Molecular Biology of the Cell - Volume 2 - Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson J. 1993

Intracellular macromolecular sorting and the maintenance of cellular compartments
Chromosome replication

Before every Cell Division, The Cell must synthesize copies of all its Chromosomes. Thus, cell division is preceded by the transition from interphase (G₁ phase) to the DNA Synthesis phase (S phase). In a typical higher Introduction/5.html">Eukaryotic Cell, the S phase lasts 8 hours. By its end, each chromosome consists of two copies that remain joined at the centromere until the M phase begins (see Fig. 9-35). Chromosome duplication requires METABOLISM/36.html">DNA Replication followed by the assembly of chromosomal Proteins, which form Chromatin, onto the newly synthesized DNA molecules. In Chapter 5, we discussed the Enzymes involved in DNA Replication and The Structure of the Replication fork that drives this synthesis (see Fig. 5-39). The transition of a cell into the S phase will be examined in Chapter 13 as part of the broader issue of Cell Cycle control. In this section, we outline the principles of the Eukaryotic chromosome replication mechanism, specify the time required for this process, and analyze the relationship between replication and Chromosome structure.

9-18

9.3.1. In Chromosomes of Higher Eukaryotes, Replication Initiation Sites Are Activated in Clusters [31]

As noted earlier, replication forks arise on the bacterial chromosome at specific DNA sequences known as replication origins (replication initiation sites) (see Section 5.3.9). At each such site, two forks form and move in opposite directions at a rate of about 500 NUCLEOTIDES per second until replication of the entire circular bacterial chromosome is complete. The bacterial genome is so small that these two replication forks can fully duplicate it in less than 40 minutes.

In the early 1960s, a method was developed to study the Selection/11.html">General features of eukaryotic chromosome replication. Human Cells in culture are pulse-labeled with a radioactive tracer (³H-thymidine), after which the cells are gently lysed, and the DNA is spread onto The surface of a Glass slide coated with a photographic emulsion. The autoradiograph is then examined under a Light Microscope. Replicated DNA is detected as a trail of silver grains. This method allows researchers to determine both the rate and Direction of replication fork movement (Fig. 9-55). Based on the rate at which the length of eukaryotic replicated DNA tracks increases with longer labeling times, replication forks in these organisms move at approximately 50 nucleotides per second—roughly ten times slower than in Bacteria. This is likely due to the fact that DNA packaged within a chromosome is more difficult to replicate.

Class="center">

Fig. 9-55. Schematic diagram of an experiment designed to analyze replication movement during the S phase. Newly synthesized DNA in human cell culture was pulse-labeled with [³H]-thymidine. A. Cells were lysed, and the DNA was stretched on a glass slide, which was then coated with a light-sensitive emulsion. After several months, the photographic emulsion was developed, revealing a series of silver grains along the radioactive DNA. B. The same experiment followed by incubation in a medium lacking the label. Additional DNA synthesized under these conditions exhibits a reduced level of radioactivity. The silver grains in the pair of dark tracks (B) appear to diverge in opposite directions, demonstrating that the fork moves outward from THE ORIGIN OF replication (see Fig. 5-49). DNA in this figure is color-coded solely to make the autoradiograph easier to interpret. In actual experiments, unlabeled DNA is not visible.

According to current models, the average human chromosome consists of a single DNA molecule containing about 150 million nucleotide pairs. To replicate such a molecule from end to end with a replication fork moving at 50 nucleotides per second would take $0.02 \times 150 \times 10^6 = 3.0 \times 10^6$ seconds (roughly 800 hours). However, autoradiographic analysis revealed that many forks move simultaneously along each eukaryotic chromosome. Moreover, multiple forks are frequently detected in the same region of DNA, while other segments of the same chromosome lack them entirely. Subsequent, more detailed experiments showed that: 1) replication origins are typically activated in clusters of 20 to 80 sites (called replication units); 2) new replication units continue to be activated throughout the S phase until all DNA is replicated; 3) individual replication origins within a replication unit are spaced 30,000 to 300,000 nucleotide pairs apart, meaning that, on average, there is roughly one such origin per chromatin loop; 4) in eukaryotes, as in bacteria, a pair of replication forks is formed initially, and as they diverge from the replication origin, a replication bubble is created. A fork stops only when it collides with a replication fork moving in the opposite direction (or when it reaches the end of the chromosome). Thus, numerous replication forks can move independently along each chromosome, ultimately producing two complete daughter DNA helices (Fig. 9-56).

