Structure of the eukaryotic replicative CMG helicase suggests a pumpjack motion for translocation

In eukaryotes, two Mcm2–7 rings are loaded onto DNA at origins of replication during G1 phase1,2. The double hexamer is inactive as a heli­ case but is activated upon entry into S phase by the Dbf4­dependent Cdc7 kinase, cyclin­dependent kinase and several transient factors that activate origins; subsequently, Cdc45 and GINS assemble onto each Mcm2–7 complex, thereby forming two active CMG helicases that surround each strand of parental DNA3,4. CMG then translo­ cates along single­stranded (ss) DNA in the 3′­to­5′ direction, thus facilitating bidirectional replication5–7. Several archaeal MCM crystal structures are known8–12, and the atomic model of the yeast double hexamer of Mcm2–7 has been determined by cryo­EM13. These struc­ tures have revealed that the Mcm2–7 core forms a two­tiered ring structure: an N­terminal domain (NTD) tier composed of a helical subdomain, a zinc­binding motif and an OB motif, and a C­terminal domain (CTD) tier containing the AAA+ motors. The structure of the 11­protein CMG complex is known only at low resolution: single­particle EM of fruit fly and yeast CMG has indicated that GINS and Cdc45 bind to one side of the Mcm2–7 hexamer, thus forming a double ring–like structure14–16. Most studies of replicative helicases have focused on homohex­ amers such as papilloma virus E1, SV40 T antigen, archaeal Mcm, Escherichia coli DnaB and bacteriophage T7 gene 4 protein17. In com­ parison, studies of the asymmetric 11­protein CMG are scarce, and many fundamental questions remain. Why is CMG an asymmetric ring? How does it translocate along DNA? Does it function differently from the rotary homohexameric helicases? Why are some ATP sites dispensable for CMG helicase activity, whereas others are essential? What is the location in CMG where double­stranded (ds) DNA is separated? There are two general models for the unwinding point: (i) the steric­exclusion model, in which the dsDNA split point is on the C surface, and leading­strand ssDNA passes through the central chan­ nel while lagging­strand ssDNA is sterically excluded to the outside of the ring and (ii) the ploughshare or side channel–extrusion model, in which the duplex is split inside the hexamer channel, and one strand is extruded from a side channel between the CTD and NTD tiers18. Translocation of a helicase along DNA is presumed to require at least two DNA­binding sites that differ in relative distance from each other during the ATP cycle, thus promoting movement along DNA17. Structural studies of papilloma virus E1 helicase have suggested the presence of six equivalent sites, the DNA­binding loops in the motor domains, that undergo changes in relative distance from one another in a rotary fashion, thereby escorting ssDNA through the ring in steps of one nucleotide per ATP10. The structure of bacterial DnaB has also revealed that the motor domain of each subunit binds ssDNA, and it has been proposed that entire protomers translocate along ssDNA in a spiral staircase–like manner, in steps of two nucleotides per ATP19. CMG lacks the six­fold symmetry of homohexamers, and whether CMG acts in a rotary fashion is unknown. To address the conformational changes during CMG transloca­ tion, and CMG’s mode of unwinding (steric exclusion or side­channel extrusion), we used cryo­EM to derive an atomic model of yeast CMG. We found that the Cdc45–GINS–Mcm architecture is sculpted to rigidify the Mcm2–7 NTD ring while still allowing movements within the Mcm2–7 CTD motor ring. The CTD ring adopts two dif­ ferent conformations, tilted and untilted, relative to a fixed NTD ring, thus giving rise to compact and extended forms of CMG with a maximum domain translocation of about 20 Å. Five of the Mcm2–7 subunits contain C­terminal winged­helix domains (WHDs) that