In vitro role of Rad54 in Rad51-ssDNA filament-dependent homology search and synaptic complexes formation

Abstract

Homologous recombination (HR) uses a homologous template to accurately repair DNA double-strand breaks and stalled replication forks to maintain genome stability. During homology search, Rad51 nucleoprotein filaments probe and interact with dsDNA, forming the synaptic complex that is stabilized on a homologous sequence. Strand intertwining leads to the formation of a displacement-loop (D-loop). In yeast, Rad54 is essential for HR in vivo and required for D-loop formation in vitro, but its exact role remains to be fully elucidated. Using electron microscopy to visualize the DNA-protein complexes, here we find that Rad54 is crucial for Rad51-mediated synaptic complex formation and homology search. The Rad54-K341R ATPase-deficient mutant protein promotes formation of synaptic complexes but not D-loops and leads to the accumulation of stable heterologous associations, suggesting that the Rad54 ATPase is involved in preventing non-productive intermediates. We propose that Rad51/Rad54 form a functional unit operating in homology search, synaptic complex and D-loop formation.

Fig. 1

D-loop assay by electron microscopy (EM) and by gel electrophoresis. a Scheme of the D-loop assay. b Representative EM images of the reaction substrates with insets of schematic drawings of the molecules: the Rad51 filament formed on 5´ DNA junction and the homologous supercoiled DNA. c Reaction substrates in the deproteinized samples suitable for EM analysis: 5´ DNA junction with 3´ ssDNA part covered with RPA (top), supercoiled homologous DNA donor (middle). d EM representative view of the D-loop reaction after 20 min incubation with homologous DNA and Rad54. e D-loop observed in the deproteinized sample, with an arrow pointing to the synaptic contact between molecules (bottom). For additional images see Supplementary Fig. 2a. f Imunolabeling of Rad54 in the D-loop reaction using rabbit anti-Rad54 primary antibody and immunolabeled anti-rabbit secondary antibody followed with DNA-protein complexes purification. The same joint molecule is imaged using dark-field mode in the top image while using brightfield (bottom image) allowing the detection of the gold beads. Seven 1% of labeled joint molecules where counted (n = 2, 100 molecules each). g D-loop assay time course after donor DNA addition using 175 nM of Rad54 h D-loop assay with Rad54 titration in 20 min reactions. D-loop yield peaks at 175 nM of Rad54, which corresponds to 7 Rad54 proteins per invading or donor dsDNA molecule. i Graph of data in c (line plot, from electrophoresis gel quantification, n = 5) and from the EM quantification (red bars plot, at three time points: 2, 20, and 30 min. n = 3, 300 molecules each). Error bars indicate standard deviations. Gray shading corresponds to the three time points chosen for further experiments. In all EM pictures, the white bar represents 100 nm. In a, d: I invading strand (blue), D donor DNA (black), and h heteroduplex (red). Source data are provided as a Source Data file

Fig. 2

Rad54 promotes the formation of synaptic complexes and D-loops as visualized by EM of DNA-protein complexes. a–c Representative EM pictures of a reactions without Rad54, b Rad51-mediated synaptic complexes and D-loops c from reaction containing Rad54. In all EM pictures, white bars represent 100 nm. For additional images see Supplementary Fig. 1b, c. d Quantification of synaptic complexes and D-loops in DNA-protein samples. Error bars in d indicate standard deviation from three independent experiments. The percentage of molecules was determined by the ratio between the population of synaptic complexes or D-loops and the sum of these two categories of molecules plus free 5´ DNA junctions. (n = 500 molecules for each experiment). Note that this division into two categories was performed by EM visual analysis therefore rare subjective analysis-associated errors cannot be rolled out. e Contact length for synaptic complexes and deproteinized D-loops at 20 min, showing median with interquartile range. Source data are provided as a Source Data file

Fig. 3

Rad54K341R mutant leads to higher yield of protein-mediated engagements with heterologous DNA than WT protein. a–c Representative EM pictures of protein-mediated engagements generated during D-loop assays with a Rad54K341R and heterologous DNA, b wild-type Rad54 and heterologous DNA, and c Rad54K341R and homologous DNA. d Quantification of joint molecules generated in assays with wild-type Rad54 or Rad54K341R mutant at 20 min (n = 3, 250 molecules each n). Error bars indicate standard deviation. White bars represent 100 nm. Source data are provided as a Source Data file

Fig. 4

RecA promotes heteroduplex DNA formation within the nucleoprotein filament without dissociating. a Time course of RecA D-loop assay. b Quantification of RecA time time courses; n = 3, error bars represent standard deviations. c Representative EM pictures of RecA-mediated joint molecules at 20 min. White bar represents 100 nm. Source data are provided as a Source Data file

Fig. 5

Duplex capture by Rad51 filaments is dependent on Rad54. a Duplex capture assay. Proteins were incubated at 30 °C, and the order of addition was 10 min Rad51 (260 nM), 9 min RPA (27 nM), 1 min Rad54 (100 nM or as indicated), 10 min 3 kb supercoiled plasmid DNA (1 nM molecules, heterologous unless noted), 1 min streptavidin-coated magnetic particles before magnetic capture, followed by deproteinization of the isolated bead-bound material for gel electrophoresis/staining to measure captured plasmid DNA. b Titration of Rad54 in the duplex capture reaction. c RecA supported duplex capture. 1 mM ATP or ATPγS was included in reactions. d Control reactions performed with 100 nM Rad54 wild-type or Rad54K341R protein. Homologous and heterologous plasmids are both ~3 kb. The homologous plasmid has 607 nt of homology to the ds98-607-ss78 (=ds98-ss685). D-loops migrate slightly slower than nicked plasmid, as previously established. Representative gel image (inverted signal) is shown, with quantitation of replicate experiments. e Time course of duplex capture by Rad51 with 100 nM Rad54. Beads were added one minute prior to the indicated time points, at which they were captured. Data in b–e are means ± standard deviation of three or more independent experiments. Source data are provided as a Source Data file

Fig. 6

Rad51 and Rad54 cooperative model for homology search and D-loop formation. (i) During homology search, Rad54 promotes DNA probing. The invading DNA (light red) uses Rad54 to bridge the Rad51 filament to dsDNA during the homology search. Rad54 ATPase activity is not required but may enhance probing. (ii) Persistent associations with heterologous DNA (blue, right arrow) may be prevented or dissociated by Rad54 in an ATPase-dependent fashion. Rad54 ATPase exerts quality control to promote homologous pairing. (iii) Rad54 is required for synaptic complex formation without strict requirement for ATPase activity, and (iv) converts such complexes into D-loops dependent on ATP hydrolysis. Rad51 left on the ssDNA outside of the heteroduplex region after removal during heteroduplex formation may able to repolymerize back into the synaptic region. Note that this is a cartoon representation not meant to model the true scale and structure of the Rad51 filament or Rad54 protein arrangement in the depicted intermediates