Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 114 (11), E2166-E2175

LEM2 Recruits CHMP7 for ESCRT-mediated Nuclear Envelope Closure in Fission Yeast and Human Cells

Affiliations

LEM2 Recruits CHMP7 for ESCRT-mediated Nuclear Envelope Closure in Fission Yeast and Human Cells

Mingyu Gu et al. Proc Natl Acad Sci U S A.

Abstract

Endosomal sorting complexes required for transport III (ESCRT-III) proteins have been implicated in sealing the nuclear envelope in mammals, spindle pole body dynamics in fission yeast, and surveillance of defective nuclear pore complexes in budding yeast. Here, we report that Lem2p (LEM2), a member of the LEM (Lap2-Emerin-Man1) family of inner nuclear membrane proteins, and the ESCRT-II/ESCRT-III hybrid protein Cmp7p (CHMP7), work together to recruit additional ESCRT-III proteins to holes in the nuclear membrane. In Schizosaccharomyces pombe, deletion of the ATPase vps4 leads to severe defects in nuclear morphology and integrity. These phenotypes are suppressed by loss-of-function mutations that arise spontaneously in lem2 or cmp7, implying that these proteins may function upstream in the same pathway. Building on these genetic interactions, we explored the role of LEM2 during nuclear envelope reformation in human cells. We found that CHMP7 and LEM2 enrich at the same region of the chromatin disk periphery during this window of cell division and that CHMP7 can bind directly to the C-terminal domain of LEM2 in vitro. We further found that, during nuclear envelope formation, recruitment of the ESCRT factors CHMP7, CHMP2A, and IST1/CHMP8 all depend on LEM2 in human cells. We conclude that Lem2p/LEM2 is a conserved nuclear site-specific adaptor that recruits Cmp7p/CHMP7 and downstream ESCRT factors to the nuclear envelope.

