Different mechanisms for translocation by monomeric and hexameric helicases

Abstract

Helicases are ATP-dependent motor proteins that translocate along single or double-stranded nucleic acids to alter base-pairing structures or molecular interactions. Helicases can be divided to monomeric and hexameric types, each with distinct ternary structures, nucleic acid-binding modes, and translocation mechanisms. It is well established that monomeric helicases translocate by the inchworm mechanism. Recent structures of different superfamilies of hexameric helicases reveal that they use a hand-over-hand mechanism for translocation. Structures of bacteriophage T7 replisome illustrate how helicase and polymerase cooperatively catalyze DNA unwinding. In this review, we survey structures of monomeric and hexameric helicases and compare different mechanisms for translocation.

Graphical Abstract

Introduction

Helicases play vital roles in all processes of nucleic acid metabolism [1]. Helicases, which literally mean separating a double-stranded helix, are actually translocases and use the energy of ATP hydrolysis to translocate uni-directionally along DNA or RNA. By translocating along a single (ss) or double-strand (ds) substrate, they separate double stranded DNA or RNA, anneal single strands to a duplex, remove proteins from DNA or RNA, such as in replication fork formation [2,3], Holliday junction migration [4,5] and remodeling of chromatin structures [6]. Helicases often form large molecular assembles together with other nucleic acid-binding proteins and enzymes to perform DNA replication [7,8], recombination and repair [4,5,9], RNA synthesis and editing [10,11], or RNA-mediated regulation and metabolism [12]. For example, during DNA replication, template unwinding by helicases is coupled to DNA synthesis by polymerases and primases in a complex called replisome [13]. Corresponding to their diverse functions, helicases have different substrate preferences and mechanistic properties [1,2]. They can translocate on single-stranded (ss) or double-stranded (ds) DNA or RNA [1,6,12]. Translocation of all helicases has a defined polarity but can be either 5′-to-3′ or 3′-to-5′ [1]. When translocating along a double-stranded substrate, helicases mainly follow the track of one strand and move uni-directionally [14,15]. Moreover, helicases exhibit different step sizes, speed, and processivity [1,16]. Interestingly, the ATP-binding sites in all helicases are located at domain or subunit interfaces, and each is composed of Walker A and B motifs and an arginine finger [1,17]. ATP binding and hydrolysis induce directional inter-domain or inter-subunit rotations, which are propagated to generate movement relative to the bound nucleic acids, thus powering helicase translocation [18].

According to sequence conservations, helicases are divided to six superfamilies (SF) [1]. Members of SF1 and SF2 are monomeric and translocate like an inchworm [1,18]. In contrast, SF3-6 helicases all form hexamers, and replicative DNA helicases are all hexameric, but distributed in SF3, SF4 and SF6 [1]. The hexameric form may have advantages in processivity, efficient energy usage, and a six-fold platform for macromolecular machine formation and regulation. Recent structures of hexameric helicases suggested that they translocate along nucleic acids in a cyclic hand-over-hand fashion [2,8,19–22]. In this review, we compare and contrast the mechanisms of monomeric and hexameric helicases, with an emphasis on the helicases in DNA replication.

Monomeric and hexameric helicases

SF1 and SF2 helicases are distantly related while each contains a unique set of conserved signature motifs [17]. Each monomeric SF1 or SF2 helicase consists of two RecA domains, which together form an ATPase site at the domain interface (Figure 1) [23–27]. Additional domains are often present as inserts and are involved in substrate binding and helicase regulation [1,28]. SF1 members are bona fide helicases and unwind and separate double-stranded DNA or RNA by translocating along one of the two strands in either the 5′-to-3′ or 3′-to-5′ direction. SF2 helicases can translocate along either ss- or ds-DNA. Even when translocating along double-strand substrates, SF2 helicases only follow the track of one strand. Most of SF2 helicases exhibit the 3′-to-5′ polarity, and others such as XPD family members are 5′-to-3′ [1,29]. Chromatin remodelers, e.g. INO80 (Snf2) and ISWI (Chd1), belong to the SF2 superfamily and translocate on dsDNA [6].

Figure 1.

