High-Precision Electrochemical Measurements of the Guanine-, Mismatch-, and Length-Dependence of Electron Transfer from Electrode-Bound DNA Are Consistent with a Contact-Mediated Mechanism

Abstract

Despite 25 years' effort, serious questions remain regarding the mechanism(s) underlying electron transfer through (or from) electrode-bound double-stranded DNA. In part this is because a control experiment regarding the putatively critical role of guanine bases in the most widely proposed transport mechanism (hopping from guanine to guanine through the π-stack) appears to be lacking from the prior literature. In response, we have employed chronoamperometry, which allows for high-precision determination of electron transfer rates, to characterize transfer to a redox reporter appended onto electrode-bound DNA duplexes. Specifically, we have measured the effects of guanines and base mismatches on the electron transfer rate associated with such constructs. Upon doing so, we find that, counter to prior reports, the transfer rate is, to within relatively tight experimental confidence intervals, unaffected by either. Parallel studies of the dependence of the electron transfer rate on the length of the DNA suggest that transfer from this system obeys a "collision" mechanism in which the redox reporter physically contacts the electrode surface prior to the exchange of electrons.

Figure 1.

To investigate the electron transfer kinetics of DNA duplexes, we designed redox-reporter-modified guanine-free, guanine-rich, and mismatch-containing constructs attached via one end to a mercaptohexanol monolayer on a gold electrode. As the redox reporter (labeled “R”), we use methylene blue, which is commonly used to monitor through-DNA electron transfer.^,,,,,,,, To measure the rate of transfer fromsuch constructs, we use chronoamperometry, an approach that determines electron transfer kinetics with improved precision than can generally be achieved by the previously employed voltammetric techniques. Shown, for example, are current decays measured for a 16-base DNA sequence in its single- and double-stranded forms. Fitting these to a single-exponential decay returns k_app = 150 ± 3 s⁻¹ for the single-stranded constructcontaining only adenine and thymine and k_app = 2.8 ± 0.9 s⁻¹ when its complement is added (at 100 nM) and the system is allowed to hybridize. (The “error bar” intervals quoted for these numbers as well as those reported or illustrated elsewhere in this paper reflect 95% confidence intervals estimated from replicate measurements conducted on 12 independently fabricated electrodes.)

Figure 2.

We have used chronoamperometry to monitor the electron transfer rates associated with surface-attached, redox-reporter-modified DNAs. (A) To ensure that the surface-attached DNA is fully in its hybridized, double-stranded state, we record chronoamperograms every second after the addition of 100 nM of the complementary strand. A plot of the resulting chronoamperograms in their “chronocoulommogram” format (charge transferred versus 1/time) clearly illustrates the evolving relative populations of single- and double-stranded molecules. Specifically, the overlaid chronocoulommograms exhibit an “isosbestic” point (point at which all the curves cross), indicating that the system only populates two states: single-stranded DNA transitions into fully double-stranded DNA without ever significantly populating any partially hybridized states. (B) If, instead, we hybridize the DNA prior to attaching it to the surface,^,, we find that the transfer rate we observe (112 ± 30 s⁻¹) is intermediate between those of the single-stranded (150 ± 3 s⁻¹) and fully duplex (2.8 ± 0.9 s⁻¹) states (see chronocoulommograms in Figure S2, SI), suggesting that single-stranded DNA is formed during the deposition process.^, Shown are data for a 16-base construct lacking guanine bases at a packing density of 2.9 pmol cm⁻².

Figure 3.

Rate of electron transfer associated with duplex DNA is not a significant function of its sequence composition, presence or absence of a mismatch, or even its relative surface coverage on the electrodes (i.e., packing density), a result that holds at both (A) lower and (B) higher packing densities. For example, the rate associated with a fully complementary, 40-base-pair guanine-free construct is effectively indistinguishable from that of a construct in which we have replaced more than half of the A−T base pairs by G−C base pairs. The introduction of a C−A mismatch (at base pair 11) in either construct likewise does not measurably alter the transfer rate. The “error bars” shown again reflect 95% confidence intervals estimated from replicate measurements conducted on 12 independently fabricated electrodes.

Figure 4.

Studies of the length dependence of transfer from guanine-free duplexes suggest that electron transfer is influenced by the position of the redox reporter with respect to the electrode surface. Specifically, the length dependence of transfer from a series of A−T duplexes fits a sinusoidal function with a period of 11.3 ± 0.7 base pairs (this is the best-fit value using an unconstrained value for the period of the sine wave). Perhaps tellingly, this periodicity is within error of the number of base pairs per turn for A−T-rich duplexes.⁵⁸ From this we hypothesized that the surface-attached DNA duplexes are either lying flat or transiently “lean over” and collide with the electrode surface, and thus, the orientation of the redox reporter relative to the face of the DNA on which the flexible surface attachment point is located defines the transfer rate. This implies that the transfer rates measured electrochemically in systems such as the ones we (and others^,,,) have employed originate from an alternate mechanism in which electrons tunnel to the redox reporter only when structural dynamics in the DNA bring the reporter in close proximity to the electrode surface.