Keywords: CHMP7; ESCRT-III; LEM2; VPS4; nuclear envelope.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.
Fig. 1.
vps4Δ cells grow slowly and have severe NE defects, which are suppressed by loss of cmp7 or lem2. (A) Tetrad dissection of vps4Δ/+ diploids, with genotypes labeled below the image. (Scale bar: 0.25 cm.) (B) Spontaneous suppressors (arrowhead) of vps4Δ (arrow) appearing after 3 d on rich medium, with colony sizes comparable to those of WT cells. (Scale bar: 1 cm.) (C) Growth curves of each genotype showing optical densities of yeast cultures, starting at OD600nm 0.06, for 32 h in 2-h intervals. Three independent isolates for each genotype were measured. The plots show mean ± SEM. (D) WT cells, showing normal NE morphologies (Ish1p-mCherry) and DNA (Hoechst staining) within the nuclei. (E) vps4Δ, showing various NE morphology defects including excess NE, fragmented NE, and DNA that appears to be outside of the nucleus. (Scale bar: 5 μm.) (F) Deletion of either cmp7 or lem2 rescues these NE morphology defects. Here and throughout, three independent isolates/strains were imaged for each genotype, and n represents the total number of scored cells. Mean ± SEM, WT: 1 ± 1%, n = 115, 121, 139; vps4Δ: 85 ± 5%, n = 62, 69, 86; cmp7Δ: 2.0 ± 0.4%, n = 151, 166, 147; vps4Δcmp7Δ: 1.1 ± 0.7%, n =161, 118, 145; lem2Δ: 8 ± 3%, n= 131, 103, 128; vps4Δlem2Δ: 12.2 ± 0.7%, n = 99, 176, 111. Two-tailed Student t tests were used here and throughout. ***P < 0.001.
Fig. S1.
Fig. S1.
vps4Δ and lem2Δ exhibit asymmetric SPB segregation. (A) WT cells showing symmetric SPB segregation. (B) vps4Δ showing asymmetric SPB segregation. (C) Quantified percentages of asymmetric SPB segregation events in each genotype. Cells with septa were selected and scored for SPB separation patterns. Nuclei were stained with Hoechst. WT: 0.7 ± 0.3%, n = 103, 101, 101; vps4Δ: 29 ± 7%, n = 88, 103, 107; cmp7Δ: 2 ± 1%, n = 105, 103, 101; vps4Δcmp7Δ: 4 ± 2%, n = 117, 100, 104; lem2Δ: 18 ± 1%, n = 102, 103, 100; vps4Δlem2Δ: 18 ± 4%, n = 102, 103, 101. *P < 0.05. N.S., not significant. (Scale bar: 5 μm.)
Fig. 2.
Fig. 2.
vps4Δ cells display a series of mitotic defects associated with failure of nuclear membrane maintenance. (A) Time lapse of normal mitotic karyokinesis in vps4Δ cells. In all cases, the nuclear membrane is marked by Ish1p-mCherry, and the SPB is marked by Cut12p-YFP. The total length of time is 100 min. (B) Karmellae formation correlates with defective separation of duplicated SPB. The total length of time is 400 min. (C) Asymmetric nuclear division. Total length of time is 100 min. (D) Failed nuclear division. Total length of time is 100 min. (E) Quantification of NE morphology during mitosis in WT and vps4Δ. Normal mitotic nuclear division, WT: 99 ± 1%, vps4Δ: 27 ± 8%; karmellae formation, WT: 0%, vps4Δ: 29 ± 3%; asymmetric nuclear division, mean ± SEM, WT: 1 ± 1%, vps4Δ: 38 ± 2%; failed nuclear division, WT: 0%, vps4Δ: 6 ± 2%. WT, n = 23, 21, 18; vps4Δ, n = 43, 51, 84. *P < 0.05; ***P < 0.001. (Scale bars: 5 μm.) (AD) Dashed lines correspond with the cell wall.
Fig. S2.
Fig. S2.
SPB separation is delayed in vps4Δ mutants. A 16-h time-lapse imaging (10-min intervals) was conducted on WT and vps4Δ cells at 32 °C, and the times required for SBP separation were scored. Data were binned and plotted by mean frequency ± SEM: WT: 10 min 96 ± 2%; 20–50 min 4 ± 2%; >50 min 0%; n = 23, 22, 22; vps4Δ: 10 min 57 ± 4%; 20–50 min 31 ± 4%; >50 min 12 ± 3%; n = 30, 30, 30. *P < 0.05; **P < 0.01; ***P < 0.001.
Fig. 3.
Fig. 3.
