Introduction

DNA replication is an essential and conserved process that is carried out by a complex and highly regulated machinery. DNA synthesis initiates at sites distributed along the chromosomes called replication origins, which are activated at different times during S phase and are used with particular frequencies, or efficiencies, in a population of cells. The timings and efficiencies of origin usage, together with their genome-wide distribution, delineate the program of DNA replication. This program is altered in response to intrinsic and extrinsic signals, with signature patterns described for distinct cell types and growth conditions (Hiratani et al. 2008, 2010; Pope et al. 2010; Ryba et al. 2011; Wu and Nurse 2014; Perrot et al. 2018; Singh and Wu 2018). However, although a large body of work has been dedicated to characterizing such profiles and investigating their regulation, it remains an open question whether and how these programs contribute to cellular function.

One context in which the control of DNA replication may be particularly critical is in situations of replication stress: challenges to DNA synthesis can lead to the generation of lesions and errors in the genetic material (Hills and Diffley 2014; Mazouzi et al. 2014; Kotsantis et al. 2018). To preserve genome integrity, cells activate checkpoint pathways that not only regulate cell cycle progression and gene expression but also modulate DNA synthesis, and in particular origin activation (Saldivar et al. 2017). Our recent study explores the interplay between the replication program, the checkpoint inhibition of origin firing, and genome instability in replication stress conditions (Gómez-Escoda and Wu 2018). Using the fission yeast Schizosaccharomyces pombe as a model system, we uncovered a critical role for the organization of DNA replication in delineating the landscape of genetic instability, shedding new light on how this architecture may modulate both genome maintenance and evolution.

The spatiotemporal program of DNA replication

The organization of DNA replication in chromosomal domains of timing and efficiency has been observed in diverse organisms, suggesting that such configurations are a common feature of replication in eukaryotes. Interestingly, replication programs are conserved between related species of yeast as well as between mouse and human cells (Ryba et al. 2010; Yaffe et al. 2010; Muller and Nieduszynski 2012). At the same time, these programs are flexible, and signature profiles have been associated with distinct developmental states (Hiratani et al. 2010). For instance, there is a clear relationship between the replication program and cell fate, as conserved replication timing domains have been identified in mouse and human embryonic stem cells, constituting a shared feature of pluripotence (Hiratani et al. 2008; Palou et al. 2017). In addition, development and differentiation are accompanied by alterations in replication patterns in organisms ranging from worms to frogs to humans (Hyrien et al. 1995; Hiratani et al. 2008, 2010; Desprat et al. 2009; Pope et al. 2010; Rodríguez-Martínez et al. 2017). Finally, the profile of origin usage is sensitive to nutritional conditions, as found in fission yeast cells that have been temporarily starved of nitrogen (Wu and Nurse 2014). Therefore, the replication program displays a dynamic specificity in response to internal and external signals.

Why might cells use different replication programs in particular contexts? One possibility is that these profiles simply arise as a consequence of the physiological changes that are occurring in distinct environments. However, the conservation of specific organizations of DNA replication described above suggests that it may be important for cellular function apart from simply ensuring the duplication of the genome. For example, links between replication and gene expression have been found in multiple systems. In mammalian cells, early replicating regions are correlated with transcriptional activity, and alterations in replication timing during differentiation are associated with changes in chromosome architecture and transcription (Rivera-Mulia et al. 2015, 2018; Pourkarimi et al. 2016; Rodríguez-Martínez et al. 2017). Along the same lines, a direct role for replication timing in the control of gene expression has been demonstrated in the budding yeast, where this parameter has been implicated in regulating the level of histone transcription (Müller and Nieduszynski 2017). Furthermore, we have previously shown that the profile of origin usage in the fission yeast determines the distribution of meiotic recombination along the chromosomes (Wu and Nurse 2014). While these observations hint toward replication being involved in diverse processes, we are only beginning to understand the biological roles of the organization of DNA replication.

Although the genetic material is particularly fragile during its duplication, little is known about whether and how the replication program contributes to genome integrity. Intriguingly, genetic analyses have established correlations between replication timing and mutation frequencies in budding yeast (Lang and Murray 2011), and whole-genome sequencing of normal and cancer cell lines has revealed differences in the types and distributions of mutations that are found in early vs. late replicating chromosomal regions (Koren et al. 2012; Liu et al. 2013; Sima and Gilbert 2014; Tomkova et al. 2018). These findings thus hint at an intimate coupling between the replication program and genome maintenance.

