A cell-cycle checkpoint is a regulatory mechanism that delays progression through the cell cycle when a prerequisite for accurate division has not been fulfilled. Checkpoints coordinate the duplication and distribution of the genome by responding to problems such as damaged DNA, incomplete replication, or chromosomes that have not attached correctly to the division apparatus. They are especially important in eukaryotic cells, where multiple chromosomes must be copied and segregated without losing genetic information.
A checkpoint is not a physical barrier or a single molecular switch. It is a signaling system that detects particular molecular conditions, transmits that information, and changes the activity of the machinery driving cell division. Some responses develop within minutes through changes in protein activity; others require altered gene expression and can persist for hours or longer. Depending on the problem and cellular context, checkpoint activation can support repair and renewed proliferation, produce a durable arrest, or contribute to cell death.
The major checkpoint systems operate before DNA synthesis, during DNA replication, before mitosis, and before chromosome separation. These systems overlap but monitor different features of the division process. Their failure can promote chromosome instability and cancer, whereas their continuing activity can also help damaged cells survive. Consequently, checkpoint biology connects fundamental cell biology with developmental biology, inherited disease, and anticancer treatment.
Historical origins and the meaning of a checkpoint
The checkpoint concept emerged from experiments distinguishing the mechanisms that execute cell-cycle events from those that impose an order on them. A cell requires machinery to copy DNA and separate chromosomes, but it also requires controls that prevent separation from beginning before copying is complete. A defect in the copying machinery and a defect in the mechanism that detects incomplete copying therefore produce different kinds of failure.
Genetic studies of budding yeast by Leland Hartwell and colleagues identified cell-division-cycle mutants that arrested at characteristic stages. Work in fission yeast by Paul Nurse and others established central regulators of entry into mitosis. In parallel, Tim Hunt’s discovery of cyclins in marine invertebrate embryos revealed proteins whose abundance oscillates during division. These findings helped establish that a conserved biochemical engine drives the eukaryotic cell cycle. Hartwell, Nurse, and Hunt received the 2001 Nobel Prize in Physiology or Medicine for discoveries concerning cell-cycle regulation.
A particularly influential experiment was reported by Ted Weinert and Hartwell in 1988. Budding yeast lacking functional RAD9 failed to impose the normal division delay after DNA damage. The distinction was important: a cell could retain the capacity to undergo division while losing the control that made division conditional on the state of its DNA. Hartwell and Weinert articulated the checkpoint framework in 1989, emphasizing mechanisms that enforce dependencies between otherwise separable events.
In this operational sense, a checkpoint is demonstrated when a perturbation normally delays a later event, but a specific regulatory defect allows that later event to proceed despite the unresolved perturbation. Such bypass does not mean that the cell has corrected the original problem. It may instead undergo division with broken, incompletely replicated, or improperly attached chromosomes. This distinction remains essential when interpreting experiments: faster cell-cycle progression can indicate failed surveillance rather than improved cellular function.
The word “checkpoint” is also used more broadly for physiological decision points, including the mammalian restriction point and controls responding to nutrient availability or cell size. These are related to surveillance but are not identical to DNA-damage or spindle checkpoints. Moreover, checkpoint responses are not confined to sharply defined moments. Replication surveillance operates throughout S phase, and damage signaling can influence several transitions. The familiar checkpoint diagram is therefore a useful map of major outputs, not a complete description of when surveillance occurs.
Molecular organization of checkpoint signaling
Checkpoint pathways are commonly described in terms of sensors, mediators, signal-transducing enzymes, and downstream effectors. These are functional categories rather than rigidly separated protein classes. A sensor-associated complex recognizes a molecular structure, such as a broken DNA end or an unattached chromosome attachment site. Mediators organize signaling proteins near that structure. Downstream reactions then inhibit a cell-cycle transition or alter the processes that generated the problem.
Much of this communication depends on protein phosphorylation, the reversible addition of phosphate groups to proteins. Protein kinases transfer phosphate from ATP, whereas phosphatases remove it. Phosphorylation can change catalytic activity, create a binding site, alter localization, or control protein stability. A checkpoint can consequently spread information from a small number of abnormal structures to many downstream targets without requiring a corresponding increase in the abundance of every signaling protein.
