Vincent Pagès

Stalled replication forks are processed in post-replication repair by homologous recombination, fork regression, and translesion DNA synthesis. However, the regulation of pathway usage is not fully understood. Rad51 protein maintains genomic stability through its roles in recombination and in protecting stalled replication forks. We report isolation of mutations in Saccharomyces cerevisiae Rad51 that shift post-replication repair from recombination to alternate pathways including mutagenic translesion synthesis. Rad51-E135D and Rad51-K305N show near normal in vitro recombination despite changes in their DNA binding profiles, in particular to dsDNA. The mutants lead to a defect in Rad51 recruitment to stalled forks in vivo as well as a defect in the protection of dsDNA from degradation by Dna2-Sgs1 and Exo1 in vitro. Together, the evidence suggests that Rad51 binding to duplex DNA is critical to control pathway usage at stalled replication forks.

Following encounter with an unrepaired DNA lesion, replication is halted and can restart downstream of the lesion leading to the formation of a single-stranded DNA (ssDNA) gap. To complete replication, this ssDNA gap is filled in by one of the two lesion tolerance pathways: the error-prone Translesion Synthesis (TLS) or the error-free Homology Directed Gap Repair (HDGR). In the present work, we evidence a role for the RecBC complex distinct from its canonical function in homologous recombination at DNA double strand breaks. We show that upon lesion encounter RecBC (independently of its catalytic activity and of the RecD subunit) is required to protect the nascent DNA in order to promote efficient lesion bypass. In the absence of RecBC, our data indicate that the nuclease ExoI can access and degrade the nascent DNA, affecting both TLS and HDGR mechanisms. We show that the recruitment of RecBC becomes particularly important at strong blocking lesions, when postreplicative ssDNA gaps persist and are covered by the ssDNA binding proteins. This protective role of RecBC is reminiscent of the role of BRCA2 in protecting the nascent DNA in human cells, highlighting once again the evolutionary conservation of DNA replication mechanisms across all living organisms.

DNA lesions are a threat to genome stability. To cope with them during DNA replication, cells have evolved lesion bypass mechanisms: Translesion Synthesis (TLS), which allows the cell to insert a nucleotide directly opposite the lesion, with the risk of introducing a mutation, and error-free damage avoidance (DA), which uses homologous recombination to retrieve the genetic information from the sister chromatid. In this study, we investigate the timing of lesion bypass in yeast and its implications for the accuracy of the process. Our findings reveal that DNA polymerase η can bypass common, UV-induced cyclobutane pyrimidine dimers at the fork, immediately after encountering the blocking lesion. In contrast, TLS at (6–4) photoproducts and bulky G-AAF adducts, mediated by Rev1 and Pol ζ, takes place behind the fork, at post-replicative gaps that are generated downstream of the lesion after repriming. We show that in this latter situation, TLS competes with the DA pathway, thus reducing overall mutagenicity of damage bypass. Additionally, our study demonstrates that Exo1 nuclease influences the balance between TLS and DA by modulating the size of the post-replicative gaps.

Processivity clamps mediate polymerase switching for translesion synthesis (TLS). All three Escherichia coli TLS polymerases interact with the β ₂ processivity clamp through a conserved clamp-binding motif (CBM), which is indispensable for TLS. Notably, Pol IV also interacts weakly with the rim of the clamp through non-CBM residues. Ablating this “rim contact” in cells results in selective sensitivity to DNA-damaging agents, raising the question how the rim contact contributes to TLS. Here, we show that the rim contact is critical for TLS past a strong replication block but barely necessary for a weak blocking lesion. Within the in vitro reconstituted E. coli replisome, the rim mutation compromises Pol IV-mediated TLS past 3-deaza-methyl dA, a strong block, whereas barely affecting TLS past N ² -furfuryl dG, a weak block. Similar observations are also made in E. coli cells bearing a single copy of these lesions in the genome. At lesion-stalled replication forks, single-stranded DNA binding protein locally enriches Pol IV, enabling it to bind the low-affinity rim site. This interaction poises Pol IV to better compete with Pol III, the replicative polymerase, which competitively inhibits Pol IV from interacting with the clamp through its CBM. We propose that this bipartite clamp interaction enables Pol IV to rapidly resolve lesion-stalled replication at a strong block through TLS, which reduces damage-induced mutagenesis.

Stalled replication forks can be processed by several distinct mechanisms collectively called post-replication repair which includes homologous recombination, fork regression, and translesion DNA synthesis. However, the regulation of the usage between these pathways is not fully understood. The Rad51 protein plays a pivotal role in maintaining genomic stability through its roles in HR and in protecting stalled replication forks from degradation. We report the isolation of separation-of-function mutations in Saccharomyces cerevisiae Rad51 that retain their recombination function but display a defect in fork protection leading to a shift in post-replication repair pathway usage from HR to alternate pathways including mutagenic translesion synthesis. Rad51-E135D and Rad51-K305N show normal in vivo and in vitro recombination despite changes in their DNA binding profiles, in particular to dsDNA, with a resulting effect on their ATPase activities. The mutants lead to a defect in Rad51 recruitment to stalled forks in vivo as well as a defect in the protection of dsDNA from degradation by Dna2-Sgs1 and Exo1 in vitro . A high-resolution cryo-electron microscopy structure of the Rad51-ssDNA filament at 2.4 Å resolution provides a structural basis for a mechanistic understanding of the mutant phenotypes. Together, the evidence suggests a model in which Rad51 binding to duplex DNA is critical to control pathway usage at stalled replication forks.

