Supplementary Materials for - Science · 2016. 3. 23. · nailed to a block of PDMS to prevent the explant from moving during imaging and also . 5 to allow flow of fresh medium. Imaging

www.sciencemag.org/cgi/content/full/science.aad7611/DC1

Supplementary Materials for ESCRT III repairs nuclear envelope ruptures during cell migration to

limit DNA damage and cell death

M. Raab, M. Gentili, H. de Belly, H. R. Thiam, P. Vargas, A. J. Jimenez, F. Lautenschlaeger, Raphaël Voituriez, A. M. Lennon-Duménil, N. Manel, M. Piel

*Corresponding author. E-mail: [email protected]

Published 24 March 2016 on Science First Release DOI: 10.1126/science.aad7611

This PDF file includes

Materials and Methods Figs. S1 to S13 Full References Movies S1 to S10

Other Supplementary Materials for this manuscript include the following: (available at www.sciencemag.org/cgi/content/full/science.aad7611/DC1)

Movies S1 to S10

2

Materials and Methods

Channels preparation and cell loading

Micro-channels were prepared as previously described (10). Briefly,

polydimethylsiloxane (PDMS, 10/1 w/w PDMS A / crosslinker B) (GE Silicones) was

used to prepare 7µm wide and 5µm high micro-channels with constrictions of varying

lengths and widths from a self-made mold. For HeLa and U2OS cells, larger channels of

12 µm width and 5 µm height were used because of their larger nuclei. Channels with

constrictions were incubated with 15 µg/mL of fibronectin for 30 min then washed with

PBS at least 3 times and finally incubated with medium (containing drugs if necessary)

for at least 5 hours before adding cells. Note that the height inside constriction varied

depending on the width: height was 2 µm for 1.5 µm wide constrictions, 3.5 µm for 2 µm

wide constrictions, 4µm for both 3 µm and 4 µm wide constrictions, and 5 µm for 5 µm

wide constrictions. 100,000 cells in 5 uL of medium was added to each reservoir, and

after 30 min of incubation, medium was added to above the PDMS construct.

Cell Confinement

For confined migration in Fig. 1, LifeAct mDCs were plated on glass bottom 6 well

plates either with or without 1.6 mg/ml bovine collagen (Advanced BioMatrix) (Filling or

No Filling) and then a 5 µm roof of PDMS was placed on top, as previously described

(12, 30). Briefly, silicon wafers were coated with SU8 2005 photoresist (Microchem) 5

µm in height and holes were made in lithography. To make the PDMS pillars as the 5 µm

height spacers, 10 mm glass coverslips were plasma treated and then placed on top of

PDMS mixture (10/1 w/w PDMS A / crosslinker B) on the wafer with the holes. After

baking at 95°C for 10 min, coverslips with PDMS pillars were carefully removed from

the wafers under isopropanol. They were then cleaned with isopropanol, well-dried,

treated with plasma for 1 min, and treated with 0.5 mg/mL pLL-PEG in 10 mM pH 7.4

HEPES buffer for 1h at room temperature. Coverslips with PDMS pillars were rinsed and

incubated in medium for at least 5 hours before confining the cells. The modified cover

lid of a multi-well plate was used to apply confining slides to cells. In this case, large

PDMS pillars were stuck on the cover lid of the multi-well plate to hold confining slides.

There are pillars that push the confining slides from the top of the plastic 6 well plate lid

to confine the cells in 6 well glass bottom plates. The process of fabrication for these

large pillars attached to the 6 well plate lid is as follows: the large PDMS pillars were

fabricated by pouring a PDMS mixture (A:B = 35:1) into a custom-made metallic mold,

removing bubbles under vacuum, then baking overnight at 80°C, and getting the pillars

out of the mold with the help of small amount of isopropanol.

For compression to break the nucleus, cells on FluoroDishes (WPI) were squeezed

with confining structures of PDMS on glass slides as previously mentioned above except

the height for HeLa cell confinement to break the nucleus was 3 µm. A suction cup was

used to press the confining PDMS pillar coverslip onto the FluoroDish. The suction cup

device was molded in PDMS by pouring the crosslinker/polymer mix (1/10 w/w) into a

3

custom-made mold and baked at 80°C for 30 min before unmolding. Applying a vacuum

causing the confining coverslip to press down onto the cells (30).

Mice

Bone marrow was taken from C57BL/6 mice, to differentiate immature mouse

dendritic cells (mDCs), described in (31). LifeAct-GFP mice were a kind gift from

Michael Sixt lab (IST, Austria), and generated as described (32). Mice were taken care of

by the guidance of the Animal Welfare Body, Research Centre, Institut Curie.

Cells

Human monocytes were isolated from peripheral adult human blood as previously

described (38). Monocytes were cultured and differentiated into dendritic cells (hDCs) in

RPMI medium with Glutamax, 10% FBS, PenStrep, Gentamicin (50 μg/ml, GIBCO), and

HEPES (GIBCO) in the presence of recombinant human GM-CSF (Miltenyi) at 10 ng/ml

and IL-4 (Miltenyi) at 50 ng/ml. For experiments involving hDCs, 3 separate human

donors were used.

Immature mouse bone-marrow derived dendritic cells (mDCs) were cultured 10-12

days in IMDM supplemented with 10% FBS, Glutamine (20mM), pen-strep (100U/mL)

and 2-mercaptonethanol (50µM), and granulocyte-macrophage colony stimulating factor

(50 ng/mL)-containing supernatant obtained from transfected J558 cells, as previously

described (31). After 4 days of differentiation, all cells are passed to a density of 10-20

million per 120 cm² and repeated again at day 7. Dendritic cells were then used for

experiments between days 10 and 12. To activate and mature mDCs to express CCR7

(C-C chemokine receptor type 7) and chemotax against gradients of chemokine CCL21,

(C-C motif) ligand 21, in collagen, immature mDCs were treated with LPS (100 ng ml−1)

for 30 min and washing 3 times with complete medium and then the experiment was

performed the next day.

HeLa cells were cultured in DMEM Glutamax (Gibco) supplemented with 10% FBS

(GE Healthcare) and 1% penicillin and streptavidin (Lonza). The HeLa cells expressing

CHMP4B-EGFP at endogenous levels were a kind gift from Antony Hymann’s lab (16).

RPE-1 cells were grown in DMEM-F12 Glutamax medium (Gibco), supplemented with

10% FBS and 1% penicillin and streptavidin (Lonza). RPE-1 cells expressing 53BP1-

GFP were obtained from the lab of René Medema (21). RPE-1 cells expressing NLS-GFP

were obtained from the lab of Martin Hetzer (28). Stable cell line of HeLa expressing

MS2-mCherry-NLS was kindly made in the Buzz Baum lab. HeLa cell line expressing

LAP2β-GFP and H2B-mCherry were a kind gift from Mark Petronczki lab (33). U2OS

cells expressing LaminA-GFP were a kind gift from the lab of Harold Hermann.

