Origin DNA melting and unwinding in DNA replication

Abstract

Genomic DNA replication is a necessary step in the life cycles of all organisms. To initiate DNA replication, the double-stranded DNA (dsDNA) at the origin of replication must be separated or melted; this melted region is propagated and a mature replication fork is formed. To accomplish origin recognition, initial DNA melting, and the eventual formation of a replication fork, coordinated activity of initiators, helicases, and other cellular factors are required. In this review, we focus on recent advances in the structural and biochemical studies of the initiators and the replicative helicases in multiple replication systems, with emphasis on the systems in archaeal and eukaryotic cells. These studies have yielded insights into the plausible mechanisms of the early stages of DNA replication.

Figure 1

Structures of initiators and helicases in eukaryotic and archaeal systems. (A) Structure of S. solfataricus Cdc6-ORC1 binding the origin DNA (in orange), bending the DNA (for comparison, B-form dsDNA is in cyan) [6]. (B) Structure of A. pernix ORC1 binding the origin DNA (in orange) through two separate domains, bending the origin DNA (B-form dsDNA is drawn in cyan) [7]. The initiator binding does not induce origin melting. (C) Side view of the apo LTag hexamer, showing the β-hairpins are on the same plane (planar arrangement, indicated by the horizontal line). To show the planar β-hairpins clearly, the subunits in the front and back of the hexamer are removed. (D) Superposition of LTag structures in ATP-bound, ADP-bound, and nucleotide-free states. Major structural shifts of the β-hairpins (β-hp, colored as indicated in the figure) in the C-to-N direction occur upon ATP binding, hydrolysis, and release. The channel diameter of these hexamers vary between ~13–17Å, too small for dsDNA to pass through. (E) The staircase-arranged six β-hairpins (hp1-6) in a E1 hexamer, showing the right-handed helical path along the central channel from C-to-N direction (indicated by a vertical bar), along which the six-nucleotide ssDNA (drawn as sticks in yellow/orange) is anchored [12]. The structure shows that the 5’ end of the ssDNA is on the N-terminal side (upper), and the 3’ on the C-terminal side (lower), which is in the opposite orientation from the model drawn in Fig 2F. (F) Alignments of E1 β-hairpins (hp1-6) in ssDNA-bound state [12] (in green) and DNA-free state [13] (in orange), showing the same staircase arrangement for the β-hairpins in the presence and absence of ssDNA and nucleotide. This same staircase arrangement was seen in the DNA-free and nucleotide-free E1 structure, suggesting that the ssDNA adopts a conformation to fit the staircase-arranged β-hairpins.

Figure 2

Models for the conversion from a melted origin to a replication fork proposed for LTag, the” squeeze-pumping” model shown in (A-D), and for E1, the “trimer assembly” model in (E-F). (A) The first step in the squeeze-pumping model, the initial melting of the origin in the AAA+ domain of a LTag hexamer, which probably is generated by the squeeze/crush of the narrow channel on the dsDNA (or the squeeze-to-open mechanism). The squeeze or crush on the dsDNA can lead to base-pair disruption and thus origin melting (shown in red bars). (B) Another round of ATP binding/hydrolysis in the AAA+ motor domain pumps the melted origin DNA toward the Zn-domain, expanding and building up the melted DNA region. (C) Further ATP binding/hydrolysis pump the DNA and push the separated ssDNA loop against the channel wall, which may eventually force the ssDNA loops through the gaps between two subunits at the Zn-domain, allowing the growing ssDNA strands to exit from the side channels. (D) A slightly different topology of the unwinding fork from that in panel C. Here, only one ssDNA exits the central channel via a side channel while the other strand passes through the Zn-domain channel. (E, F) A model previously proposed for E1 origin melting and fork formation [17,39]. The model proposes that two adjacent E1 trimers (one trimer shown) assemble at the origin to melt the origin (E). Then, a ring-shaped E1 hexamer is formed around the melted ssDNA to form a fork (F). This model allows the E1 helicase to unwind the fork by the steric exclusion unwinding model, in which E1 will translocate on ssDNA in a 3’-5’ direction to unwind the fork on the C-terminal domain end.

Figure 3

Our refined double-pump looping model [10,11] for bi-directional DNA unwinding by a double hexameric helicase (such as LTag or MCM). (A) Two hexamers stay together through N-N interactions. The C-terminal AAA+ motor domain pumping the dsDNA ahead of the fork into the double hexamer, extruding the separated ssDNA as loops. The dsDNA traversing the narrow channel in the AAA+ domain may be squeezed and melted and is poised to be separated into two strands by the next cycle of pumping toward the N-terminus. The separated ssDNA may force its way between two subunits in the Zn-domain area, extruding through two side channels on each hexamer as loops [11]. During replication, primase, polymerase, RPA, and other replication proteins dock on the side of the double hexamer near the side-channel, capturing the emerging ssDNA loops as the template for synthesizing the leading and lagging strands. (B) A slight variation of the looping topology from panel-A, showing an asymmetric looping. Here only one ssDNA exits from the helicase domain side channel, and the other ssDNA exits around the junction between two double hexamers.