Turning the Mre11/Rad50 DNA repair complex on its head: lessons from SMC protein hinges, dynamic coiled-coil movements and DNA loop-extrusion?

The maintenance of chromosomal integrity and genomic stability is central to cellular viability in all organisms. DNA double-strand breaks (DSBs), where both strands of the DNA duplex are concurrently broken, represent some of the most cytotoxic forms of genetic damage and require immediate, efficient and accurate repair in order to prevent genomic rearrangements and mutation [1,2]. Indeed, the failure to precisely repair DSBs in higher organisms is frequently associated with chromosomal instability and the development of cancer [3]. Cells can employ two broad DNA repair strategies at DSBs to restore the intact chromosome. The first method, known as non-homologous end-joining (NHEJ), reseals the fractured chromosome directly by blunt-ended ligation with minimal processing at the break [4,5]. Non-canonical ‘alternative’ NHEJ mechanisms have also been identified during which the ends of the DSB undergo limited processing prior to ligation and frequently the repair is thereby mediated via micro-homologies; such repair modes are therefore referred to as micro-homology mediated end-joining (MMEJ) [6]. In contrast, a second higher fidelity mechanism, homology-directed repair (HDR), requires a homologous chromosome as a template to restore the genetic content of the chromosome [1,7]. During this process, helicase and nuclease components are recruited to DSBs and perform long-range chromosome resection to generate the single-stranded tails required for the stand-invasion steps of homologous recombination (HR) [8–12] (Figure 1). These tails become coated by recombinase proteins of the Rad51/RadA/RecA superfamily, which are conserved across the divisions of life. The resultant proteo-filament is then able to invade the DNA duplex of the homologous chromosome and mediate HR driven DNA repair [13–15] (Figure 1). This HR pathway is, therefore, an ancient, conserved and essential DNA repair mechanism employed by all three divisions of life; the bacteria, the archaea and the eukaryotes [8]. In non-polyploid organisms, this high-fidelity mechanism is restricted to the S/G2 phases of the cell cycle when a copy of the genome is available as a homologous template, while by contrast the NHEJ pathways are possible at all cell-cycle stages [1].

It is well established that the Mre11–Rad50 (M/R) complex operates during the earliest stages of DSB repair, playing critical roles in the recognition and processing of these sites of damage and influencing which of the repair mechanisms is utilized by the cell [16–22]. The complex is also central to the DNA damage signalling pathways in eukaryotic organisms [23]. Interestingly, the M/R complex also plays roles in the MMEJ and NHEJ pathways [24]. The tight association of the enzymatically distinct Mre11 nuclease and the Rad50 ATPase at the core of this critical DNA repair complex is an arrangement conserved in all organisms [11,18], and M/R homologues are even encoded by dsDNA bacteriophages [25]. These two core protein components undergo dramatic ATP hydrolysis-mediated conformational changes that intrinsically link the nucleolytic activities of the Mre11 nuclease with the DNA tethering and chromosomal bridging functions of the Rad50 protein [18,19]. Nucleotide hydrolysis occurs within the catalytic domains of the Rad50 component, which is a divergent member of the structural maintenance of chromosomes (SMC) superfamily of ATPases [26–28]. In eukaryotic systems, a third accessory protein (Nbs1 in metazoans, or the Xrs2 homologue in yeasts) associates with M/R core (Figure 1), enabling the complex to fulfill critical functions in cell signalling and coordinating the enzymatic activities with cell-cycle checkpoints [29–31]. However, this tripartite arrangement seems to be a eukaryotic innovation, as homologues, or indeed even functional analogues of the Nbs1/Xrs2 protein, have not been identified to date in either bacterial or archaeal species.

