Although molecular mechanisms that prompt cell cycle arrest in response to DNA damage have been elucidated, the systems-level properties of DNA damage checkpoints are not understood. Here, using time-lapse microscopy and simulations that model the cell cycle as a series of Poisson processes, we characterize DNA damage checkpoints in individual, asynchronously proliferating cells. We demonstrate that within early G1 and G2, checkpoints are stringent: DNA damage triggers an abrupt, all-or-none cell cycle arrest. The duration of this arrest correlates with the severity of DNA damage. After the cell passes commitment points within G1 and G2, checkpoint stringency is relaxed. By contrast, all of S phase is comparatively insensitive to DNA damage. This checkpoint is graded: instead of halting the cell cycle, increasing DNA damage leads to slower S-phase progression. In sum, we show that a cell’s response to DNA damage depends on its exact cell cycle position and that checkpoints are phase-dependent, stringent or relaxed, and graded or all-or-none.

DNA damage checkpoints are cellular mechanisms that protect the integrity of the genome during cell cycle progression. In response to genotoxic stress, these checkpoints halt cell cycle progression until the damage is repaired, allowing cells enough time to recover from damage before resuming normal proliferation. Here, we investigate the temporal dynamics of DNA damage checkpoints in individual proliferating cells by observing cell cycle phase transitions following acute DNA damage. We find that in gap phases (G1 and G2), DNA damage triggers an abrupt halt to cell cycle progression in which the duration of arrest correlates with the severity of damage. However, cells that have already progressed beyond a proposed "commitment point" within a given cell cycle phase readily transition to the next phase, revealing a relaxation of checkpoint stringency during later stages of certain cell cycle phases. In contrast to G1 and G2, cell cycle progression in S phase is significantly less sensitive to DNA damage. Instead of exhibiting a complete halt, we find that increasing DNA damage doses leads to decreased rates of S-phase progression followed by arrest in the subsequent G2. Moreover, these phase-specific differences in DNA damage checkpoint dynamics are associated with corresponding differences in the proportions of irreversibly arrested cells. Thus, the precise timing of DNA damage determines the sensitivity, rate of cell cycle progression, and functional outcomes for damaged cells. These findings should inform our understanding of cell fate decisions after treatment with common cancer therapeutics such as genotoxins or spindle poisons, which often target cells in a specific cell cycle phase.Recent live-imaging studies in single cells have shed light on some of the underlying parameters that confer differential sensitivity to endogenous DNA damage. For example, cell-to-cell variation in p21 levels leads to differences in checkpoint stringency (i.e., the robustness of cell cycle arrest in response to DNA damage) 11,12 .However, the underlying parameters that result in differential responses within a cell cycle phase and in response to exogenous DNA damage are generally unknown. Knowledge of these collective dynamical behaviors-which we refer to as DNA damage checkpoint dynamics-is necessary for understanding the relationship between the DNA damage response and functional cellular outcomes such as permanent cell cycle arrest. From a translational perspective, understanding DNA damage checkpoint dynamics could help explain the variability in cellular responses to many chemotherapeutic drugs, which act mainly by inducing DNA damage and interfering with cancer cell proliferation.Among the various forms of DNA damage, DNA double strand breaks (DSBs) are one of the most harmful types of lesions 13,14 . DSBs can give rise to chromosome rearrangements and deletions that can subsequently lead to cancer. The functional effects of DSBs can vary depending on the cell cycle phase in which the damage was incurred. For example, it ha...