Nuclear Envelope Assembly Defects Link Mitotic Errors to Chromothripsis

Nuclear envelope assembly defects link mitotic errors to chromothripsis

Shiwei Liu*1,2,3, Mijung Kwon*,1,2,3, Mark Mannino1,2,3, Nachen Yang4, Alexey Khodjakov4 & David Pellman1,2,3,+

1Howard Hughes Medical Institute, Chevy Chase, MD 20815, USA. 2Department of Pediatric Oncology, Dana-Farber Cancer Institute, Boston, MA 02215, USA. 3Department of Cell Biology, Harvard Medical School, Boston, MA 02215, USA. 4Wadsworth Center, New York State Department of Health, Albany, NY 12201.

*Equal contribution

+Lead Contact

Correspondence should be addressed to:

David Pellman

Dana-Farber Cancer Institute

44 Binney St., Rm. M663

Boston, MA 02115

Phone: (617) 632-4918

Fax: (617) 632-6845

Email: [email protected] bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 Defects in the architecture or integrity of the nuclear envelope (NE) are associated with a 2 variety of human diseases1. Micronuclei, one common nuclear aberration, are an origin for 3 chromothripsis2,3, a catastrophic mutational process commonly observed in cancer 4 genomes and other contexts4-6. Micronuclei have a defective NE, with the extensive 5 chromosome fragmentation that generates chromothripsis occurring after abrupt, 6 spontaneous loss of NE integrity7. After NE disruption, the exposed cytoplasmic DNA can 7 additionally initiate proinflammatory signaling linked to senescence, metastasis, and the 8 immune clearance of tumor cells8. Despite its broad physiological impact, the basis for the 9 nuclear envelope fragility of micronuclei is unknown. Here we demonstrate that 10 micronuclei undergo markedly defective NE assembly: Only “core” NE proteins9,10 11 assemble efficiently on lagging chromosomes whereas “non-core” NE proteins9,10, 12 including nuclear pore complexes (NPCs), fail to properly assemble. Consequently, 13 micronuclei have impaired nuclear import, and key nuclear proteins required to maintain 14 the integrity of the NE and the genome fail to accumulate normally. We show that densely 15 bundled spindle microtubules inhibit non-core NE assembly, leading to an irreversible NE 16 assembly defect. Accordingly, experimental manipulations that position missegregated 17 chromosomes away from the spindle correct defective NE assembly, prevent spontaneous 18 NE disruption, and suppress DNA damage in micronuclei. Our findings indicate that 19 chromosome segregation and NE assembly are only loosely coordinated through the 20 timing of mitotic spindle disassembly. The absence of precise regulatory controls can 21 explain why errors during mitotic exit are frequent, and a major trigger for catastrophic 22 genome rearrangements5,6. 23 24 During normal mitotic exit, NE proteins transiently form two domains around decondensing 25 chromosomes1,9,10. The group of “core” NE proteins, which include the membrane protein emerin, 26 and BAF (barrier-to-autointegration factor), concentrate on the chromosome mass adjacent to 27 central spindle or spindle pole microtubules; the “non-core” group of proteins, which include 28 nuclear pore complex proteins (NPCs) and the Lamin B receptor (LBR), assemble on the 29 chromosome periphery away from the spindle (Extended Data Fig. 1a)9,10. After mitotic exit, these 30 domains become intermingled, with fragments of the core domain persisting as “pore-free” islands 31 that are then slowly populated by NPCs during interphase11,12. 32 33 To understand the basis for NE abnormalities of micronuclei, we asked whether 34 postmitotic NE assembly on lagging chromosomes occurs in this same spatiotemporal pattern. In

1 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

35 three cell lines, lagging chromosomes were generated in synchronized cells by recovery from 36 nocodazole-induced mitotic arrest or by short-term inhibition of the spindle assembly checkpoint 37 with a MPS1 kinase inhibitor. Immunofluorescence staining or live-cell imaging of telophase cells 38 revealed that core proteins were recruited to lagging chromosomes at equivalent or higher levels 39 than observed on the main chromosome mass. By contrast, the non-core proteins were strikingly 40 depleted from lagging chromosomes (Fig. 1a, Extended Data Fig. 1b-d). Likewise, chromatin 41 bridges formed after nocodazole release or after partial depletion of the condensin SMC2 42 displayed the same core-only NE protein composition (Extended Data Fig. 1e, f).Thus, only a 43 subset of NE proteins assembles on chromosomes that lag within the region of the central spindle 44 during telophase. 45 46 A similar reduction of non-core NE proteins was observed on micronuclei formed from 47 lagging chromosomes (Fig. 1b, Extended Data Fig. 2a-c). Consistent with the reduced assembly 48 of NPCs on micronuclei, high temporal resolution live-cell imaging of cells demonstrated 49 significant nuclear import defects in micronuclei, with some variation in the extent of the defect 50 (two import reporters in two cell lines, Fig. 1c, d, Extended Data Fig. 3a, b, Supplementary Videos 51 1 and 2). Micronuclei fail to normally accumulate key nuclear proteins, including replication protein 52 A (RPA), a key dosage-sensitive regulator of DNA replication and repair13, and B-type lamins, 53 known to be required to maintain NE integrity (Fig. 1b, e, Extended Data Figs. 2a-c, 3c, 3d)7,14. 54 For individual cells, the extent of the defect in accumulating RPA and a general import reporter 55 were strongly correlated, suggesting that micronuclei have a global import defect rather than 56 selective import pathway defects (Fig.1e). Thus, reconciling differing reports from prior 57 literature2,7,15-18, these data indicate that micronuclei undergo aberrant NE and NPC assembly, 58 leading to defective import and accumulation of many proteins, including those required for 59 genome stability and NE integrity. 60 61 We considered the possibility that the non-core recruitment defect on lagging 62 chromosomes could be explained by a previously proposed “chromosome separation 63 checkpoint”17. Under this model, the position of lagging chromosomes is monitored throughout 64 mitotic exit by the spindle midzone-centered Aurora B phosphorylation gradient19, postulated to 65 block NE assembly until membrane-free, lagging chromosomes can be incorporated into the main 66 chromosome mass, ensuring the formation of a single nucleus. However, several observations 67 were inconsistent with this model. First, core membrane proteins, which were not previously 68 examined17, assemble on lagging chromosomes and chromosome bridges20, independent of their