Fig. 9-56. Features of DNA Replication in eukaryotic chromosomes. Replication origins in most cells are spaced 30,000 to 300,000 nucleotide pairs apart. Replication forks are thought to halt only when they meet an oncoming fork moving in the opposite direction or when they reach the end of the chromosome, ensuring complete replication of the DNA.

9-19

9-21

9.3.2. Specific DNA Sequences Serve as Replication Origins [32]

In bacteria and several animal Viruses, replication origins (points of replication initiation) are defined by specific DNA sequences. One might expect that specific DNA sequences also serve as replication origins in eukaryotes. The most successful searches for such sequences have been conducted in the Yeast *Saccharomyces cerevisiae*. These searches utilized cells defective in an essential Gene. Such cells can grow on a selective medium only if they carry a plasmid containing an active copy of the missing gene. When a bacterial plasmid carrying the essential gene is transferred into a defective yeast cell, the plasmid fails to replicate because the bacterial replication origin within the plasmid DNA cannot initiate replication in a yeast cell. However, if random fragments of yeast DNA are inserted into the plasmid prior to transfer, a small fraction of the plasmid DNA molecules will acquire a yeast replication origin and thus become capable of replication. Yeast cells harboring such Plasmids can grow because they have received the essential gene (Fig. 9-57). The yeast DNA sequences present in plasmids isolated from these cells are termed autonomously replicating sequences (ARS elements). ARS sequences occur in the yeast genome with roughly the frequency postulated for replication origins (approximately one every 40,000 nucleotide pairs). ARS function requires an 11-nucleotide consensus sequence [(A or T)TTTAT(A or G)TTT(A or T)]. An artificially constructed tandem repeat of these sequences, when inserted into a plasmid, Functions as an ARS in yeast cells. This implies that the initiator protein must recognize adjacent copies of this sequence to trigger replication. Attempts have been made to identify the DNA segments that act as replication origins in higher Eukaryotic cells, but these sequences are not yet well characterized.

Fig. 9-57. Method for identifying autonomously replicating sequences (ARS elements) in yeast. These sequences function as replication origins, and plasmids containing them can replicate independently within the host cell without integrating into chromosomes.

9-20

9.3.3. SV40 Viral Chromosome Replicates in a Mammalian Cell-Free System [33]

In prokaryotic replication systems, the movement of the replication fork is driven by a multi-enzyme complex consisting of DNA polymerase, DNA primase, and DNA helicase (see Section 5.3.6). This complex assembled at the replication origin through a reaction involving an initiator protein that specifically binds to the DNA origin (see Fig. 5-50). To identify all the enzymes involved in mammalian DNA replication, numerous attempts have been made to reconstitute the process *in vitro*. The most successful *in vitro* systems developed to date have successfully replicated the small circular chromosome of the simian virus SV40. All the proteins required for its replication—with a single exception—are supplied by the host cell. The exception is the SV40 T-antigen, a large multifunctional protein (90,000 daltons) that allows the virus to bypass the block on repeated DNA replication (see Section 9.3.11) and thereby replicate faster than the host cell DNA. Multiple copies of the T-antigen bind specifically to the SV40 origin and function simultaneously as both an initiator protein and a DNA helicase to unwind the DNA double helix at this site. Next, a complex of DNA polymerase (DNA polymerase $\alpha$) and DNA primase (both host-cell enzymes) attaches to one of the strands in this region. The resulting replication fork resembles the prokaryotic replication fork, except that it utilizes two Different types of DNA polymerases: DNA polymerase $\alpha$ on the lagging strand and DNA polymerase $\delta$ on the leading strand. In contrast, prokaryotes employ different molecules of the same DNA polymerase on both sides of the replication fork (see Fig. 5-47).

