Structural basis of human PCNA sliding on DNA

Abstract

Sliding clamps encircle DNA and tether polymerases and other factors to the genomic template. However, the molecular mechanism of clamp sliding on DNA is unknown. Using crystallography, NMR and molecular dynamics simulations, here we show that the human clamp PCNA recognizes DNA through a double patch of basic residues within the ring channel, arranged in a right-hand spiral that matches the pitch of B-DNA. We propose that PCNA slides by tracking the DNA backbone via a 'cogwheel' mechanism based on short-lived polar interactions, which keep the orientation of the clamp invariant relative to DNA. Mutation of residues at the PCNA-DNA interface has been shown to impair the initiation of DNA synthesis by polymerase δ (pol δ). Therefore, our findings suggest that a clamp correctly oriented on DNA is necessary for the assembly of a replication-competent PCNA-pol δ holoenzyme.

Figure 1. Structural basis of DNA recognition by human PCNA.

(a) 2.8Å crystal structure of PCNA bound to a 10 bp DNA duplex. PCNA and DNA are shown in surface and ribbon representation, respectively. PCNA subunits are coloured in different shades of blue and DNA in orange. The expansion shows the complex in ribbon representation. Interacting PCNA side chains and DNA phosphates (interatomic side chain nitrogen—DNA phosphorus distance <5 Å) are shown as sticks and yellow spheres, respectively, and interactions as dashed lines (b) NMR analysis. Left: front- and back-face views of PCNA surface. PCNA residues whose amide chemical shifts are significantly perturbed by DNA are coloured red. The crystallographic position of DNA is shown in orange. The interacting region in the clamp channel overlaps with that seen in the crystal structure, whereas in the crystal, the side chains can be discriminated, in solution the perturbations involve the backbone amides. Right: chemical shift perturbation of the amide signal of PCNA residue T73 at different DNA concentrations. Fitting was performed using a single-site binding model. Extrapolated dissociation and exchange time constants are indicated. (c) Model interface from MD simulation. The crystallographic position of the DNA segment is shown in orange, whereas in black the DNA is shown in a position corresponding to the final state of the 100 ns MD simulation of the complex.

Figure 2. MD simulation of human PCNA bound to a 30 bp DNA.

(a) PCNA ring rotation and evolution of DNA tilting relative to the ring C3 axis (b) Time evolution of contacts between the side chain nitrogens of three representative PCNA interfacial residues and DNA phosphorus atoms at consecutive positions (c) Two views of the PCNA–DNA complex at the end of the MD trajectory. PCNA is shown as a grey surface and DNA as a black ribbon. PCNA residues whose side chains are engaged in polar contacts with DNA phosphates for >25% of the MD trajectory are labelled. Residues from different PCNA subunits are coloured in shades of blue. Residues that exchange between two or three consecutive DNA phosphates for >75% of the MD simulation are indicated by an asterisk and boxed.

Figure 3. Proposed ‘cogwheel' mechanism for PCNA sliding.

(a) Interacting side chains are able to rapidly switch between adjacent phosphates in a non-coordinated manner (illustrated by the thin and thick lines). When this stochastic process generates a state in which a sufficient number of electrostatic contacts are simultaneously established with adjacent phosphates in one direction of the DNA helical axis, a net rotation of the protein occurs and results in the advancement of one base pair (b) PCNA bi-directional tracking of the DNA backbone, which agrees with the clamp helical sliding mode inferred from diffusion data measured by single-molecule imaging. The diffusion coefficient of PCNA (1.16 μm² s⁻¹) implies that, on average, PCNA diffuses 8 bp per microsecond.

Figure 4. Possible role of PCNA–DNA interaction and sliding mechanism in lagging strand replication by pol δ.

PCNA is loaded on the P/T junction of a nascent Okazaki fragment by the clamp loader RFC. The RNA–DNA hybrid primer of the fragment (red) is blocked on both sides by Replication Protein A (RPA), which binds the template strand (black). The interaction of PCNA with DNA imparts an orientation to PCNA that is competent for pol δ binding and initiation of DNA synthesis. However, PCNA can slide off the P/T junction before binding pol δ or upon premature dissociation of pol δ from the template. The cogwheel sliding mechanism ensures that pol δ captures PCNA at the P/T junction of the fragment with the correct orientation to resume synthesis.