The conserved Mre11-Rad50 complex is crucial for the detection, signaling, end tethering and processing of DNA double-strand breaks. While it is known that Mre11-Rad50 foci formation at DNA lesions accompanies repair, the underlying molecular assembly mechanisms and functional implications remained unclear. Combining pathway reconstitution in electron microscopy, biochemical assays and genetic studies, we show that S. cerevisiae Mre11-Rad50 with or without Xrs2 forms higher-order assemblies in solution and on DNA. Rad50 mediates such oligomerization, and mutations in a conserved Rad50 beta-sheet enhance or disrupt oligomerization. We demonstrate that Mre11-Rad50-Xrs2 oligomerization facilitates foci formation, DNA damage signaling, repair, and telomere maintenance in vivo. Mre11-Rad50 oligomerization does not affect its exonuclease activity but drives endonucleolytic cleavage at multiple sites on the 5′-DNA strand near double-strand breaks. Interestingly, mutations in the human RAD50 beta-sheet are linked to hereditary cancer predisposition and our findings might provide insights into their potential role in chemoresistance.
Misfunction of the MRE11-RAD50-NBS1 (MRN) complex is associated with several pathologies, including various hereditary and sporadic cancers1. Aberrant degradation of nascent DNA by MRE11 was observed in tumor suppressor BRCA1/2-deficient cells, and abrogation of MRE11 activity is linked to PARP inhibitor resistance in cancer chemotherapy2,3. At the molecular level, human MRN and yeast Mre11-Rad50-Xrs2 (MRX) promote DNA double-strand break (DSB) repair via non-homologous end joining (NHEJ) and homologous recombination (HR) by means of DNA tethering, activation of downstream pathway components, and ATP-dependent resection of the 5′-terminated strands at DSBs4. The latter function is particularly relevant for HR and a variant of the end-joining pathway termed alternative end-joining (A-EJ)4. The MRN/X complex also activates the DNA damage checkpoint kinase ATM/Tel1, which in turn phosphorylates targets to orchestrate cell cycle arrest and DNA repair5. Moreover, MRN/X and ATM/Tel1 contribute to the maintenance of normal telomere length5.
Nucleolytic processing of DSBs by MRN/X is restricted to the vicinity of the breaks, and hence termed short-range DNA end resection4,6,7. The activity of MRN/X is versatile, as it can process DSBs with various secondary DNA structures and protein blocks, both covalently bound, such as topoisomerase cleavage complexes, or non-covalently associated, such as the NHEJ factor Ku8,9,10,11. This preference for cleaving protein-blocked DNA ends by MR homologs is conserved in evolution from archaea and bacteria to yeast and humans12,13,14,15.
In budding yeast S. cerevisiae, Mre11 nuclease activity is essential for processing protein-blocked DNA ends. This activity can be bypassed in case of free DNA ends by long-range nucleases normally acting downstream, including Exo1 and Dna216. NBS1/Xrs2 is essential for resection in humans, but largely dispensable in yeast. It is, however, needed for Tel1 activation and nuclear import of the MRX complex15,17. Due to the requirement of a sister chromatid as homologous template for repair via HR, resection is limited to the S/G2 phase of the cell cycle, which is achieved through cell cycle-regulated stimulation of Mre11 endonuclease activity by phosphorylated Sae2 (pSae2)18,19. In vitro, MR first incises the 5′-terminated strand ~15–35 bp away from the DSB end with its endonuclease function, and then exonucleolytically degrades DNA backwards to the end in 3′−5′ direction9,12,15. In cells, Mre11-dependent resection can generate 3′-overhangs of up to ~300 nt in length20. It is unclear whether the short-range resection tracts are generated by DNA cleavage that proceeds stepwise away from the DNA end, as observed in vitro21, or whether MR in most cases first cleaves DNA at greater distances from the DNA end, as observed in meiotic cells20. The single-stranded overhangs are then elongated to more than 1000 nt by the long-range resection machinery including Exo1 or Dna2-Sgs1 for homology-directed repair7,20,22.
Upon DNA damage in eukaryotic cells, MR relocates from a homogeneous, nuclear distribution to discrete foci consisting of hundreds to thousands of MR molecules23,24,25,26,27,28. This localized MR accumulation then recruits a cascade of DNA repair and signaling proteins, including ATM/Tel125,26. Given that MR foci formation at DSBs and the resultant array of protein interactions are conserved in eukaryotes, the capacity of MR to quickly accumulate at DSBs in a regulated manner appears to be an essential characteristic of efficient DSB repair25.
MR was shown to dimerize to M2R2 complexes with one or two subunits of NBS1/Xrs2, whereby the Mre11 nuclease and Rad50 ATPase domains form a globular head from which two Rad50 coiled-coils protrude, culminating at a zinc hook at the apex of the structure29,30,31,32. DNA binding occurs mainly via the head domain of Mre11 and Rad50, but Xrs2 and the Rad50 coiled-coils were also shown to bind DNA27,29,33,34. Foci formation could be facilitated by higher-order oligomerization of MR dimers [(M2R2)n] via globular head domain interactions or by zinc hook-mediated tethering of the protruding Rad50 coiled-coils35,36. Although hints of head domain oligomerization have been observed for various species in vitro27,35,37, especially for eukaryotic MR, coiled-coil tethering was predominantly investigated and functionally associated with DNA movement, condensation, and end fixation in repair27,35,36. While high-resolution structures of the MR dimer34,36,38,39,40,41 from crystallography and cryo-electron microscopy contributed significantly to the understanding of the diverse MR functions, they provided only limited information on more heterogeneous higher-order structures formed by MR dimers at DSBs. Thus, the relevance of MR head domain interactions for higher-order oligomerization and for the appearance of clearly detectable MR foci upon DNA damage remains unclear. Taken together, defining the molecular mechanisms of foci assembly and identifying the interaction interfaces underlying MR oligomerization are of great importance for understanding MR-mediated functions and mechanisms in DNA damage signaling and repair.
Here we examine DNA end resection reactions by combining biochemical reconstitution and negative staining transmission electron microscopy (TEM), and thereby provide insights into the function and architecture of active DNA end resection complexes at the single molecule level. We demonstrate that short-range DNA end resection is driven by MR oligomerization mediated by MR head domain interactions, which stabilize MR on DNA and enhance its endonuclease activity at multiple sites along the 5′-terminated DNA strand. Mutational analysis shows that a small interface in the globular part of Rad50 modulates MR oligomerization and DNA end resection, as well as MR-dependent foci formation, DNA damage signaling, and telomere maintenance in S. cerevisiae. Together, these results identify MR oligomerization as a physiologically relevant process that is pre-requisite for efficient DSB repair and suggest a molecular mechanism for foci formation at DSBs.
MR oligomerizes via head domain and forms nucleolytically active assemblies on DNA
To understand the role and regulation of MR(X) oligomerization in DSB repair, we purified recombinant S. cerevisiae MR(X) from insect cells and visualized it with negative staining TEM (Fig. 1a and Supplementary Fig. 1a, b). Since Xrs2 is largely dispensable for DNA end resection17 and integrated with variable copy numbers into purified MRX complexes, we mainly analyzed MR to reduce complexity. In accord with previous studies of MR homologs29,35,36,38, we observed the formation of MR dimers (M2R2) with a globular head domain of ~11 nm in diameter and protruding ~50 nm long Rad50 coiled-coils. The coiled-coils were either open (Fig. 1a-i, ii and Supplementary Fig. 1b-i) or closed to a clamp via zinc hooks at the apex of the structure (Fig. 1a-iii and Supplementary Fig. 1b-ii). Additionally, M2R2 also displayed the capacity to oligomerize via its head domain, allowing the formation of tetramers (M2R2)2 and larger assemblies (M2R2)n with protruding coiled-coils that could then further interact via tethering (Fig. 1a-ii–iv; Supplementary Fig. 1b-i–iii). ATP binding stimulates MR association with DNA41,42, and MR endonuclease activity is magnesium- and manganese-dependent and requires ATP hydrolysis by Rad5012,19. Accordingly, under more physiological conditions as used for in vitro MR nuclease assays (30 °C with Mg2+, Mn2+, but ATPγS instead of ATP to prevent ATP hydrolysis and stabilize Rad50 interactions), the MR oligomers were even more evident (Fig. 1a-v and Supplementary Fig. 1b-iv). Higher temperature facilitated the formation of larger MR oligomers (compare 4 °C in Supplementary Fig. 1b-i–iii, 30 °C in Supplementary Fig. 1b-iv and 42 °C in Supplementary Fig. 1d-i). These MR oligomers were stable at physiological salt concentrations (150 mM KCl or NaCl) and only dissociated into MR dimers (M2R2) at high salt (260 mM KCl) (Supplementary Fig. 1c). Oligomer size decreased linearly with increasing KCl concentration, suggesting a charge screening effect of salt on electrostatic interactions between the MR head domains (100 nM MR, Supplementary Fig. 1c). However, higher temperature (42 °C) and increased MR concentration (600 nM) could overcome this charge screening effect at high salt conditions, restoring MR oligomerization (compare 260 mM KCl in Supplementary Fig. 1c-i with Supplementary Fig. 1d-ii). Together, the high local MR concentrations at DSBs and the growth temperature of yeast may thus counteract the charge screening effect of chromatin DNA43, and allow MR oligomerization under physiological conditions.
To examine the effect of DNA on MR oligomerization, we investigated MR assembly on linear and circular plasmid DNA. At lower MR:DNA ratios (50–100 nM MR, 10–20:1 with respect to DNA molecules), the MR oligomers loaded onto DNA and dissociated into dimers that bound DNA ends (Fig. 1b-i) or internally (Fig. 1b-ii, iii and Supplementary Fig. 1e, f). A few isolated MR dimers or pairs of MR dimers bound adjacently on DNA were visible (Fig. 1b, pink and orange arrows denote examples). The majority of MR molecules, however, were non-randomly distributed and oligomerized on DNA (Fig. 1b-ii, iii; red arrows). Some of these MR molecules appeared to form “pearls-on-a-string”-like assemblies on DNA, seemingly interacting with each other through the head domains, oftentimes with DNA loops in between (Fig. 1b-ii, iii zoom-insets and cartoons). The dense accumulation of MR units at some DNA regions contrasted with other DNA stretches that were left almost protein-free, as observed both in the presence and absence of DNA ends with ATP or ATPγS (Fig. 1b-ii, iii and Supplementary Fig. 1e, f). The detected DNA loops and substrate compaction suggest that the head domain interactions and subsequent Rad50 tethering of MR oligomers can bring distant parts of DNA into close proximity, as reported previously for archaeal MR27. In electrophoretic mobility shift assays (EMSAs) we observed a distinct banded pattern at low MR concentrations (up to 100 nM), followed by a rapid shift of the protein-DNA complexes to the wells in a narrow MR concentration range (from 200 to 400 nM) (Fig. 1c-i and Supplementary Fig. 1g). This switch-like binding behavior of MR was comparable on linear and circular DNA, irrespective of free or protein-blocked ends or DNA supercoiling (Fig. 1c-ii and Supplementary Fig. 1g), and was suggestive of a stabilization of MR units bound to DNA as a protein cluster. Limited binding preference of MR to DNA ends was observed previously, as Xrs2, albeit not essential for resection in vivo, facilitates end-binding of the MRX complex17,33.
