) in

Guanidino Neo-Fluorescein (71)Boc12Guanidino6-5"-β-Mercaptoethyl-ether Neomycin B

O

HN

HNO

HOHO

OO

NH

OH

NH

OHO

HOO

S

HO

HN

NH

OHS

NH-Boc

N-Boc

NH-Boc

N-Boc

N-BocBoc-HN

N-BocBoc-HN

N-Boc

Boc-HN

Boc-HN

N-Boc

(45%, 2 steps)

TFA

O

HN

HNO

HOHO

OO

NH

OH

NH

OHO

HOO

S

HO

HN

NH

OS

NH2

NH

NH2

NH

NHH2N

NHH2N

NHH2N

H2NNH

5-IAF

O CO2H

HO

O

NH

O

287

water (0.1% TFA) (3 mL/min) (Rt = 9.3 min) to yield 1.3 mg (45%, 2 steps) of an

orange solid. All fluorescein-glycoside conjugates are slightly to moderately

hygroscopic, therefore the absorption at 496nm (in aqueous buffer pH 9.0) was

used to confirm the yield of the conjugation reaction (taking ε502 nm = 77,000 cm-1

M-1). 1H-NMR (400 MHz, D2O): δ 8.12 (d, J = 2.0 Hz, 1H), δ 7.72 (d,d J1 = 8.8 Hz,

J2 = 1.6 Hz, 1H), δ 7.22 (d, J = 8.8 Hz, 1H), δ 7.11 (d, J = 9.2 Hz, 2H), δ 6.92 (d, J

= 2.4 Hz, 2H), δ 6.76 (d,d J1 = 8.8 Hz, J2 = 2.4 Hz, 1H), δ 5.69 (d, J = 4.0 Hz, 1H),

δ 4.98 (s, 1H), δ 4.87 (s, 1H), δ 4.18-4.24 (m, 2H), δ 3.85-3.88 (m, 3H), δ 3.16-

3.62 (m, 18H), δ 2.71-2.78 (m, 4H), δ 2.42-2.52 (m, 4H), δ 2.01 (d t, J1 = 12.0 Hz,

J2 = 4.4 Hz, 1H), δ 1.48 (q, J = 12.0 Hz, 1H). MALDI TOF MS calculated for

C55H79N19O19S2:1373.53, found 1374.72 [M+H]+.

E 6.8.1 Synthesis and Characterization of BODIPY-CR9 (72):

ace-CRRRRRRRRR-am · TFA9. Standard Fmoc solid-phase synthesis was

manually employed. Fmoc PAL PEG PS resin (1 g, 0.15 mmole, PerSeptive

Biosystems), was deprotected using 20% piperidine in DMF for 20 min at RT,

washed with DMF (3x7 mL), diethyl ether (2x7 mL), and DMF (3x7 mL), then

treated with TBTU (97 mg, 0.3 mmoles), Fmoc Arg (Pbf)-OH (195 mg, 0.3

mmoles), HOBt (46 mg, 0.3 mmoles), 2,4,6 collidine (0.4 mL, 3.0 mmoles), in

PEG NH

O

O

OCH3

OCH3

NH-Fmoc

piperidine/DMF

Fmoc-Arg(Pbf)-OH

HOBt/TBTU9

Fmoc PAL PEG PS Resin

HOBt/TBTU

Fmoc-Arg(Pbf)-OH

piperidine/DMF (Ac)2O

TFA aceCRRRRRRRRRam

288

DMF (7 mL), for at least 1 hr at RT on shaker. The resin is then washed, and the

deprotection and coupling processes were repeated nine times total, the final

coupling reaction utilized Fmoc Cys (Trt)-OH (264 mg, 0.45 mmoles), TBTU (145

mg, 0.45 mmoles), HOBt (46 mg, 0.45 mmoles), 2,4,6 collidine (0.6 mL, 4.5

mmoles), in DMF (7 mL) and lasted for 2 h at RT on a shaker. Following

deprotection and washes (as above), the N-terminus was acylated using HOBt

(80 mg, 0.78 mmoles), diisopropyl ethyl amine (0.9 mL), acetic anhydride (1.9

mL) in DMF (5mL) for 1 h at RT on a shaker. The resin was then washed with

DMF (3x7 mL), diethyl ether (2x7 mL), CHCl3 (4x7 mL) and deprotected with TFA

(9 mL) in the presence of triisopropy silane (400 µL, 2 moles) and 1,2-

ethanedithiol (0.2 mL, 6.4 mmoles) for 2.5 h at RT on a shaker. The solution was

then drained into 1% acetic acid/water (180 mL) and washed with CHCl3 (3x80

mL) and diethyl ether (3x80 mL). The aqueous layer was then concentrated to a

solid and lyophilized from 0.1% TFA in water. The crude peptide was purified

using a 9% acetonitrile/water (0.1% TFA) isocratic mixture on a C-18 reversed

phase HPLC column (3 mL/min retention time 10-12 min). and lyophilized to

afford a white solid (40 mg, 10%). MALDI TOF MS calculated for C59H118N38O115:

1566.96, found 1567.82 [M+H]+.

289

BODIPY-CR9 · HCl (72). The purified peptide ace-CRRRRRRRRR-am · TFA9 (10

mg, 3.86 umoles) was dissolved in an aqueous degassed buffer (1 mL of 100

mM NaCl, 10 mM phosphate pH 7.5, Ar sparged) and treated (dropwise) with a

solution of BODIPY C1-IA (1.3 mg, 3.1 µmoles, 0.81 equiv., Molecular Probes)

dissolved in DMSO (1.25 mL) and allowed to react in the dark for 1h at RT, then

diluted with 8 mL of water and loaded onto an activated C-18 reversed-phase

cartridge (Waters, Sep-pack), the column was then washed with 5% acetonitrile

(5 mL containing 100 mM NaCl in water), 1mL pure water, the product eluted

between 0 – 20% acetonitrile (in water) and lyophilized to yield 3.8 mg (56%

yield) of a red powder. All BODIPY-glycoside conjugates are slightly to

moderately hygroscopic, therefore the absorption at 502nm (in methanol) of each

compound was used to calculate the yield of the conjugation reaction (taking ε502

nm = 76,000 cm-1 M-1). 1H-NMR (400 MHz, D2O) δ 7.43 (s, 1H), δ 6.91 (d, J = 3.6

Hz, 1H), δ 6.29 (d, J = 3.6 Hz, 1H), δ 6.25 (s, 1H), δ 4.53 (s, 2H), δ 4.32 (t, J = 7.2

Hz, 1H), δ 4.13-4.23 (m, 10H), δ 3.34 (s, 2H), δ 2.96-3.08 (m, 18H), δ 2.86 (d, J =

6.8 Hz, 2H), δ 2.42 (s, 3H), δ 2.17 (s, 3H), δ 1.89 (s, 3H), δ 1.50-1.71 (m, 36H).

The 1H NMR suggests better than 95 % purity. MALDI TOF MS calculated for

C73H132BF2N41O12S: 1856, found 1857 [M+H]+.

NNB

FFI

NH

OaceCRRRRRRRRRam aceCRRRRRRRRRam

NNB

FF NH

OS

DMSO, water (pH 7.5)(56%)

BODIPY-CR9 (72)

290

E 6.8.2 Synthesis and Characterization of Fluorescein-CR9 (73):

Fluorescein-CR9 · TFA9 (73). The purified peptide ace-CRRRRRRRRR-am ·

TFA9 (10 mg, 3.86 µmoles) was dissolved in a an aqueous degassed buffer (1

mL of 100 mM NaCl, 10 mM phosphate pH 7.5, Ar sparged) and treated

(dropwise) with a solution of BODIPY C1-IA (1.3 mg, 3.1 µmoles, 0.81 equiv.,

Molecular Probes) pre-dissolved in DMSO (1.25 mL) and allowed to react in the

dark for 1h at RT. 0.1M HCl was then added drop-wise until the solution turned

from orange to yellow. The reaction was diluted into water (8 mL) and loaded

onto an activated C-18 reversed-phase cartridge (Waters, Sep-pack), the column

was washed with pure water (10 mL), and the product eluted with 30%

acetonitrile/water, lyophilized, and found to be >85% pure by HLPC. The final

product was purified to >98% purity using a C-18 reversed phase HPLC column

with an isocratic mixture of 20% acetonitrile (0.1% TFA) in water (0.1% TFA) (3

mL/min) (Rt = 6.5 min) to yield 4.2 mg (37%) of an orange solid. All fluorescein-

glycoside conjugates are slightly to moderately hygroscopic, therefore the

absorption at 496nm (in aqueous buffer pH 9.0) of each compound is used to

confirm the yield of the conjugation reaction (taking ε502 nm = 77,000 cm-1 M-1).

I

O CO2H

HO

O

NH

OaceCRRRRRRRRRam aceCRRRRRRRRRam

DMSO, water (pH 7.5)(37%) Fluorescein-CR9 (73)

S

O CO2H

HO

O

NH

O

291

1H-NMR (400 MHz, D2O) is consistent with the desired structure, but the

expected number of aromatic peaks are multiplied, suggesting partial protonation

of fluorescein. MALDI TOF MS calculated for C81H131N39O17S: 1954.03,

observed 1955.18 [M+H]+.

