Continuous nuclear envelope surveillance is required for DNA double strand break repair

Abstract

Precise double-strand break (DSB) repair is paramount for genome stability. Homologous recombination (HR) is preferred to repair DSBs when a nearby sister chromatid ensures an error-free template. In Saccharomyces cerevisiae, this preference extends into anaphase and telophase (late mitosis; late-M) despite sister chromatids having been pulled apart. Previously, we identified the nuclear envelope (NE) protein Msc1 as important for late-M DSB repair. Here, we report that Msc1 faces the NE lumen, and its depletion leads to DSB-independent over-compartmentalization of the nucleus and nuclear pore complex (NPC) mislocation. These phenotypes are partly shared by the highly conserved NE healing complex ESCRT-III. Finally, we show that these abnormal NE phenotypes are found throughout the cell cycle and in the DSB-driven G2/M arrest. These findings highlight the essential role of NE homeostasis in DSB repair.

Msc1 is a yeast protein that maintains nuclear envelope and nuclear pore complex integrity for DNA repair. Msc1 synergistically works with the ESCRT-III complex for this purpose, underscoring NE homeostasis as key to genome stability.

Introduction

DNA double-strand breaks (DSBs) jeopardize cell survival and genome stability, crucially contributing to carcinogenesis^1,2. Cells employ two primary DNA repair mechanisms against DSBs: non-homologous end joining (NHEJ) and homologous recombination (HR). NHEJ relies on error-prone pathways to mend broken DNA ends, while HR utilizes intact homologous sequences for precise restoration. The choice between NHEJ and HR hinges on cyclin-dependent kinase (CDK) activity, dictating preference based on the cell cycle phase^{3, 4–5}. HR demands a well-aligned sister chromatid to be error-free, which makes it optimal for S and G2 phases. However, CDK activity remains high well into mitosis (M phase), even after anaphase onset breaks down the proper alignment of sister chromatids. In complex eukaryotes, extensive chromosomal condensation in early M phase impedes HR^{6, 7–8}. Conversely, in simpler eukaryotes like yeast, HR remains active in late mitosis (late-M; anaphase/telophase) and appears to repair some DSBs through re-alignment of segregated sister chromatids^{8, 9–10}. One crucial step for this realignment is a change in the late-M dynamics of the microtubule cytoskeleton, both the spindle and astral microtubules, that forces the relaxation of the elongated nucleus and thus the re-approximation of the segregated sisters¹⁰. This way, the ends of the DSB might realign with the unbroken sister for an effective HR-driven DSB repair.

An important yet frequently overlooked event during late-M is the massive remodeling of the nucleus. Eukaryotic chromosomes are surrounded by the nuclear envelope (NE), a double membrane that is continuous with the endoplasmic reticulum (ER). The NE can either remain intact or disassemble prior to chromosome segregation, defining two major forms of mitosis termed closed and open mitosis, respectively¹¹. Closed mitosis is considered the most ancient form and is found in many fungi and unicellular protists, including important human parasites. In contrast, most plant and animal cells undergo open mitosis. Although the NE remains during the closed mitosis, it still undergoes significant changes. In particular, its shape transitions from a sphere to an hourglass in early anaphase and to a dumbbell in late anaphase. In the latter morphology, a thin bridge of nucleoplasmic material connects the segregating nuclei until cytokinesis¹². Cytokinesis is preceded by karyokinesis, yet how this mechanistically occurs and is coordinated with the transition to a new cell cycle is poorly understood. In the fission yeast Schizosaccharamyces pombe, the NE protein Les1 has a direct role in modulating karyokinesis by ensuring the local corralling of a subset of nuclear pore complexes (NPCs) at the mid-bridge¹³. This appears to correctly position karyokinesis when these NPCs disassemble in late-M to trigger a local NE breakdown reminiscent of the general NE breakdown observed in open mitosis. Completion of karyokinesis must then occur through membrane fission/fusion events, a process that is likely executed by the highly conserved ESCRT-III complex^13,14. This complex also deals with many other physiological and accidental wounds to the NE that pose a risk for cell survival or genetic stability as nucleoplasm leaks into the cytoplasm and vice versa^14,15. In fact, the knockout mutant for SpLes1 not only causes karyokinesis malposition but also leads to transient NE ruptures that must be repaired by ESCRT-III to preserve cell viability^13,16. Another role of ESCRT-III on the yeast NE homeostasis is the control of the quality and distribution of NPCs¹⁷, which suggests not only a tight link with the nucleus-cytoplasm shuffling but also with DSB repair. Indeed, eroded telomeres and other one-ended DSBs are recruited to NPCs to be repaired by the specific HR subpathway known as break-induced replication (BIR)^18,19. NPCs also tether DSBs that are difficult to repair by HR and undergo NHEJ instead²⁰.

In a previous work, we found through proteomics that the only SpLes1 ortholog in S. cerevisiae, Msc1, plays a particular role in facilitating DSB repair in late-M²¹. In the present work, we show that Msc1 controls proper NE shape and NPC distribution, which establishes a novel connection between NE homeostasis and DSB repair in late-M. In addition, we also show that the highly conserved NE healing complex ESCRT-III partly phenocopies these defects and cooperates with Msc1 in ensuring cell recovery from DSBs.

Results

Msc1 faces the NE lumen and does not translocate after DNA damage

Msc1 is the only putative ortholog in S. cerevisiae of two S. pombe proteins, Les1 and Ish1. Little is known about the function of these three proteins; however, Ish1 and Les1 are known to be distributed along the inner and outer nuclear membranes (INM and ONM) of the NE and to form homo- and heterodimers²². Both Ish1 and Les1 are characterized by the presence of several Ish1 motifs (pfam PF10281), whose molecular function(s) is unknown, a putative C2H2 zinc finger (ZnF) DNA-binding domain, and a unique transmembrane (TM) domain at the very N-terminus (Fig. 1a)^13,22. The predicted number of Ish1 motifs differs between papers, from 2 to 5–6, but 4 is the most likely number based on recent in silico predictions and motif alignments²², with the third Ish1 motif overlapping with the C2H2 ZnF. Like its putative S. pombe orthologs, Msc1 is predicted to contain four Ish1 motifs (https://www.genome.jp/tools/motif/; independent E-values < 10⁻¹⁰) and a single N-terminal TM domain. However, unlike SpLes1 and SpIsh1, the putative C2H2 ZnF in Msc1 does not overlap with any Ish1 motif (Fig. 1a).

[See PDF for image]

Fig. 1

Spatial arrangement of Msc1 in the nuclear envelope.

a Protein structure of ScMsc1 and its close orthologs SpIsh1 and SpLes1. The different protein motifs are depicted as boxes. Proteins and motif lengths are to scale, with delimiting amino acids numbered under or in the boxes. b Schematic of the system used to determine the NE face of the Msc1 C-terminus. The system is based on three strains with different mCherry-GFP₁₁ chimeras reporting for INM facing the nucleoplasm (GFP₁₁-mCherry-Pus1), ONM facing the cytosol (GFP₁₁-mCherry-Scs2TM), and INM/ONM facing the NE lumen (mCherry-Scs2TM-GFP₁₁). The GFP₁₁ is a small portion (3 kDa) of the GFP that complements the larger GFP_1-10 fragment (24 kDa), which in turn is tagged to the protein of interest (Msc1 in our case). Only if GFP11 binds to GFP1-10, a full and functional GFP is reconstituted. c Based on such a system, the spatial location of the functional globular tract of Msc1 was determined in the late-M NE both before and after DSB generation (1 h phleomycin). In all scenarios, reconstituted GFP was seen only in the strain that reports for the NE lumen. Representative micrographs are shown.

The TM domain determines that the N- and the C-termini of the protein must face different compartments. The close proximity of the TM to the N-terminus implies that all protein functions reside between the TM sequence and the C-terminus. Interestingly, the ultrastructural analysis of both SpIsh1 and SpLes1 by electron microscopy has shown that the C-terminal faces the lumen of the NE²². This configuration has profound implications as, for instance, the putative C2H2 ZnF DNA-binding domain cannot interact with the chromatin. It is worth noting that C2H2 ZnF proteins can organize chromosome architecture²³, so Msc1 could directly modulate chromosome organization during DNA damage provided that its C2H2 ZnF faces the nucleoplasm. Likewise, the Ish1 motifs share similarity with the conserved HeH/LEM (pfam PF12949) and SAP (pfam PF02037) domains¹³, which have DNA-binding and chromatin-anchoring properties as well.

To check the spatial configuration of Msc1, we used a system based on the reconstitution of a split GFP when a target protein and a localization reporter reside in the same compartment²⁴. We C-terminal tagged Msc1 with one half of the GFP (GFP_1-10) and introduced the construct into three strains that differ in the localization of the other half of the GFP reporter (GFP₁₁) (Fig. 1b). One reporter, Pus1, is set for the nucleoplasm and allows GFP reconstitution when the protein of interest is located in the nucleoplasm or facing the nucleoplasm from its location at the INM. A second reporter, Scs2, is an integral ONM/ER protein with the N-terminus facing the cytoplasmic side and the C-terminus facing the space between INM and ONM (lumen). In this manner, N- vs C-terminal tagging with GFP₁₁ makes Scs2 a reporter for ONM/cytoplasm and lumen, respectively. In addition to the GFP₁₁ half, all these reporters carry a fused mCherry as an internal control (Fig. 1b). Of the three reporters, GFP reconstitution with Msc1-GFP_1-10 in asynchronous log-phase cells was only possible with the Scs2TM where the GFP₁₁ was facing the NE/ER lumen (Fig. S1a, left panels). This indicates that all protein motifs reside in the lumen, consistent with what has been reported for the S. pombe orthologs. Importantly, the Msc1-GFP_1-10 chimera was as functional as Msc1 in its protective role against DSBs generated by the radiomimetic drug phleomycin (Fig. S1b,c). Proper luminal location of the C-terminus was critical for Msc1 function, as constitutive flipping of the N- and C-terminus rendered a non-functional chimera (Fig. S1b,c). We achieved this goal by taking advantage of a construct that contains a signal peptide (from Kar2) preceding GFP_1-10 that is then tagged at N-terminal of Msc1 (Kar2ss-GFP_1-10-Msc1). As expected, this non-functional chimera flipped Msc1 orientation and placed the N-terminus in the lumen (Fig. S1d, left panels).

Next, we determined whether the lumen targeting of the functional Msc1-GFP_1-10 chimera was maintained in late-M blocked cells. To synchronize the cell culture in late-M (Fig. S2), we used the thermosensitive cdc15-2 allele of the kinase Cdc15, which is critical for the telophase-G1 transition. At 34 °C, this mutant arrests cells in late-M with sister chromatids segregated and the NE elongated in hourglass/dumbbell shapes^12,25,26. As in asynchronous cells, late-M cells maintained a lumen-oriented Msc1 in the elongated NE (Fig. 1c, left panels). This implies that the putative C2H2 ZnF domain cannot be in contact with the chromatin during the late-M arrest. However, it could still be possible that Msc1 translocates or flips exclusively upon the generation of DSBs, resulting in the functional part of the protein, including the ZnF, facing the nucleoplasm. Thus, we performed GFP reconstitution experiments after DSB generation by phleomycin, yet we still found that Msc1 was entirely facing the lumen (Fig. 1c, right panels). The same outcome was obtained when phleomycin was added to an asynchronous culture, which causes a G2/M arrest through the DNA damage checkpoint (DDC) (Fig. S1a,d; right panels).

