Why then is leading strand DNA synthesis reduced relative
Why then is leading-strand DNA synthesis reduced relative to lagging-strand synthesis in rad53-1 mutant glycine receptors under replication stress? To gain insight into this question, we first determined whether the firing of late origins in rad53-1 mutant cells contributes to compromised leading-strand synthesis. Rad53 phosphorylates multiple residues on Sld3 and Dbf4 in response to replication stress, which inhibits the firing of late replication origins (Lopez-Mosqueda et al., 2010, Zegerman and Diffley, 2010). Thus, late replication origins fire in cells expressing the Rad53 phosphorylation-site mutants, sld3-A dbf4-A, in the presence of HU (Sheu et al., 2016, Zegerman and Diffley, 2010) (Figure S4A). However, sld3-A dbf4-A mutant cells exhibited no significant bias in the distribution of BrdU-IP-ssSeq peaks at either early or late replication origins (Figures 4A–4C), suggesting that late replication origin firing is not the reason for the compromised leading-strand synthesis in rad53-1 mutant cells.
Sml1 is an inhibitor of RNR, which regulates dNTP synthesis. In WT cells, Sml1 is degraded in response to replication stress in a Rad53-dependent manner. Deletion of SML1 results in increased dNTP pools and suppresses the lethality of RAD53 deletion mutant cells (Zhao et al., 1998). Therefore, we next tested whether deletion of SML1 had any impact on uncoupled DNA synthesis in rad53-1 mutant cells. Synthesis of leading and lagging strands was similar in sml1Δ cells (Figures 4A and 4B). Remarkably, deletion of SML1 in rad53-1 mutant cells completely suppressed the lagging-strand bias of BrdU-IP-ssSeq peaks at early replication origins (Figures 4A–4C and Figure S4B), and this suppression is unlikely to be due to cell synchrony (Figures S4C and S4D). We noticed that a small lagging-strand bias exists at late replication origins in rad53-1 sml1Δ cells, suggesting that synthesis at leading and lagging strands remains partially uncoupled at late replication origins in the rad53-1 sml1Δ cells. Taken together, these results indicate that the uncoupled DNA synthesis is due in part to reduced dNTP levels in rad53-1 mutant cells.
While Pol δ likely has a role in leading-strand DNA synthesis (Johnson et al., 2015, Yeeles et al., 2017), Pol ε performs the bulk of leading-strand synthesis in vivo (Kunkel and Burgers, 2008). Therefore, we analyzed the DNA synthesis activity of Pol ε and Pol δ at different dNTP concentrations using a recently reconstituted origin-dependent budding yeast in vitro DNA replication system (Devbhandari et al., 2017). In this system, primer extension by Pol α is restricted by DNA polymerase clamp loader and clamp (RFC/PCNA), while both Pol ε and Pol δ can contribute to the synthesis of leading and lagging strands. Specific DNA synthesis by Pol ε can be monitored by omission of Pol δ from the system, while specific DNA synthesis by Pol δ can be monitored in the presence of the catalytic polymerase mutant, Pol εpol−, which maintains the DNA polymerase-independent origin activation function of Pol ε. Using constant dNTP ratios that mimic those measured in vivo (Sabouri et al., 2008), we observed that the Pol ε requires ∼2- to 4-fold higher dNTP levels for maximum extension of nascent leading strands compared to Pol δ (Figure 4D and Figure S4E). Synthesis of the short lagging-strand products by either Pol δ or Pol ε, on the other hand, was not differentially affected at any of the dNTP concentrations tested. These results support the notion that leading-strand DNA synthesis by Pol ε is more susceptible to reduced dNTP levels than lagging-strand synthesis by Pol δ.
Discussion We show that in rad53-1 mutant cells, synthesis of leading and lagging strands is uncoupled, which leads to the exposure of long stretches of single-stranded leading-strand template coated with RPA under replication stress. We observed that the replicative DNA helicase, MCM, and associated Pol ε move ahead of the site of actual DNA synthesis in rad53-1 mutant cells under replication stress. Moreover, the uncoupled synthesis in rad53-1 mutant cells is suppressed by increased dNTP levels, and Pol ε DNA synthesis is specifically compromised at low dNTP concentrations in vitro. These results support a model whereby Rad53 couples leading- and lagging-strand synthesis under replication stress at least in part through the inhibition of excessive template DNA unwinding by the MCM helicase at stalled replication forks as well as the upregulation of dNTP levels needed for DNA synthesis by Pol ε.