E 6.8.3 Synthesis and Characterization of Amino Neo-pyridyl disulfide (74):

Boc6 5'' ββββ-mercaptoethyl ether pyridyl disulfide neomycin B. Synthesized as

described (Hai Wang Ph. D. Thesis, University of California, San Diego, 1998).

Amino Neo-pyridyl disulfide · TFA6 (74). Boc6 5'' β-mercaptoethyl ether pyridyl

disulfide neomycin B (80 mg, 55 µmoles), CH2Cl2 (5 mL), triisopropyl silane (120

µL), TFA (6 mL) were reacted at RT for 15 min then partitioned into 0.1% acetic

acid (80 mL in water) and diethyl ether (50 mL). The aqueous layer was then

washed with diethyl ether (40 mL), CH2Cl2 (40 mL), CHCl3 (40 mL), ethyl acetate

(20 mL), and diethyl ether (40 mL). The aqueous layer was then concentrated to

a solid, under reduced pressure. The product was lyophilized from 0.01% TFA in

O

NHBoc

BocHNO

HOHO

OO

BocHN

OH

NHBoc

HOO

HOO

HO

NHBoc

BocNH

Boc6 5'' β-Mercaptoethyl-ether pyridyl disulfideNeomycin B

SO

S

S

NO

NH2

H2NO

HOHO

OO

H2N

OH

NH2

HOO

HOO

HO

NH2

NH2

SO

S

S

N

Amino Neo-pyridyl disulfide (74)

TFA

(76%)

292

water (2 mL) to yield 64 mg (76%) of a white solid. 1H-NMR (400 MHz, d6-

MeOD): 8.43 (d, J = 4.0 Hz, 1H), δ 7.84-7.93 (m, 2H), δ 7.24 (d,d,d J1 = 7.2 Hz,

J2 = 4.4 Hz, J3 = 1.2 Hz, 1H), δ 6.03 (d, J = 4.0 Hz, 1H), δ 5.44 (d, J = 4.0 Hz,

1H), δ 5.34 (d, J = 1.6 Hz, 1H), δ 4.42 (t, J = 3.6 Hz, 1H), δ 4.38 (t, J = 3.6 Hz,

1H), δ 4.27-4.34 (m, 2H), δ 4.13-4.19 (m, 2H), δ 4.03 (d,d J1 = 10.4 Hz, J2 = 8.8

Hz, 1H), δ 3.93 (d,t J1 = 9.2 Hz, J2 = 3.2 Hz, 1H), δ 3.86 (t, J = 8.8 Hz, 1H), δ 3.74

(t, J = 6.4 Hz, 2H), δ 3.55-3.62 (m, 3H), δ 3.58 (d,d J1 = 10.0 Hz, J2 = 9.8 Hz, 1H),

δ 3.20-3.50 (m, 8H), δ 3.08-3.16 (m, 2H), δ 3.04 (t, J = 6.0 Hz, 2H), δ 2.77-2.99

(m, 3H), δ 2.45 (d, t J1 = 12.4 Hz, J2 = 4.4 Hz, 1H), δ 2.09 (q, J = 12.8 Hz, 1H).

ESI MS calculated for C32H57N7O13S3: 843.3, found 844.2 [M+H]+, found 866.2

[M+Na]+, found 882.2 [M+K]+.

E 6.8.4 Synthesis and Characterization of Guanidino Neo-pyridyl disulfide(75):

Guanidino Neo-pyridyl disulfide · TFA6 (75). Amino Neo-pyridyl disulfide ·

TFA6 (Section E 6.8.3) (13 mg, 8.8 µmoles), methanol (1 mL), 1,2-dixoane (3

mL), N,N′-di-Boc-N′′ -trifylguanidine (75 mg, 0.19 mmoles, 21 equiv.), and triethyl

O

NH2

H2NO

HOHO

OO

H2N

OH

NH2

HOO

HOO

HO

NH2

NH2

SO

S

S

N

Amino Neo-pyridyl disulfide (74)

N-Tf

NH-BocBoc-HN

(20%)

TFA

O

HN

HNO

HOHO

OO

NH

OH

NH

OHO

HOO

HO

HN

NH

NH2

NH

NH2

NH

NHH2N

NHH2N

NH

H2N

H2NNH

SO

S

S

N

Guanidino Neo-pyridyl disulfide (75)

293

amine (100 µl, 0.72 mmoles, 80 eqiv.) were stirred together for 96 h at RT under

Ar. All volatiles were then removed and the oil was treated with CHCl3 (1 mL),

N,N′-di-Boc-N′′ -trifylguanidine (15 mg, 0.038 mmoles, 4 equiv.), and triethyl

amine (60 µL, 0.43 mmoles, 49 eqiv.) for 24 h at RT under Ar. All volatiles were

then removed under reduced pressure and the resulting oil was dissolved CHCl3

(100 mL) and washed with 1 M NaH2PO4 (40 mL), water (40 mL), and brine (40

mL). The organic phase was then dried over sodium sulfate, concentrated to a

solid, and purified using silica gel (40 mL) and flash chromatography with a 0 –

3% methanol/CH2Cl2 gradient to yield ~5 mg of a white solid. Rf = 0.2 (2.5%

methanol/CH2Cl2). All of this product was deprotected by adding CHCl3 (1 mL),

TFA (2 mL) and stirring at RT for 2 hr. The reaction was then diluted into

anhydrous toluene and concentrated to a solid under reduced pressure (at 45

°C). HLPC was used to purify the final product using a C18 reversed phase semi-

prep column at 3.5 mL/min and an isocratic mixture of 12% acetonitrile in water

(0.1% TFA), fractions were collected (Rt = 10.2 min) and lyophilized to yield 3 mg

of a white solid (20%). 1H-NMR (400 MHz, d6-MeOD): δ 8.41 (d, J = 4.4 Hz), δ

7.81-7.89 (m, 2H), δ 7.24 (d,d,d J1 = 7.2 Hz, J2 = 4.8 Hz, J3 = 1.2 Hz, 1H), δ 6.06

(d, J = 3.2 Hz, 1H), δ 5.15 (d, J = 2.8 Hz, 1H), δ 5.07 (d, J = 1.2 Hz, 1H), δ 4.30-

4.32 (m, 2H), δ 4.05-4.11 (m, 2H), δ 3.99 (t, J = 3.4 Hz, 1H), δ 3.71-3.80 (m,

5H), δ 3.41-3.63 (m, 12H), δ 3.03 (t, J = 6.2 Hz, 2H), δ 2.95 (d,d J1 = 13.2 Hz, J2

= 4.4 Hz, 1H), δ 2.67-2.76 (m, 3H), δ 2.12 (d, t J1 = 12.4 Hz, J2 = 4.4 Hz, 1H), δ

1.72 (q, J = 12.4 Hz, 1H). ESI MS calculated for C38H69N19O13S3: 1095.45, found

1096.41 [M+H]+.

294

References and Footnotes

1. For historical accounts regarding the evolution of modern molecularbiology, see (a) F. H. C . Crick, What Mad Pursuit, Nasic Books, 1988. (b)H. F. Judson, The Eigth Day of Creation, Simon & Schuster Pubs., 1980.

2. A. Kornberg, T. Baker, DNA Replication, W.H. Freeman and Co, 1992.

3. E. H. Blackburn, Telomerase RNA structure and function. In: RNAStructure and Function, Cold Spring Harbor Laboratory Press, 1998, 669-693.

4. T. W. Nilsen, RNA-RNA Interactions in nuclear pre-mRNA splicing. In:RNA Structure and Function, Cold Spring Harbor Laboratory Press, 1998,279-307.

5. J. B. Hartford and T. A. Rouault, RNA structure and function in cellular ironhomeostasis. In: RNA Structure and Function, Cold Spring HarborLaboratory Press, 1998, 575-602.

6. B. N. Fields, D. M. Knipe, P. M. Howley (Eds.), Fundamental Virology,Lippincott-Raven Pubs., 1996.

7. R. Green, H. F. Noller, Ribosomes and translation. Annu. Rev. Biochem.1997, 66, 679-716.

8. H. F. Noller, V. Hoffarth, L. Zimniak, Unusual resistance of peptidyltransferase to protein extraction procedures. Science, 1992, 256, 1416-1419.

9. T. R. Cech, The ribosome is a ribozyme. Science, 2000, 289, 878-879.

10. (a) P. Nissen, J. Hansen, N. Ban, P. B. Moore, T. A. Steitz, The structuralbasis of ribosome activity in peptide bond synthesis. Science, 2000, 289,920-930. (b) G. W. Muth, L. Ortoleva-Donnelly, S. A. Strobel, A singleadenosine with a neutral pKa in the ribosomal peptidyl transferase center.Science, 2000, 289, 947-950.

11. C. G. Kurland, Translational accuracy and the fitness of bacteria. Annu.Rev. Genet. 1992, 26, 29-50.

12. (a) H. Siomi, G. Dreyfuss, RNA-binding proteins as regulators of geneexpression. Curr. Opin. Gen. Dev. 1997, 7, 345-353. (b) G. Varani, K.Nagai, RNA recognition by RNP proteins during RNA processing. Annu.Rev. Biophys. Biomol. Struct. 1998, 27, 407-445. (c) N. K. Gray, M.

