Bioconjugate Chem. 1997, 8, 31−37
31
Electrochemistry of Methylene Blue Bound to a DNA-Modified Electrode Shana O. Kelley and Jacqueline K. Barton* Beckman Institute, Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125
Nicole M. Jackson and Michael G. Hill* Department of Chemistry, Occidental College, Los Angeles, California 90041. Received August 1, 1996X
Gold surfaces have been derivatized with 15-base-pair double-stranded DNA oligonucleotides containing a pendant 5′ hexanethiol linker. The electrochemistry of intercalated methylene blue has been investigated at these modified electrodes. Chronocoulometry, cyclic voltammetry, ellipsometry, and quantitation via 32P labeling are all consistent with a surface coverage of g75% with the DNA helices stacked at an angle from the electrode surface. Cyclic voltammetry at low methylene blue/ duplex stoichiometries yields well-behaved surface waves with E° ) -0.25 V (vs SCE), a value 0.03 V negative of that in aqueous solution. A binding isotherm for methylene blue at an electrode derivatized with the double-stranded sequence 5′ SH-(CH2)6-p-AGTACAGTCATCGCG 3′ was obtained from coulometric titrations and gave an affinity constant equal to 3.8(5) × 106 M-1 with a saturation value of 1.4(2) methylene blue intercalators per DNA duplex. Taken together, these experiments support a model for the surface morphology in which DNA duplexes are densely packed; methylene blue therefore reversibly binds to sites in the DNA that are close to the bulk solution. Electrochemistry at DNA-derivatized electrodes provides a valuable methodology to examine DNA-bound redox reactions and may offer new insight into DNA-mediated electron transfers.
The π-stack of aromatic heterocycles contained within the DNA helix presents a unique medium in which to explore electron-transfer reactions (1). As the hydrophobic interior of DNA differs substantially from aqueous solution, the redox properties of molecules intercalatively bound (and of the nucleic acid bases themselves) may vary significantly from those in solution. Redox reactions in the π-stack are particularly important for understanding charge delocalization in DNA and its effect on base damage (2-4). However, the irreversible electrochemistry of the nucleic acid bases has made it difficult to monitor directly the redox behavior of bases in either their monomeric or extended π-stacked forms where electronic interactions perturb the π-systems of adjoining base steps (5-6). Both organic molecules and transitionmetal complexes have been employed as indirect probes of the electronic properties of DNA (1, 7-15). Photophysical studies on DNA have shown that the π-stack mediates ultrafast electron transfer (kET ∼ 1010 s-1) (1). Although various systems are being explored (912), fast photoinduced electron-transfer kinetics are only observed when both donor and acceptor are intercalated into DNA (13-15); under such conditions, efficient luminescence quenching has been observed over 40 Å (13). Moreover, recent photochemical studies suggest that the extended DNA π-stack facilitates long-range oxidative base damage (4). These results all point to the base stack of DNA as an effective pathway for electron transfer. Electrochemical studies of small molecule/DNA complexes have focused primarily on solution-phase phenomena, in which DNA-induced changes in redox potentials and/or diffusion constants of organic and inorganic spe* Authors to whom correspondence should be addressed. X Abstract published in Advance ACS Abstracts, December 15, 1996.
S1043-1802(96)00070-5 CCC: $14.00
cies have been analyzed to yield association constants (16, 17). In addition, rates of guanine oxidation catalyzed by electrochemically oxidized transition-metal complexes have been used to evaluate the solvent accessibility of bases for the detection of mismatches (18). Electrochemical signals triggered by the association of small molecules with DNA have also been applied in the design of other novel biosensors. Toward this end, oligonucleotides have been immobilized on electrode surfaces by a variety of linkages [e.g., thiols on gold (19, 20), carbodiimide coupling of guanine residues on glassy carbon (21), and alkanebisphosphonate films on Al3+-treated gold (22)] for use in hybridization assays. Both direct changes in mass (measured at a quartz crystal microbalance) (20) and changes in current (19, 21) or electrogenerated chemiluminescence (22) due to duplex-binding molecules have been used as reporters for double-stranded DNA. Gold surfaces modified with thiolated polynucleotides have also been used for the detection of metal ions and DNAbinding drugs (23). To investigate the redox chemistry of molecules within the DNA environment, we have employed gold surfaces modified with double-stranded, thiol-terminated DNA (24). The covalent attachment of DNA directly to an electrode surface provides a controlled environment in which kinetic and thermodynamic parameters of DNAbound species can be evaluated. To probe DNA on these modified surfaces, we have examined the electrochemistry of methylene blue (MB) (25), an aromatic heterocycle that binds strongly to DNA via intercalation (26). Here we report the electrochemistry, binding affinity, and electron-transfer dynamics of intercalated MB at a gold electrode modified with 15-base-pair oligonucleotide duplexes, singly derivatized with a 5′-hexylthiol tether. © 1997 American Chemical Society
32 Bioconjugate Chem., Vol. 8, No. 1, 1997
Kelley et al.
