11.3 MECHANICS OF THE DNA REPLICATION FORK

The advance of a replication fork requires the coordinated action of several different types of proteins, in addition to DNA polymerase. Many of these proteins work together as a highly dynamic “replisome” machine. Some proteins at a replication fork dissociate from the DNA and are replaced by new ones every few seconds, whereas others are left behind at intervals to direct clean-up processes after the replication fork has passed. In addition, unwinding of the parental DNA requires it to be broken and resealed hundreds of thousands of times, to relieve torsional stress.

Replication fork mechanics are best understood in the E. coli system, where the process is streamlined and illustrates the fundamental mechanism of a moving replication fork. Even so, more than a dozen different proteins are involved (Table 11-2). We describe here the individual replication proteins and how they act together at a replication fork in E. coli, then describe replication mechanisms in eukaryotes. Cellular processes in eukaryotes are generally more complex, and thus it is not surprising that several more proteins are involved at the replication fork.

DNA Polymerase III Is the Replicative Polymerase in E. coli

Pol III is responsible for replicating the E. coli chromosome. The Pol III core is a heterotrimer composed of three subunits called α, ε, and θ. The DNA polymerase activity is in the α subunit; the ε subunit has the proofreading 3′→5′ exonuclease activity. The role of the θ subunit is currently unknown. The crystal structure of the Pol III α subunit, shown in Figure 11-13, reveals the hand shape common to all DNA polymerases and the presence of a polymerase domain and a histidinol phosphate (PHP) domain. The PHP domain, unique to bacterial Pol III, has a chain folding pattern similar to that of certain phosphodiesterases, suggesting that this domain may be a vestigial 3′→5′ exonuclease. In E. coli, as in many bacteria, the PHP domain is not enzymatically active, and the α subunit recruits the ε subunit for 3′→5′ proofreading activity. In some bacteria, however, the PHP domain in the α subunit is, in fact, the 3′→5′exonuclease proofreading activity, and these α subunits might not recruit an ε subunit.

Figure 11-13: The E. coli Pol III α subunit. A space-filling model of the Pol III α subunit, with the palm, fingers, thumb, and PHP domains labeled.

378

The Pol III core is just one part of a much larger protein assembly called the Pol III holoenzyme, which replicates both leading and lagging strands (Table 11-3). The Pol III holoenzyme includes three Pol III cores, two ring-shaped β sliding clamps that improve processivity, and one clamp loader (also called τ complex) that assembles β clamps onto the DNA. The clamp loader includes three τ (tau) subunits with C-terminal domains that protrude from the clamp loader and bind to the Pol III cores (Figure 11-14).

Figure 11-14: The architecture of E. coli Pol III holoenzyme. The C-termini of the τ subunits protrude from the clamp loader and bind the Pol III cores. Each Pol III core also attaches to a β sliding clamp.

The Pol III core itself is capable of DNA synthesis at a slow rate, but DNA synthesis by the Pol III holoenzyme is exceedingly rapid, nearly 1 kb/s. The β clamps and clamp loader help maintain contact between the Pol III core and DNA, making the Pol III core highly processive (∼100 kb per binding event). Having two DNA polymerases in one holoenzyme assembly facilitates the coordinated synthesis of the leading and lagging strands at the replication fork. The β sliding clamps are assembled onto both DNA strands by the single clamp loader.

379

A DNA Sliding Clamp Increases the Speed and Processivity of the Chromosomal Replicase

The processivity (1 to 10 nucleotides) and rate of synthesis (10 nucleotides per second) of the Pol III core are dramatically enhanced by the β sliding clamp. The β clamp is a homodimer, shaped like a ring that encircles the duplex DNA (Figure 11-15a). The β clamp has sixfold symmetry; it is constructed from a single domain repeated three times in each monomer. The topological binding of the β clamp to DNA allows the clamp to slide along the duplex while staying tightly attached (see the How We Know section at the end of this chapter). The β clamp binds to the Pol III core (Figure 11-15b) and thereby holds Pol III to the DNA while sliding along the duplex. This converts Pol III core from a distributive enzyme, which moves on and off DNA as it works, to a processive enzyme that stays attached to DNA during repetitive cycles of dNMP incorporation (Figure 11-16).

