Electrical Detection of DNA Binding Proteins Citation Gorodetsky, Alon A. (2009) Electrical Detection of DNA Binding Proteins. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/VPKN-CV38. https://resolver.caltech.edu/CaltechETD:etd-12102008-101354 Abstract The base pair stack of double helical DNA has proven to be an effective medium for charge transport. The π-stacked DNA base pairs can mediate charge transport (CT) chemistry over distances as long as 20 nm, and the reaction is exquisitely sensitive to DNA sequence-dependent conformation and dynamics. This sensitivity to perturbations in DNA structure and base pair stacking makes DNA-mediated charge transport chemistry an ideal methodology for the electrical detection of base mismatches, lesions, and protein binding. Efforts toward expanding the scope of electrochemistry at DNA-modified surfaces for biosensing applications are presented here. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: DNA; electrochemistry; monolayers; proteins; self-assembly Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Barton, Jacqueline K. Thesis Committee: Heath, James R. (chair) Gray, Harry B. Barton, Jacqueline K. Lewis, Nathan Saul Defense Date: 1 December 2008 Record Number: CaltechETD:etd-12102008-101354 Persistent URL: https://resolver.caltech.edu/CaltechETD:etd-12102008-101354 DOI: 10.7907/VPKN-CV38 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 5261 Collection: CaltechTHESIS Deposited By: Imported from ETD-db Deposited On: 03 Jun 2009 Last Modified: 08 Nov 2023 00:14 Thesis Files Preview PDF (Acknowledgements.pdf) - Final Version See Usage Policy. 95kB Preview PDF (Chapter1.pdf) - Final Version See Usage Policy. 404kB Preview PDF (Chapter2.pdf) - Final Version See Usage Policy. 640kB Preview PDF (Chapter3.pdf) - Final Version See Usage Policy. 454kB Preview PDF (Chapter4.pdf) - Final Version See Usage Policy. 431kB Preview PDF (Chapter5.pdf) - Final Version See Usage Policy. 521kB Preview PDF (Chapter6.pdf) - Final Version See Usage Policy. 417kB Preview PDF (Chapter7.pdf) - Final Version See Usage Policy. 443kB Preview PDF (Chapter8.pdf) - Final Version See Usage Policy. 594kB Preview PDF (Chapter9.pdf) - Final Version See Usage Policy. 407kB Preview PDF (Chapter10.pdf) - Final Version See Usage Policy. 795kB Repository Staff Only: item control page

ELECTRICAL DETECTION OF DNA BINDING PROTEINS Thesis by Alon A. Gorodetsky In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 2009 (Defended December 1, 2008) ii © 2009 Alon A. Gorodetsky All Rights Reserved iii ACKNOWLEDGEMENTS First and foremost, I would like to thank Jackie for her constant encouragement and contagious enthusiasm. As an advisor, Jackie is incredibly supportive and generous with her time. Her open door policy makes it easy to walk into her office and discuss anything, virtually at any time. I have been fortunate enough to receive an amazing amount of personal attention from Jackie, for which I am very grateful. I feel that my future career as a scientist will be greatly influenced by Jackie’s mentorship during my PhD. I would also like to thank Maureen Renta for all of her help and support. Maureen ensures that all of the day-to-day affairs of the group run smoothly, thereby making everyone’s life easier. In the Barton group, the students are truly lucky to have her help throughout their graduate/postdoctoral careers. Without Maureen, I believe it would have taken me a much longer time to graduate. I want to give a great deal of credit to Mike Hill at Occidental College. Without Mike, none of the DNA electrochemistry experiments described in this thesis would be possible. He has been an amazing resource, and I have often sought out his advice. I urge future generations of graduate students to do the same. During my time at Caltech, I have been fortunate enough to benefit from the assistance of some truly wonderful colleagues. I would like to thank Donato Ceres for training when I entered the Barton group. Amie Boal provided many stocks of Endonuclease III for a variety of experiments. Omar Green and Eylon Yavin helped hammer out some excellent strategies for DNA synthesis. Lars Dietrich was invaluable iv during SoxR electrochemistry experiments, which were also performed by Paul Lee on gold. Ali Ebrahim, a wonderful undergraduate with a bright future ahead of him, was immensely helpful in the two years he worked for me. William Hammond and Krzysztof Slowinski were essential for the successful completion of scanning electrochemical microscopy experiments. Natalie Muren is a wonderful colleague, and it has been gratifying to watch her develop as a graduate student. Russell Ernst was helpful in my cell lysate experiments. Jason Slinker is a friend from Cornell whose contributions to Nile Blue electrochemistry and multiplexing have been absolutely essential; Jason’s work will certainly impact the Barton group for many years. This list is far from complete, and I have benefited from the experience and advice of many members of the Barton Group. I would also like to thank my undergraduate advisor at Cornell University, George Malliaras. The freedom that I was afforded in the Malliaras group was unprecedented for an undergraduate researcher. I have great memories of the Malliaras group, where many of my scientific skills were initially forged. I want to thank the many amazing friends from Cornell with whom I have maintained contact. My experience as an undergraduate was truly unique, and I feel blessed to have attended Cornell University. I underwent a tremendous amount of personal growth at Cornell, and I will always look back fondly upon my time there. Finally, I want thank my parents, Alex and Stella, and my sister, Dana, for their constant encouragement and support. My current career is a testament to their strength. I also want to thank my wonderful girlfriend Irina for always being there for me and making me a better person. v Overall, my experience at Caltech has been very interesting. I have made some great friends here, and, on a professional level, I have undergone tremendous growth. I can honestly say that Caltech is a great place to work if science is your only motivation. vi ABSTRACT The base pair stack of double helical DNA has proven to be an effective medium for charge transport. The π-stacked DNA base pairs can mediate charge transport (CT) chemistry over distances as long as 20 nm, and the reaction is exquisitely sensitive to DNA sequence-dependent conformation and dynamics. This sensitivity to perturbations in DNA structure and base pair stacking makes DNA-mediated charge transport chemistry an ideal methodology for the electrical detection of base mismatches, lesions, and protein binding. Efforts toward expanding the scope of electrochemistry at DNAmodified surfaces for biosensing applications are presented here. The first conclusive study of the single-molecule conductivity of DNA is demonstrated. A pristine single DNA duplex in its native conformation can be utilized to covalently bridge an oxidatively etched, single-walled carbon nanotube gap. The effective resistance is on the order of ~ 1 MΩ, which is similar to the c-axis resistance of kish graphite. Furthermore, the inclusion of either the CA or GT mismatches within the DNA duplex is seen to increase the resistance of DNA by over two orders of magnitude. This study underscores the fact that conductivity is an intrinsic property of each individual DNA molecule and not, in fact, an ensemble property of the DNA monolayer. The characterization of DNA monolayers at high oriented pyrolytic graphite is described. Graphite electrodes are modified with duplex DNA containing a pendant pyrene; the yield of electrochemical reduction/oxidation to DNA-bound probes on graphite provides an indicator of base stacking within the DNA. The inclusion of even a single thermodynamically stable yet poorly stacked GT mismatch dramatically attenuates vii CT through the monolayer. Graphite also affords a potential window which is not limited by the chemistry/stability of thiols on gold. This broad window facilitates the exploration of DNA monolayers containing either disulfides or a TEMPO moiety. In addition, graphite offers the opportunity of interrogating the electrochemistry of DNA-binding proteins containing iron sulfur clusters both in the presence and absence of DNA. Indeed, upon DNA binding, the redox potential of the Endo III, a base excision repair enzyme, shifts by –200 mV, and the redox potential of SoxR, a transcription factor, shifts by +500 mV. Such explorations underscore the importance of investigating DNA-binding enzymes in their biologically relevant conformations. The electrochemistry of several phenoxazine- and anthraquinone-based probes of DNA structure is presented. Particular emphasis is given to a Nile Blue (NB) derivative which is connected to the base pair stack via a partially saturated linker and has a midpoint potential of ~ 0 mV versus NHE. NB can be attached to the DNA in a simple fashion and offers the dual advantages of excellent stability and electrocatalytic activity in the presence of ferricyanide. These features facilitate the extensive exploration of DNA-mediated charge transfer to NB at DNA-modified microelectrodes, with a scanning electrochemical microscope, and in a multiplexed format. The application of NB in the detection of transcription factor binding at DNA monolayers is also presented. Nanomolar concentrations of the ubiquitous TATA binding protein transcription factor can be rapidly and sensitively detected from complicated mixtures, including whole-cell lysates. The efficient transport of charge through self-assembled DNA monolayers on both gold and graphite therefore offers an extremely sensitive tool for investigations of viii DNA integrity. As such, novel methodologies for assaying protein/DNA interactions at DNA modified surfaces are now possible. ix TABLE OF CONTENTS Chapter 1: Introduction to DNA-Mediated Electrochemistry 1 1.1 Introduction 2 1.2 Characterization of DNA Monolayers 5 1.2.1 Thiolated-DNA Monolayers on Gold 1.2.2 Pyrenated-DNA Monolayers on Graphite 1.3 Electrochemistry of Small Molecules Non-covalently Bound to DNA 10 1.3.1 DNA Mediated Electrochemistry on Gold 1.3.2 DNA Mediated Electrochemistry on Graphite 1.4 DNA-Mediated Electrochemistry of Small Molecules Covalently Bound to DNA 17 1.4.1 Electrochemistry of Daunomycin 1.4.2 Electrochemistry of Anthraquinone 1.4.3 Electrochemistry of Nile Blue and Redmond Red 1.5 Electrochemical Monitoring of Protein/DNA Interactions 22 1.5.1 Detection of Base Flipping Enzymes 1.5.2 Detection of Transcription Factors that Distort DNA 1.5.3 Electrical Monitoring of Photolyase in Real Time 1.5.4 Electrochemistry of Repair Enzymes Containing Iron Sulfur Clusters 1.5.5 Electrochemistry of an Iron Sulfur Transcription Factor Bound to DNA 1.6 Conclusion 33 References 35 x Chapter 2: Conductivity of a Single DNA Duplex Bridging a Carbon Nanotube Gap 41 Abstract 42 2.1 Introduction 43 2.2 Experimental Section 45 2.2.1 Fabrication of Cut Carbon Nanotube Devices 2.2.2 Synthesis of Amine-Modified DNA 2.2.3 Reconnection of Cut Carbon Nanotube Devices with DNA 2.2.4 Dehybridization/Rehybridization of DNA Bridging the SWNT Gap 2.2.5 Enzyme Restriction of DNA Bridging the SWNT Gap 2.3 Results and Discussion 49 References 66 xi Chapter 3: Electrochemistry Using Self-Assembled DNA Monolayers on Highly Oriented Pyrolytic Graphite 69 Abstract 70 3.1 Introduction 71 3.2 Experimental Section 73 3.2.1 Materials 3.2.2 Synthesis of DNA and Redox-Active Probes 3.2.3 Preparation of DNA-Modified Electrodes 3.2.4 Electrochemical Experiments 3.2.5 Atomic Force Microscopy Experiments 3.2.6 Radioactive Labeling and Quantification 3.3 Results and Discussion 76 3.3.1 Surface Characterization 3.3.2 Electrochemistry at a DNA-Modified HOPG Surface 3.3.3 Electrochemistry on HOPG Is DNA-Mediated 3.3.4 DNA Electrochemistry with Various Intercalators 3.4 Implications 91 References 92 xii Chapter 4: DNA-Mediated Electrochemistry of Disulfides on Graphite 96 Abstract 97 4.1 Introduction 98 4.2 Experimental Section 99 4.2.1 Materials 4.2.2 General Oligonucleotide Synthesis 4.2.3 Synthesis of 5’-Pyrene-Modified Oligonucleotides 4.2.4 Synthesis of 5’-Thiol-Modified Oligonucleotides 4.2.5 Synthesis of 3’-Thiol-Modified and 5’-Pyrene-Modified Oligonucleotides 4.2.6 Preparation of Duplex DNA 4.2.7 Melting Temperatures 4.2.8 Preparation of DNA-Modified Electrodes 4.2.9 Electrochemical Experiments 4.2.10 X-Ray Photoelectron Spectroscopy Experiments 4.3 Results and Discussion 102 4.3.1 Surface Characterization by X-Ray Photoelectron Spectroscopy 4.3.2 Thermal Stability of Thiol Modified Duplexes 4.3.3 Electrochemistry of Disulfides Incorporated in the Phosphate Backbone 4.4 Implications 112 References 113 xiii Chapter 5: Coupling into the Base Pair Stack is Necessary for DNA-Mediated Electrochemistry 117 Abstract 118 5.1 Introduction 119 5.2 Experimental Section 122 5.2.1 Materials 5.2.2 Oligonucleotide Synthesis 5.2.3 Synthesis of Pyrene-Modified Oligonucleotides 5.2.4 Synthesis of Thiol-Modified Oligonucleotides 5.2.5 Synthesis of TEMPO-Modified Oligonucleotides 5.2.6 Synthesis of Anthraquinone-Modified Oligonucleotides 5.2.7 Formation of DNA Monolayers 5.2.8 Electrochemical Measurements 5.2.9 Surface Characterization 5.3 Results 128 5.3.1 Electrochemical Probes and DNA Sequences 5.3.2 Electrochemistry of the Acetylene-Linked Probes 5.3.3 Electrochemistry of the Poorly Coupled Probes 5.3.4 Surface Characterization 5.4 Discussion 139 5.4.1 Electrochemical Reduction of the Acetylene-Linked Probes Is Mediated by the DNA Duplex xiv 5.4.2 Electrochemical Reduction of Alkane-Linked Probes Do Not Depend on DNA-Mediated Electrochemistry 5.4.3 Electronic Coupling into the Base Pair Stack Governs the Size of the Electrochemical Signal 5.4.4 DNA Charge Transfer with Acetylene and Alkane-Linked Probes 5.5 Implications 145 References 147 xv Chapter 6: Direct Electrochemistry of Endonuclease III in the Presence and Absence of DNA 152 Abstract 153 6.1 Introduction 154 6.2 Experimental Section 156 6.2.1 Materials 6.2.2 General DNA Synthesis and Preparation 6.2.3 Expression and Purification of Endonuclease III 6.2.4 Preparation of Bare and DNA-Modified HOPG Electrodes 6.2.5 Preparation of Didodecyldimethylammonium-Modified Pyrolytic Graphite Electrodes 6.2.6 Electrochemical Measurements on HOPG 6.2.7 Electrochemical Measurements on Pyrolytic Graphite 6.3 Results 161 6.3.1 Electrochemistry of Free and DN- Bound Endo III on HOPG 6.3.2 Electrochemistry of Free and DN- Bound Endo III on Pyrolytic Graphite 6.4 Implications and Discussion 171 References 174 xvi Chapter 7: DNA Binding Shifts the Redox Potential of the Transcription Factor SoxR 179 Abstract 180 7.1 Introduction 181 7.2 Experimental Procedures 185 7.2.1 Materials 7.2.2 Oligonucleotide Synthesis 7.2.3 Expression and Purification of SoxR 7.2.4 Formation of DNA Monolayers and Electrochemical Measurements 7.3 Results 188 7.3.1 Experimental Strategy Used to Probe SoxR Electrochemically 7.3.2 SoxR Binding Is Reported Through the Redmond Red Electrochemical Signal 7.3.3 Electrochemistry of P. aeruginosa SoxR 7.3.4 Comparison of the Voltammetry of E. coli and P. aeruginosa SoxR 7.3.5 Electrochemistry of the P. Aeruginosa S95L SoxR Mutant 7.3.6 Comparison of the Voltammetry of SoxR to that of Free Iron 7.4 Discussion and Implications 199 References 205 xvii Chapter 8: Scanning Electrochemical Microscopy of DNA Monolayers Modified with Nile Blue 210 Abstract 211 8.1 Introduction 212 8.2 Experimental 214 8.2.1 Materials 8.2.2 Synthesis of Thiol and Nile Blue-Modified DNA 8.2.3 Preparation of Backfilled DNA/11-mercaptoundecylphosphoric Acid Monolayers 8.2.4 Electrochemistry/Scanning Electrochemical Microscopy Experiments 8.2.5 TATA Binding Protein Experiments 8.3 Results 220 8.3.1 Electrochemistry of Backfilled DNA Monolayers Modified with Nile Blue 8.3.2 SECM of DNA Monolayers Modified with Nile Blue in Feedback Mode 8.3.3 SECM Imaging of DNA Monolayer Morphology 8.3.4 Detection of TATA Binding Protein by SECM 8.4 Discussion 238 References 241 xviii Chapter 9: Electrical Detection of TATA-Binding Protein at DNA-Modified Microelectrodes 246 Abstract 247 9.1 Introduction 248 9.2 Experimental 250 9.2.1 Materials 9.2.2 Synthesis of Thiol and Nile Blue-Modified DNA 9.2.3 Protein Preparation and Storage 9.2.4 Preparation of Backfilled DNA Monolayers 9.2.5 Electrochemistry and Protein Detection Experiments 9.3 Results 254 9.4 Implications 265 References 266 xix Chapter 10: Electrochemical Detection of TATA-Binding Protein from a Cell Lysate 270 Abstract 271 10.1 Introduction 272 10.2 Experimental 275 10.2.1 Materials 10.2.2 Synthesis of Thiol and Nile Blue-Modified DNA 10.2.3 Preparation of Epithelial Cell Lysates 10.2.4 Gel Electrophoresis and Immunoblot Analysis 10.2.5 Manufacture of Gold Electrodes 10.2.6 Preparation of Backfilled DNA Monolayers 10.2.7 Electrochemical Experiments 10.3 Results 281 10.3.1 Experimental Strategy for the Detection of TATA Binding Protein 10.3.2 Nile Blue as a Redox Probe for the Detection of Protein Binding 10.3.3 10.3.3 Self Assembly of Nile Blue-DNA Monolayers 10.3.4 Detection of Human TATA Binding Protein from Cell Lysate Solutions on a Single Substrate 10.3.5 Detection of Murine TATA Binding Protein from Whole Cell Lysates 10.4 Implications and Conclusion 292 References 294 xx LIST OF TABLES AND FIGURES Chapter 1: Introduction to DNA-Mediated Electrochemistry 1 Figure 1.1: Schematic illustration of DNA duplexes in a close-packed monolayer 4 Figure 1.2: Schematic illustration of the modulation of the height of a DNA monolayer 9 Figure 1.3: Chemical structures of redox-active probes 12 Figure 1.4: Schematic illustration of the reduction of methylene blue 13 Figure 1.5: Schematic illustration of the methylene blue/ferricyanide electrocatalytic cycle at a DNA monolayer 15 Figure 1.6: A schematic illustration of DNA duplexes containing covalently crosslinked daunomycin 18 Figure 1.7: Illustration of the fabrication of DNA monolayers on gold and the subsequent electrochemical analysis of protein binding 24 Figure 1.8: Schematic illustration of the electrochemical monitoring of the repair of thymine dimers in a DNA monolayer by photolyase 30 xxi Chapter 2: Conductivity of a Single DNA Duplex Bridging a Carbon Nanotube Gap 41 Figure 2.1: A method to cut and functionalize individual SWNTs with DNA strands 46 Figure 2.2: Device characteristics for individual SWNTs connected with DNA. 50 Table 2.1: Summary of DNA resistance values 52 Figure 2.3: Device characteristics for individual SWNTs reconnected with well matched DNA 53 Figure 2.4: Device characteristics for individual SWNTs reconnected CA mismatched DNA 54 Figure 2.5: Mismatches have a large effect on DNA conductance 56 Figure 2.6: Effect of CA mismatches on DNA conductance 57 Figure 2.7: Effect of GT mismatches on DNA conductance 58 Figure 2.8: The electrical characteristics of a device rejoined with single-stranded DNA 60 Figure 2.9: Control experiments using partially cut devices 61 Figure 2.10: Enzymes can be used to cleave the DNA between the ends of the SWNTs 63 Figure 2.11: Restriction by enzymes is sequence specific 64 xxii Chapter 3: Electrochemistry Using Self-Assembled DNA Monolayers on Highly Oriented Pyrolytic Graphite 69 Figure 3.1: Schematic illustration of DNA duplexes modified with a pyrene moiety assembled on an HOPG surface 78 Figure 3.2: AFM characterization of DNA-modified HOPG 79 Figure 3.3: Cyclic voltammetry of methylene blue 81 Figure 3.4: Cyclic voltammetry of methylene blue at DNA monolayers with and without a mismatch 84 Figure 3.5: Cyclic voltammetry of methylene blue and ferricyanide at DNA monolayers with and without a mismatch 86 Table 3.1: Summary of mismatch detection with various linkers on DNA-modified HOPG electrodes 87 Figure 3.6: Cyclic voltammetry of metallointercalators on bare and DNA-modified HOPG 89 Figure 3.7: Cyclic voltammetry of metallointercalators in the presence and absence of mismatches 90 xxiii Chapter 4: DNA-Mediated Electrochemistry of Disulfides on Graphite 96 Figure 4.1: Schematic illustration of charge transfer to thiols incorporated into the DNA backbone 103 Figure 4.2: XPS spectra of HOPG modified with pyrenated DNA 105 Figure 4.3: Melting temperatures of DNA duplexes containing thiols 106 Figure 4.4: Square wave voltammetry at various DNA monolayers modified with disulfides 108 Figure 4.5: Cyclic voltammetry of an electrode modified with duplex DNA featuring a disulfide at various scan rates 109 Figure 4.6: Cyclic voltammetry of well-matched DNA featuring a disulfide under various buffer conditions 111 xxiv Chapter 5: Coupling into the Base Pair Stack is Necessary for DNA-Mediated Electrochemistry 117 Scheme 5.1: Conditions for synthesis of 2-Iodoanthraquinone and its coupling to ethynyluracil-modified DNA 126 Figure 5.1: Sequences and probes used during the course of these studies 131 Figure 5.2: Square wave voltammetry of AQ and TEMPO-modified DNA 132 Figure 5.3: Cyclic voltammetry of AQ-modified DNA at various scan rates 133 Figure 5.4: Square wave voltammetry of well-matched DNA modified with an acetylene-TEMPO label on HOPG 135 Table 5.1: Electrochemical characteristics of DNA monolayers modified with electroactive moieties 136 Figure 5.5: Tapping mode AFM characterization of DNA-modified and bare Graphite 138 Figure 5.6: Schematic representation of two possible modes of charge transfer from an electroactive label incorporated into DNA 143 xxv Chapter 6: Direct Electrochemistry of Endonuclease III in the Presence and Absence of DNA 152 Figure 6.1: Schematic representation of electrochemistry for Endo III on HOPG with and without modification with DNA 157 Figure 6.2: Cyclic and square wave voltammetry of Endo III in the presence and absence of DNA 162 Figure 6.3: Square wave voltammetry of Endo III on bare HOPG with and without backfilling 163 Figure 6.4: Cyclic voltammetry on DNA-modified HOPG with and without Endonuclease III 165 Figure 6.5: Cyclic voltammetry of higher Endo III concentrations in the absence of DNA 166 Figure 6.6: Cyclic voltammetry of free iron on bare HOPG in the presence and absence of thiols 168 Figure 6.7: Cyclic voltammetry of Endo III on bare HOPG at slow scan rates 169 Figure 6.8: Cyclic voltammetry of free didodecyldimethylammonium-modified graphite and DNA-bound Endo III on 170 Figure 6.9: Illustration of the potentials versus NHE of the couples of Endo III in the presence and absence of DNA 172 xxvi Chapter 7: DNA Binding Shifts the Redox Potential of the Transcription Factor SoxR 179 Figure 7.1: Schematic illustration of the self-assembly/backfilling of a DNA monolayer followed by incubation with protein 189 Figure 7.2: Cyclic voltammetry of SoxR at electrodes modified with DNA 192 Figure 7.3: Binding of SoxR to the DNA-modified film 194 Figure 7.4: Square wave voltammetry of P. aeruginosa and E. coli SoxR at DNAmodified graphite electrodes 196 Figure 7.5: Cyclic voltammetry of a SoxR mutant at an electrode modified with DNA 198 Figure 7.6: Cyclic voltammetry of free iron in SoxR storage buffer on bare and DNAmodified HOPG 200 Figure 7.7: Redox potentials of free and DNA-bound SoxR along with those of cellular oxidants/reductants 202 xxvii Chapter 8: Scanning Electrochemical Microscopy of DNA Monolayers Modified with Nile Blue 210 Figure 8.1: Schematic illustration of the coupling of Nile Blue to the NHS-CarboxydT-modified DNA on solid support 216 Figure 8.2: Schematic illustration of SECM imaging of DNA monolayers modified with Nile Blue 221 Figure 8.3: Cyclic voltammetry of Nile Blue-DNA monolayers 222 Figure 8.4: Cyclic voltammetry of DNA monolayers modified with Nile Blue in the presence of ferricyanide 223 Figure 8.5: Successive cyclic voltammograms of ferricyanide at a bare Au electrode in the presence of mercaptoundecylphosphoric acid 224 Figure 8.6: SECM approach curves obtained for DNA monolayers modified with Nile Blue 226 Figure 8.7: SECM approach curves at DNA monolayers in the presence and absence of Methylene Blue 229 Figure 8.8: SECM images obtained for Nile Blue-DNA monolayers 232 Figure 8.9: Schematic illustration of the SECM-based, sequence-specific detection of TBP binding at DNA monolayers 234 Figure 8.10: SECM approach curves taken for Nile Blue-DNA monolayers with and without TBP 235 Figure 8.11: Scanning electrochemical microscopy images of a Nile Blue-DNA monolayer in the presence and absence of various proteins 236 xxviii Chapter 9: Electrical Detection of TATA-Binding Protein at DNA Modified Microelectrodes 246 Figure 9.1: Schematic illustration of the addition of TBP to a DNA monolayer modified with Nile Blue 255 Figure 9.2: Cyclic and square wave voltammetry at macroelectrodes and microelectrodes before and after addition of TBP 256 Figure 9.3: Cyclic voltammetry of a Nile Blue-DNA modified microelectrode of a 25 µm diameter for two different buffers 257 Figure 9.4: The effect of oxygen on the voltammetry at a DNA modified Microelectrode 259 Figure 9.5: Square wave voltammetry of DNA-modified macroelectrodes and microelectrodes prepared on a single substrate 260 Figure 9.6: Cyclic voltammetry of a Nile Blue-DNA modified microelectrodes showing reversible binding of TBP 262 Figure 9.7: Detection of TBP binding at microelectrodes using various concentrations of TBP 263 Figure 9.8: Detection of TBP at microelectrodes in the presence of various other proteins 264 xxix Chapter 10: Electrochemical Detection of TATA-Binding Protein from a Cell Lysate 270 Figure 10.1: Illustration of the self-assembly of DNA monolayers on gold and the subsequent electrochemical analysis of protein binding 280 Figure 10.2: Cyclic and square wave voltammetry of a Nile Blue-DNA monolayer 282 Figure 10.3: An illustration of the distribution of integrated surface charge found for NB-DNA monolayers on gold electrodes 284 Figure 10.4: Square wave voltammetry of Nile Blue-DNA monolayers in the presence and absence of cell lysate/TBP 286 Figure 10.5: A schematic illustration of the growth and lysis of murine epithelial cells 288 Figure 10.6: A representation of the signal loss found at Nile Blue-DNA modified surfaces after addition of epithelial cell lysate under various conditions 289 Figure 10.7: Visualization of TBP levels in cell lysates via western blotting Analysis 291  Chapter 1: Introduction to DNA-Mediated Electrochemistry Adapted from: Gorodetsky, A. A., Buzzeo, M., Barton, J. K. (2008) Bioconjugate Chem. ASAP. 1 1.1 Introduction Since the elucidation of the double helical structure of DNA, scientists have been fascinated by the possibility that the stacked aromatic base pairs of the duplex may promote charge transport (CT) over significant distances (1, 2). Consequently, the nature of the conductive properties of duplex DNA has attracted substantial interest (3–7). Over the past two decades, a wide-ranging collection of experiments has both revealed the fundamental details of DNA-mediated CT and illustrated its potential for sensing applications. Initial solution experiments featured photoinduced DNA-mediated CT between well-defined donor and acceptor sites (8, 9). While long-range CT was shown to yield oxidative damage in DNA up to 200 Å away from the bound oxidant (10, 11), DNA CT was also found to be exquisitely sensitive to not only the integrity of the base pair stack (9, 12) but also the coupling of the donors and acceptors with the DNA (13). Indeed, this sensitivity prompted both the consideration of biological roles for DNA CT (14) and the construction of electrochemical DNA-based sensors for mutations, base lesions, and protein-binding (15). Electrochemical techniques provide a particularly convenient means for the study of heterogeneous electron transfer at solid surfaces (16). Typically, redox-active molecules are modified with thiol-terminated alkyl chains and self-assembled as wellordered monolayers on metallic surfaces (17). Under an applied potential, electrons or holes are then transferred to pendant, redox-active head groups with the rates and yield of charge transfer (as measured electrochemically) providing information on the structure of 2 the thiol-terminated linker. In our laboratory, we have extended this methodology to the study and application of DNA-modified surfaces where charge transfer is mediated by the DNA monolayers, which are self-assembled onto conductive substrates. Prior to a discussion of the applications of this methodology, it is important to have a clear understanding of the architecture of DNA-modified surfaces. To facilitate monolayer formation, duplex DNA is modified at the 5’ end of a single strand with a linker (an alkanethiol in the case of gold and pyrene in the case of graphite), which allows the DNA molecules to interact with the electrode surface (Figure 1.1). Electrodes are then incubated with modified duplexes and an organized monolayer is allowed to self-assemble. A redox-active molecule is introduced either by exposure to a solution or through covalent attachment to the 5’ terminus of the single strand lacking the linker. Subsequent addition of an inert backfilling agent, such as mercaptohexanol, promotes the removal of non-specifically adsorbed duplexes and single strands. The uniformity and reliability of this structure effectively enables the duplexes to serve as an extension of the active electrode surface, thus allowing DNA-bound, redox-active molecules to be reduced via the π-stack. Moreover, since the electrochemistry of the probe is remarkably sensitive to the integrity of the base pair stack, any perturbation of the base pair stack, which intervenes between the probe and the surface, dramatically attenuates the electrochemical yield of CT. It is noteworthy that the charge transport properties of DNA monolayers are not simply of utility in the context of electrochemistry, as illustrated in Chapter 2. Previous experiments, which found a staggering range of value for the conductivity of DNA, 3 Figure 1.1: Schematic illustration of DNA duplexes in a close-packed monolayer which is immobilized on gold via an alkanethiol tether 4 nonetheless hinted that charge transport in such monolayers was not simply an ensemble property of an aligned film (18–20). Indeed, our more recent experiments, where a single DNA molecule was utilized to bridge a carbon nanotube gap, confirmed that it is the base pair stack of each individual duplex which serves as a conduit for charge transfer (21). In fact, when considered in the context of an electrical circuit, the DNA duplexes and bound redox-active probes can be viewed as transducer elements with extraordinary gain; biochemical binding events such as hybridization are sensitively transduced into electrical signals. This emphasizes the role of DNA as a remarkably unique medium that provides unprecedented opportunities for studies and applications of long-range CT. The thesis work presented here concerns the development of new techniques and methodologies for the electrochemical monitoring of DNA binding proteins, with a particular emphasis on applications. This encompasses the characterization of alternative surfaces such as graphite, development of new electrochemical probes, explorations of electrical measurements at the micro- and nanoscales, and extensions to multiplexing. Taken together, these developments constitute a firm foundation for novel biosensors which are based on the exquisite sensitivity of DNA-mediated CT to perturbations of the base pair stack. 1.2 Characterization of DNA Monolayers For proper interpretation of experimental observations, extensive physical characterization of these self-assembled DNA monolayers has been crucial. A gamut of experimental techniques have been utilized for investigating the structure and 5 characteristics of DNA monolayers on gold including (but not limited to) radioactive labeling (22–25), fluorescence self-interference (26–28), scanning tunneling microscopy (29–31), atomic force microscopy (23, 25, 32–36), scanning electrochemical microscopy (37–39), and surface plasmon resonance (40, 41). When dense DNA monolayers are desired, the monolayers are assembled in the presence of Mg2+ to allow for close packing of the duplexes (schematically illustrated in Figure 1.1). Our group, in particular, has focused on surface characterization via radioactive labeling and scanning probe techniques. These investigations have indicated that DNA monolayers adopt an upright orientation and have predictable surface coverages, thereby allowing us to develop a detailed picture of DNA monolayer morphology. 1.2.1 Thiolated-DNA Monolayers on Gold Radioactive labeling of the duplexes provides a reliable measure of surface coverage for DNA monolayers. In a typical radioactive labeling experiment on gold, the 5’ end of the DNA complementary to the thiolated strand is labeled with polynucleotide kinase, and the amount of DNA is quantified directly via scintillation counting after selfassembly. On gold, these measurements indicated surface coverages of ~ 40 pmol/cm2 for well-packed DNA monolayers self-assembled in the presence of Mg2+ (22) and coverages of ~ 12 pmol/cm2 for loosely-packed DNA monolayers in the absence of Mg2+ (23). Interestingly, the well packed monolayer coverages obtained via radioactive labeling correspond to only ~ 75 % of the theoretical coverages of less than 60 pmol/cm2, with DNA helices packed perpendicular to the surface. This fact supports the notion that the DNA is oriented at an angle with respect to the surface. 6 While radioactive labeling affords a quantitative measurement of the overall surface coverage, the electronic structure of local areas of a DNA monolayer can be probed via scanning tunneling microscopy (STM), providing atomic level information (29). Indeed, STM allows for the investigation of the electronic properties of DNA as a function of duplex orientation. Therefore, the images obtained yield information on not only the morphology but also the local density of states of the sample. At negative potentials, where the DNA adopts an upright orientation, agglomerates of duplexes are visible because the tip can electronically access the oriented DNA in an effective metalmolecule-metal junction. However, at positive potentials, the DNA is encouraged to lie down and the DNA film is effectively not visible to the STM tip. These data indicated that the DNA duplexes comprising a DNA monolayer aggregate in small hexagonal groups of ~ 10 nm diameter. Furthermore, the inclusion of a mismatch within the monolayer dramatically attenuates communication between the tip and the surface, demonstrating that the STM is accessing the local density of states of the full DNA duplex. The morphology of self-assembled DNA monolayers can be further explored at the nanometer scale with atomic force microscopy (AFM). These investigations have indicated that 5’ tethered films are essentially smooth and featureless within the resolution of the AFM tip with the DNA dispersed in a uniform manner (32). AFM measurements also provide important information about the morphology and depth of the DNA monolayer at various substrate biases. By applying a large downward force to the monolayer with the AFM tip, a bare spot can be uncovered on the DNA-modified surface. Height contrast measurements between this bare spot and the surrounding DNA 7 monolayer indicate a film depth which ranges from 4.2 to 4.6 nm for 15 mer duplexes at open circuit on gold. In fact, the height of these films can be reversibly modulated from 2 nm (the diameter of the duplex) at positive potentials to 5.5 nm (the full length of the duplex when normal to the surface) at negative potentials (Figure 1.2). These data, in conjunction with the above experiments, indicate that the duplexes adopt an upright orientation at negative potentials and are oriented at a ~ 45 º angle with respect to the surface in the absence of an applied bias (25, 32). 1.2.2 Pyrenated-DNA Monolayers on Graphite Physical characterization of DNA monolayers on graphite is crucial for the development of this material as an alternative surface for investigations of DNA CT. The preparation and investigation of DNA monolayers on highly oriented pyrolytic graphite (HOPG) is described in Chapter 3; the monolayers are formed in an identical fashion to those on gold, but the DNA is anchored to the surface via a hydrophobic, pyrenated tether (25). Radioactive labeling indicates similar DNA densities on graphite as on gold. In addition, AFM measurements reveal that DNA monolayers on graphite are smooth and featureless with a film depth of 4.4 nm for monolayers consisting of 15 mer duplexes. Taken together, AFM measurements and radioactive labeling studies indicate that wellpacked DNA monolayers on graphite are oriented at an angle with respect to the surface and are essentially identical to thiolated DNA monolayers formed on gold. 8 Figure 1.2: Schematic illustration of the modulation of the height of a DNA monolayer at different substrate potentials. At open circuit, imaging studies have shown that duplexes are oriented upright and at a ~ 45º angle relative to the surface. At positive potentials, the polyanionic DNA duplexes are attracted to the surface, while at negative potentials the duplexes are in an upright orientation with the helix axes perpendicular to the electrode surface. 9 1.3 Electrochemistry of Small Molecules Non-covalently Bound to DNA Numerous redox-active probes have been utilized for investigations of DNA CT, and, at first glance, many molecules appear adequate for these studies. The large collection of experiments conducted in our laboratory over the past decade, however, has taught us essential lessons regarding the selection of an appropriate redox probe. Three criteria are paramount: the species must (i) be electronically well coupled to the base pair stack, (ii) be reliably linked to the duplex, and (iii) undergo stable reduction or oxidation within a potential window that does not compromise the integrity of the DNA monolayer. Furthermore, in addition to affording insight into the optimum probe design, our systematic studies clearly indicate that no one probe is ideally suited for every experiment. Instead, different markers offer distinct advantages and must be individually evaluated for specific applications. The behavior of the redox probes employed in our studies do, however, display some common features when monitored electrochemically. When bound to DNA, small molecules appear to act as surface-bound species, as indicated by a linear relationship between the peak current and scan rate. This relationship is not strictly linear for DNAbound redox proteins, where the proteins are not situated in one fixed position or conformation and thus exhibit a slight diffusive component. Nonetheless, covalent probes bound to DNA typically do not display a square-root-dependence of the peak current on the scan rate that is indicative of free diffusion. In addition, DNA binding typically shifts the redox potential of the probe by 20–50 mV from direct reduction at the bare surface and the voltammetry adopts a skewed shape relative to a classically adsorbed couple; the 10 reductive and oxidative peaks are separated slightly due to the intervening DNA/alkanethiol spacer. Electrochemical reversibility and stability are also specific to each probe, but they can be interrogated simply by repetitive cycling over the potential window of interest. Taken together, all of these features allow us to distinguish DNAmediated processes from ones involving direct electrochemistry. 1.3.1 DNA-Mediated Electrochemistry on Gold Early work on gold employed probes that were not covalently attached to DNA but instead interacted with the base pair stack electrostatically or via intercalation (22). Although the majority of our studies have involved the intercalator methylene blue, we have also studied the interaction between DNA and anionic ferricyanide, cationic ruthenium hexammine, the tris-heteroleptic intercalator Ir(bpy)(phen)(phi)3+, and the antitumor agent Daunomycin (Figure 1.3). While these non-covalent markers have afforded a significant foundation for the study of DNA-mediated CT, they are intrinsically limited in the mechanistic detail and experimental control they can provide. As a result, latter experiments focused on developing covalently bound probes, specifically those that are electronically well coupled to the base pair stack. Methylene blue (MB), a well-studied aromatic intercalator that undergoes a 2e/1H+ reduction in aqueous systems, has served as a workhorse reporter of DNA-mediated processes (22). Figure 1.4 shows a schematic illustration of the reduction of MB at a DNA monolayer; a reversible redox couple is seen with a midpoint potential centered at 10 to -50 mV vs. NHE, shifted by ~ 30 mV from bare gold. The measured peak currents scale linearly with scan rate, indicating that the redox molecule is surface bound, and a combination of electrochemical and spectroscopic measurements give a saturation value 11 Figure 1.3: Chemical structures of redox-active probes which have been utilized to explore DNA-mediated CT. Non-covalent probes are on the top row (from left to right): Methylene blue, Daunomycin, and Ir(bpy)(phen)(phi)3+. Covalent probes are on the bottom row (from left to right): Anthraquinone, Redmond Red, and Nile Blue. 12 Figure 1.4: Schematic illustration of the reduction of methylene blue (blue) via DNAmediated CT at a well-packed DNA monolayer. The probe is constrained to intercalate near the top of the monolayer. 13 of 1.4(2) molecules per duplex at high MB concentrations. This latter finding suggests that, for well-packed monolayers, MB is constrained to the top of the film. Any observed reduction, therefore, can be attributed to a DNA-mediated process, as the surface is inaccessible to the molecule. To further prove that the reduction of distally bound probes proceeds in a DNAmediated fashion, the electrochemistry of MB and other small molecules was explored at DNA monolayers containing single base mismatches positioned between the electrode surface and the probe (45). It was theorized that if the DNA base pairs combined to form a continuous, π-stacked conduit for charge transfer, that the disruption of a single base pair, as in a mismatch, could interrupt the CT pathway. The presence of a single base mismatch causes no global change in the duplex structure, and yet, the inclusion of an intervening CA mismatch proved to have a dramatic effect on the efficiency of charge transfer to MB. These results are in excellent agreement with STM studies of matched and mismatched duplexes and with studies of photoinduced DNA CT in solution (9) (vide supra). Interestingly, not only could several mismatches be detected but the DNAmediated chemistry was also found to be independent of sequence context. Furthermore, this exquisite sensitivity to base pair stacking was observed for numerous intercalating (but not groove-binding) probes, supporting the notion that the observed electrochemistry is DNA-mediated and a general characteristic of the DNA. To increase the sensitivity of mismatch detection and further improve discrimination, MB was coupled to freely diffusing Fe(CN)63- in an electrocatalytic cycle (Figure 1.5). In this catalytic process, MB undergoes a two e- reduction via the base pair 14 Figure 1.5: Schematic illustration of the methylene blue/ferricyanide electrocatalytic cycle at a DNA monolayer. DNA-bound methylene blue is reduced in a DNA-mediated reaction to leucomethylene blue (LB), which dissociates from the monolayer and is reoxidized by ferricyanide in solution; the reoxidized methylene blue returns to the DNA for another round of DNA-mediated electrochemistry. This electrochemical amplification leads to enhanced discrimination of all single base mismatches in DNA, since the intervening mismatch shuts down the DNA-mediated reaction and thus the electrocatalytic cycle. 15 stack, and upon reduction to leucomethylene blue, loses some affinity for DNA, thus dissociating from the duplex (45, 46). In turn, the cycle continues when leucomethylene blue is reoxidized by Fe(CN)63- in solution, and binds again to DNA for another cycle of DNA-mediated reduction. It should be noted that Fe(CN)63- cannot be reduced directly at the negatively charged monolayer owing to repulsion from the polyanionic DNA monolayer. The entire process is effectively governed by the on/off kinetics of the methylene blue/leucomethylene blue redox couple, as confirmed by rotating disk electrode experiments (48). In fact, during the course of a single voltammogram, the DNA is repeatedly sampled, amplifying the absolute signal of MB and increasing the signal attenuation associated with a mismatch. The detection of all naturally occurring mismatches and nearly all common lesions was achieved with the MB/Fe(CN)63- electrocatalytic couple via cyclic voltammetry and chronocoulometry. In addition, low detection limits were demonstrated in a chip-based format, with mismatch discrimination possible at electrodes as small as 30 µm in diameter (46). It should also be noted that the relatively stable GA mismatch could not be detected without the aid of the described catalytic cycle, emphasizing the utility of electrocatalysis as an electrochemical tool. Furthermore, even base lesions that cause minor thermodynamic destabilization to the duplex could be detected with electrocatalysis: a systematic investigation revealed that the efficiency of CT was dramatically decreased even by subtle base modifications that altered base structure, steric bulk, or hydrogen bonding (49). Perhaps not surprisingly, modifications with methyl groups that do not participate in hydrogen-bonding had little-to-no effect on the DNA-mediated electrochemistry. The sensitive detection of all single base mismatches 16 independent of sequence context is remarkable and has proven to be valuable in the application of DNA electrochemistry for the diagnosis of single nucleotide polymorphisms. 1.3.2 DNA-Mediated Electrochemistry on Graphite To electrochemically characterize HOPG, MB and DM electrochemistry is also explored at self-assembled DNA monolayers on this surface, as described in Chapter 3. The behavior of both intercalators on graphite is quite similar to that on gold, and evidence that the charge transfer is DNA-mediated is provided by the facile detection of the CA and GT mismatches. In addition, as on gold, mismatch discrimination with MB on graphite can be dramatically improved by couple MB to ferricyanide in a catalytic cycle. Notably, as a surface, HOPG provides an attractive alternative since its accessible potential window is no longer limited by the oxidative reduction or desorption of thiols. In fact, the wider potential window afforded by HOPG allows for the study of the voltammetry of Ru(bpy)2dppz2+ and Os(phen)2dppz2+. As expected, the electrochemistry of both intercalators is DNA-mediated. 