295

Wickens, Control of translation initiation in animals. Annu. Rev. Cell. Dev.Biol. 1998, 14, 399-458.

13. T. R. Cech and B. L. Golden, Building a catalytic active site using onlyRNA. In: The RNA World, Cold Spring Harbor Laboratory Press, 1999,321-349.

14. M. W. Hentze, L. C. Kuhn, Molecular control of vertebrate iron metabolism- mRNA-based regulatory circuits operated by iron, nitric oxide, andoxidative stress. Proc. Natl. Acad. Sci. USA, 1996, 93, 8175-8182.

15. N. C. Lau, L. P. Lim, E. G. Weinstein, D. P. Bartel, An abundant class oftiny RNAs with probable regulatory roles in Caenorhabditis elegans.Science, 2001, 294, 858-862.

16. (a) G. Werstuck, M. R. Green, Controlling gene expression in living cellsthrough small molecule-RNA interactions. Science, 1998, 282, 296-298.(b) W. Winkler, A. Nahvi. R. R. Breaker RR. Thiamine derivatives bindmessenger RNAs directly to regulate bacterial gene expression. Nature,2002, 419, 952-956. (c) A. Nahvi, N. Sudarsan, M. S. Ebert, X. Zou, K. L.Brown, R. R. Breaker, Genetic control by a metabolite binding mRNA.Chem. Biol. 2002, 9, 1043-1049.

17. International Human Genome Sequencing Consortium, Initial sequencingand analysis of the human genome. Nature, 2001, 409, 860-921.

18. R. F. Gesteland, T. R. Cech, J. F. Atkins (Eds.), The RNA World, ColdSpring Harbor Laboratory Press, 1999.

19. J. Davies, U. Ahsen, R. Schroeder, Antibiotics and the RNA world: A rolefor low-molecular-weight effectors in biochemical evolution? The RNAWorld, Cold Spring Harbor Laboratory Press, 1993, 185-204.

20. Aptamers are RNA sequences that are selected from large libraries usinga pre-determined small molecule. Through multiple rounds of mutation,selection, and amplification, the evolution of an RNA sequence for apredetermined small molecule can be effected. Selected sequences canexhibit extremely high affinity and specificity for their cognate smallmolecule.

21. R. T. Batey, J. A. Doudna, The parallel universe of RNA folding. Nat.Struct. Biol. 1998, 5, 337-40.

22. (a) R. Russell, D. Herschlag, Probing the folding landscape of theTetrahymena ribozyme: Commitment to form the native conformation islate in the folding pathway. J. Mol. Biol. 2001, 308, 839-851. (b) S. K.

296

Silverman, M. L. Deras, S. A. Woodson, S. A. Scaringe, T. R. Cech,Multiple folding pathways for the P4-P6 RNA domain. Biochemistry, 2000,39, 12465-75. (c) M. S. Rook, D. K. Treiber, J. R. Williamson, An optimalMg2+ concentration for kinetic folding of the Tetrahymena ribozyme. Proc.Natl. Acad. Sci. U S A, 1999, 96, 12471-6.

23. (a) M. Zuker, On finding all suboptimal foldings of an RNA molecule.Science, 1989, 244 48-52. (b) C. Gaspin, E. Westhof, An interactiveframework for RNA secondary structure prediction with a dynamicaltreatment of constraints. J. Mol. Biol. 1995 229, 1049-1064. (c) A. R.Banerjee, J. A. Jaeger, D. H. Turner, Thermal unfolding of a Group-Iribozyme – the low-temperature transition is primarily disruption of tertiarystructure. Biochemistry, 1993, 32, 153-163.

24. M. Wu, I. Tinoco, Jr., RNA folding causes secondary structurerearrangement. Proc. Natl. Acad. Sci. USA, 1998, 95, 11555-60.

25. (a) A. L. Feig, O. C. Uhlenbeck, The role of metal ions in RNAbiochemistry. In: The RNA World, Cold Spring Harbor Laboratory Press,1999, 287-319. (b) V. K. Misra, D. E. Draper, On the role of magnesiumions in RNA stability. Biopolymers, 1998, 48,113-135.

26. (a) T. Hermann, E. Westhof, Exploration of metal ion binding sites in RNAfolds by Brownian-dynamics simulations. Structure, 1998, 6, 1303-14. (b)V. K. Misra, D. E. Draper, A thermodynamic framework for Mg2+ bindingto RNA. Proc. Natl. Acad. Sci. USA, 2001, 98, 12456-12461.

27. M. M. Teeter, G. J. Quigley, A. Rich. In Nucleic Acid – Metal IonInteractions; T. G. Spiro Ed. Wiley-Interscience: New York, 1980, 146-177.

28. (a) D. J. Earnshaw, M. J. Gait, Aminoglycosides and cleavage of thehairpin ribozyme. In: RNA-Binding Antibiotics, Molecular BiologyIntelligence Unit 13, Eurekah.com, 2001, 35-55. (b) B. M. Chowrira, A.Berzal-Herranz, J. M. Burke, Ionic requirements for RNA binding,cleavage and ligation by the hairpin ribozyme. Biochemistry, 1993,32,1088-95. (c) F. Walter, Q. Vicens, E. Westhof, Aminoglycoside-RNAinteractions. Curr. Opin. Chem. Biol. 1999, 3, 694-704.

29. A. D. Dahlberg, F. Horodyski, P. Keller, Interaction of neomycin withribosomes and ribosomal ribonucleic acid. Antimicrob. Agents Chemother.1978, 13, 331-39.

30. D. Moazed, H. F. Noller, Interaction of antibiotics with functional sites in 16S ribosomal RNA. Nature, 1987, 327, 389-394.

297

31. (a) T. Pape, W. Wintermeyer, M. V. Rodnina, Conformational switch in thedecoding region of 16S rRNA during aminoacyl-tRNA selection on theribosome. Nature Struct. Biol., 2000, 7, 104-107. (b) R. Karimi, M.Ehrenberg, Dissociation rate of cognate peptidyl tRNA from the A-site ofhyper-accurate and error-prone ribosomes. Eur. J. Biochem. 1994, 226,355-360.

32. (a) J. Davies, B. D. Davis, Misreading of ribonucleic acid code wordsinduced by aminoglycoside antibiotics. J. Biol. Chem. 1968, 243, 3312-3316. (b) S. R. Lynch, J. D. Puglisi. Structural origins of aminoglycosidespecificity for prokaryotic ribosomes. J. Mol. Biol. 2001, 306, 1037-1058.

33. A. P. Carter, W. M. Clemons, D. E. Brodersen, R. J. Morgan-Warren, B. T.Wimberly, V. Ramakrishnan, Functional insights from the structure of the30S ribosomal subunit and its interactions with antibiotics. Nature, 2000,407, 340 – 348.

34. R. E. Botto, B. Coxon, Nitrogen-15 nuclear magnetic resonancespectroscopy of neomycin B and related aminoglycosides. J. Am. Chem.Soc. 1983, 105, 1021-1028.

35. M. Hendrix, P. B. Alper, E. S. Priestley, C-H. Wong, Hydroxyamines as anew motif for the molecular recognition of phosphodiesters: implicationsfor aminoglycoside-RNA interactions. Angew. Chem. Intl. Ed. 1997, 36,95-98.

36. H. Wang and Y. Tor, Electrostatic interactions in RNA aminoglycosidesbinding. J. Am. Chem. Soc. 1997, 119, 8734-8735.

37. (a) A. Whelton, H. C. Neu (Eds.), The Aminoglycosides: Microbiology,Clinical Use, and Toxicology, Marcel Dekker, Inc., 1982. (b) T. Koeda, K.Umemura, M. Yokota, Toxicology and pharmacology of aminoglycosideantibiotics. In: Aminoglycoside Antibiotics, Springer-Verlag, 1982, 293-356.

38. For membrane-related toxicology issues, see: N. Tanaka, Mechanism ofaction of aminoglycoside antibiotics. In: Aminoglycoside Antibiotics.Handbook of Experimental Pharmacology. Vol 62, Springer-Verlag, 1982,221-266.

39. K. Michael, Y. Tor, Designing novel RNA binders. Chem. Eur. J. 1998, 4,2091-98.

40. J. R. Williamson, Induced fit in RNA-protein recognition. Nat. Struct. Biol.2000, 7, 834-837.

298

41. For reviews on RNA-small molecule recognition, see: (a) W. D. Wilson, K.Li, Targeting RNA with small molecules. Curr. Med. Chem., 2000, 7, 73-98. (b) S. J. Sucheck, C-H. Wong, RNA as a target for small molecules.Curr. Opin. Chem. Biol. 2000, 4, 678-686. (c) A. C. Cheng, V. Calabro, A.D. Frankel, Design of RNA-binding proteins and ligands. Curr. Opin.Struct. Biol. 2001, 11, 478-484.

42. (a) A. D. Frankel, J. A. T. Young, HIV-1: Fifteen proteins and an RNA.Annu. Rev. Biochem. 1998, 67, 1-25. (b) M. J. Gait, J. Karn, RNArecognition by the human immunodeficiency virus Tat and Rev proteins.Trends Biochem. Sci. 1993, 18, 255-259.