EXPERIMENTAL SECTION
Materials. Phosphoramidite reagents (including the C6 S-S thiol modifier) were obtained from Glen Research. Methylene blue (Sigma Chemical Co.), ferrocene carboxaldehyde, and octadecyl mercaptan (Aldrich) were used as received. Potassium ferrocyanide (Fisher) was recrystallized from aqueous solution prior to use. [γ-32P]dATP was obtained from NEN-DuPont. Synthesis of Derivatized Duplexes. 5′ Mercaptohexyloligonucleotides and underivatized complements were synthesized according to automated solid-phase techniques (27), using a disulfide protected linker (C6 S-S thiol modifier). Sequences were purified by reversed phase HPLC, deprotected using dithiothreitol, and repurified before hybridization to unmodified complements. Single-stranded oligonucleotides were characterized by matrix-assisted laser desorption ionization time-of-flight mass spectrometry and HPLC retention times. Sequences prepared according to this method include 5′ SH(CH2)6-p-AGTGCGAA GCTGCGT 3′, 5′ SH-(CH2)6-pAGTACAGTCATCGCG 3′, and 5′ SH-(CH2)6-p-AGTACAGTCATCAGT 3′. Duplexes were hybridized in deoxygenated 5 mM phosphate/50 mM NaCl (pH 7) by heating to 90 °C followed by slow cooling to room temperature. Unprotected duplexes were stored frozen under argon to prevent oxidation of the thiol. Derivatization of Gold Electrodes. Bulk gold electrodes were polished successively with 0.3- and 0.05-µm alumina (Buhler), sonicated for 30 min, and etched in 1.0 M sulfuric acid. Au(111) surfaces (28) were prepared by vapor deposition onto mica or glass. Electrodes were then modified by incubation in 0.1 mM solutions of derivatized DNA duplexes in 5 mM phosphate/50 mM NaCl (pH 7) for 12-48 h at ambient temperature. Modified electrodes were rinsed in buffer prior to use. For the investigation of DNA-modified electrodes at high MB loadings, deposition of the thiol-terminated DNA was performed in the presence of 0.6 mM MB; these electrodes were rinsed in 10 µM MB/buffer solutions. CH3(CH2)17SH-modified electrodes were prepared by incubation in ethanol solutions for 48 h. Electrochemistry. All electrochemical experiments were performed with a Bioanalytical Systems (BAS) Model CV-50W electrochemical analyzer. Cyclic voltammetry (CV) and chronocoulometry (CC) were carried out at 20 ( 2 °C with a normal three-electrode configuration consisting of either a modified gold-disk or a hanging drop mercury working electrode and a saturated calomel reference electrode (SCE, Fisher Scientific). The working compartment of the electrochemical cell was separated from the reference compartment by a modified Luggin capillary. Potentials are reported versus SCE. Chronocoulometric measurements were corrected for double-layer charge as determined in buffer solutions. Heterogeneous electron-transfer rates were determined by cyclic voltammetry and analyzed as described previously (29). The electron-transfer kinetics of several previously reported systems (including benzo[c]cinnoline on gold and methylene blue on Hg) (30) were investigated using this method with our instrumentation, and in each case, excellent agreement with the literature values was observed. Ellipsometry. Optical ellipsometry (λ ) 632.8 nm) was carried out on dried samples at 25 °C using a Gaertner Model L116C ellipsometer. 32 P Labeling and Quantitation. A 15-base-pair oligonucleotide (5′ CGCGATGACTGTACT 3′) was 5′ labeled with [γ-32P]ATP and hybridized to a thiol-modified complement. Samples of Au(111) on mica (diameter
Figure 1. Cyclic voltammograms for 0.1 mM K4[Fe(CN)6] in 5 mM phosphate/50 mM NaCl (pH 7) at a DNA-modified (s) [5′ SH-(CH2)6-p-AGTACAGTCATCGCG 3′ + complement] and bare gold electrode (- - -). Scan rate ) 100 mV/s; electrode area (A) ) 0.02 cm2.