Loss of Msc1 results in aberrant NE morphologies in late mitosis

The fact that Msc1 always faces the NE lumen implies that its role on DSB repair must be indirect; i.e., Msc1 physically interacts with neither DSBs nor HR factors. SpLes1 has been proposed as a modulator of karyokinesis¹³. When SpLes1 is depleted, karyokinesis is misplaced and the NE is transiently damaged¹³. In this context, we reasoned that the nucleoplasmic bridge of S. cerevisiae cells could become compromised in Δmsc1 during the protracted arrest in late-M that precedes DSB generation. For instance, premature karyokinesis in cdc15-2 Δmsc1 could irreversibly preclude the search for the sister chromatid during the partial anaphase regression. Hence, we sought signs of premature karyokinesis in Δmsc1. For this purpose, we labeled the NE with Sec61 and Nup49. The former is an NE/ER protein that continuously and strongly labels the NE, whereas the latter is an NPC component that specifically labels the NE, although it tends to give a more punctate (and thus imprecise) pattern²⁷. In general, signs of premature karyokinesis in Δmsc1, seen as a partial or complete absence of the bridge, were rather modest or absent. A Sec61 bridge was observed in ~95% of cdc15-2 MSC1 cells and ~90% of cdc15-2 Δmsc1 cells, whereas bridging Nup49 was observed in similar proportions in both strains (~80%) (Fig. S3a,b; mock experiments). In addition, we also directly examined the continuity of the nucleoplasmic bridge using the freely circulating nucleoplasmic TetR-YFP reporter. With this marker, we found only a modest increase (~10%) of signs that can be compatible with premature karyokinesis in cdc15-2 Δmsc1 cells (Fig. S3c). Next, we looked at whether karyokinesis was instead accelerated during DSB generation, as the stress caused by DNA damage and the general mobilization of the chromatin for its repair may destabilize the NE bridge in cdc15-2 Δmsc1. In this case, and in addition to phleomycin, we also used a pair of DSBs generated at a single location on chromosome III, one per sister chromatid. The system to generate these DSBs is based on the β-estradiol inducible expression of the HO endonuclease (Fig. S2), which cuts the HOcs sequence at the MAT locus^28,29. With either strategy, a similarly high percentage of cells retain the bridge (Fig. S3).

Even though premature karyokinesis was not observed in Δmsc1, during the course of these experiments we notice that a higher proportion of cells presented an aberrant morphology in one or both segregated nuclear bodies, which appeared to consist of a partition within the nucleus (Fig. 2a, b). Although these aberrant shapes were already observed in cdc15-2 cells, the percentage in late-M was significantly higher in cdc15-2 Δmsc1 (Fig. 2a; ~15% in WT vs ~30% in Δmsc1). Airyscan superresolution analysis of cdc15-2 Δmsc1 co-expressing Sec61-CFP and Nup49-mCherry revealed that the septum separating the apparently partitioned late-M nucleus is in fact made of NE ( ~90% of these cells have a septum where Sec61 and Nup49 colocalize, n = 237 cells) (Figs. 2c, S4 and movie S1). Strikingly, the septum often appeared to be closed, as far as the Airyscan superresolution can determine. More strikingly, both compartments contain DNA in nearly 90% of these examples (n = 143 cells) (Figs. 2d and S5). The proportion of aberrant late-M nuclei did not change after DSB generation in cdc15-2 MSC1 and cdc15-2 Δmsc1 strains (Fig. 2a).

[See PDF for image]

Fig. 2

Role of Msc1 in late-M NE shape.

a and b Abnormal NE blebs and partitions in the cdc15-2 MSC1 and cdc15-2 Δmsc1 late-M cells without (mock) and with (HO or phle) DSBs. The experimental setup is described in Fig. S2. a Quantification of blebs and partitions (mean ± s.e.m., n = 3). b Three representative micrographs of normal dumbbell late-M shapes (top), and three representative micrographs of abnormal shapes (bottom) with partitions indicated by arrowheads. c and d High resolution of nuclear partitions in cdc15-2 Δmsc1 late-M cells. Note that the NE septum contains NPCs and partitions split the nuclear DNA mass. c Cells expressing Sec61-eCFP and Nup49-mCherry were fixed and visualized by confocal microscopy with airyscan 2 superresolution. A central slice of a representative cell with one of the segregating nuclear bodies presenting a partition indicated by the arrowhead (left) and a 3D reconstruction of the same cell (right). d Cells expressing Sec61-eYFP were stained with DAPI and visualized as above. A central slice of a representative cell with a partition indicated by the arrowhead that contains DNA in both lobes (left). 3D reconstruction of the same cell (right) e Msc1 forms patches at NE blebs. A short time-lapse series of a late-M cdc15-2 MSC1:eYFP cell after the DSB generation. Note how Msc1-eYFP concentrates on NE blebs (arrowheads) and dynamically corrects this aberration. f and g Short time-lapse series of late-M cdc15-2 cells expressing Sec61-eCFP and Nup49-mCherry. Arrowheads point at nuclear partitions when they are present. f Illustrative transient segmented nucleus observed in a cdc15-2 MSC1 cell. g Example of stable segmented nucleus observed in a cdc15-2 Δmsc1 cell. h Quantification of cells displaying a segmented late-M nucleus throughout the entire 5-min time-lapse movie in mock- and phle-treated cdc15-2 MSC1 and cdc15-2 Δmsc1 cells (mean ± s.e.m., n = 3; ~100 cells per condition). Scale bars correspond to 3 μm; BF, bright field. The unpaired t test was used for statistical comparisons (*** for p < 0.001, ** for p < 0.01 and * for p < 0.05).

In the previous work, we reported that Msc1 is not uniformly localized throughout the NE but forms concentrated patches, with a correlation between the presence of patches and DSB repair factories²¹. When we looked at abnormal nuclear shapes in the cdc15-2 MSC1, we found that Msc1-YFP appears enriched in NE blebs (or nuclear bulges), and short time lapse movies suggest that Msc1 may participate in correcting these herniations, thus re-establishing a rounder NE and preventing partitions (Figs. 2e, S6 and movie S2). Indeed, we observed that NE blebs and partitions are dynamically formed and resolved during the cdc15-2 MSC1 arrest, while remaining more stable in the cdc15-2 Δmsc1 strain (Figs. 2f–h, S7 and S8; movies S3 and S4).

Msc1 favors back migration of sister chromatids during DNA repair in late mitosis

One logical consequence of the observed nuclear aberrations is that a proportion of late-M cells, which is higher in Δmsc1, is physically disabled from finding the correct template for HR once DSBs occur. Since this step requires the back migration of sister loci through the bud neck¹⁰, we compared both back migration and coalescence events in cdc15-2 Δmsc1 versus cdc15-2 MSC1 after phleomycin (Fig. 3). In cdc15-2 MSC1, we observed a significant increase in these events at the right telomere of chromosome XII (cXIIr-Tel; from 5 to 15% in late-M cells after phleomycin). However, this increase was about half in cdc15-2 Δmsc1 (Fig. 3a–c). Likewise, the formation of retrograde chromatin bridges after phleomycin was largely impaired in cdc15-2 Δmsc1 (Fig. 3d–f). Chromatin was visualized through the histone A2 (Hta2-mCherry), and it went from more than a two-fold increase in bridges in cdc15-2 after DSB generation (from ~25 to ~60%) to half that increase in cdc15-2 Δmsc1 (from ~20 to 30%). These experiments confirmed that Δmsc1 late-M cells are less capable of bringing together the segregated sister chromatids upon DSBs.

[See PDF for image]

Fig. 3

Back migration and coalescence of sister loci is hampered in Δmsc1.

acdc15-2 MSC1 and cdc15-2 Δmsc1 cells carrying a TetR-YFP/tetOs construct that labels the right telomere of chromosome XII (cXIIr-Tel) were treated as in Fig. S2. Samples were taken 1 h after phle (or mock) addition, and late-M cells classified in three categories: a, segregated sister cXIIr-Tels; b, back migration of one sister cXIIr-Tel towards the other; c, coalescence between sister cXIIr-Tels (mean ± s.e.m., n = 3). Categories b and c are represented together because they are rapidly interchangeable. b Phle:mock ratios of cXIIr-Tel retrograde events in cdc15-2 MSC1 and cdc15-2 Δmsc1 late-M cells. c Representative micrographs of the three quantified categories. A short frame of a 6” movie is shown to better appreciate the dynamism of coalescence (bottom line). Arrowheads point at the cXIIr-Tel sisters. d As in (a) but with strains that label the bulk of chromatin with Hta2-mCherry (mean ± s.e.m., n = 3). Late-M cells were classified in three categories again: a, no chromatin bridge; b, thin bridge; c, bridge with bulges. Categories b and c are represented together. e Phle:mock ratios of retrograde chromatin bridges in MSC1 and Δmsc1 late-M cells. f Representative micrographs of the three quantified categories. Two examples of each are shown. Arrowheads point to a chromatin bulge often seen in retrograde bridges. Scale bars correspond to 3 μm; BF, bright field. The unpaired t test was used for statistical comparisons (*** for p < 0.001, ** for p < 0.01, * for p < 0.05, and n.s. for p > 0.05).

We have previously shown that DSBs in late-M exert a partial regression of anaphase, with shortened distances between segregated sister centromeres, dismantlement of the anaphase spindle and redistribution of the spindle motor type 5 kinesin Cin8 toward the poles¹⁰. Thus, we also checked whether these anaphase regression phenotypes were affected in Δmsc1. We looked at these three markers of regression, observing no differences between cdc15-2 MSC1 and the cdc15-2 Δmsc1 strains before and after DSB generation (Fig. S9). First, the anaphase spindle (as reported by GFP-Tub1) was seen in 70-80% of late-M cells, whether MSC1 or Δmsc1 (Fig. S9a). DSB generation by phleomycin reduced these percentages to 30%, although this reduction occurred equally in both strains. A similar pattern was seen for Cin8-mCherry; nearly 90% was found on the spindle in late-M cells for cdc15-2 MSC1 and cdc15-2 Δmsc1 cells, with an abrupt drop to just 20% after phleomycin, with no distinction between MSC1 and Δmsc1 (Fig. S9b). Lastly, the distance between sister chromosome XII centromeres was shortened after phleomycin, but to a similar extent in MSC1 and Δmsc1 (Fig. S9c). Thus, we concluded that the reduction in chromatin bridges and coalescence of sister loci at chromosome arms is not due to the inability to relieve the late anaphase segregating tension by the spindle apparatus. This conclusion supports in turn the potential role of NE partitions in this defect. In agreement with this, we found instances where cXIIr-Tel sisters were trapped within different partitions (Fig. S5e).

The NE maintenance complex ESCRT-III is important for DSB repair

In addition to its role in karyokinesis, SpLes1 might have a more general role in the NE homeostasis. For instance, the nuclear-cytoplasmic barrier is compromised in les1Δ, which is in turn synthetically lethal with several components of the evolutionarily conserved ESCRT-III complex in S. pombe^13,16. The basis of this extreme negative genetic interaction is that ESCRT-III repairs NE wounds that occur in the absence of SpLes1¹³. Indeed, this is one of the most important functions of ESCRT-III in eukaryotes, i.e., surveys NE integrity and seals unwanted NE pores^14,15,30. We envisioned that a similar scenario might be taking place in S. cerevisiae without Msc1, with a likely aggravated circumstance due to the overstretched NE at late-M.

