Direct unfolding of RuvA-HJ complex at the single-molecule level

Abstract

The repair of double-stranded DNA breaks via homologous recombination involves a four-way cross-strand intermediate known as Holliday junction (HJ), which is recognized, processed, and resolved by a specific set of proteins. RuvA, a prokaryotic HJ-binding protein, is known to stabilize the square-planar conformation of the HJ, which is otherwise a short-lived intermediate. Despite much progress being made regarding the molecular mechanism of RuvA-HJ interactions, the mechanochemical aspect of this protein-HJ complex is yet to be investigated. Here, we employed an optical-tweezers-based, single-molecule manipulation assay to detect the formation of RuvA-HJ complex and determined its mechanical and thermodynamic properties in a manner that would be impossible with traditional ensemble techniques. We found that the binding of RuvA increases the unfolding force (F_unfold) of the HJ by ∼2-fold. Compared with the ΔG_unfold of the HJ alone (54 ± 13 kcal/mol), the increased free energy of the RuvA-HJ complex (101 ± 20 kcal/mol) demonstrates that the RuvA protein stabilizes HJs. Interestingly, the protein remains bound to the mechanically melted HJ, facilitating its refolding at an unusually high force when the stretched DNA molecule is relaxed. These results suggest that the RuvA protein not only stabilizes the HJs but also induces refolding of the HJs. The single-molecule platform that we employed here for studying the RuvA-HJ interaction is broadly applicable to study other HJ-binding proteins involved in the critical DNA repair process.

Figure 1

Preparation and characterization of DNA construct. (A) Flow chart illustrating the key steps involved in the making of the DNA constructs, which includes generation of two DNA fragments via PCR and annealing and ligation of these fragments with immobile HJ central portion obtained by annealing of synthetic single-stranded DNAs. Biotin and digoxigenin were incorporated to the opposite ends of the DNA construct via modified primers. (B) Agarose gel characterization of the full-length DNA construct (4305 bp). The DNA fragments from purified PCR are shown in lanes 1 and 2, and the product of the three-piece ligation (two PCR fragments along with HJ central portion) using T4 DNA ligase is shown in lane 3. The ligated product is highlighted with an arrow. (C) Schematic of optical tweezers setup. The DNA is tethered between the two surface-functionalized beads via streptavidin/biotin linkage on one end and digoxigenin/anti-digoxigenin antibody on the other. The HJ analog is highlighted in blue, which can reversibly unfold/refold during mechanical stretch/relax cycle. (D) Typical force versus extension (F-X) curve for the HJ construct with the HJ unfolding/refolding events. Note the unfolding and refolding events, which correspond to the melting and rehybridization of the HJ, indicated with arrows. To see this figure in color, go online.

Figure 2

Typical force-extension (F-X) curves for the HJ alone (A) and for the HJ in the presence of RuvA (B). Several curves are shown to demonstrate the reproducibility of observation. The F-X curves are shifted horizontally for clarity. The unfolding/refolding portion of the curves is highlighted with a shaded-gray background. The F-X curves for RuvA-HJ showed two unfolding events, one small feature at around the 10–22-pN range and another more obvious unfolding feature at around 40 pN. The smaller feature is highlighted in a zoomed-in view (C). The observation of a clear hysteresis due to the unfolding and refolding events between the stretching and relaxing curve, respectively, allowed us to visually pick these transitions. To see this figure in color, go online.

Figure 3

Force histograms. (A) Unfolding (top) and refolding (bottom) force histogram for HJ only (n = 130). (B) Unfolding (top) and refolding force histograms in the presence of RuvA (30 nM, n = 50). The value of “n” represents the total number of F-X curves used in this data set. The number of single molecules used for the analysis of HJ and RuvA-HJ complex were 31 and 11, respectively, with a maximum of eight unfolding/refolding events per molecule. The black curve represents Gaussian fitting of the data. To see this figure in color, go online.

Figure 4

Change-in-contour length (ΔL) due to unfolding of HJ alone (n = 86) (A) or in the presence of RuvA (n = 25) (B–D). (B) is the total of the individual ΔL-values of the two unfolding events on each +RuvA F-X curve. These individual event values are separated into low-force (C) and high-force (D) histograms. The ΔL-value was obtained from the change-in-extension (Δx) due to unfolding of the HJ at a given force using Eq. 1. The black curves represent the Gaussian fit of the data. To see this figure in color, go online.

Figure 5

(A) Two types of F-X curves observed for RuvA-HJ complex. Type I curve represents two-step unfolding, whereas type II curve represents one-step unfolding of HJ. (B) Plausible model for the RuvA-HJ interaction based on type I and type II unfolding events. In type I, refolding event occurs at a higher force than that of the HJ alone, suggesting RuvA-assisted refolding of the HJ (88% population). This observation also suggests that the protein RuvA remains bound to the mechanically stretched DNA, at least up to ∼39 pN. In contrast, type II refolding occurs at a similar force level to the HJ alone, indicating the full dissociation of protein from the DNA in a small fraction of molecules (∼8%). In addition, a small fraction (∼4%) of molecules showed only the low-force (∼20 pN) unfolding event consistent with HJs with no bound RuvA. To see this figure in color, go online.