Constructs

The plasmids pSIV3+, psPAX2, pCMV-VSV-G and pTRIP-CMV were previously

described (34, 35). Human cGAS WT open reading frame was amplified by PCR from

4

cDNA prepared from monocyte-derived dendritic cells. Human cGAS E225A/D227A

mutant was obtained by overlapping PCR mutagenesis. The mutation E225A/D227A

inactivates the catalytic site of cGAS impeding cGAMP production with consequent IFN

production, but leaves intact the protein ability to bind DNA.FLAG-cGAS and FLAG-

cGAS E225A/D227 were cloned in pTRIP-CMV in frame with GFP to obtain pTRIP-

CMV-GFP-FLAG-cGAS and pTRIP-CMV-GFP-FLAG-cGAS E225A/D227A. pTRIP-

CMV-tagRFP was generated by substituting the tagRFP sequence (RFP hereafter;

Evrogen) to the EGFP sequence of pTRIP-CMV. pTRIP-CMV-tagRFP-FLAG-cGAS

was generated by cloning FLAG-cGAS in frame. pTRIP-SFFV was generated by

substitution of the CMV promoter with the SFFV promoter from GAE-SFFV-GFP-

WPRE (36). pTRIP-SFFV-EGFP-NLS (NLS-GFP hereafter) was generated introducing

the SV40 NLS sequence (PKKKRKVEDP) by overlapping PCR at the C-terminal of

GFP in pTRIP-SFFV.

Plasmids and transfections

Transfections on HeLa cells were done using Lipofectamine 2000 (Thermo Fisher)

with a ratio of 1 ug DNA per 1 uL lipofectamine reagent. RPE-1 cells were transfected

with LTX (Thermo Fisher) according to the manufacturer’s instructions. NES-2xGFP

was provided by the Hetzer lab (28) and the NLS-MS2-mCherry was a gift from the lab

of Alessandro Marcello. MS2 signifies coat protein of phage.

Collagen gel with CCL21 gradient

Adapted from (37): mDCs transduced with NLS-GFP were mixed at 4 °C with 1.6%

bovine collagen type I (Advanced BioMatrix) at basic pH. Forty microliters of the mix

was deposited on a 35 mm glass -bottom dish and the drop was homogenized while

covered with a 12 mm glass slide. The sample was incubated at 37 °C for 20 min to allow

collagen polymerization. To generate the CCL21 gradient, 2 ml of BMDC medium

containing 200 ng ml−1 of CCL21 (R&D systems) was added to the plate. Cells were

imaged at a frequency of 1 image per 30 sec using a 20x objective using a spinning disk

microscope, with a z stack of 50 µm total height and 5 µm step size.

mDC Migration in Mouse Ear Explants

Ears from C57BL/6 mice were excised and a pair of forceps was used to create a

hole on the skin. The ventral and dorsal sides of the explant were separated by peeling.

The ventral sheet was kept and immunostained with anti-LYVE-1 (R&D Systems)

primary antibody to mark the lymphatic vessels. After washing with media, a secondary

antibody against rat (Jackson Immunoresearch) was used. The ear sheet was then spread

flat in a 6 well plate and a PDMS block with a central hole of diameter 8 mm was placed

on top of each explant with the ventral side up. 200,000 DCs expressing NLS-GFP (see

below) were added in 100 µL of culture medium inside the hole. After 1 h of incubation,

the ear sheet was washed with culture medium and then placed with the face on which

cells were incubating against the bottom glass slide in a FluoroDish. Ear explants were

nailed to a block of PDMS to prevent the explant from moving during imaging and also

5

to allow flow of fresh medium. Imaging was performed on an inverted spinning disk

confocal microscope, at 37°C and with 5 % CO2, taking z-stacks 50 µm in total height

using a 5 µm step size with a 20x objective.

For quantification of nuclear deformation in ear explants in Fig. 1, dendritic cells

were prelabeled with CFSE (Carboxyfluorescein succinimidyl ester) (ThermoScientific)

according the manufacturer’s instructions and then allowed to migrate in the ear explants

for 5 hours before fixation with paraformaldehyde.

Quantifications

In Fig. 1D,E, nuclear circularity was quantified by thresholding the Hoechst signal,

as = 4π(area/perimeter²), and the minimum diameter was taken as the minimum diameter

of the pinched shaped nucleus.

For nuclear leakage quantification, nuclear circularity was quantified by

thresholding the Hoechst or NLS-GFP signal and taking circularity=4π*(area/perimeter²).

To quantify nuclear leakage of NLS-GFP, a small ROI was put in front of the nucleus at

each time frame. For measuring leakage in channels as ‘NLS-GFP Intensity

Cytoplasm/Nucleoplasm’, the average intensity of the ROI of the cytoplasm in front of

the nucleus was divided by the average intensity of the nuclear NLS-GFP signal before

the nucleus entered the constriction. This takes into account differences in expression

levels of NLS-GFP for each cell. For quantifying leaking of NLS-MS2-mCherry after

photoablation, a small ROI next to the nucleus near the site of wounding was measured

for average intensity for each time frame after background subtraction.

To quantify the ‘survival after one constriction’ a cell death was counted if the

nucleus apoptosed after passing a constriction during the rest of overnight imaging

movie. If a cell moved through another constriction during the movie, then the cell was

counted as surviving another constriction. The ‘frequency of nuclear opening’ was

quantified by assessment of entry of tagged cGAS or exit of NLS during nuclear

engagement with the constriction.

To quantify ‘NLS signal recovery’ in comparing the NLS-GFP leakage with the

knockdowns of LMNA and CHMP3, the time was measured from after the nucleus exits

the constriction until full recovery of NLS-GFP back into the nucleus as observed with

20x microscopy. An empty circle indicates that full recovery of NLS-GFP occurred

before the nucleus has completely exited the constriction.

For laser wounding experiments, the ‘Timing for first appearance’ was measured as

the time after photoablation until the time of a fluorescent signal appeared above the

background intensity.

siRNA

siRNA 5’-AAA GCA UGG ACG AUC AGG AAG-3’ was used to deplete CHMP3

(13), , both 5’-GGUGGUGACGAUCUGGGCU-3’; and 5’-

6

AACUGGACUUCCAGAAGAACAUC-3’to target LMNA (4). Non-targeting siRNA

(Dharmacon, GE Life Technologies) was used as the control. siRNA was transfected with

Lipofectamine RNAiMAX (Invitrogen). Cells were transfected with 120nM siRNA 48 h

and again 24 h before the experiment.

Lentivector transductions

Transduced hDCs were obtained by infecting monocytes purified from blood with

pTRIP-CMV-GFP-FLAG-cGASE225A/D227A lentiviral vectors. Transduction of

freshly isolated monocytes from blood has been adapted from (35). 2 million monocytes

were seeded in a 6 well plate in 2 ml of medium. 2ml of fresh virus and 2 ml of SIV-VLP

were added to each well in presence of GM-CSF, IL-4 and 8µg/ml of Protamine

(SIGMA). At day 4 or 5 of differentiated cells were resuspended in fresh medium with

fresh cytokines and used in assays.