Much of our current understanding of the DNA-repair mechanisms of the M/R complex has arisen from structural observations garnered from studies of thermophilic archaeal and bacterial M/R complexes, in particular from the species Pyrococcus furiosus, Methanocaldicoccus janaschii and Thermotoga maritima [32–35]. It is often advantageous to investigate the ancestral, biochemically tractable and structurally rigid thermophilic homologues of eukaryotic proteins and in the case of the M/R complex these analyzes have been particularly illuminating [36]. Such studies have revealed that the catalytic core of the complex forms a tetrameric arrangement consisting of a dimer of Mre11 in conjunction with two Rad50 subunits [32–35]. The Mre11 dimer organizes via four-helix bundle interactions between the core nuclease domains, while juxtaposed ‘nuclease-capping’ domains, which regulate the accessibility of DNA substrates to the Mre11 active sites, extend from the catalytic domains, and these connect to C-terminal helix–loop–helix domains, which mediate binding to the Rad50 partners within the tetrameric core [18,19] (Figure 2A–E). The Rad50 ATPase, like the other members of the SMC superfamily, is a dumbbell-shaped protein consisting of two compact globular domains, which harbour the Walker A phosphate binding-loop and Walker B ATPase motifs respectively, linked by a long coiled-coil region typically 600–900 amino acids in length [27,28,37]. The V-shaped Rad50 molecule is angled around a cysteine-containing, zinc-coordinating motif (CXXC) that is located at the centre of the coiled-coil region [38]; this permits the two globular domains to approach each other and unite to form a functional ATPase catalytic site [18,19] (Figure 2A). The zinc-binding element, distal to the catalytic domains, is commonly referred to as the ‘zinc-hook’ and is analogous to the centrally located ‘hinge’ region of other SMC proteins [38–40]. M/R tetramers can remain associated via these zinc hook regions when the globular domains have separated (Figure 2B) [41–43]. It has also been suggested, based on atomic force microscopy (AFM) and electron microscopy studies that these zinc-hook regions play critical roles in mediating chromosomal bridging via intermolecular associations, with the hooks associating between reciprocal Rad50 molecules [33,38,41,42,44,45]. In the absence of a bound DNA substrate, it has been observed that the coiled-coils of a Rad50 homodimer associate intramolecularly, flexing to form a ring-like conformation, reminiscent of the arrangements seen in canonical SMC family proteins [41,44–47] (Figure 2D). These intramolecular linkages can be broken upon DNA binding to the globular domains and the seemingly flexible coiled-coils convert into inflexible linear rods. Such changes permit the zinc hooks at the apices to associate intermolecularly with other M/R complexes in a manner that has been proposed to support DNA bridging during the repair events [41] (Figure 2E). Interestingly, the coiled-coil regions in bacterial species are somewhat shorter than their counterparts in the archaea, with reported lengths, when extended, of ∼250 Å compared with ∼300 Å, respectively [42,48]. Notably, bacteriophage homologues possess even shorter coiled-coils of ∼100 Å [49]. By comparison, the region in eukaryotic homologues is substantially longer, extending almost 600 Å [39,42] (Figure 2E). These variations in coiled-coil lengths are presumably reflective of the inherent inter-chromosomal distances and differences in chromosomal organization and complexity across the three divisions of life [39,48].