In S. cerevisiae, ESCRT-III comprises up to 8 subunits, with Snf7 being the core structural protein, and the Did4/Vps24 dimer forming a major regulatory player^14,31. The ESCRT-III machinery is completely absent in the Δsnf7 mutant, whereas it is malformed and non-functional in either Δdid4 or Δvps24. In addition, the ESCRT-III complex is targeted to NE repair sites by Chm7, presumably through binding to NE blebs and herniations³². To check whether ESCRT-III and Chm7 are involved in DSB repair, we determined sensitivity to phleomycin through both spot assays and growth curves (Figs. 4a and S10). Loss of function of the complex led to hypersensitivity to phleomycin, with Δsnf7 giving the strongest phenotype. In spot assays, Δsnf7 was less sensitive than Δmsc1 in our YPH499 genetic background, with Δmsc1 phenocopying the HR-deficient Δrad52 (Figs. 4a and S10). Similar sensitivities between Δsnf7 and Δmsc1 were found for DSBs generated at the HO locus (Fig. S11). Interestingly, in the BY4741/S288C background, phleomycin hypersensitivity of Δsnf7 was stronger than Δmsc1 and Δrad52 (Fig. S12). The other two ESCRT-III mutants gave a milder, yet consistent, sensitivity in both genetic backgrounds. Surprisingly, Δchm7 was not sensitive to phleomycin (Fig. 4a).

[See PDF for image]

Fig. 4

Mutants for the NE healing complex ESCRT-III partly phenocopy Δmsc1.

a Spot assay against phleomycin of ESCRT-III mutants. The HR mutant Δrad52 and Δmsc1 were included as references. Note that the Δsnf7 (the core component of ESCRT-III) is more sensitive to DSBs than the others, whereas Δchm7 (ESCRT-III recruiter to NE) is not. b Phle:mock ratios of retrograde chromatin bridges in cdc15-2 (WT) and ESCRT-III-deficient late-M cells (mean ± s.e.m., n = 3). c Abnormal late-M NE morphologies in ESCRT-III mutants (mean ± s.e.m., n = 3). Four categories were assessed based on Sec61-eYFP: a, normal hourglass shape; b, partitioned; c, blurred; d, multivesiculated. The strain also carries the Nup49-eCFP construct. d Representative micrographs of the four categories. All strains, including those labeled as “WT”, harbor the cdc15-2 allele. Scale bars correspond to 3 μm; BF, bright field. The unpaired t test was used for statistical comparisons (**** for p < 0.0001, ** for p < 0.01).

ESCRT-III preserves the NE in late mitosis and facilitates anaphase retrograde events after DSBs

To check whether the DSB hypersensitivity of ESCRT-III mutants was mechanistically related to that of Δmsc1, we next phenotypically examined late-M cells after DSBs in these mutants. Firstly, we quantified retrograde events for cXIIr-Tel, and found that the phleomycin vs. mock ratio was similar between cdc15-2 (WT reference in the figures) and the mutants (~2.5) (Fig. S13a,b). On the contrary, the increase in retrograde chromatin bridges (Hta2-mCherry) after phleomycin was more modest in the ESCRT-III mutants (1.5 vs. 2.5 ratios), although cdc15-2 Δchm7 was similar to cdc15-2 (Figs. 4b and S13c). Thus, with these mutants, we found for the first time a loss in the positive correlation between sister telomere retrograde events and the presence of de novo chromatin bridges. One possible explanation for this is that only the retrograde movement of chromatin not bound to the NE is affected since cXIIr-Tel is attached to the NE. However, it is important to note that this is based on the comparison of phle:mock ratios. In terms of global cXIIr-Tel retrograde events after phleomycin, there were fewer events in Δsnf7 (Fig. S13a).

Next, we checked NE morphologies in these ESCRT-III mutants to determine if they phenocopied that of Δmsc1. Unexpectedly, partitions were as rare as in the cdc15-2 (WT) strain (Fig. 4c, d). Instead, two other aberrant morphologies were observed. On the one hand, around ~20% of late-M cells presented blurred Sec61 and Nup49 signals, which gave these extended nuclei the appearance of being formed by discontinuous patches of NE (Fig. 4d, category “c”). Late-M blurred nuclei were seen in all ESCRT-III mutants, including cdc15-2 Δchm7. On the other hand, in ~5% of late-M cdc15-2 Δsnf7 cells, Sec61 appeared multivesiculated (Fig. 4d, category “d”). These aberrant morphologies preceded the induction of DSBs and their relative percentages remained after adding phleomycin (Fig. 4c). Thus, if these NE aberrations affect DSB sensitivity, they likely yield an inadequate nuclear architecture for later DSB repair.

ESCRT-III core component Snf7 is synergistic with Msc1 in DSB repair and late-M DSB-associated events

Having shown that Δmsc1 and ESCRT-III mutants, especially Δsnf7, individually hypersensitized to DSBs, and that they both coincide in having aberrant yet distinct late-M nuclei, we next tested whether they synergically change these phenotypes. Noticeably, S. pombe les1 is synthetically lethal with deletions in ESCRT-III genes¹⁶. Thus, to combine depletion of Msc1 and Snf7 in the same strain we opted for tagging Snf7 with the auxin-inducible degron minimal sequence (aid*) in the Δmsc1 strain. Snf7-aid* was quickly degraded after adding the auxin indol-acetic acid (IAA), and virtually no protein could be detected by Western blotting after 1 hour (Fig. S14a). Importantly, the Snf7-aid* chimera was functional, as the SNF7:AID* strain grew as efficiently as the strain without the C-terminal tag and better than the Δsnf7 mutant (Fig. S14b). Likewise, SNF7:AID* strains were resistant to phleomycin (Fig. S14b–e). We then tested whether there exists a negative genetic interaction between MSC1 and SNF7. In spot assays, we found that the strain carrying Δmsc1 SNF7-AID* OsTIR1 (OsTIR1 is required for auxin to target aid* for degradation³³) was unaffected by IAA, and cells grew as well as in isogenic strains without the OsTIR1 (Fig. S14d). This result is thus in marked contrast with what has been reported in S. pombe¹⁶. Likewise, a complete lack of negative genetic interaction was seen for Δmsc1 and depletion of either Did4 or Vps24 (Fig. S15).

Next, we investigated through spot assays if a negative effect in these double mutants occurred upon DNA damage with phleomycin (Fig. S14e and S15b-d). We found that, in conditions where DNA damage did not increase the sensitivity of MSC1 cells depleted of Snf7-aid* relative to cells having both Msc1 and Snf7 (MSC1 SNF7-AID* OsTIR1 vs. MSC1 SNF7-AID*; 5 mM IAA, Phle 0.05–0.1 μg/mL), cells defective in both proteins (Δmsc1 SNF7-AID* OsTIR1) were more sensitive than those only defective in Msc1 (Δmsc1 SNF7-AID*) (Fig. S14e). Thus, these spot assays point out that there is a synergistic effect of depleting both proteins in relation to sensitivity to DSBs. This genetic interaction was not observed when depleting the two other ESCRT-III subunits, Did4 and Vps24 (Fig. S15c,d). However, the combination of auxin and phleomycin diminished the relative toxicity of the latter in all strains (Fig. S14e and S15c,d). Therefore, and considering the absence of negative genetic interactions between Δmsc1 and ESCRT-III depletions (Fig. S14d, S15c,d and S16), we also constructed double knockout mutants for MSC1 and the ESCRT-III genes, including CHM7 (Fig. 5a). With these double mutants we recapitulated the synergistic hypersensitivity to phleomycin in the Δmsc1 Δsnf7 combination, but also in Δmsc1 Δdid4 and Δmsc1 Δvps24 (see 0.25 μg/mL plate). Notably, this synergism was absent in Δmsc1 Δchm7.

[See PDF for image]

Fig. 5

Synergism profiles of Msc1 and ESCRT-III in DSB sensitivity and late-M events.

a Spot assay against phleomycin of ESCRT-III Δmsc1 double mutants. The reference strain (cdc15-2) and the corresponding single mutants were included to assess genetic interactions. b Phle:mock ratios of retrograde chromatin bridges visualized with Hta2-mCherry in MSC1 SNF7, Δmsc1 SNF7, MSC1 Δsnf7 and Δmsc1 Δsnf7 late-M cells (mean ± s.e.m., n = 3). c As in (b) but following abnormal late-M NE morphologies through Sec61-eYFP (mean ± s.e.m., n = 3). Categories as in Fig. 4c. d Cells displaying a segmented nucleus throughout the entire 5 min short time-lapse. As in Fig. 2h, but including the double mutant Δmsc1 Δsnf7 (mean ± s.e.m., n = 3; ~100 cells per condition).

To check whether synergistic effects are also present for retrograde events in late-M, we determined approximation of sister cXIIr-Tel loci and retrograde chromatin bridges after phleomycin, and compared Δmsc1 cells depleted or not of Snf7. We did this with both the double knockout mutant Δmsc1 Δsnf7 and the Δmsc1 SNF7-AID* OsTIR1 strain (Figs. 5b, S17 and S18a–c). In these experimental conditions, the reference cultures just without Msc1 (Δmsc1 SNF7 or -iaa subcultures in Δmsc1 SNF7:AID*) had a behavior similar to Δmsc1 strains in previous experiments, with a ~ 2- and ~1.5-fold increase in cXIIr-Tel approximation and chromatin bridges, respectively, in phleomycin relative to the mock (Figs. 5b, S17 and S18b,c). However, all retrograde events were fully prevented when cells were further depleted from Snf7 (Δmsc1 Δsnf7 or +iaa subcultures in Δmsc1 SNF7:AID*). Thus, Msc1 and Snf7 (ESCRT-III) co-operate in bringing together sister loci in late-M after DSBs.

Lastly, we investigated whether there was also synergism in the aberrant late-M NE structures seen for each mutant separately (Figs. 5c, d and S18d). In this case, the partition phenotype observed in Δmsc1 fully dominated over those seen after Snf7 depletion. Indeed, neither blurred nor multivesiculated Sec61/Nup49 signals were seen in Snf7 Msc1 co-depletion. This result suggests that these Δsnf7 NE signatures depend on Msc1. On the other hand, the Δmsc1 NE signature, i.e., partitions, was not only present but slightly increased after the co-depletion (Figs. 5c and S18d). The dynamics of partitions in Δmsc1 Δsnf7 was equivalent to that of Δmsc1 SNF7, with up to 80% of partitioned late-M nuclei stable in the short term (Fig. 5d). Once again, these NE phenotypes manifested upstream DSBs (Figs. 5c, d and S18d).