Transduced mDCs were obtained by transduction of murine bone marrow from

C57BL/6 mice. 1.8 million bone marrow cells were plated in a 6 well plate at day of

purification (day 0) in 2 ml of medium. At day 1, 40ml of fresh pTRIP-SFFV-GFP-NLS

lentivector supernatant were loaded in Ultra-Clear Centrifuge tubes (Beckman Coulter)

and ultracentrifuged at 100,000g in a SW32 rotor (Beckman coulter) for 90 minutes at

4°C and resuspended in 400µl of in RPMI medium with Glutamax, 10% FBS, PenStrep,

Gentamicin (50 μg/ml, GIBCO). 200µl of ultracentrifuged virus were used to infect one

well of cells in presence of 8µg/ml of Protamine. Cells were then differentiated for 11

days and split as described above.

For HeLa and RPE-1 cells transduction of XFP-cGAS, 0.5 million cells were plated

in a 6 well plate in 1ml and infected with 2ml of fresh lentivector, pTRIP-CMV-GFP-

FLAG-cGAS for HeLa H2B-mCherry cells and pTRIP-CMV-tagRFP-FLAG-cGAS for

RPE-1 53BP1-GFP cells in presence of 8µg/ml of Protamine. The cells were then FACS-

sorted by gating on the brightest GFP/RFP-positive cells.

Lentiviral particles production in 293FT cells

Lentiviral particles were produced as previously described from 293FT cells (38).

Lentiviral viral particles and viral-like particles were produced by transfecting 1μg of

psPAX2 and 0.4μg of pCMV-VSV-G together with 1.6μg of a lentiviral vector plasmid

per well of a 6-well plate.

Antibodies and Reagents

For imaging the nucleus, cells were incubated with 200 ng/mL of Hoechst 33342

(Life Technologies) or 34580 (Thermo Fisher) for confocal imaging with 405 nm lasers

for 30 minutes at 37 °C and 5% CO2 before washing with fresh medium. The following

primary antibodies were used for immunoblotting: LmnA/C (H110, Santa Cruz), anti-

actin (Millipore), anti-CHMP3 (Santa Cruz), and for immunofluorescence; monoclonal

mouse Anti-phospho-Histone H2A.X (Ser139) (Millipore), anti-Lamin-B1 Nuclear

7

Envelope marker (Abcam), Anti-Nuclear Pore Complex Proteins antibody [Mab414]

(Abcam), anti-CHMP2B (Abcam).

For immunofluorescence, secondary antibodies anti-mouse-Alexa488, rabbit-Cy5,

and anti-Goat-Alexa488 (Jackson ImmunoResearch Laboratories) were used. ATMi

(KU-55933, TOCRIS) was used at 10 µM to inhibit DNA repair.

Immunoblotting

Cells were lysed on ice for 45 – 60 minutes in a buffer containing 100 mM Tris, 150

mM NaCl, 0.5% NP-40, and 1:25 of protease inhibitor cocktail (Roche). Thirty

micrograms of soluble extracts were loaded onto a 4-20% TGX gradient gel (BioRad)

and transferred onto a Trans-Blot Turbo PVDF/Nitrocellulose membrane (BioRad). The

membrane was blocked, incubated with the appropriate antibodies and revealed with

SuperSignal West Dura substrate (Thermo Scientific).

Photodamage and Time-Lapse imaging

Cells were cultured in Leibovitz’s L-15 CO2-independent medium (GIBCO) with

10% FBS during the wounding and acquisition. Spinning-disc confocal microscopy was

carried out with a Yokogawa CSU-X1 spinning-disc head on a Nikon Eclipse Ti inverted

microscope equipped with an EMCCD camera (Evolve, Photometrics), a NanoScanZ

piezo focusing stage (Prior Scientific) and a motorized scanning stage (Marzhauser) and a

Nikon S Fluor 100X/1.3 NA objective. The UV-laser damage experiments were

performed with a pulsed 355nm ultraviolet laser (Roper Scientific) driven by iLas

software. This microscope was operated with Metamorph and images were processed

with FIJI.

8

9

Fig. S1.

Dendritic cells deform their nucleus when migrating through collagen and tissue and

NLS-GFP leaks from nucleus to cytoplasm. (A) Large field view of the image of mouse

dendritic cells (mDCs) expressing LifeAct (green) and DNA labeled with Hoechst (blue).

White box indicates cell magnified in Figure1A. Cylindrical pillars of PDMS maintain

10

the 5µm height of 2D confinement. mDCs were free to migrate around the empty space

which was filled with collagen for the ‘collagen filling’ condition. (B) Large field view of

the image of mDCs shown in Figure1C (white box), labeled with CFSE (green) and

added to ear explant. LYVE1 antibody staining (red) shows location of lymphatic

vessels. Hoechst staining of entire ear explant shows the crowded environment of

endogenous cells. (C) Large field view of the image of mDCs in collagen moving across

a gradient of CCL21. Box indicates magnified cell, shown as raw images before

conversion to false colors. (D) Large field view of the image of mDCs expressing NLS-

GFP (green) added to ear explant. Lymphatic vessels indicated with LYVE1 staining

(red). White box depicts magnified cell, shown as raw images before conversion to false

colors. (E) Another example of mDCs migrating in an ear explant expressing NLS-GFP

but with an additional stain of Hoechst to stain the nuclei. Scale bars for large fields view

are 50 µm. Other scale bars are 10 µm.

11

Fig. S2

HeLa and RPE-1 cells slowly migrating in microchannels with constrictions also show

NLS-GFP leakage from nucleus to cytoplasm when the nucleus passes through the

constriction. (A) Large field of view image of the reservoir where cells were added and

allowed to migrate spontaneously into channels. The channel contains successive

constrictions. Scale bar is 100 µm. (B) Higher magnification image of channels with

constrictions. Below: Phase contrast image of a mDC migrating through a constriction 15

µm length and 2 µm width. Scale bar 20 µm. (C) HeLa cell expressing NLS-MS2-

mCherry slowly passing a constriction 15 µm length and 2 µm width. False color was

applied to better show the increase in NLS-MS2-mCherry in the cytoplasm. (D) RPE-1

cell expressing NLS-GFP migrating through a constriction of 15 µm length and 1.5 µm

width. Scale bars 10 µm.

12

Fig. S3

Nucleoplasm leaks out of the nucleus as bursts of varying size from nuclear constriction.