The principal cell-cycle engines are complexes of cyclins and cyclin-dependent kinases (CDKs). In a simplified mammalian sequence, cyclin D–CDK4/6 supports early G1 progression, cyclin E–CDK2 contributes to entry into S phase, cyclin A–associated kinases support replication and preparation for mitosis, and cyclin B–CDK1 drives mitotic entry. These assignments overlap, and the importance of individual complexes varies among cell types. Checkpoint pathways restrain this engine by inducing CDK inhibitors, maintaining inhibitory CDK phosphorylation, or preventing the destruction of proteins that must remain present until a transition is safe.
DNA-damage signaling prominently involves ATM and ATR, members of the phosphatidylinositol 3-kinase-related kinase family. ATM is strongly associated with responses to DNA double-strand breaks, whereas ATR responds especially to replication-associated structures containing single-stranded DNA coated by replication protein A. ATM frequently signals through CHK2, and ATR through CHK1, although neither pairing is exclusive. Break processing can convert a lesion initially associated with ATM signaling into a substrate that activates ATR, illustrating how the pathways cooperate.
Other checkpoint outputs depend on controlled protein destruction through the ubiquitin–proteasome system. By attaching ubiquitin chains to selected substrates, cells can rapidly remove an inhibitor or an activator. This creates directional transitions: destroying a mitotic regulator is fundamentally different from temporarily inhibiting it by phosphorylation. Checkpoint circuits combine these reversible and effectively irreversible mechanisms with feedback loops, allowing a weak or local signal to delay a global transition while preserving the capacity for coordinated release.
DNA damage control before S phase
The G1 DNA-damage response reduces the likelihood that a damaged template will be copied. Replicating damaged DNA can convert a relatively limited lesion into a stalled replication fork, a chromosome break, or a heritable sequence change. In mammalian cells, a major route from damage detection to arrest involves the transcription factor p53, encoded by TP53. Damage-dependent signaling stabilizes and activates p53, allowing it to alter the expression of genes involved in arrest, repair, and cell fate.
Under unstressed conditions, the ubiquitin ligase MDM2 promotes p53 turnover. DNA-damage signaling changes the regulation of p53 and its interacting proteins, reducing this suppression. One important p53 target is CDKN1A, which encodes the CDK inhibitor p21. Increased p21 can inhibit cyclin–CDK activities needed for S-phase entry. The resulting response is not an instantaneous universal switch: its strength and timing depend on damage type, cell-cycle position, and the state of the regulatory network.
A central downstream component is the retinoblastoma protein, usually abbreviated RB. In its growth-restraining state, RB binds members of the E2F family of transcription factors and limits expression of genes needed for DNA synthesis. Phosphorylation by cell-cycle CDKs progressively changes this regulation. When p21 restrains CDK activity, RB can continue to suppress the E2F-dependent transcriptional program, delaying the transition into S phase. This connects the damage response to the same machinery that integrates proliferative signals.
The G1 damage checkpoint must be distinguished from the restriction point. The restriction point concerns the acquisition of relative independence from external mitogenic stimulation during G1; DNA-damage surveillance concerns whether progression should be delayed because the genome is compromised. The two controls intersect through cyclin–CDK activity and RB–E2F regulation, but they are not interchangeable. A cell can have abundant growth stimulation and still arrest because of DNA damage, while another cell may stop proliferating without detectable DNA lesions because mitogenic support is absent.
G1 arrest also differs among biological systems. Budding yeast can delay the Start transition through damage-induced inhibition of G1 cyclin function, without using the mammalian p53–p21–RB arrangement. Early embryonic divisions in several animals have abbreviated gap phases and limited versions of some somatic checkpoint responses. These differences show why the textbook phrase “the G1 checkpoint” describes a functional problem rather than one universally conserved molecular circuit. In mammalian tissues, loss of the p53 or RB pathways weakens an important barrier to the propagation of damaged genomes, but it does not abolish every remaining checkpoint.
Replication stress and the intra-S-phase checkpoint
The intra-S-phase checkpoint coordinates surveillance with ongoing DNA replication. Replication forks encounter obstacles including DNA lesions, difficult-to-replicate sequences, tightly bound proteins, and collisions with transcription. Fork progression can also be impaired by inadequate nucleotide supply or excessive initiation of replication. These conditions are collectively described as replication stress, although they produce several distinct molecular structures and do not all have identical consequences.