Figure 11-15: The E. coli β sliding clamp. (a) Frontal view (left) and side view (right) of the E. coli β clamp bound to DNA. The DNA is tilted 22° from the perpendicular. Subunits of β are shades of purple; the DNA is blue. (b) Model of E. coli Pol III α bound to the β clamp.
Figure 11-16: Processivity conferred by the β clamp. (a) In distributive synthesis, the Pol III core would extend a primed site by a few nucleotides and then dissociate from the DNA. It must rebind the primed site to continue synthesis. (b) Binding of a Pol III core to the β clamp tethers the core to the DNA for processive synthesis. When the Pol III core detaches from the DNA, it stays attached to the β clamp and rapidly reattaches to the DNA.

An example of the remarkable speed and processivity conferred on the Pol III core by the β clamp is shown in Figure 11-17. The DNA substrate used in this experiment was the large (5.4 kb), single-stranded, circular DNA genome of phage φX174. With the Pol III core (within the holoenzyme) bound to the β clamp at a primed site, [32P]dNTPs labeled on the α phosphate were added to start the polymerase reaction. At different times, reaction samples were analyzed by agarose gel electrophoresis and autoradiography to visualize the newly synthesized radioactive DNA. The results show that Pol III goes full circle within a few seconds. When the β clamp was omitted in a control reaction, no DNA synthesis was observed, because without the clamp, Pol III runs into secondary structures that block synthesis (data not shown). However, even in the absence of these secondary structures, Pol III would require about 9 minutes to complete replication of the 5.4 kb DNA without the help of the β clamp.

Figure 11-17: Rapid DNA synthesis by E. coli Pol III. The β sliding clamp is loaded onto a primed site by a clamp loader in the Pol III holoenzyme. Synthesis of the 5.4 kb circular DNA of bacteriophage φX174 is completed within 11 seconds, as shown in the autoradiograph from the experiment described in the text.

The β sliding clamp does not assemble onto DNA by itself; it requires the multiprotein clamp loader to open and close the ring around the DNA. The clamp loader within the Pol III holoenzyme is the τ complex, which consists of several subunits, τ3 (τ trimer), δ, and δ′, arranged in a circular pentamer (see Figure 11-14), and two small subunits, χ (chi) and ψ (psi), that connect to other proteins at the replication fork. The clamp loader uses the energy of ATP binding to open the β sliding clamp. In this operation, the clamp loader binds to one of the flat surfaces of the clamp and forces it to open (Figure 11-18a). The gene that encodes the τ subunit (dnaX) also produces a shorter version, called γ, through a translation frameshift event. During translation frameshifting, the mRNA slips and places the mRNA in a different reading frame. In the case of dnaX, the new reading frame results in a stop codon after one amino acid, thus yielding a δ subunit that is about 2/3 the size of τ. The τ and γ subunits form homotrimers (τ3 and γ3), and each can assemble with δ, δ′, χ, and ψ to form a clamp loader, either a τ complex or a γ complex. Only the τ complex binds to Pol III cores for DNA replication. As described in Chapter 12, sliding clamps are used by many different proteins for a variety of purposes, not just for replication. The γ complex is used to assemble β clamps onto DNA for use by enzymes other than Pol III.

380

The clamp loader has an inner chamber that binds primed DNA, and this positions the DNA through the sliding clamp. The clamp-loading activity is specific to a primed site, because the DNA must bend out of a gap in the side of the clamp loader, and only single-stranded DNA, not double-stranded DNA, has the flexibility to make this sharp bend. The γ33, δ, and δ′ subunits of the clamp loader all have a homologous region that binds ATP, a domain found in a common class of ATPase proteins called AAA+ proteins (ATPases associated with a variety of cellular activities). The clamp-loading reaction is an unusual enzyme-catalyzed process because the DNA and protein substrates are not converted to a new product—they are simply intertwined. Many of the proteins involved in DNA replication have AAA+ domains and drive protein and DNA conformational changes (further discussed in Section 11.4).