69 position relative to the spindle midzone (Extended Data Fig. 4a). Thus, an NPC-depleted NE does 70 form prior to the completion of chromosome segregation. Second, chromosome bridges are 71 uniformly depleted for non-core proteins, showing no obvious gradient (Extended Data Figs. 1e, 72 f, 4a). Finally, we inhibited Aurora B transport to the spindle midzone by siRNA-mediated 73 knockdown of the kinesin MKLP221. After MKLP2 knockdown, cells lacking detectable midzone 74 Aurora B nevertheless failed to recruit non-core proteins to lagging chromosomes (Extended Data 75 Fig. 4b), indicating that the gradient is not required for inhibition of non-core NE assembly. 76 77 We designed an experiment to directly test the main prediction of the checkpoint model 78 that the position of lagging chromosomes is monitored continuously throughout mitotic exit by 79 Aurora B. Cells were released from a mitotic arrest, followed by live-cell imaging at high temporal 80 resolution, treated with an Aurora B inhibitor (ZM447439), and then fixed and labeled to assess 81 NE assembly (Fig. 2a). Because cells release from the mitotic block with only partial synchrony, 82 this approach enabled us to identify cells exposed to Aurora B inhibition at all intervals from 83 metaphase through telophase. Strikingly, we found that Aurora B inhibition restored non-core 84 protein assembly to lagging chromosomes only if it occurred early, at or shortly after anaphase 85 onset. However, there was no effect if Aurora B inhibition occurred later, ~6-8 min after anaphase 86 onset (Fig. 2b). Thus, global Aurora B activity inhibits NE assembly up until early anaphase, but 87 has no effect during the critical time period where a chromosome separation checkpoint would 88 need to operate. Instead, the lagging chromosomes appear to become irreversibly defective for 89 non-core NE recruitment in telophase. This irreversible effect is unlikely to be caused by persistent, 90 Aurora B-mediated association of condensin I with lagging chromosomes17 because siRNA- 91 mediated knockdown of SMC2 did not rescue non-core NE assembly (Extended Data Fig. 4c). 92 93 The formation of a nearly continuously enclosed, core-only NE on the lagging 94 chromosome could explain the irreversible defect in the recruitment of NPCs and other non-core 95 proteins. This idea is consistent with the observation that Xenopus egg extracts lacking the 96 Nup107-160 complex assemble NPCs only if the purified complex is added back prior to the 97 formation of a continuous NPC-free NE22. Importantly, telophase lagging chromosomes are 98 already ensheathed by an apparently continuous layer of the core-membrane protein emerin (Fig. 99 2c). Concomitantly, ESCRT-III components, which are thought to seal small membrane gaps 100 during NE assembly23,24, associate and dissociate from lagging chromosomes approximately on 101 schedule (Extended Data Fig. 5a, b, Supplementary Video 3). These ESCRT-III kinetics suggest 102 that the micronuclear membrane likely undergoes significant membrane fusion. Accordingly,