Processivity clamps mediate polymerase switching for translesion synthesis (TLS). All three E. coli TLS polymerases interact with the β ₂ processivity clamp through a conserved clamp-binding motif (CBM), which is indispensable for TLS. Notably, Pol IV also makes a unique secondary contact with the clamp through non-CBM residues. However, the role of this “rim contact” in Pol IV-mediated TLS remains poorly understood. Here we show that the rim contact is critical for TLS past strong replication blocks. In in vitro reconstituted Pol IV-mediated TLS, ablating the rim contact compromises TLS past 3-methyl dA, a strong block, while barely affecting TLS past N ² -furfuryl dG, a weak block. Similar observations are also made in E. coli cells bearing a single copy of these lesions in the genome. Within lesion-stalled replication forks, the rim interaction and ssDNA binding protein cooperatively poise Pol IV to better compete with Pol III for binding to a cleft through its CBM. We propose that this bipartite clamp interaction enables Pol IV to rapidly resolve lesion-stalled replication through TLS at the fork, which reduces damage induced mutagenesis.

DNA lesions are a threat to genome stability. To cope with them during DNA replication, cells have evolved lesion bypass mechanisms: Translesion Synthesis (TLS), which allows the cell to insert a nucleotide directly opposite the lesion, with the risk of introducing a mutation, and error-free Damage Avoidance (DA), which uses homologous recombination to retrieve the genetic information from the sister chromatid. In this study, we investigate the timing of lesion bypass and its implications for the accuracy of the process. Our findings reveal that DNA polymerase η can bypass common, UV-induced TT-cyclobutane pyrimidine dimers at the fork, immediately after encountering the blocking lesion. In contrast, TLS at TT(6-4) photoproducts and bulky G-AAF adducts, mediated by Rev1 and Pol ζ, takes place behind the fork, at post-replicative gaps that are generated downstream of the lesion after repriming. We show that in this latter situation, TLS competes with the DA pathway, thus reducing overall mutagenicity of damage bypass. Additionally, our study demonstrates that Exo1 nuclease influences the balance between TLS and DA by modulating the size of the post-replicative gaps.

Following encounter with an unrepaired DNA lesion, replication is halted and can restart downstream of the lesion leading to the formation of a single-stranded DNA (ssDNA) gap. To complete replication, this ssDNA gap is filled in by one of the two lesion tolerance pathways: the error-prone Translesion Synthesis (TLS) or the error-free Homology Directed Gap Repair (HDGR). In the present work, we evidence a new role for the RecBC complex distinct from its canonical function in homologous recombination at DNA double strand breaks. We show that upon lesion encounter RecBC (independently of its catalytic activity and of the RecD subunit) is required to protect the nascent DNA in order to promote efficient lesion bypass. In the absence of RecBC, our data indicate that the nuclease ExoI can access and degrade the nascent DNA, affecting both TLS and HDGR mechanisms. We show that the recruitment of RecBC becomes particularly important at strong blocking lesions, when post-replicatively ssDNA gaps persist and are covered by the single-stranded DNA binding protein (SSB). This protective role of RecBC is reminiscent of the role of BRCA2 in protecting the nascent DNA in human cells, highlighting once again the evolutionary conservation of DNA replication mechanisms across all living organisms. SIGNIFICANCE STATEMENT: In the presence of unrepaired lesions, stalled replication can restart ahead of the lesion creating a single stranded DNA gap. To preserve genome stability, this gap needs to be later filled in by one of the two lesion tolerance pathways, Translesion Synthesis or Homology Directed Gap Repair. In this study, we unveil a role for the RecBC complex in protecting the nascent DNA blocked at the lesion site from degradation by the ExoI nuclease. This protection is essential for an efficient repair of the single-stranded DNA gap. In the absence of RecBC, both lesion tolerance pathways are affected and genome stability is compromised.

DNA Damage Tolerance (DDT) functions to bypass replication-blocking lesions and is divided into two distinct pathways: error-prone Translesion Synthesis (TLS) and error-free Damage Avoidance (DA). Rad5 is a multifunctional protein that is involved in these DDT processes. Saccharomyces cerevisiae Rad5 contains three well defined domains: a RING domain that promotes PCNA polyubiquitination, a ssDNA-dependent ATPase/helicase domain, and a Rev1-binding domain. Both the RING domain and the ATPase/helicase domain are conserved in human Rad5 ortholog HLTF. In this study we used domain-specific mutants to address the contribution of each of the Rad5 domains to the lesion tolerance. We demonstrate that the two critical functions of Rad5 during DNA damage tolerance are the activation of template switching through polyubiquitination of PCNA and the recruitment of TLS polymerases, and that loss of one of those functions can be compensated by increased usage of the other. We also show that, unlike previously suggested, the helicase activity does not play any role in lesion tolerance.

Processivity clamps tether DNA polymerases to DNA, allowing their access to the primer–template junction. In addition to DNA replication, DNA polymerases also participate in various genome maintenance activities, including translesion synthesis (TLS). However, owing to the error-prone nature of TLS polymerases, their association with clamps must be tightly regulated. Here we show that fork-associated ssDNA-binding protein (SSB) selectively enriches the bacterial TLS polymerase Pol IV at stalled replication forks. This enrichment enables Pol IV to associate with the processivity clamp and is required for TLS on both the leading and lagging strands. In contrast, clamp-interacting proteins (CLIPs) lacking SSB binding are spatially segregated from the replication fork, minimally interfering with Pol IV-mediated TLS. We propose that stalling-dependent structural changes within clusters of fork-associated SSB establish hierarchical access to the processivity clamp. This mechanism prioritizes a subset of CLIPs with SSB-binding activity and facilitates their exchange at the replication fork.