1.4 DNA-Mediated Electrochemistry of Small Molecules Covalently Bound to DNA To facilitate the systematic investigation of charge transduction of DNA, it became necessary to employ redox-active probes that are covalently conjugated to the base pair stack. These initial efforts focused on covalently attached daunomycin (DM), which can be site-specifically cross-linked to the exocyclic amine of guanine, forming a covalent bond, as shown in Figure 1.6 (50–52). Most importantly, this allows for 17 Figure 1.6: A schematic illustration of DNA duplexes with daunomycin covalently cross-linked at various separations from the electrode surface. Also depicted are the DM-guanine cross-link (bottom right) and the thiol-terminated tether which connects the duplex to the electrode surface (bottom left). 18 controlled placement of the probe within the sequence. Importantly, the covalent attachment of this probe for DNA electrochemistry has made possible several seminal conclusions about DNA-mediated CT involving ground-state reactants. While DM has afforded several key findings, it does possess some inherent limitations. Since functionalization with DM occurs after the synthesis of the DNA, it does necessitate that all guanines other than the one to be linked are replaced by inosine, a close base analogue. For the purpose of electrochemical measurements, this substitution has no influence on the results, but this requirement limits its applicability in protein detection schemes and biosensors. Moreover, the stability of the linkage between DM and DNA is heat sensitive. As a result, we have been focused on developing a new generation of probes based on phenoxazine and anthraquinone cores that can be covalently attached to DNA with synthetic flexibility, display excellent stability, and are electronically well coupled to the base pair stack. 1.4.1 Electrochemistry of Daunomycin In initial experiments on gold, the separation between the gold surface and DM adduct was varied from 15 Å to 45 Å, and no effect on the yield or rate of electron transfer was observed (50). The inclusion of a mismatch in the intervening DNA, however, completely shuts off the electrochemistry of DM, conclusively demonstrating that CT is DNA-mediated. Subsequently, with DM held in a fixed position in the duplex, the length of the thiolated alkyl tether was varied to probe the effect of tether length on the rate of charge transfer (51). While the yield of CT remained the same regardless of chain length (the number of intervening methylene units was increased from 4 to 9), significantly, the rate of electron transfer decreased with increased tether length with a βn 19 of 1.0 per –CH2– unit, as expected for the variation in coupling with σ-bonded systems. Therefore, CT through the alkyl tether, not through 30 Å of π-stacked DNA, was the ratelimiting step of the DNA-mediated reduction. Taken together, these findings illustrated that the DNA-DM construct behaves as a single redox-active entity with dramatically different rates of CT through the stacked DNA and the σ-bonded alkyl chain. In a later study, DM was employed to prove that DNA-mediated electrochemistry occurs via the base pair stack and not the sugar-phosphate backbone (52). With two DM moieties crosslinked to guanine residues, voltammograms of duplexes containing a nick in the backbone, a nick and a mismatch, a nick on both strands, and no modifications were compared. Nicks in the phosphate backbone did not attenuate CT, but, consistent with previous results, mismatches significantly attenuated the electrochemistry of DM. Interestingly, Chapter 4 shows that the electrochemistry of thiols incorporated within the sugar-phosphate backbone is DNA-mediated but proceeds at low yield (53). This demonstrates the utility of graphite for the exploration of redox chemistry which is not accessible on gold. In addition, this work reveals that although DNA electrochemistry proceeds through the base pair stack rather than through the sugar-phosphate backbone, it can promote reactions on the DNA backbone. 1.4.2 Electrochemistry of Anthraquinone Through a variety of experiments conducted both in our laboratory and others, it has become apparent that efficient coupling of the probe to the base stack is critical for efficient DNA CT. Recently, the groups of Saito and Gooding introduced thymines modified with anthraquinone as effective, covalently attached probes of charge transfer at DNA-modified surfaces (54–57). Using this technology, Gooding and coworkers 20 demonstrated the application of these modified probes to the detection of primer extension reactions (54, 55), and Saito and coworkers demonstrated the photostimulated detection of mismatches with applications to genotyping of single nucleotide polymorphisms (56, 57). However, the low current densities obtained in these studies led us to conduct a comparative study of anthraquinone linked to DNA through conjugated and saturated tethers, as described in Chapter 5 (58). Importantly, we found efficient redox chemistry and mismatch sensitivity for anthraquinone only when linked through an alkyne and not when tethered by an alkyl chain. Thus not only the choice of probe but also how the probe is coupled to the base pair stack are key to effective DNA-mediated CT. 1.4.3 Electrochemistry of Nile Blue and Redmond Red More recently, we have investigated two phenoxazine-based probes, Redmond Red (RR) and Nile Blue (NB), as covalent reporters of DNA CT (Figure 1,3). Both of these probes can be incorporated into DNA on solid support, have virtually no sequence restrictions, are stable to light/heat, and support DNA-mediated electrochemistry. Accordingly, they are preferable to DM for routine use and high throughput experiments, and the continued development of these probes holds great promise for future protein detection experiments. As described in Chapter 7, Redmond Red is commercially available as a phosphoramidite, affording facile incorporation into DNA (59). Although this probe is not perfectly coupled into the DNA π-stack, it can be placed opposite an abasic site to provide a more intimate interaction with the DNA and obtain larger electrochemical signals. In addition, the remarkable stability of this probe makes it particularly useful as 21 an internal standard for the electrochemistry of the iron-sulfur-containing transcription factor SoxR (vide supra). NB, as described in detail in Chapters 8–10, can be attached to DNA on solid support by reacting its exocylic amine with a carboxy-NHS-ester-thymine, ensuring that NB is connected to the DNA via a partially saturated linkage (60). Such a linkage affords relatively large electrochemical signals and is similar to the one described by Gooding and coworkers for anthraquinone (54, 55). Furthermore, the electrochemical signal from NB can be amplified via electrocatalysis with oxygen and ferricyanide. This catalytic activity facilitates the characterization of NB-modified DNA monolayers with scanning electrochemical microscopy (SECM) in Chapter 8, affording great insight into the morphology of probe modified films. Finally, as an added benefit, NB indicates its own degradation through the appearance of a new redox signature. All of these features make NB the probe of choice for protein detection experiments at both small scales (Chapter 9) and in a multiplexed format (Chapter 10). 1.5 Electrochemical Monitoring of Protein/DNA Interactions Self-assembled DNA monolayers provide especially convenient platforms for electrical monitoring of protein/DNA interactions. These interactions play crucial roles in many cellular processes, such as transcription, repair, and replication. In particular, the association of transcription factors with DNA is an important area for exploration in proteomics and genomics; it is these interactions that control the developmental and regulatory responses of the cell, often in a complicated fashion. Therefore, the 22 development of convenient and inexpensive methodologies for monitoring protein-DNA interactions remains of critical importance. In a typical experiment, a loosely packed DNA monolayer is self-assembled in the absence of magnesium, with or without a covalently appended probe, and the surface is backfilled (Figure 1.7). Such sparse DNA films are necessary for improved accessibility of proteins; prior to backfilling, they possess a distinct “island” morphology since the DNA has room to lie flat on the gold surface in the absence of an applied potential (23). However, monolayers backfilled with a short chain alkanethiol such as mercaptohexanol have been demonstrated to adopt an upright orientation relative to the gold surface even at open circuit (35, 36). This well-defined morphology is a crucial for accurate sequencespecific detection of DNA binding proteins. After self-assembly and backfilling, voltammetry of the DNA-modified electrode is recorded in the absence and presence of protein (Figure 1.7). Detection of protein binding can be achieved in two ways: 1) proteins that perturb the base pair stack will attenuate the yield of DNA CT to a distally-bound electroactive probe such as DM, and 2) proteins featuring a redox-active cofactor such as an iron-sulfur cluster will cause the appearance of a new DNA-dependent electrochemical signal. These experiments have greatly benefited from the development of covalently attached probes (vide supra) which have allowed for conclusive quantification of the yield of DNA CT before and after protein binding. Furthermore, such electrochemical assays are general and based on exquisitely sensitive DNA-mediated electrochemistry, so a wide range of protein/DNA interactions and binding motifs can be sequence-specifically interrogated in real time. 23 Figure 1.7: Illustration of the fabrication of DNA monolayers on gold and the subsequent electrochemical analysis of protein binding. First, loosely packed DNA monolayers are formed and the surface is backfilled with mercaptohexanol for passivation. In the absence of protein, DNA-mediated redox chemistry to the DNA-bound probe is observed. Binding of a protein that interrupts the base pair stack also interrupts the DNA-mediated redox chemistry, leading to attenuation of the redox signal. 24 Previous studies at DNA-DM monolayers demonstrated that protein-induced distortions of the base pair stack could be easily detected (23). The perturbations included DNA bending by the TATA-binding protein transcription factor (TBP), base flipping by the Hha I methyltransferase (M.Hha I) and uracil deglycosylase (UDG), and cutting by the Pvu II restriction enzyme (R.Pvu II). These proteins were chosen for detection experiments since they had all been extensively characterized from a biochemical standpoint. These early efforts hinted at the potential of assays based on DNA CT for the high throughput, multiplexed electrical monitoring of numerous DNA binding proteins on a single chip. 1.5.1 Detection of Base Flipping Enzymes The electrical detection of methyltransferases, base flipping enzymes which catalyze the transfer of a methyl group from S-adenosylmethionine to adenine or cytosine, is particularly noteworthy. Establishing a base flipping mechanism was historically quite difficult, usually requiring a crystal structure, but the electrical detection of base flipping using DNA electrochemistry has been found to be a sensitive, rapid, and attractive method for such characterization. While methyl substitution does not appreciably perturb the base pair stack, methyltransferases, in carrying out the base flip, significantly attenuate DNA-mediated CT. Based upon this attenuation, we have conducted a detailed exploration of base flipping by HhaI methyltransferase upon binding to its cognate sequences (23). At a surface featuring a DNA-DM monolayer, for example, the addition of M.HhaI disrupts the integrity of the base pair stack with a concomitant decrease in the DM redox signal. From the crystal structure of the protein/DNA complex, it is apparent 25 that upon base flipping the internal cytosine of the recognition sequence 5’-GCGC-3’, M.HhaI also intercalates the non-aromatic residue Gln 237 into the base pair stack, filling the space occupied by the flipped out cytosine and interrupting π-stacking within the duplex. With this in mind, electrochemistry experiments were conducted with a Q237W mutant enzyme, since it was expected that the mutant protein would insert the aromatic tryptophan residue into the π-stack upon base flipping, leading to little signal loss. Indeed, minimal attenuation of the DM signal upon binding the mutant protein was observed. Moreover, to establish that the lack of attenuation was the result of restoration of the πstack and not simply poor DNA binding by the mutant, we also examined binding of both wild type and mutant M.HhaI to a DNA-DM monolayer containing an abasic site at the position of what would be the flipped out cytosine. On this DNA-DM monolayer, the DM redox signal is only weakly detected both in the absence of protein and in the presence of wild type M.HhaI, owing to the poor stacking of the duplex with an intervening abasic site. Nonetheless, in the presence of the Q237W mutant, the DM redox signal is restored. This enhanced signal for DM on binding the mutant reflects insertion of the tryptophan from the mutant protein within the base pair stack, so as to restore proper stacking in the DNA-DM duplex. Hence DNA-binding proteins are seen to modulate DNA CT both positively and negatively, depending upon how they affect the conformation of the DNA. 1.5.2 Detection of Transcription Factors that Distort DNA Our early studies at DNA-DM monolayers demonstrated that we could easily detect TATA-binding protein (TBP), a transcription factor that bends duplex DNA by ~ 90º upon binding (23). The addition of TBP to a DNA-modified surface results in distortion of the DNA and lowers the yield of DNA CT (Figure 1.7). Consequently, this 26 transcription factor was chosen as a model system because of its central role in the eukaryotic/prokaryotic transcription machinery, excellent stability, simple 5’-TATA-3’ binding site, and well-known structure. In Chapter 8, the detection of TBP with a scanning probe technique is demonstrated (61). This assay exploits the interruption of a DNA-mediated, bimolecular reaction between DNA-bound NB and ferricyanide, which is generated at the SECM tip. The SECM facilitates the exploration of the morphology of NB-modified DNA monolayers in the presence and absence of proteins which do not bind DNA, such as BSA, and proteins which distort the DNA, such as TBP. In fact, attomoles of TBP, as estimated from the footprint of the SECM tip, are detected only at DNA monolayers which feature the 5’-TATA-3’ binding site. The detection of TBP due to interruption of DNA CT at DNA-modified microelectrodes is described in Chapter 9 (60). At DNA monolayers modified with Nile Blue, TBP can be readily detected at both the macro- and microscales. However, microelectrodes allow for the rapid detection of nanomolar concentrations of this transcription factor even from complicated mixtures containing bovine serum albumin, EndonucleaseIII, or Bam HI methyltransferase. The sequence-specific detection of TBP from a cell lysate is explored in Chapter 10 via a multiplexed format. The manufacture and individual addressing of multiple electrodes on a single chip affords high throughput experiments and excellent statistics. The detection of TBP is accomplished in two ways: 1) a cell lysate solution is intentionally supplemented with TBP and 2) the transcription factor is overproduced in mouse epithelial cells through treatment with epidermal growth factor. In both cases, 27 highly accurate detection can be achieved, emphasizing the potential of this methodology for rapid electrical detection of numerous DNA binding proteins from complicated, real world mixtures. 1.5.3 Electrical Monitoring of Photolyase in Real Time Another advantage of monitoring protein binding to DNA electrically comes from the fact that this technique allows for the sensitive detection of events in real time; this is best exemplified by the electrical monitoring of thymine dimer repair by photolyase (62). Photolyase is a repair enzyme from E. coli containing a flavin cofactor that binds pyrimidine dimers in duplex DNA and repairs them, upon photoactivation, in a reaction involving electron transfer from the reduced flavin cofactor to the dimer flipped out of the DNA duplex by the enzyme. Importantly, in monitoring photolyase reactions electrochemically on DNA monolayers, an externally bound DM redox probe was not utilized; the redox signal of the flavin cofactor, which is intimately associated with the DNA, was monitored instead. Upon binding of photolyase to a DNA monolayer containing the thymine dimer, a weak redox signal arising at 40 mV versus NHE was first observed. Through control studies with apoprotein and a mutant, this redox signal was assigned to the flavin, and its small size was attributed to poor coupling with the DNA, owing to the intervening destacked thymine dimer. The repair reaction was subsequently carried out by illuminating the electrode surface. With irradiation, an increase of almost an order of magnitude in the integrated flavin signal was observed, and repair of the thymine dimer was confirmed via HPLC analysis. Thus, the repair of the thymine dimer by photolyase was monitored in real time, as a function of irradiation; as the photolyase repairs the 28 thymine dimer, the base pair stack is restored, the flavin cofactor becomes well coupled into the base stack, and its signal is enhanced (Figure 1.8). Subsequently, it was observed that once the DNA is repaired, the protein dissociates from the monolayer, and the intensity of the signal is diminished. Overall, the photolyase experiments illustrated not only how DNA electrochemistry could be used to assess DNA conformational changes associated with protein binding, but also the utility of DNA electrochemistry for measuring the redox potentials of DNA-binding proteins in their biologically relevant forms. 1.5.4 Electrochemistry of Repair Enzymes Containing Iron Sulfur Clusters DNA electrochemistry has also been more generally applied in assessing the redox characteristics of other DNA-binding proteins in their active, DNA-bound form. An illustrative example is found in the characterization of the redox chemistry associated with two base excision repair (BER) enzymes: MutY, a glycosylase that excises adenine from 8-oxoguanine:adenine mispairs, and Endonuclease III, a glycosylase that targets oxidized pyrimidines (63–70). Crystal structures of these enzymes indicated that they both contain an iron-sulfur cluster with the C-X6-C-X2-C-X5-C motif in close proximity to the DNA backbone (64–67). This Fe4S4 motif, moreover, is conserved in base excision repair proteins from bacteria to man. Interestingly, in MutY, this cluster is not necessary for protein folding or stability but is crucial for DNA binding and in vivo activity (68). EPR experiments have demonstrated that the cluster is in its [4Fe-4S]2+ form when isolated from the cell, but the protein cannot be reduced within a physiologically relevant window nor oxidized by ferricyanide without degradation 29 Figure 1.8: Schematic illustration of the electrochemical monitoring of the repair of thymine dimers in a DNA monolayer by photolyase. In the absence of photolyase, no signal is observed, but when the protein is added, a weak signal is observed due to binding of the T^T dimer. Control studies show that the signal corresponds to the flavin bound within the photolyase. Upon irradiation, the thymine dimer is repaired, the integrity of the base pair stack is restored, and the signal grows in intensity. Finally, because the protein has a lower affinity for undamaged DNA, it dissociates, leading to the slow loss of the protein redox signature. 30 (69, 70). However, these studies, which suggested a lack of redox activity associated with the reactivity of the proteins, were conducted in the absence of DNA. Significantly, we found that EndoIII and MutY display stable, quasi-reversible signals with potentials of ~ 90 mV versus NHE (71, 72), at DNA-modified and backfilled surfaces, and importantly, these signals were not observed at surfaces backfilled with mercaptohexanol but lacking DNA. Furthermore, weak signals were found on surfaces modified with DNA duplexes containing an abasic site, demonstrating that the observed redox processes for the proteins are DNA-mediated. As illustrated in Chapter 6, experiments performed on DNA-modified HOPG allow for the direct comparison of the electrochemistry of EndoIII in the presence and absence of DNA (73). Qualitatively similar results are obtained in the presence of DNA on gold and HOPG. Due to its extended potential window, however, graphite allows for the study of the electrochemistry of EndoIII also in the absence of DNA, a measurement made impossible by the potential window limit on gold. The ability to measure the redox potential of both free and bound EndoIII on the same surface demonstrates that binding to DNA not only stabilizes the cluster but also shifts its redox potential. 1.5.5 Electrochemistry of an Iron Sulfur Transcription Factor Bound to DNA SoxR is a 17 kDa transcription factor that binds DNA as a dimer with a [2Fe-2S] cluster in each monomer (59, 74–76). The loss of the cluster does not affect protein folding, DNA binding, or promoter affinity, but oxidation of the cluster triggers the transcription of SoxS and the subsequent expression of over 100 other genes responsible for the oxidative stress response in bacteria. Interestingly, the redox potential of this cluster in the absence of DNA was determined as -290 mV vs. NHE. Importantly, this 31 measurement of the potential of SoxR led to a conundrum: given the reducing environment of the cell, the cluster would always be in its oxidized form, thereby making SoxR essentially constituitively activated in vivo. How does the cell keep this protein reduced and transcriptionally silent unless required? Since DNA-modified surfaces allow us to determine the redox potential of proteins in their DNA-bound forms, we explore the electrochemistry of the SoxR transcription factor on DNA-modified HOPG in Chapter 7. Such an exploration is particularly relevant given that significant structural differences were recently found between the free (low energy) and DNA-bound (high energy) forms of SoxR (76). In addition, to facilitate these experiments, self-assembled DNA monolayers are prepared on graphite with an appended redox-active probe, Redmond Red (59). Redmond Red provides a convenient internal reference for the voltammetry of the SoxR [2Fe-2S] cluster. Therefore, the redox potential of SoxR is conclusively determined in its physiologically relevant DNA-bound form, shedding light on how the cell regulates the activity of this protein. These data further underscore the importance of characterizing the redox chemistry of DNA-binding proteins in their biologically relevant, DNA-bound forms. Base excision repair enzymes and transcription factors both represent an interesting target for study, given the ubiquity of the iron-sulfur clusters in these enzymes with no apparent function. Importantly, these are only two examples, and other DNA-binding enzymes that contain a potential redox cofactor may also take advantage of this chemistry. Electrochemistry at self-assembled DNA monolayers thus provides a first assay to 32 characterize redox chemistry associated with DNA-binding, which may have relevance within the cell. 1.6 Conclusion The intrinsic ability of DNA to conduct charge along its base pair stack, and the remarkable sensitivity of this chemistry to the integrity of the DNA duplex has been an intriguing topic of study for over two decades. A combination of detailed structural characterization and careful experimental design has allowed the investigation of this chemistry on DNA-modified surfaces. Through a host of electrochemical studies, we have not only uncovered several fundamental properties of this chemistry but have also demonstrated its potential use in a variety of biosensing and biochemical applications. DNA-modified surfaces offer an ideal platform on which to investigate CT. Structural studies have provided information on the morphology of these self-assembled monolayers, which are found to orient themselves upright and away from the surface, thereby creating an accessible conduit for the transport of charge. We have previously employed several different redox-active, DNA-bound molecules to serve as sensitive reporters of this process. While earlier work largely involved non-covalent intercalating moieties, subsequent studies have taken advantage of the precise control and improved electronic coupling afforded by covalent attachment. In all cases, DNA-mediated CT displays an exquisite sensitivity to the stacking of the base pairs that intervene between the electrode surface and the pendant probe, allowing for the electrochemical discrimination of single base mismatches, lesions, and conformational perturbations. 33 Finally, due to the possible DNA conformational changes and redox chemistry associated with protein binding, we have also been able to demonstrate the sensitive detection of protein/DNA interactions. In this thesis, we build upon these previous explorations of DNA CT. In Chapter 2, we demonstrate the single molecule conductivity of DNA, and its applications to the detection of protein binding. DNA monolayers on HOPG are introduced in Chapters 3–7. New covalent probes of DNA CT are explored in Chapters 5, 7, and 8–10. The electrochemistry of proteins containing redox active cofactors is described in Chapters 6 and 7. In addition, the detection of transcription factors that perturb the base pair stack at self-assembled DNA monolayers is explored in Chapters 8–10. Taken together, these studies highlight even further the utility of DNA monolayers for the exploration of DNA CT. Much remains to be discovered with regard to the scope and fundamentals of this process. Can we harness this chemistry in the construction of highly sensitive nanoscale devices to detect not only proteins but also RNA in real time in a multiplexed chip format? Are we, in fact, simply mimicking what Nature already exploits for long-range signaling among proteins in the cell that utilize DNA CT? Many more questions remain to be addressed, yet it is abundantly clear that DNA CT and, more specifically, DNA electrochemistry, represent a remarkably rich field for study. 34 REFERENCES (1) Szent-Gyorgyi, A., Isenberg, I., Baird, S. L. (1960) Proc. Natl. Acad. Sci. U. S. A. 46, 1444–1449. (2) Eley, D. D., Spivey, D. I. (1962) Trans. Faraday Soc. 58, 411–415. (3) Boon, E. M., Barton, J. K. (2002) Curr. Opin. Struc. Biol. 12, 320–329. (4) Delaney, S., Barton, J. K. (2003) J. Org. Chem. 68, 6475–6483. (5) O’Neill, M. A., Barton, J. K. (2005) Charge Transfer in DNA: From Mechanism to Application (H.-A. Wagenknecht) pp. 27–75, Wiley–VCH, Hoboken. (6) Giese, B. (2000) Acc. Chem. Res. 33, 631–636. (7) Schuster, G. B. (2000) Acc. Chem. Res. 33, 253–260. (8) Murphy, C. J., Arkin, M. R., Jenkins, Y., Ghatlia, N. D., Turro, N. J., Barton, J. K. (1993) Science 262, 1025–1029. (9) Kelley, S. O., Holmlin, R. E., Stemp, E. D. A., Barton, J. K. (1997) J. Am Chem. Soc. 119, 9861–9870. (10) Hall, D. B., Holmlin, R. E., Barton, J. K. (1996) Nature 382, 731–735. (11) Nunez, M. E., Hall, D. B., Barton, J. K. (1999) Chem. Biol. 6, 85–97. (12) Bhattacharya, P. K., Barton, J. K. (2001) J. Am. Chem. Soc. 123, 8649–8656. (13) Williams, T. T., Dohno, C., Stemp, E. D. A., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 8148–8158. (14) Merino, E. J., Boal, A. K., Barton, J. K. (2008) Curr. Opin. Chem. Biol. 12, 229–237. 35 (15) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nature Biotechnol. 21, 1192–1199. (16) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. (17) Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G., Whitesides, G. M. (2005) Chem. Rev. 105, 1103–1169. (18) Fink, H. W., Schonenberger, C. (1999) Nature 398, 407–410. (19) Porath, D., Bezryadin, A., de Vries, S., Dekker, C. (2000) Nature 403, 635– 638. (20) Kasumov, A. Y., Kociak, M., Gueron, S., Reulet, B., Volkov, V. T., Klinov, D. V., Bouchiat, H. (2000) Science 291, 280–282. (21) Guo, X., Gorodetsky, A. A., Hone, J., Barton, J. K., Nuckolls, C. (2008) Nature Nanotechnol. 3, 163–167. (22) Kelley, S. O., Barton, J. K., Jackson, N. M., Hill, M. G. (1997) Bioconjugate Chem. 8, 31–37. (23) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nature Biotechnol. 20, 282– 286. (24) Gore, M. R., Szalai, V. A., Ropp, P. A., Yang, I. V., Silverman, J. S., Thorp, H. H. (2003) Anal. Chem. 75, 6586–6592. (25) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. (26) Rant, U., Arinaga, K., Fujita, S., Yokoyama, N., Abstreiter, G., Tornow, M. (2004) Nano Lett. 4, 2441–2445. 36 (27) Rant, U., Arinaga, K., Fujita, S., Yokoyama, N., Abstreiter, G., Tornow, M. (2004) Langmuir 20, 10086–10092. (28) Arinaga, K., Rant, U., Tornow, M., Fujita, S., Abstreiter, G., Yokoyama, N. (2006) Langmuir 22, 5560–5562. (29) Ceres, D. M., Barton, J. K. (2003) J. Am. Chem. Soc. 125, 14964–14965. (30) Hihath, J., Xu, B., Zhang, P., Tao, N. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 16979–16983. (31) Wierzbinski, E.; Arndt, J.; Hammond, W.; Slowinski, K. (2006) Langmuir 22, 2426–2429. (32) Kelley, S. O., Barton, J. K., Jackson, N. M., McPherson, L. D., Potter, A. B., Spain, E. M., Allen, M. J., Hill, M. G. (1998) Langmuir 14, 6781–6784. (33) Boon, E. M., Sam, M., Barton, J. K., Hill, M. G., Spain, E. M. (2001) Langmuir 17, 5727–5730. (34) Zhou, D., Sinniah, K., Abell, C., Rayment, T. (2002) Langmuir 18, 8278– 8281. (35) Zhou, D., Sinniah, K., Abell, C., Rayment, T. (2003) Angew. Chem. Int. Ed. 42, 4394–4397. (36) Erts, D., Polyakov, B., Olin, H., Tuite, E. (2003) J. Phys. Chem. B. 107, 3591– 3597. (37) Turcu, F., Schulte, A., Hartwich, G., Schuhmann, W. (2004) Angew. Chem. Int. Ed. 43, 3482–3485. (38) Liu, B., Li, C.-Z., Kraatz, H.-B., Bard, A. J. (2005) J. Phys. Chem. B. 109, 5193–5198. 37 (39) Wain, A. J., Zhou, F. (2008) Langmuir 24, 5155–5160. (40) Yang, X., Wang, Q., Wang, K., Tan, W., Yao, J., Li, H. (2006) Langmuir 22, 5654–5659. (41) Li, Y.-J., Xiang, J., Zhou, F. (2007) Plasmonics 2, 79–87. (42) Steel, A. B., Herne, T. M., Tarlov, M. J. (1998) Anal. Chem. 70, 4670–4677. (43) Yu, H.-Z., Luo, C.-Y., Sankar, D. G., Sen, D. (2003) Anal. Chem. 75, 3902– 3907. (44) Ceres, D. M., Udit, A. K., Hill, H. D., Hill, M. G., Barton, J. K. (2007) J. Phys. Chem. B. 111, 663–668. (45) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucl. Acids Res. 27, 4830–4837. (46) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (47) Boon, E. M., Jackson, N. M., Wightman, M. D., Kelley, S. O., Barton, J. K., Hill, M. G. (2003) J. Phys. Chem. B 107, 11805–11812. (48) Boon, E. M., Barton, J. K., Bhagat, V., Nersissian, M., Wang, W., Hill, M. G. (2003) Langmuir 19, 9255–9259. (49) Boal, A. K., Barton, J. K. (2005) Bioconjugate Chem. 16, 312–321. (50) Kelley, S. O., Jackson, N. M., Hill, M. G., Barton, J. K. (1999) Angew. Chem. Int. Ed. 38, 941–945. (51) Drummond, T. G., Hill, M. G., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 15010–15011. (52) Liu, T., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 10160–10161. 38 (53) Gorodetsky, A. A., Barton, J. K. (2007) J. Am. Chem. Soc. 129, 6074–6075. (54) Di Giusto, D. A., Wlasoff, W. A., Giesebrecht, S., Gooding, J. J., King, G. C. (2004) J. Am. Chem. Soc. 126, 4120–4121. (55) Di Giusto, D. A., Wlasoff, W. A., Giesebrecht, S., Gooding, J. J., King, G. C. (2004) Angew. Chem., Int. Ed. 116, 2809–2812. (56) Okamoto, A., Kamei, T., Tanaka, K., Saito, I. (2004) J. Am. Chem. Soc., 126, 14732–14733. (57) Okamoto, A., Kamei, T., Saito, I. (2006) J. Am. Chem. Soc. 128, 658–662. (58) Gorodetsky, A. A., Green, O., Yavin, E., Barton, J. K. (2007) Bioconjugate Chem. 18, 1434–1441. (59) Gorodetsky, A. A., Dietrich, L. E. P., Lee, P. E. Demple, B., Newman, D. K., Barton, J. K. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 3684–3689. (60) Gorodetsky, A. A., Ebrahim, A., Barton, J. K. (2008) J. Am. Chem. Soc. 130, 2924–2925. (61) Gorodetsky, A. A., Hammond, W., Hill, M. G., Slowinski, K., Barton, J. K. (2008) Langmuir ASAP. (62) Derosa, M. C., Sancar, A., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 10788–10792. (63) David, S.S., Williams, S.D. (1998) Chem. Rev. 98, 1221–1262. (64) Kuo, C. F., McRee, D. E., Fisher, C. L., O’Handley, S. F., Cunningham, R. P., Tainer, J. A. (1992) Science 258, 434–440. (65) Thayer, M. M., Ahern, H., Xing, D., Cunningham, R. P., Tainer, J. A. (1995) EMBO J. 14, 4108–4120. 39 (66) Fromme, J. C., Verdine, G. L. (2003) EMBO J. 22, 3461–3471. (67) Fromme, J. C., Banerjee, A., Huang, S. J., Verdine, G. L. (2004) Nature 427, 652–656. (68) Porello, S. L., Cannon, M. J., David, S. S. (1998) Biochemistry 37, 6465– 6475. (69) Cunningham, R. P., Asahara, H., Bank, J. F., Scholes, C. P., Salerno, J. C., Surerus, K., Munck, E., McCracken, J., Peisach, J., Emptage, M. H. (1989) Biochemistry 28, 4450–4455. (70) Fu, W., O’Handley, S., Cunningham, R. P., Johnson, M. K. (1992) J. Biol. Chem. 267, 16135–16137. (71) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 12543–12547. (72) Boal, A. K., Yavin, E., Lukianova, O. A., O’Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (73) Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. (74) Pomposiello, P. J., Demple, B. (2001) Trends Biotechnol. 19, 109–114. (75) Demple, B., Ding, H., Jorgensen, M. (2002) Methods Enzymol. 348, 355–364. (76) Watanabe, S., Kita, A., Kobayashi, K., Miki, K. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 4121–4126. 40  Chapter 10: Electrochemical Detection of TATA-Binding Protein from a Cell Lysate This work was performed in collaboration with Jason Slinker, Ali Ebrahim, and Russell Ernst. The multiplexed arrays were fabricated by Jason Slinker. 270 ABSTRACT A simple and high-throughput method for the sequence-specific electrochemical detection of TATA-binding protein from complicated mixtures is demonstrated at DNAmodified gold electrodes. The assay is general and based on the interruption of DNAmediated charge transport to Nile Blue (NB), a redox-active probe covalently attached to the DNA base pair stack. In cell lysate solutions containing nanomolar conentrations of human TBP, we observe a signal attenuation of the NB signal for DNA-modified surfaces containing the TBP binding site; the attenuation is diminished at GC-rich DNA monolayers. Furthermore, in solutions containing murine epithelial cell lysate, we demonstrate the differential detection of upregulated cellular TBP levels with high sequence-specificity. This methodology provides a basis for the sensitive electrical detection of numerous DNA-binding proteins from clinical isolates and on a single DNA chip. 271 10.1 Introduction Protein-DNA interactions play a central role in the regulation of nearly all cellular processes including transcription, replication, recombination, and repair. Therefore, the development of sensitive, high-throughput methods to detect the specific binding of proteins to DNA is essential in monitoring pathways for gene expression and in the development of new diagnostics of disease states (1,2). In fact, the examination of gene expression profiles in tumor cells (relative to healthy tissue) can identify “signatures” which forecast the onset of disease and guide the implementation of cancer treatments (3). As such, the reliable detection and identification of properly/improperly functioning DNA binding proteins becomes a crucial area for research and development. Currently, the most technologically viable methodologies for the detection of protein binding to DNA are based on fluorescence or mass spectrometry. Mass spectrometry, the workhorse of proteomics research, is a harsh technique which degrades the sample during the measurement, making it unsuitable for monitoring delicate, noncovalent protein-DNA interactions (4). By comparison, fluorescent techniques are more gentle and include 1) immunoprecipitation of reversibly crosslinked chromatin/protein complexes (5), 2) capture of fluorescent antibodies by epitope-tagged DNA-binding proteins (6), and 3) amplification of sequences methylated by enzyme/DNA adeninemethyltransferase fusions (7). However, all of these fluorescent methods rely on surface chemistry featuring high lot-to-lot variability and cumbersome optical detection (all within the confines of microbead, microwell, or glass slide formats). In addition, both DNA amplification and extensive labeling of proteins/antibodies are frequently necessary 272 for fluorescence and mass spectrometry, adding to the expense of these techniques and making them poorly suited for high-throughput assays. The electrical monitoring of protein/DNA interactions provides a convenient alternative to technologies based on both fluorescence and mass spectrometry (8–13). In fact, a number of recent studies which utilized DNA-functionalized surfaces have demonstrated the feasibility of this approach at small scales (10–13). Within this context, electrochemical assays based on DNA-mediated charge transport (DNA CT) represent a particularly reliable detection strategy for monitoring protein binding at self-assembled DNA monolayers (14-18). Importantly, such assays allow for the interrogation of proteins in their DNA-bound and, therefore, biologically relevant, forms (18). In a typical electrochemical protein detection experiment, loosely packed DNA monolayers with a covalently appended, electroactive probe are self-assembled in the absence of Mg2+, and the surface is backfilled with an alkanethiol such as mercaptohexanol (9). The resulting DNA films have been characterized extensively with both atomic force microscopy (9, 19) and scanning electrochemical microscopy (20), and the films possess a well-defined morphology with the DNA in an upright orientation, thereby leaving room for proteins to bind. Typically, voltammetry of the DNA-modified electrode is recorded in the absence and presence of protein. Since DNA CT is exquisitely sensitive to perturbation of the base pair stack, detection of protein binding is confirmed by a decrease in the electrochemical yield of charge transfer to a distallybound probe. Previously, we have demonstrated the in vitro electrical detection of nanomolar concentrations of TATA-binding protein (TBP) not only at DNA-modified macro- and 273 micro-electrodes (9, 10) but also with scanning electrochemical microscopy (20). TBP is a ubiquitous eukaryotic/prokaryotic protein which plays a central role in the transcriptional machinery as part of the transcription factor IID (TFIID) multi-subunit complex (21–23). Crystal structures of TBP indicate that it resembles a molecular saddle whose concave and hydrophobic underside interacts with the minor groove of DNA, thereby bending the duplex by ~ 90º upon binding (21–23). This unusual binding mode ensures that TBP specifically binds flexible AT-rich tracts, such as its 5’-TATAAAG3’consensus sequence, and is particularly amenable to detection through attenuation of DNA CT upon binding DNA. Here we exploit the sensitivity of DNA-modified electrodes for probing protein/DNA interactions to detect the induction of TBP from mouse epithelial cell lysates. Differential expression of TBP is achieved via treatment of epithelial cells with epidermal growth factor (EGF), mimicking the situation encountered in tumorigenesis. By using a multiplexed array which can be independently addressed with DNA featuring a pendant Nile Blue (NB) redox-active probe, we demonstrate the sequence specific detection of the differential expression of TBP. These data underscore the unique technological opportunities presented by electrochemistry at self-assembled DNA monolayers for the reliable detection of DNA-binding proteins. 274 10.2 Experimental 10.2.1 Materials All reagents for DNA synthesis were purchased from Glen Research. All reagents for gel electrophoresis were purchased from BioRad, Inc. All antibodies were obtained from AbCam, Inc., and all other reagents for immunoblotting were obtained from Invitrogen, Inc. Mercaptohexanol and buffer preparation reagents were purchased from Sigma in the highest purity available and used as received. Nile Blue perchlorate was purchased from Acros in laser grade purity. Mammalian protein extraction reagent was purchased from Pierce, Inc. Human TATA-binding protein was purchased from Protein One, Inc. Silicon wafers were purchased from Silicon Quest International, Inc. (Santa Clara, CA). Sodium phosphate buffers were prepared with Milli-Q water and pH adjustments were made with sodium hydroxide, if necessary. 10.2.2 Synthesis of Thiol and Nile Blue-Modified DNA Oligonucleotides were prepared on solid support using standard phosphoramidite chemistry on an Applied Biosystems 394 DNA synthesizer. The pure oligonucleotides were characterized with matrix-assisted laser desorption (MALDI) mass spectrometry, UV-visible (UV-Vis) spectrophotometry, and multiple rounds of high-performance liquid chromatography (HPLC). Thiol-terminated oligonucleotides were synthesized according to established protocols from Glen Research, Inc., using the C6 S-S thiol modifier. After deprotection and cleavage from solid support with ammonium hydroxide (60 ºC for 8 hours), the disulfide-containing DNA was purified by HPLC. The disulfide was subsequently 275 reduced with an excess of dithiothreitol in ammonium acetate buffer at pH = 8, and the free thiol containing single-stranded DNA was then purified with a second round of HPLC. DNA modified with NB at the 5’ terminus was prepared according to ultra-mild protocols (Glen Research, Inc.) to avoid degradation of the NB moiety. A 17-mer sequence (5’-UGC GTG CTT TAT ATC TC-3’ or 5’-UGC GCG CCC GGC GCC TC-3’) was prepared on solid support with a 5-[3-acrylate NHS Ester]-deoxy uridine as the terminal 5’ base. The beads were then removed from the synthesizer and dried thoroughly. The solid support were reacted with a 10 mg/mL Nile Blue perchlorate solution in either 9:1 N,N-dimethylformamide/N,N-diisopropylethylamine or 9:1 dichloromethane/N,Ndiisopropylethylamine for 12–48 hours. The beads were subsequently washed up to three times with dichloromethane or N,N-dimethylformamide, methanol, and acetonitrile. Subsequently, the Nile Blue-containing sequence was simultaneously cleaved from the support and deprotected with 0.05 M potassium carbonate in methanol at ambient temperature for 12–14 hours. The overall yields of the reaction ranged from 30% to 80%. Both thiol- and NB-modified DNA was quantified using the extinction coefficient of the single-stranded DNA at 260 nm on a Beckman UV-Vis spectrophotometer. Equimolar amounts of each strand were combined before the resulting solution was purged with argon. Duplexes were formed by thermally annealing to 90 ºC in deoxygenated pH = 7.1 phosphate buffer containing 5 mM NaPi and 50 mM NaCl, followed by cooling to ambient temperature. 276 10.2.3 Preparation of Epithelial Cell Lysates Murine epithelial (JB6) cell lysates were prepared according to the method of Johnson and coworkers (24). In brief, JB6 cells were grown to 80% confluency in 5% fetal bovine serum (FBS) in MEM and then serum deprived using 0.1% FBS in MEM for 3 hours. To induce upregulation of TATA binding protein, cells were then incubated with 50 ng/mL of EGF for 30 to 60 minutes prior to lysis. The cells were washed with phosphate-buffered saline twice and lysed with ~ 1–3 mL of the mammalian protein extraction reagent according to the manufacturer’s instructions. 10.2.4 Gel Electrophoresis and Immunoblot Analysis Denaturing polyacrylamide gel electrophoresis was performed according to established protocols from BioRad. Fresh lysates were loaded onto a denaturing 12% Tris-HCl gel in a 1:1 ratio with Laemmli loading buffer. The gel was electrophoresed for 35 minutes at 200 V using the standard SDS Tris-HCl buffer system. Western blotting was performed according to established protocols from Invitrogen and AbCam. A freshly electrophoresed gel was transferred to a nitrocellulose membrane via electroblotting for 7 minutes at 30 V with the iBlot gel transfer apparatus available from Invitrogen, Inc. After transfer, the membrane was blocked for 1 hour by incubation in a 5% milk/Tris-buffered saline with Tween solution (TBS-T). The membrane was washed with TBS-T and incubated with an anti-TBP antibody for 1–2 hours at room temperature. After several washes with TBS-T, the membrane was incubated with fluorescein-conjugated goat anti-mouse IgG for 1–2 hours in TBS-T. The membrane was then imaged with a Typhoon 3600 fluorescence imaging system. As a 277 loading control, the membrane was then blotted with a fluorescein-conjugated anti-β actin antibody for 1–2 hours, washed, and imaged again. 10.2.5 Fabrication of Gold Electrodes Silicon chips bearing arrays of 16 Au electrodes (.02 cm2 area each) were fabricated by a two–layer process. In the first layer, the gold electrodes were patterned by a metal lift-off process. The second layer involved patterning Microchem SU-8 photoresist as an insulator to isolate the working areas of the electrodes on the chip surface. Microchem SPR 220 3.0 photoresist was spin-cast at 4000 rpm and baked. The photoresist was patterned with a Karl Suss MA6 contact aligner and a chrome photomask. Following post-exposure baking, wafers were developed in AZ 300 MIF developer and rinsed thoroughly with DI water. A 15 Å Ti adhesion layer and a 1000 Å Au layer were evaporated with a CHA Mark 50 electron beam evaporator. Wafers were then immersed in Microchem 1165 Remover overnight and sonicated as needed to complete lift-off. Subsequently, the wafers were thoroughly baked and cleaned by UV ozone treatment or oxygen plasma etching. Microchem SU-8 2002 was spin cast at 3000 rpm, baked and photopatterned as above. Wafers were developed in Microchem SU-8 Developer and baked for a permanent set of the photoresist. Wafers were diced into 1-inch by 1-inch chips by hand with a diamond scribe. 10.2.6 Preparation of Backfilled DNA Monolayers Immediately prior to incubation with DNA, gold surfaces were cleaned by sonication in acetone (3 rounds of five minutes each) and isopropanol (1 round of five minutes), followed by treatment with UV-ozone for 3 minutes. Multi-level wells over the electrodes were defined with custom-made viton rubber gaskets and a polypropylene 278 clamp secured to a test mount. The lower level of the well assembly provided a compression seal that divided the chip into four quadrants of four electrodes near the surface of the chip. This allowed for incubation with up to four distinct sequence of 25– 50 µM duplex NB-DNA solution in phosphate storage buffer. Monolayer formation was typically allowed to proceed in a humidified environment for a period of 24–48 hours. Upon completion of film formation, the cell was rinsed thoroughly with phosphate buffer to remove residual DNA before the DNA-modified surface was backfilled with 1.0 mM 1-mercaptohexanol in phosphate storage buffer/5% glycerol for 30–60 minutes. The electrodes and cells were rinsed thoroughly prior to electrochemistry experiments to ensure removal of residual alkanethiols. 