Late-M nuclear pore complex distribution is affected by Msc1 and ESCRT-III

Alongside the defects seen in the NE by Sec61, NPCs (Nup49) were largely mis-distributed and tend to concentrate in spots in the subset of late-M cells affected by the blurred NE in ESCRT-III (Fig. 4d). Interestingly, this uneven distribution was also observed in a proportion of cells in which the NE could be outlined (Fig. 6a). The total percentage of late-M cells with an uneven distribution of Nup49 was as high as 60% in cdc15-2 Δsnf7 (and ~40% in cdc15-2 Δdid4 and cdc15-2 Δvps24) compared to less than 10% in cdc15-2 (WT) (Fig. 6b). Strikingly, proper NPC distribution was seen in cdc15-2 Δchm7 (only ~15% misdistribution). On the other hand, the level of misdistribution was also high in cdc15-2 Δmsc1 (~40%) The percentage of these aberrant nuclei did not change in any of these mutants after generating DSBs with phleomycin. We next checked whether this defective distribution of NPCs could be synergistically increased in the Δmsc1 Δsnf7 and Δmsc1 SNF7-AID* OsTIR1 strains (Figs. 6b and S19). Depletion of Snf7 in this Δmsc1 background increased Nup49 misdistribution to ~75% (in cdc15-2 Δmsc1 Δsnf7). This increase is slightly lower than expected from an additive effect of each individual mutation (and was not observed upon +iaa depletion of Snf7-aid*; Fig. S19), which suggests partially independent contributions of Msc1 and ESCRT-III, or even a putative positive epistasis.

[See PDF for image]

Fig. 6

Nuclear pore complex distribution in strains depleted of the ESCRT-III complex and Msc1.

a Representative micrographs of even (a), single uneven (b), and multiple uneven (c) distribution of Nup49 across the late-M NE. The arrowhead points to the characteristic patch of enriched Nup49-eCFP seen when distribution was uneven. b Late-M distribution of Nup49 in the cdc15-2 (WT), cdc15-2 Δmsc1, four ESCRT-III mutants and the cdc15-2 Δmsc1 Δsnf7 double mutant with (phle) or without (mock) concomitant DSBs (mean ± s.e.m., n = 3). c Representative micrographs of NE-localized (a), cytoplasmic focus (b) and cytoplasmic foci (c) of Nup49 in late-M cells. The arrowheads indicate each cytoplasmic focus of Nup49. d Late-M localization of Nup49 in same strains and conditions as in panel (b). All strains, including the one labeled as “WT”, harbor the cdc15-2 allele. Scale bars correspond to 3 μm; BF, bright field.

Interestingly, in addition to NE misdistribution, we also observed Nup49-eCFP spots in the cytoplasm (Fig. 6c), as if a subset of nucleoporins could not assemble into NPCs at the NE. In fact, these two phenotypes have been reported together for mutants that disrupt proper NPC formation and stoichiometry^34,35. Cytosolic Nup49 puncta were present in ~10% of cdc15-2 late-M arrested cells, and this percentage increased to ~20% in cdc15-2 Δmsc1, whereas this level was significantly higher in ESCRT-III mutants, including cdc15-2 Δchm7 (Fig. 6d). Similar to the NPC clustering phenotype, cytosolic nucleoporins were highest in the cdc15-2 Δsnf7 and cdc15-2 Δmsc1 Δsnf7 strains.

NE partitions and NPC misdistribution also occur in both asynchronous and G2/M-arrested cells in mutants for MSC1 and the ESCRT-III complex

So far, we have studied the phenotypes present in cells arrested in late-M, since Msc1 was identified in a screen for DSB repair in this cell cycle stage²¹. We next asked what happens to the NE and NPCs in asynchronous cells and during the phleomycin-induced G2/M (metaphase-like) arrest³⁶. In asynchronous cells, NE partitioning was again the major abnormality observed in Δmsc1 (~20%) (Fig. 7a). Interestingly the ratio between MSC1 and Δmsc1 was even higher than in late-M ( ~ 5:1). Partitions were also observed in the ESCRT-III mutants, especially in Δsnf7, although to a lesser extent. Interestingly, the other two phenotypes described in late-M, blurred and multivesiculated Sec61, were completely absent. The proportion of partitions in Δmsc1 Δsnf7 is equal to that in Δmsc1 SNF7 (i.e., MSC1 is epistatic to SNF7). Regarding NPC distribution, both Δmsc1 and Δsnf7 showed four times more NPC clustering than the reference strain (~40 vs. ~10%), with the double mutant peaking at ~60% (positive epistasis) (Fig. 7b). Intermediate values were obtained after mutating the other two ESCRT-III subunits.

[See PDF for image]

Fig. 7

Nuclear envelope abnormalities in the absence of Msc1 and the ESCRT-III complex are also present in asynchronous populations and the G2/M arrest elicited by DSBs.

a Abnormal asynchronous NE morphologies in MSC1 DID4 SNF7 VPS24, Δmsc1, Δdid4, Δsnf7, Δvps24 and Δmsc1 Δsnf7 (mean ± s.e.m., n = 3). Four categories were assessed based on Sec61-eYFP; from left to right in the X-axis: normal round or bow-tie shapes; partitioned; blurred; multivesiculated. b Abnormal Nup49 distribution in MSC1 DID4 SNF7 VPS24, Δmsc1, Δdid4, Δsnf7, Δvps24 and Δmsc1 Δsnf7 asynchronous cells (mean ± s.e.m., n = 3). A distinction is made between categories of even and uneven distribution. c Percentage of G2/M-arrested cells in the WT and ESCRT-III mutants after 3 h and 5 h of incubation with phleomycin. (mean ± s.e.m., n = 3). Note that the ESCRT-III does not affect the arrest efficiency. G2/M-arrested cells were assessed by the presence of mononucleated dumbbell-shaped cells. The mononucleated nucleus could be either in one cell body or across the neck (bow-tie phenotype). d Representative micrograph with two G2/M-arrested cells with the two distinct NE shapes, round nucleus in one body (top) and with the bow-tie shape (bottom). e Cell viability of Δmsc1 and ESCRT-III mutants after short-term G2/M blocks by DSBs (mean ± s.e.m., n = 3). Percentage of survivors is relative to the isogenic WT. f Abnormal G2/M NE morphologies in Δmsc1 and ESCRT-III mutants at the G2/M arrest elicited by DSBs (mean ± s.e.m., n = 3). Categories as in (a). Representative micrographs for the normal and partitioned shapes are included underneath. The arrowhead points to the constriction that makes the partition. g As in (f) but looking at the NPC (Nup49) distribution (mean ± s.e.m., n = 3). Representative micrographs for the even and uneven distribution are shown below. Scale bars correspond to 3 μm; BF, bright field. ANOVA followed by the Dunnett’s test against WT (cdc15-2) was used for statistical comparisons (**** for p < 0.0001, *** for p < 0.001, * for p < 0.05).

Next, we examined NE partitions and NPC clustering after generating DSBs in the asynchronous culture, which induces a G2/M arrest by the DDC³⁶. We first checked whether the G2/M arrest was as effective in the mutants as is in the WT, observing similar cell percentages after 3 h and 5 h of phleomycin treatment (70-80%) (Fig. 7c). In this G2/M arrest, cells appear mononucleated with either a round nucleus close to the bud neck or traversing it. In the latter case, the NE appears as bilobed with the neck making the constriction (bow-tie phenotype) (Fig. 7d). When G2/M cells were released from DNA damage (phleomycin removal), there was a drop in viability for all mutants relative to the WT (Fig. 7e). Only 60% of Δmsc1 cells survived compared to the WT, whereas this percentage dropped to 20% for Δsnf7. Intermediate survival was obtained with the other two ESCRT-III subunits. This sensitivity profile fitted well with what was observed in spot assays above and emphasizes the importance that both Msc1 and ESCRT-III has in an effective DSB repair, beyond the particular cell cycle stage the cells are in when they receive the insult. We next addressed the aberrant NE phenotypes described above in late-M. Partition was again the major abnormality observed in Δmsc1, reaching up to 30% of G2/M cells (Fig. 7f). This led to a MSC1/Δmsc1 ratio even higher than in late-M and asynchronous (~7:1 after 3 h in phle). Partition was also observed in the ESCRT-III mutants and blurred and multivesiculated Sec61 signals were also absent in G2/M (Fig. 7f). Because both the asynchronous and the G2/M arrest experiments were performed at 25 °C to keep Cdc15-2 active, whereas the late-M experiments were performed at 34 °C, we checked the influence of the temperature in the NE in the Δsnf7 mutant. We did not observe the blurred NE either (Figure S20). This points out that the blurred NE is a late-M specific phenotype of the Δsnf7 mutant. On the other hand, G2/M NPC misdistribution was also observed for all mutants (Fig. 7g). In this case, Δsnf7 surpasses Δmsc1 (~50 vs. ~35%), with values for both highly above the WT ( <10%) and the other two ESCRT-III mutants (~20%).

Finally, we investigated the subcellular location and abundance of Snf7 before and after DSBs (Fig. S21). For the subcellular location, we C-terminally tagged Snf7 with eYFP in a strain that also bore Nup49-mCherry as a NE reporter and Hta2-eCFP as a reporter of chromatin. Snf7-eYFP appeared concentrated in cytosolic spots but was not functional (Fig. S21a, b), making us cautious about drawing conclusions about Snf7 location. Interestingly though, single cell quantification showed that Snf7-eYFP abundance was rather variable in asynchronous cultures, and its levels increased after phleomycin treatment (Fig. S21b, c). This increase in Snf7 levels was further confirmed by Western blot against a Snf7 tagged with the HA epitope, which is functional (Fig. S21a, d, e).

In conclusion, both Msc1 and Snf7 are not only important for a proper repair of DSBs in late-M but also in other phases of the cell cycle, in which DSBs elicit a G2/M arrest. In this arrest, NE and NPCs abnormalities are also seen in cells depleted from Msc1 and/or Snf7.

Rad52 DSB repair factories are not increased in ESCRT-III-deficient late-M cells, yet there is a loss of nuclear Rad52

The hypersensitivity of late-M cdc15-2 Δmsc1 cells to DSBs appears to be due to slower HR repair, which includes fewer Rad52 repair factories (seen as nuclear foci) than the reference cdc15-2 strain²¹. We thus investigated whether a similar phenotype could be seen in ESCRT-III mutants. However, neither cdc15-2 Δsnf7 nor cdc15-2 Δchm7 showed fewer Rad52 foci relative to the reference after DSBs (Fig. 8a, b). In all cases, ~10% of late-M arrested cells presented a Rad52 focus, whereas Rad52 factories increased to ~40% after DSBs (~20% with one focus and ~20% with two or more foci). There were even more factories in cdc15-2 Δsnf7 (~50%), indicating more persistent damage.

[See PDF for image]

Fig. 8

Depletion of Chm7 and Snf7 affects the subcelular localization of Rad52 but does not prevent the formation of repair factories.

a Rad52 foci in CHM7 SNF7, Δchm7 and Δsnf7 strains with (phle) and without (mock) DSBs generation after the late-M arrest. Three categories were distinguished: a no focus; b a single focus in the cell; c two or more foci in the cell (mean ± s.e.m., n = 3). b Representative micrographs of the three quantified categories. Arrowheads indicate Rad52 foci. c Subcellular Rad52 localization in CHM7 SNF7, Δchm7 and Δsnf7 strains with (phle) and without (mock) DSBs generation after the late-M arrest. Cells were classified based on exclusively nucleoplasmic (a) or nucleocytoplasmic (b) localization, with the latter category characterized by the presence of both nuclear and cytoplasmic signals (mean ± s.e.m., n = 3). d Representative micrographs of the two quantified categories. Scale bars correspond to 3 μm; BF, bright field. ANOVA followed by the Dunnett’s test against cdc15-2 CHM7 SNF7 was used for statistical comparisons (**** for p < 0.0001, *** for p < 0.001).