In the absence of ATP, the clamp loader cannot bind the β clamp, because the subunits are oriented in a way that blocks their interaction with the clamp. ATP binding to the τ (or γ) subunits induces a conformational change that enables the clamp loader to bind and open the clamp (as shown for the γ complex in Figure 11-18b). ATP is also needed for the clamp loader to bind DNA. ATP hydrolysis causes the clamp loader to revert to the form that cannot bind the β clamp or DNA, thereby ejecting the clamp loader and allowing the clamp to close around DNA. It is important that the clamp loader be ejected at the end of the reaction because it binds to the same spot on β to which the Pol III core must attach.

Figure 11-18: The E. coli clamp loader. The γ complex is shown here. (a) The five clamp-loading subunits are arranged in a circle, with a gap between parts of two subunits, δ′ and δ. The β clamp (purple) docks onto the clamp loader. (b) The clamp loading mechanism. ATP binding to the γ subunits powers a conformational change that enables the binding and opening of the β clamp. The combined γ complex–ATP–β clamp binds primed DNA in a central chamber, and the single-stranded template DNA passes through the gap in the side of the clamp loader. ATP hydrolysis ejects the clamp loader, allowing β to close again around the DNA.

381

Many Different Proteins Advance a Replication Fork

The simultaneous replication of both DNA strands at a replication fork requires the interaction of many proteins in addition to the Pol III core and β clamp (Figure 11-19a). Here we give particular attention to the E. coli system, but all cells contain these basic protein components for chromosome replication.

Figure 11-19: Activities required at a DNA replication fork. (a) DNA polymerase and the β sliding clamp are required for processive DNA synthesis. (b) Helicases that function at replication forks are hexamers that encircle single-stranded DNA and translocate on the DNA strand to separate the strands of the parental duplex. (c) A topoisomerase removes twists in the DNA. In E. coli, the topoisomerase is DNA gyrase. (d) Primase (an RNA polymerase) makes short RNA primers to initiate DNA synthesis. Primases typically bind the helicase, thus localizing primers to the replication fork. (e) Ligase seals DNA nicks, joining Okazaki fragments together (after removal of the RNA primers). (f) Single-stranded DNA–binding protein (SSB) binds cooperatively to single-stranded DNA (ssDNA), removing secondary structure in the DNA strand and protecting it from the action of nucleases.

DNA Helicase The two strands of the parental DNA duplex are separated by a class of enzymes known as DNA helicases, which harness the energy of NTP hydrolysis (usually ATP) to drive strand separation (Figure 11-19b). Helicases are used in a wide variety of DNA and RNA transactions. DNA helicases usually load onto DNA at a single-strand gap in the duplex and move along the DNA strand in one direction (fueled by NTP hydrolysis), unwinding the duplex as they move. The direction of translocation along the DNA is characteristic of the particular helicase (see Chapter 5).

KEY CONVENTION

A helicase translocates along a single strand of DNA in one direction, parting the duplex as it moves. The direction of movement is specified, by convention, as the direction along the strand to which the enzyme is bound. If the helicase binds to a DNA strand and progresses from the 5′ end toward the 3′ end, it is said to be a 5′→3′ helicase.

Helicases that function at a replication fork are typically ring-shaped hexamers that encircle one DNA strand. The ring shape of replicative helicases is thought to enhance their grip on DNA for processive unwinding. The E. coli replicative helicase is a hexamer of DnaB protein; it encircles the lagging strand and translocates in the 5′→3′ direction. An assay that demonstrates the action of DnaB is shown in Figure 11-20. In this experiment, the DNA substrate included a central single-stranded DNA region for helicase assembly, flanked on each side by duplex DNA of different lengths; one strand of each duplex was radioactively labeled. The direction of helicase translocation was determined by observing which DNA fragment was unwound from the substrate, as analyzed on a polyacrylamide gel. The resulting autoradiograph shows that DnaB translocates in the 5′→3′ direction along a single strand of DNA and unwinds the duplex at only one of the two ends. Helicases require NTPs for translocation, and a control reaction conducted in the absence of ATP showed no unwinding.