10.2.7 Electrochemical Experiments Cyclic and square wave voltammetry experiments were performed by automated measurement with a CH760B Electrochemical Analyzer and a 16-channel multiplexer module from CH Instruments (Austin, TX). The chips were interfaced with these instruments with a custom built device mount bearing spring-loaded probe pins. Electrochemistry was recorded at ambient temperature in pH 7.1 phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol (storage buffer). The custom-manufactured well assembly encompassed all quadrants and electrodes and provided an additional 500 µL of solution capacity. Gold working electrodes in this custom-built cell allowed for experiments with a common Pt auxiliary electrode and a common silver/silver chloride (Ag/AgCl) reference electrode. 279 Figure 10.1: Illustration of the self-assembly of DNA monolayers on gold and the subsequent electrochemical analysis of protein binding. First, loosely packed DNA monolayers are formed and the surface is backfilled with mercaptohexanol for passivation. In the absence of TBP, DNA-mediated redox chemistry to the DNA-bound NB is observed. Binding of a protein interrupts the base pair stack and the DNAmediated redox chemistry, leading to attenuation of the redox signal. 280 10.3 Results 10.3.1 Experimental Strategy for the Detection of TATA Binding Protein Figure 10.1 illustrates the strategy for detection of TATA binding protein from cell lysates at DNA monolayers modified with NB-DNA. DNA duplexes containing the 5’-TATAAAG-3’ TBP binding site are prepared by hybridizing thiol-modified singlestranded DNA with its NB-modified complement. The duplexes are self-assembled on a clean gold surface in the absence of Mg2+ and the surface is backfilled; the resulting monolayer leaves room for the protein to bind. Background voltammetry is recorded initially, and the freshly prepared electrode is subsequently incubated with a small amount of a cell lysate sample before the voltammetry is recorded again. The solution is then supplemented with either human TBP or a lysate sample containing a controlled amount of murine TBP, and the voltammetry is recorded for a third time. 10.3.2 Nile Blue as a Redox Probe for the Detection of Protein Binding Figure 10.2 shows the cyclic voltammetry (CV) of a DNA monolayer modified at its periphery with NB, which provides a particularly convenient internal standard for the detection of protein binding. DNA monolayers modified with NB display a DNAmediated redox couple at ~ -200 mV versus Ag/AgCl due to a surface bound species (20). Notably, modification of the exocyclic amine of NB perturbs its aromatic core with an accompanying shift in its redox potential of ~ 200 mV relative to free NB (25, 26). Therefore, DNA monolayer stability can be assessed in situ, and the appearance of a NB signal at ~ -400 mV indicates DNA degradation or proteolysis. 281 A. B. Figure 10.2: CV (A) and SWV (B) of a NB-DNA-modified electrode with a 0.02 cm2 area at 50 mV/s and 15 Hz, respectively. The electrochemistry buffer is pH=7.1, 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, and 10% glycerol. The sequence is thiol-5'-GAGATATAAAGCACGCA-3' plus NB-modified complement. Note that the integrated surface coverage found here is ~ 6 pmol/cm2. The potentials are reported vs. Ag/AgCl. 282 For somewhat imperfect NB-DNA monolayers, a small signal at ~ -100 mV is observed in addition to the DNA-mediated redox peak at ~ -200 mV. The signal at -100 mV is due to non-specifically bound NB-DNA, as indicated in titration experiments with DNA featuring NB but lacking a thiol (27). This minor side peak provides an additional indicator of film quality prior to protein detection experiments. The redox couple of NB also serves as an internal standard for changes in buffer conditions due to increased concentrations of lysate. Upon incubation of the electrode with a cell lysate solution, we observe shifts of ~ 20 to ~ 60 mV towards more negative potentials. The pH sensitivity of DNA-bound NB affords a high degree of consistency among different detection experiments and further underscores the attractiveness of NB as an electrochemical probe. 10.3.3 Self Assembly of Nile Blue-DNA Monolayers Self-assembly of NB-modified DNA monolayers on the gold surfaces of the chips fabricated here is highly reliable and reproducible, as shown in Figure 10.3. In sampling more than 140 electrodes, we find an essentially Gaussian distribution of signal sizes by integration of the cathodic CV peak of NB. The integrated charge ranges from 0 to 6.5 nC with an average value of 4 nC, which corresponds a DNA surface coverage of 1.0 (± 0.4) pmol/cm2. For comparison, the distribution of signal sizes found for standard rod electrodes cleaned via acid etching and tested under identical conditions is also shown in Figure 10.3. Relative to the freshly deposited gold surfaces, the average DNA coverage drops by a factor of two for rod electrodes (sample size of 14). Surprisingly, although we have observed surface coverages of up to ~ 6 pmol/cm2 for NB-DNA monolayers (Figure 10.2), our typical coverage falls far short of the coverage of 12 pmol/cm2 found by 283 20 15 10 5 6 6. 5 5 5. 5 4. 5 4 3 3. 5 2. 5 2 1 1. 5 0 0 0. 5 Number of Electrodes 25 Integrated Charge (nC) Figure 10.3: An illustration of the distribution of integrated surface charge (cathodic wave) found for NB-DNA monolayers on different gold electrodes prepared under identical conditions. More than 140 electrodes prepared by evaporation of gold on silicon are shown in blue, and 14 electrodes prepared by acid etching of rod electrodes are shown in red. The average value is 4 nC. For ease of comparison, integrated charges have been rounded to the nearest half coulomb. Note that on average a much higher signal and fewer failures are found for gold electrodes fabricated on silicon. 284 radioactive labeling (9). Nonetheless, the uniform signals obtained here afford excellent statistics which are essential for the proper interpretation of sequence-specific protein detection experiments. 10.3.4 Detection of Human TATA Binding Protein from Cell Lysate Solutions on a Single Substrate Figure 10.4 shows the effect of a cell lysate on the voltammetry of the NBmodified DNA monolayers on a single chip; a uniform signal loss is observed for films both with and without the TBP binding site. In general, treatment of DNA-modified surfaces with solutions containing high cell lysate concentrations (20% or greater) causes significant initial signal losses which vary from ~ 20 to ~ 50% but are sequence independent. In fact, at lysate concentrations of 20% or greater, a small but significant portion of the DNA films undergoes complete failure. This failure mode is not time dependent and does not typically correlate with DNA degradation; instead, large changes in the background capacitance accompanied are accompanied by complete signal loss. Given that the lysates utilized here were not processed or purified in any fashion, these failures likely result from catastrophic protein adsorption/precipitation onto the electrode surface. Figure 10.4 also illustrates the effect of 100 nM TBP on DNA monolayers with (TA-DNA) and without (GC-DNA) the TBP binding site in a 20% cell lysate solution and on a single chip. It is important to note that detection experiments with DNA monolayers on a single substrate afford a high degree of reliability and internal consistency. For TA-DNA monolayers featuring the recognition site, the voltammetry of DNA-bound NB is significantly attenuated, but for GC-DNA monolayers missing the 285 Figure 10.4: SWV of NB-DNA monolayers 15 Hz in dialysis buffer (black), after addition of 20% JB6 cell lysate (blue), and after addition of 100 nM TBP (red). The dialysis buffer is pH=7.1, 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, and 10% glycerol. The sequence is 5’-UGC GTG CTT TAT ATC TC-3’ plus thiol-modified complement (GC rich, left panel) and 5’-UGC GCG CCC GGC GCC TC3’ plus thiol-modified complement in (TA rich, right panel). The DNA-modified electrodes were prepared on the same substrate (center); half of the chip was modified with TA-rich DNA and half of the chip was modified with GC-rich DNA. Note that a greater signal attenuation is observed for the sequence containing the 5’-TATAAAG-3’ TBP binding site. 286 TBP recognition site, the voltammetry is affected to a lesser extent. These observations indicate the sequence-specific recognition of TBP. Upon protein binding, the DNA containing the recognition site is kinked, inhibiting DNA-mediated CT to NB, and a sequence-specific decrease is observed for the NB signal. 10.3.5 Detection of Murine TATA Binding Protein from Whole Cell Lysates To test the feasibility of our approach for the detection of TATA binding protein produced in vivo, detection experiments were also performed with higher lysate concentrations, as schematically illustrated in Figure 10.5. JB6 epithelial cells were treated with epidermal growth factor (EGF) to upregulate TATA binding protein, presumably through activation of the epidermal growth factor receptor (EGFR) (24). The cells were subsequently lysed, and the resulting solution was analyzed electrochemically. The effect of various cell lysates on TA-DNA and GC-DNA monolayers is summarized in Figure 10.6. To minimize the initial effect of non-specific adsorption, all of the films were initially tested in a 20% JB6 cell lysate solution. Subsequently, the addition of an 80 % by volume solution of EGF-treated epithelial lysate causes a statistically significant signal loss of 63 (±2) % at TA-DNA electrodes (sample size of 15 electrodes), and incubation of GC-DNA electrodes results in a diminished signal loss of 31 (±8) % (sample size of 15 electrodes). By comparison, incubation of TA-DNA- and GC-DNA-modified electrodes with an untreated epithelial lysate does not show a differential effect with signal losses of 18 (±4)% and 17 (±7)%, respectively (sample size of 15 electrodes each). As an additional control, an EGF-treated but heat inactivated cell lysate solution also yields consistent results. For a heat treated lysate, a signal decrease of 287 Lyse (MPER) Detection of TBP Treated With EGF Grow Murine Epithelial Cells Untreated Lyse (MPER) No Detection of TBP Figure 10.5: A schematic illustration of the growth and lysis of murine epithelial cells. A single batch of cells is split with half of the cells undergoing EGF treatment to upregulate TBP. The cell lysate is subsequently utilized for detection experiments. Sequence specific detection of TBP is achieved only in lysates containing upregulated levels of this transcription factor. 288 Percent Loss 80% 60% 40% 20% 0% TATA 80% EGF TATA 80% Untreated TATA 80% Heat EGF GC 80% EGF GC 80% Untreated GC 80% Heat EGF Figure 10.6: A representation of the signal loss found at NB-DNA-modified surfaces after addition of 80% JB6 epithelial cell lysate under various conditions. In each case, the buffer is an 8:2 ratio of cell lysate to pH=7.1, 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, and 10% glycerol. The sequence is 5’-UGC GTG CTT TAT ATC TC-3’ plus thiol-modified complement (blue, purple, and gray) and 5’-UGC GCG CCC GGC GCC TC-3’ plus thiol-modified complement in (red, teal, and peach). For the blue and red graphs, overexpression of TBP was induced with EGF. For the purple and teal graphs, overexpression of TBP was not induced. For the gray and peach graphs, overexpression of TBP was induced with EGF but the lysate was heat treated prior to experiments to inactivate TBP. The data was obtained from square wave voltammetry peak currents and represents a minimum of fifteen electrodes (blue, red, purple, teal) or seven electrodes (gray, peach). 289 25 (±9)% is found for GC-DNA-modified electrodes (sample size of 7), and a signal decrease of 13 (±2)% is found for TA-DNA-modified electrodes (sample size of 7). It is likely that differential effect observed upon heating simply represents the small sample size. Overall, these data strongly indicate the detection of upregulated TBP from a whole cell lysate. As confirmation of the overexpression of TBP in EGF-treated JB6 epithelial cells, four batches of lysate from independently grown cells underwent Western blotting analysis, as shown in Figure 10.7. The lysates utilized for the majority of the electrochemistry experiments presented above are shown in lanes 3 and 4. As expected, there is a faint band, which corresponds to the human TBP standard in lane 5, for the EGF-treated lysate in lane 4 of the right panel of Figure 10.7. In addition, the TBP band (MW = 42 kDa) is in close proximity to the β-actin loading control (MW = 38 kDa) in the left panel of Figure 10.7. Interestingly, the cells lysates for bands 1 and 2 were treated in an identical fashion to those of lanes 3 and 4, yet the β-actin loading controls indicate lower protein concentrations in lanes 1 and 2, making it difficult to quantify TBP. This indicates that some variation in TBP cellular concentrations is to be expected among lysates. Analysis of TBP levels by Western blotting also allows us to estimate the concentrations of TBP we are detecting. In untreated cells, if we assume a copy number of ~ 105 for TBP (28) and a cell count of ~ 5 million/mL, we obtain an effective TBP concentration of approximately ~ 1 nM, which is similar to the reported binding affinity of TBP (29, 30). The in vivo concentration is certain to be higher not only in EGF-treated 290 Figure 10.7: Visualization of TBP levels in cell lysates by Western blotting analysis of an SDS-PAGE Tris-HCl gel. A molecular weight ladder is shown for comparison, and a human TBP standard is shown in lane 5. Lanes 1 and 2 correspond to lysate from a single batch of cells, and lanes 3 and 4 correspond to the lysate from another batch of cells. The cells lysed for lanes 2 and 4 were subjected to EGF treatment, while the cells in lanes 1 and 3 were not treated. Note the increased level of TBP in lane 4, and the smaller amounts of actin (relative to lanes 3 and 4) in lanes 1 and 2. The left and right panels are from a single gel, and the black lines are intended as a guide to the eye. 291 cells but also in more concentrated lysates. Indeed, immunoblotting of TBP standards at various concentrations indicates that the TBP concentrations for the lysates utilized in our experiments may approach the ~ 10 nM range, lending further credence to our observations. 10.4 Implications and Conclusion We have now demonstrated the electrical detection of TBP binding from a cell lysate at DNA-modified surfaces. In lysate solutions supplemented with human TBP, a dramatic attenuation is observed at DNA-modified surfaces featuring the 5’-TATAAAG3’ consensus sequence, while little attenuation is found for surfaces modified with GC rich DNA. For epithelial cell lysates with upregulated levels of murine TBP, a differential effect of 30% was also observed. Taken together, these data support the assertion of a sequence specific recognition event. To emphasize the technological feasibility of electrical detection, we can compare the amount of cell lysate required for Western blotting and DNA electrochemistry. Each electrochemical experiment necessitates ~ 10–20 µL of sample per electrode, and each Western blot requires ~ 15 µL of cell lysate per lane. Therefore, the material consumed is comparable in both electrical and fluorescent detection, which is significant given the lower cost of electrical techniques. Through optimization of the electrode size and the electrode housing, it should even be possible to reduce the sample sizes to less than ~ 5 µL per electrode. In addition, some processing of the lysate prior to experiments would also improve on the observed differential detection (relative to background signal losses). 292 Such experimental challenges are certainly important for electrical detection and warrant further attention in future experiments. The results presented here are also particularly significant given that TBP plays a central role in the transcriptional machinery and changes in the in vivo levels of TBP dramatically influence gene expression. In epithelial cells, increased expression of TBP is triggered by EGFR, and consequently, TBP has been implicated in cellular transformation and oncogenesis (24, 31). Furthermore, higher levels of TBP have even been discovered in human tumors (relative to healthy tissue) indicating that it may play a role in human tumorigenesis (31, 32). Therefore, the development of an electrical strategy for the reliable detection of “native” TBP cellular levels (and by extension the levels of other transcription factors) has implications for the early detection of cancer. Our data generally underscore the utility of DNA-modified electrodes in detection schemes which take advantage of DNA CT. When compared to traditional fluorescence based detection techniques, electrochemical techniques present attractive features such as lower cost, portable equipment, and simplified multiplexing (8). In addition, the methodology outlined here is based upon an interruption of exquisitely sensitive DNAmediated electrochemistry, so it is applicable for the detection of a variety of proteins that perturb the base pair stack. Therefore, given our ability to rapidly and reliably detect protein binding to DNA from complicated mixtures, the sequence-specific electrical monitoring of numerous DNA binding proteins on a single platform may now be considered. 293 REFERENCES (1) Sikder, D., Kodadek, T. (2005) Curr. Opin. Chem. Biol. 9, 38–45. (2) Bulyk, M. L. (2006) Curr. Opin. Biotechnol. 17, 422–430. (3) van’t Veer, L. J., Bernards, R. (2008) Nature 452, 564–570. (4) Hanash, S. M., Pitteri, S. J., Faca, V. M. (2008) Nature 452, 571–579. (5) Lee, T. I., Johnstone, S. E., Young, R. A. (2006) Nat. Protoc. 1, 729–748. (6) Bulyk, M. L. (2007) Adv. Biochem. Engin./Biotechnol. 104, 65–85. (7) Vogel, M. J., Peric-Hupkes, D., van Steensel, B. (2007) Nat. Protoc. 2, 1467– 1478. (8) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nat. Biotechnol. 21, 1192–1199. (9) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nat. Biotechnol. 20, 282–286. (10) Gorodetsky, A. A., Ebrahim, A., Barton, J. K. (2008) J. Am. Chem. Soc. 130, 2924–2925. (11) Zheng, G., Patosky, F., Cui, Y., Wang, W. U., Lieber, C. M. (2005) Nat. Biotechnol. 23, 1294–1301. (12) Patolsky, F., Zheng, G., Lieber, C. M. (2006) Nanomedicine 1, 51–65. (13) Gore, M. R., Szalai, V. A., Ropp, P. A., Yang, I. V., Silverman, J. S., Thorp, H. H. (2003) Anal. Chem. 75, 6586–6592. (14) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U.S.A. 100, 12543–12547. 294 (15) Boal, A. K., Yavin, E., Lukianova, O. A., O'Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (16) DeRosa, M. C., Sancar, A., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 10788–10792. (17) Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. (18) Gorodetsky, A. A., Dietrich, L. E. P., Lee, P. E. Demple, B., Newman, D. K., Barton, J. K. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 3684–3689. (19) Erts, D., Polyakov, B., Olin, H., Tuite, E. (2003) J. Phys. Chem. B. 107, 3591– 3597. (20) Gorodetsky, A. A., Hammond, W., Hill, M. G., Slowinski, K., Barton, J. K. (2008) Langmuir ASAP. (21) Kim, Y., Geiger, J. H., Hahn, S. H., Sigler, P. B. (1993) Nature 365, 512–520. (22) Kim, J. L., Nikolov, D. B., Burley, S. K. (1993) Nature 365, 520–527. (23) Patikoglou, G. A., Kim, J. L., Sun, L., Yang, S.–H., Kodadek, T. K., Burley, S. K. (1999) Genes Dev. 13, 3217–3230 (24) Zhong, S; Zhang, C; Johnson, D. L. (2004) Molec. Cell. Biol. 2004 24 5119– 5129. (25) Schlereth, D. D., Schmidt, H.–L. (1995) J. Electroanal. Chem. 380, 117–125. (26) Sugawara, K., Yamauchi, Y., Hoshi, S., Akatsuka, K., Yamamoto, F., Tanaka, S., Nakamura, H. (1996) Bioelectrochem. Bioenerg. 41, 167–172. (27) Slinker, J., Barton, J. K. Unpublished results. 295 (28) Kato, K., Makino, Y., Kishimoto, T., Yamauchi, J., Kato, S., Muramatsu, M., Tamura, T. (1994) Nucleic Acids Res. 22, 1179–1185. (29) Coleman, R. A., Taggart, A. K. P., Benjamin, L. R., Pugh, B. F. (1995) J. Biol. Chem. 270, 13842–13849. (30) Gilfillan, S., Stelzer, G., Piaia, E., Hofmann, M. G., Meisterernst, M. (2005) J. Biol. Chem. 280, 6222–6230. (31) Johnson, S. A. S., Dubeau, L., White, R. J., Johnson, D. L. (2003) Cell Cycle 2, 442–444. (32) Johnson, S. A. S., Dubeau, L., Kawalek, M., Dervan, A., Schonthal, A. H., Dang, C. V., Johnson, D. L. (2003) Mol. Cell. Biol. 23, 3043–3051. 296  Chapter 2: Conductivity of a Single DNA Duplex Bridging a Carbon Nanotube Gap Adapted from: Guo, X., Gorodetsky, A. A., Hone, J., Barton, J. K., Nuckolls, C. (2008) Nature Nanotechnol. 3, 163–167. Xuefeng Guo fabricated the SWNT-FET devices. 41 ABSTRACT We describe a general method to integrate DNA strands between single-walled carbon nanotube electrodes and to measure their electrical properties. We modified DNA sequences with amines on either the 5' terminus or both the 3' and 5' termini and coupled these to the single-walled carbon nanotube electrodes through amide linkages, enabling the electrical properties of complementary and mismatched strands to be measured. Wellmatched duplex DNA in the gap between the electrodes exhibits a resistance on the order of 1 MΩ. A single GT or CA mismatch in a DNA 15-mer increases the resistance of the duplex ~ 300-fold relative to a well-matched one. Certain DNA sequences oriented within this gap are substrates for Alu I, a blunt end restriction enzyme. This enzyme cuts the DNA and eliminates the conductive path, supporting the supposition that the DNA is in its native conformation when bridging the ends of the single-walled carbon nanotubes. 42 2.1 Introduction Since the elucidation of the double helical structure of DNA, scientists have been fascinated by the possibility that the stacked aromatic base pairs of DNA may enable charge transport (CT) over significant distances (1–5). The nature of the conductive properties of duplex DNA has consequently attracted substantial interest. Initial solution experiments featured photoinduced DNA-mediated CT between well-defined donor and acceptor sites (6–13). Long-range CT has been shown to lead to oxidative damage in DNA over 200 Å away from the bound oxidant (14, 15) and DNA CT has also been found to be exquisitely sensitive to the integrity of the base-pair stack and to the coupling of the donors and acceptors with the DNA (10). Furthermore, DNA CT can be attenuated by a single base mismatch (16). Indeed, this sensitivity of DNA CT to the integrity of the base-pair duplex has prompted both the consideration of roles for DNA CT within the cell (17, 18) and the construction of electrochemical DNA-based sensors for mutations, base lesions, and protein binding (19, 20). Previously, we have described a system for measuring the conductivity of a single molecule covalently immobilized within a nanotube gap (21–24). In this system, gaps are formed in single-walled carbon nanotubes (SWNTs) that can be reconnected by one or a few molecules attached to both sides of the gap through amide bond formation. This strategy allows molecules to be wired into metal electrodes by means of robust amide linkages, avoiding the problems commonly associated with suspending DNA between electrodes and forming irreproducible electrical contacts. Moreover, the devices are 43 sufficiently robust that aqueous environments can be used. Using this method we have made molecular devices that are able to change their conductance as a function of pH (21), and others that are sensitive to the binding between protein and substrate (22), or that switch their conductance when the bridging molecules are photoswitched (24). Here we describe the first measurements of the conductivity of a single DNA duplex when it is wired into a carbon electrode through covalent bonds. Numerous CT measurements on DNA strands bridging two electrodes have also been carried out in an effort to establish the conductivity of DNA. These measurements have yielded a remarkably wide range of resistance values of 1 to 1 x 107 MΩ (25–31). For example, DNA ropes suspended on a metallic grid were found to behave as a semiconductor with a resistance on the order of 1 MΩ (25). However, one group initially found wide bandgap semiconducting behaviour for DNA duplexes set between two nanoelectrodes using high-voltage electrostatic trapping (26), but later found insulating behaviour for longer strands (27). In contrast, in other research it was found possible to induce superconductivity at low temperatures in dehydrated DNA bundles on rhenium/carbon electrodes (28). In general, the variability in the results obtained can be understood by considering the solution experiments, which show that DNA CT depends sensitively upon the integrity of the base-pair stack, the absence of damage within the duplex, and the electrical connections to the duplex. It should be noted that measurements using both conducting atomic force microscopy (AFM) (29) and scanning tunnelling microscopy (STM) (30–32) under aqueous conditions have recently been carried out and show that well-matched DNA exhibits a low resistance of 1–10 MΩ, as well as an increase in 44 resistance with an intervening base mismatch (31, 32). Also, STM measurements on DNA monolayers have shown effective charge transport for well-matched DNA oriented by the STM tip (33). However, none of these measurements was of a single duplex, but instead they were carried out for a collection of duplexes on the surface below the AFM or STM tip. Thus, in the conductivity measurements carried out so far, the integrity of the DNA was not well established, the connections to the duplex were not well defined, or the measurement was not definitively of a single DNA duplex. 2.2 Experimental Section 2.2.1 Fabrication of Cut Carbon Nanotube Devices Fabrication of the cut SWNT devices has been previously described in detail (21– 24). In brief, SWNTs were grown using chemical vapour deposition (CVD) on highly doped silicon wafers with 300 nm of thermally grown silicon oxide on their surface. Metal electrodes consisting of 5 nm of Cr overlaid with 50 nm of Au were deposited through a shadow mask onto the carbon nanotubes. The silicon wafer serves as a global back gate for the devices. We then spin-cast a layer of polymethylmethacrylate (PMMA) over the entire device structure and used ultrahigh-resolution electron beam lithography to open a window in the PMMA. This process exposed a section of the SWNT only a few nanometres in length, which was excised with an oxygen ion plasma. The oxidative etching of the carbon nanotube generated carboxylic acid functionalities on both sides of the gap (Figure 2.1a), which can be bridged with amine-terminated molecules. 45 Figure 2.1: A method to cut and functionalize individual SWNTs with DNA strands. (A) Functionalized point contacts made through the oxidative cutting of a SWNT wired into a device. (B) Bridging by functionalization of both strands with amine functionality. (C) Bridging by functionalization of one strand with amines on either end. 46 2.2.2 Synthesis of Amine-Modified DNA Unmodified oligonucleotides were prepared using standard phosphoramidite chemistry on an Applied Biosystems 394 DNA synthesizer, purified by HPLC and characterized by mass spectrometry. Two strategies were used to synthesize DNA modified with only a single amine at the 5’ terminus and DNA modified with amines at both the 3’ and 5’ termini. Oligonucleotides modified with an amine on the 5’ terminus only were prepared via solid phase synthesis on a CPG resin with an unprotected hydroxyl group at the 5’ terminus. The 5’-OH was treated with a 120 mg/mL solution of carbonyldiimidazole in dioxane for two hours followed by an 80 mg/mL solution of 1,3diaminopropane. The beads were thoroughly washed with dioxane, acetonitrile, and methanol leaving a free amine at the 5’ end. Oligonucleotides modified with amines on both the 3’ and 5’ termini were prepared via solid phase synthesis using reagents purchased from Glen Research, Inc. The solid phase synthesis was performed on 3’-PTAmino-Modifier C3 CPG with the 5’-Amino-Modifier C3-TFA phosphoramidite added in the final step of the solid phase synthesis to leave protected amines at both the 3’ and 5’ ends. In all instances, the oligonucleotides were cleaved from the resin with concentrated ammonium hydroxide before being stringently purified by HPLC with a C18 column. The purified oligonucleotides were quantified via UV-visible spectroscopy. Complementary single-strand DNA was hybridized with its complement by heating equimolar amounts of each strand in buffer containing 5 mM phosphate, pH = 7.1, 50 mM NaCl to 90 ºC, followed by cooling to ambient temperature. 47 2.2.3 Reconnection of Cut Carbon Nanotube Devices with DNA The devices were bridged with DNA via a simple two-step process. For carboxylic acid activation, newly cut devices were incubated overnight in the BupHTM MES buffered saline solution (pH 4.7, Pierce Biotech.) containing 5 mM EDCI and 10 mM Sulfo-NHS. The devices were then removed from the solution, washed with fresh buffer solution, and dried with a stream of nitrogen gas for device characterization. For subsequent amide bond formation, the as-formed devices were incubated in the BupHTM phosphate buffered saline solution (pH 7.2, Pierce Biotech.) containing 10 µM duplex or single stranded DNA. The devices were then removed from the solution, washed with fresh buffer solution, and dried with a stream of nitrogen gas for device characterization. 2.2.4 Dehybridization/Rehybridization of DNA Bridging the SWNT Gap The reconnected devices were immersed in a 50% formamide/DI water solution at 30 0C for one hour. Then the devices were removed from the solution, washed with DI water, and dried with a stream of nitrogen gas. Subsequently, the above devices were incubated in BupHTM phosphate buffered saline solution containing 10 µM of the corresponding single-stranded DNA. After one hour, the devices were removed from the solution, washed with fresh buffer, and dried with a stream of Nitrogen gas for device characterization. 2.2.5 Enzyme Restriction of DNA Bridging the SWNT Gap For restriction enzyme epxeriments, the devices rejoined with duplex DNA were incubated in NEBuffer solution (pH 7.4, New England Biolabs Inc.) containing 100 units of the enzyme Alu I (New England Biolabs Inc.) at 37 0C for 6 hours. Then the devices 48 were removed from the solution, washed with fresh buffer, and dried with a stream of N2 gas for device characterization. 2.3 Results and Discussion We reconnected the carbon nanotube gap with single DNA molecules terminated with amines using a two-step strategy. First, the freshly cut carbon nanotubes were immersed in a buffer solution containing standard amide coupling and activating agents (Sulfo-NHS, EDCI). Then, the activated carbon nanotube termini were reacted with amine-modified DNA to covalently bridge the gap with a single molecule. Given that the cross-sectional area of duplex DNA (~ 3 nm2) is comparable to that of the SWNTs grown here, it is unlikely that more than one DNA duplex can fit lengthwise within the gap. We explored two different methods to bridge these gaps. In one method (Figure 2.1b), one end each of the two strands of the DNA duplex are bound to the SWNT electrodes. In a second method (Figure 2.1c), a single strand is bound between the ends of the SWNT electrodes. The strategy in Figure 2.1c is inherently more interesting because it allows for dehybridization/rehydridization with mismatched strands. The measurements were carried out under ambient conditions. Figure 2.2 shows representative I–V curves for these two different methods of DNA attachment. We could not determine any significant difference between the conductance measurements when using these two connection strategies. In Figure 2.2a we used a DNA duplex functionalized on both strands with an amine at the 5' end. In the 49 Figure 2.2: Device characteristics for individual SWNTs connected with DNA. (A), (B) Source-drain current versus VG at a constant source-drain voltage (50 mV) before cutting (black curve: 1), after cutting (red curve: 2), and after connection with the DNA sequence shown (green curve: 3), for a semiconducting SWNT device (A) and a metallic SWNT device (B). Guanine, G; cytosine, C; adenine, A; thymine, T. 50 other case (Figure 2.2b), we used a DNA duplex containing a strand functionalized at both the 5' and 3' ends. The black curves show the source–drain current (ISD) as a function of the gate voltage (VG) at a constant source–drain bias of 50 mV for the pristine nanotube. Before cutting of the SWNT, the device in Figure 2.2a is a hole transporting semiconducting device, and the one in Figure 2.2b is a metallic device. After cutting and initial treatment of the gap with coupling agents, the devices show no measurable current (as indicated by the red curves). The green curves in Figure 2.2a,b illustrate the conductance of the two devices after reconnection with the two amine-modified DNAs. In both cases shown here, the reconnected carbon nanotube devices recover their original p-type semiconducting or metallic properties. Note that the gate voltage that can be applied to the reconnected devices is limited; device breakdown is observed for gate voltages greater than 6 V. Over time, at these higher gate biases, the DNA bridges became poorer and poorer conductors until, ultimately, the current levels are at the noise level of the measurement (Figure 2.3 and Figure 2.4). It is difficult to determine if this effect could be due to a hydration layer of water around the devices. Table 2.1 summarizes the device characteristics measured in the course of this study for the devices before cutting, after cutting, and after reconnection with amine-terminated DNA sequences. Using this method we obtained 10 working devices out of 370 that were tested. Devices were also reconnected with mismatched DNA, as it has been shown in a variety of experiments that single-base mismatches dramatically attenuate CT. The DNA duplexes and mismatches explored here are shown in Figure 2.5a. The mismatched devices were found to have higher resistance than corresponding devices reconnected with well-matched DNA. These results could not be compared quantitatively with those 51 Table 2.1: Summary of the resistance values obtained before cutting and after reconnection 52 Figure 2.3: Device characteristics for individual SWNTs reconnected with well-matched DNA. Source-drain current versus VG at a constant source-drain voltage (50 mV) before/after cutting (A) and after connection with DNA (B). Device breakdown is observed when the gate voltage gradually increases, starting at ~ 6 V. The sequence is the same as in Figure 2.2. 53 Figure 2.4: Device characteristics for individual SWNTs reconnected CA mismatched DNA. Source-drain current versus VG at a constant source-drain voltage (50 mV) before/after cutting (A) and after connection with DNA (B). The sequence was aminemodified NH2(CH2)3NHCO-5'-AGT ACA GCC ATC GCG-3' and its amine-modified complement 3'-TCA TGT CAG TAG CGC-5'-OCHN (CH2)3NH2 (the mismatch is in italics). 54 on the well-matched duplex, because different devices were fabricated to test the different duplexes. A device was therefore first reconnected with well-matched DNA duplexes functionalized with the amines on the 5' and 3' termini of one strand, and then the duplex was dehybridized using a 1:1 solution of formamide and deionized water at 30 °C and rehybridized with different complements (Figure 2.5 a–c). Figure 2.5b shows the corresponding current-voltage curves for this sequence of experiments, and Figure 2.5c shows the current at VG = -3 V for this sequence at a constant source-drain bias of 50 mV. Rehybridization with the complement so as to generate a CA mismatch reduced the current significantly and yielded an increase in the on-state resistance of nearly 300-fold from 0.5 MΩ to 155 MΩ (Figure 2.5c). Replacing the complement featuring a CA mismatch with a complement featuring a GT mismatch yielded no changes in the device characteristics. However, the original on-state resistance and nanoamp current levels could then be recovered by replacing the GT mismatched complement with the original well-matched sequence. Importantly, the device could be taken through multiple dehybridization/rehybridization cycles, as shown in Figure 2.5 b,c. As further confirmation that CT in the carbon nanotube gap is DNA-mediated, we reconnected separate devices first with DNA featuring a GT mismatch or DNA featuring a CA mismatch (Figure 2.6 and Figure 2.7). Dehybridization of the mismatched DNA and replacement with well-matched DNA yielded an increase in the current and a decrease in the on-state resistance in both instances. It is important to note that the thermodynamically stable GT mismatch produced an effect that is identical to that found with the thermodynamically destabilizing CA mismatch. As we have found in solution experiments, the attenuation in DNA CT seen with mismatches does not correlate with 55 Figure 2.5: Mismatches have a large effect on DNA conductance. (A) Replacing wellmatched (WM) duplexes with CA and GT mismatches. (B) Source-drain current versus VG at a constant source-drain voltage (50 mV) for a SWNT device taken through the sequence 1 through 6. The current levels for points 2, 3, 5, and 6 are ~ 300 times lower. (C) Source-drain current at VG = –3 V at a constant source-drain voltage (50 mV) for the sequence 1 through 6. 56 Figure 2.6: Mismatches have a large effect on DNA conductance. The electrical characteristics of a device rejoined with CA-mismatched DNA. Source–drain current versus VG at a constant source-drain voltage (50 mV) for a SWNT device before/after cutting (A) and when well-matched DNA is replaced with mismatched DNA (B). 57 Figure 2.7: Mismatches have a large effect on DNA conductance. The electrical characteristics of a device rejoined with GT-mismatched DNA. Source-drain current versus VG at a constant source-drain voltage (50 mV) for a SWNT device before/after cutting (A) and when well-matched DNA is replaced with mismatched DNA (B). 58 thermodynamic stability of the duplex (34). Ultrafast spectroscopic experiments indicate that DNA CT depends upon the sequence-dependent dynamics of DNA (35). Certainly, the changes observed here in the electrical characteristics of the device with mismatches cannot be due to poorer stability of the DNA. Although the mismatch experiments provide strong evidence that the observed signals do not result from ionic conduction from the DNA molecules, as an additional control, newly cut devices were subjected to the same reconnection conditions but with the DNA excluded. After removal from the solution and rinsing, all of the devices treated in this manner remained at open circuit with no measurable current. Devices were also reconnected with single-stranded DNA featuring amines at both the 5' and 3' ends but without its complement. Although carbon nanotube gaps could be bridged with the single-stranded DNA, the resulting devices were found to be highly unstable (Figure 2.8). After three voltage cycles, the current passing through singlestranded DNA degraded to open-circuit levels. Such instability may result from voltageinduced oxidation of the exposed nucleobases and was not observed with duplex DNA. Additional control experiments were performed to determine if non-specific absorption of DNA could be responsible for the conduction changes during dehybridization/rehybridization. Devices were partially cut with a shorter oxygen plasma treatment before being taken through the sequential steps of reconnection and exchanges from matched to mismatched sequences. In essence, the SWNT is only nicked, not cut completely through, so the electrical connection is maintained. The devices treated in this way displayed little change in either the resistance or threshold voltage (Figure 2.9). 59 Figure 2.8: The electrical characteristics of a device rejoined with single stranded DNA. Source-drain current versus VG at a constant source-drain voltage (50 mV) for a SWNT device before/after cutting (A) and reconnected with single-stranded DNA (B). 60 Figure 2.9: Control experiments using partially cut devices that underwent the treatment of cutting, reconnection, dehybridization, and rehybridization. None of the devices displayed any obvious changes in resistance or threshold voltage for each step of the procedure. The sequence and conditions were identical to those utilized for steps 1, 2, and 3 in Figure 2.3. 61 As a final test that the duplex DNA within the gap adopts a native conformation under the conditions of the experiment, we used Alu I, a blunt end restriction enzyme, to cut the DNA (Figure 2.10). Alu I only cuts DNA that is in its native conformation. Devices were reconnected with duplex DNA containing the restriction sequence 5'AGCT-3'. The device was subsequently incubated with Alu I, resulting in a concomitant decrease in the current to the noise limits of the measurement. As another control, a device reconnected with a nearly identical sequence that featured the sequence 5'-AGTC3' in place of the restriction site was incubated with Alu I. In this instance, no significant change was observed in the electrical characteristics of the device (Figure 2.11). These data support the observation of a sequence-specific restriction event. The enzyme is able to cleave its target sequence, yielding no detectable current in the device. Under the experimental conditions presented, then, the DNA duplex is intact, and the results suggest that it adopts a native conformation. We can now place the values found here in the proper context to establish a range for the conductivity of a single, intact DNA duplex. Our measurements place the resistance of well-matched DNA duplexes with ~ 6 nm length in the range of 0.1–5 MΩ (Table 2.1). For comparison, based on the bulk c-axis resistance, highly oriented pyrolytic graphite (HOPG) with similar dimensions should also have a resistance of ~ 1 MΩ (36). We estimate this value by substituting a HOPG stack of equivalent diameter for the double-stranded DNA. Thus it appears that DNA, in its well-matched and wellstacked duplex form, behaves electrically much like an array of stacked aromatic graphite planes. Importantly, just as we have seen in solution, the presence of intervening mismatches attenuates DNA-mediated CT. 62 Figure 2.10: Enzymes can be used to cleave the DNA between the ends of the SWNTs. Source-drain current versus VG at a constant source-drain voltage (50 mV) for a SWNT device after cutting and reconnection with the shown DNA sequence before (green curve: 1) and after reaction with Alu 1 (red curve: 2). 63 Figure 2.11: Restriction by enzymes is sequence specific. Source-drain current versus VG at a constant source-drain voltage (50 mV) for a SWNT device before/after cutting in (A), and reconnection with the shown DNA sequence in (B) before and after reaction with Alu 1. 64 This attenuation leads to a ~ 300-fold increase in resistance. Such an increase in the resistance of mismatched DNA is consistent with previously reported STM measurements (31–33). Also, it should be noted that even within our own measurements, the covalent σ-bonded linkages at the termini of the DNA duplex must also decrease the conductivity observed versus the conductivity expected with coupling directly into the base-pair stack. Therefore, the values obtained represent the upper limits of the resistance of the DNA pi-stack. In conclusion, we have outlined a method to integrate a single DNA duplex within an electrical device. The DNA molecules are covalently wired into electrical circuits through robust amide linkages that are stable over a wide range of chemistries and conditions. The experiments presented here illustrate the ability of DNA to mediate CT over significant distances and allow for the direct measurement of the resistance of a single well-matched DNA molecule. It is perhaps not surprising that DNA, if in its native conformation, and containing a stack of aromatic heterocycles in its core, resembles the aromatic stacked planes of graphite with respect to electrical characteristics. However, significantly, CT through a DNA assembly is sensitive to perturbations that arise in the base-pair stack. As illustrated directly here, as well as in solution experiments, single base mismatches attenuate CT through DNA. A molecular π-stacked array in solution is necessarily less robust than a solid-state conductive material. Thus, although the sensitivity of DNA CT to perturbations in stacking suggests that DNA may not be appropriate to serve as a robust wire for nanoelectronic circuits, DNA molecules bridging nanodevices can surely serve as uniquely powerful reporters to transduce biochemical events into electrical signals at the single-molecule level. 65 REFERENCES (1) Eley, D. D., Spivey, D. I. (1962) Trans. Faraday Soc. 58, 411–415. (2) O’Neill, M. A., Barton, J. K. (2005) Charge Transfer in DNA: From Mechanism to Application (H.-A. Wagenknecht), pp. 27–75, Wiley-VCH, Hoboken. (3) O'Neill, M. A.; Barton, J. K. (2004) Topics in Current Chemistry: Electron Transfer in DNA I (G. B. Schuster), pp. 67–115, Springer-Verlag, New York. (4) Giese, B. (2002) Ann. Rev. Biochem. 71, 51–70. (5) Delaney, S., Barton, J. K. (2003) J. Org. Chem. 68, 6475–6483. (6) Murphy, C. J., Arkin, M. R., Jenkins, Y., Ghatlia, N. D., Turro, N. J., Barton, J. K. (1993) Science 262, 1025–1029. (7) Lewis, F. D., Wu, T., Zhang, Y., Letsinger, R. L., Greenfield, S. R., Wasielewski, M. R. (1997) Science 277, 673–676. (8) Hall, D. B., Holmlin, R. E., Barton, J. K. (1996) Nature 382, 731–735. (9) Dandliker, P. J., Holmlin, R. E., Barton, J. K. (1997) Science 275, 1465–1468. (10) Kelley, S. O., Barton, J. K. (1999) Science 283, 375–381. (11) Gasper, S. M., Schuster, G. B. (1997) J. Am. Chem. Soc. 119, 12762–12771. (12) Giese, B., Amaudrut, J., Kohler, A. K., Spormann, M., Wesselly, S. (2001) Nature 412, 318–320. (13) Wan, C., Fiebig, T., Kelley, S. O., Treadway, C. R., Barton, J. K. (1999) Proc. Natl Acad. Sci. U.S.A. 96, 6014–6019. (14) Nunez, M. E., Hall, D. B., Barton, J. K. (1999) Chem. Biol. 6, 85–97. 