103 correlated light and electron microscopy (CLEM) demonstrated that the NPC-deficient NE formed 104 around lagging chromosomes persists on micronuclei throughout interphase (Fig. 2d). As 105 expected for defective NPC assembly and therefore defective import, chromatin within these 106 structures was hyper-condensed (Fig. 2d)25. 107 108 Although our experiments argue against spatial regulation of NE assembly by the Aurora 109 B gradient (above and Extended Data Fig. 4a, b), we are still left with the question of why global 110 inhibition of Aurora B in early anaphase allows non-core NE assembly on lagging chromosomes 111 (Fig. 2b). We considered the possibility that this effect on NE assembly could be an indirect 112 consequence of Aurora B’s well-characterized role in stabilizing spindle microtubules, which 113 increases both the mass and bundling of spindle microtubules (Extended Data Fig. 6a)26. 114 Supporting the idea that microtubules adjacent to chromosomes can prevent non-core NE 115 assembly, in Xenopus egg extracts, microtubule stabilization has been shown to irreversibly 116 inhibit the formation of an NE containing NPCs27. Accordingly, we found that nocodazole and 117 Aurora B inhibition had a remarkably similar effect on non-core NE assembly: Early anaphase 118 microtubule disassembly by nocodazole reversed the non-core assembly defect on lagging 119 chromosomes whereas there was minimal effect if nocodazole was added later (Extended Data 120 Fig. 6b, c). Finally, during normal NE assembly, non-core NE is typically excluded from the region 121 of the chromosome mass adjacent to dense microtubule bundles from the central spindle9,28. We 122 found that this exclusion occurs whether Aurora B localizes to the spindle midzone or is forced to 123 remain on the main chromosome mass (Extended Data Fig. 6d). 124 125 Although consistent with the hypothesis that microtubules inhibit non-core NE assembly, 126 the above experiments do not conclusively separate the interdependent effects of Aurora B and 127 microtubules26. To make this distinction, we used paclitaxel to prevent microtubule disassembly 128 after Aurora B inhibition (Fig. 3a) and determined whether microtubules can still exclude the non- 129 core NE from central spindle region, in the absence of active Aurora B (Extended Data Figs. 1b, 130 6d). In untreated controls, core (emerin) and non-core (Nup-133) proteins become intermingled 131 on the main chromosome mass after mitotic exit as expected. By contrast, in paclitaxel-treated 132 cells there was a persistent non-core NE gap and/or an exaggerated core NE domain at this stage 133 (Fig. 3a, arrowheads). Importantly, this effect of paclitaxel was completely independent of Aurora 134 B because co-addition of paclitaxel and the Aurora B inhibitor had the identical effect as paclitaxel 135 treatment alone (Fig. 3a). Thus, microtubules inhibit non-core NE assembly irrespective of Aurora 136 B activity.

137 138 We next asked whether the inhibitory effect of microtubules on non-core NE assembly 139 was local and depended on the degree of microtubule bundling near the lagging chromosome. 140 To address this, we “loosened” central spindle microtubule bundling by siRNA-mediated 141 knockdown of the kinesin KIF4A29. As expected, KIF4A knockdown preserved general spindle 142 organization and central spindle localization of Aurora B (Extended Data Fig. 7a)29. Consistent 143 with our hypothesis, KIF4A depletion increased the recruitment of non-core NE to lagging 144 chromosomes (Fig. 3b-d, Extended Data Fig. 7b, c). Interestingly, the NE at these sites often 145 displayed a small-scale separation of core (emerin, white arrowheads) and non-core proteins 146 (Nup133, white arrows in Fig. 3d). These “mini” core/non-core subdomains formed at any location 147 within the central spindle, including the spindle midzone where Aurora B concentrates. Structured 148 illumination microscopy (SIM) of the lagging chromosomes from KIF4A-deleted cells revealed that 149 the regions of microtubule-chromosome contact, marked by the ESCRT-III component CHM4B24, 150 were depleted for NPCs (Nup133) (Fig. 3e, Extended Data Fig. 8). Together, these experiments 151 suggest that NE subdomain assembly is primarily controlled by the organization of microtubules, 152 independent of chromosome position within the spindle. 153 154 A direct prediction of our microtubule-inhibition model is that positioning of chromosomes 155 away from the central spindle should normalize NE assembly and restore nuclear function. To 156 achieve a peripheral localization of missegregated chromosomes, cells were exposed to a high 157 concentration of a MPS1 inhibitor, causing chromosome missegregation prior to the completion 158 of chromosome congression. In HeLa cells, MPS1 inhibition was combined with the knockdown 159 of tubulin tyrosine ligase (TTL), which further enhances peripheral chromosome localization as 160 previously reported (Fig. 4a, Extended Data Fig. 9a)30. High temporal resolution, live-cell imaging 161 was used to identify chromosomes that either remained consistently peripheral or consistently 162 within the central spindle from anaphase onset until mitotic exit (Fig. 4a). Unlike lagging 163 chromosomes within the spindle, the peripheral chromosomes recruited both core and non-core 164 NE proteins, including NPCs (Fig. 4b, c, Extended Data Fig. 9a, Supplementary Video 4). As 165 expected, microtubule and ESCRT-labeling24 suggested that peripheral chromosomes had less 166 contact with microtubules than central spindle lagging chromosomes (Extended Data Fig. 9b). 167 Strikingly, using a combination of live and fixed cell imaging, we observed that micronuclei derived 168 from peripheral chromosomes had normal levels of nuclear import, a normal accumulation of 169 RPA2 and lamin B1, and a normal extent of DNA replication (Fig. 4d and Extended Data Fig. 9c- 170 e). The restoration of NE assembly and function on peripheral micronuclei, which occurred