Figure 11-20: An assay for determining the direction of DNA helicase activity. (a) DnaB helicase is added to a long, single-stranded DNA that has short 32P-labeled DNA strands of different sizes annealed to its ends (shown here are a 796-mer and 722-mer). Each annealed DNA strand has a short single-stranded tail to mimic a replication fork. DnaB initially binds to the single-stranded DNA region, then translocates in one direction to unwind one of the two annealed-DNA duplexes. (b) The DNA-unwinding products are analyzed in a polyacrylamide gel. The result here shows that DnaB displaced only the 722-mer, revealing that DnaB translocates in the 5′→3′ direction along single-stranded DNA.

382

383

Topoisomerase As helicase separates the two strands of DNA, the duplex DNA ahead of the helicase becomes overwound, and this creates superhelical tension ahead of the fork. This occurs because the helicase cannot untwist the DNA strands. The action of untwisting DNA requires a special type of enzyme that cuts one or both strands of DNA to unwind it, then reseals it. This untwisting action is performed by topoisomerases that act on duplex DNA ahead of the replication fork (Figure 11-19c). There are many different kinds of topoisomerase. In E. coli, gyrase is the primary replicative topoisomerase, although topoisomerase IV also participates (see Chapter 9).

Primase As discussed earlier, DNA polymerase requires a preformed primer from which to elongate. Cells contain specialized RNA polymerases called primases that synthesize short RNA primers specifically for initiating DNA polymerase action. In E. coli, an RNA primer of 11 to 13 nucleotides is synthesized by the DnaG primase. RNA primers are needed to initiate each of the thousands of Okazaki fragments on the lagging strand. The leading strand is initiated by primase at a replication origin. E. coli DnaG primase must bind the DNA helicase for activity, and this localizes primase action to the replication fork (Figure 11-19d). RNA synthesis is less accurate than DNA synthesis (see Chapter 15), and the use of RNA to prime DNA synthesis provides a way for DNA polymerase I to recognize and remove the less accurate primer before Okazaki fragments are joined together. Note that although DNA polymerases will extend DNA primers as well as an RNA primer, in cells, only RNA is used to initiate DNA synthesis. It is not clear why this is the case. One possibility is that the task of forming the first phosphodiester bond, to create a dinucleotide, is more difficult than forming subsequent bonds, because two NTPs must be held by the enzyme at the same time, and the much higher concentration of rNTPs than dNTPs in the cell may be required for this first step.

Pol I and Ligase RNA primers must be removed at the end of each Okazaki fragment and replaced with DNA. This is achieved through the nick translation activity of Pol I (see Figure 11-7), which removes the ribonucleotides of the primer while simultaneously replacing them with deoxyribonucleotides. A ribonuclease called RNaseH can also remove RNA that is base-paired to DNA, but it cannot remove the last rNMP attached to the DNA. Hence, in E. coli, RNaseH may help remove RNA, but another enzyme (e.g., Pol I) is needed to complete the task.

Once all RNA is replaced with DNA, the nick in the phosphodiester backbone is sealed by DNA ligase in a reaction that requires ATP (or NAD+ in E. coli) (Figure 11-19e; see also Figure 5-12). Ligase acts only on a 5′ terminus of DNA, not on RNA. This specificity ensures that all the RNA at the end of an Okazaki fragment is removed before the nick is sealed. Both ligase and Pol I interact with the β sliding clamp.

384

SSB Single-stranded DNA produced by helicase-catalyzed unwinding is quickly bound by single-stranded DNA–binding protein (SSB), protecting the DNA from endonucleases (see Figure 5-3 for a more detailed look at how SSB binds DNA). SSB stimulates DNA polymerase activity by melting small DNA hairpin structures (i.e., separating base pairs) in the single-stranded template (Figure 11-19f). SSB is found in all cell types and binds DNA in a sequence-independent fashion. E. coli SSB is a homotetramer, but other SSBs range from monomers (e.g., gene 32 protein in T4 phage) to heterotrimers (e.g., RPA in eukaryotes, discussed shortly).