To ascertain that MR oligomers are active entities, we performed in vitro nuclease assays21 in the presence of phosphorylated Sae2 (pSae2) to trigger MR endonuclease activity (Fig. 1d-i). Resection was monitored with a radio-labeled oligonucleotide annealing to the produced 3′-overhang. Efficient DNA end resection was observed at the MR concentrations that led to the binding of multiple MR entities on DNA in EMSA assays (Fig. 1c-i and Supplementary Fig. 1g). Interestingly, negative staining TEM analysis of the MR:DNA ratio with the highest resection activity in the nuclease assay (200 nM MR with 2.5 nM DNA, MR:DNA ratio 80:1; Fig. 1d-i) revealed that the MR dimers formed large oligomers on DNA, in which the individual MR dimers could no longer be easily distinguished (Fig. 1d-ii), in contrast to the assemblies with distinguishable MR dimers on DNA observed at lower MR:DNA ratios (Fig. 1b-ii, iii and Supplementary Fig. 1e). The combination of imaging and biochemical experiments demonstrates that the observed MR oligomers are not inactive assemblies, but rather functional units that allow resection to proceed through multiple endonucleolytic cuts beyond the vicinity of DNA ends, as observed in vivo6,20.
A high MR:DNA ratio (60:1; Supplementary Fig. 1h) also stabilized MR on DNA against higher salt concentration, thus counteracting the expected salt sensitivity of MR-DNA association34. Moreover, MR preferentially bound longer DNA substrates, as shown by the trend of apparent dissociation constants (KD) derived from EMSA experiments with oligonucleotides, which decreased with increasing DNA length (Fig. 1e and Supplementary Fig. 2a). EMSA experiments also revealed that longer DNA molecules stabilized MR on DNA against high salt concentrations (Supplementary Fig. 2b-d), similarly to higher MR:DNA ratios (Supplementary Fig. 1h). As MR-DNA species remaining in the wells may bias the KD values, we performed mass photometry, in which the large MR-DNA assemblies stay in solution. These measurements confirmed better MR binding to longer oligonucleotides, as the peak area between the two dashed lines, corresponding to one MR dimer binding to one DNA molecule, grew with increasing DNA length (Fig. 1f; theoretical molecular weight of M2R2 is 460 kDa). Moreover, a higher molecular weight peak at ~1000 kDa could be detected and properly fitted for longer oligonucleotides, corresponding to two MR dimers bound per DNA molecule (zoom-insets for 106 and 197 bp). In addition, larger MR oligomers were likely formed on longer DNA that were beyond the detection limit of the instrument44, as indicated by the sudden decrease in the observed MR-DNA peak areas with 106 or 197 bp, compared to 73 bp-long DNA (Fig. 1f). Taken together, these observations suggest that MR dimers form oligomers in solution, which dissociate to bind DNA and reassemble in a switch-like manner to larger MR assemblies on DNA (Fig. 1g). As expected, purified MRX behaved comparably in TEM, EMSAs, and mass photometry experiments (Supplementary Fig. 3a–e), demonstrating that Xrs2 binding does not diminish MR oligomerization. MR(X) oligomerization may allow a rapid loading of multiple MR(X) units onto DNA, thereby facilitating short-range resection of DNA ends and foci formation.
MR oligomerization on DNA accompanies end resection
Next, we reconstituted short-range (MR and pSae2) and long-range (Exo1) resection in vitro using recombinant S. cerevisiae proteins and linear plasmid DNA with free or streptavidin-blocked ends. Gel-based assays confirmed that DNA with free ends could be rapidly processed by Exo1 alone, as visualized by the characteristic pattern of resection products, depending on the extent of resected DNA and whether resection was initiated at one or both DNA ends (Fig. 2a, lane 2, see cartoons on the right; Supplementary Fig. 4a). Adding MR-pSae2 to the Exo1 reaction moderately enhanced resection (Fig. 2a, compare lane 2 with lanes 4, 6 and 8 and Supplementary Fig. 4a), as observed previously21,45,46. In contrast, MR-pSae2 alone resulted in slow resection that was limited to the vicinity of DNA ends (Fig. 2a, lanes 3, 5 and 7 and Supplementary Fig. 4a)6,21. As expected, long-range resection by Exo1 alone was strongly reduced on linear DNA with streptavidin-blocked ends (Fig. 2b, lane 2 and Supplementary Fig. 4a, b), but this activity could be restored by the addition of MR-pSae2 (Fig. 2b, lanes 4, 6 and 8 and Supplementary Fig. 4a). Thereby, MR-pSae2 was able to overcome the protein block and resect blocked DNA ends similarly to free DNA (Fig. 2b, lanes 3, 5 and 7 and Supplementary Fig. 4a), creating an entry site for Exo1-mediated long-range resection. Using shorter, 3′-labeled oligonucleotide-based DNA with one streptavidin-block at the 5′-terminated end revealed a gradual shortening of the 5′-strand by MR-pSae2 and the generation of DNA fragments of similar length (Fig. 2c). This indicates the sequential formation of several endonucleolytic incisions and the stepwise DNA end resection by many MR-pSae2 molecules that together produce the single-stranded overhang observed21. MR oligomerization may promote such joint DNA end resection21. Together, our reconstituted assays demonstrate that Exo1 requires MR-pSae2-dependent short-range resection at protein-blocked DNA ends to initiate long-range resection.
We next set out to visualize the architecture of DNA end resection complexes, using the same recombinant proteins and reaction conditions, in negative staining TEM. As shown in Fig. 2d, we readily observed MR oligomers in conjunction with pSae2 bound to DNA with free or streptavidin-blocked ends (Fig. 2d-i, II, III, Fig. 2d-iii, II and Supplementary Fig. 4c; magenta arrows and MR-pSae2 labels). While free ends of linear DNA were not clearly visible and likely buried within the MR-pSae2 clusters (Fig. 2d-i, II), streptavidin-blocked ends of linear DNA could be easily distinguished (Fig. 2d-iii,II, dark blue arrows). MR-pSae2 assemblies could bind DNA at internal sites and harbor several DNA molecules, as apparent from the number of streptavidin-blocked ends in particular protein-DNA complexes (Fig. 2d-iii, II, left panel, dark blue arrows), reminiscent of the assembly of repair foci involving several broken DNA molecules in vivo25,47. Interestingly, resection of DNA ends was observed in TEM independently of protein blocks, producing up to ~300 nt-long stretches of single-stranded DNA, consistent with cellular6,20 and previous biochemical data21. These single-stranded tracts were readily coated by RPA, thus appearing more electron-dense than double-stranded DNA (Fig. 2d-i, III, Fig. 2d-iii, II, right panel; arrows in light blue). In accordance with the biochemical assay in Fig. 2a, MR-pSae2 activity was not required to trigger Exo1-dependent long-range resection on DNA templates with free ends, as visualized by the appearance of plasmid-length RPA-coated, single-stranded DNA upon incubation with Exo1 alone (Fig. 2d-ii, II, arrows in light blue). In contrast, MR-pSae2-dependent end processing was required for subsequent tail elongation by Exo1 when using DNA substrates with blocked ends (Fig. 2d-iii, II–IV). Accordingly, no RPA-coated, single-stranded DNA overhangs were observed when imaging streptavidin-blocked DNA with Exo1 without prior short-range resection by MR-pSae2 to remove the protein block (Fig. 2d-iii, I). Taken together, these single-molecule assays recapitulated that short-range resection by MR-pSae2 is essential for initial processing of protein-blocked ends to allow handover to the Exo1 exonuclease for subsequent long-range resection. Strikingly, our TEM analysis visualized large MR-pSae2 assemblies at DNA ends undergoing resection.
Mutagenesis of conserved Rad50 interface regulates MR oligomerization
To define the functional relevance of MR oligomerization on DNA, we next set out to identify the interaction interface(s) between the MR head domains. As a basis for mutational analysis, we employed a recent crystallographic structure showing a eukaryotic C. thermophilum (Ct) Rad50 dimer bound to a DNA molecule that forms a “quasi-continuous” structure with another Rad50 dimer-bound DNA molecule adjacent in the crystal, indicating a possible arrangement of two Rad50 dimers on DNA (Fig. 3a; RCSB Protein Data Bank PDB identifier: 5DAC48, https://doi.org/10.2210/pdb5dac/pdb). We hypothesized that an anti-parallel, two-stranded beta-sheet (black rectangle, β) and its connecting loop (black rectangle, L) protruding from the head domain of one Rad50 dimer in one asymmetric unit towards the identical structure of another Rad50 dimer in the adjacent asymmetric unit of the crystal could represent a possible interaction site, especially on a continuous DNA molecule and in solution when MR molecules could come even closer on DNA than in the crystal structure shown. The residues constituting this small beta-sheet and its connecting loop, or their physicochemical characteristics, are conserved between C. thermophilum, S. cerevisiae, and H. sapiens Rad50 (Fig. 3b), which is intriguing given the limited overall sequence similarity of MR between species49. Interestingly, mutations derived from cancer predisposed individuals (* in Fig. 3b, from ClinVar/MedGen databases) cluster in this conserved beta-sheet and could be relevant for Rad50 functions in DNA repair and cancer avoidance.
To investigate whether beta-sheet interactions are indeed involved in MR oligomerization, we stained S. cerevisiae (Sc) MR oligomers with Thioflavin T (ThT), a dye used to detect beta-sheet-rich structures50. Indeed, a ThT signal was detected in the presence of ScMR oligomers at 4 °C and enhanced at higher temperature (42 °C), suggesting the presence of extended beta-sheet interactions between MR molecules (Fig. 3c). In order to test the possible role of the identified protruding beta-sheet and its connecting loop in MR oligomerization, we engineered two ScRad50 variants, each with several point mutations in the beta-sheet to disrupt either electrostatic or hydrophobic interactions between the Rad50 subunits of MR dimers. Addressing the electrostatic component of MR oligomerization (Supplementary Fig. 1c), we substituted a polar residue in the potential interaction interface with an uncharged, non-polar alanine (Sc N121A, Ct N122) and removed a negative charge (Sc D124, Ct E125) to disrupt an attractive salt bridge between the D124 and R125 residues (Ct E125, R126) or to alleviate a potentially strong repulsive force between the aspartate residues D124 (Ct E125) of interacting ScRad50 molecules (red residues mutated in Fig. 3a zoom-inset and Fig. 3b; Rad50ho). In the second ScRad50 variant, four hydrophobic beta-sheet residues were mutated to alanine to disrupt or destabilize the beta-sheet structure51 (ScRad50L116A/I119A/T127A/L128A; CtRad50L118/L120/T127/I129; blue residues mutated in Fig. 3a zoom-inset and Fig. 3b; Rad50lo). These two ScRad50 mutants were co-expressed and co-purified with Mre11, showing that the physical interaction between Mre11 and Rad50 was not notably affected (Supplementary Fig. 5a).