Early structural investigations of the M/R complex revealed the nature of the tetrameric assembly and also identified how the complex might bind to DNA ends [33,34]. However, it initially seemed unclear how the separable enzymatic functions of the Rad50 ATPase and the Mre11 nuclease might be coordinated during the recognition and subsequent repair of sites of DNA damage. Considerable insight was subsequently provided by a series of crystal structures and small-angle X-ray scattering (SAXS) analyzes performed in the presence or absence of ATP, or non-hydrolysable analogues of this nucleotide [23,32,37,50,51]. These studies unexpectedly revealed that the globular head domains of the Mre11/Rad50 complex experience extensive conformational shifts (Figure 3). Upon ATP binding a compact arrangement is formed, wherein the two ATPase domain are united, drawing all four of the globular domains of the M/R tetramer together in a closed conformation. However, the Rad50 ATPase domains separate dramatically upon nucleotide hydrolysis and the conformation opens into a distinctive ‘W’-shaped structure, joined centrally by the Mre11 dimer, with the two Rad50 ATPase domains poised at opposite ends of the arrangement [18,19] (Figure 3). In this configuration the two Rad50 subunits attach to each of the Mre11 subunits via the helix–loop–helix region (Figure 3A). When arranged this way the Mre11 nuclease domains are accessible for DNA processing, but by contrast, these nucleolytic sites become partially masked in the more compact arrangement. When ATP binds to the Rad50 subunits the resulting conformational changes are transmitted to the Mre11 dimer through rotation of the helix–loop–helix region and this results in the M/R tetramer forming a clamp-like arrangement with increased affinity for DNA substrates [50]. Interestingly, in eukaryotic cells, it has been revealed that the Mre11-dependent nucleolytic activity of M/R complex plays a key role in preparing a DSB for HR-dependent repair [32,50,52,53]. In eukaryotes, DSBs are initially recognized and bound by the Ku DNA repair proteins. It has been revealed that before the G1/S phase transition of the cell cycle the endonucleolytic activity of the M/R complex, stimulated by the activity of the CtIP(Sae2) cofactor [54] (Figure 1), cleaves DNA proximal to the break, thereby removing Ku protein [52–56]. This permits the recruitment of the long-range DNA end-resection machinery, such as Dna2 and Exo1, effectively resulting in the selection of the HR pathway rather than the NHEJ mechanisms that are mediated by the end-binding Ku proteins in conjunction with DNA-PK and ligase [9,57].

Subsequent experimental interrogation of the distinct arrangements of the M/R complex revealed that the ATP-dependent actuation of the compact state promotes DNA-end recognition and also stimulates the chromosomal tethering activities of the M/R complex at DSBs [11,18,58]. In contrast, following ATP hydrolysis and the release of the complex into the more accessible state, the Mre11 catalytic sites are exposed potentiating the nucleolytic capability of the complex, facilitating processing of the DSB and initiating the end-resection process [11,18,58]. Thus, it has been proposed that the Mre11/Rad50 complex operates as an ATP dependent ‘molecular-switch’ providing a method of regulation that controls the separation of function between the disparate DNA break-recognition and end-resection roles [11,19,58]. Indeed, the seemingly low reported rate of in vitro ATP-hydrolysis by the M/R complex appears consistent with the energetic demands required by a molecular switch. However, it has been observed that this rate of nucleotide turnover can be dramatically stimulated upon binding to double-stranded DNA, at least in prokaryotic or viral examples of the M/R complex [32,59]. It therefore currently remains undetermined if the energy derived from ATP hydrolysis by the M/R complex might also be utilized to drive the active manipulation of chromosomal substrates in addition to the well-characterized actions as a functional switch. Interestingly, recent insights into the mechanism of action of the canonical SMC family proteins, of which Rad50 is a member, may provide intriguing clues as to how the energy-dependent movement of the long coiled-coil regions could plausibly operate to organize and manipulate chromosomes [46].