Strikingly, we observed in the mutants a leakage of nuclear Rad52 into the cytoplasm (Fig. 8c, d). This leakage was present before DSB generation and did not change after phleomycin. This leakage phenotype is consistent with NE damage, confirming that cdc15-2 Δsnf7 has this damage, although levels were lower than in late-M cdc15-2 Δchm7 cells (~20 vs. 35%).

Discussion

In S. cerevisiae, HR remains active despite sister chromatids being already segregated^9,10. This predicts that there is a cell cycle window in late mitosis (anaphase and telophase) in which HR-driven DSB repair could be highly mutagenic. However, it appears that yeast cells might still use the intact sister as a HR template by partly reversing segregation¹⁰. The molecular players that reinforce HR with the sister in late-M are loosely defined. In a previous work, we identified the spindle machinery and the kinesin 5 Cin8 as important players in ensuring partial regression. Since yeast performed a closed mitosis and sisters must transverse back the narrow bud neck during regression, we theorized that the NE ought to be important as well. Accordingly, we subsequently identified through proteomics the poorly-characterized NE protein Msc1 as a late-M player for HR and DSB repair²¹. Msc1 levels increase after DSBs in late-M, and Δmsc1 mutants delay DSB repair. Mechanistically, Msc1 appears to affect either the formation of the presynaptic filament or the homology search afterward. Regardless, Msc1 functions downstream of the DSB end resection and the establishment of the DDC²¹. We herewith show that the Msc1 functional globular part faces the perinuclear space (lumen) (Fig. 1 and S1), like its S. pombe orthologs²². This fact implies that the action of Msc1 on HR must be molecularly indirect (i.e., Msc1 physically interacts with neither DNA/chromatin nor HR proteins). In this sense, the morphological changes of the late-M nucleus in this mutant are remarkable (Figs. 2, S4 and S5), and Msc1 appears to dynamically correct these aberrations in the elongated late-M nucleus (Figs. 2e, S6 and S7). It is not lost on us that the most striking aberration, the partition/septation/segmentation (inner over-compartimentalization) of the late-M nucleus (Figa. 2a–c and S4), which also involves the compartmentalization of part of the segregated DNA (Figa. 2d and S5), should retard or completely prevent the full mobilization of chromatids that ensues the DSB, resulting in the inability of the resected DSB ends to find the sister chromatid (Figs. 3 and S18). This scenario would explain the late-M special role, which could stem from the need to recruit a sister template that is no longer aligned (Fig. 9 for a model). In addition, the elongated and stressed late-M nucleus may have a greater need to protect the NE structure, with Msc1 being recruited to compromised areas to restore the hemisphericity of the two nuclear halves (Figs. 2e and S6). This need may be less urgent in the more spherical G2/M nucleus, which is likely less prone to such severe partition, although we still observed this phenotype in an asynchronous population and the G2/M arrest elicited by DSBs (Fig. 7). A seemingly obvious corollary of this model for late-M cells would be premature karyokinesis, which could indeed stem from the partition itself, but would also be expected based on the previous results obtained with SpLes1¹³. However, we found modest evidence of karyokinesis completion in cdc15-2 Δmsc1, only 5–10% higher than in cdc15-2 MSC1 (Fig. S3). Even in extreme cases, nuclear compartments and bulges/blebs still seem connected to each other in the late-M nucleus (Figa. 2, S4–S8; movies S1–S4). Alternatively, the nuclear bridge could be affected in other ways that prevent sister chromatids from moving across it. For example, SpLes1 localizes to the bridge stalk before karyokinesis and corrals NPC at the midzone of the bridge¹³. Thus, the Msc1 deficiency could also interfere with the movement of NPC-attached DSBs in their search for a template. In support of this, we found signs of NPC mislocation in Δmsc1 (Figs. 7 and S19). In addition to defects in the mobilization of the whole chromatin, free circulation of the repair factors/factories could also be lessened. Indeed, physical interference with the location of repair factors has been suggested as the cause of DNA repair impairments in Hutchinson-Gilford progeria syndrome fibroblasts, which are also characterized by abnormal nuclei with lobes, invaginations and compartmentalization, creating unfavorable regions in the nucleus for DSB repair³⁷.

[See PDF for image]

Fig. 9

Models of how Msc1 and ESCRT-III may influence DSB repair in late-M.

a In late-M, the nucleus assumes an extended dumbbell shape with already segregated sister chromatids (a single pair of sisters is depicted for simplicity). (1) After a DSB occurs in one of the sisters, the extended nucleus becomes shorter and the bridge becomes thicker. The sisters also move closer and can coalesce for HR repair of the DSB. The broken ends can also get attached to the NE to facilitate retrograde events, and NPCs may participate in this attachment. (2) If segregated sisters are physically partitioned, as it occurs in Δmsc1, there are late-M cells in which sister chromatid approximation is challenged. A similar scenario may happen in NE invaginations in ESCRT-III mutants. (3) Alternatively, or in addition to partitions, NPC maldistribution and aggregation, which are observed in both mutants, could challenge DSB processing and template search. b A dynamic model for the role of Msc1 in preventing late-M nuclear partitions. Msc1 would remove NE blebs before they segment nuclear areas.

ScMsc1 and SpLes1 are poorly-characterized NE proteins, in part due to the apparent lack of orthologs in higher eukaryotes. Nevertheless, SpLes1 is synthetically lethal with the highly conserved ESCRT-III complex¹⁶. The fact that the absence of SpLes1 leads to transient NE ruptures, and the subsequent loss of the nuclear-cytoplasmic barrier in the extended late-M nucleus, suggests that the underlying nature of this negative genetic interaction resides in the need of sealing the ruptures by the ESCRT-III complex^13,14. It is fairly known that NE healing is one of the main cell functions of this versatile complex. For instance, HeLa cells that elongate their interphase nucleus to migrate through narrow spaces heavily rely on ESCRT-III to preserve NE integrity^38,39. Likewise, deregulation of ESCRT-III causes aberrant nuclear structures and micronuclei in HeLa and D. melanogaster cells^{39, 40–41}, as well as invaginations of the INM in C. elegans⁴². Because of the genetic interaction in S. pombe, and because the NE phenotypes resemble in part what we observed in the absence of Msc1, we decided to study the role of ESCRT-III in late-M DSB repair and its relationship with Msc1 (Figs. 4–8 and S10–S21). Indeed, mutants for the core (Δsnf7) and regulatory (Δdid4 and Δvps24) components of ESCRT-III were hypersensitive to DSBs (Figs. 4a, S10–S12, S14, S15), and less capable of forming retrograde anaphase bridges after DSBs in late-M (Figs. 4b, S13 and S15). These mutants also rendered partitioned nuclei, but only in an asynchronous population and at the G2/M arrest that ensues DSBs (Fig. 7). In late-M, partition was not seen and, instead, the major abnormal phenotype was a blurred NE (Fig. 5c, d), which in turn was absent in asynchronous and G2/M arrests (Fig. 7). The most likely explanation for this difference is that partitioned nuclei do form in ESCRT-III mutants but, unlike in Δmsc1, they quickly collapse into the blurred nucleus in late-M. On the other hand, we looked into DSB sensitivity and NE reshaping when both ESCRT-III and Msc1 were absent (Figs. 5, S14–S16 and S18). We approached this by conditionally degrading ESCRT-III components in the Δmsc1 background as well as by double knockout mutants. In the absence of DSBs, we found no negative genetic interaction between MSC1 and ESCRT-III genes (Figs. 5a, S14, S16 and S17), in contrast to what has been reported for S. pombe¹⁶. Remarkably, the Δmsc1 background suppresses NE phenotypes seen in the ESCRT-III knockouts (e.g., the blurred nuclei; Fig. 5c). Regardless, there was synergism in DSB sensitivity and suppression of late-M retrograde events in the double mutants (Figs. 5a, b and S15, S18), which suggest a cooperation of both factors in DSB repair.

Another major phenotype of ESCRT-III mutants, also shared in the Δmsc1 mutant, was the abnormal distribution of NPCs. This was seen in both the elongated NE of late-M cells and the more rounded NE of the G2/M arrest and asynchronous populations (Figs. 6 and 7). It has been reported that ESCRT-III surveys for and/or eliminates non-functional NPCs, which otherwise aggregate into the so-called storage of improperly assembled NPCs (SINC) compartment at the NE⁴². Thus, the Msc1/ESCRT-III deficiency could also interfere with the axis that biochemically regulates the repair of DSBs via NPCs, which involves post-translational modifications of HR factors at the NPC and attachment of irreparable DSBs to the pores^{20,43, 44, 45–46}. In addition, the SINC can also cause NE breakdown. Whether through its direct NE healing activity or through SINC- or Msc1-mediated NE breakdown, ESCRT-III could also facilitate DSB repair by simply ensuring the nuclear-cytoplasmic barrier. In this alternative model, disruption of the NE as a barrier, rather than the NE partition or NPC malfunction, would leak in and out the cytoplasmic and nuclear proteins that undermine and ensure DNA repair, respectively. In agreement with this hypothesis, we observed Rad52 leakage in Δsnf7 (Fig. 8). Remarkably, a strong association between major nuclear misshaping and SINC formation has been observed in other yeast mutants, including for processes as diverse as NPC assembly, chromatin remodeling, linkage of the nucleoskeleton to the cytoskeleton, and ER biology^{17,47, 48–49}; and this association extends to both human and plant cells for at least some of these processes^50,51.

To recapitulate, cells deficient in either the novel NE protein Msc1 or the membrane healing complex ESCRT-III show at least three NE abnormalities that could explain their hypersensitivity to DSBs: (i) NE blebs and partitions, (ii) NPC clustering, and (iii) signs of NE overpermeability. Any one of these phenotypes alone could be associated with impaired DSB repair as explained above (Fig. 9), combinations of them could further diminish DSB repair, and they could be linked to each other as well (i.e., NE partitioning could lead to NPC clustering and vice versa). The question remains on whether any one of the three overtakes the others in terms of its role in DSB repair. In this sense, the results with Chm7 depletion are puzzling but clarifying. Chm7 defines the canonical mechanism whereby ESCRT-III is recruited to the NE^14,52. Late-M cdc15-2 Δchm7 cells also show blurred NE and nucleocytoplasmic leakage (Figs. 4c and 8a), but are not more sensitive to DSBs (Fig. 4a). Strikingly, and compared to Δmsc1 and the other three ESCRT-III mutants, NPC clustering is less pronounced in Δchm7 (Fig. 6a). This would posit NPC abnormalities at the core of DSB repair defects in all of these mutants, although further research is required to confirm this hypothesis, since Δchm7 still exhibits nucleoporin abnormalities (cytosolic puncta), and we cannot rule out that ESCRT-III/Snf7 is recruited to the NE via non-canonical Chm7-independent pathways⁵³.

In conclusion, the NE protein Msc1 and the NE sealing complex ESCRT-III are critical for DSB repair in late-M and G2/M. Because the major phenotypes of their mutants affect the structure, integrity and shape of the NE, as well as the distribution of NPCs, we conclude that surveillance of the NE by these proteins is essential for DSB repair. To the best of our knowledge, this work presents for the first time the concept of active maintenance of the correct nuclear shape during DSB repair in yeast, and opens a new avenue of research in genome instability.