(A) Quantification of cytoplasmic NLS-GFP for the hDC shown in (Fig. 2A) with the

nuclear circularity determined from Hoechst labeling. (B) Quantification of NLS leakage

into the cytoplasm for the HeLa cell shown in Fig. 2B. 4 burst events are indicated on the

13

graph, with events 2 and 4 depicted in the images to the right of the graph with colored

dots to indicate location in time on the graph. The size of the blebs vary during nuclear

passage, and so an example of a HeLa nucleus passing with very small blebs which

constantly form and burst is shown in (C). The increase in NLS-MS2-mCherry in the

cytoplasm is quantified in the graph and appears to continuously increase because there

are many small blebs rupturing. Constriction L=15 µm w=2µm. (D) Example of a

nucleus passing that forms larger blebs. Rupture of these blebs cause dramatic increases

in the NLS-MS2-mCherry in the cytoplasm, with one example shown in the graph. Also

to note, when the nucleus is becoming deformed when first entering the constriction, the

NLS-MS2-mCherry in the cytoplasm can increase very slightly, pointed out during the

early phase of the graph before plateauing, which could be attributed to an increase in

nuclear pressure. Constriction L=15 µm w=4µm. (E) During the formation of blebs,

there was no increase in NLS-MS2-mCherry in the cytoplasm, only after rupture. The

time between the formation of blebs and their rupture was measured as ‘Lifetime of bleb’

in the histogram plot. Most blebs rupture in less than 10 min. n=43, N=2. (F) Viability of

cells after passing a constriction for constrictions of 15 µm in length and 4 µm in width

(less nuclear constriction than Fig. 2E of constriction width=2 µm). n=10 for HeLa, n>20

for hDCs and RPE-1, N=2. Scale bars are 10 µm.

14

Fig. S4

Compression of the nucleus and migration through constrictions cause entering of

cytoplasmic GFP-cGAS into the nucleus. (A) hDC expressing GFP-cGAS (green) and

15

DNA labeled with Hoechst (red). The cell was compressed with a PDMS roof until

nuclear blebs were observed (dark region between DNA and cytoplasm in green with

GFP-cGAS, noted with white arrow). The bleb then burst and –GFP-cGAS bound to

DNA along the nuclear edge locally where the bleb originated (B) Frequency of GFP-

cGAS entry in the nucleus for varying constriction dimensions for hDCs. , n>25 for each

condition, N=3 (C) Position along the constriction at which GFP-cGAS entered the

nucleus for various constriction dimensions. Dashed lines indicate the exiting end of the

constriction, thus points after the dashed line signify breaks occurring after the front

nuclear tip had passed the constriction. (D) Location of GFP-cGAS entry in the nucleus

for HeLa cells expressing GFP-cGAS migrating through constrictions, L=15 µm, w=2

µm. n=19, N=2. (E) Plot of total intensity of GFP-cGAS which has entered the nucleus

in the HeLa cell depicted in Fig. 2I. while passing the constriction. Green arrows

indicate sequential events of entry of GFP-cGAS. (F) Additional example of pronounced

rupture events. R indicates a rupture event with a sharp increase in GFP-cGAS,

quantified in (G). The gradual decrease in GFP-cGAS after the ruptures indicate the slow

export of GFP-cGAS back out of the nucleus into the cytoplasm. Step increases larger

than 2 standard deviations are colored red. Scale bars are 10 µm.

16

17

Fig. S5

Migration through constrictions causes rupturing of the lamina which is then reformed.

(A) U2OS cell expressing LaminA-GFP migrating through a constriction. The Lamina

ruptures and reforms constantly at the tip of the nucleus. (B) mDC with nuclei inside

constrictions fixed and stained for nuclear pore complex protein (NPC) (green), DNA

18

with Hoechst (blue) and LaminB (red). (C) mDC stained for NPC and DNA. The tip of

the nucleus appears devoid of nuclear pores or LaminB staining (white arrows). (D-F)

Three representative examples of HeLa BAC cells expressing LAP2β-GFP and H2B-

mCherry migrating through constrictions (L=15 µm, w=2 µm). New LAP2β-GFP

quickly reforms around DNA herniations. In some cases, threads of chromatin could

extend in front of the nucleus (E), but the nuclear envelope quickly reformed around

them. Multiple blebs on top of other blebs could also be observed (F). Scale bars are 10

µm.

19

Fig. S6

Minimal physical model for nuclear deformation during cell migration through a narrow

pore.

The cell nucleus has a complex rheology. It is often described effectively as a linear

elastomer, whose elastic component is largely due to the lamina layer (4). For slow

deformations, as involved in cell migration, a viscous behavior of the nucleoplasm is

expected. Based on these observations, we assume that the nucleus can be modeled as a

viscous fluid enclosed by the NE, a double membrane (ONE: outer nuclear membrane,

INE: inner nuclear membrane). The lamina layer is modeled as a thin elastic membrane

connected to the NE by linkers such as LAP2β and SUN proteins.

We consider a cell migrating through a narrow pore or constriction. When the nucleus is

engaged in the constriction, it induces a physical barrier between the back and the front of

the cell cytoplasm (defined by the direction of migration). This allows for a difference of

hydrostatic pressure in the cytoplasm between the back (Pb) and front (Pf) compartments.

We assume that the contractility of the actomyosin cortex at the back of the cell increases

upon engagement of the nucleus in the pore; we will therefore assume that Pb increases

over the process, and in particular Pb> Pf. Force balance at the NE is then expressed by

Laplace law, which relates the pressure in the nucleus Pn and the total surface tension

𝛾𝑛 of the NE and lamina to Pb and Pf :

𝑃𝑛 − 𝑃𝑏 = 𝛾𝑛𝐶𝑏 , 𝑃𝑛 − 𝑃𝑓 = 𝛾𝑛𝐶𝑓 (1)

20

where Cb (resp. Cf) denotes the curvature of the nucleus at the back (resp. the tip). For

very narrow pores, as in experiments, one has 𝐶𝑓 ≫ 𝐶𝑏 so that

𝑃𝑛 − 𝑃𝑓 ≈ 𝑃𝑏 − 𝑃𝑓 ≈ 𝛾𝑛𝐶𝑓. (2)

This description suggests the following mechanism. Upon the build-up of 𝑃𝑏 − 𝑃𝑓 by

tension of the actomyosin cortex, the pressure difference 𝑃𝑛 − 𝑃𝑓 at the tip of the nucleus

is increased, leading to a nuclear protrusion in the pore. This deformation eventually

increases the total area of the NE and therefore its tension.

The specific rheology of the lamina layer is still to be determined; we assume here that it

can be described as a linear elastic membrane. We therefore write 𝛾𝑛 ≈ 𝛾0𝜀, where ε is

the strain of the lamina layer and 𝛾0 a constant. In the case of microfabricated pores of

small rectangular cross-section 𝑙 × ℎ, the larger strain is found in the extension of the

nuclear protrusion in the pore; it can be written 𝜀 ≈∆𝐿

𝐿0 where L0 denotes the initial length

of the nuclear protrusion length when strain starts increasing, and ∆L is the increase of

the length of the nuclear protrusion. In turn the curvature at the front nuclear interface can

be estimated as

𝐶𝑓 ≈2

𝑙+

2

ℎ. Equation (3) then related the deformation ∆L observed in response to a given

pressure difference 𝑃𝑏 − 𝑃𝑓:

𝑃𝑏 − 𝑃𝑓 ≈ 𝛾0 (2

𝑙+

2

ℎ) ∆𝐿/𝐿0 (3)

This shows in particular that for narrow pores (l,h small), a larger pressure difference is

needed to deform the nucleus through the full pore length.