The SMC superfamily is conserved across the bacterial, archaeal and eukaryotic divisions of life and includes the condensins, cohesins, SMC-5/6 DNA repair complexes, and prokaryotic SMCs [46]. Like the Rad50 members, these proteins arrange as dimers and adopt a distinctive ring-shaped arrangement mediated by the long and flexible coiled-coils, which encircle their DNA substrates. While Rad50 proteins form homodimers, the other SMC family members are heterodimeric formed of two distinct but closely related SMC subunits. One end of the ring is formed by the dimerization of the two SMC subunits at the hinge region, which is analogous to the zinc-hook region of Rad50 homologues. At the opposing end, the coiled-coil toroid is sealed by the association of the globular ATPase domains, stabilized by an additional kleisin subunit, which is conserved in the majority of prokaryotic and eukaroytic SMC complexes. In the Rad50 homologues, however, this ATPase domain stabilization is mediated by the unrelated Mre11 dimer [27,60] (Figure 4A). Although it was initially unclear how these proteins acted to manipulate their DNA substrates, it is now evident that SMC complexes operate as unconventional molecular machines, utilizing the energy liberated by ATP hydrolysis to drive critical chromosomal maintenance events [47,61,62]. This allows the complexes to actively sculpt the chromosomal landscape during their essential roles in gene expression, DNA repair and chromosome cohesion, compaction and organization. A growing body of experimental evidence has revealed that these complexes slide along DNA substrates in a process dependent on ATP hydrolysis. This substrate translocation is frequently linked to a dramatic extrusion of extensive DNA loops, which are critical for the organization of chromosomal architecture [47,63–72]. SMC complexes had been implicated in the formation of these critical looped chromosomal structures for some time, but there was initially some confusion as to how these unusual DNA fastening complexes could drive such dramatic events. It was initially hard to consolidate the chromosomal sculpting events with the inherently low ATP hydrolysis rates observed in SMC proteins, which ranges from less than 0.1 to 2 molecules of ATP per SMC dimer per second [63]. In contrast, most conventional DNA translocases display a considerably higher rate of ATP turnover frequently turning over thousands of ATP molecules per second [73]. Due to this disparity, it was originally suspected that additional molecular motors such as the powerful RNA polymerase of the transcriptional apparatus might provide the energy necessary to drive the formation of loops, with the SMC proteins simply topologically capturing the conformational changes. However, recent advancements in single molecule experimentation have now visualized the progressive extrusion of DNA loops mediated by SMC proteins alone [68,69,74–76]. These studies are now starting to reveal the mechanisms by which the SMC superfamily orchestrate chromosomal organization by DNA loop formation. Indeed, unique and unconventional translocation mechanisms have been observed, dependent on nucleotide hydrolysis in the ATPase domains of the SMC subunits [47,63–70,77].

Further efforts are required to elucidate the mechanisms by which the SMC complexes actively modulate chromosomal structures, although several models have been posited, including an ‘inchworm’ translocation mechanism, a ‘hand-over-head’ myosin-like stepping system and dynamic ‘pumping’ and ‘scrunching’ models [47,63,67]. These models have been formulated by taking into account the various experimental observations in the single molecule studies [68,69,74–76]. Currently, the ‘pumping’ and ‘scrunching’ models appear the most plausible and it remains possible that SMC action could work via a combination of these modes [47,63,67]. Furthermore, it has not yet been determined if the different members of the SMC superfamily operate via a common mechanism, or whether substantially different modes of action are employed by different SMC complexes.

It is widely accepted that the ATP-dependent engagement and release of the globular head domains is the driving force behind the dynamic and energy-dependent transactions that SMC superfamily proteins direct towards their chromosomal substrates [46] (Figure 4A). Indeed, early experimental observations of the prokaryotic SMCs have proposed models of action that suggest that architectural rearrangements transmitted between the ATPase head and hinge domains via the long coiled-coil arms appear critical for their respective roles in chromosomal organization [47,63,78–83]. Considering the clear evolutionary relationship between Rad50 and the other SMC family members, observable at both the level of amino-acid homology, and also in the overall architecture of these distinctive ATPases [28], it is tempting to speculate that commonalties might be retained in their respective modes of action. Whilst the globular catalytic domains of all the SMC family members are evolutionarily related to the overarching superfamily of ABC (ATP binding cassette) membrane transporters [27], the broad SMC class is distinguishable by the arrangement of the two globular ATPase domains at the N- and C-termini of the molecule, separated by the long coiled-coil regions. Indeed, the entire SMC class of proteins, which include the Rad50 homologues in addition to canonical SMCs, all form the characteristic V-shaped arrangement, which closes to form the toroid that can encircle the DNA substrates [46,47,84]. All of these SMC superfamily proteins share the common mechanistic feature of the engagement and disengagement of the disparate globular head domains as a result of ATP binding, hydrolysis and release, which results in conformational changes of the long coiled-coil regions [46,78,81,82,85] (Figure 4A). Indeed, studies of the archaeal Rad50 proteins have revealed that nucleotide hydrolysis results in a 30° rotation of the Rad50 globular lobes that effect the positioning of the coiled-coils, the accessibility of Mre11 nucleolytic sites, and may even culminate in localized unwinding of the DNA duplex at the DSB [32,37,86].

