The Inability to Disassemble Rad51 Nucleoprotein Filaments Leads to Aberrant Mitosis and Cell Death

Abstract

The proper maintenance of genetic material is essential for the survival of living organisms. One of the main safeguards of genome stability is homologous recombination involved in the faithful repair of DNA double-strand breaks, the restoration of collapsed replication forks, and the bypass of replication barriers. Homologous recombination relies on the formation of Rad51 nucleoprotein filaments which are responsible for the homology-based interactions between DNA strands. Here, we demonstrate that without the regulation of these filaments by Srs2 and Rad54, which are known to remove Rad51 from single-stranded and double-stranded DNA, respectively, the filaments strongly inhibit damage-associated DNA synthesis during DNA repair. Furthermore, this regulation is essential for cell survival under normal growth conditions, as in the srs2Δ rad54Δ mutants, unregulated Rad51 nucleoprotein filaments cause activation of the DNA damage checkpoint, formation of mitotic bridges, and loss of genetic material. These genome instability features may stem from the problems at stalled replication forks as the lack of Srs2 and Rad54 in the presence of Rad51 nucleoprotein filaments impedes cell recovery from replication stress. This study demonstrates that the timely and efficient disassembly of recombination machinery is essential for genome maintenance and cell survival.

Keywords: Rad51 nucleoprotein filament disassembly; Rad54 and Srs2; budding yeast; mitotic bridges; stalled replication forks.

Figure 1

Rad54 and Srs2 are both required for DSB repair by SSA. (A) The conditional srs2Δ and rad54Δ synthetic lethality system. Cells pre-grown on YPRAF (YP + raffinose) were serially diluted (5× step) in YP and frogged on YPGAL (YP + galactose) and YPD (YP + glucose) agar plates. Strains used: NK6933, NK7200, NK7204, and NK7208. (B) SSA system before the induction of a DSB (top) and after its repair by SSA (bottom). Light grey boxes indicate the homologies predominantly used for SSA. As one of the direct repeats (grey boxes) used for the DSB repair and the sequence between the repeats are lost during SSA, the KAN marker in the SSA system allows to distinguish between SSA and Non-Homologous End Joining (NHEJ) repair events: KAN is always lost during the former but rarely during the latter. Double-headed arrows span the DNA fragments formed by the EcoRI + SalI restriction digest as well as the HO cleavage in vivo and monitored by the Southern blotting shown in panel E, using a probe (red boxes) that matches the homologies highlighted in grey. (C) A schematic of the DNA dynamics during SSA. After the initial resection of the DNA around a break, the direct DNA repeats are annealed to each other and non-homologous DNA ends are cleaved off, leaving two ssDNA gaps which are then filled in by DNA polymerases. Direct repeats are shown in red. The DNA synthesis at the post-annealing step is represented by dotted lines. (D) The SSA efficiency in P_GAL1-RAD55, P_GAL1-RAD55 srs2Δ, and P_GAL1-RAD55 rad54Δ mutants determined by plating. Average ± SD of at least three biological repeats is shown for each genotype. Strains used: NK6724-NK6726; NK7291-NK7293; NK7424-NK7428; NK7188-NK7190. (E) Representative images of the Southern blots obtained using the samples collected during the time-course experiments analysed in panels (F,G). Blots were hybridised to two probes. One was specific to the URA3 locus (see panel (B), red boxes) and allowed to monitor the DNA dynamics at the repair site on CHRV, while the other one hybridised to the ARS1 locus on CHRIV and was used for normalisation. (F) Quantitative analysis of the SSA repair product formation over the time course in the P_GAL1-RAD55 background using Southern blotting. The average ± SD of at least three biological repeats is shown for each time point of each genotype. Asterisks describe the statistical significance of the difference between the value of the P_GAL1-RAD55 control and the value of each mutant derivative at the time point 4 h. Strains used: NK6724-NK6726; NK7188-NK7190; NK7291-NK7293; NK7295-NK7297. (G) Quantitative analysis of the SSA repair product formation over the time course in the P_GAL1-RAD51 background using Southern blotting. Same colour-coding was used as in panel (F). The average ± SD of at least three biological repeats is shown for each time point of each genotype. Asterisks describe the statistical significance of the difference between the value of the P_GAL1-RAD51 control and the value of each mutant derivative at the time point 4 h. The p values of the two-sample Student’s t-test are presented as follows: ns (p > 0.05), ** (p ≤ 0.01), *** (p ≤ 0.001). Strains used: NK5858-NK5860; NK5861-NK5863; NK5864-NK5866; NK5868-NK5870.