A major signaling intermediate forms when DNA unwinding and DNA synthesis become uncoupled. Exposed single-stranded DNA is coated by replication protein A, which helps recruit ATR through its partner ATRIP. Additional structures and proteins are required for effective signaling: the RAD9–RAD1–HUS1 clamp and TOPBP1 provide one important route to ATR activation, while ETAA1 provides another. Activated ATR phosphorylates targets including CHK1, coordinating replication dynamics with cell-cycle delay.
One prominent output is restraint of additional replication-origin firing. In mammalian cells, CHK1 can promote inhibition and degradation of CDC25A, reducing CDK activity needed for initiation. This limits the number of new forks competing for nucleotides and replication factors. The response is spatially and temporally nuanced, however. While broad initiation programs can be suppressed, nearby dormant origins may still help complete regions threatened by slow or stalled forks. The checkpoint therefore reorganizes replication rather than simply stopping every origin.
Replication surveillance also preserves the conditions needed for fork stabilization and restart. ATR-dependent networks regulate fork-associated proteins, nucleotide metabolism, and the processing of abnormal DNA intermediates. A stalled fork is not automatically a broken fork: it can remain recoverable. Prolonged or improperly processed arrest can instead generate double-strand breaks and dependence on homologous recombination. Checkpoint failure may increase both the number of stalled structures and the likelihood that those structures deteriorate into more dangerous lesions.
The distinction between replication slowing and checkpoint action is experimentally important. Hydroxyurea reduces deoxyribonucleotide availability by inhibiting ribonucleotide reductase, thereby slowing DNA synthesis directly. A cell with defective surveillance may therefore still replicate slowly after hydroxyurea exposure, even though it fails to suppress new origin firing or protect stressed forks appropriately. Similarly, a reduction in bulk DNA synthesis does not by itself establish that a checkpoint is functional. The intra-S-phase response is defined by regulated changes to replication and cell-cycle control, not merely by the physical inability of polymerases to proceed through an unfavorable template.
The G2 checkpoint and control of mitotic entry
The G2 checkpoint delays mitosis when DNA damage or unresolved replication problems remain after most genome duplication has occurred. Entry into mitosis requires activation of cyclin B–CDK1, which initiates chromosome condensation, nuclear-envelope breakdown in organisms with open mitosis, and reorganization of the cytoskeleton. Starting these processes before DNA is ready can transform repairable lesions into chromosome fragmentation or segregation failure.
CDK1 activity depends on more than cyclin binding. WEE1 and, in relevant contexts, MYT1 place inhibitory phosphates on CDK1, while CDC25 phosphatases remove them. Checkpoint signaling shifts this balance toward inhibition. ATR–CHK1 signaling is particularly important when replication-associated lesions persist, while ATM-dependent signaling contributes strongly to responses to double-strand breaks. Regulation of CDC25 proteins and maintenance of inhibitory CDK1 phosphorylation help keep the mitotic engine below its activation threshold.
Once sufficient CDK1 activity is reached, positive feedback sharpens the transition. Active cyclin B–CDK1 promotes CDC25 activity and inhibits opposing regulators, helping activate more CDK1. This produces switch-like behavior rather than a gradual, uncoordinated entry into mitosis. The checkpoint acts on this feedback-controlled system, preventing or postponing its activation. In some mammalian contexts, p53-dependent transcriptional responses reinforce a longer arrest, including through p21 and other targets, whereas rapid inhibition relies more heavily on changes to existing proteins.
G2 is also a phase in which a replicated sister chromatid is available as a template for homologous recombination. This does not mean that every G2 lesion is repaired through that pathway; repair choice depends on DNA-end processing, chromatin context, and other regulatory factors. Nevertheless, checkpoint-dependent delay can provide time for repair before chromosomes become highly condensed and separated. The surveillance pathway and the repair machinery therefore cooperate, although checkpoint activity is not itself equivalent to completed DNA repair.
