Charge Transport Communication through DNA by Protein Fe-S Clusters: How Far Is Not Too Far?

The key role of DNA in storing genetic instructions for producing proteins and RNA constitutes the foundational chemistry for life itself. However, DNA also has the more enigmatic and provocative chemical characteristic that it can mediate charge transport (CT) through the continuum of pi electron density formed by double-helical base stacking.[1] Now, measurements by a team led by Jacqueline Barton imply ∼3.5 kilobase pairs (kbp) is a duplex DNA length that is not too far for protein iron–sulfur (Fe–S) clusters to mediate DNA CT.[2]

Fe–S clusters are central to all life as ancient and uniquely powerful cofactors for accepting or donating electrons. However, their inherent susceptibility to oxidation and degradation enables iron-mediated DNA damage via the Fenton reaction,[3] making the clusters a mechanism for oxygen cytotoxicity that threatens genetic inheritance. The recent recognition that human DNA replication, transcription, and repair enzymes have Fe–S cluster cofactors whose fundamental chemistry endangers DNA is a paradox that may be explained by evolutionary selection of Fe–S cofactors for efficient CT communication along DNA that breaks classical diffusion-limits for protein communication and exceeds known distances for electron transfer in proteins.[4]

Importantly, protein Fe–S clusters can be oxidized upon DNA binding or reduced on DNA by accepting an electron lost from a neighboring Fe–S containing enzyme, and enzyme affinity decreases upon reduction of Fe–S cluster promoting dissociation from the DNA.[5] Thus, DNA repair enzyme Endonuclease III (EndoIII) with oxidized [4Fe–4S]3+ cluster binds DNA 550-fold stronger than one with a reduced [4Fe–4S]2+ cluster.[5] Furthermore, Fe–S proteins can influence each other’s DNA-binding affinity through CT. Upon binding to DNA, EndoIII is oxidized to the [4Fe–4S]3+ state and binds DNA with higher affinity by losing an electron that can travel through DNA and reduce a neighboring DNA-bound Fe–S protein to its [Fe–S]2+ state leading to its reduced DNA binding and consequent release from DNA.[6] In the presence of CT, DNA will therefore have fewer bound Fe–S proteins due to efficient CT-mediated release compared to conditions without CT where more than one protein with an oxidized [Fe–S]3+ cluster can retain high affinity. Thus, experimentally examining DNA-binding enzymes with redox-active Fe–S clusters identified DNA CT as a means for proteins to communicate over distances.

Using these properties in combination with single-molecule atomic force microscopy (AFM), Barton and co-workers compared Fe–S containing enzyme distribution on matched DNA duplex versus mismatched duplex indicating mismatch defects interrupts CT.[6] They previously found that similar CT rates between different Fe–S containing enzymes on DNA duplex may occur over kbp,[7] but is there an effective CT limit (Figure )? That is, if we consider DNA as a wire capable of transmitting messages from one Fe–S containing enzyme to another, how far can a CT signal be transported through well-packed DNA? Prior experiments with short DNA lengths on electrodes suggested minimal distance effects with similar CT rates seen for 17 bp vs 100 bp, which seem in conflict with the Marcus theory.[8,9] In this new paper, the Barton team sought to answer this key question by employing DNA duplexes with sizes varying from 1.6 to 8.9 kbp and visualizing Fe–S enzyme binding by single-molecule AFM.

Figure 1

Charge transport (CT) communication between Fe–S proteins (cartoon figures) occurs along the DNA duplex (double-helical lines). Within the CT range (top), CT allows a DNA-bound [Fe–S]+3 protein to decrease the DNA binding affinity of neighboring Fe–S proteins by reducing the Fe–S cluster (green and gold spheres) to a [Fe–S]+2 oxidation state. This communication is lost when proteins are outside the effective CT limit (bottom).

For a CT-proficient wild-type (WT) EndoIII at an ambient temperature, the authors found that the curve representing bound proteins versus DNA length has two regions: one with a shallow slope and another with a steep slope (Figure ). The shallow slope region suggests an in-range DNA length for CT whereas the steep slope region implies an out-of-range DNA length for CT, and their intersection (∼3.5 kbp) provides a proposed measure of the effective DNA length through which CT may proceed. This ∼3.5 kbp (Figure , blue arrow) region exceeds the DNA length in prior reports on CT through DNA.[7] However, for a CT-deficient mutant EndoIII Y82A, the binding curve showed only a single region with steep slope consistent with an absence of in-range DNA CT. Interestingly, decreasing the temperature to 4 °C or adding a metallointercalator to the experiment with WT EndoIII increases the effective length (∼5.3 kbp or 1.8 μm) of in-range DNA length for CT (Figure , green arrow) by increased base stacking and decreased DNA floppiness.

Figure 2

Effective CT distance as defined by plotting the number of Fe–S proteins bound to variable length DNA duplexes. Lines fit to the binding data (green and blue) show two distinct regions: one with a shallow slope (solid lines) for lengths enabling CT and another with a steep slope (dashed lines) for lengths beyond the effective CT limit. A sharp change in slope marks the CT boundary (colored arrows) that is absent for a CT defective mutant (black line). Data is plotted from ref (2). Copyright 2018 American Chemical Society.

Readers will recognize caveats in defining the current effective DNA length that supports CT. First, for human biology one would wish to measure CT regions at 37 °C. Second, as the authors note, the experimental setup hinders directly measuring the distance (or DNA length) through which CT can proceed between protein pairs, as the setup does not control where proteins land on DNA duplex. Similarly, random selection of DNA lengths for CT measurements complicates robust comparisons among experiments. In the future increasing the number of data points might improve accuracy in identifying in-range and out-of-range CT distances. However, many tests may prove experimentally cumbersome, supporting the value of a theoretical treatment as suggested by the authors. As Fe–S proteins have similar reduction potentials upon binding to DNA, CT occurs between the same and also between different proteins. Also, since CT depends primarily upon base stacking, it proceeds through nucleosomal DNA.[10] Consequently, the authors note that Fe–S cluster containing DNA repair proteins should aid in damage sensing by allowing efficient scanning for damage within the CT feasible range compared to protein diffusion within the crowded environment of the nucleus.

In sum, DNA CT chemistry enables long-distance electron transfer among Fe–S proteins via the DNA duplex by a mechanism exquisitely sensitive to defects in base pair stacking that arise from DNA damage and during transcription and replication. CT is provocative, as it seems to change the rules for protein communication, DNA–protein transactions, and possible chemically directed coordination of DNA replication, transcription, and repair processes. Importantly, while refinements are expected, having a specific CT length should aid design of experiments to test CT in the possible regulation and coordination of DNA transactions in cells. In particular, as human replicative polymerases contain Fe–S clusters, the proposed ∼3.5 kbp CT region appears suitable to coordinate and regulate multiple origin firing for efficient DNA replication despite long human chromosomes. Disrupting polymerase CT may therefore be detectable as a decrease in coordinated origin firing patterns in cells. In general, CT in DNA metabolism warrants more investigation for defining the chemical activities of Fe–S cluster enzymes on DNA. While Tse et al. focus on damage searching for DNA base repair, the measurement of an effective CT region in DNA should provoke multiple tests of DNA CT functions in cells with possible broad implications for biology and nanomedicine including the design of DNA-based sensors and single-molecule devices.