Figure 2

In the absence of Rad54 and Srs2, Rad51 accumulates on DNA. (A) Analysis of the Rad51 presence in the chromatin fraction by Western blotting. Actin was used as a non-chromatin protein marker and histone H2B as a chromatin marker. The faint band observed in the rad51Δ samples at the position where Rad51 normally runs comes from a cross-reacting protein (i.e., background) with a slightly lower gel mobility than Rad51. A representative set of Western blot images from one of the repeats. (B) A data summary plot for the relative Rad51 levels in the total cell lysates and chromatin fractions of the galactose-induced P_GAL1-RAD55 srs2Δ rad54Δ cells and the appropriate control strains. Histone H2B was used as a normaliser to determine the relative amounts of Rad51 in different samples. The values were then normalised to the average value of P_GAL1-RAD55 strains for the total protein and the chromatin fractions separately. The average ±SD of three biological repeats is shown for each strain with the P_GAL1-RAD55 background and three technical repeats for the rad52Δ control. The red dotted line corresponds to the non-specific presence of Rad51 in the chromatin fraction based on the fact that Rad51 filaments cannot be formed in rad52Δ cells. The p values of the two-sample Student’s t-test are presented as follows: ns (p > 0.05), *** (p ≤ 0.001). Strains used: NK6933-NK6935; NK7200-NK7202; NK7204-NK7206; NK7208-NK7210; NK81.

Figure 3

A single round of DNA replication in the cells with unregulated Rad51 results in the activation of the DNA damage checkpoint and aberrant mitoses. (A) A schematic of the time-course experiments performed to investigate the cell cycle progression and the DNA damage checkpoint dynamics in the P_GAL1-RAD55 srs2Δ rad54Δ rad53-13Myc mutants. The time points correspond to the time (in h) before (negative) and after (positive) the release from the first α-factor arrest. (B) FACS profiles of the samples collected during the time-course experiment described in panel A. Vertical yellow lines specify the positions of the 1N and 2N peaks from left to right, respectively. Strains used: NK9134, NK9135. (C) Rad53 Western blotting analysis. The hyper-phosphorylation of Rad53 characteristic of the DNA damage checkpoint activation is seen in the galactose-induced P_GAL1-RAD55 srs2Δ rad54Δ cells but not in the control culture grown in raffinose. The time points correspond to the ones in the panels (A,B). (D) Representative images of the mitotic bridges observed in the P_GAL1-RAD55 srs2Δ rad54Δ htb2-mCherry cells grown in the presence of galactose. Two sets of images are shown. In each set, the image on the left shows the overlap of the bright field with the red channel while the image with the red channel on its own is on the right. (E) Quantification analysis of the nuclear divisions for the mitotic bridge formation. Only the first cell divisions after the P_GAL1-RAD55 induction were scored. The data bars show average ± SD based on three biological repeats for each genotype. A minimum of 75 mitoses were scored for each genotype. Strains used: NK10689-NK10691; NK10692-NK10694.

Figure 4

Presence of the Rad51 nucleoprotein filaments confers HU sensitivity in srs2Δ and srs2Δ rad54Δ yeast. (A) A schematic outline of the experiment testing cell recovery from the HU-induced replication stress, in the presence and absence of Rad51 nucleoprotein filaments. (B) Cell survival after a transient HU treatment with or without the expression of RAD55. The survival is calculated as the ratio of the colonies grown after the treatment either in raffinose (YPRAF + HU/YPRAF, left bars in each coloured pair) or in galactose (YPGAL + HU/YPGAL, right bars in each coloured pair). The data show average ± SD based on three biological repeats for each genotype. The p values of the two-sample Student’s t-test are presented as follows: ns (p > 0.05), * (p ≤ 0.05), *** (p ≤ 0.001). Strains used: NK6933-NK6935; NK7200-NK7202; NK7204-NK7206; NK7208-NK7210.

Figure 5

A model explaining the complementary functions of Srs2 and Rad54 in the damage-associated DNA synthesis.

Figure 6

Two models (A,B) explaining how unregulated Rad51 nucleoprotein filaments might permanently stall challenged replication forks and cause mitotic bridges during cell divisions.