5 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

171 independent of their chromosome number, likely contributes to their relatively larger size 172 (Extended Data Fig. 9c-h). 173 174 We next determined whether micronuclei from peripheral chromosomes have a lower 175 frequency of spontaneous NE disruption and a consequently lower frequency of DNA damage 176 (Fig. 4e). In HeLa cells, peripheral micronuclei indeed had a lower frequency of NE disruption, 177 verified with two independent assays, and a significantly lower frequency of DNA damage (Fig. 178 4e, f, Extended Data Fig. 10a). In RPE-1 cells, peripheral chromosome localization also 179 significantly reduced NE disruption and DNA damage (Extended Data Fig. 10b, Supplementary 180 Video 5). However, in RPE-1 cells, the effect of peripheral chromosomes on NE disruption was 181 partly masked because the large micronuclei from peripheral chromosomes in these cells are 182 subject to actin-dependent NE breakage, as can occur transiently on primary nuclei14 (see 183 discussion in Extended Data Fig. 10b, c). Thus, positioning chromosomes away from the spindle 184 restores NE assembly and function of the resulting micronuclei. 185 186 Here, we have uncovered the basis of NE defects in micronuclei. The findings define an 187 important mechanism underlying chromothripsis and suggest a new model for the coordination of 188 chromosome segregation and nuclear envelope assembly during normal cell division 189 (Supplementary Video 6). We propose that in telophase, densely bundled spindle microtubules 190 inhibit the recruitment of non-core NE to lagging chromosomes, with the consequence that lagging 191 chromosomes become enclosed by an NE primarily composed of core proteins. Based on prior 192 in vitro studies22,27, core-only NE assembly should be a nearly irreversible barrier for the assembly 193 of NE containing NPCs and other non-core proteins. Because micronuclei are generated with a 194 largely-isolated core NE, they have little or no access to the nuclear transport-dependent 195 interphase pathway for NPC assembly12. Defective nucleocytoplasmic transport then leads to 196 defects in the accumulation of numerous proteins, including those necessary for normal DNA 197 replication, DNA repair, and the maintenance of NE integrity. The resulting spontaneous NE 198 disruption then leads to DNA damage, chromosome fragmentation, and ultimately 199 chromothripsis2,3,7. Why microtubules inhibit non-core NE assembly remains to be determined, 200 but could occur through a simple physical barrier effect. Non-core NE, including NPCs, was 201 recently reported to assemble with fenestrated ER sheets31 that might less readily penetrate 202 dense bundles of spindle microtubules than vesicles or tubules, which we speculate could be the 203 main source of core membrane proteins.

204 Together, our findings demonstrate that altered NE assembly on lagging chromosomes is 205 not the consequence of a beneficial checkpoint delay, but rather a pathological outcome. 206 Consequently, during normal cell division, instead of precise and continuous monitoring of 207 chromosome position, it appears that there is only loose coordination, with normal non-core NE 208 assembly being dependent on timely spindle microtubule disassembly. Loose coordination, 209 coupled with the irreversibility of NE assembly errors during mitotic exit, provides one explanation 210 of why chromothripsis is common, with frequencies recently reported to be as high as 65% in 211 some cancers5,6.

7 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Acknowledgements

We thank I. Cheeseman, T. Rapoport, N. Umbreit, T. Walther, and K. Xie for comments on the manuscript; E. Jackson and A. Spektor for preliminary experiments; J. Ellenberg, D. Gerlich, E. Hatch, M. Hetzer, A. Hyman, and T. Kuroda for reagents; J. Waters and T. Lambert of the Nikon Imaging Center at Harvard Medical School for advice and use of microscopes; M. Cicconet and C. Yapp from the Image and Data Analysis Core at Harvard Medical School for assistance with 3D rendering; Lin Shao (Yale University) for SIM reconstruction code. We acknowledge the use of the Wadsworth Center’s Electron Microscopy Core Facility. A. K. is supported by the NIH grant GM059363. D.P. is a HHMI investigator and is supported by R37 GM61345-14.

Author Contributions

D.P., S.L., and M.K. designed the experiments. D.P., S.L., and M.K. wrote the manuscript, with edits from all authors. S.L., and M.K. performed most experiments and analysis. M.M. assisted with several experiments and contributed Extended Data Fig. 10b, c. N. Y. and A. K. performed the electron microscopy.