Challenges to DNA synthesis: causes and consequences

During DNA replication, the genetic material is particularly vulnerable to damage and to acquiring errors. On one hand, even during an “unperturbed” S phase, cells encounter a variety of insults that impede DNA synthesis. For instance, markers of genotoxic stress are detected during replication in embryonic stem cells (Ahuja et al. 2016). On the other hand, replication stress can be induced by extrinsic DNA damaging agents, and it can also arise due to intracellular challenges such as limiting levels of nucleotides and replication factors, difficult to replicate sites with secondary DNA structures, chromatin accessibility, or collisions between the replication and transcription machineries (Zeman and Cimprich 2014; Mazouzi et al. 2014; Tubbs and Nussenzweig 2017). These challenges can then lead to the slowing, stalling, or collapsing of replication forks. If not properly managed, replication stress may, therefore, result in DNA breaks as well as incomplete DNA synthesis with cells entering mitosis prior to completing replication.

Notably, the genome instability that arises from replication stress is considered to be a key contributor to cancer. Genetic alterations are a fundamental feature of cancer cells and a driving force in tumorigenesis. Recent studies have identified thousands of coding mutations that are heterogeneous within and between tumors (McGranahan and Swanton 2017), with mutations in established cancer genes present in only a subset of cells within a tumor. This suggests that tumor fitness may be modulated by the numerous other genetic alterations that are present, which may be generated in part by replication stress-associated instability (Negrini et al. 2010). These findings thus highlight how our understanding of cancer progression will benefit from investigating the cellular response to replication challenges and the mechanisms that modulate these pathways.

The checkpoint response to replication stress

Eukaryotic cells utilize a number of mechanisms to protect the genome from endogenous and exogenous sources of DNA damage (Palou et al. 2017; Saldivar et al. 2017). Challenges to DNA synthesis that disrupt the coordination between replicative helicase and DNA polymerase lead to the formation of single-stranded DNA and the activation of the S phase/ATR checkpoint (Byun et al. 2005). The subsequent modification of a number of substrates by checkpoint kinases brings about specific changes in gene expression, cell cycle arrest, replication fork protection and subsequent restart, and inhibition of replication initiation (Saldivar et al. 2017; Mikolaskova et al. 2018; Villa-Hernández and Bermejo 2018). These mechanisms are conserved throughout eukaryotes, and the importance of much of this regulation has been previously described. However, the impact of the checkpoint control of origin firing, first reported over 20 years ago (Santocanale and Diffley 1998; Shirahige et al. 1998), has remained surprisingly elusive. This inhibition of origins occurs through modulation of the CDK (cyclin-dependent kinase) and DDK (Dbf4-dependent kinase) pathways (Lopez-Mosqueda et al. 2010; Zegerman and Diffley 2010), and initial studies in budding yeast suggested that inefficient and late-firing origins are specifically targeted for inhibition (Santocanale and Diffley 1998; Shirahige et al. 1998; Santocanale et al. 1999). However, the selection of these target origins, the extent of inappropriate initiation in the absence of the checkpoint, and the significance of this regulation remained contested. As the conservation of the checkpoint regulation of origin activation from yeast to human indicates a central role for this mechanism in the DNA damage response, we have taken advantage of this process to investigate the interplay between the organization of DNA replication and genome maintenance (Gómez-Escoda and Wu 2018).

