Fast and efficient DNA replication with purified human proteins

Abstract

Chromosome replication is performed by a complex and intricate ensemble of proteins termed the replisome, where the DNA polymerases Polδ and Polε, DNA polymerase α-primase (Polα) and accessory proteins including AND-1, CLASPIN and TIMELESS-TIPIN (respectively known as Ctf4, Mrc1 and Tof1-Csm3 in Saccharomyces cerevisiae) are organized around the CDC45-MCM-GINS (CMG) replicative helicase^1-7. Because a functional human replisome has not been reconstituted from purified proteins, how these factors contribute to human DNA replication and whether additional proteins are required for optimal DNA synthesis are poorly understood. Here we report the biochemical reconstitution of human replisomes that perform fast and efficient DNA replication using 11 purified human replication factors made from 43 polypeptides. Polε, but not Polδ, is crucial for optimal leading-strand synthesis. Unexpectedly, Polε-mediated leading-strand replication is highly dependent on the sliding-clamp processivity factor PCNA and the alternative clamp loader complex CTF18-RFC. We show how CLASPIN and TIMELESS-TIPIN contribute to replisome progression and demonstrate that, in contrast to the budding yeast replisome⁸, AND-1 directly augments leading-strand replication. Moreover, although AND-1 binds to Polα^9,10, the interaction is dispensable for lagging-strand replication, indicating that Polα is functionally recruited via an AND-1-independent mechanism for priming in the human replisome. Collectively, our work reveals how the human replisome achieves fast and efficient leading-strand and lagging-strand DNA replication, and provides a powerful system for future studies of the human replisome and its interactions with other DNA metabolic processes.

Extended Data Fig. 1. Purified human DNA replication proteins.

Coomassie stained SDS-PAGE of human DNA replication proteins. Individual lanes from the gel in Fig. 1a are shown with each subunit labelled.

Extended Data Fig. 2. Leading-strand synthesis.

a, Standard replication reaction on the 9.7 kbp template performed with the indicated proteins and analysed by native and denaturing agarose gel electrophoresis as indicated. In the absence of RFC and PCNA the predominant replication products (intermediates) migrate above the position of full length in the native gel. As indicated, in the native gel template labelling products and complete full-length replication products migrate in the same position. b, Denaturing agarose gel analysis of a pulse chase experiment on the 9.7 kbp template with the indicated proteins. Unless otherwise stated, in this and all pulse chase experiments, the chase was added at 50 s. c, Denaturing agarose gel analysis of a replication reaction on the 9.7 kbp template with the indicated proteins. d, Denaturing agarose gel analysis of a pulse chase experiment on the 9.7 kbp template with the indicated proteins.

Extended Data Fig. 3. CTF18-RFC is required for optimal leading-strand synthesis.

a, Denaturing agarose gel analysis of a replication reaction on the 9.7 kbp template with the indicated proteins. b, Denaturing agarose gel analysis of a pulse chase experiment on the 9.7 kbp template with the indicated proteins. c, Silver-stained SDS-PAGE analysis of glycerol gradients performed with the indicated proteins. For clarity, only CMG, CTF18-RFC and Polε subunits are annotated. d, Coomassie stained SDS-PAGE of CTF18-RFC and Pol ε interaction mutants. e, Silver-stained SDS-PAGE analysis of a pull-down experiment with the indicated proteins showing that mutation of CTF18 and POLE1 disrupt the interaction between the two proteins. f, Denaturing agarose gel analysis of a replication reaction performed for 3 min on the 9.7 kbp template with the indicated proteins.

Extended Data Fig. 4. CTF18-RFC accelerates established replication forks.

a, Lane profiles of the 165 s timepoints in Fig. 3e where CTF18-RFC was absent or added in the chase (lanes 2 and 8 respectively). b, Denaturing agarose gel analysis of a pulse chase experiment on the 15.8 kbp template with the indicated proteins. Where indicated CTF18-RFC or CTF18-RFC^RAA were added with the chase. c, Lane profiles of the 240 s timepoints in (b). For lane profiles (a, c), product intensities were normalised by dividing each value by the relative intensity of the total signal in a given lane.