The G2 checkpoint is not perfectly sensitive to every residual lesion. Cells may enter mitosis with some unresolved damage, and certain under-replicated regions can undergo specialized processing or DNA synthesis during early mitosis. Such events do not eliminate the importance of G2 surveillance; they show that cells possess successive layers of risk management rather than a single absolute certification step. Likewise, arrest can be temporary even when repair is incomplete. Whether a cell remains in G2, resumes division, or undergoes a durable change of state depends on damage burden, checkpoint duration, and the broader regulatory context.
The spindle assembly checkpoint
The spindle assembly checkpoint delays the onset of anaphase until chromosomes have established appropriate connections with the mitotic spindle. Its signaling centers are kinetochores, multiprotein assemblies built on chromosomal centromeres. Kinetochores attach chromosomes to spindle microtubules, permitting replicated sister chromatids to move toward opposite poles. The checkpoint prevents premature separation when kinetochore attachment remains incomplete.
Unattached kinetochores promote a signaling pathway involving MPS1, KNL1, BUB proteins, and MAD proteins. MPS1-dependent phosphorylation creates recruitment sites for checkpoint components. MAD1-associated MAD2 promotes capture of CDC20 in an inhibitory assembly. A principal product is the mitotic checkpoint complex, containing MAD2, BUBR1, BUB3, and CDC20. This complex inhibits the CDC20-dependent activity of the anaphase-promoting complex/cyclosome (APC/C), a ubiquitin ligase that normally triggers anaphase and mitotic exit.
Two crucial APC/C substrates are securin and cyclin B. Securin restrains separase, the protease that cleaves the cohesin complexes holding sister chromatids together. Cyclin B maintains mitotic CDK1 activity. When checkpoint inhibition is relieved, APC/C-dependent destruction of securin permits separase activation, while cyclin B destruction promotes mitotic exit. By controlling these substrates, the checkpoint links the physical state of chromosome attachment to a coordinated, effectively irreversible segregation program.
Attachment surveillance must be distinguished from attachment-error correction. Aurora B kinase helps destabilize inappropriate kinetochore–microtubule connections, promoting new attempts at attachment. Correct biorientation generally places sister kinetochores under opposing forces and favors stabilization of productive attachments. However, the checkpoint should not be reduced to a simple tension meter. Unattached or insufficiently occupied kinetochores are major signaling platforms, while tension and attachment geometry influence signaling through several mechanisms. Error correction can maintain checkpoint activity indirectly by repeatedly generating unattached kinetochores.
Some errors can escape detection. A merotelic attachment, in which one kinetochore connects to microtubules from both spindle poles, may satisfy attachment-dependent checkpoint requirements while still causing a lagging chromosome during anaphase. Thus, checkpoint satisfaction does not guarantee perfect segregation. If checkpoint signaling persists, cells may remain in mitosis for an extended period, die during the arrest, or eventually exit through mitotic slippage as cyclin B levels decline. These outcomes depend on the balance between continued checkpoint inhibition, residual protein degradation, and cell-death signaling.
Checkpoint silencing, recovery, and alternative cell fates
A successful checkpoint response requires termination as well as activation. Once the initiating problem has been sufficiently resolved, cells must dismantle signaling complexes, reverse regulatory modifications, and restore cell-cycle activity. Persistent arrest after repair would itself impair tissue maintenance. Conversely, premature silencing would defeat surveillance. Checkpoint recovery is therefore an actively regulated process, not simply the passive disappearance of a damage signal.
In DNA-damage responses, repair removes the structures sustaining kinase activation, while phosphatases reverse phosphorylation on checkpoint components and substrates. In mammalian cells, WIP1, encoded by PPM1D, can attenuate several damage-response signals. Proteolysis and changes in localization also contribute. During recovery from G2 arrest, the Aurora A–BORA–PLK1 network can help reactivate the mitotic-entry machinery. Because many of these regulators have additional cell-cycle functions, their effects depend on timing and on the state of the cell.
Spindle checkpoint silencing likewise involves multiple processes. Stable microtubule attachment reduces the ability of kinetochores to recruit and activate checkpoint proteins. Phosphatases including PP1 and PP2A oppose signaling phosphorylations, while molecular disassembly mechanisms dismantle inhibitory complexes. The AAA+ ATPase TRIP13 and its partner p31comet participate in remodeling MAD2-containing complexes. These reactions help restore APC/C activity only after inhibitory signaling has fallen sufficiently, coordinating anaphase across the cell rather than allowing each chromosome pair to separate independently.