Materials and methods

Yeast strains and experimental conditions

Strains used in this work are listed in Tables 1 and S1. Strain construction was undertaken through transformation, and mostly involved gene deletion and C-terminal tagging⁵⁴.

Table 1. Strains used in this work

Name1	Genotype2	Origin
SLJ1838	MATα can1Δ::STE2p:SpHIS5 lyp1Δ his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2	Jaspersen lab
SLJ7465	MATα can1Δ::STE2pr-SpHIS5 lyp1Δ his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2; [NOP1p:GFP11:mCherry:SCS2TM::LEU2 (pSJ1568)]	Jaspersen lab
SLJ7859	MATα can1Δ::STE2p:SpHIS5 lyp1Δ his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2; [NOP1p:GFP11:mCherry:PUS1::LEU2 (pSJ1321)]	Jaspersen lab
SLJ8344	MATα can1Δ::STE2p:SpHIS5 lyp1Δ his3Δ1 leu2Δ0 ura3Δ0 met15Δ0 LYS2; [NOP1p:mCherry:SCS2TM:GFP11::LEU2 (pSJ1602)]	Jaspersen lab
FM2995	SLJ7859; MSC1:yGFP1–10::CaURA3MX	This study
FM2996	SLJ1838; MSC1:yGFP1–10::CaURA3MX	This study
FM2997	SLJ8344; MSC1:yGFP1–10::CaURA3MX	This study
FM2999	SLJ7465; MSC1:yGFP1–10::CaURA3MX	This study
AS499*	MATa bar1∆ leu2-3,112 ura3-52 his3-∆200 trp1-∆63 ade2-1 lys2-801	Strunnikov lab
FM2531*	AS499; cdc15-2:9myc::Hph; HMLα ∆hmr::HIS3MX; leu-2-3::lexOp:HO:FLAG:CYC1t::ACT1p:LexA-ER-B112:CYC1t::LEU2MX (pRG646)	Machín lab
FM2942	FM2531; SEC61:eYFP::KanMX	This study
FM2974	FM2531; Δmsc1::NatNT2; SEC61:eYFP::KanMX	This study
FM588	AS499; ade2-1::TetR:YFP::ADE2; tetO(5.6Kb)::1061Kb-ChrXII::HIS3; cdc15-2:9myc::HphNT1	Machín lab
FM2976	FM588; NUP49:mCherry::KanMX6; Δmsc1::NatNT2; SEC61:CFP::TRP1	This study
FM2831	FM2531; MSC1:eYFP::KanMX	This study
FM3149	FM2942; NUP49-eCFP::TRP1	This study
FM2807	FM588; Δmsc1::NatNT2	This study
FM2354	AS499; ade2-1::TetR:YFP::ADE2; tetO(5.6Kb)::194Kb-ChrXII::HIS3; cdc15-2:9myc::HphNT1; HTA2:mCHERRY::KanMX6	Machín lab
FM3127	FM2354; Δmsc1::NatNT2	This study
FM82	AS499; Δrad52::kanMX4	Machín lab
FM3139	FM588; Δdid4::NatNT2	This study
FM3142	FM588; Δsnf7::NatNT2	This study
FM3145	FM588; Δvps24::NatNT2	This study
FM3247	FM588; Δchm7::NatNT2	This study
FM3140	FM2354; Δdid4::NatNT2	This study
FM3144	FM2354; Δsnf7::NatNT2	This study
FM3147	FM2354; Δvps24::NatNT2	This study
FM3251	FM2354; Δchm7::NatNT2	This study
FM3129	FM2942; Δdid4::NatNT2	This study
FM3130	FM2942; ΔSnf7::NatNT2	This study
FM3131	FM2942; ΔVps24::NatNT2	This study
FM3151	FM3129; NUP49-eCFP::TRP1	This study
FM3152	FM3130; NUP49-eCFP::TRP1	This study
FM3162	FM3131; NUP49-eCFP::TRP1	This study
FM3252	FM3149; Δchm7::NatNT2	This study
FM3254	FM2807; Δchm7::KanMX	This study
FM3259	FM2807; Δsnf7::KanMX	This study
FM3258	FM2807; Δdid4::KanMX	This study
FM3263	FM2807; Δvps24::KanMX	This study
FM3281	FM3127; Δsnf7::TRP1	This study
FM3298	FM2976; Δsnf7::URA3	This study
FM2956	FM2531; Δmsc1::NatNT2; RAD52:mCherry::KanMX; NUP49:eCFP::TRP1	This study
FM2878	FM2831; RAD52:mCherry::KanMX	This study
FM3285	FM2878; Δchm7::TRP1	This study
FM3274	FM2878; Δsnf7::TRP1	This study
FM3001	SLJ7465; CDC42p:KAR2SS:GFP1-10:MSC1::NatMX	This study
FM3006	SLJ8344; CDC42p:KAR2SS:GFP1-10:MSC1::NatMX	This study
FM3009	SLJ7859; CDC42p:KAR2SS:GFP1-10:MSC1::NatMX	This study
FM3007	SLJ1838; CDC42p:KAR2SS:GFP1-10:MSC1::NatMX	This study
FM2947	FM2531; RAD52:mCherry::NatNT2; NUP49:eGFP::TRP1	This study
FM2954	FM2531; Δmsc1::NatNT2; RAD52:mCherry::KanMX; NUP49:eGFP::TRP1	This study
FM2381	AS499; GFP:TUB1::URA3; cdc15-2:9myc::HphNT1	F. Machín lab
FM2789	FM2381; Δmsc1::NatNT2	This Study
FM2317	AS499; ade2-1::TetR:YFP::ADE2; tetO(5.6Kb)::194Kb-ChrXII::HIS3; cdc15-2:9myc::HphNT1; CIN8:mCherry::KanMX6	Machín lab
FM2804	FM2317; Δmsc1::NatNT2	This study
BY4741	MATa his3∆1 leu2∆0 met15∆0 ura3∆0	Euroscarf
Y00540	BY4741; ∆rad52::kanMX4	Euroscarf
FM3002	BY4741; Δdid4::HphNT1	This study
FM3004	BY4741; Δsnf7::NatNT2	This study
FM3003	BY4741; Δvps24::NatNT2	This study
FM2865	MATα his3∆1 leu2∆0 lys2∆0 ura3∆0; Δmsc1::NatNT2	This study
FM1391	Mata ura3-52::ADH1p:OsTIR1:9myc::URA3 ade2-1 his3-11,15 leu2-3,112 trp1-1 can1-100; NET1:GFP::LEU2; CDC14:AID*::KanMX	Machín lab
FM3170	FM588; SNF7:AID*:9myc::KanMX	This study
FM3188	FM3170; ura3-52::ADH1p:OsTIR1:9myc::URA3	This study
FM3172	FM2807; SNF7:AID*:9myc::KanMX	This study
FM3189	FM3172; ura3-52::ADH1p:OsTIR1:9myc::URA3	This study
FM3173	FM588; VPS24:AID*:9myc::KanMX	This study
FM 3175	FM2807; VPS24:AID*:9myc::KanMX	This study
FM3190	FM3173; ura3-52::ADH1p:OsTIR1:9myc::URA3	This study
FM3191	FM3175; ura3-52::ADH1p:OsTIR1:9myc::URA3	This study
FM3166	FM588; DID4:AID*:9myc::KanMX	This study
FM3168	FM2807; DID4:AID*:9myc::KanMX	This study
FM3186	FM3166; ADH1p:OsTIR1-9myc::URA3	This study
FM3187	FM3168; ADH1:OsTIR1-9myc::URA3	This study
FM3198	FM3189; HTA2:eCFP::TRP1	This study
FM3202	FM3189; SEC61:eCFP::TRP1	This study
FM3200	FM3189; NUP49:eCFP::TRP1	This study
FM3184	AS499; NUP49:mCherry::NatNT2; HTA2:eCFP::TRP1	This study
FM3238	FM3184; SEC61:eYFP::KanMX; Δsnf7::URA	This study
FM3194	FM3184; SNF7::eYFP::HphNT1	This study
FM3246	FM3184; SNF7:6HA::HIS3	This study

¹The order of the strains is as they first appear in the main figures, then in the Supplementary Figs. The correspondence between the strains and the figures can be found in Table S1. The asterisk (*) indicates that the strain was not used in the figures, but is parental to others, thus simplifying genotype annotations below.

²Semicolon (“;”) separates genetic modifications accomplished sequentially through transformation. Intermediate strains are omitted. A non-italic beginning refers to a parental strain above.

Cell cultures were grown overnight in air orbital incubators at 25 °C in YPD media (10 g·L⁻¹ yeast extract, 20 g·L⁻¹ peptone and 20 g·L⁻¹ glucose). To arrest cells in late-M, log-phase asynchronous cultures were adjusted to OD₆₀₀ ~ 0.4 and the temperature was shifted to 34 °C for 3 h. In most experiments, the arrested culture was split into two subcultures: one subculture was treated with phleomycin (2 µg·mL⁻¹; Sigma-Aldrich, P9564), and the one was left untreated (mock). In a few experiments, a third subculture was treated with β-estradiol (2 µM; Sigma-Aldrich, E8875) for the induction of HO endonuclease (lexO:HO/ LexA-ER-B112 system), which generates DSBs at the HO locus^21,55. Samples were collected at the arrest and 1 and 2 h after treatment. To synchronize cells in G2/M by the DDC, phleomycin was added to the log-phase asynchronous population and the culture was held at 25 °C for 3 h. In experiments with conditional degron variants for the auxin system (AID* tags), 5 mM 3-indol-acetic acid (IAA; Sigma-Aldrich, I2886) was added 1 h prior to adding phleomycin or β-estradiol.

Western blots

Western blotting was carried out as reported before with minor modifications^10,21. Briefly, 5 ml samples were pelleted and fixed in 1 mL of 20% (w/v) trichloroacetic acid TCA. Cells were broken by vortexing for 3 min with ~200 mg of glass beads. After centrifuging, pellets were resuspended in 150 µL of PAGE Laemmli Sample Buffer 1X (Bio-Rad, 1610747), Tris HCl 0.75 M pH 8.0 and β-mercaptoethanol 2.5% (Sigma-Aldrich, M3148), and tubes were boiled at 95 °C for 3 min and pelleted again. Total protein in the supernatant was quantified using a Qubit 4 Fluorometer (Thermo Fisher Scientific, Q33227). Proteins were resolved in 7.5% SDS-PAGE gels and transferred to PVFD membranes (Pall Corporation, PVM020C099). The membrane was stained with Ponceau S solution (PanReac AppliChem, A2935) as a loading reference.

The following antibodies were used for immunoblotting: The HA epitope was detected with a primary mouse monoclonal anti-HA (1:2500; Sigma-Aldrich, H9658); the Myc epitope was detected with a primary mouse monoclonal anti-Myc (1:5000; Sigma-Aldrich, M4439); the Pgk1 protein was recognized with a primary mouse monoclonal anti-Pgk1 (1:5000; Thermo Fisher Scientific, 22C5D8), and Rad53 was recognized with a primary mouse monoclonal anti-Rad53 (1:2500; Abcam, ab166859). A polyclonal goat anti-mouse conjugated to horseradish peroxidase (1:5000 or 1:10,000; Promega, W4021) was used as the secondary antibody. Antibodies were diluted in 5% skimmed milk TBST (TBS pH 7.5 plus 0.1% Tween 20). Proteins were detected by using the ECL reagent (GE Healthcare, RPN2232) and visualized in a Vilber-Lourmat Fusion Solo S chamber. Protein bands were quantified using BioProfile Bio1D software (Vilber-Lourmat) and normalized to the housekeeping Pgk1.