Structure and mechanism of monomeric helicases. (a-b) Cut-open views of SF1 UvrD-DNA complexes without (a) or with ATP (b). DNA interaction residues are highlighted in stick representation and the separation pin is indicated by a black arrow. The alternative tight and loose interactions are indicated by the hand icons in the schemes below each panel. (c) Structure of SF1 RecD2 complexed with ssDNA and ATP analog. The tight and loose interactions are indicated and are reverse of that in UvrD. (d) Structures of HCV NS3 (SF2 RNA helicase) with ssDNA and ATP analog. (e) Structure of Snf2 (SF2) complexed with nucleosome and ATP analog. The RecA domains of HCV NS3 are overlaid (transparent light green and blue), and ssDNA bound to HCV NS3 is shown in magenta. In all panels, domains 1A and 2A are colored green and blue, respectively, and inserted domains are shown in light orange. The ATP (or its analogs) and Mg²⁺ are depicted as sticks and balls. The tracking DNA strand bound to UvrD, RecD2, HCV NS3, and Snf2 are colored orange, pink, ruby and cyan, respectively. The directions of helicase translocation are marked by grey arrows in panel a-e. (f) Superpositions of ssDNA or dsDNA substrates bound to SF1 and SF2 helicases. (g) Diagram of ATPase dependent inchworm movement of helicase translocation.

SF3-6 helicases form ring or lock-washer shaped hexamers (Figure 2) and are divided to either RecA-like ATPase (SF4 and SF5) or AAA+ ATPase (SF3 and SF6) [1,2]. Although with distinct subunit orientations within hexamers, both ATPase folds belong to the P loop ATPase super group and share a conserved structure core [20,30,31]. Six subunits share an increased substrate-binding avidity, and the circular arrangement prevents nucleic acid from dissociation and thus likely enhances the processivity of translocation. Specific factors may be required to load the ring shaped helicases onto a continuous substrate [32]. Helicases in SF4 serve as replicative helicases in bacteria and bacteriophages [8,21]. Transcription terminator Rho (SF5) separates RNA from a DNA template [20]. SF3 and SF6 helicases translocate in the 3′-to-5′ direction, opposite to that of SF4 and SF5 [19,22,33,34]. The SF3 helicases are encoded by small DNA or RNA viruses for copying DNA or RNA, with E1 helicase from papillomavirus as a representative member [1,19]. Archaeal and eukaryotic replicative MCM helicases belong to SF6. Eukaryotic MCM2-7 form a hetero-hexamer and requires accessory factors GINS and Cdc45 to function as the holo-helicase, known as CMG [35].

Figure 2.

Structure of hexameric helicases. (a-f) Orthogonal views of hexameric helicase-DNA complexes of gp4 (a-b), E1 (c-d), and CMG (e-f). The N-terminal domains (NTD, or primase (Pri) domain in gp4) are shown as semi-transparent molecular surfaces, and the C-terminal helicase (Hel) domains are shown in cartoons. The subunits in gp4 and E1 are marked as A-F following the order from the 3′-to-5′ end of the DNA and colored in blue, light magenta, green, cyan, yellow, and red, respectively. The MCM subunits in CMG complexes followed the same color scheme while the MCM4 and MCM7 are in light blue and red to indicate that they are not involved in DNA binding. HelF subunits in gp4 and E1 as well as MCM4 and MCM7 in CMG are omitted in panels a, c, e, for clarity. The translocated position of HelF in gp4 are indicated by red outline or oval. The directions of helicase translocation are marked by grey arrows in panel a, c, e. (g) Superposition of DNAs in these helicase complexes with B-DNA (black and grey). (h) The hand-over-hand translocation mechanism of hexameric helicase base on gp4 helicase structures.

Structures and mechanism of monomeric helicases

Monomeric helicases share a conserved structural core (Figure 1). The ATPase site is formed between Walker A and B motifs from the N-terminal RecA domain (domain 1A) and an arginine finger from the C-terminal RecA domain (domain 2A) [1,17,23–26,36–38]. Substrate-binding motifs are present in both domains, and together they bind the backbone of 5–6 nucleotides (Figure 1, a–e), which lie orthogonal to the 1A-2A domain interface. Regardless of whether the direction of translocation is 5′-to-3′ or 3′-to-5′, domain 1A is always at the 3′ end relative to domain 2A (Figure 1b and c) [24–26,38]. ATP binding induces a ~20° rotation between the two RecA domains and tighten the domain interface [25,26,36]. Therefore, the apo state is “open”, and ATP-bound state is “closed” (Figure 1a and b).