vps4Δ cells have leaky nuclei and nuclear integrity is restored by loss of either cmp7 or lem2. (A) GFP signals in the nuclear lumen of WT cells expressing NLS-GFP-LacZ. (B) vps4Δ cells expressing NLS-GFP-LacZ with moderate (arrowheads) or severe (arrow) nuclear leaking. (Scale bar: 10 μm.) (C) Nuclear enrichment of NLS-GFP-LacZ (nucleus/cytoplasm). Mean ± SD, WT: 21.6 ± 8.9, n = 205; vps4Δ: 8.6 ± 4.7, n = 180; cmp7Δ: 19.4 ± 7.7, n = 175; vps4Δcmp7Δ: 24.5 ± 11.7, n = 202; lem2Δ: 19.1 ± 10.4, n = 198; vps4Δlem2Δ: 21.5 ± 8.8, n = 189. (D) Percentage of cells in different nuclear leaky phenotype categories. Mean ± SEM, normal (>10-fold GFP nuclear enrichment) WT: 97.9 ± 1.3%, vps4Δ: 36.2 ± 6.5%, cmp7Δ: 94.1 ± 1.1%, vps4Δcmp7Δ: 93.2 ± 1.6%, lem2Δ: 86.6 ± 2.3%, vps4Δlem2Δ: 93.4 ± 2.0%; partial leaking (2- to 10-fold GFP nuclear enrichment) WT: 1.6 ± 0.9%, vps4Δ: 54.4 ± 8.1%, cmp7Δ: 5.9 ± 1.1%, vps4Δcmp7Δ: 6.8 ± 1.6%, lem2Δ: 11.5 ± 2.3%, vps4Δlem2Δ: 5.9 ± 1.4%; severe leaking (<2-fold GFP nuclear enrichment) WT: 0%, vps4Δ: 9.4 ± 2.4%, cmp7Δ: 0%, vps4Δcmp7Δ: 0%, lem2Δ: 2.0 ± 1.3%, vps4Δlem2Δ: 0.7 ± 0.7%; WT: n = 82,56,67, vps4Δ: n = 60,59,61, cmp7Δ: n = 62,63,50, vps4Δcmp7Δ: n = 82,59,61, lem2Δ: n = 68,61,69, vps4Δlem2Δ: n = 76,50,63. *P < 0.05; **P < 0.01; ***P < 0.001.
Fig. S3.
Fig. S3.
Time-lapse analysis of NLS-GFP-LacZ localization in vps4Δ. (A) Time-lapse images showing normal nuclear localization of NLS-GFP-LacZ converting to partial cytoplasmic localization. A pair of newly divided cells is shown over a time course of 100 min with 10-min interval. (B) Time lapse images showing partial cytoplasmic localization of NLS-GFP-LacZ converting to a homogenous distribution throughout the cell. A single cell is shown over a time course of 100 min with 10-min intervals. (Scale bars: 5 μm.)
Fig. 4.
Fig. 4.
vps4Δ NEs have persistent fenestrations, karmellae, and disorganized tubular extensions. (A, Left) Single tomographic slice showing the WT NE. NPCs are marked by asterisks. Segmented NE (green) with NPCs (red) reconstructed from 150 tomographic slices is shown in a merged view (A, Left Center), a front view (A, Right Center), and a side view (A, Right). (B, Left) Single tomographic slice showing the vps4Δ NE. The following features and defects are highlighted: a fenestration (bracket), an NPC (asterisk), karmellae layers (white arrowheads), and a disorganized whorl of tubules (arrows). (B, Right) Segmented model of NE reconstructed from 100 tomographic slices is shown in a merged view with karmellae in gold and a whorl of tubules in purple. (C) A segmented model of vps4Δ NE reconstructed from 200 tomographic slices is shown from the front (Left), back (Left Center), top (Right Center), and bottom (Right). (D, Left) Single vps4Δ tomographic slice showing NPCs are absent from karmellae region. The following features and defects are highlighted: a fenestration (bracket), an NPC (asterisk), and karmellae layers (white arrowheads). Segmented NE (green) with NPCs (red) and SPB (yellow) reconstructed from 150 tomographic slices shown in a merged view (D, Left Center), a front view (D, Right Center), and a side view (D, Right). (Scale bars: 200 nm.)
Fig. 5.
Fig. 5.
vps4Δ NEs have persistent fenestrations, karmellae, and disorganized extensions, and these defects are suppressed by loss of cmp7 or lem2. (A) WT cells with normal NE. (BD) vps4Δ cells with a variety of NE phenotypes, including karmellae (B), multiple holes (arrows) (C), and discontinuities (black arrowhead) that can contain apparently intranuclear NPCs (D, arrowheads). (EH) cmp7Δ and lem2Δ single mutants (E and G) vs. vps4Δcmp7Δ and vps4Δlem2Δ double mutants (F and H). (Scale bars: 500 nm.)
Fig. S4.
Fig. S4.
Karmellae and tubular extension are connected NE structures in vps4Δ. Selected tomographic slices of vps4Δ show connected karmellae and tubular extensions. The relevant membranes are shaded in light green, and the junction is marked by black arrows. (Scale bar: 200 nm.)