We first characterized the oligomerization capacity of the ScMR variants in negative staining TEM in the absence of DNA. This analysis revealed that the quadruple mutant MRL116A/I119A/T127A/L128A formed only dimers (thus henceforth denoted low oligomerization mutant, MRlo), while the MRN121A/D124N mutant (henceforth denoted high oligomerization mutant, MRho) assembled to even larger oligomers than wild-type MR (MRwt) (Fig. 3d). Mass photometry confirmed that the MRho mutant formed larger oligomers than MRwt, as in the absence of DNA only a small dimeric fraction of MRho could be detected (−DNA, zoom-inset ~460 kDa; Fig. 3e and Supplementary Fig. 5b). The MRho mutant was also more resistant towards high salt (240 mM KCl), as significantly more MRho oligomer assemblies could be harvested in a pelleting assay compared to MRwt after incubation at 1.2 µM MR and higher temperature (42 °C) (42 °C pellet; Fig. 3f and Supplementary Fig. 5c). Consistently, mass photometry detected less MRho dimers at ~460 kDa after 30 min incubation at 250 mM KCl and 30 °C compared to MRwt, suggesting the formation of larger MRho assemblies beyond the detection limit (Supplementary Fig. 5d). Similar to MRwt, MRho was stabilized on longer oligonucleotide substrates (Figs. 1f and 3e). However, MRho formed larger oligomers on longer oligonucleotide DNA compared to MRwt, as the peak at ~1000 kDa of two MRho dimers binding to one DNA molecule could be detected and properly fitted already for 73 bp with a larger peak area than for MRwt (compare 73 bp zoom-insets in Figs. 1f and 3e). Moreover, the areas of the MR-DNA peaks decreased more with 106 and 197 bp DNA for MRho than MRwt, indicating the formation of larger MRho nucleoprotein assemblies beyond the detection limit (compare peak areas for 106 and 197 bp in Figs. 1f and 3e). For MRlo, on the other hand, binding to oligonucleotide DNA could not be observed in mass photometry at the low MR and DNA concentrations required for measuring (197 bp in Supplementary Fig. 5b).
MR oligomerization influences DNA binding behavior and endonuclease activity
EMSA experiments revealed a moderate DNA binding impairment of the MRlo mutant compared to MRwt (Fig. 4a and Supplementary Figs. 2a and 5f); however, at ~90 nM, MRlo efficiently bound the 100 bp-long DNA substrate. Similarly, 90 nM MRlo was also bound to linear plasmid-length DNA (35 mM KCl, MRlo:DNA 36:1) in negative staining TEM, appearing as dispersed single dimers (Fig. 4b). Surprisingly, the DNA binding of the MRho mutant was also reduced in EMSA experiments (Fig. 4a and Supplementary Figs. 2a and 5f), indicating that higher oligomeric structures may interfere with disassembly to smaller MRho complexes including dimers and thus efficient loading onto DNA. Mass photometry measurements at low salt (25 mM KCl, zoom-insets in Supplementary Fig. 5g) revealed that once MRho loaded onto plasmid DNA, it assembled as one or maximally two MRho dimers per DNA molecule (in 1500–2500 kDa range, zoom-inset in Supplementary Fig. 5g) and then rapidly formed large nucleoprotein assemblies beyond the detection limit. Thus, the tail of events >2500 kDa, corresponding to smaller nucleoprotein assemblies, was significantly reduced for MRho compared to MRwt (25 mM KCl, zoom-insets in Supplementary Fig. 5g), and even more so after incubation (25 mM KCl, 30 min, zoom-insets in Supplementary Fig. 5g). Moreover, TEM analysis demonstrated that MRho assembled into defined, condensed “worm”-like structures once it bound plasmid DNA (90 nM MRho, 35 mM KCl, MRho:DNA 36:1; Fig. 4b), while MRwt formed much smaller oligomers at comparable conditions (100 nM MRwt; Fig. 1b-ii and Supplementary Fig. 1e). We next expressed the Mre11wt and Rad50wt/ho/lo subunits individually and observed that Mre11wt alone did not oligomerize on DNA (Fig. 4c). In contrast, oligomerization of the Rad50wt/ho/lo subunits recapitulated the behavior of the respective MR heterodimers (Figs. 1b-ii and 4b, c): Rad50ho formed larger oligomeric assemblies on DNA compared to Rad50wt, while Rad50lo bound DNA dispersedly. Taken together, these results suggest that Rad50 is necessary and sufficient to drive oligomerization of the MR complex, most likely through a conserved beta-sheet motif.
We next performed clipping assays using oligonucleotide-based DNA with blocked ends to assess if MR oligomerization affects its endonuclease activity in conjunction with pSae2. Interestingly, the MRlo mutant was entirely endonuclease-deficient, even when used at a concentration (200 nM) that supports DNA binding (Fig. 4d, e and Supplementary Fig. 6a). Conversely, MRho was slightly more active than MRwt (Fig. 4d, e), which was particularly apparent upon short incubation times during kinetic experiments with oligonucleotide-based DNA (Fig. 4f, g). Using plasmid-length substrates, we observed that conditions resulting in large “worm”-like assemblies support DNA end resection (90–100 nM MRho, Fig. 4b, h). A further increase of the MRho concentration to 200 nM additionally enhanced resection activity (Fig. 4h), suggesting that the assemblies are sites of efficient DNA end resection. The initial resection by MRho could efficiently be extended by Exo1 (Supplementary Fig. 6b–d), similarly to MRwt (Fig. 2a, b), indicating that the large protein assembly does not interfere with downstream long-range resection.
Notably, both MRlo and MRho were proficient as exonucleases and functionally interacted with ATP (Supplementary Fig. 6e). Moreover, both MR variants were able to bind pSae2 (Fig. 4i), as shown in a protein interaction assay using the conserved C-terminal part of Sae2 (pSae2 \(\triangle\)N169)19. These experiments demonstrate that although endonuclease-deficient, the Rad50lo variant is clearly distinct from the Rad50S mutants, which fail to physically and functionally interact with Sae219,52. Taken together, these data suggest that the conserved beta-sheet motif in Rad50 regulates MR oligomerization, and that oligomerization in turn promotes MR resection activity.
Foci formation and DNA repair require MRX oligomerization in vivo
In order to probe the function of MRX oligomerization in vivo, we introduced Rad50wt or the Rad50ho/lo mutants in an S. cerevisiae Rad50 deletion (rad50\(\triangle\)) strain together with a C-terminal yEVenus- tag (rad50wt-yEVenus, rad50ho: rad50N121A/D124N-yEVenus, rad50lo: rad50L116A/I119A/T127A/L128A-yEVenus), while Mre11 and Xrs2 were unaltered. DNA damage sensitivity assays with the topoisomerase inhibitor and DSB-inducer camptothecin (CPT), as well as the replication inhibitor hydroxyurea (HU), revealed that cells expressing Rad50ho coped with DNA lesions similarly to wild-type, while cells with Rad50lo displayed a phenotype comparable to rad50\(\triangle\) (Fig. 5a). This defect was recapitulated by proliferation assays with 5 or 25 µM CPT: whereas the doubling times of wild-type and Rad50ho-expressing cells only marginally increased upon CPT-treatment, cells with Rad50lo or rad50\(\triangle\) were more sensitive to DNA damage, especially at higher CPT concentrations (Fig. 5b). We then engineered wild-type yeast strains overexpressing untagged Rad50wt/ho/lo from the galactose-inducible GAL1-promoter (Fig. 5c, d). We found that overexpression of Rad50lo on top of endogenous Rad50wt results in a pronounced growth defect and DNA damage sensitivity (Fig. 5d). Although we cannot exclude that the introduced mutations affect the endonuclease activity by altering the local structure, these results suggest that Rad50lo retains the capacity to act in a dominant-negative manner, most likely by inhibiting oligomerization of the wild-type MRX pool.
Live-cell imaging of yEVenus-tagged Rad50 variants revealed the formation of small Rad50wt/ho foci at potential DNA damage sites in the nucleus of CPT-treated cells (arrows, Fig. 5e). Interestingly, the Rad50lo mutant instead failed to assemble distinguishable foci (Fig. 5e), although we cannot exclude that its slightly reduced protein levels contribute to this defect (Fig. 5f). Conversely, Rad50ho formed foci even in the absence of CPT-induced DNA damage (Fig. 5e, g and Supplementary Fig. 7a). These data support the notion that MRX oligomerization promotes stable DNA binding required for the assembly of large repair structures detectable as foci in vivo.
DNA damage checkpoint activation and telomere maintenance
Upon DNA damage, the checkpoint-kinase Tel1 is recruited by MRX to repair foci at DNA lesions and phosphorylates target proteins including Rad53, which delays cell cycle progression until the DNA damage is repaired5. A single unrepaired DSB is sufficient to trigger the DNA damage checkpoint in yeast53. We hypothesized that MRX oligomerization at DNA lesions could have a function in amplifying the checkpoint signal. Indeed, the phosphorylation-mediated mobility shift of Rad53 upon DNA damage was reduced in cells expressing untagged Rad50lo compared to Rad50wt or Rad50ho in western blotting; however, the limited Rad53 phosphorylation was still sufficient to allow full arrest of Rad50lo-expressing cells after HU treatment, as demonstrated by flow cytometry (Fig. 6a and Supplementary Fig. 7b). MRX and Tel1 are, however, also required for the maintenance of telomere length — a function that does not require the nuclease activity of the MRX complex54. A characteristic feature of rad50\(\triangle\) cells are thus shortened telomeres (Fig. 6b). Telomeric PCR experiments revealed that yeast cells carrying untagged Rad50lo suffered from short telomeres similarly to rad50\(\triangle\), while cells with Rad50ho maintained normal telomere length (Fig. 6b, c and Supplementary Fig. 7c). Taken together, these data suggest that MRX oligomerization at DNA damage sites, governed by the conserved Rad50 beta-sheet motif, facilitates the assembly of protein foci, which is required for the DNA repair, signaling, and telomere maintenance functions of MRX in vivo.
Here, we show that S. cerevisiae MR(X) dimers oligomerize via the head domain of Rad50 to nucleolytically active assemblies at DNA ends. The formation of these higher-order MR complexes is governed by a conserved beta-sheet motif in the Rad50 head domain. Mutational analysis of the beta-sheet motif identified specific mutations that preserve MR dimerization but decrease or enhance MR head domain oligomerization in vitro. Biochemical reconstitutions coupled with TEM imaging revealed that MR oligomerization is required for pSae2-stimulated endonucleolytic DNA cleavage and hence short-range resection but is dispensable for MR exonuclease activity. Subsequent analysis of the Rad50 mutants in vivo confirmed the importance of MRX oligomerization for its cellular functions in DNA repair, signaling, and telomere maintenance.
