How the Eukaryotic Replisome Achieves Rapid and Efficient DNA Replication

Abstract

The eukaryotic replisome is a molecular machine that coordinates the Cdc45-MCM-GINS (CMG) replicative DNA helicase with DNA polymerases α, δ, and ε and other proteins to copy the leading- and lagging-strand templates at rates between 1 and 2 kb min^-1. We have now reconstituted this sophisticated machine with purified proteins, beginning with regulated CMG assembly and activation. We show that replisome-associated factors Mrc1 and Csm3/Tof1 are crucial for in vivo rates of replisome progression. Additionally, maximal rates only occur when DNA polymerase ε catalyzes leading-strand synthesis together with its processivity factor PCNA. DNA polymerase δ can support leading-strand synthesis, but at slower rates. DNA polymerase δ is required for lagging-strand synthesis, but surprisingly also plays a role in establishing leading-strand synthesis, before DNA polymerase ε engagement. We propose that switching between these DNA polymerases also contributes to leading-strand synthesis under conditions of replicative stress.

Graphical abstract

Figure 1

Replication Reactions on Soluble Plasmid Templates (A) Reaction scheme for soluble replication reactions is shown. Firing factors: Sld3/7, Sld2, Dpb11, S-CDK, GINS, Cdc45, Pol ε, Mcm10. Replication proteins: Topo II, Pol α, RPA, Ctf4. (B) Time course of a reaction performed as in (A) is shown. (C) Quantitation of leading- and lagging-strand products in (B) is shown. (D) Coomassie-stained SDS-PAGE of proteins involved in lagging-strand replication is shown. (E) Replication performed as in (B) is shown. In this and all subsequent figures, the protein constituents of the reactions are listed above each figure. Min replisome encompasses the minimum set of proteins required for origin firing together with Ctf4 and Topo II (see A for details).

Figure 2

Reconstitution of In Vivo Replication Rates with Purified Proteins (A) Coomassie-stained SDS-PAGE of RPC components is shown. (B) Replication time course conducted as in Figures 1A and 1B but including RFC, PCNA, and the additional RPC components shown in (A). The additional proteins were added together with the firing factors and replication proteins. (C) Pulse-chase experiment to measure replication rates in the presence of RPC components. FACT and Topo II were omitted. The chase was added at 2 min 20 s. (D) Maximum (front) and peak product lengths plotted against time for pulse-chase experiments performed as in (C). Error bars represent the SEM from two experiments. Data were fit to a linear regression to derive the maximum and bulk leading-strand synthesis rates.

Figure 3

Csm3/Tof1 and Mrc1 Are Required for Maximum Rates (A) Replication reactions performed with the proteins illustrated. In lane 1, All refers to Csm3/Tof1, Mrc1, FACT, and Topo I. (B and C) Reactions were performed as in (A) except that FACT was omitted. (A) and (B) were incubated for 15 min. The potassium glutamate concentrations in (C) were 100, 150, 200, and 250 mM.

Figure 4

PCNA Has a Major Role in Pol ε-Catalyzed Leading-Strand Synthesis (A) Time course reaction performed using the same experimental conditions as Figure 3C, lane 8. PCNA was omitted where indicated. (B) Pulse-chase experiment performed with the same compliment of proteins as in (A) except that RFC, PCNA, and Topo II were omitted. The chase was added at 3 min 50 s. (C) Maximum product lengths plotted against time for the pulse-chase experiment in (B). Data were fit to a linear regression to derive the maximum leading-strand synthesis rate. (D) Experiment performed as in (A) for 8 min. The dNTP concentrations are the concentrations of the individual dNTPs in the reaction.

Figure 5

The Effect of Pol δ on Replication (A) A 20-min reaction performed with the same set of proteins as Figure 4A with PCNA. The Pol α concentration was 40 nM and the Pol δ concentration was 10 nM. (B) Lane profiles of the data in (A) are shown. (C) Quantitation of leading- and lagging-strand replication products for experiments performed as in (A). Data were normalized to the sum of leading and lagging strands in the reaction containing Pol δ. Error bars represent the SEM from two experiments. (D) Reaction performed as in (A) with 10 nM Pol δ for 20 min. Pol α concentrations were 5, 10, 20, 40, and 80 nM. (E and F) Pulse-chase experiment (E) was performed and analyzed (F) as in Figures 2C and 2D but with the inclusion of 10 nM Pol δ. (G) Normalized product-length distribution for the leading-strand products in (A). To account for the continuous incorporation of radiolabel, product intensities were divided by product lengths.

Figure 6

Leading-Strand Synthesis Catalyzed by Pol δ Is Slow (A and B) The roles of Pol δ (A and B) and Mrc1 (B) in replication with Pol ε-Δcat. Reactions performed for 15 min with the same set of proteins as in Figure 4A with PCNA. Where indicated, wild-type Pol ε was substituted with Pol ε-Δcat. (C) Quantitation of a pulse-chase reaction performed as in Figure 5E except that Pol ε was substituted with Pol ε-Δcat Figure S6C. The chase was added at 2 min 50 s. (D) Primer extension reaction with Pol δ. The primed template was incubated with PCNA and RFC for 5 min before reactions were initiated by the addition of Pol δ. (E) Quantitation of the data in (D) plotting the peak of the product distributions. Data were fit to a linear regression. (F) Pulse-chase reactions were performed as in Figure 2C but with varying concentrations of Pol δ added immediately after the 2-min 30-s time point.

Figure 7

Pol δ Promotes the Establishment of Leading-Strand Synthesis (A) Pulse-chase reactions conducted as in Figure 2C with 10 nM Pol δ included where indicated. When Pol δ was included in the chase, it was added immediately after the 2-min 30-s time point was removed. (B) Model for eukaryotic leading-strand synthesis. (i) Following helicase activation the replisome advances slowly, unwinding the template to generate a priming site on the leading strand for Pol α. (ii) Following priming, RFC assembles PCNA around the primer terminus. Pol δ rapidly binds to the primer and commences elongation. The elongation rate of Pol δ is considerably faster than the advancing replisome, so Pol δ quickly catches up with the replication fork. (iii) Once Pol δ has made contact with the replisome, the rate of synthesis is limited by the template-unwinding rate of the replisome. (iv) A polymerase switch transfers the 3′ end of the leading strand together with PCNA from Pol δ to Pol ε. Pol ε-dependent leading-strand synthesis stimulates the template-unwinding rate of the replisome, and DNA synthesis rates of ∼2 kb min⁻¹ are established. (v) In the absence of Pol δ, Pol ε can take over leading-strand synthesis directly from Pol α, although this process is less efficient than the pathway involving Pol δ.