Helicase Activity Is Stimulated by Its Connection to the DNA Polymerase

The E. coli DnaB helicase connects to the Pol III holoenzyme through the τ subunits of the clamp loader within the Pol III holoenzyme. Without this connection to polymerase, DnaB helicase is slow, unwinding about 35 bp/s. On connection of DnaB to the Pol III holoenzyme, unwinding proceeds at a rate of approximately 700 bp/s. The complex of Pol III holoenzyme, DnaB helicase, and primase forms a replisome. The three τ subunits of the τ complex clamp loader bind three Pol III cores, and these same τ subunits also bind the DnaB helicase. Figure 11-21 shows one Pol III core associated with the leading strand and one with the lagging strand; as we will see later, the third Pol III core also participates in replication of the lagging strand. The leading-strand Pol III–β clamp complex moves continuously with DnaB helicase, while the lagging-strand Pol III–β clamp complex repeatedly moves on and off the DNA to extend the multiple RNA primers made by primase.

Figure 11-21: The architecture of the E. coli replisome at a replication fork. DnaB helicase encircles the lagging strand, and the Pol III holoenzyme connects to DnaB via the τ subunits of the clamp loader. The holoenzyme contains three Pol III cores. In the illustration, two Pol III cores connect to β clamps on the leading and lagging strands. The function of the third Pol III is discussed in the text. Primase (DnaG) transiently associates with DnaB for the synthesis of an RNA primer on the lagging strand. The lagging-strand Pol III core–β clamp travels with the replisome, yet extends DNA in the 5′→3′ direction, resulting in a DNA loop. The lagging strand is bound by SSB. The clamp loader is shown bound to a β clamp that it has opened in preparation for loading the clamp onto the RNA primer.

DNA Loops Repeatedly Grow and Collapse on the Lagging Strand

The lagging-strand polymerase must repeatedly extend RNA primers into full-length, 1 to 2 kb Okazaki fragments. As we have seen, however, the direction of chain growth on the lagging strand is opposite to that on the leading strand. How can the lagging-strand Pol III synthesize DNA in the opposite direction to replication fork movement, yet remain tethered to the replisome? To accommodate these opposed directions, the lagging-strand template is pulled up through the polymerase during chain extension to form a loop (Figure 11-22). As an Okazaki fragment is extended, the double-stranded portion of the loop grows longer. As the replication fork generates more single-stranded DNA, it also adds to the growing DNA loop. When the Okazaki fragment is complete, the polymerase bumps into the fragment it made previously and lets go of the DNA loop so that it can extend a new RNA primer for the next Okazaki fragment. The process of repeated loop growth and disassembly is often referred to as the trombone model of replication, because it resembles movement of the slide of a trombone. The trombone model was first proposed in 1980 by Bruce Alberts. At a replication fork speed of about 700 nucleotides per second, and with Okazaki fragments of 1 to 2 kb, a new loop is formed every 2 to 3 seconds. Two of the three Pol III cores of the Pol III holoenzyme can act on the lagging strand, so there are sometimes two loops on the lagging strand.

Figure 11-22: The trombone model of replication fork function. The lagging-strand polymerase extends the 3′ terminus of an Okazaki fragment in the opposite direction to fork movement, yet this polymerase is part of the replisome and thus moves with the fork. The opposed directions result in formation of a DNA loop for each Okazaki fragment. As multiple Okazaki fragments are synthesized, loops repeatedly grow and are released, similar to the movement of a trombone slide as the instrument is played. For simplicity, the participation of the third Pol III core in lagging-strand synthesis is not shown here.

385

When the lagging-strand polymerase finishes an Okazaki fragment, it must dissociate from the DNA in order to transfer to a new RNA primer and extend the next fragment. It does so by detaching from the β clamp, leaving it behind on the DNA. Once released, the lagging-strand Pol III core can associate with a new RNA-primed site on which a new β clamp has been assembled (see Figure 11-22). This process results in a buildup of β clamps on replicated DNA. These leftover clamps perform additional functions, as we will see shortly, but ultimately they, too, must be removed and recycled.