) 0.41 cm) were treated with this duplex in the manner described above. After thorough rinsing in buffer solutions, dried samples were counted on a Beckman LS60001C scintillation counter, adjusted for attenuation by the presence of the mica, and compared to calibration standards that were prepared from known quantities of labeled oligonucleotide. The values obtained by scintillation counting and the surface areas of the modified mica were confirmed by analysis of the radioactivity of each sample by phosphorimagery (Molecular Dynamics). RESULTS
Surface Analysis. Modification of gold surfaces with 5′-thiol-terminated DNA duplexes was established by cyclic voltammetry, ellipsometry, and the direct quantitation of 32P-labeled oligonucleotides on Au(111). The electrochemical window of 50 mM NaCl (pH 7) at DNAmodified gold extends from 0.70 to -0.70 V (versus SCE). The cyclic voltammograms of Fe(CN)64- at bare and DNAmodified electrodes are shown in Figure 1. As ferrocyanide does not bind to the DNA polyanion, the lack of signal at the modified electrode implies essentially complete coverage by the thiol-terminated DNA. To ensure that Fe(CN)64- was not merely electrostatically repelled from holes in the monolayer, a neutral probe, ferrocene carboxaldehyde, was investigated and gave a similar result. Gold surface waves (generated by the formation and stripping of gold oxide) at bare and derivatized electrodes were also compared by cyclic voltammetry as a qualitative measure of the surface coverage (31, 32). These results suggest a very high loading (>85%). However, thiol-gold linkages are known to undergo oxidative desorption at high potentials (33), so an alternative method was used to quantitate the coverage. DNA adsorbed on Au(111) was directly quantitated in a radioactive tagging (32P) experiment. This assay yielded an average of 5.7(3) × 10-12 mol of duplex DNA on a 0.14-cm2 surface after 24 h of modification. From this value and the cross-sectional area of DNA (3.14 × 10-12 cm2), the surface coverage was calculated as Γ ) 4.1(2) × 10-11 mol/cm2, which corresponds to a closepacked fractional coverage of 0.75(3). Ellipsometry of a 15-base-pair thiol-derivatized duplex on Au(111) was carried out and yielded an estimate for
Electrochemistry of Methylene Blue
Bioconjugate Chem., Vol. 8, No. 1, 1997 33
Figure 3. Plot of E - Epc vs log(ν) for 0.1 µM methylene blue at a gold electrode derivatized with 5′ SH-(CH2)6-p-AGTACAGTCATCGCG 3′ hybridized to its complement in 25 mM phosphate/ 75 mM NaCl (pH 7) .
Figure 2. (A, top) Cyclic voltammetry of 1.0 µM methylene blue in 50 mM phosphate (pH 7) at a DNA-modified [5′ SH(CH2)6-p-AGTGCGAAGCTGCGT 3′ + complement] electrode (A ) 0.7 cm2; scan rate ) 5, 10, 20, 50, 100, 200, and 500 mV/s). (B, bottom) Plot of ipc vs scan rate.
the average monolayer thickness of 35 Å. As this measurement corresponds to an average thickness on the gold surface, it could reflect anywhere from 100% surface coverage with derivatized DNA duplexes stacked at an angle of 32° from the gold surface, to DNA helices oriented perpendicular to the surface with only 55% coverage, given a cylindrical DNA 15-mer duplex of 66-Å height (including the fully extended linker) and 20-Å diameter (34). Cyclic Voltammetry of Methylene Blue. The cyclic voltammetry of 1.0 µM MB at a DNA-modified electrode is shown in Figure 2. The pronounced electrochemical response at such a low concentration is strong evidence that MB binds tightly and is electronically well coupled to the modified electrode surface. The reduction potential of MB at the DNA-modified electrode is -0.25 V (vs SCE), compared to -0.22 V at bare gold. A plot of cathodic peak current (ipc) vs scan rate (ν) is linear (Figure 2B), establishing that MB is strongly adsorbed to the DNAmodified surface (35). Qualitatively, ∆Ep increases as a function of scan rate (Figure 3), indicating slow electron-transfer kinetics on the CV time scale (29). For comparison, the peak separations are much less pronounced for MB adsorbed to a mercury surface, where the rate constant is reported as 1500 s-1 (25); at bare gold we observed essentially no peak splitting up to our fastest scan rates (50 V/s). Importantly, these measurements were all made at very low loadings of MB; thus, artifacts due to lateral charge migration or ohmic (iR) drop were minimized (at the largest currents used, iR was