43. For a review on the Rev-RRE interaction see: (a) V. W. Pollard, M. H.Malim, The HIV-1 Rev protein. Annu. Rev. Microbiol. 1998, 52, 491-532.

44. (a) M. H. Malim, L. S. Tiley, D. F. McCarn, J. R. Rusche, J. Hauber, B. R.Cullen, HIV-1 structural gene expression requires binding of the Revtrans-activator to its RNA-target sequence. Cell, 1990, 60, 675-683. (b) T.J. Daly, K. S. Cook, G. S. Gray, T. E. Maione, J. R. Rusche, Specificbinding of HIV-1 recombinant Rev protein to the Rev-response element invitro. Nature, 1989, 342, 816-819. (c) C. A. Rosen, E. Terwilliger, A.Dayton, J. G. Sodroski, W. A. Haseltine, Intragenic cis-acting art gene-responsive sequences of the human immunodeficiency virus sequences ifthe human immunodeficiency virus. Proc. Natl. Acad. Sci. USA, 1988, 85,2071-2075. (d) T. J. Hope, D. McDonald, X. Huang, J. Low, T. G. Parslow,Minimal Rev-response element for type 1 human immunodeficiency virus.J. Virol. 1991, 65, 2131-2134. (e) M. L. Zapp, T. J. Hope, T. G. Parslow,M. R. Green, Oligomerization and RNA binding domains of the type 1human immunodeficiency virus Rev protein: a dual function for anarginine-rich binding motif. Proc. Natl. Acad. Sci. USA, 1991, 88, 7734-7738.

45. T. J. Hope, The ins and outs of HIV Rev. Arch. Biochem. Biophys. 1999,365, 186-191.

46. R. Tan, L. Chen, J. A. Buettner, D. Hudson, A. D. Frankel, RNArecognition by an isolated alpha-helix. Cell, 1993, 73, 1031-1040.

47. D. A. Mann, I. Mikaelian, R. W. Zemmel, S. M. Green, A. D. Lowe, T.Kimura, M. Singh, P. J. Butler, M. J. Gait, J. Karn, A molecular rheostat –cooperative binding to stem-1 of the Rev-response element modulateshuman immunodeficiency virus type-1 late gene expression. J. Mol. Biol.1994, 241, 193-207.

299

48. H. Schaal, M. Klein, P. Gehrmann, O. Adams, A. Scheid, Requirement ofN-terminal amino acid residues of gp41 for human immunodeficiency virustype 1-mediates cell fusion. J. Virol. 1995, 69, 3308-3314.

49. R. C. Gallo, HIV- The cause of AIDS – An overview of its biology,mechanisms of disease induction, and our attempts to control it. J. Acquir.Immune Defic. Syndr. 1988, 1, 521-535.

50. J-H. Chen, S-Y Le, J. V. Maizel, Prediction of common secondarystructures of RNAs: a genetic algorithm approach. Nucleic Acids Res.2000, 28, 991-999.

51. B. Charpentier, F. Stutz, M. Rosbash, A Dynamic in Vivo View of theHIV-1 Rev-RRE Interaction. J. Mol. Biol. 1997, 266, 950-962.

52. J. Kjems, B. J. Calnan, A. D. Frankel, P. A. Sharp, Specific Binding of aBasic Peptide from HIV-1 Rev. EMBO J. 1992, 11, 1119-1129.

53. J. L. Battiste, H. Mao, N. S. Rao, R. Tan, D. R. Muhandiram, L. E. Kay, A.D. Frankel, J. R. Williamson, Alpha helix-RNA major groove recognition inan HIV-1 rev peptide-RRE RNA complex. Science, 1996, 273, 1547-1551.

54. K. Li, M. Fernandez-Saiz, C. T. Rigl, A. Kumar, K. G. Ragunathan, A. W.McConnaughie, D. W. Boykin, H. J. Schneider, W. D. Wilson. Design andanalysis of molecular motifs for specific recognition of RNA. Bioorg. Med.Chem. 1997, 5, 1157-72.

55. K. M. Weeks, C. Ampe, S. C. Schultz, T. A. Steitz, D. M. Crothers,Fragments of the HIV-1 Tat Protein Specifically Bind TAR RNA. Science,1990, 249, 1281-1285.

56. (a) E. Kikuta, S. Aoki, E. Kimura, J. Am. Chem. Soc. 2001, 123, 7911-7912. (b) A. Litovchick, A. G. Evdokimov, A. Lapidot, Biochemistry, 2000,39, 2838-2852; (c) N. Tamilarasu, I. Huq, T. M. Rana, Bioorg. Med. Chem.Lett. 2000, 10, 971-974; (d) N. Gelus, C. Bailly, F. Hamy, T. Klimkait, W.D. Wilson, D. W. Boykin, Bioorg. Med. Chem. 1999, 7, 1089-1096; (e) H-Y. Mei, M. Cui, A. Heldsinger, S. M. Lemrow, J. A. Loo, K. A. Sannes-Lowery, L. Sharmeen, A. W. Czarnik, Biochemistry, 1998, 37, 14204-14212. (f) F. Hamy, V. Brondani, A. Floersheimer, W. Stark, M. Blommers,T. Klimkait, Biochemistry 1998, 37, 5086-5095. (g) S. Hwang, N.Tamilarasu, K. Ryan, I. Huq, S. Richter, W. C. Still, T. M. Rana, Proc. Natl.Acad. Sci. USA, 1999, 23, 12997-13002.

57. J. A. Ippolito, T. A. Steitz, A 1.3 Å resolution crystal structure of the HIV-1trans-activation response region RNA stem reveals a metal ion-dependentbulge conformation. Proc. Natl. Acad. Sci. USA, 1998, 95, 9819-9824.

300

58. J. D. Puglisi, R. Tan, B. J. Calnan, A. D. Frankel, J. R. Williamson,Conformation of the TAR RNA-Arginine Complex by NMR Spectroscopy.Science, 1992, 257,76-80.

59. J. A. Ippolito, T. A. Steitz, The structure of the HIV-1 RRE high affinity revbinding site at 1.6 Å resolution. J. Mol. Biol. 2000, 295, 711-717.

60. L. W. Hung, E. L. Holbrook, S. R. Holbrook SR. The crystal structure ofthe Rev binding element of HIV-1 reveals novel base pairing andconformational variability. Proc. Natl. Acad. Sci. USA, 2000, 97, 5107-5112.

61. N. K. Tanner, Ribozymes: the characteristics and properties of catalyticRNAs. FEMS Microbiology Rev. 1999, 23, 257-275.

62. M. Chastain, I. Tinoco Jr. Structural elements in RNA. Prog. Nucl. AcidRes. 1991, 41, 131-177.

63. O. C. Uhlenbeck, A small catalytic oligoribonucleotide. Nature, 1987, 328,596-600.

64. N. K. Vaish, A. R. Kore, F. Eckstein, Recent developments in thehammerhead ribozyme field. Nucl. Acids Res. 1998, 26, 5237-5242.

65. T. K. Stage, K. J. Hertel and O. C. Uhlenbeck, Inhibition of thehammerhead ribozyme by neomycin. RNA, 1995, 1, 95-101.

66. S. R. Kirk, N. W. Luedtke, Y. Tor, 2-Aminopurine as a real-time probe ofenzymatic cleavage and inhibition of hammerhead ribozymes. Bioorg. &Med. Chem. 2001, 9, 2295-2301.

67. Y. Tor, T. Hermann, E. Westhof, Deciphering RNA Recognition:Aminoglycoside Binding to the Hammerhead Ribozyme. Chem. Biol. 1998,5, R277-R283.

68. D. Fourmy, M. I. Recht, S. C. Blanchard, J. D. Puglisi. Structure of the Asite of Escherichia coli 16S ribosomal RNA complexed with anaminoglycoside antibiotic. Science, 1996, 274, 1367-1371.

69. (a) W. A. Greenberg, E. S. Priestley, P. S. Sears, P. B. Alper, C.Rosenbohm, M. Hendrix, S.-C. Hung, and C.-H. Wong. Design andSynthesis of New Amino Glycoside Antibiotics Containing Neamine as anOptimal Core Structure: Correlation of Antibiotic Activity with in VitroInhibition of Translation. J. Am. Chem. Soc. 1999, 121, 6527-6541. Seealso the supporting information of: (b) S. J. Sucheck, A. L. Wong, K. M.

301

Koeller, D. D. Boehr, K.-a. Draker, P. Sears, G. D. Wright, C.-H. Wong.Design of Bifunctional Antibiotics that Target Bacterial rRNA and InhibitResistance-Causing Enzymes. J. Am Chem. Soc. 2000, 122, 5230-5231.

70. S. R. Kirk, Y. Tor. tRNAPhe Binds Aminoglycoside Antibiotics. Bioor. Med.Chem. 1999, 7, 1979-1991.

72. (a) K. Hamasaki, R. R. Rando, A high-throughput fluorescence screen tomonitor the specific binding of antagonists to RNA targets. AnalyticalBiochemistry, 1998, 261, 183-190. (b) D. L. Boger, B. E. Fink, S. R.Brunette, W. C. Tse, M. P. Hedrick, A simple, high-resolution method forestablishing DNA binding affinity and sequence selectivity. J. Am. Chem.Soc. 2001, 123, 5878-91. (c) M. Auer, J-M. Seifert, S. Wallace, R. Sleigh,RNA in a miniaturized lead discovery process. In: RNA-BindingAntiobiotics, Molecular Biology Intelligence Unit 13, Eurekah.com, 2001,164-176.