66 (15) Henderson, P. T., Jones, D., Hampikian, G., Kan, Y., Schuster, G. B. (1999) Proc. Natl Acad. Sci. U.S.A. 96, 8353–8358. (16) Kelley, S. O., Homlin, R. E., Stemp, E. D. A., Barton, J. K. (1997) J. Am. Chem. Soc. 119, 9861–9870. (17) Boal, A. K., Yavin, E., Lukianova, O. A., O’Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (18) Nunez, M. E., Holmquist, G. P., Barton, J. K. (2001) Biochemistry 40, 12465– 12471. (19) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nature Biotechnol. 20, 282– 286. (20) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (21) Guo, X., Small, J. P., Klare, J. E., Wang, Y., Purewal, M. S., Tam, I. W., Hong, B. H., Caldwell, R., Huang, L., O'Brien, S., Yan, J., Breslow, R., Wind, S. J., Hone, J., Kim, P., Nuckolls, C. (2006) Science 311, 356–359. (22) Guo, X., Whalley, A., Klare, J. E., Huang, L., O’Brien, S., Steigerwald, M., Nuckolls, C. (2007) Nano Lett. 7, 1119–1122. (23) Guo, X., Myers, M., Xiao, X., Lefenfeld, M., Steiner, R., Tulevski, G. S., Tang, J., Baumert, J., Leibfarth, F., Yardley, J. T., Steigerwald, M. L., Kim, P., Nuckolls, C. (2006) Proc. Natl Acad. Sci U.S.A. 103, 11452–11456. (24) Whalley, A. C., Steigerwald, M. L., Guo, X., Nuckolls, C. (2007) J. Am. Chem. Soc. 129, 12590–12591. (25) Fink, H. W., Schonenberger, C. (1999) Nature 398, 407–410. 67 (26) Porath, D., Bezryadin, A., de Vries, S., Dekker, C. (2000) Nature 403, 635– 638. (27) Storm, A. J., van Noort, J., de Vries, S., Dekker, C. S. (2001) Appl. Phys. Lett. 79, 3881–3883. (28) Kasumov, A. Y., Kociak, M., Gueron, S., Reulet, B., Volkov, V. T., Klinov, D. V., Bouchiat, H. (2000) Science 291, 280–282. (29) Gohen, H., Nogues, C., Naaman, R., Porath, D. (2005) Proc. Natl Acad. Sci. U.S.A. 102, 11589–11593. (30) van Zalinge, H., Schiffrin, D. J., Bates, A. D., Starikov, E. B., Wenzel, W., Nichols, R. J. (2006) Angew. Chem. Int. Ed. 45, 5499–5502. (31) Hihath, J., Xu, B., Zhang, P., Tao, N. (2005) Proc. Natl Acad. Sci. U.S.A. 102, 16979–16983. (32) Wierzbinski, E., Arndt, J., Hammond, W., Slowinski, K. (2006) Langmuir 22, 2426–2429. (33) Ceres, D. M., Barton, J. K. (2003) J. Am. Chem. Soc. 125, 14964–14965. (34) Bhattacharya, P. K., Barton, J. K. (2001) J. Am. Chem. Soc. 123, 8649–8656. (35) O'Neill, M. A., Becker, H.-C., Wan, C., Barton, J. K., Zewail, A. H. (2003) Angew. Chem. Int. Ed. 42, 5896–5900. (36) Tsang, D. Z., Dresselhaus, M. S. (1976) Carbon 14, 43–46. 68  Chapter 3: Electrochemistry Using SelfAssembled DNA Monolayers on Highly Oriented Pyrolytic Graphite Adapted from: Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. 69 ABSTRACT Duplex DNA functionalized with pyrene has been utilized to fabricate DNA-modified electrodes on highly oriented pyrolytic graphite (HOPG). Films have been characterized using AFM and radioactive labeling as well as electrochemically. The data obtained are consistent with a close-packed structure in the film with helices oriented in a nearly upright orientation, as seen earlier with the fabrication of thiol-tethered duplexes on gold. Also as on gold, we observe the reduction of DNA-bound intercalators in a DNAmediated reaction. The reduction of the intercalator is attenuated in the presence of the single-base mismatches, CA and GT, independent of the sequence composition of the oligonucleotide. This sensitivity to single-base mismatches is enhanced when methylene blue reduction is coupled in an electrocatalytic cycle with ferricyanide. The extended potential range afforded by the HOPG surface has allowed us also to investigate the electrochemistry of previously inaccessible metallointercalators, Ru(bpy)2dppz2+ and Os(phen)2dppz2+, at the DNA-modified HOPG surface. These results support the application of DNA-modified HOPG as a convenient and reproducible surface for electrochemical DNA sensors using DNA-mediated charge transport. 70 3.1 Introduction Numerous studies have shown that DNA can efficiently transport charge in solution (1–6). Self-assembled DNA monolayers on gold have also been shown to mediate charge transport (CT) over significant distances (7–9). Either in solution or at a surface, DNA-mediated CT exhibits exquisite sensitivity to perturbations in base stacking as occur with single-base mismatches (10, 11), lesions (12, 13), and protein binding (14– 17). Consequently, electrochemistry on DNA-modified electrodes has been useful in tools to detect DNA mutations, lesions, and protein/DNA interactions (5–7). We have previously investigated the electrochemical properties of DNA monolayers assembled on gold (8, 9, 18–20). Electrochemical experiments have revealed that redox-active intercalators bound near the periphery of the film can serve as probes of DNA-mediated CT. The presence of even a single mismatch leads to dramatic attenuation of the electrochemical signal. This sensitivity to mutations within the base pair stack has been exploited for the detection of genomic sequence variations in a chip-based format (10). The ability to manufacture DNA chips for large-scale mutational analysis has led us to investigate self-assembly of DNA monolayers at surfaces other than gold. Alternative surfaces may provide a route to disposable and inexpensive DNA chips with improved performance. We have chosen to focus our efforts on HOPG (highly oriented pyrolytic graphite) since this material provides a well-defined and clean surface possessing few defects (21). 71 Previous experiments with DNA on graphite have focused on the use of glassy carbon, pyrolytic graphite, and carbon paste electrodes. For example, Wang and coworkers have detected hybridization by electrochemical stripping of colloidal gold tags at screen-printed carbon paste electrodes (22–26). Additionally, Rusling and co-workers have detected both oxidative damage and nucleobase adducts using catalytic osmium and ruthenium square wave voltammetry (27–29). These methods have been applied to both DNA in solution and DNA immobilized at the electrode surface. Pyrene modifications of both proteins and DNA have been made for surface functionalization of various graphite electrodes (30–33). For example, 1-pyrenebutyric acid N-hydroxysuccinimidyl ester has been utilized as a bifunctional reagent to immobilize a photosynthetic reaction center on pyrolytic graphite, and a cytochrome P450 mutant has been adsorbed onto a graphite surface via a pyrene-modified cysteine residue (30–32). Furthermore, proteins and oligonucleotides have been noncovalently attached to the graphitic side walls of carbon nanotubes via a pyrene moiety (33). Utilizing this approach, we have synthesized oligonucleotides derivatized with pyrene for noncovalent immobilization of DNA duplexes on a graphitic surface. DNA sequences plus complements are prepared by standard solid-phase techniques with a pyrene moiety incorporated at the 5'-phosphate terminus. The pyrene-modified duplexes are then used to fabricate a DNA monolayer adsorbed onto HOPG. Using these films, we are able to detect single-base mismatches in oligonucleotides. DNA CT is mediated by the base pair stack on this surface. 72 3.2 Experimental Section 3.2.1 Materials All reagents for DNA synthesis were purchased from Glen Research. Methylene blue, daunomycin, potassium ferricyanide, and all reagents used in the synthesis of pyrene-modified DNA were purchased from Sigma in the highest purity available and used as received. SPI-1 and SPI-3 grade HOPG was purchased from SPI, Inc. Buffers were prepared with Milli-Q water and filtered with a 0.2 µm filter. 3.2.2 Synthesis of DNA and Redox-Active Probes Oligonucleotides were prepared using standard phosphoramidite chemistry on an Applied Biosystems 394 DNA synthesizer, purified by HPLC, and characterized by mass spectrometry. Pyrene-terminated oligonucleotides were synthesized in the following two ways: In the first strategy, oligonucleotides were prepared by solid-phase synthesis on a CPG resin with an unprotected hydroxyl group at the 5' terminus. The 5'-OH was treated with a 120 mg/mL solution of carbonyldiimidazole in dioxane for 90 min followed by an 80 mg/mL solution of 1,6-diaminohexane. For tether variations, ethylenediamine or 1,10diaminononane was used instead. The free amine was then treated with 1-pyrenebutyric acid N-hydroxysuccinimidyl ester, leaving a pyrene moiety at the 5' terminus. In the second strategy, the free hydroxyl at the 5' terminus of the oligonucleotide was condensed with a 2-cyanoethyl N,N-diisopropylphosphoramidite of 1-pyrenebutanol on the DNA synthesizer. The phosphoramidite was prepared using a modified literature procedure (34). In brief, 500 mg of N,N-diisopropylethylamine was added to 5 mL of a 20 mg/mL solution of 1-pyrenebutanol in acetonitrile under an inert argon atmosphere. 73 While the reaction mixture was gently stirred, 90 mg of N,N- diisopropylphosphoramidochloridite was added dropwise. The reaction progress was monitored by TLC before the reaction mixture was moved to an extraction funnel and washed twice with 5% (by volume) NaHCO3 and 3 M NaCl. The organic layer was collected, dried with Na2SO4, and rotovapped to a yellow oil. The phosphoramidite was dissolved in 1 mL of acetonitrile and used immediately for DNA synthesis. The modified oligonucleotides were cleaved from the resin with concentrated ammonium hydroxide before being stringently purified by HPLC with a C18 column. Incorporation of the pyrene moiety was confirmed by HPLC retention times, UV-vis spectra, and MALDI-TOF mass spectrometry. Pyrene-modified single-strand DNA was hybridized with its complement by heating equimolar amounts of each strand (100 µM) in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl to 90 º C followed by cooling to ambient temperature. Since single-stranded DNA can adsorb to graphite, great care was taken during quantification of DNA to ensure that equimolar amounts of each strand were annealed to form duplexes. [Ru(bpy)2dppz]Cl2 and [Os(phen)2dppz]Cl2 (dppz = dipyridophenazine) were prepared and purified according to established protocols (33–38). Oligonucleotides and metallointercalators were quantified by UV-vis spectroscopy as previously described (4, 8, 9, 33–38). 3.2.3 Preparation of DNA-Modified Electrodes Electrodes were constructed using a procedure modified from the literature (39). Briefly, a 10 mm × 10 mm × 1 mm HOPG block was affixed to a copper square with silver paste and mounted in a Teflon holder. The active area of this working electrode 74 was defined by a Viton or silicon O-ring mounted in a Teflon cap. To minimize mechanical stress on the HOPG block, electrical contact to the copper square was made with a spring-loaded pin. A clean and reproducible surface with few defects was crucial to the success of all electrochemistry experiments. Consequently, a fresh HOPG surface was cleaved with 3-M scotch tape before each experiment. The exposed surface was immediately used for electrochemistry experiments or immediately modified with DNA. Duplex DNA containing a pyrene moiety was deposited onto the HOPG surface in buffer, pH 7.1, containing 5 mM Pi 50 mM NaCl, and 100 mM MgCl2; Mg2+ is included to ensure a closely packed film (8–11, 18–20). DNA films were allowed to form in a humidified chamber over a period of 4–24 hours. 3.2.4 Electrochemical Experiments Cyclic voltammetry and chronocoulometry experiments were performed using a CH Instruments electrochemical analyzer (Austin, TX). Unless otherwise noted, experiments were performed at ambient temperature in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl under an argon atmosphere. For daunomycin experiments, pH = 8. A custom-built three-compartment electrochemical cell was used for all experiments with a Pt auxiliary electrode, an SPI-1 grade HOPG working electrode, and a saturated calomel (SCE) or silver/silver chloride (Ag/AgCl) reference electrode. The compartment containing the reference electrode was separated from the rest of the cell with a Luggin capillary. For metallointercalators, the redox potentials on bare HOPG are reported versus Ag/AgCl at a 500 mV/s scan rate for a freshly cleaved surface featuring some defects. The potentials of transition-metal complexes on HOPG have previously been shown to be 75 highly dependent on the defect density of the HOPG surface (21). Consequently, variations of ±50 mV were observed for cathodic peak potentials on bare HOPG. 3.2.5 Atomic Force Microscopy Experiments All AFM measurements were performed using an MFP-3D atomic force microscope from Asylum Research, Inc. (Santa Barbara, CA). The HOPG block was mounted in a custom-designed glass chamber before being modified with DNA as described. AFM measurements were performed in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl in either contact (constant force) or tapping mode. Commercially available silicon tips were utilized for all experiments. 3.2.6 Radioactive Labeling and Quantification Before being hybridized to its pyrene-modified complements, the oligonucleotide was labeled at the 5' end with [gamma-32P]ATP and polynucleotide kinase. A 0.25 cm2 surface area was defined on an SPI-3 grade HOPG block and then modified with DNA as described. Samples were rinsed thoroughly and dried before being counted on an LS 5000TD scintillation counter (counts were adjusted for attenuation by the HOPG). Quantification was performed by comparing samples to calibration standards prepared from known quantities of labeled oligonucleotide. 3.3 Results and Discussion 3.3.1 Surface Characterization After hybridization, densely packed DNA monolayers are self-assembled on a freshly cleaved HOPG surface with excess Mg2+ to permit close packing. Well-packed 76 films are crucial to the success of these experiments, so noncovalent intercalator binding is restricted to the upper segments of DNA. Thus, extensive characterization of the DNA monolayers is necessary. Quantitative characterization of the DNA-modified surface does reveal that the fabrication on HOPG resembles that found earlier on gold, as schematically illustrated in Figure 3.1 (40, 41). AFM measurements were utilized to explore the morphology of DNA monolayers self-assembled on HOPG. Previous experiments on Au have focused on the effect of applied potential and linker structure on the morphology of 15-mer DNA films (40, 41). Additionally, AFM measurements on DNA films in solution have been used for labelfree detection of hybridization (42, 43). In this work, we have chosen to examine the effect of sequence composition and duplex length on film formation at a HOPG surface. On the basis of AFM measurements, we find that duplexes of 10-, 14-, 15-, and 20-mer lengths form smooth, featureless films. Regardless of duplex length, these films appear to completely cover the HOPG surface. To determine the film depth, a small patch of DNA may be removed from the HOPG surface by applying a vertical force with the AFM tip (40–43). Subsequent height contrast measurements between the resulting square and the surrounding covered surface are used to determine the film depth. AFM measurements for 14- and 15-mer sequences show film depths of 4.3 and 4.4 nm, respectively. The values found for 15-mer duplexes closely agree with previous experiments on Au, where a film depth of 4.5 nm was observed. In addition, height contrast measurements reveal that 10-mer duplexes result in shallower films whereas 20-mer duplexes result in deeper films. The observed thicknesses correspond to those expected for duplexes oriented at a ~ 45 º angle with 77 Figure 3.1: Schematic illustration of DNA duplexes modified with a pyrene moiety assembled on an HOPG surface. Surface characterization is consistent with DNA duplexes bound to HOPG in a nearly upright orientation and closely packed. 78 Figure 3.2: AFM characterization of DNA-modified HOPG: (A) typical AFM image of a patch of DNA removed from a monolayer of 20-mer duplexes on HOPG (tapping mode); (B) typical depth profile of a patch of DNA removed from a monolayer of 14-mer duplexes on HOPG (contact mode). All measurements were taken in buffer at pH 7.1, containing 5 mM Pi and 50 mM NaCl. Sequences were for (A) pyrene(CH2)3CONH(CH2)6NHCO-5'-AGT ACA GAG TAC AGT ACG CG-3' plus complement and for (B) pyrene-(CH2)3CONH(CH2)6NHCO-5'-ATT ATA TAA TTG CT-3' plus complement. 79 respect to the surface assuming a fully extended alkane tether (40–43). A typical film with a DNA patch removed and a typical height profile are shown in Figure 3.2. Surface coverage is also quantitated by radioactive labeling. Here, on HOPG, we find a DNA duplex coverage of 47 ± 18 pmol/cm2; this corresponds to a fractional surface coverage of 85%, which corresponds to close packing of DNA duplexes oriented at ~ 45 º. Previous radioactive labeling experiments on Au found a surface coverage of 41 pmol/cm2, corresponding to a fractional surface coverage of 75% (8–11, 18–20). The intrinsically flat HOPG surface (relative to gold) may allow for slightly denser films corresponding to the higher observed surface coverage. Additionally, the flat "footprint" attachment of the pyrene may aid in close packing. Quantitation of surface coverage may also be performed electrochemically. Assay using ruthenium hexammine (42–44) yields coverages similar to those found by radioactive labeling and AFM, a value of 44.6 ± 14 pmol/cm2. It should be noted that, as a companion to all experiments, films are routinely screened using ferricyanide as an anionic probe. At a close-packed DNA-modified surface, ferricyanide remains electrochemically silent, consistent with complete coverage (8–11, 18–20). These results therefore all are consistent with results seen earlier on Au. Fabrication of DNA duplexes on HOPG with high concentrations of Mg2+ leads to closely packed helices arranged at a ~ 45 º angle in the absence of an applied potential. 3.3.2 Electrochemistry at a DNA-Modified HOPG Surface Shown in Figure 3.3 is the cyclic voltammetry for methylene blue (MB) on bare HOPG and HOPG modified with duplex DNA featuring a pyrene moiety. At the DNA-modified surface, MB shows a redox couple at -280 mV versus Ag/AgCl, shifted by 30 mV from 80 Figure 3.3: Cyclic voltammetry of 0.5 µM methylene blue at a 100 mV/s scan rate with a 0.18 cm2 working area in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl. Cyclic voltammograms are shown versus Ag/AgCl for a bare electrode in blue and a DNAmodified electrode in red. The sequence was pyrene-(CH2)4Pi-5'-AGT ACA GTC ATC GCG-3'. 81 that found on bare HOPG. A similar shift is observed for MB bound to DNA on Au (18– 20). Since aromatic molecules tightly adsorb onto bare pyrolytic graphite, the electrochemistry of MB on both DNA-modified and bare HOPG surfaces leads to linear plots of peak current as a function of scan rate, although with different slopes; the redoxactive species is surface-bound in both cases (45). It should also be noted that control experiments with MB on HOPG surfaces using DNA duplexes lacking the pyrene moiety yield broadened and irreproducible electrochemical signals. This is most likely due to the disordered orientation of duplexes lacking a pyrene moiety on the HOPG surface. We also examined a range of tether lengths for the pyrene linked to DNA in these studies. In earlier studies of daunomycin covalently bound to thiolated DNA on Au, we found that electron-transfer rates were limited by tunneling through the alkanethiol tether (46). Here different tether lengths have been explored in studies using noncovalent MB; because the noncovalent MB may bind to several sites near the top of the film, detailed information concerning rates as a function of distance can therefore not be obtained. Instead we were interested in determining the optimum length for close packing of the film and reproducible electrochemistry. Good reproducibility is in fact found with tethers (CH2)4Pi, (CH2)3CONH(CH2)2NHCO, (CH2)3CONH(CH2)6NHCO, and (CH2)3CONH(CH2)10NHCO. For all of these linkers, larger peak splittings between cathodic and anodic waves are evident in comparison to those for bare HOPG (15 mV versus 5 mV for MB in Figure 3.3). Extending the length of the tether between the pyrene and the duplex leads to even greater peak splittings at these scan rates, a result which is consistent with electron-transfer rates being limited by tether length. For example, cyclic voltammograms at pyrene-(CH2)3CONH(CH2)2NHCO-DNA-modified surfaces exhibit a 82 peak splitting of 23 ± 7 mV for MB, while surfaces modified with pyrene(CH2)3CONH(CH2)10NHCO-DNA exhibit a peak splitting of 52 ± 10 mV for MB. These observations are consistent with the structural model for the DNA film shown in Figure 3.1 where the tether intervenes between the pyrene bound to HOPG and the DNA duplex. 3.3.3 Electrochemistry on HOPG Is DNA-Mediated If CT is DNA-mediated, the introduction of a mismatch within the base pair stack should lead to a significant attenuation of the electrochemical signal at the DNA film; we have found in a range of experiments that the perturbation in the base pair stack associated with a mismatch inhibits CT (1–4, 8–11). We find that electrochemistry on DNA-modified HOPG with either daunomycin or MB as the redox-active probe leads to a significant diminution in the integrated cathodic charge (Qc) for fully base paired (WT) and mismatch-containing (MM) DNA films. For the duplex sequence pyrene-(CH2)4Pi-5'AGT ACA GTC ATC GCG-3', QMM/QWT = 0.51 with daunomycin as the redox probe and QMM/QWT = 0.41 for MB, where the italicized base indicates the position of a CA mismatch. We also examined the electrochemistry using DNA duplexes of different sequences. As evident in Figure 3.4 and as previously shown for DNA monolayers on gold, mismatch detection is independent of sequence context (8, 9). In particular, AT-rich sequences yield results equivalent to those of sequences rich in GC content. The sequence pyrene-(CH2)3CONH(CH2)6NHCO-5'-ATT ATA TAA TTG CT-3' gives QMM/QWT = 0.60 with daunomycin as the redox probe and QMM/QWT = 0.46 for MB. The observed variations in QMM and QWT with daunomycin and MB between various sequences and linker lengths are well within experimental error. 83 Figure 3.4: Cyclic voltammetry of 0.5 µM methylene blue at a 100 mV/s scan rate versus SCE in a buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl for an electrode modified with fully base paired DNA duplexes in blue and an electrode modified with duplexes featuring a CA mismatch in red. Sequences were for (A) pyrene- (CH2)3CONH(CH2)6NHCO-5'-AGT ACA GTC ATC GCG-3' and for (B) pyrene(CH2)3CONH(CH2)6NHCO-5'-ATT ATA TAA TTG CT-3', where the italicized base indicates the position of the CA mismatch. 84 Coupling MB to an electrocatalytic cycle involving ferricyanide allows for even greater discrimination between fully base paired and mismatch-containing DNA films (10, 11). Figure 3.5 shows cyclic voltammetry for 0.5 µM MB in the presence of 2 mM ferricyanide on HOPG with pyrene-tethered duplex DNA with and without a CA or GT mismatch. Without the mismatch, a highly asymmetric cyclic voltammogram is evident, consistent with electrocatalytic behavior (45, 47–49). As expected for an electrocatalytic process, the Ep of the electrocatalytic peak matches the E1/2 of the MB peak at a DNAmodified surface without K3Fe(CN)6 (45, 47–49). Moreover, note the complete absence of a ferricyanide signal due to repulsion of the anionically charged probe by the DNA monolayer. Greater sensitivity in discriminating the single-base mismatch is obtained via electrocatalysis. We can, for example, easily detect the thermodynamically stable GT wobble base pair (Figure 3.5). Chronocoulometry using a -350 mV step with ferricyanide and MB gave an integrated charge of 111.5 ± 12.0 µC for fully base paired DNA compared to maximal charges at 5 s of 49.5 ± 8.1 and 55.4 ± 20.3 for duplexes featuring a CA or GT mismatch, respectively. These results underscore that mismatch discrimination depends on CT through the base pair stack rather than on thermodynamic differences associated with hybridization events. Results for different tethers and different sequences, with and without electrocatalysis, are summarized in Table 3.1. Significantly, these results, taken together, all fully support the conclusion that CT on HOPG modified with pyrene-tethered DNA is DNA-mediated. The results found here for HOPG compare closely to those found for DNA films fabricated on Au. 85 Figure 3.5: Cyclic voltammetry of 0.5 µM methylene blue and 2 mM K3Fe(CN)6 at a 100 mV/s scan rate versus SCE in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl for an electrode modified with fully base paired DNA duplexes in blue and an electrode modified with duplexes featuring a mismatch in red. Sequences utilized were for (A) pyrene-(CH2)3CONH(CH2)6NHCO-5'-AGT ACA GTC ATC GCG-3' and for (B) pyrene(CH2)3CONH(CH2)6NHCO-5'- AGT ACA GTC ATC GCG-3', where the italicized base indicates the position of the GT or CA mismatch, respectively. 86 Table 3.1: Summary of mismatch detection with various linkers on DNA-modified HOPG electrodes. 87 3.3.4 DNA Electrochemistry with Various Intercalators Electrochemistry with DNA films on HOPG affords an extended potential range compared to that for thiolated DNA monolayers on gold (18–20, 50, 51). Thus, exploiting this characteristic, Os(phen)2dppz2+ as DNA films intercalating were investigated redox-active using probes Ru(bpy)2dppz2+ and (Figure 3.6). These metallointercalators exhibit redox couples at the edge of the potential window of Au, with larger applied biases leading to thiol desorption (50, 51). The extended potential range on HOPG allows us also to test the electrochemical activity of DNA-bound Ru(bpy)2dppz2+ and Os(phen)2dppz2+, two well-characterized metallointercalators (35–38). Surfacebound Ru(bpy)2dppz2+ and Os(phen)2dppz2+ display quasi-irreversible redox couples at the DNA-modified surface, which can be assigned to reduction of the dppz ligand (35– 38). Moreover, all of these intercalators bound to the DNA-modified electrodes exhibit linear plots of peak current as a function of scan rate, as expected for surface-bound species. Mismatch detection is also possible with metallointercalators bound to DNA at the HOPG surface (Figure 3.7). When DNA films containing a CA mismatch are electrochemically interrogated with Os(phen)2dppz2+, a significant diminution compared to fully matched DNA films in both the cathodic and anodic waves is observed. This result provides further evidence that charge transfer to the metallointercalators is in fact also DNA-mediated on HOPG. Suprisingly, electrocatalytic activity for Os(phen)2dppz2+ is evident in conjunction with ferricyanide, yielding amplification of the DNA-mediated electrochemical signal. 88 Figure 3.6: Cyclic voltammetry versus Ag/AgCl for electrodes modified with fully base paired duplexes at a 500 mV/s scan rate in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl. Bare electrodes are in blue, and DNA-modified electrodes are in red for (A) 2 µM Ru(bpy)2dppz2+ and (B) 2 µM Os(phen)2dppz2+. The sequence was pyrene-(CH2)4Pi5'-AGT ACA GTC ATC GCG-3'. 89 Figure 3.7: Cyclic voltammetry of 2 µM Os(phen)2dppz2+ at a 100 mV/s scan rate versus SCE in a buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl. The electrode modified with fully base paired DNA duplexes is in blue, and the electrode modified with duplexes featuring a CA mismatch is in red. The sequence was pyrene- (CH2)3CONH(CH2)6NHCO-5'-AGT ACA GTC ATC GCG-3', where the italicized base indicates the position of the CA mismatch. 90 It is noteworthy, however, that secondary peaks were observed to grow in after repeated cycling of the Ru or Os intercalators at the DNA-modified electrodes; this may indicate some decomposition of the DNA-bound metallointercalators upon reduction of the dppz ligand. Nonetheless, in sum, these data, with and without electrocatalysis, certainly indicate that HOPG surfaces offer an opportunity to explore the rich electrochemistry of this family of metallointercalators bound to DNA. 3.4 Implications DNA-mediated electrochemistry provides a convenient route to the detection of perturbations within the base pair stack. Electrochemical detection avoids expensive equipment associated with fluorescence measurements and radioactive labeling. Additionally, stringent hybridization conditions are not necessary, and the electrochemical signal can be easily amplified via electrocatalysis. Here we establish this same DNA-mediated electrochemistry at a graphite surface. The methodology developed here provides a convenient strategy to fabricate electroactive DNA chips on graphite. Moreover, the extended potential range on HOPG allows us to probe and exploit previously inaccessible redox-active species bound to DNA. 91 REFERENCES (1) Boon, E. M., Barton, J. K. (2002) Curr. Opin. Struct. Biol. 12, 320–329. (2) O'Neill, M. A.; Barton, J. K. (2004) Topics in Current Chemistry: Electron Transfer in DNA I (G. B. Schuster), pp. 67–115, Springer-Verlag, New York. (3) O’Neill, M. A., Barton, J. K. (2005) Charge Transfer in DNA: From Mechanism to Application (H.-A. Wagenknecht), pp. 27–75, Wiley-VCH, Hoboken. (4) Delaney, S., Barton, J. K. (2003) J. Org. Chem. 68, 6475–6483. (5) Giese, B. (2000) Acc. Chem. Res. 33, 631–636. (6) Schuster, G. B. (2000) Acc. Chem. Res. 33, 253–260. (7) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nat. Biotechnol. 21, 1192–1199. (8) Kelley, S. O.; Jackson, N. M.; Hill, M. G.; Barton, J. K. (1999) Angew. Chem., Int. Ed. 38, 941–945. (9) Liu, T., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 10160–10161. (10) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nat. Biotechnol. 18, 1096–1100. (11) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucleic Acids Res. 27, 4830–4837. (12) Boal, A. K., Barton, J. K. (2005) Bioconjugate Chem. 16, 312–321. 92 (13) Boon, E. M., Barton, J. K., Pradeepkumar, P. I., Isaksson, J., Petit, C. Chattopadhyaya, J. (2002) Angew. Chem., Int. Ed. 41, 3402–3405. (14) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nat. Biotechnol. 20, 282–286. (15) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U.S.A. 100, 12543–12547. (16) DeRosa, M. C., Sancar, A., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 10788–10792. (17) Boal, A. K., Yavin, E., Lukianova, O. A., O'Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (18) Kelley, S. O., Jackson, N. M., Hill, M. G., Barton, J. K. (1997) Bioconjugate Chem. 8, 31–37. (19) Boon, E. M., Barton, J. K. (2003) Bioconjugate Chem. 14, 1140–1147. (20) Boon, E. M., Jackson, N. M., Wightman, M. D., Kelley, S. O., Barton, J. K., Hill, M. G. (2003) J. Phys. Chem. B 107, 11805–11812. (21) McCreery, R. L. (1990) Electroanalytical Chemistry (A. J. Bard,), pp. 221– Marcel Dekker, New York. (22) Palecek, E., Jelen, F., Teijeiro, C., Fucik, V., Jovin, T. M. (1993) Anal. Chim. Acta 273, 175–186. (23) Wang, J., Palecek, E., Nielsen, P. E., Rivas, G., Cai, X., Shiraishi, H., Dontha, N., Luo, D., Farias, P. A. M. (1996) J. Am. Chem. Soc. 118, 7667–7670. (24) Wang, J., Xu, D., Kawde, A., Polsky, R. (2001) Anal. Chem. 73, 5576–5581. (25) Wang, J. (2002) Anal. Chim. Acta 469, 63–71. (26) Wang, J., Liu, G., Jan, M. R. (2004) J. Am. Chem. Soc. 126, 3010–3011. 93 (27) Zhou, L., Rusling, J. F. (2001) Anal. Chem. 73, 4780–4786. (28) Mugweru, A., Wang, B., Rusling, J. F. (2004) Anal. Chem. 76, 5557–5563. (29) Dennany, L., Forster, R. J., White, B., Smyth, M., Rusling, J. F. (2004) J. Am. Chem. Soc. 126, 8835–8841. (30) Willner, I., Katz, E. (2000) Angew. Chem., Int. Ed. 39, 1180–1218. (31) Katz, E. (1994) J. Electroanal. Chem. 365, 157–164. (32) Udit, A. K., Hill, M. G., Bittner, V. G., Arnold, F. H., Gray, H. B. (2004) J. Am. Chem. Soc. 126, 10218–10219. (33) Taft, B. J., Lazareck, A. D., Withey, G. D., Yin, A., Xu, J. M., Kelley, S. O. (2004) J. Am. Chem. Soc. 126, 12750–12751. (34) Mann, J. S., Shibata, Y., Meehan, T. (1992) Bioconjugate Chem. 3, 554–558. (35) Freidman, A. E., Chambron, J. C., Sauvage, J. P., Turro, N. J., Barton, J. K. (1990) J. Am. Chem. Soc. 112, 4960–4962. (36) Delaney, S., Pascaly, M., Bhattacharya, P. K., Han, K., Barton, J. K. (2002) Inorg. Chem. 41, 1966–1974. (37) Holmlin, R. E., Barton, J. K. (1995) Inorg. Chem. 34, 7–8. (38) Holmlin, R. E., Yao, J. A. Barton, J. K. (1999) Inorg. Chem. 38, 174–189. (39) Bowler, R., Davies, T. J., Hyde, M. E., Compton, R. G. (2005) Anal. Chem. 77, 1916–1919. (40) Sam, M., Boon, E. M., Barton, J. K., Hill, M. G., Spain, E. M. (2001) Langmuir 17, 5727–5730. (41) Kelley, S. O., Barton, J. K., Jackson, N. M., McPherson, L. D., Potter, A. B., Spain, E. M., Allen, M. J., Hill, M. G. (1998) Langmuir 14, 6781–6784. 94 (42) Zhou, D., Sinniah, K., Abell, C., Rayment. T. (2002) Langmuir 18, 8278– 8281. (43) Zhou, D., Sinniah, K., Abell, C., Rayment, T. (2003) Angew. Chem. 42, 4934– 4397. (44) Steel, A. B., Herne, T. M., Tarlov, M. J. (1998) Anal. Chem. 70, 4670–4677. (45) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. (46) Drummond, T. G., Hill, M. G., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 15010–15011. (47) Boon, E. M., Barton, J. K., Bhagat, V., Nersissian, M., Wang, W., Hill, M. G. (2003) Langmuir 19, 9255–9259. (48) Anson, F. C. (1980) J. Phys. Chem. 84, 3336–3338. (49) Oyama, N., Anson, F. C. (1980) Anal. Chem. 52, 1192–1198. (50) Wackerbarth, H., Grubb, M., Zhang, J., Hansen, A. G., Ulstrup, J. (2004) Langmuir 20, 1647–1655. (51) Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G., Whitesides, G. M. (2005) Chem. Rev. 105, 1103–1169. 95  Chapter 4: DNA-Mediated Electrochemistry of Disulfides on Graphite Adapted from: Gorodetsky, A. A., Barton, J. K. (2006) J. Am. Chem. Soc. 129, 6074– 6075. 96 ABSTRACT The electrochemistry of disulfides incorporated into the sugar-phosphate backbone of pyrene-modified DNA has been examined on highly oriented pyrolytic graphite (HOPG). Two signals, one irreversible, one reversible, are observed at -160 and -290 mV versus normal hydrogen electrode, respectively. The inclusion of a single base mismatch in the DNA duplex below the location of the disulfide leads to a significant attenuation in the electrochemical signal, while the inclusion of a mismatch above the disulfide has little effect on the electrochemistry observed. Thus disulfide reduction in the DNA backbone appears to be DNA-mediated. The redox couples found show a strong pH dependence consistent with formation of a disulfide radical anion or 2e- reduction of the disulfide to the two thiols. These data demonstrate that DNA electrochemistry can be utilized to promote disulfide reduction at a distance mediated by the DNA duplex. 97 4.1 Introduction The DNA base pair stack has been shown to mediate charge transport over significant distances (1, 2) using a variety of experiments (3–5). Both in solution and at DNA-modified surfaces, DNA-mediated charge transport (CT) chemistry exhibits exquisite sensitivity to perturbations of the intervening base pair stack such as single base mismatches (6), bulges (7, 8), and protein binding (9–11). However, the mechanisms underlying this chemistry remain poorly understood (12–16). Nonetheless, an overwhelming body of evidence indicates that DNA charge transport provides an efficient route to carry out both oxidative and reductive reactions at a distance (3–8, 17– 22). We have reported that DNA CT can lead to oxidation of thiols incorporated into a DNA duplex with concomitant disulfide bond formation in solution (23). Although the reaction proceeded on the DNA backbone, rather than within the base pair stack, DNA CT was dramatically affected by the inclusion of both proximal and distal single base mismatches. These findings indicated that oxidation of thiols did, in fact, occur via the base pair stack. Furthermore, these findings were particularly significant given the importance of thiol chemistry in vivo. Electrochemistry at DNA-modified gold surfaces has been utilized in the development of new methodologies for mismatch detection (24, 25) and in probing DNA-protein interactions (26). These investigations rely on reductive DNA CT chemistry and have frequently made use of DNA-bound, redox active substrates whose electrochemistry is exquisitely sensitive to the integrity of the base pair stack. In 98 particular, mechanistic studies of CT at DNA-modified surfaces have greatly benefited from the use of probes which are covalently attached to the DNA (27–29). Recently, DNA-modified highly oriented pyrolytic graphite (HOPG) has been developed as an attractive alternative to gold (30). DNA monolayers on HOPG have been extensively characterized via electrochemistry, AFM, and radioactive labeling. As on gold, dense monolayers with surface densities of ~ 40 pmol/cm2 are assembled in the presence of Mg2+ on graphite. Relative to gold, graphite is certainly better suited for exploring the electrochemistry of disulfide bond formation, which would be complicated by the ease of S-Au bond formation (31). Therefore, this work explores the reversible electrochemical reduction of disulfides covalently incorporated within the sugarphosphate backbone at self-assembled DNA monolayers on HOPG. 4.2 Experimental Section 4.2.1 Materials All reagents for DNA synthesis were purchased from Glen Research. Reagents used in the synthesis of pyrene- and thiol-modified DNA were purchased from Sigma in the highest purity available and used as received. SPI-1 HOPG was purchased from SPI, Inc. Buffers were prepared with Milli-Q water and filtered with a 0.2 µm filter. 4.2.2 General Oligonucleotide Synthesis Oligonucleotides were prepared using standard phosphoramidite chemistry on an ABI 392 DNA Synthesizer. DNA was purified by HPLC on a reverse phase C18 column with acetonitrile and ammonium acetate as the eluents. Unless otherwise noted, all 99 oligonucleotides were cleaved from the resin with concentrated ammonium hydroxide (NH4OH) prior to purification. The desired products were characterized by UV-visible spectroscopy and MALDI-TOF mass spectrometry. 4.2.3 Synthesis of 5’-Pyrene-Modified Oligonucleotides DNA was modified with pyrene at the 5' terminus by following a previously reported procedure (30). In brief, oligonucleotides were prepared by solid-phase synthesis with an unprotected hydroxyl group at the 5' terminus. The 5'-OH was treated with a 120 mg/mL solution of carbonyldiimidazole in dioxane for 2 h followed by an 80 mg/mL solution of 1,6-diaminohexane for 30 min. Subsequently, the free amine was treated with 1-pyrenebutyric acid, N-hydroxysuccinimide ester, resulting in the desired pyrene moiety linked to the 5' terminus. 4.2.4 Synthesis of 5’-Thiol-Modified Oligonucleotides The synthesis of DNA featuring a thiol at the 5' terminus followed previously established procedures (32). In brief, oligonucleotides were prepared by solid-phase synthesis with the commercially available C6 S-S linker phosphoramidite incorporated at the 5' terminus. After deprotection and an initial round of HPLC purification, dithiothreitol was utilized to reduce the disulfide, leaving a free thiol at the 5' terminus. The thiol terminated oligonucleotides were then purified by a second round of HPLC. 4.2.5 Synthesis of 3’-Thiol-Modified and 5’-Pyrene-Modified Oligonucleotides The synthesis of DNA featuring a thiol at the 5' terminus followed previously established procedures (32). In brief, oligonucleotides were prepared by solid-phase synthesis by utilizing the commercially available 3' C3 thiol modifier S-S CPG. The 3' thiolated sequence was further modified with pyrene at its 5' end as described above. 100 After deprotection and an initial round of HPLC purification, dithiothreitol was utilized to reduce the disulfide, leaving a free thiol at the 3' terminus and a pyrene moiety at the 5’ terminus. The thiol-terminated oligonucleotides were then purified by a second round of HPLC. 4.2.6 Preparation of Duplex DNA DNA oligonucleotides were suspended in a pH 7.1 buffer containing 5 mM Pi and 50 mM NaCl and quantified via UV-visible spectroscopy. Duplex DNA was then prepared by combining equimolar amounts of both thiolated strands with a complementary 24-mer strand. In the presence of oxygen, the DNA was hybridized by heating equimolar amounts of the three strands to 90 º C, followed by cooling to ambient temperature. In an aerobic buffer for duplex DNA, a disulfide bond is formed between two contiguous thiols incorporated within the DNA backbone in quantitative yields. The reaction is templated by the complementary strand and occurs in negligible yields for the two single strands alone (23). 4.2.7 Melting Temperatures Melting temperatures (Tm) of all the duplexes were measured using a Beckman DU 7400 spectrophotometer with a temperature control attachment. Absorption at 260 nm of equimolar DNA complements was measured every 1 ºC from 90 to 15 ºC with rate 1 ºC/min. The reverse temperature traces were measured under the same conditions to confirm the reversibility of the DNA annealing process. 4.2.8 Preparation of DNA-Modified Electrodes DNA modified surfaces were prepared as previously described (30). Briefly, the HOPG surface was cleaned with 3-M scotch tape before each experiment and mounted in 101 the electrode housing. Duplex DNA containing a pyrene moiety was deposited onto the HOPG surface in buffer, pH 7.1, containing 5 mM Pi 50 mM NaCl, and 100 mM MgCl2; Mg2+ is included to ensure a closely packed film. DNA films were allowed to form in a humidified chamber over a period of 24–48 hours. 4.2.9 Electrochemical Experiments Cyclic and square wave voltammetry experiments were performed using a CH Instruments electrochemical analyzer (Austin, TX). Unless otherwise noted, experiments were performed at ambient temperature in buffer, pH 7.1, containing 5 mM Pi and 50 mM NaCl under an argon atmosphere. A custom-built three-compartment electrochemical cell was used for all experiments with a Pt auxiliary electrode, an SPI-1 grade HOPG working electrode, and a saturated calomel (SCE) reference electrode. 4.2.10 X-Ray Photoelectron Spectroscopy Experiments Pyrene-modified duplexes were self-assembled on a clean HOPG surface and subsequently dried. XPS data were acquired using an M-Probe X-ray photoelectron spectrometer operated under ultrahigh vacuum (UHV) at 10−9−10−8 Torr, as previously described (33). 4.3 Results and Discussion 4.3.1 Surface Characterization by X-Ray Photoelectron Spectroscopy A DNA modified graphite surface is schematically illustrated in Figure 4.1. Previous AFM characterization found that DNA monolayers on graphite are smooth and featureless with the DNA oriented at a ~ 45 º angle with respect to the surface. For proper 102 Figure 4.1: Schematic illustration of charge transfer to thiols incorporated into the DNA backbone 103 interpretation of electrochemical data, some surface characterization is a prerequisite, so the presence of thiols within the DNA monolayer was confirmed here by X-ray photoelectron spectroscopy (XPS) measurements. DNA monolayers featuring thiols incorporated into the sugar-phosphate backbone reveal a broad S2p peak centered at 164 eV (Figure 4.2). The signal observed is fully consistent with previous XPS measurements of polymer films on gold featuring bound thiols (31, 34). In addition, no signal is observed at a surface modified with DNA which features no backbone modifications. These data provide supporting evidence for the presence of disulfides within the DNA film on graphite. 4.3.2 Thermal Stability of Thiol Modified Duplexes To further characterize the DNA assembly utilized for subsequent electrochemistry experiments, thermal denaturation experiments were performed. Figure 4.3 shows the sequence and melting temperature of pyrene-modified DNA duplexes featuring no thiols (red), a disulfide (black), only a 3’ thiol (light blue), a CA mismatch above the disulfide (dark blue), and a CA mismatch below the disulfide (gray). Wellmatched DNA with no thiols has a melting temperature of ~ 70 ºC, and when two thiols are included within the backbone, the duplex is destabilized to a melting temperature of ~ 60 ºC. The incorporation of a CA mismatch, regardless of location, further lowers the melting temperature by another ~ 10 ºC relative to well-matched DNA containing a disulfide (from ~ 60 ºC to ~ 50 ºC). Interestingly, the incorporation of a solitary 3’ thiol dramatically destabilizes the unmodified DNA by ~ 25 ºC while the incorporation of both the 5’ and 3’ thiols only destabilizes the unmodified DNA by ~ 10 ºC. The added thermal stability observed for 104 Figure 4.2: XPS spectra of HOPG modified with pyrenated DNA with (blue) and without (red) disulfides incorporated within the sugar phosphate backbone. Note the presence of a small signal at ~ 164 eV for the thiol-modified DNA. 105 Absorban 1.2 WM Control 1 3' Thiol Only CAMismatch Below CAMismatch Above 0.8 WM Disulfide 0.6 10 30 50 70 90 Temperature (C) Figure 4.3: Melting temperatures of DNA duplexes as monitored by a change in absorbance at 260 nm. (Red) Well-matched DNA featuring no thiols. (Black) Wellmatched DNA featuring two thiols. (Light Blue) Well-matched DNA featuring only a 3’ thiol. (Dark Blue) DNA with a CA mismatch above the disulfide. (Gray) DNA with a CA mismatch below the disulfide. The 24 mer sequence utilized in the course of these experiments was pyrene-(CH2)3-CONH-(CH2)6-NHCO-5'-ATG CAT CGA C-S-S-CA CAG TGC TGT CGT-3' plus unmodified complement. The locations of the CA mismatches are in bold italics. 106 two contiguous thiol modifications provides strong evidence for disulfide bond formation. In the absence of such a covalent bond, the opposite effect would be observed after incorporation of two potentially destabilizing modifications. 4.3.3 Electrochemistry of Disulfides Incorporated in the Phosphate Backbone The DNA-mediated electrochemical reduction of a disulfide incorporated within the DNA sugar-phosphate backbone is schematically illustrated in Figure 4.1. The corresponding electrochemistry of a well-matched DNA monolayer featuring a disulfide is illustrated in Figure 4.4 by square wave voltammetry (SWV) and Figure 4.5 by cyclic voltammetry (CV). Two signals centered at -160 ± 10 mV and at -290 ± 10 mV versus the normal hydrogen electrode (NHE) are observed. The first peak is electrochemically irreversible, and the second peak is electrochemically reversible. Neither signal is found for monolayers lacking the 3' thiol (Figure 4.4). Furthermore, a plot of peak current as a function of scan rate is linear for the cathodic waves of the reversible signal (Figure 4.5), as expected for a surface bound species (35). The reaction is mediated by the base pair stack, as evidenced by the effect of base mismatches on the second, reversible peak. For well-matched DNA featuring a disulfide, the reversible electrochemical signal exhibits a peak current of 150 ± 30 nA (Figure 4.4). The incorporation of a CA mismatch below the thiols leads to significant attenuation of the electrochemical signal with a resulting peak current of 4 ± 7 nA. On the other hand, the incorporation of a CA mismatch above the thiols has little effect on the signal with a corresponding peak current of 124 ± 7 nA. It is interesting to note that both sequences which feature the CA mismatch are destabilized relative to the well-matched DNA. Therefore, the dramatic attenuation of the signal observed with the CA mismatch below 107 Figure 4.4: SWV at 15 Hz (left) and schematic (right) for various DNA monolayers. Top: Well-matched DNA featuring two thiols (black) and well-matched DNA featuring one thiol (red). Bottom: DNA featuring a CA mismatch above the disulfide (black) and DNA featuring a CA mismatch below the disulfide (red). The 24 mer sequence utilized in the course of these experiments was pyrene-(CH2)3-CONH-(CH2)6-NHCO-5'-ATG CAT CGA C-S-S-CA CAG TGC TGT CGT-3' plus unmodified complement. The locations of the CA mismatches are in bold italics. 