Two distinct conformational states define the interaction of human RAD51‐ATP with single‐stranded DNA

Abstract An essential mechanism for repairing DNA double‐strand breaks is homologous recombination (HR). One of its core catalysts is human RAD51 (hRAD51), which assembles as a helical nucleoprotein filament on single‐stranded DNA, promoting DNA‐strand exchange. Here, we study the interaction of hRAD51 with single‐stranded DNA using a single‐molecule approach. We show that ATP‐bound hRAD51 filaments can exist in two different states with different contour lengths and with a free‐energy difference of ~4 kBT per hRAD51 monomer. Upon ATP hydrolysis, the filaments convert into a disassembly‐competent ADP‐bound configuration. In agreement with the single‐molecule analysis, we demonstrate the presence of two distinct protomer interfaces in the crystal structure of a hRAD51‐ATP filament, providing a structural basis for the two conformational states of the filament. Together, our findings provide evidence that hRAD51‐ATP filaments can exist in two interconvertible conformational states, which might be functionally relevant for DNA homology recognition and strand exchange.

Disassembly rate (s ) -1 Figure EV1. Disassembly of hRAD51 from ssDNA does not depend on the presence of ATP or ADP in the filament.
A Fluorescence images and kymographs of hRAD51 disassembling from ssDNA for filaments assembled in 20 mM Tris pH 7.5, 40 mM KCl, 10 mM Mg(OAc) 2 , 2 mM ATP and 10 mM DTT and disassembled in 20 mM Tris pH 7.5, 100 mM KCl, 1 mM MgCl 2 , 1 mM ATP and 10 mM DTT (right panel) or filaments assembled in 20 mM Tris pH 7.5, 40 mM KCl, 10 mM Mg(OAc) 2 , 2 mM ADP and 10 mM DTT and disassembled in 20 mM Tris pH 7.5, 100 mM KCl, 1 mM MgCl 2 , 1 mM ADP and 10 mM DTT (right panel). Images are typical examples of 16 (for the ATP condition) and 11 (for the ADP condition). Scale bars: 2 lm (horizontal) and 5 s (vertical). B Normalized integrated fluorescence intensity of the images shown in (A) over time. Exponential fits to these traces give disassembly rates of (6.7 AE 0.3) 10 À4 s À1 (black dataset; in the presence of ATP) and (17 AE 4) 10 À4 s À1 (red dataset; in the presence of ADP). Coloured edges in (A) show colour of corresponding force curve. C Disassembly rate is correlated with the initial coverage of the DNA molecule. Red closed circles: measured in ATP at 5 pN; red open circles: measured in ATP at 20 pN; red closed triangles: measured in ATP at 50 pN; red open triangles: measured in ATP at 75 pN; black closed circles; measured in ADP in 100 mM KCl and 1 mM MgCl 2 ; black open circles; measured in ADP in 100 mM KCl and 10 mM MgCl 2 ; black triangles: measured in 1 mM MgCl 2 . Blue line: linear fit with a slope of (À21 AE 3) 10 À10 s À1 /AU (Pearson's correlation coefficient of the fit: À0.84) that was used to correct observed disassembly rates for differences in initial coverage. D Average disassembly rates in ATP or ADP conditions after correcting for differences in initial coverage. Since the error bars of the black and red datasets overlap, there is no significant difference between the disassembly rate in the presence of ATP or ADP. All coverages were scaled to 400,000 using a linear approximation based on the fit in (C). and 5 s (vertical). B Normalized integrated fluorescence intensity of the images shown in (A) over time. Exponential fits to these traces give disassembly rates of (8 AE 1) 10 À4 s À1 (black dataset), (6.7 AE 0.3) 10 À4 s À1 (red dataset) and (6.7 AE 0.2) 10 À4 s À1 (blue dataset). Coloured edges in ( Force-extension and force-relaxation curve measured in a buffer containing ATP and Mg 2+ . Under these conditions, ATP hydrolysis and NPF disassembly can occur. We observe a slight hysteresis between extension and relaxation curves. However, the disassembly rate under these conditions is relatively high, such that the assumption that the amount of hRAD51 bound remains constant during one extension-relaxation cycle is no longer valid. Therefore, under these conditions, no quantitative analysis (such as shown for other conditions in Fig 4) of the hysteresis and the structural transitions can be performed. Scale bar: 2 lm.
Source data are available online for this figure.
The . Under these conditions, ATP hydrolysis can occur, and thus, hRAD51 can disassemble from the ssDNA. Therefore, the difference between the extension and relaxation curves (A), the total fluorescence intensity (B) and hysteresis area (C) decrease over time. In (A), the blue curve is indistinguishable from the grey curve of bare ssDNA. The disassembly rate can be determined either by an exponential fit to the fluorescence data (B), yielding, after correcting for photobleaching, a rate of (3.5 AE 0.3) 10 À4 s À1 , or by an exponential fit to the hysteresis data (C), yielding a rate of (4 AE 1) 10 À4 s À1 . Data shown is a representative example of six identical experiments. D-F Same experiments as in (A-C) in the presence of Ca 2+ and Mg 2+ (20 mM Tris pH 7.5, 2 mM CaCl 2 , 10 mM Mg(OAc) 2 , 1 mM DTT). NPFs were formed in the presence of ATP, but there was no ATP or ADP in the observation channel. Under these conditions, ATP hydrolysis can occur, and thus, hRAD51 can disassemble from the ssDNA, but reloading of ATP to the NPF after ATP hydrolysis and ADP release is impossible. The differences between the extension and relaxation curves (D), the total fluorescence intensity (E) and hysteresis area (F) decrease over time. In (D), blue curve is indistinguishable from grey curve of bare ssDNA. The disassembly rate can be determined by an exponential fit to the data in (E), giving a rate of (4 AE 1) 10 À4 s À1 , or by an exponential fit to the data in (F), giving a rate of (4 AE 1) 10 À4 s À1 .
Source data are available online for this figure.  Figure EV5. Lengths of the different hRAD51-ssDNA NPF states. The EMBO Journal Two conformations of RAD51 on ssDNA Ineke Brouwer et al