Low bias charge transport in DNA – Scientific Reports

Charge transport (CT) is a powerful tool for investigating chemical processes that are highly sensitive to the molecular structure and shape of molecules. Its application to single-stranded DNA (ssDNA) offers numerous potential applications across various fields. For instance, in DNA sequencing1, the development of biosensors that detect various biomolecules and pathogens by monitoring changes in charge transport through ssDNA when it interacts with target molecules2. This technology also extends to the fabrication of DNA-based electronic components3 and molecular junctions, which are integral to molecular-scale electronics3,4. Furthermore, engineering DNA-based nanomaterials with specific charge transport properties can revolutionize fields like drug delivery, tissue engineering, and medical imaging, theranostics are just a few of the many applications of charge transport investigations in DNA. The study of charge transport in ssDNA enhances our understanding of biotechnology and nanoelectronics and has implications for energy, environmental monitoring, and fundamental scientific research, highlighting its diverse potential across numerous disciplines.

The purpose of our studies outlined in this article was to assess how sensitive charge transport methods are for studying the physicochemical processes within ssDNA chains attached to gold electrodes. A thiol modification, widely used recently in various applications5 and having a high affinity for the gold surface, was used to link ssDNA to gold electrodes6. This will allow, in the next steps, the application of these methods to the study of ssDNA mutations and damage that contribute to cancer and other diseases. Despite significant advancements, the mechanisms and nature of charge transport in DNA remain incompletely understood. Many factors affect the electric transport phenomenon in DNA: the type of DNA (whether ssDNA or double-stranded (dsDNA), strand length, ambient conditions (either in vacuum or liquid), and DNA modifications with compounds (e.g., metal electrode-thiol-DNA).

We also discovered that DNA’s electronic properties are not solely dependent on electrons but also involve hole charge carriers. Low-energy electron microscopy has revealed both types of carriers in ssDNA, showing correlated charge redistribution over distances of tens of nanometers7. This finding is crucial for applications like using DNA as a nanowire or an electrochemical biosensor, where the transfer processes of electrons or holes along the strand are integral. Understanding the redox properties and the charge distribution along the strand is of great importance for designing new devices that effectively transmit electric current. We found that nucleobases primarily facilitate the charge transfer process in the aqueous phase, with neutral nucleobases typically more likely to be oxidized than reduced, and as a result, the charge transfer will usually be carried out by a hole rather than by an electron. Particularly, adenine (A) and guanine (G) display stronger tendencies to facilitate hole transport compared to cytosine (C) and thymine (T). Therefore, they impose the hole type of charge transport in DNA. For example, when guanine is one of the components of the system, its properties are dominated by those of guanine, and it has been computed that around 80% of the hole charge is located on just one of the guanine moieties8,9. The localization or delocalization of charge also strongly depends on the nucleotide sequence. For example, charge is more delocalized in sequences with AG compared to those with GT or GC, due to the similar reducing characteristics of adenine and guanine. We know that guanine and thymine/cytosine. It is known that G and A are responsible for hole transfer whereas C and T are for electron transfer in the strand. Hopping and tunneling are the primary mechanisms for transporting these charge carriers in DNA10,11. Hopping is a process where the charge moves across several nucleobases at once, highly dependent on the π-stacking interactions and the distance between the nucleobases. This mechanism is considered a long-range process where the charge hops from one nucleobase to another, propagating the hole along the DNA strand. It shows a smoother dependence with the distance between nucleobases, but a stronger influence of the DNA-strand sequence since nucleobases with identical one-electron oxidation potentials. Our findings indicate that the rate of hole charge transfer significantly increases with the presence of multiple guanine nucleobases and decreases as the distance between them expands9. Hopping conduction occurs between π-stacked bases of the DNA and is limited by the length of the DNA strand12,13,14. The phenomenon of hopping conduction, well established in organic materials, demonstrates that electron hopping rates in DNA are influenced by the sequence of the DNA, its temperature, and the presence of specific chemical groups on the strand. Quantum tunneling, particularly proton tunneling in dsDNA, frequently occurs and can lead to the presence of hundreds of thousands of tautomers within a cell’s genome15 at any given moment. Even though these structures are transient, the numerous tunneling events they facilitate can become a significant source of mutations. This suggests that quantum-mechanical instability plays a crucial role in DNA mutation dynamics.

In ssDNA, while proton tunneling is not observed, electron and hole tunneling are still relevant phenomena. Hole tunneling involves the delocalization of holes across several nucleobases until reaching the target component10. Electron tunneling, observed in both ssDNA and dsDNA, shows that the conductivity of dsDNA is 30 times higher than that of ssDNA. This increased conductivity in dsDNA is attributed to the formation of highly ordered self-assembled monolayers, which facilitate efficient and coherent charge tunneling. Conversely, ssDNA is characterized as a more disordered chain where charge tunneling occurs more incoherently14. Tunneling can manifest as either coherent or incoherent. Coherent tunneling has been observed between AT and CG base pairs15,16,17,18,19, whereas a crossover between intermediate incoherent hopping and coherent tunneling processes has been observed exclusively in dsDNA only20. Studies have shown that the resistance of dsDNA stretched between two gold electrodes exhibits a periodically modulated linear dependency on the length of the DNA chain, indicating a hybrid regime of charge transport combining coherent tunneling and hopping. In ssDNA, the tunneling phenomenon is associated with charge carriers traversing the potential barrier separating the bases. The tunneling current heavily depends on the base-to-base distance, the chemical nature of the bases, and the orientation of the DNA strand. Advanced experimental techniques such as CAFM and STM have been pivotal in demonstrating charge transport and tunneling currents in ssDNA. For example, Fink and Schönenberger21 used STM to show that electrons can effectively tunnel through DNA bases with a decay length of approximately 1 nm. Similarly, Cui et al.22 utilized CAFM to explore and confirm the conductivity of both ssDNA and dsDNA, indicating that the conductivity of ssDNA can be modulated by altering the DNA sequence or attaching specific chemical groups to the strand.