108 A. B. Figure 4.5: A: Cyclic voltammetry at various scan rates of an electrode modified with duplex DNA featuring a disulfide in the sugar phosphate in a pH=7.1 buffer containing 5 mM Na+ Pi, 50 mM NaCl. B: The corresponding plot of peak current as a function of scan rate for the reversible second wave. 109 cannot be simply explained by differences in thermal stabilities between DNA with and without a mismatch. We can interrogate these redox signals further by varying the solution pH (Figure 4.6). Two proton coupled steps are observed electrochemically. These signals are similar to the two separate one-electron reduction steps that have been previously reported for the electrochemistry of the cysteine active site of thioredoxins at similar potentials (36, 37). As the pH, is changed, the amplitude of the reversible signal remains nearly constant, and the midpoint potential of this signal shows a linear pH dependence with a slope of 44 ± 5 mV per pH unit. Although a slope of 59 mV per pH unit is expected for the protoncoupled transfer of one electron, similar nonidealized pH responses have also been observed for DNA polyion films on graphite (38). On the other hand, the amplitude of the irreversible signal is significantly affected by changes in pH. At acidic pH, the irreversible cathodic wave is almost completely suppressed, while, at basic pH, the irreversible cathodic wave is substantially enhanced. A plausible reaction scheme accounts for these electrochemical data. Although the reduction of a disulfide to two thiols often proceeds through many intermediates, it is a net 2 e-, 2 H+ process (39–41). For our system, we can postulate that the addition of the first electron leads to a disulfide radical anion, which we assign to the irreversible peak; at acidic pH, this irreversible reduction is suppressed (42, 43). The reversible addition of two electrons to the disulfide results in disproportionation with concomitant free thiol formation. Thus, the reversible formation of free thiols may be a concerted 2 e-, 2 H+ process. 110 A. Potential (V vs. NHE) B. -0.3 -0.2 -0.1 4 5 6 7 8 9 pH Figure 4.6: (A) Cyclic voltammetry of well-matched DNA featuring a disulfide in 5 mM Na+ Pi, 50 mM NaCl at a 50 mV/s scan rate. The black, blue, and red traces represent pH 7.8, 6.6, and 4.7, respectively. The blue and red traces have been offset for clarity. (B) The corresponding plot of peak potential as a function of pH. 111 4.4 Implications We have therefore shown that DNA-mediated electrochemistry can promote reactions at a distance on the DNA sugar-phosphate backbone. We had earlier seen that breaks in the backbone cause little attenuation in DNA-mediated charge transport through the base stack. We can reconcile these observations by noting relative current densities for DNA-mediated disulfide reduction, 1.8 µA/cm2, versus well-stacked intercalator reduction, ~ 80 µA/cm2 for daunomycin (27–29). DNA-mediated reactions neighboring but not coupled into the sugar-phosphate backbone are therefore less efficient. Nonetheless, these results expand the reactions that can be achieved through DNAmediated charge transport chemistry. When one notes that many DNA regulatory proteins utilize disulfide switches in close proximity to the DNA backbone (44–46), these results may be important to consider in a biological context. 112 REFERENCES (1) Delaney, S., Barton, J. K. (2003) J. Org. Chem. 68, 6475–6483. (2) O'Neill, M. A., Barton, J. K. (2005) Charge Transfer in DNA: From Mechanism to Application (H.-A. Wagenknecht), pp. 27–75, Wiley-VCH, Hoboken. (3) Hall, D. B., Holmlin, R. E., Barton, J. K. (1996) Nature 382, 731–735. (4) Nunez, M. E., Hall, D. B., Barton, J. K. (1999) Chem. Biol. 6, 85–97. (5) Henderson, P. T., Jones, D., Hampikian, G., Kan, Y., Schuster, G. B. (1999) Proc. Natl. Acad. Sci. U.S.A. 96, 8353–8358. (6) Bhattacharya, P. K., Barton, J. K. (2001) J. Am. Chem. Soc. 123, 8649–8656. (7) Hall, D. B., Barton, J. K. (1997) J. Am. Chem. Soc. 119, 5045–5046. (8) Dandliker, P. J., Holmlin, R. E., Barton, J. K. (1997) Science 275, 1465–1468. (9) Rajski, S. R., Kumar, S., Roberts, R. J., Barton, J. K. (1999) J. Am. Chem. Soc. 121, 5615–5616. (10) Rajski, S. R., Barton, J. K. (2001) Biochemistry 40, 5556–5564. (11) Boon, E. M., Pope, M. A., Williams, S. D., David, S. S., Barton, J. K. (2002) Biochemistry 41, 8464–8470. (12) Meggers, E., Michel-Beyerle, M. E., Giese, B. (1998) J. Am. Chem. Soc. 120, 12950–12955. (13) Lewis, F. D., Liu, X. Y., Liu, J. Q., Miller, S. E., Hayer, R. T., Wasielewski, M. R. (2000) Nature 406, 51–53. 113 (14) Takada, T., Kawai, K., Fujitsuka, M., Majima, T. (2004) Proc. Natl. Acad. Sci. U.S.A. 101, 14002–14006. (15) O'Neill, M. A., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 11471–11483. (16) Shao, F. W., O' Neill, M. A., Barton, J. K. (2004) Proc. Natl. Acad. Sci. U.S.A. 101, 17914–17919. (17) Vicic, D. A., Odom, D. T., Nunez, M. E., Gianolio, D. A., McLaughlin, L. W., Barton, J. K. (2000) J. Am. Chem. Soc. 122, 8603–8611. (18) Giese, B., Carl, B., Carl, T., Carell, T., Behrens, C., Hennecke, U., Schiemann, O., Feresin, E. (2004) Angew. Chem. Int. Ed. 2004, 43, 1848– 1851. (19) Ito, T., Rokita, S. E. (2004) Angew. Chem. Int. Ed. 43, 1839–1842. (20) Wagner, C., Wagenknecht, H.-A. (2005) Chem.-Eur. J. 11, 1871–1876. (21) Kaden, P., Mayer-Enthart, E., Trifonov, A., Fiebig, T., Wagenknecht, H.-A. (2005) Angew. Chem. Int. Ed. 44, 1636–1639. (22) Manetto, A., Breeger, S., Chatgilialoglu, C., Carell, T. (2006) Angew. Chem., Int. Ed. 45, 318–321. (23) Takada, T., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 12204–12205. (24) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucl. Acids Res. 27, 4830–4837. (25) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nat. Biotechnol. 18, 1096–1100. (26) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nat. Biotechnol. 20, 282–286. 114 (27) Kelley, S. O., Jackson, N. M., Hill, M. G., Barton, J. K. (1999) Angew. Chem. Int. Ed. 38, 941–945. (28) Drummond, T. G., Hill, M. G., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 15010–15011. (29) Liu, T., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 10160–10161. (30) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. (31) Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G., Whitesides, G. M. (2005) Chem. Rev. 105, 1103–1170. (32) Kelley, S. O., Barton, J. K., Jackson, N. M., Hill, M. G. (1997) Bioconjugate Chem. 8, 31–37. (33) Bansal, A., Li, X. L., Yi, S. I., Weinberg, W. H., Lewis, N. S. (2001) J. Phys. Chem. B 105, 10266–10277. (34) Johnson, P. A., Levicky, R. (2004) Langmuir 20, 9621–9627. (35) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. (36) Salamon, Z., Tollin, G., Hirasawa, M., Gardet-Salvi, L., Stritt-Etter, A.-L., Knaff, D. B., Schurmann, P. (1995) Biochim. Biophys. Acta 1230, 114–118. (37) Johnson, D. L., Polyak, S. W., Wallace, J. C., Martin, L. L. (2004) Lett. Pept. Sci. 10, 495–500. (38) Munge, B., Estavillo, C., Schenkman, J. B., Rusling, J. F. (2003) ChemBioChem 4, 82–89. (39) Gilbert, H. F. (1995) Methods Enzymol. 251, 8–28. (40) Schoenich, C. (1995) Methods Enzymol. 251, 45–55. 115 (41) Houee-Levin, C. (2002) Methods Enzymol. 353, 35–44. (42) Goebl, M., Bonifacic, M., Asmus, K.-D. (1984) J. Am. Chem. Soc. 106, 5984– 5988. (43) Lmoumene, C., Conte, D., Jacquot, J.-P., Houee-Levin, C. (2000) Biochemistry 39, 9295–9301. (44) Das, K. C., White, C. W. (2002) Proc. Natl. Acad. Sci. U.S.A. 99, 9617–9618. (45) Sevier, C. S., Kaiser, C. A. (2002) Nat. Rev. Mol. Cell Biol. 3, 836–847 (46) Tooze, J., Branden, C. (1998) Introduction to Protein Structure, 2nd ed., Garland Publishing: New York. 116  Chapter 5: Coupling into the Base Pair Stack is Necessary for DNA-Mediated Electrochemistry Adapted from: Gorodetsky, A. A., Green, O., Yavin, E., Barton, J. K. (2007) Bioconjugate Chem. 18, 1434–1441. Iodoanthraquinone was synthesized by Omar Green, and oligonucleotides containing the TEMPO label were prepared by Eylon Yavin. 117 ABSTRACT The electrochemistry of DNA films modified with different redox probes linked to DNA through saturated and conjugated tethers was investigated. Experiments feature two redox probes bound to DNA on two surfaces: anthraquinone (AQ)-modified uridines incorporated into thiolated DNA on gold (Au) and 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO)-modified uridines in pyrene-labeled DNA on highly oriented pyrolytic graphite (HOPG). The electrochemistry of these labels when incorporated into DNA has been examined in DNA films containing both well-matched and mismatched DNA. DNA-mediated electrochemistry is found to be effective for the TEMPO probe linked with an acetylene linker but not for a saturated TEMPO connected through an ethylenediamine linker. For the AQ probe, DNA-mediated electrochemistry is found with an acetylene linker to uridine but not with an alkyl chain to the 5' terminus of the oligonucleotide. Large electrochemical signals and effective discrimination of intervening base mismatches are achieved for the probes connected through the acetylene linkages, while probes connected through saturated linkages exhibit small electrochemical signals associated only with direct surface-to-probe charge transfer and poor mismatch discrimination. Thus DNA electrochemistry with these probes is dramatically influenced by the chemical nature of their linkage to DNA. These results highlight the importance of effective coupling into the π-stack for long-range DNAmediated electrochemistry. 118 5.1 Introduction Numerous studies in solution have shown that the DNA duplex can serve as an effective medium to transport charge (1–5). DNA-mediated charge transport (CT) in solution exhibits exquisite sensitivity to perturbations in the intervening base stack such as single base mismatches (6), bulges (7, 8), and protein binding (9–11). Previous work has shown that self-assembled DNA monolayers on both gold and graphite can also mediate charge transport over significant distances (12–14). In particular, electrochemistry at DNA-modified electrodes has been demonstrated in both a chipbased format for mismatch detection (15, 16, 17) and as a tool for monitoring protein/DNA interactions (18–23). Assays based on DNA CT have frequently made use of solution-borne substrates that interact with duplex DNA at the electrode surface (12–15). To further develop electrochemical detection strategies based on DNA CT, it is necessary to explore electroactive probes that are covalently tethered to duplex DNA. Electrochemistry of DNA monolayers modified with daunomycin, ferrocene, methylene blue, and anthraquinone has shown that these redox-active reporters covalently linked to DNA can function as DNA-bound electrochemical probes (24–34). However, important questions remain regarding the nature of charge transfer to these probes, which may proceed either via the base pair stack of DNA or by direct contact of the probe with the surface. Daunomycin has displayed particularly large electrochemical signals when used as a covalently attached probe (24–26). The size of the signals associated with this probe likely reflect its intimate intercalation within the base pair stack (35, 36), as can be seen 119 in crystal structures of daunomycin bound to DNA. Furthermore, breaks in the sugar phosphate backbone were shown to have no significant effect on the electrochemistry of covalently attached daunomycin (26); base mismatches, on the other hand, cause a dramatic attenuation of the signal (24). These observations are consistent with DNA CT occurring through the base pair stack rather than through the sugar-phosphate backbone. Ferrocene/DNA conjugates have been employed in DNA electrochemistry experiments and, in particular, to examine the importance of the linkage between DNA and the redox active probe (27). Interrupting conjugation of the ferrocene label with the base pair stack results in an almost complete loss of the electrochemical signal. Furthermore, the location of the ferrocene label within a DNA monolayer was found to affect dramatically the magnitude of the electrochemical signals (29, 31). A well coupled ferrocene-uridine derivative exhibited a larger electrochemical signal at the top of the DNA monolayer than in close proximity to the surface (29). However, the opposite effect was observed for a poorly coupled ferrocene; an alkyl chain-derivatized ferrocene connected to the top of a DNA monolayer exhibited a smaller signal than the same ferrocene moiety incorporated in close proximity to the surface (31). There remains some controversy about the effects of DNA conformation on electrochemical behavior of ferrocene, however. Recently, Anne and Demaille explored the electrochemistry of a ferrocene label attached to the 5' terminus of duplex DNA (37). These ferrocene labels were not conjugated into the base pair stack but still afforded strong electrochemical signals. Elastic bending of 5' ferrocene-labeled DNA was suggested as the cause of the electrochemical signals observed. It was concluded that bending of the DNA could account for the electrochemistry of all DNA-bound probes. 120 This conclusion was in direct contrast to those based upon the ferrocene data already described (29, 31). Further, these investigations did not account for the role of the linkage between DNA and ferrocene, nor the effect of sequence context on electrochemical signals. Experiments with covalently tethered methylene blue have underscored the importance of intercalative stacking for DNA-mediated electrochemistry (28). In low salt buffers, intercalation is detected electrochemically in DNA films containing methylene blue tethered distally at the top of the film. Simply changing to a high ionic strength buffer leads to the complete loss of electrochemical signal in the DNA film; methylene blue, though still covalently attached, cannot bind by intercalation under these conditions. Thus intercalative stacking appears necessary for efficient reduction of the attached probe by DNA-mediated CT. Notably, recent studies have also demonstrated the effectiveness of methylene blue in assays based on surface-to-probe charge transfer without DNA mediation (38, 39). In this work we focus on two established probes of DNA CT to explore in more detail how the linkage to DNA affects the efficiency of DNA-mediated electrochemistry. We have utilized two different redox active markers that can be covalently attached to DNA. Anthraquinone (AQ) has been extensively studied in electrochemical assays based on DNA-mediated CT. For example, AQ has been applied to the electrochemical monitoring of primer extension reactions at Au surfaces (29). Additionally, AQmodified uridines have been used for detection of single nucleotide polymorphisms (32, 33) and of duplex-to-triplex conversions at DNA-modified Au surfaces (34). Recent studies in our group have also examined uridine modified with 2,2,6,6-tetramethyl-3,4- 121 dehydro-piperidine-N-oxyl (3,4-dehydro-TEMPO) as a probe of DNA CT from a redoxactive protein to DNA (40). Notably, the acetylene linkage connecting the probe to the uridine was shown to be crucial for determining its effectiveness as an electron trap for DNA CT using electron paramagnetic resonance (EPR) spectroscopy. Here we present the electrochemistry of DNA duplexes featuring AQ-modified uridines on Au and TEMPO-modified uridines on highly oriented pyrolytic graphite (HOPG). The observations reported for the AQ and TEMPO probes make them particularly well suited for establishing a correlation between the linkage connecting the probe to the DNA and the effectiveness of the probe in assays based on DNA-mediated electrochemistry. On DNA-modified HOPG, we examine the difference between a modified 3,4-dehydro-TEMPO moiety connected to the base pair stack via an acetylene linker and a saturated TEMPO connected via an ethylenediamine linker. On DNAmodified Au, we examine the difference between an AQ moiety connected to the base pair stack via an acetylene linker and an AQ linked via an alkyl chain to the 5' terminus of the DNA. Electrochemistry with these probes is dramatically influenced by the chemical nature of their linkage to DNA. 5.2 Experimental Section 5.2.1 Materials All phosphoramidites and reagents for DNA synthesis were purchased from Glen Research with the exception of the 5-ethynyluracil phosphoramidite, which was purchased from Berry and Associates. 122 All organic reagents (acetonitrile, dichloromethane, N,N-dimethylformamide, pyridine, tetrahydrofuran, triethylamine, and 40% aqueous methylamine solution) were purchased from Aldrich in the highest available purity and used as received. All buffers were freshly prepared and filtered using a 0.45 µm filter prior to use. Tetrakis(triphenylphosphine)Pd(0) was purchased from Strem Chemicals. Technical grade 2-aminoanthraquinone was obtained from Aldrich, and anthraquinone-2-carbonyl chloride was purchased from TCI America. The synthesis of 2-iodoanthraquinone has been previously described (41). TEMPO was purchased from Aldrich. The preparation of 4-diamoethylene-2,2,6,6-tetramethyl-3,4dehydropiperidine-N-oxyl (ethylenediamine TEMPO) (42) and 4-ethynyl-2,2,6,6tetramethyl-3,4-dehydropiperidine-N-oxyl (acetylene-3,4-dehydro-TEMPO) (40, 43) followed established literature procedures. 5.2.2 Oligonucleotide Synthesis Oligonucleotides were prepared using standard phosphoramidite chemistry on an ABI 392 DNA Synthesizer. DNA was purified by HPLC on a reverse phase C18 column with acetonitrile and ammonium acetate as the eluents. Unless otherwise noted, all oligonucleotides were cleaved from the resin with concentrated ammonium hydroxide (NH4OH) prior to purification. The desired products were characterized by UV-visible spectroscopy and MALDI-TOF mass spectrometry. 5.2.3 Synthesis of Pyrene-Modified Oligonucleotides DNA was modified with pyrene at the 5' terminus by following a previously reported procedure (13). In brief, oligonucleotides were prepared by solid-phase synthesis with an unprotected hydroxyl group at the 5' terminus. The 5'-OH was treated with a 120 mg/mL solution of carbonyldiimidazole in dioxane for 2 hours followed by 123 an 80 mg/mL solution of 1,6-diaminohexane for 30 min. Subsequently, the free amine was treated with 1-pyrenebutyric acid, N-hydroxysuccinimide ester, resulting in the desired pyrene moiety linked to the 5' terminus. The oligonucleotides were deprotected with concentrated NH4OH at 60 ºC for 6 hours. 5.2.4 Synthesis of Thiol-Modified Oligonucleotides The synthesis of DNA featuring a thiol at the 5' terminus followed previously established procedures (44). In brief, oligonucleotides were prepared by solid-phase synthesis with the commercially available C6 S-S linker phosphoramidite incorporated at the 5' terminus. The oligonucleotides were deprotected in NH4OH at 60 C for 12 h. Dithiothreitol (100 mM) was subsequently added to reduce the disulfide, leaving a free thiol at the 5' terminus. 5.2.5 Synthesis of TEMPO-Modified Oligonucleotides The synthesis of acetylene 3,4-dehydro-TEMPO-modified DNA has been previously described (40). The 5' DMT-protected 5-iodouridine was Pd coupled to the acetylene-modified 3,4-dehydro-TEMPO in THF. The 3'-OH of the product of this reaction was subsequently converted to its phosphoramidite form. The phosphoramidite was dissolved in acetonitrile and used immediately for DNA synthesis. The synthesis of the ethylenediamine TEMPO-modified DNA was carried out according to a protocol adapted from the literature (42). DNA featuring a 5bromouridine at the position targeted for modification was synthesized on a solid support. A ~ 0.5 M aqueous solution of the ethylenediamine TEMPO was added to the 5-bromouridine-modified DNA on solid support with the resulting slurry incubated at 60 ºC overnight. This incubation resulted in the ethylenediamine TEMPO replacing the 124 bromine, and in cleaving the DNA from the bead and deprotecting the bases. Prior to HPLC purification, the aqueous phase was washed with dichloromethane to remove excess ethylenediamine TEMPO. 5.2.6 Synthesis of Anthraquinone-Modified Oligonucleotides The modification of DNA with anthraquinone was carried out as shown in Scheme 5.1. 2-Iodoanthraquinone was synthesized according to literature precedent (41). DNA containing a 5-ethynyluracil at the position targeted for modification was synthesized on a solid support. The beads were placed in an oven-dried flask which had been charged with 2.5 mL of DMF/Et3N (3.5:1.5). 2-Iodoanthraquinone (60 mmol) and CuI (60 mmol) were added to the flask, and the solution was flushed with argon. Pd(PPh3)4 (30 mmol) was then added to the flask, and the solution was again flushed with argon. The mixture was allowed to stir at ambient temperature for 3 hours under argon. Subsequently, the beads were washed with a 5% (w/v) EDTA solution, DMF, CH3CN, and CH2Cl2 in succession to remove excess reagents. The DNA was cleaved from the resin and deprotected at 60 ºC for 20 minutes with a 1:1 mixture of 40% aqueous methylamine and a saturated ammonium hydroxide solution. m/z calculated 5417, found 5416. λmax nm (M-1 cm-1) 260 ( 158 400), 320 ( 58 000), 394 ( 43 800). Attachment of AQ at the 5' terminus of DNA was carried out using procedures modified from the literature and Glen Research protocols (44-47). DNA with a sixmethyl carbon chain terminated with a monomethoxytrityl (MMT)-protected amine was incorporated at the 5' terminus on the solid support. The MMT-protecting group was removed from the 5' amine with 3% trichloroacetic acid in CH2Cl2. A 0.5 M solution of 125 Scheme 5.1: Conditions for synthesis of 2-Iodoanthraquinone and its coupling to ethynyluracil-modified DNA. The synthesis of 2-iodoanthraquinone was performed according to literature procedure (41). The coupling of the 2-iodoanthraquinone to the ethynyluracil-modified DNA was performed while the DNA was still attached to controlled pore glass (CPG) solid support. 126 anthraquinone-2-carbonyl chloride in CH2Cl2/Et3N (10:1) was added to the aminemodified DNA on solid support. The resulting slurry was incubated overnight. The beads were subsequently washed with copious amounts of CH2Cl2 to remove excess reagents. The DNA was cleaved from the resin at 60 ºC for 20 minutes with a 1:1 mixture of 40% aqueous methylamine and a saturated ammonium hydroxide solution. m/z calculated 5615, found 5614. 5.2.7 Formation of DNA Monolayers DNA films were formed on SPI-1 grade HOPG electrodes (SPI, Inc.) with a surface area of 0.08 cm2, or on commercially available Au electrodes (BAS, Inc.) with a surface area of 0.02 cm2, as previously described (12, 13). Duplex DNA (in 5 mM sodium phosphate, 50 mM NaCl, pH 7.1) was formed by combining equimolar amounts of the pyrene or thiol-modified strand with TEMPO or AQ-labeled complements, respectively. Duplex DNA (25-50 µM) was deposited onto HOPG or Au over a period of 24-48 hours. Mg2+ was included to minimize electrostatic repulsion between the duplexes and ensure well-packed monolayers on both HOPG and Au (12, 13). Longer incubation times (48 hours) and lower incubation temperatures (4 ºC) generally improved film formation. Electrodes being compared were prepared under equivalent conditions. Also, electrodes were brought to ambient temperature prior to electrochemical analysis. 5.2.8 Electrochemical Measurements Electrochemical data were collected using a CH Instruments 760B potentiostat. Electrochemical experiments were performed at ambient temperature in 5 mM sodium phosphate, 50 mM NaCl, pH 7.1, under anaerobic conditions. All measurements 127 reported for working electrodes were taken versus a platinum (Pt) auxiliary. Either a saturated calomel electrode or a silver/silver chloride electrode (Ag/AgCl) was used as the reference electrode. Square wave voltammetry was performed at 15 Hz, and cyclic voltammetry was performed at 50 mV/s. 5.2.9 Surface Characterization All atomic force microscopy (AFM) measurements were performed using an MFP-3D atomic force microscope from Asylum Research, Inc. (Santa Barbara, CA). The HOPG block was mounted in a custom designed glass chamber before being modified with DNA as described. AFM measurements were performed in buffer containing 5 mM sodium phosphate, 50 mM NaCl, pH 7.1, either in contact (constant force) or tapping mode. Commercially available silicon tips were utilized for all experiments. 5.3 Results 5.3.1 Electrochemical Probes and DNA Sequences The electrochemical probes and sequences examined are illustrated in Figure 5.1. Two redox active probes were tested, AQ and TEMPO-based moieties, and the probes were each linked to the base pair stack in two ways. The AQ was attached to a modified uridine via an acetylene linker and to the 5' phosphate terminus via a six-carbon alkane linker. The synthesis of the acetylene-linked AQ moiety involved coupling of 2iodoanthraquinone to ethylnyl-modified DNA (Scheme 5.1). Although DNA has previously been modified via Sonogashira conditions (48, 49), these syntheses involved 128 halogenated bases incorporated into DNA coupled to ethynyl-containing molecules. 3,4Dehydro-TEMPO was attached to a modified uridine by an acetylene linker, and TEMPO was connected via an ethylenediamine linker. The acetylene linkage and the 3,4-dehydro-TEMPO probe were designed to allow for efficient coupling of the probe into the base pair stack, whereas the alkane linkage was intended to insulate the probe electronically from the base pair stack. TEMPO experiments were performed with a 15mer sequence modified with a pyrene on the complementary strand. The 15-mer sequence selected to compare the acetylene and saturated TEMPO probes is that utilized extensively by our laboratory (12, 15) and others (27) in characterizing different redox probes for DNA-mediated redox chemistry. Many experiments have been used to demonstrate, however, for a variety of probes that DNA-mediated electrochemistry is independent of DNA sequence and length (< 30 bp) (12, 13, 15, 18, 24–26). It should be noted that electrochemistry experiments were also carried out on the 10-mer TEMPO probes that had been utilized in solution experiments with similar results (data not shown) (40). Anthraquinone experiments are described using a 17-mer oligonucleotide sequence containing a 5'-thiolate on the strand complementary to that modified with AQ. Several different AQ-modified DNA sequences 14–17 base pairs in length were prepared with comparable results obtained. Since neither of these redox probes intercalate directly into the DNA base stack, they depend upon their links to the stacked uridine within the helix for π-coupling. Hence parallel experiments with these assemblies should give insight into the effect of DNA-probe linkages on DNA-mediated electrochemistry. Two different surfaces, Au and graphite, are utilized for these electrochemical experiments. Studies involving AQ-modified DNA were performed on standard Au 129 electrodes since the aromatic AQ moiety may directly adsorb to the graphite surface, impeding proper formation of DNA monolayers. Experiments utilizing TEMPOmodified DNA take advantage of the extended potential range afforded by HOPG. Since thiols are known to undergo oxidation at the large positive biases necessary for oxidation of TEMPO, HOPG is more suitable for exploring the electrochemistry of this marker. Electrochemical measurements were used to assess the CT yields from the surface to the redox active probe. The efficiency of DNA CT was determined by direct comparison of peak currents and integrated charge obtained using square wave and cyclic voltammetry. Peak potentials were obtained using SWV and are reported versus NHE. 5.3.2 Electrochemistry of the Acetylene-Linked Probes Large electrochemical signals are observed for probes linked to the DNA via an acetylene unit. Figure 5.2 shows the SWV of the acetylene-linked AQ and TEMPO moieties. A reversible electrochemical signal is found at -301 (±7) mV versus NHE for AQ-modified DNA on Au, and the signal is stable for a minimum of two hundred scans under anaerobic conditions. The magnitudes and potentials of the AQ signals are similar to those previously reported (29-34). As expected for a surface-bound species, a plot of peak current as a function of scan rate is linear for the DNA-bound AQ as shown in Figure 5.3 (50), and the electron-transfer rate was estimated to be approximately 30 s-1. This electron-transfer rate is consistent with previous electrochemical studies of DNAmodified surfaces using similar alkylthiol linkages to bind the DNA to the Au surface (24, 26). A reversible electrochemical signal is also observed at 810 (±20) mV versus 130 Figure 5.1: Sequences and probes used during the course of these studies. The DNA sequence used for each electrochemical marker is shown below the structures of the markers. Top: structures of well- and poorly coupled AQ; -SH indicates the position of thiol modification on DNA. Bottom: structures of well- and poorly coupled TEMPO and its derivative; Py- indicates the position of pyrene modification on DNA. The location of the mismatch is underlined, and the location of the electrochemical probe is indicated by uracil (U). 131 Figure 5.2: Square wave voltammetry of AQ and TEMPO-modified DNA. Voltammograms for DNA modified with the acetylene-linked AQ (top left), the acetylene-linked 3,4-dehydro-TEMPO (top right), the alkane-linked AQ (bottom left), and the alkane-linked TEMPO (bottom right) are shown. Voltammograms of wellmatched DNA are shown in black, and voltammograms of DNA featuring a CA mismatch are shown in red. All experiments were performed in 5 mM sodium phosphate, 50 mM NaCl, pH 7.1 buffer. 132 A. B. 200 Peak Current (n 180 160 140 120 100 80 60 40 20 0 0 100 200 300 400 500 600 700 800 Scan Rate (mV) Figure 5.3: Cyclic voltammetry of AQ-modified DNA at various scan rates (A) and the corresponding plot of peak current as a function of scan rate (B). The experiment was performed in 5 mM sodium phosphate, 50 mM NaCl, pH 7.1 buffer. 133 NHE for the 3,4-dehydro-TEMPO-modified DNA on graphite. The potential of the signal observed for the 3,4-dehydro-TEMPO-modified DNA is similar to that previously reported for the free label in solution (40). It is important to note that the TEMPO signal is electrochemically unstable and rapidly decays upon repeated scanning when attached to an electrode or to graphite felt (51, 52). Therefore, a plot of peak current as a function of scan rate could not be easily obtained for this marker. As expected, controls run with well-matched, unmodified DNA reveal no electrochemical signals at the potentials reported for the 3,4-dehydro-TEMPO and AQ-modified DNA. The introduction of a mismatch below the two labels causes a significant attenuation of the electrochemical signal (Figure 5.2 and Table 5.1). For AQ-modified DNA, the peak current is reduced from 160 (±50) nA for well-matched DNA to 40 (±20) nA for DNA featuring a CA mismatch. For the 3,4-dehydro-TEMPO-modified DNA, the peak current is reduced from 250 (±100) nA for well-matched DNA to 30 (±15) nA for DNA featuring a CA mismatch below the probe. It is noteworthy that the introduction of a CA mismatch above the 3,4-dehydro-TEMPO probe has little effect on the magnitude of the peak current (Figure 5.4). 5.3.3 Electrochemistry of the Poorly Coupled Probes Small electrochemical signals are observed for probes connected to the DNA via an alkane linkage. Figure 5.2 also shows SWVs for the alkane-linked AQ and TEMPO moieties. A small signal centered at -330 (±60) mV versus NHE, whose reversibility cannot be judged, is observed for the AQ-modified DNA on Au. The size of this signal makes it difficult to obtain a plot of peak current as a function of scan rate. The signal is reduced by almost 2 orders of magnitude in comparison to the acetylene-linked AQ 134 Figure 5.4: Square wave voltammetry of a surface featuring well-matched DNA modified with an acetylene-TEMPO label (black) and DNA with a CA mismatch above the acetylene-TEMPO label (red). The sequence was pyrene- (CH2)3CONH(CH2)6NHCO-5’-CTA CAG TCG T-3’ where the italicized base indicates the location of the TEMPO and the bold base indicates the location of the mismatch. The experiments were performed in 5 mM sodium phosphate, 50 mM NaCl, pH 7.1 buffer. 135 Table 5.1: Electrochemical characteristics of DNA monolayers modified with electroactive moieties 136 probe (Table 5.1) with a drop in peak current from 160 (±50) nA to 9 (±4) nA. For the ethylenediamine-linked TEMPO label, a reversible signal is found at 800 (±10) mV versus NHE. Unlike the AQ, the magnitude of the TEMPO signal is reduced by only 4fold upon altering the linkage from an acetylene to an alkane; the peak current drops from 250 (±100) nA for the acetylene-linked 3,4-dehydro-TEMPO to 60 (±20) nA for the ethylenediamine-linked TEMPO. As with the AQ case, the smaller size and instability of the TEMPO signal make it difficult to plot peak current as a function of scan rate. In contrast to that seen for the acetylene-linked probes, the introduction of a mismatch below the alkane-linked AQ and TEMPO labels does not lead to substantial reductions in the electrochemical signals (Figure 5.2). Incorporation of a CA mismatch below the alkane AQ has virtually no effect on the electrochemical signal. The electrochemical signals are small for both well-matched and mismatched DNA making quantitative evaluation difficult. Similarly, incorporation of a CA mismatch below the ethylenediamine-linked TEMPO does not lead to a diminution in the electrochemical signal. Well-defined electrochemical signals are observed with and without the mismatches for these ethylenediamine-linked, TEMPO-modified DNAs. Peak currents of 60 (±20) nA are found for well-matched TEMPO-modified DNA, and peak currents of 70 (±10) nA are found for DNA featuring a CA mismatch below the TEMPO. 5.3.4 Surface Characterization To aid in the proper interpretation of electrochemical data, surface characterization is necessary. DNA monolayers on both HOPG and Au have previously been examined using scanning probe microscopy, radioactive labeling, and 137 Figure 5.5: Tapping mode AFM characterization of DNA-modified and bare HOPG in 5 mM sodium phosphate, 50 mM NaCl, pH 7.1 buffer. (A) Well-matched and unmodified DNA on HOPG. (B) Well-matched DNA featuring an acetylene-TEMPO uridine. (C) Unmodified HOPG. The sequences were pyrene-(CH2)3CONH(CH2)6NHCO-5’-CTA CAG TCG T-3’ plus unmodified complement (A) or plus TEMPO modified complement (B). 138 electrochemical techniques (13, 53, 54, 55). Previous AFM measurements found smooth films with depths corresponding to duplexes oriented at a ~45 angle with respect to the surface. As a complement to our electrochemical studies, we also examined DNA monolayers with or without appended electrochemical markers on HOPG (Figure 5.5). These monolayers were found to be indistinguishable from those previously examined. 5.4 Discussion 5.4.1 Electrochemical Reduction of the Acetylene-Linked Probes Is Mediated by the DNA Duplex For efficient DNA-mediated CT, it is necessary for the electrochemical probe to be directly coupled into the base pair stack. Our data provide strong evidence that the acetylene-linked probes are reduced in a DNA-mediated fashion. The largest signals are found for 3,4-dehydro-TEMPO and AQ moieties connected to the DNA by an acetylene linker. The inclusion of a CA mismatch below either of these probes significantly attenuates the electrochemical signal, a hallmark of DNA-mediated CT. Interestingly, previous reports of DNA modified with AQ-labeled uridines allow for a strong correlation to be drawn between the nature of the DNA-probe linkage and the size of the electrochemical signal. The signals obtained here for an unsaturated tether integrate to current densities of 7900 nA/cm2. Gooding and co-workers have previously reported the electrochemistry of a partially saturated tether whose current densities integrate to a value of ~ 1500 nA/cm2 (29, 30). Further, Okamoto and co-workers have found current densities on the order of ~ 300 nA/cm2 for a fully saturated linkage (32, 139 33). A clear trend emerges from these values: a linkage which allows for conjugation of the probe into the π-stack leads to larger redox signals and more effective electrochemistry. DNA-mediated electrochemistry is observed for the acetylene-linked probes at both positive and negative potentials. DNA is expected to be repelled from the surface at negative potentials and attracted to the surface at positive potentials (53, 56). Assays based on DNA-mediated electrochemistry rely on DNA maintaining an upright orientation; CT can then take place through the base pair stack. However, at positive potentials direct contact of the probe with the surface is possible if sufficient space surrounding each duplex is available. It has been shown that in a close-packed monolayer, the DNA is not able to make direct contact with the surface (56). Therefore, assays based on DNA CT may be possible even for probes which are reduced at positive potentials. 5.4.2 Electrochemical Reduction of Alkane-Linked Probes Do Not Depend on DNA Mediated Electrochemistry Electrochemical experiments with alkane-tethered TEMPO and AQ labels support the notion of direct surface-to-probe charge transfer for these assemblies. Changing from an acetylene linkage to an alkane linkage minimizes electronic coupling, and the signals observed for such linkages are associated with pathways of chargetransfer that are not DNA-mediated. The lack of mismatch discrimination observed for either of these probes strongly supports the notion of direct surface-to-probe CT for these cases. Since there is no effect upon inclusion of a mismatch in the DNA sequence, this electrochemistry cannot be mediated by the base pair stack. 140 The current densities obtained for the alkane-linked TEMPO and AQ labels are much smaller than those of the acetylene-linked probes. The peak current densities obtained from SWV are 400 nA/cm2 and 200 nA/cm2 for the TEMPO and AQ, respectively. Anthraquinone is a poor intercalator which should not interact strongly with the base pair stack, and TEMPO does not intercalate. Therefore the smaller signals for these probes are reasonable since the CT pathway to each of these probes involves poor electronic coupling. Furthermore, it is not unexpected that the TEMPO label exhibits larger current densities compared to the AQ label. The AQ is reduced at negative potentials, where DNA is repelled from the surface, making reduction of the probe by direct CT from the surface less likely (53, 56). On the other hand, the TEMPO is oxidized at positive potentials where the DNA can electrostatically contact the surface, making direct surface-to-probe charge-transfer more likely (53, 56). 5.4.3 Electronic Coupling into the Base Pair Stack Governs the Size of the Electrochemical Signal Electrochemical signals that are far smaller than expected based on surface coverage are observed for the acetylene-linked probes versus that expected when compared to the electrochemical signal of intercalated daunomycin (24). Previous reports for daunomycin have revealed surface coverages of 35–45 pmol/cm2 based upon charge integration (24, 25). These values are similar to the theoretical maximum coverage of the surface by a DNA monolayer. However, far lower apparent surface coverages are obtained for the two acetylene-linked probes presented here; cathodic waves give peak currents (Table 5.1) which roughly translate to surface coverages of 6 and 8 pmol/cm2 for the AQ- and TEMPO-modified DNA, respectively. 141 How can the dramatic difference in signal intensities be reconciled? Covalently attached daunomycin has a particularly intimate association with the base pair stack and raises the melting temperature of the duplex by 20 ºC upon crosslinking (35, 36). On the other hand, the bulky AQ moiety only raises the melting temperature of the DNA duplex by 4 ºC compared to unmodified DNA. Therefore, it is reasonable to assume that the AQ is not as well coupled into the DNA base stack. This lack of electronic coupling would account for the smaller electrochemical signals observed relative to DM. In order to investigate if the apparent discrepancy in surface coverage is due to the AQ and TEMPO moieties interfering with surface formation, the monolayers were examined by scanning probe microscopy. Previous work has shown that even subtle structural variations in the DNA monolayer can be easily detected with a high-resolution AFM tip (54). However, we observed no obvious differences between the morphology of monolayers formed with acetylene-TEMPO-modified DNA and unmodified, well matched DNA (Figure 5.5). 5.4.4 DNA Charge Transfer with Acetylene and Alkane-Linked Probes The two linkages may provide different electron-transfer pathways from the electrode to the marker, as illustrated schematically in Figure 5.6. When the probes are effectively coupled into the base pair stack through an acetylene linkage, the yield of DNA-mediated electrochemistry is high and charge transfer through the DNA base pair stack takes place. Large electrochemical signals result, and perturbations within the base pair stack can be easily detected. When the probes are insulated from the base pair stack through a σ-bonded alkane linkage, the yield of DNA-mediated electrochemistry is low 142 A. Figure 5.6: Schematic representation of two possible modes of charge transfer from an electroactive label incorporated into DNA. Direct interaction of the diaminoethanelinked TEMPO probe with the surface is shown on the left, and DNA-mediated electrochemistry to the well-coupled probe is shown on the right. 143 and alternative charge-transfer pathways are dominant. Small electrochemical signals are observed and perturbations within the base pair stack are not easily detected. These results can now be used to reexamine some of the data obtained from experiments involving long-range oxidative damage to DNA. Adequate sensitivity to mismatches or lesions can only be achieved with probes that are well coupled into the DNA base pair stack. Consequently, solution studies with covalently attached anthraquinone provide a particularly interesting example of the necessity of coupling into the DNA base pair stack. For example, in direct agreement with the electrochemical data presented here, anthraquinone linked to the 5' terminus of DNA by an alkane chain has exhibited poor mismatch sensitivity in some contexts (46). Additionally, a direct comparison of the ability of various photo-oxidants to promote DNA-mediated CT showed that even changing the length of the linker connecting the anthraquinone had a significant effect on the yields of oxidative damage. In the same study, significant differences were found between poorly coupled probes that bind by an end-capping mode and well-coupled probes that bind by an intercalative mode. Therefore, it is important to note that DNA CT reveals the characteristics not only of the DNA sequence employed but also of the redox probe. Consideration of the inherent sensitivity of DNA CT to the linkage of the redox probe to the DNA base pair stack can also be extended to solution studies of protein/DNA interactions. A uridine modified with the TEMPO moiety was utilized as an electron trap for DNA-bound repair enzymes (40). Notably, the linkage connecting the TEMPO to the uridine was shown to be crucial for determining its effectiveness as an electron trap in EPR experiments. When the TEMPO moiety is connected to a uridine 144 by an unsaturated acetylene linkage, large EPR signals result, consistent with the larger electrochemical signals observed for this probe. On the other hand, when the TEMPO probe is connected by a saturated linkage, smaller EPR and electrochemical signals result. 5.5 Implications Subtle structural changes can have a significant effect on the sensitivity of redox active reporters in DNA-mediated electrochemistry. Clearly, the utility of labels that are covalently attached to DNA depends upon their interaction with the base pair stack of DNA. Assays based on DNA-mediated CT necessitate well-coupled probes that are highly sensitive to even subtle perturbations in the base pair stack. This need is clearly illustrated by the electrochemical characteristics of AQ and TEMPO when coupled to DNA via a saturated linker versus being conjugated into the base stack. The electrochemistry of the acetylene-linked 3,4-dehydro-TEMPO and AQ is shown to be DNA-mediated. In contrast, charge transfer to the alkane TEMPO on HOPG and the alkane AQ on Au must proceed through an alternate direct surface-to-marker pathway. The data presented makes it clear that linkers conjugated into the base pair stack afford larger electrochemical signals, revealing a direct connection between the electronic structures of DNA-bound probes and their electrochemical properties. DNA-modified electrodes are a powerful tool for directly comparing the electrochemistry of various redox active probes that are bound to DNA. DNA-mediated electrochemistry clearly allows the base pair stack to be probed directly. However, 145 surface-to-probe charge transfer, while useful in other contexts, does not allow for the DNA base pair stack to be interrogated. In fact, the impressive sensitivity of DNA CT chemistry allows for the transduction of small structural changes into large electrochemical differences. As observed in the current study, the linkage directly governs the electrochemical characteristics of the probe. For electrochemical assays to be successful, it is crucial to tailor the probe to the desired application. 146 REFERENCES (1) Boon, E. M., Barton, J. K. (2002) Curr. Opin. Struct. Biol. 12, 320–329. (2) Delaney, S., Barton, J. K. (2003) J. Org. Chem. 68, 6475–6483. (3) O'Neill, M. A., Barton, J. K. (2005) Charge Transfer in DNA: From Mechanism to Application (H.