Following the mechanism of formation of plasma membrane blebs, we hypothesize that

the pressure difference between the nucleoplasm and the cytoplasm, when it reaches a

threshold value Pt, leads to the breakage of the links connecting the lamina layer to the

NE, resulting in the formation of a nuclear bleb.

The above analysis then implies that: (i) nuclear blebs are formed preferentially at the

front of the nucleus, where the pressure difference is the largest (𝑃𝑛 − 𝑃𝑓 ≫ 𝑃𝑛 − 𝑃𝑏).

This is consistent with observations (Fig 2F, I). (ii) Formation of nuclear blebs are more

likely for narrower pores, for which a larger pressure increase 𝑃𝑏 − 𝑃𝑓 is needed for

nuclear translocation through the full pore length. This is consistent with observations

(Fig. 2G). (iii) The length ∆L of the nuclear protrusion when blebbing starts is larger for

larger pores (see Eq. (3)), as observed (Fig 2H).

It should be noted that the increase in tension 𝛾𝑛 could also induce the rupture of the

lamina layer, as is observed when chromatin invades the bleb (Fig S5). This is however

not the rule since a physical barrier is often preserved between small blebs and the

chromatin, when the lamina remains intact.

21

Last, when the bleb is formed, the pressure difference 𝑃𝑛 − 𝑃𝑓is supported locally by the

NE only, inducing an increase of its tension and potentially its rupture.

Overall, this simple model of nuclear passage through narrow pores and nuclear blebs

formation is therefore fully consistent with experimental observations.

22

Fig. S7

Recruitment of ESCRT-III occurs at ruptured nuclear blebs during compression and also

at sites of laser wounding at nuclear edge. (A) Images of a representative HeLa cell

expressing CHMP4B-EGFP (green) and DNA stained with Hoechst (blue) which was

23

compressed and held to 3 µm in height. This compression caused formation of nuclear

blebs which could be detected by the elevated CHMP4B-EGFP signal in the nucleus. The

blebs protruded away from the DNA and after they burst, a transient recruitment in

CHMP4B-EGFP was observed along the edge of the DNA, white arrow. Scale bar for

low magnification image 10 µm, cropped images 5 µm. (B) Representative images of a

HeLa cell expressing CHMP4B-EGFP before and after UV laser wounding at 2 spots.

One along the nuclear edge and a second one farther randomly into the cytoplasm as a

control (yellow circles). Localized enhancement in CHMP4B-EGFP was observed for the

region along the nuclear edge but not for the second ablated region in the cytoplasm. (C)

Quantification of the kinetics of the increase in CHMP4B-EGFP after photoablation, n=8

cells, N=3 (D) Images of a representative HeLa cell expressing both CHMP4B-EGFP and

NLS-MS2-mCherry before and after laser wounding. The XZ plane shows that the

CHMP4B-EGFP is recruited along the contour of the nuclear edge. (B, D) Scale bars 10

µm.

24

Fig. S8

Endogenous ESCRT-III is also recruited to sites of laser wounding at nuclear edge. (A)

Images of HeLa cell expressing CHMP4B-EGFP before laser wounding on NE. White

box shows the ROI in (B) before and time sequence after laser wounding. Site of

wounding indicated with yellow arrow. After recruitment of CHMP4B-EGFP was

25

visualized in live micrsocopy, the cell was fixed and immunostained for endogenous

CHMP2B, shown in (C). CHMP2B (red) co-localized with CHMP4B-EGFP (green) at

the nuclear edge, yellow arrow. (D) An additional example of this experiment. (E)

Enhancement of ROI. First, a laser wounding was made and confirmed with live imaging

for CHMP4B-EGFP recruitment, and then a second wounding was made after 7 min and

the cell was then fixed shortly after. The woundings are marked ‘1’ and ‘2’. (F) The same

cell was found with the cell shape from the CHMP2B immunofluorescence. Z stack was

taken in the higher region above the cell surface where the wounding took place on the

nuclear edge. The immunostain for CHMP2B reveals weak recruitment to the first site of

wounding and strong accumulation as the second wounding. This agrees well with the

concept that after wounding, ESCRT III is recruited transiently for repair and slowly

diminishes over time.

26

Fig. S9

Laser wounding at nuclear edge causes leakage of NLS-MS-mCherry out of the nucleus.

(A) Images of HeLa cell expressing CHMP4B-EGFP before and after laser wounding on

NE. ROI denotes the recruitment of CHMP4B-EGFP. (B) NLS-MS2-mCherry images

27

before and after laser wounding and their corresponding false color images in (C). Both

images where montaged together as one image to keep the same brightness and contrast

settings. (D) Quantification of NLS-MS2-mCherry of a small ROI of cytoplasm adjacent

to the nucleus after laserablation of the nuclear edge. n=4, N=2. Scale bar is 10 µm.

28

Fig. S10

ESCRT-III is involved in resealing the NE. (A) Quantification of subsequent opening

events as the nucleus continued to pass through the constriction for the cell shown in Fig.

3A. (B) Time for rise to maximum intensity for RFP-cGAS and CHMP4B-EGFP for

several break and resealing events (n=7, N=2). (C) Time delay between laser wounding

and first appearance of CHMP4B-EGFP at the nuclear edge for HeLa cells treated with a

29

control SiRNA or a CHMP3 SiRNA. n≥6 for each condition, N=2. (D) Images of a

representative RPE1 cell expressing NLS-GFP, after SiRNA against LMNA/C, migrating

in a channel without constrictions, showing spontaneous leakage events without nuclear

constriction. (E-G) RPE-1 cells expressing NLS-GFP passing through constrictions

(L=20µm, w=3µm), after various treatments with siRNA. Less nuclear constriction with

constriction width=3µm was chosen to more clearly detect more NLS-GFP leakage, refer

to Fig. 2C.

30

Fig. S11

Controls for siRNA knockdowns and efficacy of ATMi drug. (A-B) Quantification of

efficiency for siRNA knockdown of RPE-1 cells for CHMP3 (A) and LMNA/C (B),

normalized to β-actin, quantified in bar plots underneath. Non-targeting siRNA was used

as siCTL. (C) Control for efficacy of ATMi (KU-55933) treatment. mDCs in culture

were irradiated (IR) with 10 Gγ with or without 10 µM of ATMi. ATMi reduced the total

intensity of γH2AX immunostaining by nearly 50% in cells fixed 30 min after irradiation.

N>30 for each condition, N=2. Scale bar is 10 µm.