Fig. S5.
Fig. S5.
A cartoon model for the NE phenotypes in vps4Δ, lem2Δ, and cmp7Δ in S. pombe. (Top) In WT, the exposed Lem2p recruits Cmp7p to initiate an ESCRT-III–mediated NE sealing event. That process is regulated and/or completed by the AAA ATPase, Vps4p. (Middle) in vps4Δ, unrestricted Lem2p/Cmp7p activities induce the formation of karmellae and large gaps in NE so that the proper progress of mitosis is disrupted. (Bottom) In lem2Δ or cmp7Δ, mutant cells tolerate a limited number of small breaches and therefore show little or no growth defect.
Fig. 6.
Fig. 6.
Live imaging of LEM2-dependent recruitment of CHMP7 to reforming nuclei during mammalian anaphase. (A) Montage of a representative cell expressing LEM2-mCherry and GFP-CHMP7 progressing through anaphase before complete furrow ingression (designated as t = 0′). LEM2-mCherry makes initial contacts with chromatin (white arrow), and GFP-CHMP7 localizes to sites of LEM2-mCherry enrichment (t = −2′, −1′). (B) Illustrative cells, at t = −1′, treated with siRNA as indicated and expressing GFP-CHMP7. (C) Quantification of GFP-CHMP7 peaks at chromatin disks of cells treated with siRNA (mean ± SD; siControl: 12 ± 6%, n = 28; siLEM2-1: 5 ± 4%, n = 32; siLEM2-2: 3 ± 3%, n = 38). ***P < 0.001. (Scales bar: 10 μm.)
Fig. S6.
Fig. S6.
LEM2 and CHMP7 colocalization is restricted to a brief window specifically during anaphase. (A) Additional montages of cells expressing LEM2-mCherry and GFP-CHMP7 progressing through anaphase before complete furrow ingression (designated as t = 0′). (B) Representative image of an interphase cell expressing LEM2-mCherry and GFP-CHMP7. (Scale bars: 10 μm.)
Fig. 7.
Fig. 7.
Recruitment of CHMP2A and IST1/CHMP8 during mammalian nuclear reformation depends on LEM2, whose targeting is independent of CHMP7. (A) Confocal images illustrating IST1/CHMP8 localization in anaphase B cells after siControl, siLEM2-1, or siLEM2-2 treatment. (B) Quantification of IST1/CHMP8 recruitment to chromatin disks during anaphase B. IST1/CHMP8 recruitment was scored as robust, weak, or no chromatin-associated foci. The robust category was graphed and statistical analysis was performed, comparing the siControl dataset to each depletion condition dataset (siControl: 63 ± 6%, n = 18, 58, 24; siLEM2-1: 0 ± 0%, n = 40, 44, 12; siLEM2-2: 6 ± 6%, n = 34, 22, 6; siCHMP7-1: 0 ± 0%, n = 42, 20, 12; siCHMP7-2: 0 ± 0%, n = 22, 20, 2). (C) Widefield images illustrating CHMP2A localization in anaphase B cells after siControl, siLEM2-1, or siLEM2-2 treatment. (D) CHMP2A recruitment to chromatin disks at anaphase B, assessed as in B (siControl: 70 ± 9%, n = 48, 48, 52; siLEM2-1: 9 ± 5%, n = 108, 86, 38; siLEM2-2: 11 ± 5%, n = 98, 52, 42; siCHMP7-1: 4 ± 1%, n = 78, 102, 47; siCHMP7-2: 21 ± 8%, n = 112, 58, 56). Analysis of parallel samples confirmed that LEM2 and CHMP7 depletion profoundly disrupted IST1/CHMP8 recruitment as before and is shown in Fig. S7G. (E) Widefield images of cells costained for IST1/CHMP8 and LEM2 illustrates the differential sensitivity of their localization at anaphase chromatin disks after siCHMP7-1 or siCHMP7-2 treatment. Signal detected at the midzone with LEM2 antibody is likely nonspecific (Fig. S7A). (F) Quantification of LEM2 recruitment to chromatin disks at anaphase B. Images were used for blind scoring the presence of chromatin-associated LEM2 (siControl: 95 ± 4%, n = 28, 35, 39; siCHMP7-1: 90 ± 6%, n = 18, 72, 45; siCHMP7-2: 99 ± 1%, n = 35, 56, 7). Analysis of parallel samples confirmed that CHMP7 depletion profoundly disrupted IST1/CHMP8 recruitment as before and is shown in Fig. S7D. *P < 0.05; **P < 0.01; ***P < 0.001. N.S., not significant. (Scale bars: 10 μm.) All graphs plot mean ± SEM.