-A. Wagenknecht), pp. 27–75, Wiley-VCH, Hoboken. (4) Giese, B. (2000) Acc. Chem. Res. 33, 631–636. (5) Schuster, G. B. (2000) Acc. Chem. Res. 33, 253–260. (6) Bhattacharya, P. K., Barton, J. K. (2001) J. Am. Chem. Soc. 123, 8649–8656. (7) Hall, D. B., Barton, J. K. (1997) J. Am. Chem. Soc. 119, 5045–5046. (8) Dandliker, P. J., Holmlin, R. E., Barton, J. K. (1997) Science 275, 1465– 1468. (9) Rajski, S. R., Kumar, S., Roberts, R. J., Barton, J. K. (1999) J. Am. Chem. Soc. 121, 5615–5616. (10) Rajski, S. R., Barton, J. K. (2001) Biochemistry 40, 5556–5564. (11) Boon, E. M., Pope, M. A., Williams, S. D., David, S. S., Barton, J. K. (2002) Biochemistry 41, 8464–8470. (12) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucl. Acids Res. 27, 4830–4837. (13) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. (14) Li, X., Song, H., Nakatani, K., Kraatz, H.-B. (2007) Anal. Chem. 79, 2552– 2555. 147 (15) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (16) Li, X., Lee, J. S., Kraatz, H.-B. (2006) Anal. Chem. 78, 6096–6101. (17) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nature Biotechnol. 21, 1192–1199. (18) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nature Biotechnol. 20, 282– 286. (19) Derosa, M. C., Sancar, A., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 10788–10792. (20) Boal, A. K., Yavin, E., Lukianova, O. A., O'Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (21) Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. (22) Li, C. Z., Long, Y. T., Lee, J. S., Kraatz, H. B. (2004) Chem. Commun. 7, 574–575. (23) Masarik, M., Cahova, K., Kizek, R., Palecek, E., Fojta, M. (2007) Anal. Bioanal. Chem. 388, 259–270. (24) Kelley, S. O., Jackson, N. M., Hill, M. G., Barton, J. K. (1999) Angew. Chem. Int. Ed. 38, 941–945. (25) Drummond, T. G., Hill, M. G., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 15010–15011. (26) Liu, T., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 10160–10161. 148 (27) Inouye, M., Ikeda, R., Takase, M., Tsuri, T., Chiba, J. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 11606–11610. (28) Boon, E. M., Jackson, N. M., Wightman, M. D., Kelley, S. O., Barton, J. K., Hill, M. G. (2003) J. Phys. Chem. B 107, 11805–11812. (29) Di Giusto, D. A., Wlasoff, W. A., Giesebrecht, S., Gooding, J. J., King, G. C. (2004) J. Am. Chem. Soc. 126, 4120–4121. (30) Di Giusto, D. A., Wlasoff, W. A., Giesebrecht, S., Gooding, J. J., King, G. C. (2004) Angew. Chem. Int. Ed. 43, 2809–2812. (31) Zuo, X., Song, S., Zhang, J., Pan, D., Wang, L., Fan, C. (2007) J. Am. Chem. Soc. 129, 1042–1043. (32) Okamoto, A., Kamei, T., Tanaka, K., Saito, I. (2004) J. Am. Chem. Soc. 126, 14732–14733. (33) Okamoto, A., Kamei, T., and Saito, I. (2006) J. Am. Chem. Soc. 128, 658– 662. (34) Tanabe, K., Iida, H., Haruna, K., Kamei, T., Okamoto, A., Nishimoto, S. (2006) J. Am. Chem. Soc. 128, 692–693. (35) Wang, A. H.-J., Gao, Y.-G., Liaw, Y.-C., Li, Y.-K. (1991) Biochemistry 30, 3812–3815. (36) Leng, F., Savkur, R., Fokt, I, Przewloka, T, Priebe, W., Chaires, J. B. (1996) J. Am. Chem. Soc. 118, 4731–4738. (37) Anne, A., Demaille, C. (2006) J. Am. Chem. Soc. 128, 542–557. (38) Lai, R. Y., Lagally, E. T., Lee, S.-H., Soh, H. T., Plaxco, K. W., Heeger, A. J. (2006) Proc. Natl. Acad. Sci. U.S.A. 103, 4017–4021. 149 (39) Xiao, Y., Lubin, A. A., Baker, B. R., Plaxco, K. W., Heeger, A. J. (2006) Proc. Natl. Acad. Sci. U.S.A. 103, 16677–16680. (40) Yavin, E., Stemp, E. D. A., O'Shea, V. L., David, S. S., Barton, J. K. (2006) Proc. Natl. Acad. Sci. U.S.A. 103, 3610–3614. (41) Baik, W., Luan, W., Lee, H. J., Yoon, C. H., Koo, S., Kim, B. H. (2005) Can. J. Chem. 83, 213–219. (42) Bondarev, G. N., Burzina, T. S., Krasotskaya, G. I., Kachurin, A. M., Kleiner, A. R. (1985) Izvestiya Akademii Nauk SSSR: Seriya Khimicheskaya. 1850–1855. (43) Gannett, P. M., Darian, E., Powell, J. H., Johnson, E. M. (2001) Synth. Commun. 31, 2137–2141. (44) Kelley, S. O., Barton, J. K., Jackson, N. M., Hill, M. G. (1997) Bioconjugate Chem. 8, 31–37. (45) Gasper, S. M., Schuster, G. B. (1997) J. Am. Chem. Soc. 119, 12762–12771. (46) Williams, T. T., Dohno, C., Stemp, E. D. A., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 8148–8158. (47) Shao, F., Augustyn, K., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 17445– 17452. (48) Rist, M., Amann, N., Wagenknecht, H.-A. (2003) Eur. J. Org. Chem. 13, 2498–2504. (49) Abou-Elkhair, R. A. I., Netzel, T. L. (2005) Nucleosides Nucleotides Nucleic Acids 24, 85–110. 150 (50) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. (51) Geneste, F., Moinet, C. (2005) New J. Chem. 29, 269–271. (52) Geneste, F., Moinet, C., Ababou-Girard, S., Solal, F. (2005) New J. Chem. 29, 1520–1526. (53) Kelley, S. O., Barton, J. K., Jackson, N. M., McPherson, L. D., Potter, A. B., Spain, E. M., Allen, M. J., Hill, M. G. (1998) Langmuir 14, 6781–6784. (54) Boon, E. M., Sam, M., Barton, J. K., Hill, M. G., Spain, E. M. (2001) Langmuir 17, 5727–5730. (55) Oliveira Brett, A. M., Chiorcea, A.-M. (2003) Langmuir 19, 3830–3839. (56) Rant, U., Arinaga, K., Fujita, S., Yokoyama, N., Abstreiter, G., Tornow, M. (2004) Nano Lett. 4, 2441–2445. 151  Chapter 6: Direct Electrochemistry of Endonuclease III in the Presence and Absence of DNA Adapted from: Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. Endonuclease III was expressed and purified by Amie Boal. 152 ABSTRACT The electrochemistry of the base excision repair enzyme Endonuclease III (Endo III) in the presence and absence of DNA has been examined on highly oriented pyrolytic graphite (HOPG). At the surface modified with pyrenated DNA, a reversible signal is observed at 20 mV versus NHE for the [4Fe-4S]3+/2+ couple of Endo III, similar to Au. Without DNA modification, oxidative and reductive signals for the [4Fe-4S] cluster of Endo III are found on bare HOPG, allowing a direct comparison between DNA-bound and free redox potentials. These data indicate a shift of approximately -200 mV in the 3+/2+ couple upon binding of Endo III to DNA. This potential shift reflects a difference in affinity for DNA of more than 3 orders of magnitude between the oxidized 3+ and reduced 2+ protein and provides quantitative support for our model utilizing DNAmediated charge transport to redistribute base excision repair enzymes in the vicinity of damaged DNA. 153 6.1 Introduction In vivo, DNA is constantly being assaulted and damaged (1, 2). To protect the integrity of the genome, an impressive detection and repair network has evolved. However, macromolecular crowding, low repair enzyme copy number, and small structural differences in DNA lesions are dramatic challenges faced by the cell in damage detection. Consequently, the mechanisms by which DNA binding proteins can locate their substrates in rapid fashion are not well understood. Base excision repair (BER) is one component of the DNA detection and repair machinery (3). In BER, glycosylase enzymes must scan a large and topologically complex genome for chemically modified bases and catalyze their excision. While much is known about the catalysis and substrate discrimination steps, very little is known about the daunting initial search of the genome these enzymes must undertake. Since glycosylase enzymes move processively along DNA, processive searches along DNA may represent one component of detection (4–7), but the in vivo relevance of this mechanism has been brought into question (8–10). Therefore, we have proposed DNAmediated charge transport as the first step in damage detection since it provides a means to redistribute base excision repair proteins in the vicinity of damage rapidly and efficiently (11–14). Endonuclease III and MutY are glycosylase enzymes which repair damaged pyrimidines and have been extensively characterized from a structural and biophysical perspective (3, 15–32). Crystal structures of these enzymes indicate that they both contain an iron-sulfur cluster with the C-X6-C-X2-C-X5-C motif in close proximity to the DNA 154 backbone (15–18). EPR experiments demonstrated that the cluster is in its [4Fe-4S]2+ form when isolated from the cell, but the protein cannot be reduced within a physiologically relevant window nor oxidized by ferricyanide without degradation (20, 21). Although these experiments indicated a structural role for the cluster, they were conducted in the absence of DNA. Notably, the [4Fe-4S] cluster was later found not necessary for protein folding or stability in MutY; this argued against a simple structural role, especially because the cluster was crucial for DNA binding and in vivo glycosylase activity (30–32). Subsequently, we demonstrated for both Endo III and MutY that the cluster is activated toward oxidation upon enzyme binding to DNA, and this DNAdependent redox activity promotes charge transport through DNA (11–14). DNA-modified electrodes have proven particularly useful for probing the redox properties of BER enzymes in their DNA-bound, biologically relevant forms (11, 12, 33). Electrochemistry of MutY and Endo III on DNA-modified gold electrodes shows a redox potential of ~ 60 mV and ~ 90 mV versus NHE respectively for the [4Fe-4S]3+/2+ couple. Importantly, these signals were not observed at surfaces backfilled with mercaptohexanol but lacking DNA (11). Furthermore, weak signals were found on surfaces modified with DNA duplexes containing an abasic site, demonstrating that the observed redox processes for the proteins are DNA-mediated (11). Overall, DNA binding appears to shift the potential of Endo III and MutY, so that the proteins bound to DNA are more similar to high-potential iron proteins than ferredoxins (34–36). We have recently explored the electrochemical properties of highly oriented pyrolytic graphite modified with pyrenated DNA as an attractive alternative to gold (37). The DNA monolayers formed on HOPG have been extensively characterized via 155 electrochemistry, AFM, and radioactive labeling and are quite similar to thiolated DNA films on gold (38–41). However, the accessible potential window is significantly larger on graphite, and moreover, graphite electrodes are particularly well suited for protein electrochemistry (42–49). Since previous work had shown that, without DNA binding, the [4Fe-4S]2+ cluster is not readily oxidized or reduced within a physiological range of potentials (20, 21), we compared the electrochemical properties of Endo III in the presence and absence of DNA (Figure 6.1). Here, we directly demonstrate the shift in potential associated with DNA binding using HOPG electrodes. 6.2 Experimental Section 6.2.1 Materials All reagents for DNA synthesis were purchased from Glen Research. All other reagents used in DNA and electrode preparation/processing were purchased from Sigma in the highest purity available and used as received. SPI-1 grade HOPG was purchased from SPI, Inc., and pyrolytic graphite was purchased from Union Carbide, Inc. Buffers were prepared with Milli-Q water and filtered with a 0.2 µm filter. 6.2.2 General DNA Synthesis and Preparation Oligonucleotides were prepared using standard phosphoramidite chemistry on an ABI 392 DNA Synthesizer. DNA was purified by HPLC on a reverse phase C18 column with acetonitrile and ammonium acetate as the eluents. Unless otherwise noted, all oligonucleotides were cleaved from the resin with concentrated ammonium hydroxide 156 Figure 6.1: Schematic representation of electrochemistry for Endo III on HOPG with and without modification with DNA 157 (NH4OH) prior to purification. The desired products were characterized by UV-visible spectroscopy and MALDI-TOF mass spectrometry. As previously described (37), pyrene-terminated oligonucleotides with a long pyrene-to-DNA linker were prepared by solid-phase synthesis on a CPG resin with an unprotected hydroxyl group at the 5' terminus. The 5'-OH was treated with a 120 mg/mL solution of carbonyldiimidazole in dioxane for 90 minutes followed by an 80 mg/mL solution of 1,6-diaminohexane. The free amine was then treated with 1-pyrenebutyric acid N-hydroxysuccinimidyl ester, leaving a pyrene moiety at the 5' terminus. Pyrene-terminated oligonucleotides with a short pyrene-to-DNA linker were also prepared as previously described (37). In brief, 500 mg of N,N-diisopropylethylamine was added to 5 mL of a 20 mg/mL solution of 1-pyrenebutanol in acetonitrile under an inert argon atmosphere. While the reaction mixture was gently stirred, 90 mg of N,Ndiisopropylphosphoramidochloridite was added dropwise. The reaction progress was monitored by TLC before the reaction mixture was moved to an extraction funnel and washed twice with 5% (by volume) NaHCO3 and 3 M NaCl. The organic layer was collected, dried with Na2SO4, and rotovapped to a yellow oil. The phosphoramidite was dissolved in 1 mL of acetonitrile and used immediately. On the synthesizer, the free hydroxyl at the 5' terminus of the oligonucleotide was condensed with a 2-cyanoethyl N,N-diisopropylphosphoramidite of 1-pyrenebutanol. Pyrene-modified single-strand DNA was hybridized with its complement by heating equimolar amounts of each strand (typically 100 µM) in pH 7.1 buffer containing 5 mM Pi and 50 mM NaCl to 90 ºC, followed by cooling to ambient temperature. Since 158 single-stranded DNA can adsorb to graphite, great care was taken during quantification of DNA to ensure that equimolar amounts of each strand were annealed to form duplexes. 6.2.3 Expression and Purification of Endonuclease III Eight separate batches of protein were utilized for the electrochemical experiments described. Endo III was expressed and purified according to previously reported procedures (50, 51). MutY was 1) generously donated by the David Lab (UCDavis) and 2) purchased from Trevigen, Inc. and dialyzed against Endo III storage buffer. 6.2.4 Preparation of Bare and DNA-Modified HOPG Electrodes DNA-modified surfaces were prepared on custom-built highly oriented pyrolytic graphite electrodes as previously described (37, 52). Briefly, the HOPG surface was cleaned with 3-M scotch tape before each experiment and mounted in the electrode housing. Duplex DNA containing a pyrene moiety was deposited onto the HOPG surface in pH 7.1 buffer containing 5 mM Pi 50 mM NaCl in the absence of Mg2+ to ensure a loosely packed film. DNA films were then allowed to form in a humidified chamber over a period of 24–48 hours. To ensure experimental consistency, bare electrodes were prepared in an identical fashion with the DNA omitted in the steps above. When necessary, the bare or DNA-modified graphite surfaces were backfilled with short chain alkanes (53). The surface was rinsed thoroughly with Endo III storage buffer (pH= 7.5, 20 mM NaPi, 100 mM NaCl, 1 mM EDTA) before being backfilled for ~ 1–2 hours with a 5% by volume octane solution in a 1:1 Endo III storage buffer/glycerol solution. Large percentages of glycerol or small percentages of ethanol were used to improve the poor solubility of alkanes in the Endo III storage buffer. The electrodes were then thoroughly cleaned with multiple rinses of storage buffer. 159 6.2.5 Preparation of Didodecyldimethylammonium-Modified Pyrolytic Graphite Electrodes Didodecyldimethylammonium bromide (DDAB) modified pyrolytic graphite (PG) surfaces were prepared as previously described (47–49). Briefly, electrodes were made from the basal plane of pyrolytic graphite and pretreated by abrasion with sandpaper, brief sonication in deionized H2O, and drying with a heat gun. Didodecyldimethylammonium films were applied by incubating the surface in a saturated DDAB solution in a humidified environment overnight. 6.2.6 Electrochemical Measurements on HOPG Prior to incubation with protein, bare or DNA-modified HOPG electrodes were cooled to 4 ºC. Although this was not done in all experiments, cooling of the electrodes was found to dramatically improve protein stability. Bare or DNA modified electrodes were incubated with protein for ~ 1–4 hours prior to electrochemical measurements. Blank electrodes were treated in an identical fashion with the protein omitted. Cyclic and square wave voltammetry experiments were subsequently recorded on either a CH Instruments electrochemical analyzer (Austin, TX) or a Bioanalytical Systems CV-50W potentiostat using the inverted drop cell configuration. All measurements reported for the working electrode were taken versus a platinum (Pt) auxiliary and a silver/silver chloride (Ag/AgCl) reference. The Ag/AgCl reference was frequently standardized versus SCE. Unless otherwise noted, experiments were performed at ambient temperature in Endo III storage buffer. 160 6.2.7 Electrochemical Measurements on Pyrolytic Graphite The electrodes were incubated with a 50 µM Endo III solution for ~ 6 hours in the presence or absence of an excess of DNA at 4 º C. The protein-modified electrodes were rinsed thoroughly prior to electrochemistry experiments. Cyclic voltammetry experiments were subsequently recorded on either a CH Instruments electrochemical analyzer (Austin, TX) or a Bioanalytical Systems CV-50W potentiostat. Experiments were performed at ambient temperature in pH 8.0 buffer containing 20 mM KPi and 100 mM KCl under an argon atmosphere. A custom-built three-compartment electrochemical cell was used for all experiments with a Pt auxiliary electrode, a PG working electrode, and a saturated calomel (SCE) reference electrode. 6.3 Results 6.3.1 Electrochemistry of Free and DNA-Bound Endo III on HOPG Figure 6.2 shows cyclic voltammetry (CV) and square wave voltammetry (SWV) of Endo III on HOPG with and without DNA modification. For the DNA-modified electrode, a quasi-reversible redox couple is observed with a midpoint potential of 20 ± 10 mV versus NHE. Backfilling the DNA electrode with octane has no effect on this signal, while backfilling HOPG without DNA leads to the loss of any protein signal (Figure 6.3). To establish that this signal is DNA-mediated, we also examined an electrode modified with DNA featuring an abasic site prepared under identical conditions; 161 Figure 6.2: CV (left, 50 mV/s scan rate) and SWV (right, 15 Hz) of 50 µM Endo III in 20 mM Na phosphate, 100 mM NaCl, 1 mM EDTA, 20% glycerol, pH 7.5 buffer. The top two panels show electrochemistry of Endo III at a HOPG electrode modified with the sequence pyrene-(CH2)4-Pi-5'-AGT ACA GTC ATC GCG-3' plus complement. Cyclic voltammetry of a HOPG electrode modified with DNA featuring an abasic site is in red (top left), where the abasic position corresponds to the complement of the italicized base. The bottom two panels show electrochemistry of Endo III on bare HOPG. 162 Figure 6.3: SWV (15 Hz) of 50 µM Endo III in 20 mM Na phosphate, 100 mM NaCl, 1 mM EDTA, 20% glycerol, pH 7.5 buffer showing the 2+/1+ couple. An electrode backfilled with octane prior to incubation with protein is shown in blue. The two electrodes have been normalized by their geometric areas for clarity. 163 DNA-mediated charge transport has been shown to be inhibited by the abasic site owing to the disruption in base stacking (11, 12). As seen in Figure 6.2, a complete loss of signal for Endo III is observed at the electrode modified with DNA containing an abasic site. Thus the DNA does not serve to locally concentrate the protein on the graphite surface; the duplex with an abasic site would serve a similar function. Instead it is the DNA-bound protein that is probed electrochemically on HOPG in a DNA-mediated reaction, as long as the DNA duplex is well stacked. Note that at the DNA-modified surface, we observe only one redox signal, with no other peaks evident in the range of 600 to -400 mV versus NHE. This signal is not observed for a DNA-modified electrodes in Endo III dialysis buffer (Figure 6.4). The only couple we observe features a cathodic peak at -30 ± 30 mV versus NHE whose shape and magnitude indicate slow diffusive kinetics, as found for MutY (11, 12). Indeed, in all respects, this couple resembles that found for Endo III at a DNA-modified Au surface (11, 12) and is assigned to the [4Fe-4S]3+/2+ couple. EPR experiments in solution on DNA- bound MutY reveal g values characteristic of a 3+ cluster and support this assignment (13, 14). Significantly, on HOPG versus Au, we may explore the electrochemistry of Endo III at a larger range of applied biases (37), and thus we may directly compare the electrochemistry of Endo III in the presence and absence of DNA. Oxidative scans of Endo III on bare HOPG reveal an irreversible anodic peak at 250 ± 30 mV versus NHE and no couple either at 20 mV as with DNA (Figure 6.2) or in Endo III dialysis buffer (Figure 6.4). However, we do observe a quasi-reversible wave at higher protein concentrations (Figure 6.5). Successive positive scans lead to new broad, irregular signals 164 A. B. Figure 6.4: CV (left, 50 mV/s scan rate) and SWV (right, 15 Hz) of 50 µM Endo III in 20 mM Na phosphate, 100 mM NaCl, 1 mM EDTA, 20% glycerol, pH 7.5 buffer. The top panel shows electrochemistry of Endo III at a HOPG electrode modified with the sequence pyrene-(CH2)4-Pi-5'-AGT ACA GTC ATC GCG-3' plus complement (A). The bottom panel shows electrochemistry of Endo III at bare HOPG electrode (B). Voltammetry of the protein dialysis buffer in the absence of protein is shown in red for comparison. 165 Figure 6.5: CV (50 mV/s scan rate) of 200 µM Endo III in 20 mM Na phosphate, 100 mM NaCl, 1 mM EDTA, 50% glycerol, pH 7.5 buffer on bare HOPG showing the 3+/2+ couple. Note that both cathodic and anodic peaks are observed here. 166 at approximately -80 and -710 mV versus NHE; additionally, the yellow color of the protein solution is lost. These results are fully consistent with oxidative decomposition of the cluster in Endo III without DNA and similar peaks are found with FeCl3 controls (Figure 6.6). Indeed, these redox signals are commonly associated with ferredoxin [3Fe4S] clusters (34, 35, 54). It is noteworthy that on bare HOPG, we observe also the 2+/1+ couple of the [4Fe-4S] cluster during reductive scans with a cathodic peak at approximately -300 ± 80 mV versus NHE (Figure 6.3 and Figure 6.7). The peak is near the edge of our potential window, and this redox signal also contains a small oxidative wave at slow scan rates (Figure 6.7). In our experiments, the potential difference between 3+/2+ and 2+/1+ couples is somewhat smaller than expected (20, 21). 6.3.2 Electrochemistry of Free and DNA-Bound Endo III on Pyrolytic Graphite Due to the frequently observed instability of Endo III on bare HOPG, the electrochemistry of Endo III was also investigated at DDAB modified pyrolytic graphite electrodes. Biomembrane-like films are commonly utilized in protein electrochemistry experiments (47–49). They are formed via the incubation of a graphite electrode with surfactants bearing a polar head group and a long, non-polar tail consisting of hydrocarbon chain. Such films possess the dual advantage of preventing adsorptive denaturation of proteins at the electrode surface and orienting the proteins for efficient electron transfer (47–49). The electrochemistry of DDAB electrodes modified with Endo III in the presence and absence of DNA is shown in Figure 6.8. In the absence of DNA, two weak and irreversible signals are observed with cathodic waves at -180 mV and -370 mV versus 167 Figure 6.6: Typical CV (100 mV/s scan rate) of FeCl3 on bare HOPG in 20 mM Na phosphate, 100 mM NaCl, 1 mM EDTA, 50% glycerol, pH 7.5 buffer. The voltammetry of free iron was recorded in the absence (blue) and presence (red) of thiol. The potential and reversibility of the free iron peak were both found to vary depending on the identity of the added alkanethiol (mercaptohexanol in this case). 168 Figure 6.7: CV (20 mV/s scan rate) of 50 µM Endo III in 20 mM Na phosphate, 100 mM NaCl, 1 mM EDTA, 20% glycerol, pH 7.5 buffer on bare HOPG showing the 2+/1+ couple. A plot of peak current as a function of scan rate is inset. 169 A. B. Figure 6.8: CV (20 mV/s scan rate) of Endo III in 20 mM K phosphate, 100 mM KCl, pH 8 buffer on a DDAB modified pyrolytic graphite. An electrode incubated with DNA bound Endo III is shown on the top panel (A), and an electrode modified with free Endo III is shown on the bottom panel. (B). 170 NHE, which can be assigned to the 2+/3+ and 2+/1+ couples, respectively. In the presence of DNA, Endo III is dramatically stabilized, and a single, large, and reversible redox couple is observed with a cathodic wave at -290 mV versus NHE, which can be assigned to its 2+/3+ couple. It is noteworthy that the potentials for free and DNA-bound Endo III in DDAB are different from the ones found on gold or graphite, which may indicate that Endo III does not maintain its native conformation in the positively charged, DDAB surfactant films. However, these experiments do provide support for the dramatically different electrochemical properties of Endo III observed in the presence and absence of DNA. 6.4 Implications and Discussion Figure 6.9 summarizes the potentials we have observed for Endo III on HOPG over several trials. Although SWV gives a shift of 280 mV between the cathodic DNAbound potential and the anodic potential on bare HOPG, the shift in midpoint potentials should be slightly smaller. Given the lack of reversibility and the supporting results obtained for Endo III in DDAB films on pyrolytic graphite, we can provide a lower estimate of ~ 200 mV for the change in redox potential by the 3+/2+ couple upon DNA binding; the corresponding change in the 2+/1+ couple cannot be easily determined. The observed shift is understandable based upon the sensitivity of [4Fe-4S] cluster potentials to their environment (34–36). DNA binding clearly stabilizes the oxidized 3+ form of the cluster, whereas without DNA, it is [4Fe-4S]2+ that is more stable. Crystal structures of Endo III with and without DNA reveal that the cluster is located 171 Figure 6.9: Illustration of the potentials versus NHE of the couples of Endo III in the presence and absence of DNA. These values are based upon SWV on HOPG and are averages of at least four trials each. 172 near amino acid residues that contact DNA (15–17). DNA binding takes the cluster to a more hydrophobic environment compared to the exposed and polar environment in the absence of DNA. Importantly, the resultant shift in potential is not associated with significant conformational changes in the protein; the structures of the bound and free proteins are remarkably similar. Instead, then, the ~ 200 mV shift in potential must correspond to a decrease in DNA binding affinity of more than 3 orders of magnitude between the 2+ and 3+ forms of the cluster. While previous evidence qualitatively indicated a lessened DNA binding affinity for the reduced protein (12), these data provide a more quantitative estimate. In the context of our model of DNA-mediated signaling for damage detection, it is this difference in DNA binding affinity for the reduced versus oxidized state that leads to the dissociation of protein from the DNA upon reduction. This allows for the effective redistribution of BER proteins onto sites near damage, perhaps through long range electron transfer to other redox-active, DNA-binding proteins (13, 14). We have, therefore, now identified the electrochemistry of Endo III both with and without DNA on HOPG electrodes. DNA binding is seen to promote a shift in redox potential, activating the protein toward oxidation; subsequent reduction of the cluster to the 2+ form leads to dissociation from the duplex. These results provide strong support for the detection strategy we have proposed for BER enzymes. Furthermore, these data underscore the importance of the outer sphere environment in regulating potentials of [4Fe-4S] proteins (34–36, 42–45), as well as the utility of DNA-modified electrodes in probing the redox characteristics of proteins that bind to DNA. 173 REFERENCES (1) Friedberg, E. C., Walker, G. C., Siede, W. (1995) DNA Repair and Mutagenesis, ASM Press, Washington, D. C. (2) Krokan, H. E., Standal, R., Slupphaug, G. (1997) Biochem. J. 325, 1–16. (3) David, S. S., Williams, S. D. (1998) Chem. Rev. 98, 1221–1262. (4) Blainey, P. C., van Oijen, A. M., Banerjee, A., Verdine, G. L., Xie, X. S. (2006) Proc. Natl. Acad. Sci. U. S. A. 103, 5752–5757. (5) Roberts, R. J., Cheng, X. (1998) Annu. Rev. Biochem. 67, 181–198. (6) Verdine, G. L., Bruner, S. D. (1997) Chem. Biol. 4, 329–334. (7) Francis, A. W., David, S. S. (2003) Biochemistry 42, 801–810. (8) Halford, S. E., Szczelkun, M. D. (2002) Eur. Biophys. J. 31, 257–267. (9) Gowers, D. M., Halford, S. E. (2003) EMBO J. 22, 1410–1418. (10) Gerland, U., Moroz, J. D., Hwa, T. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 12015–12020. (11) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 12543–12547. (12) Boal, A. K., Yavin, E., Lukianova, O. A., O'Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (13) Yavin, E., Boal, A. K., Stemp, E. D. A., Boon, E. M., Livingston, A. L., O'Shea, V. L., David, S. S., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 3546–3551. 174 (14) Yavin, E., Stemp, E. D. A., O'Shea, V. L., David, S. S., Barton, J. K. (2006) Proc. Natl. Acad. Sci. U. S. A. 103, 3610–3614. (15) Kuo, C. F., McRee, D. E., Fisher, C. L., O’Handley, S. F., Cunningham, R. P., Tainer, J. A. (1992) Science 258, 434–440. (16) Thayer, M. M., Ahern, H., Xing, D., Cunningham, R. P., Tainer, J. A. (1995) EMBO J. 14, 4108–4120. (17) Fromme, J. C., Verdine, G. L. (2003) EMBO J. 22, 3461–3471. (18) Fromme, J. C., Banerjee, A., Huang, S. J., Verdine, G. L. (2004) Nature 427, 652–656. (19) Asahara, H., Wistort, P. M., Bank, J. F., Bakerian, R. H., Cunningham, R. P. (1989) Biochemistry 28, 4444–4449. (20) Cunningham, R. P., Asahara, H., Bank, J. F., Scholes, C. P., Salerno, J. C., Surerus, K., Munck, E., McCracken, J., Peisach, J., Emptage, M. H. (1989) Biochemistry 28, 4450–4455. (21) Fu, W., O’Handley, S., Cunningham, R. P., Johnson, M. K. (1992) J. Biol. Chem. 267, 16135–16137. (22) O'Handley, S., Scholes, C. P., Cunningham, R. P. (1995) Biochemistry 34, 2528–2536. (23) Xing, D., Dorr, R., Cunningham, R. P., Scholes, C. P. (1995) Biochemistry 34, 2537–2544. (24) Porello, S. L., Cannon, M. J., David, S. S. (1998) Biochemistry 37, 6465– 6475. 175 (25) Michaels, M. L., Pham, L., Nghiem, Y., Cruz, C., Miller, J. H. (1990) Nucleic Acids Res. 18, 3841–3845. (26) Tsai–Wu, J. J., Liu, H. F., Lu, A. L. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 8779–8783. (27) Lu, A. L., Tsai-Wu, J. J., Cillo, J. (1995) J. Biol. Chem. 270, 23582–23589. (28) Bulychev, N. V., Varaprasad, C. V., Dorman, G., Miller, J. H., Eisenberg, M., Grollman, A. P., Johnson, F. (1996) Biochemistry 35, 13147–13156. (29) Porello, S. L., Williams, S. D., Kuhn, H., Michaels, M. L., David, S. S. (1996) J. Am. Chem. Soc. 118, 10684–10692. (30) Porello, S. L., Cannon, M. J., David, S. S. (1998) Biochemistry 37, 6465– 6475. (31) Golinelli, M. P., Chmiel, N. H., David, S. S. (1999) Biochemistry 38, 6997– 7007. (32) Messick, T. E., Chmiel, N. H., Golinelli, M. P., Langer, M. R., Joshua-Tor, L., David, S.S. (2002) Biochemistry 41, 3931–3942. (33) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nature Biotechnol. 21, 1192–1199. (34) Cowan, J. A., Lui, S. M. (1998) Adv. Inorg. Chem. 45, 313–350. (35) Beinert, H. (2000) J. Biol. Inorg. Chem. 5, 2–18. (36) Carter, C. W., Kraut, J., Freer, S. T., Alden, R. A., Sieker, L. C., Adman, E., Jensen, L. H. (1972) Proc. Natl. Acad. Sci. U. S. A. 69, 3526–3529. (37) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. 176 (38) Kelley, S. O., Barton, J. K., Jackson, N. M., Hill, M. G. (1997) Bioconjugate Chem. 8, 31–37. (39) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucleic Acids Res. 27, 4830–4837. (40) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (41) Kelley, S. O., Barton, J. K., Jackson, N. M., McPherson, L. D., Potter, A. B., Spain, E. M., Allen, M. J., Hill, M. G. (1998) Langmuir 14, 6781–6784. (42) Guo, L. H., Hill, H. A. O. (1991) Adv. Inorg. Chem. 36, 341–375. (43) Armstrong, F. A. (1992) Adv. Inorg. Chem. 38, 117–163. (44) Armstrong, F. A., Heering, H. A., Hirst, J. (1997) Chem. Soc. Rev. 26, 169– 188. (45) Camba, R., Armstrong, F. A. (2000) Biochemistry 39, 10587–10598. (46) Armstrong, F. A., Wilson, G. S. (2000) Electrochim. Acta 45, 2623–2645. (47) Nassar, A. F., Rusling, J. F., Nakashima, N. (1996) J. Am. Chem. Soc. 118, 3043–3044. (48) Rusling, J. F. (1998) Acc. Chem. Res. 31, 363–369. (49) Udit, A. K., Hill, M. G., Gray, H. B. (2006) Langmuir 22, 10854–10857. (50) Boiteux, S.; O'Connor, T. R.; Laval, J. (1987) EMBO J. 6, 3177–3183. (51) O'Connor, T. R. (1993) Nucleic Acids Res. 21, 5561–5569. (52) Bowler, R., Davies, T. J., Hyde, M. E., Compton, R. G. (2005) Anal. Chem. 77, 1916–1919. 177 (53) Kim, Y., Long, E. C., Barton, J. K., Lieber, C. M. (1992) Langmuir 8, 496– 500. (54) Thomson, A. J., Breton, J., George, S. J., Butt, J. N., Armstrong, F. A., Hatchikian, E. C. (1991) Biochem. Soc. Trans. 19, 594–599. 178  Chapter 7: DNA Binding Shifts the Redox Potential of the Transcription Factor SoxR Adapted from: Gorodetsky, A. A., Dietrich, L. E. P., Lee, P. E., Demple, B., Newman, D. K., Barton, J. K. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 3684–3689. SoxR was expressed and purified by Lars Dietrich. 179 ABSTRACT Electrochemistry measurements on DNA-modified electrodes are used to probe the effects of binding to DNA on the redox potential of SoxR, a transcription factor that contains a [2Fe-2S] cluster and is activated through oxidation. A DNA-bound potential of +200 mV versus NHE (normal hydrogen electrode) is found for SoxR isolated from Escherichia coli and Pseudomonas aeruginosa. This potential value corresponds to a dramatic shift of +490 mV versus values found in the absence of DNA. Using Redmond Red as a covalently bound redox reporter affixed above the SoxR binding site, we also see, associated with SoxR binding, an attenuation in the Redmond Red signal compared with that for Redmond Red attached below the SoxR binding site. This observation is consistent with a SoxR-binding-induced structural distortion in the DNA base stack that inhibits DNA-mediated charge transport to the Redmond Red probe. The dramatic shift in potential for DNA-bound SoxR compared with the free form is thus reconciled based on a high-energy conformational change in the SoxR–DNA complex. The substantial positive shift in potential for DNA-bound SoxR furthermore indicates that, in the reducing intracellular environment, DNA-bound SoxR is primarily in the reduced form; the activation of DNA-bound SoxR would then be limited to strong oxidants, making SoxR an effective sensor for oxidative stress. These results more generally underscore the importance of using DNA electrochemistry to determine DNA-bound potentials for redox-sensitive transcription factors because such binding can dramatically affect this key protein property. 180 7.1 Introduction SoxR belongs to the MerR family of transcriptional regulators. The members of this family are defined by an N-terminal helix–turn–helix DNA binding motif, a coiledcoil dimerization region, and a C-terminal sensory domain (1–3). Although the DNA binding and dimerization regions are conserved among MerR-type regulators, their sensory domains are diverse (2). Typically, MerR type transcription factors occupy suboptimally spaced 19 ± 1-bp promoter elements in the inactivated state, often inducing a slight bend of the promoter DNA. Upon activation, these proteins are thought to undergo a conformational change that unwinds the promoter region, thereby allowing RNA polymerase to initiate transcription (2). SoxR regulates an oxidative stress response to superoxide in the enterics Escherichia coli and Salmonella enterica (4, 5). This unique transcription factor is a 17kDa polypeptide that binds DNA as a dimer and contains a [2Fe-2S] cluster in each monomer (4). Loss of this cluster does not affect protein folding, DNA binding, or promoter affinity (6–8), but oxidation of this cluster by either oxygen or superoxidegenerating agents (e.g., methyl viologen) triggers expression of the transcription factor SoxS (8, 9). Subsequently, SoxS controls the expression of > 100 genes in the SoxRS regulon that collectively act to repair or avoid oxidative damage (10). The role of SoxR appears to vary dramatically across organisms. Whereas SoxR is conserved in both Gram-negative and Gram-positive bacteria, soxS is exclusively found in enterics, indicating that SoxR can be part of different regulatory networks (11, 12). Indeed, Pseudomonas putida and Pseudomonas aeruginosa do not rely on SoxR for 181 an oxidative-stress response (13, 14). Instead, P. aeruginosa SoxR responds to phenazines, endogenous redox-active pigments, and activates transcription of two probable efflux pumps and a putative monooxygenase (15) that might aid in phenazine transport and modification. Considering that SoxR shows functional diversity between pseudomonads and enterics, it is surprising that the transcription factor is biochemically conserved: (i) Expression of P. putida SoxR in E. coli can complement a soxR deletion mutant (14), and (ii) the redox potentials of soluble SoxR from E. coli and P. aeruginosa in vitro are both approximately −290 mV (7, 8, 15). That SoxR requires oxidation for its transcriptional activity seems biologically reasonable but also leads to a conundrum. Under normal physiological conditions, it is assumed that SoxR is kept in its reduced, inactive state by the intracellular NADPH/NADP+ redox potential of approximately −340 mV versus NHE (16, 17). Furthermore, it has been reported that NADPH-dependent SoxR reduction is enzymemediated, allowing for a rapid adjustment to changes in cellular conditions, although direct enzymatic interaction with SoxR has not yet been demonstrated (18, 19). The conundrum does not lie in the mechanism of SoxR reduction but rather in the specificity of its oxidation: At a low redox potential of −290 mV versus NHE (7, 8, 15), many cellular oxidants could react with SoxR, in particular glutathione (20), and therefore SoxR would be primarily in an oxidized form, even without imposing oxidative stress. Since this is not the case, how is SoxR maintained in its reduced and transcriptionally silent form? The mechanism underlying the oxidation/activation of SoxR is also not well understood. For E. coli SoxR, it was first suggested that superoxide directly oxidizes the 182 iron-sulfur cluster, but this has not been established (21, 22). Alternatively, the redox state of SoxR might be coupled to changes in the equilibrium of biologically relevant redox couples, such as NADPH or glutathione (16, 23). Recently, we have shown that the activation of SoxR in P. aeruginosa can occur in an oxygen-independent manner (14). Considering that both E. coli and P. aeruginosa SoxR can transfer electrons to the mediator safranin O, a phenazine derivative (7, 8, 15), it seems reasonable that endogenous phenazines may oxidize SoxR in pseudomonads. Alternatively, given that pseudomonad phenazines can also modulate the intracellular NADH/NAD+ ratio, the possibility that phenazines activate SoxR indirectly must also be considered (24). One interesting possibility that has been suggested but never addressed experimentally is the effect of DNA binding on the redox potential of SoxR. The published redox potentials for SoxR were measured in the absence of DNA (4, 7, 8). This is particularly significant because SoxR activates transcription only in its DNA-bound state, so determining the redox potential of the DNA-bound form of SoxR becomes critical. We have previously explored DNA-modified electrodes as flexible platforms for the study of DNA-mediated charge transport chemistry (25–27). Typically, selfassembled DNA monolayers on gold or graphite are interrogated electrochemically with the efficiency of charge transfer to an electroactive probe yielding information on the integrity of the intervening base pair stack. In fact, duplexes that are covalently modified with redox-active reporters at a fixed position provide particularly well defined systems for study of DNA charge transport at electrode surfaces, allowing for the electrochemical detection of even small perturbations in the intervening base pair stack (28–31). 183 In addition, DNA-modified electrodes have proven useful for probing redox centers within proteins bound to DNA (32–34). We have used DNA monolayers to probe the redox potential of MutY and Endonuclease III, base excision repair glycosylases that contain a [4Fe-4S] cluster. Initial studies of these enzymes had found no clear role for the clusters because, in the absence of DNA, they did not display redox activity within a physiologically relevant range of potentials (35–37). We found, however, that at DNAmodified Au surfaces, these repair enzymes display reversible, DNA-mediated electrochemistry with redox potentials of ≈ 90 mV (33). Moreover, experiments comparing directly the electrochemistry of Endonuclease III on bare and DNA-modified graphite demonstrated that binding to DNA shifts the redox potential of the protein by ≈ 200 mV into a physiologically relevant range, activating the cluster for oxidation (34). DNA binding thus changes the redox properties of the enzymes from being similar to ferredoxins to instead resembling high potential iron proteins. Based on these data, we have proposed a redox role for the [4Fe-4S] clusters in long-range DNA-mediated signaling as a first step in detecting damaged sites that are to be repaired in the genome (32–34, 38). Our ability to alter the redox states of these proteins in a DNA-mediated manner further suggests that DNA may be a medium through which oxidation/reduction reactions occur. This mechanism may also be important to consider in the context of SoxR. Given the sensitivity of DNA-modified electrodes in probing redox centers of proteins bound to DNA and the precedent that DNA binding can alter redox potentials of the bound protein, here we explore the redox properties of the DNA-bound form of SoxR. Model studies have shown repeatedly the sensitivity of redox potentials of iron-sulfur 184 clusters to environmental perturbations, which are expected to be significant for SoxR (39). Here, using self-assembled DNA monolayers on highly oriented pyrolytic graphite (HOPG), we address the effect of DNA binding on the redox potential of both E. coli and P. aeruginosa SoxR. The DNA-bound potential provides convincing evidence for the mechanism the cell uses to maintain SoxR in its reduced form in vivo. 7.2 Experimental Procedures 7.2.1 Materials All phosphoramidites and reagents for DNA synthesis were purchased from Glen Research. 1,6-Diaminohexane was obtained from Acros Organics. All organic solvents and other reagents were purchased from Aldrich in the highest available purity. 7.2.2 Oligonucleotide Synthesis Oligonucleotides were prepared using standard phosphoramidite chemistry on an ABI 394 DNA synthesizer. DNA was purified by HPLC on a reverse-phase C18 column with acetonitrile and ammonium acetate as eluents. The desired products were characterized by HPLC, UV-visible spectroscopy, and MALDI-TOF mass spectrometry. For experiments on HOPG, DNA was modified with pyrene at the 5′ terminus by following the procedure reported in reference 27. In brief, oligonucleotides were prepared by solid phase synthesis using standard reagents with an unprotected hydroxyl group at the 5′ terminus. The 5′-OH was treated with a 120 mg/ml solution of carbonyldiimidazole in dioxane for 2 hours followed by an 80 mg/ml solution of 1,6-diaminohexane for 30 minutes. Subsequently, the free amine was treated with 1-pyrenebutyric acid, N- 185 hydroxysuccinimide ester, resulting in the desired pyrene moiety linked to the 5′ terminus. The oligonucleotides were deprotected with concentrated NH4OH at 60 °C for 8 hours. DNA modified with Redmond Red at the 3′ terminus or 3 bases in from the 5′ terminus was prepared according to the ultra-mild protocols outlined on the Glen Research web site (www.glenres.com). Pac-protected bases and ultra mild reagents were used. The oligonucleotides were deprotected in 0.05 M potassium carbonate in methanol at room temperature for 12–14 hours to prevent degradation of the Redmond Red moiety under harsh conditions. 7.2.3 Expression and Purification of SoxR E. coli SoxR was prepared as described in reference 66. N-terminally histidinetagged SoxR from P. aeruginosa PA14 were expressed from plasmid pET16b in E. coli strain BL21 (DE3). Cells were grown in 1 liter of LB medium with 100 µg/ml ampicillin at 37 °C. At an OD600nm of 0.3, protein expression was induced by the addition of 1 mM IPTG and the cultures were incubated for an additional 10 hours at 16 °C. All subsequent steps were performed at 4°C. Cells were pelleted, resuspended in buffer A [50 mM NaH2PO4 (pH 8.0), 300 mM NaCl, 10% glycerol] containing 10 mM imidazole and PIC (protease inhibitor mixture without EDTA; Roche), and lysed using a French Press. The cell extract was centrifuged at 14,000 × g for 20 minutes. The supernatant was incubated with Talon-beads (Clontech) for 30 minutes and then transferred to a column. The beads were washed with buffer A containing 50 mM imidazole and PIC. Histidine-tagged SoxR was eluted from the column with buffer A containing 250 mM imidazole and PIC. Peak fractions and purity were determined by SDS/PAGE with Coomassie Blue staining. 186 Purified protein was dialyzed against SoxR storage buffer [50 mM Pi (pH 8.0), 500 mM NaCl, 20% glycerol]. To generate expression plasmid pET16b-soxR, soxR (PA14_35170) was PCRamplified from genomic DNA of P. aeruginosa PA14 using primers A (CGC catatg AAG AAT TCC TGC GCA TC) and B (GGC gga tcc CTA GCC GTC GTG CTC G). Primer A contains an NdeI restriction site (small letters) and soxR's start codon (italicized). Primer B contains a BamHI site (small letters) and soxR's stop codon. The PCR fragment was ligated into NdeI/BamHI-digested pET16b. 7.2.4 Formation of DNA Monolayers and Electrochemical Measurements DNA films were self-assembled on SPI-1 grade HOPG electrodes (SPI) with an estimated surface area of 0.08 cm2 defined by an o-ring. Duplex DNA was formed in (pH 8) 50 mM Pi/500 mM NaCl/20% glycerol buffer (SoxR storage buffer) by combining equimolar amounts of the pyrene-modified strand with its complement. Loosely packed DNA monolayers were allowed to form over a period of 24–48 hours. The electrodes were then thoroughly rinsed with SoxR storage buffer before being backfilled for 2–4 hours with 10% by volume octane or decane solutions in SoxR storage buffer. The electrodes were then thoroughly rinsed with SoxR storage buffer again and moved into a nitrogen atmosphere for electrochemistry experiments. Electrochemical data were collected with a Bioanalytical Systems CV-50W potentiostat using the inverted drop cell configuration. All measurements reported for the working electrode were taken versus a platinum (Pt) auxiliary and a silver/silver chloride (Ag/AgCl) reference. The Ag/AgCl reference was frequently standardized versus SCE, and all reported potentials have an experimental uncertainty of < 40 mV. Electrochemical 187 experiments were performed at ambient temperature and under an anaerobic atmosphere in SoxR storage buffer. In a typical experiment, background electrochemical scans were performed before SoxR was added to the storage buffer, resulting in an ≈ 15–35 µM monomer concentration within the cell. Further scans were then performed in the presence of SoxR, typically over a period of 30–45 minutes. 