Long stretches of ssDNA have also been
Long stretches of ssDNA have also been observed in rad53-1 mutant angiogenic under replication stress using EM images (Sogo et al., 2002). Deletion of EXO1, a 5′-to-3′ exonuclease, can suppress the generation of ssDNA detected in rad53-1 mutant cells (Cotta-Ramusino et al., 2005). It was thus proposed that ssDNA is generated through the nucleolytic processing of nascent lagging strands. Intriguingly, deletion of EXO1 in rad53-1 mutant cells had no effect on the bias in leading- and lagging-strand synthesis in rad53-1 mutant cells under replication stress observed here (Figure S2). Therefore, ssDNA detected using BrdU-IP-ssSeq and RPA-ssSeq appears to differ from ssDNA detected previously. It has been shown that the balance between DNA synthesis mode and exonuclease mode of Pol ε shifts toward the 3′-to-5′ exonuclease mode when dNTP concentration is lowered in vitro (Ganai et al., 2015). To test if nascent leading strands are processed by a 3′-to-5′ exonuclease in rad53-1 mutant cells, we analyzed how inactivation of the 3′-to-5′ exonuclease activities of Mre11, Pol2 (catalytic subunit of Pol ε), or Pol3 (catalytic subunit of Pol δ) may affect leading- and lagging-strand DNA synthesis in rad53-1 mutant cells. Inactivation of the exonuclease activity of Mre11, Pol2, or Pol3 did not affect the biased pattern of BrdU-IP-ssSeq peaks in rad53-1 mutant cells dramatically (Figure S3). These results indicate that the generation of ssDNA in rad53-1 mutant cells is not due to nucleolytic processing of newly synthesized leading-strand DNA by the exonuclease activity of Mre11, Pol2, or Pol3. Mrc1 is involved in Rad53 activation (Alcasabas et al., 2001), and in mrc1Δ mutant cells, Cdc45 and other proteins at DNA replication forks move ahead of the site of actual DNA synthesis (Katou et al., 2003). Therefore, we tested the hypothesis that the replicative DNA helicase, MCM, may unwind the double-strand template DNA while leading-strand synthesis is uncoupled from lagging-strand synthesis in rad53-1 cells under replication stress. To do this, we analyzed the chromatin distribution of the MCM subunit, Mcm6, in WT and rad53-1 cells under replication stress. Briefly, WT or rad53-1 mutant yeast cells were first arrested in G1 phase and then released into S phase in the presence of both HU and BrdU. ChIP-seq was performed to monitor Mcm6 distribution (Figure 3A). To correlate Mcm6 chromatin binding with new DNA synthesis, we simultaneously performed BrdU-IP-Seq to mark replicated regions. As shown in Figures 3B and 3C, and consistent with published observations (Yu et al., 2014), the BrdU track length was similar between WT and rad53-1 mutant cells. In contrast, Mcm6 ChIP-seq peaks in rad53-1 mutant cells were significantly wider than the ones in WT cells (Figures 3B and 3D). To exclude the possibility that this is due to differences in aberrant DNA synthesis in rad53-1 mutant cells, we normalized Mcm6 ChIP-seq peak width to the corresponding BrdU-IP-Seq peak width at each replication origin. The box-dot plots indicated that the ratios of Mcm6 ChIP-seq over BrdU-IP-Seq at early replication origins in rad53-1 mutant cells were significantly larger than in WT cells (Figure 3E). If we assume that the MCM helicase and DNA synthesis progress over similar distances in WT cells, these results indicate that MCMs, which translocate on the leading-strand template (Fu et al., 2011), moves ahead of the site of actual DNA synthesis in rad53-1 mutant cells, suggesting that the DNA template is unwound in the absence of leading-strand synthesis in rad53-1 mutant cells under replication stress. We also analyzed how the leading-strand DNA polymerase, Pol ε (Pursell et al., 2007), is distributed around HU-stalled forks in WT and rad53-1 mutant cells using ChIP-seq. In WT cells, Pol ε ChIP-seq peaks exhibited bifurcated peaks surrounding replication origins, indicating that Pol ε moves away from DNA replication origins bi-directionally. This pattern was compromised in rad53-1 mutant cells (Figures 3B and 3F). Importantly, we observed that the Pol ε ChIP-seq peak track length in rad53-1 mutant cells was much longer than that in WT cells (Figures 3F and 3G). These results suggest that, in rad53-1 mutant cells, Pol ε associate with HU-stalled DNA replication forks and travel with the MCM helicase under replication stress.