Extended Data Fig. 5. PCNA loading by CTF18-RFC is required for optimal leading-strand synthesis.

a, Coomassie stained SDS-PAGE of CTF18-1-8 module complexes. b, Silver-stained SDS-PAGE analysis of a pull-down experiment with the indicated proteins showing that the CTF18-1-8 module interacts specifically with Pol ε. c, Coomassie stained SDS-PAGE of WT and CTF18^K380E complexes. d, Silver-stained SDS-PAGE analysis of a pull-down experiment with the indicated proteins showing that CTF18^K380E-RFC retains the capacity to interact with Pol ε. e, f, Primer extension reactions on M13mp18 single-strand DNA with Pol δ and Pol ε showing that CTF18^K380E-RFC has a severe defect in supporting PCNA-dependent DNA synthesis by both polymerases. g, (left) Denaturing agarose gel analysis of a pulse-chase experiment on the 15.8 kbp template with the indicated proteins. The chase was added at 1 min 45 s. Where indicated CTF18-RFC or CTF18^K380E-RFC were added with the chase. (right) Lane profiles for the 5 min timepoint. Data were normalised by dividing each intensity value by the relative total signal at the 2 min timepoint.

Extended Data Fig. 6. TIM-TIPIN, CLASPIN and AND-1 enhance leading-strand replication.

a, Denaturing agarose gel analysis of a time course experiment on the 15.8 kbp template with the indicated proteins at two concentrations of potassium glutamate (K-Glu). b, c, Lane profiles of the data in Fig. 4a, b respectively. d, Denaturing agarose gel analysis (top) and lane profiles (bottom) of a 3 min 45 s replication reaction on the 15.8 kbp template with the indicated proteins. e, Lane profiles of the data in Fig. 4c. TT, TIM-TIPIN. f, Denaturing agarose gel analysis (top) and lane profile (bottom) of a 3 min 45 s replication reaction on the 15.8 kbp template with the indicated proteins. TT, TIM-TIPIN. g, Denaturing agarose gel analysis of a 3.5 min replication reaction on the 15.8 kbp template with the indicated proteins. In d, f, g, the potassium glutamate concentration was 250 mM. For lane profiles (b–f), product intensities were normalised by dividing each value by the relative intensity of the total signal in a given lane.

Extended Data Fig. 7. CLASPIN and AND-1 truncations.

a, Cartoon representation of the core human replisome (PDB:7PFO) showing the region of CLASPIN (E284–K319) that interacts with the TIM α-solenoid. b, Coomassie stained SDS-PAGE of CLASPIN truncation mutants. c, d Lane profiles of the data in Fig. 4e. e, Coomassie stained SDS-PAGE of AND-1 truncation mutants. f, Lane profiles of the data in Fig. 4g. For lane profiles (c, d, f), product intensities were normalised by dividing each value by the relative intensity of the total signal in a given lane.

Extended Data Fig. 8. Reconstitution of lagging-strand replication.

a, Denaturing agarose gel analysis of a 20 min replication on the 9.7 kbp template with the indicated proteins. b, Lane profiles from lanes 3 and 4 in (a). c, Schematic showing the possible replication products (–/+ Pol δ) if lagging strands are extended by Pol δ and are constituents of both replication intermediates and full-length products. d, e, Two-dimensional agarose gel analysis of 20 min replication reactions performed with the indicated proteins on the 9.7 kbp template in the absence (d) and presence (e) of Pol δ. In all reactions, the concentration of potassium glutamate was 250 mM.

Extended Data Fig. 9. Lagging-strand replication occurs at all replication forks.

a, Schematic of a replication reaction on the cyclobutane pyrimidine dimer (CPD) template. b, Native and denaturing gel analysis of a time course experiment on undamaged and CPD templates with the indicated proteins. c, Native and denaturing gel analysis of a 60 min reaction on the CPD template with different concentrations of Pol α as indicated. d, e, Two-dimensional agarose gel analysis of 30 min replication reactions performed with the indicated proteins on the CPD template in the absence (d) and presence (e) of Pol α. In all reactions, the concentration of potassium glutamate was 250 mM.

Extended Data Fig. 10. Role of AND-1 in lagging-strand replication.

a, Lane profiles from Fig. 5d, lanes 5 and 6. b, Lane profiles from the experiment in Fig. 5e, lanes 2, 3, 5, 7, 8 and 10. c, Denaturing agarose gel analysis of a 30 min reaction on the 9.7 kbp template with the indicated proteins. d, Silver-stained SDS-PAGE analysis of glycerol gradients performed with the indicated proteins demonstrating complex formation between Pol α and CMG in the absence of replication fork DNA.