Besides single-strand nucleic acid binding by two RecA domains, SF1 helicases also interact with the duplex portion to facilitate unwinding [25,26]. For example, in DNA helicases UvrD and PcrA, a β hairpin structure is positioned at the ss-ds junction and acts as a separation pin to split the dsDNA (Figure 1a and b) [25,26]. A similar β hairpin structure has also been identified in SF2 helicases [37]. UvrD contacts ~15 bp of dsDNA upstream of the ss-ds junction (Figure 1a and b). The single- and double-stranded regions of DNA are approximately orthogonal to each other when bound to UvrD (Fig. 1a–b), thus ssDNA translocation and duplex unwinding can occur autonomously [25].

SF2 helicases bind ss- or ds substrate alike (Figure 1d and e) [36,39,40]. The tracking strand in DNA duplex bound to Snf2 (DNA helicase) is superimposable with the single strand substrate bound to SF1 and SF2 helicases (Figure 1f). The similarity indicates that SF1 and SF2 helicases rely on a conserved mechanism for ss and ds substrate binding and translocation in either the 5′−3′ or 3′−5′ direction. Uni-directional movement by following one strand has been observed during nucleosome remodeling by Snf2 [23] and transcription-coupled DNA repair mediated by Cockayane syndrome group B (CSB helicase) [15]. Additional domains and subunits enable these SF2 helicases to recognize histones in nucleosomes and stabilize detached DNA to achieve nucleosome remodeling [41–44].

Monomeric helicases in SF1 and SF2 are shown to translocate along DNA/RNA by the inchworm mechanism (Figure 1g) [1,18,25,26,45]. Each helicase has at least two ssDNA attaching points (in UvrD there are two more attaching points for dsDNA) analogous to the fore and hindleg of an inchworm. As shown in the 3′−5′ UvrD helicases (Figure 1a and b) [25], ATP binding induces domain closure and at the same time, domain 1A (the hindleg) and the associated DNA exiting channel have loose contacts with the 3′ half of ssDNA and release one nucleotide, while domain 1A (foreleg) still holds the 5′ half of the ssDNA near the ss-ds junction. During ADP and Pi releases, the RecA domains open up and the foreleg releases the ssDNA and allows it to move towards the exit by one nucleotide. The net translocation is one nucleotide per ATP hydrolysis. The direction of movement can be reversed if the two DNA interfaces switch their tight and loose contacts with DNA, as shown in the case of RecD (Figure 1c) [38].

Structures of hexameric helicases

Most hexameric helicases have an N- and C-terminal domain, and the ATPase domain is always at the C terminus (Figure 2) [2,8,19–22,33,34,46,47]. The N-terminal domains are often required for helicase translocation by helping DNA or RNA binding and hexamer formation. Although bacterial and eukaryotic replicative helicases translocate in opposite directions along ssDNA, the N-terminal domains are always at the 5′ end of DNA relative to the C domains (at the 3′ end) (Figure 2, a, c and e).

DnaB and gp4 helicases both belong to SF4 and adopt a lock-washer shape when bound to ssDNA (Figure 2a and b) [8,21,48]. The subunits at two ends of each hexamer are over 20 Å apart along a ssDNA coil, leaving a gap between them, and a domain-swapped linker N-terminal to the helicase domain cyclically links hexamer together (Figure 2a and b). In contrast, E1 (SF3) and Rho (SF5) helicases are ring-shaped with small openings between two ATPase subunits at two ends of DNA (E1) or RNA (Rho) (Figure 2d) [19,20]. Archaeal MCM and eukaryotic CMG are overall ring-shaped [22,33,34]. However, five subunits within MCM and three or four subunits within CMG form spirals, and the remaining subunits bridge the ends of each spiral (represented by CMG structure, Figure 2e and f).