Interplay between the replication program and the response to replication stress

In the fission yeast Schizosaccharomyces pombe, replication stress activates the conserved Rad3/ATR checkpoint kinase, which then signals through Cds1/CHK1 to modify a number of substrates that include replication factors and cell cycle regulators (Bentley et al. 1996; Lindsay et al. 1998; Edwards et al. 1999; Labib and De Piccoli 2011; Willis et al. 2016). Previous work in this organism reported differing results for the checkpoint inhibition of origin firing, identifying 2–28% of the origins in the genome as sensitive to this control (Feng et al. 2006; Heichinger et al. 2006; Hayashi et al. 2007; Mickle et al. 2007). For our study, we therefore began by evaluating origin activity in Rad3/ATR-deleted cells following exposure to the ribonucleotide reductase inhibitor hydroxyurea (HU), which triggers replication stress through depletion of nucleotide pools. As we surmised that the detection of checkpoint-inhibited initiation sites in earlier studies may have been impeded by the use of high HU concentrations, we applied a moderate level of the drug that permits slow progression through S phase. Our results showed that in rad3Δ cells treated with 6 mM HU, origin firing is significantly deregulated at a subset of initiation sites in the genome. Specifically, we identified 176 deregulated sites out of 876 total origins that are increased by up to 25% efficiency in these conditions. In contrast to the previous findings, these checkpoint-inhibited origins (CIOs) span a broad spectrum of activities. However, regardless of their individual characteristics, they are clearly clustered in late-firing and inefficient regions of the genome, with a large portion of the origins in such domains (up to ~ 70%) being deregulated in HU-treated rad3Δ cells. Together with this signature distribution of CIOs, we established the quantitative profile of the extent of inappropriate firing in HU-treated rad3Δ cells. This then allowed us to probe how the organization of genome duplication impacts genome maintenance.

First, we evaluated the consequences of deregulated origin firing for genome instability, using the single-stranded DNA (ssDNA) binding protein Ssb1/RPA and the DNA repair protein Rad52 as molecular markers. Our data demonstrated abnormal ssDNA formation and high levels of Rad52 recruitment at CIOs. We further found that upon HU exposure, the changes in RPA and Rad52 binding between rad3Δ vs. wild type were correlated with the extent of origin deregulation: the increase in activity of a CIO in rad3Δ was directly proportional to the increase in RPA and Rad52 occupancy at this site. We then evaluated the relationship between the genome-wide profiles of RPA and Rad52 with respect to inappropriate replication initiation. Taking into account all origins as well as all RPA- and Rad52-binding sites in HU-treated cells, our analysis revealed a clear correspondence between the domains of origin deregulation vs. the densities of RPA and Rad52 loci along the chromosomes in a rad3Δ mutant (Fig. 1, middle and bottom panels). Our results thus suggested unscheduled origin firing as a critical determinant of the overall pattern of genome instability. Strikingly, the profile of origin deregulation in rad3Δ cells under replication stress showed a strong negative correlation with the wild-type program of origin efficiencies, consistent with our observation that origins in late-firing and inefficient regions of the genome are inhibited by the checkpoint (Fig. 1, compare top panel with middle and bottom panels). Taken together, these results led us to propose that the replication program defines the checkpoint regulation of origin firing in replication stress conditions, with consequences for the landscape of genome instability.

Fig. 1
figure 1

Schematic representation of the relationship between the replication program, origin deregulation, and genome instability. Top: replication program in wild-type cells. Origin activity (dashed black line) represents the efficiency of origin usage, which is measured by the frequency of initiation at a given origin in a population of cells. Middle: change in origin efficiency (dashed red line) between rad3Δ and wild-type cells in replication stress conditions. Bottom: density of DNA damage hotspots (dashed blue line), as assayed by Rad52 binding, in rad3Δ cells exposed to replication stress. x-axis: chromosome coordinates, y-axis: origin efficiency (top), changes in origin usage (middle), and density of DNA damage hotspots (bottom). The replication program shows strong negative correlations with origin deregulation and the level of genome instability. Inappropriate initiation and DNA damage hotspots are clustered in inefficient and late-firing replication domains

Full size image

To test this model, we took advantage of the rif1Δ genetic background in which replication initiation is altered genome-wide (Hayano et al. 2012; Cornacchia et al. 2012; Yamazaki et al. 2012). We first determined the effect of this modified replication program on origin deregulation in the absence of rad3 and found a complete redistribution of checkpoint-inhibited origins: these rif1-CIOs are more uniformly positioned along the chromosomes and no longer clustered in distinct domains (Fig. 2). We then assessed whether the changes in checkpoint inhibition are accompanied by alterations in Rad52 binding. Our data showed that the correlation between origin deregulation and Rad52 accumulation observed in HU-treated checkpoint-defective cells is maintained in the rif1Δ rad3Δ genetic background. Strikingly, Rad52 sites in rif1Δrad3Δ cells were no longer clustered in the same genomic domains as in rad3Δ, revealing a reorganization of Rad52 loci. Collectively, these findings support our model that the replication program governs the checkpoint regulation of origin firing and the profile of genome instability when cells are challenged for DNA synthesis.