1 Ungricht, R. & Kutay, U. Mechanisms and functions of nuclear envelope remodelling. Nature Reviews Molecular Cell Biology 18, 229-245, doi:10.1038/nrm.2016.153 (2017). 2 Crasta, K. et al. DNA breaks and chromosome pulverization from errors in mitosis. Nature 482, 53-58, doi:10.1038/nature10802 (2012). 3 Zhang, C.-Z. et al. Chromothripsis from DNA damage in micronuclei. Nature 522, 179-184, doi:10.1038/nature14493 (2015). 4 Stephens, P. J. et al. Massive Genomic Rearrangement Acquired in a Single Catastrophic Event during Cancer Development. Cell 144, 27-40, doi:10.1016/j.cell.2010.11.055 (2011). 5 Notta, F. et al. A renewed model of pancreatic cancer evolution based on genomic rearrangement patterns. Nature 538, 378-382, doi:10.1038/nature19823 (2016). 6 Ly, P. & Cleveland, D. W. Rebuilding Chromosomes After Catastrophe: Emerging Mechanisms of Chromothripsis. Trends Cell Biol 27, 917-930, doi:10.1016/j.tcb.2017.08.005 (2017). 7 Hatch, Emily M., Fischer, Andrew H., Deerinck, Thomas J. & Hetzer, Martin W. Catastrophic Nuclear Envelope Collapse in Cancer Cell Micronuclei. Cell 154, 47-60, doi:10.1016/j.cell.2013.06.007 (2013). 8 Umbreit, N. T. & Pellman, D. Cancer biology: Genome jail-break triggers lockdown. Nature, doi:10.1038/nature24146 (2017). 9 Haraguchi, T. et al. Live cell imaging and electron microscopy reveal dynamic processes of BAF- directed nuclear envelope assembly. Journal of Cell Science 121, 2540-2554, doi:10.1242/jcs.033597 (2008).

8 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

10 Dechat, T. et al. LAP2alpha and BAF transiently localize to telomeres and specific regions on chromatin during nuclear assembly. J Cell Sci 117, 6117-6128, doi:10.1242/jcs.01529 (2004). 11 Maeshima, K. et al. Cell-cycle-dependent dynamics of nuclear pores: pore-free islands and lamins. J Cell Sci 119, 4442-4451, doi:10.1242/jcs.03207 (2006). 12 Otsuka, S. et al. Nuclear pore assembly proceeds by an inside-out extrusion of the nuclear envelope. Elife 5, doi:10.7554/eLife.19071 (2016). 13 Toledo, L. I. et al. ATR prohibits replication catastrophe by preventing global exhaustion of RPA. Cell 155, 1088-1103, doi:10.1016/j.cell.2013.10.043 (2013). 14 Hatch, E. M. & Hetzer, M. W. Nuclear envelope rupture is induced by actin-based nucleus confinement. J Cell Biol 215, 27-36, doi:10.1083/jcb.201603053 (2016). 15 Hoffelder, D. R. et al. Resolution of anaphase bridges in cancer cells. Chromosoma 112, 389-397, doi:10.1007/s00412-004-0284-6 (2004). 16 Terradas, M., Martin, M., Hernandez, L., Tusell, L. & Genesca, A. Nuclear envelope defects impede a proper response to micronuclear DNA lesions. Mutat Res 729, 35-40, doi:10.1016/j.mrfmmm.2011.09.003 (2012). 17 Afonso, O. et al. Feedback control of chromosome separation by a midzone Aurora B gradient. Science 345, 332-336, doi:10.1126/science.1251121 (2014). 18 Karg, T., Warecki, B. & Sullivan, W. Aurora B-mediated localized delays in nuclear envelope formation facilitate inclusion of late-segregating chromosome fragments. Mol Biol Cell 26, 2227- 2241, doi:10.1091/mbc.E15-01-0026 (2015). 19 Fuller, B. G. et al. Midzone activation of aurora B in anaphase produces an intracellular phosphorylation gradient. Nature 453, 1132-1136, doi:10.1038/nature06923 (2008). 20 Steigemann, P. et al. Aurora B-mediated abscission checkpoint protects against tetraploidization. Cell 136, 473-484, doi:10.1016/j.cell.2008.12.020 (2009). 21 Gruneberg, U., Neef, R., Honda, R., Nigg, E. A. & Barr, F. A. Relocation of Aurora B from centromeres to the central spindle at the metaphase to anaphase transition requires MKlp2. The Journal of Cell Biology 166, 167-172, doi:10.1083/jcb.200403084 (2004). 22 Walther, T. C. et al. The conserved Nup107-160 complex is critical for nuclear pore complex assembly. Cell 113, 195-206 (2003). 23 Olmos, Y., Hodgson, L., Mantell, J., Verkade, P. & Carlton, J. G. ESCRT-III controls nuclear envelope reformation. Nature 522, 236-239, doi:10.1038/nature14503 (2015). 24 Vietri, M. et al. Spastin and ESCRT-III coordinate mitotic spindle disassembly and nuclear envelope sealing. Nature 522, 231-235, doi:10.1038/nature14408 (2015). 25 Franz, C. et al. MEL-28/ELYS is required for the recruitment of nucleoporins to chromatin and postmitotic nuclear pore complex assembly. EMBO Rep 8, 165-172, doi:10.1038/sj.embor.7400889 (2007). 26 Hochegger, H., Hegarat, N. & Pereira-Leal, J. B. Aurora at the pole and equator: overlapping functions of Aurora kinases in the mitotic spindle. Open Biol 3, 120185, doi:10.1098/rsob.120185 (2013). 27 Xue, John Z., Woo, Eileen M., Postow, L., Chait, Brian T. & Funabiki, H. Chromatin-Bound Xenopus Dppa2 Shapes the Nucleus by Locally Inhibiting Microtubule Assembly. Developmental Cell 27, 47-59, doi:10.1016/j.devcel.2013.08.002 (2013). 28 Lu, L., Ladinsky, M. S. & Kirchhausen, T. Formation of the postmitotic nuclear envelope from extended ER cisternae precedes nuclear pore assembly. The Journal of Cell Biology 194, 425-440, doi:10.1083/jcb.201012063 (2011). 29 Kurasawa, Y., Earnshaw, W. C., Mochizuki, Y., Dohmae, N. & Todokoro, K. Essential roles of KIF4 and its binding partner PRC1 in organized central spindle midzone formation. EMBO J 23, 3237- 3248, doi:10.1038/sj.emboj.7600347 (2004).