73. Topics in Fluorescence Spectroscopy, Vol. 2: Principals. J. R. LakowiczEd., Plenum, New York, 1991.

74. K. L. Bentley, L. K. Thompson, R. J. Klebe, P. M. Horowitz, Fluorescencepolarization: A general method for measuring ligand binding andmembrane microviscosity. BioTechniques, 1985, 3, 356-365.

75. A. J. Pope, U. M. Haupts, K. J. Moore, Homogeneous fluorescencereadouts for miniaturized high-throughput screening: theory and practice.Drug Discovery Today, 1999, 4, 350-362.

76. (a) Y. Wang, K. Hamasaki, R. R. Rando, Specificity of aminoglycosidebinding to RNA constructs derived from the 16S rRNA decoding regionand the HIV-RRE activator region. Biochemistry, 1997, 36, 768-779.

77. D. I. Van Ryk, S. Venkatesan, Real-time kinetics of HIV-1 Rev-RevResponse Element Interactions. J. Biol. Chem. 1999, 274, 17452-17463.

78. C. Ma, N. A. Baker, S. Joseph, J. A. McCammon, Binding ofaminoglycoside antibiotics to the small ribosomal subunit: a continuumelectrostatics investigation. J. Am. Chem. Soc. 2002, 124, 1438-1442.

79. B. J. Calnan, S. Biancalana, D. Hudson, A. D. Frankel, Analysis ofarginine-rich peptides from the HIV Tat protein reveals unusual features ofRNA protein recognition. Genes Dev. 1991, 5, 201-210.

80. M. L. Zapp, T. J. Hope, T. G. Parslow, M. R. Green, Oligomerization andRNA binding of the type-1 Human Immunodeficiency Virus Rev protein - a

302

dual function for an arginine-rich binding motif. Proc. Natl. Acad. Sci. USA,1991, 88, 7734-7738.

81. N. W. Luedtke, Y. Tor, A novel solid-phase assembly for identifying potentand selective RNA ligands. Angew. Chem. Intl. Ed. 2000, 39, 1788-1790.

82. M. L. Zapp, S. Stern, M. R. Green, Small molecules that selectively blockRNA binding of HIV-1 Rev protein inhibit Rev function and viral production.Cell, 1993, 74, 969-978.

83. (a) K. Li, G. Xiao, T. Rigl, A. Kumar, D. W. Boykin, W. D. Wilson, InStructure, Motion, Interaction and Expression of BiologicalMacromolecules (Eds.; R.H. Sarma, M.H. Sarma), Adenine Press, 1998,137-145. (b) G. Xiao, A. Kumar, K. Li, C. T. Rigl, M. Bajic, T. M. Davis, D.W. Boykin, W. D. Wilson, Inhibition of the HIV-1 Rev-RRE complexformation by unfused aromatic cations. Bioorg. Med. Chem. 2001, 9,1097-1113.

84. K. Li, T. M. Davis, C. Bailly, A. Kumar, D. W. Boykin, W. D. Wilson, Aheterocyclic inhibitor of the Rev-RRE complex binds to RRE as a dimer.Biochemistry, 2001, 40, 1150-1158.

85. N.W. Luedtke, T. J. Baker, M. Goodman, Y. Tor, Guanidinoglycosides: Anovel family of RNA ligands. J. Am. Chem. Soc. 2000, 122, 12035-12036.

86. J-B. LePecq, C. J. Paoletti, A fluorescent complex between ethidiumbromide and nucleic acids. Physical-chemical characterization. J. Mol.Biol. 1967, 27, 87-106.

87. J. L. Bresloff, D. M. Crothers, Equilibrium studies of ethidium-polynucleotide interactions. Biochemistry, 1981, 20, 3547-3553.

88. D. D. Perrin, Dissociation Constants of Organic Bases in AqueousSolution, Butterworths: London, 1965.

89. K. A. Lacourciere, J. T. Strivers, J, P. Marino, Mechanism of neomycinand Rev binding to the Rev responsive element of HIV-1 as determined byfluorescence and NMR spectroscopy. Biochemistry, 2000, 39, 5630-5641.

90. W. D. Wilson, L. Ratmeyer, M. Zhao, L. Strekowski, D. Boykin, The searchfor structure-specific nucleic acid interactive drugs - effects of compoundstructure on RNA versus DNA interaction strength. Biochemistry, 1993,32, 4098-4104.

91. For reviews see: (a) D. P. Millar, Fluorescence studies of DNA and RNAstructure and dynamics. Curr. Opin. Struct. Biol. 1996, 6, 322-326. (b) D.

303

M. Jameson, J. F. Eccleston, Fluorescent Nucleotide Analogs: Synthesisand Applications. Methods Enzymology, 1997, 278, 363-390.

92. D. C. Ward, E. Reich, L. Stryer. Fluorescence studies of nucleotides andpolynucleotides. I. Formycin, 2-aminopurine riboside, 2,6-diaminopurineriboside, and their derivatives. J. Biol. Chem. 1969, 10, 1228-1237.

93. T. M. Nordlund, S. Andersson, L. Nilsson, R. Rigler, Strucutre andDynamics of a fluorescent DNA oligomer containing the EcoRI recognitionsequence: fluorescence, molecular dynamics and NMR studies.Biochemistry, 1989, 28, 9095-9103.

94. M. Rist, J. Marino. Association of an RNA kissing complex analyzed using2-aminopurine fluorescence. Nucleic Acids Res. 2001, 29, 2401-2408.

95. D. A. Harris, D. Rueda, N. G. Walter, Local conformational changes in thecatalytic core of the trans-acting hepatitis delta virus ribozyme accompanycatalysis. Biochemistry, 2002, 41, 12051-12061.

96. M. Menger, T. Tuschl, F. Eckstein, D. Porschke, Mg2+-dependentconformational changes in the hammerhead ribozyme. Biochemistry,1996, 35, 14710-14716.

97. S.R. Kirk, N.W. Luedtke, Y. Tor, 2-aminopurine as a real-time probe ofenzymatic cleavage and inhibition of hammerhead ribozymes. Bioorg.Med. Chem., 2001, 9, 2295-2301.

98. I. R. Klotz, Ligand-Receptor Energetics, John Wiley & Sons, Inc. NewYork, 1997.

99. T. Heyduk T, J. C. Lee, Application of fluorescence energy transfer andpolarization to monitor Escherichia coli cAMP receptor protein and lacpromoter interaction. Proc. Natl. Acad. Sci. U S A, 1990, 87, 1744-1748.

100. F. W. Sevenich, J. Langowski, V. Weiss, K. Rippe, DNA binding andoligomerization of NtrC studied by fluorescence anisotropy andfluorescence correlation spectroscopy. Nucleic Acids Res. 1998, 26, 1373-1381.

101. A. Rudi, Y. Benayahu, I. Goldberg, Y. Kashman, Tetrahedron Lett. 1988,29, 6655-6656.

102. N. R. Shochet, A. Rudi, Y. Kashman, H. Yaacov, M. R. El-Maghrabi, I.Spector, Novel marine alkaloids from the tunicate Eudistoma sp. arepotent regulators of cellular growth and differentiation and affect cAMP-mediated processes. J. Cell Physiol. 1993, 157, 481-492.

304

103. M. Einat, M. Lishner, A. Amiel, A. Nagler, S. Yarkorli, A. Rudi, Y.Kashman, D. Markel, I. Fabian, Eilatin: a novel marine alkaloid inhibits invitro proliferation of progenitor cells in chronic myeloid leukemia patients.Exp. Hematol. 1995, 23, 1439-1444.

104. L. A. McDonald, G. S. Eldredge, L. R. Barrows, C. M. Ireland, Inhibition oftopoisomerase II catalytic activity by pyridoacridine alkaloids from aCystodytes sp. ascidian: a mechanism for the apparent intercalator-induced inhibition of topoisomerase II. J. Med. Chem. 1994, 37, 3819-3827.

105. A. Rudi, Y. Kashman, D. Gut, F. Lellouche, M. Kol, . Rutheniumcomplexes of eilatin: Face selectivity in octahedral geometry; Synthesis of[Ru(bpy)2(eilatin)]2+ and [Ru(phen)2(eilatin)]2+. Chem. Commun. 1997,1,17-18.

106. D. Gut, A. Rudi, J. Kopilov, I. Goldberg, M. Kol, Pairing of propellers:dimerization of octahedral ruthenium(II) and osmium(II) complexes ofeilatin via pi-pi stacking featuring heterochiral recognition. J. Am. Chem.Soc. 2002, 124, 5449-5456.

107. E. C. Glazer, Y. Tor, Ru(II) complexes of “Large Surface” Ligands, Angew.Chem. Intl. Ed. 2002, 41, 4022-4026.

108. B. Bosnich, The application of exiton theory to the determination of theabsolute configurations of inorganic complexes. Acc. Chem. Res. 1969, 2,266-273.

