DNA replication takes place during the S phase (synthesis phase) of the cell cycle, a dedicated window sandwiched between two gap phases. In a typical human cell, S phase lasts roughly 6 to 8 hours, during which the entire genome is copied once and only once. The cell cycle enforces this timing through an elaborate set of molecular switches that prepare origins of replication before S phase begins and then dismantle the preparation machinery once copying starts, making a second round of replication all but impossible until the cell divides again. That single-copy rule is one of the most tightly controlled events in biology, and when it breaks down, the consequences range from chromosomal abnormalities to cancer.
How S Phase Fits into the Larger Cell Cycle
The cell cycle is traditionally divided into four consecutive stages. During G1 (gap 1), the cell grows, produces proteins, and monitors its environment for signals telling it whether to commit to division. If conditions are favorable, the cell crosses a decision point and enters S phase, where all of its chromosomal DNA is duplicated. After S phase comes G2 (gap 2), a shorter interval during which the cell checks the new copies for errors and prepares for mitosis. Finally, during M phase (mitosis), the duplicated chromosomes separate and the cell physically splits into two daughter cells.
G1, S, and G2 together make up what is called interphase, the period between one mitosis and the next. In rapidly dividing human cells, the whole cycle takes about 24 hours, with S phase occupying roughly a third of that time. G1 is the most variable phase; it can last just a few hours in fast-growing tissue or stretch to days or even indefinitely in cells that have paused division. G2 and M phase are comparatively short and consistent.
Preparing for Replication in G1
DNA replication does not simply switch on when S phase begins. The groundwork is laid during late mitosis and G1, when the cell “licenses” its replication origins. Thousands of specific sites along each chromosome are marked as future starting points for copying. At each of these origins, a group of proteins called the pre-replicative complex assembles, with a ring-shaped complex known as MCM2–7 loaded onto the DNA. This loading step is the license: it signals that a particular stretch of DNA is cleared to be copied once S phase arrives.
The licensing step depends on helper proteins like Cdt1 and Cdc6, which recruit MCM2–7 to origins during G1. Once S phase starts, this preparation machinery is actively dismantled. MCM2–7 moves away from the origin as the replication fork progresses, and both Cdt1 and Cdc6 are broken down or exported out of the nucleus.
Why DNA Is Copied Exactly Once Per Cycle
If a stretch of DNA were accidentally copied twice in one S phase, the resulting cell would have three copies of that region instead of two. This kind of over-replication can cause breaks, rearrangements, and gene amplification, all of which are hallmarks of genomic instability. Cells prevent this by splitting replication into two non-overlapping phases: licensing (which only happens in late mitosis and G1) and firing (which only happens in S phase).
The key insight is that the conditions allowing origin licensing are mutually exclusive with the conditions allowing origin firing. During G1, the enzymes that would trigger replication forks are kept inactive, so origins get loaded but do not fire. When S phase begins, rising levels of specific protein kinases flip the switch: they activate the loaded origins to start copying but simultaneously destroy the licensing factors so no new MCM2–7 complexes can be loaded onto already-fired origins.
Cyclin E paired with CDK2 and cyclin A paired with CDK2, along with another kinase called CDC7, phosphorylate components of the pre-replicative complex to initiate replication.
What Actually Happens During S Phase
Once origins fire, two replication forks emerge from each origin and move in opposite directions, unzipping the double helix and synthesizing a new complementary strand on each template. Human cells have tens of thousands of replication origins, but they do not all fire simultaneously. Instead, origins fire in a coordinated sequence that unfolds over the entire duration of S phase.
Regions of the genome that are actively used by the cell tend to replicate early in S phase, while more tightly packed or less active regions replicate later. This replication timing program is spatially organized inside the nucleus: early-replicating regions cluster in the interior of the nucleus, while late-replicating regions concentrate near the nuclear periphery. The timing pattern is remarkably stable from one cell division to the next and is thought to protect the genome by ensuring that the copying machinery does not overwhelm the cell’s resources at any single moment.
As the replication forks advance, the cell must also produce enormous quantities of histone proteins. Histones are the spool-like proteins around which DNA is wound, and every newly copied stretch of DNA needs a fresh set of them. Cells restrict the bulk of histone production to S phase itself; accumulating excess histones outside of S phase is toxic.