Nucleic acids reside in the central channel of the hexameric ATPase ring or lock-washer (Figure 2, a–f). Nucleic-acid interacting loops from six subunits interact with the backbone of DNA or RNA and form a spiral staircase surrounding the DNA or RNA helical coil. DNA bound to bacterial DnaB, phage T7 gp4, archaeal MCM and eukaryotic CMG helicases adopts an A-like structure, with a diameter of around 23 Å. Each subunit interacts with the backbone of two nucleotides (represented by gp4 and CMG structure, Figure 2g) [8,21,22,33,34], and each hexamer covers 12 nts in a full turn. The ssDNA coil complexed with MCM has a shorter helicase pitch (~12Å) than that bound to gp4 and DnaB (~ 28Å). In contrast, ssDNA in E1 and ssRNA in Rho complexes have a much narrower diameter (15 Å or less) and short pitch (~18 Å) made of a 6-nt turn, as each helicase subunit interacts with one nucleotide (represented by E1 structure, Figure 2g) [19,20].

ATP-binding sites occur at each subunit interface, and each is composed of Walker A and B motifs from one subunit and an arginine finger from the other [1,8,19–22,34]. In the lock-washer shaped gp4 and DnaB, two end subunits do not form an interface for ATP-binding site (Figure 2, b, d and f) [8,21,48]. Although six ATP-analog molecules have been found in E1, Rho and archaeal MCM hexamer [19,20,22], the ATP-binding site between two end subunits is rather open and ATP can easily diffuse out [49]. In CMG, three ATP-analog molecules are found between four adjacent DNA-interacting subunits [34]. The ATPases sites at the subunit interfaces show gradual conformation changes along the ssDNA-bound hexamers [8,19,20,22]. The directions of gradual changes are opposite in the 5′-to-3′ and 3′-to-5′ helicases and correlated with the direction of translocation.

The hand-over-hand mechanism for hexameric helicases

In hexameric helicase, six DNA binding loops (one in each subunit) align side-by-side along a spiral DNA backbone (Figure 2). The strength of DNA interaction is modulated by the ATPase cycle. ATP binding stabilizes the subunit interface and strengthens DNA binding; whereas upon ATP hydrolysis and product release, both the subunit interface and the DNA binding become unstable. In the gp4-ssDNA complex structures, the gradual changes at ATP-binding sites from loose (3′) to tight (5′) and DNA binding from tight (3′) to loose (5′) suggest ATP hydrolysis occurs first at the 5′ end, while the other five subunits still hold ATP and DNA [8,19,20,22]. Following ATP hydrolysis and product release, the 5′-end subunit is released from its neighboring subunit and DNA and translocates to the 3′ end, where it gains tight DNA binding and forms a new subunit interface to bind ATP (Figure 2h). In the gp4 cryo-electron microscope structures, the gp4 subunit bound to the 5′-end of DNA adopts three distinct positions, which represent its move from the 5′ to the 3′-end of DNA over the five stationary subunits [8]. Each of the remaining five subunits will sequentially translocate from one end of the lock-washer (spiral) to the other. This translocation mechanism has been likened to hand-over-hand movement [8]. N-terminal domains and domain-swapped linkers keep the hexamer intact during subunit translocation. Depending on the direction of the gradual changes, the helicase may move in the 3´−5´ or 5´−3´ direction [8,19,20,22]. The A-like DNA in hexameric helicases indicates dsDNA can be accommodated similarly as ssDNA with slight bending and deformation [8,21,34]. Indeed, CMG has been shown to mainly follow the track of one strand during translocation along dsDNA [14,50]. Similar sequential hand-over-hand translocation have been widely proposed for AAA+ type peptide translocases [51].

The hand-over-hand translocation may be large-scale movements of entire domains as observed with DnaB [21] and gp4 [8] or relatively small shifts of the DNA-binding loops coupled with subunit rotations as observed in E1 [19] and Rho [20]. However, in CMG complex, not all ATPase sites are essential for DNA unwinding [35], although sequential ATPase hydrolysis and hand-over-hand movement is evident in its homolog archaeal MCM [22]. It is possible that instead one subunit moving at a time, two or three subunits may move en bloc in CMG and thus not all ATPase sites are necessary [8].

During the hand-over-hand movement, each of the six subunits within a hexamer alternatively takes the lead position (Figure 2h). This is in contrast to the inchworm movement of monomeric helicases, in which the two ATPase domains maintain the same front and back positions (Figure 1g). The step size for all monomeric helicases is 1-nt per ATPase cycle. The step size of hexameric helicases is determined structurally by the number of nucleic acids bound to each subunit, resulting in a 2-nt step per ATP hydrolyzed by gp4, DnaB, MCM, and CMG helicases and a 1-nt step for E1 and Rho. Biochemical and biophysical analyses of hexameric helicases, however, are inconclusive, and both 2-nt [52–55] and 1-nt [56–58] step sizes have been proposed, even for a single helicase [52,53,55–57].