Whilst the canonical SMC superfamily proteins were originally envisaged as static DNA-clasping structures it has become evident that dynamic architectural rearrangements of the long coiled-coil regions occur during association with DNA substrates. Conformational changes of the coiled-coil arms are also coincident with the binding and hydrolysis of ATP at the head domains [47,63–67,78,81,85]. These allosterically regulated architectural arrangements appear critical for the established roles of the SMC family proteins in a variety of biological processes, which include sister-chromosome cohesion, chromosome condensation and DNA repair [28,47,60,79,81,85] (Figure 4A). When comparing the Rad50 homologues to other SMC family members it is evident that the hinged apex of the coiled coils is the most divergent region. This feature is considerably shorter in Rad50 proteins at around thirty amino acids in Rad50 [38] compared with around one-hundred and fifty amino acids in the SMC hinges [39] (Figure 4B,C). The Rad50 region also bears the unique ‘zinc-hook’ motif (Figure 4C) that mediates zinc-dependent coiled-coil dimerization and promotes the intermolecular transactions that are believed to support chromosomal bridging [38]. Atomic force microscopy (AFM) and electron microscopy (EM) observations of the Rad50 complexes from both archaeal and eukaryotic species have revealed that the Rad50 coiled-coil regions undergo dramatic mesoscale conformational changes as a result of both DNA binding and ATP hydrolysis. [41,42,44,45,87] These events switch the conformation of the coiled-coils from seemingly flexible arrangements that encourage ring-shaped intramolecular associations via the zinc hooks, to rigid rod arrangements where the intramolecular interactions of the zinc hooks are broken [41,42,44,45] (Figure 2). Interestingly, the canonical SMC family proteins also adopt conformations that alternate between rod and ring shapes, [47,65] although the more extensive interaction interface observed at the canonical SMC hinge appears to promote and maintain the intramolecular associations at the hinge interface in both of these arrangements [78,79] (Figure 4A). Nevertheless, mounting experimental evidence appears suggestive that the mechanisms of chromosome organization and manipulation by both the classical SMC representatives and also the Rad50 proteins involve transitions between rod and ring shapes of the long coiled-coil regions.

When considering canonical SMC family proteins, it has been proposed that DNA substrates can enter or exit the SMC ring structure at one of two principal sites. These binding events are mediated by either the opening of the globular ATPase domains at the head, or alternatively occur following separation of heterodimer at the hinge region at the opposite end of the coiled-coils [88–90]. Certainly, in both prokaryotic and eukaryotic SMC proteins, the hinge region has been suggested to play essential roles in DNA-binding and act as a DNA entry gate, in addition to the established functions in hetero-dimerization [40,46,85,89–92]. In contrast, the analogous zinc hook region in the Rad50 has thus far only been implicated in either the intra-molecular coiled-coil associations, forming the closed ring-like or rod-like conformations [93,94], or alternatively in the inter-molecular interactions between disparate Rad50 molecules, which appear more prevalent following DNA binding [38,41]. However, little consideration has been given to putative roles of the zinc-hook region in direct DNA binding, in a manner evocative of the canonical SMC hinge. Indeed, the majority of experimental studies describe modes of M/R-mediated nucleic acid interaction that exclusively involve the globular ATPase domains and the localized coiled-coil regions that abut these globular catalytic termini [95,96]. Despite this commonly held perception, it perhaps should be noted that there is experimental evidence to suggest that the absolutely conserved Rad50 zinc-hook region might plausibly play an important role in direct DNA binding. Might the zinc-hook region of the M/R complex function as a counterpart of the SMC hinge, not only in terms of a coiled-coil interface, but also as an important site for mediating interactions with DNA substrates? In support of this view, it is particularly notable that mutations in the zinc hook region from Saccharomyces cerevisiae lead to phenotypes that are comparable with Rad50 null mutants, highlighting the essential nature of these regions [97]. Furthermore, it has been observed in mammalian cells that these zinc-hooks are essential for guiding the M/R complex to chromosomal DSBs following exogenously induced chromosomal damage [98]. Interestingly, it has also been revealed that the zinc-hook region from the archaeon P. furiosus can be substituted for the SMC hinge region of the prokaryotic condensin homologue in Bacillus subtilis, seemingly without affecting cellular viability under nutrient-rich conditions [80]. Similarly, it has been shown that substitution of the zinc-hook of the yeast M/R complex for the hinge region of the MukB bacterial condensin has minimal physiological effects on the resulting mutant yeast cells and thus the zinc-hook and hinge dimerization domains appear interchangeable [100].

