Replication stress at centromeres biases the segregation of DNA damage

Abstract

Replication-associated errors can cause DNA damage to accumulate on the newly synthesized strand over time. In specific cases such as stem cells, retention of the immortal strand used as template preserves one daughter cell into pluripotency while correlating with terminal differentiation of the damage one. In somatic cells, DNA damage distribution after mitosis remains unclear. Here, we uncovered a mechanism of non-random segregation of the DNA damage marker gH2AX occurring during a single cell division cycle. Replication stress using hydroxyurea (HU) upon release into S phase in RPE-1, BJ, hCEC D29 and fibroblasts showed reproducible Non-Random Segregation (NRS) of gH2AX in the ensuing G1, a phenotype not observed in any of the cancer cell lines analyzed. Notably, removal of R-loops led to a reduction of cells with NRS, whether RNaseH1 was over-expressed globally or exclusively targeted to centromeres, indicating that centromeric DNA-RNA hybrids contribute to NRS of the damage. In line with our previous evidence of centromeric chromatin disruption leading to R-loops, rapid removal of the histone H3 variant CENP-A causes damage and NRS, although to a lower extent than HU alone. This implies that additional mechanisms contribute to centromeric R-loops and NRS of damage in the daughter cells upon mitotic exit. Mechanistically, chemical inhibition of the catalytic activity of Rad51 led to a significant drop in NRS without a change in the total amount of damaged cells, implying involvement of the Homologous Recombination (HR) pathway to accumulation of gH2AX to only one chromatid. In turn, this affects the spindle-kinetochore with a measurable length asymmetry, inducing mechanical and/or epigenetic signals that affect the orientation of the sister chromatids on the metaphase plate to bias segregation. Altogether, we found replication-induced asymmetric segregation of DNA damage during mitosis that is influenced by centromeric R-loops, Rad51 activity and spindle dynamics, with implications on cell fate, chromosome and genome stability in the daughter cells.