The protein that initiates replication on eukaryotic chromosomes has not yet been discovered. Consequently, it remains unknown whether this protein acts as a DNA helicase like the T-antigen or whether, as in prokaryotes, distinct proteins assume the initiator and helicase roles.

9.3.4. As DNA Replicates, New Histones Form Chromatin [34]

The formation of new chromatin in each cell cycle requires a massive amount of histones, with a total mass roughly equal to that of the newly synthesized DNA. Accordingly, most organisms possess multiple gene copies for each histone type. For instance, vertebrate cells contain approximately 20 identical gene clusters, each housing all five histone genes.

Unlike most cellular proteins, which are synthesized throughout interphase, histones are produced predominantly during the S phase. During this period, enhanced Transcription coupled with reduced histone mRNA turnover leads to a roughly fivefold increase in intracellular mRNA levels. Due to the unusual structure of their 3' ends (see Section 10.4.13), bulk histone mRNAs are extremely unstable and degrade within minutes after DNA synthesis ceases at the end of the S phase (or upon The addition of a DNA synthesis inhibitor). Conversely, the histone proteins themselves are very stable and persist throughout the life of the cell. The tight coupling between histone and DNA synthesis may be mediated, at least in part, by a feedback mechanism that monitors the pool of free histones to ensure that histone production matches The amount of newly synthesized DNA.

Once nucleosomes are assembled, histone molecules rarely, if ever, dissociate from the DNA they are bound to. Consequently, as the replication fork advances, it must negotiate parental nucleosomes, which appear structurally adapted to avoid hindering Transcription and Replication. According to one hypothesis, each nucleosome temporarily disassembles into two "half-nucleosomes" during DNA replication, allowing DNA polymerase to copy the unwound template (Fig. 9-58). The newly synthesized DNA left in the wake of the replication fork inherits the old histones and, to complete chromatin packaging, requires an equivalent Complement of new histones (see Fig. 5-58). New histones typically bind within the first few minutes after the passage of the replication fork. However, complete maturation of newly assembled chromatin takes up to an entire hour. During this window, new histones exhibit greater sensitivity to modifying enzymes than old histones. The exact chemical differences between mature and immature chromatin remain unclear; it is believed that maturation requires both covalent histone modifications and alterations in the capacity to bind other chromatin-associated proteins.

Fig. 9-58. A hypothetical diagram showing how a nucleosome might open up for DNA replication and then reassemble after the passage of the replication fork. In this model, the histones that make up the nucleosome remain continuously bound to the DNA. Here, the old nucleosome is inherited intact by the DNA helix formed on the leading strand, although how it is actually inherited remains unknown. This represents just one of many possible models.

9-5

9.3.5. Telomeres consist of short G-rich repeats added to the ends of chromosomes [35]

As noted above, the inability of DNA polymerase to completely replicate the ends of linear DNA molecules led to The Emergence of specific DNA sequences called telomeres at the ends of eukaryotic chromosomes (see Section 9.1.2). These regions have the same structure in organisms as diverse as Protozoa, Fungi, plants, and mammals. They consist of multiple tandem repeats of a short sequence containing a block of G-nucleotides (Fig. 9-59, A). The G-rich telomeric sequence is always located at the 3' end of the DNA molecule and apparently folds into a specialized structure that protects the chromosome end. A putative mechanism for the replication of telomeric DNA in the ciliated protozoan Tetrahymena is shown in Fig. 9-59, B.

9.3.6. Different Regions of the same chromosome replicate at different times during the S phase [36]

Given The rate of movement of the replication fork and the distance separating two adjacent origins of replication, we can conclude that DNA synthesis in this region normally takes about an hour. Despite this, the S phase in mammalian cells typically lasts approximately 8 hours. It follows that not all replication origins are activated simultaneously, and the replication time of each group of these sites (20–80 sites) occupies only a small fraction of the entire S phase.

