Pol is responsible for the bulk of leading strand
Pol ε is responsible for the bulk of leading-strand synthesis in vivo (Daigaku et al., 2015, Nick McElhinny et al., 2008, Pursell et al., 2007) and physically associates with CMG (Langston et al., 2014, Sengupta et al., 2013, Sun et al., 2015, Zhou et al., 2017). Furthermore, leading-strand synthesis rates matching those observed in vivo can only be attained by a reconstituted replisome when Pol ε is synthesizing the leading strand in conjunction with PCNA (Yeeles et al., 2017). Therefore, once the primer for leading-strand replication has been synthesized, the 3ʹ end must be coupled to CMG-bound Pol ε (CMGE) before rapid and efficient leading-strand replication can commence. Multiple non-mutually exclusive mechanisms might account for this process (Kunkel and Burgers, 2017). The simplest involves direct primer transfer from Pol α to CMGE. Support for this mechanism comes from observations that Pol α can prime the leading-strand template at model replication forks with CMG (Georgescu et al., 2015), and rapid and efficient leading-strand synthesis is observed in in vitro replication reactions where Pol α and Pol ε are the only DNA polymerases (Yeeles et al., 2017). In contrast, in vivo (Daigaku et al., 2015, Garbacz et al., 2018) and in vitro (Yeeles et al., 2017) experiments have indicated that, in addition to its role in lagging-strand synthesis, Pol δ might also participate in the e-mail of leading-strand replication via a polymerase switch mechanism, with the 3ʹ end of the nascent leading strand sequentially transferred from Pol α to Pol δ to CMGE. Why such an elaborate mechanism may be required is unknown, as is the frequency by which the two pathways are utilized.
In this study, we have addressed these outstanding questions by mapping start sites for leading-strand replication at two S. cerevisiae replication origins using a reconstituted replication system (Taylor and Yeeles, 2018, Yeeles et al., 2015, Yeeles et al., 2017), determining the basis of Pol α recruitment to these sites, and defining the pathway by which the 3ʹ end of the nascent leading strand is connected to CMGE following primer synthesis. This has enabled us to elucidate the mechanism of bidirectional leading-strand synthesis establishment at eukaryotic DNA replication origins.
Discussion In this study, we have determined the pathway by which the 3ʹ end of the nascent leading strand is connected with CMGE after priming, revealing that Pol δ likely plays a crucial role in establishing all continuously synthesized leading strands at eukaryotic replication origins (Figures 1 and 2). Initiation-site mapping experiments have identified start sites for leading-strand replication at two S. cerevisiae origins. Synthesis is predominantly initiated outside the origin sequence; Left leading strands are started to the right, and Right leading strands are started to the left (Figures 3 and 4). This distribution strongly suggests that leading strands are established from lagging-strand primers synthesized at replication forks on opposite sides of the origin. We provide direct evidence to support this conclusion: first, delaying Pol α addition to reactions lengthened, rather than shortened, leading-strand products (Figure 5); second, the two replisomes remain interdependent downstream of CMG activation, because placing a single CPD in the lagging-strand template of the leftward fork blocked the establishment of rightward leading strands (Figure 6). The mechanism of priming at origins that we have uncovered provides a clear mechanistic basis for Pol δ function in establishing continuous leading-strand synthesis.
Acknowledgments We thank J. Diffley for protein-expression strains and J. Sale, H. Williams, and members of the Yeeles lab for helpful discussions and comments on the manuscript. This work was supported by the Medical Research Council (MC_UP_1201/12).