Using reconstituted biochemical reactions, we could visualize intermediates of DNA end resection in vitro. Our biochemical assays coupled with TEM analysis recapitulated MR-pSae2-dependent short-range processing and subsequent tail elongation by Exo1. TEM imaging revealed that oligomerization of MR and pSae2 complexes to higher-order assemblies can occur internally of DNA, but resection is restricted to DNA ends. By sliding along DNA as previously shown55, MR could find and accumulate at the DNA end, where the MR endonuclease is locally activated by pSae2. The resulting short-range processing of blocked ends is prerequisite for subsequent handover to Exo1-mediated long-range resection. After cleavage, MR-pSae2 assemblies dissociate from RPA-coated single-stranded DNA ends to allow subsequent steps in the DSB repair pathway.
Previous work showed that an initial endonucleolytic cut by MRX-pSae2 is positioned 15–35 nt away from the DNA end when using oligonucleotide-based DNA12. On longer DNA templates, however, additional MRX-pSae2 incisions were found that led to the resection of several hundred nucleotides in length21, consistent with observations from meiotic cells lacking the long-range resection pathways6,20. Accordingly, a stepwise DNA cleavage model for 5′-DNA end resection by MRX-Sae2 was proposed21. Here, our experiments revealed that MR molecules form large assemblies, which provide a structural basis to explain the multiple endonucleolytic cuts along the 5′-terminated DNA strand during short-range resection.
Mutations in a protruding beta-sheet in the Rad50 subunit allowed us to construct two MR mutants, one with enhanced (MRho) and one with reduced oligomerization (MRlo). Although available results demonstrate that MR oligomerization is dependent on the Rad50 head domain and beta-sheet interactions, the resolution of the TEM images does not allow to prove that the beta-sheet motif directly mediates this effect. Nevertheless, MRho shows slightly increased resection activity compared to MRwt, and enhanced foci formation in living cells. As MRho displays enhanced oligomerization despite the loss of an attractive salt bridge between Sc D124 and R125 (Ct E125 and R126) due to the D124N mutation, we propose that the abolished charge repulsion of D124N residues between MRho molecules could drive this behavior. Although in the crystal structure model with two separate and only “quasi-continuous” DNA molecules (Fig. 3a) the Ct E125 (Sc D124) residues of adjacent Rad50 dimers are ~5 Å apart, they may be closer on a continuous DNA molecule in solution. Moreover, the hydrophobic residues in the beta-sheet could come closer when charge interactions are alleviated, which could explain the lower salt-sensitivity of MRho oligomers and the more pronounced beta-sheet signal in ThT staining. Yet, other head domain interactions are possible and high-resolution structures of MR dimers adjacently bound on DNA will ultimately be needed to demonstrate the exact residues mediating beta-sheet-driven MR oligomerization.
In contrast to MRwt and MRho, MRlo was endonuclease-deficient in vitro and failed to form DNA damage-induced nuclear foci in vivo, resulting in cellular sensitivity to DNA damage, an altered checkpoint response and short telomeres. Moreover, Rad50lo acts as a dominant-negative factor on endogenous Rad50wt in overexpression experiments in yeast cells. Although we cannot exclude that the mutations in MRlo negatively affect MRlo activity beyond disrupting oligomerization, we note that oligomerization accompanies resection in MRwt, and that the Rad50lo mutant still binds pSae2, Mre11, and Xrs2 that mediates the observed nuclear localization in cells17. Moreover, MRlo shows normal levels of exonuclease activity, which is inhibited by ATPγS, as observed for wild-type MR21, indicating that the nuclease active site of MRlo is sufficiently intact, given that the endo- and exonuclease activities of MR are mediated through the same active site56. The defects of MRlo are also unlikely explained with reduced DNA binding, as increased MRlo concentrations restored DNA binding but not its enzymatic activity. In addition, MRho was more active as endonuclease than MRwt, despite being impaired in DNA binding as well. We thus propose that MR oligomerization on DNA might facilitate the drastic conformational changes required for nuclease activation38,57, and promote activation of adjacent MR molecules on the 5′-terminated DNA strand, thereby creating a “domino-like” effect close to the DNA ends that explains step-wise short-range resection.
Repair foci with hundreds to thousands of MRX/N molecules were found at DNA breaks in yeast and mammalian cells23,24,25,26,27,28. ChIP-seq data of human MRN show an accumulation of the complex across a DSB on both sides of broken DNA, covering ~1 kb of DNA, with the strongest MRN localization at the DNA ends, indicating MRN focus formation around a DSB and tethering of broken ends58. In accord, our TEM analysis revealed that MR clusters can hold several DNA molecules together, mediated by oligomerization of the MR head domains and interactions via the zinc hooks. We thus propose that MR head domain oligomerization together with DNA end tethering may explain the coordinated resection observed at two-ended DNA breaks59. Consistent with this interpretation, no nuclear foci were detected upon DNA damage for yEVenus-tagged oligomerization-defective MRloX in S. cerevisiae, while the hyper-oligomerizing MRhoX mutant formed nuclear foci even in the absence of exogenous DNA lesions. Moreover, pelleting and ThT assays detected stable MR assemblies, which could build a rather rigid core inside potentially liquid-like repair foci, as formed, e.g., by the NHEJ-factor 53BP160 or the strand-exchange protein Rad5261, and act as scaffold for transiently joining downstream repair and signaling components including pSae2/pCtIP and Tel1/ATM.
Interestingly, database (ClinVar/MedGen) and literature reports62,63 suggest that human individuals with a hereditary predisposition for BRCA1/2- and PALB2-associated cancers carry germline point mutations of predominantly hydrophobic residues in the human RAD50 beta-sheet motif. The tumor suppressor BRCA1 interacts with PALB2 and BRCA2, and localizes to repair foci at DNA lesions similarly to MR64,65,66. Loss of BRCA1/2 function leads to aberrant degradation of nascent DNA by MRE11 upon replication stress, and a dramatic increase in the lifetime risk for ovarian and breast cancer3,65,66,67. The abrogation of MRE11 activity in BRCA1/2-deficient cells, on the other hand, was linked to PARP inhibitor resistance in cancer chemotherapy2. It is intriguing that MRlo, bearing beta-sheet-disturbing alanine mutations of four hydrophobic residues altered in cancer predisposed individuals, shows impaired endonuclease activity in vitro. These loss-of-function RAD50 mutations could restrict MRE11-mediated degradation of nascent DNA at challenged replication forks in PARP-resistant, BRCA1/2-PALB2-mutated cells, and may help to explain tumor chemoresistance. However, the importance of these RAD50 mutations for cancer risk and survival is not well understood and thus more extensive studies are needed.
Preparation of recombinant proteins
Yeast Mre11-Rad50 complexes were co-expressed in S. frugiperda 9 (Sf9) cells with an optimal ratio of baculoviruses producing Mre11wt-6xHis and untagged Rad50wt, Rad50ho, or Rad50lo that was codon-optimized for Sf9 expression to achieve 1:1 MR subunit stoichiometry. MR complexes were purified by NiNTA affinity and ion-exchange chromatography17. MRlo was expressed at considerably lower amounts compared to MRwt or MRho, indicating that MR oligomerization may stabilize the complex. The MRX complex was expressed using vectors for His-tagged Mre11, untagged Rad50, and FLAG-tagged Xrs2 and purified via NiNTA and FLAG affinity chromatography45. Recombinant Mre11 was purified by amylose and NiNTA affinity chromatography, while Xrs2 and Rad50wt/ho/lo were purified with FLAG affinity resin21. Exo1 was purified by FLAG affinity and ion-exchange chromatography45. Recombinant phosphorylated Sae2 (pSae2) was purified from Sf9 cells by amylose and NiNTA affinity chromatography in the presence of phosphatase inhibitors19. Yeast RPA was expressed from the p11d-scRPA vector (kind gift from M. Wold, University of Iowa) in E. coli (BL21 DE3 pLysS) and purified as described for human RPA15.
Oligonucleotides used in this study are listed in Supplementary Table 1. To prepare the DNA substrates (20, 31, 70, 100, and 197 bp) for DNA binding and endonuclease experiments, the respective oligonucleotides were labeled at the 3′-end with [α−32P]dCTP (Perkin Elmer) and terminal transferase (New England BioLabs) following the manufacturer’s recommendations. The enzyme was inactivated, and the reaction purified on Micro Bio-Spin P-30 Gel Columns (Bio-Rad). The labeled oligonucleotide was then annealed with twofold excess of the related unlabeled oligonucleotide in 1× PNK buffer (New England BioLabs). The 5′-labeled 70 bp substrate for exonuclease assays was prepared similarly, except that the PC210 oligonucleotide was labeled at the 5′-end with [γ−32P]ATP (PerkinElmer) and T4 polynucleotide kinase (New England BioLabs). Oligonucleotides for hot annealing (5′_OligoA and 3′_OligoA) were labeled at the 5′-end with [γ−32P]ATP (PerkinElmer) and T4 polynucleotide kinase (New England BioLabs) following the manufacturer’s recommendations. For mass photometry, oligonucleotides of 20, 50, 73, 106, and 197 bp (Supplementary Table 1) and 2.7 kbp-long pUC19 were used. For TEM, circular 5.3 kbp-long pFBDM (Addgene; Fig. 1b-iii and Supplementary Fig. 1f), linear 2.7 kbp-long pUC19 with free 4 nt-overhangs or a 2.8 kbp-long, pUC19-derived linear plasmid with streptavidin-blocked ends were used. Where indicated, circular, relaxed or linear 2.7 kbp-long pUC19 with free 4 nt-overhangs, and a 2.8 kbp-long, pUC19-derived blunt linear plasmid with streptavidin-blocked ends were used for EMSAs and nuclease assays. The circular, supercoiled form of pUC19 was relaxed with Topoisomerase I from E. coli (New England Biolabs) and purified using a MinElute PCR purification kit (Qiagen). Linearized pUC19 was prepared with restriction digestion using KpnI-HF (New England BioLabs) and 0.8% agarose-gel purification. The 2.8 kbp-long, pUC19-derived linear plasmid substrate with biotin at the ends was prepared using pAttP-S vector, annealed PC210 and PC211 oligonucleotides and ΦC31 integrase68. After incubation with monovalent streptavidin (kind gift from M. Howarth, University of Oxford), streptavidin-blocked 2.8 kbp-long pUC19 was purified with Chromaspin TE-200 columns (Takara). For Telomeric PCR, the 100 µM stocks of oligonucleotides listed in Supplementary Table 1 were diluted in sterile water to a working concentration of 10 µM. oBL359 was purchased from Microsynth; oBL358, oBL360, and oBL361 from Merck Millipore.