31

32

Fig. S12

Deciphering the precise mechanism by which nuclear deformation in migrating cells lead

to NE opening would require further investigation, but our observations suggest the

following model (see also Fig S6). Migrating cells, when facing a constriction, exert a

force on their nucleus to deform it; this produces an increase in nuclear surface tension

and in intra-nuclear pressure (as seen in experiments in which the nucleus is deformed by

application of an external compression, 12); increased intra-nuclear pressure would lead

to formation of nuclear envelope blebs (separation of the inner and outer NE membranes

from the underlying nuclear lamina), which would eventually rupture, inducing a

transient leakage of nuclear proteins in the cytoplasm and likewise allow cytoplasmic

proteins to enter the nucleus. The rupture of the double membrane would spontaneously

lead to the formation of pores, to which ESCRT complex components would be recruited,

leading to rapid sealing, similar to what happens at the end of mitosis during NE

reformation (14, 15) or plasma membrane repair (13). The fact that similar mechanisms

underlie membrane repair in these different contexts is consistent with the similar kinetics

observed.

Schematic of the process: as the nucleus flows through the constriction, the NE (blue

lines) dissociates from the lamina layer (red line), and forms a nuclear bleb containing

nucleoplasm (light blue) but not chromatin (not depicted here). The NE bleb then

collapses as the NE breaks and opens. This is because NE rupture releases the NE surface

tension. While the NE remains open, nuclear proteins (light blue) can diffuse in the

cytoplasm (the whole cell is now light blue). Similarly, cytoplasmic proteins (not

depicted here) can enter the nucleus. This is visualized by the binding of the cytoplasmic

DNA binding protein (cGAS, green) at the tip of the nucleus.

After the rupture of the two nuclear membranes occurred, likely due to increased tension

following the bleb formation, blunt lipid membrane edges are formed; this is a highly

unstable state, which would rapidly result in the fusion of the inner and outer membranes,

to minimize lipid-water free energy.

This would result in the formation of a pore in the double membrane. Such a topology is

highly favorable for recruitment of the ESCRT III complex (dark blue helix). The

ESCRT III complex would allow fusion of the membranes in the pore, thus closing it.

33

Fig. S13

Influx of NES-GFP from cytoplasm to nucleus during nuclear constriction. (A) RPE-1

cell expressing NES-GFP migrating through a constriction L=15µm, w=3µm. Brightfield

image and confocal plane through the nucleus inside the constriction. The nucleus

normally is dark when the NES-GFP is exported out of the nucleus. (B) R1 and R2 note

34

the times when NES-GFP fluxes into the nuclear tip, as 2 separate rupture events. Line

scans of the average intensity of the NES-GFP were taken from the cytoplasm on the left

side of the nucleus, through the nucleus, and until the cytoplasm on the right side of the

nucleus, and plotted to the right of each picture. During rupture events, line scans show a

gradient in NES-GFP concentration along the nucleus. (C) Quantification of the average

cytoplasmic NES-GFP on the left of the nucleus for each time normalized to the initial

average intensity, black curve. The A similar quantification was made for the NES-GFP

intensity inside the nucleus, again normalized to the initial time before rupture. Average

intensity of NES-GFP in the nucleus increases 3-fold for R1 and R2. As expected,

cytoplasmic NES-GFP does not deviate much from 1. Scale bar is 10 µm.

How would NE opening lead to DNA damage? The cell cytoplasm contains several

enzymes that degrade DNA, some of which are involved in protection against

cytoplasmic DNA (23), others in mitotic processing of chromatin (24). Mitotic control of

the activity of these enzymes has been reported; this control prevents chromosomal

damage after nuclear envelope breakdown at mitotic entry (25). But when such proteins

are artificially engineered to accumulate in the nucleus in interphase cells, they induce a

large number of DNA breaks (26). Upon opening of the nuclear envelope during cell

migration through constrictions in interphase cells, such proteins could diffuse inside the

nucleus and damage it. Alternatively, cytoplasmic ROS (27) could have a similar effect.

Support for this hypothesis comes from reports of massive DNA damage in micronuclei

after nuclear envelope rupture, although the precise factors inducing this damage have

not been identified (28). This makes DNA repair an essential process for the survival of

cells migrating through dense environments, as they are constantly opening their NE.

Consistent with this observation, nuclear rupture induced by compression was shown to

induce expression of genes associated with DNA repair (12).

35

Movie legends

Movie S1. Two successive image sequences—fig. S1, C and D. The sequences correspond to images shown in fig. S1C, mDC expressing NLS-GFP migrating in a collagen gel, and fig. S1D, mDC expressing NLSGFP migrating in an ear explant. Time is hours:minutes. Scale bars, 20 μm.

Movie S2. Three successive image sequences—Fig. 2, A and B, and figs. S3B and S2D. The sequences correspond to images shown in Fig. 2A, hDC expressing NLS-GFP migrating in a microchannel with a constriction of length L = 15 μm and width W = 2 μm; Fig. 2B, HeLa cell expressing NLS-GFP migrating in a microchannel with a constriction (L = 15 μm, W = 2 μm); fig. S3B, same cell with slow motion on nuclear blebs; and fig. S2D, RPE1 cell expressing NLS-GFP migrating in a microchannel with a constriction (L = 15 μm W = 1.5 μm). Time is hours:minutes. Scale bars, 10 μm.

Movie S3. A sequence corresponding to images in fig. S4A. An hDC cell expressing GFP-cGAS (green, labeled DNA_BP in the movie, for DNA binding protein) and stained with Hoechst (red), is first uncompressed then compressed, which forms a nuclear bleb that bursts. Scale bar, 10 μm.

Movie S4. Two successive image sequences— from Fig. 2, F and I. The sequences correspond to images shown in Fig. 2F, hDC expressing GFP-cGAS (green, labeled DNA_BP in the movie) stained with Hoechst (red) migrating through a constriction (L = 15 μm W = 2 μm), and Fig. 2I, HeLa cell expressing DNA_BP-GFP (green) and H2B-mCherry (red) migrating through a constriction (L = 15 μm W = 2 μm). Time is hours:minutes. Scale bars, 10 μm.

Movie S5. Three successive image sequences—fig. S5, D to F. The sequences correspond to images shown in fig. S5, D, E, and F, three examples of HeLa cells expressing Lap2β-GFP (green) and H2B-mCherry (red) migrating through a constriction (L = 15 μm W = 2 μm). Time is hours:minutes. Scale bars, 10 μm.

Movie S6. A sequence corresponding to image in fig. S7A .HeLa cell expressing CHMP4B-EGFP (green) stained with Hoechst (blue) first not compressed and then compressed, with nuclear blebs that burst. Time is minutes:seconds. Scale bar, 5 μm.

Movie S7. A sequence corresponding to images in fig. S7B. HeLa cell expressing CHMP4B-EGFP laser wounded at two spots (arrows). Time is minutes:seconds. Scale bar, 10 μm.

Movie S8. A sequence corresponding to images shown in Fig. 3A. HeLa cell expressing RFP-cGAS (green, labeled DNA_BP in the movie) and CHMP4B-EGFP (red) migrating through a constriction (L = 15 μm, W = 2 μm). Time is hours:minutes:seconds. Scale bar, 20 μm.