Recovery is distinct from checkpoint adaptation. Recovery generally describes resumption after the initiating problem has been resolved sufficiently; adaptation describes renewed progression despite a persistent problem. The terminology is most sharply established in experimental systems such as yeast, and its application varies across studies. Neither adaptation nor mitotic slippage should be interpreted as evidence of accurate repair. Both can permit survival in some circumstances but can also produce cells with substantial genetic abnormalities.
If stress remains severe or prolonged, cells may enter cellular senescence, a durable proliferation arrest accompanied by broader changes in cellular organization and gene expression. Alternatively, they may activate apoptosis. p53 can contribute to both arrest and cell-death programs, but no simple rule assigns one outcome to a fixed damage threshold. Cell type, proliferative history, metabolic conditions, and signaling dynamics all matter. Moreover, a prolonged checkpoint arrest and senescence are not synonymous: demonstrating senescence requires evidence beyond observing that a population has temporarily stopped dividing.
Meiotic surveillance and developmental variation
Checkpoint systems are modified for meiosis, the specialized division program that produces haploid cells from a diploid precursor. Meiosis includes one round of DNA replication followed by two chromosome-segregation divisions. Its first division separates homologous chromosomes rather than sister chromatids, creating surveillance requirements that differ from those of ordinary mitosis. Programmed DNA breaks and their repair are part of normal meiotic development rather than simply accidental damage.
During meiotic prophase, the SPO11 protein initiates double-strand breaks that support recombination between homologous chromosomes. A subset of repair events becomes crossovers, which help maintain the physical connections needed for homolog segregation. Surveillance mechanisms monitor persistent recombination intermediates and defects in chromosome synapsis, the close alignment of homologs along the synaptonemal complex. When these processes fail, cells can delay progression or be eliminated rather than proceeding with an unsuitable chromosome configuration.
In budding yeast, meiotic recombination surveillance includes the chromosome-associated kinase Mek1, which helps coordinate recombination with prophase progression. Mammalian meiotic surveillance involves ATM, ATR, and several downstream pathways, with outcomes differing between male and female germ cells and among developmental stages. Defects that prevent normal synapsis or repair can cause germ-cell loss and infertility. These responses are not simply copies of the somatic G2 checkpoint: they operate within a program in which controlled DNA breakage and specialized chromosome structures are expected.
The spindle checkpoint must also accommodate meiosis-specific attachment patterns. During meiosis I, sister kinetochores generally function together toward the same pole, while homologous chromosomes attach toward opposite poles. During meiosis II, sister chromatids separate in a pattern more similar to mitosis. Checkpoint signaling and error correction operate within these distinct geometries. In mammalian oocytes, spindle assembly is prolonged and lacks the conventional centrosomal organization of many somatic cells. Errors can persist despite checkpoint activity, contributing to aneuploidy, an abnormal chromosome number.
Developmental stage also changes checkpoint behavior outside the germ line. Early embryos in several animal species undergo rapid divisions with abbreviated or absent gap phases, often supported by maternally supplied proteins and RNAs. The capacity to delay division in response to damage or incomplete replication changes as embryonic cell cycles are remodeled. Stem cells, proliferating progenitors, and differentiated cells likewise differ in their checkpoint dependencies. A terminally differentiated cell outside the proliferative cycle cannot be understood simply as a cell waiting at a checkpoint; it occupies a different regulatory state, with distinct requirements for maintaining its genome.
Genome instability, inherited disorders, and cancer
Checkpoint failure can increase the survival and propagation of cells carrying DNA lesions or segregation errors. Its effects range from small sequence changes to chromosome breaks, rearrangements, and changes in chromosome number. These abnormalities can alter gene dosage and disrupt regulatory networks. However, genomic instability is not caused by checkpoint defects alone: impaired repair, excessive replication stress, abnormal chromosome architecture, and defects in spindle organization can all generate errors even when parts of surveillance remain functional.
Cancer frequently involves alterations in the p53 and RB pathways, weakening controls over proliferation and damage responses. Loss of a checkpoint does not automatically produce a malignant cell, because transformation generally requires additional changes affecting survival, growth signaling, tissue organization, and other properties. Nor do tumor cells necessarily lose every checkpoint. Many remain strongly dependent on selected surveillance pathways because oncogenic signaling and abnormal replication programs create chronic endogenous stress.