Might these experimental observations imply that Rad50 zinc-hook regions can interact directly with DNA substrates? To date, it remains undetermined precisely how the M/R complex initially locates and then associates with sites of DNA damage. Indeed, it is particularly unclear which functional domains of the complex might be involved in the detection of DSBs. However, a recent single-molecule study has revealed that the M/R complex is considerably more mobile on chromosomes than was previously anticipated [100] This study also demonstrated that the M/R complex is able to translocate along even histone-occupied chromatin, effectively scanning for broken DNA ends via facilitated diffusion along the substrate [100]. This unexpected DNA-tracking mechanism appears dependent on the Rad50 globular domains of the M/R complex. Presumably, the coiled-coils arms of the Rad50 dimer also encircle the chromosome during this scanning process in a manner evocative of the previously described movements of other canonical SMC proteins upon chromosomal substrates [61,63,77,88,101]. Certainly, other studies are indicative of the coiled-coil regions associating with DNA substrates [60,102]. It therefore appears that the M/R complex tracks chromosomes via associations with the Rad50 globular domains utilizing an as yet undetermined interaction mode. It has also been revealed that upon arrival at the DSB the Mre11 component of the M/R complex then recognizes the broken DNA terminus, which it binds with high-affinity [94,100]. It should be noted that previous models for DSB recognition have almost exclusively depicted the M/R complex DNA loading events occurring directly at the break without prior chromosomal translocation [11,18]. This dynamic DNA end-surveillance system, dependent on chromosomal-tracking revolutionized our views of DSB recognition by the M/R complex. Furthermore, in support of this model, a recent study of the M/R complex from the thermophilic archaeon Sulfolobus acidocaldarius, using real-time AFM, has demonstrated that the zinc-hook region of Rad50 appears to associate transiently with DNA substrates [43]. This study also predicted, by classical molecular dynamics simulations of the archaeal Rad50 hinged region on double-stranded DNA, that the zinc-hook region and juxtaposed coiled coils are seemingly able to track the minor groove of a DNA duplex until an end is encountered [43]. It therefore seems plausible that the Rad50 zinc-hook and associated coiled-coils may aid the facilitated diffusion along DNA substrates in addition to established transactions with the Rad50 globular domains [32,50,95,96]. It remains to be determined if any DNA binding activity can also occur close to the zinc-hooks when the region is in the dimerized conformation.