Negative staining TEM
Purified proteins were mixed at the indicated concentrations with or without plasmid substrates to 5, 10, or 15 µl reactions on ice in a reaction buffer containing 25 mM Tris acetate pH 7.5, 1 mM dithiothreitol (DTT) and the co-factors 5 mM magnesium acetate (Sigma-Aldrich, M5661), 1 mM manganese chloride (Sigma-Aldrich, M3634) and 1 mM ATPγS (Sigma-Aldrich, A1388), unless indicated otherwise. In Supplementary Fig. 1f-ii and for resection reactions in Fig. 2d, 1 mM ATP (Sigma-Aldrich, A2383) was used instead of ATPγS. KCl in phosphate buffer (storage buffer of MR; 20 mM phosphate solution pH 7.4 [KH2PO4, K2HPO4], 10% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF], 1 mM DTT) was added to adjust salt concentrations as indicated. Assessment of oligomer size in Figs. 1a and 3d was performed at 120 mM KCl. For MRwt/ho/lo/MRwt-pSae2/Mre11wt/Rad50wt/ho/lo binding to plasmid DNA, 30 mM KCl was used for Fig. 1b and Supplementary Fig. 1e, f; 80 mM KCl for Fig. 1d-ii; 35 mM KCl for Fig. 4b; 50 mM KCl for Fig. 4c; 40 mM KCl for Supplementary Fig. 4c; otherwise specified in the Figure itself. The reactions were incubated at 30 °C for 30 min, unless indicated otherwise.
For Fig. 2d-i, I and 2d-ii, I, 5 nM 2.7 kbp-long, linear plasmid DNA (in molecules) was incubated for 2.5 h with 75 nM RPA at 30 °C and 40 mM KCl. For Fig. 2d-i, II, III, 50 nM MRwt and 200 nM pSae2 were added and incubated for 2.5 h at 30 °C with 40 mM KCl. For Fig. 2d-ii, II, 5 nM 2.7 kbp-long, linear plasmid DNA (in molecules) was incubated with 50 nM Exo1 and 200 nM RPA for 5 min at 30 °C and 50 mM KCl. For Fig. 2d-iii, I, 5 nM 2.8 kbp-long, linear plasmid DNA (in molecules) carrying two streptavidin blocks per end was incubated with 50 nM Exo1 and 75 nM RPA for 2.5 h at 30 °C and 40 mM KCl. For Fig. 2d-iii, II, 50 nM MRwt and 200 nM pSae2 were incubated with 5 nM of the substrate (in molecules) for 2.5 h at 30 °C and 40 mM KCl. For Fig. 2d-iii, III, IV, 50 nM Exo1 and RPA to a final concentration of 670 nM were added to the reaction with MRwt-pSae2 (Fig. 2d-iii, II) for 10 min at 30 °C and 50 mM KCl.
For TEM grid preparation, 5 µl of the reaction mixture were incubated for 1 min at room temperature on a carbon film 300 mesh copper grid (CF300-CU from Electron Microscopy Sciences) that was negatively glow discharged with the Emitech K100X glow discharge system for 45 s with 25 mA. After blotting away the excess sample with Whatman filter paper, the grid was washed twice in EM buffer (20 mM HEPES pH 7.55, 130 mM NaCl, 2% glycerol, 0.5 mM DTT), stained in two droplets of 2% uranyl acetate, and air dried. TEM micrographs were acquired with a FEI Morgagni 268 microscope at 100 kV using the Morgagni User Interface 3.0, iTEM 5.2 software and a CCD 1376 × 1032 pixel camera at different magnifications. All samples were imaged at least three times, and image analysis was performed using Fiji/ImageJ 1.52t, Excel 2016, and GraphPad Prism 9.
DNA binding experiments with oligonucleotide-based DNA were carried out in buffer containing 25 mM Tris acetate pH 7.5, 1 mM DTT, 0.25 mg/ml bovine serum albumin (BSA), 5 mM magnesium acetate, and 1 mM ATPγS. To keep the total DNA amount constant, 5 nM 20 bp, 3.22 nM 31 bp, 1.43 nM 70 bp, 1 nM 100 bp, and 0.51 nM 197 bp DNA were used, such that 100 nM (in base pairs) of 3′-labeled DNA substrate was incubated with the indicated amount of protein at 30 °C for 15 min in 15 µl reaction volume at 30 mM NaCl. Other indicated final salt concentrations were reached by NaCl addition. After incubation, the samples were supplemented with 5 µl loading dye (50% glycerol with bromophenol blue) and loaded on a 6% polyacrylamide gel (TAE). The gel was run for 90 min at 80 V (2.7 V/cm) on ice (75 min for the 20 bp substrate) in a Mini-PROTEAN® Tetra Cell system (BioRad). After separation, the gels were dried and exposed to a phosphor screen. The screen was then imaged using a Typhoon FLA 9000 imager with Typhoon FLA 9500 software (GE Healthcare), quantitated with Fiji/ImageJ 1.52t and Excel 2016, and statistically analyzed with GraphPad Prism 9.
For DNA binding experiments with plasmid-based DNA 2.5 nM (in molecules) circular 2.7 kbp-long Topoisomerase I-relaxed or KpnI-linearized pUC19 plasmid was used, as well as 2.5 nM (in molecules) streptavidin-blocked pUC19-derived 2.8 kbp substrate. Reactions of 20 µl were mixed in buffer containing 25 mM Tris acetate pH 7.5, 1 mM DTT, 80 mM KCl, 5 mM magnesium acetate, 1 mM manganese chloride, and 1 mM ATPγS. After incubation for 1 h at 30 °C, the reactions were mixed with TriTrack DNA loading dye 6× (Thermo ScientificTM) to 1× and loaded on a 0.5% agarose gel (TB). The gel was run for 1 h at 140 V (11.7 V/cm) on ice. After running, the gels were stained with GelRed Nucleic Acid Gel Stain (Biotium), imaged, and quantified as above.
Endonuclease assays (15 μl volume, unless indicated otherwise) were performed with 1 nM (in molecules) 70 bp-long DNA substrate labeled at the 3′-end in reaction buffer containing 25 mM Tris acetate pH 7.5, 1 mM DTT, 80 U/ml pyruvate kinase (Sigma-Aldrich), 1 mM phosphoenolpyruvate (PEP), 0.25 mg/ml BSA, 5 mM magnesium acetate, 1 mM manganese acetate, and 1 mM ATP. Monovalent streptavidin (30 nM) was added and incubated for 5 min at room temperature. The indicated proteins were added and incubated for 30 min at 30 °C. After incubation, the reaction was stopped with 0.5 μl of 14–22 mg/ml proteinase K (Roche), 0.5 μl of 10% (w/v) sodium dodecyl sulfate (SDS), and 0.5 μl of 0.5 M ethylenediaminetetraacetic acid (EDTA) and incubated at 50 °C for 30 min. For time-course experiments, proteins were added to a single master mix and incubated at 30 °C. Aliquots (15 μl) of the reaction were collected at the indicated time and stopped as described above. Each sample was mixed with equal volume of loading dye (95% formamide, 20 mM EDTA, 1 mg/ml bromophenol blue) and boiled for 4 min at 95 °C. The products were separated by denaturing polyacrylamide electrophoresis using a 15% polyacrylamide gel with 7 M urea. The gels were fixed for 30 min in fixing solution (40% methanol, 10% acetic acid, 5% glycerol) and then dried on 3MM paper (Whatman), exposed to phosphor screen, and imaged using a Typhoon FLA 9000 imager with Typhoon FLA 9500 software (GE Healthcare). The resulting images were analyzed with Fiji/ImageJ 1.52t, Excel 2016, and GraphPad Prism 9.
Exonuclease assays were performed with 1 nM (in molecules) 70 bp-long DNA substrate labeled at the 5′-end in reaction buffer containing 25 mM Tris acetate pH 7.5, 1 mM DTT, 0.25 mg/ml BSA, 5 mM manganese acetate, and 1 mM ATP or ATPγS where indicated. After protein addition, the reactions were incubated for 30 min at 30 °C, and subsequently stopped and processed as described above for the endonuclease assays.
Nuclease assays on plasmid-length substrate were performed with 1 nM (in molecules) 2.8 kbp-long, pUC19-derived substrate obtained with ΦC31 integrase (as described in the “DNA substrates” section) in reaction buffer as used for endonuclease assays containing 25 mM Tris acetate pH 7.5, 1 mM DTT, 80 U/ml pyruvate kinase, 1 mM PEP, 0.25 mg/ml BSA, 5 mM magnesium acetate, 1 mM manganese acetate, and 1 mM ATP. The substrate was incubated, where indicated, for 5 min at room temperature with 30 nM monovalent streptavidin. The indicated proteins were added and incubated at 30 °C. For Fig. 1d-i, the salt concentration was adjusted with KCl to 80 mM to match the TEM imaging conditions. At different time points (0, 15, and 30 min), 14 µl of the reaction were collected and transferred into tubes containing 1 µl of Exo1 or an equal amount of storage buffer and further incubated for 10 min at 30 °C. The samples without MRwt/ho-pSae2 were collected from the master mix before the addition of MRwt/MRho and pSae2 and processed with or without Exo1 as described above. After the additional 10 min, the reactions were stopped by the addition of 5 µl of 2% STOP solution (30 mM EDTA, 2% SDS, 30% glycerol, and 1 mg/ml bromophenol blue) and 1 µl Proteinase K (14–22 mg/ml). The reactions were deproteinated at 37 °C for 1 h, then 15 µl of each sample were mixed with a threefold excess of the indicated 22 nt-long radio-labeled probe compared to the plasmid-length substrate (in molecules). For Figs. 1d-i, 2a,b, 4h and Supplementary Figs. 4a and 6b, c, the radio-labeled probe anneals 160 nt away from the 5′-DNA end where the 5′-terminated strand was resected. For Supplementary Figs. 4b and 6d, the radio-labeled probe anneals 160 nt away from the 3′-DNA end where the complementary strand was resected. The mix was incubated for 3 min at 60 °C in a water bath and subsequently cooled down overnight to room temperature. The annealing reactions were loaded on a 1% agarose gel. The gels were then dried on DE81 chromatography paper (Whatman) and processed as described above.
Mass photometry measurements were carried out on a OneMP device (Refeyn Ltd). Glass coverslips (No. 1.5H thickness, 24 × 50 mm, VWR) were cleaned by sonication, first in isopropanol and then in deionized water for 15 min. After cleaning, the coverslips were dried under a clean stream of nitrogen. For measurements with the 2.7 kbp-long pUC19 plasmid, coverslips were coated with a 1% poly-lysine solution. For each measurement, a clean coverslip was placed onto the objective. For sample delivery, a silicone gasket (CultureWellTM Reusable Gasket, Grace Bio-Labs) with four wells was fixed on the surface of the coverslip. Before measurements, samples were incubated in buffer containing 25 mM Tris acetate pH 7.5, 1 mM DTT, 0.1 mM ATPγS, 5 mM magnesium acetate, and 1 mM manganese chloride at 30 °C for 30 min.
