Discovery at the VICB
Electrifying Control of DNA Replication
By: Carol A. Rouzer, VICB Communications
Published: March 6, 2017
Charge transfer through DNA triggers redox chemistry at the [4Fe4S] center of DNA primase that modulates enzymatic function.
A remarkable, but rather obscure property of DNA is its ability to transfer an electrical charge over a great distance. Attributed to the pi stacking of bases in the core of the double helix, this charge transfer (CT) capability opens the possibility that DNA can participate in redox-based signaling involving molecules bound to the duplex. Consistent with this hypothesis is the discovery that a substantial number of DNA-processing enzymes contain [4Fe4S] clusters that can serve as redox centers within the proteins. Cells expend considerable energy and resources to construct these clusters, so they likely play an important role in the regulation of protein function. To gain new insight into the redox-mediated control of one of these [4Fe4S] cluster-containing enzymes, Vanderbilt Institute of Chemical Biology member Walter Chazin, his collaborator Jacqueline Barton (California Institute of Technology), and their laboratories launched a detailed study of the human DNA primase. They now report that DNA CT modulates the binding affinity and activity of DNA primase through redox chemistry at the [4Fe4S] cluster contained in the enzyme's regulatory p58 subunit [E. O'Brien et al., (2017) Science, 355, eaag1789].
DNA replication always starts with the synthesis of a short (8 to 12 nucleotides) RNA primer that is complementary to the DNA template. DNA primase is the enzyme responsible for constructing this primer. Primase works in collaboration with DNA polymerase α (pol α), which adds ~20 nucleotides of DNA to the primer before handing off the task of further extension to one of the more processive DNA polymerases. Primase and pol α act together as a heterotetrameric complex, to which each enzyme contributes a catalytic subunit and a regulatory subunit containing a [4Fe4S] cluster. The Chazin and Barton laboratories hypothesized that the [4Fe4S] clusters serve as redox-dependent switches to regulate primer synthesis and/or hand-off, and that DNA CT is the mechanism by which oxidation and reduction of the clusters occur.
To test their hypothesis, the investigators constructed gold-surfaced electrodes that they incubated with DNA duplexes bearing a terminal alkane-thiol moiety. The thiol functionality facilitated binding of the DNA to the gold surface, while a 3-nucleotide 5′ single-stranded overhang at the opposite end provided a binding site for the [4Fe4S]-containing C-terminal domain of human DNA primase (p58C). The resulting electrodes enabled the investigators to deliver an electric charge directly to associated p58C via CT through the DNA bound to the electrode surface. Using this device the investigators could apply a constant potential of +412 mV [versus the normal hydrogen electrode (NHE)] to a solution containing p58C to convert the [4Fe4S] center to the fully oxidized state ([4Fe4S]3+), or they could apply a potential of -188 mV (versus the NHE) to fully reduce the center to the [4Fe4S]2+ state. Alternatively, the use of cyclic voltametry to apply a gradually increasing or decreasing potential enabled them to monitor redox activity at the [4Fe4S] center as detected by changes in current flow.
Initial studies using cyclic voltametry of p58C in the presence of ATP revealed a reversible oxidation and reduction of the [4Fe4S] center in the range of 145-150 mV (versus the NHE). This was not observed in the absence of ATP. The investigators hypothesized that, in the presence of ATP which is a substrate for primase, the p58C subunit becomes tightly associated to the electrode-bound DNA, enabling a facile oxidation and reduction of the [4Fe4S] center via CT.
The p58C appeared to be unresponsive to CT-mediated redox reactions in the absence of ATP. However, the investigators discovered that, following bulk oxidation of the [4Fe4S] center, cyclic voltametry revealed a single reduction event occurring at -130 to -140 mV (versus the NHE). This reduction was irreversible, and did not recur on subsequent cycles. No similar redox event occurred upon cyclic voltametry following bulk reduction of p58C (Figure 1). However, the process of bulk oxidation followed by cyclic voltametry could be repeated multiple times with the same single reduction event observed every time. An interesting observation was that the magnitude of the reduction event increased with each repeated oxidation and reduction. To explain these results, the researchers proposed that oxidation of the [4Fe4S] cluster leads to an increase in affinity of p58C for DNA and that the resulting tight association with DNA renders the protein susceptible to CT-dependent reduction of the center. It is this reduction that is observed during cyclic voltametry, and it produces a loss of affinity and associated resistance to CT-dependent processes. Thus, it requires bulk oxidation to return p58C to its CT-susceptible state. However, with each succeeding oxidation, the investigators proposed that more of the protein is oxidized, thereby explaining the increased signal with each successive repeat of the process.
FIGURE 1. Overview of the experimental design. (A) Sixteen DNA-coated gold electrodes (circles, center) are located on a chip that is divided into four sections. This arrangement allows simultaneous data collection from four distinct experimental conditions. In this case, one quadrant each is undergoing bulk oxidation, bulk reduction, no electrolysis, or cyclic voltametry as labeled. (B) Schematic showing the effect of redox modulation on p58C binding. p58C (brown) containing a reduced [4Fe4S] center (red) exhibits only moderate affinity for the DNA attached to the gold electrode. Bulk oxidation of the [4Fe4S] center (blue) by application of a positive potential results in an increased affinity and tight binding to the DNA. (C) Following bulk oxidation, cyclic voltametry carried out with increasingly negative potential leads to a reduction event as observed by a peak of current between -100 and -200 mV (blue). No such event occurs when the potential is cycled back to positive, or on subsequent cycles. (D) No redox events are observed upon cyclic voltametry of p58C following bulk reduction. Figure reprinted by permission from AAAS from E. O'Brien et al., (2017) Science, 355, eaag1789. Copyright 2017, AAAS.