Following the recently described method of DNA translocation by the M/R complex, it seems logical to question whether this scanning for broken DNA ends might be limited to facilitated diffusion. Could, for instance, the active architectural rearrangements in the M/R complex brought about by ATP hydrolysis be harnessed to actively drive these translocation events along DNA duplexes? While the single-molecule study of the S. cerevisiae M/R proteins did not establish a link between the Rad50 mediated diffusion and ATP hydrolysis [100], it is noteworthy that the AFM study of the S. acidocaldarius M/R complex identified unexpected and unprecedented evidence of ATP-dependent DNA duplex strand separation [43]. This dramatic DNA unwinding, which extended over several hundred base-pairs, was observable as ‘bubble’ structures on double-stranded DNA substrates. These duplex strand-separations might conceivably arise as a consequence of supercoiling or writhe generated from the energy-dependent manipulation of the DNA substrate by the archaeal M/R complex [43]. Fast-scan AFM analyzes in fluid also revealed that archaeal M/R complex can manipulate DNA substrates in a manner consistent with translocation and loop formation [43]. While further experimental interrogation is required to verify how the M/R complexes are involved in the active manipulation of DNA substrates, the mounting data provided by studies of the canonical SMC proteins indicate that the ATP-dependent architectural contortions induced by these complexes are indeed linked to DNA translocation and chromosomal loop extrusion [43,68,69,74–77,80]. Furthermore, the growing body of evidence from studies of prokaryotic and eukaryotic SMC proteins indicates that the transitions between closed rod and ring conformations of the coiled-coil arms appear to mediate the DNA loop extrusion (Figure 4A); these events involve allosteric communication between the globular domains and the hinge regions via the coiled-coils of the SMC proteins [78–81]. Indeed, the ATPase globular domains appear to be able to exist in at least three arrangements: a so-called engaged ‘J-state’ rod-like arrangement, or fully disengaged in the absence of ATP, or the engaged ring-like ‘E-state’ in the presence of ATP [47]. Conformational changes in the vicinity of the SMC hinge, alternating between juxtaposed ‘J-state’ coils and the ring-like ‘E-states’, as a result of either DNA or ATP binding at the globular domains, appear critical for the actuation of these structural shifts [47]. These events seem to be key for powering the translocation of the SMC complexes along chromosomal substrates and loop extrusion. Indeed, after the pioneering single-molecule study of the S. cerevisiae condensin revealed that this class of SMC proteins are indeed able to translocate along immobilized DNA substrates in vitro [61], it has subsequently become clear that similar events appear to be associated with the formation of DNA loops in condensins, cohesins and bacterial SMCs alike [63,66,67]. While the mechanism of action is still currently being investigated and debated, ‘scrunching’ models, where the coiled coils bend at non-helical ‘elbow’ regions located in the middle of coiled-coils [65,70,103,104], seem a likely scenario [47,63,65,66]. It therefore appears that canonical SMC family of proteins do indeed utilize a previously uncharacterized ATP-dependent and processive mechano-chemical motor mechanism, which is essential for the extrusion of DNA loops through the SMC arms [47,61,63].

It is noteworthy that recently the structure of both the human and E. coli Rad50 zinc-hook regions have been determined and these can form rigid rod-like structures upon DNA binding [93,94] (Figure 4C), distinct from the previously described splayed zinc hook structure that was revealed using the hinged region from the thermophilic archaeon Pyrococcus furiosus Rad50 [38] (Figure 4C). The observation of these two different conformations at the Rad50 zinc-hook region, reminiscent of the two characterized analogous arrangements of the canonical SMC hinges adds weight to the intriguing possibility that the M/R complex may also be an unconventional chromosome manipulating motor. Indeed, models have suggested communication between the catalytic head and zinc-hook regions, via transmission through the long Rad50 coiled-coils [87,94].