Mass photometry experiments with oligonucleotides in Figs. 1f, 3e and Supplementary Fig. 3d were performed at 25 mM KCl with 5 nM MRwt/ho or MRX and 27.7 nM 20 bp, 9.8 nM 50 bp, 6.7 nM 73 bp, 4.6 nM 106 bp, and 2.5 nM 197 bp DNA to keep the total DNA amount constant (100 nM base pairs). Peaks were detected and fitted with Gaussian functions where possible. For Supplementary Figs. 3c and 5b, 5 nM Mre11/Rad50/Xrs2/MRwt/ho/lo/MRX were measured with or without 2.5 nM 197 bp-long oligonucleotide DNA at 25 mM KCl. For mass photometry experiments in Supplementary Fig. 5d without DNA, 10 nM MRwt/ho was incubated at 250 mM KCl or 25 mM KCl and 30 °C for 30 min with co-factors as afore-mentioned. Binding was monitored for 40s.
For Supplementary Figs. 3e and 5g with 2.7 kbp plasmid DNA, 20 nM MRwt/ho/MRX was first incubated with 1 nM substrate (in molecules) for 30 min at 250 mM KCl and 30 °C and subsequently measured (labeled as 250 mM KCl, 30 min). MRwt (~460 kDa) and plasmid DNA (~1300 kDa) appeared mostly unbound, while MRX and MRho formed oligomers without DNA beyond the detection limit such that less free MRX or MRho dimers could be detected at ~550 or ~460 kDa, respectively. Secondly, the salt in the sample was diluted to 25 mM KCl, while keeping the protein and DNA concentrations constant, and the sample was immediately analyzed (labeled 25 mM KCl). Much less unbound MRX, MRwt or MRho and plasmid DNA could be detected. For MRX and MRho, however, the decrease in unbound plasmid DNA was smaller than for MRwt, indicating that they form larger oligomers per DNA molecule. For MRwt, a tail of counts at high molecular weight >2500 kDa, corresponding to higher-order MR-DNA assemblies, could be detected. MRX and MRho, on the other hand, formed much larger nucleoprotein assemblies beyond the detection limit (smaller tail of counts >2500 kDa). Third, the sample with 25 mM KCl was incubated for 30 min at 30 °C and then measured (labeled 25 mM KCl, 30 min). For both MRwt and MRho, even fewer higher-order assemblies on DNA could be detected >2500 kDa, as they surpassed the detection limit, while for MRX the tail of counts >2500 kDa remained similar. Binding was monitored for 100 s.
For data acquisition, the gasket well was filled with 9 µl measuring buffer for adjustment of surface focusing. Afterwards, 1 µl of sample was added to the buffer in the well resulting in the final sample concentration. Protein and complex binding to the coverslip surface was monitored using AcquireMP (Refeyn Ltd, version 2.3.0). Data analysis was performed by DiscoverMP (Refeyn Ltd, version 2.3.0), OriginPro 2017 and a custom-written Python program (Supplementary Software). For contrast to mass conversion, a known protein mass standard calibrant (NativeMarkTM Unstained Protein Standard; Invitrogen) was measured on the same day. Due to calibration with respect to protein mass, the pUC19 peak in plasmid measurements appeared at ~1300 kDa. All samples were measured at least three times.
Thioflavin T (ThT) fluorescence
For ThT measurements, 20 μl reactions of 1.2 µM MRwt or MRho were incubated for 30 min at 4 or 42 °C in phosphate buffer containing 20 mM phosphate solution pH 7.4 (KH2PO4, K2HPO4), 10% glycerol, 1 mM PMSF, 1 mM DTT, and 240 mM KCl. Thioflavin T (ThT; Sigma-Aldrich, T3516) was dissolved in water to a final concentration of 2.5 mM and filtered (0.2 µm; Millipore). ThT solution was added (1:10 dilution) to the samples, and fluorescence intensity measured in a 384-well plate (Corning Life Sciences) using a CLARIOstar plate reader (BMG Labtech, 5.40 R3 software). Excitation wavelength was set at 450 nm and emission recorded at 490 nm. Statistical analysis was performed with GraphPad Prism 9.
For pelleting assays, 60 μl reactions of 1.2 µM MRwt or MRho were incubated for 30 min at 4 or 42 °C in phosphate buffer as for ThT assays. After incubation, the reactions were centrifuged at 21,000g for 10 min at 4 °C. The resulting pellets were washed twice with phosphate buffer, resuspended in urea sample buffer (116 mM Tris-HCl pH 6.8, 5% glycerol, 8 M urea, 5% SDS, 1% beta-mercaptoethanol, 0.1% bromophenol blue) and loaded likewise to the supernatant fractions onto a denaturing 4–12% BisTris SDS-PAGE (Thermo Fisher Scientific) that was then Coomassie-stained. MRwt/ho bands were normalized in intensity to their input using Fiji/ImageJ 1.52t and Excel 2016; statistical analysis was performed with GraphPad Prism 9.
pSae2-Rad50 protein interaction assay
For interaction assays, 50 µl of soluble extract from Sf9 cells, expressing the MBP-tagged C-terminal part of pSae2 (pSae2 \(\triangle\)N169)19 with phosphatase inhibitors, were incubated with 50 µl Amylose resin (New England Biolabs) in 450 µl lysis buffer (50 mM Tris-HCl pH 7.5, 2 mM beta-mercaptoethanol, 1 mM EDTA, 15.6 % glycerol, 305 mM NaCl, 30 µg/ml leupeptin, 1 mM PMSF, and protease inhibitor cocktail (1:400, Sigma-Aldrich, P8340)) for 1 h at 4 °C with agitation. The resin was then centrifuged at 2000g for 2 min at 4 °C and washed five times with wash buffer containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.2% NP40, 2 mM EDTA, and 1 mM PMSF in a total volume of 1.3 ml. After the addition of 1.5 µg purified MRwt/ho/lo and 250 µl wash buffer, the mix was further incubated for 1 h at 4 °C with agitation. MRwt/ho/lo controls were also incubated with amylose resin without pSae2 \(\triangle\)N169. The resin was then centrifuged and washed five times as described above. MBP-tagged pSae2 \(\triangle\)N169 bound to pulled-down MRwt/ho/lo via Rad50 interactions was eluted with 100 µl wash buffer containing 20 mM maltose (Carl Roth). The eluate (20 µl) was then loaded with 1× NuPAGETM LDS Sample buffer on a denaturing 4–12% Bis-Tris SDS-PAGE (both Thermo Fisher Scientific). Western blots were performed with mouse α-MBP antibody (1:1000, Abcam, ab49923), rabbit α-Rad50 antibody (1:1000, ThermoScientific, PA5-32176), and goat α-rabbit or α-mouse HRP-conjugated secondary antibody (1:2000, BioRad, 170-6515 or 170-6516, respectively). The blots were imaged using a Vilber Fusion FX6 system and FusionCapt Adv FX7 software.
Yeast plasmids and strains
Construction of yeast plasmids (Supplementary Table 2) and strains (Supplementary Table 3) was performed according to standard molecular biology methods. Yeast strains used in this study are derivatives of BY4741. Cells were grown at 30 °C in yeast extract pentose dextrose, YPD (2% glucose, 2% peptone, 1% yeast extract) or synthetic SD media (2% glucose, 0.17% yeast nitrogen base, 0.5% NH4 sulfate, and amino acids). For galactose-induced overexpression experiments, overnight grown cells in SD media (2% raffinose) were diluted in SD 2% galactose or SD 2% glucose for 5 h before spotting assays or harvesting protein samples.
DNA damage sensitivity spotting and growth assays
Serial dilutions of exponentially growing cells were spotted on YPD or SD-URA (with 2% glucose or 2% galactose) agar plates without or with DNA damaging agents: Camptothecin, CPT 25 µM (Sigma-Aldrich, C9911) and Hydroxyurea, HU 25 mM (Sigma-Aldrich, H8627). Agar plates were incubated at 30 °C and imaged after 2 days using a Vilber Fusion FX6 system and FusionCapt Adv FX7 software. For growth rates, OD600 of exponentially growing cells was measured in the absence or presence of CPT (5 or 25 µM) in a 24-well plate using a plate reader (CLARIOStar BMG Labtech, 5.40 R3 software). Doubling times for exponential growth were calculated using the following Eq. (1):
Results from independent replicates (n = 3) were averaged and displayed using GraphPad Prism 9.
Live-cell fluorescence microscopy and image analysis
Yeast cells were grown overnight in SD media (2% glucose) and then diluted into fresh growth medium to resume exponential phase growth. Before imaging, cells were left untreated or treated with 200 µM CPT for 2 h. Live-cell microscopy was performed using a spinning disk confocal fluorescence microscope (Visitron system) equipped with an inverted microscope (Nikon Eclipse TiE), an EM-CCD camera (Andor iXon Ultra, Andor), a motorized XY stage, and a piezo Z drive. The system is controlled by VisiVIEW software 5.0.0 (Visitron system). Cells in 96-well plates were imaged at 30 °C in a temperature-controlled incubator with a ×100 objective lens (NA = 1.4, Nikon CFI Plan Apo). For image analysis, the percentage of cells containing Rad50 foci in the nucleus, based on histone Hta2-mCherry signal, was quantified using the Cell Counter plugin in Fiji/ImageJ 1.52t software. Results of independent biological replicates (n = 3) were averaged and displayed with GraphPad Prism 9. Furthermore, automated single-cell image analysis was performed using YeastQuant 2021 software69 on raw images running in Matlab R2017b. Cells were segmented based on brightfield channel images and nuclei were segmented based on Hta2-mCherry. Relocalization and foci formation of Rad50-yEVenus were quantified as normalized standard deviation of yEVenus average intensity signal. At least 50 cells per condition were analyzed and quantified.
Rad50/Rad53 western blots and flow cytometry analysis
Total protein lysates were prepared according to standard procedures. Briefly, cells of OD600 = 1–2 were resuspended in 150 µl lysis solution (1.85 M NaOH, 1.09 M beta-mercaptoethanol) and precipitated in 50% TCA on ice for 10 min. After centrifugation, the pellet was resuspended with 1 ml acetone and centrifuged again. Then the pellet was resuspended in 100 μl urea buffer (120 mM Tris-HCl pH 6.8, 5% glycerol, 8 M urea, 143 mM beta-mercaptoethanol, 8% SDS, bromophenol blue) and loaded on a 7.5% acrylamide gel (Rad50 blot) or on a 7.5% TGX SDS-PAGE gel (BioRad) (Rad53 blot). The following primary antibodies were used for immunoblotting: α-Rad50 (1:1000; ThermoScientific, PA5-32176), α-Tubulin (1:20,000; Sigma-Aldrich T5168), α-Pgk1 (1:10,000; Invitrogen, 459250), and α-Rad53 (1:1000; Abcam, ab166859). Goat α-rabbit or α-mouse HRP-conjugated secondary antibody (1:2000; BioRad, 170-6515 or 170-6516 in Rad50 blot/170-5047 in Rad53 blot, respectively) were used as secondary antibodies. Blots were developed using the SuperSignal West Pico PLUS Chemiluminescence substrate (Thermo Fisher Scientific, 34579), and images were acquired using a Vilber Fusion FX6 system (FusionCapt Adv FX7 software).