Bruce Alberts

Recent studies have shown that two of the three Pol III cores of the Pol III holoenzyme can function to replicate the lagging strand while the third continuously replicates the leading strand. These actions are illustrated in Figure 11-23. First, two polymerases are shown, one on each strand (step 1). As the Okazaki fragment is being extended by the first lagging-strand Pol III core, the primase synthesizes an RNA primer for a second Okazaki fragment (step 2), and the clamp loader assembles a clamp onto the new primer (step 3). Then the second lagging-strand polymerase can associate with the new clamp and extend a second Okazaki fragment while the first lagging-strand polymerase is still active, forming two lagging-strand DNA loops (step 4). On completing the first Okazaki fragment, the first lagging-strand Pol III core dissociates from its clamp and is now ready for another cycle of Okazaki fragment synthesis, while the second lagging-strand Pol III core continues to extend the second Okazaki fragment (step 5).

Figure 11-23: Function of three polymerases during the Okazaki fragment cycle. Two Pol III cores are attached to β clamps on the leading and lagging strands, creating a loop. Primase binds to DnaB helicase and synthesizes an RNA primer (purple). The clamp loader assembles a β clamp onto the new RNA primer. The second lagging-strand Pol III core (third polymerase) assembles with the new clamp while the first lagging-strand Pol III core is extending an Okazaki fragment, creating two DNA loops. The first lagging-strand Pol III core ejects from the β clamp on the first Okazaki fragment, leaving the clamp on the DNA.

Experiments using a simple model system first suggested that Pol III hops from one β clamp to another to cycle among Okazaki fragments on the lagging strand (Figure 11-24). The Pol III holoenzyme was assembled on a 5.4 kb DNA substrate (M13mp18), then mixed with two competing DNA substrates of different sizes (M13Gori and φX174), each with one site primed for DNA synthesis. In each of two experiments, only one of the competing DNAs included a β clamp at the primed site. Replication was initiated using [32P]dNTPs, and timed aliquots were analyzed by agarose gel electrophoresis and autoradiography. As the result in Figure 11-24b shows, Pol III replicated the initial 5.4 kb DNA, then transferred to the DNA substrate that included the preassembled clamp. Because the competing DNA with a preassembled β clamp was preferentially replicated compared with the DNA lacking a clamp, the result suggests that Pol III leaves the β clamp behind on the first template, then hops to another DNA that has a new β clamp. Further studies confirmed that Pol III indeed hops from one clamp to another, leaving clamps behind on the DNA as it does so.

Figure 11-24: Transfer of Pol III from an old β clamp to a new β clamp. In these experiments, the Pol III holoenzyme (just the β clamp and Pol III core are shown for simplicity) is assembled onto a primed donor DNA circle (M13mp18), then the Pol III–DNA is mixed with two competing primed acceptor DNA circles of different sizes, only one of which includes a β clamp. In the experiment shown on the left side, the M13Gori DNA has a β clamp and the other (φX174) DNA does not. In the experiment on the right, φX174 DNA has a β clamp and the other (M13Gori) does not. Replication is initiated and timed. Aliquots of the reaction mixtures are analyzed on an agarose gel. The acceptor DNA with a β clamp is replicated in preference to the acceptor DNA without a β clamp.

386

Okazaki Fragments Require Removal of RNA and Ligase-Mediated Joining of DNA

The RNA at the 5′ terminus of each Okazaki fragment must be removed and the gap filled in with DNA. This job is performed by the nick translation activity of Pol I (see Figure 11-7). As noted earlier, the cell also has a backup enzyme, RNaseH, which can remove the RNA, in which case the single-strand gap must be filled in by a DNA polymerase. Processed fragments are then joined together by ligase to form a continuous duplex. Both Pol I and ligase interact with the β clamp, and the β clamps left behind by the replisome are thought to attract Pol I and ligase for the removal of RNA and sealing of the fragments.

387

Okazaki fragments outnumber β clamps in the cell by about 10 to 1, so clamps must be recycled during chromosome replication. Clamps are removed from DNA by the δ subunit of the clamp loader, which is produced in excess in the cell (relative to the other clamp loader subunits). By itself, the δ subunit can open and unload a β clamp from DNA, but it cannot assemble a clamp onto DNA.