Quality Control During Replication
Copying three billion base pairs of DNA in a matter of hours inevitably risks errors and encounters obstacles like damaged bases or tangled strands. To handle this, cells have an internal alarm system called the intra-S checkpoint. When the replication machinery hits a problem, this checkpoint slows or pauses fork progression, delays the firing of origins that have not yet started, and activates DNA repair pathways.
The intra-S checkpoint protects the genome by regulating origin firing, fork progression, and gene expression in response to damage encountered during copying. Without it, cells would barrel through DNA lesions, generating breaks and mutations at a far higher rate. Defects in this checkpoint are a common feature of cancer cells, which helps explain their characteristic genomic chaos.
When Replication Spills Outside S Phase
Although S phase is the designated window for DNA replication, some copying can occur outside it under stress. When replication forks stall and cannot be rescued during S phase, certain difficult-to-copy regions of the genome remain under-replicated as the cell enters mitosis. To deal with this, cells can activate a backup process called mitotic DNA synthesis, or MiDAS, which attempts to finish copying those stubborn regions during early mitosis.
MiDAS is a last resort, not a routine pathway. It operates at chromosomal fragile sites, regions that are inherently hard to replicate because of their unusual sequence or structure. When even MiDAS cannot finish the job, the result is threads of DNA connecting the two separating chromosome sets, known as ultra-fine anaphase bridges. If these bridges are not resolved before the cell completes division, they lead to mis-segregated chromosomes, micronuclei, and reduced cell survival.
Variations on the Theme
The textbook G1-S-G2-M cycle is common in most dividing cells, but nature has modified the pattern in several contexts.
In early embryonic development, the gap phases are effectively eliminated. After fertilization, early cleavage divisions cycle rapidly between S phase and M phase with virtually no G1 or G2. The embryo inherits a large stockpile of maternal proteins and messenger RNA from the egg, so it does not need the growth and checkpoint time that gap phases provide. These truncated cycles allow the single-celled zygote to divide into hundreds of cells in a matter of hours.
Some specialized cell types use a variant called the endoreplication cycle, in which cells duplicate their DNA without dividing at all. The result is a cell with extra copies of the entire genome. This strategy is widespread in plants and in certain animal tissues. In fruit flies, for example, larval salivary gland cells undergo many rounds of endoreplication to produce the giant polytene chromosomes that biologists have studied for over a century. In mammals, liver cells and the giant cells of the placenta called trophoblasts also undergo endoreplication to increase their DNA content and metabolic output.
Meiotic S Phase Is Different Too
Cells destined to become eggs or sperm go through a specialized form of division called meiosis, which ultimately halves the chromosome number. Before the first meiotic division, these cells pass through their own S phase to duplicate their DNA. Meiotic S phase is considerably longer than the mitotic version. In the roundworm C. elegans, for example, meiotic S phase is at least twice as long as the mitotic S phase in the same organism’s germ cells.
The extended duration likely reflects the extra work the cell must do during meiotic S phase. Homologous chromosomes need to be prepared for the pairing and recombination events that will follow in meiotic prophase. The replication machinery may also move more slowly or fire fewer origins at once to accommodate these preparations. Regardless of the mechanism, the principle remains the same: DNA is copied during an S phase, just a slower and more elaborate one.
Bacteria Handle Replication Differently
Bacteria lack a cell cycle in the eukaryotic sense, but they still face the fundamental challenge of copying their DNA before dividing. In many bacteria, replication initiation is tightly coupled to cell growth. According to a longstanding model, a bacterial cell initiates a new round of chromosome replication when it reaches a critical mass. Nutrient availability directly affects growth rate, which in turn sets the pace of replication initiation.
Fast-growing bacteria like E. coli can actually have multiple overlapping rounds of replication running simultaneously. When nutrients are abundant and cells divide faster than the time it takes to copy the whole chromosome, a new round of replication begins before the previous one finishes. This means a rapidly dividing bacterial cell can contain four or even eight partially replicated copies of its chromosome at once. The concept of a discrete “S phase” does not apply in the same way; replication is continuous and overlapping rather than confined to a single window.