Over the last three decades, considerable progress has been made towards elucidating the action of the M/R complex. Despite these experimental insights, many mechanistic details regarding how this intriguing DNA repair complex targets DSBs and regulates chromosomal stability remain unexplained. The recent discovery that the M/R complex is considerably more mobile on chromatin than was previously anticipated, coupled with the idea that the essential zinc-hook region may play crucial roles in chromosomal transactions, provides us with tantalizing clues that suggest that Rad50 may share unanticipated similarities to the other canonical SMC-like proteins during chromosomal processing events. Considering the recent advances in our understanding of the modes of action of the canonical SMCs, it remains an intriguing possibility that this dynamic and enigmatic complex may utilize the ATPase activity of Rad50 activity to drive an energy-dependent search for DSBs. Interestingly, it has also been observed that the M/R complex in yeast appears to co-migrate along a chromosomal substrate with the Exo1 exonuclease, which is required for DNA end-resection [8,105]. It will therefore be important to explore in the future how the M/R machinery is involved in the recruitment of the long-range DNA end resection machinery. Perhaps energy-dependent chromosomal manipulation by the M/R complex may even facilitate the critical downstream processing of the DSB to produce the single-stranded tails required for strand invasion [43] (see Figure 5 for proposed models).

Given the inherent complexities associated with studying the dynamic SMC superfamily of proteins it has become clear that to completely ascertain the intricacies of the action of these proteins it will likely be necessary to employ the latest state-of-the-art single molecule approaches in combination with classical structural and biophysical approaches. Experimentally tractable thermophilic archaeal homologues, which exhibit the intrinsic catalytic activities, even in the absence of accessory factors such as CtIP, will no doubt play an important role in the future experimental pursuit of our understanding of these mechanisms. However, to ultimately comprehend the complicated chromosomal arrangements mediated by Rad50 and the other SMC family proteins it may be necessary to study these events in the context of the localized cellular environment. Indeed, it seems likely that the development of new technologies for analyzing chromosome organization within cells will doubtlessly be important for enhancing our understanding of the wide range of essential biological events mediated by the SMC superfamily of proteins.

• The Mre11/Rad50 complex is an essential DNA repair machinery used by all forms of life with homologues identified in the bacteria, archaea, and eukarya. The disparate catalytic functions of the Mre11 nuclease and the Rad50 ATPase (a member of the Structural Maintenance of Chromosomes (SMC) superfamily of ATPases that include the condensins and cohesins) work together to drive the DNA repair events that maintain chromosomal integrity.

• Several decades of research have provided critical insights into how the Mre11/Rad50 complex orchestrates these essential DNA repair events, although many questions remain still unanswered concerning the modes of action that this chromosomal maintenance machinery utilizes ensure the fidelity of the genetic code. There has been considerable emphasis on the study of the catalytic ‘head’ domains of the Rad50 ATPase, while the coordinated movements of associated long coil-coiled regions of that connect the ‘head’ and ‘zinc hooks’ (analogous to SMC hinge regions) of Rad50 are less well understood.

• Recent pioneering studies have advanced our understanding of how canonical SMC complexes, such as the condensins, orchestrate dramatic ATP-driven conformational changes within their coiled-coil regions, acting as molecular machines that physically extrude DNA loops. Similar modes of activity may be involved during DNA repair events mediated by the Mre11/Rad50 complex, perhaps by ‘turning the M/R complex on its head’, using analagous movements within the coiled-coiled regions to those that have recently been described for the canonical SMC proteins.

The authors declare that there are no competing interests associated with the manuscript.

The research performed in the N.P.R. and R.M.H. laboratories described in this review was funded by the Medical Research Council [Career Development Award G0701443] an Isaac Newton Trust Research Grant (Trinity College and Department of Biochemistry, Cambridge) and start-up funds from the Division of Biomedical and Life Sciences (Lancaster University) and a Biotechnology and Biological Sciences Research Council Grant BB/J018236/1, which provided funds for the purchase of the Bruker FastScan atomic force microscope. N.P.R. is currently funded by the Leverhulme Trust (RPG-2019-297).

Open access for this article was enabled by the participation of Lancaster University in an all-inclusive Read & Publish pilot with Portland Press and the Biochemical Society under a transformative agreement with JISC.

All authors were involved in the preparation of this review manuscript.

ABC

ATP binding cassette

AFM

atomic force microscopy

DSBs

double-strand breaks

HR

homologous recombination

MMEJ

micro-homology mediated end-joining

NHEJ

non-homologous end-joining

SMC

structural maintenance of chromosomes