The β clamp binds many proteins, including all five E. coli DNA polymerases, ligase, the clamp loader complex, and several proteins not described in this chapter. These proteins all bind to the same spot on β. Therefore, when the β clamp is being used by a DNA polymerase or another protein, the recycling of β is blocked. Only when the clamp is no longer bound by other proteins is it available to be recycled (Figure 11-25).

Figure 11-25: β clamp recycling in E. coli.

The Replication Fork Is More Complex in Eukaryotes Than in Bacteria

Many eukaryotic replication proteins have counterparts in bacteria (e.g., the clamps and clamp loader), but the replication fork machinery of eukaryotes includes more proteins beyond those used in the comparatively simple bacterial machinery (Figure 11-26). New eukaryotic replication factors are still being identified, and details of the replication fork in eukaryotic cells are only now coming into focus.

Figure 11-26: A hypothetical model of the eukaryotic replication fork. Eukaryotes contain all the proteins that function in a bacterial replisome, but most components have more subunits than the bacterial proteins, and several additional proteins function at the eukaryotic replication fork. The components and their functions are described in the text.

The MCM complex is thought to function at the replication fork. Like E. coli DnaB, the MCM subunits form a ring-shaped hexamer, but each subunit is different (named Mcm2 through Mcm7, or Mcm2–7). The six subunits are homologous AAA+ proteins (each subunit Mr ∼ 100,000). The Mcm2–7 complex associates with a heterotetramer called GINS and with the Cdc45 protein, forming a complex referred to as the CMG complex (for Cdc45-MCM-GINS). The CMG complex is an active helicase, while the MCM complex alone has only feeble helicase activity. The mechanism by which GINS and Cdc45 activate the CMG helicase is not yet understood, and these proteins have no sequence-related homologs in bacteria. The direction of activity of the CMG helicase is 3′→5′, opposite to that of E. coli DnaB, and therefore the CMG helicase must act on the leading strand to unwind the parental DNA. Some of the eukaryotic replication proteins, including the CMG helicase, are targets of cell cycle kinases (cyclin-dependent protein kinases), enzymes that phosphorylate specific proteins and are active at certain phases of the cell cycle. For example, phosphorylation of a replication protein may activate it, and this modification may occur only on entering S phase of the cell cycle. Even though most eukaryotic chromosomes are linear, they are still topologically constrained, and helicase unwinding creates torsional stress that is relieved by topoisomerases.

388

The eukaryotic primase is a four-subunit complex called DNA polymerase α (Pol α). Priming activity is located in the smallest subunit, and it makes an RNA primer of about a dozen nucleotides. The largest subunit of Pol α is a DNA polymerase that extends the RNA primer with DNA, to a total length of 25 to 40 nucleotides. As in bacteria, the RNA is excised before Okazaki fragments are joined. The DNA made by Pol α may contain errors, because the enzyme has no 3′→5′ proofreading exonuclease, and it is thought that the DNA made by Pol α is also replaced or is corrected by repair proteins.

Eukaryotes have two different chromosomal replicases: DNA polymerase δ (Pol δ) and DNA polymerase ε (Pol ε). Both Pol δ and Pol ε are four-subunit enzymes in higher eukaryotes, and the largest subunit of each has both a DNA polymerase and a 3′→5′ exonuclease activity (Pol δ in yeast has only three subunits). Current research suggests that Pol δ and Pol ε operate on different strands at the replication fork: Pol ε on the leading strand and Pol δ on the lagging strand.

389

Both Pol δ and Pol ε interact with a DNA sliding clamp called PCNA (proliferating cell nuclear antigen). PCNA looks remarkably like the E. coli β clamp (Figure 11-27a). The two proteins share no sequence homology, but the three-dimensional structures are nearly superposable. Both proteins are constructed from a domain that is repeated six times around the ring. The three monomer units of PCNA have only two domains and trimerize to form a ring, while the two monomer units of the β clamp consist of three domains and dimerize to form the ring. The eukaryotic clamp loader, replication factor C (RFC), has five subunits similar in shape and function to those of the E. coli clamp loader (Figure 11-27b). In a fascinating twist, eukaryotes contain alternative forms of RFC in which one of the subunits is replaced by another protein. These alternative clamp loaders usually function with PCNA, and their intracellular role is not entirely clear. In one case, the alternative clamp loader loads an entirely different clamp onto the DNA.