When Cells Stop Replicating Entirely
Not every cell in your body is actively cycling. Neurons, mature muscle fibers, and most cells in adult tissues sit in a quiescent state called G0, a resting phase branching off from G1. Quiescent cells are alive and metabolically active but have exited the cell cycle and do not replicate their DNA. Many can re-enter the cycle if stimulated by the right growth signals.
Senescence is a more permanent exit. Senescent cells have entered an essentially irreversible state of proliferative arrest in which they remain metabolically active but do not respond to signals that would normally push them back into the cycle. Senescent cells accumulate with age and contribute to chronic inflammation by secreting a cocktail of inflammatory and tissue-degrading molecules. Because these cells are stuck permanently before S phase, their DNA is never replicated again, yet they persist in tissues and influence the behavior of their neighbors.
Cancer Drugs That Target S Phase and Cell Cycle Entry
The tight link between S phase and cell proliferation makes it a prime target for cancer therapy. Many traditional chemotherapy agents work by disrupting DNA replication directly. Drugs like hydroxyurea deplete the raw materials that the replication machinery needs, stalling forks and triggering the intra-S checkpoint. Platinum-based drugs and alkylating agents damage the DNA template so that replication forks cannot proceed normally.
A newer class of drugs called CDK4/6 inhibitors takes a different approach. Instead of attacking S phase itself, these drugs block the cell from ever entering it. By inhibiting the kinases that drive the transition from G1 into S phase, CDK4/6 inhibitors keep cancer cells locked in G1 with unreplicated DNA. Their chief mechanism is preventing phosphorylation of the retinoblastoma protein, a key gatekeeper of S phase entry. These drugs have become a standard treatment for certain breast cancers and are being tested in other tumor types as well.
Researchers are also exploring strategies that specifically exploit the vulnerabilities of cells already in S phase. One recent approach combines a checkpoint kinase 1 inhibitor with low-dose hydroxyurea to push cancer cells into S-phase arrest and trigger a form of immunogenic cell death. The idea is that by forcing replication stress to an extreme, the dying cancer cells release signals that attract the immune system, turning a targeted killing strategy into one that also trains the body’s own defenses.
How Scientists Measure S Phase
Much of what we know about S phase timing comes from pulse-labeling experiments. Researchers expose cells to a modified nucleotide, a building block that gets incorporated into newly synthesized DNA but not into existing strands. By adding this label for a brief window and then tracking which cells took it up, scientists can identify exactly which cells were in S phase during the pulse.
Modern versions of this technique use molecules like EdU (5-ethynyl-2′-deoxyuridine) that can be detected with click chemistry, a fast and sensitive reaction that lights up labeled DNA under a fluorescence microscope or in a flow cytometer. In a typical flow cytometry experiment, researchers can separate cells into three populations: those with the normal amount of DNA (in G1), those with double the DNA (in G2 or M), and those with intermediate amounts of DNA that have incorporated the label (in S phase). The S-phase population forms a distinct arc between the G1 and G2 peaks.
These labeling methods have been adapted for use in living organisms and even in plants, where measuring S-phase duration is especially challenging because cell cycles vary widely between tissues and species. By following the labeled cohort of cells as they progress through S phase and into G2, researchers can estimate how long S phase lasts in a given cell type, data that feeds into models of tissue growth, development, and disease.
Replication Timing and Gene Activity
The order in which different parts of the genome are copied during S phase is not random, and it correlates with how genes in those regions are used. Actively transcribed genes, the ones a cell is reading to make proteins, sit in regions that replicate early. Genes that are silenced or packed away tend to sit in late-replicating regions. This correlation holds across many cell types and organisms, though the causal relationship is still debated: does early replication cause a region to be more active, or does activity cause early replication?
What is clear is that the replication timing program can shift when a cell changes its identity. When a skin cell is reprogrammed into a stem cell, for instance, the replication timing of hundreds of genomic regions flips. Regions that replicated late in the skin cell begin replicating early in the stem cell, and vice versa. These shifts often correspond to changes in which genes are switched on or off, reinforcing the idea that replication timing is not a fixed property of the DNA sequence but rather a reflection of how the cell is using its genome at that moment.
Disruptions to replication timing have been observed in many cancers, where large swaths of the genome shift from their normal early or late patterns. Whether these shifts are drivers of cancer or merely passengers of broader genomic disorganization is an active area of research, but the observation underscores how fundamental the S-phase timing program is to normal cell behavior.