Microscopy

A Zeiss Axio Observer.Z1/7 was used for both wide field and confocal microscopy as reported before²¹. The microscope was equipped for super-resolution confocal microscopy with live cell capability (LSM980 with Airyscan 2) together with the settings for wide field microscopy against CFP, YFP/GFP, and mCherry without crosstalk (Axiocam 702 sCMOS camera, the Colibri-7 LED excitation system, and narrow-band filter cubes). All images were obtained with a Plan-Apochromat 63x/NA 1.40 Oil M27 DIC objective. For each field, a stack of 10–20 z-focal planes (0.2–0.3 µm depth) was collected. In general, the images were taken from freshly harvested cells without further processing. For super-resolution images and 3D reconstructions, cells were first fixed with 3.7% w/v formaldehyde (Sigma-Aldrich, 47,608) for 15 min. The Zen Blue (Zeiss) and Fiji-ImageJ (NIH) software were used for image processing and quantification.

Growth curves, long-term viability spot assays, and short-term viability clonogenic assays

For real-time growth curves, log-phase overnight cultures were adjusted to an initial OD₆₀₀ = 0.05 in YPD without or with phleomycin (2 µg·mL⁻¹). Three replicates of each culture were aliquoted in a flat-bottomed 96-well plate and the real-time growth was measured in a Spark TECAN incubator by reading the OD₆₀₀ every 15 min for 24 h with shaking (96 rpm and 6 mm of orbital amplitude). The mean of the three replicates was calculated to obtain the final growth curves. Two independent experiments were performed but only one is shown since the s.e.m was less than 0.1 OD₆₀₀ for each time point.

For spot sensitivity assays cultures were grown exponentially and adjusted to an OD₆₀₀ = 0.5 and then five-fold serially diluted in YPD in 96-well plates. A 48-pin replica plater (Sigma-Aldrich, R2383) was used to spot ~3 µL onto the corresponding plates, which were incubated at 25 °C for 3–4 days before taking photographs.

For clonogenic survival assays, log-phase asynchronous cultures were adjusted to OD₆₀₀ = 0.4 before the corresponding treatment. After that, 100 µL of 4:10,000 dilutions were spread onto YPD plates. Viability was determined based on the number of colonies grown on the plates after 3 days at 25 °C and normalized to the WT. The mock treatments yielded ~500 colonies per plate.

Statistics and reproducibility

Error bars in all graphs represent the standard error of the mean (s.e.m.) of independent biological replicates performed in different days. The number of replicates (n) is given in the figure legend. Graphpad Prism 10 was used for statistical tests. Differences between experimental data points were generally estimated using the Mann-Whitney U test, the unpaired t-test or ANOVA; the test used in each specific case is indicated in the figure caption.

In general, we used three types of graphs to represent the data: bar charts, XY line graphs and box plots. In box plots, the center line represents the medians, box limits represent the 25 and 75th percentile, the whiskers extend to the 5 and 95th percentiles, and the dots represent outliers.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Acknowledgements

We would like to kindly thank Sue Jaspersen, Sarah Zanders, Rachel Helston and Lorraine Symington for strains and plasmids. This research was funded by the Ministerio de Ciencia, Innovación y Universidades (MICIU/AEI/10.13039/501100011033) through the grant PID2021-123716OB-I100 to F.M., which is co-funded by the EU-ERDF “A way of making Europe”. The Agencia Canaria de Investigación, Innovación y Sociedad de la Información (ACIISI) supported S. M-S. through a predoctoral fellowship (TESIS2020010028), co-funded by the EU-ESF + .

Author contributions

S.M.-S. performed all experiments shown in the main and supplementary figures and tables, constructed strains and plasmids, prepared figures and tables, gave critical insights as to the direction and development of the study, and aided with writing; F.M. conceived the work, supervised S.M.-S., gave critical insights as to the direction and development of the study, prepared figures and tables, was responsible for funding acquisition and project administration, and wrote the manuscript.

Peer review

Peer review information

Communications Biology thanks Akira Shinohara and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editors: Mengtan Xing and Christina Karlsson Rosenthal. A peer review file is available.

Data availability

All data is contained within the manuscript and/or supplementary files. Uncropped Western blot membranes and agarose gels are provided in Fig. S22. The raw data supporting the findings of this study are available on Figshare (https://doi.org/10.6084/m9.figshare.28899602)⁵⁶. Further data and materials are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s42003-025-08416-x.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

1. Aparicio, T; Baer, R; Gautier, J. DNA double-strand break repair pathway choice and cancer. DNA Repair; 2014; 19, pp. 169-175. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/24746645][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4051845][DOI: https://dx.doi.org/10.1016/j.dnarep.2014.03.014] 1:CAS:528:DC%2BC2cXmtlCltLo%3D

2. Jeggo, PA; Löbrich, M. DNA double-strand breaks: their cellular and clinical impact?. Oncogene; 2007; 26, pp. 7717-7719. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18066083][DOI: https://dx.doi.org/10.1038/sj.onc.1210868] 1:CAS:528:DC%2BD2sXhsVSmur7O

3. Trovesi, C; Manfrini, N; Falcettoni, M; Longhese, MP. Regulation of the DNA damage response by cyclin-dependent kinases. J. Mol. Biol.; 2013; 425, pp. 4756-4766. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23603016][DOI: https://dx.doi.org/10.1016/j.jmb.2013.04.013] 1:CAS:528:DC%2BC3sXnt1amtbg%3D

4. Kciuk, M; Gielecińska, A; Mujwar, S; Mojzych, M; Kontek, R. Cyclin-dependent kinases in DNA damage response. Biochim. Biophys. Acta Rev. Cancer; 2022; 1877, 188716. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35271993][DOI: https://dx.doi.org/10.1016/j.bbcan.2022.188716] 1:CAS:528:DC%2BB38Xms1ygtrg%3D

5. Scully, R; Panday, A; Elango, R; Willis, NA. DNA double-strand break repair-pathway choice in somatic mammalian cells. Nat. Rev. Mol. Cell Biol.; 2019; 20, pp. 698-714. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31263220][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7315405][DOI: https://dx.doi.org/10.1038/s41580-019-0152-0] 1:CAS:528:DC%2BC1MXht1yktL3I

6. Thompson, R; Gatenby, R; Sidi, S. How cells handle DNA breaks during mitosis: detection, signaling, repair, and fate choice. Cells; 2019; 8, 1049. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31500247][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6770852][DOI: https://dx.doi.org/10.3390/cells8091049] 1:CAS:528:DC%2BB3cXmtVSlsLs%3D

7. Audrey, A; de Haan, L; van Vugt, MATM; de Boer, HR. Processing DNA lesions during mitosis to prevent genomic instability. Biochem. Soc. Trans.; 2022; 50, pp. 1105-1118. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36040211][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9444068][DOI: https://dx.doi.org/10.1042/BST20220049] 1:CAS:528:DC%2BB38XisV2qu7zE

8. Machín, F; Ayra-Plasencia, J. Are anaphase events really irreversible? The endmost stages of cell division and the paradox of the dna double-strand break repair. BioEssays; 2020; 42, 2000021. [DOI: https://dx.doi.org/10.1002/bies.202000021]

9. Ayra-Plasencia, J; Symington, L; Machín, F. Cohesin still drives homologous recombination repair of DNA double-strand breaks in late mitosis. Elife; 2024; 13, RP92706.

10. Ayra-Plasencia, J; Machín, F. DNA double-strand breaks in telophase lead to coalescence between segregated sister chromatid loci. Nat. Commun.; 2019; 10, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31253793][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6598993][DOI: https://dx.doi.org/10.1038/s41467-019-10742-8] 2862.

11. Boettcher, B; Barral, Y. The cell biology of open and closed mitosis. Nucleus; 2013; 4, pp. 160-165. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/23644379][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3720745][DOI: https://dx.doi.org/10.4161/nucl.24676]

12. Quevedo, O; García-Luis, J; Matos-Perdomo, E; Aragón, L; Machín, F. Nondisjunction of a single chromosome leads to breakage and activation of DNA damage checkpoint in g2. PLoS Genet; 2012; 8, e1002509. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22363215][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3280967][DOI: https://dx.doi.org/10.1371/journal.pgen.1002509] 1:CAS:528:DC%2BC38Xjt1Sjt7k%3D

13. Dey, G et al. Closed mitosis requires local disassembly of the nuclear envelope. Nature; 2020; 585, pp. 119-123. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32848252][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7610560][DOI: https://dx.doi.org/10.1038/s41586-020-2648-3] 1:CAS:528:DC%2BB3cXhs12jt7bP

14. Stoten, CL; Carlton, JG. ESCRT-dependent control of membrane remodelling during cell division. Semin. Cell Dev. Biol.; 2018; 74, pp. 50-65. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/28843980][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6015221][DOI: https://dx.doi.org/10.1016/j.semcdb.2017.08.035] 1:CAS:528:DC%2BC2sXhsVeisL7L

15. Vietri, M; Radulovic, M; Stenmark, H. The many functions of ESCRTs. Nat. Rev. Mol. Cell Biol.; 2020; 21, pp. 25-42. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/31705132][DOI: https://dx.doi.org/10.1038/s41580-019-0177-4] 1:CAS:528:DC%2BC1MXitFars7jJ

16. Frost, A et al. Functional repurposing revealed by comparing S. pombe and S. cerevisiae genetic interactions. Cell; 2012; 149, pp. 1339-1352. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22682253][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3613983][DOI: https://dx.doi.org/10.1016/j.cell.2012.04.028] 1:CAS:528:DC%2BC38XosVals7g%3D

17. Webster, BM; Colombi, P; Jäger, J; Patrick Lusk, C. Surveillance of nuclear pore complex assembly by ESCRT-III/Vps4. Cell; 2014; 159, pp. 388-401. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/25303532][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4194032][DOI: https://dx.doi.org/10.1016/j.cell.2014.09.012] 1:CAS:528:DC%2BC2cXhslGjsbjE

18. Nagai, S et al. Functional targeting of DNA damage to a nuclear pore–associated SUMO-dependent ubiquitin ligase. Science; 2008; 322, 597. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18948542][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3518492][DOI: https://dx.doi.org/10.1126/science.1162790] 1:CAS:528:DC%2BD1cXht1Krt7vP

19. Rodriguez-Berriguete, G et al. Nucleoporin 54 contributes to homologous recombination repair and post-replicative DNA integrity. Nucleic Acids Res; 2018; 46, pp. 7731-7746. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/29986057][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6125679][DOI: https://dx.doi.org/10.1093/nar/gky569] 1:CAS:528:DC%2BC1MXosFams7s%3D

20. Gasser, SM; Stutz, F. SUMO in the regulation of DNA repair and transcription at nuclear pores. FEBS Lett.; 2023; 597, pp. 2833-2850. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/37805446][DOI: https://dx.doi.org/10.1002/1873-3468.14751] 1:CAS:528:DC%2BB3sXit1Cjtb3N

21. Medina-Suárez, S; Ayra-Plasencia, J; Pérez-Martínez, L; Butter, F; Machín, F. Msc1 is a nuclear envelope protein that reinforces DNA repair in late mitosis. iScience; 2024; 27, 110250. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/39021806][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11253511][DOI: https://dx.doi.org/10.1016/j.isci.2024.110250]