109. A. Sylwester, S. Murphy, D. Shutt, D. R. Soll HIV-induced T cell syncytiaare self-perpetuation and the primary cause of T cell death in culture. J.Immunol. 1997, 158, 3996-4007.

110. D. D. Richman, V. A. Johnson, D. L. Mayers, T. Shirasaka, M. C. O'Brien,H. Mitsuya, In Current Protocols in Immunology; J. E. Coligan, A. M.Kruisbeek, D. H. Margulies, E. M. Shevach, W. Strober, Eds.; John Wiley& Sons: Brooklyn, NY, 1993; Suppl 8, Unit 12.9, pp 1-21.

111. (a) O. Novakova, J. Kasparkova, O. Vrana, P. M. van Vliet, J. Reedijk, V.Brabec, Correlation between cytotoxicity and DNA binding of polypyridylruthenium complexes. Biochemistry, 1995, 34, 12369-12378. (b) L.Mishra, R. Sinha, H. Itokawa, K. F. Bastow, Y. Tachibana, Y .Nakanishi,N. Kilgore, K. H. Lee, Anti-HIV and cytotoxic activities of Ru(II)/Ru(III)polypyridyl complexes containing 2,6-(2'-benzimidazolyl)-pyridine/chalcone as co-ligand. Bioorg. Med. Chem. 2001, 9, 1667-1671.

305

112. C. V. Kumar, J. K. Barton, N. J. Turro, Photophysical Properties ofRuthenium Complexes Bound to Double Helical DNA. J. Am. Chem. Soc.1985, 107, 5518-5523.

113. Estimated twist angle is based upon the crystal structure an analogous un-fused system [Ru(bpy)2bisisoquinoline]2+, see M. T. Ashby, G. N.Govindan, A. K. Grafton, Metal-assisted racemization of the atropisomersof a 1,1'- binaphthyl skeleton via a syn transition state. J. Am. Chem. Soc.1994, 116, 4801-4809.

114. Some metal complexes containing non-planar ligands have been shown tobind to DNA. For examples, see (a) R. J. Morgan, S. Chatterjee, A. D.Baker, T. C. Strekas, Effects of ligand planarity and peripheral charge onintercalative binding of Ru(2,2' –bipyridine)2L

2+ to calf thymus DNA. Inorg.Chem. 1991, 30, 2687-2692. (b) Y. Xiong, X-F. He, X-H. Zou, J-Z. Wu, X-M. Chen, L-N. Ji, R-H. Li, J-Y. Zhou, K-B. Yu, Interaction of polypyridylruthenium(II) complexes containing non-planar ligands with DNA. J.Chem. Soc. Dalton Trans., 1999, 19-24.

115. [7] (a) K. E. Erkkila, D. T. Odom, J. K. Barton, Recognition and reaction ofmetallointercalators with DNA. Chem. Rev. 1999, 99, 2777-2796. (b) J. K.Barton, A. T. Danishefsky, J. M. Goldberg, Tris(phenanthroline) Ru(II):Stereoselectivity in Binding to DNA. J. Am. Chem. Soc. 1984, 106, 2172-2176. (c) I. Haq, P. Lincoln, D. Suh, B. Norden, B. Z. Chowdhry, J. B.Chaires, Interaction of delta- and lambda- [Ru(phen)2 DPPZ] with DNA – Acalorimetric and equilibrium binding study. J. Am. Chem. Soc. 1995, 117,4788-4796; (d) J. G. Collins, J. R. Aldrich-Wright, I. D. Greguric, P. A.Pellegrini, Binding of the delta- and lambda- enantiomers of[Ru(dmphen)2+ to the hexanucleotide d(GTCGAC)2. Inorg. Chem. 1999,38, 5502-5509. (e) C. V. Kumar, J. K. Barton, N. J. Turro, Photophysics ofruthenium complexes bound to double helical DNA. J. Am. Chem. Soc.1985, 107, 5518-5523. (f) L-N. Ji, X-H. Zou, J-G Liu, Shape- andenantioselective interaction of Ru(II)/Co(III) polypyridyl complexes withDNA Coord. Chem. Rev. 2001, 216, 513-536. (g) C. Chow, J. K. Barton,Shaper-selective cleavage of tRNAPhe by transition-metal complexes. J.Am. Chem. Soc. 1990, 112, 2839-2841. (h) A. C. Lim, J. K. Barton,Targeting the Tat-binding site of bovine immunodeficiency virus TAR RNAwith a shape-selective rhodium complex. Bioorg. Med. Chem. 1997, 6,1131-1136. (i) S. Satyanarayana, J. C. Dabrowiak, J. B. Chaires, Neitherdelta- nor lambda tris(phenanthroline) ruthenium(II) binds to DNA byclassical intercalation. Biochemistry, 1992, 31, 9319-9324. (j) D. Z.Coggan, I. S. Haworth, P. J. Bates, A. Robinson, A. Rodger, DNA Bindingof Ruthenium Tris(1,10-phenanthroline): Evidence for the Dependence ofBinding Mode on Metal Complex Concentration. Inorg. Chem. 1999, 38,4486-4497. (k) J. G. Liu, Q. L. Zhang, X. F. Shi, L. N. Ji, Interaction of[Ru(dmp)2(dppz)]2+ and [Ru(dmb)2(dppz)]2+ with DNA: effects of the

306

ancillary ligands on the DNA-binding behaviors. Inorg Chem. 2001, 40,5045-5050.

116. L. S. Lerman, Structural considerations in the interaction of DNA andacridines. J. Mol. Biol. 1961, 3, 18-30.

117. G. Cohen, H. Eisenberg, Viscosity and sedimentation study of sonicatedDNA-proflavin complexes. Biopolymers, 1969, 8, 45-55.

118. E. Jin, V. Katritch, W. K. Olson, M. Kharatisvili, R. Abagyan, D. S. Pilch,Aminoglycoside binding in the major groove of duplex RNA: thethermodynamic and electrostatic forces that govern recognition. J. Mol.Biol. 2000, 298, 95-110.

119. J. K. Barton, S. J. Lippard, Cooperative binding of a platinummetallointercalation reagent to poly(A).poly(U). Biochemistry, 1979, 18,2661-2668.

120. D. H. Turner, In Nucleic Acids Structures Properties and Functions, V. A.Bloomfield, D. M. Crothers, I. Tinoco Jr. (Eds.), University Science Books,2000, 259-334.

121. I. Piantanida, B. S. Palm, M. Zinic, H-J, Schneider, A new 4,9-diazapyrenium intercalator for single- and double-stranded nucleic acids;distinct differences from related diazapyrenium compounds and ethidiumbromide. J. Chem. Soc., Perkin Trans. 2, 2001, 1808-1816.

122. G. R. Jones, M. E. Oliveira, R. B. Cundall, The interactions ofpyrenylmethyl tributylphosphonium bromide with single strandpolynucleotides. Photochem. Photobiol. 1990, 52, 451-460.

123. J. L. Battiste, R. Tan, A. D. Frankel, J. R. Williamson, Binding of an HIVRev peptide to Rev responsive element RNA induces formation of purine-purine base pairs. Biochemistry, 1994, 33, 2741-2747.

124. T. I. Watkins, G. Woolfe, Nature, 1952, 169, 506.

125. T. I. Watkins, Trypanocides of the phenanthridine series. Part I. The effectof changing the quaternary grouping in dimidium bromide. J. Chem. Soc.1952, 3059-3064.

126. J. Sambrook, E. F. Fritsch, T. Maniatis, In: Molecular Cloning A LaboratoryManual Second Edition, Cold Spring Harbor Press, 1989.

127. (a) H. Nishiwaki, M. Miura, K. Imai, R. Ohno, K. Kawashima,

307

Experimental studies on the antitumor effect of ethidium bromide andrelated substances. Cancer Res. 1974, 34, 2699-2703 (b) B. -R. Balda, G.D. Birkmayer, Drug effects in melanoma: tumor specific interactions ofproflavine and ethidium bromide. Yale J. Biol. Med. 1973, 46, 464-470.

128. C. L. Courchesne, J. A. Bantle, Analysis of the activity of DNA, RNA, andprotein synthesis inhibitors on Xenopus embryo development. Teratog.Carcinog. Mutagen. 1985, 5, 177-193.

129. (a) J. R. MacGregor, I. J. Johnson, In vitro metabolic activation of ethidiumbromide and other phenanthridinium compounds: mutagenic activity inSalmonella typhimurium. Mutation Res. 1977, 48, 103-108. (b) I. E.Mattern, Mutagenicity of ethidium bromide after metabolic activation invitro, Mutation Res. 1976, 38, 120.

130. http://www.laprovet.fr/Trypanosomoses.html

131. M. M. K. Nass, Differential effects of ethidium bromide on mitochondrialand nuclear DNA synthesis in vivo in cultured mammalian cells.Experimental Cell Research, 1972, 72, 211-222.

132. A. Peña, E. Chávez, A. Cárabez, M. T. De Gómez-Puyou, M.T. Themetabolic effects and uptake of ethidium bromide by rat liver mitochondria.Arch. Biochem. Biophys. 1977, 180, 522-529.

133. Anon. Material Safety Data Sheet (MSDS): Ethidium bromide. MDLInformation Systems Inc., San Leandro, CA (on-line retrieval through STNdatabase: MSDS-OHS) 1994.

