Description

DNA topoisomerases play a vital role in solving topological constrains during DNA replication and transcription (Pommier et al. 2016). Top1p is a DNA topoisomerase that relaxes DNA supercoiling by nicking the DNA, creating a covalent bond between the enzyme and the 3’ end of the DNA, called Top1p cleavage complex (Top1cc). Once the DNA is relaxed, Top1p religates the break by reversing its covalent binding. Top1p activity can be inhibited by drugs such as camptothecin (CPT), whose derivatives (topotecan, irinotecan) are largely used in cancer therapy. CPT binds to the Top1cc and delay the religation reaction, thus blocking Top1p on DNA (Figure 1A) (Pommier et al. 2006). It has been shown that treatment of cells with CPT induces DNA double-strand breaks (DSBs) specifically during DNA replication. One model proposes that the DNA nick is converted into a DSB by the passage of the replication fork, which separates the parental duplex DNA (Strumberg et al. 2000). Because CPT prevents Top1p ability to remove topological stress (Koster et al. 2007), another model proposed that the accumulation of topological constrains would block replication fork advance and induce DSB formation as a consequence of fork cleavage by the MUS81 nuclease (Regairaz et al. 2011). These DSBs must be repaired, otherwise mutations could arise and affect genome stability. DSBs are mainly repaired by Homologous Recombination (HR) during the S and G₂ phases of the cell cycle, the Rad52p protein being a key player in this pathway in yeast Saccharomyces cerevisiae (Pardo et al. 2009). Rad52p proteins relocalize into discrete sub nuclear foci upon DSB formation (Lisby et al. 2003). In response to CPT treatment, yeast cells accumulate Rad52p foci both in the nucleus and the nucleolus (Stuckey et al. 2015). These results suggest that CPT-induced DSBs arise in the ribosomal DNA (rDNA) genes, which reside in the nucleolus. These genes are organized on the chromosome XII in a single cluster of 150-200 tandem repeats (Kobayashi 2011). Each rDNA repeat includes a Replication Fork Barrier (rRFB) bound by Fob1p, whose function is to avoid collisions between replication and transcription machineries by stalling the movement of replication forks in only one direction (Kobayashi and Horiuchi 1996). In the rDNA, it has been shown that Top1ccs accumulate naturally and specifically at the rRFB, and that this accumulation is completely lost in the absence of Fob1p (Krawczyk et al. 2014). DSBs have also been detected at the rRFB and are dependent on Top1p and Fob1p. These DSB signals were increased upon the inhibition of Top1p by CPT (Krawczyk et al. 2014), suggesting that the accumulation of Fob1p-dependent Top1ccs at the rRFB leads to DSB formation. Since the rDNA represents about 10% of the genome size of Saccharomyces cerevisiae and these DSBs could be detrimental for cell survival, we thus investigated if the absence of Fob1p could improve the resistance of yeast cells to CPT.

In order to test if FOB1 deletion could improve cell resistance to CPT, we combined the rad52∆ mutation, which confers hypersensitivity to CPT, to the fob1Δ mutation. fob1∆ cells did not show sensitivity to CPT at all tested concentrations (Figure 1B). Moreover, the hypersensitivity of rad52Δ cells was not alleviated by fob1Δ (Figure 1B). These results suggest that the presence of Fob1p and the consequent stabilization of Top1ccs at the rRFB is not the main cause of cell death in the absence of Rad52p when exposed to CPT.

We then analyzed the progression of cells by flow cytometry during a single cell cycle in response to CPT exposure. After the synchronization in G₁ (1C DNA content), cells were incubated with either DMSO (control) or CPT in a minimal medium increasing cell permeability to CPT (Pardo et al., 2020) and released into S phase. WT cells treated with DMSO entered the S phase at 30 minutes and reached the G₂ phase (2C DNA content) at 90 minutes. At 150 minutes, cells finished mitosis and progressed to the following G₁, where they remained arrested by a second addition of the synchronizing agent α-factor (Figure 1C). WT cells exposed to CPT progressed slower through S phase and reached the G₂ phase 30 minutes later than control cells. Consequently, these cells reached the following G₁ with a 30 minutes delay (Figure 1C). Cell cycle progression of fob1∆ cells was comparable to WT, either incubated with DMSO or CPT, respectively. On the contrary, rad52∆ cells exposed to CPT showed an increased S-phase delay compared to WT (Figure 1C), consistent with a previous observation (Pardo et al. 2020). Once they reached the G₂ phase 150 minutes after the release from the G₁ synchronization, they remained blocked in G₂/M until the end of the time course experiment, at 240 minutes (Figure 1C). These results confirm that Rad52p is important for cells to progress through S phase and complete mitosis when they are exposed to CPT (Pardo et al. 2020). FOB1 gene deletion did not change the cell cycle progression of rad52∆ cells, whether treated with DMSO or with CPT (Figure 1C). These data indicate that the absence of Fob1p does not impact the global cell cycle progression in cells exposed to CPT.

Finally, we used a Rad52 protein tagged with YFP to follow by fluorescence microscopy the formation of Rad52p-YFP foci in response to CPT exposure in WT and fob1∆ cells. After G₁ synchronization, cells were released into S phase with DMSO/CPT. In WT cells treated with DMSO, Rad52p foci formed specifically during S-phase (30-60 min after release) and decreased until mitosis completion (Figure 1E). The foci were localized both in the nucleolus (marked by Nop1p-CFP) and in the remaining of the nucleus (marked by mCherry-Pus1p) (Figure 1D-E). The percentage of WT cells containing Rad52p-YFP foci increased about 3-fold in response to CPT and followed the same kinetics as control cells (Figure 1E). fob1∆ cells showed similar results (Figure 1F), suggesting that DNA lesions, presumably DSBs, still form in the rDNA in the absence of Fob1p.

Altogether, these results show that FOB1 deletion does not suppress the phenotypes induced by CPT exposure in WT or rad52∆ cells. The fact that Rad52p foci formation increased upon CPT exposure (Figure 1E) suggests that Rad52p foci are related to Top1ccs. Rad52p foci still formed in the nucleolus in fob1∆ cells exposed to CPT (Figure 1F), suggesting that Top1ccs can still accumulate in the rDNA, likely at sequences other than the rRFB, in the absence of Fob1p. This is consistent with the low amount of Top1ccs detected in the 35S gene, located upstream of the rRFB, in the presence or absence of Fob1p without CPT treatment (Krawczyk et al. 2014). Even if present in a lower amount, these Top1ccs further stabilized by CPT do not impact the cell cycle progression, nor the cell viability in fob1∆ cells (Figure 1B-C). One explanation could be that they do not give rise to DSBs, which are lethal DNA lesions. Indeed, the DNA fragments detected in the rDNA and interpreted as being the consequence of DSBs (Fritsch et al. 2010; Krawczyk et al. 2014; Sasaki and Kobayashi 2017) could also correspond to the reversal of replication forks (Kara et al. 2021), an event in which the fork goes backward by annealing the newly-synthetized DNA strands together. This hypothesis is supported by the lack of “DSB” signals accumulation at the rRFB in the absence of DSB repair factors (Fritsch et al. 2010; Sasaki and Kobayashi 2017; Kara et al. 2021) and by the fact that fork reversal events are increased in cells exposed to CPT (Ray Chaudhuri et al. 2012; Menin et al. 2018). Then, Rad52p foci formation could originate from recombination events initiated from the tip of reversed forks instead of DSBs.