The results thus far strongly supported a role for CT-dependent redox signaling in the control of p58C function. For this to be true, however, requires a mechanism for charge transfer from the DNA-binding domain of the protein to the [4Fe4S] cluster, 25 Å away. The investigators proposed that a potential pathway for this transfer could be formed by three highly conserved tyrosine residues (Tyr-309, Tyr-345, and Tyr-347) that lie between the DNA binding site and the redox cluster (Figure 2). The low ionization potential of tyrosine, combined with their key location suggested that these residues could serve as the required charge-transfer network. To test this hypothesis, the researchers constructed three p58C proteins, each one bearing a mutation of one of the tyrosine residues to phenylalanine. They used circular dichroism and UV/visible spectroscopy of all three proteins along with X-ray crystallography of the Y345F mutant to show that the alterations had no significant effect on the overall structural integrity of the proteins. They also used a fluorescence anisotropy-based DNA binding assay to show that the mutations did not affect the baseline ability of the proteins to interact with a DNA substrate. Having established these important control parameters, the investigators went on to show that, although the mutants exhibited the same pattern of CT-dependent redox behavior as the wild-type protein, the size of the reduction-dependent signal on cyclic voltametry was markedly reduced in all three, and the percentage of the electrolysis charge recovered in the cyclic voltametry peak was also much lower in the mutants than the wild-type protein. These results clearly demonstrated that all three tyrosine mutations resulted in a CT-deficient state, confirming that they play a role in conveying electrical charge between the DNA binding site and the [4Fe4S] center of p58C.
FIGURE 2. (A) Diagrammatic representation of the X-ray crystal structure of p58C, showing the DNA-binding region, the location of the [4F34S] center (colored spheres), and the tyrosine residues proposed to serve as a charge-transfer network. (B) Close-up view of the three tyrosine residues. Figure reprinted by permission from AAAS from E. O'Brien et al., (2017) Science, 355, eaag1789. Copyright 2017, AAAS.
Although the results strongly suggested that CT-mediated redox events at the [4Fe4S] cluster of p58C affect DNA binding, they did not directly address whether or not this phenomenon plays a significant role in the physiological modulation of DNA primase function. To explore this possibility, the investigators expressed a complete wild-type DNA primase heterodimer in addition to heterodimers containing a mutation at one of the critical charge-transfer tyrosine residues. They first tested these enzymes in an assay that measured their ability to initiate primer synthesis on a single-stranded DNA template. The results showed that the tyrosine mutations each resulted in a 65-85% reduction in the enzymes' ability to initiate primer synthesis. The researchers then evaluated the proteins for their ability to elongate a pre-existing primer. In this case they found that the mutants were much more likely than the wild-type enzyme to elongate the primer to full-length, suggesting a defect in primer truncation, a process necessary to prevent the primer from becoming too long. They confirmed that CT-mediated redox signaling is required for correct primer truncation by showing that a template designed to result in a mismatch in the sequence of the newly synthesized DNA was much less likely to be properly truncated than a substrate designed to lead to a correctly matched sequence. The mismatch blocks CT in the DNA strand by disrupting the pi-stacking in the core of the double helix. These results indicate that CT-dependent redox signaling is required both for correct initiation of primer synthesis and for proper truncation. However, primer elongation does not appear to require control by the redox-dependent processes studied here.
From their data, the investigators proposed a model for CT-dependent regulation of primase function (Figure 3). They suggest that oxidation of the [4Fe4S] center is required to convert the p58C heterodimer to a high affinity state that facilitates its association with DNA and initiation of primer synthesis. After the primer becomes ~7 to 10 nucleotides long, there is sufficient room for binding the reduced regulatory subunit of pol α (or another redox signaling partner). CT results in movement of electrons from the [4Fe4S] center of pol α to that of primase. Upon reduction of its [4Fe4S] center, primase reverts to a lower affinity form, and RNA primer synthesis ceases, while pol α takes its place upon oxidation of its redox center. This enables pol α to elongate the RNA primer with a stretch of DNA before hand off to the more processive polymerase that will finish the job. Considering the large number of DNA-processing enzymes that contain an [4Fe4S] center, it is quite likely that this fascinating mechanism may apply, in some form, to the regulation of many key reactions involving DNA replication, transcription, and repair.
FIGURE 3. Proposed mechanism for DNA-CT-dependent modulation of RNA primase. Oxidation of the [4Fe4S] center of p58C leads to high affinity binding to the DNA template and efficient initiation of primer synthesis. When the primer has reached sufficient length, the regulatory subunit of a second enzyme bearing a reduced [4Fe4S] center can bind. Transfer of electrons from the reduced to the oxidized [4Fe4S] centers results in reduced affinity/activity for primase, and activation of the new enzyme. Figure reprinted by permission from AAAS from E. O'Brien et al., (2017) Science, 355, eaag1789. Copyright 2017, AAAS.