Fig. 9-59. A. Repetitive G-rich sequences that form the telomeres of various eukaryotic organisms. B. A potential pathway for telomere formation (based on data obtained from Tetrahymena). An incomplete strand is synthesized on the lagging strand of the DNA template molecule (see Fig. 5-39). Upon binding to the telomere, Tetrahymena telomerase recognizes the structural feature of the G-rich sequence. Since this enzyme uses nucleoside triphosphates as substrates, the nucleotide at the 3' end must signal to the telomerase which nucleotide, G or T, to add at that stage. The reactions that shorten and lengthen the telomere sequence are not entirely balanced, with the result that chromosome ends consist of a variable number of repeats (roughly hundreds of nucleotide pairs).

Fig. 9-60. Light micrograph of stained mitotic chromosomes. Replicating DNA was labeled in the S phase preceding mitosis. To do this, cells were grown in the presence of the synthetic nucleoside 5-bromodeoxyuridine (BrdU) and then pulsed with thymidine for short intervals in the early, middle, or late S phase. Because DNA synthesized during the pulse-labeling period has thymidine in one strand and BrdU in the other, these segments stain more intensely than the rest of the DNA, which incorporates BrdU into both strands. Dashed lines connect the corresponding regions on the three copies of the chromosome. (Courtesy of Elton Stubblefield.)

Are different replication units activated at random, or is there a strict order according to which specific regions of The Genome are duplicated? To answer this question, a series of experiments was performed. A synchronized cell culture in the S phase was pulse-labeled with the thymidine analog 5-bromodeoxyuridine (BrdU). Regions of mitotic chromosomes that incorporated the label into their DNA can be detected in the M phase by a decrease in staining with a specific dye or by binding with specific Antibodies. Such experiments have shown that chromosomal regions replicate as large units, and the order of their duplication in the S phase is strictly regulated for each chromosome (Fig. 9-60).

9.3.7. Highly condensed chromatin replicates in the late S phase [37]

In the cells of higher eukaryotes, some DNA regions are more condensed than others. For example, in interphase, heterochromatin remains in a heavily condensed state (resembling the conformation of chromosomes in mitosis), whereas active chromatin becomes less compact, which is apparently necessary for RNA Synthesis (see Section 10.3.11).

To understand the mechanism that determines the timing of replication for a specific region, data showing that the heterochromatin block (including centromeric regions) remains condensed in interphase and replicates in the late S phase are of great importance. This late replication is apparently linked to the packaging of DNA into chromatin. This Conclusion is supported by the timing of replication of the two X chromosomes in mammalian female cells. These two chromosomes are known to contain identical DNA sequences, but one is active and the other is not (see Section 10.3.9). Almost all inactive X chromosomes consist of heterochromatin, and their DNA replicates in the late S phase; the active X chromosome is less condensed and replicates throughout the S phase. The obvious conclusion is that chromatin that is less condensed in interphase is more accessible to the replication machinery, and the regions composed of it replicate first.

On the other hand, autoradiography data indicate that all replication forks move at comparable speeds throughout the S phase; thus, the degree of chromosome Condensation probably does not affect the operation of the fork once it has formed. However, the order of activation of replication origins appears to depend, at least in part, on the Chromatin Structure of the given region.

9.3.8. Genes within active chromatin replicate in the early S phase [38]

The Link Between chromatin structure and DNA replication timing is supported by data on the replication timing of individual genes. A population of growing cells was labeled with bromodeoxyuridine, after which the cells were immediately separated by size using centrifugation. Because cell growth is coupled to the cell cycle, larger cells are "older," and their DNA will therefore be labeled at a later stage of the S phase. BrdU-labeled DNA was isolated from each cell type and hybridized with a series of known DNA probes to determine which genes it contained. Because BrdU-containing DNA is denser, it can be easily separated from normal DNA by cesium chloride density gradient centrifugation (see Section 4.4.1).

Fig. 9-61. One of the models explaining why active chromatin (highlighted in color) replicates in the early S phase, whereas inactive chromatin replicates in the late S phase. It is suggested that different initiator proteins act on active and inactive chromatin at the replication origin.

According to another hypothesis, the Two Types of replication initiation sites use the same molecular mechanism but at different times, simply because the condensed structure of conventional chromatin hinders access for the proteins required for replication.