Fig. 1. PCNA is required for efficient leading-strand synthesis by CMG-Polε.

a, Coomassie-stained SDS–PAGE analysis of purified human DNA replication proteins. b, Schematic illustrating the reaction scheme for CMG-based DNA replication assays using forked DNA templates. The CMG complex is coloured grey and Polε is coloured purple. c, Denaturing agarose gel analysis of a time course experiment performed as in b in the absence of RFC and PCNA on a 9.7-kbp forked DNA template. The CMG-independent primer and template labelling products are indicated. d, Experiment performed as in c with RFC and PCNA included where indicated. e, Quantification of the maximal leading-strand synthesis rate from pulse-chase experiments (Extended Data Fig. 2b). Linear regression is fit to the mean of three experiments. The error bars represent the s.e.m., the mean is indicated by filled circles and individual data points are represented as crosses. In all figures, the proteins present in each reaction are shown above the gel or graph. For gel source data in this and all subsequent figures, see Supplementary Fig. 1.

Fig. 2. Reconstitution of cellular DNA replication rates with purified proteins.

a, Replication reaction performed as in Fig. 1b, c but including RFC and PCNA. TIM–TIPIN, CLASPIN and AND-1 were included where indicated. The additional proteins were added after the CMG-binding step (Fig. 1b). b, Replication reaction performed in the presence of TIM–TIPIN, CLASPIN and AND-1, with CTF18–RFC added where indicated. c, Quantification of the maximal leading-strand synthesis rate from pulse-chase experiments (Extended Data Fig. 2d). Linear regression is fit to the mean of three experiments. The error bars represent the s.e.m., the mean is indicated by filled circles and the individual data points are represented as crosses.

Fig. 3. PCNA loading by CTF18–RFC coupled to Polε is required for maximal replication rates.

a, Replication reaction performed on the 9.7-kbp template for 3 min 30 s with PCNA, RFC and CTF18–RFC added or omitted where indicated. b, Quantification of the maximal leading-strand synthesis rate from pulse-chase experiments (Extended Data Fig. 3b). Linear regression is fit to the mean of three experiments. The error bars represent the s.e.m., the mean is indicated by filled circles and the individual data points are represented as crosses. c, Silver-stained SDS-PAGE of a peak fraction from a glycerol sedimentation gradient performed with the indicated proteins (Extended Data Fig. 3c; +CMG, fraction 13). d, Denaturing agarose gel of a 3-min replication reaction comparing Polε and CTF18–RFC interaction mutants. WT, wild type. e, Pulse-chase experiment on the 15.8-kbp template where CTF18–RFC was either added in the pulse or with the chase. Unless stated otherwise, in this and all pulse-chase experiments, the chase was added after 50 s. f, Denaturing agarose gel analysis of a replication reaction performed for 3 min 40 s on the 15.8-kbp template with the indicated proteins.

Fig. 4. How TIM–TIPIN, CLASPIN and AND-1 facilitate leading-strand replication.

a-c, Denaturing agarose gel analysis of replication reactions performed on the 15.8-kbp forked DNA template for 3 min 30 s with the indicated proteins. d, Diagrams of the primary structure for full-length CLASPIN and CLASPIN truncation mutants. Regions of CLASPIN that bind to MCM and TIMELESS are indicated. e, Denaturing agarose gel analysis of a replication reaction performed on the 15.8-kbp forked DNA template for 3 min 40 s with the indicated proteins. f, Diagrams of the primary structure for full-length AND-1 and AND-1 truncation mutants. g, Denaturing agarose gel analysis of a replication reaction performed on the 15.8-kbp forked DNA template for 3 min 40 s with the indicated proteins. In all reactions, the concentration of potassium glutamate was 250 mM.

Fig. 5. Reconstitution of lagging-strand DNA replication with purified human proteins.

a, Schematic illustrating the expected DNA replication products in reactions with or without the primase Polα. b, Time course experiment on the 9.7-kbp template with or without Polα as indicated. Products were analysed by native and denaturing agarose gel electrophoresis. c-e, Denaturing agarose gel analysis of replication reactions performed on the 9.7-kbp template for 20 min with the indicated proteins. In all reactions, the concentration of potassium glutamate was 250 mM. f, Silver-stained SDS–PAGE of a peak fraction from a glycerol sedimentation gradient performed with the indicated proteins (Extended Data Fig. 10d; CMG + Polα, fraction 10).