Figure 11-27: The eukaryotic PCNA clamp and RFC clamp loader. (a) PCNA is a homotrimer; the monomer units are shown in different colors. (b) The RFC clamp loader is homologous to the bacterial clamp loader. Compare this with Figure 11-18.

The eukaryotic replication machinery also includes several other proteins associated with the CMG complex. The complex was isolated from cells using highly selective antibodies directed against CMG subunits. The large protein assemblage is referred to as the replisome progression complex (RPC), and its subunit composition has been identified by mass spectrometry. Among the proteins identified in this fashion are Ctf4 and Mcm10, proteins that move with replication forks in vivo and are essential for cell viability (except in yeast, where Ctf4 is not essential). Ctf4 binds to both GINS and Pol α, acting as a bridge to affix Pol α in the replisome. The role of Mcm10 is presently unknown. RPCs also contain several nonessential proteins that are thought to control the rate of replication during times of cellular stress.

Eukaryotic replication forks proceed at a rate of about 30 to 50 nucleotides per second, far slower than bacterial forks. Also, eukaryotic Okazaki fragments are considerably shorter than bacterial Okazaki fragments, only 100 to 200 nucleotides. The heterotrimeric replication protein A (RPA) is the functional equivalent of E. coli SSB.

Additional proteins of the eukaryotic replication machinery have functions in finishing and sealing Okazaki fragments. On completing an Okazaki fragment, Pol δ performs limited strand displacement, lifting the RNA primer synthesized by Pol α. The RNA is then excised by a 5′→3′ nuclease called Fen1, and DNA ligase I then joins the fragments. This process often removes some of the DNA made by Pol α. An alternative pathway for RNA primer removal comes into play when strand displacement has proceeded farther than normal; this is accomplished by the Dna2 nuclease, which can excise longer tracts of displaced RNA and DNA than Fen1.

The identity and arrangement of proteins that function at the replication fork in eukaryotes are still the subject of intense investigation. Several proteins with functions that, as yet, lack clear definition seem to be involved in the architecture of the eukaryotic replication fork. The numbers and types of proteins currently thought to participate in eukaryotic chromosome replication are summarized in Table 11-4.

390

SECTION 11.3 SUMMARY

  • The E. coli chromosomal replicase, the Pol III core, connects to the ring-shaped β sliding clamp that encircles DNA for processive DNA synthesis. The β clamp is assembled onto DNA by a multiprotein clamp loader. Three Pol III cores, three β clamps, and one clamp loader complex form the Pol III holoenzyme assembly.

  • The Pol III holoenzyme, DnaB helicase, and DnaG primase form the replisome complex. The hexameric DnaB helicase encircles the lagging strand and uses ATP to unwind DNA at the replication fork. DnaG primase forms RNA primers to initiate DNA synthesis.

  • Topoisomerases act ahead of the replication fork to remove superhelical tension generated by DNA unwinding. SSB binds the single-stranded DNA created by the unwinding action of helicase, preventing the formation of secondary structures in the DNA and protecting it from endonucleases.

  • RNA primers are removed from finished Okazaki fragments by the nick translation action of Pol I, and the processed fragments are joined by DNA ligase.

  • Simultaneous replication of the two antiparallel strands of duplex DNA by two Pol III cores in the replisome requires loops to form on the lagging strand that repeatedly grow and reset for each Okazaki fragment.

    391

  • Eukaryotes have two different multiprotein DNA polymerases (Pol ε and Pol δ) that function on the leading and lagging strands. These DNA polymerases connect to PCNA sliding clamps that are loaded onto the DNA by the RFC clamp loader.

  • Eukaryotes have functional counterparts for each E. coli replication fork protein, but the eukaryotic replisome is more complex. The eukaryotic primase is a four-subunit enzyme (Pol α) that has both DNA polymerase and primase activities. The CMG helicase has 11 subunits: an Mcm2–7 heterohexamer, GINS heterotetramer, and Cdc45. The eukaryotic SSB homolog, RPA, has three different subunits. Several other eukaryotic proteins that travel with the fork have no known homologs in bacteria.