Inherited disorders demonstrate the importance of these networks in normal physiology. Biallelic pathogenic variants in ATM cause ataxia-telangiectasia, a disorder associated with neurological disease, immune abnormalities, radiosensitivity, and increased cancer susceptibility. Certain hypomorphic ATR variants cause an ATR-related form of Seckel syndrome, illustrating the developmental importance of replication surveillance. Biallelic pathogenic variants in BUB1B, which encodes BUBR1, can cause mosaic variegated aneuploidy syndrome. These conditions involve broader protein functions as well as checkpoint defects, so their clinical features cannot be attributed to cell-cycle arrest alone.
Checkpoint biology has also shaped anticancer drug development. Inhibitors targeting ATR, CHK1, and WEE1 have been investigated as ways to exploit tumor dependencies on replication and G2 surveillance. The underlying principle is that a cell already burdened by replication problems may be less able to tolerate further loss of protective signaling. This can overlap with synthetic lethality, in which combined defects produce a much stronger loss of viability than either defect alone. Selectivity is not guaranteed, because proliferating normal tissues also depend on these pathways.
Several important distinctions prevent overgeneralization. PARP inhibitors primarily exploit aspects of DNA-repair vulnerability rather than simply switching off a cell-cycle checkpoint. CDK4/6 inhibitors restrain a central cell-cycle engine through the RB pathway rather than reproducing every component of DNA-damage signaling. Drugs that disrupt microtubules can activate the spindle checkpoint indirectly by preventing productive attachments. The term “checkpoint inhibitor” is therefore ambiguous in oncology: it can refer to cell-cycle surveillance targets or to an immune checkpoint, which regulates immune responses through a biologically different set of mechanisms.
Experimental analysis and unresolved questions
Checkpoint function is established by combining measurements of the initiating defect, signaling activity, cell-cycle progression, and cellular outcome. No single marker provides all four. Phosphorylated histone H2AX, commonly called γH2AX, can indicate a chromatin response to DNA damage or replication stress, but it is not a direct count of unrepaired double-strand breaks under every condition. Likewise, phosphorylation of CHK1 or CHK2 shows pathway engagement without necessarily proving that a particular cell-cycle transition has been delayed.
Flow cytometry provides population-level measurements of DNA content and DNA synthesis. In a conventional diploid population, G1 cells have approximately DNA, cells in S phase range between and , and G2 and mitotic cells have approximately . Here denotes the DNA content of one unreplicated haploid genome, not chromosome number. DNA content alone cannot distinguish G2 from mitosis, and tetraploid G1 cells can overlap with diploid G2 cells. Incorporation of nucleotide analogues and additional markers are therefore needed to resolve these states.
Live-cell microscopy reveals whether an apparent population arrest represents a uniform delay or several different behaviors. Fluorescent reporters of CDK activity, cell-cycle-regulated protein degradation, and damage-associated structures can track individual cells through arrest, recovery, division, or death. For the spindle checkpoint, investigators can visualize kinetochore attachment, chromosome alignment, and the interval between nuclear-envelope breakdown and anaphase. Single-cell trajectories are particularly valuable because a population average can conceal a mixture of cells that recover promptly and cells that never resume proliferation.
Genetic experiments test causality by disrupting a signaling component and asking whether an abnormal condition still delays progression. Pharmacological inhibitors can provide rapid perturbations, but their selectivity and indirect effects require controls. Synchronization procedures also complicate interpretation: thymidine blocks, microtubule poisons, and related methods can themselves create the stress being studied. Orthogonal approaches, including rescue with a functional gene and direct examination of chromosome integrity, help distinguish checkpoint bypass from correction of the original defect.
Current questions concern how local lesions are translated into a cell-wide delay, how signaling thresholds vary among cell types, and why genetically similar cells choose different fates after comparable stress. Quantitative models examine feedback, signal duration, protein turnover, and competition between arrest and death pathways. Studies in organoids and intact tissues investigate how extracellular signals and tissue architecture modify responses established in cultured cells. An important experimental challenge is to distinguish a checkpoint’s capacity to delay division from its capacity to improve the genetic integrity of the descendants that eventually survive.