Researchers thus have a powerful METHOD FOR DETERMINING the replication timing of any gene for which a DNA probe is available. It turned out that in all studied cells, housekeeping genes (i.e., those active in all cells) replicate in the early S phase. Conversely, genes active in certain cell types replicate early only in the cells where they are active, and later in other cell types. For example, when studying long sequences of the 300,000-nucleotide-pair immunoglobulin gene in this manner, it was found that in cells where this gene is active, all chromatin regions containing it complete their replication at the beginning of the S phase. This is likely due to the presence of multiple replication origins within the gene that are activated at approximately the same time. A different picture is observed in cells lacking immunoglobulin synthesis. Using DNA probes, a single replication fork was detected that originated at one end of this chromosomal region about an hour after THE START OF the S phase and then moved along the DNA at a constant rate of about 3,000 nucleotides per minute.

Fig. 9-61 presents a simple model explaining these results. According to this model, all replication origins located in active chromatin begin to function at the very beginning of the S phase. Subsequently, the replication forks arising at these sites inevitably move into adjacent chromosomal regions where the chromatin has a more condensed structure; therefore, any gene located less than a million nucleotide pairs from the replication origin will replicate in the middle of the S phase. To understand how inactive chromatin regions located far from active chromatin can replicate, we must assume that a second group of replication origins is activated in the middle or late S phase, which can initiate the formation of replication forks in chromatin of any structure.

9.3.9. Late-replicating regions coincide with AT-rich bands on metaphase chromosomes [23]

Presumably, many replication units correspond to the various chromosomal segments revealed by the special fixation and staining Methods used in karyotyping.

As noted above (see Section 9.2.3), about 2,000 dark AT-rich segments (G bands), separated by light GC-rich regions (R bands), can be detected by staining in the haploid set of mammalian mitotic chromosomes. Notably, AT-rich and GC-rich DNA differ in their replication timing during the S phase. Experiments similar to the one shown in Fig. 9-60 have demonstrated that the majority of GC-rich bands replicate in the first half of the S phase, whereas most AT-rich bands replicate In the second half. From this, we can conclude that housekeeping (constitutive) genes are located primarily in GC-rich bands, whereas genes whose activity is characteristic of specialized cells belong to AT-rich segments. It remains a mystery why the mammalian genome should be subdivided into such large alternating blocks of chromatin, many of which approach the size of an entire bacterial genome. It is also unknown how many replication origins from each replication unit are activated simultaneously. Perhaps in late-replicating replicons, the chromatin remains in a condensed state even after the M phase ends and decondenses only in the middle of the S phase, exposing all replication origins in the replicon at once. In this case, coordinated DNA replication within a single replication unit, proceeding on an "all-or-none" principle, may reflect the cooperative nature of chromatin decondensation (see Section 9.1.21).

9.3.10. Why is the control of replication origin firing time necessary? [39]

The S phase is completed exceptionally rapidly in the cleaving eggs of many species, which store a large supply of chromatin components (such as histones) to rapidly provision new nuclei. As shown in Fig. 9-62, a short S phase requires The Use of an extremely large number of replication origins spaced only a few thousand nucleotide pairs apart (rather than the tens or hundreds of thousands of nucleotides separating the corresponding sites at later developmental stages). Because any foreign DNA injected into a fertilized Xenopus egg (including small circular fragments of bacterial DNA) is replicated, we can conclude that a very short, simple sequence present in any DNA molecule functions as a replication origin in this cell. DNA replication is a process that can proceed at a remarkably high rate and, in most cells, is regulated by a complex system that ensures a pattern of replication fork generation in which different PARTS OF THE genome replicate at quite different times. It is possible that the timing of replication influences chromatin structure. For example, according to one hypothesis, early-replicating chromatin is assembled from a special pool of chromosomal proteins synthesized during the G₁ phase, which help maintain the chromatin in an active state and thereby facilitate the transcription of expressed genes.

Fig. 9-62. Transmission electron micrograph of extended chromatin from a Drosophila embryo. Note that replication bubbles are located very close to one another (indicated by arrows). In these embryos, only 10 minutes elapse between nuclear divisions. The replication origins visible in the micrograph are spaced so closely (separated by a few thousand nucleotide pairs) that only one minute is required to replicate the DNA enclosed between them. (Courtesy of Victoria Foe.)