134. J. Warring, Complex formation between ethidium bromide and nucleicacids. J. Mol. Biol. 1965, 13, 269-282.

135. J. B. LePecq, C. Paoletti, A fluorescent complex between ethidiumbromide and nucleic acids. Physical-chemical characterization. J. Mol.Biol. 1967, 27, 87-106.

136. M. Liebman, J. Rubin, M. Sundaralingam, Nonintercalative binding ofethidium bromide to nucleic acids: Crystal structure of an ethidium-tRNAmolecular complex. Proc. Natl. Acad. Sci. USA, 1977, 74, 4821-4825.

137. W. C. Chu, J. C. Liu, J. Horowitz J. Localization of the major ethidiumbromide binding site on tRNA. Nucleic Acids Res. 1997, 25, 3944-3949.

138. (a) C. C. Tsai, S. C. Jain, H. M. Sobell, Visualization of drug-nucleic acidinteractions at atomic resolution. I. Structure of an ethidium/dinucleosidemonophosphate crystalline complex, ethidium:5-iodouridylyl (3'-5')

308

adenosine. J. Mol. Biol. 1977, 114, 301-315. (b) S. C. Jain, C. C. Tsai, H.M. Sobell HM. Related Articles, Visualization of drug-nucleic acidinteractions at atomic resolution. II. Structure of an ethidium/dinucleosidemonophosphate crystalline complex, ethidium:5-iodocytidylyl (3'-5')guanosine. J. Mol. Biol. 1977, 114, 317-331.

139. W. D. Wilson, L. Ratmeyer, M. T. Cegla, J. Spychala, D. Boykin, M.Demeunynck, J. Lhomme, G. Krishnan, D. Kennedy, R. Vinayak, G. Zon,Bulged-base nucleic acids as potential targets for antiviral drug action.New. J. Chem. 1994, 18, 419-423.

140. H. Jacquemin-Sablon, M. Le Bret, A. Jacquemin-Sablon, C. Paoletti,Yeast mitochondrial deoxyribonuclease stimulated by ethidium bromide. 2.Mechanism of enzyme activation. Biochemistry, 1979, 18, 128-134.

141. W. J. Firth, C. L. Watkins, D. E. Graves, David E.; L. W. Yielding,Synthesis and characterization of ethidium analogs: emphasis on aminoand azido substituents. J. Heterocycl. Chem. 1983, 20, 759-765.

142. L. P. Wakelin, M. J. Waring, The unwinding of circular deoxyribonucleicacid by phenanthridinium drugs: structure-activity relations for theintercalation reaction. Mol. Pharmacol. 1974, 10, 544-561.

143. D. E. Graves, C. L. Watkins, L. W. Yielding, Ethidium bromide and itsphotoreactive analogues: spectroscopic analysis of deoxyribonucleic acidbinding properties. Biochemistry, 1981, 20, 1887-1892.

144. L. W. Yielding, K. L. Yielding, J. E. Donoghue, Ethidium Binding toDeoxyribonucleic Acid: Spectrophotometric Analysis of Analogs withAmino, Azido, and Hydrogen Substituents. Biopolymers, 1984, 23, 83-110.

145. J. Loccufier, E. Schacht, Regioselective acetylation of 3,8-diamino 5-ethyl6-phenyl-phenanthridinium bromide, preparation of potential newtrypanocides. Tetrahedron, 1989, 45, 3385-3396.

146. V. Peytou, R. Condom, N. Patino, R. Guedj, A. M. Aubertin, N. Gelus, C.Bailly, R. Terreux, D. Cabrol-Bass, Synthesis and antiviral activity ofethidium-arginine conjugates directed against the TAR RNA of HIV-1. J.Med. Chem. 1999, 42, 4042-4053.

147. M. Le Bret, O. Chalvet, Optical Spectrum of ethidium intercalated intoDNA a CNDO/S study. J. Mol. Struct. 1977, 37, 299-319.

148. C. Medhi, J. B. Mitchell, S. L. Price, A. B. Tabor. Electrostatic factors inDNA intercalation. Biopolymers, 1999, 52, 84-93.

309

149. D. Reha, M. Kabelac, F. Ryjacek, J. Sponer, J. E. Sponer, M. Elstner, S.Suhai, P. Hobza P. Intercalators. 1. Nature of stacking interactionsbetween intercalators (ethidium, daunomycin, ellipticine, and 4',6-diaminide-2-phenylindole) and DNA base pairs. Ab initio quantumchemical, density functional theory, and empirical potential study. J. Am.Chem. Soc. 2002, 124, 3366-3376.

150. S. E. Patterson, J. M. Coxon, L. Strekowski L. Intercalation of ethidiumand analogues with nucleic acids: a molecular orbital study. Bioorg. Med.Chem. 1997, 5, 277-281.

151. C. A. Hunter, K. R. Lawson, J. Perkins, C. J. Urch, Aromatic interactions.J. Chem. Soc. Perkin Trans. 2, 2001, 651-669.

152. D. A. Bondarev, W. J. Skawinski, C. A. Venanzi, Nature of intercalatoramiloride-nucelobase stacking. An empirical potential and ab initioelectron correlation study. J. Phys. Chem. 2000, 104, 815-822.

153. J. Sponer, J. Leszczynski, P. Hobza, Electronic properties, hydrogenbonding, stacking, and cation binding of DNA and RNA bases.Biopolymers, 2001, 61, 3-31.

154. F. Cozzi, M. Cinquini, R. Annuziata, J. S. Siegel, Dominance of polar/πover charge-transfer effects in stacked phenyl interactions. J. Am. Chem.Soc. 1993, 115, 5330-5331.

155. F. Cozzi, M. Cinquini, R. Annuziata, T. Dwyer, J. S. Siegel, Polar/πinteractions between stacked aryls in 1,8-diarylnaphthalenes. J. Am.Chem. Soc. 1992, 114, 5729-5733.

156. P. U. Giacomoni, M. Le Bret, Electronic structure of ethidium bromide.FEBS Lett. 1973, 29, 227-230.

157. G. P. Kreishman, S. I. Chan, W. Bauer, Proton magnetic resonance studyof the interaction of ethidium bromide with several uracil residues, uridylyl(3' leads to 5') uridine and polyuridylic acid. J. Mol. Biol. 1971, 61, 45-58.

158. G. Dougherty, A comparison of the base-pair specificities of threephenanthridine drugs using solution spectroscopy. Int. J. Biochem.1982,14, 493-504.

159. I. Tinoco Jr., K. Sauer, J. C. Wang, In Physical Chemistry, Principals andApplications in Biological Sciences (third edition), Prentice Hall, NewJersey, 1995.

310

160, All attempts to obtain NMR data for an ethidium analog that lacks the 6-phenyl ring have so far proven fruitless.

161. B. G. Griggs, M. W. Davidson, W. D. Wilson, D. W. Boykin, Assignment ofthe 13C chemical shifts of ethidium bromide and analogs in D2O solution.Org. Mag. Reson. 1980, 14, 371-373.

162. (a) H. Spiesecke, W. G. Schneider, the determination of π-electrondensities in azulene from C13 and H1 nuclear resonance shifts.Tetrahedron Lett. 1961, 14, 468-472. (b) G. A. Olah, G. P. Mateescu,Stable carbonium ions. The Tetraphenylcyclobutadiene dication. J. Am.Chem. Soc. 1970, 92, 1430-1432.

163. W. A. Brett, P. Rademacher, R. Boese, Redetermination of the structure ofphenanthridine. Acta. Cryst. 1993, C49, 1564-1566.

164. R. Kiralk, B. Kojic-Prodic, M. Zinic, S. Alihodzic, N. Trinajstic, Bond length-bond order relations and calculated geometries for some bezenoidaromatics, including phenanthridine. Structures of 5,6-dimethylphenanathridinium triflate, [N-(6-phenanthridinylmethyl)-aza-18-crown-6-κ5 O,O’,O’’, O’’’, O’’’’](picrate-κ2O,O’)potassium, and [N,N’-bis96-phenanthridinyl-κN-methyl)-7,16-diaza-18-crown-6-κ4 O,O’,O’’, O’’’]sodiumiodine dichloromethane solvate. Acta. Cryst. 1996, B52, 823-837.

165. N. W. Luedtke, Q. Liu, Y. Tor, The ground state electronic structure ofethidium reveals charge separation in an organic cation. In preparation

166. (a) M. Hospital, B. Busetta, C. R. Acad. Sci. Paris, 1969, 268, 1232-1235.(b) P. C. Courseille, B. Busetta, M. Hospital, Acta Cryst. 1974, B30, 2631-2639.

167. E. Subramanian, J. Trotter, C. E. Bugg, Crystal structure of ethidiumbromide. J. Cryst. Mol. Struct. 1971, 1, 3-15.

168. I. Zimmerman, H. W. Zimmerman, pKa values of ethidium bromide and 7-amino-9-phenyl-10-ethyl-phenanthridinium bromide. Z. Naturforsch. 1976,31, 656.

169. S. R. Kirk, N. W. Luedtke, Y. Tor, Neomycin-acridine conjugate: A potentinhibitor of Rev-RRE binding. J. Am. Chem. Soc., 2000, 122, 980-981.