Movie S9. Four successive image sequences—fig. S11C and fig. S10, C, E, and F. Sequences of RPE1 cells expressing NLS-GFP corresponding to images shown in fig. S11C. SiLMNA/C in a straight channel showing spontaneous nuclear leakage; fig. S10C, SiCTRL through a constriction (L = 15 μm W = 3 μm) showing transient partial nuclear leakage; fig. S10E, SiLMNA/C through a constriction (L = 15 μm W = 3 μm) showing repeated total nuclear leakage and cell death; and fig. S10F, SiCHMP3 through a constriction (L = 1 μm W = 3 μm) showing total nuclear leakage and delayed recovery but no death. Time is hours:minutes. Scale bars, 20 μm.

Movie S10. A sequence corresponding to images shown in Fig. 4B. RPE1 cell expressing 53BP1-GFP (gray) and RFP-cGAS (red, labeled DNA_BP in the movie) migrating through a constriction (L = 15 μm W = 2 μm). Time is hours:minutes. Scale bars, 20 μm.

36

REFERENCES AND NOTES References and Notes 1. C. P. Lee, P. T. Liu, H. N. Kung, M. T. Su, H. H. Chua, Y. H. Chang, C. W. Chang, C. H.

Tsai, F. T. Liu, M. R. Chen, The ESCRT machinery is recruited by the viral BFRF1 protein to the nucleus-associated membrane for the maturation of Epstein-Barr Virus. PLOS Pathog. 8, e1002904 (2012). Medline doi:10.1371/journal.ppat.1002904

2. E. Hatch, M. Hetzer, Breaching the nuclear envelope in development and disease. J. Cell Biol. 205, 133–141 (2014). Medline doi:10.1083/jcb.201402003

3. A. Fidziańska, Z. T. Bilińska, F. Tesson, T. Wagner, M. Walski, J. Grzybowski, W. Ruzyłło, I. Hausmanowa-Petrusewicz, Obliteration of cardiomyocyte nuclear architecture in a patient with LMNA gene mutation. J. Neurol. Sci. 271, 91–96 (2008). Medline doi:10.1016/j.jns.2008.03.017

4. T. Harada, J. Swift, J. Irianto, J. W. Shin, K. R. Spinler, A. Athirasala, R. Diegmiller, P. C. Dingal, I. L. Ivanovska, D. E. Discher, Nuclear lamin stiffness is a barrier to 3D migration, but softness can limit survival. J. Cell Biol. 204, 669–682 (2014). Medline doi:10.1083/jcb.201308029

5. A. C. Rowat, D. E. Jaalouk, M. Zwerger, W. L. Ung, I. A. Eydelnant, D. E. Olins, A. L. Olins, H. Herrmann, D. A. Weitz, J. Lammerding, Nuclear envelope composition determines the ability of neutrophil-type cells to passage through micron-scale constrictions. J. Biol. Chem. 288, 8610–8618 (2013). Medline doi:10.1074/jbc.M112.441535

6. P. M. Davidson, C. Denais, M. C. Bakshi, J. Lammerding, Nuclear deformability constitutes a rate-limiting step during cell migration in 3-D environments. Cell Mol Bioeng 7, 293–306 (2014). Medline doi:10.1007/s12195-014-0342-y

7. M. Sixt, Interstitial locomotion of leukocytes. Immunol. Lett. 138, 32–34 (2011). Medline doi:10.1016/j.imlet.2011.02.013

8. K. Wolf, M. Te Lindert, M. Krause, S. Alexander, J. Te Riet, A. L. Willis, R. M. Hoffman, C. G. Figdor, S. J. Weiss, P. Friedl, Physical limits of cell migration: Control by ECM space and nuclear deformation and tuning by proteolysis and traction force. J. Cell Biol. 201, 1069–1084 (2013). Medline doi:10.1083/jcb.201210152

9. H. Pflicke, M. Sixt, Preformed portals facilitate dendritic cell entry into afferent lymphatic vessels. J. Exp. Med. 206, 2925–2935 (2009). Medline

10. M. L. Heuzé, O. Collin, E. Terriac, A. M. Lennon-Duménil, M. Piel, Cell migration in confinement: A micro-channel-based assay. Methods Mol. Biol. 769, 415–434 (2011). Medline doi:10.1007/978-1-61779-207-6_28

11. M. Krause, K. Wolf, Cancer cell migration in 3D tissue: Negotiating space by proteolysis and nuclear deformability. Cell Adhes. Migr. 9, 357–366 (2015). Medline doi:10.1080/19336918.2015.1061173

12. M. Le Berre, J. Aubertin, M. Piel, Fine control of nuclear confinement identifies a threshold deformation leading to lamina rupture and induction of specific genes. Integr. Biol. (Camb.) 4, 1406–1414 (2012). Medline doi:10.1039/c2ib20056b

37

13. A. J. Jimenez, P. Maiuri, J. Lafaurie-Janvore, S. Divoux, M. Piel, F. Perez, ESCRT machinery is required for plasma membrane repair. Science 343, 1247136 (2014). Medline doi:10.1126/science.1247136

14. Y. Olmos, L. Hodgson, J. Mantell, P. Verkade, J. G. Carlton, ESCRT-III controls nuclear envelope reformation. Nature 522, 236–239 (2015). Medline doi:10.1038/nature14503

15. M. Vietri, K. O. Schink, C. Campsteijn, C. S. Wegner, S. W. Schultz, L. Christ, S. B. Thoresen, A. Brech, C. Raiborg, H. Stenmark, Spastin and ESCRT-III coordinate mitotic spindle disassembly and nuclear envelope sealing. Nature 522, 231–235 (2015). Medline doi:10.1038/nature14408

16. I. Poser, M. Sarov, J. R. Hutchins, J. K. Hériché, Y. Toyoda, A. Pozniakovsky, D. Weigl, A. Nitzsche, B. Hegemann, A. W. Bird, L. Pelletier, R. Kittler, S. Hua, R. Naumann, M. Augsburg, M. M. Sykora, H. Hofemeister, Y. Zhang, K. Nasmyth, K. P. White, S. Dietzel, K. Mechtler, R. Durbin, A. F. Stewart, J. M. Peters, F. Buchholz, A. A. Hyman, BAC TransgeneOmics: A high-throughput method for exploration of protein function in mammals. Nat. Methods 5, 409–415 (2008). Medline doi:10.1038/nmeth.1199

17. L. L. Scheffer, S. C. Sreetama, N. Sharma, S. Medikayala, K. J. Brown, A. Defour, J. K. Jaiswal, Mechanism of Ca²⁺-triggered ESCRT assembly and regulation of cell membrane repair. Nat. Commun. 5, 5646 (2014). Medline doi:10.1038/ncomms6646

18. W. H. De Vos, F. Houben, M. Kamps, A. Malhas, F. Verheyen, J. Cox, E. M. Manders, V. L. Verstraeten, M. A. van Steensel, C. L. Marcelis, A. van den Wijngaard, D. J. Vaux, F. C. Ramaekers, J. L. Broers, Repetitive disruptions of the nuclear envelope invoke temporary loss of cellular compartmentalization in laminopathies. Hum. Mol. Genet. 20, 4175–4186 (2011). Medline doi:10.1093/hmg/ddr344