9 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

30 Barisic, M. et al. Mitosis. Microtubule detyrosination guides chromosomes during mitosis. Science 348, 799-803, doi:10.1126/science.aaa5175 (2015). 31 Otsuka, S. et al. Postmitotic nuclear pore assembly proceeds by radial dilation of small membrane openings. Nature Structural & Molecular Biology 25, 21-28, doi:10.1038/s41594-017- 0001-9 (2018).

10 a b Nocodazole Release Nocodazole Release

bioRxiv6 hrs preprint doi: https://doi.org/10.1101/26339260 mins ; this version posted February 11,6 hrs 2018. The copyright60 mins holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. anaphase/telophase interphase DNA LAP2 α LBR Merge DNA RFP-NLS LAP2α Nup133LAP2 Merge

DNA emerin Nup133 Merge DNA RFP-NLS emerin LBR LAP2 Merge

100

75 15 50 10 RPE-1 25 5 0

% Laggards % Laggards Labeled BAF LBR Tpr ELYS LAP2α emerin Nup62 1.5 Nup133mAb414 Nup153Nup358 lamin A/C

Core Non-core MN:PN 1.0

100 0.5 100 75 75 0.0 50 5hr 5hr 5hr 5hr 5hr 50 5hr mi n mi n mi n mi n mi n mi n 18hr 18hr 18hr 18hr 25 18hr U2OS HeLa K 90 90 90 90 90 25 90 0 0 LAP2α emerin lamin A/C LBR lamin B1 lamin B2 % % Laggards Labeled

% % Laggards Labeled emerin LBR emerin Core Non-core Core Non-core Core Non-core c d A normal import

RFP-NLS import B no import 6 min C < 35% normal import

GFP-H2B RFP-NLS D delayed import

40 40 0 AO 30 30 2 20 20 RPE-1 4 U2OS 10 10 6 % total micronuclei % total micronuclei 0 0 8 ABCD ABCD 10 Import defects Import defects 12

14 PN import e 1.25 1.25 16 18 1.00 1.00 20 0.75 0.75 22 0.50 MN:PN 0.50

24 MN:PN (RPA2) 0.25 0.25 26 0.00 28 0.25 0.50 0.75 1.00 1.25 RFP-NLS RPA 2 MN:PN (RFP-NLS) Figure 1 | Micronuclei undergo defective NE assembly.

11 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 1. Micronuclei undergo defective NE assembly.

a, b. Defective non-core NE protein recruitment to lagging chromosomes (a) or micronuclei (MN, b). Top panels: Experimental scheme. Middle panels: Representative images of RPE-1 cells with lagging chromosomes or MN (arrows, insets for enlarged images). Red letters: core NE proteins; Green: non-core proteins. (a) Bottom: quantification of the results (n > 50; ≥ 2 experiments). (b) Bottom: The fluorescence intensity (FI) ratio of the indicated proteins in intact MN relative to primary nucleus (PN) in RPE-1 cells at indicated timepoints after release from nocodazole block (mean with 95% CI, n > 100; 2 experiments). mAb414 detects nucleoporins. c, d, Impaired nuclear import in MN. c, Kymograph of RPE-1 cell after synchronization as in a, 2 min intervals (t=0 is anaphase onset, AO). Kymograph is from the boxed region of top image. Merged image at bottom shows poor accumulation of the import reporter, RFP fused to a nuclear localization signal (RFP-NLS) in the MN, which persisted for at least 2h (not shown). Arrowheads: newly formed MN. d, Top: Cartoon depicting patterns of import to MN: (A) Normal import kinetics; (B) No detectable import; (C) Delayed import with a persistent defect (up to 35% of PN) and (D) Delayed import with eventual normal RFP-NLS accumulation. Bottom: Percentage of cells corresponding to the categories above in the indicated cell lines (n=21 for RPE-1 and n=17 for U2OS). e, Left: The MN/PN FI ratio of RPA2 and RFP-NLS in RPE-1 cells ~1h post AO (mean with 95% CI, n > 200, 2 experiments). Right: Defective accumulation of RPA2 and RFP-NLS are correlated (r = 0.9039, P < 0.0001, Spearman’s correlation analysis). Scale bars, 10 µm.

bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was a not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Nocodazole Release Drug addition Fix AO 6 hrs Immunofluorescence Live-cell imaging staining

ABCDEFHI 12 mins

ABCDEFHI

Gridded coverslip b Live DNA pT232 LBR Merge

0 min 0 1.00 DMSO

0.75 **** NS 0.50 0 min 0 0.25 MN:PN (LBR)

0.00

ZM447439 0 2 4 6 8 10 DMSO

6 min 6 ZM addition (min post AO)

c d

DNA PN

5 µm

emerin-GFP Lagging

2 µm GFP-H2B RFP-NLS 2 µm

Primary nucleus NLS-positive MN NLS-negative MN Nup133 PN

2 µm

1 µm

Figure 2 | NE assembly defect of lagging chromosomes becomes irreversible in telophase.