7.3 Results 7.3.1 Experimental Strategy Used to Probe SoxR Electrochemically Figure 7.1 illustrates the experimental strategy used to investigate the electrochemistry of DNA-bound SoxR from E. coli and P. aeruginosa PA14. DNA duplexes are prepared by hybridizing pyrene-modified single-stranded DNA with its complement (with or without covalently attached Redmond Red). The duplexes are then self-assembled on HOPG in the absence of Mg2+ to form a loosely packed DNA monolayer, leaving room for SoxR to bind (32–34). The surface is backfilled with octane or decane to prevent direct charge transfer from the surface to the protein (40, 41). The electrode is subsequently incubated with protein, and electrochemical experiments are performed before and after protein addition. The DNA binding sites for SoxR are 18-bp symmetrical sequences that are conserved across species (15). For P. aeruginosa experiments, we chose the SoxR binding site found upstream of an operon that encodes the efflux pump MexGHI-OpmD in P. aeruginosa PA14. 188 Figure 7.1: Schematic illustration of the self-assembly/backfilling of a DNA monolayer followed by incubation with protein 189 7.3.2 SoxR Binding Is Reported Through the Redmond Red Electrochemical Signal We can observe protein binding in electrochemistry experiments by monitoring DNA-mediated transport between the electrode and the redox active probe Redmond Red that is attached at either end of the DNA duplex (Figure 7.2). The midpoint potential of Redmond Red is −160 mV versus NHE, and the linearity of the plot of peak current as a function of scan rate indicates that Redmond Red behaves as a surface-bound species (42). Although small potential shifts (≈ 20 mV) in the Redmond Red signal are occasionally observed upon addition of SoxR, Redmond Red provides a convenient and reliable internal standard. It is expected that a redox-active probe located at the top of the DNA monolayer will report on perturbations of the base pair stack that intervene between the redox probe and the electrode, whereas the same probe located at the bottom of the monolayer near the electrode surface will not be affected by disruptions in base stacking above the probe. Previously, we have reported attenuation in charge accumulation by chronocoulometry for daunomycin covalently attached near the top of a DNA film due to perturbations in the intervening DNA structure by the base-flipping methylase M.HhaI and TATA binding protein (43). Here, as shown in Figure 7.2, when Redmond Red is incorporated above the SoxR binding site, a 16% decrease in the integrated cathodic charge of Redmond Red is observed upon addition of SoxR. In contrast, when Redmond Red is incorporated at the bottom of the DNA duplex below the SoxR binding site, there is little detectable change in the Redmond Red signal in the presence of SoxR. Although the loss of signal observed upon addition of SoxR is far smaller than that found for TATA 190 binding protein or M.HhaI, the decrease in electrochemical signal when the binding site is positioned between the probe and the electrode does provide evidence for SoxR binding. 7.3.3 Electrochemistry of P. aeruginosa SoxR As is evident in Figure 7.2, besides the Redmond Red probe, we also observe a second distinct and quasi-reversible electrochemical signal at +200 mV versus NHE upon addition of SoxR. The signal is observed only after SoxR addition and is not affected by Redmond Red because it is also present in the absence of the probe. Note that no redox signature is observed at −290 mV versus NHE, the potential previously reported for SoxR in solution (Figure 7.2). Incubation of the DNA-modified surface with PA2274, a control protein that lacks an iron-sulfur cluster, does not result in the appearance of any redox signature. Furthermore, experiments with SoxR stocks featuring low iron-sulfur content after purification do not lead to appreciable cyclic voltammetric signals (data not shown). Therefore, we can assign the new signal observed to the [2Fe-2S] cluster of SoxR. In a typical experiment, the observed SoxR signal increases over a period of ≈ 15 minutes and is stable for a minimum of 18 scans before slowly decaying (Figure 7.3), although we have found a high variability in electrode stability upon addition of protein. High concentrations of protein (> 10 µM) are required for these experiments, certainly concentrations higher than is required for site-specific binding, and both the high protein concentrations and long DNA sequences used make DNA/protein film formation difficult. It is important to note that the Redmond Red signal is highly stable and exhibits no noticeable degradation during typical electrochemistry experiments. 191 Figure 7.2: Cyclic voltammetry at 50 mV/s of electrodes modified with DNA featuring Redmond Red at the bottom (A), Redmond Red above the binding site (B), and no Redmond Red (C). Voltammograms before addition of SoxR are blue whereas those after addition of SoxR are red. The sequences used in the course of these experiments are illustrated with the binding sequence for SoxR highlighted, the location of Redmond Red indicated by an “R,” and the location of abasic sites underlined. 192 As can be seen in Figures 7.2 and 7.3, the cathodic and anodic waves observed for SoxR are asymmetric: The oxidation wave is pronounced and substantially less broad compared with the reduction wave. In an ideal quasi-reversible system, the anodic to cathodic peak current ratio is unity (42, 44), but this is certainly not the case for SoxR. We find an anodic to cathodic peak current ratio of 3.0 for SoxR in Figure 7.2, strongly indicative of a non-ideal and quasi-reversible electrochemical response. In contrast, the anodic to cathodic peak current ratio is 1.3 for the 3′-Redmond Red on the same film, far closer to the ideal value for a fully reversible system. These data show that the electrochemistry of SoxR is complicated, hardly surprising given that SoxR binds DNA as a dimer. Interestingly, the asymmetries in the reduction and oxidation waves of SoxR are qualitatively distinct from those previously observed for the DNA repair enzymes MutY and Endo III; the electrochemistry of those enzymes featured a reduction wave that was somewhat more pronounced than the oxidation wave (32). The better resolved anodic wave of SoxR integrates to very low surface coverages of 0.5 pmol/cm2. This apparent low coverage is comparable to that of 2 pmol/cm2 previously found for MutY at DNA monolayers on gold (33) and may reflect poor coupling of the iron-sulfur cluster with the base pair stack. However, the Redmond Red probe at the bottom of the DNA monolayer integrates to surface coverages of 1 pmol/cm2 whereas the Redmond Red probe at the top the film integrates to coverages that are 3-fold lower (over sample sizes of at least 10 electrodes). These values are far less than the ideal DNA surface coverage of 10 pmol/cm2 expected for a loosely packed DNA monolayer and indicate that the amount of DNA on the surface is the main determinant of the size of the SoxR signal. 193 Figure 7.3: Binding of SoxR to the DNA-modified film. (Left) Background subtracted cyclic voltammetry of P. aeruginosa SoxR at DNA-modified graphite electrodes at a 50 mV/s scan rate immediately after addition of SoxR (light) and 20 min after addition of SoxR (dark) revealing the signal observed. (Right) Integrated anodic charge for SoxR showing the growth of the signal as a function of time. 194 Despite the broad cathodic wave, we can calculate an upper bound for the number of electrons transferred for the oxidation of SoxR. For an ideal surface-bound species, the slope derived from the plot of peak current as a function of scan rate divided by the integrated charge Q at any scan rate is equal to nF/4RT, where F is Faraday's constant, R is the gas constant, and T is the temperature (44); performing this operation for the Redmond Red probe at the bottom of the monolayer (Q anodic/cathodic = ≈ 17 nC at any scan rate) yields a value of n = 2, as expected for 2 e− transfer to the resorufin moiety. The integrated charge for the anodic wave of SoxR on the same film varies from 3 to 9 nC, indicating that SoxR receives at most half the number of electrons transferred to the Redmond Red. If we assume that all of the DNA is bound and that the Redmond Red signal at the bottom of the monolayer corresponds perfectly to the number of DNA molecules on the surface (highly likely for the sparse films obtained), we can estimate that each DNA-bound SoxR dimer undergoes at most a one electron oxidation/reduction. In fact, all of these observations are consistent with titrations of free SoxR, which deviate from ideal reductions, but appear also to yield values of n = 1 (8, 45). 7.3.4 Comparison of the Voltammetry of E. coli and P. aeruginosa SoxR To expand our work to multiple organisms, we directly compared the electrochemistry of E. coli and P. aeruginosa SoxR. Only weak cyclic voltammetry for E. coli SoxR is obtained, irrespective of the source, likely because of poor solubility. Therefore, the comparison was made using square wave voltammetry, which is a more sensitive technique and allows for better discrimination of small signals. As can be seen in Figure 7.4, the potentials, referenced to the Redmond Red internal standard, are nearly identical for P. aeruginosa SoxR and E. coli SoxR. This observation is consistent with 195 Figure 7.4: Square wave voltammetry of P. aeruginosa SoxR (Left) and E. coli (Right) at DNA-modified graphite electrodes at a frequency of 15 Hz showing both the Redmond Red and SoxR signals. The 5′ Redmond Red-modified sequence was 5′-AGR GTA AAA CCT CAA GCA AAC TTG AGG TCA AGC CAA-3′ plus pyrene-modified complement, where “R” indicates the position of the probe. 196 the in vitro redox titrations of SoxR in the absence of DNA that found their potentials to differ from one another by ≈ 10 mV (7, 8, 15). Here, we see that the DNA-bound potentials are indistinguishable. The identical redox potentials of the free and DNAbound E. coli and P. aeruginosa SoxR should allow both to activate transcription upon oxidation. In fact, previous work has shown that P. putida SoxR is functional and can complement an E. coli SoxR deletion mutant, lending credence to the in vivo importance of these observations (13). 7.3.5 Electrochemistry of the P. Aeruginosa S95L SoxR Mutant To further explore the surrounding environment of the iron sulfur cluster of P. Aeruginosa SoxR, the voltammetry of a S95L SoxR mutant was also investigated at a DNA-modified surface (Figure 7.5). The serine 95 residue (or its neighbor) may be important for the in vivo regulation of SoxR activity since it is conserved for SoxR homologues in over one hundred organisms (46). As can be seen in Figure 7.5, the redox signal of the S95L mutant is very qualitatively and qunatitatively simialr to that of wild type P. aeruginosa SoxR in Figure 7.2. When referenced to the Redmond Red internal standard, the midpoint potential of the S95L mutant is shifted by ≈ -15 mV, and its cathodic wave is shifted by ≈ -40 mV (relative to the wild type, DNA-bound SoxR). These observations are consistent with the in vitro reductive titrations of wild type SoxR and a transcriptionally active S95L mutant in E. coli; the potentials of the two proteins differed from one another by ≈ -60 mV (47). The data is also consistent with higher transcriptional activity by the S95L E. coli SoxR mutant (due to an enhanced sensitivity to oxidation), again supporting the in vivo relevance of these observations (47). 197 Figure 7.5: Cyclic voltammetry at 50 mV/s of an electrode modified with DNA featuring Redmond Red above the binding site. Voltammograms before addition of the P. aeruginosa SoxR S95L mutant are black whereas those after addition of S95L SoxR are red. The sequence context is the same as in Figure 7.2. 198 7.3.6 Comparison of the Voltammetry of SoxR to that of Free Iron As an important control, we explored the electrochemistry of free iron in the presence of excess L-cysteine under conditions identical to those used for SoxR experiments. As can be seen in Figure 7.6, the voltammetry of free iron on either bare or DNA-modified HOPG is vastly different from that found for SoxR; the midpoint potential is 0 mV versus NHE, rather than +200 mV versus NHE. In fact, the backfilled DNA monolayer dramatically broadens and attenuates the electrochemical signal, presumably by providing reduced access to the surface. These data conclusively indicate that we are, indeed, observing the voltammetry of a protein-bound iron sulfur cluster cofactor. 7.4 Discussion and Implications The activity of SoxR, a transcription factor containing an Fe-S cofactor, is regulated via a redox switch: SoxR triggers transcription in its oxidized state (4, 5). However, the redox potentials of free E. coli and P. aeruginosa SoxR have previously been determined to be approximately −290 mV in solution (7, 8, 15). Although the redox potential of SoxR can explain how it is maintained in its reduced state by coupling it to the cellular NADPH/NADP+ pool (−40 mV), it was unclear how the relatively low potential of −290 mV would allow for specificity in vivo. To understand the activation of SoxR at a mechanistic level, it is crucial to determine its redox potential within an appropriate context. 199 A. B. Figure 7.6: Cyclic voltammetry of ≈ 50 µM FeCl3 with an excess of dithionite at 50 mV/s in SoxR storage buffer. An unmodified HOPG electrode (A) and an electrode modified with DNA featuring Redmond Red below the binding site (B) are shown. Voltammograms before addition of free iron are black whereas those after addition of free iron are red. The sequence context is the same as in Figure 7.2. 200 Here, using DNA-modified HOPG electrodes, we have demonstrated that DNA association positively shifts the redox potential of SoxR to 200 mV versus NHE. The +490-mV shift between the free and DNA-bound states of SoxR is functionally crucial because it keeps SoxR reduced across a range of intracellular potential. For example, Figure 7.7 shows standard and free midpoint potentials for a variety of cellular redox pairs, and where DNA-bound and free SoxR are positioned along this series. Although numerous redox couples, ranging from glutathione to FADH, are oxidants for soluble SoxR, they are reductants to DNA-bound SoxR. In fact, the positive shift in potential associated with DNA binding means that DNA-bound SoxR is primarily in the reduced, transcriptionally silent form in vivo. Oxidative stress serves to promote oxidation of DNA-bound SoxR, activating the numerous genes required to protect the organism. This provides a rationale for how DNA-bound SoxR can serve as an effective sensor of oxidative stress in E. coli. In P. aeruginosa, the paradigm for SoxR activation may be different. Here, activation may be promoted by pyocyanin. When we consider the standard potential of the phenazine pyocyanin (E m = −34 mV at pH 7 and E m = −110 mV at pH 8), we predict it would also act as a reductant for DNA-bound SoxR (48). However, pyocyanin is an extracellular electron shuttle that reacts readily with oxygen, as indicated by the bright blue color of P. aeruginosa cultures. Uptake of oxidized pyocyanin increases the intracellular ratio of the oxidized versus the reduced form and thus favors the oxidation of SoxR. Considering a one-electron transfer under physiological conditions (pH 7 and 37 °C), to shift the redox equilibrium of DNA-bound SoxR toward its oxidized state would 201 Figure 7.7: Redox potentials of free and DNA-bound SoxR along with those of cellular oxidants/reductants at pH 7. 202 require a ratio of oxidized-to-reduced pyocyanin of at least 6,500:1. It remains to be determined whether this is of physiological relevance in P. aeruginosa. The substantial shift in SoxR potential on DNA binding of ≈ 500 mV is striking but understandable. The significance of the molecular environment for tuning the redox potentials of Fe-S clusters is well documented (49–51): Each hydrogen-bonding interaction with the cluster can cause a potential shift of ≈ 80 mV. Moreover, for [4Fe-4S] clusters in proteins, all with the same ligating residues, cluster potentials vary from approximately −600 mV for ferredoxins to approximately +400 mV for high potential iron proteins (52). We have previously observed a negative shift of at least ≈ 200 mV for Endo III in the presence of DNA (34). Because the structures of Endo III with and without DNA were known and showed no significant distortion in the protein (53–55), thermodynamically this shift was interpreted as a favorable shift in the binding affinity of the protein in the oxidized form relative to the reduced form (34), perhaps not so surprising on binding to the DNA polyanion. By contrast, although the binding affinities of oxidized and reduced SoxR are comparable (6, 7), SoxR and other MerR-type transcriptional regulators have been shown to induce conformational changes of the promoter region (1, 56–61). In particular, copper phenanthroline footprinting studies have provided strong evidence that SoxR significantly distorts its promoter sequence (62, 63). Although this experiment only reports on the reduced form of SoxR, the observed loss of signal for Redmond Red found here strongly supports a DNA-distortion mechanism. We propose that the more positive reduction potential for DNA-bound SoxR yields a higher energy complex, which may 203 drive a conformational change in the protein/DNA complex. If this were the case, it would constitute an effective means of allosteric regulation in vivo. It is important to note that the crystal structure of SoxR in any form has not been reported (64). Therefore, the structural difference between the free (low energy) and DNA-bound (high energy) complexes is not clear, but a positive shift of the magnitude we observe has been associated with bulk folding of other metalloproteins from P. aeruginosa (65). These energetic differences have been attributed to burying the cofactor in a more hydrophobic environment. Consequently, we predict a large structural difference between the free and DNA-bound SoxR to provide a rationale for the dramatic shift in potential associated with binding. Within a broader context, these data illustrate that it is critical to take the effect of DNA binding into account when considering the redox characteristics of DNA binding proteins. It is also likely that it is the DNA-bound potential of these proteins that is most relevant within the crowded environment of the cell, and this potential may be altered even further upon recruitment of RNA polymerase. Therefore, in many cases, as with SoxR, it is perhaps the redox characteristics within a multiprotein/DNA complex that must be considered. In fact, because several transcription factors feature iron-sulfur clusters as sensor elements, a change in the redox potential of these cofactors upon binding DNA may generally be an important trait to consider. 204 REFERENCES (1) Brown, N. L., Stoyanov, J. V., Kidd, S. P., Hobman, J. L. (2003) FEMS Microbiol. Rev. 27, 145–163. (2) Hobman, J. L., Wilkie, J., Brown, N. L. (2005) Biometals 18, 429–436. (3) Giedroc, D. P., Arunkumar, A. I. (2007) J. Chem. Soc. Dalton Trans. 3107– 3120. (4) Pomposiello, P. J., Demple, B. (2001) Trends Biotechnol. 19, 109–114. (5) Demple, B., Ding, H., Jorgensen, M. (2002) Methods Enzymol. 348, 355–364. (6) Hidalgo, E., Demple, B. (1994) EMBO J. 13, 138–146. (7) Gaudu, P., Weiss, B. (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 10094–10098. (8) Ding, H., Hidalgo, E., Demple, B. (1996) J. Biol. Chem. 271, 33173–33175. (9) Ding, H., Demple, B. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 8445–8449. (10) Pomposiello, J., Bennik, M. H. J., Demple, B. (2001) J. Bacteriol. 183, 3890– 3902. (11) Ha, U., Jin, S. (1999) Infect. Immun. 67, 5324–5331. (12) Stover, C. K., et al. (2000) Nature 406, 959–964. (13) Park, W., Pena-Llopis, S., Lee, Y., Demple, B. (2006) Biochem. Biophys. Res. Commun. 341, 51–56. (14) Dietrich, L. E. P., Price-Whelan, A., Petersen, A., Whiteley, M, Newman, D. K. (2006) Mol. Microbiol. 61, 1308–1321. (15) Kobayashi, K., Tagawa, S. (2004) J. Biochem. 136, 607–615. (16) Zheng, M., Storz, G. (2000) Biochem. Pharmacol. 59, 1–6. 205 (17) Liochev, S. I., Fridovich, I. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 5892– 5896. (18) Kobayashi, K., Tagawa, S. (1999) FEBS Lett. 451, 227–230. (19) Koo, M.S., Lee, J. H., Rah, S. Y., Yeo, W. S., Lee, J. W., Lee, K. L., Koh, Y. S., Kang, S. O., Roe, J. H. (2003) EMBO J. 22, 2614–2622. (20) Hwang. C., Lodish, H. F., Sinskey, A. J. (1995) Methods Enzymol. 251, 212– 221. (21) Kobayashi, K., Tagawa, S. (1997) J. Inorg. Biochem. 67, 258. (22) Liochev, S. I., Benov, L., Touati, D., Fridovich, I. (1999) J. Biol. Chem. 274, 9479–9481. (23) Ding, H., Demple, B. (1998) Biochemistry 37, 17280–17286. (24) Price-Whelan, A., Dietrich, L. E. P., Newman, D. K. (2007) J. Bacteriol. 189, 6372–6381. (25) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucleic Acids Res. 27, 4830–4837. (26) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (27) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. (28) Kelley, S. O., Jackson, N. M., Hill, M. G., Barton, J. K. (1999) Angew. Chem. Int. Ed. 38, 941–945. (29) Inouye, M., Ikeda, R., Takase, M., Tsuri, T., Chiba, J. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 11606–11610. (30) Okamoto, A., Kamei, T., Saito, I. (2006) J. Am. Chem. Soc. 128, 658–662. 206 (31) Gorodetsky, A. A., Green, O., Yavin, E., Barton, J. K. (2007) Bioconjugate Chem. 18, 1434–1441. (32) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 12543–12547. (33) Boal, A. K., Yavin, E., Lukianova, O. A., O'Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (34) Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. (35) Cunningham, R. P., Asahara, H., Bank, J. F., Scholes, C. P., Salerno, J. C., Surerus, K., Munck, E., McCracken, J., Peisach, J., Emptage, M. H. (1989) Biochemistry 28, 4450–4455. (36) Asahara, H., Wistort, P. M., Bank, J. F., Bakerian, R. H., Cunningham, R. P. (1989) Biochemistry 28, 4444–4449. (37) Fu, W., O'Handley, S., Cunningham, R. P., Johnson, M. K. (1992) J. Biol. Chem. 267, 16135–16137. (38) Boal, A. K., Yavin, E., Barton, J. K. (2007) J. Inorg. Biochem. 101, 1913– 1921. (39) Stephens, P. J., Jollie, D. R., Warshel, A. (1996) Chem. Rev. 96, 2491–2513. (40) Kim, Y., Long, E. C., Barton, J. K., Lieber, C. M. (1992) Langmuir 8, 496– 500. (41) Giancarlo, L. C., Flynn, G. W. (1998) Ann. Rev. Phys. Chem. 49, 297–336. (42) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. 207 (43) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nature Biotechnol. 20, 282– 286. (44) Udit, A. K., Hill, M. G., Bittner, V. G., Arnold, F. M., Gray, H. B. (2004) J. Am. Chem. Soc. 126, 10218–10219. (45) Hidalgo, E., Ding, H., Demple, B. (1997) Trends Biochem. Sci. 22, 207–210. (46) Chander, M., Raducha-Grace, L., Demple, B. (2003) J. Bacteriol. 185, 2441– 2450. (47) Hidalgo, E., Ding, H., Demple, B. (1997) Cell 88, 121–129. (48) Price-Whelan, A., Dietrich, L. E. P., Newman, D. K. (2006) Nature Chem. Biol. 2, 71–78. (49) Rusling, J. F. (1998) Acc. Chem. Res. 31, 363–369. (50) Carter, C. W., Jr., Kraut, J., Freer, S. T., Alden, R. A., Sieker, L. C., Adman, E., Jensen, L. H. (1972) Proc. Natl. Acad. Sci. U. S. A. 69, 3526–3529. (51) Heering, H. A., Bulsink, Y. B. M., Hagen, W. R., Meyer, T. E. (1995) Biochemistry 34, 14675–14686. (52) Beinert, H. (2000) J. Biol. Inorg. Chem. 5, 2–15. (53) Kuo, C. F., McRee, D. E., Fisher, C. L., O'Handley, S. F., Cunningham, R. P., Tainer, J. A. (1992) Science 258, 434–440. (54) Thayer, M. M., Ahern, H., Xing, D., Cunningham, R. P., Tainer, J. A. (1995) EMBO J. 14, 4108–4120. (55) Fromme, J. C., Verdine, G. L. (2003) EMBO J. 22, 3461–3471. (56) Ansari, A. Z., Bradner, J. E., O'Halloran, T. V. (1995) Nature 374, 371–375. 208 (57) Outten, C. E., Outten, F. W., O'Halloran, T. V. (1999) J. Biol. Chem. 274, 37517–37524. (58) Zheleznova-Heldwein, E. E., Brennan, R. G. (2001) Nature 409, 378–382. (59) Godsey, M. H., Baranova, N. N., Neyfakh, A. A., Brennan, R. G. (2001) J. Biol. Chem. 276, 47178–47184. (60) Changela, A., Chen, K., Xue, Y., Holschen, J., Outten, C. E., O’Halloran, T. V., Mondragon, A. (2003) Science 301, 1383–1387. (61) Newberry, K. J., Brennan, R. G. (2004) J. Biol. Chem. 279, 20356–20362. (62) Hidalgo, E., Bollinger, J. M., Jr,, Bradley, T. M., Walsh, C. T., Demple, B. (1995) J. Biol. Chem. 270, 28908–28914. (63) Hidalgo, E., Demple, B. (1997) EMBO J. 16, 1056–1065. (64) Watanabe, S., Kita, A., Kobayashi, K., Takahashi, Y., Miki, K. (2006) Act. Crystal. F 62, 1275–1277. (65) Wittung-Stafshede P., Hill, M. G., Gomez, E., Di Bilio, A. J., Karlsson, B. G., Leckner, J., Winkler, J. R., Gray, H. B., Malmstrom, B. G. (1998) J. Biol. Inorg. Chem. 3, 367–370. (66) Chander, M., Demple, B. (2004) J. Biol. Chem. 279, 41603–41610. 209  Chapter 8: Scanning Electrochemical Microscopy of DNA Monolayers Modified with Nile Blue Adapted from: Gorodetsky, A. A., Hammond, W., Hill, M. G., Slowinski, K., Barton, J. K. (2008) Langmuir ASAP. Electrochemical measurements were performed in collaboration with William Hammond. 210 ABSTRACT Scanning electrochemical microscopy (SECM) is used to probe long range charge transport (CT) through DNA monolayers containing the redox-active Nile Blue (NB) intercalator covalently affixed at a specific location in the DNA film. At substrate potentials negative of the formal potential of covalently attached NB, the electrocatalytic reduction of Fe(CN)63- generated at the SECM tip is observed only when NB is located at the DNA/solution interface; for DNA films containing NB in close proximity to the DNA/electrode interface, the electrocatalytic effect is absent. This behavior is consistent with both rapid DNA-mediated CT between the NB intercalator and the gold electrode as well as a rate-limiting electron transfer between NB and the solution phase Fe(CN)63-. The DNA-mediated nature of the catalytic cycle is confirmed through sequence-specific and localized detection of attomoles of TATA-binding protein, a transcription factor which severely distorts DNA upon binding. Importantly, the strategy outlined here is general and allows for the local investigation of the surface characteristics of DNA monolayers both in the absence and presence of DNA-binding proteins. These experiments highlight the utility of DNA-modified electrodes as versatile platforms for SECM detection schemes that take advantage of CT mediated by the DNA base-pair stack. 211 8.1 Introduction Numerous studies have shown that DNA efficiently mediates long range charge transport both in solution (1–5) and at DNA-modified surfaces (6–7). The electrochemical response at DNA self-assembled monolayers (SAMs) is remarkably sensitive to base stacking perturbations, with the efficiency of CT to an electroactive probe reflecting the integrity of the base pair stack. Consequently, DNA electrochemistry has been exploited successfully in the development of simple assays for the rapid detection of single-nucleotide polymorphisms (8–10), base lesions (11), and DNA/protein binding (12–18). Controlling the orientation and packing of DNA helices within these SAMs is critical for many analytical applications, yet assessing film morphology and homogeneity, particularly in aqueous media, remains a difficult challenge. Previously, the morphology of DNA SAMs on gold has been investigated by scanning tunneling microscopy (STM) (19–21), atomic force microscopy (AFM) (22–24), and most recently, scanning electrochemical microscopy (SECM) (25–28). SECM provides submicron visualization of DNA films through in situ imaging of chemical reactivity at modified insulating or conducting substrates (29, 30). SECM is also inherently compatible with aqueous environments, thereby allowing for the direct electrochemical profiling of DNA arrays with micron resolution (29, 30). In addition, SECM has been used successfully to monitor DNA hybridization and to detect DNA lesions (29–34). One advantage of this technique over conventional bulk electrochemistry is that the micron-level diameter of the SECM 212 tip inherently minimizes contributions from imperfections on large-sized electrodes by sampling a local area of the film at a specific substrate potential. Electrochemical assays based on DNA CT have frequently utilized non-covalent mediators to collect global information (8–11, 35, 36), but the electrochemistry of these mediators can be significantly complicated by contributions from CT pathways that are not DNA-mediated (37). The use of covalent probes at fixed positions within the film (38–45), particularly in conjunction with backfilling of the DNA monolayer with a suitable straight-chain diluent molecule, minimizes such contributions from defects, pinholes, and other imperfections. Therefore, SECM investigations of DNA monolayers can particularly benefit from the use of covalently attached probes. Here, we report the application of SECM in probing long range CT across a DNA monolayer containing the redox-active intercalator Nile Blue (NB) covalently attached at discrete sites along the individual DNA helices. Nile Blue is particularly useful for such studies since it can be easily incorporated into DNA on solid support and displays catalytic activity (13). Therefore, based on our previous investigations of the catalytic reduction of ferricyanide by methylene blue (MB) at DNA-modified electrodes (8–11, 46), we have replaced MB with covalently bound NB to reduce ferricyanide, which is generated at the SECM tip in our current experiments. Only Nile Blue attached at the solution-exposed periphery of the film is capable of supporting catalytic regeneration of ferrocyanide present in solution. Furthermore, this cycle can be interrupted in a sequencespecific manner by addition of TATA Binding Protein (TBP), a transcription factor that bends DNA by ~ 90º upon binding (12, 13). Taken together, these studies establish the upright orientation of individual helices within a relatively homogeneous dilute 213 monolayer and demonstrate the critical role of efficient long-range electrochemical charge transport through the π-stack. They additionally suggest a sensitive and highly selective SECM method for assaying specific DNA/protein interactions with even dilute and imperfect assemblies of probe DNA. 8.2 Experimental 8.2.1 Materials All reagents for DNA synthesis were purchased from Glen Research. Methylene blue, ruthenium hexammine chloride, potassium ferricyanide, potassium ferrocyanide, and 11-mercaptoundecylphosphoric acid were purchased from Sigma in the highest purity available and used as received. Nile Blue perchlorate was purchased from Acros in laser grade purity. TATA Binding protein was custom ordered from Protein One, Inc. Bovine Serum Albumin (BSA) was obtained from New England Biolabs, Inc. Platinum scanning electrochemical microscopy tips were purchased from CH Instruments, Inc. as used as received. Au (111) on mica substrates packaged under argon were purchased from Agilent, Inc., and utilized immediately. Sodium phosphate buffers were prepared with Milli-Q water and pH adjustments were made with sodium hydroxide, if necessary. 8.2.2 Synthesis of Thiol and Nile Blue-Modified DNA Oligonucleotides were prepared on solid support using standard phosphoramidite chemistry on an Applied Biosystems 394 DNA synthesizer. All of the prepared sequences were purified by multiple rounds of reverse-phased, high-performance liquid chromatography (HPLC). The pure oligonucleotides were characterized with matrix- 214 assisted laser desorption (MALDI) mass spectrometry and UV-visible (UV-Vis) spectrophotometry. Thiol-terminated oligonucleotides were synthesized according to established protocols from Glen Research, Inc. using the C6 S-S thiol modifier. After deprotection and cleavage from solid support with ammonium hydroxide (60 ºC for 8 hours), the disulfide containing DNA was purified by HPLC. The disulfide was subsequently reduced with an excess of dithiothreitol in ammonium acetate buffer at pH=8, and the free thiol containing single-stranded DNA was then purified with a second round of HPLC. As schematically illustrated in Figure 8.1, DNA modified with NB at the 5’ terminus was prepared according to ultra-mild protocols (Glen Research, Inc.) to avoid degradation of the NB moiety. Additionally, Pac-protected bases and ultra mild reagents were utilized during the synthesis to prevent undesirable capping of the protecting groups. A 17-mer sequence (either 5’-UGC GTG CTT TAT ATC TC-3’ or 5’-UGC GCG CCC GGC GCC TC-3’) was prepared on solid support with a 5-[3-acrylate NHS Ester]-deoxy uridine as the terminal 5’ base. The beads were then removed from the synthesizer and dried thoroughly. The solid support were reacted with a 10 mg/mL Nile Blue perchlorate solution in either 9:1 N,N-dimethylformamide/N,N-diisopropylethylamine or 9:1 dichloromethane/N,N-diisopropylethylamine for 12–48 hours. The beads were subsequently washed up to three times with dichloromethane or N,N-dimethylformamide, methanol, and acetonitrile. Subsequently, the Nile Blue-containing sequence was simultaneously cleaved from the support and deprotected with 0.05 M potassium 215 Figure 8.1: Schematic illustration of the coupling of Nile Blue to the NHS-Carboxy-dT modified DNA on solid support (CPG) 216 carbonate in methanol at room temperature for 12–14 hours. The overall yields of the reaction ranged from 30% to 80%. DNA modified with NB one base in from the 3’ terminus was prepared according to ultra-mild protocols in a similar fashion. A 2-mer sequence containing a 5-[3-acrylate NHS Ester]-deoxy uridine as the terminal 5’ base was prepared on solid support. The beads were removed from the synthesizer and reacted with a 10 mg/mL Nile Blue perchlorate solution in either 9:1 N,N-dimethylformamide/N,N-diisopropylethylamine or 9:1 dichloromethane/N,N-diisopropylethylamine for 12–48 hours. After thorough washing of the beads, the Nile Blue containing 2-mer was placed back on the synthesizer and the remaning 15 bases were added to the sequence, yielding the oligonucleotide 5’TGC GTG CTT TAT ATC UC-3’. The resulting sequence contained an internal Nile Blue and was deprotected with 0.05 M potassium carbonate in methanol at room temperature for 12-14 hours. The overall yields of the reaction ranged from 30% to 80%. Both thiol- and NB-modified DNA was quantified using the extinction coefficient of the single-stranded DNA at 260 nm on a Beckman UV-Vis spectrophotometer. Equimolar amounts of each strand (50 µM) were combined before the resulting solution was purged with argon. Duplexes were formed by thermally annealing in deoxygenated buffer containing 5 mM NaPi, pH = 7.1, 50 mM NaCl to 90 ºC, followed by cooling to ambient temperature. 8.2.3 Preparation of Backfilled DNA/11-mercaptoundecylphosphoric Acid Monolayers Electrode areas were defined with a Viton o-ring with estimated surface areas of either 0.1 cm2 or 0.2 cm2 in either a scanning electrochemical microscope teflon cell from 217 CH Instruments, Inc., or a custom-built teflon cell. To ensure successful experiments, great care was taken to minimize mechanical damage to the gold substrate upon mounting in the cell. After cell assembly, a fresh 25–50 µM duplex NB-DNA solution was deposited onto the gold surface. Monolayer formation was allowed to proceed in a humidified environment for a period of 24–72 hours. Incubation times outside of this window (either shorter or longer) were found to be detrimental to proper monolayer formation. Upon completion of film formation, the cell was rinsed thoroughly with phosphate buffer to remove residual DNA before the DNA-modified surface was backfilled with 11-mercaptoundecylphosphoric acid for 10–20 minutes (47). Mercaptoundecylphosphoric acid rapidly displaces DNA from the surface, so backfilling times were short and the cell was rinsed thoroughly to eliminate any residual alkanethiols. 8.2.4 Electrochemistry/Scanning Electrochemical Microscopy Experiments Cyclic voltammetry and scanning electrochemical microscopy experiments were performed using a CH Instruments electrochemical analyzer (Austin, TX). Unless otherwise noted, experiments were performed at ambient temperature in phosphate buffer at pH 7 containing either 5 mM NaPi and 50 mM NaCl or 20 mM NaPi and 80 mM NaCl under an Argon atmosphere (the oxygen content was constantly monitored). The buffer was supplemented with K4Fe(CN)6 for SECM experiments. A teflon electrochemical cell was used for all experiments with a Pt auxiliary electrode, a gold working electrode, and a silver/silver chloride (Ag/AgCl) reference electrode. Since we have observed some variation in the potential of covalently attached NB depending on the type of gold surface and the buffer conditions, the voltammetric potentials reported in this work have a maximum uncertainty of 50 mV. 218 In a typical SECM experiment, total or partial passivation of the substrate was confirmed through cyclic voltammetry of the ferri/ferrocyanide couple at the substrate. Approach curves were obtained with either 2 µm or 10 µm diameter platinum tips which were polarized between 600 mV and 1 V, and identical results were obtained regardless of the tip bias. To aid interpretation of the results, substrates featuring a significant free NB signal or virtually no passivation against ferricyanide were discarded. Initial approach curves to the substrate were recorded at a 0 mV substrate bias to confirm the insulating behavior of the NB-DNA monolayer. Subsequent approach curves were recorded with the substrate bias modulated between 0 mV and -400 mV. For imaging of the substrate, the SECM tip was moved to within less than one tip radius of the substrate before being retracted by 2–5 microns prior to scans. Great care was taken during all experiments to avoid damaging the DNA monolayer through direct contact of the tip with the surface. 8.2.5 TATA Binding Protein Experiments In a typical TBP detection experiment by SECM, approach curves or substrate scans were initially recorded with the bias modulated between 0 mV and -400 mV to confirm switching from negative to positive feedback. Subsequently, TBP was added to the SECM cell from a concentrated ~16 µM stock in phosphate buffer at pH = 7 containing 5 mM KPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol. The solution was equilibrated via gentle pipetting, and approach curves or substrate scans were again recorded with the bias modulated between 0 mV and -400 mV. 219 8.3 Results 8.3.1 Electrochemistry of Backfilled DNA Monolayers Modified with Nile Blue Figure 8.2 shows a schematic illustration of DNA monolayers modified with NB either at the top or bottom (Figure 8.1 shows the structure of the nucleotide). DNA monolayers modified with NB display a pH dependent redox couple at ~ -200 mV versus Ag/AgCl due to the 2e- reduction of the bound probe (13, 48, 49). Since modification of the exocyclic amine of NB perturbs its aromatic core with an accompanying shift in its redox potential of ~ 200 mV relative to free NB (50, 51), DNA monolayer stability can be assessed in situ with the appearance of a NB signal at ~ -400 mV indicating film degradation. Coulometric measurements of the NB signals for monolayers labeled at either the top or bottom yield typical DNA surface coverages of ~1–3 pmol/cm2 (compared to ~ 40 pmol/cm2 or greater for densely packed monolayers) (10). As expected for a surface-bound species, plots of peak current as a function of scan rate are linear, and the ratios of the cathodic to anodic charge are approximately unity (Figure 8.3) (52). Overall, the electrochemical characteristics of the two types of monolayers (functionalized at either the top or the bottom) are very similar, despite the ~ 45 Å difference in gold/NB separations. This observation is consistent with previous studies involving site-specific labeling of DNA monolayers with redox-active intercalators (13, 39, 40). Based on previous work involving electrocatalytic reduction of ferricyanide at DNA-modified substrates via intercalated Methylene Blue (MB) (8, 46), we explored the electrochemistry of covalently bound NB in the presence of ferricyanide (Figure 8.4). 220 Figure 8.2: Schematic illustration of SECM imaging of DNA monolayers modified with Nile Blue (NB) at the top or the bottom; the negatively charged backfilling underlayer is also shown. The sequence was 5’-UGC GTG CTT TAT ATC TC-3’ (top NB) and 5’TGC GTG CTT TAT ATC UC-3’ (bottom NB) where the italicized U indicates the location of the NB moiety. 221 Figure 8.3: Cyclic voltammetry of DNA monolayers at various scan rates modified with NB at the bottom (A) and top (B) in pH =7.1, 5 mM NaPi, 50 mM NaCl buffer. The corresponding plots of peak current as a function of scan rate are shown on the right. The sequence was 5’-TGC GTG CTT TAT ATC UC-3’ (bottom NB) and 5’-UGC GTG CTT TAT ATC TC-3’ (top NB) where the italicized U indicates the location of the NB moiety. 222 Figure 8.4: Cyclic voltammetry at 100 mV/s of DNA monolayers modified with NB at the top (blue) and bottom (red) in 80 mM NaCl, 20 mM NaPi, pH= 7.2, 5 mM K3Fe(CN)6 buffer. The sequence was 5’-UGC GTG CTT TAT ATC TC-3’ (top NB) and 5’-TGC GTG CTT TAT ATC UC-3’ (bottom NB) where the italicized U indicates the location of the NB moiety. Note that a catalytic peak is only observed for a DNA monolayer modified with NB at the top. 223 Figure 8.5: Successive cyclic voltammograms of ferricyanide at a bare Au electrode in pH =7.1, 5 mM NaPi, 50 mM NaCl buffer before (red) and after (other colors) addition of 1 mM 11-mercaptoundecylphosphoric acid to the buffer. The voltammograms correspond to different exposure times in the 11-mercaptoundecylphoshoric acid containing buffer. Complete attenuation of the ferricyanide signal is observed within 15 minutes. 224 The DNA monolayers were backfilled with 11-mercaptoundecylphoshoric acid, a negatively charged alkanethiol which completely passivates against ferricyanide (Figure 8.4 and Figure 8.5). Importantly, the presence of this underlayer makes it more likely that the DNA is electrostatically repelled by the negatively charged surface (47), ensuring an upright orientation in a dilute film. For a DNA monolayer modified with NB at the top, only an asymmetric cathodic peak is observed at the reduction potential of NB. However, for a DNA monolayer modified with NB at the bottom, the catalytic peak is absent. This data strongly indicates that only solution-exposed NB is capable of reducing ferricyanide. 8.3.2 SECM of DNA Monolayers Modified with Nile Blue in Feedback Mode To visualize the NB/ferricyanide catalytic cycle both locally and directly (8, 46), we carried out SECM measurements at DNA monolayers covalently modified with NB (Figure 8.1). In these experiments, Fe(CN)64- is oxidized in solution at a platinum SECM tip, yielding a steady-state, diffusion-controlled current. The SECM tip is then physically lowered toward the DNA-modified gold substrate at a constant rate, and the currentdistance feedback curve is recorded at various substrate bias potentials. If reduction of tip-generated Fe(CN)63- at the underlying substrate is blocked by the monolayer, the steady-state tip current drops upon approach (negative feedback) due to restricted diffusion of bulk ferrocyanide. If, on the other hand, the rate of catalytic Fe(CN)63reduction via reduced NB at the modified substrate is rapid, redox cycling between the tip and substrate occurs, yielding increased tip currents (positive feedback). As shown in Figure 8.6, films containing NB attached at the bottom of the DNA helices result in complete negative feedback at substrate bias potentials of both 0 mV and -400 mV. At 0 mV, bound NB remains in its oxidized state and there is a relatively small 225 Figure 8.6: SECM approach curves obtained for DNA monolayers modified with NB at the top and bottom in 80 mM NaCl, 20 mM NaPi, pH= 7.2, 5 mM K4Fe(CN)6 buffer with a tip-approach speed of 1 µm/s. The tip approach current has been normalized by the tip steady-state current at an infinite distance from the substrate. The normalized current is plotted against d/a where d is the tip/substrate separation and a represents the tip radius. For the DNA monolayer functionalized with NB at the top, positive feedback is observed at -400 mV (A), and negative feedback is observed at 0 mV (B). For the DNA monolayer modified with NB at the bottom, negative feedback is observed both at 0 mV (C) and -400 mV (D). Theoretical curves are shown for ideal negative feedback at an insulating substrate (black solid line) and for finite heterogeneous kinetics with a rate constant of kapp=4.6*10-2 cm/s for this specific substrate (red solid line). 226 (~ 200 mV) overpotential for ferricyanide reduction. At -400 mV, however, the NB is fully reduced and the substrate overpotential increases to ~ 600 mV. Despite this high driving force and the ability of reduced NB to act as a redox catalyst, the SECM response mimics that expected for a purely insulating substrate. This indicates sufficient electrostatic repulsion at the DNA/phosphate-terminated surface to effectively block tipgenerated ferricyanide not only from the underlying gold electrode (26, 28) but also from the reduced NB buried deep within the DNA film. A markedly different response is observed for DNA monolayers modified with NB at the top of the film. At substrate bias potentials of 0 mV, pure negative feedback is again observed. However, when the substrate bias voltage is held at -400 mV (a value negative of the formal reduction potential of NB) the response switches to positive feedback. Such a response signals rapid CT between tip-generated Fe(CN)63- and reduced NB, which is now bound at the solvent-exposed periphery of the monolayer. Indeed, assuming that the tip current is controlled by NB-mediated CT through the DNA film, the approach curve can be modeled using well-known analytical approximations for finite heterogeneous kinetics (53–55). The fits in Figure 8.6 show excellent agreement with the approximations both for the insulating monolayers (black solid line for negative feedback) and for the finite heterogeneous kinetics of ferricyanide reduction (red solid line for positive feedback). The one-parameter fit shown in Figure 8.6A for the DNA film modified at the top corresponds to a surprisingly rapid apparent rate constant for regeneration of ferricyanide of kapp= 6 ± 3 * 10-2 cm/s as measured at -400 mV versus SSCE (over four independent trials). 