For flow cytometry analysis, cells with OD600 = 0.5 were fixed in 70% ethanol overnight. Pellets were treated with RNase A (Thermo Fisher Scientific, 10753721) at 37 °C for 2 h and Proteinase K (Biofroxx, 1151ML010) at 50 °C for 2 h in 50 mM Tris-HCl pH 7.5 buffer. The cell suspension was sonified using a Branson sonifier 450 for 5 s with output control 1 and duty cycle constant. Then, cells were stained with 2.4 µM SYTOX Green (ThermoFisher Scientific, 1076273). Measurement was performed on the BD LSRFortessa flow cytometer (BD Biosciences) using the BD FACSDiva software (v9.0.1). In total, 20,000 events were recorded. Analysis was performed with FlowJo (v10.8.0) using the gating strategy as indicated in Supplementary Fig. 7b and the Reporting summary.
Isolation of genomic DNA and Telomeric PCR
Untagged strains were grown in YPD medium until the cultures reached the exponential growth phase and OD600 = 0.8–1. Then, 10 ml of the cell suspension were harvested for genomic DNA (gDNA) isolation using the Gentra Yeast/Bact. PureGene Kit (Qiagen) according to the yeast protocol section with the following minor modifications: 10 ml cells were used as input, 10 µl Zymolyase 20T (5 units/µl, amsbio 120491-1) were added and incubated for 1 h at 37 °C. The gDNA concentration was measured using the Nanodrop2000c Spectrophotometer (ThermoFisher). The isolated gDNA was stored at 4 °C.
Telomeric PCR was performed as described in Förstemann et al.70 with minor modifications. All steps were carried out on ice and in the C1000 Touch Thermal Cycler (Bio-Rad). The C-tailing reaction (7.1 µl sterile water, 0.9 µl 10× NEBuffer 4 (New England Biolabs, B7004S), 1 µl of 100 ng/µl gDNA) was incubated for 10 min at 96 °C to denature the DNA, then the mix was cooled down to 4 °C. Next, 1 µl of 10× tailing mix (2 µl 2 units/µl Terminal transferase (New England Biolabs, M0315S), 1 µl 10× NEBuffer 4, 1 µl 10 mM dCTP (Thermo Fisher, R0151), 6 µl sterile water) was added and incubated for 30 min at 37 °C, 10 min at 65 °C, 5 min at 96 °C, and kept at 65 °C. Then, 30 µl of 65 °C preheated PCR-mix (21 µl sterile water, 4 µl 10× Telo-PCR buffer (670 mM Tris-HCl (pH 8.8), 160 mM (NH4)2SO4, 50% glycerol, 0.1% Tween-20), 4 µl 2 mM dNTP mix (ThermoFisher, R0191), 0.3 µl 10 µM forward oligonucleotide (1L oBL358, 6Y′ oBL361, or 6R oBL360), 0.3 µl 10 µM reverse oligonucleotide (G18 oBL359), and 0.4 µl Phusion High Fidelity DNA polymerase (2 U/µl) (New England Biolabs, M0530S)) were added to the C-tailing reaction. The distance of the subtelomeric oligonucleotides to the first TG repeat is 119 bp for 6Y′ oBL361, 39 bp for 1L oBL358, and 88 bp for 6R oBL360. The Telo-PCR program was run with the following steps: initial denaturation 3 min 95 °C, 45 cycles: 30 s 95 °C, 15 s 63 °C, 20 s 72 °C, and a final extension of 5 min at 72 °C. Finally, 8 µl of 6× Ficoll orange loading dye (15% Ficoll Type 400, 10 mM EDTA pH 8.0, Orange G) were mixed with each Telo-PCR sample and the entire PCR product was separated on a 1.8% 1× TBE-agarose gel using 0.5× SYBR Safe DNA gel stain (ThermoFisher, S33102) and the 100 bp DNA ladder (N3231S, New England Biolabs). Gel images were taken with the ChemiDoc Touch Imaging System (BioRad) and the analysis of the mean telomere length and distribution was executed using the ImageLab software version 5.2.1 (BioRad). Graphs were generated using the Graph Pad Prism 9 software.
All data shown in this study are representative of at least three independent experiments if not stated otherwise in the figure legends. See respective figure legends and “Methods” section for details to plotted data points, error bars, and specific data analyses. Statistical unpaired, two-tailed significance t-tests were performed using GraphPad Prism 9. For Fig. 1e, the t-value was 12.63 with 5 degrees of freedom. For Fig. 3c, the t-value was 22.51 with 6 degrees of freedom. For Fig. 3f, the t-values were 3.37 (Pellet) and 3.05 (Sup.), each with 4 degrees of freedom. For Fig. 4e, the t-values were 2.95 (12.5 nM), 4.18 (25 nM), and 6.04 (50 nM), each with 4 degrees of freedom. For Fig. 5g, the t-value was 3.03 with 4 degrees of freedom. For the non-significant t-test in Supplementary Fig. 1c-iv, the t-value was 1.47 with 198 degrees of freedom. For Supplementary Fig. 4a, the t-values were 4.28 (**), 6.6 (***), and 22.86 (****) with 6 degrees of freedom each. For Supplementary Fig. 5e, the t-values were 3.56 (*), 6.25 (***), and 22.51 (****) with 6 degrees of freedom each.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The data supporting this study are available from the corresponding authors upon reasonable request. The Rad50 crystal structure (identifier 5DAC https://doi.org/10.2210/pdb5DAC/pdb) was obtained from the RCSB Protein Data Bank (PBD, https://www.rcsb.org/). Rad50 patient mutations were found in the ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) and MedGen (https://www.ncbi.nlm.nih.gov/medgen/) Databases. Source data are provided with this paper.
An exemplary custom-written Python code for the analysis of mass photometer data is provided with this study as a Supplementary Software file.
Bian, L., Meng, Y., Zhang, M. & Li, D. MRE11-RAD50-NBS1 complex alterations and DNA damage response: implications for cancer treatment. Mol. Cancer 18, 169 (2019).
Chaudhuri, A. R. et al. Replication fork stability confers chemoresistance in BRCA-deficient cells. Nature 535, 382–387 (2016).
Schlacher, K. et al. Double-strand break repair-independent role for BRCA2 in blocking stalled replication fork degradation by MRE11. Cell 145, 529–542 (2011).
Reginato, G. & Cejka, P. The MRE11 complex: a versatile toolkit for the repair of broken DNA. DNA Repair 91–92, 102869 (2020).
Oh, J. & Symington, L. S. Role of the Mre11 complex in preserving genome integrity. Genes 9, 589 (2018).
Zhu, Z., Chung, W.-H., Shim, E. Y., Lee, S. E. & Ira, G. Sgs1 helicase and two nucleases Dna2 and Exo1 resect DNA double strand break ends. Cell 134, 981–994 (2008).
Mimitou, E. P. & Symington, L. S. Sae2, Exo1 and Sgs1 collaborate in DNA double-strand break processing. Nature 455, 770–774 (2008).
Lobachev, K. S., Gordenin, D. A. & Resnick, M. A. The Mre11 complex is required for repair of hairpin-capped double-strand breaks and prevention of chromosome rearrangements. Cell 108, 183–193 (2002).
Reginato, G., Cannavo, E. & Cejka, P. Physiological protein blocks direct the Mre11-Rad50-Xrs2 and Sae2 nuclease complex to initiate DNA end resection. Genes Dev. 31, 2325–2330 (2017).
Wang, W., Daley, J. M., Kwon, Y., Krasner, D. S. & Sung, P. Plasticity of the Mre11–Rad50–Xrs2–Sae2 nuclease ensemble in the processing of DNA-bound obstacles. Genes Dev. 31, 2331–2336 (2017).
Deshpande, R. A. et al. DNA-dependent protein kinase promotes DNA end processing by MRN and CtIP. Sci. Adv. 6, eaay0922 (2020).
Cannavo, E. & Cejka, P. Sae2 promotes dsDNA endonuclease activity within Mre11-Rad50-Xrs2 to resect DNA breaks. Nature 514, 122–125 (2014).
Connelly, J. C., de Leau, E. S. & Leach, D. R. F. Nucleolytic processing of a protein-bound DNA end by the E. coli SbcCD (MR) complex. DNA Repair 2, 795–807 (2003).
Hopkins, B. B. & Paull, T. T. The P. furiosus Mre11/Rad50 complex promotes 5′ strand resection at a DNA double-strand break. Cell 135, 250–260 (2008).
Anand, R., Ranjha, L., Cannavo, E. & Cejka, P. Phosphorylated CtIP functions as a co-factor of the MRE11-RAD50-NBS1 endonuclease in DNA end resection. Mol. Cell 64, 940–950 (2016).
Llorente, B. & Symington, L. S. The Mre11 nuclease is not required for 5’ to 3’ resection at multiple HO-induced double-strand breaks. Mol. Cell. Biol. 24, 9682–9694 (2004).
Oh, J., Al-Zain, A., Cannavo, E., Cejka, P. & Symington, L. S. Xrs2 dependent and independent functions of the Mre11-Rad50 complex. Mol. Cell 64, 405–415 (2016).
Huertas, P., Cortés-Ledesma, F., Sartori, A. A., Aguilera, A. & Jackson, S. P. CDK targets Sae2 to control DNA-end resection and homologous recombination. Nature 455, 689–692 (2008).
Cannavo, E. et al. Regulatory control of DNA end resection by Sae2 phosphorylation. Nat. Commun. 9, 4016 (2018).
Garcia, V., Phelps, S. E. L., Gray, S. & Neale, M. J. Bidirectional resection of DNA double-strand breaks by Mre11 and Exo1. Nature 479, 241–244 (2011).
Cannavo, E., Reginato, G. & Cejka, P. Stepwise 5′ DNA end-specific resection of DNA breaks by the Mre11-Rad50-Xrs2 and Sae2 nuclease ensemble. Proc. Natl Acad. Sci. USA 116, 5505–5513 (2019).
Reczek, C. R., Szabolcs, M., Stark, J. M., Ludwig, T. & Baer, R. The interaction between CtIP and BRCA1 is not essential for resection-mediated DNA repair or tumor suppression. J. Cell Biol. 201, 693–707 (2013).
Djuzenova, C., Mühl, B., Schakowski, R., Oppitz, U. & Flentje, M. Normal expression of DNA repair proteins, hMre11, Rad50 and Rad51 but protracted formation of Rad50 containing foci in X-irradiated skin fibroblasts from radiosensitive cancer patients. Br. J. Cancer 90, 2356–2363 (2004).
Gatei, M., Kijas, A. W., Biard, D., Dörk, T. & Lavin, M. F. RAD50 phosphorylation promotes ATR downstream signaling and DNA restart following replication stress. Hum. Mol. Genet. 23, 4232–4248 (2014).
