(A) Schematic representation of known domains in SWI2/SNF2 and ISWI ATPases. The helicase-like region (H) is highly conserved between the two types of ATPases whereas the flanking sequences are not conserved. (B) Model for monomeric chromatin remodeling ATPases: (i) ATP binding by the Helicase domain (H) causes DNA distortion at SHL−2 and stabilization of unpeeled DNA “slack” by the DNA Binding Domain (DBD). The H domain binds tightly in this state; (ii) ATP hydrolysis causes translocation of the H domain on the disrupted DNA at SHL−2, generating a DNA loop; (iii) Phosphate release weakens affinity of H domain increasing the probability of falling off. (iv) Dissociation of enzyme allows stochastic collapse of disrupted nucleosome into multiple distinct products. (C) Model for dimeric chromatin remodeling ATPases: (i) Only one monomer can bind ATP via its H domain at a given time. ATP can rapidly switch between the two monomers. The monomer that contacts the longer flanking DNA hydrolyzes ATP faster. Here this is the black monomer; (ii) ATP hydrolysis causes translocation of the H domain on the DNA at SHL−2, generating a DNA loop (iii) Phosphate release from the black H domain weakens the affinity of the black H domain but at the same time allows ATP binding to the white H domain switching it to a tight binding state at SHL+2. This allows propagation of the DNA loop without enzyme dissociation. In this state the white DBD does not properly engage the flanking DNA preventing ATP hydrolysis by the white H domain; (iv) Loss of ADP from the black H domain allows the white DBD to engage the flanking DNA. At this stage ATP can rapidly switch between the two monomers and further DNA movement can occur or the enzyme can dissociate resulting in a repositioned nucleosome.

(A) Inchworm mechanism proposed for monomeric SF1 helicases. Translocation occurs in two steps: (i) The helicase binds at the junction between single-stranded and double-stranded DNA. (ii) Domain 1A translocates towards 2A upon ATP binding because the binding site is formed by the cleft between 1A and 2A. In this state, domains 1B and 2B tightly bind and destabilize duplex DNA. (iii) Phosphate release allows translocation of 2A through the destabilized duplex. The absence of unhydrolyzed ATP reduces the interactions between 1A and 2A, allowing for this translocation. (B) Rolling mechanism proposed for dimeric SF1 helicases. (i) In the absence of ATP only one monomer binds at the junction between single stranded and double stranded DNA. (ii) ATP binding promotes tight binding of the leading helicase (black) to the ssDNA-dsDNA fork while the lagging helicase (white) remains bound to ssDNA. (iii) ATP hydrolysis results in dsDNA unpairing and subsequent phosphate release allows dissociation of the lagging helicase (white), which can then become the leading helicase in the next ATP cycle. In both (A) and (B), for ease of visualization three base pairs are shown to be disrupted in each ATPase cycle. In reality most helicases appear to unpair one base pair per ATPase cycle.