170. W. D. Wilson, L. Ratmeyer, M. Zhao, L. Strekowski, D. W. Boykin, Thesearch for structure-specific nucleic acid-interactive drugs: effects ofcompound structure on RNA versus DNA interaction strength.Biochemistry, 1993, 32, 4098-4104.

311

171. R. H. Griffey, S. A. Hofstadler, K. A. Sannes-Lowery, D. J. Ecker, S. T.Crooke, Determinants of aminoglycoside-binding specificity for rRNA byusing mass spectrometry. Proc. Natl. Acad. Sci. USA, 1999, 96, 10129-10133.

172. S. J. Sucheck, W. A. Greenberg, T. J. Tolbert, C. -H. Wong. Design ofsmall molecules that recognize RNA: Development of aminoglycosides aspotential antitumor agents that target oncogenic RNA sequences. Angew.Chem. Int. Ed. 2000, 39, 1080-1084.

173. F. Walter, Q. Vicens, E. Westhof, Aminoglycoside-RNA interactions.Curr. Opin. Chem. Biol. 1999, 3, 694-704.

174. Y. Tor, T. Hermann, E. Westhof, Deciphering RNA recognition:aminoglycoside binding to the hammerhead ribozyme. Chem Biol. 1998,5, R277-R283.

175. M. Hendrix, E. S. Priestley, G. F. Joyce, C. -H Wong, Direct observation ofaminoglycoside-RNA interactions by surface plasmon resonance. J. Am.Chem. Soc. 1997, 119, 3641-3648.

176. Q. Chen, R. H. Shafer, I. D. Kuntz, Structure-based discovery of ligandstargeted to the RNA double helix. Biochemistry 1997, 36, 11402-11407.

177. S. R. Kirk, N. W. Luedtke, Y. Tor, Neomycin-acridine conjugate: A potentinhibitor of Rev-RRE binding. J. Am. Chem. Soc. 2000, 122, 980-981.

178. J. B. H. Tok, J. H. Cho, R. R. Rando, Aminoglycoside hybrids as potentRNA antagonists. Tetrahedron, 1999, 55, 5741-5758.

179. K. Hamasaki, A. Ueno, Aminoglycoside antibiotics, neamine and itsderivatives as potent inhibitors for the RNA-protein interactions derivedfrom HIV-1 activators. Bioorg. Med. Chem. Lett. 2001, 26, 591-594.

180. H. Wang, Y. Tor, Dimeric aminoglycosides: Design, synthesis and RNAbinding. Bioorg. Med. Chem. Lett. 1997, 7, 1951-1956.

181. J. B. Tok, G. R. Huffman, Enhanced binding of aminoglycoside dimers to a"dimerized" A-site 16S rRNA construct. Bioorg. Med. Chem. Lett. 2000,10, 1593-1595.

182. K. Michael, H. Wang, Y. Tor, Enhanced RNA binding of dimerizedaminoglycosides. Bioorg. Med. Chem. 1999, 7, 1361-1371.

312

183. J. B. Tok, L. J. Dunn, R. C. Des Jean, Binding of dimeric aminoglycosidesto the HIV-1 rev responsive element (RRE) RNA construct. Bioorg. Med.Chem. Lett. 2001, 11, 1127-1131.

184. (a) M. A. Weiss, N. Narayana, RNA recognition by arginine-rich peptidemotifs. Biopolymers, 1998, 48, 167-180. (b) R. N. De Guzman, R. B.Turner RB, M. F. Summers, Protein-RNA recognition. Biopolymers, 1998,48, 181-195.

185. R. Tan, A. D. Frankel, Costabilization of peptide and RNA structure in anHIV Rev peptide-RRE complex. Biochemistry, 1994, 33, 14579-14585.

186. N.W. Luedtke, T. J. Baker, M. Goodman, Y. Tor, Guanidinoglycosides: Anovel family of RNA ligands. J. Am. Chem. Soc. 2000, 122, 12035-12036.

187. T. J. Baker, N.W. Luedtke, Y. Tor, M. Goodman, Synthesis and anti-HIVactivity of guanidinoglycosides. J. Org. Chem. 2000, 65, 9054-9058.

188. Q. Liu, N.W. Luedtke, Y. Tor, A simple conversion of amines intomonosubstituted ureas in organic and aqueous solvents. Tetrahedron Lett.2001, 42, 1445-1447.

189. A. Litovchick, A. G. Evdokimov, A. Lapidot, Aminoglycoside-arginineconjugates that bind TAR RNA: Synthesis, characterization, and antiviralactivity. Biochemistry, 2000, 39, 2838-2852.

190. E. R. Jamieson, S. J. Lippard, Structure, recognition, and processing ofCisplatin-DNA adducts. Chem. Rev. 1999, 99, 2467–2498.

191. (a) A. M. J. Fichtinger-Schepman, J. L. van der Veer, J. H. J. den Hartog,P. H. M. Lohman, J. Reedijk, Adducts of the antitumor drug cis-diamminedichloroplatinum(II) with DNA: formation, identification, andquantitation. Biochemistry, 1985, 24, 707–713. (b) A. Eastman,Reevaluation of interaction of cis-dichloro(ethylenediamine)platinum(II)with DNA. Biochemistry, 1986, 25, 3912–3915.

192, (a) H. Huang, J. Woo, S. C. Alley, P. B. Hopkins, DNA-DNA interstrandcross-linking by cis-diamminedichloroplatinum(II): N7(dG)-to-N7(dG)cross-linking at 5'-d(GC) in synthetic oligonucleotides. Bioorg. Med. Chem.1995, 3, 659–669. (b) H Huang, L. Zhu, B. R. Reid, G. P. Drobny, P. B.Hopkins, Solution structure of a cisplatin-induced DNA interstrand cross-link. Science, 1995, 270 (5243), 842-845.

193. R. N. Bose, Biomolecular targets for platinum antitumor drugs. Mini. Rev.Med. Chem. 2002, 2, 103-111.

313

194. G. Speelmans, R. W. Staffhorst, K. Versluis, J. Reedijk, B. de Kruijff,Cisplatin complexes with phosphatidylserine in membranes. Biochemistry,1997, 36, 10545-10550.

195. T. Ishikawa, F. Ali-Osman, Glutathione-associated cis-diamminedichloro-platinum(II) metabolism and ATP-dependent efflux from leukemia cells.Molecular characterization of glutathione-platinum complex and itsbiological significance. J. Biol. Chem. 1993, 268, 20116-20125.

196. Selected examples: (a) B. E. Bowler, S. J. Lippard, Modulation of platinumantitumor drug binding to DNA by linked and free intercalators.Biochemistry, 1986, 25, 3031-3038. (b) B. D. Palmer, H. H. Lee, P.Johnson, B. C. Baguley, G. Wickham, L. P.G. Wakelin, W. D. McFadyen,W. A. Denny, DNA-directed alkylating agents. 2. Synthesis and biologicalactivity of platinum complexes linked to 9-anilinoacridine. J. Med. Chem.1990, 33, 3008–3014. (c) D. Gibson, K. F. Gean, J. Katzhendle, R. Ben-Shoshan, A. Ramu, I. Ringel, Preparation, characterization, andanticancer activity of a series of cis-PtCl2 complexes linked toanthraquinone intercalators. J. Med. Chem. 1991, 34, 414–420. (d) J.Whittaker, W. D. McFadyen, G. Wickham, L. P. Wakelin, V. Murray, Theinteraction of DNA-targeted platinum phenanthridinium complexes withDNA. Nucl. Acids. Res. 1998, 26, 3933–3939. (e) S. K. Sharma, L. W.McLaughlin, Cross-linking of a DNA conjugate tethering a cis-bifunctionalplatinated complex to a target DNA duplex. J. Am. Chem. Soc. 2002, 124,9658-9659.

197. J. Boer, N. W. Luedtke, Y. Tor, RNA-selective covalent modification byNeo-platin (A neomycin-cisplatin conjugate). J. Am. Chem. Soc.submitted.

198. For a recent review, see: S. Futaki, Arginine-rich peptides: potential forintracellular delivery of macromolecules and the mystery of thetranslocation mechanisms. Int. J. Pharm. 2002, 245, 1-7.

199. A. Ho, S. R. Schwarze, S. J. Mermelstein, G. Waksman, S. F. Dowdy.Synthetic protein transduction domains: enhanced transduction potentialin vitro and in vivo. Cancer Res. 2001, 61, 474-477.

200. D. J. Mitchell, D. T. Kim, L. Steinman, C. G. Fathman, J. B. Rothbard,Polyarginine enters cells more efficiently than other polycationichomopolymers. J. Pept. Res. 2000, 56, 318-325.

201. P. Belmont A. Aissaoui, M. Hauchecorne, N. Oudrhiri, L. Petit, J. P.Vigneron, J. M. Lehn, P. Lehn, Aminoglycoside-derived cationic lipids asefficient vectors for gene transfection in vitro and in vivo. J. Gene. Med.2002, 4, 517-526.

314

202. For a recent review describing the potential applications see: L. Bonetta,Getting proteins into cells, The Scientist, 2002, 16, 38-40.

203. O. C. Uhlenbeck, J. F. Milligan, Synthesis of small RNAs using T7 RNApolymerase. Methods Enzymol. 1989, 180, 51-62.