13 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 2. NE assembly defect of lagging chromosomes becomes irreversible in telophase.

a, Scheme of the live-cell/ fixed cell imaging protocol. RFP-H2B-expressing RPE-1 cells were plated on gridded dishes to identify cells of interest. Cells were imaged at 2 min intervals during the experiments. Images of two live cells (red and blue boxes) upon (left) or after (right, prior to fixation) ZM447439 addition. b, Aurora B inhibition after late-anaphase fails to restore non-core (LBR) assembly to lagging chromosomes (arrows). Left column: representative cells (0 or 6 min post AO) at the time of ZM447439 (ZM) or DMSO addition. Right columns: Cells labeled for DNA (blue), phospho-T232 Aurora B (red) and LBR (green). For the control (DMSO) samples, lagging chromosomes fail to recruit LBR whether they are proximal (white arrowhead) or distal (red arrowhead) to activated Aurora B (pT232). Scale bars, 10 µm. Graph on right: MN/PN LBR FI ratio in cells exposed to ZM at the indicated times (mean with 95% CI, n > 50 each timepoint, ≥ 3 experiments, for simplicity, lagging chromosomes are designated as “MN”). **** P < 0.0001, NS: not significant, Mann-Whitney test. c, Near-continuous assembly of core membrane protein around the lagging chromosome. Images of 3D-SIM (structured illumination microscopy, left) and Imaris surface renderings (right) of an emerin-GFP-expressing RPE-1 cell. Enlarged images show lagging chromosomes (yellow box) and PN (white box): DNA (blue), emerin (red), Nup133 (green) (representative of 22 cells). d, Correlative light and electron microscopy (CLEM) showing membranous NE (enlarged images in middle panels, cross sections) but reduced NPCs (yellow arrows in bottom panels, tangential sections) on intact (RFP-NLS- positive) and newly disrupted MN (RPF-NLS negative, red arrows). Left, top: DIC image of an RPE-1 cell just after loss of NE integrity of one MN. Right, top: EM image; fixation was < 20 min after MN disruption. Enlarged images (middle and bottom panels) are at the same magnification.

a bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was Nocodazolenot Release certified by peer review)Drug is the author/funder. All rights reserved. No reuse allowed without permission. AO 6 hrs 4 mins 14 mins Fix Live-cell imaging

DNA emerin Nup133 Merge DMSO 100 NS **** NS NS **** NS 80 gt h (A.U) 60 ZM

20 Normalized Len 0 PTX ZM ZM PTX PTX PTX PTX DMSO DMSO ZM + ZM +

Nup133 gap emerin enrichment ZM +PTXZM b c DNA Tubulin LBR Merge **** **** 1.0 1.2 **** 1.0 0.8 0.8 0.6

ControlRNAi 0.6 0.4 0.4

MN:PN (LBR) MN:PN 0.2 0.2 0.0 0.0

RNAi Control KIF4A % Circumference (Nup133) RNAi Control KIF4A MKLP2 KIF4A RNAi KIF4A d e DNA emerin Nup133 Aurora B-GFP MergedMerge DNA ESCRT (CHMP4B) NPCs (Nup133) Microtubules Control siRNA Control ControlRNAi Kif4 siRNA Kif4 KIF4A RNAi KIF4A

2 µm

Figure 3. Independent of Aurora B, bundled microtubules inhibit non-core NE recruitment to lagging chromosomes.

15 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 3. Independent of Aurora B, bundled microtubules inhibit non-core NE recruitment to lagging chromosomes.

a, Paclitaxel (PTX) inhibits non-core assembly. Left, top: Experimental scheme. Left, bottom: representative images of cells after drug treatments. Merged image: DNA (blue), emerin (red) and Nup133 (green). Enlarged images from the boxed regions show the non-core (Nup133) gap (orange arrowheads). Right: quantification of the results (mean with 95% CI, n > 40, 2 experiments). **** P < 0.0001, NS: not significant, Mann-Whitney test. b, Restoration of LBR to some lagging chromosomes after KIF4A depletion (RPE-1 cells). Synchronization as in Fig. 1a. Left: Representative images: DNA (blue), tubulin (red) and LBR (green). Right: MN/PN LBR FI ratio (mean with 95% CI, n > 100, 3 experiments). **** P < 0.0001, Mann-Whitney test. c, d, Small-scale core/non-core domain separation on lagging chromosomes after KIF4A depletion. Synchronization as in Extended Data Fig. 1c. c, The percentage of the lagging chromosome circumference with Nup133 in HeLa K cells (mean with 95% CI, n > 60, 2 experiments). *** P < 0.0001, Mann-Whitney test. d, Representative images of Aurora B-GFP-expressing HeLa K cells. Merged and enlarged image: emerin (red, white arrowheads) Nup133 (green, white arrows). Scale bars, 10 µm. e, Imaris surface three-dimensional renderings from SIM image showing recruitment of Nup133 (white) to the region of a lagging chromosome depleted of CHMP4B (green) and microtubules (red) from a KIF4A-depleted HeLa K cell (representative of 4 lagging chromosomes).

bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. a d 2.5 **** **** **** TTL siRNA (HeLa K) RO-3306 Release into MPS1i 2.0 24 hrs 19 hrs Live-cell imaging (1-1.5h) Fix 1.5 Central +/- TTL siRNA chromosome 1.0 MN:PN 0.5

0.0 Mps1i Peripheral chromosome Central Central Central Peripheral Peripheral Peripheral

Nup133 RFP-NLS RPA2 b e DNA emerin LBR Merge TTL siRNA RO-3306 Release into MPS1i

24 hrs 19 hrs Live-cell imaging (16-18h) Fix

GFP-H2B RFP-NLS γ H2AX lamin B1

C-MN C-MN DNA LAP2 Nup133 Merge Central

2.0 **** 1.50 **** 1.25 Peripheral P-MN 1.5 P-MN 1.00 1.0 0.75 0.50 0.5 f

LBR (MN:PN) 0.25 NE disruption DNA damage ***RFP- NL S loss ( 5 comb I MF +N A) 0.0 0.00 **** G**** FP -B AF ( 2 co mb IMF***DNA ) da ma ge ( 3 co mb I MF by gH 2AX)

% Circumference (Nup133) n=187 lag gi ng Central Peripheral lag gi ng Central Peripheral 5 0 6 0 n =32 5 0 p ol ar 5 0 p ol ar 4 0 4 0 c 4 0

(γH2AX) ** 3 0 3 0 3 0 p <0 .0 05 * * n=89 2 0 2 0 n=181 2 0 p ai rwis e t-te st for m ean p <0 .0 2 ** 1 0 n =111 1 0 % total MN ruptur e % total MN ruptur e 1 0 n=58

% total MN total % p ai rwis e t-te st for m ean % total MN (GFP-BAF) MN total % % total MN (NLS loss) (NLS MN total % 0 0 0 % total MN with D NA dam age

p ol ar p ol ar p ol ar p ol ar Centrallag gi ng lag gi ng Centrallag gi ng lag gi ng Central GFP-Nup107 Peripheral Peripheral Peripheralp ol ar H 2AX p ol ar H 2AX lag gi ng H2A X lag gi ng H2A X 0 2 4 6 8 10 12 14 8 0 62 5 0 70 6 0 90 7

n =2 9 n =3 6 n =2 4 n =2 2 n =3 2 n =4 GFP-Nup107

AO RFP-H2B NE assembly

Figure 4. Peripheral localization of missegregated chromosomes corrects defects of micronuclei.

17 bioRxiv preprint doi: https://doi.org/10.1101/263392; this version posted February 11, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 4. Peripheral localization of missegregated chromosomes corrects defects of micronuclei.

a, Experimental scheme to generate missegregated chromosomes (MPS1i with or without TTL knockdown) positioned within or away from the spindle (microtubules: red; non-core NE: green). Cells were imaged at 4 min intervals throughout mitosis. b, c, Non-core NE assembles on peripheral chromosomes but not to central chromosomes. b, Top: Representative images of RPE-1 cells with central (yellow arrows) and peripheral (red arrows) chromosomes labeled with core (red) or non-core (green) proteins. Bottom: Quantification as in Fig. 3b, c (mean with 95% CI, n > 50, 2 experiments). **** P < 0.0001, Mann-Whitney test. c, Similar result as in b from live-cell imaging of HeLa K cell expressing RFP-H2B (red) and GFP-Nup107 (green) (Supplementary Video 4). AO is t=0 (min). d-f, Restoration of function for MN from peripheral chromosomes. d, FI ratios for the indicated proteins, comparing MN from peripheral with central chromosomes (RPE-1 cells, scheme as in a) (mean with 95% CI, n > 50, 2 experiments). **** P < 0.0001, Mann-Whitney test. e, Top: Experimental scheme. Bottom: Representative images of HeLa K cells (C-MN from central chromosomes; P-MN from peripheral chromosomes). f, Graphs of the results. Left: NE integrity monitored by loss of NLS-RFP (5 experiments). Middle: NE integrity monitored by hyper-accumulation of GFP-BAF (3 experiments). Right: DNA damage (3 experiments). ** P < 0.05, **** P < 0.0001, Fisher’s exact Test. Scale bars, 5 µm.