Helicase within replisome

Helicases act as central organizers during DNA replication by connecting to both polymerase on the leading strand and the primase on the lagging strand [7,13]. The helicase unwinding is tightly coupled to the leading strand DNA synthesis in bacteria as well in eukaryotes [57,59,60]. In the newly solved bacteriophage T7 replisome structure, the three enzymes form a three-tier architecture with hexameric gp4 helicase in the middle along the ssDNA substrate (Figure 3a) [8]. The leading-strand polymerase interacts with the C-terminus of gp4 helicase, and the parental dsDNA is situated between the helicase and polymerase. The primase located in the N-terminal domain of gp4 captures the lagging strand DNA exiting from the helicase. The leading-strand DNA synthesis can take place without the lagging-strand synthesis [45]. At the ds-ss junction the leading- and lagging-strand templates are at 90 degree to the parental dsDNA, forming a T-shaped fork. A hairpin structure from the leading-strand polymerase stacks with the double-strand end of the parental DNA and facilitates strand unwinding, which is analogous to the separation pin in monomeric helicases (Figure 3b) [25]. The concerted pulling of daughter strands off the T-shaped fork and the protein hairpin explains how the leading-strand polymerase stimulates DNA unwinding by the helicase. The dsDNA unwinding by the helicase-polymerase complex share mechanistic similarities with that of monomeric helicase, with motor domains to translocate along ssDNA, separation pin to separate basepairs, and the orthogonal single- and double-strand arrangement to unspool the DNA coil [18,25]. Although bearing no sequence homologs, partial structures of eukaryotic replisomes also suggested a similar three tier structure as T7 (Figure 3c) [61,62]. The lagging strand primase attaches to the N-terminal domain of CMG through Ctf4, while the leading strand Pol ε interacts with the C-terminal ATPase domains of CMG. The difference is that the eukaryotic CMG translocates towards the 5′-end on the leading strand and the DNA fork is situated between the primase and helicase (Figure 3d). The DNA synthesis by Pol ε is shown to help CMG helicase to unwind DNA processively [37], perhaps by preventing DNA back tracking.

Figure 3.

Helicases in bacterial and yeast replisomes. (a) Structure of the T7 replisome. The “T” shaped DNA fork is highlighted by an overlaid semi-transparent T. (b) A zoom-in view of the T7 replication fork. A β hairpin in the leading-strand DNA polymerase is at the fork to promote DNA strand separation. (c) The cryoEM image of yeast CMG, Pol ε and Pol α (primosome) assembled on a DNA fork reveals a three-tier-core structure [14], similar to that of the phage T7 replisome shown in (a). (d) Diagrams of the DNA fork substrate in the three-tier-core replisome from bacteria and eukaryotic systems. The conserved DNA part are shown in orange, and the different downstream parental DNA in bacterial and eukaryotic replisome are shown in pink and yellow, respectively.

Conclusion

Structures of various helicase ternary complexes provide mechanistic details of how helicase couples ATP hydrolysis to linear translocation along nucleic acids. Two distinct mechanisms, inchworm and hand-over-hand, appear to be employed by monomeric and hexameric helicases, respectively. In both cases, the direction of translocation can be reversed depending on the order of tight and loose contacts. Both types of helicases can bind and translocate along ss- and ds-substrates in similar fashions. An orthogonal arrangement of ss-ds junctions and a protein hairpin at the ss-ds substrate junction promote duplex unwinding by both monomeric and hexameric helicases. The basepair “separation” hairpin in the T7 DNA replisome is provided by the leading-strand polymerase. As helicases are often parts of large molecular machineries, structures of the T7 replisome begin to shed light on how helicase actions are coupled to other enzymatic activities.

In the future, we need to develop new biophysical and biochemical assays that enable accurate measurement of helicase step sizes and determination of their translocation mechanisms when complexed with accessary components of natural biological processes. Moreover, investigations of how helicases are recruited and regulated in native assemblies and how helicase translocation is coupled to other enzymatic activities are needed. Many helicases are drug targets for various diseases, and development of inhibitors targeting specific helicases has great therapeutic potential.