227 Two pieces of experimental evidence further substantiate the proposed DNAmediated pathway for the SECM response. First, DNA quantification by ruthenium hexammine assay (56–58) for DNA films modified with NB either at the top or bottom yields very similar surface coverages (a minor contribution to the ruthenium hexammine signal is expected from 11-mercaptoundecylphosphoric acid due to its single phosphate group). Although this assay provides only a qualitative measure of surface coverage (58), it indicates nevertheless that both electrodes are covered with similar amounts of negative charge. Thus, for substrate potentials of -400 mV, the large discrepancy in ET rate across DNA films modified with NB at the top versus that modified at the bottom cannot be ascribed to relative differences in electrostatic screening of ferricyanide. Second, DNA films modified with NB at the bottom can support electrocatalytic reduction of ferricyanide, but only upon addition of a redox mediator that has access to the film/solution interface. Thus, as an additional control experiment, the approach curve in SECM feedback mode was recorded for DNA monolayers modified with NB at the bottom before and after addition of 1.5 µM Methylene Blue (MB) to the solution (Figure 8.7). While the addition of MB does not change the observed negative feedback at a substrate potential of 0 mV, measurable positive feedback is found at substrate potentials negative of the MB redox couple. Analogous behavior was recently observed by Zhou for DNA-modified electrodes in the presence of MB in solution (28). The positive feedback observed in our experiments with MB is notably weaker versus that found for the monolayers covalently modified with NB at the top, and yields an apparent rate constant for ferricyanide reduction by non-covalent MB of kapp=4.8 * 10-3 cm/s as measured at 300 mV versus SSCE, which is similar to that found earlier (28). 228 Figure 8.7: SECM approach curves taken for Nile Blue-DNA monolayers before (C, D) and after (A, B) addition of 1.5 µM Methylene Blue at substrate biases of 0 mV (B, D) and -300 mV (A, C). Approach curves were taken in pH= 7.2, 20 mM Na2HPO4, 80 mM NaCl, and 5 mM K4Fe(CN)6 buffer. The sequence was 5’-UGC GTG CTT TAT ATC TC-3’ where the italicized U indicates the location of the NB moiety. Theoretical fits for finite heterogeneous kinetics with kapp=4.8*10-3 cm/s and negative feedback are shown as solid lines for comparison. 229 Moreover, the MB controls yield essentially identical results with and without backfilling, revealing that MB cannot be simply incorporated within the 11- mercaptoundecylphosphoric acid underlayer. This experiment indicates that the DNA film modified with NB at the bottom supports long-range DNA-mediated CT, but measurable CT occurs only as long as the Fe(CN)63- in solution has access to the redoxactive intercalator within the film. 8.3.3 SECM Imaging of DNA Monolayer Morphology Previously, undiluted DNA monolayers of the type utilized during the course of this study have been characterized by atomic-force microscopy (22–24), fluorescence spectroscopy (59–61), surface-plasmon resonance (62), and radio-active tagging (12, 63). These experiments (along with electrochemical measurements) confirm that dense DNA films adopt an upright orientation with respect to the surface. However, a distinct geometric orientation of a dilute submonolayer film has been difficult to characterize conclusively with such techniques. To gain further insight into the morphology of DNA monolayers modified with NB, SECM in imaging mode was utilized to examine the DNA films. In general, the images indicate that the backfilled monolayers investigated here are electrostatically smooth within the one micron resolution of the SECM tip. The analysis of a large collection of images (hundreds collected over numerous substrates) indicates that some electrodes are completely and uniformly covered by the DNA film. On the other hand, some images feature domains on the electrode surface that are either DNA-free or contain thinner and less organized DNA assemblies. Indeed, such partially covered or imperfect DNA monolayers were found to correlate well with incomplete 230 passivation of ferricyanide at the substrate, allowing for a detailed analysis of even these imperfect surfaces with SECM. Figure 8.8 shows images obtained for DNA substrates modified with NB at the top or bottom over an 85 µm by 85 µm scan area. The entire image is essentially flat and uniform with the exception of the small area in the rear corner, which was generated by close approach of the tip to the substrate upon initiation of the scan thereby resulting in a partially uncovered surface. Significantly, it is apparent that changing the substrate bias has only a small effect on the steady-state current at the tip for films functionalized with NB at the bottom, yielding nearly identical images at 0 mV and -400 mV. However, for the film functionalized with NB at the top, changing the substrate bias modulates or “switches” the entire area under the scan with a resulting significant enhancement of the steady-state current at -400 mV. These observations parallel those obtained via the feedback mode of the SECM and illustrate the powerful imaging capabilities of this technique. Notably, the most efficient modulation occurs within regions of the DNA film that exhibit the lowest currents at substrate potentials of 0 mV (Figure 8.8A, 8.8B). This can be best illustrated by mapping the switching ratios i.e. the currents obtained at a substrate bias of -400 mV versus a substrate bias of 0 mV for both types of films (Figure 8.8C, 8.8D). For example, the DNA monolayer modified with NB at the top exhibits low currents at 0 mV over the entire scan area, except for the partially uncovered corner (rear corner in Figure 8.8A, 8.8C); at a substrate bias of -400 mV, the smallest current amplification is observed in this uncovered spot. On the other hand, the DNA monolayer 231 Figure 8.8: SECM images obtained for DNA monolayers modified with NB at the top (A, C) and NB at the bottom (B, D). The tip was polarized to 1 V, and the scans were performed in 80 mM NaCl, 20 mM NaPi, pH= 7.2, 5 mM K4Fe(CN)6 buffer at a speed of 10 µm/s. The substrate bias was -400 mV (green in A, B) and 0 mV (blue in A, B). The switching ratios (the current obtained at -400 mV divided by the current obtained at 0 mV) are shown for monolayers with NB at the top (C) and monolayers with NB at the bottom (D). The poorly covered areas of the DNA monolayer are in the rear corners of A, B, C, and D. \ 232 modified with NB at the bottom exhibits low currents both at 0 mV and -400 mV over the well-covered areas of the scan. However, the poorly passivated region in the corner of the film with NB at the bottom (rear corner in Figure 8.8B, 8.8D) exhibits partial modulation when biased at -400 mV, hinting that the tip-generated Fe(CN)63- can access either some of the NB moieties or the bare gold surface. These observations demonstrate that the SECM allows for investigation of imperfect DNA films since it affords the ability to differentiate between catalytically active DNA spots and poorly passivated defects within the film. 8.3.4 Detection of TATA Binding Protein by SECM As confirmation of the DNA-mediated nature of this cycle, approach curves were recorded at DNA monolayers featuring NB at the top before and after addition of TBP, a transcription factor which bends the DNA by ~ 90º thereby attenuating DNA CT upon binding (12, 13). Figure 8.9 shows a schematic illustration of the sequence-specific detection of TBP during an SECM experiment. As previously described, the tip is polarized at a potential sufficient for direct oxidation of Fe(CN)64- while the potential of the substrate is varied. Prior to addition of TBP, the ferricyanide/NB catalytic cycle is “turned on” at a substrate bias of -400 mV resulting in positive feedback (Figure 8.10A). Upon addition of TBP to a monolayer containing its binding site (5’-TATA-3’), the ferricyanide/NB catalytic cycle is “turned off” at a substrate bias of -400 mV resulting in negative feedback (Figure 8.10D). On the other hand, the feedback observed at DNA monolayers lacking the binding site is unaffected by the addition of TBP (Figure 8.10B, 10C). Therefore, the sequence-specific distortion of the DNA spacer by TBP inhibits the DNA-mediated electrochemistry of Nile Blue. 233 Figure 8.9: Schematic illustration of the SECM-based, sequence-specific detection of TBP binding at DNA monolayers modified with NB. The addition of TBP in the presence of its binding site (right) attenuates the NB/ferricyanide catalytic cycle. Without the binding site for TBP, no protein binding occurs and DNA-mediated CT proceeds. 234 Figure 8.10: SECM approach curves taken for Nile Blue-DNA monolayers with (A, D) and without (B, C) the TBP binding site before (A, B) and after (C, D) addition of 100 nM TBP. Note that the addition of TBP eliminates the positive feedback observed at the DNA monolayer containing the 5’-TATA-3’ binding site but has no effect on the other sequence. Approach curves were obtained in 20 mM Na2HPO4, 80 mM NaCl, and 5 mM K4Fe(CN)6, pH= 7.2, at an approach speed of 1 µm/s. The tip was polarized at 1 V with the substrate polarized at -400 mV. The sequences were 5’-UGC GTG CTT TAT ATC TC-3’ and 5’-UGC GCG CCC GGC GCC TC-3’ where the italicized U indicates the location of the NB moiety. 235 Figure 8.11: Scanning electrochemical microscopy images of a Nile Blue-DNA monolayer with a substrate bias of -400 mV (yellow) and 0 mV (blue). The tip was polarized at 1 V, and the scan rate was 10 µm/s. Images were obtained in pH= 7.2, 20 mM Na2HPO4, 80 mM NaCl, and 5 mM K4Fe(CN)6 buffer before protein addition (A), after addition of 1 µM BSA (B), and after addition of 300 nM TBP (C). After addition of TBP, the scans are identical regardless of substrate bias (the plots in C completely overlap). 236 To further explore the analytical capabilities of the SECM, TBP binding at DNA monolayers was also investigated in imaging mode (Figure 8.11). Before addition of protein, a ~ 30 % increase in the steady-state current is observed over nearly the entire 85 µm by 85 µm scan area upon modulation of the substrate potential from 0 mV to -400 mV (note that the partially uncovered area in the rear corner exhibits the poorest current modulation in Figure 8.11A). Upon addition of 1 µM BSA to the solution, the steadystate tip current is slightly lowered (Figure 8.11B). However, there is no effect on the switching behavior of the substrate with the degree of modulation remaining at ~ 30 %, and these observations are consistent with some non-specific adsorption of BSA at the tip and substrate (12). Subsequently, the addition of 300 nM TBP then further lowers the steady-state tip current, presumably through non-specific adsorption (Figure 8.11C), and importantly, this transcription factor eliminates “switching” over nearly the entire scan area, indicating reliable detection of TBP binding in a localized area of the monolayer. These results provide strong support for the DNA-mediated nature of the Nile Blue/ferricyanide catalytic cycle. It is particularly striking that TBP binding eliminates positive feedback given that bending the DNA brings NB (and presumably ferricyanide) closer to the surface. In fact, upon binding, TBP should shield the negatively charged DNA and reduce the electrostatic repulsion of ferricyanide from the negatively charged monolayer. If significant direct charge transfer between the surface and tip generated ferricyanide were occurring (as observed by Bard and coworkers) (26), there would be an enhancement in the tip current at -400 mV upon addition of TBP, yet the opposite effect is observed here. 237 8.4 Discussion We have now utilized the SECM to demonstrate electrocatalytic signal amplification at DNA monolayers modified with NB. Positive feedback was observed for DNA monolayers modified with NB at the top and the substrate biased at potentials sufficiently negative to reduce NB. However, only negative feedback was observed for DNA monolayers modified with NB at the bottom, regardless of the substrate bias. These measurements have allowed us to extract an apparent heterogeneous rate constant for ferricyanide reduction by DNA-bound NB of kapp= 6 ± 3*10-2 cm/s, indicating that such probe-modified DNA monolayers may behave somewhat like conducting films (26). Furthermore, we have shown the sequence-specific interruption of the ferricyanide/NB electrocatalytic cycle upon addition of the TBP transcription factor: positive feedback is completely eliminated upon addition of this protein. Taken together, these data strongly indicate that the redox chemistry of NB is DNA-mediated. Previous studies (22, 24, 59, 60) have indicated that the individual duplexes within dense DNA SAMs stand upright at potentials negative of the PZC, particularly in the presence of an alkanethiol underlayer (61, 62). These findings were hinted by our observations for the electrochemistry of MB at dense DNA monolayers, where only the MB constrained to the periphery of the film was available for the MB/ferricyanide catalytic cycle (8, 46). The present results for sparse films are fully consistent with an upright orientation of the DNA. By fixing the location of the NB covalent probe, we have conclusively demonstrated that only those monolayers which feature NB at the top of the 238 film and close to the film/solution interface can mediate catalytic ferrocyanide recycling at the probe tip. The images obtained over micron areas of DNA monolayers modified with NB are consistent with the approach curves in feedback mode. Notably, the imaging mode of the SECM allows for the reliable investigation of imperfect DNA monolayers in multiple areas of the substrate, making it an attractive complement to bulk voltammetric techniques. In addition, the SECM can distinguish between areas with and without DNA by carefully monitoring the ratio of the currents at various substrate biases. Therefore, unlike substrate voltammetry, the SECM ensures that the observed electrochemical response is due solely to DNA-mediated electrochemistry. This feature of SECM allows for the highly reliable detection of protein binding, minimizing the possibility of false positives that can arise in experiments involving only substrate voltammetry. As a surface characterization technique, the SECM also provides important information which is intermediate between centimeter scale electrochemical measurements and nanometer scale AFM/STM measurements. For example, an SECM tip with a 10 µm footprint will access attomoles of DNA molecules (for uniform surface coverages of 1–3 pmol/cm2), thereby simultaneously sampling a large number of molecules (relative to an AFM/STM) and avoiding overwhelming contributions from defects in the monolayer (relative to substrate electrochemistry). In addition, assuming a 1:1 DNA/protein stoichiometry, we are consequently able to detect attomole quantities of the TBP transcription factor with the SECM, underscoring the potential of this methodology for highly sensitive, sequence-specific detection. Therefore, the scale and sensitivity of the instrument makes the SECM inherently compatible with multiplexed 239 technologies such as DNA microarrays, indicating that it is a potential alternative to fluorescence-based microarray technologies. Overall, these results indicate that the investigation of DNA-mediated electrochemistry with scanning electrochemical microscopy is a powerful and flexible methodology for the localized investigation of DNA monolayers. Here, by backfilling with a negatively charged alkanethiol and coupling NB with ferricyanide in a catalytic cycle, we have ensured that the DNA duplexes adopt an upright orientation and that DNA CT dominates the electrochemical response observed by SECM. In fact, if non-specific protein adsorption at the tip can be minimized, the micron size and reliability of the SECM could lead to sensitive monitoring of DNA/protein interactions in a multiplexed format. 240 REFERENCES (1) Boon, E. M., Barton, J. K. (2002) Curr. Opin. Struct. Biol. 12, 320–329. (2) Delaney, S., Barton, J. K. (2003) J. Org. Chem. 68, 6475–6483. (3) O’Neill, M. A., Barton, J. K. (2005) Charge Transfer in DNA: From Mechanism to Application (H.-A. Wagenknecht), pp. 27–75, Wiley-VCH, Hoboken. (4) Giese, B. (2000) Acc. Chem. Res. 33, 631–636. (5) Schuster, G. B. (2000) Acc. Chem. Res. 33, 253–260. (6) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nature Biotechnol. 21, 1192–1199. (7) Odenthal, K. J., Gooding, J. J. (2007) Analyst 132, 603–610. (8) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucleic Acids Res. 27, 4830–4837. (9) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (10) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. (11) Boal, A. K., Barton, J. K. (2005) Bioconjugate Chem. 16, 312–321. (12) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nature Biotechnol. 20, 282– 286. (13) Gorodetsky, A. A., Ebrahim, A., Barton, J. K. (2008) J. Am. Chem. Soc. 130, 2924–2925. 241 (14) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U. S. A. 100, 12543–12547. (15) Boal, A. K., Yavin, E., Lukianova, O. A., O’Shea, V. L., David, S. S., Barton, J. K. (2005) Biochemistry 44, 8397–8407. (16) Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. (17) DeRosa, M. C., Sancar, A., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 10788–10792. (18) Gorodetsky, A. A., Dietrich, L. E. P., Lee, P. E., Demple, B., Newman, D. K., Barton, J. K. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 3684–3689. (19) D. M. Ceres, J. K. Barton. (2003) J. Am. Chem. Soc. 125, 14964–14965. (20) Hihath, J., Xu, B., Zhang, P., Tao, N. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 16979–16983. (21) Wierzbinski, E., Arndt, J., Hammond, W., Slowinski, K. (2006) Langmuir 22, 2426–2429. (22) Kelley, S. O., Barton, J. K., Jackson, N. M., McPherson, L. D., Potter, A. B., Spain, E. M., Allen, M. J., Hill, M. G. (1998) Langmuir 14, 6781–6784. (23) Zhou, D., Sinniah, K., Abell, C., Rayment, T. (2002) Langmuir 18, 8278– 8281. (24) Erts, D., Polyakov, B., Olin, H., Tuite, E. (2003) J. Phys. Chem. B. 107, 3591– 3597. (25) Turcu, F., Schulte, A., Hartwich, G., Schuhmann, W. (2004) Angew. Chem. Int. Ed. 43, 3482–3485. 242 (26) Liu, B., Li, C.-Z., Kraatz, H.-B., Bard, A. J. (2005) J. Phys. Chem. B. 109, 5193–5198. (27) Wang, K., Goyer, C., Anne, A., Demaille, C. (2007) J. Phys. Chem. B. 111, 6051–6058. (28) Wain, A. J., Zhou, F. (2008) Langmuir 24, 5155–5160. (29) Stoica, L., Neugebauer, S., Schuhmann, W. (2008) Adv. Biochem. Eng. Biot. 109, 455–492. (30) Bard, A. J., Mirkin, M. V., eds. (2001) Scanning Electrochemical Microscopy, Marcel Dekker: New York. (31) Wang, J., Zhou, F. (2002) J. Electroanal. Chem. 537, 95–102. (32) Komatsu, M., Yamashita, K., Uchida, K., Kondo, H., Takenaka, S. (2006) Electrochim. Acta 51, 2023–2029. (33) Lie, L. H., Mirkin, M. V., Hakkarainen, S., Houlten, A., Horrocks, B. R. (2007) J. Electroanal. Chem. 603, 67–80. (34) Palchetti, I., Laschi, S., Marrazza, G., Mascini, M. (2007) Anal. Chem. 79, 7206–7213. (35) Wong, E. L. S., Gooding, J. J. (2006) Anal. Chem. 78, 2138–2144. (36) Li, X., Song, H., Nakatani, K., Kraatz, H.-B. (2007) Anal. Chem. 79, 2552– 2555. (37) Wong, E. L. S., Gooding, J. J. (2007) J. Am. Chem. Soc. 129, 8950–8951. (38) Di Giusto, D. A., Wlasoff, W. A., Giesebrecht, S., Gooding, J. J., King, G. C. (2004) J. Am. Chem. Soc. 126, 4120–4121. 243 (39) Kelley, S. O., Jackson, N. M., Hill, M. G., Barton, J. K. (1999) Angew. Chem., Int. Ed. 38, 941–945. (40) Drummond, T. G., Hill, M. G., Barton, J. K. (2004) J. Am. Chem. Soc. 126, 15010–15011. (41) Liu, T., Barton, J. K. (2005) J. Am. Chem. Soc. 127, 10160–10161. (42) Gorodetsky, A. A., Green, O., Yavin, E., Barton, J. K. (2007) Bioconjugate Chem. 18, 1434–1441. (43) Okamoto, A., Kamei, T., Tanaka, K., Saito, I. (2004) J. Am. Chem. Soc. 126, 14732–14733. (44) Okamoto, A., Kamei, T., Saito, I. (2006) J. Am. Chem. Soc. 128, 658–662. (45) Inouye, M., Ikeda, R., Takase, M., Tsuri, T., Chiba, J. (2005) Proc. Natl. Acad. Sci. U. S. A. 102, 11606–11610. (46) Boon, E. M., Barton, J. K., Bhagat, V., Nersissian, M., Wang, W., Hill, M. G. (2003) Langmuir 19, 9255–9259. (47) Zhang, J., Kirkham, J., Robinson, C., Wallwork, M. L., Smith, D. A., Marsh, A., Wong, M. (2000) Anal. Chem. 72, 1973–1978. (48) Liu, H.-H., Lu, J.-L., Zhang, M., Pang, D.-W. (2002) Anal. Sci. 18, 1339– 1344. (49) Ju, H., Ye, Y., Zhu, Y. (2005) Electrochim. Acta 50, 1361–1367. (50) Schlereth, D. D., Schmidt, H.-L. (1995) J. Electroanal. Chem. 380, 117–125. (51) Sugawara, K., Yamauchi, Y., Hoshi, S., Akatsuka, K., Yamamoto, F., Tanaka, S., Nakamura, H. (1996) Bioelectrochem. Bioenerg. 41, 167–172. 244 (52) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. (53) Mirkin, M.V., Arca, M., Bard, A. J. (1993) J. Phys. Chem. 97, 10790–10795. (54) Wei, C., Bard, A.J., Mirkin, M.V. (1995) J. Phys. Chem. 99, 16033–16042. (55) Lefrou, C. (2007) J. Electroanal. Chem. 601, 94–100. (56) Steel, A. B., Herne, T. M., Tarlov, M. J. (1998) Anal. Chem. 70, 4670–4677. (57) Yu, H.-Z., Luo, C.-Y., Sankar, D. G., Sen, D. (2003) Anal. Chem. 75, 3902– 3907. (58) Ceres, D. M., Udit, A. K., Hill, H. D., Hill, M. G., Barton, J. K. (2007) J. Phys. Chem. B. 111, 663–668. (59) Rant, U., Arinaga, K., Fujita, S., Yokoyama, N., Abstreiter, G., Tornow, M. (2004) Nano Lett. 4, 2441–2445. (60) Rant, U., Arinaga, K., Fujita, S., Yokoyama, N., Abstreiter, G., Tornow, M. (2004) Langmuir 20, 10086–10092. (61) Arinaga, K., Rant, U., Tornow, M., Fujita, S., Abstreiter, G., Yokoyama, N. (2006) Langmuir 22, 5560–5562. (62) Yang, X., Wang, Q., Wang, K., Tan, W., Yao, J., Li, H. (2006) Langmuir 22, 5654–5659. (63) Gore, M. R., Szalai, V. A., Ropp, P. A., Yang, I. V., Silverman, J. S., Thorp, H. H. (2003) Anal. Chem. 75, 6586–6592. 245  Chapter 9: Electrical Detection of TATABinding Protein at DNA-Modified Microelectrodes Adapted from: Gorodetsky, A. A., Ebrahim, A. E., Barton, J. K. (2008) J. Am. Chem. Soc. 130, 2924–2925. 246 ABSTRACT A simple method for the electrochemical detection of TATA-binding protein is demonstrated at DNA-modified microelectrodes. The assay is general and based on the interruption of DNA-mediated charge transport to Nile Blue, a redox-active probe covalently attached to the DNA base pair stack. Nanomolar quantities of TATA binding protein can be detected on the microelectrodes even in the presence of micromolar amounts of bovine serum albumin, EndonucleaseIII, or Bam HI methyltransferase. The scheme outlined provides a basis for the sensitive electrical detection of numerous proteins on a single DNA chip. 247 9.1 Introduction Sensitive methods to detect the specific binding of proteins to DNA are essential in monitoring pathways for gene expression and in the development of new diagnostics of disease states (1, 2). There are several high-throughput, array-based methodologies currently available for monitoring transcription factor binding to DNA including immunoprecipitation of reversibly crosslinked chromatin/protein complexes (3, 4), capture of fluorescent antibodies by epitope-tagged DNA-binding proteins (5, 6), and amplification of sequences methylated by enzyme/DNA adenine-methyltransferase fusions (7, 8). However, these methods rely on fluorescence detection and require DNA amplification and/or protein labeling. Electrochemical assays based on DNA-mediated charge transport (CT) through self-assembled DNA monolayers offer an alternative approach (9). DNA electrochemistry has been useful not only in assays to detect mutations and lesions (10– 17) but also for the detection of protein binding at self-assembled DNA monolayers (18– 25). Moreover, DNA electrochemistry provides the opportunity for highly sensitive detection in a multiplexed format (9). However, although we have demonstrated the detection of all possible single base pair mismatches at gold microelectrodes (10), there are few reports of the electrical detection of protein binding utilizing DNA monolayers at small scales (9, 18–25). In a typical protein detection experiment, a loosely packed DNA monolayer with a covalently appended, electroactive probe is self-assembled in the absence of Mg2+ and the surface is backfilled with an alkanethiol such as mercaptohexanol (9, 18, 22). Such 248 sparse DNA films are necessary for improved accessibility of the protein and possess a well-defined morphology; the DNA oriented in an upright orientation (26, 27). After self assembly and backfilling, voltammetry of the DNA-modified electrode is recorded in the absence and presence of protein. Detection of protein binding is confirmed by an attenuation of the yield of DNA CT to a distally-bound probe, such as DM. Overall, this electrochemical assay is highly general and based on exquisitely sensitive DNA-mediated electrochemistry, so a wide range of protein/DNA interactions and binding motifs can be interrogated in rapid fashion. Our previous studies at DNA-modified macroelectrodes demonstrated that we could easily detect micromolar concentrations of TATA-binding protein (TBP), a transcription factor that bends duplex DNA by ~ 90º upon binding and attenuates DNA CT (18, 28–30). Importantly, this transcription factor not only plays a central role in the eukaryotic/prokaryotic transcription machinery but also possesses an extensively explored crystal structure and a simple 5’-TATAAAG-3’ binding site (30). Therefore, TBP was chosen as the model system for our explorations at DNA-modified microelectrodes. Here, we demonstrate the rapid and specific detection of nanomolar concentrations of TBP at microelectrodes modified with DNA featuring Nile Blue (NB), an easily prepared and stable redox probe. 249 9.2 Experimental 9.2.1 Materials All reagents for DNA synthesis were purchased from Glen Research. Mercaptohexanol and buffer preparation reagents were purchased from Sigma in the highest purity available and used as received. Nile Blue perchlorate was purchased from Acros in laser grade purity. Gold microelectrodes were purchased from Bioanalytical Systems Inc., and silver/silver chloride reference electrodes were purchased from Cypress Systems, Inc. Nanocloth and aluminum oxide slurry were purchased from Buehler, Inc. Sodium phosphate buffers were prepared with Milli-Q water and pH adjustments were made with sodium hydroxide, if necessary. 9.2.2 Synthesis of Thiol and Nile Blue-Modified DNA Oligonucleotides were prepared on solid support using standard phosphoramidite chemistry on an Applied Biosystems 394 DNA synthesizer. The pure oligonucleotides were characterized with matrix-assisted laser desorption (MALDI) mass spectrometry, UV-visible (UV-Vis) spectrophotometry, and multiple rounds of high performance liquid chromatography (HPLC). Thiol-terminated oligonucleotides were synthesized according to established protocols from Glen Research, Inc. using the C6 S-S thiol modifier. After deprotection and cleavage from solid support with ammonium hydroxide (60 ºC for 8 hours), the disulfide containing DNA was purified by HPLC. The disulfide was subsequently reduced with an excess of dithiothreitol in ammonium acetate buffer at pH = 8, and the 250 free thiol containing single-stranded DNA was then purified with a second round of HPLC. DNA modified with NB at the 5’ terminus was prepared according to ultra-mild protocols (Glen Research, Inc.) to avoid degradation of the NB moiety. Additionally, Pac-protected bases and ultra mild reagents were utilized during the synthesis to prevent undesirable capping of the protecting groups. A 17-mer sequence (5’-UGC GTG CTT TAT ATC TC-3’) was prepared on solid support with a 5-[3-acrylate NHS Ester]-deoxy uridine as the terminal 5’ base. The beads were then removed from the synthesizer and dried thoroughly. The solid support were reacted with a 10 mg/mL Nile Blue perchlorate solution in either 9:1 N,N-dimethylformamide/N,N-diisopropylethylamine or 9:1 dichloromethane/N,N-diisopropylethylamine for 12–48 hours. The beads were subsequently washed up to three times with dichloromethane or N,N-dimethylformamide, methanol, and acetonitrile. Subsequently, the Nile Blue-containing sequence was simultaneously cleaved from the support and deprotected with 0.05 M potassium carbonate in methanol at room temperature for 12–14 hours. The overall yields of the reaction ranged from 30% to 80%. Both thiol- and NB-modified DNA was quantified using the extinction coefficient of the single-stranded DNA at 260 nm on a Beckman UV-Vis spectrophotometer. Equimolar amounts of each strand were combined before the resulting solution was purged with argon. Duplexes were formed by thermally annealing in deoxygenated phosphate storage buffer containing 5 mM NaPi, pH = 7.1, 50 mM NaCl to 90 ºC, followed by cooling to ambient temperature. 251 9.2.3 Protein Preparation and Storage Human TATA binding protein (TBP) was custom ordered from ProteinOne, Inc. in pH = 7, 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10 % glycerol buffer. Due to conflicting concentration reports from ProteinOne, Inc., the protein stock concentration (~ 16 µM) was determined using the Bradford Assay Kit from Bio-Rad Laboratories. Stock solutions of the protein were stored at -80 ºC and aliquoted under an argon atmosphere. The TATA binding protein utilized for detection experiments was taken through no more than two freeze/thaw cycles (including the aliquoting step). In addition to TBP, bovine serum albumin (BSA), Endonuclease III (Endo III), and Bam HI methyltransferase (Bam HI) were utilized for electrochemistry experiments. Bovine serum albumin was purchased from New England Biolabs, Inc., and utilized for experiments as received. Bam HI Methyltransferase was purchased from New England Biolabs, Inc. Prior to electrochemistry experiments, Bam HI was dialyzed overnight in pH = 7, 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10 % glycerol buffer to remove excess dithiothreitol. Endo III was prepared according to established literature procedures and quantified by UV-visible spectroscopy (20). 9.2.4 Preparation of Backfilled DNA Monolayers Gold microelectrodes possessing either a 10 µm or a 25 µm diameter were polished in 0.05 µm alumina slurry on nanocloth. The electrodes were subsequently briefly polished in H2O on nanocloth, rinsed thoroughly in H2O, and washed with 1 N sulfuric acid to remove residual alumina particles. The electrodes were then sonicated in deionized water. Finally, to remove gold oxide from the exposed surface, the 252 microelectrodes were etched in 1 N H2SO4 before being briefly rinsed in deionized water and phosphate storage buffer. Immediately after cleaning, the microelectrodes were incubated in a 25–50 µM duplex NB-DNA solution in phosphate storage buffer. Monolayer formation was allowed to proceed in a humidified environment for a period of 24–48 hours. Upon completion of film formation, the cell was rinsed thoroughly with phosphate buffer to remove residual DNA before the DNA-modified surface was backfilled with 1.0 mM 1-mercaptohexanol in phosphate storage buffer/5% glycerol for 30–60 minutes or 0.1 mM 11mercaptoundecylphosphoric acid in phosphate buffer/25% glycerol for 5–10 minutes. Alkanethiols rapidly displace DNA from the electrode surface, so the microelectrodes were rinsed thoroughly in buffer to remove residual 1-mercaptohexanol or 11mercaptoundecylphosphoric acid. 9.2.5 Electrochemistry and Protein Detection Experiments Cyclic and square wave voltammetry experiments were recorded using a CH Instruments electrochemical analyzer (Austin, TX). When necessary, the data was smoothed for clarity using the CH instruments software. Experiments were performed at ambient temperature in pH 7 phosphate buffer containing either 5 mM NaPi and 50 mM NaCl (standard phosphate buffer) or in pH 7 phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol (TBP dialysis buffer) under an Argon atmosphere. An eppendorf tube electrochemical cell was used for all experiments with a Pt auxiliary electrode, a gold working electrode, and a silver/silver chloride (Ag/AgCl) reference electrode. In a typical detection experiment, background scans were initially recorded in TBP dialysis buffer, which was thoroughly purged with 253 Argon. Great care was taken to ensure that similar degassing times were utilized in all experiments. Subsequently, TBP or other proteins were added to the electrochemical cell from a stock solution in phosphate buffer at pH = 7 containing 5 mM KPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol. The solution was equilibrated via gentle pipetting, stirred with a brief argon purge, and voltammetry was recorded again. 9.3 Results Figure 9.1 schematically illustrates the strategy utilized for electrochemistry experiments, and Figure 9.2 shows the corresponding cyclic voltammetry (CV) and square wave voltammetry (SWV) for electrodes modified with NB-DNA duplexes in TBP dialysis buffer. For macroelectrodes, the CV is reversible with a midpoint potential of -220 mV (±50) versus Ag/AgCl. We have found that the observed midpoint potentials of NB-DNA monolayers vary somewhat depending upon the type of gold surface utilized and the buffer conditions, leading to an uncertainty of ~ 50 mV in the potentials, although they have a higher certainty for any given surface. In addition, the cathodic and anodic waves are symmetric (though not overlapping), and a plot of peak current as a function of scan rate is linear, as expected for a surface-bound species (31). For microelectrodes with critical dimensions ≥ 25 µm, the CV differs in shape from that observed for a macroelectrode in TBP dialysis buffer. For example, in pH = 7 phosphate buffer containing only 5 mM NaPi and 50 mM NaCl, the microelectrode voltammetry more closely resembles that typically found at a macroelectrode (Figure 9.3). 254 Figure 9.1: Schematic illustration of addition of TBP to a DNA monolayer modified with Nile Blue (NB), followed by DNA kinking and attenuation of DNA-mediated electrochemistry 255 Figure 9.2: CV (top) and SWV (bottom) of NB-DNA at macroelectrodes of 2 mm diameter (left) and microelectrodes of 25 µm diameter (right). Voltammetry without TBP (blue) and with 300 nM TBP (red) is shown at pH 7 in a degassed phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol. The sequence is thiol-5'-GAGATATAAAGCACGCA-3' plus NB-modified complement. The potentials are reported versus Ag/AgCl with a CV scan rate of 50 mV/s and an SWV frequency of 15 Hz. 256 A. B. C. D. Figure 9.3: CV of a NB-DNA modified microelectrode of a 25 µm diameter for two different buffers. Voltammetry in a pH 7 degassed phosphate buffer containing 5 mM NaPi and 50 mM NaCl, is shown in (A). The electrode is moved to a pH 7 degassed phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol with the recorded voltammetry shown in (B). The electrode is subsequently backfilled with 11-mercaptoundecylphosphoric acid. Voltammetry of the backfilled electrode in a pH 7 degassed phosphate buffer containing 5 mM NaPi and 50 mM NaCl, is shown in (C). The subsequent voltammetry in a pH 7 degassed phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol is shown in (D). The sequence is thiol-5'- GAGATATAAAGCACGCA-3' plus NB-modified complement. The potentials are reported versus Ag/AgCl with a CV scan rate of 50 mV/s. 257 The more sigmoidal voltammetry can be restored simply by moving the microelectrode from standard phosphate buffer to TBP dialysis buffer. In fact, similar buffer effects have been previously documented for the electrochemistry ferrocene-peptide conjugates at microelectrodes (32). Nonetheless, regardless of the buffer conditions utilized, the observed cyclic voltammetry at microelectrodes is different from the nearly ideal voltammograms found at macroelectrodes. Nile Blue has displayed catalytic activity with a number of substrates (33), and we attribute this phenomenon to a small catalytic contribution to the NB signal from trace oxygen (Figure 9.4) (34). For microelectrodes, the midpoint potential appears to be systematically shifted by ~ 100 mV. Interestingly, the potential shifts we observe here upon moving from the macro to the micro scale are comparable to those previously observed by others (35, 36). However, this shift in potential persists even when DNA modified macroelectrodes and microelectrodes are fabricated in parallel on a single substrate and run minutes apart with a common reference (Figure 9.5). The potential shift may reflect the different coverages and therefore geometries of DNA monolayers on microelectrodes, as indicated by both radioactive labeling and ruthenium hexammine assay (36). Upon addition of 300 nM TBP, as illustrated in Figure 9.1 and the red traces of Figure 9.2, reduction of DNA-bound NB is dramatically attenuated at both the macroand microscales. For macroelectrodes, addition of the protein causes a signal loss of 60% with the signal slowly decreasing over a period of 15 minutes. For microelectrodes, 100% signal loss occurs in less than 30 seconds. The consistently greater signal loss and appreciably faster equilibration on the microelectrode may reflect better access of the protein to the NB-modified DNA on the smaller surface. It is also noteworthy that 258 Figure 9.4: The effect of oxygen on the CV of a NB-DNA-modified microelectrode of a 10 µm diameter. Voltammetry in an oxygen saturated pH 7 phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol is shown in black with voltammetry after ten to fifteen more minutes of vigorous argon bubbling into the solution in blue. Trace amounts of oxygen have a significant effect on the voltammetry observed at microelectrodes. The gray-to-orange traces represent successive voltammograms taken on the same electrode open to air in an initially degassed solution with a slow, steady increase observed over thirty minutes. The sequence is thiol-5'-GAGATATAAAGCACGCA-3' plus NB-modified complement. The potentials are reported versus Ag/AgCl with a CV scan rate of 50 mV/s. 259 Figure 9.5: SWV of a NB-DNA-modified Au on silicon electrodes of 130 µm radius (A) and 13 µm radius (B) in pH = 7 phosphate buffer containing 5 mM Pi and 50 mM NaCl. Note that the two electrodes were fabricated together in parallel and run minutes apart yet still did not yield identical potentials. The GAGATATAAAGCACGCA-3' plus NB-modified complement. 260 sequence is thiol-5'- binding of TBP, in kinking the DNA (19), may bring NB in closer proximity to the surface, yet a decrease is observed in the NB signal. This observation is fully consistent not with contact of NB to the surface but instead a DNA-mediated reduction that is inhibited by bending the intervening DNA. TBP binding and the resulting signal attenuation are fully reversible on a microelectrode (Figure 9.6). A microelectrode incubated with TBP can be washed in 1 M KCl, promoting protein dissociation (37, 38) and rapid recovery of the NB signal. Electrodes can be taken through this cycle up to three times with no loss of response. In addition, detection of low concentrations of TBP is easily achieved (Figure 9.7). At 30 and 15 nM protein concentrations, complete NB signal loss is observed. At 3 nM protein, the NB signal is partially attenuated, as expected based upon the affinity of the protein for its target sequence (39, 40). The detection event is furthermore specific to TBP (Figure 9.8). Bovine serum albumin (BSA), a protein that does not bind DNA, has little effect on DNA-mediated electrochemistry or on signal attenuation associated with TBP binding (18). BSA is known to adsorb nonspecifically on self-assembled alkanethiol monolayers (41). Treatment of the electrode with BSA effectively passivates against further non-specific adsorption. The inclusion of micromolar Endonuclease III (EndoIII), a glycosylase that targets damaged pyrimidines (not unlike the NB-modified thymine), also has little effect on the DNA-mediated NB signal (42–44). Finally, the addition of BamHI methyltransferase (BAM), a base-flipping methylase that targets 5'-GGATCC-3' (45), yields little change in the voltammetry or the signal decrease associated with TBP binding. Although EndoIII and BAM likely bind nonspecifically to the DNA film at these 261 Figure 9.6: CV of a NB-DNA-modified 10 µm diameter electrode before addition of TBP in blue and after addition of 300 nM TBP in red. Initial background voltammetry of the electrode is shown in pH 7 phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol (A) with significant signal loss observed after addition of TBP. The electrode is subsequently rinsed in succession (5 minutes each) with 1 M KCl in MQ H2O, 1 M KCl in Tris buffer, and TATA dialysis buffer. Background voltammograms are recorded for the same electrode before it is used to detect TBP again (B). The electrode can then be taken through another KCl rinse cycle before being used for detection a third time (C). 262 Figure 9.7: CV of a NB-DNA-modified 25 µm diameter electrode before addition of TBP in blue and after addition of TBP in red. Initial background voltammetry of the electrode is shown in pH 7 phosphate buffer containing 5 mM NaPi, 50 mM NaCl, 4 mM MgCl2, 4 mM spermidine, 50 µM EDTA, 10% glycerol. Significant signal attenuation is found after addition of (A) 3 nM, (B) 15 nM, and (C) 30 nM TBP. The sequence is thiol5'-GAGATATAAAGCACGCA-3' plus NB-modified complement, and the potentials are reported versus Ag/AgCl with a CV scan rate of 50 mV/s. 263 Figure 9.8: CV of three NB-DNA-modified microelectrodes (10 µm) before and after addition of proteins. CVs before protein addition are in blue. From left to right, the black CVs represent addition of BSA (2 µM); BSA (1 µM) + EndoIII (1 µM); and BSA (1 µM) + BAM (1 µM). CVs in red reflect the subsequent addition of 30 nM TBP to each electrode. The sequence is thiol-5'-GAGATATAAAGCACGCA-3' plus NB-modified complement, and the potentials are reported versus Ag/AgCl with a CV scan rate of 50 mV/s. 264 concentrations, they do not inhibit specific binding by TBP. These results are therefore consistent with observing a sequence-specific TBP binding event. 9.4 Implications We have thus demonstrated the electrochemical detection of nanomolar TBP binding to DNA in the presence of other proteins with a rapid and sensitive electrochemical assay. This methodology is based upon an interruption of DNA-mediated electrochemistry associated with protein binding and therefore may be generally applied in detecting other proteins that perturb the DNA base stack. Our ability to detect protein binding to DNA reliably at small scales provides a foundation for rapid, electrical assays of DNA-binding proteins on a single chip. 265 REFERENCES (1) Sikder, D., Kodadek, T. (2005) Curr. Opin. Chem. Biol. 9, 38–45. (2) Bulyk, M. L. (2006) Curr. Opin. Biotechnol. 17, 422–430. (3) Kim, J., Iyer, V. R. (2004) Mol. Cell. Biol. 24, 8104–8112. (4) Lee, T. I., Johnstone, S. E., Young, R. A. (2006) Nature Protoc. 1, 729–748. (5) Berger, M. F., Philippakis, A. A., Qureshi, A. M., He, F. S., Estop, P. W., III, Bulyk, M. L. (2006) Nature Biotechnol. 24, 1429–1435. (6) Bulyk, M. L. (2007) Adv. Biochem. Engin./Biotechnol. 104, 65–85. (7) van Steensel, B., Delrow, J., Henikoff, S. (2001) Nature Genet. 27, 304–308. (8) Vogel, M. J., Peric-Hupkes, D., van Steensel, B. (2007) Nature Protoc. 2, 1467–1478. (9) Drummond, T. G., Hill, M. G., Barton, J. K. (2003) Nature. Biotechnol. 21, 1192–1199. (10) Boon, E. M., Ceres, D. M., Drummond, T. G., Hill, M. G., Barton, J. K. (2000) Nature Biotechnol. 18, 1096–1100. (11) Kelley, S. O., Boon, E. M., Barton, J. K., Jackson, N. M., Hill, M. G. (1999) Nucl. Acids Res. 27, 4830–4837. (12) Gorodetsky, A. A., Barton, J. K. (2006) Langmuir 22, 7917–7922. (13) Boal, A. K., Barton, J. K. (2005) Bioconjugate Chem. 16, 312–321. (14) Inouye, M., Ikeda, R., Takase, M., Tsuri, T., Chiba, J. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 11606–11610. (15) Okamoto, A., Kamei, T., Saito, I. (2006) J. Am. Chem. Soc. 128, 658–662. 266 (16) Li, X., Song, H., Nakatani, K., Kraatz, H.-B. (2007) Anal. Chem. 79, 2552– 2555. (17) Wong, E. L. S., Gooding, J. J. (2007) J. Am. Chem. Soc. 129, 8950–8951. (18) Boon, E. M., Salas, J. E., Barton, J. K. (2002) Nature Biotechnol. 20, 282– 286. (19) Boon, E. M., Livingston, A. L., Chmiel, N. H., David, S. S., Barton, J. K. (2003) Proc. Natl. Acad. Sci. U.S.A. 100, 12543–12547. (20) Boal, A. K.; Yavin, E.; Lukianova, O. A.; O'Shea, V. L.; David, S. S.; Barton, J. K. (2005) Biochemistry 44, 8397–8407. (21) Gorodetsky, A. A., Boal, A. K., Barton, J. K. (2006) J. Am. Chem. Soc. 128, 12082–12083. (22) DeRosa, M. C., Sancar, A., Barton, J. K. (2005) Proc. Natl. Acad. Sci. U.S.A. 102, 10788–10792. (23) Yeung, S. S. W., Lee, T. M. H., Hsing, I.-M. (2006) J. Am. Chem. Soc. 128, 13374–13375. (24) Anne, A., Bonnaudat, C., Demaille, C., Wang, K. (2007) J. Am. Chem. Soc. 129, 2734–2735. (25) Li, C.-Z., Long, Y.-T., Lee, J. S., Kraatz, H.-B. (2004) Chem. Comm. 5, 574– 575. (26) Zhou, D., Sinniah, K., Abell, C., Rayment, T. (2003) Angew. Chem. Int. Ed. 42, 4394–4397. (27) Erts, D., Polyakov, B., Olin, H., Tuite, E. (2003) J. Phys. Chem. B. 107, 3591– 3597. 267 (28) Kim, Y., Geiger, J. H., Hahn, S. H., Sigler, P. B. (1993) Nature 365, 512–520. (29) Kim, J. L., Nikolov, D. B., Burley, S. K. (1993) Nature 365, 520–527. (30) Patikoglou, G. A., Kim, J. L., Sun, L., Yang, S.-H., Kodadek, T. K., Burley, S. K. (1999) Genes Dev. 13, 3217–3230 (31) Bard, A. J., Faulkner, L. R. (2001) Electrochemical Methods, 2nd ed., John Wiley & Sons: New York. (32) Orlowski, G. A., Chowdhury, S., Kraatz, H.-B. (2007) Electrochim. Acta 53, 2034–2039. (33) Liu, H.-H., Lu, J.-L., Zhang, M., Pang, D.-W. (2002) Anal. Sci. 18, 1339– 1344. (34) Ju, H., Shen, C. (2001) Electroanalysis 13, 789–793. (35) Gore, M. R., Szalai, V. A., Ropp, P. A., Yang, I. V., Silverman, J. S., Thorp, H. H. (2003) Anal. Chem. 75, 6586–6592. (36) Lapierre-Devlin, M. A., Asher, C. L., Taft, B. J., Gasparac, R., Roberts, M. A., Kelley, S. O. (2005) Nano Lett. 5, 1051–1055. (37) Daugherty, M. A., Brenowitz, M., Fried, M. G. (1999) J. Mol. Biol. 285, 1389–1399. (38) Khrapunov, S., Brenowitz, M. (2007) Biochemistry 46, 4876–4887. (39) Coleman, R. A., Taggart, A. K. P., Benjamin, L. R., Pugh, B. F. (1995) J. Biol. Chem. 270, 13842–13849. (40) Gilfillan, S., Stelzer, G., Piaia, E., Hofmann, M. G., Meisterernst, M. (2005) J. Biol. Chem. 280, 6222–6230. 268 (41) Holmlin, R. E., Chen, X., Chapman, R. G., Takayama, S., Whitesides, G. M. (2001) Langmuir 17, 2841–2950. (42) Asahara, H., Wistort, P. M., Bank, J. F., Bakerian, R. H., Cunningham, R. P. (1989) Biochemistry 28, 4444–4449. (43) Cunningham, R. P., Asahara, H., Bank, J. F., Scholes, C. P., Salerno, J. C., Surerus, K., Munck, E., McCracken, J., Peisach, J., Emptage, M. H. (1989) Biochemistry 28, 4450–4455. (44) David, S. S., Williams, S. D. (1998) Chem. Rev. 98, 1221–1262. (45) Lindstrom, W. M., Jr., Malygin, E. G., Ovechkina, L. G., Zinoviev, V. V., Reich, N. O. (2003) J. Mol. Biol. 325, 711–720. 269