9-21

9-22

9.3.11. Chromatin-Associated Factors Ensure That DNA Replicates Only Once per S Phase by Blocking Re-Replication [40]

In a normal cell, the entire genome must be replicated precisely once during the S phase. As discussed above, DNA replication in most eukaryotic cells occurs asynchronously over a specific period. Because replication origins in different regions of the chromosome are activated at different times, by the middle of the S phase some parts of the chromosomes have not yet begun to replicate, whereas others have already finished. Thus, by mid-to-late S phase, a complex "bookkeeping" problem arises. By this time, some replication origins have already doubled, whereas others (presumably with a similar DNA sequence) have not yet been activated. Yet, we know that each replication origin can be used only once per S phase. How is this achieved?

The key to solving this problem came from Cell Fusion experiments. If cells in the S phase are fused with cells in the G₁ phase, DNA synthesis is induced in the G₁ nucleus. This suggests that the transition from the G₁ to the S phase is driven by a diffusible activator of DNA synthesis. A different picture emerges when cells in the S phase are fused with cells at the G₂ stage (i.e., those that have just completed the S phase). In this case, DNA synthesis is not initiated in the G₂ nucleus, while it continues unabated in the S-phase nucleus. Thus, a nucleus at the G₂ stage (whose DNA has already replicated) behaves as though it is protected from entering subsequent replication cycles by a nondiffusible inhibitor tightly bound to the DNA. Such an inhibitor, if it truly exists, could help each replication fork solve the "bookkeeping" problem: by modifying the chromatin of newly synthesized DNA, it would prevent the re-replication of already replicated DNA within the same S phase (Fig. 9-63, A). Alternatively, one can envision a mechanism based on initiator proteins that are tightly bound to DNA and become inactivated after the passage of the replication fork (Fig. 9-63, B). Whatever the mechanism that blocks DNA re-replication, it must disappear by or during mitosis, because once cell division is complete, the DNA delivered to a daughter cell whose nucleus is in the G₁ stage is no longer protected. This re-replication block has been shown to operate on bacterial DNA fragments injected into a fertilized frog egg as well. Consequently, the mechanism responsible for the block does not require a highly specific replication origin. The block does not affect SV40 virus, presumably because its T antigen possesses both initiator and helicase activities, thereby replacing analogous host cell components whose function is not yet clear (see Section 9.3.3).

Fig. 9-63. Two possible mechanisms for blocking DNA re-replication, which prevents already replicated DNA from doubling within the same cell cycle. The absence of such a block in certain cell types (such as Drosophila salivary gland cells) leads to the formation of giant polytene chromosomes. (A) A model in which an inhibitor binds to newly synthesized chromatin. (B) A model proposing the existence of single-use initiator proteins that bind to DNA only during mitosis.

Summary

Studies using various model systems for DNA replication have shown that in eukaryotes, as in prokaryotes, the process begins with the loading of a helicase onto the DNA via an initiator protein that binds to the replication origin. As the two replication forks move apart, a replication bubble is formed. In higher eukaryotes during the S phase, adjacent replication origins appear to be activated in clusters (termed replication units). Because a replication fork moves at a rate of about 50 nucleotides per second, completing DNA replication within a single unit takes approximately one hour. Over the course of the 8-hour S phase, different clusters of replication origins are activated in a set sequence. This activation order is determined in part by their chromatin structure, with the most condensed chromosomal regions replicating last. The relationship between replication units and the segments observed in eukaryotic mitotic chromosomes suggests that replication units may correspond to structurally distinct domains of interphase chromatin.

Following the passage of a replication fork, the chromatin structure is reassembled by adding new histones and other chromosomal proteins to the legacy histones inherited by the daughter DNA molecule. To prevent a second round of replication from following immediately on the heels of the first (i.e., before the chromosome enters mitosis), the cell employs a block. This safeguard ensures that every region of the DNA is replicated precisely once per S phase.