22. Asakawa, H; Hirano, Y; Shindo, T; Haraguchi, T; Hiraoka, Y. Fission yeast Ish1 and Les1 interact with each other in the lumen of the nuclear envelope. Genes Cells; 2022; 27, pp. 643-656. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/36043331][DOI: https://dx.doi.org/10.1111/gtc.12981] 1:CAS:528:DC%2BB38XisVSnsrnM

23. Fedotova, AA; Bonchuk, AN; Mogila, VA; Georgiev, PG. C2H2 zinc finger proteins: the largest but poorly explored family of higher eukaryotic transcription factors. Acta Nat.; 2017; 9, pp. 47-58. [DOI: https://dx.doi.org/10.32607/20758251-2017-9-2-47-58] 1:STN:280:DC%2BC1cfgvVCntQ%3D%3D

24. Smoyer, CJ et al. Analysis of membrane proteins localizing to the inner nuclear envelope in living cells. J. Cell Biol.; 2016; 215, pp. 575-590. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27831485][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5119940][DOI: https://dx.doi.org/10.1083/jcb.201607043] 1:CAS:528:DC%2BC2sXktVWns7o%3D

25. D’Amours, D; Stegmeier, F; Amon, A. Cdc14 and condensin control the dissolution of cohesin-independent chromosome linkages at repeated DNA. Cell; 2004; 117, pp. 455-469. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15137939][DOI: https://dx.doi.org/10.1016/S0092-8674(04)00413-1]

26. Machín, F; Torres-Rosell, J; Jarmuz, A; Aragón, L. Spindle-independent condensation-mediated segregation of yeast ribosomal DNA in late anaphase. J. Cell Biol.; 2005; 168, pp. 209-219. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15657393][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2171579][DOI: https://dx.doi.org/10.1083/jcb.200408087]

27. Matos-Perdomo, E; Santana-Sosa, S; Ayra-Plasencia, J; Medina-Suárez, S; Machín, F. The vacuole shapes the nucleus and the ribosomal DNA loop during mitotic delays. Life Sci. Alliance; 2022; 5, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35961781][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9375157][DOI: https://dx.doi.org/10.26508/lsa.202101161] 1:CAS:528:DC%2BB38XitlWksr7E e202101161.

28. Haber, JE. Mating-Type Genes and MAT Switching in Saccharomyces cerevisiae. Genetics; 2012; 191, pp. 33-64. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22555442][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3338269][DOI: https://dx.doi.org/10.1534/genetics.111.134577] 1:CAS:528:DC%2BC3sXkt1ersw%3D%3D

29. Gnügge, R; Symington, LS. Efficient DNA double-strand break formation at single or multiple defined sites in the Saccharomyces cerevisiae genome. Nucleic Acids Res.; 2020; 48, pp. e115-e115. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/33053188][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7672422][DOI: https://dx.doi.org/10.1093/nar/gkaa833]

30. Thaller, DJ et al. An escrt-lem protein surveillance system is poised to directly monitor the nuclear envelope and nuclear transport system. Elife; 2019; 8, pp. 1-36. [DOI: https://dx.doi.org/10.7554/eLife.45284]

31. Babst, M; Katzmann, DJ; Estepa-Sabal, EJ; Meerloo, T; Emr, SD. Escrt-III. Dev. Cell; 2002; 3, pp. 271-282. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/12194857][DOI: https://dx.doi.org/10.1016/S1534-5807(02)00220-4] 1:CAS:528:DC%2BD38Xmtl2it7c%3D

32. Thaller, D. J. et al. Direct binding of ESCRT protein Chm7 to phosphatidic acid-rich membranes at nuclear envelope herniations. J. Cell Biol. 220, 4222 (2021).

33. Nishimura, K; Fukagawa, T; Takisawa, H; Kakimoto, T; Kanemaki, M. An auxin-based degron system for the rapid depletion of proteins in nonplant cells. Nat. Methods; 2009; 6, pp. 917-922. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19915560][DOI: https://dx.doi.org/10.1038/nmeth.1401] 1:CAS:528:DC%2BD1MXhsVWlsbzM

34. Chadrin, A et al. Pom33, a novel transmembrane nucleoporin required for proper nuclear pore complex distribution. J. Cell Biol.; 2010; 189, pp. 795-811. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20498018][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2878943][DOI: https://dx.doi.org/10.1083/jcb.200910043] 1:CAS:528:DC%2BC3cXnt1Gns74%3D

35. Onischenko, E; Stanton, LH; Madrid, AS; Kieselbach, T; Weis, K. Role of the NDCL interaction network in yeast nuclear pore complex assembly and maintenance. J. Cell Biol.; 2009; 185, pp. 475-491. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/19414609][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2700399][DOI: https://dx.doi.org/10.1083/jcb.200810030] 1:CAS:528:DC%2BD1MXlvFSrurY%3D

36. Weinert, TA; Hartwell, LH. The RAD9 gene controls the cell cycle response to DNA damage in Saccharomyces cerevisiae. Science; 1988; 241, pp. 317-322. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/3291120][DOI: https://dx.doi.org/10.1126/science.3291120] 1:CAS:528:DyaL1cXlt1Glt70%3D

37. Constantinescu, D; Csoka, AB; Navara, CS; Schatten, GP. Defective DSB repair correlates with abnormal nuclear morphology and is improved with FTI treatment in Hutchinson-Gilford progeria syndrome fibroblasts. Exp. Cell Res.; 2010; 316, pp. 2747-2759. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20599958][DOI: https://dx.doi.org/10.1016/j.yexcr.2010.05.015] 1:CAS:528:DC%2BC3cXht1Wmtb%2FF

38. Denais, CM et al. Nuclear envelope rupture and repair during cancer cell migration. Science; 2016; 352, pp. 353-358. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/27013428][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4833568][DOI: https://dx.doi.org/10.1126/science.aad7297] 1:CAS:528:DC%2BC28Xlsl2isbw%3D

39. Vietri, M et al. Unrestrained ESCRT-III drives micronuclear catastrophe and chromosome fragmentation. Nat. Cell Biol.; 2020; 22, pp. 856-867. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/32601372][DOI: https://dx.doi.org/10.1038/s41556-020-0537-5] 1:CAS:528:DC%2BB3cXht1yjtbbJ

40. Willan, J et al. ESCRT-III is necessary for the integrity of the nuclear envelope in micronuclei but is aberrant at ruptured micronuclear envelopes generating damage. Oncogenesis; 2019; 8, [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/30988276][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6465242][DOI: https://dx.doi.org/10.1038/s41389-019-0136-0] 29.

41. Warecki, B., Ling, X., Bast, I. & Sullivan, W. ESCRT-III–mediated membrane fusion drives chromosome fragments through nuclear envelope channels. J. Cell Biol. 219, 5091 (2020).

42. Shankar, R; Lettman, MM; Whisler, W; Frankel, EB; Audhya, A. The ESCRT machinery directs quality control over inner nuclear membrane architecture. Cell Rep.; 2022; 38, 110263. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/35045304][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8801257][DOI: https://dx.doi.org/10.1016/j.celrep.2021.110263] 1:CAS:528:DC%2BB38XhtlOlurc%3D

43. Mekhail, K; Seebacher, J; Gygi, SP; Moazed, D. Role for perinuclear chromosome tethering in maintenance of genome stability. Nature; 2008; 456, pp. 667-670. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/18997772][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2596277][DOI: https://dx.doi.org/10.1038/nature07460] 1:CAS:528:DC%2BD1cXhsVGlur%2FP

44. Lemaître, C et al. The nucleoporin 153, a novel factor in double-strand break repair and DNA damage response. Oncogene; 2012; 31, pp. 4803-4809. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22249246][DOI: https://dx.doi.org/10.1038/onc.2011.638]

45. Géli, V; Lisby, M. Recombinational DNA repair is regulated by compartmentalization of DNA lesions at the nuclear pore complex. Bioessays; 2015; 37, pp. 1287-1292. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/26422820][DOI: https://dx.doi.org/10.1002/bies.201500084]

46. Loeillet, S et al. Genetic network interactions among replication, repair and nuclear pore deficiencies in yeast. DNA Repair; 2005; 4, pp. 459-468. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/15725626][DOI: https://dx.doi.org/10.1016/j.dnarep.2004.11.010] 1:CAS:528:DC%2BD2MXhslWksro%3D

47. Titus, LC; Dawson, TR; Rexer, DJ; Ryan, KJ; Wente, SR. Members of the RSC chromatin-remodeling complex are required for maintaining proper nuclear envelope structure and pore complex localization. Mol. Biol. Cell; 2010; 21, pp. 1072-1087. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/20110349][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2836959][DOI: https://dx.doi.org/10.1091/mbc.e09-07-0615] 1:CAS:528:DC%2BC3cXjs1eqsb4%3D

48. Wente, SR; Blobel, G. A temperature-sensitive NUP116 null mutant forms a nuclear envelope seal over the yeast nuclear pore complex thereby blocking nucleocytoplasmic traffic. J. Cell Biol.; 1993; 123, pp. 275-284. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/7691829][DOI: https://dx.doi.org/10.1083/jcb.123.2.275] 1:STN:280:DyaK2c%2Fht12ktg%3D%3D

49. Liu, Q et al. Functional association of Sun1 with nuclear pore complexes. J. Cell Biol.; 2007; 178, pp. 785-798. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/17724119][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2064544][DOI: https://dx.doi.org/10.1083/jcb.200704108] 1:CAS:528:DC%2BD2sXhtVSgtrjN

50. Clever, M; Funakoshi, T; Mimura, Y; Takagi, M; Imamoto, N. The nucleoporin ELYS/Mel28 regulates nuclear envelope subdomain formation in HeLa cells. Nucleus; 2012; 3, pp. 187-199. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/22555603][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3383574][DOI: https://dx.doi.org/10.4161/nucl.19595]

51. Tamura, K; Hara-Nishimura, I. Involvement of the nuclear pore complex in morphology of the plant nucleus. Nucleus; 2011; 2, pp. 168-172. [PubMed: https://www.ncbi.nlm.nih.gov/pubmed/21818409][PubMedCentral: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3149876][DOI: https://dx.doi.org/10.4161/nucl.2.3.16175]

52. Gu, M. et al. LEM2 recruits CHMP7 for ESCRT-mediated nuclear envelope closure in fission yeast and human cells. Proc. Natl. Acad. Sci. USA114, 166–175 (2017).

53. Lee, I.-J. et al. Factors promoting nuclear envelope assembly independent of the canonical ESCRT pathway. J. Cell Biol. 219, 8232 (2020).

54. Dunham, M., Gartenberg, M. & Brown, G. W. Methods in Yeast Genetics and Genomics, 2015 Edition: A CSHL Course Manual. (Cold Spring Harbor Laboratory Press, 2015).

55. Ottoz, D. S. M., Rudolf, F. & Stelling, J. Inducible, tightly regulated and growth condition-independent transcription factor in Saccharomyces cerevisiae. Nucleic Acids Res. 42, e130 (2014).

56. Medina-Suárez, S. & Machín, F. Continuous nuclear envelope surveillance is required for DNA double-strand break repair [Data set]. Figsharehttps://doi.org/10.6084/m9.figshare.28899602 (2025).