Initially, the tunneling effect was understood as the peculiar electrical behavior observed in some semiconductors, where the electric current decreases as the applied voltage increases within a specific range, leading to what is known as negative differential resistance (NDR). This phenomenon is significant in the development of low-power memory or logic devices like Esaki diodes23 and resonant tunneling diodes24, where NDR characteristics have been widely utilized. More recently, NDR traits have also been observed in organic molecules25,26,27, broadening the scope of materials demonstrating this behavior. Various mechanisms behind NDR have been proposed depending on the molecular device systems. These include charging reduction25,28, redox reactions27, structural changes29, chemical reactions30, and association–dissociation processes31 of molecules. Interestingly, this phenomenon has been recorded in DNA as well, particularly in metal/molecule junctions where charge transport often involves the Fowler-Nordheim (FN) tunneling process. FN tunneling, a quantum mechanical phenomenon, involves electrons tunneling through a potential barrier, typically observed in metal-insulator-metal structures. The current resulting from FN tunneling is notably influenced by factors such as the thickness of the insulating layer (or tunneling barrier) and the applied bias voltage. While FN tunneling can significantly contribute to charge transport in metal/molecule junctions, other mechanisms like charge transfer and hopping transport also play essential roles. FN tunneling exhibits a certain degree of temperature independence within a specific range; however, higher temperatures can influence the tunneling process by increasing electron thermal excitation, a regime known as “thermionic emission”32. Furthermore, the tunneling current in FN tunneling is also impacted by the thickness of the tunneling barrier and the voltage applied. At low-bias limits, the shape of the tunneling barrier is typically rectangular, and the formula for calculating the direct tunneling current (I) through such a barrier is approximated by Ref.32:

$$:lnleft(frac{I}{{V}^{2}}right)propto:lnleft(frac{I}{V}right)-frac{2dsqrt{2{m}_{e}varphi:}}{{hslash:}}$$

(1)

where V is the applied bias voltage, d is the barrier width, me is the mass of an electron, f is the barrier height, me – mass of an electron.

However, as the biasing voltage increases in our experiments, the potential barrier initially resembling a rectangle shifts to a trapezoidal shape, leading to “direct quantum tunneling” becoming the predominant mechanism for charge transport across the barrier. Further increase in bias voltage transforms the barrier into an approximately triangular shape, facilitating charge transport primarily through FN tunneling:

$$:lnleft(frac{I}{{V}^{2}}right)propto:-left(frac{I}{V}right)frac{4dsqrt{2{m}_{e}{varphi:}^{3}}}{3q{hslash:}}$$

(2)

where q is the electron charge.

In our study, the devices constructed consisted of ssDNA chains linked to gold nanoelectrodes (Fig. 1). Measured resistance of DNA hybrids (ssDNA/thiol/Au) depends on many factors. The most important are: type of DNA (dsDNA or ssDNA)14, base-pairs sequencies1, contact resistance of electrode/molecule junction33, humidity34,35, and chain’s length1,36, 37. Taking into account above mentioned factors, for DNA chains of length of 100 nm the expected resistance of our samples should be between few mega ohms (see ref.13, Table 2) till tens of mega ohms36.

Fig. 1
figure 1

(a) A schematic of the ssDNA ligation procedure. (b) An SEM image showing Au/Ti electric contacts between which ssDNA was stretched. (c) AFM topography of ssDNA deposited on the surface of gold electrodes without forming a connection between the electrodes. (d) AFM topography of the ssDNA (red arrow) connecting both electrodes.

We prepared 30 such devices. A few days prior to the scheduled measurements, fresh ssDNA synthesis was performed and after thiol functionalization deposited on gold nanostructures. Drying of the samples prepared in this way was performed in a vacuum desiccator. Devices stored in vacuum and measured immediately after removal from the desiccator showed similar resistance and current–voltage (IV) characteristics. In the main text we presented results of the sample that exhibited the best quality data with the least variation in differential resistance. Although other samples displayed similar differential resistance (see fig. S1) characteristics, their data quality was lower, likely due to less optimal contact resistance between the DNA and the gold electrodes, as well as slight variations in the DNA strand lengths and formation of secondary structures during the anchoring process on the gold surface. Therefore, we chose to focus on the sample with the most reliable and consistent results to ensure the accuracy of our analysis.

The target 300 nt ssDNA was synthesized by ligating three 100 nt ssDNA oligonucleotides via two distinct ligation reactions, detailed in the Supporting Information and illustrated in Fig. 1a. Initially, two 100 nt ssDNA strands were connected using T4 DNA ligase in an in vitro ligation reaction, aided by a 30 nt bracket ssDNA complementary to the 3ʹ and 5ʹ ends of the first and second ssDNA, respectively. The resulting 200 nt ssDNA product was then ligated with a third 100 nt ssDNA in a similar in vitro reaction, facilitated by appropriate bracket ssDNA. Two of the ssDNA oligonucleotides used in these reactions were chemically modified: ssDNA_01 was modified at the 5ʹ end with a thiol modifier C6 S–S (disulfide), and ssDNA_03 was modified at the 3ʹ end with a thiol modifier C3 S–S (disulfide), both linked via a diester phosphate bond to an oligonucleotide. Consequently, the finally 300 nt ssDNA product of this ligation presented disulfide groups at both ends (Fig. 1a). Reduction with DTT reagent transformed these disulfide groups into thiol groups, allowing the attachment of the ssDNA to the gold electrodes by it ends.