Lisby, M., Barlow, J. H., Burgess, R. C. & Rothstein, R. Choreography of the DNA damage response: spatiotemporal relationships among checkpoint and repair proteins. Cell 118, 699–713 (2004).
Maser, R. S., Monsen, K. J., Nelms, B. E. & Petrini, J. H. J. hMre11 and hRad50 nuclear foci are induced during the normal cellular response to DNA double-strand breaks. Mol. Cell. Biol. 17, 6087–6096 (1997).
Zabolotnaya, E. et al. Modes of action of the archaeal Mre11/Rad50 DNA-repair complex revealed by fast-scan atomic force microscopy. Proc. Natl Acad. Sci. USA 117, 14936–14947 (2020).
Usui, T. et al. Complex formation and functional versatility of Mre11 of budding yeast in recombination. Cell 95, 705–716 (1998).
Hopfner, K.-P. et al. Structural biochemistry and interaction architecture of the DNA double-strand break repair Mre11 nuclease and Rad50-ATPase. Cell 105, 473–485 (2001).
Chen, L., Trujillo, K., Ramos, W., Sung, P. & Tomkinson, A. E. Promotion of Dnl4-Catalyzed DNA end-joining by the Rad50/Mre11/Xrs2 and Hdf1/Hdf2 complexes. Mol. Cell 8, 1105–1115 (2001).
Tisi, R., Vertemara, J., Zampella, G. & Longhese, M. P. Functional and structural insights into the MRX/MRN complex, a key player in recognition and repair of DNA double-strand breaks. Comput. Struct. Biotechnol. J. 18, 1137–1152 (2020).
Tatebe, H. et al. Rad50 zinc hook functions as a constitutive dimerization module interchangeable with SMC hinge. Nat. Commun. 11, 370 (2020).
Trujillo, K. M. et al. Yeast Xrs2 binds DNA and helps target Rad50 and Mre11 to DNA ends. J. Biol. Chem. 278, 48957–48964 (2003).
Rojowska, A. et al. Structure of the Rad50 DNA double‐strand break repair protein in complex with DNA. EMBO J. 33, 2847–2859 (2014).
de Jager, M. et al. Human Rad50/Mre11 is a flexible complex that can tether DNA ends. Mol. Cell 8, 1129–1135 (2001).
Hopfner, K.-P. et al. The Rad50 zinc-hook is a structure joining Mre11 complexes in DNA recombination and repair. Nature 418, 562–566 (2002).
de Jager, M. et al. Differential arrangements of conserved building blocks among homologs of the Rad50/Mre11 DNA repair protein complex. J. Mol. Biol. 339, 937–949 (2004).
Möckel, C., Lammens, K., Schele, A. & Hopfner, K.-P. ATP driven structural changes of the bacterial Mre11:Rad50 catalytic head complex. Nucleic Acids Res. 40, 914–927 (2012).
Käshammer, L. et al. Mechanism of DNA end sensing and processing by the Mre11-Rad50 complex. Mol. Cell 76, 382–394 (2019).
Liu, Y. et al. ATP‐dependent DNA binding, unwinding, and resection by the Mre11/Rad50 complex. EMBO J. 35, 743–758 (2016).
Lammens, K. et al. The Mre11:Rad50 structure shows an ATP-dependent molecular clamp in DNA double-strand break repair. Cell 145, 54–66 (2011).
Raymond, W. E. & Kleckner, N. RAD50 protein of S. cerevisiae exhibits ATP-dependent DNA binding. Nucleic Acids Res. 21, 3851–3856 (1993).
Saecker, R. M. Protein-DNA interactions: polyelectrolyte effects. Encyclopedia of Life Sciences. https://doi.org/10.1002/9780470015902.a0001350.pub2 (2007).
Sonn-Segev, A. et al. Quantifying the heterogeneity of macromolecular machines by mass photometry. Nat. Commun. 11, 1772 (2020).
Cannavo, E., Cejka, P. & Kowalczykowski, S. C. Relationship of DNA degradation by Saccharomyces cerevisiae Exonuclease 1 and its stimulation by RPA and Mre11-Rad50-Xrs2 to DNA end resection. Proc. Natl Acad. Sci. USA 110, 1661–1668 (2013).
Nicolette, M. L. et al. Mre11–Rad50–Xrs2 and Sae2 promote 5′ strand resection of DNA double-strand breaks. Nat. Struct. Mol. Biol. 17, 1478–1485 (2010).
Lisby, M., Mortensen, U. H. & Rothstein, R. Colocalization of multiple DNA double-strand breaks at a single Rad52 repair center. Nat. Cell Biol. 5, 572–577 (2003).
Seifert, F. U., Lammens, K., Stoehr, G., Kessler, B. & Hopfner, K. Structural mechanism of ATP-dependent DNA binding and DNA end bridging by eukaryotic Rad50. EMBO J. 35, 759–772 (2016).
Hopfner, K.-P. et al. Mre11 and Rad50 from Pyrococcus furiosus: cloning and biochemical characterization reveal an evolutionarily conserved multiprotein machine. J. Bacteriol. 182, 6036–6041 (2000).
Biancalana, M., Makabe, K., Koide, A. & Koide, S. Molecular mechanism of Thioflavin-T binding to the surface of β-rich peptide self-assemblies. J. Mol. Biol. 385, 1052–1063 (2009).
Fujiwara, K., Toda, H. & Ikeguchi, M. Dependence of α-helical and β-sheet amino acid propensities on the overall protein fold type. BMC Struct. Biol. 12, 18 (2012).
Alani, E., Padmore, R. & Kleckner, N. Analysis of wild-type and rad50 mutants of yeast suggests an intimate relationship between meiotic chromosome synapsis and recombination. Cell 61, 419–436 (1990).
Sandell, L. L. & Zakian, V. A. Loss of a yeast telomere: arrest, recovery, and chromosome loss. Cell 75, 729–739 (1993).
Keener, R., Connelly, C. J. & Greider, C. W. Tel1 activation by the MRX complex is sufficient for telomere length regulation but not for the DNA damage response in Saccharomyces cerevisiae. Genetics 213, 1271–1288 (2019).
Myler, L. R. et al. Single-molecule imaging reveals how Mre11-Rad50-Nbs1 initiates DNA break repair. Mol. Cell 67, 891–898 (2017).
Rahman, S., Beikzadeh, M., Canny, M. D., Kaur, N. & Latham, M. P. Mutation of conserved Mre11 residues alter protein dynamics to separate nuclease functions. J. Mol. Biol. 432, 3289–3308 (2020).
Lim, H. S., Kim, J. S., Park, Y. B., Gwon, G. H. & Cho, Y. Crystal structure of the Mre11-Rad50-ATPγS complex: understanding the interplay between Mre11 and Rad50. Genes Dev. 25, 1091–1104 (2011).
Wienert, B. et al. Unbiased detection of CRISPR off-targets in vivo using DISCOVER-Seq. Science 364, 286–289 (2019).
Pham, N. et al. Mechanisms restraining break‐induced replication at two‐ended DNA double‐strand breaks. EMBO J. 40, e104847 (2021).
Kilic, S. et al. Phase separation of 53BP1 determines liquid‐like behavior of DNA repair compartments. EMBO J. 38, e101379 (2019).
Oshidari, R. et al. DNA repair by Rad52 liquid droplets. Nat. Commun. 11, 695 (2020).
Kim, H. et al. Frequency of pathogenic germline mutation in CHEK2, PALB2, MRE11, and RAD50 in patients at high risk for hereditary breast cancer. Breast Cancer Res.Treat. 161, 95–102 (2017).
Tommiska, J. et al. Evaluation of RAD50 in familial breast cancer predisposition. Int. J. Cancer 118, 2911–2916 (2006).
Scully, R. et al. Dynamic changes of BRCA1 subnuclear location and phosphorylation state are initiated by DNA damage. Cell 90, 425–435 (1997).
Wu, J., Lu, L.-Y. & Yu, X. The role of BRCA1 in DNA damage response. Protein Cell 1, 117–123 (2010).
Feng, L., Fong, K.-W., Wang, J., Wang, W. & Chen, J. RIF1 counteracts BRCA1-mediated end resection during DNA repair. J. Biol. Chem. 288, 11135–11143 (2013).
Lemaçon, D. et al. MRE11 and EXO1 nucleases degrade reversed forks and elicit MUS81-dependent fork rescue in BRCA2-deficient cells. Nat. Commun. 8, 860 (2017).
Pinto, C., Anand, R. & Cejka, P. Methods to study DNA end Resection II: biochemical reconstitution assays. Methods Enzymol. 600, 67–106 (2018).
Pelet, S., Dechant, R., Lee, S. S., van Drogen, F. & Peter, M. An integrated image analysis platform to quantify signal transduction in single cells. Integr. Biol. 4, 1274–1282 (2012).
Förstemann, K., Höss, M. & Lingner, J. Telomerase-dependent repeat divergence at the 3′ ends of yeast telomeres. Nucleic Acids Res. 28, 2690–2694 (2000).
We thank Marc Wold (University of Iowa, US) for the p11d-scRPA vector, Mark Howarth (University of Oxford, UK) for monovalent streptavidin, and Susan Gasser and Kenji Shimada (FMI, Basel) for Rad53 antibodies and protocols. We are grateful to Maryna Levikova-Denzler, Cosimo Pinto, Radoslav Enchev, Weaam Mohamed and Anne Schreiber for technical advice and assistance; Peter Tittmann, Stephan Handschin, Miroslav Peterek, ScopeM, and the ETH Cryo-EM Knowledge Hub for assistance and access to electron microscopy; the Institute of Molecular Biology (IMB gGmBH) Flow Cytometry Core Facility for support with flow cytometry and Thomas M. Wismer for help with ClinVar/MedGen database analysis. We thank Josef Jiricny and members of the Cejka and Peter laboratories for fruitful discussions, and Jana R. Kissling, Silvia Napolitano, and Alicia Smith for critical reading of the manuscript. Work in the Cejka laboratory profits from funding by the Swiss National Science Foundation (SNSF, 31003A_175444, 310030_205199), the Swiss Cancer League, the Helmut Horten Foundation and the European Research Council (ERC, 681-630), while the Peter laboratory was supported by the SNSF, the Swiss Cancer league and ETH Zurich. The work of the Seidel laboratory was supported by the European Research Council (ERC). Work in the Luke lab was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) Project-ID 3935478939-SFB1631 and the Heisenberg program of the DFG, LU 1709-2-1.
The authors declare no competing interests.
Peer review information
Nature Communications thanks Karl-Peter Hopfner, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Kissling, V.M., Reginato, G., Bianco, E. et al. Mre11-Rad50 oligomerization promotes DNA double-strand break repair. Nat Commun 13, 2374 (2022). https://doi.org/10.1038/s41467-022-29841-0
This article is cited by
Journal of Biomolecular NMR (2023)