Fig. S7.
Fig. S7.
Control experiments for knockdown analysis. (A) Endogenous LEM2 detected at anaphase chromatin or disks. Spinning disk confocal microscopy of representative anaphase cells 72 h after treatment with siLEM2-1 or -2 confirmed that detection of signal at the chromatin surface is specific. Signal detected at the midzone with LEM2 antibody is nonspecific, persisting after knockdown of LEM2 with two independent oligos. (B) Widefield microscopy of representative interphase cells comparing detection of endogenous LEM2 at the nuclear rim 72 h after treatment with siControl or LEM2-specific oligos. (C) Specific depletion of LEM2 and CHMP7 with respective oligos used for RNAi is confirmed by immunoblot. # indicates a smaller protein product likely derived from CHMP7. (D) IST1/CHMP8 recruitment, assessed and graphed as described in Fig. 7B, performed in parallel with LEM2 analysis in Fig. 7F (siControl: 67 ± 10%, n = 104, 14, 40; siCHMP7-1: 0 ± 0%, n = 68, 26, 24; siCHMP7-2: 0 ± 0%, n = 24, 10, 48). (E) Widefield microscopy of representative interphase cells comparing detection of endogenous SUN1 at the nuclear rim 72 h after treatment with siControl or SUN1-directed oligos. (F) Specific depletion of SUN1 by RNAi confirmed by immunoblot. (G) IST1/CHMP8 recruitment, assessed and graphed as described in Fig. 7B, performed in parallel with CHMP2A analysis in Fig. 7D (siControl: 60 ± 8%, n = 80, 42, 48; siLEM2-1: 8.2 ± 4%, n = 80, 82, 30; siLEM2-2: 17 ± 7%, n = 120, 82, 78; siCHMP7-1: 5 ± 3%, n = 76, 50, 58; siCHMP7-2: 8 ± 4%, n = 58, 46, 48; siSUN1A: 53 ± 18%, n = 58, 18, 20; siSUN1B: 50 ± 7%, n = 78, 44, 36). All graphs plot mean ± SEM. *P < 0.05; **P < 0.01. N.S., not significant. Scale bars, 10 μm.
Fig. S8.
Fig. S8.
LEM2 plays a role in the recruitment of CHMP7 to anaphase chromatin disks. Additional montages of cells expressing GFP-CHMP7 and treated with siRNA, as indicated, progressing through anaphase are shown. Scale bars, 10 μm.
Fig. S9.
Fig. S9.
The CTD of LEM2 binds full-length CHMP7 in vitro. (A) Protein domain structure of human LEM2. TM, transmembrane domain. (B and C) In vitro binding assay of full-length human CHMP7 with immobilized human His-SUMO-LEM2(NTD) (amino acids 1–208; B) and immobilized His-SUMO-LEM2(CTD) (amino acids 398–503; C). Samples are SDS-buffer eluates analyzed by SDS/PAGE and Coomassie staining. Note that CHMP7 showed weak unspecific binding to the resin (B, lane 2) that was not increased by immobilized His-SUMO-LEM2(NTD) (B, lane 3). In contrast, immobilized His-SUMO-LEM2(CTD), increased CHMP7 binding significantly (C, compare lanes 3 and 4). Finally, LEM2 (NTD) did not bind immobilized LEM2 (CTD) (C, lane 5), and the presence of LEM2 (NTD) did not enhance CHMP7 binding to immobilized His-SUMO-LEM2 (CTD) (C, compare lanes 4 and 6).

Similar articles

See all similar articles

Cited by 20 articles

See all "Cited by" articles

References

    1. Devos DP, Gräf R, Field MC. Evolution of the nucleus. Curr Opin Cell Biol. 2014;28:8–15. - PMC - PubMed
    1. Madhani HD. The frustrated gene: Origins of eukaryotic gene expression. Cell. 2013;155(4):744–749. - PMC - PubMed
    1. Heath IB. Variant mitoses in lower eukaryotes: Indicators of the evolution of mitosis. Int Rev Cytol. 1980;64:1–80. - PubMed
    1. Byers B, Goetsch L. Behavior of spindles and spindle plaques in the cell cycle and conjugation of Saccharomyces cerevisiae. J Bacteriol. 1975;124(1):511–523. - PMC - PubMed
    1. Kubai DF. The evolution of the mitotic spindle. Int Rev Cytol. 1975;43:167–227. - PubMed

Publication types

MeSH terms

Substances

Feedback