19. J. D. Vargas, E. M. Hatch, D. J. Anderson, M. W. Hetzer, Transient nuclear envelope rupturing during interphase in human cancer cells. Nucleus 3, 88–100 (2012). Medline doi:10.4161/nucl.18954

20. S. Gonzalo, R. Kreienkamp, DNA repair defects and genome instability in Hutchinson-Gilford Progeria Syndrome. Curr. Opin. Cell Biol. 34, 75–83 (2015). Medline doi:10.1016/j.ceb.2015.05.007

21. A. Janssen, M. van der Burg, K. Szuhai, G. J. Kops, R. H. Medema, Chromosome segregation errors as a cause of DNA damage and structural chromosome aberrations. Science 333, 1895–1898 (2011). Medline doi:10.1126/science.1210214

22. I. Hickson, Y. Zhao, C. J. Richardson, S. J. Green, N. M. Martin, A. I. Orr, P. M. Reaper, S. P. Jackson, N. J. Curtin, G. C. Smith, Identification and characterization of a novel and specific inhibitor of the ataxia-telangiectasia mutated kinase ATM. Cancer Res. 64, 9152–9159 (2004). Medline doi:10.1158/0008-5472.CAN-04-2727

23. J. Maciejowski, Y. Li, N. Bosco, P. J. Campbell, T. de Lange, Chromothripsis and Kataegis Induced by Telomere Crisis. Cell 163, 1641–1654 (2015). Medline doi:10.1016/j.cell.2015.11.054

24. S. Sarbajna, D. Davies, S. C. West, Roles of SLX1-SLX4, MUS81-EME1, and GEN1 in avoiding genome instability and mitotic catastrophe. Genes Dev. 28, 1124–1136 (2014). Medline doi:10.1101/gad.238303.114

38

25. D. Gritenaite, L. N. Princz, B. Szakal, S. C. Bantele, L. Wendeler, S. Schilbach, B. H. Habermann, J. Matos, M. Lisby, D. Branzei, B. Pfander, A cell cycle-regulated Slx4-Dpb11 complex promotes the resolution of DNA repair intermediates linked to stalled replication. Genes Dev. 28, 1604–1619 (2014). Medline doi:10.1101/gad.240515.114

26. Y. W. Chan, S. C. West, Spatial control of the GEN1 Holliday junction resolvase ensures genome stability. Nat. Commun. 5, 4844 (2014). Medline doi:10.1038/ncomms5844

27. T. Shimi, R. D. Goldman, Nuclear lamins and oxidative stress in cell proliferation and longevity. Adv. Exp. Med. Biol. 773, 415–430 (2014). Medline doi:10.1007/978-1-4899-8032-8_19

28. E. M. Hatch, A. H. Fischer, T. J. Deerinck, M. W. Hetzer, Catastrophic nuclear envelope collapse in cancer cell micronuclei. Cell 154, 47–60 (2013). Medline doi:10.1016/j.cell.2013.06.007

29. C. Soria-Valles, F. G. Osorio, A. Gutiérrez-Fernández, A. De Los Angeles, C. Bueno, P. Menéndez, J. I. Martín-Subero, G. Q. Daley, J. M. Freije, C. López-Otín, NF-κB activation impairs somatic cell reprogramming in ageing. Nat. Cell Biol. 17, 1004–1013 (2015). Medline doi:10.1038/ncb3207

30. Y. J. Liu, M. Le Berre, F. Lautenschlaeger, P. Maiuri, A. Callan-Jones, M. Heuzé, T. Takaki, R. Voituriez, M. Piel, Confinement and low adhesion induce fast amoeboid migration of slow mesenchymal cells. Cell 160, 659–672 (2015). Medline doi:10.1016/j.cell.2015.01.007

31. G. Faure-André, P. Vargas, M. I. Yuseff, M. Heuzé, J. Diaz, D. Lankar, V. Steri, J. Manry, S. Hugues, F. Vascotto, J. Boulanger, G. Raposo, M. R. Bono, M. Rosemblatt, M. Piel, A. M. Lennon-Duménil, Regulation of dendritic cell migration by CD74, the MHC class II-associated invariant chain. Science 322, 1705–1710 (2008). Medline doi:10.1126/science.1159894

32. J. Riedl, K. C. Flynn, A. Raducanu, F. Gärtner, G. Beck, M. Bösl, F. Bradke, S. Massberg, A. Aszodi, M. Sixt, R. Wedlich-Söldner, Lifeact mice for studying F-actin dynamics. Nat. Methods 7, 168–169 (2010). Medline doi:10.1038/nmeth0310-168

33. L. Holmer, H. J. Worman, Inner nuclear membrane proteins: Functions and targeting. Cell. Mol. Life Sci. 58, 1741–1747 (2001). Medline doi:10.1007/PL00000813

34. N. Manel, B. Hogstad, Y. Wang, D. E. Levy, D. Unutmaz, D. R. Littman, A cryptic sensor for HIV-1 activates antiviral innate immunity in dendritic cells. Nature 467, 214–217 (2010). Medline doi:10.1038/nature09337

35. T. Satoh, N. Manel, Gene transduction in human monocyte-derived dendritic cells using lentiviral vectors. Methods Mol. Biol. 960, 401–409 (2013). Medline doi:10.1007/978-1-62703-218-6_30

36. D. Nègre, P. E. Mangeot, G. Duisit, S. Blanchard, P. O. Vidalain, P. Leissner, A. J. Winter, C. Rabourdin-Combe, M. Mehtali, P. Moullier, J. L. Darlix, F. L. Cosset, Characterization of novel safe lentiviral vectors derived from simian immunodeficiency virus (SIVmac251) that efficiently transduce mature human dendritic cells. Gene Ther. 7, 1613–1623 (2000). Medline doi:10.1038/sj.gt.3301292

39

37. P. Vargas, P. Maiuri, M. Bretou, P. J. Sáez, P. Pierobon, M. Maurin, M. Chabaud, D. Lankar, D. Obino, E. Terriac, M. Raab, H. R. Thiam, T. Brocker, S. M. Kitchen-Goosen, A. S. Alberts, P. Sunareni, S. Xia, R. Li, R. Voituriez, M. Piel, A. M. Lennon-Duménil, Innate control of actin nucleation determines two distinct migration behaviours in dendritic cells. Nat. Cell Biol. 18, 43–53 (2016). Medline doi:10.1038/ncb3284

38. X. Lahaye, T. Satoh, M. Gentili, S. Cerboni, C. Conrad, I. Hurbain, A. El Marjou, C. Lacabaratz, J. D. Lelièvre, N. Manel, The capsids of HIV-1 and HIV-2 determine immune detection of the viral cDNA by the innate sensor cGAS in dendritic cells. Immunity 39, 1132–1142 (2013). Medline doi:10.1016/j.immuni.2013.11.002