Fig. 2
figure 2

Alteration of the replication program has direct consequences for the checkpoint regulation of origin firing. Left: schematic representation of the replication program in wild-type (top) vs. rif1Δ (bottom) cells. Origin activity (dashed black lines) represents the efficiency of origin usage. x-axis: chromosome coordinates, y-axis: origin efficiency. The distinct replication domains apparent in wild-type cells are lost in rif1Δ. Right: Inappropriate initiation in rad3Δ (top, dashed red line) and rif1Δ rad3Δ (bottom, dashed green line) cells. Consistent with the changes in the replication program, the signature pattern of origin deregulation in rad3Δ is reorganized in rif1Δ rad3Δ. This is accompanied by changes in the profiles of DNA damage hotspots (not shown). x-axis: chromosome coordinates, y-axis: changes in origin efficiency

Full size image

Perspectives for genome organization and evolution

Our results demonstrate that the control of origin usage by the Rad3/ATR pathway in replication stress conditions is a key aspect of its function in genome maintenance. Why might genome duplication need to be organized for the response to such challenges? One possibility is that the use of particular replication programs may promote elevated local concentrations of checkpoint factors that may result in a more efficient response, either for checkpoint signaling or for DNA repair (Rivera-Mulia et al. 2015). Related to this idea, particular regions may be more accessible to the recombination machinery or for the search for homology, which has been proposed to be affected by the spatial organization of the genome (Misteli and Soutoglou 2009; Nagai et al. 2010). Moreover, evolution may have favored an arrangement in which potential sites of DNA damage are enriched in regions, where mutations have a lower likelihood to be deleterious. The replication program may also be important in the context of an “unperturbed” cell cycle, where cells are in fact constantly exposed to endogenous challenges to replication (Zeman and Cimprich 2014). In support of this idea, ATR function is essential in vertebrate cells (Eykelenboom et al. 2013), and embryonic stem cells display markers of replication stress that are dependent on ATR (Ahuja et al. 2016). Therefore, how the genome is organized for its duplication may ultimately have consequences for its integrity in the context of normal and challenging situations.

While our study has assessed the immediate consequences of origin deregulation, it is also interesting to consider the long-term impacts of the replication program. For instance, our work may provide insight into how this aspect of genome architecture may contribute to genome evolution. Indeed, our demonstration that replication origins are hotspots of DNA damage when they are fired inappropriately in stress conditions is consistent with comparative genomic analyses and laboratory evolution experiments in budding yeast that indicate a co-localization of chromosomal breakpoints with origins (Di Rienzi et al. 2009; Gordon et al. 2009). The coupling between the replication program and genome instability that we uncovered may also help us understand the correlations between replication timing and the differences in the types and frequencies of mutations that are found in different genomic regions (Lang and Murray 2011; Koren et al. 2012; Liu et al. 2013; Lu et al. 2014; Polak et al. 2015; Tomkova et al. 2018). Furthermore, our results may be pertinent in the context of natural situations in which checkpoints are inefficient or inactive. This is, for instance, the case in the rapid and synchronous embryonic cell cycles observed in many metazoa (Kermi et al. 2017). Finally, checkpoint deficiencies, replication stress, and altered patterns of origin usage have also all been identified as features of cancer cells (Halazonetis et al. 2008; Ciccia and Elledge 2010; Bester et al. 2011; Donley and Thayer 2013). Our results may therefore have implications for how the replication program may critically contribute to heterogeneity, somatic evolution, and cancer development. All together, given the flexibility of the organization of DNA replication in response to internal and external stimuli, it is tempting to speculate that the way cells duplicate their genetic material plays crucial roles in adaptation and evolution in the long term.

Conclusion

The organization of DNA replication is a conserved feature of eukaryotic genomes. Its remarkable similarity between related species highlights potential evolutionary constraints and suggests a biological importance for this architecture. Future studies in this area of research will open exciting frontiers in our understanding of how the essential process of DNA replication may not only serve to copy the genome but also to modulate cellular pathways and shape genome evolution.